1FFMPEG-ALL(1) FFMPEG-ALL(1)
2
6 ffmpeg - ffmpeg video converter
7
9 ffmpeg [global_options] {[input_file_options] -i input_url} ...
10 {[output_file_options] output_url} ...
11
13 ffmpeg is a very fast video and audio converter that can also grab from
14 a live audio/video source. It can also convert between arbitrary sample
15 rates and resize video on the fly with a high quality polyphase filter.
16 ffmpeg reads from an arbitrary number of input "files" (which can be
18 regular files, pipes, network streams, grabbing devices, etc.),
19 specified by the "-i" option, and writes to an arbitrary number of
20 output "files", which are specified by a plain output url. Anything
21 found on the command line which cannot be interpreted as an option is
22 considered to be an output url.
23 Each input or output url can, in principle, contain any number of
25 streams of different types (video/audio/subtitle/attachment/data). The
26 allowed number and/or types of streams may be limited by the container
27 format. Selecting which streams from which inputs will go into which
28 output is either done automatically or with the "-map" option (see the
29 Stream selection chapter).
30 To refer to input files in options, you must use their indices
32 (0-based). E.g. the first input file is 0, the second is 1, etc.
33 Similarly, streams within a file are referred to by their indices. E.g.
34 "2:3" refers to the fourth stream in the third input file. Also see the
35 Stream specifiers chapter.
36 As a general rule, options are applied to the next specified file.
38 Therefore, order is important, and you can have the same option on the
39 command line multiple times. Each occurrence is then applied to the
40 next input or output file. Exceptions from this rule are the global
41 options (e.g. verbosity level), which should be specified first.
42 Do not mix input and output files -- first specify all input files,
44 then all output files. Also do not mix options which belong to
45 different files. All options apply ONLY to the next input or output
46 file and are reset between files.
47 • To set the video bitrate of the output file to 64 kbit/s:
49 ffmpeg -i input.avi -b:v 64k -bufsize 64k output.avi
51 • To force the frame rate of the output file to 24 fps:
53 ffmpeg -i input.avi -r 24 output.avi
55 • To force the frame rate of the input file (valid for raw formats
57 only) to 1 fps and the frame rate of the output file to 24 fps:
58 ffmpeg -r 1 -i input.m2v -r 24 output.avi
60 The format option may be needed for raw input files.
62
64 The transcoding process in ffmpeg for each output can be described by
65 the following diagram:
66 _______ ______________
68 | | | |
69 | input | demuxer | encoded data | decoder
70 | file | ---------> | packets | -----+
71 |_______| |______________| |
72 v
73 _________
74 | |
75 | decoded |
76 | frames |
77 |_________|
78 ________ ______________ |
79 | | | | |
80 | output | <-------- | encoded data | <----+
81 | file | muxer | packets | encoder
82 |________| |______________|
83 ffmpeg calls the libavformat library (containing demuxers) to read
85 input files and get packets containing encoded data from them. When
86 there are multiple input files, ffmpeg tries to keep them synchronized
87 by tracking lowest timestamp on any active input stream.
88 Encoded packets are then passed to the decoder (unless streamcopy is
90 selected for the stream, see further for a description). The decoder
91 produces uncompressed frames (raw video/PCM audio/...) which can be
92 processed further by filtering (see next section). After filtering, the
93 frames are passed to the encoder, which encodes them and outputs
94 encoded packets. Finally those are passed to the muxer, which writes
95 the encoded packets to the output file.
96 Filtering
98 Before encoding, ffmpeg can process raw audio and video frames using
99 filters from the libavfilter library. Several chained filters form a
100 filter graph. ffmpeg distinguishes between two types of filtergraphs:
101 simple and complex.
102 Simple filtergraphs
104 Simple filtergraphs are those that have exactly one input and output,
106 both of the same type. In the above diagram they can be represented by
107 simply inserting an additional step between decoding and encoding:
108 _________ ______________
110 | | | |
111 | decoded | | encoded data |
112 | frames |\ _ | packets |
113 |_________| \ /||______________|
114 \ __________ /
115 simple _\|| | / encoder
116 filtergraph | filtered |/
117 | frames |
118 |__________|
119 Simple filtergraphs are configured with the per-stream -filter option
121 (with -vf and -af aliases for video and audio respectively). A simple
122 filtergraph for video can look for example like this:
123 _______ _____________ _______ ________
125 | | | | | | | |
126 | input | ---> | deinterlace | ---> | scale | ---> | output |
127 |_______| |_____________| |_______| |________|
128 Note that some filters change frame properties but not frame contents.
130 E.g. the "fps" filter in the example above changes number of frames,
131 but does not touch the frame contents. Another example is the "setpts"
132 filter, which only sets timestamps and otherwise passes the frames
133 unchanged.
134 Complex filtergraphs
136 Complex filtergraphs are those which cannot be described as simply a
138 linear processing chain applied to one stream. This is the case, for
139 example, when the graph has more than one input and/or output, or when
140 output stream type is different from input. They can be represented
141 with the following diagram:
142 _________
144 | |
145 | input 0 |\ __________
146 |_________| \ | |
147 \ _________ /| output 0 |
148 \ | | / |__________|
149 _________ \| complex | /
150 | | | |/
151 | input 1 |---->| filter |\
152 |_________| | | \ __________
153 /| graph | \ | |
154 / | | \| output 1 |
155 _________ / |_________| |__________|
156 | | /
157 | input 2 |/
158 |_________|
159 Complex filtergraphs are configured with the -filter_complex option.
161 Note that this option is global, since a complex filtergraph, by its
162 nature, cannot be unambiguously associated with a single stream or
163 file.
164 The -lavfi option is equivalent to -filter_complex.
166 A trivial example of a complex filtergraph is the "overlay" filter,
168 which has two video inputs and one video output, containing one video
169 overlaid on top of the other. Its audio counterpart is the "amix"
170 filter.
171 Stream copy
173 Stream copy is a mode selected by supplying the "copy" parameter to the
174 -codec option. It makes ffmpeg omit the decoding and encoding step for
175 the specified stream, so it does only demuxing and muxing. It is useful
176 for changing the container format or modifying container-level
177 metadata. The diagram above will, in this case, simplify to this:
178 _______ ______________ ________
180 | | | | | |
181 | input | demuxer | encoded data | muxer | output |
182 | file | ---------> | packets | -------> | file |
183 |_______| |______________| |________|
184 Since there is no decoding or encoding, it is very fast and there is no
186 quality loss. However, it might not work in some cases because of many
187 factors. Applying filters is obviously also impossible, since filters
188 work on uncompressed data.
189
191 ffmpeg provides the "-map" option for manual control of stream
192 selection in each output file. Users can skip "-map" and let ffmpeg
193 perform automatic stream selection as described below. The "-vn / -an /
194 -sn / -dn" options can be used to skip inclusion of video, audio,
195 subtitle and data streams respectively, whether manually mapped or
196 automatically selected, except for those streams which are outputs of
197 complex filtergraphs.
198 Description
200 The sub-sections that follow describe the various rules that are
201 involved in stream selection. The examples that follow next show how
202 these rules are applied in practice.
203 While every effort is made to accurately reflect the behavior of the
205 program, FFmpeg is under continuous development and the code may have
206 changed since the time of this writing.
207 Automatic stream selection
209 In the absence of any map options for a particular output file, ffmpeg
211 inspects the output format to check which type of streams can be
212 included in it, viz. video, audio and/or subtitles. For each acceptable
213 stream type, ffmpeg will pick one stream, when available, from among
214 all the inputs.
215 It will select that stream based upon the following criteria:
217 • for video, it is the stream with the highest resolution,
219 • for audio, it is the stream with the most channels,
221 • for subtitles, it is the first subtitle stream found but there's a
223 caveat. The output format's default subtitle encoder can be either
224 text-based or image-based, and only a subtitle stream of the same
225 type will be chosen.
226 In the case where several streams of the same type rate equally, the
228 stream with the lowest index is chosen.
229 Data or attachment streams are not automatically selected and can only
231 be included using "-map".
232 Manual stream selection
234 When "-map" is used, only user-mapped streams are included in that
236 output file, with one possible exception for filtergraph outputs
237 described below.
238 Complex filtergraphs
240 If there are any complex filtergraph output streams with unlabeled
242 pads, they will be added to the first output file. This will lead to a
243 fatal error if the stream type is not supported by the output format.
244 In the absence of the map option, the inclusion of these streams leads
245 to the automatic stream selection of their types being skipped. If map
246 options are present, these filtergraph streams are included in addition
247 to the mapped streams.
248 Complex filtergraph output streams with labeled pads must be mapped
250 once and exactly once.
251 Stream handling
253 Stream handling is independent of stream selection, with an exception
255 for subtitles described below. Stream handling is set via the "-codec"
256 option addressed to streams within a specific output file. In
257 particular, codec options are applied by ffmpeg after the stream
258 selection process and thus do not influence the latter. If no "-codec"
259 option is specified for a stream type, ffmpeg will select the default
260 encoder registered by the output file muxer.
261 An exception exists for subtitles. If a subtitle encoder is specified
263 for an output file, the first subtitle stream found of any type, text
264 or image, will be included. ffmpeg does not validate if the specified
265 encoder can convert the selected stream or if the converted stream is
266 acceptable within the output format. This applies generally as well:
267 when the user sets an encoder manually, the stream selection process
268 cannot check if the encoded stream can be muxed into the output file.
269 If it cannot, ffmpeg will abort and all output files will fail to be
270 processed.
271 Examples
273 The following examples illustrate the behavior, quirks and limitations
274 of ffmpeg's stream selection methods.
275 They assume the following three input files.
277 input file 'A.avi'
279 stream 0: video 640x360
280 stream 1: audio 2 channels
281 input file 'B.mp4'
283 stream 0: video 1920x1080
284 stream 1: audio 2 channels
285 stream 2: subtitles (text)
286 stream 3: audio 5.1 channels
287 stream 4: subtitles (text)
288 input file 'C.mkv'
290 stream 0: video 1280x720
291 stream 1: audio 2 channels
292 stream 2: subtitles (image)
293 Example: automatic stream selection
295 ffmpeg -i A.avi -i B.mp4 out1.mkv out2.wav -map 1:a -c:a copy out3.mov
297 There are three output files specified, and for the first two, no
299 "-map" options are set, so ffmpeg will select streams for these two
300 files automatically.
301 out1.mkv is a Matroska container file and accepts video, audio and
303 subtitle streams, so ffmpeg will try to select one of each type.For
304 video, it will select "stream 0" from B.mp4, which has the highest
305 resolution among all the input video streams.For audio, it will select
306 "stream 3" from B.mp4, since it has the greatest number of channels.For
307 subtitles, it will select "stream 2" from B.mp4, which is the first
308 subtitle stream from among A.avi and B.mp4.
309 out2.wav accepts only audio streams, so only "stream 3" from B.mp4 is
311 selected.
312 For out3.mov, since a "-map" option is set, no automatic stream
314 selection will occur. The "-map 1:a" option will select all audio
315 streams from the second input B.mp4. No other streams will be included
316 in this output file.
317 For the first two outputs, all included streams will be transcoded. The
319 encoders chosen will be the default ones registered by each output
320 format, which may not match the codec of the selected input streams.
321 For the third output, codec option for audio streams has been set to
323 "copy", so no decoding-filtering-encoding operations will occur, or can
324 occur. Packets of selected streams shall be conveyed from the input
325 file and muxed within the output file.
326 Example: automatic subtitles selection
328 ffmpeg -i C.mkv out1.mkv -c:s dvdsub -an out2.mkv
330 Although out1.mkv is a Matroska container file which accepts subtitle
332 streams, only a video and audio stream shall be selected. The subtitle
333 stream of C.mkv is image-based and the default subtitle encoder of the
334 Matroska muxer is text-based, so a transcode operation for the
335 subtitles is expected to fail and hence the stream isn't selected.
336 However, in out2.mkv, a subtitle encoder is specified in the command
337 and so, the subtitle stream is selected, in addition to the video
338 stream. The presence of "-an" disables audio stream selection for
339 out2.mkv.
340 Example: unlabeled filtergraph outputs
342 ffmpeg -i A.avi -i C.mkv -i B.mp4 -filter_complex "overlay" out1.mp4 out2.srt
344 A filtergraph is setup here using the "-filter_complex" option and
346 consists of a single video filter. The "overlay" filter requires
347 exactly two video inputs, but none are specified, so the first two
348 available video streams are used, those of A.avi and C.mkv. The output
349 pad of the filter has no label and so is sent to the first output file
350 out1.mp4. Due to this, automatic selection of the video stream is
351 skipped, which would have selected the stream in B.mp4. The audio
352 stream with most channels viz. "stream 3" in B.mp4, is chosen
353 automatically. No subtitle stream is chosen however, since the MP4
354 format has no default subtitle encoder registered, and the user hasn't
355 specified a subtitle encoder.
356 The 2nd output file, out2.srt, only accepts text-based subtitle
358 streams. So, even though the first subtitle stream available belongs to
359 C.mkv, it is image-based and hence skipped. The selected stream,
360 "stream 2" in B.mp4, is the first text-based subtitle stream.
361 Example: labeled filtergraph outputs
363 ffmpeg -i A.avi -i B.mp4 -i C.mkv -filter_complex "[1:v]hue=s=0[outv];overlay;aresample" \
365 -map '[outv]' -an out1.mp4 \
366 out2.mkv \
367 -map '[outv]' -map 1:a:0 out3.mkv
368 The above command will fail, as the output pad labelled "[outv]" has
370 been mapped twice. None of the output files shall be processed.
371 ffmpeg -i A.avi -i B.mp4 -i C.mkv -filter_complex "[1:v]hue=s=0[outv];overlay;aresample" \
373 -an out1.mp4 \
374 out2.mkv \
375 -map 1:a:0 out3.mkv
376 This command above will also fail as the hue filter output has a label,
378 "[outv]", and hasn't been mapped anywhere.
379 The command should be modified as follows,
381 ffmpeg -i A.avi -i B.mp4 -i C.mkv -filter_complex "[1:v]hue=s=0,split=2[outv1][outv2];overlay;aresample" \
383 -map '[outv1]' -an out1.mp4 \
384 out2.mkv \
385 -map '[outv2]' -map 1:a:0 out3.mkv
386 The video stream from B.mp4 is sent to the hue filter, whose output is
388 cloned once using the split filter, and both outputs labelled. Then a
389 copy each is mapped to the first and third output files.
390 The overlay filter, requiring two video inputs, uses the first two
392 unused video streams. Those are the streams from A.avi and C.mkv. The
393 overlay output isn't labelled, so it is sent to the first output file
394 out1.mp4, regardless of the presence of the "-map" option.
395 The aresample filter is sent the first unused audio stream, that of
397 A.avi. Since this filter output is also unlabelled, it too is mapped to
398 the first output file. The presence of "-an" only suppresses automatic
399 or manual stream selection of audio streams, not outputs sent from
400 filtergraphs. Both these mapped streams shall be ordered before the
401 mapped stream in out1.mp4.
402 The video, audio and subtitle streams mapped to "out2.mkv" are entirely
404 determined by automatic stream selection.
405 out3.mkv consists of the cloned video output from the hue filter and
407 the first audio stream from B.mp4.
408
410 All the numerical options, if not specified otherwise, accept a string
411 representing a number as input, which may be followed by one of the SI
412 unit prefixes, for example: 'K', 'M', or 'G'.
413 If 'i' is appended to the SI unit prefix, the complete prefix will be
415 interpreted as a unit prefix for binary multiples, which are based on
416 powers of 1024 instead of powers of 1000. Appending 'B' to the SI unit
417 prefix multiplies the value by 8. This allows using, for example: 'KB',
418 'MiB', 'G' and 'B' as number suffixes.
419 Options which do not take arguments are boolean options, and set the
421 corresponding value to true. They can be set to false by prefixing the
422 option name with "no". For example using "-nofoo" will set the boolean
423 option with name "foo" to false.
424 Stream specifiers
426 Some options are applied per-stream, e.g. bitrate or codec. Stream
427 specifiers are used to precisely specify which stream(s) a given option
428 belongs to.
429 A stream specifier is a string generally appended to the option name
431 and separated from it by a colon. E.g. "-codec:a:1 ac3" contains the
432 "a:1" stream specifier, which matches the second audio stream.
433 Therefore, it would select the ac3 codec for the second audio stream.
434 A stream specifier can match several streams, so that the option is
436 applied to all of them. E.g. the stream specifier in "-b:a 128k"
437 matches all audio streams.
438 An empty stream specifier matches all streams. For example, "-codec
440 copy" or "-codec: copy" would copy all the streams without reencoding.
441 Possible forms of stream specifiers are:
443 stream_index
445 Matches the stream with this index. E.g. "-threads:1 4" would set
446 the thread count for the second stream to 4. If stream_index is
447 used as an additional stream specifier (see below), then it selects
448 stream number stream_index from the matching streams. Stream
449 numbering is based on the order of the streams as detected by
450 libavformat except when a program ID is also specified. In this
451 case it is based on the ordering of the streams in the program.
452 stream_type[:additional_stream_specifier]
454 stream_type is one of following: 'v' or 'V' for video, 'a' for
455 audio, 's' for subtitle, 'd' for data, and 't' for attachments. 'v'
456 matches all video streams, 'V' only matches video streams which are
457 not attached pictures, video thumbnails or cover arts. If
458 additional_stream_specifier is used, then it matches streams which
459 both have this type and match the additional_stream_specifier.
460 Otherwise, it matches all streams of the specified type.
461 p:program_id[:additional_stream_specifier]
463 Matches streams which are in the program with the id program_id. If
464 additional_stream_specifier is used, then it matches streams which
465 both are part of the program and match the
466 additional_stream_specifier.
467 #stream_id or i:stream_id
469 Match the stream by stream id (e.g. PID in MPEG-TS container).
470 m:key[:value]
472 Matches streams with the metadata tag key having the specified
473 value. If value is not given, matches streams that contain the
474 given tag with any value.
475 u Matches streams with usable configuration, the codec must be
477 defined and the essential information such as video dimension or
478 audio sample rate must be present.
479 Note that in ffmpeg, matching by metadata will only work properly
481 for input files.
482 Generic options
484 These options are shared amongst the ff* tools.
485 -L Show license.
487 -h, -?, -help, --help [arg]
489 Show help. An optional parameter may be specified to print help
490 about a specific item. If no argument is specified, only basic (non
491 advanced) tool options are shown.
492 Possible values of arg are:
494 long
496 Print advanced tool options in addition to the basic tool
497 options.
498 full
500 Print complete list of options, including shared and private
501 options for encoders, decoders, demuxers, muxers, filters, etc.
502 decoder=decoder_name
504 Print detailed information about the decoder named
505 decoder_name. Use the -decoders option to get a list of all
506 decoders.
507 encoder=encoder_name
509 Print detailed information about the encoder named
510 encoder_name. Use the -encoders option to get a list of all
511 encoders.
512 demuxer=demuxer_name
514 Print detailed information about the demuxer named
515 demuxer_name. Use the -formats option to get a list of all
516 demuxers and muxers.
517 muxer=muxer_name
519 Print detailed information about the muxer named muxer_name.
520 Use the -formats option to get a list of all muxers and
521 demuxers.
522 filter=filter_name
524 Print detailed information about the filter named filter_name.
525 Use the -filters option to get a list of all filters.
526 bsf=bitstream_filter_name
528 Print detailed information about the bitstream filter named
529 bitstream_filter_name. Use the -bsfs option to get a list of
530 all bitstream filters.
531 protocol=protocol_name
533 Print detailed information about the protocol named
534 protocol_name. Use the -protocols option to get a list of all
535 protocols.
536 -version
538 Show version.
539 -buildconf
541 Show the build configuration, one option per line.
542 -formats
544 Show available formats (including devices).
545 -demuxers
547 Show available demuxers.
548 -muxers
550 Show available muxers.
551 -devices
553 Show available devices.
554 -codecs
556 Show all codecs known to libavcodec.
557 Note that the term 'codec' is used throughout this documentation as
559 a shortcut for what is more correctly called a media bitstream
560 format.
561 -decoders
563 Show available decoders.
564 -encoders
566 Show all available encoders.
567 -bsfs
569 Show available bitstream filters.
570 -protocols
572 Show available protocols.
573 -filters
575 Show available libavfilter filters.
576 -pix_fmts
578 Show available pixel formats.
579 -sample_fmts
581 Show available sample formats.
582 -layouts
584 Show channel names and standard channel layouts.
585 -dispositions
587 Show stream dispositions.
588 -colors
590 Show recognized color names.
591 -sources device[,opt1=val1[,opt2=val2]...]
593 Show autodetected sources of the input device. Some devices may
594 provide system-dependent source names that cannot be autodetected.
595 The returned list cannot be assumed to be always complete.
596 ffmpeg -sources pulse,server=192.168.0.4
598 -sinks device[,opt1=val1[,opt2=val2]...]
600 Show autodetected sinks of the output device. Some devices may
601 provide system-dependent sink names that cannot be autodetected.
602 The returned list cannot be assumed to be always complete.
603 ffmpeg -sinks pulse,server=192.168.0.4
605 -loglevel [flags+]loglevel | -v [flags+]loglevel
607 Set logging level and flags used by the library.
608 The optional flags prefix can consist of the following values:
610 repeat
612 Indicates that repeated log output should not be compressed to
613 the first line and the "Last message repeated n times" line
614 will be omitted.
615 level
617 Indicates that log output should add a "[level]" prefix to each
618 message line. This can be used as an alternative to log
619 coloring, e.g. when dumping the log to file.
620 Flags can also be used alone by adding a '+'/'-' prefix to
622 set/reset a single flag without affecting other flags or changing
623 loglevel. When setting both flags and loglevel, a '+' separator is
624 expected between the last flags value and before loglevel.
625 loglevel is a string or a number containing one of the following
627 values:
628 quiet, -8
630 Show nothing at all; be silent.
631 panic, 0
633 Only show fatal errors which could lead the process to crash,
634 such as an assertion failure. This is not currently used for
635 anything.
636 fatal, 8
638 Only show fatal errors. These are errors after which the
639 process absolutely cannot continue.
640 error, 16
642 Show all errors, including ones which can be recovered from.
643 warning, 24
645 Show all warnings and errors. Any message related to possibly
646 incorrect or unexpected events will be shown.
647 info, 32
649 Show informative messages during processing. This is in
650 addition to warnings and errors. This is the default value.
651 verbose, 40
653 Same as "info", except more verbose.
654 debug, 48
656 Show everything, including debugging information.
657 trace, 56
659 For example to enable repeated log output, add the "level" prefix,
661 and set loglevel to "verbose":
662 ffmpeg -loglevel repeat+level+verbose -i input output
664 Another example that enables repeated log output without affecting
666 current state of "level" prefix flag or loglevel:
667 ffmpeg [...] -loglevel +repeat
669 By default the program logs to stderr. If coloring is supported by
671 the terminal, colors are used to mark errors and warnings. Log
672 coloring can be disabled setting the environment variable
673 AV_LOG_FORCE_NOCOLOR, or can be forced setting the environment
674 variable AV_LOG_FORCE_COLOR.
675 -report
677 Dump full command line and log output to a file named
678 "program-YYYYMMDD-HHMMSS.log" in the current directory. This file
679 can be useful for bug reports. It also implies "-loglevel debug".
680 Setting the environment variable FFREPORT to any value has the same
682 effect. If the value is a ':'-separated key=value sequence, these
683 options will affect the report; option values must be escaped if
684 they contain special characters or the options delimiter ':' (see
685 the ``Quoting and escaping'' section in the ffmpeg-utils manual).
686 The following options are recognized:
688 file
690 set the file name to use for the report; %p is expanded to the
691 name of the program, %t is expanded to a timestamp, "%%" is
692 expanded to a plain "%"
693 level
695 set the log verbosity level using a numerical value (see
696 "-loglevel").
697 For example, to output a report to a file named ffreport.log using
699 a log level of 32 (alias for log level "info"):
700 FFREPORT=file=ffreport.log:level=32 ffmpeg -i input output
702 Errors in parsing the environment variable are not fatal, and will
704 not appear in the report.
705 -hide_banner
707 Suppress printing banner.
708 All FFmpeg tools will normally show a copyright notice, build
710 options and library versions. This option can be used to suppress
711 printing this information.
712 -cpuflags flags (global)
714 Allows setting and clearing cpu flags. This option is intended for
715 testing. Do not use it unless you know what you're doing.
716 ffmpeg -cpuflags -sse+mmx ...
718 ffmpeg -cpuflags mmx ...
719 ffmpeg -cpuflags 0 ...
720 Possible flags for this option are:
722 x86
724 mmx
725 mmxext
726 sse
727 sse2
728 sse2slow
729 sse3
730 sse3slow
731 ssse3
732 atom
733 sse4.1
734 sse4.2
735 avx
736 avx2
737 xop
738 fma3
739 fma4
740 3dnow
741 3dnowext
742 bmi1
743 bmi2
744 cmov
745 ARM
746 armv5te
747 armv6
748 armv6t2
749 vfp
750 vfpv3
751 neon
752 setend
753 AArch64
754 armv8
755 vfp
756 neon
757 PowerPC
758 altivec
759 Specific Processors
760 pentium2
761 pentium3
762 pentium4
763 k6
764 k62
765 athlon
766 athlonxp
767 k8
768 -cpucount count (global)
769 Override detection of CPU count. This option is intended for
770 testing. Do not use it unless you know what you're doing.
771 ffmpeg -cpucount 2
773 -max_alloc bytes
775 Set the maximum size limit for allocating a block on the heap by
776 ffmpeg's family of malloc functions. Exercise extreme caution when
777 using this option. Don't use if you do not understand the full
778 consequence of doing so. Default is INT_MAX.
779 AVOptions
781 These options are provided directly by the libavformat, libavdevice and
782 libavcodec libraries. To see the list of available AVOptions, use the
783 -help option. They are separated into two categories:
784 generic
786 These options can be set for any container, codec or device.
787 Generic options are listed under AVFormatContext options for
788 containers/devices and under AVCodecContext options for codecs.
789 private
791 These options are specific to the given container, device or codec.
792 Private options are listed under their corresponding
793 containers/devices/codecs.
794 For example to write an ID3v2.3 header instead of a default ID3v2.4 to
796 an MP3 file, use the id3v2_version private option of the MP3 muxer:
797 ffmpeg -i input.flac -id3v2_version 3 out.mp3
799 All codec AVOptions are per-stream, and thus a stream specifier should
801 be attached to them:
802 ffmpeg -i multichannel.mxf -map 0:v:0 -map 0:a:0 -map 0:a:0 -c:a:0 ac3 -b:a:0 640k -ac:a:1 2 -c:a:1 aac -b:2 128k out.mp4
804 In the above example, a multichannel audio stream is mapped twice for
806 output. The first instance is encoded with codec ac3 and bitrate 640k.
807 The second instance is downmixed to 2 channels and encoded with codec
808 aac. A bitrate of 128k is specified for it using absolute index of the
809 output stream.
810 Note: the -nooption syntax cannot be used for boolean AVOptions, use
812 -option 0/-option 1.
813 Note: the old undocumented way of specifying per-stream AVOptions by
815 prepending v/a/s to the options name is now obsolete and will be
816 removed soon.
817 Main options
819 -f fmt (input/output)
820 Force input or output file format. The format is normally auto
821 detected for input files and guessed from the file extension for
822 output files, so this option is not needed in most cases.
823 -i url (input)
825 input file url
826 -y (global)
828 Overwrite output files without asking.
829 -n (global)
831 Do not overwrite output files, and exit immediately if a specified
832 output file already exists.
833 -stream_loop number (input)
835 Set number of times input stream shall be looped. Loop 0 means no
836 loop, loop -1 means infinite loop.
837 -recast_media (global)
839 Allow forcing a decoder of a different media type than the one
840 detected or designated by the demuxer. Useful for decoding media
841 data muxed as data streams.
842 -c[:stream_specifier] codec (input/output,per-stream)
844 -codec[:stream_specifier] codec (input/output,per-stream)
845 Select an encoder (when used before an output file) or a decoder
846 (when used before an input file) for one or more streams. codec is
847 the name of a decoder/encoder or a special value "copy" (output
848 only) to indicate that the stream is not to be re-encoded.
849 For example
851 ffmpeg -i INPUT -map 0 -c:v libx264 -c:a copy OUTPUT
853 encodes all video streams with libx264 and copies all audio
855 streams.
856 For each stream, the last matching "c" option is applied, so
858 ffmpeg -i INPUT -map 0 -c copy -c:v:1 libx264 -c:a:137 libvorbis OUTPUT
860 will copy all the streams except the second video, which will be
862 encoded with libx264, and the 138th audio, which will be encoded
863 with libvorbis.
864 -t duration (input/output)
866 When used as an input option (before "-i"), limit the duration of
867 data read from the input file.
868 When used as an output option (before an output url), stop writing
870 the output after its duration reaches duration.
871 duration must be a time duration specification, see the Time
873 duration section in the ffmpeg-utils(1) manual.
874 -to and -t are mutually exclusive and -t has priority.
876 -to position (input/output)
878 Stop writing the output or reading the input at position. position
879 must be a time duration specification, see the Time duration
880 section in the ffmpeg-utils(1) manual.
881 -to and -t are mutually exclusive and -t has priority.
883 -fs limit_size (output)
885 Set the file size limit, expressed in bytes. No further chunk of
886 bytes is written after the limit is exceeded. The size of the
887 output file is slightly more than the requested file size.
888 -ss position (input/output)
890 When used as an input option (before "-i"), seeks in this input
891 file to position. Note that in most formats it is not possible to
892 seek exactly, so ffmpeg will seek to the closest seek point before
893 position. When transcoding and -accurate_seek is enabled (the
894 default), this extra segment between the seek point and position
895 will be decoded and discarded. When doing stream copy or when
896 -noaccurate_seek is used, it will be preserved.
897 When used as an output option (before an output url), decodes but
899 discards input until the timestamps reach position.
900 position must be a time duration specification, see the Time
902 duration section in the ffmpeg-utils(1) manual.
903 -sseof position (input)
905 Like the "-ss" option but relative to the "end of file". That is
906 negative values are earlier in the file, 0 is at EOF.
907 -isync input_index (input)
909 Assign an input as a sync source.
910 This will take the difference between the start times of the target
912 and reference inputs and offset the timestamps of the target file
913 by that difference. The source timestamps of the two inputs should
914 derive from the same clock source for expected results. If "copyts"
915 is set then "start_at_zero" must also be set. If either of the
916 inputs has no starting timestamp then no sync adjustment is made.
917 Acceptable values are those that refer to a valid ffmpeg input
919 index. If the sync reference is the target index itself or -1, then
920 no adjustment is made to target timestamps. A sync reference may
921 not itself be synced to any other input.
922 Default value is -1.
924 -itsoffset offset (input)
926 Set the input time offset.
927 offset must be a time duration specification, see the Time duration
929 section in the ffmpeg-utils(1) manual.
930 The offset is added to the timestamps of the input files.
932 Specifying a positive offset means that the corresponding streams
933 are delayed by the time duration specified in offset.
934 -itsscale scale (input,per-stream)
936 Rescale input timestamps. scale should be a floating point number.
937 -timestamp date (output)
939 Set the recording timestamp in the container.
940 date must be a date specification, see the Date section in the
942 ffmpeg-utils(1) manual.
943 -metadata[:metadata_specifier] key=value (output,per-metadata)
945 Set a metadata key/value pair.
946 An optional metadata_specifier may be given to set metadata on
948 streams, chapters or programs. See "-map_metadata" documentation
949 for details.
950 This option overrides metadata set with "-map_metadata". It is also
952 possible to delete metadata by using an empty value.
953 For example, for setting the title in the output file:
955 ffmpeg -i in.avi -metadata title="my title" out.flv
957 To set the language of the first audio stream:
959 ffmpeg -i INPUT -metadata:s:a:0 language=eng OUTPUT
961 -disposition[:stream_specifier] value (output,per-stream)
963 Sets the disposition for a stream.
964 By default, the disposition is copied from the input stream, unless
966 the output stream this option applies to is fed by a complex
967 filtergraph - in that case the disposition is unset by default.
968 value is a sequence of items separated by '+' or '-'. The first
970 item may also be prefixed with '+' or '-', in which case this
971 option modifies the default value. Otherwise (the first item is not
972 prefixed) this options overrides the default value. A '+' prefix
973 adds the given disposition, '-' removes it. It is also possible to
974 clear the disposition by setting it to 0.
975 If no "-disposition" options were specified for an output file,
977 ffmpeg will automatically set the 'default' disposition on the
978 first stream of each type, when there are multiple streams of this
979 type in the output file and no stream of that type is already
980 marked as default.
981 The "-dispositions" option lists the known dispositions.
983 For example, to make the second audio stream the default stream:
985 ffmpeg -i in.mkv -c copy -disposition:a:1 default out.mkv
987 To make the second subtitle stream the default stream and remove
989 the default disposition from the first subtitle stream:
990 ffmpeg -i in.mkv -c copy -disposition:s:0 0 -disposition:s:1 default out.mkv
992 To add an embedded cover/thumbnail:
994 ffmpeg -i in.mp4 -i IMAGE -map 0 -map 1 -c copy -c:v:1 png -disposition:v:1 attached_pic out.mp4
996 Not all muxers support embedded thumbnails, and those who do, only
998 support a few formats, like JPEG or PNG.
999 -program
1001 [title=title:][program_num=program_num:]st=stream[:st=stream...]
1002 (output)
1003 Creates a program with the specified title, program_num and adds
1004 the specified stream(s) to it.
1005 -target type (output)
1007 Specify target file type ("vcd", "svcd", "dvd", "dv", "dv50"). type
1008 may be prefixed with "pal-", "ntsc-" or "film-" to use the
1009 corresponding standard. All the format options (bitrate, codecs,
1010 buffer sizes) are then set automatically. You can just type:
1011 ffmpeg -i myfile.avi -target vcd /tmp/vcd.mpg
1013 Nevertheless you can specify additional options as long as you know
1015 they do not conflict with the standard, as in:
1016 ffmpeg -i myfile.avi -target vcd -bf 2 /tmp/vcd.mpg
1018 The parameters set for each target are as follows.
1020 VCD
1022 <pal>:
1024 -f vcd -muxrate 1411200 -muxpreload 0.44 -packetsize 2324
1025 -s 352x288 -r 25
1026 -codec:v mpeg1video -g 15 -b:v 1150k -maxrate:v 1150k -minrate:v 1150k -bufsize:v 327680
1027 -ar 44100 -ac 2
1028 -codec:a mp2 -b:a 224k
1029 <ntsc>:
1031 -f vcd -muxrate 1411200 -muxpreload 0.44 -packetsize 2324
1032 -s 352x240 -r 30000/1001
1033 -codec:v mpeg1video -g 18 -b:v 1150k -maxrate:v 1150k -minrate:v 1150k -bufsize:v 327680
1034 -ar 44100 -ac 2
1035 -codec:a mp2 -b:a 224k
1036 <film>:
1038 -f vcd -muxrate 1411200 -muxpreload 0.44 -packetsize 2324
1039 -s 352x240 -r 24000/1001
1040 -codec:v mpeg1video -g 18 -b:v 1150k -maxrate:v 1150k -minrate:v 1150k -bufsize:v 327680
1041 -ar 44100 -ac 2
1042 -codec:a mp2 -b:a 224k
1043 SVCD
1045 <pal>:
1047 -f svcd -packetsize 2324
1048 -s 480x576 -pix_fmt yuv420p -r 25
1049 -codec:v mpeg2video -g 15 -b:v 2040k -maxrate:v 2516k -minrate:v 0 -bufsize:v 1835008 -scan_offset 1
1050 -ar 44100
1051 -codec:a mp2 -b:a 224k
1052 <ntsc>:
1054 -f svcd -packetsize 2324
1055 -s 480x480 -pix_fmt yuv420p -r 30000/1001
1056 -codec:v mpeg2video -g 18 -b:v 2040k -maxrate:v 2516k -minrate:v 0 -bufsize:v 1835008 -scan_offset 1
1057 -ar 44100
1058 -codec:a mp2 -b:a 224k
1059 <film>:
1061 -f svcd -packetsize 2324
1062 -s 480x480 -pix_fmt yuv420p -r 24000/1001
1063 -codec:v mpeg2video -g 18 -b:v 2040k -maxrate:v 2516k -minrate:v 0 -bufsize:v 1835008 -scan_offset 1
1064 -ar 44100
1065 -codec:a mp2 -b:a 224k
1066 DVD
1068 <pal>:
1070 -f dvd -muxrate 10080k -packetsize 2048
1071 -s 720x576 -pix_fmt yuv420p -r 25
1072 -codec:v mpeg2video -g 15 -b:v 6000k -maxrate:v 9000k -minrate:v 0 -bufsize:v 1835008
1073 -ar 48000
1074 -codec:a ac3 -b:a 448k
1075 <ntsc>:
1077 -f dvd -muxrate 10080k -packetsize 2048
1078 -s 720x480 -pix_fmt yuv420p -r 30000/1001
1079 -codec:v mpeg2video -g 18 -b:v 6000k -maxrate:v 9000k -minrate:v 0 -bufsize:v 1835008
1080 -ar 48000
1081 -codec:a ac3 -b:a 448k
1082 <film>:
1084 -f dvd -muxrate 10080k -packetsize 2048
1085 -s 720x480 -pix_fmt yuv420p -r 24000/1001
1086 -codec:v mpeg2video -g 18 -b:v 6000k -maxrate:v 9000k -minrate:v 0 -bufsize:v 1835008
1087 -ar 48000
1088 -codec:a ac3 -b:a 448k
1089 DV
1091 <pal>:
1093 -f dv
1094 -s 720x576 -pix_fmt yuv420p -r 25
1095 -ar 48000 -ac 2
1096 <ntsc>:
1098 -f dv
1099 -s 720x480 -pix_fmt yuv411p -r 30000/1001
1100 -ar 48000 -ac 2
1101 <film>:
1103 -f dv
1104 -s 720x480 -pix_fmt yuv411p -r 24000/1001
1105 -ar 48000 -ac 2
1106 The "dv50" target is identical to the "dv" target except that the
1108 pixel format set is "yuv422p" for all three standards.
1109 Any user-set value for a parameter above will override the target
1111 preset value. In that case, the output may not comply with the
1112 target standard.
1113 -dn (input/output)
1115 As an input option, blocks all data streams of a file from being
1116 filtered or being automatically selected or mapped for any output.
1117 See "-discard" option to disable streams individually.
1118 As an output option, disables data recording i.e. automatic
1120 selection or mapping of any data stream. For full manual control
1121 see the "-map" option.
1122 -dframes number (output)
1124 Set the number of data frames to output. This is an obsolete alias
1125 for "-frames:d", which you should use instead.
1126 -frames[:stream_specifier] framecount (output,per-stream)
1128 Stop writing to the stream after framecount frames.
1129 -q[:stream_specifier] q (output,per-stream)
1131 -qscale[:stream_specifier] q (output,per-stream)
1132 Use fixed quality scale (VBR). The meaning of q/qscale is codec-
1133 dependent. If qscale is used without a stream_specifier then it
1134 applies only to the video stream, this is to maintain compatibility
1135 with previous behavior and as specifying the same codec specific
1136 value to 2 different codecs that is audio and video generally is
1137 not what is intended when no stream_specifier is used.
1138 -filter[:stream_specifier] filtergraph (output,per-stream)
1140 Create the filtergraph specified by filtergraph and use it to
1141 filter the stream.
1142 filtergraph is a description of the filtergraph to apply to the
1144 stream, and must have a single input and a single output of the
1145 same type of the stream. In the filtergraph, the input is
1146 associated to the label "in", and the output to the label "out".
1147 See the ffmpeg-filters manual for more information about the
1148 filtergraph syntax.
1149 See the -filter_complex option if you want to create filtergraphs
1151 with multiple inputs and/or outputs.
1152 -filter_script[:stream_specifier] filename (output,per-stream)
1154 This option is similar to -filter, the only difference is that its
1155 argument is the name of the file from which a filtergraph
1156 description is to be read.
1157 -reinit_filter[:stream_specifier] integer (input,per-stream)
1159 This boolean option determines if the filtergraph(s) to which this
1160 stream is fed gets reinitialized when input frame parameters change
1161 mid-stream. This option is enabled by default as most video and all
1162 audio filters cannot handle deviation in input frame properties.
1163 Upon reinitialization, existing filter state is lost, like e.g. the
1164 frame count "n" reference available in some filters. Any frames
1165 buffered at time of reinitialization are lost. The properties
1166 where a change triggers reinitialization are, for video, frame
1167 resolution or pixel format; for audio, sample format, sample rate,
1168 channel count or channel layout.
1169 -filter_threads nb_threads (global)
1171 Defines how many threads are used to process a filter pipeline.
1172 Each pipeline will produce a thread pool with this many threads
1173 available for parallel processing. The default is the number of
1174 available CPUs.
1175 -pre[:stream_specifier] preset_name (output,per-stream)
1177 Specify the preset for matching stream(s).
1178 -stats (global)
1180 Print encoding progress/statistics. It is on by default, to
1181 explicitly disable it you need to specify "-nostats".
1182 -stats_period time (global)
1184 Set period at which encoding progress/statistics are updated.
1185 Default is 0.5 seconds.
1186 -progress url (global)
1188 Send program-friendly progress information to url.
1189 Progress information is written periodically and at the end of the
1191 encoding process. It is made of "key=value" lines. key consists of
1192 only alphanumeric characters. The last key of a sequence of
1193 progress information is always "progress".
1194 The update period is set using "-stats_period".
1196 -stdin
1198 Enable interaction on standard input. On by default unless standard
1199 input is used as an input. To explicitly disable interaction you
1200 need to specify "-nostdin".
1201 Disabling interaction on standard input is useful, for example, if
1203 ffmpeg is in the background process group. Roughly the same result
1204 can be achieved with "ffmpeg ... < /dev/null" but it requires a
1205 shell.
1206 -debug_ts (global)
1208 Print timestamp information. It is off by default. This option is
1209 mostly useful for testing and debugging purposes, and the output
1210 format may change from one version to another, so it should not be
1211 employed by portable scripts.
1212 See also the option "-fdebug ts".
1214 -attach filename (output)
1216 Add an attachment to the output file. This is supported by a few
1217 formats like Matroska for e.g. fonts used in rendering subtitles.
1218 Attachments are implemented as a specific type of stream, so this
1219 option will add a new stream to the file. It is then possible to
1220 use per-stream options on this stream in the usual way. Attachment
1221 streams created with this option will be created after all the
1222 other streams (i.e. those created with "-map" or automatic
1223 mappings).
1224 Note that for Matroska you also have to set the mimetype metadata
1226 tag:
1227 ffmpeg -i INPUT -attach DejaVuSans.ttf -metadata:s:2 mimetype=application/x-truetype-font out.mkv
1229 (assuming that the attachment stream will be third in the output
1231 file).
1232 -dump_attachment[:stream_specifier] filename (input,per-stream)
1234 Extract the matching attachment stream into a file named filename.
1235 If filename is empty, then the value of the "filename" metadata tag
1236 will be used.
1237 E.g. to extract the first attachment to a file named 'out.ttf':
1239 ffmpeg -dump_attachment:t:0 out.ttf -i INPUT
1241 To extract all attachments to files determined by the "filename"
1243 tag:
1244 ffmpeg -dump_attachment:t "" -i INPUT
1246 Technical note -- attachments are implemented as codec extradata,
1248 so this option can actually be used to extract extradata from any
1249 stream, not just attachments.
1250 Video Options
1252 -vframes number (output)
1253 Set the number of video frames to output. This is an obsolete alias
1254 for "-frames:v", which you should use instead.
1255 -r[:stream_specifier] fps (input/output,per-stream)
1257 Set frame rate (Hz value, fraction or abbreviation).
1258 As an input option, ignore any timestamps stored in the file and
1260 instead generate timestamps assuming constant frame rate fps. This
1261 is not the same as the -framerate option used for some input
1262 formats like image2 or v4l2 (it used to be the same in older
1263 versions of FFmpeg). If in doubt use -framerate instead of the
1264 input option -r.
1265 As an output option:
1267 video encoding
1269 Duplicate or drop frames right before encoding them to achieve
1270 constant output frame rate fps.
1271 video streamcopy
1273 Indicate to the muxer that fps is the stream frame rate. No
1274 data is dropped or duplicated in this case. This may produce
1275 invalid files if fps does not match the actual stream frame
1276 rate as determined by packet timestamps. See also the "setts"
1277 bitstream filter.
1278 -fpsmax[:stream_specifier] fps (output,per-stream)
1280 Set maximum frame rate (Hz value, fraction or abbreviation).
1281 Clamps output frame rate when output framerate is auto-set and is
1283 higher than this value. Useful in batch processing or when input
1284 framerate is wrongly detected as very high. It cannot be set
1285 together with "-r". It is ignored during streamcopy.
1286 -s[:stream_specifier] size (input/output,per-stream)
1288 Set frame size.
1289 As an input option, this is a shortcut for the video_size private
1291 option, recognized by some demuxers for which the frame size is
1292 either not stored in the file or is configurable -- e.g. raw video
1293 or video grabbers.
1294 As an output option, this inserts the "scale" video filter to the
1296 end of the corresponding filtergraph. Please use the "scale" filter
1297 directly to insert it at the beginning or some other place.
1298 The format is wxh (default - same as source).
1300 -aspect[:stream_specifier] aspect (output,per-stream)
1302 Set the video display aspect ratio specified by aspect.
1303 aspect can be a floating point number string, or a string of the
1305 form num:den, where num and den are the numerator and denominator
1306 of the aspect ratio. For example "4:3", "16:9", "1.3333", and
1307 "1.7777" are valid argument values.
1308 If used together with -vcodec copy, it will affect the aspect ratio
1310 stored at container level, but not the aspect ratio stored in
1311 encoded frames, if it exists.
1312 -display_rotation[:stream_specifier] rotation (input,per-stream)
1314 Set video rotation metadata.
1315 rotation is a decimal number specifying the amount in degree by
1317 which the video should be rotated counter-clockwise before being
1318 displayed.
1319 This option overrides the rotation/display transform metadata
1321 stored in the file, if any. When the video is being transcoded
1322 (rather than copied) and "-autorotate" is enabled, the video will
1323 be rotated at the filtering stage. Otherwise, the metadata will be
1324 written into the output file if the muxer supports it.
1325 If the "-display_hflip" and/or "-display_vflip" options are given,
1327 they are applied after the rotation specified by this option.
1328 -display_hflip[:stream_specifier] (input,per-stream)
1330 Set whether on display the image should be horizontally flipped.
1331 See the "-display_rotation" option for more details.
1333 -display_vflip[:stream_specifier] (input,per-stream)
1335 Set whether on display the image should be vertically flipped.
1336 See the "-display_rotation" option for more details.
1338 -vn (input/output)
1340 As an input option, blocks all video streams of a file from being
1341 filtered or being automatically selected or mapped for any output.
1342 See "-discard" option to disable streams individually.
1343 As an output option, disables video recording i.e. automatic
1345 selection or mapping of any video stream. For full manual control
1346 see the "-map" option.
1347 -vcodec codec (output)
1349 Set the video codec. This is an alias for "-codec:v".
1350 -pass[:stream_specifier] n (output,per-stream)
1352 Select the pass number (1 or 2). It is used to do two-pass video
1353 encoding. The statistics of the video are recorded in the first
1354 pass into a log file (see also the option -passlogfile), and in the
1355 second pass that log file is used to generate the video at the
1356 exact requested bitrate. On pass 1, you may just deactivate audio
1357 and set output to null, examples for Windows and Unix:
1358 ffmpeg -i foo.mov -c:v libxvid -pass 1 -an -f rawvideo -y NUL
1360 ffmpeg -i foo.mov -c:v libxvid -pass 1 -an -f rawvideo -y /dev/null
1361 -passlogfile[:stream_specifier] prefix (output,per-stream)
1363 Set two-pass log file name prefix to prefix, the default file name
1364 prefix is ``ffmpeg2pass''. The complete file name will be
1365 PREFIX-N.log, where N is a number specific to the output stream
1366 -vf filtergraph (output)
1368 Create the filtergraph specified by filtergraph and use it to
1369 filter the stream.
1370 This is an alias for "-filter:v", see the -filter option.
1372 -autorotate
1374 Automatically rotate the video according to file metadata. Enabled
1375 by default, use -noautorotate to disable it.
1376 -autoscale
1378 Automatically scale the video according to the resolution of first
1379 frame. Enabled by default, use -noautoscale to disable it. When
1380 autoscale is disabled, all output frames of filter graph might not
1381 be in the same resolution and may be inadequate for some
1382 encoder/muxer. Therefore, it is not recommended to disable it
1383 unless you really know what you are doing. Disable autoscale at
1384 your own risk.
1385 Advanced Video options
1387 -pix_fmt[:stream_specifier] format (input/output,per-stream)
1388 Set pixel format. Use "-pix_fmts" to show all the supported pixel
1389 formats. If the selected pixel format can not be selected, ffmpeg
1390 will print a warning and select the best pixel format supported by
1391 the encoder. If pix_fmt is prefixed by a "+", ffmpeg will exit
1392 with an error if the requested pixel format can not be selected,
1393 and automatic conversions inside filtergraphs are disabled. If
1394 pix_fmt is a single "+", ffmpeg selects the same pixel format as
1395 the input (or graph output) and automatic conversions are disabled.
1396 -sws_flags flags (input/output)
1398 Set SwScaler flags.
1399 -rc_override[:stream_specifier] override (output,per-stream)
1401 Rate control override for specific intervals, formatted as
1402 "int,int,int" list separated with slashes. Two first values are the
1403 beginning and end frame numbers, last one is quantizer to use if
1404 positive, or quality factor if negative.
1405 -psnr
1407 Calculate PSNR of compressed frames. This option is deprecated,
1408 pass the PSNR flag to the encoder instead, using "-flags +psnr".
1409 -vstats
1411 Dump video coding statistics to vstats_HHMMSS.log.
1412 -vstats_file file
1414 Dump video coding statistics to file.
1415 -vstats_version file
1417 Specifies which version of the vstats format to use. Default is 2.
1418 version = 1 :
1420 "frame= %5d q= %2.1f PSNR= %6.2f f_size= %6d s_size= %8.0fkB time=
1422 %0.3f br= %7.1fkbits/s avg_br= %7.1fkbits/s"
1423 version > 1:
1425 "out= %2d st= %2d frame= %5d q= %2.1f PSNR= %6.2f f_size= %6d
1427 s_size= %8.0fkB time= %0.3f br= %7.1fkbits/s avg_br= %7.1fkbits/s"
1428 -top[:stream_specifier] n (output,per-stream)
1430 top=1/bottom=0/auto=-1 field first
1431 -vtag fourcc/tag (output)
1433 Force video tag/fourcc. This is an alias for "-tag:v".
1434 -qphist (global)
1436 Show QP histogram
1437 -vbsf bitstream_filter
1439 Deprecated see -bsf
1440 -force_key_frames[:stream_specifier] time[,time...] (output,per-stream)
1442 -force_key_frames[:stream_specifier] expr:expr (output,per-stream)
1443 -force_key_frames[:stream_specifier] source (output,per-stream)
1444 -force_key_frames[:stream_specifier] source_no_drop (output,per-stream)
1445 force_key_frames can take arguments of the following form:
1446 time[,time...]
1448 If the argument consists of timestamps, ffmpeg will round the
1449 specified times to the nearest output timestamp as per the
1450 encoder time base and force a keyframe at the first frame
1451 having timestamp equal or greater than the computed timestamp.
1452 Note that if the encoder time base is too coarse, then the
1453 keyframes may be forced on frames with timestamps lower than
1454 the specified time. The default encoder time base is the
1455 inverse of the output framerate but may be set otherwise via
1456 "-enc_time_base".
1457 If one of the times is ""chapters"[delta]", it is expanded into
1459 the time of the beginning of all chapters in the file, shifted
1460 by delta, expressed as a time in seconds. This option can be
1461 useful to ensure that a seek point is present at a chapter mark
1462 or any other designated place in the output file.
1463 For example, to insert a key frame at 5 minutes, plus key
1465 frames 0.1 second before the beginning of every chapter:
1466 -force_key_frames 0:05:00,chapters-0.1
1468 expr:expr
1470 If the argument is prefixed with "expr:", the string expr is
1471 interpreted like an expression and is evaluated for each frame.
1472 A key frame is forced in case the evaluation is non-zero.
1473 The expression in expr can contain the following constants:
1475 n the number of current processed frame, starting from 0
1477 n_forced
1479 the number of forced frames
1480 prev_forced_n
1482 the number of the previous forced frame, it is "NAN" when
1483 no keyframe was forced yet
1484 prev_forced_t
1486 the time of the previous forced frame, it is "NAN" when no
1487 keyframe was forced yet
1488 t the time of the current processed frame
1490 For example to force a key frame every 5 seconds, you can
1492 specify:
1493 -force_key_frames expr:gte(t,n_forced*5)
1495 To force a key frame 5 seconds after the time of the last
1497 forced one, starting from second 13:
1498 -force_key_frames expr:if(isnan(prev_forced_t),gte(t,13),gte(t,prev_forced_t+5))
1500 source
1502 If the argument is "source", ffmpeg will force a key frame if
1503 the current frame being encoded is marked as a key frame in its
1504 source.
1505 source_no_drop
1507 If the argument is "source_no_drop", ffmpeg will force a key
1508 frame if the current frame being encoded is marked as a key
1509 frame in its source. In cases where this particular source
1510 frame has to be dropped, enforce the next available frame to
1511 become a key frame instead.
1512 Note that forcing too many keyframes is very harmful for the
1514 lookahead algorithms of certain encoders: using fixed-GOP options
1515 or similar would be more efficient.
1516 -copyinkf[:stream_specifier] (output,per-stream)
1518 When doing stream copy, copy also non-key frames found at the
1519 beginning.
1520 -init_hw_device type[=name][:device[,key=value...]]
1522 Initialise a new hardware device of type type called name, using
1523 the given device parameters. If no name is specified it will
1524 receive a default name of the form "type%d".
1525 The meaning of device and the following arguments depends on the
1527 device type:
1528 cuda
1530 device is the number of the CUDA device.
1531 The following options are recognized:
1533 primary_ctx
1535 If set to 1, uses the primary device context instead of
1536 creating a new one.
1537 Examples:
1539 -init_hw_device cuda:1
1541 Choose the second device on the system.
1542 -init_hw_device cuda:0,primary_ctx=1
1544 Choose the first device and use the primary device context.
1545 dxva2
1547 device is the number of the Direct3D 9 display adapter.
1548 d3d11va
1550 device is the number of the Direct3D 11 display adapter.
1551 vaapi
1553 device is either an X11 display name or a DRM render node. If
1554 not specified, it will attempt to open the default X11 display
1555 ($DISPLAY) and then the first DRM render node
1556 (/dev/dri/renderD128).
1557 vdpau
1559 device is an X11 display name. If not specified, it will
1560 attempt to open the default X11 display ($DISPLAY).
1561 qsv device selects a value in MFX_IMPL_*. Allowed values are:
1563 auto
1565 sw
1566 hw
1567 auto_any
1568 hw_any
1569 hw2
1570 hw3
1571 hw4
1572 If not specified, auto_any is used. (Note that it may be
1574 easier to achieve the desired result for QSV by creating the
1575 platform-appropriate subdevice (dxva2 or d3d11va or vaapi) and
1576 then deriving a QSV device from that.)
1577 Alternatively, child_device_type helps to choose platform-
1579 appropriate subdevice type. On Windows d3d11va is used as
1580 default subdevice type.
1581 Examples:
1583 -init_hw_device qsv:hw,child_device_type=d3d11va
1585 Choose the GPU subdevice with type d3d11va and create QSV
1586 device with MFX_IMPL_HARDWARE.
1587 -init_hw_device qsv:hw,child_device_type=dxva2
1589 Choose the GPU subdevice with type dxva2 and create QSV
1590 device with MFX_IMPL_HARDWARE.
1591 opencl
1593 device selects the platform and device as
1594 platform_index.device_index.
1595 The set of devices can also be filtered using the key-value
1597 pairs to find only devices matching particular platform or
1598 device strings.
1599 The strings usable as filters are:
1601 platform_profile
1603 platform_version
1604 platform_name
1605 platform_vendor
1606 platform_extensions
1607 device_name
1608 device_vendor
1609 driver_version
1610 device_version
1611 device_profile
1612 device_extensions
1613 device_type
1614 The indices and filters must together uniquely select a device.
1616 Examples:
1618 -init_hw_device opencl:0.1
1620 Choose the second device on the first platform.
1621 -init_hw_device opencl:,device_name=Foo9000
1623 Choose the device with a name containing the string
1624 Foo9000.
1625 -init_hw_device
1627 opencl:1,device_type=gpu,device_extensions=cl_khr_fp16
1628 Choose the GPU device on the second platform supporting the
1629 cl_khr_fp16 extension.
1630 vulkan
1632 If device is an integer, it selects the device by its index in
1633 a system-dependent list of devices. If device is any other
1634 string, it selects the first device with a name containing that
1635 string as a substring.
1636 The following options are recognized:
1638 debug
1640 If set to 1, enables the validation layer, if installed.
1641 linear_images
1643 If set to 1, images allocated by the hwcontext will be
1644 linear and locally mappable.
1645 instance_extensions
1647 A plus separated list of additional instance extensions to
1648 enable.
1649 device_extensions
1651 A plus separated list of additional device extensions to
1652 enable.
1653 Examples:
1655 -init_hw_device vulkan:1
1657 Choose the second device on the system.
1658 -init_hw_device vulkan:RADV
1660 Choose the first device with a name containing the string
1661 RADV.
1662 -init_hw_device
1664 vulkan:0,instance_extensions=VK_KHR_wayland_surface+VK_KHR_xcb_surface
1665 Choose the first device and enable the Wayland and XCB
1666 instance extensions.
1667 -init_hw_device type[=name]@source
1669 Initialise a new hardware device of type type called name, deriving
1670 it from the existing device with the name source.
1671 -init_hw_device list
1673 List all hardware device types supported in this build of ffmpeg.
1674 -filter_hw_device name
1676 Pass the hardware device called name to all filters in any filter
1677 graph. This can be used to set the device to upload to with the
1678 "hwupload" filter, or the device to map to with the "hwmap" filter.
1679 Other filters may also make use of this parameter when they require
1680 a hardware device. Note that this is typically only required when
1681 the input is not already in hardware frames - when it is, filters
1682 will derive the device they require from the context of the frames
1683 they receive as input.
1684 This is a global setting, so all filters will receive the same
1686 device.
1687 -hwaccel[:stream_specifier] hwaccel (input,per-stream)
1689 Use hardware acceleration to decode the matching stream(s). The
1690 allowed values of hwaccel are:
1691 none
1693 Do not use any hardware acceleration (the default).
1694 auto
1696 Automatically select the hardware acceleration method.
1697 vdpau
1699 Use VDPAU (Video Decode and Presentation API for Unix) hardware
1700 acceleration.
1701 dxva2
1703 Use DXVA2 (DirectX Video Acceleration) hardware acceleration.
1704 d3d11va
1706 Use D3D11VA (DirectX Video Acceleration) hardware acceleration.
1707 vaapi
1709 Use VAAPI (Video Acceleration API) hardware acceleration.
1710 qsv Use the Intel QuickSync Video acceleration for video
1712 transcoding.
1713 Unlike most other values, this option does not enable
1715 accelerated decoding (that is used automatically whenever a qsv
1716 decoder is selected), but accelerated transcoding, without
1717 copying the frames into the system memory.
1718 For it to work, both the decoder and the encoder must support
1720 QSV acceleration and no filters must be used.
1721 This option has no effect if the selected hwaccel is not available
1723 or not supported by the chosen decoder.
1724 Note that most acceleration methods are intended for playback and
1726 will not be faster than software decoding on modern CPUs.
1727 Additionally, ffmpeg will usually need to copy the decoded frames
1728 from the GPU memory into the system memory, resulting in further
1729 performance loss. This option is thus mainly useful for testing.
1730 -hwaccel_device[:stream_specifier] hwaccel_device (input,per-stream)
1732 Select a device to use for hardware acceleration.
1733 This option only makes sense when the -hwaccel option is also
1735 specified. It can either refer to an existing device created with
1736 -init_hw_device by name, or it can create a new device as if
1737 -init_hw_device type:hwaccel_device were called immediately before.
1738 -hwaccels
1740 List all hardware acceleration components enabled in this build of
1741 ffmpeg. Actual runtime availability depends on the hardware and
1742 its suitable driver being installed.
1743 -fix_sub_duration_heartbeat[:stream_specifier]
1745 Set a specific output video stream as the heartbeat stream
1746 according to which to split and push through currently in-progress
1747 subtitle upon receipt of a random access packet.
1748 This lowers the latency of subtitles for which the end packet or
1750 the following subtitle has not yet been received. As a drawback,
1751 this will most likely lead to duplication of subtitle events in
1752 order to cover the full duration, so when dealing with use cases
1753 where latency of when the subtitle event is passed on to output is
1754 not relevant this option should not be utilized.
1755 Requires -fix_sub_duration to be set for the relevant input
1757 subtitle stream for this to have any effect, as well as for the
1758 input subtitle stream having to be directly mapped to the same
1759 output in which the heartbeat stream resides.
1760 Audio Options
1762 -aframes number (output)
1763 Set the number of audio frames to output. This is an obsolete alias
1764 for "-frames:a", which you should use instead.
1765 -ar[:stream_specifier] freq (input/output,per-stream)
1767 Set the audio sampling frequency. For output streams it is set by
1768 default to the frequency of the corresponding input stream. For
1769 input streams this option only makes sense for audio grabbing
1770 devices and raw demuxers and is mapped to the corresponding demuxer
1771 options.
1772 -aq q (output)
1774 Set the audio quality (codec-specific, VBR). This is an alias for
1775 -q:a.
1776 -ac[:stream_specifier] channels (input/output,per-stream)
1778 Set the number of audio channels. For output streams it is set by
1779 default to the number of input audio channels. For input streams
1780 this option only makes sense for audio grabbing devices and raw
1781 demuxers and is mapped to the corresponding demuxer options.
1782 -an (input/output)
1784 As an input option, blocks all audio streams of a file from being
1785 filtered or being automatically selected or mapped for any output.
1786 See "-discard" option to disable streams individually.
1787 As an output option, disables audio recording i.e. automatic
1789 selection or mapping of any audio stream. For full manual control
1790 see the "-map" option.
1791 -acodec codec (input/output)
1793 Set the audio codec. This is an alias for "-codec:a".
1794 -sample_fmt[:stream_specifier] sample_fmt (output,per-stream)
1796 Set the audio sample format. Use "-sample_fmts" to get a list of
1797 supported sample formats.
1798 -af filtergraph (output)
1800 Create the filtergraph specified by filtergraph and use it to
1801 filter the stream.
1802 This is an alias for "-filter:a", see the -filter option.
1804 Advanced Audio options
1806 -atag fourcc/tag (output)
1807 Force audio tag/fourcc. This is an alias for "-tag:a".
1808 -absf bitstream_filter
1810 Deprecated, see -bsf
1811 -guess_layout_max channels (input,per-stream)
1813 If some input channel layout is not known, try to guess only if it
1814 corresponds to at most the specified number of channels. For
1815 example, 2 tells to ffmpeg to recognize 1 channel as mono and 2
1816 channels as stereo but not 6 channels as 5.1. The default is to
1817 always try to guess. Use 0 to disable all guessing.
1818 Subtitle options
1820 -scodec codec (input/output)
1821 Set the subtitle codec. This is an alias for "-codec:s".
1822 -sn (input/output)
1824 As an input option, blocks all subtitle streams of a file from
1825 being filtered or being automatically selected or mapped for any
1826 output. See "-discard" option to disable streams individually.
1827 As an output option, disables subtitle recording i.e. automatic
1829 selection or mapping of any subtitle stream. For full manual
1830 control see the "-map" option.
1831 -sbsf bitstream_filter
1833 Deprecated, see -bsf
1834 Advanced Subtitle options
1836 -fix_sub_duration
1837 Fix subtitles durations. For each subtitle, wait for the next
1838 packet in the same stream and adjust the duration of the first to
1839 avoid overlap. This is necessary with some subtitles codecs,
1840 especially DVB subtitles, because the duration in the original
1841 packet is only a rough estimate and the end is actually marked by
1842 an empty subtitle frame. Failing to use this option when necessary
1843 can result in exaggerated durations or muxing failures due to non-
1844 monotonic timestamps.
1845 Note that this option will delay the output of all data until the
1847 next subtitle packet is decoded: it may increase memory consumption
1848 and latency a lot.
1849 -canvas_size size
1851 Set the size of the canvas used to render subtitles.
1852 Advanced options
1854 -map [-]input_file_id[:stream_specifier][?] | [linklabel] (output)
1855 Create one or more streams in the output file. This option has two
1856 forms for specifying the data source(s): the first selects one or
1857 more streams from some input file (specified with "-i"), the second
1858 takes an output from some complex filtergraph (specified with
1859 "-filter_complex" or "-filter_complex_script").
1860 In the first form, an output stream is created for every stream
1862 from the input file with the index input_file_id. If
1863 stream_specifier is given, only those streams that match the
1864 specifier are used (see the Stream specifiers section for the
1865 stream_specifier syntax).
1866 A "-" character before the stream identifier creates a "negative"
1868 mapping. It disables matching streams from already created
1869 mappings.
1870 A trailing "?" after the stream index will allow the map to be
1872 optional: if the map matches no streams the map will be ignored
1873 instead of failing. Note the map will still fail if an invalid
1874 input file index is used; such as if the map refers to a non-
1875 existent input.
1876 An alternative [linklabel] form will map outputs from complex
1878 filter graphs (see the -filter_complex option) to the output file.
1879 linklabel must correspond to a defined output link label in the
1880 graph.
1881 This option may be specified multiple times, each adding more
1883 streams to the output file. Any given input stream may also be
1884 mapped any number of times as a source for different output
1885 streams, e.g. in order to use different encoding options and/or
1886 filters. The streams are created in the output in the same order in
1887 which the "-map" options are given on the commandline.
1888 Using this option disables the default mappings for this output
1890 file.
1891 Examples:
1893 map everything
1895 To map ALL streams from the first input file to output
1896 ffmpeg -i INPUT -map 0 output
1898 select specific stream
1900 If you have two audio streams in the first input file, these
1901 streams are identified by 0:0 and 0:1. You can use "-map" to
1902 select which streams to place in an output file. For example:
1903 ffmpeg -i INPUT -map 0:1 out.wav
1905 will map the second input stream in INPUT to the (single)
1907 output stream in out.wav.
1908 create multiple streams
1910 To select the stream with index 2 from input file a.mov
1911 (specified by the identifier 0:2), and stream with index 6 from
1912 input b.mov (specified by the identifier 1:6), and copy them to
1913 the output file out.mov:
1914 ffmpeg -i a.mov -i b.mov -c copy -map 0:2 -map 1:6 out.mov
1916 create multiple streams 2
1918 To select all video and the third audio stream from an input
1919 file:
1920 ffmpeg -i INPUT -map 0:v -map 0:a:2 OUTPUT
1922 negative map
1924 To map all the streams except the second audio, use negative
1925 mappings
1926 ffmpeg -i INPUT -map 0 -map -0:a:1 OUTPUT
1928 optional map
1930 To map the video and audio streams from the first input, and
1931 using the trailing "?", ignore the audio mapping if no audio
1932 streams exist in the first input:
1933 ffmpeg -i INPUT -map 0:v -map 0:a? OUTPUT
1935 map by language
1937 To pick the English audio stream:
1938 ffmpeg -i INPUT -map 0:m:language:eng OUTPUT
1940 -ignore_unknown
1942 Ignore input streams with unknown type instead of failing if
1943 copying such streams is attempted.
1944 -copy_unknown
1946 Allow input streams with unknown type to be copied instead of
1947 failing if copying such streams is attempted.
1948 -map_channel
1950 [input_file_id.stream_specifier.channel_id|-1][?][:output_file_id.stream_specifier]
1951 This option is deprecated and will be removed. It can be replaced
1952 by the pan filter. In some cases it may be easier to use some
1953 combination of the channelsplit, channelmap, or amerge filters.
1954 Map an audio channel from a given input to an output. If
1956 output_file_id.stream_specifier is not set, the audio channel will
1957 be mapped on all the audio streams.
1958 Using "-1" instead of input_file_id.stream_specifier.channel_id
1960 will map a muted channel.
1961 A trailing "?" will allow the map_channel to be optional: if the
1963 map_channel matches no channel the map_channel will be ignored
1964 instead of failing.
1965 For example, assuming INPUT is a stereo audio file, you can switch
1967 the two audio channels with the following command:
1968 ffmpeg -i INPUT -map_channel 0.0.1 -map_channel 0.0.0 OUTPUT
1970 If you want to mute the first channel and keep the second:
1972 ffmpeg -i INPUT -map_channel -1 -map_channel 0.0.1 OUTPUT
1974 The order of the "-map_channel" option specifies the order of the
1976 channels in the output stream. The output channel layout is guessed
1977 from the number of channels mapped (mono if one "-map_channel",
1978 stereo if two, etc.). Using "-ac" in combination of "-map_channel"
1979 makes the channel gain levels to be updated if input and output
1980 channel layouts don't match (for instance two "-map_channel"
1981 options and "-ac 6").
1982 You can also extract each channel of an input to specific outputs;
1984 the following command extracts two channels of the INPUT audio
1985 stream (file 0, stream 0) to the respective OUTPUT_CH0 and
1986 OUTPUT_CH1 outputs:
1987 ffmpeg -i INPUT -map_channel 0.0.0 OUTPUT_CH0 -map_channel 0.0.1 OUTPUT_CH1
1989 The following example splits the channels of a stereo input into
1991 two separate streams, which are put into the same output file:
1992 ffmpeg -i stereo.wav -map 0:0 -map 0:0 -map_channel 0.0.0:0.0 -map_channel 0.0.1:0.1 -y out.ogg
1994 Note that currently each output stream can only contain channels
1996 from a single input stream; you can't for example use
1997 "-map_channel" to pick multiple input audio channels contained in
1998 different streams (from the same or different files) and merge them
1999 into a single output stream. It is therefore not currently
2000 possible, for example, to turn two separate mono streams into a
2001 single stereo stream. However splitting a stereo stream into two
2002 single channel mono streams is possible.
2003 If you need this feature, a possible workaround is to use the
2005 amerge filter. For example, if you need to merge a media (here
2006 input.mkv) with 2 mono audio streams into one single stereo channel
2007 audio stream (and keep the video stream), you can use the following
2008 command:
2009 ffmpeg -i input.mkv -filter_complex "[0:1] [0:2] amerge" -c:a pcm_s16le -c:v copy output.mkv
2011 To map the first two audio channels from the first input, and using
2013 the trailing "?", ignore the audio channel mapping if the first
2014 input is mono instead of stereo:
2015 ffmpeg -i INPUT -map_channel 0.0.0 -map_channel 0.0.1? OUTPUT
2017 -map_metadata[:metadata_spec_out] infile[:metadata_spec_in]
2019 (output,per-metadata)
2020 Set metadata information of the next output file from infile. Note
2021 that those are file indices (zero-based), not filenames. Optional
2022 metadata_spec_in/out parameters specify, which metadata to copy. A
2023 metadata specifier can have the following forms:
2024 g global metadata, i.e. metadata that applies to the whole file
2026 s[:stream_spec]
2028 per-stream metadata. stream_spec is a stream specifier as
2029 described in the Stream specifiers chapter. In an input
2030 metadata specifier, the first matching stream is copied from.
2031 In an output metadata specifier, all matching streams are
2032 copied to.
2033 c:chapter_index
2035 per-chapter metadata. chapter_index is the zero-based chapter
2036 index.
2037 p:program_index
2039 per-program metadata. program_index is the zero-based program
2040 index.
2041 If metadata specifier is omitted, it defaults to global.
2043 By default, global metadata is copied from the first input file,
2045 per-stream and per-chapter metadata is copied along with
2046 streams/chapters. These default mappings are disabled by creating
2047 any mapping of the relevant type. A negative file index can be used
2048 to create a dummy mapping that just disables automatic copying.
2049 For example to copy metadata from the first stream of the input
2051 file to global metadata of the output file:
2052 ffmpeg -i in.ogg -map_metadata 0:s:0 out.mp3
2054 To do the reverse, i.e. copy global metadata to all audio streams:
2056 ffmpeg -i in.mkv -map_metadata:s:a 0:g out.mkv
2058 Note that simple 0 would work as well in this example, since global
2060 metadata is assumed by default.
2061 -map_chapters input_file_index (output)
2063 Copy chapters from input file with index input_file_index to the
2064 next output file. If no chapter mapping is specified, then chapters
2065 are copied from the first input file with at least one chapter. Use
2066 a negative file index to disable any chapter copying.
2067 -benchmark (global)
2069 Show benchmarking information at the end of an encode. Shows real,
2070 system and user time used and maximum memory consumption. Maximum
2071 memory consumption is not supported on all systems, it will usually
2072 display as 0 if not supported.
2073 -benchmark_all (global)
2075 Show benchmarking information during the encode. Shows real,
2076 system and user time used in various steps (audio/video
2077 encode/decode).
2078 -timelimit duration (global)
2080 Exit after ffmpeg has been running for duration seconds in CPU user
2081 time.
2082 -dump (global)
2084 Dump each input packet to stderr.
2085 -hex (global)
2087 When dumping packets, also dump the payload.
2088 -readrate speed (input)
2090 Limit input read speed.
2091 Its value is a floating-point positive number which represents the
2093 maximum duration of media, in seconds, that should be ingested in
2094 one second of wallclock time. Default value is zero and represents
2095 no imposed limitation on speed of ingestion. Value 1 represents
2096 real-time speed and is equivalent to "-re".
2097 Mainly used to simulate a capture device or live input stream (e.g.
2099 when reading from a file). Should not be used with a low value
2100 when input is an actual capture device or live stream as it may
2101 cause packet loss.
2102 It is useful for when flow speed of output packets is important,
2104 such as live streaming.
2105 -re (input)
2107 Read input at native frame rate. This is equivalent to setting
2108 "-readrate 1".
2109 -vsync parameter (global)
2111 -fps_mode[:stream_specifier] parameter (output,per-stream)
2112 Set video sync method / framerate mode. vsync is applied to all
2113 output video streams but can be overridden for a stream by setting
2114 fps_mode. vsync is deprecated and will be removed in the future.
2115 For compatibility reasons some of the values for vsync can be
2117 specified as numbers (shown in parentheses in the following table).
2118 passthrough (0)
2120 Each frame is passed with its timestamp from the demuxer to the
2121 muxer.
2122 cfr (1)
2124 Frames will be duplicated and dropped to achieve exactly the
2125 requested constant frame rate.
2126 vfr (2)
2128 Frames are passed through with their timestamp or dropped so as
2129 to prevent 2 frames from having the same timestamp.
2130 drop
2132 As passthrough but destroys all timestamps, making the muxer
2133 generate fresh timestamps based on frame-rate.
2134 auto (-1)
2136 Chooses between cfr and vfr depending on muxer capabilities.
2137 This is the default method.
2138 Note that the timestamps may be further modified by the muxer,
2140 after this. For example, in the case that the format option
2141 avoid_negative_ts is enabled.
2142 With -map you can select from which stream the timestamps should be
2144 taken. You can leave either video or audio unchanged and sync the
2145 remaining stream(s) to the unchanged one.
2146 -frame_drop_threshold parameter
2148 Frame drop threshold, which specifies how much behind video frames
2149 can be before they are dropped. In frame rate units, so 1.0 is one
2150 frame. The default is -1.1. One possible usecase is to avoid
2151 framedrops in case of noisy timestamps or to increase frame drop
2152 precision in case of exact timestamps.
2153 -adrift_threshold time
2155 Set the minimum difference between timestamps and audio data (in
2156 seconds) to trigger adding/dropping samples to make it match the
2157 timestamps. This option effectively is a threshold to select
2158 between hard (add/drop) and soft (squeeze/stretch) compensation.
2159 "-async" must be set to a positive value.
2160 -apad parameters (output,per-stream)
2162 Pad the output audio stream(s). This is the same as applying "-af
2163 apad". Argument is a string of filter parameters composed the same
2164 as with the "apad" filter. "-shortest" must be set for this output
2165 for the option to take effect.
2166 -copyts
2168 Do not process input timestamps, but keep their values without
2169 trying to sanitize them. In particular, do not remove the initial
2170 start time offset value.
2171 Note that, depending on the vsync option or on specific muxer
2173 processing (e.g. in case the format option avoid_negative_ts is
2174 enabled) the output timestamps may mismatch with the input
2175 timestamps even when this option is selected.
2176 -start_at_zero
2178 When used with copyts, shift input timestamps so they start at
2179 zero.
2180 This means that using e.g. "-ss 50" will make output timestamps
2182 start at 50 seconds, regardless of what timestamp the input file
2183 started at.
2184 -copytb mode
2186 Specify how to set the encoder timebase when stream copying. mode
2187 is an integer numeric value, and can assume one of the following
2188 values:
2189 1 Use the demuxer timebase.
2191 The time base is copied to the output encoder from the
2193 corresponding input demuxer. This is sometimes required to
2194 avoid non monotonically increasing timestamps when copying
2195 video streams with variable frame rate.
2196 0 Use the decoder timebase.
2198 The time base is copied to the output encoder from the
2200 corresponding input decoder.
2201 -1 Try to make the choice automatically, in order to generate a
2203 sane output.
2204 Default value is -1.
2206 -enc_time_base[:stream_specifier] timebase (output,per-stream)
2208 Set the encoder timebase. timebase is a floating point number, and
2209 can assume one of the following values:
2210 0 Assign a default value according to the media type.
2212 For video - use 1/framerate, for audio - use 1/samplerate.
2214 -1 Use the input stream timebase when possible.
2216 If an input stream is not available, the default timebase will
2218 be used.
2219 >0 Use the provided number as the timebase.
2221 This field can be provided as a ratio of two integers (e.g.
2223 1:24, 1:48000) or as a floating point number (e.g. 0.04166,
2224 2.0833e-5)
2225 Default value is 0.
2227 -bitexact (input/output)
2229 Enable bitexact mode for (de)muxer and (de/en)coder
2230 -shortest (output)
2232 Finish encoding when the shortest output stream ends.
2233 Note that this option may require buffering frames, which
2235 introduces extra latency. The maximum amount of this latency may be
2236 controlled with the "-shortest_buf_duration" option.
2237 -shortest_buf_duration duration (output)
2239 The "-shortest" option may require buffering potentially large
2240 amounts of data when at least one of the streams is "sparse" (i.e.
2241 has large gaps between frames – this is typically the case for
2242 subtitles).
2243 This option controls the maximum duration of buffered frames in
2245 seconds. Larger values may allow the "-shortest" option to produce
2246 more accurate results, but increase memory use and latency.
2247 The default value is 10 seconds.
2249 -dts_delta_threshold
2251 Timestamp discontinuity delta threshold.
2252 -dts_error_threshold seconds
2254 Timestamp error delta threshold. This threshold use to discard
2255 crazy/damaged timestamps and the default is 30 hours which is
2256 arbitrarily picked and quite conservative.
2257 -muxdelay seconds (output)
2259 Set the maximum demux-decode delay.
2260 -muxpreload seconds (output)
2262 Set the initial demux-decode delay.
2263 -streamid output-stream-index:new-value (output)
2265 Assign a new stream-id value to an output stream. This option
2266 should be specified prior to the output filename to which it
2267 applies. For the situation where multiple output files exist, a
2268 streamid may be reassigned to a different value.
2269 For example, to set the stream 0 PID to 33 and the stream 1 PID to
2271 36 for an output mpegts file:
2272 ffmpeg -i inurl -streamid 0:33 -streamid 1:36 out.ts
2274 -bsf[:stream_specifier] bitstream_filters (output,per-stream)
2276 Set bitstream filters for matching streams. bitstream_filters is a
2277 comma-separated list of bitstream filters. Use the "-bsfs" option
2278 to get the list of bitstream filters.
2279 ffmpeg -i h264.mp4 -c:v copy -bsf:v h264_mp4toannexb -an out.h264
2281 ffmpeg -i file.mov -an -vn -bsf:s mov2textsub -c:s copy -f rawvideo sub.txt
2284 -tag[:stream_specifier] codec_tag (input/output,per-stream)
2286 Force a tag/fourcc for matching streams.
2287 -timecode hh:mm:ssSEPff
2289 Specify Timecode for writing. SEP is ':' for non drop timecode and
2290 ';' (or '.') for drop.
2291 ffmpeg -i input.mpg -timecode 01:02:03.04 -r 30000/1001 -s ntsc output.mpg
2293 -filter_complex filtergraph (global)
2295 Define a complex filtergraph, i.e. one with arbitrary number of
2296 inputs and/or outputs. For simple graphs -- those with one input
2297 and one output of the same type -- see the -filter options.
2298 filtergraph is a description of the filtergraph, as described in
2299 the ``Filtergraph syntax'' section of the ffmpeg-filters manual.
2300 Input link labels must refer to input streams using the
2302 "[file_index:stream_specifier]" syntax (i.e. the same as -map
2303 uses). If stream_specifier matches multiple streams, the first one
2304 will be used. An unlabeled input will be connected to the first
2305 unused input stream of the matching type.
2306 Output link labels are referred to with -map. Unlabeled outputs are
2308 added to the first output file.
2309 Note that with this option it is possible to use only lavfi sources
2311 without normal input files.
2312 For example, to overlay an image over video
2314 ffmpeg -i video.mkv -i image.png -filter_complex '[0:v][1:v]overlay[out]' -map
2316 '[out]' out.mkv
2317 Here "[0:v]" refers to the first video stream in the first input
2319 file, which is linked to the first (main) input of the overlay
2320 filter. Similarly the first video stream in the second input is
2321 linked to the second (overlay) input of overlay.
2322 Assuming there is only one video stream in each input file, we can
2324 omit input labels, so the above is equivalent to
2325 ffmpeg -i video.mkv -i image.png -filter_complex 'overlay[out]' -map
2327 '[out]' out.mkv
2328 Furthermore we can omit the output label and the single output from
2330 the filter graph will be added to the output file automatically, so
2331 we can simply write
2332 ffmpeg -i video.mkv -i image.png -filter_complex 'overlay' out.mkv
2334 As a special exception, you can use a bitmap subtitle stream as
2336 input: it will be converted into a video with the same size as the
2337 largest video in the file, or 720x576 if no video is present. Note
2338 that this is an experimental and temporary solution. It will be
2339 removed once libavfilter has proper support for subtitles.
2340 For example, to hardcode subtitles on top of a DVB-T recording
2342 stored in MPEG-TS format, delaying the subtitles by 1 second:
2343 ffmpeg -i input.ts -filter_complex \
2345 '[#0x2ef] setpts=PTS+1/TB [sub] ; [#0x2d0] [sub] overlay' \
2346 -sn -map '#0x2dc' output.mkv
2347 (0x2d0, 0x2dc and 0x2ef are the MPEG-TS PIDs of respectively the
2349 video, audio and subtitles streams; 0:0, 0:3 and 0:7 would have
2350 worked too)
2351 To generate 5 seconds of pure red video using lavfi "color" source:
2353 ffmpeg -filter_complex 'color=c=red' -t 5 out.mkv
2355 -filter_complex_threads nb_threads (global)
2357 Defines how many threads are used to process a filter_complex
2358 graph. Similar to filter_threads but used for "-filter_complex"
2359 graphs only. The default is the number of available CPUs.
2360 -lavfi filtergraph (global)
2362 Define a complex filtergraph, i.e. one with arbitrary number of
2363 inputs and/or outputs. Equivalent to -filter_complex.
2364 -filter_complex_script filename (global)
2366 This option is similar to -filter_complex, the only difference is
2367 that its argument is the name of the file from which a complex
2368 filtergraph description is to be read.
2369 -accurate_seek (input)
2371 This option enables or disables accurate seeking in input files
2372 with the -ss option. It is enabled by default, so seeking is
2373 accurate when transcoding. Use -noaccurate_seek to disable it,
2374 which may be useful e.g. when copying some streams and transcoding
2375 the others.
2376 -seek_timestamp (input)
2378 This option enables or disables seeking by timestamp in input files
2379 with the -ss option. It is disabled by default. If enabled, the
2380 argument to the -ss option is considered an actual timestamp, and
2381 is not offset by the start time of the file. This matters only for
2382 files which do not start from timestamp 0, such as transport
2383 streams.
2384 -thread_queue_size size (input/output)
2386 For input, this option sets the maximum number of queued packets
2387 when reading from the file or device. With low latency / high rate
2388 live streams, packets may be discarded if they are not read in a
2389 timely manner; setting this value can force ffmpeg to use a
2390 separate input thread and read packets as soon as they arrive. By
2391 default ffmpeg only does this if multiple inputs are specified.
2392 For output, this option specified the maximum number of packets
2394 that may be queued to each muxing thread.
2395 -sdp_file file (global)
2397 Print sdp information for an output stream to file. This allows
2398 dumping sdp information when at least one output isn't an rtp
2399 stream. (Requires at least one of the output formats to be rtp).
2400 -discard (input)
2402 Allows discarding specific streams or frames from streams. Any
2403 input stream can be fully discarded, using value "all" whereas
2404 selective discarding of frames from a stream occurs at the demuxer
2405 and is not supported by all demuxers.
2406 none
2408 Discard no frame.
2409 default
2411 Default, which discards no frames.
2412 noref
2414 Discard all non-reference frames.
2415 bidir
2417 Discard all bidirectional frames.
2418 nokey
2420 Discard all frames excepts keyframes.
2421 all Discard all frames.
2423 -abort_on flags (global)
2425 Stop and abort on various conditions. The following flags are
2426 available:
2427 empty_output
2429 No packets were passed to the muxer, the output is empty.
2430 empty_output_stream
2432 No packets were passed to the muxer in some of the output
2433 streams.
2434 -max_error_rate (global)
2436 Set fraction of decoding frame failures across all inputs which
2437 when crossed ffmpeg will return exit code 69. Crossing this
2438 threshold does not terminate processing. Range is a floating-point
2439 number between 0 to 1. Default is 2/3.
2440 -xerror (global)
2442 Stop and exit on error
2443 -max_muxing_queue_size packets (output,per-stream)
2445 When transcoding audio and/or video streams, ffmpeg will not begin
2446 writing into the output until it has one packet for each such
2447 stream. While waiting for that to happen, packets for other streams
2448 are buffered. This option sets the size of this buffer, in packets,
2449 for the matching output stream.
2450 The default value of this option should be high enough for most
2452 uses, so only touch this option if you are sure that you need it.
2453 -muxing_queue_data_threshold bytes (output,per-stream)
2455 This is a minimum threshold until which the muxing queue size is
2456 not taken into account. Defaults to 50 megabytes per stream, and is
2457 based on the overall size of packets passed to the muxer.
2458 -auto_conversion_filters (global)
2460 Enable automatically inserting format conversion filters in all
2461 filter graphs, including those defined by -vf, -af, -filter_complex
2462 and -lavfi. If filter format negotiation requires a conversion, the
2463 initialization of the filters will fail. Conversions can still be
2464 performed by inserting the relevant conversion filter (scale,
2465 aresample) in the graph. On by default, to explicitly disable it
2466 you need to specify "-noauto_conversion_filters".
2467 -bits_per_raw_sample[:stream_specifier] value (output,per-stream)
2469 Declare the number of bits per raw sample in the given output
2470 stream to be value. Note that this option sets the information
2471 provided to the encoder/muxer, it does not change the stream to
2472 conform to this value. Setting values that do not match the stream
2473 properties may result in encoding failures or invalid output files.
2474 -stats_enc_pre[:stream_specifier] path (output,per-stream)
2476 -stats_enc_post[:stream_specifier] path (output,per-stream)
2477 -stats_mux_pre[:stream_specifier] path (output,per-stream)
2478 Write per-frame encoding information about the matching streams
2479 into the file given by path.
2480 -stats_enc_pre writes information about raw video or audio frames
2482 right before they are sent for encoding, while -stats_enc_post
2483 writes information about encoded packets as they are received from
2484 the encoder. -stats_mux_pre writes information about packets just
2485 as they are about to be sent to the muxer. Every frame or packet
2486 produces one line in the specified file. The format of this line is
2487 controlled by -stats_enc_pre_fmt / -stats_enc_post_fmt /
2488 -stats_mux_pre_fmt.
2489 When stats for multiple streams are written into a single file, the
2491 lines corresponding to different streams will be interleaved. The
2492 precise order of this interleaving is not specified and not
2493 guaranteed to remain stable between different invocations of the
2494 program, even with the same options.
2495 -stats_enc_pre_fmt[:stream_specifier] format_spec (output,per-stream)
2497 -stats_enc_post_fmt[:stream_specifier] format_spec (output,per-stream)
2498 -stats_mux_pre_fmt[:stream_specifier] format_spec (output,per-stream)
2499 Specify the format for the lines written with -stats_enc_pre /
2500 -stats_enc_post / -stats_mux_pre.
2501 format_spec is a string that may contain directives of the form
2503 {fmt}. format_spec is backslash-escaped --- use \{, \}, and \\ to
2504 write a literal {, }, or \, respectively, into the output.
2505 The directives given with fmt may be one of the following:
2507 fidx
2509 Index of the output file.
2510 sidx
2512 Index of the output stream in the file.
2513 n Frame number. Pre-encoding: number of frames sent to the
2515 encoder so far. Post-encoding: number of packets received from
2516 the encoder so far. Muxing: number of packets submitted to the
2517 muxer for this stream so far.
2518 ni Input frame number. Index of the input frame (i.e. output by a
2520 decoder) that corresponds to this output frame or packet. -1 if
2521 unavailable.
2522 tb Encoder timebase, as a rational number num/den. Note that this
2524 may be different from the timebase used by the muxer.
2525 tbi Timebase for ptsi, as a rational number num/den. Available when
2527 ptsi is available, 0/1 otherwise.
2528 pts Presentation timestamp of the frame or packet, as an integer.
2530 Should be multiplied by the timebase to compute presentation
2531 time.
2532 ptsi
2534 Presentation timestamp of the input frame (see ni), as an
2535 integer. Should be multiplied by tbi to compute presentation
2536 time. Printed as (2^63 - 1 = 9223372036854775807) when not
2537 available.
2538 t Presentation time of the frame or packet, as a decimal number.
2540 Equal to pts multiplied by tb.
2541 ti Presentation time of the input frame (see ni), as a decimal
2543 number. Equal to ptsi multiplied by tbi. Printed as inf when
2544 not available.
2545 dts Decoding timestamp of the packet, as an integer. Should be
2547 multiplied by the timebase to compute presentation time. Post-
2548 encoding only.
2549 dt Decoding time of the frame or packet, as a decimal number.
2551 Equal to dts multiplied by tb.
2552 sn Number of audio samples sent to the encoder so far. Audio and
2554 pre-encoding only.
2555 samp
2557 Number of audio samples in the frame. Audio and pre-encoding
2558 only.
2559 size
2561 Size of the encoded packet in bytes. Post-encoding only.
2562 br Current bitrate in bits per second. Post-encoding only.
2564 abr Average bitrate for the whole stream so far, in bits per
2566 second, -1 if it cannot be determined at this point. Post-
2567 encoding only.
2568 The default format strings are:
2570 pre-encoding
2572 {fidx} {sidx} {n} {t}
2573 post-encoding
2575 {fidx} {sidx} {n} {t}
2576 In the future, new items may be added to the end of the default
2578 formatting strings. Users who depend on the format staying exactly
2579 the same, should prescribe it manually.
2580 Note that stats for different streams written into the same file
2582 may have different formats.
2583 Preset files
2585 A preset file contains a sequence of option=value pairs, one for each
2586 line, specifying a sequence of options which would be awkward to
2587 specify on the command line. Lines starting with the hash ('#')
2588 character are ignored and are used to provide comments. Check the
2589 presets directory in the FFmpeg source tree for examples.
2590 There are two types of preset files: ffpreset and avpreset files.
2592 ffpreset files
2594 ffpreset files are specified with the "vpre", "apre", "spre", and
2596 "fpre" options. The "fpre" option takes the filename of the preset
2597 instead of a preset name as input and can be used for any kind of
2598 codec. For the "vpre", "apre", and "spre" options, the options
2599 specified in a preset file are applied to the currently selected codec
2600 of the same type as the preset option.
2601 The argument passed to the "vpre", "apre", and "spre" preset options
2603 identifies the preset file to use according to the following rules:
2604 First ffmpeg searches for a file named arg.ffpreset in the directories
2606 $FFMPEG_DATADIR (if set), and $HOME/.ffmpeg, and in the datadir defined
2607 at configuration time (usually PREFIX/share/ffmpeg) or in a ffpresets
2608 folder along the executable on win32, in that order. For example, if
2609 the argument is "libvpx-1080p", it will search for the file
2610 libvpx-1080p.ffpreset.
2611 If no such file is found, then ffmpeg will search for a file named
2613 codec_name-arg.ffpreset in the above-mentioned directories, where
2614 codec_name is the name of the codec to which the preset file options
2615 will be applied. For example, if you select the video codec with
2616 "-vcodec libvpx" and use "-vpre 1080p", then it will search for the
2617 file libvpx-1080p.ffpreset.
2618 avpreset files
2620 avpreset files are specified with the "pre" option. They work similar
2622 to ffpreset files, but they only allow encoder- specific options.
2623 Therefore, an option=value pair specifying an encoder cannot be used.
2624 When the "pre" option is specified, ffmpeg will look for files with the
2626 suffix .avpreset in the directories $AVCONV_DATADIR (if set), and
2627 $HOME/.avconv, and in the datadir defined at configuration time
2628 (usually PREFIX/share/ffmpeg), in that order.
2629 First ffmpeg searches for a file named codec_name-arg.avpreset in the
2631 above-mentioned directories, where codec_name is the name of the codec
2632 to which the preset file options will be applied. For example, if you
2633 select the video codec with "-vcodec libvpx" and use "-pre 1080p", then
2634 it will search for the file libvpx-1080p.avpreset.
2635 If no such file is found, then ffmpeg will search for a file named
2637 arg.avpreset in the same directories.
2638
2640 Video and Audio grabbing
2641 If you specify the input format and device then ffmpeg can grab video
2642 and audio directly.
2643 ffmpeg -f oss -i /dev/dsp -f video4linux2 -i /dev/video0 /tmp/out.mpg
2645 Or with an ALSA audio source (mono input, card id 1) instead of OSS:
2647 ffmpeg -f alsa -ac 1 -i hw:1 -f video4linux2 -i /dev/video0 /tmp/out.mpg
2649 Note that you must activate the right video source and channel before
2651 launching ffmpeg with any TV viewer such as
2652 <http://linux.bytesex.org/xawtv/> by Gerd Knorr. You also have to set
2653 the audio recording levels correctly with a standard mixer.
2654 X11 grabbing
2656 Grab the X11 display with ffmpeg via
2657 ffmpeg -f x11grab -video_size cif -framerate 25 -i :0.0 /tmp/out.mpg
2659 0.0 is display.screen number of your X11 server, same as the DISPLAY
2661 environment variable.
2662 ffmpeg -f x11grab -video_size cif -framerate 25 -i :0.0+10,20 /tmp/out.mpg
2664 0.0 is display.screen number of your X11 server, same as the DISPLAY
2666 environment variable. 10 is the x-offset and 20 the y-offset for the
2667 grabbing.
2668 Video and Audio file format conversion
2670 Any supported file format and protocol can serve as input to ffmpeg:
2671 Examples:
2673 • You can use YUV files as input:
2675 ffmpeg -i /tmp/test%d.Y /tmp/out.mpg
2677 It will use the files:
2679 /tmp/test0.Y, /tmp/test0.U, /tmp/test0.V,
2681 /tmp/test1.Y, /tmp/test1.U, /tmp/test1.V, etc...
2682 The Y files use twice the resolution of the U and V files. They are
2684 raw files, without header. They can be generated by all decent
2685 video decoders. You must specify the size of the image with the -s
2686 option if ffmpeg cannot guess it.
2687 • You can input from a raw YUV420P file:
2689 ffmpeg -i /tmp/test.yuv /tmp/out.avi
2691 test.yuv is a file containing raw YUV planar data. Each frame is
2693 composed of the Y plane followed by the U and V planes at half
2694 vertical and horizontal resolution.
2695 • You can output to a raw YUV420P file:
2697 ffmpeg -i mydivx.avi hugefile.yuv
2699 • You can set several input files and output files:
2701 ffmpeg -i /tmp/a.wav -s 640x480 -i /tmp/a.yuv /tmp/a.mpg
2703 Converts the audio file a.wav and the raw YUV video file a.yuv to
2705 MPEG file a.mpg.
2706 • You can also do audio and video conversions at the same time:
2708 ffmpeg -i /tmp/a.wav -ar 22050 /tmp/a.mp2
2710 Converts a.wav to MPEG audio at 22050 Hz sample rate.
2712 • You can encode to several formats at the same time and define a
2714 mapping from input stream to output streams:
2715 ffmpeg -i /tmp/a.wav -map 0:a -b:a 64k /tmp/a.mp2 -map 0:a -b:a 128k /tmp/b.mp2
2717 Converts a.wav to a.mp2 at 64 kbits and to b.mp2 at 128 kbits.
2719 '-map file:index' specifies which input stream is used for each
2720 output stream, in the order of the definition of output streams.
2721 • You can transcode decrypted VOBs:
2723 ffmpeg -i snatch_1.vob -f avi -c:v mpeg4 -b:v 800k -g 300 -bf 2 -c:a libmp3lame -b:a 128k snatch.avi
2725 This is a typical DVD ripping example; the input is a VOB file, the
2727 output an AVI file with MPEG-4 video and MP3 audio. Note that in
2728 this command we use B-frames so the MPEG-4 stream is DivX5
2729 compatible, and GOP size is 300 which means one intra frame every
2730 10 seconds for 29.97fps input video. Furthermore, the audio stream
2731 is MP3-encoded so you need to enable LAME support by passing
2732 "--enable-libmp3lame" to configure. The mapping is particularly
2733 useful for DVD transcoding to get the desired audio language.
2734 NOTE: To see the supported input formats, use "ffmpeg -demuxers".
2736 • You can extract images from a video, or create a video from many
2738 images:
2739 For extracting images from a video:
2741 ffmpeg -i foo.avi -r 1 -s WxH -f image2 foo-%03d.jpeg
2743 This will extract one video frame per second from the video and
2745 will output them in files named foo-001.jpeg, foo-002.jpeg, etc.
2746 Images will be rescaled to fit the new WxH values.
2747 If you want to extract just a limited number of frames, you can use
2749 the above command in combination with the "-frames:v" or "-t"
2750 option, or in combination with -ss to start extracting from a
2751 certain point in time.
2752 For creating a video from many images:
2754 ffmpeg -f image2 -framerate 12 -i foo-%03d.jpeg -s WxH foo.avi
2756 The syntax "foo-%03d.jpeg" specifies to use a decimal number
2758 composed of three digits padded with zeroes to express the sequence
2759 number. It is the same syntax supported by the C printf function,
2760 but only formats accepting a normal integer are suitable.
2761 When importing an image sequence, -i also supports expanding shell-
2763 like wildcard patterns (globbing) internally, by selecting the
2764 image2-specific "-pattern_type glob" option.
2765 For example, for creating a video from filenames matching the glob
2767 pattern "foo-*.jpeg":
2768 ffmpeg -f image2 -pattern_type glob -framerate 12 -i 'foo-*.jpeg' -s WxH foo.avi
2770 • You can put many streams of the same type in the output:
2772 ffmpeg -i test1.avi -i test2.avi -map 1:1 -map 1:0 -map 0:1 -map 0:0 -c copy -y test12.nut
2774 The resulting output file test12.nut will contain the first four
2776 streams from the input files in reverse order.
2777 • To force CBR video output:
2779 ffmpeg -i myfile.avi -b 4000k -minrate 4000k -maxrate 4000k -bufsize 1835k out.m2v
2781 • The four options lmin, lmax, mblmin and mblmax use 'lambda' units,
2783 but you may use the QP2LAMBDA constant to easily convert from 'q'
2784 units:
2785 ffmpeg -i src.ext -lmax 21*QP2LAMBDA dst.ext
2787
2789 This section documents the syntax and formats employed by the FFmpeg
2790 libraries and tools.
2791 Quoting and escaping
2793 FFmpeg adopts the following quoting and escaping mechanism, unless
2794 explicitly specified. The following rules are applied:
2795 • ' and \ are special characters (respectively used for quoting and
2797 escaping). In addition to them, there might be other special
2798 characters depending on the specific syntax where the escaping and
2799 quoting are employed.
2800 • A special character is escaped by prefixing it with a \.
2802 • All characters enclosed between '' are included literally in the
2804 parsed string. The quote character ' itself cannot be quoted, so
2805 you may need to close the quote and escape it.
2806 • Leading and trailing whitespaces, unless escaped or quoted, are
2808 removed from the parsed string.
2809 Note that you may need to add a second level of escaping when using the
2811 command line or a script, which depends on the syntax of the adopted
2812 shell language.
2813 The function "av_get_token" defined in libavutil/avstring.h can be used
2815 to parse a token quoted or escaped according to the rules defined
2816 above.
2817 The tool tools/ffescape in the FFmpeg source tree can be used to
2819 automatically quote or escape a string in a script.
2820 Examples
2822 • Escape the string "Crime d'Amour" containing the "'" special
2824 character:
2825 Crime d\'Amour
2827 • The string above contains a quote, so the "'" needs to be escaped
2829 when quoting it:
2830 'Crime d'\''Amour'
2832 • Include leading or trailing whitespaces using quoting:
2834 ' this string starts and ends with whitespaces '
2836 • Escaping and quoting can be mixed together:
2838 ' The string '\'string\'' is a string '
2840 • To include a literal \ you can use either escaping or quoting:
2842 'c:\foo' can be written as c:\\foo
2844 Date
2846 The accepted syntax is:
2847 [(YYYY-MM-DD|YYYYMMDD)[T|t| ]]((HH:MM:SS[.m...]]])|(HHMMSS[.m...]]]))[Z]
2849 now
2850 If the value is "now" it takes the current time.
2852 Time is local time unless Z is appended, in which case it is
2854 interpreted as UTC. If the year-month-day part is not specified it
2855 takes the current year-month-day.
2856 Time duration
2858 There are two accepted syntaxes for expressing time duration.
2859 [-][<HH>:]<MM>:<SS>[.<m>...]
2861 HH expresses the number of hours, MM the number of minutes for a
2863 maximum of 2 digits, and SS the number of seconds for a maximum of 2
2864 digits. The m at the end expresses decimal value for SS.
2865 or
2867 [-]<S>+[.<m>...][s|ms|us]
2869 S expresses the number of seconds, with the optional decimal part m.
2871 The optional literal suffixes s, ms or us indicate to interpret the
2872 value as seconds, milliseconds or microseconds, respectively.
2873 In both expressions, the optional - indicates negative duration.
2875 Examples
2877 The following examples are all valid time duration:
2879 55 55 seconds
2881 0.2 0.2 seconds
2883 200ms
2885 200 milliseconds, that's 0.2s
2886 200000us
2888 200000 microseconds, that's 0.2s
2889 12:03:45
2891 12 hours, 03 minutes and 45 seconds
2892 23.189
2894 23.189 seconds
2895 Video size
2897 Specify the size of the sourced video, it may be a string of the form
2898 widthxheight, or the name of a size abbreviation.
2899 The following abbreviations are recognized:
2901 ntsc
2903 720x480
2904 pal 720x576
2906 qntsc
2908 352x240
2909 qpal
2911 352x288
2912 sntsc
2914 640x480
2915 spal
2917 768x576
2918 film
2920 352x240
2921 ntsc-film
2923 352x240
2924 sqcif
2926 128x96
2927 qcif
2929 176x144
2930 cif 352x288
2932 4cif
2934 704x576
2935 16cif
2937 1408x1152
2938 qqvga
2940 160x120
2941 qvga
2943 320x240
2944 vga 640x480
2946 svga
2948 800x600
2949 xga 1024x768
2951 uxga
2953 1600x1200
2954 qxga
2956 2048x1536
2957 sxga
2959 1280x1024
2960 qsxga
2962 2560x2048
2963 hsxga
2965 5120x4096
2966 wvga
2968 852x480
2969 wxga
2971 1366x768
2972 wsxga
2974 1600x1024
2975 wuxga
2977 1920x1200
2978 woxga
2980 2560x1600
2981 wqsxga
2983 3200x2048
2984 wquxga
2986 3840x2400
2987 whsxga
2989 6400x4096
2990 whuxga
2992 7680x4800
2993 cga 320x200
2995 ega 640x350
2997 hd480
2999 852x480
3000 hd720
3002 1280x720
3003 hd1080
3005 1920x1080
3006 2k 2048x1080
3008 2kflat
3010 1998x1080
3011 2kscope
3013 2048x858
3014 4k 4096x2160
3016 4kflat
3018 3996x2160
3019 4kscope
3021 4096x1716
3022 nhd 640x360
3024 hqvga
3026 240x160
3027 wqvga
3029 400x240
3030 fwqvga
3032 432x240
3033 hvga
3035 480x320
3036 qhd 960x540
3038 2kdci
3040 2048x1080
3041 4kdci
3043 4096x2160
3044 uhd2160
3046 3840x2160
3047 uhd4320
3049 7680x4320
3050 Video rate
3052 Specify the frame rate of a video, expressed as the number of frames
3053 generated per second. It has to be a string in the format
3054 frame_rate_num/frame_rate_den, an integer number, a float number or a
3055 valid video frame rate abbreviation.
3056 The following abbreviations are recognized:
3058 ntsc
3060 30000/1001
3061 pal 25/1
3063 qntsc
3065 30000/1001
3066 qpal
3068 25/1
3069 sntsc
3071 30000/1001
3072 spal
3074 25/1
3075 film
3077 24/1
3078 ntsc-film
3080 24000/1001
3081 Ratio
3083 A ratio can be expressed as an expression, or in the form
3084 numerator:denominator.
3085 Note that a ratio with infinite (1/0) or negative value is considered
3087 valid, so you should check on the returned value if you want to exclude
3088 those values.
3089 The undefined value can be expressed using the "0:0" string.
3091 Color
3093 It can be the name of a color as defined below (case insensitive match)
3094 or a "[0x|#]RRGGBB[AA]" sequence, possibly followed by @ and a string
3095 representing the alpha component.
3096 The alpha component may be a string composed by "0x" followed by an
3098 hexadecimal number or a decimal number between 0.0 and 1.0, which
3099 represents the opacity value (0x00 or 0.0 means completely transparent,
3100 0xff or 1.0 completely opaque). If the alpha component is not specified
3101 then 0xff is assumed.
3102 The string random will result in a random color.
3104 The following names of colors are recognized:
3106 AliceBlue
3108 0xF0F8FF
3109 AntiqueWhite
3111 0xFAEBD7
3112 Aqua
3114 0x00FFFF
3115 Aquamarine
3117 0x7FFFD4
3118 Azure
3120 0xF0FFFF
3121 Beige
3123 0xF5F5DC
3124 Bisque
3126 0xFFE4C4
3127 Black
3129 0x000000
3130 BlanchedAlmond
3132 0xFFEBCD
3133 Blue
3135 0x0000FF
3136 BlueViolet
3138 0x8A2BE2
3139 Brown
3141 0xA52A2A
3142 BurlyWood
3144 0xDEB887
3145 CadetBlue
3147 0x5F9EA0
3148 Chartreuse
3150 0x7FFF00
3151 Chocolate
3153 0xD2691E
3154 Coral
3156 0xFF7F50
3157 CornflowerBlue
3159 0x6495ED
3160 Cornsilk
3162 0xFFF8DC
3163 Crimson
3165 0xDC143C
3166 Cyan
3168 0x00FFFF
3169 DarkBlue
3171 0x00008B
3172 DarkCyan
3174 0x008B8B
3175 DarkGoldenRod
3177 0xB8860B
3178 DarkGray
3180 0xA9A9A9
3181 DarkGreen
3183 0x006400
3184 DarkKhaki
3186 0xBDB76B
3187 DarkMagenta
3189 0x8B008B
3190 DarkOliveGreen
3192 0x556B2F
3193 Darkorange
3195 0xFF8C00
3196 DarkOrchid
3198 0x9932CC
3199 DarkRed
3201 0x8B0000
3202 DarkSalmon
3204 0xE9967A
3205 DarkSeaGreen
3207 0x8FBC8F
3208 DarkSlateBlue
3210 0x483D8B
3211 DarkSlateGray
3213 0x2F4F4F
3214 DarkTurquoise
3216 0x00CED1
3217 DarkViolet
3219 0x9400D3
3220 DeepPink
3222 0xFF1493
3223 DeepSkyBlue
3225 0x00BFFF
3226 DimGray
3228 0x696969
3229 DodgerBlue
3231 0x1E90FF
3232 FireBrick
3234 0xB22222
3235 FloralWhite
3237 0xFFFAF0
3238 ForestGreen
3240 0x228B22
3241 Fuchsia
3243 0xFF00FF
3244 Gainsboro
3246 0xDCDCDC
3247 GhostWhite
3249 0xF8F8FF
3250 Gold
3252 0xFFD700
3253 GoldenRod
3255 0xDAA520
3256 Gray
3258 0x808080
3259 Green
3261 0x008000
3262 GreenYellow
3264 0xADFF2F
3265 HoneyDew
3267 0xF0FFF0
3268 HotPink
3270 0xFF69B4
3271 IndianRed
3273 0xCD5C5C
3274 Indigo
3276 0x4B0082
3277 Ivory
3279 0xFFFFF0
3280 Khaki
3282 0xF0E68C
3283 Lavender
3285 0xE6E6FA
3286 LavenderBlush
3288 0xFFF0F5
3289 LawnGreen
3291 0x7CFC00
3292 LemonChiffon
3294 0xFFFACD
3295 LightBlue
3297 0xADD8E6
3298 LightCoral
3300 0xF08080
3301 LightCyan
3303 0xE0FFFF
3304 LightGoldenRodYellow
3306 0xFAFAD2
3307 LightGreen
3309 0x90EE90
3310 LightGrey
3312 0xD3D3D3
3313 LightPink
3315 0xFFB6C1
3316 LightSalmon
3318 0xFFA07A
3319 LightSeaGreen
3321 0x20B2AA
3322 LightSkyBlue
3324 0x87CEFA
3325 LightSlateGray
3327 0x778899
3328 LightSteelBlue
3330 0xB0C4DE
3331 LightYellow
3333 0xFFFFE0
3334 Lime
3336 0x00FF00
3337 LimeGreen
3339 0x32CD32
3340 Linen
3342 0xFAF0E6
3343 Magenta
3345 0xFF00FF
3346 Maroon
3348 0x800000
3349 MediumAquaMarine
3351 0x66CDAA
3352 MediumBlue
3354 0x0000CD
3355 MediumOrchid
3357 0xBA55D3
3358 MediumPurple
3360 0x9370D8
3361 MediumSeaGreen
3363 0x3CB371
3364 MediumSlateBlue
3366 0x7B68EE
3367 MediumSpringGreen
3369 0x00FA9A
3370 MediumTurquoise
3372 0x48D1CC
3373 MediumVioletRed
3375 0xC71585
3376 MidnightBlue
3378 0x191970
3379 MintCream
3381 0xF5FFFA
3382 MistyRose
3384 0xFFE4E1
3385 Moccasin
3387 0xFFE4B5
3388 NavajoWhite
3390 0xFFDEAD
3391 Navy
3393 0x000080
3394 OldLace
3396 0xFDF5E6
3397 Olive
3399 0x808000
3400 OliveDrab
3402 0x6B8E23
3403 Orange
3405 0xFFA500
3406 OrangeRed
3408 0xFF4500
3409 Orchid
3411 0xDA70D6
3412 PaleGoldenRod
3414 0xEEE8AA
3415 PaleGreen
3417 0x98FB98
3418 PaleTurquoise
3420 0xAFEEEE
3421 PaleVioletRed
3423 0xD87093
3424 PapayaWhip
3426 0xFFEFD5
3427 PeachPuff
3429 0xFFDAB9
3430 Peru
3432 0xCD853F
3433 Pink
3435 0xFFC0CB
3436 Plum
3438 0xDDA0DD
3439 PowderBlue
3441 0xB0E0E6
3442 Purple
3444 0x800080
3445 Red 0xFF0000
3447 RosyBrown
3449 0xBC8F8F
3450 RoyalBlue
3452 0x4169E1
3453 SaddleBrown
3455 0x8B4513
3456 Salmon
3458 0xFA8072
3459 SandyBrown
3461 0xF4A460
3462 SeaGreen
3464 0x2E8B57
3465 SeaShell
3467 0xFFF5EE
3468 Sienna
3470 0xA0522D
3471 Silver
3473 0xC0C0C0
3474 SkyBlue
3476 0x87CEEB
3477 SlateBlue
3479 0x6A5ACD
3480 SlateGray
3482 0x708090
3483 Snow
3485 0xFFFAFA
3486 SpringGreen
3488 0x00FF7F
3489 SteelBlue
3491 0x4682B4
3492 Tan 0xD2B48C
3494 Teal
3496 0x008080
3497 Thistle
3499 0xD8BFD8
3500 Tomato
3502 0xFF6347
3503 Turquoise
3505 0x40E0D0
3506 Violet
3508 0xEE82EE
3509 Wheat
3511 0xF5DEB3
3512 White
3514 0xFFFFFF
3515 WhiteSmoke
3517 0xF5F5F5
3518 Yellow
3520 0xFFFF00
3521 YellowGreen
3523 0x9ACD32
3524 Channel Layout
3526 A channel layout specifies the spatial disposition of the channels in a
3527 multi-channel audio stream. To specify a channel layout, FFmpeg makes
3528 use of a special syntax.
3529 Individual channels are identified by an id, as given by the table
3531 below:
3532 FL front left
3534 FR front right
3536 FC front center
3538 LFE low frequency
3540 BL back left
3542 BR back right
3544 FLC front left-of-center
3546 FRC front right-of-center
3548 BC back center
3550 SL side left
3552 SR side right
3554 TC top center
3556 TFL top front left
3558 TFC top front center
3560 TFR top front right
3562 TBL top back left
3564 TBC top back center
3566 TBR top back right
3568 DL downmix left
3570 DR downmix right
3572 WL wide left
3574 WR wide right
3576 SDL surround direct left
3578 SDR surround direct right
3580 LFE2
3582 low frequency 2
3583 Standard channel layout compositions can be specified by using the
3585 following identifiers:
3586 mono
3588 FC
3589 stereo
3591 FL+FR
3592 2.1 FL+FR+LFE
3594 3.0 FL+FR+FC
3596 3.0(back)
3598 FL+FR+BC
3599 4.0 FL+FR+FC+BC
3601 quad
3603 FL+FR+BL+BR
3604 quad(side)
3606 FL+FR+SL+SR
3607 3.1 FL+FR+FC+LFE
3609 5.0 FL+FR+FC+BL+BR
3611 5.0(side)
3613 FL+FR+FC+SL+SR
3614 4.1 FL+FR+FC+LFE+BC
3616 5.1 FL+FR+FC+LFE+BL+BR
3618 5.1(side)
3620 FL+FR+FC+LFE+SL+SR
3621 6.0 FL+FR+FC+BC+SL+SR
3623 6.0(front)
3625 FL+FR+FLC+FRC+SL+SR
3626 hexagonal
3628 FL+FR+FC+BL+BR+BC
3629 6.1 FL+FR+FC+LFE+BC+SL+SR
3631 6.1 FL+FR+FC+LFE+BL+BR+BC
3633 6.1(front)
3635 FL+FR+LFE+FLC+FRC+SL+SR
3636 7.0 FL+FR+FC+BL+BR+SL+SR
3638 7.0(front)
3640 FL+FR+FC+FLC+FRC+SL+SR
3641 7.1 FL+FR+FC+LFE+BL+BR+SL+SR
3643 7.1(wide)
3645 FL+FR+FC+LFE+BL+BR+FLC+FRC
3646 7.1(wide-side)
3648 FL+FR+FC+LFE+FLC+FRC+SL+SR
3649 7.1(top)
3651 FL+FR+FC+LFE+BL+BR+TFL+TFR
3652 octagonal
3654 FL+FR+FC+BL+BR+BC+SL+SR
3655 cube
3657 FL+FR+BL+BR+TFL+TFR+TBL+TBR
3658 hexadecagonal
3660 FL+FR+FC+BL+BR+BC+SL+SR+WL+WR+TBL+TBR+TBC+TFC+TFL+TFR
3661 downmix
3663 DL+DR
3664 22.2
3666 FL+FR+FC+LFE+BL+BR+FLC+FRC+BC+SL+SR+TC+TFL+TFC+TFR+TBL+TBC+TBR+LFE2+TSL+TSR+BFC+BFL+BFR
3667 A custom channel layout can be specified as a sequence of terms,
3669 separated by '+'. Each term can be:
3670 • the name of a single channel (e.g. FL, FR, FC, LFE, etc.), each
3672 optionally containing a custom name after a '@', (e.g. FL@Left,
3673 FR@Right, FC@Center, LFE@Low_Frequency, etc.)
3674 A standard channel layout can be specified by the following:
3676 • the name of a single channel (e.g. FL, FR, FC, LFE, etc.)
3678 • the name of a standard channel layout (e.g. mono, stereo, 4.0,
3680 quad, 5.0, etc.)
3681 • a number of channels, in decimal, followed by 'c', yielding the
3683 default channel layout for that number of channels (see the
3684 function "av_channel_layout_default"). Note that not all channel
3685 counts have a default layout.
3686 • a number of channels, in decimal, followed by 'C', yielding an
3688 unknown channel layout with the specified number of channels. Note
3689 that not all channel layout specification strings support unknown
3690 channel layouts.
3691 • a channel layout mask, in hexadecimal starting with "0x" (see the
3693 "AV_CH_*" macros in libavutil/channel_layout.h.
3694 Before libavutil version 53 the trailing character "c" to specify a
3696 number of channels was optional, but now it is required, while a
3697 channel layout mask can also be specified as a decimal number (if and
3698 only if not followed by "c" or "C").
3699 See also the function "av_channel_layout_from_string" defined in
3701 libavutil/channel_layout.h.
3702
3704 When evaluating an arithmetic expression, FFmpeg uses an internal
3705 formula evaluator, implemented through the libavutil/eval.h interface.
3706 An expression may contain unary, binary operators, constants, and
3708 functions.
3709 Two expressions expr1 and expr2 can be combined to form another
3711 expression "expr1;expr2". expr1 and expr2 are evaluated in turn, and
3712 the new expression evaluates to the value of expr2.
3713 The following binary operators are available: "+", "-", "*", "/", "^".
3715 The following unary operators are available: "+", "-".
3717 The following functions are available:
3719 abs(x)
3721 Compute absolute value of x.
3722 acos(x)
3724 Compute arccosine of x.
3725 asin(x)
3727 Compute arcsine of x.
3728 atan(x)
3730 Compute arctangent of x.
3731 atan2(x, y)
3733 Compute principal value of the arc tangent of y/x.
3734 between(x, min, max)
3736 Return 1 if x is greater than or equal to min and lesser than or
3737 equal to max, 0 otherwise.
3738 bitand(x, y)
3740 bitor(x, y)
3741 Compute bitwise and/or operation on x and y.
3742 The results of the evaluation of x and y are converted to integers
3744 before executing the bitwise operation.
3745 Note that both the conversion to integer and the conversion back to
3747 floating point can lose precision. Beware of unexpected results for
3748 large numbers (usually 2^53 and larger).
3749 ceil(expr)
3751 Round the value of expression expr upwards to the nearest integer.
3752 For example, "ceil(1.5)" is "2.0".
3753 clip(x, min, max)
3755 Return the value of x clipped between min and max.
3756 cos(x)
3758 Compute cosine of x.
3759 cosh(x)
3761 Compute hyperbolic cosine of x.
3762 eq(x, y)
3764 Return 1 if x and y are equivalent, 0 otherwise.
3765 exp(x)
3767 Compute exponential of x (with base "e", the Euler's number).
3768 floor(expr)
3770 Round the value of expression expr downwards to the nearest
3771 integer. For example, "floor(-1.5)" is "-2.0".
3772 gauss(x)
3774 Compute Gauss function of x, corresponding to "exp(-x*x/2) /
3775 sqrt(2*PI)".
3776 gcd(x, y)
3778 Return the greatest common divisor of x and y. If both x and y are
3779 0 or either or both are less than zero then behavior is undefined.
3780 gt(x, y)
3782 Return 1 if x is greater than y, 0 otherwise.
3783 gte(x, y)
3785 Return 1 if x is greater than or equal to y, 0 otherwise.
3786 hypot(x, y)
3788 This function is similar to the C function with the same name; it
3789 returns "sqrt(x*x + y*y)", the length of the hypotenuse of a right
3790 triangle with sides of length x and y, or the distance of the point
3791 (x, y) from the origin.
3792 if(x, y)
3794 Evaluate x, and if the result is non-zero return the result of the
3795 evaluation of y, return 0 otherwise.
3796 if(x, y, z)
3798 Evaluate x, and if the result is non-zero return the evaluation
3799 result of y, otherwise the evaluation result of z.
3800 ifnot(x, y)
3802 Evaluate x, and if the result is zero return the result of the
3803 evaluation of y, return 0 otherwise.
3804 ifnot(x, y, z)
3806 Evaluate x, and if the result is zero return the evaluation result
3807 of y, otherwise the evaluation result of z.
3808 isinf(x)
3810 Return 1.0 if x is +/-INFINITY, 0.0 otherwise.
3811 isnan(x)
3813 Return 1.0 if x is NAN, 0.0 otherwise.
3814 ld(var)
3816 Load the value of the internal variable with number var, which was
3817 previously stored with st(var, expr). The function returns the
3818 loaded value.
3819 lerp(x, y, z)
3821 Return linear interpolation between x and y by amount of z.
3822 log(x)
3824 Compute natural logarithm of x.
3825 lt(x, y)
3827 Return 1 if x is lesser than y, 0 otherwise.
3828 lte(x, y)
3830 Return 1 if x is lesser than or equal to y, 0 otherwise.
3831 max(x, y)
3833 Return the maximum between x and y.
3834 min(x, y)
3836 Return the minimum between x and y.
3837 mod(x, y)
3839 Compute the remainder of division of x by y.
3840 not(expr)
3842 Return 1.0 if expr is zero, 0.0 otherwise.
3843 pow(x, y)
3845 Compute the power of x elevated y, it is equivalent to "(x)^(y)".
3846 print(t)
3848 print(t, l)
3849 Print the value of expression t with loglevel l. If l is not
3850 specified then a default log level is used. Returns the value of
3851 the expression printed.
3852 Prints t with loglevel l
3854 random(x)
3856 Return a pseudo random value between 0.0 and 1.0. x is the index of
3857 the internal variable which will be used to save the seed/state.
3858 root(expr, max)
3860 Find an input value for which the function represented by expr with
3861 argument ld(0) is 0 in the interval 0..max.
3862 The expression in expr must denote a continuous function or the
3864 result is undefined.
3865 ld(0) is used to represent the function input value, which means
3867 that the given expression will be evaluated multiple times with
3868 various input values that the expression can access through ld(0).
3869 When the expression evaluates to 0 then the corresponding input
3870 value will be returned.
3871 round(expr)
3873 Round the value of expression expr to the nearest integer. For
3874 example, "round(1.5)" is "2.0".
3875 sgn(x)
3877 Compute sign of x.
3878 sin(x)
3880 Compute sine of x.
3881 sinh(x)
3883 Compute hyperbolic sine of x.
3884 sqrt(expr)
3886 Compute the square root of expr. This is equivalent to "(expr)^.5".
3887 squish(x)
3889 Compute expression "1/(1 + exp(4*x))".
3890 st(var, expr)
3892 Store the value of the expression expr in an internal variable. var
3893 specifies the number of the variable where to store the value, and
3894 it is a value ranging from 0 to 9. The function returns the value
3895 stored in the internal variable. Note, Variables are currently not
3896 shared between expressions.
3897 tan(x)
3899 Compute tangent of x.
3900 tanh(x)
3902 Compute hyperbolic tangent of x.
3903 taylor(expr, x)
3905 taylor(expr, x, id)
3906 Evaluate a Taylor series at x, given an expression representing the
3907 ld(id)-th derivative of a function at 0.
3908 When the series does not converge the result is undefined.
3910 ld(id) is used to represent the derivative order in expr, which
3912 means that the given expression will be evaluated multiple times
3913 with various input values that the expression can access through
3914 ld(id). If id is not specified then 0 is assumed.
3915 Note, when you have the derivatives at y instead of 0,
3917 "taylor(expr, x-y)" can be used.
3918 time(0)
3920 Return the current (wallclock) time in seconds.
3921 trunc(expr)
3923 Round the value of expression expr towards zero to the nearest
3924 integer. For example, "trunc(-1.5)" is "-1.0".
3925 while(cond, expr)
3927 Evaluate expression expr while the expression cond is non-zero, and
3928 returns the value of the last expr evaluation, or NAN if cond was
3929 always false.
3930 The following constants are available:
3932 PI area of the unit disc, approximately 3.14
3934 E exp(1) (Euler's number), approximately 2.718
3936 PHI golden ratio (1+sqrt(5))/2, approximately 1.618
3938 Assuming that an expression is considered "true" if it has a non-zero
3940 value, note that:
3941 "*" works like AND
3943 "+" works like OR
3945 For example the construct:
3947 if (A AND B) then C
3949 is equivalent to:
3951 if(A*B, C)
3953 In your C code, you can extend the list of unary and binary functions,
3955 and define recognized constants, so that they are available for your
3956 expressions.
3957 The evaluator also recognizes the International System unit prefixes.
3959 If 'i' is appended after the prefix, binary prefixes are used, which
3960 are based on powers of 1024 instead of powers of 1000. The 'B' postfix
3961 multiplies the value by 8, and can be appended after a unit prefix or
3962 used alone. This allows using for example 'KB', 'MiB', 'G' and 'B' as
3963 number postfix.
3964 The list of available International System prefixes follows, with
3966 indication of the corresponding powers of 10 and of 2.
3967 y 10^-24 / 2^-80
3969 z 10^-21 / 2^-70
3971 a 10^-18 / 2^-60
3973 f 10^-15 / 2^-50
3975 p 10^-12 / 2^-40
3977 n 10^-9 / 2^-30
3979 u 10^-6 / 2^-20
3981 m 10^-3 / 2^-10
3983 c 10^-2
3985 d 10^-1
3987 h 10^2
3989 k 10^3 / 2^10
3991 K 10^3 / 2^10
3993 M 10^6 / 2^20
3995 G 10^9 / 2^30
3997 T 10^12 / 2^40
3999 P 10^15 / 2^50
4001 E 10^18 / 2^60
4003 Z 10^21 / 2^70
4005 Y 10^24 / 2^80
4007
4009 libavcodec provides some generic global options, which can be set on
4010 all the encoders and decoders. In addition each codec may support so-
4011 called private options, which are specific for a given codec.
4012 Sometimes, a global option may only affect a specific kind of codec,
4014 and may be nonsensical or ignored by another, so you need to be aware
4015 of the meaning of the specified options. Also some options are meant
4016 only for decoding or encoding.
4017 Options may be set by specifying -option value in the FFmpeg tools, or
4019 by setting the value explicitly in the "AVCodecContext" options or
4020 using the libavutil/opt.h API for programmatic use.
4021 The list of supported options follow:
4023 b integer (encoding,audio,video)
4025 Set bitrate in bits/s. Default value is 200K.
4026 ab integer (encoding,audio)
4028 Set audio bitrate (in bits/s). Default value is 128K.
4029 bt integer (encoding,video)
4031 Set video bitrate tolerance (in bits/s). In 1-pass mode, bitrate
4032 tolerance specifies how far ratecontrol is willing to deviate from
4033 the target average bitrate value. This is not related to min/max
4034 bitrate. Lowering tolerance too much has an adverse effect on
4035 quality.
4036 flags flags (decoding/encoding,audio,video,subtitles)
4038 Set generic flags.
4039 Possible values:
4041 mv4 Use four motion vector by macroblock (mpeg4).
4043 qpel
4045 Use 1/4 pel motion compensation.
4046 loop
4048 Use loop filter.
4049 qscale
4051 Use fixed qscale.
4052 pass1
4054 Use internal 2pass ratecontrol in first pass mode.
4055 pass2
4057 Use internal 2pass ratecontrol in second pass mode.
4058 gray
4060 Only decode/encode grayscale.
4061 psnr
4063 Set error[?] variables during encoding.
4064 truncated
4066 Input bitstream might be randomly truncated.
4067 drop_changed
4069 Don't output frames whose parameters differ from first decoded
4070 frame in stream. Error AVERROR_INPUT_CHANGED is returned when
4071 a frame is dropped.
4072 ildct
4074 Use interlaced DCT.
4075 low_delay
4077 Force low delay.
4078 global_header
4080 Place global headers in extradata instead of every keyframe.
4081 bitexact
4083 Only write platform-, build- and time-independent data. (except
4084 (I)DCT). This ensures that file and data checksums are
4085 reproducible and match between platforms. Its primary use is
4086 for regression testing.
4087 aic Apply H263 advanced intra coding / mpeg4 ac prediction.
4089 ilme
4091 Apply interlaced motion estimation.
4092 cgop
4094 Use closed gop.
4095 output_corrupt
4097 Output even potentially corrupted frames.
4098 time_base rational number
4100 Set codec time base.
4101 It is the fundamental unit of time (in seconds) in terms of which
4103 frame timestamps are represented. For fixed-fps content, timebase
4104 should be "1 / frame_rate" and timestamp increments should be
4105 identically 1.
4106 g integer (encoding,video)
4108 Set the group of picture (GOP) size. Default value is 12.
4109 ar integer (decoding/encoding,audio)
4111 Set audio sampling rate (in Hz).
4112 ac integer (decoding/encoding,audio)
4114 Set number of audio channels.
4115 cutoff integer (encoding,audio)
4117 Set cutoff bandwidth. (Supported only by selected encoders, see
4118 their respective documentation sections.)
4119 frame_size integer (encoding,audio)
4121 Set audio frame size.
4122 Each submitted frame except the last must contain exactly
4124 frame_size samples per channel. May be 0 when the codec has
4125 CODEC_CAP_VARIABLE_FRAME_SIZE set, in that case the frame size is
4126 not restricted. It is set by some decoders to indicate constant
4127 frame size.
4128 frame_number integer
4130 Set the frame number.
4131 delay integer
4133 qcomp float (encoding,video)
4134 Set video quantizer scale compression (VBR). It is used as a
4135 constant in the ratecontrol equation. Recommended range for default
4136 rc_eq: 0.0-1.0.
4137 qblur float (encoding,video)
4139 Set video quantizer scale blur (VBR).
4140 qmin integer (encoding,video)
4142 Set min video quantizer scale (VBR). Must be included between -1
4143 and 69, default value is 2.
4144 qmax integer (encoding,video)
4146 Set max video quantizer scale (VBR). Must be included between -1
4147 and 1024, default value is 31.
4148 qdiff integer (encoding,video)
4150 Set max difference between the quantizer scale (VBR).
4151 bf integer (encoding,video)
4153 Set max number of B frames between non-B-frames.
4154 Must be an integer between -1 and 16. 0 means that B-frames are
4156 disabled. If a value of -1 is used, it will choose an automatic
4157 value depending on the encoder.
4158 Default value is 0.
4160 b_qfactor float (encoding,video)
4162 Set qp factor between P and B frames.
4163 codec_tag integer
4165 bug flags (decoding,video)
4166 Workaround not auto detected encoder bugs.
4167 Possible values:
4169 autodetect
4171 xvid_ilace
4172 Xvid interlacing bug (autodetected if fourcc==XVIX)
4173 ump4
4175 (autodetected if fourcc==UMP4)
4176 no_padding
4178 padding bug (autodetected)
4179 amv
4181 qpel_chroma
4182 std_qpel
4183 old standard qpel (autodetected per fourcc/version)
4184 qpel_chroma2
4186 direct_blocksize
4187 direct-qpel-blocksize bug (autodetected per fourcc/version)
4188 edge
4190 edge padding bug (autodetected per fourcc/version)
4191 hpel_chroma
4193 dc_clip
4194 ms Workaround various bugs in microsoft broken decoders.
4195 trunc
4197 trancated frames
4198 strict integer (decoding/encoding,audio,video)
4200 Specify how strictly to follow the standards.
4201 Possible values:
4203 very
4205 strictly conform to an older more strict version of the spec or
4206 reference software
4207 strict
4209 strictly conform to all the things in the spec no matter what
4210 consequences
4211 normal
4213 unofficial
4214 allow unofficial extensions
4215 experimental
4217 allow non standardized experimental things, experimental
4218 (unfinished/work in progress/not well tested) decoders and
4219 encoders. Note: experimental decoders can pose a security
4220 risk, do not use this for decoding untrusted input.
4221 b_qoffset float (encoding,video)
4223 Set QP offset between P and B frames.
4224 err_detect flags (decoding,audio,video)
4226 Set error detection flags.
4227 Possible values:
4229 crccheck
4231 verify embedded CRCs
4232 bitstream
4234 detect bitstream specification deviations
4235 buffer
4237 detect improper bitstream length
4238 explode
4240 abort decoding on minor error detection
4241 ignore_err
4243 ignore decoding errors, and continue decoding. This is useful
4244 if you want to analyze the content of a video and thus want
4245 everything to be decoded no matter what. This option will not
4246 result in a video that is pleasing to watch in case of errors.
4247 careful
4249 consider things that violate the spec and have not been seen in
4250 the wild as errors
4251 compliant
4253 consider all spec non compliancies as errors
4254 aggressive
4256 consider things that a sane encoder should not do as an error
4257 has_b_frames integer
4259 block_align integer
4260 rc_override_count integer
4261 maxrate integer (encoding,audio,video)
4262 Set max bitrate tolerance (in bits/s). Requires bufsize to be set.
4263 minrate integer (encoding,audio,video)
4265 Set min bitrate tolerance (in bits/s). Most useful in setting up a
4266 CBR encode. It is of little use elsewise.
4267 bufsize integer (encoding,audio,video)
4269 Set ratecontrol buffer size (in bits).
4270 i_qfactor float (encoding,video)
4272 Set QP factor between P and I frames.
4273 i_qoffset float (encoding,video)
4275 Set QP offset between P and I frames.
4276 dct integer (encoding,video)
4278 Set DCT algorithm.
4279 Possible values:
4281 auto
4283 autoselect a good one (default)
4284 fastint
4286 fast integer
4287 int accurate integer
4289 mmx
4291 altivec
4292 faan
4293 floating point AAN DCT
4294 lumi_mask float (encoding,video)
4296 Compress bright areas stronger than medium ones.
4297 tcplx_mask float (encoding,video)
4299 Set temporal complexity masking.
4300 scplx_mask float (encoding,video)
4302 Set spatial complexity masking.
4303 p_mask float (encoding,video)
4305 Set inter masking.
4306 dark_mask float (encoding,video)
4308 Compress dark areas stronger than medium ones.
4309 idct integer (decoding/encoding,video)
4311 Select IDCT implementation.
4312 Possible values:
4314 auto
4316 int
4317 simple
4318 simplemmx
4319 simpleauto
4320 Automatically pick a IDCT compatible with the simple one
4321 arm
4323 altivec
4324 sh4
4325 simplearm
4326 simplearmv5te
4327 simplearmv6
4328 simpleneon
4329 xvid
4330 faani
4331 floating point AAN IDCT
4332 slice_count integer
4334 ec flags (decoding,video)
4335 Set error concealment strategy.
4336 Possible values:
4338 guess_mvs
4340 iterative motion vector (MV) search (slow)
4341 deblock
4343 use strong deblock filter for damaged MBs
4344 favor_inter
4346 favor predicting from the previous frame instead of the current
4347 bits_per_coded_sample integer
4349 aspect rational number (encoding,video)
4350 Set sample aspect ratio.
4351 sar rational number (encoding,video)
4353 Set sample aspect ratio. Alias to aspect.
4354 debug flags (decoding/encoding,audio,video,subtitles)
4356 Print specific debug info.
4357 Possible values:
4359 pict
4361 picture info
4362 rc rate control
4364 bitstream
4366 mb_type
4367 macroblock (MB) type
4368 qp per-block quantization parameter (QP)
4370 dct_coeff
4372 green_metadata
4373 display complexity metadata for the upcoming frame, GoP or for
4374 a given duration.
4375 skip
4377 startcode
4378 er error recognition
4379 mmco
4381 memory management control operations (H.264)
4382 bugs
4384 buffers
4385 picture buffer allocations
4386 thread_ops
4388 threading operations
4389 nomc
4391 skip motion compensation
4392 cmp integer (encoding,video)
4394 Set full pel me compare function.
4395 Possible values:
4397 sad sum of absolute differences, fast (default)
4399 sse sum of squared errors
4401 satd
4403 sum of absolute Hadamard transformed differences
4404 dct sum of absolute DCT transformed differences
4406 psnr
4408 sum of squared quantization errors (avoid, low quality)
4409 bit number of bits needed for the block
4411 rd rate distortion optimal, slow
4413 zero
4415 0
4416 vsad
4418 sum of absolute vertical differences
4419 vsse
4421 sum of squared vertical differences
4422 nsse
4424 noise preserving sum of squared differences
4425 w53 5/3 wavelet, only used in snow
4427 w97 9/7 wavelet, only used in snow
4429 dctmax
4431 chroma
4432 subcmp integer (encoding,video)
4433 Set sub pel me compare function.
4434 Possible values:
4436 sad sum of absolute differences, fast (default)
4438 sse sum of squared errors
4440 satd
4442 sum of absolute Hadamard transformed differences
4443 dct sum of absolute DCT transformed differences
4445 psnr
4447 sum of squared quantization errors (avoid, low quality)
4448 bit number of bits needed for the block
4450 rd rate distortion optimal, slow
4452 zero
4454 0
4455 vsad
4457 sum of absolute vertical differences
4458 vsse
4460 sum of squared vertical differences
4461 nsse
4463 noise preserving sum of squared differences
4464 w53 5/3 wavelet, only used in snow
4466 w97 9/7 wavelet, only used in snow
4468 dctmax
4470 chroma
4471 mbcmp integer (encoding,video)
4472 Set macroblock compare function.
4473 Possible values:
4475 sad sum of absolute differences, fast (default)
4477 sse sum of squared errors
4479 satd
4481 sum of absolute Hadamard transformed differences
4482 dct sum of absolute DCT transformed differences
4484 psnr
4486 sum of squared quantization errors (avoid, low quality)
4487 bit number of bits needed for the block
4489 rd rate distortion optimal, slow
4491 zero
4493 0
4494 vsad
4496 sum of absolute vertical differences
4497 vsse
4499 sum of squared vertical differences
4500 nsse
4502 noise preserving sum of squared differences
4503 w53 5/3 wavelet, only used in snow
4505 w97 9/7 wavelet, only used in snow
4507 dctmax
4509 chroma
4510 ildctcmp integer (encoding,video)
4511 Set interlaced dct compare function.
4512 Possible values:
4514 sad sum of absolute differences, fast (default)
4516 sse sum of squared errors
4518 satd
4520 sum of absolute Hadamard transformed differences
4521 dct sum of absolute DCT transformed differences
4523 psnr
4525 sum of squared quantization errors (avoid, low quality)
4526 bit number of bits needed for the block
4528 rd rate distortion optimal, slow
4530 zero
4532 0
4533 vsad
4535 sum of absolute vertical differences
4536 vsse
4538 sum of squared vertical differences
4539 nsse
4541 noise preserving sum of squared differences
4542 w53 5/3 wavelet, only used in snow
4544 w97 9/7 wavelet, only used in snow
4546 dctmax
4548 chroma
4549 dia_size integer (encoding,video)
4550 Set diamond type & size for motion estimation.
4551 (1024, INT_MAX)
4553 full motion estimation(slowest)
4554 (768, 1024]
4556 umh motion estimation
4557 (512, 768]
4559 hex motion estimation
4560 (256, 512]
4562 l2s diamond motion estimation
4563 [2,256]
4565 var diamond motion estimation
4566 (-1, 2)
4568 small diamond motion estimation
4569 -1 funny diamond motion estimation
4571 (INT_MIN, -1)
4573 sab diamond motion estimation
4574 last_pred integer (encoding,video)
4576 Set amount of motion predictors from the previous frame.
4577 precmp integer (encoding,video)
4579 Set pre motion estimation compare function.
4580 Possible values:
4582 sad sum of absolute differences, fast (default)
4584 sse sum of squared errors
4586 satd
4588 sum of absolute Hadamard transformed differences
4589 dct sum of absolute DCT transformed differences
4591 psnr
4593 sum of squared quantization errors (avoid, low quality)
4594 bit number of bits needed for the block
4596 rd rate distortion optimal, slow
4598 zero
4600 0
4601 vsad
4603 sum of absolute vertical differences
4604 vsse
4606 sum of squared vertical differences
4607 nsse
4609 noise preserving sum of squared differences
4610 w53 5/3 wavelet, only used in snow
4612 w97 9/7 wavelet, only used in snow
4614 dctmax
4616 chroma
4617 pre_dia_size integer (encoding,video)
4618 Set diamond type & size for motion estimation pre-pass.
4619 subq integer (encoding,video)
4621 Set sub pel motion estimation quality.
4622 me_range integer (encoding,video)
4624 Set limit motion vectors range (1023 for DivX player).
4625 global_quality integer (encoding,audio,video)
4627 slice_flags integer
4628 mbd integer (encoding,video)
4629 Set macroblock decision algorithm (high quality mode).
4630 Possible values:
4632 simple
4634 use mbcmp (default)
4635 bits
4637 use fewest bits
4638 rd use best rate distortion
4640 rc_init_occupancy integer (encoding,video)
4642 Set number of bits which should be loaded into the rc buffer before
4643 decoding starts.
4644 flags2 flags (decoding/encoding,audio,video,subtitles)
4646 Possible values:
4647 fast
4649 Allow non spec compliant speedup tricks.
4650 noout
4652 Skip bitstream encoding.
4653 ignorecrop
4655 Ignore cropping information from sps.
4656 local_header
4658 Place global headers at every keyframe instead of in extradata.
4659 chunks
4661 Frame data might be split into multiple chunks.
4662 showall
4664 Show all frames before the first keyframe.
4665 export_mvs
4667 Export motion vectors into frame side-data (see
4668 "AV_FRAME_DATA_MOTION_VECTORS") for codecs that support it. See
4669 also doc/examples/export_mvs.c.
4670 skip_manual
4672 Do not skip samples and export skip information as frame side
4673 data.
4674 ass_ro_flush_noop
4676 Do not reset ASS ReadOrder field on flush.
4677 icc_profiles
4679 Generate/parse embedded ICC profiles from/to colorimetry tags.
4680 export_side_data flags (decoding/encoding,audio,video,subtitles)
4682 Possible values:
4683 mvs Export motion vectors into frame side-data (see
4685 "AV_FRAME_DATA_MOTION_VECTORS") for codecs that support it. See
4686 also doc/examples/export_mvs.c.
4687 prft
4689 Export encoder Producer Reference Time into packet side-data
4690 (see "AV_PKT_DATA_PRFT") for codecs that support it.
4691 venc_params
4693 Export video encoding parameters through frame side data (see
4694 "AV_FRAME_DATA_VIDEO_ENC_PARAMS") for codecs that support it.
4695 At present, those are H.264 and VP9.
4696 film_grain
4698 Export film grain parameters through frame side data (see
4699 "AV_FRAME_DATA_FILM_GRAIN_PARAMS"). Supported at present by
4700 AV1 decoders.
4701 threads integer (decoding/encoding,video)
4703 Set the number of threads to be used, in case the selected codec
4704 implementation supports multi-threading.
4705 Possible values:
4707 auto, 0
4709 automatically select the number of threads to set
4710 Default value is auto.
4712 dc integer (encoding,video)
4714 Set intra_dc_precision.
4715 nssew integer (encoding,video)
4717 Set nsse weight.
4718 skip_top integer (decoding,video)
4720 Set number of macroblock rows at the top which are skipped.
4721 skip_bottom integer (decoding,video)
4723 Set number of macroblock rows at the bottom which are skipped.
4724 profile integer (encoding,audio,video)
4726 Set encoder codec profile. Default value is unknown. Encoder
4727 specific profiles are documented in the relevant encoder
4728 documentation.
4729 level integer (encoding,audio,video)
4731 Possible values:
4732 unknown
4734 lowres integer (decoding,audio,video)
4735 Decode at 1= 1/2, 2=1/4, 3=1/8 resolutions.
4736 mblmin integer (encoding,video)
4738 Set min macroblock lagrange factor (VBR).
4739 mblmax integer (encoding,video)
4741 Set max macroblock lagrange factor (VBR).
4742 skip_loop_filter integer (decoding,video)
4744 skip_idct integer (decoding,video)
4745 skip_frame integer (decoding,video)
4746 Make decoder discard processing depending on the frame type
4747 selected by the option value.
4748 skip_loop_filter skips frame loop filtering, skip_idct skips frame
4750 IDCT/dequantization, skip_frame skips decoding.
4751 Possible values:
4753 none
4755 Discard no frame.
4756 default
4758 Discard useless frames like 0-sized frames.
4759 noref
4761 Discard all non-reference frames.
4762 bidir
4764 Discard all bidirectional frames.
4765 nokey
4767 Discard all frames excepts keyframes.
4768 nointra
4770 Discard all frames except I frames.
4771 all Discard all frames.
4773 Default value is default.
4775 bidir_refine integer (encoding,video)
4777 Refine the two motion vectors used in bidirectional macroblocks.
4778 keyint_min integer (encoding,video)
4780 Set minimum interval between IDR-frames.
4781 refs integer (encoding,video)
4783 Set reference frames to consider for motion compensation.
4784 trellis integer (encoding,audio,video)
4786 Set rate-distortion optimal quantization.
4787 mv0_threshold integer (encoding,video)
4789 compression_level integer (encoding,audio,video)
4790 bits_per_raw_sample integer
4791 channel_layout integer (decoding/encoding,audio)
4792 Possible values:
4793 request_channel_layout integer (decoding,audio)
4795 Possible values:
4796 rc_max_vbv_use float (encoding,video)
4798 rc_min_vbv_use float (encoding,video)
4799 ticks_per_frame integer (decoding/encoding,audio,video)
4800 color_primaries integer (decoding/encoding,video)
4801 Possible values:
4802 bt709
4804 BT.709
4805 bt470m
4807 BT.470 M
4808 bt470bg
4810 BT.470 BG
4811 smpte170m
4813 SMPTE 170 M
4814 smpte240m
4816 SMPTE 240 M
4817 film
4819 Film
4820 bt2020
4822 BT.2020
4823 smpte428
4825 smpte428_1
4826 SMPTE ST 428-1
4827 smpte431
4829 SMPTE 431-2
4830 smpte432
4832 SMPTE 432-1
4833 jedec-p22
4835 JEDEC P22
4836 color_trc integer (decoding/encoding,video)
4838 Possible values:
4839 bt709
4841 BT.709
4842 gamma22
4844 BT.470 M
4845 gamma28
4847 BT.470 BG
4848 smpte170m
4850 SMPTE 170 M
4851 smpte240m
4853 SMPTE 240 M
4854 linear
4856 Linear
4857 log
4859 log100
4860 Log
4861 log_sqrt
4863 log316
4864 Log square root
4865 iec61966_2_4
4867 iec61966-2-4
4868 IEC 61966-2-4
4869 bt1361
4871 bt1361e
4872 BT.1361
4873 iec61966_2_1
4875 iec61966-2-1
4876 IEC 61966-2-1
4877 bt2020_10
4879 bt2020_10bit
4880 BT.2020 - 10 bit
4881 bt2020_12
4883 bt2020_12bit
4884 BT.2020 - 12 bit
4885 smpte2084
4887 SMPTE ST 2084
4888 smpte428
4890 smpte428_1
4891 SMPTE ST 428-1
4892 arib-std-b67
4894 ARIB STD-B67
4895 colorspace integer (decoding/encoding,video)
4897 Possible values:
4898 rgb RGB
4900 bt709
4902 BT.709
4903 fcc FCC
4905 bt470bg
4907 BT.470 BG
4908 smpte170m
4910 SMPTE 170 M
4911 smpte240m
4913 SMPTE 240 M
4914 ycocg
4916 YCOCG
4917 bt2020nc
4919 bt2020_ncl
4920 BT.2020 NCL
4921 bt2020c
4923 bt2020_cl
4924 BT.2020 CL
4925 smpte2085
4927 SMPTE 2085
4928 chroma-derived-nc
4930 Chroma-derived NCL
4931 chroma-derived-c
4933 Chroma-derived CL
4934 ictcp
4936 ICtCp
4937 color_range integer (decoding/encoding,video)
4939 If used as input parameter, it serves as a hint to the decoder,
4940 which color_range the input has. Possible values:
4941 tv
4943 mpeg
4944 MPEG (219*2^(n-8))
4945 pc
4947 jpeg
4948 JPEG (2^n-1)
4949 chroma_sample_location integer (decoding/encoding,video)
4951 Possible values:
4952 left
4954 center
4955 topleft
4956 top
4957 bottomleft
4958 bottom
4959 log_level_offset integer
4960 Set the log level offset.
4961 slices integer (encoding,video)
4963 Number of slices, used in parallelized encoding.
4964 thread_type flags (decoding/encoding,video)
4966 Select which multithreading methods to use.
4967 Use of frame will increase decoding delay by one frame per thread,
4969 so clients which cannot provide future frames should not use it.
4970 Possible values:
4972 slice
4974 Decode more than one part of a single frame at once.
4975 Multithreading using slices works only when the video was
4977 encoded with slices.
4978 frame
4980 Decode more than one frame at once.
4981 Default value is slice+frame.
4983 audio_service_type integer (encoding,audio)
4985 Set audio service type.
4986 Possible values:
4988 ma Main Audio Service
4990 ef Effects
4992 vi Visually Impaired
4994 hi Hearing Impaired
4996 di Dialogue
4998 co Commentary
5000 em Emergency
5002 vo Voice Over
5004 ka Karaoke
5006 request_sample_fmt sample_fmt (decoding,audio)
5008 Set sample format audio decoders should prefer. Default value is
5009 "none".
5010 pkt_timebase rational number
5012 sub_charenc encoding (decoding,subtitles)
5013 Set the input subtitles character encoding.
5014 field_order field_order (video)
5016 Set/override the field order of the video. Possible values:
5017 progressive
5019 Progressive video
5020 tt Interlaced video, top field coded and displayed first
5022 bb Interlaced video, bottom field coded and displayed first
5024 tb Interlaced video, top coded first, bottom displayed first
5026 bt Interlaced video, bottom coded first, top displayed first
5028 skip_alpha bool (decoding,video)
5030 Set to 1 to disable processing alpha (transparency). This works
5031 like the gray flag in the flags option which skips chroma
5032 information instead of alpha. Default is 0.
5033 codec_whitelist list (input)
5035 "," separated list of allowed decoders. By default all are allowed.
5036 dump_separator string (input)
5038 Separator used to separate the fields printed on the command line
5039 about the Stream parameters. For example, to separate the fields
5040 with newlines and indentation:
5041 ffprobe -dump_separator "
5043 " -i ~/videos/matrixbench_mpeg2.mpg
5044 max_pixels integer (decoding/encoding,video)
5046 Maximum number of pixels per image. This value can be used to avoid
5047 out of memory failures due to large images.
5048 apply_cropping bool (decoding,video)
5050 Enable cropping if cropping parameters are multiples of the
5051 required alignment for the left and top parameters. If the
5052 alignment is not met the cropping will be partially applied to
5053 maintain alignment. Default is 1 (enabled). Note: The required
5054 alignment depends on if "AV_CODEC_FLAG_UNALIGNED" is set and the
5055 CPU. "AV_CODEC_FLAG_UNALIGNED" cannot be changed from the command
5056 line. Also hardware decoders will not apply left/top Cropping.
5057
5059 Decoders are configured elements in FFmpeg which allow the decoding of
5060 multimedia streams.
5061 When you configure your FFmpeg build, all the supported native decoders
5063 are enabled by default. Decoders requiring an external library must be
5064 enabled manually via the corresponding "--enable-lib" option. You can
5065 list all available decoders using the configure option
5066 "--list-decoders".
5067 You can disable all the decoders with the configure option
5069 "--disable-decoders" and selectively enable / disable single decoders
5070 with the options "--enable-decoder=DECODER" /
5071 "--disable-decoder=DECODER".
5072 The option "-decoders" of the ff* tools will display the list of
5074 enabled decoders.
5075
5077 A description of some of the currently available video decoders
5078 follows.
5079 av1
5081 AOMedia Video 1 (AV1) decoder.
5082 Options
5084 operating_point
5086 Select an operating point of a scalable AV1 bitstream (0 - 31).
5087 Default is 0.
5088 rawvideo
5090 Raw video decoder.
5091 This decoder decodes rawvideo streams.
5093 Options
5095 top top_field_first
5097 Specify the assumed field type of the input video.
5098 -1 the video is assumed to be progressive (default)
5100 0 bottom-field-first is assumed
5102 1 top-field-first is assumed
5104 libdav1d
5106 dav1d AV1 decoder.
5107 libdav1d allows libavcodec to decode the AOMedia Video 1 (AV1) codec.
5109 Requires the presence of the libdav1d headers and library during
5110 configuration. You need to explicitly configure the build with
5111 "--enable-libdav1d".
5112 Options
5114 The following options are supported by the libdav1d wrapper.
5116 framethreads
5118 Set amount of frame threads to use during decoding. The default
5119 value is 0 (autodetect). This option is deprecated for libdav1d >=
5120 1.0 and will be removed in the future. Use the option
5121 "max_frame_delay" and the global option "threads" instead.
5122 tilethreads
5124 Set amount of tile threads to use during decoding. The default
5125 value is 0 (autodetect). This option is deprecated for libdav1d >=
5126 1.0 and will be removed in the future. Use the global option
5127 "threads" instead.
5128 max_frame_delay
5130 Set max amount of frames the decoder may buffer internally. The
5131 default value is 0 (autodetect).
5132 filmgrain
5134 Apply film grain to the decoded video if present in the bitstream.
5135 Defaults to the internal default of the library. This option is
5136 deprecated and will be removed in the future. See the global option
5137 "export_side_data" to export Film Grain parameters instead of
5138 applying it.
5139 oppoint
5141 Select an operating point of a scalable AV1 bitstream (0 - 31).
5142 Defaults to the internal default of the library.
5143 alllayers
5145 Output all spatial layers of a scalable AV1 bitstream. The default
5146 value is false.
5147 libdavs2
5149 AVS2-P2/IEEE1857.4 video decoder wrapper.
5150 This decoder allows libavcodec to decode AVS2 streams with davs2
5152 library.
5153 libuavs3d
5155 AVS3-P2/IEEE1857.10 video decoder.
5156 libuavs3d allows libavcodec to decode AVS3 streams. Requires the
5158 presence of the libuavs3d headers and library during configuration.
5159 You need to explicitly configure the build with "--enable-libuavs3d".
5160 Options
5162 The following option is supported by the libuavs3d wrapper.
5164 frame_threads
5166 Set amount of frame threads to use during decoding. The default
5167 value is 0 (autodetect).
5168 QSV Decoders
5170 The family of Intel QuickSync Video decoders (VC1, MPEG-2, H.264, HEVC,
5171 JPEG/MJPEG, VP8, VP9, AV1).
5172 Common Options
5174 The following options are supported by all qsv decoders.
5176 async_depth
5178 Internal parallelization depth, the higher the value the higher the
5179 latency.
5180 gpu_copy
5182 A GPU-accelerated copy between video and system memory
5183 default
5185 on
5186 off
5187 HEVC Options
5189 Extra options for hevc_qsv.
5191 load_plugin
5193 A user plugin to load in an internal session
5194 none
5196 hevc_sw
5197 hevc_hw
5198 load_plugins
5199 A :-separate list of hexadecimal plugin UIDs to load in an internal
5200 session
5201 v210
5203 Uncompressed 4:2:2 10-bit decoder.
5204 Options
5206 custom_stride
5208 Set the line size of the v210 data in bytes. The default value is 0
5209 (autodetect). You can use the special -1 value for a strideless
5210 v210 as seen in BOXX files.
5211
5213 A description of some of the currently available audio decoders
5214 follows.
5215 ac3
5217 AC-3 audio decoder.
5218 This decoder implements part of ATSC A/52:2010 and ETSI TS 102 366, as
5220 well as the undocumented RealAudio 3 (a.k.a. dnet).
5221 AC-3 Decoder Options
5223 -drc_scale value
5225 Dynamic Range Scale Factor. The factor to apply to dynamic range
5226 values from the AC-3 stream. This factor is applied exponentially.
5227 The default value is 1. There are 3 notable scale factor ranges:
5228 drc_scale == 0
5230 DRC disabled. Produces full range audio.
5231 0 < drc_scale <= 1
5233 DRC enabled. Applies a fraction of the stream DRC value.
5234 Audio reproduction is between full range and full compression.
5235 drc_scale > 1
5237 DRC enabled. Applies drc_scale asymmetrically. Loud sounds are
5238 fully compressed. Soft sounds are enhanced.
5239 flac
5241 FLAC audio decoder.
5242 This decoder aims to implement the complete FLAC specification from
5244 Xiph.
5245 FLAC Decoder options
5247 -use_buggy_lpc
5249 The lavc FLAC encoder used to produce buggy streams with high lpc
5250 values (like the default value). This option makes it possible to
5251 decode such streams correctly by using lavc's old buggy lpc logic
5252 for decoding.
5253 ffwavesynth
5255 Internal wave synthesizer.
5256 This decoder generates wave patterns according to predefined sequences.
5258 Its use is purely internal and the format of the data it accepts is not
5259 publicly documented.
5260 libcelt
5262 libcelt decoder wrapper.
5263 libcelt allows libavcodec to decode the Xiph CELT ultra-low delay audio
5265 codec. Requires the presence of the libcelt headers and library during
5266 configuration. You need to explicitly configure the build with
5267 "--enable-libcelt".
5268 libgsm
5270 libgsm decoder wrapper.
5271 libgsm allows libavcodec to decode the GSM full rate audio codec.
5273 Requires the presence of the libgsm headers and library during
5274 configuration. You need to explicitly configure the build with
5275 "--enable-libgsm".
5276 This decoder supports both the ordinary GSM and the Microsoft variant.
5278 libilbc
5280 libilbc decoder wrapper.
5281 libilbc allows libavcodec to decode the Internet Low Bitrate Codec
5283 (iLBC) audio codec. Requires the presence of the libilbc headers and
5284 library during configuration. You need to explicitly configure the
5285 build with "--enable-libilbc".
5286 Options
5288 The following option is supported by the libilbc wrapper.
5290 enhance
5292 Enable the enhancement of the decoded audio when set to 1. The
5293 default value is 0 (disabled).
5294 libopencore-amrnb
5296 libopencore-amrnb decoder wrapper.
5297 libopencore-amrnb allows libavcodec to decode the Adaptive Multi-Rate
5299 Narrowband audio codec. Using it requires the presence of the
5300 libopencore-amrnb headers and library during configuration. You need to
5301 explicitly configure the build with "--enable-libopencore-amrnb".
5302 An FFmpeg native decoder for AMR-NB exists, so users can decode AMR-NB
5304 without this library.
5305 libopencore-amrwb
5307 libopencore-amrwb decoder wrapper.
5308 libopencore-amrwb allows libavcodec to decode the Adaptive Multi-Rate
5310 Wideband audio codec. Using it requires the presence of the
5311 libopencore-amrwb headers and library during configuration. You need to
5312 explicitly configure the build with "--enable-libopencore-amrwb".
5313 An FFmpeg native decoder for AMR-WB exists, so users can decode AMR-WB
5315 without this library.
5316 libopus
5318 libopus decoder wrapper.
5319 libopus allows libavcodec to decode the Opus Interactive Audio Codec.
5321 Requires the presence of the libopus headers and library during
5322 configuration. You need to explicitly configure the build with
5323 "--enable-libopus".
5324 An FFmpeg native decoder for Opus exists, so users can decode Opus
5326 without this library.
5327
5329 libaribb24
5330 ARIB STD-B24 caption decoder.
5331 Implements profiles A and C of the ARIB STD-B24 standard.
5333 libaribb24 Decoder Options
5335 -aribb24-base-path path
5337 Sets the base path for the libaribb24 library. This is utilized for
5338 reading of configuration files (for custom unicode conversions),
5339 and for dumping of non-text symbols as images under that location.
5340 Unset by default.
5342 -aribb24-skip-ruby-text boolean
5344 Tells the decoder wrapper to skip text blocks that contain half-
5345 height ruby text.
5346 Enabled by default.
5348 dvbsub
5350 Options
5351 compute_clut
5353 -2 Compute clut once if no matching CLUT is in the stream.
5354 -1 Compute clut if no matching CLUT is in the stream.
5356 0 Never compute CLUT
5358 1 Always compute CLUT and override the one provided in the
5360 stream.
5361 dvb_substream
5363 Selects the dvb substream, or all substreams if -1 which is
5364 default.
5365 dvdsub
5367 This codec decodes the bitmap subtitles used in DVDs; the same
5368 subtitles can also be found in VobSub file pairs and in some Matroska
5369 files.
5370 Options
5372 palette
5374 Specify the global palette used by the bitmaps. When stored in
5375 VobSub, the palette is normally specified in the index file; in
5376 Matroska, the palette is stored in the codec extra-data in the same
5377 format as in VobSub. In DVDs, the palette is stored in the IFO
5378 file, and therefore not available when reading from dumped VOB
5379 files.
5380 The format for this option is a string containing 16 24-bits
5382 hexadecimal numbers (without 0x prefix) separated by commas, for
5383 example "0d00ee, ee450d, 101010, eaeaea, 0ce60b, ec14ed, ebff0b,
5384 0d617a, 7b7b7b, d1d1d1, 7b2a0e, 0d950c, 0f007b, cf0dec, cfa80c,
5385 7c127b".
5386 ifo_palette
5388 Specify the IFO file from which the global palette is obtained.
5389 (experimental)
5390 forced_subs_only
5392 Only decode subtitle entries marked as forced. Some titles have
5393 forced and non-forced subtitles in the same track. Setting this
5394 flag to 1 will only keep the forced subtitles. Default value is 0.
5395 libzvbi-teletext
5397 Libzvbi allows libavcodec to decode DVB teletext pages and DVB teletext
5398 subtitles. Requires the presence of the libzvbi headers and library
5399 during configuration. You need to explicitly configure the build with
5400 "--enable-libzvbi".
5401 Options
5403 txt_page
5405 List of teletext page numbers to decode. Pages that do not match
5406 the specified list are dropped. You may use the special "*" string
5407 to match all pages, or "subtitle" to match all subtitle pages.
5408 Default value is *.
5409 txt_default_region
5411 Set default character set used for decoding, a value between 0 and
5412 87 (see ETS 300 706, Section 15, Table 32). Default value is -1,
5413 which does not override the libzvbi default. This option is needed
5414 for some legacy level 1.0 transmissions which cannot signal the
5415 proper charset.
5416 txt_chop_top
5418 Discards the top teletext line. Default value is 1.
5419 txt_format
5421 Specifies the format of the decoded subtitles.
5422 bitmap
5424 The default format, you should use this for teletext pages,
5425 because certain graphics and colors cannot be expressed in
5426 simple text or even ASS.
5427 text
5429 Simple text based output without formatting.
5430 ass Formatted ASS output, subtitle pages and teletext pages are
5432 returned in different styles, subtitle pages are stripped down
5433 to text, but an effort is made to keep the text alignment and
5434 the formatting.
5435 txt_left
5437 X offset of generated bitmaps, default is 0.
5438 txt_top
5440 Y offset of generated bitmaps, default is 0.
5441 txt_chop_spaces
5443 Chops leading and trailing spaces and removes empty lines from the
5444 generated text. This option is useful for teletext based subtitles
5445 where empty spaces may be present at the start or at the end of the
5446 lines or empty lines may be present between the subtitle lines
5447 because of double-sized teletext characters. Default value is 1.
5448 txt_duration
5450 Sets the display duration of the decoded teletext pages or
5451 subtitles in milliseconds. Default value is -1 which means infinity
5452 or until the next subtitle event comes.
5453 txt_transparent
5455 Force transparent background of the generated teletext bitmaps.
5456 Default value is 0 which means an opaque background.
5457 txt_opacity
5459 Sets the opacity (0-255) of the teletext background. If
5460 txt_transparent is not set, it only affects characters between a
5461 start box and an end box, typically subtitles. Default value is 0
5462 if txt_transparent is set, 255 otherwise.
5463
5465 Encoders are configured elements in FFmpeg which allow the encoding of
5466 multimedia streams.
5467 When you configure your FFmpeg build, all the supported native encoders
5469 are enabled by default. Encoders requiring an external library must be
5470 enabled manually via the corresponding "--enable-lib" option. You can
5471 list all available encoders using the configure option
5472 "--list-encoders".
5473 You can disable all the encoders with the configure option
5475 "--disable-encoders" and selectively enable / disable single encoders
5476 with the options "--enable-encoder=ENCODER" /
5477 "--disable-encoder=ENCODER".
5478 The option "-encoders" of the ff* tools will display the list of
5480 enabled encoders.
5481
5483 A description of some of the currently available audio encoders
5484 follows.
5485 aac
5487 Advanced Audio Coding (AAC) encoder.
5488 This encoder is the default AAC encoder, natively implemented into
5490 FFmpeg.
5491 Options
5493 b Set bit rate in bits/s. Setting this automatically activates
5495 constant bit rate (CBR) mode. If this option is unspecified it is
5496 set to 128kbps.
5497 q Set quality for variable bit rate (VBR) mode. This option is valid
5499 only using the ffmpeg command-line tool. For library interface
5500 users, use global_quality.
5501 cutoff
5503 Set cutoff frequency. If unspecified will allow the encoder to
5504 dynamically adjust the cutoff to improve clarity on low bitrates.
5505 aac_coder
5507 Set AAC encoder coding method. Possible values:
5508 twoloop
5510 Two loop searching (TLS) method. This is the default method.
5511 This method first sets quantizers depending on band thresholds
5513 and then tries to find an optimal combination by adding or
5514 subtracting a specific value from all quantizers and adjusting
5515 some individual quantizer a little. Will tune itself based on
5516 whether aac_is, aac_ms and aac_pns are enabled.
5517 anmr
5519 Average noise to mask ratio (ANMR) trellis-based solution.
5520 This is an experimental coder which currently produces a lower
5522 quality, is more unstable and is slower than the default
5523 twoloop coder but has potential. Currently has no support for
5524 the aac_is or aac_pns options. Not currently recommended.
5525 fast
5527 Constant quantizer method.
5528 Uses a cheaper version of twoloop algorithm that doesn't try to
5530 do as many clever adjustments. Worse with low bitrates (less
5531 than 64kbps), but is better and much faster at higher bitrates.
5532 aac_ms
5534 Sets mid/side coding mode. The default value of "auto" will
5535 automatically use M/S with bands which will benefit from such
5536 coding. Can be forced for all bands using the value "enable", which
5537 is mainly useful for debugging or disabled using "disable".
5538 aac_is
5540 Sets intensity stereo coding tool usage. By default, it's enabled
5541 and will automatically toggle IS for similar pairs of stereo bands
5542 if it's beneficial. Can be disabled for debugging by setting the
5543 value to "disable".
5544 aac_pns
5546 Uses perceptual noise substitution to replace low entropy high
5547 frequency bands with imperceptible white noise during the decoding
5548 process. By default, it's enabled, but can be disabled for
5549 debugging purposes by using "disable".
5550 aac_tns
5552 Enables the use of a multitap FIR filter which spans through the
5553 high frequency bands to hide quantization noise during the encoding
5554 process and is reverted by the decoder. As well as decreasing
5555 unpleasant artifacts in the high range this also reduces the
5556 entropy in the high bands and allows for more bits to be used by
5557 the mid-low bands. By default it's enabled but can be disabled for
5558 debugging by setting the option to "disable".
5559 aac_ltp
5561 Enables the use of the long term prediction extension which
5562 increases coding efficiency in very low bandwidth situations such
5563 as encoding of voice or solo piano music by extending constant
5564 harmonic peaks in bands throughout frames. This option is implied
5565 by profile:a aac_low and is incompatible with aac_pred. Use in
5566 conjunction with -ar to decrease the samplerate.
5567 aac_pred
5569 Enables the use of a more traditional style of prediction where the
5570 spectral coefficients transmitted are replaced by the difference of
5571 the current coefficients minus the previous "predicted"
5572 coefficients. In theory and sometimes in practice this can improve
5573 quality for low to mid bitrate audio. This option implies the
5574 aac_main profile and is incompatible with aac_ltp.
5575 profile
5577 Sets the encoding profile, possible values:
5578 aac_low
5580 The default, AAC "Low-complexity" profile. Is the most
5581 compatible and produces decent quality.
5582 mpeg2_aac_low
5584 Equivalent to "-profile:a aac_low -aac_pns 0". PNS was
5585 introduced with the MPEG4 specifications.
5586 aac_ltp
5588 Long term prediction profile, is enabled by and will enable the
5589 aac_ltp option. Introduced in MPEG4.
5590 aac_main
5592 Main-type prediction profile, is enabled by and will enable the
5593 aac_pred option. Introduced in MPEG2.
5594 If this option is unspecified it is set to aac_low.
5596 ac3 and ac3_fixed
5598 AC-3 audio encoders.
5599 These encoders implement part of ATSC A/52:2010 and ETSI TS 102 366, as
5601 well as the undocumented RealAudio 3 (a.k.a. dnet).
5602 The ac3 encoder uses floating-point math, while the ac3_fixed encoder
5604 only uses fixed-point integer math. This does not mean that one is
5605 always faster, just that one or the other may be better suited to a
5606 particular system. The ac3_fixed encoder is not the default codec for
5607 any of the output formats, so it must be specified explicitly using the
5608 option "-acodec ac3_fixed" in order to use it.
5609 AC-3 Metadata
5611 The AC-3 metadata options are used to set parameters that describe the
5613 audio, but in most cases do not affect the audio encoding itself. Some
5614 of the options do directly affect or influence the decoding and
5615 playback of the resulting bitstream, while others are just for
5616 informational purposes. A few of the options will add bits to the
5617 output stream that could otherwise be used for audio data, and will
5618 thus affect the quality of the output. Those will be indicated
5619 accordingly with a note in the option list below.
5620 These parameters are described in detail in several publicly-available
5622 documents.
5623 *<<http://www.atsc.org/cms/standards/a_52-2010.pdf>>
5625 *<<http://www.atsc.org/cms/standards/a_54a_with_corr_1.pdf>>
5626 *<<http://www.dolby.com/uploadedFiles/zz-_Shared_Assets/English_PDFs/Professional/18_Metadata.Guide.pdf>>
5627 *<<http://www.dolby.com/uploadedFiles/zz-_Shared_Assets/English_PDFs/Professional/46_DDEncodingGuidelines.pdf>>
5628 Metadata Control Options
5630 -per_frame_metadata boolean
5632 Allow Per-Frame Metadata. Specifies if the encoder should check for
5633 changing metadata for each frame.
5634 0 The metadata values set at initialization will be used for
5636 every frame in the stream. (default)
5637 1 Metadata values can be changed before encoding each frame.
5639 Downmix Levels
5641 -center_mixlev level
5643 Center Mix Level. The amount of gain the decoder should apply to
5644 the center channel when downmixing to stereo. This field will only
5645 be written to the bitstream if a center channel is present. The
5646 value is specified as a scale factor. There are 3 valid values:
5647 0.707
5649 Apply -3dB gain
5650 0.595
5652 Apply -4.5dB gain (default)
5653 0.500
5655 Apply -6dB gain
5656 -surround_mixlev level
5658 Surround Mix Level. The amount of gain the decoder should apply to
5659 the surround channel(s) when downmixing to stereo. This field will
5660 only be written to the bitstream if one or more surround channels
5661 are present. The value is specified as a scale factor. There are 3
5662 valid values:
5663 0.707
5665 Apply -3dB gain
5666 0.500
5668 Apply -6dB gain (default)
5669 0.000
5671 Silence Surround Channel(s)
5672 Audio Production Information
5674 Audio Production Information is optional information describing the
5676 mixing environment. Either none or both of the fields are written to
5677 the bitstream.
5678 -mixing_level number
5680 Mixing Level. Specifies peak sound pressure level (SPL) in the
5681 production environment when the mix was mastered. Valid values are
5682 80 to 111, or -1 for unknown or not indicated. The default value is
5683 -1, but that value cannot be used if the Audio Production
5684 Information is written to the bitstream. Therefore, if the
5685 "room_type" option is not the default value, the "mixing_level"
5686 option must not be -1.
5687 -room_type type
5689 Room Type. Describes the equalization used during the final mixing
5690 session at the studio or on the dubbing stage. A large room is a
5691 dubbing stage with the industry standard X-curve equalization; a
5692 small room has flat equalization. This field will not be written
5693 to the bitstream if both the "mixing_level" option and the
5694 "room_type" option have the default values.
5695 0
5697 notindicated
5698 Not Indicated (default)
5699 1
5701 large
5702 Large Room
5703 2
5705 small
5706 Small Room
5707 Other Metadata Options
5709 -copyright boolean
5711 Copyright Indicator. Specifies whether a copyright exists for this
5712 audio.
5713 0
5715 off No Copyright Exists (default)
5716 1
5718 on Copyright Exists
5719 -dialnorm value
5721 Dialogue Normalization. Indicates how far the average dialogue
5722 level of the program is below digital 100% full scale (0 dBFS).
5723 This parameter determines a level shift during audio reproduction
5724 that sets the average volume of the dialogue to a preset level. The
5725 goal is to match volume level between program sources. A value of
5726 -31dB will result in no volume level change, relative to the source
5727 volume, during audio reproduction. Valid values are whole numbers
5728 in the range -31 to -1, with -31 being the default.
5729 -dsur_mode mode
5731 Dolby Surround Mode. Specifies whether the stereo signal uses Dolby
5732 Surround (Pro Logic). This field will only be written to the
5733 bitstream if the audio stream is stereo. Using this option does NOT
5734 mean the encoder will actually apply Dolby Surround processing.
5735 0
5737 notindicated
5738 Not Indicated (default)
5739 1
5741 off Not Dolby Surround Encoded
5742 2
5744 on Dolby Surround Encoded
5745 -original boolean
5747 Original Bit Stream Indicator. Specifies whether this audio is from
5748 the original source and not a copy.
5749 0
5751 off Not Original Source
5752 1
5754 on Original Source (default)
5755 Extended Bitstream Information
5757 The extended bitstream options are part of the Alternate Bit Stream
5759 Syntax as specified in Annex D of the A/52:2010 standard. It is grouped
5760 into 2 parts. If any one parameter in a group is specified, all values
5761 in that group will be written to the bitstream. Default values are
5762 used for those that are written but have not been specified. If the
5763 mixing levels are written, the decoder will use these values instead of
5764 the ones specified in the "center_mixlev" and "surround_mixlev" options
5765 if it supports the Alternate Bit Stream Syntax.
5766 Extended Bitstream Information - Part 1
5768 -dmix_mode mode
5770 Preferred Stereo Downmix Mode. Allows the user to select either
5771 Lt/Rt (Dolby Surround) or Lo/Ro (normal stereo) as the preferred
5772 stereo downmix mode.
5773 0
5775 notindicated
5776 Not Indicated (default)
5777 1
5779 ltrt
5780 Lt/Rt Downmix Preferred
5781 2
5783 loro
5784 Lo/Ro Downmix Preferred
5785 -ltrt_cmixlev level
5787 Lt/Rt Center Mix Level. The amount of gain the decoder should apply
5788 to the center channel when downmixing to stereo in Lt/Rt mode.
5789 1.414
5791 Apply +3dB gain
5792 1.189
5794 Apply +1.5dB gain
5795 1.000
5797 Apply 0dB gain
5798 0.841
5800 Apply -1.5dB gain
5801 0.707
5803 Apply -3.0dB gain
5804 0.595
5806 Apply -4.5dB gain (default)
5807 0.500
5809 Apply -6.0dB gain
5810 0.000
5812 Silence Center Channel
5813 -ltrt_surmixlev level
5815 Lt/Rt Surround Mix Level. The amount of gain the decoder should
5816 apply to the surround channel(s) when downmixing to stereo in Lt/Rt
5817 mode.
5818 0.841
5820 Apply -1.5dB gain
5821 0.707
5823 Apply -3.0dB gain
5824 0.595
5826 Apply -4.5dB gain
5827 0.500
5829 Apply -6.0dB gain (default)
5830 0.000
5832 Silence Surround Channel(s)
5833 -loro_cmixlev level
5835 Lo/Ro Center Mix Level. The amount of gain the decoder should apply
5836 to the center channel when downmixing to stereo in Lo/Ro mode.
5837 1.414
5839 Apply +3dB gain
5840 1.189
5842 Apply +1.5dB gain
5843 1.000
5845 Apply 0dB gain
5846 0.841
5848 Apply -1.5dB gain
5849 0.707
5851 Apply -3.0dB gain
5852 0.595
5854 Apply -4.5dB gain (default)
5855 0.500
5857 Apply -6.0dB gain
5858 0.000
5860 Silence Center Channel
5861 -loro_surmixlev level
5863 Lo/Ro Surround Mix Level. The amount of gain the decoder should
5864 apply to the surround channel(s) when downmixing to stereo in Lo/Ro
5865 mode.
5866 0.841
5868 Apply -1.5dB gain
5869 0.707
5871 Apply -3.0dB gain
5872 0.595
5874 Apply -4.5dB gain
5875 0.500
5877 Apply -6.0dB gain (default)
5878 0.000
5880 Silence Surround Channel(s)
5881 Extended Bitstream Information - Part 2
5883 -dsurex_mode mode
5885 Dolby Surround EX Mode. Indicates whether the stream uses Dolby
5886 Surround EX (7.1 matrixed to 5.1). Using this option does NOT mean
5887 the encoder will actually apply Dolby Surround EX processing.
5888 0
5890 notindicated
5891 Not Indicated (default)
5892 1
5894 on Dolby Surround EX Off
5895 2
5897 off Dolby Surround EX On
5898 -dheadphone_mode mode
5900 Dolby Headphone Mode. Indicates whether the stream uses Dolby
5901 Headphone encoding (multi-channel matrixed to 2.0 for use with
5902 headphones). Using this option does NOT mean the encoder will
5903 actually apply Dolby Headphone processing.
5904 0
5906 notindicated
5907 Not Indicated (default)
5908 1
5910 on Dolby Headphone Off
5911 2
5913 off Dolby Headphone On
5914 -ad_conv_type type
5916 A/D Converter Type. Indicates whether the audio has passed through
5917 HDCD A/D conversion.
5918 0
5920 standard
5921 Standard A/D Converter (default)
5922 1
5924 hdcd
5925 HDCD A/D Converter
5926 Other AC-3 Encoding Options
5928 -stereo_rematrixing boolean
5930 Stereo Rematrixing. Enables/Disables use of rematrixing for stereo
5931 input. This is an optional AC-3 feature that increases quality by
5932 selectively encoding the left/right channels as mid/side. This
5933 option is enabled by default, and it is highly recommended that it
5934 be left as enabled except for testing purposes.
5935 cutoff frequency
5937 Set lowpass cutoff frequency. If unspecified, the encoder selects a
5938 default determined by various other encoding parameters.
5939 Floating-Point-Only AC-3 Encoding Options
5941 These options are only valid for the floating-point encoder and do not
5943 exist for the fixed-point encoder due to the corresponding features not
5944 being implemented in fixed-point.
5945 -channel_coupling boolean
5947 Enables/Disables use of channel coupling, which is an optional AC-3
5948 feature that increases quality by combining high frequency
5949 information from multiple channels into a single channel. The per-
5950 channel high frequency information is sent with less accuracy in
5951 both the frequency and time domains. This allows more bits to be
5952 used for lower frequencies while preserving enough information to
5953 reconstruct the high frequencies. This option is enabled by default
5954 for the floating-point encoder and should generally be left as
5955 enabled except for testing purposes or to increase encoding speed.
5956 -1
5958 auto
5959 Selected by Encoder (default)
5960 0
5962 off Disable Channel Coupling
5963 1
5965 on Enable Channel Coupling
5966 -cpl_start_band number
5968 Coupling Start Band. Sets the channel coupling start band, from 1
5969 to 15. If a value higher than the bandwidth is used, it will be
5970 reduced to 1 less than the coupling end band. If auto is used, the
5971 start band will be determined by the encoder based on the bit rate,
5972 sample rate, and channel layout. This option has no effect if
5973 channel coupling is disabled.
5974 -1
5976 auto
5977 Selected by Encoder (default)
5978 flac
5980 FLAC (Free Lossless Audio Codec) Encoder
5981 Options
5983 The following options are supported by FFmpeg's flac encoder.
5985 compression_level
5987 Sets the compression level, which chooses defaults for many other
5988 options if they are not set explicitly. Valid values are from 0 to
5989 12, 5 is the default.
5990 frame_size
5992 Sets the size of the frames in samples per channel.
5993 lpc_coeff_precision
5995 Sets the LPC coefficient precision, valid values are from 1 to 15,
5996 15 is the default.
5997 lpc_type
5999 Sets the first stage LPC algorithm
6000 none
6002 LPC is not used
6003 fixed
6005 fixed LPC coefficients
6006 levinson
6008 cholesky
6009 lpc_passes
6010 Number of passes to use for Cholesky factorization during LPC
6011 analysis
6012 min_partition_order
6014 The minimum partition order
6015 max_partition_order
6017 The maximum partition order
6018 prediction_order_method
6020 estimation
6021 2level
6022 4level
6023 8level
6024 search
6025 Bruteforce search
6026 log
6028 ch_mode
6029 Channel mode
6030 auto
6032 The mode is chosen automatically for each frame
6033 indep
6035 Channels are independently coded
6036 left_side
6038 right_side
6039 mid_side
6040 exact_rice_parameters
6041 Chooses if rice parameters are calculated exactly or approximately.
6042 if set to 1 then they are chosen exactly, which slows the code down
6043 slightly and improves compression slightly.
6044 multi_dim_quant
6046 Multi Dimensional Quantization. If set to 1 then a 2nd stage LPC
6047 algorithm is applied after the first stage to finetune the
6048 coefficients. This is quite slow and slightly improves compression.
6049 opus
6051 Opus encoder.
6052 This is a native FFmpeg encoder for the Opus format. Currently its in
6054 development and only implements the CELT part of the codec. Its quality
6055 is usually worse and at best is equal to the libopus encoder.
6056 Options
6058 b Set bit rate in bits/s. If unspecified it uses the number of
6060 channels and the layout to make a good guess.
6061 opus_delay
6063 Sets the maximum delay in milliseconds. Lower delays than 20ms will
6064 very quickly decrease quality.
6065 libfdk_aac
6067 libfdk-aac AAC (Advanced Audio Coding) encoder wrapper.
6068 The libfdk-aac library is based on the Fraunhofer FDK AAC code from the
6070 Android project.
6071 Requires the presence of the libfdk-aac headers and library during
6073 configuration. You need to explicitly configure the build with
6074 "--enable-libfdk-aac". The library is also incompatible with GPL, so if
6075 you allow the use of GPL, you should configure with "--enable-gpl
6076 --enable-nonfree --enable-libfdk-aac".
6077 This encoder has support for the AAC-HE profiles.
6079 VBR encoding, enabled through the vbr or flags +qscale options, is
6081 experimental and only works with some combinations of parameters.
6082 Support for encoding 7.1 audio is only available with libfdk-aac 0.1.3
6084 or higher.
6085 For more information see the fdk-aac project at
6087 <http://sourceforge.net/p/opencore-amr/fdk-aac/>.
6088 Options
6090 The following options are mapped on the shared FFmpeg codec options.
6092 b Set bit rate in bits/s. If the bitrate is not explicitly specified,
6094 it is automatically set to a suitable value depending on the
6095 selected profile.
6096 In case VBR mode is enabled the option is ignored.
6098 ar Set audio sampling rate (in Hz).
6100 channels
6102 Set the number of audio channels.
6103 flags +qscale
6105 Enable fixed quality, VBR (Variable Bit Rate) mode. Note that VBR
6106 is implicitly enabled when the vbr value is positive.
6107 cutoff
6109 Set cutoff frequency. If not specified (or explicitly set to 0) it
6110 will use a value automatically computed by the library. Default
6111 value is 0.
6112 profile
6114 Set audio profile.
6115 The following profiles are recognized:
6117 aac_low
6119 Low Complexity AAC (LC)
6120 aac_he
6122 High Efficiency AAC (HE-AAC)
6123 aac_he_v2
6125 High Efficiency AAC version 2 (HE-AACv2)
6126 aac_ld
6128 Low Delay AAC (LD)
6129 aac_eld
6131 Enhanced Low Delay AAC (ELD)
6132 If not specified it is set to aac_low.
6134 The following are private options of the libfdk_aac encoder.
6136 afterburner
6138 Enable afterburner feature if set to 1, disabled if set to 0. This
6139 improves the quality but also the required processing power.
6140 Default value is 1.
6142 eld_sbr
6144 Enable SBR (Spectral Band Replication) for ELD if set to 1,
6145 disabled if set to 0.
6146 Default value is 0.
6148 eld_v2
6150 Enable ELDv2 (LD-MPS extension for ELD stereo signals) for ELDv2 if
6151 set to 1, disabled if set to 0.
6152 Note that option is available when fdk-aac version
6154 (AACENCODER_LIB_VL0.AACENCODER_LIB_VL1.AACENCODER_LIB_VL2) >
6155 (4.0.0).
6156 Default value is 0.
6158 signaling
6160 Set SBR/PS signaling style.
6161 It can assume one of the following values:
6163 default
6165 choose signaling implicitly (explicit hierarchical by default,
6166 implicit if global header is disabled)
6167 implicit
6169 implicit backwards compatible signaling
6170 explicit_sbr
6172 explicit SBR, implicit PS signaling
6173 explicit_hierarchical
6175 explicit hierarchical signaling
6176 Default value is default.
6178 latm
6180 Output LATM/LOAS encapsulated data if set to 1, disabled if set to
6181 0.
6182 Default value is 0.
6184 header_period
6186 Set StreamMuxConfig and PCE repetition period (in frames) for
6187 sending in-band configuration buffers within LATM/LOAS transport
6188 layer.
6189 Must be a 16-bits non-negative integer.
6191 Default value is 0.
6193 vbr Set VBR mode, from 1 to 5. 1 is lowest quality (though still pretty
6195 good) and 5 is highest quality. A value of 0 will disable VBR, and
6196 CBR (Constant Bit Rate) is enabled.
6197 Currently only the aac_low profile supports VBR encoding.
6199 VBR modes 1-5 correspond to roughly the following average bit
6201 rates:
6202 1 32 kbps/channel
6204 2 40 kbps/channel
6206 3 48-56 kbps/channel
6208 4 64 kbps/channel
6210 5 about 80-96 kbps/channel
6212 Default value is 0.
6214 Examples
6216 • Use ffmpeg to convert an audio file to VBR AAC in an M4A (MP4)
6218 container:
6219 ffmpeg -i input.wav -codec:a libfdk_aac -vbr 3 output.m4a
6221 • Use ffmpeg to convert an audio file to CBR 64k kbps AAC, using the
6223 High-Efficiency AAC profile:
6224 ffmpeg -i input.wav -c:a libfdk_aac -profile:a aac_he -b:a 64k output.m4a
6226 libmp3lame
6228 LAME (Lame Ain't an MP3 Encoder) MP3 encoder wrapper.
6229 Requires the presence of the libmp3lame headers and library during
6231 configuration. You need to explicitly configure the build with
6232 "--enable-libmp3lame".
6233 See libshine for a fixed-point MP3 encoder, although with a lower
6235 quality.
6236 Options
6238 The following options are supported by the libmp3lame wrapper. The
6240 lame-equivalent of the options are listed in parentheses.
6241 b (-b)
6243 Set bitrate expressed in bits/s for CBR or ABR. LAME "bitrate" is
6244 expressed in kilobits/s.
6245 q (-V)
6247 Set constant quality setting for VBR. This option is valid only
6248 using the ffmpeg command-line tool. For library interface users,
6249 use global_quality.
6250 compression_level (-q)
6252 Set algorithm quality. Valid arguments are integers in the 0-9
6253 range, with 0 meaning highest quality but slowest, and 9 meaning
6254 fastest while producing the worst quality.
6255 cutoff (--lowpass)
6257 Set lowpass cutoff frequency. If unspecified, the encoder
6258 dynamically adjusts the cutoff.
6259 reservoir
6261 Enable use of bit reservoir when set to 1. Default value is 1. LAME
6262 has this enabled by default, but can be overridden by use --nores
6263 option.
6264 joint_stereo (-m j)
6266 Enable the encoder to use (on a frame by frame basis) either L/R
6267 stereo or mid/side stereo. Default value is 1.
6268 abr (--abr)
6270 Enable the encoder to use ABR when set to 1. The lame --abr sets
6271 the target bitrate, while this options only tells FFmpeg to use ABR
6272 still relies on b to set bitrate.
6273 libopencore-amrnb
6275 OpenCORE Adaptive Multi-Rate Narrowband encoder.
6276 Requires the presence of the libopencore-amrnb headers and library
6278 during configuration. You need to explicitly configure the build with
6279 "--enable-libopencore-amrnb --enable-version3".
6280 This is a mono-only encoder. Officially it only supports 8000Hz sample
6282 rate, but you can override it by setting strict to unofficial or lower.
6283 Options
6285 b Set bitrate in bits per second. Only the following bitrates are
6287 supported, otherwise libavcodec will round to the nearest valid
6288 bitrate.
6289 4750
6291 5150
6292 5900
6293 6700
6294 7400
6295 7950
6296 10200
6297 12200
6298 dtx Allow discontinuous transmission (generate comfort noise) when set
6299 to 1. The default value is 0 (disabled).
6300 libopus
6302 libopus Opus Interactive Audio Codec encoder wrapper.
6303 Requires the presence of the libopus headers and library during
6305 configuration. You need to explicitly configure the build with
6306 "--enable-libopus".
6307 Option Mapping
6309 Most libopus options are modelled after the opusenc utility from opus-
6311 tools. The following is an option mapping chart describing options
6312 supported by the libopus wrapper, and their opusenc-equivalent in
6313 parentheses.
6314 b (bitrate)
6316 Set the bit rate in bits/s. FFmpeg's b option is expressed in
6317 bits/s, while opusenc's bitrate in kilobits/s.
6318 vbr (vbr, hard-cbr, and cvbr)
6320 Set VBR mode. The FFmpeg vbr option has the following valid
6321 arguments, with the opusenc equivalent options in parentheses:
6322 off (hard-cbr)
6324 Use constant bit rate encoding.
6325 on (vbr)
6327 Use variable bit rate encoding (the default).
6328 constrained (cvbr)
6330 Use constrained variable bit rate encoding.
6331 compression_level (comp)
6333 Set encoding algorithm complexity. Valid options are integers in
6334 the 0-10 range. 0 gives the fastest encodes but lower quality,
6335 while 10 gives the highest quality but slowest encoding. The
6336 default is 10.
6337 frame_duration (framesize)
6339 Set maximum frame size, or duration of a frame in milliseconds. The
6340 argument must be exactly the following: 2.5, 5, 10, 20, 40, 60.
6341 Smaller frame sizes achieve lower latency but less quality at a
6342 given bitrate. Sizes greater than 20ms are only interesting at
6343 fairly low bitrates. The default is 20ms.
6344 packet_loss (expect-loss)
6346 Set expected packet loss percentage. The default is 0.
6347 fec (n/a)
6349 Enable inband forward error correction. packet_loss must be non-
6350 zero to take advantage - frequency of FEC 'side-data' is
6351 proportional to expected packet loss. Default is disabled.
6352 application (N.A.)
6354 Set intended application type. Valid options are listed below:
6355 voip
6357 Favor improved speech intelligibility.
6358 audio
6360 Favor faithfulness to the input (the default).
6361 lowdelay
6363 Restrict to only the lowest delay modes.
6364 cutoff (N.A.)
6366 Set cutoff bandwidth in Hz. The argument must be exactly one of the
6367 following: 4000, 6000, 8000, 12000, or 20000, corresponding to
6368 narrowband, mediumband, wideband, super wideband, and fullband
6369 respectively. The default is 0 (cutoff disabled).
6370 mapping_family (mapping_family)
6372 Set channel mapping family to be used by the encoder. The default
6373 value of -1 uses mapping family 0 for mono and stereo inputs, and
6374 mapping family 1 otherwise. The default also disables the surround
6375 masking and LFE bandwidth optimzations in libopus, and requires
6376 that the input contains 8 channels or fewer.
6377 Other values include 0 for mono and stereo, 1 for surround sound
6379 with masking and LFE bandwidth optimizations, and 255 for
6380 independent streams with an unspecified channel layout.
6381 apply_phase_inv (N.A.) (requires libopus >= 1.2)
6383 If set to 0, disables the use of phase inversion for intensity
6384 stereo, improving the quality of mono downmixes, but slightly
6385 reducing normal stereo quality. The default is 1 (phase inversion
6386 enabled).
6387 libshine
6389 Shine Fixed-Point MP3 encoder wrapper.
6390 Shine is a fixed-point MP3 encoder. It has a far better performance on
6392 platforms without an FPU, e.g. armel CPUs, and some phones and tablets.
6393 However, as it is more targeted on performance than quality, it is not
6394 on par with LAME and other production-grade encoders quality-wise.
6395 Also, according to the project's homepage, this encoder may not be free
6396 of bugs as the code was written a long time ago and the project was
6397 dead for at least 5 years.
6398 This encoder only supports stereo and mono input. This is also CBR-
6400 only.
6401 The original project (last updated in early 2007) is at
6403 <http://sourceforge.net/projects/libshine-fxp/>. We only support the
6404 updated fork by the Savonet/Liquidsoap project at
6405 <https://github.com/savonet/shine>.
6406 Requires the presence of the libshine headers and library during
6408 configuration. You need to explicitly configure the build with
6409 "--enable-libshine".
6410 See also libmp3lame.
6412 Options
6414 The following options are supported by the libshine wrapper. The
6416 shineenc-equivalent of the options are listed in parentheses.
6417 b (-b)
6419 Set bitrate expressed in bits/s for CBR. shineenc -b option is
6420 expressed in kilobits/s.
6421 libtwolame
6423 TwoLAME MP2 encoder wrapper.
6424 Requires the presence of the libtwolame headers and library during
6426 configuration. You need to explicitly configure the build with
6427 "--enable-libtwolame".
6428 Options
6430 The following options are supported by the libtwolame wrapper. The
6432 twolame-equivalent options follow the FFmpeg ones and are in
6433 parentheses.
6434 b (-b)
6436 Set bitrate expressed in bits/s for CBR. twolame b option is
6437 expressed in kilobits/s. Default value is 128k.
6438 q (-V)
6440 Set quality for experimental VBR support. Maximum value range is
6441 from -50 to 50, useful range is from -10 to 10. The higher the
6442 value, the better the quality. This option is valid only using the
6443 ffmpeg command-line tool. For library interface users, use
6444 global_quality.
6445 mode (--mode)
6447 Set the mode of the resulting audio. Possible values:
6448 auto
6450 Choose mode automatically based on the input. This is the
6451 default.
6452 stereo
6454 Stereo
6455 joint_stereo
6457 Joint stereo
6458 dual_channel
6460 Dual channel
6461 mono
6463 Mono
6464 psymodel (--psyc-mode)
6466 Set psychoacoustic model to use in encoding. The argument must be
6467 an integer between -1 and 4, inclusive. The higher the value, the
6468 better the quality. The default value is 3.
6469 energy_levels (--energy)
6471 Enable energy levels extensions when set to 1. The default value is
6472 0 (disabled).
6473 error_protection (--protect)
6475 Enable CRC error protection when set to 1. The default value is 0
6476 (disabled).
6477 copyright (--copyright)
6479 Set MPEG audio copyright flag when set to 1. The default value is 0
6480 (disabled).
6481 original (--original)
6483 Set MPEG audio original flag when set to 1. The default value is 0
6484 (disabled).
6485 libvo-amrwbenc
6487 VisualOn Adaptive Multi-Rate Wideband encoder.
6488 Requires the presence of the libvo-amrwbenc headers and library during
6490 configuration. You need to explicitly configure the build with
6491 "--enable-libvo-amrwbenc --enable-version3".
6492 This is a mono-only encoder. Officially it only supports 16000Hz sample
6494 rate, but you can override it by setting strict to unofficial or lower.
6495 Options
6497 b Set bitrate in bits/s. Only the following bitrates are supported,
6499 otherwise libavcodec will round to the nearest valid bitrate.
6500 6600
6502 8850
6503 12650
6504 14250
6505 15850
6506 18250
6507 19850
6508 23050
6509 23850
6510 dtx Allow discontinuous transmission (generate comfort noise) when set
6511 to 1. The default value is 0 (disabled).
6512 libvorbis
6514 libvorbis encoder wrapper.
6515 Requires the presence of the libvorbisenc headers and library during
6517 configuration. You need to explicitly configure the build with
6518 "--enable-libvorbis".
6519 Options
6521 The following options are supported by the libvorbis wrapper. The
6523 oggenc-equivalent of the options are listed in parentheses.
6524 To get a more accurate and extensive documentation of the libvorbis
6526 options, consult the libvorbisenc's and oggenc's documentations. See
6527 <http://xiph.org/vorbis/>, <http://wiki.xiph.org/Vorbis-tools>, and
6528 oggenc(1).
6529 b (-b)
6531 Set bitrate expressed in bits/s for ABR. oggenc -b is expressed in
6532 kilobits/s.
6533 q (-q)
6535 Set constant quality setting for VBR. The value should be a float
6536 number in the range of -1.0 to 10.0. The higher the value, the
6537 better the quality. The default value is 3.0.
6538 This option is valid only using the ffmpeg command-line tool. For
6540 library interface users, use global_quality.
6541 cutoff (--advanced-encode-option lowpass_frequency=N)
6543 Set cutoff bandwidth in Hz, a value of 0 disables cutoff. oggenc's
6544 related option is expressed in kHz. The default value is 0 (cutoff
6545 disabled).
6546 minrate (-m)
6548 Set minimum bitrate expressed in bits/s. oggenc -m is expressed in
6549 kilobits/s.
6550 maxrate (-M)
6552 Set maximum bitrate expressed in bits/s. oggenc -M is expressed in
6553 kilobits/s. This only has effect on ABR mode.
6554 iblock (--advanced-encode-option impulse_noisetune=N)
6556 Set noise floor bias for impulse blocks. The value is a float
6557 number from -15.0 to 0.0. A negative bias instructs the encoder to
6558 pay special attention to the crispness of transients in the encoded
6559 audio. The tradeoff for better transient response is a higher
6560 bitrate.
6561 mjpeg
6563 Motion JPEG encoder.
6564 Options
6566 huffman
6568 Set the huffman encoding strategy. Possible values:
6569 default
6571 Use the default huffman tables. This is the default strategy.
6572 optimal
6574 Compute and use optimal huffman tables.
6575 wavpack
6577 WavPack lossless audio encoder.
6578 Options
6580 The equivalent options for wavpack command line utility are listed in
6582 parentheses.
6583 Shared options
6585 The following shared options are effective for this encoder. Only
6587 special notes about this particular encoder will be documented here.
6588 For the general meaning of the options, see the Codec Options chapter.
6589 frame_size (--blocksize)
6591 For this encoder, the range for this option is between 128 and
6592 131072. Default is automatically decided based on sample rate and
6593 number of channel.
6594 For the complete formula of calculating default, see
6596 libavcodec/wavpackenc.c.
6597 compression_level (-f, -h, -hh, and -x)
6599 Private options
6601 joint_stereo (-j)
6603 Set whether to enable joint stereo. Valid values are:
6604 on (1)
6606 Force mid/side audio encoding.
6607 off (0)
6609 Force left/right audio encoding.
6610 auto
6612 Let the encoder decide automatically.
6613 optimize_mono
6615 Set whether to enable optimization for mono. This option is only
6616 effective for non-mono streams. Available values:
6617 on enabled
6619 off disabled
6621
6623 A description of some of the currently available video encoders
6624 follows.
6625 a64_multi, a64_multi5
6627 A64 / Commodore 64 multicolor charset encoder. "a64_multi5" is extended
6628 with 5th color (colram).
6629 Cinepak
6631 Cinepak aka CVID encoder. Compatible with Windows 3.1 and vintage
6632 MacOS.
6633 Options
6635 g integer
6637 Keyframe interval. A keyframe is inserted at least every "-g"
6638 frames, sometimes sooner.
6639 q:v integer
6641 Quality factor. Lower is better. Higher gives lower bitrate. The
6642 following table lists bitrates when encoding akiyo_cif.y4m for
6643 various values of "-q:v" with "-g 100":
6644 "-q:v 1" 1918 kb/s
6646 "-q:v 2" 1735 kb/s
6647 "-q:v 4" 1500 kb/s
6648 "-q:v 10" 1041 kb/s
6649 "-q:v 20" 826 kb/s
6650 "-q:v 40" 553 kb/s
6651 "-q:v 100" 394 kb/s
6652 "-q:v 200" 312 kb/s
6653 "-q:v 400" 266 kb/s
6654 "-q:v 1000" 237 kb/s
6655 max_extra_cb_iterations integer
6656 Max extra codebook recalculation passes, more is better and slower.
6657 skip_empty_cb boolean
6659 Avoid wasting bytes, ignore vintage MacOS decoder.
6660 max_strips integer
6662 min_strips integer
6663 The minimum and maximum number of strips to use. Wider range
6664 sometimes improves quality. More strips is generally better
6665 quality but costs more bits. Fewer strips tend to yield more
6666 keyframes. Vintage compatible is 1..3.
6667 strip_number_adaptivity integer
6669 How much number of strips is allowed to change between frames.
6670 Higher is better but slower.
6671 GIF
6673 GIF image/animation encoder.
6674 Options
6676 gifflags integer
6678 Sets the flags used for GIF encoding.
6679 offsetting
6681 Enables picture offsetting.
6682 Default is enabled.
6684 transdiff
6686 Enables transparency detection between frames.
6687 Default is enabled.
6689 gifimage integer
6691 Enables encoding one full GIF image per frame, rather than an
6692 animated GIF.
6693 Default value is 0.
6695 global_palette integer
6697 Writes a palette to the global GIF header where feasible.
6698 If disabled, every frame will always have a palette written, even
6700 if there is a global palette supplied.
6701 Default value is 1.
6703 Hap
6705 Vidvox Hap video encoder.
6706 Options
6708 format integer
6710 Specifies the Hap format to encode.
6711 hap
6713 hap_alpha
6714 hap_q
6715 Default value is hap.
6717 chunks integer
6719 Specifies the number of chunks to split frames into, between 1 and
6720 64. This permits multithreaded decoding of large frames,
6721 potentially at the cost of data-rate. The encoder may modify this
6722 value to divide frames evenly.
6723 Default value is 1.
6725 compressor integer
6727 Specifies the second-stage compressor to use. If set to none,
6728 chunks will be limited to 1, as chunked uncompressed frames offer
6729 no benefit.
6730 none
6732 snappy
6733 Default value is snappy.
6735 jpeg2000
6737 The native jpeg 2000 encoder is lossy by default, the "-q:v" option can
6738 be used to set the encoding quality. Lossless encoding can be selected
6739 with "-pred 1".
6740 Options
6742 format integer
6744 Can be set to either "j2k" or "jp2" (the default) that makes it
6745 possible to store non-rgb pix_fmts.
6746 tile_width integer
6748 Sets tile width. Range is 1 to 1073741824. Default is 256.
6749 tile_height integer
6751 Sets tile height. Range is 1 to 1073741824. Default is 256.
6752 pred integer
6754 Allows setting the discrete wavelet transform (DWT) type
6755 dwt97int (Lossy)
6757 dwt53 (Lossless)
6758 Default is "dwt97int"
6760 sop boolean
6762 Enable this to add SOP marker at the start of each packet. Disabled
6763 by default.
6764 eph boolean
6766 Enable this to add EPH marker at the end of each packet header.
6767 Disabled by default.
6768 prog integer
6770 Sets the progression order to be used by the encoder. Possible
6771 values are:
6772 lrcp
6774 rlcp
6775 rpcl
6776 pcrl
6777 cprl
6778 Set to "lrcp" by default.
6780 layer_rates string
6782 By default, when this option is not used, compression is done using
6783 the quality metric. This option allows for compression using
6784 compression ratio. The compression ratio for each level could be
6785 specified. The compression ratio of a layer "l" species the what
6786 ratio of total file size is contained in the first "l" layers.
6787 Example usage:
6789 ffmpeg -i input.bmp -c:v jpeg2000 -layer_rates "100,10,1" output.j2k
6791 This would compress the image to contain 3 layers, where the data
6793 contained in the first layer would be compressed by 1000 times,
6794 compressed by 100 in the first two layers, and shall contain all
6795 data while using all 3 layers.
6796 librav1e
6798 rav1e AV1 encoder wrapper.
6799 Requires the presence of the rav1e headers and library during
6801 configuration. You need to explicitly configure the build with
6802 "--enable-librav1e".
6803 Options
6805 qmax
6807 Sets the maximum quantizer to use when using bitrate mode.
6808 qmin
6810 Sets the minimum quantizer to use when using bitrate mode.
6811 qp Uses quantizer mode to encode at the given quantizer (0-255).
6813 speed
6815 Selects the speed preset (0-10) to encode with.
6816 tiles
6818 Selects how many tiles to encode with.
6819 tile-rows
6821 Selects how many rows of tiles to encode with.
6822 tile-columns
6824 Selects how many columns of tiles to encode with.
6825 rav1e-params
6827 Set rav1e options using a list of key=value pairs separated by ":".
6828 See rav1e --help for a list of options.
6829 For example to specify librav1e encoding options with
6831 -rav1e-params:
6832 ffmpeg -i input -c:v librav1e -b:v 500K -rav1e-params speed=5:low_latency=true output.mp4
6834 libaom-av1
6836 libaom AV1 encoder wrapper.
6837 Requires the presence of the libaom headers and library during
6839 configuration. You need to explicitly configure the build with
6840 "--enable-libaom".
6841 Options
6843 The wrapper supports the following standard libavcodec options:
6845 b Set bitrate target in bits/second. By default this will use
6847 variable-bitrate mode. If maxrate and minrate are also set to the
6848 same value then it will use constant-bitrate mode, otherwise if crf
6849 is set as well then it will use constrained-quality mode.
6850 g keyint_min
6852 Set key frame placement. The GOP size sets the maximum distance
6853 between key frames; if zero the output stream will be intra-only.
6854 The minimum distance is ignored unless it is the same as the GOP
6855 size, in which case key frames will always appear at a fixed
6856 interval. Not set by default, so without this option the library
6857 has completely free choice about where to place key frames.
6858 qmin qmax
6860 Set minimum/maximum quantisation values. Valid range is from 0 to
6861 63 (warning: this does not match the quantiser values actually used
6862 by AV1 - divide by four to map real quantiser values to this
6863 range). Defaults to min/max (no constraint).
6864 minrate maxrate bufsize rc_init_occupancy
6866 Set rate control buffering parameters. Not used if not set -
6867 defaults to unconstrained variable bitrate.
6868 threads
6870 Set the number of threads to use while encoding. This may require
6871 the tiles or row-mt options to also be set to actually use the
6872 specified number of threads fully. Defaults to the number of
6873 hardware threads supported by the host machine.
6874 profile
6876 Set the encoding profile. Defaults to using the profile which
6877 matches the bit depth and chroma subsampling of the input.
6878 The wrapper also has some specific options:
6880 cpu-used
6882 Set the quality/encoding speed tradeoff. Valid range is from 0 to
6883 8, higher numbers indicating greater speed and lower quality. The
6884 default value is 1, which will be slow and high quality.
6885 auto-alt-ref
6887 Enable use of alternate reference frames. Defaults to the internal
6888 default of the library.
6889 arnr-max-frames (frames)
6891 Set altref noise reduction max frame count. Default is -1.
6892 arnr-strength (strength)
6894 Set altref noise reduction filter strength. Range is -1 to 6.
6895 Default is -1.
6896 aq-mode (aq-mode)
6898 Set adaptive quantization mode. Possible values:
6899 none (0)
6901 Disabled.
6902 variance (1)
6904 Variance-based.
6905 complexity (2)
6907 Complexity-based.
6908 cyclic (3)
6910 Cyclic refresh.
6911 tune (tune)
6913 Set the distortion metric the encoder is tuned with. Default is
6914 "psnr".
6915 psnr (0)
6917 ssim (1)
6918 lag-in-frames
6919 Set the maximum number of frames which the encoder may keep in
6920 flight at any one time for lookahead purposes. Defaults to the
6921 internal default of the library.
6922 error-resilience
6924 Enable error resilience features:
6925 default
6927 Improve resilience against losses of whole frames.
6928 Not enabled by default.
6930 crf Set the quality/size tradeoff for constant-quality (no bitrate
6932 target) and constrained-quality (with maximum bitrate target)
6933 modes. Valid range is 0 to 63, higher numbers indicating lower
6934 quality and smaller output size. Only used if set; by default only
6935 the bitrate target is used.
6936 static-thresh
6938 Set a change threshold on blocks below which they will be skipped
6939 by the encoder. Defined in arbitrary units as a nonnegative
6940 integer, defaulting to zero (no blocks are skipped).
6941 drop-threshold
6943 Set a threshold for dropping frames when close to rate control
6944 bounds. Defined as a percentage of the target buffer - when the
6945 rate control buffer falls below this percentage, frames will be
6946 dropped until it has refilled above the threshold. Defaults to
6947 zero (no frames are dropped).
6948 denoise-noise-level (level)
6950 Amount of noise to be removed for grain synthesis. Grain synthesis
6951 is disabled if this option is not set or set to 0.
6952 denoise-block-size (pixels)
6954 Block size used for denoising for grain synthesis. If not set, AV1
6955 codec uses the default value of 32.
6956 undershoot-pct (pct)
6958 Set datarate undershoot (min) percentage of the target bitrate.
6959 Range is -1 to 100. Default is -1.
6960 overshoot-pct (pct)
6962 Set datarate overshoot (max) percentage of the target bitrate.
6963 Range is -1 to 1000. Default is -1.
6964 minsection-pct (pct)
6966 Minimum percentage variation of the GOP bitrate from the target
6967 bitrate. If minsection-pct is not set, the libaomenc wrapper
6968 computes it as follows: "(minrate * 100 / bitrate)". Range is -1
6969 to 100. Default is -1 (unset).
6970 maxsection-pct (pct)
6972 Maximum percentage variation of the GOP bitrate from the target
6973 bitrate. If maxsection-pct is not set, the libaomenc wrapper
6974 computes it as follows: "(maxrate * 100 / bitrate)". Range is -1
6975 to 5000. Default is -1 (unset).
6976 frame-parallel (boolean)
6978 Enable frame parallel decodability features. Default is true.
6979 tiles
6981 Set the number of tiles to encode the input video with, as columns
6982 x rows. Larger numbers allow greater parallelism in both encoding
6983 and decoding, but may decrease coding efficiency. Defaults to the
6984 minimum number of tiles required by the size of the input video
6985 (this is 1x1 (that is, a single tile) for sizes up to and including
6986 4K).
6987 tile-columns tile-rows
6989 Set the number of tiles as log2 of the number of tile rows and
6990 columns. Provided for compatibility with libvpx/VP9.
6991 row-mt (Requires libaom >= 1.0.0-759-g90a15f4f2)
6993 Enable row based multi-threading. Disabled by default.
6994 enable-cdef (boolean)
6996 Enable Constrained Directional Enhancement Filter. The libaom-av1
6997 encoder enables CDEF by default.
6998 enable-restoration (boolean)
7000 Enable Loop Restoration Filter. Default is true for libaom-av1.
7001 enable-global-motion (boolean)
7003 Enable the use of global motion for block prediction. Default is
7004 true.
7005 enable-intrabc (boolean)
7007 Enable block copy mode for intra block prediction. This mode is
7008 useful for screen content. Default is true.
7009 enable-rect-partitions (boolean) (Requires libaom >= v2.0.0)
7011 Enable rectangular partitions. Default is true.
7012 enable-1to4-partitions (boolean) (Requires libaom >= v2.0.0)
7014 Enable 1:4/4:1 partitions. Default is true.
7015 enable-ab-partitions (boolean) (Requires libaom >= v2.0.0)
7017 Enable AB shape partitions. Default is true.
7018 enable-angle-delta (boolean) (Requires libaom >= v2.0.0)
7020 Enable angle delta intra prediction. Default is true.
7021 enable-cfl-intra (boolean) (Requires libaom >= v2.0.0)
7023 Enable chroma predicted from luma intra prediction. Default is
7024 true.
7025 enable-filter-intra (boolean) (Requires libaom >= v2.0.0)
7027 Enable filter intra predictor. Default is true.
7028 enable-intra-edge-filter (boolean) (Requires libaom >= v2.0.0)
7030 Enable intra edge filter. Default is true.
7031 enable-smooth-intra (boolean) (Requires libaom >= v2.0.0)
7033 Enable smooth intra prediction mode. Default is true.
7034 enable-paeth-intra (boolean) (Requires libaom >= v2.0.0)
7036 Enable paeth predictor in intra prediction. Default is true.
7037 enable-palette (boolean) (Requires libaom >= v2.0.0)
7039 Enable palette prediction mode. Default is true.
7040 enable-flip-idtx (boolean) (Requires libaom >= v2.0.0)
7042 Enable extended transform type, including FLIPADST_DCT,
7043 DCT_FLIPADST, FLIPADST_FLIPADST, ADST_FLIPADST, FLIPADST_ADST,
7044 IDTX, V_DCT, H_DCT, V_ADST, H_ADST, V_FLIPADST, H_FLIPADST. Default
7045 is true.
7046 enable-tx64 (boolean) (Requires libaom >= v2.0.0)
7048 Enable 64-pt transform. Default is true.
7049 reduced-tx-type-set (boolean) (Requires libaom >= v2.0.0)
7051 Use reduced set of transform types. Default is false.
7052 use-intra-dct-only (boolean) (Requires libaom >= v2.0.0)
7054 Use DCT only for INTRA modes. Default is false.
7055 use-inter-dct-only (boolean) (Requires libaom >= v2.0.0)
7057 Use DCT only for INTER modes. Default is false.
7058 use-intra-default-tx-only (boolean) (Requires libaom >= v2.0.0)
7060 Use Default-transform only for INTRA modes. Default is false.
7061 enable-ref-frame-mvs (boolean) (Requires libaom >= v2.0.0)
7063 Enable temporal mv prediction. Default is true.
7064 enable-reduced-reference-set (boolean) (Requires libaom >= v2.0.0)
7066 Use reduced set of single and compound references. Default is
7067 false.
7068 enable-obmc (boolean) (Requires libaom >= v2.0.0)
7070 Enable obmc. Default is true.
7071 enable-dual-filter (boolean) (Requires libaom >= v2.0.0)
7073 Enable dual filter. Default is true.
7074 enable-diff-wtd-comp (boolean) (Requires libaom >= v2.0.0)
7076 Enable difference-weighted compound. Default is true.
7077 enable-dist-wtd-comp (boolean) (Requires libaom >= v2.0.0)
7079 Enable distance-weighted compound. Default is true.
7080 enable-onesided-comp (boolean) (Requires libaom >= v2.0.0)
7082 Enable one sided compound. Default is true.
7083 enable-interinter-wedge (boolean) (Requires libaom >= v2.0.0)
7085 Enable interinter wedge compound. Default is true.
7086 enable-interintra-wedge (boolean) (Requires libaom >= v2.0.0)
7088 Enable interintra wedge compound. Default is true.
7089 enable-masked-comp (boolean) (Requires libaom >= v2.0.0)
7091 Enable masked compound. Default is true.
7092 enable-interintra-comp (boolean) (Requires libaom >= v2.0.0)
7094 Enable interintra compound. Default is true.
7095 enable-smooth-interintra (boolean) (Requires libaom >= v2.0.0)
7097 Enable smooth interintra mode. Default is true.
7098 aom-params
7100 Set libaom options using a list of key=value pairs separated by
7101 ":". For a list of supported options, see aomenc --help under the
7102 section "AV1 Specific Options".
7103 For example to specify libaom encoding options with -aom-params:
7105 ffmpeg -i input -c:v libaom-av1 -b:v 500K -aom-params tune=psnr:enable-tpl-model=1 output.mp4
7107 libsvtav1
7109 SVT-AV1 encoder wrapper.
7110 Requires the presence of the SVT-AV1 headers and library during
7112 configuration. You need to explicitly configure the build with
7113 "--enable-libsvtav1".
7114 Options
7116 profile
7118 Set the encoding profile.
7119 main
7121 high
7122 professional
7123 level
7124 Set the operating point level. For example: '4.0'
7125 hielevel
7127 Set the Hierarchical prediction levels.
7128 3level
7130 4level
7131 This is the default.
7132 tier
7134 Set the operating point tier.
7135 main
7137 This is the default.
7138 high
7140 qmax
7141 Set the maximum quantizer to use when using a bitrate mode.
7142 qmin
7144 Set the minimum quantizer to use when using a bitrate mode.
7145 crf Constant rate factor value used in crf rate control mode (0-63).
7147 qp Set the quantizer used in cqp rate control mode (0-63).
7149 sc_detection
7151 Enable scene change detection.
7152 la_depth
7154 Set number of frames to look ahead (0-120).
7155 preset
7157 Set the quality-speed tradeoff, in the range 0 to 13. Higher
7158 values are faster but lower quality.
7159 tile_rows
7161 Set log2 of the number of rows of tiles to use (0-6).
7162 tile_columns
7164 Set log2 of the number of columns of tiles to use (0-4).
7165 svtav1-params
7167 Set SVT-AV1 options using a list of key=value pairs separated by
7168 ":". See the SVT-AV1 encoder user guide for a list of accepted
7169 parameters.
7170 libjxl
7172 libjxl JPEG XL encoder wrapper.
7173 Requires the presence of the libjxl headers and library during
7175 configuration. You need to explicitly configure the build with
7176 "--enable-libjxl".
7177 Options
7179 The libjxl wrapper supports the following options:
7181 distance
7183 Set the target Butteraugli distance. This is a quality setting:
7184 lower distance yields higher quality, with distance=1.0 roughly
7185 comparable to libjpeg Quality 90 for photographic content. Setting
7186 distance=0.0 yields true lossless encoding. Valid values range
7187 between 0.0 and 15.0, and sane values rarely exceed 5.0. Setting
7188 distance=0.1 usually attains transparency for most input. The
7189 default is 1.0.
7190 effort
7192 Set the encoding effort used. Higher effort values produce more
7193 consistent quality and usually produces a better quality/bpp curve,
7194 at the cost of more CPU time required. Valid values range from 1 to
7195 9, and the default is 7.
7196 modular
7198 Force the encoder to use Modular mode instead of choosing
7199 automatically. The default is to use VarDCT for lossy encoding and
7200 Modular for lossless. VarDCT is generally superior to Modular for
7201 lossy encoding but does not support lossless encoding.
7202 libkvazaar
7204 Kvazaar H.265/HEVC encoder.
7205 Requires the presence of the libkvazaar headers and library during
7207 configuration. You need to explicitly configure the build with
7208 --enable-libkvazaar.
7209 Options
7211 b Set target video bitrate in bit/s and enable rate control.
7213 kvazaar-params
7215 Set kvazaar parameters as a list of name=value pairs separated by
7216 commas (,). See kvazaar documentation for a list of options.
7217 libopenh264
7219 Cisco libopenh264 H.264/MPEG-4 AVC encoder wrapper.
7220 This encoder requires the presence of the libopenh264 headers and
7222 library during configuration. You need to explicitly configure the
7223 build with "--enable-libopenh264". The library is detected using pkg-
7224 config.
7225 For more information about the library see <http://www.openh264.org>.
7227 Options
7229 The following FFmpeg global options affect the configurations of the
7231 libopenh264 encoder.
7232 b Set the bitrate (as a number of bits per second).
7234 g Set the GOP size.
7236 maxrate
7238 Set the max bitrate (as a number of bits per second).
7239 flags +global_header
7241 Set global header in the bitstream.
7242 slices
7244 Set the number of slices, used in parallelized encoding. Default
7245 value is 0. This is only used when slice_mode is set to fixed.
7246 loopfilter
7248 Enable loop filter, if set to 1 (automatically enabled). To disable
7249 set a value of 0.
7250 profile
7252 Set profile restrictions. If set to the value of main enable CABAC
7253 (set the "SEncParamExt.iEntropyCodingModeFlag" flag to 1).
7254 max_nal_size
7256 Set maximum NAL size in bytes.
7257 allow_skip_frames
7259 Allow skipping frames to hit the target bitrate if set to 1.
7260 libtheora
7262 libtheora Theora encoder wrapper.
7263 Requires the presence of the libtheora headers and library during
7265 configuration. You need to explicitly configure the build with
7266 "--enable-libtheora".
7267 For more information about the libtheora project see
7269 <http://www.theora.org/>.
7270 Options
7272 The following global options are mapped to internal libtheora options
7274 which affect the quality and the bitrate of the encoded stream.
7275 b Set the video bitrate in bit/s for CBR (Constant Bit Rate) mode.
7277 In case VBR (Variable Bit Rate) mode is enabled this option is
7278 ignored.
7279 flags
7281 Used to enable constant quality mode (VBR) encoding through the
7282 qscale flag, and to enable the "pass1" and "pass2" modes.
7283 g Set the GOP size.
7285 global_quality
7287 Set the global quality as an integer in lambda units.
7288 Only relevant when VBR mode is enabled with "flags +qscale". The
7290 value is converted to QP units by dividing it by "FF_QP2LAMBDA",
7291 clipped in the [0 - 10] range, and then multiplied by 6.3 to get a
7292 value in the native libtheora range [0-63]. A higher value
7293 corresponds to a higher quality.
7294 q Enable VBR mode when set to a non-negative value, and set constant
7296 quality value as a double floating point value in QP units.
7297 The value is clipped in the [0-10] range, and then multiplied by
7299 6.3 to get a value in the native libtheora range [0-63].
7300 This option is valid only using the ffmpeg command-line tool. For
7302 library interface users, use global_quality.
7303 Examples
7305 • Set maximum constant quality (VBR) encoding with ffmpeg:
7307 ffmpeg -i INPUT -codec:v libtheora -q:v 10 OUTPUT.ogg
7309 • Use ffmpeg to convert a CBR 1000 kbps Theora video stream:
7311 ffmpeg -i INPUT -codec:v libtheora -b:v 1000k OUTPUT.ogg
7313 libvpx
7315 VP8/VP9 format supported through libvpx.
7316 Requires the presence of the libvpx headers and library during
7318 configuration. You need to explicitly configure the build with
7319 "--enable-libvpx".
7320 Options
7322 The following options are supported by the libvpx wrapper. The
7324 vpxenc-equivalent options or values are listed in parentheses for easy
7325 migration.
7326 To reduce the duplication of documentation, only the private options
7328 and some others requiring special attention are documented here. For
7329 the documentation of the undocumented generic options, see the Codec
7330 Options chapter.
7331 To get more documentation of the libvpx options, invoke the command
7333 ffmpeg -h encoder=libvpx, ffmpeg -h encoder=libvpx-vp9 or vpxenc
7334 --help. Further information is available in the libvpx API
7335 documentation.
7336 b (target-bitrate)
7338 Set bitrate in bits/s. Note that FFmpeg's b option is expressed in
7339 bits/s, while vpxenc's target-bitrate is in kilobits/s.
7340 g (kf-max-dist)
7342 keyint_min (kf-min-dist)
7343 qmin (min-q)
7344 Minimum (Best Quality) Quantizer.
7345 qmax (max-q)
7347 Maximum (Worst Quality) Quantizer. Can be changed per-frame.
7348 bufsize (buf-sz, buf-optimal-sz)
7350 Set ratecontrol buffer size (in bits). Note vpxenc's options are
7351 specified in milliseconds, the libvpx wrapper converts this value
7352 as follows: "buf-sz = bufsize * 1000 / bitrate", "buf-optimal-sz =
7353 bufsize * 1000 / bitrate * 5 / 6".
7354 rc_init_occupancy (buf-initial-sz)
7356 Set number of bits which should be loaded into the rc buffer before
7357 decoding starts. Note vpxenc's option is specified in milliseconds,
7358 the libvpx wrapper converts this value as follows:
7359 "rc_init_occupancy * 1000 / bitrate".
7360 undershoot-pct
7362 Set datarate undershoot (min) percentage of the target bitrate.
7363 overshoot-pct
7365 Set datarate overshoot (max) percentage of the target bitrate.
7366 skip_threshold (drop-frame)
7368 qcomp (bias-pct)
7369 maxrate (maxsection-pct)
7370 Set GOP max bitrate in bits/s. Note vpxenc's option is specified as
7371 a percentage of the target bitrate, the libvpx wrapper converts
7372 this value as follows: "(maxrate * 100 / bitrate)".
7373 minrate (minsection-pct)
7375 Set GOP min bitrate in bits/s. Note vpxenc's option is specified as
7376 a percentage of the target bitrate, the libvpx wrapper converts
7377 this value as follows: "(minrate * 100 / bitrate)".
7378 minrate, maxrate, b end-usage=cbr
7380 "(minrate == maxrate == bitrate)".
7381 crf (end-usage=cq, cq-level)
7383 tune (tune)
7384 psnr (psnr)
7385 ssim (ssim)
7386 quality, deadline (deadline)
7387 best
7388 Use best quality deadline. Poorly named and quite slow, this
7389 option should be avoided as it may give worse quality output
7390 than good.
7391 good
7393 Use good quality deadline. This is a good trade-off between
7394 speed and quality when used with the cpu-used option.
7395 realtime
7397 Use realtime quality deadline.
7398 speed, cpu-used (cpu-used)
7400 Set quality/speed ratio modifier. Higher values speed up the encode
7401 at the cost of quality.
7402 nr (noise-sensitivity)
7404 static-thresh
7405 Set a change threshold on blocks below which they will be skipped
7406 by the encoder.
7407 slices (token-parts)
7409 Note that FFmpeg's slices option gives the total number of
7410 partitions, while vpxenc's token-parts is given as
7411 log2(partitions).
7412 max-intra-rate
7414 Set maximum I-frame bitrate as a percentage of the target bitrate.
7415 A value of 0 means unlimited.
7416 force_key_frames
7418 "VPX_EFLAG_FORCE_KF"
7419 Alternate reference frame related
7421 auto-alt-ref
7422 Enable use of alternate reference frames (2-pass only). Values
7423 greater than 1 enable multi-layer alternate reference frames
7424 (VP9 only).
7425 arnr-maxframes
7427 Set altref noise reduction max frame count.
7428 arnr-type
7430 Set altref noise reduction filter type: backward, forward,
7431 centered.
7432 arnr-strength
7434 Set altref noise reduction filter strength.
7435 rc-lookahead, lag-in-frames (lag-in-frames)
7437 Set number of frames to look ahead for frametype and
7438 ratecontrol.
7439 min-gf-interval
7441 Set minimum golden/alternate reference frame interval (VP9
7442 only).
7443 error-resilient
7445 Enable error resiliency features.
7446 sharpness integer
7448 Increase sharpness at the expense of lower PSNR. The valid range
7449 is [0, 7].
7450 ts-parameters
7452 Sets the temporal scalability configuration using a :-separated
7453 list of key=value pairs. For example, to specify temporal
7454 scalability parameters with "ffmpeg":
7455 ffmpeg -i INPUT -c:v libvpx -ts-parameters ts_number_layers=3:\
7457 ts_target_bitrate=250,500,1000:ts_rate_decimator=4,2,1:\
7458 ts_periodicity=4:ts_layer_id=0,2,1,2:ts_layering_mode=3 OUTPUT
7459 Below is a brief explanation of each of the parameters, please
7461 refer to "struct vpx_codec_enc_cfg" in "vpx/vpx_encoder.h" for more
7462 details.
7463 ts_number_layers
7465 Number of temporal coding layers.
7466 ts_target_bitrate
7468 Target bitrate for each temporal layer (in kbps). (bitrate
7469 should be inclusive of the lower temporal layer).
7470 ts_rate_decimator
7472 Frame rate decimation factor for each temporal layer.
7473 ts_periodicity
7475 Length of the sequence defining frame temporal layer
7476 membership.
7477 ts_layer_id
7479 Template defining the membership of frames to temporal layers.
7480 ts_layering_mode
7482 (optional) Selecting the temporal structure from a set of pre-
7483 defined temporal layering modes. Currently supports the
7484 following options.
7485 0 No temporal layering flags are provided internally, relies
7487 on flags being passed in using "metadata" field in
7488 "AVFrame" with following keys.
7489 vp8-flags
7491 Sets the flags passed into the encoder to indicate the
7492 referencing scheme for the current frame. Refer to
7493 function "vpx_codec_encode" in "vpx/vpx_encoder.h" for
7494 more details.
7495 temporal_id
7497 Explicitly sets the temporal id of the current frame to
7498 encode.
7499 2 Two temporal layers. 0-1...
7501 3 Three temporal layers. 0-2-1-2...; with single reference
7503 frame.
7504 4 Same as option "3", except there is a dependency between
7506 the two temporal layer 2 frames within the temporal period.
7507 VP9-specific options
7509 lossless
7510 Enable lossless mode.
7511 tile-columns
7513 Set number of tile columns to use. Note this is given as
7514 log2(tile_columns). For example, 8 tile columns would be
7515 requested by setting the tile-columns option to 3.
7516 tile-rows
7518 Set number of tile rows to use. Note this is given as
7519 log2(tile_rows). For example, 4 tile rows would be requested
7520 by setting the tile-rows option to 2.
7521 frame-parallel
7523 Enable frame parallel decodability features.
7524 aq-mode
7526 Set adaptive quantization mode (0: off (default), 1: variance
7527 2: complexity, 3: cyclic refresh, 4: equator360).
7528 colorspace color-space
7530 Set input color space. The VP9 bitstream supports signaling the
7531 following colorspaces:
7532 rgb sRGB
7534 bt709 bt709
7535 unspecified unknown
7536 bt470bg bt601
7537 smpte170m smpte170
7538 smpte240m smpte240
7539 bt2020_ncl bt2020
7540 row-mt boolean
7541 Enable row based multi-threading.
7542 tune-content
7544 Set content type: default (0), screen (1), film (2).
7545 corpus-complexity
7547 Corpus VBR mode is a variant of standard VBR where the
7548 complexity distribution midpoint is passed in rather than
7549 calculated for a specific clip or chunk.
7550 The valid range is [0, 10000]. 0 (default) uses standard VBR.
7552 enable-tpl boolean
7554 Enable temporal dependency model.
7555 ref-frame-config
7557 Using per-frame metadata, set members of the structure
7558 "vpx_svc_ref_frame_config_t" in "vpx/vp8cx.h" to fine-control
7559 referencing schemes and frame buffer management. Use a
7560 :-separated list of key=value pairs. For example,
7561 av_dict_set(&av_frame->metadata, "ref-frame-config", \
7563 "rfc_update_buffer_slot=7:rfc_lst_fb_idx=0:rfc_gld_fb_idx=1:rfc_alt_fb_idx=2:rfc_reference_last=0:rfc_reference_golden=0:rfc_reference_alt_ref=0");
7564 rfc_update_buffer_slot
7566 Indicates the buffer slot number to update
7567 rfc_update_last
7569 Indicates whether to update the LAST frame
7570 rfc_update_golden
7572 Indicates whether to update GOLDEN frame
7573 rfc_update_alt_ref
7575 Indicates whether to update ALT_REF frame
7576 rfc_lst_fb_idx
7578 LAST frame buffer index
7579 rfc_gld_fb_idx
7581 GOLDEN frame buffer index
7582 rfc_alt_fb_idx
7584 ALT_REF frame buffer index
7585 rfc_reference_last
7587 Indicates whether to reference LAST frame
7588 rfc_reference_golden
7590 Indicates whether to reference GOLDEN frame
7591 rfc_reference_alt_ref
7593 Indicates whether to reference ALT_REF frame
7594 rfc_reference_duration
7596 Indicates frame duration
7597 For more information about libvpx see: <http://www.webmproject.org/>
7599 libwebp
7601 libwebp WebP Image encoder wrapper
7602 libwebp is Google's official encoder for WebP images. It can encode in
7604 either lossy or lossless mode. Lossy images are essentially a wrapper
7605 around a VP8 frame. Lossless images are a separate codec developed by
7606 Google.
7607 Pixel Format
7609 Currently, libwebp only supports YUV420 for lossy and RGB for lossless
7611 due to limitations of the format and libwebp. Alpha is supported for
7612 either mode. Because of API limitations, if RGB is passed in when
7613 encoding lossy or YUV is passed in for encoding lossless, the pixel
7614 format will automatically be converted using functions from libwebp.
7615 This is not ideal and is done only for convenience.
7616 Options
7618 -lossless boolean
7620 Enables/Disables use of lossless mode. Default is 0.
7621 -compression_level integer
7623 For lossy, this is a quality/speed tradeoff. Higher values give
7624 better quality for a given size at the cost of increased encoding
7625 time. For lossless, this is a size/speed tradeoff. Higher values
7626 give smaller size at the cost of increased encoding time. More
7627 specifically, it controls the number of extra algorithms and
7628 compression tools used, and varies the combination of these tools.
7629 This maps to the method option in libwebp. The valid range is 0 to
7630 6. Default is 4.
7631 -quality float
7633 For lossy encoding, this controls image quality. For lossless
7634 encoding, this controls the effort and time spent in compression.
7635 Range is 0 to 100. Default is 75.
7636 -preset type
7638 Configuration preset. This does some automatic settings based on
7639 the general type of the image.
7640 none
7642 Do not use a preset.
7643 default
7645 Use the encoder default.
7646 picture
7648 Digital picture, like portrait, inner shot
7649 photo
7651 Outdoor photograph, with natural lighting
7652 drawing
7654 Hand or line drawing, with high-contrast details
7655 icon
7657 Small-sized colorful images
7658 text
7660 Text-like
7661 libx264, libx264rgb
7663 x264 H.264/MPEG-4 AVC encoder wrapper.
7664 This encoder requires the presence of the libx264 headers and library
7666 during configuration. You need to explicitly configure the build with
7667 "--enable-libx264".
7668 libx264 supports an impressive number of features, including 8x8 and
7670 4x4 adaptive spatial transform, adaptive B-frame placement, CAVLC/CABAC
7671 entropy coding, interlacing (MBAFF), lossless mode, psy optimizations
7672 for detail retention (adaptive quantization, psy-RD, psy-trellis).
7673 Many libx264 encoder options are mapped to FFmpeg global codec options,
7675 while unique encoder options are provided through private options.
7676 Additionally the x264opts and x264-params private options allows one to
7677 pass a list of key=value tuples as accepted by the libx264
7678 "x264_param_parse" function.
7679 The x264 project website is at
7681 <http://www.videolan.org/developers/x264.html>.
7682 The libx264rgb encoder is the same as libx264, except it accepts packed
7684 RGB pixel formats as input instead of YUV.
7685 Supported Pixel Formats
7687 x264 supports 8- to 10-bit color spaces. The exact bit depth is
7689 controlled at x264's configure time.
7690 Options
7692 The following options are supported by the libx264 wrapper. The
7694 x264-equivalent options or values are listed in parentheses for easy
7695 migration.
7696 To reduce the duplication of documentation, only the private options
7698 and some others requiring special attention are documented here. For
7699 the documentation of the undocumented generic options, see the Codec
7700 Options chapter.
7701 To get a more accurate and extensive documentation of the libx264
7703 options, invoke the command x264 --fullhelp or consult the libx264
7704 documentation.
7705 b (bitrate)
7707 Set bitrate in bits/s. Note that FFmpeg's b option is expressed in
7708 bits/s, while x264's bitrate is in kilobits/s.
7709 bf (bframes)
7711 g (keyint)
7712 qmin (qpmin)
7713 Minimum quantizer scale.
7714 qmax (qpmax)
7716 Maximum quantizer scale.
7717 qdiff (qpstep)
7719 Maximum difference between quantizer scales.
7720 qblur (qblur)
7722 Quantizer curve blur
7723 qcomp (qcomp)
7725 Quantizer curve compression factor
7726 refs (ref)
7728 Number of reference frames each P-frame can use. The range is from
7729 0-16.
7730 sc_threshold (scenecut)
7732 Sets the threshold for the scene change detection.
7733 trellis (trellis)
7735 Performs Trellis quantization to increase efficiency. Enabled by
7736 default.
7737 nr (nr)
7739 me_range (merange)
7740 Maximum range of the motion search in pixels.
7741 me_method (me)
7743 Set motion estimation method. Possible values in the decreasing
7744 order of speed:
7745 dia (dia)
7747 epzs (dia)
7748 Diamond search with radius 1 (fastest). epzs is an alias for
7749 dia.
7750 hex (hex)
7752 Hexagonal search with radius 2.
7753 umh (umh)
7755 Uneven multi-hexagon search.
7756 esa (esa)
7758 Exhaustive search.
7759 tesa (tesa)
7761 Hadamard exhaustive search (slowest).
7762 forced-idr
7764 Normally, when forcing a I-frame type, the encoder can select any
7765 type of I-frame. This option forces it to choose an IDR-frame.
7766 subq (subme)
7768 Sub-pixel motion estimation method.
7769 b_strategy (b-adapt)
7771 Adaptive B-frame placement decision algorithm. Use only on first-
7772 pass.
7773 keyint_min (min-keyint)
7775 Minimum GOP size.
7776 coder
7778 Set entropy encoder. Possible values:
7779 ac Enable CABAC.
7781 vlc Enable CAVLC and disable CABAC. It generates the same effect as
7783 x264's --no-cabac option.
7784 cmp Set full pixel motion estimation comparison algorithm. Possible
7786 values:
7787 chroma
7789 Enable chroma in motion estimation.
7790 sad Ignore chroma in motion estimation. It generates the same
7792 effect as x264's --no-chroma-me option.
7793 threads (threads)
7795 Number of encoding threads.
7796 thread_type
7798 Set multithreading technique. Possible values:
7799 slice
7801 Slice-based multithreading. It generates the same effect as
7802 x264's --sliced-threads option.
7803 frame
7805 Frame-based multithreading.
7806 flags
7808 Set encoding flags. It can be used to disable closed GOP and enable
7809 open GOP by setting it to "-cgop". The result is similar to the
7810 behavior of x264's --open-gop option.
7811 rc_init_occupancy (vbv-init)
7813 preset (preset)
7814 Set the encoding preset.
7815 tune (tune)
7817 Set tuning of the encoding params.
7818 profile (profile)
7820 Set profile restrictions.
7821 fastfirstpass
7823 Enable fast settings when encoding first pass, when set to 1. When
7824 set to 0, it has the same effect of x264's --slow-firstpass option.
7825 crf (crf)
7827 Set the quality for constant quality mode.
7828 crf_max (crf-max)
7830 In CRF mode, prevents VBV from lowering quality beyond this point.
7831 qp (qp)
7833 Set constant quantization rate control method parameter.
7834 aq-mode (aq-mode)
7836 Set AQ method. Possible values:
7837 none (0)
7839 Disabled.
7840 variance (1)
7842 Variance AQ (complexity mask).
7843 autovariance (2)
7845 Auto-variance AQ (experimental).
7846 aq-strength (aq-strength)
7848 Set AQ strength, reduce blocking and blurring in flat and textured
7849 areas.
7850 psy Use psychovisual optimizations when set to 1. When set to 0, it has
7852 the same effect as x264's --no-psy option.
7853 psy-rd (psy-rd)
7855 Set strength of psychovisual optimization, in psy-rd:psy-trellis
7856 format.
7857 rc-lookahead (rc-lookahead)
7859 Set number of frames to look ahead for frametype and ratecontrol.
7860 weightb
7862 Enable weighted prediction for B-frames when set to 1. When set to
7863 0, it has the same effect as x264's --no-weightb option.
7864 weightp (weightp)
7866 Set weighted prediction method for P-frames. Possible values:
7867 none (0)
7869 Disabled
7870 simple (1)
7872 Enable only weighted refs
7873 smart (2)
7875 Enable both weighted refs and duplicates
7876 ssim (ssim)
7878 Enable calculation and printing SSIM stats after the encoding.
7879 intra-refresh (intra-refresh)
7881 Enable the use of Periodic Intra Refresh instead of IDR frames when
7882 set to 1.
7883 avcintra-class (class)
7885 Configure the encoder to generate AVC-Intra. Valid values are
7886 50,100 and 200
7887 bluray-compat (bluray-compat)
7889 Configure the encoder to be compatible with the bluray standard.
7890 It is a shorthand for setting "bluray-compat=1 force-cfr=1".
7891 b-bias (b-bias)
7893 Set the influence on how often B-frames are used.
7894 b-pyramid (b-pyramid)
7896 Set method for keeping of some B-frames as references. Possible
7897 values:
7898 none (none)
7900 Disabled.
7901 strict (strict)
7903 Strictly hierarchical pyramid.
7904 normal (normal)
7906 Non-strict (not Blu-ray compatible).
7907 mixed-refs
7909 Enable the use of one reference per partition, as opposed to one
7910 reference per macroblock when set to 1. When set to 0, it has the
7911 same effect as x264's --no-mixed-refs option.
7912 8x8dct
7914 Enable adaptive spatial transform (high profile 8x8 transform) when
7915 set to 1. When set to 0, it has the same effect as x264's
7916 --no-8x8dct option.
7917 fast-pskip
7919 Enable early SKIP detection on P-frames when set to 1. When set to
7920 0, it has the same effect as x264's --no-fast-pskip option.
7921 aud (aud)
7923 Enable use of access unit delimiters when set to 1.
7924 mbtree
7926 Enable use macroblock tree ratecontrol when set to 1. When set to
7927 0, it has the same effect as x264's --no-mbtree option.
7928 deblock (deblock)
7930 Set loop filter parameters, in alpha:beta form.
7931 cplxblur (cplxblur)
7933 Set fluctuations reduction in QP (before curve compression).
7934 partitions (partitions)
7936 Set partitions to consider as a comma-separated list of. Possible
7937 values in the list:
7938 p8x8
7940 8x8 P-frame partition.
7941 p4x4
7943 4x4 P-frame partition.
7944 b8x8
7946 4x4 B-frame partition.
7947 i8x8
7949 8x8 I-frame partition.
7950 i4x4
7952 4x4 I-frame partition. (Enabling p4x4 requires p8x8 to be
7953 enabled. Enabling i8x8 requires adaptive spatial transform
7954 (8x8dct option) to be enabled.)
7955 none (none)
7957 Do not consider any partitions.
7958 all (all)
7960 Consider every partition.
7961 direct-pred (direct)
7963 Set direct MV prediction mode. Possible values:
7964 none (none)
7966 Disable MV prediction.
7967 spatial (spatial)
7969 Enable spatial predicting.
7970 temporal (temporal)
7972 Enable temporal predicting.
7973 auto (auto)
7975 Automatically decided.
7976 slice-max-size (slice-max-size)
7978 Set the limit of the size of each slice in bytes. If not specified
7979 but RTP payload size (ps) is specified, that is used.
7980 stats (stats)
7982 Set the file name for multi-pass stats.
7983 nal-hrd (nal-hrd)
7985 Set signal HRD information (requires vbv-bufsize to be set).
7986 Possible values:
7987 none (none)
7989 Disable HRD information signaling.
7990 vbr (vbr)
7992 Variable bit rate.
7993 cbr (cbr)
7995 Constant bit rate (not allowed in MP4 container).
7996 x264opts (N.A.)
7998 Set any x264 option, see x264 --fullhelp for a list.
7999 Argument is a list of key=value couples separated by ":". In filter
8001 and psy-rd options that use ":" as a separator themselves, use ","
8002 instead. They accept it as well since long ago but this is kept
8003 undocumented for some reason.
8004 For example to specify libx264 encoding options with ffmpeg:
8006 ffmpeg -i foo.mpg -c:v libx264 -x264opts keyint=123:min-keyint=20 -an out.mkv
8008 a53cc boolean
8010 Import closed captions (which must be ATSC compatible format) into
8011 output. Only the mpeg2 and h264 decoders provide these. Default is
8012 1 (on).
8013 udu_sei boolean
8015 Import user data unregistered SEI if available into output. Default
8016 is 0 (off).
8017 x264-params (N.A.)
8019 Override the x264 configuration using a :-separated list of
8020 key=value parameters.
8021 This option is functionally the same as the x264opts, but is
8023 duplicated for compatibility with the Libav fork.
8024 For example to specify libx264 encoding options with ffmpeg:
8026 ffmpeg -i INPUT -c:v libx264 -x264-params level=30:bframes=0:weightp=0:\
8028 cabac=0:ref=1:vbv-maxrate=768:vbv-bufsize=2000:analyse=all:me=umh:\
8029 no-fast-pskip=1:subq=6:8x8dct=0:trellis=0 OUTPUT
8030 Encoding ffpresets for common usages are provided so they can be used
8032 with the general presets system (e.g. passing the pre option).
8033 libx265
8035 x265 H.265/HEVC encoder wrapper.
8036 This encoder requires the presence of the libx265 headers and library
8038 during configuration. You need to explicitly configure the build with
8039 --enable-libx265.
8040 Options
8042 b Sets target video bitrate.
8044 bf
8046 g Set the GOP size.
8047 keyint_min
8049 Minimum GOP size.
8050 refs
8052 Number of reference frames each P-frame can use. The range is from
8053 1-16.
8054 preset
8056 Set the x265 preset.
8057 tune
8059 Set the x265 tune parameter.
8060 profile
8062 Set profile restrictions.
8063 crf Set the quality for constant quality mode.
8065 qp Set constant quantization rate control method parameter.
8067 qmin
8069 Minimum quantizer scale.
8070 qmax
8072 Maximum quantizer scale.
8073 qdiff
8075 Maximum difference between quantizer scales.
8076 qblur
8078 Quantizer curve blur
8079 qcomp
8081 Quantizer curve compression factor
8082 i_qfactor
8084 b_qfactor
8085 forced-idr
8086 Normally, when forcing a I-frame type, the encoder can select any
8087 type of I-frame. This option forces it to choose an IDR-frame.
8088 udu_sei boolean
8090 Import user data unregistered SEI if available into output. Default
8091 is 0 (off).
8092 x265-params
8094 Set x265 options using a list of key=value couples separated by
8095 ":". See x265 --help for a list of options.
8096 For example to specify libx265 encoding options with -x265-params:
8098 ffmpeg -i input -c:v libx265 -x265-params crf=26:psy-rd=1 output.mp4
8100 libxavs2
8102 xavs2 AVS2-P2/IEEE1857.4 encoder wrapper.
8103 This encoder requires the presence of the libxavs2 headers and library
8105 during configuration. You need to explicitly configure the build with
8106 --enable-libxavs2.
8107 The following standard libavcodec options are used:
8109 • b / bit_rate
8111 • g / gop_size
8113 • bf / max_b_frames
8115 The encoder also has its own specific options:
8117 Options
8119 lcu_row_threads
8121 Set the number of parallel threads for rows from 1 to 8 (default
8122 5).
8123 initial_qp
8125 Set the xavs2 quantization parameter from 1 to 63 (default 34).
8126 This is used to set the initial qp for the first frame.
8127 qp Set the xavs2 quantization parameter from 1 to 63 (default 34).
8129 This is used to set the qp value under constant-QP mode.
8130 max_qp
8132 Set the max qp for rate control from 1 to 63 (default 55).
8133 min_qp
8135 Set the min qp for rate control from 1 to 63 (default 20).
8136 speed_level
8138 Set the Speed level from 0 to 9 (default 0). Higher is better but
8139 slower.
8140 log_level
8142 Set the log level from -1 to 3 (default 0). -1: none, 0: error, 1:
8143 warning, 2: info, 3: debug.
8144 xavs2-params
8146 Set xavs2 options using a list of key=value couples separated by
8147 ":".
8148 For example to specify libxavs2 encoding options with
8150 -xavs2-params:
8151 ffmpeg -i input -c:v libxavs2 -xavs2-params RdoqLevel=0 output.avs2
8153 libxvid
8155 Xvid MPEG-4 Part 2 encoder wrapper.
8156 This encoder requires the presence of the libxvidcore headers and
8158 library during configuration. You need to explicitly configure the
8159 build with "--enable-libxvid --enable-gpl".
8160 The native "mpeg4" encoder supports the MPEG-4 Part 2 format, so users
8162 can encode to this format without this library.
8163 Options
8165 The following options are supported by the libxvid wrapper. Some of the
8167 following options are listed but are not documented, and correspond to
8168 shared codec options. See the Codec Options chapter for their
8169 documentation. The other shared options which are not listed have no
8170 effect for the libxvid encoder.
8171 b
8173 g
8174 qmin
8175 qmax
8176 mpeg_quant
8177 threads
8178 bf
8179 b_qfactor
8180 b_qoffset
8181 flags
8182 Set specific encoding flags. Possible values:
8183 mv4 Use four motion vector by macroblock.
8185 aic Enable high quality AC prediction.
8187 gray
8189 Only encode grayscale.
8190 gmc Enable the use of global motion compensation (GMC).
8192 qpel
8194 Enable quarter-pixel motion compensation.
8195 cgop
8197 Enable closed GOP.
8198 global_header
8200 Place global headers in extradata instead of every keyframe.
8201 trellis
8203 me_method
8204 Set motion estimation method. Possible values in decreasing order
8205 of speed and increasing order of quality:
8206 zero
8208 Use no motion estimation (default).
8209 phods
8211 x1
8212 log Enable advanced diamond zonal search for 16x16 blocks and half-
8213 pixel refinement for 16x16 blocks. x1 and log are aliases for
8214 phods.
8215 epzs
8217 Enable all of the things described above, plus advanced diamond
8218 zonal search for 8x8 blocks, half-pixel refinement for 8x8
8219 blocks, and motion estimation on chroma planes.
8220 full
8222 Enable all of the things described above, plus extended 16x16
8223 and 8x8 blocks search.
8224 mbd Set macroblock decision algorithm. Possible values in the
8226 increasing order of quality:
8227 simple
8229 Use macroblock comparing function algorithm (default).
8230 bits
8232 Enable rate distortion-based half pixel and quarter pixel
8233 refinement for 16x16 blocks.
8234 rd Enable all of the things described above, plus rate distortion-
8236 based half pixel and quarter pixel refinement for 8x8 blocks,
8237 and rate distortion-based search using square pattern.
8238 lumi_aq
8240 Enable lumi masking adaptive quantization when set to 1. Default is
8241 0 (disabled).
8242 variance_aq
8244 Enable variance adaptive quantization when set to 1. Default is 0
8245 (disabled).
8246 When combined with lumi_aq, the resulting quality will not be
8248 better than any of the two specified individually. In other words,
8249 the resulting quality will be the worse one of the two effects.
8250 ssim
8252 Set structural similarity (SSIM) displaying method. Possible
8253 values:
8254 off Disable displaying of SSIM information.
8256 avg Output average SSIM at the end of encoding to stdout. The
8258 format of showing the average SSIM is:
8259 Average SSIM: %f
8261 For users who are not familiar with C, %f means a float number,
8263 or a decimal (e.g. 0.939232).
8264 frame
8266 Output both per-frame SSIM data during encoding and average
8267 SSIM at the end of encoding to stdout. The format of per-frame
8268 information is:
8269 SSIM: avg: %1.3f min: %1.3f max: %1.3f
8271 For users who are not familiar with C, %1.3f means a float
8273 number rounded to 3 digits after the dot (e.g. 0.932).
8274 ssim_acc
8276 Set SSIM accuracy. Valid options are integers within the range of
8277 0-4, while 0 gives the most accurate result and 4 computes the
8278 fastest.
8279 MediaFoundation
8281 This provides wrappers to encoders (both audio and video) in the
8282 MediaFoundation framework. It can access both SW and HW encoders.
8283 Video encoders can take input in either of nv12 or yuv420p form (some
8284 encoders support both, some support only either - in practice, nv12 is
8285 the safer choice, especially among HW encoders).
8286 mpeg2
8288 MPEG-2 video encoder.
8289 Options
8291 profile
8293 Select the mpeg2 profile to encode:
8294 422
8296 high
8297 ss Spatially Scalable
8298 snr SNR Scalable
8300 main
8302 simple
8303 level
8304 Select the mpeg2 level to encode:
8305 high
8307 high1440
8308 main
8309 low
8310 seq_disp_ext integer
8311 Specifies if the encoder should write a sequence_display_extension
8312 to the output.
8313 -1
8315 auto
8316 Decide automatically to write it or not (this is the default)
8317 by checking if the data to be written is different from the
8318 default or unspecified values.
8319 0
8321 never
8322 Never write it.
8323 1
8325 always
8326 Always write it.
8327 video_format integer
8329 Specifies the video_format written into the sequence display
8330 extension indicating the source of the video pictures. The default
8331 is unspecified, can be component, pal, ntsc, secam or mac. For
8332 maximum compatibility, use component.
8333 a53cc boolean
8335 Import closed captions (which must be ATSC compatible format) into
8336 output. Default is 1 (on).
8337 png
8339 PNG image encoder.
8340 Private options
8342 dpi integer
8344 Set physical density of pixels, in dots per inch, unset by default
8345 dpm integer
8347 Set physical density of pixels, in dots per meter, unset by default
8348 ProRes
8350 Apple ProRes encoder.
8351 FFmpeg contains 2 ProRes encoders, the prores-aw and prores-ks encoder.
8353 The used encoder can be chosen with the "-vcodec" option.
8354 Private Options for prores-ks
8356 profile integer
8358 Select the ProRes profile to encode
8359 proxy
8361 lt
8362 standard
8363 hq
8364 4444
8365 4444xq
8366 quant_mat integer
8367 Select quantization matrix.
8368 auto
8370 default
8371 proxy
8372 lt
8373 standard
8374 hq
8375 If set to auto, the matrix matching the profile will be picked. If
8377 not set, the matrix providing the highest quality, default, will be
8378 picked.
8379 bits_per_mb integer
8381 How many bits to allot for coding one macroblock. Different
8382 profiles use between 200 and 2400 bits per macroblock, the maximum
8383 is 8000.
8384 mbs_per_slice integer
8386 Number of macroblocks in each slice (1-8); the default value (8)
8387 should be good in almost all situations.
8388 vendor string
8390 Override the 4-byte vendor ID. A custom vendor ID like apl0 would
8391 claim the stream was produced by the Apple encoder.
8392 alpha_bits integer
8394 Specify number of bits for alpha component. Possible values are 0,
8395 8 and 16. Use 0 to disable alpha plane coding.
8396 Speed considerations
8398 In the default mode of operation the encoder has to honor frame
8400 constraints (i.e. not produce frames with size bigger than requested)
8401 while still making output picture as good as possible. A frame
8402 containing a lot of small details is harder to compress and the encoder
8403 would spend more time searching for appropriate quantizers for each
8404 slice.
8405 Setting a higher bits_per_mb limit will improve the speed.
8407 For the fastest encoding speed set the qscale parameter (4 is the
8409 recommended value) and do not set a size constraint.
8410 QSV Encoders
8412 The family of Intel QuickSync Video encoders (MPEG-2, H.264, HEVC,
8413 JPEG/MJPEG and VP9)
8414 Ratecontrol Method
8416 The ratecontrol method is selected as follows:
8418 • When global_quality is specified, a quality-based mode is used.
8420 Specifically this means either
8421 - CQP - constant quantizer scale, when the qscale codec flag is
8423 also set (the -qscale ffmpeg option).
8424 - LA_ICQ - intelligent constant quality with lookahead, when the
8426 look_ahead option is also set.
8427 - ICQ -- intelligent constant quality otherwise. For the ICQ
8429 modes, global quality range is 1 to 51, with 1 being the best
8430 quality.
8431 • Otherwise, a bitrate-based mode is used. For all of those, you
8433 should specify at least the desired average bitrate with the b
8434 option.
8435 - LA - VBR with lookahead, when the look_ahead option is
8437 specified.
8438 - VCM - video conferencing mode, when the vcm option is set.
8440 - CBR - constant bitrate, when maxrate is specified and equal to
8442 the average bitrate.
8443 - VBR - variable bitrate, when maxrate is specified, but is
8445 higher than the average bitrate.
8446 - AVBR - average VBR mode, when maxrate is not specified, both
8448 avbr_accuracy and avbr_convergence are set to non-zero. This
8449 mode is available for H264 and HEVC on Windows.
8450 Note that depending on your system, a different mode than the one you
8452 specified may be selected by the encoder. Set the verbosity level to
8453 verbose or higher to see the actual settings used by the QSV runtime.
8454 Global Options -> MSDK Options
8456 Additional libavcodec global options are mapped to MSDK options as
8458 follows:
8459 • g/gop_size -> GopPicSize
8461 • bf/max_b_frames+1 -> GopRefDist
8463 • rc_init_occupancy/rc_initial_buffer_occupancy -> InitialDelayInKB
8465 • slices -> NumSlice
8467 • refs -> NumRefFrame
8469 • b_strategy/b_frame_strategy -> BRefType
8471 • cgop/CLOSED_GOP codec flag -> GopOptFlag
8473 • For the CQP mode, the i_qfactor/i_qoffset and b_qfactor/b_qoffset
8475 set the difference between QPP and QPI, and QPP and QPB
8476 respectively.
8477 • Setting the coder option to the value vlc will make the H.264
8479 encoder use CAVLC instead of CABAC.
8480 Common Options
8482 Following options are used by all qsv encoders.
8484 async_depth
8486 Specifies how many asynchronous operations an application performs
8487 before the application explicitly synchronizes the result. If zero,
8488 the value is not specified.
8489 preset
8491 This option itemizes a range of choices from veryfast (best speed)
8492 to veryslow (best quality).
8493 veryfast
8495 faster
8496 fast
8497 medium
8498 slow
8499 slower
8500 veryslow
8501 forced_idr
8502 Forcing I frames as IDR frames.
8503 low_power
8505 For encoders set this flag to ON to reduce power consumption and
8506 GPU usage.
8507 Runtime Options
8509 Following options can be used durning qsv encoding.
8511 global_quality
8513 i_quant_factor
8514 i_quant_offset
8515 b_quant_factor
8516 b_quant_offset
8517 Supported in h264_qsv and hevc_qsv. Change these value to reset
8518 qsv codec's qp configuration.
8519 max_frame_size
8521 Supported in h264_qsv and hevc_qsv. Change this value to reset qsv
8522 codec's MaxFrameSize configuration.
8523 gop_size
8525 Change this value to reset qsv codec's gop configuration.
8526 int_ref_type
8528 int_ref_cycle_size
8529 int_ref_qp_delta
8530 int_ref_cycle_dist
8531 Supported in h264_qsv and hevc_qsv. Change these value to reset
8532 qsv codec's Intra Refresh configuration.
8533 qmax
8535 qmin
8536 max_qp_i
8537 min_qp_i
8538 max_qp_p
8539 min_qp_p
8540 max_qp_b
8541 min_qp_b
8542 Supported in h264_qsv. Change these value to reset qsv codec's
8543 max/min qp configuration.
8544 low_delay_brc
8546 Supported in h264_qsv and hevc_qsv. Change this value to reset qsv
8547 codec's low_delay_brc configuration.
8548 framerate
8550 Change this value to reset qsv codec's framerate configuration.
8551 bit_rate
8553 rc_buffer_size
8554 rc_initial_buffer_occupancy
8555 rc_max_rate
8556 Change these value to reset qsv codec's bitrate control
8557 configuration.
8558 pic_timing_sei
8560 Supported in h264_qsv and hevc_qsv. Change this value to reset qsv
8561 codec's pic_timing_sei configuration.
8562 H264 options
8564 These options are used by h264_qsv
8566 extbrc
8568 Extended bitrate control.
8569 recovery_point_sei
8571 Set this flag to insert the recovery point SEI message at the
8572 beginning of every intra refresh cycle.
8573 rdo Enable rate distortion optimization.
8575 max_frame_size
8577 Maximum encoded frame size in bytes.
8578 max_frame_size_i
8580 Maximum encoded frame size for I frames in bytes. If this value is
8581 set as larger than zero, then for I frames the value set by
8582 max_frame_size is ignored.
8583 max_frame_size_p
8585 Maximum encoded frame size for P frames in bytes. If this value is
8586 set as larger than zero, then for P frames the value set by
8587 max_frame_size is ignored.
8588 max_slice_size
8590 Maximum encoded slice size in bytes.
8591 bitrate_limit
8593 Toggle bitrate limitations. Modifies bitrate to be in the range
8594 imposed by the QSV encoder. Setting this flag off may lead to
8595 violation of HRD conformance. Mind that specifying bitrate below
8596 the QSV encoder range might significantly affect quality. If on
8597 this option takes effect in non CQP modes: if bitrate is not in the
8598 range imposed by the QSV encoder, it will be changed to be in the
8599 range.
8600 mbbrc
8602 Setting this flag enables macroblock level bitrate control that
8603 generally improves subjective visual quality. Enabling this flag
8604 may have negative impact on performance and objective visual
8605 quality metric.
8606 low_delay_brc
8608 Setting this flag turns on or off LowDelayBRC feautre in qsv
8609 plugin, which provides more accurate bitrate control to minimize
8610 the variance of bitstream size frame by frame. Value: -1-default
8611 0-off 1-on
8612 adaptive_i
8614 This flag controls insertion of I frames by the QSV encoder. Turn
8615 ON this flag to allow changing of frame type from P and B to I.
8616 adaptive_b
8618 This flag controls changing of frame type from B to P.
8619 p_strategy
8621 Enable P-pyramid: 0-default 1-simple 2-pyramid(bf need to be set to
8622 0).
8623 b_strategy
8625 This option controls usage of B frames as reference.
8626 dblk_idc
8628 This option disable deblocking. It has value in range 0~2.
8629 cavlc
8631 If set, CAVLC is used; if unset, CABAC is used for encoding.
8632 vcm Video conferencing mode, please see ratecontrol method.
8634 idr_interval
8636 Distance (in I-frames) between IDR frames.
8637 pic_timing_sei
8639 Insert picture timing SEI with pic_struct_syntax element.
8640 single_sei_nal_unit
8642 Put all the SEI messages into one NALU.
8643 max_dec_frame_buffering
8645 Maximum number of frames buffered in the DPB.
8646 look_ahead
8648 Use VBR algorithm with look ahead.
8649 look_ahead_depth
8651 Depth of look ahead in number frames.
8652 look_ahead_downsampling
8654 Downscaling factor for the frames saved for the lookahead analysis.
8655 unknown
8657 auto
8658 off
8659 2x
8660 4x
8661 int_ref_type
8662 Specifies intra refresh type. The major goal of intra refresh is
8663 improvement of error resilience without significant impact on
8664 encoded bitstream size caused by I frames. The SDK encoder achieves
8665 this by encoding part of each frame in refresh cycle using intra
8666 MBs. none means no refresh. vertical means vertical refresh, by
8667 column of MBs. horizontal means horizontal refresh, by rows of MBs.
8668 slice means horizontal refresh by slices without overlapping. In
8669 case of slice, in_ref_cycle_size is ignored. To enable intra
8670 refresh, B frame should be set to 0.
8671 int_ref_cycle_size
8673 Specifies number of pictures within refresh cycle starting from 2.
8674 0 and 1 are invalid values.
8675 int_ref_qp_delta
8677 Specifies QP difference for inserted intra MBs. This is signed
8678 value in [-51, 51] range if target encoding bit-depth for luma
8679 samples is 8 and this range is [-63, 63] for 10 bit-depth or [-75,
8680 75] for 12 bit-depth respectively.
8681 int_ref_cycle_dist
8683 Distance between the beginnings of the intra-refresh cycles in
8684 frames.
8685 profile
8687 unknown
8688 baseline
8689 main
8690 high
8691 a53cc
8692 Use A53 Closed Captions (if available).
8693 aud Insert the Access Unit Delimiter NAL.
8695 mfmode
8697 Multi-Frame Mode.
8698 off
8700 auto
8701 repeat_pps
8702 Repeat pps for every frame.
8703 max_qp_i
8705 Maximum video quantizer scale for I frame.
8706 min_qp_i
8708 Minimum video quantizer scale for I frame.
8709 max_qp_p
8711 Maximum video quantizer scale for P frame.
8712 min_qp_p
8714 Minimum video quantizer scale for P frame.
8715 max_qp_b
8717 Maximum video quantizer scale for B frame.
8718 min_qp_b
8720 Minimum video quantizer scale for B frame.
8721 scenario
8723 Provides a hint to encoder about the scenario for the encoding
8724 session.
8725 unknown
8727 displayremoting
8728 videoconference
8729 archive
8730 livestreaming
8731 cameracapture
8732 videosurveillance
8733 gamestreaming
8734 remotegaming
8735 avbr_accuracy
8736 Accuracy of the AVBR ratecontrol (unit of tenth of percent).
8737 avbr_convergence
8739 Convergence of the AVBR ratecontrol (unit of 100 frames)
8740 The parameters avbr_accuracy and avbr_convergence are for the
8742 average variable bitrate control (AVBR) algorithm. The algorithm
8743 focuses on overall encoding quality while meeting the specified
8744 bitrate, target_bitrate, within the accuracy range avbr_accuracy,
8745 after a avbr_Convergence period. This method does not follow HRD
8746 and the instant bitrate is not capped or padded.
8747 skip_frame
8749 Use per-frame metadata "qsv_skip_frame" to skip frame when
8750 encoding. This option defines the usage of this metadata.
8751 no_skip
8753 Frame skipping is disabled.
8754 insert_dummy
8756 Encoder inserts into bitstream frame where all macroblocks are
8757 encoded as skipped.
8758 insert_nothing
8760 Similar to insert_dummy, but encoder inserts nothing into
8761 bitstream. The skipped frames are still used in brc. For
8762 example, gop still include skipped frames, and the frames after
8763 skipped frames will be larger in size.
8764 brc_only
8766 skip_frame metadata indicates the number of missed frames
8767 before the current frame.
8768 HEVC Options
8770 These options are used by hevc_qsv
8772 extbrc
8774 Extended bitrate control.
8775 recovery_point_sei
8777 Set this flag to insert the recovery point SEI message at the
8778 beginning of every intra refresh cycle.
8779 rdo Enable rate distortion optimization.
8781 max_frame_size
8783 Maximum encoded frame size in bytes.
8784 max_frame_size_i
8786 Maximum encoded frame size for I frames in bytes. If this value is
8787 set as larger than zero, then for I frames the value set by
8788 max_frame_size is ignored.
8789 max_frame_size_p
8791 Maximum encoded frame size for P frames in bytes. If this value is
8792 set as larger than zero, then for P frames the value set by
8793 max_frame_size is ignored.
8794 max_slice_size
8796 Maximum encoded slice size in bytes.
8797 mbbrc
8799 Setting this flag enables macroblock level bitrate control that
8800 generally improves subjective visual quality. Enabling this flag
8801 may have negative impact on performance and objective visual
8802 quality metric.
8803 low_delay_brc
8805 Setting this flag turns on or off LowDelayBRC feautre in qsv
8806 plugin, which provides more accurate bitrate control to minimize
8807 the variance of bitstream size frame by frame. Value: -1-default
8808 0-off 1-on
8809 adaptive_i
8811 This flag controls insertion of I frames by the QSV encoder. Turn
8812 ON this flag to allow changing of frame type from P and B to I.
8813 adaptive_b
8815 This flag controls changing of frame type from B to P.
8816 p_strategy
8818 Enable P-pyramid: 0-default 1-simple 2-pyramid(bf need to be set to
8819 0).
8820 b_strategy
8822 This option controls usage of B frames as reference.
8823 dblk_idc
8825 This option disable deblocking. It has value in range 0~2.
8826 idr_interval
8828 Distance (in I-frames) between IDR frames.
8829 begin_only
8831 Output an IDR-frame only at the beginning of the stream.
8832 load_plugin
8834 A user plugin to load in an internal session.
8835 none
8837 hevc_sw
8838 hevc_hw
8839 load_plugins
8840 A :-separate list of hexadecimal plugin UIDs to load in an internal
8841 session.
8842 look_ahead_depth
8844 Depth of look ahead in number frames, available when extbrc option
8845 is enabled.
8846 profile
8848 Set the encoding profile (scc requires libmfx >= 1.32).
8849 unknown
8851 main
8852 main10
8853 mainsp
8854 rext
8855 scc
8856 tier
8857 Set the encoding tier (only level >= 4 can support high tier).
8858 This option only takes effect when the level option is specified.
8859 main
8861 high
8862 gpb 1: GPB (generalized P/B frame)
8863 0: regular P frame.
8865 tile_cols
8867 Number of columns for tiled encoding.
8868 tile_rows
8870 Number of rows for tiled encoding.
8871 aud Insert the Access Unit Delimiter NAL.
8873 pic_timing_sei
8875 Insert picture timing SEI with pic_struct_syntax element.
8876 transform_skip
8878 Turn this option ON to enable transformskip. It is supported on
8879 platform equal or newer than ICL.
8880 int_ref_type
8882 Specifies intra refresh type. The major goal of intra refresh is
8883 improvement of error resilience without significant impact on
8884 encoded bitstream size caused by I frames. The SDK encoder achieves
8885 this by encoding part of each frame in refresh cycle using intra
8886 MBs. none means no refresh. vertical means vertical refresh, by
8887 column of MBs. horizontal means horizontal refresh, by rows of MBs.
8888 slice means horizontal refresh by slices without overlapping. In
8889 case of slice, in_ref_cycle_size is ignored. To enable intra
8890 refresh, B frame should be set to 0.
8891 int_ref_cycle_size
8893 Specifies number of pictures within refresh cycle starting from 2.
8894 0 and 1 are invalid values.
8895 int_ref_qp_delta
8897 Specifies QP difference for inserted intra MBs. This is signed
8898 value in [-51, 51] range if target encoding bit-depth for luma
8899 samples is 8 and this range is [-63, 63] for 10 bit-depth or [-75,
8900 75] for 12 bit-depth respectively.
8901 int_ref_cycle_dist
8903 Distance between the beginnings of the intra-refresh cycles in
8904 frames.
8905 max_qp_i
8907 Maximum video quantizer scale for I frame.
8908 min_qp_i
8910 Minimum video quantizer scale for I frame.
8911 max_qp_p
8913 Maximum video quantizer scale for P frame.
8914 min_qp_p
8916 Minimum video quantizer scale for P frame.
8917 max_qp_b
8919 Maximum video quantizer scale for B frame.
8920 min_qp_b
8922 Minimum video quantizer scale for B frame.
8923 scenario
8925 Provides a hint to encoder about the scenario for the encoding
8926 session.
8927 unknown
8929 displayremoting
8930 videoconference
8931 archive
8932 livestreaming
8933 cameracapture
8934 videosurveillance
8935 gamestreaming
8936 remotegaming
8937 avbr_accuracy
8938 Accuracy of the AVBR ratecontrol (unit of tenth of percent).
8939 avbr_convergence
8941 Convergence of the AVBR ratecontrol (unit of 100 frames)
8942 The parameters avbr_accuracy and avbr_convergence are for the
8944 average variable bitrate control (AVBR) algorithm. The algorithm
8945 focuses on overall encoding quality while meeting the specified
8946 bitrate, target_bitrate, within the accuracy range avbr_accuracy,
8947 after a avbr_Convergence period. This method does not follow HRD
8948 and the instant bitrate is not capped or padded.
8949 skip_frame
8951 Use per-frame metadata "qsv_skip_frame" to skip frame when
8952 encoding. This option defines the usage of this metadata.
8953 no_skip
8955 Frame skipping is disabled.
8956 insert_dummy
8958 Encoder inserts into bitstream frame where all macroblocks are
8959 encoded as skipped.
8960 insert_nothing
8962 Similar to insert_dummy, but encoder inserts nothing into
8963 bitstream. The skipped frames are still used in brc. For
8964 example, gop still include skipped frames, and the frames after
8965 skipped frames will be larger in size.
8966 brc_only
8968 skip_frame metadata indicates the number of missed frames
8969 before the current frame.
8970 MPEG2 Options
8972 These options are used by mpeg2_qsv
8974 profile
8976 unknown
8977 simple
8978 main
8979 high
8980 VP9 Options
8982 These options are used by vp9_qsv
8984 profile
8986 unknown
8987 profile0
8988 profile1
8989 profile2
8990 profile3
8991 tile_cols
8992 Number of columns for tiled encoding (requires libmfx >= 1.29).
8993 tile_rows
8995 Number of rows for tiled encoding (requires libmfx >= 1.29).
8996 AV1 Options
8998 These options are used by av1_qsv (requires libvpl).
9000 profile
9002 unknown
9003 main
9004 tile_cols
9005 Number of columns for tiled encoding.
9006 tile_rows
9008 Number of rows for tiled encoding.
9009 adaptive_i
9011 This flag controls insertion of I frames by the QSV encoder. Turn
9012 ON this flag to allow changing of frame type from P and B to I.
9013 adaptive_b
9015 This flag controls changing of frame type from B to P.
9016 b_strategy
9018 This option controls usage of B frames as reference.
9019 extbrc
9021 Extended bitrate control.
9022 look_ahead_depth
9024 Depth of look ahead in number frames, available when extbrc option
9025 is enabled.
9026 low_delay_brc
9028 Setting this flag turns on or off LowDelayBRC feautre in qsv
9029 plugin, which provides more accurate bitrate control to minimize
9030 the variance of bitstream size frame by frame. Value: -1-default
9031 0-off 1-on
9032 max_frame_size
9034 Set the allowed max size in bytes for each frame. If the frame size
9035 exceeds the limitation, encoder will adjust the QP value to control
9036 the frame size. Invalid in CQP rate control mode.
9037 snow
9039 Options
9040 iterative_dia_size
9042 dia size for the iterative motion estimation
9043 VAAPI encoders
9045 Wrappers for hardware encoders accessible via VAAPI.
9046 These encoders only accept input in VAAPI hardware surfaces. If you
9048 have input in software frames, use the hwupload filter to upload them
9049 to the GPU.
9050 The following standard libavcodec options are used:
9052 • g / gop_size
9054 • bf / max_b_frames
9056 • profile
9058 If not set, this will be determined automatically from the format
9060 of the input frames and the profiles supported by the driver.
9061 • level
9063 • b / bit_rate
9065 • maxrate / rc_max_rate
9067 • bufsize / rc_buffer_size
9069 • rc_init_occupancy / rc_initial_buffer_occupancy
9071 • compression_level
9073 Speed / quality tradeoff: higher values are faster / worse quality.
9075 • q / global_quality
9077 Size / quality tradeoff: higher values are smaller / worse quality.
9079 • qmin
9081 • qmax
9083 • i_qfactor / i_quant_factor
9085 • i_qoffset / i_quant_offset
9087 • b_qfactor / b_quant_factor
9089 • b_qoffset / b_quant_offset
9091 • slices
9093 All encoders support the following options:
9095 low_power
9097 Some drivers/platforms offer a second encoder for some codecs
9098 intended to use less power than the default encoder; setting this
9099 option will attempt to use that encoder. Note that it may support
9100 a reduced feature set, so some other options may not be available
9101 in this mode.
9102 idr_interval
9104 Set the number of normal intra frames between full-refresh (IDR)
9105 frames in open-GOP mode. The intra frames are still IRAPs, but
9106 will not include global headers and may have non-decodable leading
9107 pictures.
9108 b_depth
9110 Set the B-frame reference depth. When set to one (the default),
9111 all B-frames will refer only to P- or I-frames. When set to
9112 greater values multiple layers of B-frames will be present, frames
9113 in each layer only referring to frames in higher layers.
9114 async_depth
9116 Maximum processing parallelism. Increase this to improve single
9117 channel performance. This option doesn't work if driver doesn't
9118 implement vaSyncBuffer function. Please make sure there are enough
9119 hw_frames allocated if a large number of async_depth is used.
9120 max_frame_size
9122 Set the allowed max size in bytes for each frame. If the frame size
9123 exceeds the limitation, encoder will adjust the QP value to control
9124 the frame size. Invalid in CQP rate control mode.
9125 rc_mode
9127 Set the rate control mode to use. A given driver may only support
9128 a subset of modes.
9129 Possible modes:
9131 auto
9133 Choose the mode automatically based on driver support and the
9134 other options. This is the default.
9135 CQP Constant-quality.
9137 CBR Constant-bitrate.
9139 VBR Variable-bitrate.
9141 ICQ Intelligent constant-quality.
9143 QVBR
9145 Quality-defined variable-bitrate.
9146 AVBR
9148 Average variable bitrate.
9149 Each encoder also has its own specific options:
9151 av1_vaapi
9153 profile sets the value of seq_profile. tier sets the value of
9154 seq_tier. level sets the value of seq_level_idx.
9155 tiles
9157 Set the number of tiles to encode the input video with, as
9158 columns x rows. (default is auto, which means use minimal tile
9159 column/row number).
9160 tile_groups
9162 Set tile groups number. All the tiles will be distributed as
9163 evenly as possible to each tile group. (default is 1).
9164 h264_vaapi
9166 profile sets the value of profile_idc and the
9167 constraint_set*_flags. level sets the value of level_idc.
9168 coder
9170 Set entropy encoder (default is cabac). Possible values:
9171 ac
9173 cabac
9174 Use CABAC.
9175 vlc
9177 cavlc
9178 Use CAVLC.
9179 aud Include access unit delimiters in the stream (not included by
9181 default).
9182 sei Set SEI message types to include. Some combination of the
9184 following values:
9185 identifier
9187 Include a user_data_unregistered message containing
9188 information about the encoder.
9189 timing
9191 Include picture timing parameters (buffering_period and
9192 pic_timing messages).
9193 recovery_point
9195 Include recovery points where appropriate (recovery_point
9196 messages).
9197 hevc_vaapi
9199 profile and level set the values of general_profile_idc and
9200 general_level_idc respectively.
9201 aud Include access unit delimiters in the stream (not included by
9203 default).
9204 tier
9206 Set general_tier_flag. This may affect the level chosen for
9207 the stream if it is not explicitly specified.
9208 sei Set SEI message types to include. Some combination of the
9210 following values:
9211 hdr Include HDR metadata if the input frames have it
9213 (mastering_display_colour_volume and content_light_level
9214 messages).
9215 tiles
9217 Set the number of tiles to encode the input video with, as
9218 columns x rows. Larger numbers allow greater parallelism in
9219 both encoding and decoding, but may decrease coding efficiency.
9220 mjpeg_vaapi
9222 Only baseline DCT encoding is supported. The encoder always uses
9223 the standard quantisation and huffman tables - global_quality
9224 scales the standard quantisation table (range 1-100).
9225 For YUV, 4:2:0, 4:2:2 and 4:4:4 subsampling modes are supported.
9227 RGB is also supported, and will create an RGB JPEG.
9228 jfif
9230 Include JFIF header in each frame (not included by default).
9231 huffman
9233 Include standard huffman tables (on by default). Turning this
9234 off will save a few hundred bytes in each output frame, but may
9235 lose compatibility with some JPEG decoders which don't fully
9236 handle MJPEG.
9237 mpeg2_vaapi
9239 profile and level set the value of profile_and_level_indication.
9240 vp8_vaapi
9242 B-frames are not supported.
9243 global_quality sets the q_idx used for non-key frames (range
9245 0-127).
9246 loop_filter_level
9248 loop_filter_sharpness
9249 Manually set the loop filter parameters.
9250 vp9_vaapi
9252 global_quality sets the q_idx used for P-frames (range 0-255).
9253 loop_filter_level
9255 loop_filter_sharpness
9256 Manually set the loop filter parameters.
9257 B-frames are supported, but the output stream is always in encode
9259 order rather than display order. If B-frames are enabled, it may
9260 be necessary to use the vp9_raw_reorder bitstream filter to modify
9261 the output stream to display frames in the correct order.
9262 Only normal frames are produced - the vp9_superframe bitstream
9264 filter may be required to produce a stream usable with all
9265 decoders.
9266 vbn
9268 Vizrt Binary Image encoder.
9269 This format is used by the broadcast vendor Vizrt for quick texture
9271 streaming. Advanced features of the format such as LZW compression of
9272 texture data or generation of mipmaps are not supported.
9273 Options
9275 format string
9277 Sets the texture compression used by the VBN file. Can be dxt1,
9278 dxt5 or raw. Default is dxt5.
9279 vc2
9281 SMPTE VC-2 (previously BBC Dirac Pro). This codec was primarily aimed
9282 at professional broadcasting but since it supports yuv420, yuv422 and
9283 yuv444 at 8 (limited range or full range), 10 or 12 bits, this makes it
9284 suitable for other tasks which require low overhead and low compression
9285 (like screen recording).
9286 Options
9288 b Sets target video bitrate. Usually that's around 1:6 of the
9290 uncompressed video bitrate (e.g. for 1920x1080 50fps yuv422p10
9291 that's around 400Mbps). Higher values (close to the uncompressed
9292 bitrate) turn on lossless compression mode.
9293 field_order
9295 Enables field coding when set (e.g. to tt - top field first) for
9296 interlaced inputs. Should increase compression with interlaced
9297 content as it splits the fields and encodes each separately.
9298 wavelet_depth
9300 Sets the total amount of wavelet transforms to apply, between 1 and
9301 5 (default). Lower values reduce compression and quality. Less
9302 capable decoders may not be able to handle values of wavelet_depth
9303 over 3.
9304 wavelet_type
9306 Sets the transform type. Currently only 5_3 (LeGall) and 9_7
9307 (Deslauriers-Dubuc) are implemented, with 9_7 being the one with
9308 better compression and thus is the default.
9309 slice_width
9311 slice_height
9312 Sets the slice size for each slice. Larger values result in better
9313 compression. For compatibility with other more limited decoders
9314 use slice_width of 32 and slice_height of 8.
9315 tolerance
9317 Sets the undershoot tolerance of the rate control system in
9318 percent. This is to prevent an expensive search from being run.
9319 qm Sets the quantization matrix preset to use by default or when
9321 wavelet_depth is set to 5
9322 - default Uses the default quantization matrix from the
9324 specifications, extended with values for the fifth level. This
9325 provides a good balance between keeping detail and omitting
9326 artifacts.
9327 - flat Use a completely zeroed out quantization matrix. This
9329 increases PSNR but might reduce perception. Use in bogus
9330 benchmarks.
9331 - color Reduces detail but attempts to preserve color at
9333 extremely low bitrates.
9334
9336 dvdsub
9337 This codec encodes the bitmap subtitle format that is used in DVDs.
9338 Typically they are stored in VOBSUB file pairs (*.idx + *.sub), and
9339 they can also be used in Matroska files.
9340 Options
9342 palette
9344 Specify the global palette used by the bitmaps.
9345 The format for this option is a string containing 16 24-bits
9347 hexadecimal numbers (without 0x prefix) separated by commas, for
9348 example "0d00ee, ee450d, 101010, eaeaea, 0ce60b, ec14ed, ebff0b,
9349 0d617a, 7b7b7b, d1d1d1, 7b2a0e, 0d950c, 0f007b, cf0dec, cfa80c,
9350 7c127b".
9351 even_rows_fix
9353 When set to 1, enable a work-around that makes the number of pixel
9354 rows even in all subtitles. This fixes a problem with some players
9355 that cut off the bottom row if the number is odd. The work-around
9356 just adds a fully transparent row if needed. The overhead is low,
9357 typically one byte per subtitle on average.
9358 By default, this work-around is disabled.
9360
9362 When you configure your FFmpeg build, all the supported bitstream
9363 filters are enabled by default. You can list all available ones using
9364 the configure option "--list-bsfs".
9365 You can disable all the bitstream filters using the configure option
9367 "--disable-bsfs", and selectively enable any bitstream filter using the
9368 option "--enable-bsf=BSF", or you can disable a particular bitstream
9369 filter using the option "--disable-bsf=BSF".
9370 The option "-bsfs" of the ff* tools will display the list of all the
9372 supported bitstream filters included in your build.
9373 The ff* tools have a -bsf option applied per stream, taking a comma-
9375 separated list of filters, whose parameters follow the filter name
9376 after a '='.
9377 ffmpeg -i INPUT -c:v copy -bsf:v filter1[=opt1=str1:opt2=str2][,filter2] OUTPUT
9379 Below is a description of the currently available bitstream filters,
9381 with their parameters, if any.
9382 aac_adtstoasc
9384 Convert MPEG-2/4 AAC ADTS to an MPEG-4 Audio Specific Configuration
9385 bitstream.
9386 This filter creates an MPEG-4 AudioSpecificConfig from an MPEG-2/4 ADTS
9388 header and removes the ADTS header.
9389 This filter is required for example when copying an AAC stream from a
9391 raw ADTS AAC or an MPEG-TS container to MP4A-LATM, to an FLV file, or
9392 to MOV/MP4 files and related formats such as 3GP or M4A. Please note
9393 that it is auto-inserted for MP4A-LATM and MOV/MP4 and related formats.
9394 av1_metadata
9396 Modify metadata embedded in an AV1 stream.
9397 td Insert or remove temporal delimiter OBUs in all temporal units of
9399 the stream.
9400 insert
9402 Insert a TD at the beginning of every TU which does not already
9403 have one.
9404 remove
9406 Remove the TD from the beginning of every TU which has one.
9407 color_primaries
9409 transfer_characteristics
9410 matrix_coefficients
9411 Set the color description fields in the stream (see AV1 section
9412 6.4.2).
9413 color_range
9415 Set the color range in the stream (see AV1 section 6.4.2; note that
9416 this cannot be set for streams using BT.709 primaries, sRGB
9417 transfer characteristic and identity (RGB) matrix coefficients).
9418 tv Limited range.
9420 pc Full range.
9422 chroma_sample_position
9424 Set the chroma sample location in the stream (see AV1 section
9425 6.4.2). This can only be set for 4:2:0 streams.
9426 vertical
9428 Left position (matching the default in MPEG-2 and H.264).
9429 colocated
9431 Top-left position.
9432 tick_rate
9434 Set the tick rate (time_scale / num_units_in_display_tick) in the
9435 timing info in the sequence header.
9436 num_ticks_per_picture
9438 Set the number of ticks in each picture, to indicate that the
9439 stream has a fixed framerate. Ignored if tick_rate is not also
9440 set.
9441 delete_padding
9443 Deletes Padding OBUs.
9444 chomp
9446 Remove zero padding at the end of a packet.
9447 dca_core
9449 Extract the core from a DCA/DTS stream, dropping extensions such as
9450 DTS-HD.
9451 dump_extra
9453 Add extradata to the beginning of the filtered packets except when said
9454 packets already exactly begin with the extradata that is intended to be
9455 added.
9456 freq
9458 The additional argument specifies which packets should be filtered.
9459 It accepts the values:
9460 k
9462 keyframe
9463 add extradata to all key packets
9464 e
9466 all add extradata to all packets
9467 If not specified it is assumed k.
9469 For example the following ffmpeg command forces a global header (thus
9471 disabling individual packet headers) in the H.264 packets generated by
9472 the "libx264" encoder, but corrects them by adding the header stored in
9473 extradata to the key packets:
9474 ffmpeg -i INPUT -map 0 -flags:v +global_header -c:v libx264 -bsf:v dump_extra out.ts
9476 dv_error_marker
9478 Blocks in DV which are marked as damaged are replaced by blocks of the
9479 specified color.
9480 color
9482 The color to replace damaged blocks by
9483 sta A 16 bit mask which specifies which of the 16 possible error status
9485 values are to be replaced by colored blocks. 0xFFFE is the default
9486 which replaces all non 0 error status values.
9487 ok No error, no concealment
9489 err Error, No concealment
9491 res Reserved
9493 notok
9495 Error or concealment
9496 notres
9498 Not reserved
9499 Aa, Ba, Ca, Ab, Bb, Cb, A, B, C, a, b, erri, erru
9501 The specific error status code
9502 see page 44-46 or section 5.5 of
9504 <http://web.archive.org/web/20060927044735/http://www.smpte.org/smpte_store/standards/pdf/s314m.pdf>
9505 eac3_core
9507 Extract the core from a E-AC-3 stream, dropping extra channels.
9508 extract_extradata
9510 Extract the in-band extradata.
9511 Certain codecs allow the long-term headers (e.g. MPEG-2 sequence
9513 headers, or H.264/HEVC (VPS/)SPS/PPS) to be transmitted either "in-
9514 band" (i.e. as a part of the bitstream containing the coded frames) or
9515 "out of band" (e.g. on the container level). This latter form is called
9516 "extradata" in FFmpeg terminology.
9517 This bitstream filter detects the in-band headers and makes them
9519 available as extradata.
9520 remove
9522 When this option is enabled, the long-term headers are removed from
9523 the bitstream after extraction.
9524 filter_units
9526 Remove units with types in or not in a given set from the stream.
9527 pass_types
9529 List of unit types or ranges of unit types to pass through while
9530 removing all others. This is specified as a '|'-separated list of
9531 unit type values or ranges of values with '-'.
9532 remove_types
9534 Identical to pass_types, except the units in the given set removed
9535 and all others passed through.
9536 Extradata is unchanged by this transformation, but note that if the
9538 stream contains inline parameter sets then the output may be unusable
9539 if they are removed.
9540 For example, to remove all non-VCL NAL units from an H.264 stream:
9542 ffmpeg -i INPUT -c:v copy -bsf:v 'filter_units=pass_types=1-5' OUTPUT
9544 To remove all AUDs, SEI and filler from an H.265 stream:
9546 ffmpeg -i INPUT -c:v copy -bsf:v 'filter_units=remove_types=35|38-40' OUTPUT
9548 hapqa_extract
9550 Extract Rgb or Alpha part of an HAPQA file, without recompression, in
9551 order to create an HAPQ or an HAPAlphaOnly file.
9552 texture
9554 Specifies the texture to keep.
9555 color
9557 alpha
9558 Convert HAPQA to HAPQ
9560 ffmpeg -i hapqa_inputfile.mov -c copy -bsf:v hapqa_extract=texture=color -tag:v HapY -metadata:s:v:0 encoder="HAPQ" hapq_file.mov
9562 Convert HAPQA to HAPAlphaOnly
9564 ffmpeg -i hapqa_inputfile.mov -c copy -bsf:v hapqa_extract=texture=alpha -tag:v HapA -metadata:s:v:0 encoder="HAPAlpha Only" hapalphaonly_file.mov
9566 h264_metadata
9568 Modify metadata embedded in an H.264 stream.
9569 aud Insert or remove AUD NAL units in all access units of the stream.
9571 pass
9573 insert
9574 remove
9575 Default is pass.
9577 sample_aspect_ratio
9579 Set the sample aspect ratio of the stream in the VUI parameters.
9580 See H.264 table E-1.
9581 overscan_appropriate_flag
9583 Set whether the stream is suitable for display using overscan or
9584 not (see H.264 section E.2.1).
9585 video_format
9587 video_full_range_flag
9588 Set the video format in the stream (see H.264 section E.2.1 and
9589 table E-2).
9590 colour_primaries
9592 transfer_characteristics
9593 matrix_coefficients
9594 Set the colour description in the stream (see H.264 section E.2.1
9595 and tables E-3, E-4 and E-5).
9596 chroma_sample_loc_type
9598 Set the chroma sample location in the stream (see H.264 section
9599 E.2.1 and figure E-1).
9600 tick_rate
9602 Set the tick rate (time_scale / num_units_in_tick) in the VUI
9603 parameters. This is the smallest time unit representable in the
9604 stream, and in many cases represents the field rate of the stream
9605 (double the frame rate).
9606 fixed_frame_rate_flag
9608 Set whether the stream has fixed framerate - typically this
9609 indicates that the framerate is exactly half the tick rate, but the
9610 exact meaning is dependent on interlacing and the picture structure
9611 (see H.264 section E.2.1 and table E-6).
9612 zero_new_constraint_set_flags
9614 Zero constraint_set4_flag and constraint_set5_flag in the SPS.
9615 These bits were reserved in a previous version of the H.264 spec,
9616 and thus some hardware decoders require these to be zero. The
9617 result of zeroing this is still a valid bitstream.
9618 crop_left
9620 crop_right
9621 crop_top
9622 crop_bottom
9623 Set the frame cropping offsets in the SPS. These values will
9624 replace the current ones if the stream is already cropped.
9625 These fields are set in pixels. Note that some sizes may not be
9627 representable if the chroma is subsampled or the stream is
9628 interlaced (see H.264 section 7.4.2.1.1).
9629 sei_user_data
9631 Insert a string as SEI unregistered user data. The argument must
9632 be of the form UUID+string, where the UUID is as hex digits
9633 possibly separated by hyphens, and the string can be anything.
9634 For example, 086f3693-b7b3-4f2c-9653-21492feee5b8+hello will insert
9636 the string ``hello'' associated with the given UUID.
9637 delete_filler
9639 Deletes both filler NAL units and filler SEI messages.
9640 display_orientation
9642 Insert, extract or remove Display orientation SEI messages. See
9643 H.264 section D.1.27 and D.2.27 for syntax and semantics.
9644 pass
9646 insert
9647 remove
9648 extract
9649 Default is pass.
9651 Insert mode works in conjunction with "rotate" and "flip" options.
9653 Any pre-existing Display orientation messages will be removed in
9654 insert or remove mode. Extract mode attaches the display matrix to
9655 the packet as side data.
9656 rotate
9658 Set rotation in display orientation SEI (anticlockwise angle in
9659 degrees). Range is -360 to +360. Default is NaN.
9660 flip
9662 Set flip in display orientation SEI.
9663 horizontal
9665 vertical
9666 Default is unset.
9668 level
9670 Set the level in the SPS. Refer to H.264 section A.3 and tables
9671 A-1 to A-5.
9672 The argument must be the name of a level (for example, 4.2), a
9674 level_idc value (for example, 42), or the special name auto
9675 indicating that the filter should attempt to guess the level from
9676 the input stream properties.
9677 h264_mp4toannexb
9679 Convert an H.264 bitstream from length prefixed mode to start code
9680 prefixed mode (as defined in the Annex B of the ITU-T H.264
9681 specification).
9682 This is required by some streaming formats, typically the MPEG-2
9684 transport stream format (muxer "mpegts").
9685 For example to remux an MP4 file containing an H.264 stream to mpegts
9687 format with ffmpeg, you can use the command:
9688 ffmpeg -i INPUT.mp4 -codec copy -bsf:v h264_mp4toannexb OUTPUT.ts
9690 Please note that this filter is auto-inserted for MPEG-TS (muxer
9692 "mpegts") and raw H.264 (muxer "h264") output formats.
9693 h264_redundant_pps
9695 This applies a specific fixup to some Blu-ray streams which contain
9696 redundant PPSs modifying irrelevant parameters of the stream which
9697 confuse other transformations which require correct extradata.
9698 hevc_metadata
9700 Modify metadata embedded in an HEVC stream.
9701 aud Insert or remove AUD NAL units in all access units of the stream.
9703 insert
9705 remove
9706 sample_aspect_ratio
9707 Set the sample aspect ratio in the stream in the VUI parameters.
9708 video_format
9710 video_full_range_flag
9711 Set the video format in the stream (see H.265 section E.3.1 and
9712 table E.2).
9713 colour_primaries
9715 transfer_characteristics
9716 matrix_coefficients
9717 Set the colour description in the stream (see H.265 section E.3.1
9718 and tables E.3, E.4 and E.5).
9719 chroma_sample_loc_type
9721 Set the chroma sample location in the stream (see H.265 section
9722 E.3.1 and figure E.1).
9723 tick_rate
9725 Set the tick rate in the VPS and VUI parameters (time_scale /
9726 num_units_in_tick). Combined with num_ticks_poc_diff_one, this can
9727 set a constant framerate in the stream. Note that it is likely to
9728 be overridden by container parameters when the stream is in a
9729 container.
9730 num_ticks_poc_diff_one
9732 Set poc_proportional_to_timing_flag in VPS and VUI and use this
9733 value to set num_ticks_poc_diff_one_minus1 (see H.265 sections
9734 7.4.3.1 and E.3.1). Ignored if tick_rate is not also set.
9735 crop_left
9737 crop_right
9738 crop_top
9739 crop_bottom
9740 Set the conformance window cropping offsets in the SPS. These
9741 values will replace the current ones if the stream is already
9742 cropped.
9743 These fields are set in pixels. Note that some sizes may not be
9745 representable if the chroma is subsampled (H.265 section
9746 7.4.3.2.1).
9747 level
9749 Set the level in the VPS and SPS. See H.265 section A.4 and tables
9750 A.6 and A.7.
9751 The argument must be the name of a level (for example, 5.1), a
9753 general_level_idc value (for example, 153 for level 5.1), or the
9754 special name auto indicating that the filter should attempt to
9755 guess the level from the input stream properties.
9756 hevc_mp4toannexb
9758 Convert an HEVC/H.265 bitstream from length prefixed mode to start code
9759 prefixed mode (as defined in the Annex B of the ITU-T H.265
9760 specification).
9761 This is required by some streaming formats, typically the MPEG-2
9763 transport stream format (muxer "mpegts").
9764 For example to remux an MP4 file containing an HEVC stream to mpegts
9766 format with ffmpeg, you can use the command:
9767 ffmpeg -i INPUT.mp4 -codec copy -bsf:v hevc_mp4toannexb OUTPUT.ts
9769 Please note that this filter is auto-inserted for MPEG-TS (muxer
9771 "mpegts") and raw HEVC/H.265 (muxer "h265" or "hevc") output formats.
9772 imxdump
9774 Modifies the bitstream to fit in MOV and to be usable by the Final Cut
9775 Pro decoder. This filter only applies to the mpeg2video codec, and is
9776 likely not needed for Final Cut Pro 7 and newer with the appropriate
9777 -tag:v.
9778 For example, to remux 30 MB/sec NTSC IMX to MOV:
9780 ffmpeg -i input.mxf -c copy -bsf:v imxdump -tag:v mx3n output.mov
9782 mjpeg2jpeg
9784 Convert MJPEG/AVI1 packets to full JPEG/JFIF packets.
9785 MJPEG is a video codec wherein each video frame is essentially a JPEG
9787 image. The individual frames can be extracted without loss, e.g. by
9788 ffmpeg -i ../some_mjpeg.avi -c:v copy frames_%d.jpg
9790 Unfortunately, these chunks are incomplete JPEG images, because they
9792 lack the DHT segment required for decoding. Quoting from
9793 <http://www.digitalpreservation.gov/formats/fdd/fdd000063.shtml>:
9794 Avery Lee, writing in the rec.video.desktop newsgroup in 2001,
9796 commented that "MJPEG, or at least the MJPEG in AVIs having the MJPG
9797 fourcc, is restricted JPEG with a fixed -- and *omitted* -- Huffman
9798 table. The JPEG must be YCbCr colorspace, it must be 4:2:2, and it must
9799 use basic Huffman encoding, not arithmetic or progressive. . . . You
9800 can indeed extract the MJPEG frames and decode them with a regular JPEG
9801 decoder, but you have to prepend the DHT segment to them, or else the
9802 decoder won't have any idea how to decompress the data. The exact table
9803 necessary is given in the OpenDML spec."
9804 This bitstream filter patches the header of frames extracted from an
9806 MJPEG stream (carrying the AVI1 header ID and lacking a DHT segment) to
9807 produce fully qualified JPEG images.
9808 ffmpeg -i mjpeg-movie.avi -c:v copy -bsf:v mjpeg2jpeg frame_%d.jpg
9810 exiftran -i -9 frame*.jpg
9811 ffmpeg -i frame_%d.jpg -c:v copy rotated.avi
9812 mjpegadump
9814 Add an MJPEG A header to the bitstream, to enable decoding by
9815 Quicktime.
9816 mov2textsub
9818 Extract a representable text file from MOV subtitles, stripping the
9819 metadata header from each subtitle packet.
9820 See also the text2movsub filter.
9822 mp3decomp
9824 Decompress non-standard compressed MP3 audio headers.
9825 mpeg2_metadata
9827 Modify metadata embedded in an MPEG-2 stream.
9828 display_aspect_ratio
9830 Set the display aspect ratio in the stream.
9831 The following fixed values are supported:
9833 4/3
9835 16/9
9836 221/100
9837 Any other value will result in square pixels being signalled
9839 instead (see H.262 section 6.3.3 and table 6-3).
9840 frame_rate
9842 Set the frame rate in the stream. This is constructed from a table
9843 of known values combined with a small multiplier and divisor - if
9844 the supplied value is not exactly representable, the nearest
9845 representable value will be used instead (see H.262 section 6.3.3
9846 and table 6-4).
9847 video_format
9849 Set the video format in the stream (see H.262 section 6.3.6 and
9850 table 6-6).
9851 colour_primaries
9853 transfer_characteristics
9854 matrix_coefficients
9855 Set the colour description in the stream (see H.262 section 6.3.6
9856 and tables 6-7, 6-8 and 6-9).
9857 mpeg4_unpack_bframes
9859 Unpack DivX-style packed B-frames.
9860 DivX-style packed B-frames are not valid MPEG-4 and were only a
9862 workaround for the broken Video for Windows subsystem. They use more
9863 space, can cause minor AV sync issues, require more CPU power to decode
9864 (unless the player has some decoded picture queue to compensate the
9865 2,0,2,0 frame per packet style) and cause trouble if copied into a
9866 standard container like mp4 or mpeg-ps/ts, because MPEG-4 decoders may
9867 not be able to decode them, since they are not valid MPEG-4.
9868 For example to fix an AVI file containing an MPEG-4 stream with DivX-
9870 style packed B-frames using ffmpeg, you can use the command:
9871 ffmpeg -i INPUT.avi -codec copy -bsf:v mpeg4_unpack_bframes OUTPUT.avi
9873 noise
9875 Damages the contents of packets or simply drops them without damaging
9876 the container. Can be used for fuzzing or testing error
9877 resilience/concealment.
9878 Parameters:
9880 amount
9882 Accepts an expression whose evaluation per-packet determines how
9883 often bytes in that packet will be modified. A value below 0 will
9884 result in a variable frequency. Default is 0 which results in no
9885 modification. However, if neither amount nor drop is specified,
9886 amount will be set to -1. See below for accepted variables.
9887 drop
9889 Accepts an expression evaluated per-packet whose value determines
9890 whether that packet is dropped. Evaluation to a positive value
9891 results in the packet being dropped. Evaluation to a negative value
9892 results in a variable chance of it being dropped, roughly inverse
9893 in proportion to the magnitude of the value. Default is 0 which
9894 results in no drops. See below for accepted variables.
9895 dropamount
9897 Accepts a non-negative integer, which assigns a variable chance of
9898 it being dropped, roughly inverse in proportion to the value.
9899 Default is 0 which results in no drops. This option is kept for
9900 backwards compatibility and is equivalent to setting drop to a
9901 negative value with the same magnitude i.e. "dropamount=4" is the
9902 same as "drop=-4". Ignored if drop is also specified.
9903 Both "amount" and "drop" accept expressions containing the following
9905 variables:
9906 n The index of the packet, starting from zero.
9908 tb The timebase for packet timestamps.
9910 pts Packet presentation timestamp.
9912 dts Packet decoding timestamp.
9914 nopts
9916 Constant representing AV_NOPTS_VALUE.
9917 startpts
9919 First non-AV_NOPTS_VALUE PTS seen in the stream.
9920 startdts
9922 First non-AV_NOPTS_VALUE DTS seen in the stream.
9923 duration
9925 d Packet duration, in timebase units.
9926 pos Packet position in input; may be -1 when unknown or not set.
9928 size
9930 Packet size, in bytes.
9931 key Whether packet is marked as a keyframe.
9933 state
9935 A pseudo random integer, primarily derived from the content of
9936 packet payload.
9937 Examples
9939 Apply modification to every byte but don't drop any packets.
9941 ffmpeg -i INPUT -c copy -bsf noise=1 output.mkv
9943 Drop every video packet not marked as a keyframe after timestamp 30s
9945 but do not modify any of the remaining packets.
9946 ffmpeg -i INPUT -c copy -bsf:v noise=drop='gt(t\,30)*not(key)' output.mkv
9948 Drop one second of audio every 10 seconds and add some random noise to
9950 the rest.
9951 ffmpeg -i INPUT -c copy -bsf:a noise=amount=-1:drop='between(mod(t\,10)\,9\,10)' output.mkv
9953 null
9955 This bitstream filter passes the packets through unchanged.
9956 pcm_rechunk
9958 Repacketize PCM audio to a fixed number of samples per packet or a
9959 fixed packet rate per second. This is similar to the asetnsamples audio
9960 filter but works on audio packets instead of audio frames.
9961 nb_out_samples, n
9963 Set the number of samples per each output audio packet. The number
9964 is intended as the number of samples per each channel. Default
9965 value is 1024.
9966 pad, p
9968 If set to 1, the filter will pad the last audio packet with
9969 silence, so that it will contain the same number of samples (or
9970 roughly the same number of samples, see frame_rate) as the previous
9971 ones. Default value is 1.
9972 frame_rate, r
9974 This option makes the filter output a fixed number of packets per
9975 second instead of a fixed number of samples per packet. If the
9976 audio sample rate is not divisible by the frame rate then the
9977 number of samples will not be constant but will vary slightly so
9978 that each packet will start as close to the frame boundary as
9979 possible. Using this option has precedence over nb_out_samples.
9980 You can generate the well known 1602-1601-1602-1601-1602 pattern of
9982 48kHz audio for NTSC frame rate using the frame_rate option.
9983 ffmpeg -f lavfi -i sine=r=48000:d=1 -c pcm_s16le -bsf pcm_rechunk=r=30000/1001 -f framecrc -
9985 pgs_frame_merge
9987 Merge a sequence of PGS Subtitle segments ending with an "end of
9988 display set" segment into a single packet.
9989 This is required by some containers that support PGS subtitles (muxer
9991 "matroska").
9992 prores_metadata
9994 Modify color property metadata embedded in prores stream.
9995 color_primaries
9997 Set the color primaries. Available values are:
9998 auto
10000 Keep the same color primaries property (default).
10001 unknown
10003 bt709
10004 bt470bg
10005 BT601 625
10006 smpte170m
10008 BT601 525
10009 bt2020
10011 smpte431
10012 DCI P3
10013 smpte432
10015 P3 D65
10016 transfer_characteristics
10018 Set the color transfer. Available values are:
10019 auto
10021 Keep the same transfer characteristics property (default).
10022 unknown
10024 bt709
10025 BT 601, BT 709, BT 2020
10026 smpte2084
10028 SMPTE ST 2084
10029 arib-std-b67
10031 ARIB STD-B67
10032 matrix_coefficients
10034 Set the matrix coefficient. Available values are:
10035 auto
10037 Keep the same colorspace property (default).
10038 unknown
10040 bt709
10041 smpte170m
10042 BT 601
10043 bt2020nc
10045 Set Rec709 colorspace for each frame of the file
10047 ffmpeg -i INPUT -c copy -bsf:v prores_metadata=color_primaries=bt709:color_trc=bt709:colorspace=bt709 output.mov
10049 Set Hybrid Log-Gamma parameters for each frame of the file
10051 ffmpeg -i INPUT -c copy -bsf:v prores_metadata=color_primaries=bt2020:color_trc=arib-std-b67:colorspace=bt2020nc output.mov
10053 remove_extra
10055 Remove extradata from packets.
10056 It accepts the following parameter:
10058 freq
10060 Set which frame types to remove extradata from.
10061 k Remove extradata from non-keyframes only.
10063 keyframe
10065 Remove extradata from keyframes only.
10066 e, all
10068 Remove extradata from all frames.
10069 setts
10071 Set PTS and DTS in packets.
10072 It accepts the following parameters:
10074 ts
10076 pts
10077 dts Set expressions for PTS, DTS or both.
10078 duration
10080 Set expression for duration.
10081 time_base
10083 Set output time base.
10084 The expressions are evaluated through the eval API and can contain the
10086 following constants:
10087 N The count of the input packet. Starting from 0.
10089 TS The demux timestamp in input in case of "ts" or "dts" option or
10091 presentation timestamp in case of "pts" option.
10092 POS The original position in the file of the packet, or undefined if
10094 undefined for the current packet
10095 DTS The demux timestamp in input.
10097 PTS The presentation timestamp in input.
10099 DURATION
10101 The duration in input.
10102 STARTDTS
10104 The DTS of the first packet.
10105 STARTPTS
10107 The PTS of the first packet.
10108 PREV_INDTS
10110 The previous input DTS.
10111 PREV_INPTS
10113 The previous input PTS.
10114 PREV_INDURATION
10116 The previous input duration.
10117 PREV_OUTDTS
10119 The previous output DTS.
10120 PREV_OUTPTS
10122 The previous output PTS.
10123 PREV_OUTDURATION
10125 The previous output duration.
10126 NEXT_DTS
10128 The next input DTS.
10129 NEXT_PTS
10131 The next input PTS.
10132 NEXT_DURATION
10134 The next input duration.
10135 TB The timebase of stream packet belongs.
10137 TB_OUT
10139 The output timebase.
10140 SR The sample rate of stream packet belongs.
10142 NOPTS
10144 The AV_NOPTS_VALUE constant.
10145 text2movsub
10147 Convert text subtitles to MOV subtitles (as used by the "mov_text"
10148 codec) with metadata headers.
10149 See also the mov2textsub filter.
10151 trace_headers
10153 Log trace output containing all syntax elements in the coded stream
10154 headers (everything above the level of individual coded blocks). This
10155 can be useful for debugging low-level stream issues.
10156 Supports AV1, H.264, H.265, (M)JPEG, MPEG-2 and VP9, but depending on
10158 the build only a subset of these may be available.
10159 truehd_core
10161 Extract the core from a TrueHD stream, dropping ATMOS data.
10162 vp9_metadata
10164 Modify metadata embedded in a VP9 stream.
10165 color_space
10167 Set the color space value in the frame header. Note that any frame
10168 set to RGB will be implicitly set to PC range and that RGB is
10169 incompatible with profiles 0 and 2.
10170 unknown
10172 bt601
10173 bt709
10174 smpte170
10175 smpte240
10176 bt2020
10177 rgb
10178 color_range
10179 Set the color range value in the frame header. Note that any value
10180 imposed by the color space will take precedence over this value.
10181 tv
10183 pc
10184 vp9_superframe
10186 Merge VP9 invisible (alt-ref) frames back into VP9 superframes. This
10187 fixes merging of split/segmented VP9 streams where the alt-ref frame
10188 was split from its visible counterpart.
10189 vp9_superframe_split
10191 Split VP9 superframes into single frames.
10192 vp9_raw_reorder
10194 Given a VP9 stream with correct timestamps but possibly out of order,
10195 insert additional show-existing-frame packets to correct the ordering.
10196
10198 The libavformat library provides some generic global options, which can
10199 be set on all the muxers and demuxers. In addition each muxer or
10200 demuxer may support so-called private options, which are specific for
10201 that component.
10202 Options may be set by specifying -option value in the FFmpeg tools, or
10204 by setting the value explicitly in the "AVFormatContext" options or
10205 using the libavutil/opt.h API for programmatic use.
10206 The list of supported options follows:
10208 avioflags flags (input/output)
10210 Possible values:
10211 direct
10213 Reduce buffering.
10214 probesize integer (input)
10216 Set probing size in bytes, i.e. the size of the data to analyze to
10217 get stream information. A higher value will enable detecting more
10218 information in case it is dispersed into the stream, but will
10219 increase latency. Must be an integer not lesser than 32. It is
10220 5000000 by default.
10221 max_probe_packets integer (input)
10223 Set the maximum number of buffered packets when probing a codec.
10224 Default is 2500 packets.
10225 packetsize integer (output)
10227 Set packet size.
10228 fflags flags
10230 Set format flags. Some are implemented for a limited number of
10231 formats.
10232 Possible values for input files:
10234 discardcorrupt
10236 Discard corrupted packets.
10237 fastseek
10239 Enable fast, but inaccurate seeks for some formats.
10240 genpts
10242 Generate missing PTS if DTS is present.
10243 igndts
10245 Ignore DTS if PTS is set. Inert when nofillin is set.
10246 ignidx
10248 Ignore index.
10249 nobuffer
10251 Reduce the latency introduced by buffering during initial input
10252 streams analysis.
10253 nofillin
10255 Do not fill in missing values in packet fields that can be
10256 exactly calculated.
10257 noparse
10259 Disable AVParsers, this needs "+nofillin" too.
10260 sortdts
10262 Try to interleave output packets by DTS. At present, available
10263 only for AVIs with an index.
10264 Possible values for output files:
10266 autobsf
10268 Automatically apply bitstream filters as required by the output
10269 format. Enabled by default.
10270 bitexact
10272 Only write platform-, build- and time-independent data. This
10273 ensures that file and data checksums are reproducible and match
10274 between platforms. Its primary use is for regression testing.
10275 flush_packets
10277 Write out packets immediately.
10278 shortest
10280 Stop muxing at the end of the shortest stream. It may be
10281 needed to increase max_interleave_delta to avoid flushing the
10282 longer streams before EOF.
10283 seek2any integer (input)
10285 Allow seeking to non-keyframes on demuxer level when supported if
10286 set to 1. Default is 0.
10287 analyzeduration integer (input)
10289 Specify how many microseconds are analyzed to probe the input. A
10290 higher value will enable detecting more accurate information, but
10291 will increase latency. It defaults to 5,000,000 microseconds = 5
10292 seconds.
10293 cryptokey hexadecimal string (input)
10295 Set decryption key.
10296 indexmem integer (input)
10298 Set max memory used for timestamp index (per stream).
10299 rtbufsize integer (input)
10301 Set max memory used for buffering real-time frames.
10302 fdebug flags (input/output)
10304 Print specific debug info.
10305 Possible values:
10307 ts
10309 max_delay integer (input/output)
10310 Set maximum muxing or demuxing delay in microseconds.
10311 fpsprobesize integer (input)
10313 Set number of frames used to probe fps.
10314 audio_preload integer (output)
10316 Set microseconds by which audio packets should be interleaved
10317 earlier.
10318 chunk_duration integer (output)
10320 Set microseconds for each chunk.
10321 chunk_size integer (output)
10323 Set size in bytes for each chunk.
10324 err_detect, f_err_detect flags (input)
10326 Set error detection flags. "f_err_detect" is deprecated and should
10327 be used only via the ffmpeg tool.
10328 Possible values:
10330 crccheck
10332 Verify embedded CRCs.
10333 bitstream
10335 Detect bitstream specification deviations.
10336 buffer
10338 Detect improper bitstream length.
10339 explode
10341 Abort decoding on minor error detection.
10342 careful
10344 Consider things that violate the spec and have not been seen in
10345 the wild as errors.
10346 compliant
10348 Consider all spec non compliancies as errors.
10349 aggressive
10351 Consider things that a sane encoder should not do as an error.
10352 max_interleave_delta integer (output)
10354 Set maximum buffering duration for interleaving. The duration is
10355 expressed in microseconds, and defaults to 10000000 (10 seconds).
10356 To ensure all the streams are interleaved correctly, libavformat
10358 will wait until it has at least one packet for each stream before
10359 actually writing any packets to the output file. When some streams
10360 are "sparse" (i.e. there are large gaps between successive
10361 packets), this can result in excessive buffering.
10362 This field specifies the maximum difference between the timestamps
10364 of the first and the last packet in the muxing queue, above which
10365 libavformat will output a packet regardless of whether it has
10366 queued a packet for all the streams.
10367 If set to 0, libavformat will continue buffering packets until it
10369 has a packet for each stream, regardless of the maximum timestamp
10370 difference between the buffered packets.
10371 use_wallclock_as_timestamps integer (input)
10373 Use wallclock as timestamps if set to 1. Default is 0.
10374 avoid_negative_ts integer (output)
10376 Possible values:
10377 make_non_negative
10379 Shift timestamps to make them non-negative. Also note that
10380 this affects only leading negative timestamps, and not non-
10381 monotonic negative timestamps.
10382 make_zero
10384 Shift timestamps so that the first timestamp is 0.
10385 auto (default)
10387 Enables shifting when required by the target format.
10388 disabled
10390 Disables shifting of timestamp.
10391 When shifting is enabled, all output timestamps are shifted by the
10393 same amount. Audio, video, and subtitles desynching and relative
10394 timestamp differences are preserved compared to how they would have
10395 been without shifting.
10396 skip_initial_bytes integer (input)
10398 Set number of bytes to skip before reading header and frames if set
10399 to 1. Default is 0.
10400 correct_ts_overflow integer (input)
10402 Correct single timestamp overflows if set to 1. Default is 1.
10403 flush_packets integer (output)
10405 Flush the underlying I/O stream after each packet. Default is -1
10406 (auto), which means that the underlying protocol will decide, 1
10407 enables it, and has the effect of reducing the latency, 0 disables
10408 it and may increase IO throughput in some cases.
10409 output_ts_offset offset (output)
10411 Set the output time offset.
10412 offset must be a time duration specification, see the Time duration
10414 section in the ffmpeg-utils(1) manual.
10415 The offset is added by the muxer to the output timestamps.
10417 Specifying a positive offset means that the corresponding streams
10419 are delayed bt the time duration specified in offset. Default value
10420 is 0 (meaning that no offset is applied).
10421 format_whitelist list (input)
10423 "," separated list of allowed demuxers. By default all are allowed.
10424 dump_separator string (input)
10426 Separator used to separate the fields printed on the command line
10427 about the Stream parameters. For example, to separate the fields
10428 with newlines and indentation:
10429 ffprobe -dump_separator "
10431 " -i ~/videos/matrixbench_mpeg2.mpg
10432 max_streams integer (input)
10434 Specifies the maximum number of streams. This can be used to reject
10435 files that would require too many resources due to a large number
10436 of streams.
10437 skip_estimate_duration_from_pts bool (input)
10439 Skip estimation of input duration when calculated using PTS. At
10440 present, applicable for MPEG-PS and MPEG-TS.
10441 strict, f_strict integer (input/output)
10443 Specify how strictly to follow the standards. "f_strict" is
10444 deprecated and should be used only via the ffmpeg tool.
10445 Possible values:
10447 very
10449 strictly conform to an older more strict version of the spec or
10450 reference software
10451 strict
10453 strictly conform to all the things in the spec no matter what
10454 consequences
10455 normal
10457 unofficial
10458 allow unofficial extensions
10459 experimental
10461 allow non standardized experimental things, experimental
10462 (unfinished/work in progress/not well tested) decoders and
10463 encoders. Note: experimental decoders can pose a security
10464 risk, do not use this for decoding untrusted input.
10465 Format stream specifiers
10467 Format stream specifiers allow selection of one or more streams that
10468 match specific properties.
10469 The exact semantics of stream specifiers is defined by the
10471 avformat_match_stream_specifier() function declared in the
10472 libavformat/avformat.h header and documented in the Stream specifiers
10473 section in the ffmpeg(1) manual.
10474
10476 Demuxers are configured elements in FFmpeg that can read the multimedia
10477 streams from a particular type of file.
10478 When you configure your FFmpeg build, all the supported demuxers are
10480 enabled by default. You can list all available ones using the configure
10481 option "--list-demuxers".
10482 You can disable all the demuxers using the configure option
10484 "--disable-demuxers", and selectively enable a single demuxer with the
10485 option "--enable-demuxer=DEMUXER", or disable it with the option
10486 "--disable-demuxer=DEMUXER".
10487 The option "-demuxers" of the ff* tools will display the list of
10489 enabled demuxers. Use "-formats" to view a combined list of enabled
10490 demuxers and muxers.
10491 The description of some of the currently available demuxers follows.
10493 aa
10495 Audible Format 2, 3, and 4 demuxer.
10496 This demuxer is used to demux Audible Format 2, 3, and 4 (.aa) files.
10498 aac
10500 Raw Audio Data Transport Stream AAC demuxer.
10501 This demuxer is used to demux an ADTS input containing a single AAC
10503 stream alongwith any ID3v1/2 or APE tags in it.
10504 apng
10506 Animated Portable Network Graphics demuxer.
10507 This demuxer is used to demux APNG files. All headers, but the PNG
10509 signature, up to (but not including) the first fcTL chunk are
10510 transmitted as extradata. Frames are then split as being all the
10511 chunks between two fcTL ones, or between the last fcTL and IEND chunks.
10512 -ignore_loop bool
10514 Ignore the loop variable in the file if set. Default is enabled.
10515 -max_fps int
10517 Maximum framerate in frames per second. Default of 0 imposes no
10518 limit.
10519 -default_fps int
10521 Default framerate in frames per second when none is specified in
10522 the file (0 meaning as fast as possible). Default is 15.
10523 asf
10525 Advanced Systems Format demuxer.
10526 This demuxer is used to demux ASF files and MMS network streams.
10528 -no_resync_search bool
10530 Do not try to resynchronize by looking for a certain optional start
10531 code.
10532 concat
10534 Virtual concatenation script demuxer.
10535 This demuxer reads a list of files and other directives from a text
10537 file and demuxes them one after the other, as if all their packets had
10538 been muxed together.
10539 The timestamps in the files are adjusted so that the first file starts
10541 at 0 and each next file starts where the previous one finishes. Note
10542 that it is done globally and may cause gaps if all streams do not have
10543 exactly the same length.
10544 All files must have the same streams (same codecs, same time base,
10546 etc.).
10547 The duration of each file is used to adjust the timestamps of the next
10549 file: if the duration is incorrect (because it was computed using the
10550 bit-rate or because the file is truncated, for example), it can cause
10551 artifacts. The "duration" directive can be used to override the
10552 duration stored in each file.
10553 Syntax
10555 The script is a text file in extended-ASCII, with one directive per
10557 line. Empty lines, leading spaces and lines starting with '#' are
10558 ignored. The following directive is recognized:
10559 "file path"
10561 Path to a file to read; special characters and spaces must be
10562 escaped with backslash or single quotes.
10563 All subsequent file-related directives apply to that file.
10565 "ffconcat version 1.0"
10567 Identify the script type and version.
10568 To make FFmpeg recognize the format automatically, this directive
10570 must appear exactly as is (no extra space or byte-order-mark) on
10571 the very first line of the script.
10572 "duration dur"
10574 Duration of the file. This information can be specified from the
10575 file; specifying it here may be more efficient or help if the
10576 information from the file is not available or accurate.
10577 If the duration is set for all files, then it is possible to seek
10579 in the whole concatenated video.
10580 "inpoint timestamp"
10582 In point of the file. When the demuxer opens the file it instantly
10583 seeks to the specified timestamp. Seeking is done so that all
10584 streams can be presented successfully at In point.
10585 This directive works best with intra frame codecs, because for non-
10587 intra frame ones you will usually get extra packets before the
10588 actual In point and the decoded content will most likely contain
10589 frames before In point too.
10590 For each file, packets before the file In point will have
10592 timestamps less than the calculated start timestamp of the file
10593 (negative in case of the first file), and the duration of the files
10594 (if not specified by the "duration" directive) will be reduced
10595 based on their specified In point.
10596 Because of potential packets before the specified In point, packet
10598 timestamps may overlap between two concatenated files.
10599 "outpoint timestamp"
10601 Out point of the file. When the demuxer reaches the specified
10602 decoding timestamp in any of the streams, it handles it as an end
10603 of file condition and skips the current and all the remaining
10604 packets from all streams.
10605 Out point is exclusive, which means that the demuxer will not
10607 output packets with a decoding timestamp greater or equal to Out
10608 point.
10609 This directive works best with intra frame codecs and formats where
10611 all streams are tightly interleaved. For non-intra frame codecs you
10612 will usually get additional packets with presentation timestamp
10613 after Out point therefore the decoded content will most likely
10614 contain frames after Out point too. If your streams are not tightly
10615 interleaved you may not get all the packets from all streams before
10616 Out point and you may only will be able to decode the earliest
10617 stream until Out point.
10618 The duration of the files (if not specified by the "duration"
10620 directive) will be reduced based on their specified Out point.
10621 "file_packet_metadata key=value"
10623 Metadata of the packets of the file. The specified metadata will be
10624 set for each file packet. You can specify this directive multiple
10625 times to add multiple metadata entries. This directive is
10626 deprecated, use "file_packet_meta" instead.
10627 "file_packet_meta key value"
10629 Metadata of the packets of the file. The specified metadata will be
10630 set for each file packet. You can specify this directive multiple
10631 times to add multiple metadata entries.
10632 "option key value"
10634 Option to access, open and probe the file. Can be present multiple
10635 times.
10636 "stream"
10638 Introduce a stream in the virtual file. All subsequent stream-
10639 related directives apply to the last introduced stream. Some
10640 streams properties must be set in order to allow identifying the
10641 matching streams in the subfiles. If no streams are defined in the
10642 script, the streams from the first file are copied.
10643 "exact_stream_id id"
10645 Set the id of the stream. If this directive is given, the string
10646 with the corresponding id in the subfiles will be used. This is
10647 especially useful for MPEG-PS (VOB) files, where the order of the
10648 streams is not reliable.
10649 "stream_meta key value"
10651 Metadata for the stream. Can be present multiple times.
10652 "stream_codec value"
10654 Codec for the stream.
10655 "stream_extradata hex_string"
10657 Extradata for the string, encoded in hexadecimal.
10658 "chapter id start end"
10660 Add a chapter. id is an unique identifier, possibly small and
10661 consecutive.
10662 Options
10664 This demuxer accepts the following option:
10666 safe
10668 If set to 1, reject unsafe file paths and directives. A file path
10669 is considered safe if it does not contain a protocol specification
10670 and is relative and all components only contain characters from the
10671 portable character set (letters, digits, period, underscore and
10672 hyphen) and have no period at the beginning of a component.
10673 If set to 0, any file name is accepted.
10675 The default is 1.
10677 auto_convert
10679 If set to 1, try to perform automatic conversions on packet data to
10680 make the streams concatenable. The default is 1.
10681 Currently, the only conversion is adding the h264_mp4toannexb
10683 bitstream filter to H.264 streams in MP4 format. This is necessary
10684 in particular if there are resolution changes.
10685 segment_time_metadata
10687 If set to 1, every packet will contain the lavf.concat.start_time
10688 and the lavf.concat.duration packet metadata values which are the
10689 start_time and the duration of the respective file segments in the
10690 concatenated output expressed in microseconds. The duration
10691 metadata is only set if it is known based on the concat file. The
10692 default is 0.
10693 Examples
10695 • Use absolute filenames and include some comments:
10697 # my first filename
10699 file /mnt/share/file-1.wav
10700 # my second filename including whitespace
10701 file '/mnt/share/file 2.wav'
10702 # my third filename including whitespace plus single quote
10703 file '/mnt/share/file 3'\''.wav'
10704 • Allow for input format auto-probing, use safe filenames and set the
10706 duration of the first file:
10707 ffconcat version 1.0
10709 file file-1.wav
10711 duration 20.0
10712 file subdir/file-2.wav
10714 dash
10716 Dynamic Adaptive Streaming over HTTP demuxer.
10717 This demuxer presents all AVStreams found in the manifest. By setting
10719 the discard flags on AVStreams the caller can decide which streams to
10720 actually receive. Each stream mirrors the "id" and "bandwidth"
10721 properties from the "<Representation>" as metadata keys named "id" and
10722 "variant_bitrate" respectively.
10723 Options
10725 This demuxer accepts the following option:
10727 cenc_decryption_key
10729 16-byte key, in hex, to decrypt files encrypted using ISO Common
10730 Encryption (CENC/AES-128 CTR; ISO/IEC 23001-7).
10731 ea
10733 Electronic Arts Multimedia format demuxer.
10734 This format is used by various Electronic Arts games.
10736 Options
10738 merge_alpha bool
10740 Normally the VP6 alpha channel (if exists) is returned as a
10741 secondary video stream, by setting this option you can make the
10742 demuxer return a single video stream which contains the alpha
10743 channel in addition to the ordinary video.
10744 imf
10746 Interoperable Master Format demuxer.
10747 This demuxer presents audio and video streams found in an IMF
10749 Composition.
10750 flv, live_flv, kux
10752 Adobe Flash Video Format demuxer.
10753 This demuxer is used to demux FLV files and RTMP network streams. In
10755 case of live network streams, if you force format, you may use live_flv
10756 option instead of flv to survive timestamp discontinuities. KUX is a
10757 flv variant used on the Youku platform.
10758 ffmpeg -f flv -i myfile.flv ...
10760 ffmpeg -f live_flv -i rtmp://<any.server>/anything/key ....
10761 -flv_metadata bool
10763 Allocate the streams according to the onMetaData array content.
10764 -flv_ignore_prevtag bool
10766 Ignore the size of previous tag value.
10767 -flv_full_metadata bool
10769 Output all context of the onMetadata.
10770 gif
10772 Animated GIF demuxer.
10773 It accepts the following options:
10775 min_delay
10777 Set the minimum valid delay between frames in hundredths of
10778 seconds. Range is 0 to 6000. Default value is 2.
10779 max_gif_delay
10781 Set the maximum valid delay between frames in hundredth of seconds.
10782 Range is 0 to 65535. Default value is 65535 (nearly eleven
10783 minutes), the maximum value allowed by the specification.
10784 default_delay
10786 Set the default delay between frames in hundredths of seconds.
10787 Range is 0 to 6000. Default value is 10.
10788 ignore_loop
10790 GIF files can contain information to loop a certain number of times
10791 (or infinitely). If ignore_loop is set to 1, then the loop setting
10792 from the input will be ignored and looping will not occur. If set
10793 to 0, then looping will occur and will cycle the number of times
10794 according to the GIF. Default value is 1.
10795 For example, with the overlay filter, place an infinitely looping GIF
10797 over another video:
10798 ffmpeg -i input.mp4 -ignore_loop 0 -i input.gif -filter_complex overlay=shortest=1 out.mkv
10800 Note that in the above example the shortest option for overlay filter
10802 is used to end the output video at the length of the shortest input
10803 file, which in this case is input.mp4 as the GIF in this example loops
10804 infinitely.
10805 hls
10807 HLS demuxer
10808 Apple HTTP Live Streaming demuxer.
10810 This demuxer presents all AVStreams from all variant streams. The id
10812 field is set to the bitrate variant index number. By setting the
10813 discard flags on AVStreams (by pressing 'a' or 'v' in ffplay), the
10814 caller can decide which variant streams to actually receive. The total
10815 bitrate of the variant that the stream belongs to is available in a
10816 metadata key named "variant_bitrate".
10817 It accepts the following options:
10819 live_start_index
10821 segment index to start live streams at (negative values are from
10822 the end).
10823 prefer_x_start
10825 prefer to use #EXT-X-START if it's in playlist instead of
10826 live_start_index.
10827 allowed_extensions
10829 ',' separated list of file extensions that hls is allowed to
10830 access.
10831 max_reload
10833 Maximum number of times a insufficient list is attempted to be
10834 reloaded. Default value is 1000.
10835 m3u8_hold_counters
10837 The maximum number of times to load m3u8 when it refreshes without
10838 new segments. Default value is 1000.
10839 http_persistent
10841 Use persistent HTTP connections. Applicable only for HTTP streams.
10842 Enabled by default.
10843 http_multiple
10845 Use multiple HTTP connections for downloading HTTP segments.
10846 Enabled by default for HTTP/1.1 servers.
10847 http_seekable
10849 Use HTTP partial requests for downloading HTTP segments. 0 =
10850 disable, 1 = enable, -1 = auto, Default is auto.
10851 seg_format_options
10853 Set options for the demuxer of media segments using a list of
10854 key=value pairs separated by ":".
10855 seg_max_retry
10857 Maximum number of times to reload a segment on error, useful when
10858 segment skip on network error is not desired. Default value is 0.
10859 image2
10861 Image file demuxer.
10862 This demuxer reads from a list of image files specified by a pattern.
10864 The syntax and meaning of the pattern is specified by the option
10865 pattern_type.
10866 The pattern may contain a suffix which is used to automatically
10868 determine the format of the images contained in the files.
10869 The size, the pixel format, and the format of each image must be the
10871 same for all the files in the sequence.
10872 This demuxer accepts the following options:
10874 framerate
10876 Set the frame rate for the video stream. It defaults to 25.
10877 loop
10879 If set to 1, loop over the input. Default value is 0.
10880 pattern_type
10882 Select the pattern type used to interpret the provided filename.
10883 pattern_type accepts one of the following values.
10885 none
10887 Disable pattern matching, therefore the video will only contain
10888 the specified image. You should use this option if you do not
10889 want to create sequences from multiple images and your
10890 filenames may contain special pattern characters.
10891 sequence
10893 Select a sequence pattern type, used to specify a sequence of
10894 files indexed by sequential numbers.
10895 A sequence pattern may contain the string "%d" or "%0Nd", which
10897 specifies the position of the characters representing a
10898 sequential number in each filename matched by the pattern. If
10899 the form "%d0Nd" is used, the string representing the number in
10900 each filename is 0-padded and N is the total number of 0-padded
10901 digits representing the number. The literal character '%' can
10902 be specified in the pattern with the string "%%".
10903 If the sequence pattern contains "%d" or "%0Nd", the first
10905 filename of the file list specified by the pattern must contain
10906 a number inclusively contained between start_number and
10907 start_number+start_number_range-1, and all the following
10908 numbers must be sequential.
10909 For example the pattern "img-%03d.bmp" will match a sequence of
10911 filenames of the form img-001.bmp, img-002.bmp, ...,
10912 img-010.bmp, etc.; the pattern "i%%m%%g-%d.jpg" will match a
10913 sequence of filenames of the form i%m%g-1.jpg, i%m%g-2.jpg,
10914 ..., i%m%g-10.jpg, etc.
10915 Note that the pattern must not necessarily contain "%d" or
10917 "%0Nd", for example to convert a single image file img.jpeg you
10918 can employ the command:
10919 ffmpeg -i img.jpeg img.png
10921 glob
10923 Select a glob wildcard pattern type.
10924 The pattern is interpreted like a glob() pattern. This is only
10926 selectable if libavformat was compiled with globbing support.
10927 glob_sequence (deprecated, will be removed)
10929 Select a mixed glob wildcard/sequence pattern.
10930 If your version of libavformat was compiled with globbing
10932 support, and the provided pattern contains at least one glob
10933 meta character among "%*?[]{}" that is preceded by an unescaped
10934 "%", the pattern is interpreted like a glob() pattern,
10935 otherwise it is interpreted like a sequence pattern.
10936 All glob special characters "%*?[]{}" must be prefixed with
10938 "%". To escape a literal "%" you shall use "%%".
10939 For example the pattern "foo-%*.jpeg" will match all the
10941 filenames prefixed by "foo-" and terminating with ".jpeg", and
10942 "foo-%?%?%?.jpeg" will match all the filenames prefixed with
10943 "foo-", followed by a sequence of three characters, and
10944 terminating with ".jpeg".
10945 This pattern type is deprecated in favor of glob and sequence.
10947 Default value is glob_sequence.
10949 pixel_format
10951 Set the pixel format of the images to read. If not specified the
10952 pixel format is guessed from the first image file in the sequence.
10953 start_number
10955 Set the index of the file matched by the image file pattern to
10956 start to read from. Default value is 0.
10957 start_number_range
10959 Set the index interval range to check when looking for the first
10960 image file in the sequence, starting from start_number. Default
10961 value is 5.
10962 ts_from_file
10964 If set to 1, will set frame timestamp to modification time of image
10965 file. Note that monotonity of timestamps is not provided: images go
10966 in the same order as without this option. Default value is 0. If
10967 set to 2, will set frame timestamp to the modification time of the
10968 image file in nanosecond precision.
10969 video_size
10971 Set the video size of the images to read. If not specified the
10972 video size is guessed from the first image file in the sequence.
10973 export_path_metadata
10975 If set to 1, will add two extra fields to the metadata found in
10976 input, making them also available for other filters (see drawtext
10977 filter for examples). Default value is 0. The extra fields are
10978 described below:
10979 lavf.image2dec.source_path
10981 Corresponds to the full path to the input file being read.
10982 lavf.image2dec.source_basename
10984 Corresponds to the name of the file being read.
10985 Examples
10987 • Use ffmpeg for creating a video from the images in the file
10989 sequence img-001.jpeg, img-002.jpeg, ..., assuming an input frame
10990 rate of 10 frames per second:
10991 ffmpeg -framerate 10 -i 'img-%03d.jpeg' out.mkv
10993 • As above, but start by reading from a file with index 100 in the
10995 sequence:
10996 ffmpeg -framerate 10 -start_number 100 -i 'img-%03d.jpeg' out.mkv
10998 • Read images matching the "*.png" glob pattern , that is all the
11000 files terminating with the ".png" suffix:
11001 ffmpeg -framerate 10 -pattern_type glob -i "*.png" out.mkv
11003 libgme
11005 The Game Music Emu library is a collection of video game music file
11006 emulators.
11007 See <https://bitbucket.org/mpyne/game-music-emu/overview> for more
11009 information.
11010 It accepts the following options:
11012 track_index
11014 Set the index of which track to demux. The demuxer can only export
11015 one track. Track indexes start at 0. Default is to pick the first
11016 track. Number of tracks is exported as tracks metadata entry.
11017 sample_rate
11019 Set the sampling rate of the exported track. Range is 1000 to
11020 999999. Default is 44100.
11021 max_size (bytes)
11023 The demuxer buffers the entire file into memory. Adjust this value
11024 to set the maximum buffer size, which in turn, acts as a ceiling
11025 for the size of files that can be read. Default is 50 MiB.
11026 libmodplug
11028 ModPlug based module demuxer
11029 See <https://github.com/Konstanty/libmodplug>
11031 It will export one 2-channel 16-bit 44.1 kHz audio stream. Optionally,
11033 a "pal8" 16-color video stream can be exported with or without printed
11034 metadata.
11035 It accepts the following options:
11037 noise_reduction
11039 Apply a simple low-pass filter. Can be 1 (on) or 0 (off). Default
11040 is 0.
11041 reverb_depth
11043 Set amount of reverb. Range 0-100. Default is 0.
11044 reverb_delay
11046 Set delay in ms, clamped to 40-250 ms. Default is 0.
11047 bass_amount
11049 Apply bass expansion a.k.a. XBass or megabass. Range is 0 (quiet)
11050 to 100 (loud). Default is 0.
11051 bass_range
11053 Set cutoff i.e. upper-bound for bass frequencies. Range is 10-100
11054 Hz. Default is 0.
11055 surround_depth
11057 Apply a Dolby Pro-Logic surround effect. Range is 0 (quiet) to 100
11058 (heavy). Default is 0.
11059 surround_delay
11061 Set surround delay in ms, clamped to 5-40 ms. Default is 0.
11062 max_size
11064 The demuxer buffers the entire file into memory. Adjust this value
11065 to set the maximum buffer size, which in turn, acts as a ceiling
11066 for the size of files that can be read. Range is 0 to 100 MiB. 0
11067 removes buffer size limit (not recommended). Default is 5 MiB.
11068 video_stream_expr
11070 String which is evaluated using the eval API to assign colors to
11071 the generated video stream. Variables which can be used are "x",
11072 "y", "w", "h", "t", "speed", "tempo", "order", "pattern" and "row".
11073 video_stream
11075 Generate video stream. Can be 1 (on) or 0 (off). Default is 0.
11076 video_stream_w
11078 Set video frame width in 'chars' where one char indicates 8 pixels.
11079 Range is 20-512. Default is 30.
11080 video_stream_h
11082 Set video frame height in 'chars' where one char indicates 8
11083 pixels. Range is 20-512. Default is 30.
11084 video_stream_ptxt
11086 Print metadata on video stream. Includes "speed", "tempo", "order",
11087 "pattern", "row" and "ts" (time in ms). Can be 1 (on) or 0 (off).
11088 Default is 1.
11089 libopenmpt
11091 libopenmpt based module demuxer
11092 See <https://lib.openmpt.org/libopenmpt/> for more information.
11094 Some files have multiple subsongs (tracks) this can be set with the
11096 subsong option.
11097 It accepts the following options:
11099 subsong
11101 Set the subsong index. This can be either 'all', 'auto', or the
11102 index of the subsong. Subsong indexes start at 0. The default is
11103 'auto'.
11104 The default value is to let libopenmpt choose.
11106 layout
11108 Set the channel layout. Valid values are 1, 2, and 4 channel
11109 layouts. The default value is STEREO.
11110 sample_rate
11112 Set the sample rate for libopenmpt to output. Range is from 1000
11113 to INT_MAX. The value default is 48000.
11114 mov/mp4/3gp
11116 Demuxer for Quicktime File Format & ISO/IEC Base Media File Format
11117 (ISO/IEC 14496-12 or MPEG-4 Part 12, ISO/IEC 15444-12 or JPEG 2000 Part
11118 12).
11119 Registered extensions: mov, mp4, m4a, 3gp, 3g2, mj2, psp, m4b, ism,
11121 ismv, isma, f4v
11122 Options
11124 This demuxer accepts the following options:
11126 enable_drefs
11128 Enable loading of external tracks, disabled by default. Enabling
11129 this can theoretically leak information in some use cases.
11130 use_absolute_path
11132 Allows loading of external tracks via absolute paths, disabled by
11133 default. Enabling this poses a security risk. It should only be
11134 enabled if the source is known to be non-malicious.
11135 seek_streams_individually
11137 When seeking, identify the closest point in each stream
11138 individually and demux packets in that stream from identified
11139 point. This can lead to a different sequence of packets compared to
11140 demuxing linearly from the beginning. Default is true.
11141 ignore_editlist
11143 Ignore any edit list atoms. The demuxer, by default, modifies the
11144 stream index to reflect the timeline described by the edit list.
11145 Default is false.
11146 advanced_editlist
11148 Modify the stream index to reflect the timeline described by the
11149 edit list. "ignore_editlist" must be set to false for this option
11150 to be effective. If both "ignore_editlist" and this option are set
11151 to false, then only the start of the stream index is modified to
11152 reflect initial dwell time or starting timestamp described by the
11153 edit list. Default is true.
11154 ignore_chapters
11156 Don't parse chapters. This includes GoPro 'HiLight' tags/moments.
11157 Note that chapters are only parsed when input is seekable. Default
11158 is false.
11159 use_mfra_for
11161 For seekable fragmented input, set fragment's starting timestamp
11162 from media fragment random access box, if present.
11163 Following options are available:
11165 auto
11167 Auto-detect whether to set mfra timestamps as PTS or DTS
11168 (default)
11169 dts Set mfra timestamps as DTS
11171 pts Set mfra timestamps as PTS
11173 0 Don't use mfra box to set timestamps
11175 use_tfdt
11177 For fragmented input, set fragment's starting timestamp to
11178 "baseMediaDecodeTime" from the "tfdt" box. Default is enabled,
11179 which will prefer to use the "tfdt" box to set DTS. Disable to use
11180 the "earliest_presentation_time" from the "sidx" box. In either
11181 case, the timestamp from the "mfra" box will be used if it's
11182 available and "use_mfra_for" is set to pts or dts.
11183 export_all
11185 Export unrecognized boxes within the udta box as metadata entries.
11186 The first four characters of the box type are set as the key.
11187 Default is false.
11188 export_xmp
11190 Export entire contents of XMP_ box and uuid box as a string with
11191 key "xmp". Note that if "export_all" is set and this option isn't,
11192 the contents of XMP_ box are still exported but with key "XMP_".
11193 Default is false.
11194 activation_bytes
11196 4-byte key required to decrypt Audible AAX and AAX+ files. See
11197 Audible AAX subsection below.
11198 audible_fixed_key
11200 Fixed key used for handling Audible AAX/AAX+ files. It has been
11201 pre-set so should not be necessary to specify.
11202 decryption_key
11204 16-byte key, in hex, to decrypt files encrypted using ISO Common
11205 Encryption (CENC/AES-128 CTR; ISO/IEC 23001-7).
11206 max_stts_delta
11208 Very high sample deltas written in a trak's stts box may
11209 occasionally be intended but usually they are written in error or
11210 used to store a negative value for dts correction when treated as
11211 signed 32-bit integers. This option lets the user set an upper
11212 limit, beyond which the delta is clamped to 1. Values greater than
11213 the limit if negative when cast to int32 are used to adjust onward
11214 dts.
11215 Unit is the track time scale. Range is 0 to UINT_MAX. Default is
11217 "UINT_MAX - 48000*10" which allows upto a 10 second dts correction
11218 for 48 kHz audio streams while accommodating 99.9% of "uint32"
11219 range.
11220 Audible AAX
11222 Audible AAX files are encrypted M4B files, and they can be decrypted by
11224 specifying a 4 byte activation secret.
11225 ffmpeg -activation_bytes 1CEB00DA -i test.aax -vn -c:a copy output.mp4
11227 mpegts
11229 MPEG-2 transport stream demuxer.
11230 This demuxer accepts the following options:
11232 resync_size
11234 Set size limit for looking up a new synchronization. Default value
11235 is 65536.
11236 skip_unknown_pmt
11238 Skip PMTs for programs not defined in the PAT. Default value is 0.
11239 fix_teletext_pts
11241 Override teletext packet PTS and DTS values with the timestamps
11242 calculated from the PCR of the first program which the teletext
11243 stream is part of and is not discarded. Default value is 1, set
11244 this option to 0 if you want your teletext packet PTS and DTS
11245 values untouched.
11246 ts_packetsize
11248 Output option carrying the raw packet size in bytes. Show the
11249 detected raw packet size, cannot be set by the user.
11250 scan_all_pmts
11252 Scan and combine all PMTs. The value is an integer with value from
11253 -1 to 1 (-1 means automatic setting, 1 means enabled, 0 means
11254 disabled). Default value is -1.
11255 merge_pmt_versions
11257 Re-use existing streams when a PMT's version is updated and
11258 elementary streams move to different PIDs. Default value is 0.
11259 max_packet_size
11261 Set maximum size, in bytes, of packet emitted by the demuxer.
11262 Payloads above this size are split across multiple packets. Range
11263 is 1 to INT_MAX/2. Default is 204800 bytes.
11264 mpjpeg
11266 MJPEG encapsulated in multi-part MIME demuxer.
11267 This demuxer allows reading of MJPEG, where each frame is represented
11269 as a part of multipart/x-mixed-replace stream.
11270 strict_mime_boundary
11272 Default implementation applies a relaxed standard to multi-part
11273 MIME boundary detection, to prevent regression with numerous
11274 existing endpoints not generating a proper MIME MJPEG stream.
11275 Turning this option on by setting it to 1 will result in a stricter
11276 check of the boundary value.
11277 rawvideo
11279 Raw video demuxer.
11280 This demuxer allows one to read raw video data. Since there is no
11282 header specifying the assumed video parameters, the user must specify
11283 them in order to be able to decode the data correctly.
11284 This demuxer accepts the following options:
11286 framerate
11288 Set input video frame rate. Default value is 25.
11289 pixel_format
11291 Set the input video pixel format. Default value is "yuv420p".
11292 video_size
11294 Set the input video size. This value must be specified explicitly.
11295 For example to read a rawvideo file input.raw with ffplay, assuming a
11297 pixel format of "rgb24", a video size of "320x240", and a frame rate of
11298 10 images per second, use the command:
11299 ffplay -f rawvideo -pixel_format rgb24 -video_size 320x240 -framerate 10 input.raw
11301 sbg
11303 SBaGen script demuxer.
11304 This demuxer reads the script language used by SBaGen
11306 <http://uazu.net/sbagen/> to generate binaural beats sessions. A SBG
11307 script looks like that:
11308 -SE
11310 a: 300-2.5/3 440+4.5/0
11311 b: 300-2.5/0 440+4.5/3
11312 off: -
11313 NOW == a
11314 +0:07:00 == b
11315 +0:14:00 == a
11316 +0:21:00 == b
11317 +0:30:00 off
11318 A SBG script can mix absolute and relative timestamps. If the script
11320 uses either only absolute timestamps (including the script start time)
11321 or only relative ones, then its layout is fixed, and the conversion is
11322 straightforward. On the other hand, if the script mixes both kind of
11323 timestamps, then the NOW reference for relative timestamps will be
11324 taken from the current time of day at the time the script is read, and
11325 the script layout will be frozen according to that reference. That
11326 means that if the script is directly played, the actual times will
11327 match the absolute timestamps up to the sound controller's clock
11328 accuracy, but if the user somehow pauses the playback or seeks, all
11329 times will be shifted accordingly.
11330 tedcaptions
11332 JSON captions used for <http://www.ted.com/>.
11333 TED does not provide links to the captions, but they can be guessed
11335 from the page. The file tools/bookmarklets.html from the FFmpeg source
11336 tree contains a bookmarklet to expose them.
11337 This demuxer accepts the following option:
11339 start_time
11341 Set the start time of the TED talk, in milliseconds. The default is
11342 15000 (15s). It is used to sync the captions with the downloadable
11343 videos, because they include a 15s intro.
11344 Example: convert the captions to a format most players understand:
11346 ffmpeg -i http://www.ted.com/talks/subtitles/id/1/lang/en talk1-en.srt
11348 vapoursynth
11350 Vapoursynth wrapper.
11351 Due to security concerns, Vapoursynth scripts will not be autodetected
11353 so the input format has to be forced. For ff* CLI tools, add "-f
11354 vapoursynth" before the input "-i yourscript.vpy".
11355 This demuxer accepts the following option:
11357 max_script_size
11359 The demuxer buffers the entire script into memory. Adjust this
11360 value to set the maximum buffer size, which in turn, acts as a
11361 ceiling for the size of scripts that can be read. Default is 1
11362 MiB.
11363
11365 Muxers are configured elements in FFmpeg which allow writing multimedia
11366 streams to a particular type of file.
11367 When you configure your FFmpeg build, all the supported muxers are
11369 enabled by default. You can list all available muxers using the
11370 configure option "--list-muxers".
11371 You can disable all the muxers with the configure option
11373 "--disable-muxers" and selectively enable / disable single muxers with
11374 the options "--enable-muxer=MUXER" / "--disable-muxer=MUXER".
11375 The option "-muxers" of the ff* tools will display the list of enabled
11377 muxers. Use "-formats" to view a combined list of enabled demuxers and
11378 muxers.
11379 A description of some of the currently available muxers follows.
11381 a64
11383 A64 muxer for Commodore 64 video. Accepts a single "a64_multi" or
11384 "a64_multi5" codec video stream.
11385 adts
11387 Audio Data Transport Stream muxer. It accepts a single AAC stream.
11388 Options
11390 It accepts the following options:
11392 write_id3v2 bool
11394 Enable to write ID3v2.4 tags at the start of the stream. Default is
11395 disabled.
11396 write_apetag bool
11398 Enable to write APE tags at the end of the stream. Default is
11399 disabled.
11400 write_mpeg2 bool
11402 Enable to set MPEG version bit in the ADTS frame header to 1 which
11403 indicates MPEG-2. Default is 0, which indicates MPEG-4.
11404 aiff
11406 Audio Interchange File Format muxer.
11407 Options
11409 It accepts the following options:
11411 write_id3v2
11413 Enable ID3v2 tags writing when set to 1. Default is 0 (disabled).
11414 id3v2_version
11416 Select ID3v2 version to write. Currently only version 3 and 4 (aka.
11417 ID3v2.3 and ID3v2.4) are supported. The default is version 4.
11418 alp
11420 Muxer for audio of High Voltage Software's Lego Racers game. It accepts
11421 a single ADPCM_IMA_ALP stream with no more than 2 channels nor a sample
11422 rate greater than 44100 Hz.
11423 Extensions: tun, pcm
11425 Options
11427 It accepts the following options:
11429 type type
11431 Set file type.
11432 tun Set file type as music. Must have a sample rate of 22050 Hz.
11434 pcm Set file type as sfx.
11436 auto
11438 Set file type as per output file extension. ".pcm" results in
11439 type "pcm" else type "tun" is set. (default)
11440 asf
11442 Advanced Systems Format muxer.
11443 Note that Windows Media Audio (wma) and Windows Media Video (wmv) use
11445 this muxer too.
11446 Options
11448 It accepts the following options:
11450 packet_size
11452 Set the muxer packet size. By tuning this setting you may reduce
11453 data fragmentation or muxer overhead depending on your source.
11454 Default value is 3200, minimum is 100, maximum is 64k.
11455 avi
11457 Audio Video Interleaved muxer.
11458 Options
11460 It accepts the following options:
11462 reserve_index_space
11464 Reserve the specified amount of bytes for the OpenDML master index
11465 of each stream within the file header. By default additional master
11466 indexes are embedded within the data packets if there is no space
11467 left in the first master index and are linked together as a chain
11468 of indexes. This index structure can cause problems for some use
11469 cases, e.g. third-party software strictly relying on the OpenDML
11470 index specification or when file seeking is slow. Reserving enough
11471 index space in the file header avoids these problems.
11472 The required index space depends on the output file size and should
11474 be about 16 bytes per gigabyte. When this option is omitted or set
11475 to zero the necessary index space is guessed.
11476 write_channel_mask
11478 Write the channel layout mask into the audio stream header.
11479 This option is enabled by default. Disabling the channel mask can
11481 be useful in specific scenarios, e.g. when merging multiple audio
11482 streams into one for compatibility with software that only supports
11483 a single audio stream in AVI (see the "amerge" section in the
11484 ffmpeg-filters manual).
11485 flipped_raw_rgb
11487 If set to true, store positive height for raw RGB bitmaps, which
11488 indicates bitmap is stored bottom-up. Note that this option does
11489 not flip the bitmap which has to be done manually beforehand, e.g.
11490 by using the vflip filter. Default is false and indicates bitmap
11491 is stored top down.
11492 chromaprint
11494 Chromaprint fingerprinter.
11495 This muxer feeds audio data to the Chromaprint library, which generates
11497 a fingerprint for the provided audio data. See
11498 <https://acoustid.org/chromaprint>
11499 It takes a single signed native-endian 16-bit raw audio stream of at
11501 most 2 channels.
11502 Options
11504 silence_threshold
11506 Threshold for detecting silence. Range is from -1 to 32767, where
11507 -1 disables silence detection. Silence detection can only be used
11508 with version 3 of the algorithm. Silence detection must be
11509 disabled for use with the AcoustID service. Default is -1.
11510 algorithm
11512 Version of algorithm to fingerprint with. Range is 0 to 4. Version
11513 3 enables silence detection. Default is 1.
11514 fp_format
11516 Format to output the fingerprint as. Accepts the following options:
11517 raw Binary raw fingerprint
11519 compressed
11521 Binary compressed fingerprint
11522 base64
11524 Base64 compressed fingerprint (default)
11525 crc
11527 CRC (Cyclic Redundancy Check) testing format.
11528 This muxer computes and prints the Adler-32 CRC of all the input audio
11530 and video frames. By default audio frames are converted to signed
11531 16-bit raw audio and video frames to raw video before computing the
11532 CRC.
11533 The output of the muxer consists of a single line of the form:
11535 CRC=0xCRC, where CRC is a hexadecimal number 0-padded to 8 digits
11536 containing the CRC for all the decoded input frames.
11537 See also the framecrc muxer.
11539 Examples
11541 For example to compute the CRC of the input, and store it in the file
11543 out.crc:
11544 ffmpeg -i INPUT -f crc out.crc
11546 You can print the CRC to stdout with the command:
11548 ffmpeg -i INPUT -f crc -
11550 You can select the output format of each frame with ffmpeg by
11552 specifying the audio and video codec and format. For example to compute
11553 the CRC of the input audio converted to PCM unsigned 8-bit and the
11554 input video converted to MPEG-2 video, use the command:
11555 ffmpeg -i INPUT -c:a pcm_u8 -c:v mpeg2video -f crc -
11557 dash
11559 Dynamic Adaptive Streaming over HTTP (DASH) muxer that creates segments
11560 and manifest files according to the MPEG-DASH standard ISO/IEC
11561 23009-1:2014.
11562 For more information see:
11564 • ISO DASH Specification:
11566 <http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip>
11567 • WebM DASH Specification:
11569 <https://sites.google.com/a/webmproject.org/wiki/adaptive-streaming/webm-dash-specification>
11570 It creates a MPD manifest file and segment files for each stream.
11572 The segment filename might contain pre-defined identifiers used with
11574 SegmentTemplate as defined in section 5.3.9.4.4 of the standard.
11575 Available identifiers are "$RepresentationID$", "$Number$",
11576 "$Bandwidth$" and "$Time$". In addition to the standard identifiers,
11577 an ffmpeg-specific "$ext$" identifier is also supported. When
11578 specified ffmpeg will replace $ext$ in the file name with muxing
11579 format's extensions such as mp4, webm etc.,
11580 ffmpeg -re -i <input> -map 0 -map 0 -c:a libfdk_aac -c:v libx264 \
11582 -b:v:0 800k -b:v:1 300k -s:v:1 320x170 -profile:v:1 baseline \
11583 -profile:v:0 main -bf 1 -keyint_min 120 -g 120 -sc_threshold 0 \
11584 -b_strategy 0 -ar:a:1 22050 -use_timeline 1 -use_template 1 \
11585 -window_size 5 -adaptation_sets "id=0,streams=v id=1,streams=a" \
11586 -f dash /path/to/out.mpd
11587 seg_duration duration
11589 Set the segment length in seconds (fractional value can be set).
11590 The value is treated as average segment duration when use_template
11591 is enabled and use_timeline is disabled and as minimum segment
11592 duration for all the other use cases.
11593 frag_duration duration
11595 Set the length in seconds of fragments within segments (fractional
11596 value can be set).
11597 frag_type type
11599 Set the type of interval for fragmentation.
11600 window_size size
11602 Set the maximum number of segments kept in the manifest.
11603 extra_window_size size
11605 Set the maximum number of segments kept outside of the manifest
11606 before removing from disk.
11607 remove_at_exit remove
11609 Enable (1) or disable (0) removal of all segments when finished.
11610 use_template template
11612 Enable (1) or disable (0) use of SegmentTemplate instead of
11613 SegmentList.
11614 use_timeline timeline
11616 Enable (1) or disable (0) use of SegmentTimeline in
11617 SegmentTemplate.
11618 single_file single_file
11620 Enable (1) or disable (0) storing all segments in one file,
11621 accessed using byte ranges.
11622 single_file_name file_name
11624 DASH-templated name to be used for baseURL. Implies single_file set
11625 to "1". In the template, "$ext$" is replaced with the file name
11626 extension specific for the segment format.
11627 init_seg_name init_name
11629 DASH-templated name to used for the initialization segment. Default
11630 is "init-stream$RepresentationID$.$ext$". "$ext$" is replaced with
11631 the file name extension specific for the segment format.
11632 media_seg_name segment_name
11634 DASH-templated name to used for the media segments. Default is
11635 "chunk-stream$RepresentationID$-$Number%05d$.$ext$". "$ext$" is
11636 replaced with the file name extension specific for the segment
11637 format.
11638 utc_timing_url utc_url
11640 URL of the page that will return the UTC timestamp in ISO format.
11641 Example: "https://time.akamai.com/?iso"
11642 method method
11644 Use the given HTTP method to create output files. Generally set to
11645 PUT or POST.
11646 http_user_agent user_agent
11648 Override User-Agent field in HTTP header. Applicable only for HTTP
11649 output.
11650 http_persistent http_persistent
11652 Use persistent HTTP connections. Applicable only for HTTP output.
11653 hls_playlist hls_playlist
11655 Generate HLS playlist files as well. The master playlist is
11656 generated with the filename hls_master_name. One media playlist
11657 file is generated for each stream with filenames media_0.m3u8,
11658 media_1.m3u8, etc.
11659 hls_master_name file_name
11661 HLS master playlist name. Default is "master.m3u8".
11662 streaming streaming
11664 Enable (1) or disable (0) chunk streaming mode of output. In chunk
11665 streaming mode, each frame will be a moof fragment which forms a
11666 chunk.
11667 adaptation_sets adaptation_sets
11669 Assign streams to AdaptationSets. Syntax is "id=x,streams=a,b,c
11670 id=y,streams=d,e" with x and y being the IDs of the adaptation sets
11671 and a,b,c,d and e are the indices of the mapped streams.
11672 To map all video (or audio) streams to an AdaptationSet, "v" (or
11674 "a") can be used as stream identifier instead of IDs.
11675 When no assignment is defined, this defaults to an AdaptationSet
11677 for each stream.
11678 Optional syntax is
11680 "id=x,seg_duration=x,frag_duration=x,frag_type=type,descriptor=descriptor_string,streams=a,b,c
11681 id=y,seg_duration=y,frag_type=type,streams=d,e" and so on,
11682 descriptor is useful to the scheme defined by ISO/IEC
11683 23009-1:2014/Amd.2:2015. For example, -adaptation_sets
11684 "id=0,descriptor=<SupplementalProperty
11685 schemeIdUri=\"urn:mpeg:dash:srd:2014\"
11686 value=\"0,0,0,1,1,2,2\"/>,streams=v". Please note that descriptor
11687 string should be a self-closing xml tag. seg_duration,
11688 frag_duration and frag_type override the global option values for
11689 each adaptation set. For example, -adaptation_sets
11690 "id=0,seg_duration=2,frag_duration=1,frag_type=duration,streams=v
11691 id=1,seg_duration=2,frag_type=none,streams=a" type_id marks an
11692 adaptation set as containing streams meant to be used for Trick
11693 Mode for the referenced adaptation set. For example,
11694 -adaptation_sets "id=0,seg_duration=2,frag_type=none,streams=0
11695 id=1,seg_duration=10,frag_type=none,trick_id=0,streams=1"
11696 timeout timeout
11698 Set timeout for socket I/O operations. Applicable only for HTTP
11699 output.
11700 index_correction index_correction
11702 Enable (1) or Disable (0) segment index correction logic.
11703 Applicable only when use_template is enabled and use_timeline is
11704 disabled.
11705 When enabled, the logic monitors the flow of segment indexes. If a
11707 streams's segment index value is not at the expected real time
11708 position, then the logic corrects that index value.
11709 Typically this logic is needed in live streaming use cases. The
11711 network bandwidth fluctuations are common during long run
11712 streaming. Each fluctuation can cause the segment indexes fall
11713 behind the expected real time position.
11714 format_options options_list
11716 Set container format (mp4/webm) options using a ":" separated list
11717 of key=value parameters. Values containing ":" special characters
11718 must be escaped.
11719 global_sidx global_sidx
11721 Write global SIDX atom. Applicable only for single file, mp4
11722 output, non-streaming mode.
11723 dash_segment_type dash_segment_type
11725 Possible values:
11726 auto
11728 If this flag is set, the dash segment files format will be
11729 selected based on the stream codec. This is the default mode.
11730 mp4 If this flag is set, the dash segment files will be in in
11732 ISOBMFF format.
11733 webm
11735 If this flag is set, the dash segment files will be in in WebM
11736 format.
11737 ignore_io_errors ignore_io_errors
11739 Ignore IO errors during open and write. Useful for long-duration
11740 runs with network output.
11741 lhls lhls
11743 Enable Low-latency HLS(LHLS). Adds #EXT-X-PREFETCH tag with current
11744 segment's URI. hls.js player folks are trying to standardize an
11745 open LHLS spec. The draft spec is available in
11746 https://github.com/video-dev/hlsjs-rfcs/blob/lhls-spec/proposals/0001-lhls.md
11747 This option tries to comply with the above open spec. It enables
11748 streaming and hls_playlist options automatically. This is an
11749 experimental feature.
11750 Note: This is not Apple's version LHLS. See
11752 <https://datatracker.ietf.org/doc/html/draft-pantos-hls-rfc8216bis>
11753 ldash ldash
11755 Enable Low-latency Dash by constraining the presence and values of
11756 some elements.
11757 master_m3u8_publish_rate master_m3u8_publish_rate
11759 Publish master playlist repeatedly every after specified number of
11760 segment intervals.
11761 write_prft write_prft
11763 Write Producer Reference Time elements on supported streams. This
11764 also enables writing prft boxes in the underlying muxer. Applicable
11765 only when the utc_url option is enabled. It's set to auto by
11766 default, in which case the muxer will attempt to enable it only in
11767 modes that require it.
11768 mpd_profile mpd_profile
11770 Set one or more manifest profiles.
11771 http_opts http_opts
11773 A :-separated list of key=value options to pass to the underlying
11774 HTTP protocol. Applicable only for HTTP output.
11775 target_latency target_latency
11777 Set an intended target latency in seconds (fractional value can be
11778 set) for serving. Applicable only when streaming and write_prft
11779 options are enabled. This is an informative fields clients can use
11780 to measure the latency of the service.
11781 min_playback_rate min_playback_rate
11783 Set the minimum playback rate indicated as appropriate for the
11784 purposes of automatically adjusting playback latency and buffer
11785 occupancy during normal playback by clients.
11786 max_playback_rate max_playback_rate
11788 Set the maximum playback rate indicated as appropriate for the
11789 purposes of automatically adjusting playback latency and buffer
11790 occupancy during normal playback by clients.
11791 update_period update_period
11793 Set the mpd update period ,for dynamic content.
11794 The unit is second.
11795 fifo
11797 The fifo pseudo-muxer allows the separation of encoding and muxing by
11798 using first-in-first-out queue and running the actual muxer in a
11799 separate thread. This is especially useful in combination with the tee
11800 muxer and can be used to send data to several destinations with
11801 different reliability/writing speed/latency.
11802 API users should be aware that callback functions (interrupt_callback,
11804 io_open and io_close) used within its AVFormatContext must be thread-
11805 safe.
11806 The behavior of the fifo muxer if the queue fills up or if the output
11808 fails is selectable,
11809 • output can be transparently restarted with configurable delay
11811 between retries based on real time or time of the processed stream.
11812 • encoding can be blocked during temporary failure, or continue
11814 transparently dropping packets in case fifo queue fills up.
11815 fifo_format
11817 Specify the format name. Useful if it cannot be guessed from the
11818 output name suffix.
11819 queue_size
11821 Specify size of the queue (number of packets). Default value is 60.
11822 format_opts
11824 Specify format options for the underlying muxer. Muxer options can
11825 be specified as a list of key=value pairs separated by ':'.
11826 drop_pkts_on_overflow bool
11828 If set to 1 (true), in case the fifo queue fills up, packets will
11829 be dropped rather than blocking the encoder. This makes it possible
11830 to continue streaming without delaying the input, at the cost of
11831 omitting part of the stream. By default this option is set to 0
11832 (false), so in such cases the encoder will be blocked until the
11833 muxer processes some of the packets and none of them is lost.
11834 attempt_recovery bool
11836 If failure occurs, attempt to recover the output. This is
11837 especially useful when used with network output, since it makes it
11838 possible to restart streaming transparently. By default this
11839 option is set to 0 (false).
11840 max_recovery_attempts
11842 Sets maximum number of successive unsuccessful recovery attempts
11843 after which the output fails permanently. By default this option is
11844 set to 0 (unlimited).
11845 recovery_wait_time duration
11847 Waiting time before the next recovery attempt after previous
11848 unsuccessful recovery attempt. Default value is 5 seconds.
11849 recovery_wait_streamtime bool
11851 If set to 0 (false), the real time is used when waiting for the
11852 recovery attempt (i.e. the recovery will be attempted after at
11853 least recovery_wait_time seconds). If set to 1 (true), the time of
11854 the processed stream is taken into account instead (i.e. the
11855 recovery will be attempted after at least recovery_wait_time
11856 seconds of the stream is omitted). By default, this option is set
11857 to 0 (false).
11858 recover_any_error bool
11860 If set to 1 (true), recovery will be attempted regardless of type
11861 of the error causing the failure. By default this option is set to
11862 0 (false) and in case of certain (usually permanent) errors the
11863 recovery is not attempted even when attempt_recovery is set to 1.
11864 restart_with_keyframe bool
11866 Specify whether to wait for the keyframe after recovering from
11867 queue overflow or failure. This option is set to 0 (false) by
11868 default.
11869 timeshift duration
11871 Buffer the specified amount of packets and delay writing the
11872 output. Note that queue_size must be big enough to store the
11873 packets for timeshift. At the end of the input the fifo buffer is
11874 flushed at realtime speed.
11875 Examples
11877 • Stream something to rtmp server, continue processing the stream at
11879 real-time rate even in case of temporary failure (network outage)
11880 and attempt to recover streaming every second indefinitely.
11881 ffmpeg -re -i ... -c:v libx264 -c:a aac -f fifo -fifo_format flv -map 0:v -map 0:a
11883 -drop_pkts_on_overflow 1 -attempt_recovery 1 -recovery_wait_time 1 rtmp://example.com/live/stream_name
11884 flv
11886 Adobe Flash Video Format muxer.
11887 This muxer accepts the following options:
11889 flvflags flags
11891 Possible values:
11892 aac_seq_header_detect
11894 Place AAC sequence header based on audio stream data.
11895 no_sequence_end
11897 Disable sequence end tag.
11898 no_metadata
11900 Disable metadata tag.
11901 no_duration_filesize
11903 Disable duration and filesize in metadata when they are equal
11904 to zero at the end of stream. (Be used to non-seekable living
11905 stream).
11906 add_keyframe_index
11908 Used to facilitate seeking; particularly for HTTP pseudo
11909 streaming.
11910 framecrc
11912 Per-packet CRC (Cyclic Redundancy Check) testing format.
11913 This muxer computes and prints the Adler-32 CRC for each audio and
11915 video packet. By default audio frames are converted to signed 16-bit
11916 raw audio and video frames to raw video before computing the CRC.
11917 The output of the muxer consists of a line for each audio and video
11919 packet of the form:
11920 <stream_index>, <packet_dts>, <packet_pts>, <packet_duration>, <packet_size>, 0x<CRC>
11922 CRC is a hexadecimal number 0-padded to 8 digits containing the CRC of
11924 the packet.
11925 Examples
11927 For example to compute the CRC of the audio and video frames in INPUT,
11929 converted to raw audio and video packets, and store it in the file
11930 out.crc:
11931 ffmpeg -i INPUT -f framecrc out.crc
11933 To print the information to stdout, use the command:
11935 ffmpeg -i INPUT -f framecrc -
11937 With ffmpeg, you can select the output format to which the audio and
11939 video frames are encoded before computing the CRC for each packet by
11940 specifying the audio and video codec. For example, to compute the CRC
11941 of each decoded input audio frame converted to PCM unsigned 8-bit and
11942 of each decoded input video frame converted to MPEG-2 video, use the
11943 command:
11944 ffmpeg -i INPUT -c:a pcm_u8 -c:v mpeg2video -f framecrc -
11946 See also the crc muxer.
11948 framehash
11950 Per-packet hash testing format.
11951 This muxer computes and prints a cryptographic hash for each audio and
11953 video packet. This can be used for packet-by-packet equality checks
11954 without having to individually do a binary comparison on each.
11955 By default audio frames are converted to signed 16-bit raw audio and
11957 video frames to raw video before computing the hash, but the output of
11958 explicit conversions to other codecs can also be used. It uses the
11959 SHA-256 cryptographic hash function by default, but supports several
11960 other algorithms.
11961 The output of the muxer consists of a line for each audio and video
11963 packet of the form:
11964 <stream_index>, <packet_dts>, <packet_pts>, <packet_duration>, <packet_size>, <hash>
11966 hash is a hexadecimal number representing the computed hash for the
11968 packet.
11969 hash algorithm
11971 Use the cryptographic hash function specified by the string
11972 algorithm. Supported values include "MD5", "murmur3", "RIPEMD128",
11973 "RIPEMD160", "RIPEMD256", "RIPEMD320", "SHA160", "SHA224", "SHA256"
11974 (default), "SHA512/224", "SHA512/256", "SHA384", "SHA512", "CRC32"
11975 and "adler32".
11976 Examples
11978 To compute the SHA-256 hash of the audio and video frames in INPUT,
11980 converted to raw audio and video packets, and store it in the file
11981 out.sha256:
11982 ffmpeg -i INPUT -f framehash out.sha256
11984 To print the information to stdout, using the MD5 hash function, use
11986 the command:
11987 ffmpeg -i INPUT -f framehash -hash md5 -
11989 See also the hash muxer.
11991 framemd5
11993 Per-packet MD5 testing format.
11994 This is a variant of the framehash muxer. Unlike that muxer, it
11996 defaults to using the MD5 hash function.
11997 Examples
11999 To compute the MD5 hash of the audio and video frames in INPUT,
12001 converted to raw audio and video packets, and store it in the file
12002 out.md5:
12003 ffmpeg -i INPUT -f framemd5 out.md5
12005 To print the information to stdout, use the command:
12007 ffmpeg -i INPUT -f framemd5 -
12009 See also the framehash and md5 muxers.
12011 gif
12013 Animated GIF muxer.
12014 It accepts the following options:
12016 loop
12018 Set the number of times to loop the output. Use -1 for no loop, 0
12019 for looping indefinitely (default).
12020 final_delay
12022 Force the delay (expressed in centiseconds) after the last frame.
12023 Each frame ends with a delay until the next frame. The default is
12024 -1, which is a special value to tell the muxer to re-use the
12025 previous delay. In case of a loop, you might want to customize this
12026 value to mark a pause for instance.
12027 For example, to encode a gif looping 10 times, with a 5 seconds delay
12029 between the loops:
12030 ffmpeg -i INPUT -loop 10 -final_delay 500 out.gif
12032 Note 1: if you wish to extract the frames into separate GIF files, you
12034 need to force the image2 muxer:
12035 ffmpeg -i INPUT -c:v gif -f image2 "out%d.gif"
12037 Note 2: the GIF format has a very large time base: the delay between
12039 two frames can therefore not be smaller than one centi second.
12040 hash
12042 Hash testing format.
12043 This muxer computes and prints a cryptographic hash of all the input
12045 audio and video frames. This can be used for equality checks without
12046 having to do a complete binary comparison.
12047 By default audio frames are converted to signed 16-bit raw audio and
12049 video frames to raw video before computing the hash, but the output of
12050 explicit conversions to other codecs can also be used. Timestamps are
12051 ignored. It uses the SHA-256 cryptographic hash function by default,
12052 but supports several other algorithms.
12053 The output of the muxer consists of a single line of the form:
12055 algo=hash, where algo is a short string representing the hash function
12056 used, and hash is a hexadecimal number representing the computed hash.
12057 hash algorithm
12059 Use the cryptographic hash function specified by the string
12060 algorithm. Supported values include "MD5", "murmur3", "RIPEMD128",
12061 "RIPEMD160", "RIPEMD256", "RIPEMD320", "SHA160", "SHA224", "SHA256"
12062 (default), "SHA512/224", "SHA512/256", "SHA384", "SHA512", "CRC32"
12063 and "adler32".
12064 Examples
12066 To compute the SHA-256 hash of the input converted to raw audio and
12068 video, and store it in the file out.sha256:
12069 ffmpeg -i INPUT -f hash out.sha256
12071 To print an MD5 hash to stdout use the command:
12073 ffmpeg -i INPUT -f hash -hash md5 -
12075 See also the framehash muxer.
12077 hls
12079 Apple HTTP Live Streaming muxer that segments MPEG-TS according to the
12080 HTTP Live Streaming (HLS) specification.
12081 It creates a playlist file, and one or more segment files. The output
12083 filename specifies the playlist filename.
12084 By default, the muxer creates a file for each segment produced. These
12086 files have the same name as the playlist, followed by a sequential
12087 number and a .ts extension.
12088 Make sure to require a closed GOP when encoding and to set the GOP size
12090 to fit your segment time constraint.
12091 For example, to convert an input file with ffmpeg:
12093 ffmpeg -i in.mkv -c:v h264 -flags +cgop -g 30 -hls_time 1 out.m3u8
12095 This example will produce the playlist, out.m3u8, and segment files:
12097 out0.ts, out1.ts, out2.ts, etc.
12098 See also the segment muxer, which provides a more generic and flexible
12100 implementation of a segmenter, and can be used to perform HLS
12101 segmentation.
12102 Options
12104 This muxer supports the following options:
12106 hls_init_time duration
12108 Set the initial target segment length. Default value is 0.
12109 duration must be a time duration specification, see the Time
12111 duration section in the ffmpeg-utils(1) manual.
12112 Segment will be cut on the next key frame after this time has
12114 passed on the first m3u8 list. After the initial playlist is
12115 filled ffmpeg will cut segments at duration equal to "hls_time"
12116 hls_time duration
12118 Set the target segment length. Default value is 2.
12119 duration must be a time duration specification, see the Time
12121 duration section in the ffmpeg-utils(1) manual. Segment will be
12122 cut on the next key frame after this time has passed.
12123 hls_list_size size
12125 Set the maximum number of playlist entries. If set to 0 the list
12126 file will contain all the segments. Default value is 5.
12127 hls_delete_threshold size
12129 Set the number of unreferenced segments to keep on disk before
12130 "hls_flags delete_segments" deletes them. Increase this to allow
12131 continue clients to download segments which were recently
12132 referenced in the playlist. Default value is 1, meaning segments
12133 older than "hls_list_size+1" will be deleted.
12134 hls_start_number_source
12136 Start the playlist sequence number ("#EXT-X-MEDIA-SEQUENCE")
12137 according to the specified source. Unless "hls_flags single_file"
12138 is set, it also specifies source of starting sequence numbers of
12139 segment and subtitle filenames. In any case, if "hls_flags
12140 append_list" is set and read playlist sequence number is greater
12141 than the specified start sequence number, then that value will be
12142 used as start value.
12143 It accepts the following values:
12145 generic (default)
12147 Set the starting sequence numbers according to start_number
12148 option value.
12149 epoch
12151 The start number will be the seconds since epoch (1970-01-01
12152 00:00:00)
12153 epoch_us
12155 The start number will be the microseconds since epoch
12156 (1970-01-01 00:00:00)
12157 datetime
12159 The start number will be based on the current date/time as
12160 YYYYmmddHHMMSS. e.g. 20161231235759.
12161 start_number number
12163 Start the playlist sequence number ("#EXT-X-MEDIA-SEQUENCE") from
12164 the specified number when hls_start_number_source value is generic.
12165 (This is the default case.) Unless "hls_flags single_file" is set,
12166 it also specifies starting sequence numbers of segment and subtitle
12167 filenames. Default value is 0.
12168 hls_allow_cache allowcache
12170 Explicitly set whether the client MAY (1) or MUST NOT (0) cache
12171 media segments.
12172 hls_base_url baseurl
12174 Append baseurl to every entry in the playlist. Useful to generate
12175 playlists with absolute paths.
12176 Note that the playlist sequence number must be unique for each
12178 segment and it is not to be confused with the segment filename
12179 sequence number which can be cyclic, for example if the wrap option
12180 is specified.
12181 hls_segment_filename filename
12183 Set the segment filename. Unless "hls_flags single_file" is set,
12184 filename is used as a string format with the segment number:
12185 ffmpeg -i in.nut -hls_segment_filename 'file%03d.ts' out.m3u8
12187 This example will produce the playlist, out.m3u8, and segment
12189 files: file000.ts, file001.ts, file002.ts, etc.
12190 filename may contain full path or relative path specification, but
12192 only the file name part without any path info will be contained in
12193 the m3u8 segment list. Should a relative path be specified, the
12194 path of the created segment files will be relative to the current
12195 working directory. When strftime_mkdir is set, the whole expanded
12196 value of filename will be written into the m3u8 segment list.
12197 When "var_stream_map" is set with two or more variant streams, the
12199 filename pattern must contain the string "%v", this string
12200 specifies the position of variant stream index in the generated
12201 segment file names.
12202 ffmpeg -i in.ts -b:v:0 1000k -b:v:1 256k -b:a:0 64k -b:a:1 32k \
12204 -map 0:v -map 0:a -map 0:v -map 0:a -f hls -var_stream_map "v:0,a:0 v:1,a:1" \
12205 -hls_segment_filename 'file_%v_%03d.ts' out_%v.m3u8
12206 This example will produce the playlists segment file sets:
12208 file_0_000.ts, file_0_001.ts, file_0_002.ts, etc. and
12209 file_1_000.ts, file_1_001.ts, file_1_002.ts, etc.
12210 The string "%v" may be present in the filename or in the last
12212 directory name containing the file, but only in one of them.
12213 (Additionally, %v may appear multiple times in the last sub-
12214 directory or filename.) If the string %v is present in the
12215 directory name, then sub-directories are created after expanding
12216 the directory name pattern. This enables creation of segments
12217 corresponding to different variant streams in subdirectories.
12218 ffmpeg -i in.ts -b:v:0 1000k -b:v:1 256k -b:a:0 64k -b:a:1 32k \
12220 -map 0:v -map 0:a -map 0:v -map 0:a -f hls -var_stream_map "v:0,a:0 v:1,a:1" \
12221 -hls_segment_filename 'vs%v/file_%03d.ts' vs%v/out.m3u8
12222 This example will produce the playlists segment file sets:
12224 vs0/file_000.ts, vs0/file_001.ts, vs0/file_002.ts, etc. and
12225 vs1/file_000.ts, vs1/file_001.ts, vs1/file_002.ts, etc.
12226 strftime
12228 Use strftime() on filename to expand the segment filename with
12229 localtime. The segment number is also available in this mode, but
12230 to use it, you need to specify second_level_segment_index hls_flag
12231 and %%d will be the specifier.
12232 ffmpeg -i in.nut -strftime 1 -hls_segment_filename 'file-%Y%m%d-%s.ts' out.m3u8
12234 This example will produce the playlist, out.m3u8, and segment
12236 files: file-20160215-1455569023.ts, file-20160215-1455569024.ts,
12237 etc. Note: On some systems/environments, the %s specifier is not
12238 available. See
12239 strftime() documentation.
12240 ffmpeg -i in.nut -strftime 1 -hls_flags second_level_segment_index -hls_segment_filename 'file-%Y%m%d-%%04d.ts' out.m3u8
12242 This example will produce the playlist, out.m3u8, and segment
12244 files: file-20160215-0001.ts, file-20160215-0002.ts, etc.
12245 strftime_mkdir
12247 Used together with -strftime_mkdir, it will create all
12248 subdirectories which is expanded in filename.
12249 ffmpeg -i in.nut -strftime 1 -strftime_mkdir 1 -hls_segment_filename '%Y%m%d/file-%Y%m%d-%s.ts' out.m3u8
12251 This example will create a directory 201560215 (if it does not
12253 exist), and then produce the playlist, out.m3u8, and segment files:
12254 20160215/file-20160215-1455569023.ts,
12255 20160215/file-20160215-1455569024.ts, etc.
12256 ffmpeg -i in.nut -strftime 1 -strftime_mkdir 1 -hls_segment_filename '%Y/%m/%d/file-%Y%m%d-%s.ts' out.m3u8
12258 This example will create a directory hierarchy 2016/02/15 (if any
12260 of them do not exist), and then produce the playlist, out.m3u8, and
12261 segment files: 2016/02/15/file-20160215-1455569023.ts,
12262 2016/02/15/file-20160215-1455569024.ts, etc.
12263 hls_segment_options options_list
12265 Set output format options using a :-separated list of key=value
12266 parameters. Values containing ":" special characters must be
12267 escaped.
12268 hls_key_info_file key_info_file
12270 Use the information in key_info_file for segment encryption. The
12271 first line of key_info_file specifies the key URI written to the
12272 playlist. The key URL is used to access the encryption key during
12273 playback. The second line specifies the path to the key file used
12274 to obtain the key during the encryption process. The key file is
12275 read as a single packed array of 16 octets in binary format. The
12276 optional third line specifies the initialization vector (IV) as a
12277 hexadecimal string to be used instead of the segment sequence
12278 number (default) for encryption. Changes to key_info_file will
12279 result in segment encryption with the new key/IV and an entry in
12280 the playlist for the new key URI/IV if "hls_flags periodic_rekey"
12281 is enabled.
12282 Key info file format:
12284 <key URI>
12286 <key file path>
12287 <IV> (optional)
12288 Example key URIs:
12290 http://server/file.key
12292 /path/to/file.key
12293 file.key
12294 Example key file paths:
12296 file.key
12298 /path/to/file.key
12299 Example IV:
12301 0123456789ABCDEF0123456789ABCDEF
12303 Key info file example:
12305 http://server/file.key
12307 /path/to/file.key
12308 0123456789ABCDEF0123456789ABCDEF
12309 Example shell script:
12311 #!/bin/sh
12313 BASE_URL=${1:-'.'}
12314 openssl rand 16 > file.key
12315 echo $BASE_URL/file.key > file.keyinfo
12316 echo file.key >> file.keyinfo
12317 echo $(openssl rand -hex 16) >> file.keyinfo
12318 ffmpeg -f lavfi -re -i testsrc -c:v h264 -hls_flags delete_segments \
12319 -hls_key_info_file file.keyinfo out.m3u8
12320 -hls_enc enc
12322 Enable (1) or disable (0) the AES128 encryption. When enabled
12323 every segment generated is encrypted and the encryption key is
12324 saved as playlist name.key.
12325 -hls_enc_key key
12327 16-octet key to encrypt the segments, by default it is randomly
12328 generated.
12329 -hls_enc_key_url keyurl
12331 If set, keyurl is prepended instead of baseurl to the key filename
12332 in the playlist.
12333 -hls_enc_iv iv
12335 16-octet initialization vector for every segment instead of the
12336 autogenerated ones.
12337 hls_segment_type flags
12339 Possible values:
12340 mpegts
12342 Output segment files in MPEG-2 Transport Stream format. This is
12343 compatible with all HLS versions.
12344 fmp4
12346 Output segment files in fragmented MP4 format, similar to MPEG-
12347 DASH. fmp4 files may be used in HLS version 7 and above.
12348 hls_fmp4_init_filename filename
12350 Set filename to the fragment files header file, default filename is
12351 init.mp4.
12352 Use "-strftime 1" on filename to expand the segment filename with
12354 localtime.
12355 ffmpeg -i in.nut -hls_segment_type fmp4 -strftime 1 -hls_fmp4_init_filename "%s_init.mp4" out.m3u8
12357 This will produce init like this 1602678741_init.mp4
12359 hls_fmp4_init_resend
12361 Resend init file after m3u8 file refresh every time, default is 0.
12362 When "var_stream_map" is set with two or more variant streams, the
12364 filename pattern must contain the string "%v", this string
12365 specifies the position of variant stream index in the generated
12366 init file names. The string "%v" may be present in the filename or
12367 in the last directory name containing the file. If the string is
12368 present in the directory name, then sub-directories are created
12369 after expanding the directory name pattern. This enables creation
12370 of init files corresponding to different variant streams in
12371 subdirectories.
12372 hls_flags flags
12374 Possible values:
12375 single_file
12377 If this flag is set, the muxer will store all segments in a
12378 single MPEG-TS file, and will use byte ranges in the playlist.
12379 HLS playlists generated with this way will have the version
12380 number 4. For example:
12381 ffmpeg -i in.nut -hls_flags single_file out.m3u8
12383 Will produce the playlist, out.m3u8, and a single segment file,
12385 out.ts.
12386 delete_segments
12388 Segment files removed from the playlist are deleted after a
12389 period of time equal to the duration of the segment plus the
12390 duration of the playlist.
12391 append_list
12393 Append new segments into the end of old segment list, and
12394 remove the "#EXT-X-ENDLIST" from the old segment list.
12395 round_durations
12397 Round the duration info in the playlist file segment info to
12398 integer values, instead of using floating point. If there are
12399 no other features requiring higher HLS versions be used, then
12400 this will allow ffmpeg to output a HLS version 2 m3u8.
12401 discont_start
12403 Add the "#EXT-X-DISCONTINUITY" tag to the playlist, before the
12404 first segment's information.
12405 omit_endlist
12407 Do not append the "EXT-X-ENDLIST" tag at the end of the
12408 playlist.
12409 periodic_rekey
12411 The file specified by "hls_key_info_file" will be checked
12412 periodically and detect updates to the encryption info. Be sure
12413 to replace this file atomically, including the file containing
12414 the AES encryption key.
12415 independent_segments
12417 Add the "#EXT-X-INDEPENDENT-SEGMENTS" to playlists that has
12418 video segments and when all the segments of that playlist are
12419 guaranteed to start with a Key frame.
12420 iframes_only
12422 Add the "#EXT-X-I-FRAMES-ONLY" to playlists that has video
12423 segments and can play only I-frames in the "#EXT-X-BYTERANGE"
12424 mode.
12425 split_by_time
12427 Allow segments to start on frames other than keyframes. This
12428 improves behavior on some players when the time between
12429 keyframes is inconsistent, but may make things worse on others,
12430 and can cause some oddities during seeking. This flag should be
12431 used with the "hls_time" option.
12432 program_date_time
12434 Generate "EXT-X-PROGRAM-DATE-TIME" tags.
12435 second_level_segment_index
12437 Makes it possible to use segment indexes as %%d in
12438 hls_segment_filename expression besides date/time values when
12439 strftime is on. To get fixed width numbers with trailing
12440 zeroes, %%0xd format is available where x is the required
12441 width.
12442 second_level_segment_size
12444 Makes it possible to use segment sizes (counted in bytes) as
12445 %%s in hls_segment_filename expression besides date/time values
12446 when strftime is on. To get fixed width numbers with trailing
12447 zeroes, %%0xs format is available where x is the required
12448 width.
12449 second_level_segment_duration
12451 Makes it possible to use segment duration (calculated in
12452 microseconds) as %%t in hls_segment_filename expression besides
12453 date/time values when strftime is on. To get fixed width
12454 numbers with trailing zeroes, %%0xt format is available where x
12455 is the required width.
12456 ffmpeg -i sample.mpeg \
12458 -f hls -hls_time 3 -hls_list_size 5 \
12459 -hls_flags second_level_segment_index+second_level_segment_size+second_level_segment_duration \
12460 -strftime 1 -strftime_mkdir 1 -hls_segment_filename "segment_%Y%m%d%H%M%S_%%04d_%%08s_%%013t.ts" stream.m3u8
12461 This will produce segments like this:
12463 segment_20170102194334_0003_00122200_0000003000000.ts,
12464 segment_20170102194334_0004_00120072_0000003000000.ts etc.
12465 temp_file
12467 Write segment data to filename.tmp and rename to filename only
12468 once the segment is complete. A webserver serving up segments
12469 can be configured to reject requests to *.tmp to prevent access
12470 to in-progress segments before they have been added to the m3u8
12471 playlist. This flag also affects how m3u8 playlist files are
12472 created. If this flag is set, all playlist files will written
12473 into temporary file and renamed after they are complete,
12474 similarly as segments are handled. But playlists with "file"
12475 protocol and with type ("hls_playlist_type") other than "vod"
12476 are always written into temporary file regardless of this flag.
12477 Master playlist files ("master_pl_name"), if any, with "file"
12478 protocol, are always written into temporary file regardless of
12479 this flag if "master_pl_publish_rate" value is other than zero.
12480 hls_playlist_type event
12482 Emit "#EXT-X-PLAYLIST-TYPE:EVENT" in the m3u8 header. Forces
12483 hls_list_size to 0; the playlist can only be appended to.
12484 hls_playlist_type vod
12486 Emit "#EXT-X-PLAYLIST-TYPE:VOD" in the m3u8 header. Forces
12487 hls_list_size to 0; the playlist must not change.
12488 method
12490 Use the given HTTP method to create the hls files.
12491 ffmpeg -re -i in.ts -f hls -method PUT http://example.com/live/out.m3u8
12493 This example will upload all the mpegts segment files to the HTTP
12495 server using the HTTP PUT method, and update the m3u8 files every
12496 "refresh" times using the same method. Note that the HTTP server
12497 must support the given method for uploading files.
12498 http_user_agent
12500 Override User-Agent field in HTTP header. Applicable only for HTTP
12501 output.
12502 var_stream_map
12504 Map string which specifies how to group the audio, video and
12505 subtitle streams into different variant streams. The variant stream
12506 groups are separated by space. Expected string format is like this
12507 "a:0,v:0 a:1,v:1 ....". Here a:, v:, s: are the keys to specify
12508 audio, video and subtitle streams respectively. Allowed values are
12509 0 to 9 (limited just based on practical usage).
12510 When there are two or more variant streams, the output filename
12512 pattern must contain the string "%v", this string specifies the
12513 position of variant stream index in the output media playlist
12514 filenames. The string "%v" may be present in the filename or in the
12515 last directory name containing the file. If the string is present
12516 in the directory name, then sub-directories are created after
12517 expanding the directory name pattern. This enables creation of
12518 variant streams in subdirectories.
12519 ffmpeg -re -i in.ts -b:v:0 1000k -b:v:1 256k -b:a:0 64k -b:a:1 32k \
12521 -map 0:v -map 0:a -map 0:v -map 0:a -f hls -var_stream_map "v:0,a:0 v:1,a:1" \
12522 http://example.com/live/out_%v.m3u8
12523 This example creates two hls variant streams. The first variant
12525 stream will contain video stream of bitrate 1000k and audio stream
12526 of bitrate 64k and the second variant stream will contain video
12527 stream of bitrate 256k and audio stream of bitrate 32k. Here, two
12528 media playlist with file names out_0.m3u8 and out_1.m3u8 will be
12529 created. If you want something meaningful text instead of indexes
12530 in result names, you may specify names for each or some of the
12531 variants as in the following example.
12532 ffmpeg -re -i in.ts -b:v:0 1000k -b:v:1 256k -b:a:0 64k -b:a:1 32k \
12534 -map 0:v -map 0:a -map 0:v -map 0:a -f hls -var_stream_map "v:0,a:0,name:my_hd v:1,a:1,name:my_sd" \
12535 http://example.com/live/out_%v.m3u8
12536 This example creates two hls variant streams as in the previous
12538 one. But here, the two media playlist with file names
12539 out_my_hd.m3u8 and out_my_sd.m3u8 will be created.
12540 ffmpeg -re -i in.ts -b:v:0 1000k -b:v:1 256k -b:a:0 64k \
12542 -map 0:v -map 0:a -map 0:v -f hls -var_stream_map "v:0 a:0 v:1" \
12543 http://example.com/live/out_%v.m3u8
12544 This example creates three hls variant streams. The first variant
12546 stream will be a video only stream with video bitrate 1000k, the
12547 second variant stream will be an audio only stream with bitrate 64k
12548 and the third variant stream will be a video only stream with
12549 bitrate 256k. Here, three media playlist with file names
12550 out_0.m3u8, out_1.m3u8 and out_2.m3u8 will be created.
12551 ffmpeg -re -i in.ts -b:v:0 1000k -b:v:1 256k -b:a:0 64k -b:a:1 32k \
12553 -map 0:v -map 0:a -map 0:v -map 0:a -f hls -var_stream_map "v:0,a:0 v:1,a:1" \
12554 http://example.com/live/vs_%v/out.m3u8
12555 This example creates the variant streams in subdirectories. Here,
12557 the first media playlist is created at
12558 http://example.com/live/vs_0/out.m3u8 and the second one at
12559 http://example.com/live/vs_1/out.m3u8.
12560 ffmpeg -re -i in.ts -b:a:0 32k -b:a:1 64k -b:v:0 1000k -b:v:1 3000k \
12562 -map 0:a -map 0:a -map 0:v -map 0:v -f hls \
12563 -var_stream_map "a:0,agroup:aud_low a:1,agroup:aud_high v:0,agroup:aud_low v:1,agroup:aud_high" \
12564 -master_pl_name master.m3u8 \
12565 http://example.com/live/out_%v.m3u8
12566 This example creates two audio only and two video only variant
12568 streams. In addition to the #EXT-X-STREAM-INF tag for each variant
12569 stream in the master playlist, #EXT-X-MEDIA tag is also added for
12570 the two audio only variant streams and they are mapped to the two
12571 video only variant streams with audio group names 'aud_low' and
12572 'aud_high'.
12573 By default, a single hls variant containing all the encoded streams
12575 is created.
12576 ffmpeg -re -i in.ts -b:a:0 32k -b:a:1 64k -b:v:0 1000k \
12578 -map 0:a -map 0:a -map 0:v -f hls \
12579 -var_stream_map "a:0,agroup:aud_low,default:yes a:1,agroup:aud_low v:0,agroup:aud_low" \
12580 -master_pl_name master.m3u8 \
12581 http://example.com/live/out_%v.m3u8
12582 This example creates two audio only and one video only variant
12584 streams. In addition to the #EXT-X-STREAM-INF tag for each variant
12585 stream in the master playlist, #EXT-X-MEDIA tag is also added for
12586 the two audio only variant streams and they are mapped to the one
12587 video only variant streams with audio group name 'aud_low', and the
12588 audio group have default stat is NO or YES.
12589 By default, a single hls variant containing all the encoded streams
12591 is created.
12592 ffmpeg -re -i in.ts -b:a:0 32k -b:a:1 64k -b:v:0 1000k \
12594 -map 0:a -map 0:a -map 0:v -f hls \
12595 -var_stream_map "a:0,agroup:aud_low,default:yes,language:ENG a:1,agroup:aud_low,language:CHN v:0,agroup:aud_low" \
12596 -master_pl_name master.m3u8 \
12597 http://example.com/live/out_%v.m3u8
12598 This example creates two audio only and one video only variant
12600 streams. In addition to the #EXT-X-STREAM-INF tag for each variant
12601 stream in the master playlist, #EXT-X-MEDIA tag is also added for
12602 the two audio only variant streams and they are mapped to the one
12603 video only variant streams with audio group name 'aud_low', and the
12604 audio group have default stat is NO or YES, and one audio have and
12605 language is named ENG, the other audio language is named CHN.
12606 By default, a single hls variant containing all the encoded streams
12608 is created.
12609 ffmpeg -y -i input_with_subtitle.mkv \
12611 -b:v:0 5250k -c:v h264 -pix_fmt yuv420p -profile:v main -level 4.1 \
12612 -b:a:0 256k \
12613 -c:s webvtt -c:a mp2 -ar 48000 -ac 2 -map 0:v -map 0:a:0 -map 0:s:0 \
12614 -f hls -var_stream_map "v:0,a:0,s:0,sgroup:subtitle" \
12615 -master_pl_name master.m3u8 -t 300 -hls_time 10 -hls_init_time 4 -hls_list_size \
12616 10 -master_pl_publish_rate 10 -hls_flags \
12617 delete_segments+discont_start+split_by_time ./tmp/video.m3u8
12618 This example adds "#EXT-X-MEDIA" tag with "TYPE=SUBTITLES" in the
12620 master playlist with webvtt subtitle group name 'subtitle'. Please
12621 make sure the input file has one text subtitle stream at least.
12622 cc_stream_map
12624 Map string which specifies different closed captions groups and
12625 their attributes. The closed captions stream groups are separated
12626 by space. Expected string format is like this "ccgroup:<group
12627 name>,instreamid:<INSTREAM-ID>,language:<language code> ....".
12628 'ccgroup' and 'instreamid' are mandatory attributes. 'language' is
12629 an optional attribute. The closed captions groups configured using
12630 this option are mapped to different variant streams by providing
12631 the same 'ccgroup' name in the "var_stream_map" string. If
12632 "var_stream_map" is not set, then the first available ccgroup in
12633 "cc_stream_map" is mapped to the output variant stream. The
12634 examples for these two use cases are given below.
12635 ffmpeg -re -i in.ts -b:v 1000k -b:a 64k -a53cc 1 -f hls \
12637 -cc_stream_map "ccgroup:cc,instreamid:CC1,language:en" \
12638 -master_pl_name master.m3u8 \
12639 http://example.com/live/out.m3u8
12640 This example adds "#EXT-X-MEDIA" tag with "TYPE=CLOSED-CAPTIONS" in
12642 the master playlist with group name 'cc', language 'en' (english)
12643 and INSTREAM-ID 'CC1'. Also, it adds "CLOSED-CAPTIONS" attribute
12644 with group name 'cc' for the output variant stream.
12645 ffmpeg -re -i in.ts -b:v:0 1000k -b:v:1 256k -b:a:0 64k -b:a:1 32k \
12647 -a53cc:0 1 -a53cc:1 1\
12648 -map 0:v -map 0:a -map 0:v -map 0:a -f hls \
12649 -cc_stream_map "ccgroup:cc,instreamid:CC1,language:en ccgroup:cc,instreamid:CC2,language:sp" \
12650 -var_stream_map "v:0,a:0,ccgroup:cc v:1,a:1,ccgroup:cc" \
12651 -master_pl_name master.m3u8 \
12652 http://example.com/live/out_%v.m3u8
12653 This example adds two "#EXT-X-MEDIA" tags with
12655 "TYPE=CLOSED-CAPTIONS" in the master playlist for the INSTREAM-IDs
12656 'CC1' and 'CC2'. Also, it adds "CLOSED-CAPTIONS" attribute with
12657 group name 'cc' for the two output variant streams.
12658 master_pl_name
12660 Create HLS master playlist with the given name.
12661 ffmpeg -re -i in.ts -f hls -master_pl_name master.m3u8 http://example.com/live/out.m3u8
12663 This example creates HLS master playlist with name master.m3u8 and
12665 it is published at http://example.com/live/
12666 master_pl_publish_rate
12668 Publish master play list repeatedly every after specified number of
12669 segment intervals.
12670 ffmpeg -re -i in.ts -f hls -master_pl_name master.m3u8 \
12672 -hls_time 2 -master_pl_publish_rate 30 http://example.com/live/out.m3u8
12673 This example creates HLS master playlist with name master.m3u8 and
12675 keep publishing it repeatedly every after 30 segments i.e. every
12676 after 60s.
12677 http_persistent
12679 Use persistent HTTP connections. Applicable only for HTTP output.
12680 timeout
12682 Set timeout for socket I/O operations. Applicable only for HTTP
12683 output.
12684 -ignore_io_errors
12686 Ignore IO errors during open, write and delete. Useful for long-
12687 duration runs with network output.
12688 headers
12690 Set custom HTTP headers, can override built in default headers.
12691 Applicable only for HTTP output.
12692 ico
12694 ICO file muxer.
12695 Microsoft's icon file format (ICO) has some strict limitations that
12697 should be noted:
12698 • Size cannot exceed 256 pixels in any dimension
12700 • Only BMP and PNG images can be stored
12702 • If a BMP image is used, it must be one of the following pixel
12704 formats:
12705 BMP Bit Depth FFmpeg Pixel Format
12707 1bit pal8
12708 4bit pal8
12709 8bit pal8
12710 16bit rgb555le
12711 24bit bgr24
12712 32bit bgra
12713 • If a BMP image is used, it must use the BITMAPINFOHEADER DIB header
12715 • If a PNG image is used, it must use the rgba pixel format
12717 image2
12719 Image file muxer.
12720 The image file muxer writes video frames to image files.
12722 The output filenames are specified by a pattern, which can be used to
12724 produce sequentially numbered series of files. The pattern may contain
12725 the string "%d" or "%0Nd", this string specifies the position of the
12726 characters representing a numbering in the filenames. If the form
12727 "%0Nd" is used, the string representing the number in each filename is
12728 0-padded to N digits. The literal character '%' can be specified in the
12729 pattern with the string "%%".
12730 If the pattern contains "%d" or "%0Nd", the first filename of the file
12732 list specified will contain the number 1, all the following numbers
12733 will be sequential.
12734 The pattern may contain a suffix which is used to automatically
12736 determine the format of the image files to write.
12737 For example the pattern "img-%03d.bmp" will specify a sequence of
12739 filenames of the form img-001.bmp, img-002.bmp, ..., img-010.bmp, etc.
12740 The pattern "img%%-%d.jpg" will specify a sequence of filenames of the
12741 form img%-1.jpg, img%-2.jpg, ..., img%-10.jpg, etc.
12742 The image muxer supports the .Y.U.V image file format. This format is
12744 special in that that each image frame consists of three files, for each
12745 of the YUV420P components. To read or write this image file format,
12746 specify the name of the '.Y' file. The muxer will automatically open
12747 the '.U' and '.V' files as required.
12748 Options
12750 frame_pts
12752 If set to 1, expand the filename with pts from pkt->pts. Default
12753 value is 0.
12754 start_number
12756 Start the sequence from the specified number. Default value is 1.
12757 update
12759 If set to 1, the filename will always be interpreted as just a
12760 filename, not a pattern, and the corresponding file will be
12761 continuously overwritten with new images. Default value is 0.
12762 strftime
12764 If set to 1, expand the filename with date and time information
12765 from strftime(). Default value is 0.
12766 atomic_writing
12768 Write output to a temporary file, which is renamed to target
12769 filename once writing is completed. Default is disabled.
12770 protocol_opts options_list
12772 Set protocol options as a :-separated list of key=value parameters.
12773 Values containing the ":" special character must be escaped.
12774 Examples
12776 The following example shows how to use ffmpeg for creating a sequence
12778 of files img-001.jpeg, img-002.jpeg, ..., taking one image every second
12779 from the input video:
12780 ffmpeg -i in.avi -vsync cfr -r 1 -f image2 'img-%03d.jpeg'
12782 Note that with ffmpeg, if the format is not specified with the "-f"
12784 option and the output filename specifies an image file format, the
12785 image2 muxer is automatically selected, so the previous command can be
12786 written as:
12787 ffmpeg -i in.avi -vsync cfr -r 1 'img-%03d.jpeg'
12789 Note also that the pattern must not necessarily contain "%d" or "%0Nd",
12791 for example to create a single image file img.jpeg from the start of
12792 the input video you can employ the command:
12793 ffmpeg -i in.avi -f image2 -frames:v 1 img.jpeg
12795 The strftime option allows you to expand the filename with date and
12797 time information. Check the documentation of the strftime() function
12798 for the syntax.
12799 For example to generate image files from the strftime()
12801 "%Y-%m-%d_%H-%M-%S" pattern, the following ffmpeg command can be used:
12802 ffmpeg -f v4l2 -r 1 -i /dev/video0 -f image2 -strftime 1 "%Y-%m-%d_%H-%M-%S.jpg"
12804 You can set the file name with current frame's PTS:
12806 ffmpeg -f v4l2 -r 1 -i /dev/video0 -copyts -f image2 -frame_pts true %d.jpg"
12808 A more complex example is to publish contents of your desktop directly
12810 to a WebDAV server every second:
12811 ffmpeg -f x11grab -framerate 1 -i :0.0 -q:v 6 -update 1 -protocol_opts method=PUT http://example.com/desktop.jpg
12813 matroska
12815 Matroska container muxer.
12816 This muxer implements the matroska and webm container specs.
12818 Metadata
12820 The recognized metadata settings in this muxer are:
12822 title
12824 Set title name provided to a single track. This gets mapped to the
12825 FileDescription element for a stream written as attachment.
12826 language
12828 Specify the language of the track in the Matroska languages form.
12829 The language can be either the 3 letters bibliographic ISO-639-2
12831 (ISO 639-2/B) form (like "fre" for French), or a language code
12832 mixed with a country code for specialities in languages (like "fre-
12833 ca" for Canadian French).
12834 stereo_mode
12836 Set stereo 3D video layout of two views in a single video track.
12837 The following values are recognized:
12839 mono
12841 video is not stereo
12842 left_right
12844 Both views are arranged side by side, Left-eye view is on the
12845 left
12846 bottom_top
12848 Both views are arranged in top-bottom orientation, Left-eye
12849 view is at bottom
12850 top_bottom
12852 Both views are arranged in top-bottom orientation, Left-eye
12853 view is on top
12854 checkerboard_rl
12856 Each view is arranged in a checkerboard interleaved pattern,
12857 Left-eye view being first
12858 checkerboard_lr
12860 Each view is arranged in a checkerboard interleaved pattern,
12861 Right-eye view being first
12862 row_interleaved_rl
12864 Each view is constituted by a row based interleaving, Right-eye
12865 view is first row
12866 row_interleaved_lr
12868 Each view is constituted by a row based interleaving, Left-eye
12869 view is first row
12870 col_interleaved_rl
12872 Both views are arranged in a column based interleaving manner,
12873 Right-eye view is first column
12874 col_interleaved_lr
12876 Both views are arranged in a column based interleaving manner,
12877 Left-eye view is first column
12878 anaglyph_cyan_red
12880 All frames are in anaglyph format viewable through red-cyan
12881 filters
12882 right_left
12884 Both views are arranged side by side, Right-eye view is on the
12885 left
12886 anaglyph_green_magenta
12888 All frames are in anaglyph format viewable through green-
12889 magenta filters
12890 block_lr
12892 Both eyes laced in one Block, Left-eye view is first
12893 block_rl
12895 Both eyes laced in one Block, Right-eye view is first
12896 For example a 3D WebM clip can be created using the following command
12898 line:
12899 ffmpeg -i sample_left_right_clip.mpg -an -c:v libvpx -metadata stereo_mode=left_right -y stereo_clip.webm
12901 Options
12903 This muxer supports the following options:
12905 reserve_index_space
12907 By default, this muxer writes the index for seeking (called cues in
12908 Matroska terms) at the end of the file, because it cannot know in
12909 advance how much space to leave for the index at the beginning of
12910 the file. However for some use cases -- e.g. streaming where
12911 seeking is possible but slow -- it is useful to put the index at
12912 the beginning of the file.
12913 If this option is set to a non-zero value, the muxer will reserve a
12915 given amount of space in the file header and then try to write the
12916 cues there when the muxing finishes. If the reserved space does not
12917 suffice, no Cues will be written, the file will be finalized and
12918 writing the trailer will return an error. A safe size for most use
12919 cases should be about 50kB per hour of video.
12920 Note that cues are only written if the output is seekable and this
12922 option will have no effect if it is not.
12923 cues_to_front
12925 If set, the muxer will write the index at the beginning of the file
12926 by shifting the main data if necessary. This can be combined with
12927 reserve_index_space in which case the data is only shifted if the
12928 initially reserved space turns out to be insufficient.
12929 This option is ignored if the output is unseekable.
12931 default_mode
12933 This option controls how the FlagDefault of the output tracks will
12934 be set. It influences which tracks players should play by default.
12935 The default mode is passthrough.
12936 infer
12938 Every track with disposition default will have the FlagDefault
12939 set. Additionally, for each type of track (audio, video or
12940 subtitle), if no track with disposition default of this type
12941 exists, then the first track of this type will be marked as
12942 default (if existing). This ensures that the default flag is
12943 set in a sensible way even if the input originated from
12944 containers that lack the concept of default tracks.
12945 infer_no_subs
12947 This mode is the same as infer except that if no subtitle track
12948 with disposition default exists, no subtitle track will be
12949 marked as default.
12950 passthrough
12952 In this mode the FlagDefault is set if and only if the
12953 AV_DISPOSITION_DEFAULT flag is set in the disposition of the
12954 corresponding stream.
12955 flipped_raw_rgb
12957 If set to true, store positive height for raw RGB bitmaps, which
12958 indicates bitmap is stored bottom-up. Note that this option does
12959 not flip the bitmap which has to be done manually beforehand, e.g.
12960 by using the vflip filter. Default is false and indicates bitmap
12961 is stored top down.
12962 md5
12964 MD5 testing format.
12965 This is a variant of the hash muxer. Unlike that muxer, it defaults to
12967 using the MD5 hash function.
12968 Examples
12970 To compute the MD5 hash of the input converted to raw audio and video,
12972 and store it in the file out.md5:
12973 ffmpeg -i INPUT -f md5 out.md5
12975 You can print the MD5 to stdout with the command:
12977 ffmpeg -i INPUT -f md5 -
12979 See also the hash and framemd5 muxers.
12981 mov, mp4, ismv
12983 MOV/MP4/ISMV (Smooth Streaming) muxer.
12984 The mov/mp4/ismv muxer supports fragmentation. Normally, a MOV/MP4 file
12986 has all the metadata about all packets stored in one location (written
12987 at the end of the file, it can be moved to the start for better
12988 playback by adding faststart to the movflags, or using the qt-faststart
12989 tool). A fragmented file consists of a number of fragments, where
12990 packets and metadata about these packets are stored together. Writing a
12991 fragmented file has the advantage that the file is decodable even if
12992 the writing is interrupted (while a normal MOV/MP4 is undecodable if it
12993 is not properly finished), and it requires less memory when writing
12994 very long files (since writing normal MOV/MP4 files stores info about
12995 every single packet in memory until the file is closed). The downside
12996 is that it is less compatible with other applications.
12997 Options
12999 Fragmentation is enabled by setting one of the AVOptions that define
13001 how to cut the file into fragments:
13002 -moov_size bytes
13004 Reserves space for the moov atom at the beginning of the file
13005 instead of placing the moov atom at the end. If the space reserved
13006 is insufficient, muxing will fail.
13007 -movflags frag_keyframe
13009 Start a new fragment at each video keyframe.
13010 -frag_duration duration
13012 Create fragments that are duration microseconds long.
13013 -frag_size size
13015 Create fragments that contain up to size bytes of payload data.
13016 -movflags frag_custom
13018 Allow the caller to manually choose when to cut fragments, by
13019 calling "av_write_frame(ctx, NULL)" to write a fragment with the
13020 packets written so far. (This is only useful with other
13021 applications integrating libavformat, not from ffmpeg.)
13022 -min_frag_duration duration
13024 Don't create fragments that are shorter than duration microseconds
13025 long.
13026 If more than one condition is specified, fragments are cut when one of
13028 the specified conditions is fulfilled. The exception to this is
13029 "-min_frag_duration", which has to be fulfilled for any of the other
13030 conditions to apply.
13031 Additionally, the way the output file is written can be adjusted
13033 through a few other options:
13034 -movflags empty_moov
13036 Write an initial moov atom directly at the start of the file,
13037 without describing any samples in it. Generally, an mdat/moov pair
13038 is written at the start of the file, as a normal MOV/MP4 file,
13039 containing only a short portion of the file. With this option set,
13040 there is no initial mdat atom, and the moov atom only describes the
13041 tracks but has a zero duration.
13042 This option is implicitly set when writing ismv (Smooth Streaming)
13044 files.
13045 -movflags separate_moof
13047 Write a separate moof (movie fragment) atom for each track.
13048 Normally, packets for all tracks are written in a moof atom (which
13049 is slightly more efficient), but with this option set, the muxer
13050 writes one moof/mdat pair for each track, making it easier to
13051 separate tracks.
13052 This option is implicitly set when writing ismv (Smooth Streaming)
13054 files.
13055 -movflags skip_sidx
13057 Skip writing of sidx atom. When bitrate overhead due to sidx atom
13058 is high, this option could be used for cases where sidx atom is not
13059 mandatory. When global_sidx flag is enabled, this option will be
13060 ignored.
13061 -movflags faststart
13063 Run a second pass moving the index (moov atom) to the beginning of
13064 the file. This operation can take a while, and will not work in
13065 various situations such as fragmented output, thus it is not
13066 enabled by default.
13067 -movflags rtphint
13069 Add RTP hinting tracks to the output file.
13070 -movflags disable_chpl
13072 Disable Nero chapter markers (chpl atom). Normally, both Nero
13073 chapters and a QuickTime chapter track are written to the file.
13074 With this option set, only the QuickTime chapter track will be
13075 written. Nero chapters can cause failures when the file is
13076 reprocessed with certain tagging programs, like mp3Tag 2.61a and
13077 iTunes 11.3, most likely other versions are affected as well.
13078 -movflags omit_tfhd_offset
13080 Do not write any absolute base_data_offset in tfhd atoms. This
13081 avoids tying fragments to absolute byte positions in the
13082 file/streams.
13083 -movflags default_base_moof
13085 Similarly to the omit_tfhd_offset, this flag avoids writing the
13086 absolute base_data_offset field in tfhd atoms, but does so by using
13087 the new default-base-is-moof flag instead. This flag is new from
13088 14496-12:2012. This may make the fragments easier to parse in
13089 certain circumstances (avoiding basing track fragment location
13090 calculations on the implicit end of the previous track fragment).
13091 -write_tmcd
13093 Specify "on" to force writing a timecode track, "off" to disable it
13094 and "auto" to write a timecode track only for mov and mp4 output
13095 (default).
13096 -movflags negative_cts_offsets
13098 Enables utilization of version 1 of the CTTS box, in which the CTS
13099 offsets can be negative. This enables the initial sample to have
13100 DTS/CTS of zero, and reduces the need for edit lists for some cases
13101 such as video tracks with B-frames. Additionally, eases conformance
13102 with the DASH-IF interoperability guidelines.
13103 This option is implicitly set when writing ismv (Smooth Streaming)
13105 files.
13106 -write_btrt bool
13108 Force or disable writing bitrate box inside stsd box of a track.
13109 The box contains decoding buffer size (in bytes), maximum bitrate
13110 and average bitrate for the track. The box will be skipped if none
13111 of these values can be computed. Default is -1 or "auto", which
13112 will write the box only in MP4 mode.
13113 -write_prft
13115 Write producer time reference box (PRFT) with a specified time
13116 source for the NTP field in the PRFT box. Set value as wallclock to
13117 specify timesource as wallclock time and pts to specify timesource
13118 as input packets' PTS values.
13119 Setting value to pts is applicable only for a live encoding use
13121 case, where PTS values are set as as wallclock time at the source.
13122 For example, an encoding use case with decklink capture source
13123 where video_pts and audio_pts are set to abs_wallclock.
13124 -empty_hdlr_name bool
13126 Enable to skip writing the name inside a "hdlr" box. Default is
13127 "false".
13128 -movie_timescale scale
13130 Set the timescale written in the movie header box ("mvhd"). Range
13131 is 1 to INT_MAX. Default is 1000.
13132 -video_track_timescale scale
13134 Set the timescale used for video tracks. Range is 0 to INT_MAX. If
13135 set to 0, the timescale is automatically set based on the native
13136 stream time base. Default is 0.
13137 Example
13139 Smooth Streaming content can be pushed in real time to a publishing
13141 point on IIS with this muxer. Example:
13142 ffmpeg -re <<normal input/transcoding options>> -movflags isml+frag_keyframe -f ismv http://server/publishingpoint.isml/Streams(Encoder1)
13144 mp3
13146 The MP3 muxer writes a raw MP3 stream with the following optional
13147 features:
13148 • An ID3v2 metadata header at the beginning (enabled by default).
13150 Versions 2.3 and 2.4 are supported, the "id3v2_version" private
13151 option controls which one is used (3 or 4). Setting "id3v2_version"
13152 to 0 disables the ID3v2 header completely.
13153 The muxer supports writing attached pictures (APIC frames) to the
13155 ID3v2 header. The pictures are supplied to the muxer in form of a
13156 video stream with a single packet. There can be any number of those
13157 streams, each will correspond to a single APIC frame. The stream
13158 metadata tags title and comment map to APIC description and picture
13159 type respectively. See <http://id3.org/id3v2.4.0-frames> for
13160 allowed picture types.
13161 Note that the APIC frames must be written at the beginning, so the
13163 muxer will buffer the audio frames until it gets all the pictures.
13164 It is therefore advised to provide the pictures as soon as possible
13165 to avoid excessive buffering.
13166 • A Xing/LAME frame right after the ID3v2 header (if present). It is
13168 enabled by default, but will be written only if the output is
13169 seekable. The "write_xing" private option can be used to disable
13170 it. The frame contains various information that may be useful to
13171 the decoder, like the audio duration or encoder delay.
13172 • A legacy ID3v1 tag at the end of the file (disabled by default). It
13174 may be enabled with the "write_id3v1" private option, but as its
13175 capabilities are very limited, its usage is not recommended.
13176 Examples:
13178 Write an mp3 with an ID3v2.3 header and an ID3v1 footer:
13180 ffmpeg -i INPUT -id3v2_version 3 -write_id3v1 1 out.mp3
13182 To attach a picture to an mp3 file select both the audio and the
13184 picture stream with "map":
13185 ffmpeg -i input.mp3 -i cover.png -c copy -map 0 -map 1
13187 -metadata:s:v title="Album cover" -metadata:s:v comment="Cover (Front)" out.mp3
13188 Write a "clean" MP3 without any extra features:
13190 ffmpeg -i input.wav -write_xing 0 -id3v2_version 0 out.mp3
13192 mpegts
13194 MPEG transport stream muxer.
13195 This muxer implements ISO 13818-1 and part of ETSI EN 300 468.
13197 The recognized metadata settings in mpegts muxer are "service_provider"
13199 and "service_name". If they are not set the default for
13200 "service_provider" is FFmpeg and the default for "service_name" is
13201 Service01.
13202 Options
13204 The muxer options are:
13206 mpegts_transport_stream_id integer
13208 Set the transport_stream_id. This identifies a transponder in DVB.
13209 Default is 0x0001.
13210 mpegts_original_network_id integer
13212 Set the original_network_id. This is unique identifier of a network
13213 in DVB. Its main use is in the unique identification of a service
13214 through the path Original_Network_ID, Transport_Stream_ID. Default
13215 is 0x0001.
13216 mpegts_service_id integer
13218 Set the service_id, also known as program in DVB. Default is
13219 0x0001.
13220 mpegts_service_type integer
13222 Set the program service_type. Default is "digital_tv". Accepts the
13223 following options:
13224 hex_value
13226 Any hexadecimal value between 0x01 and 0xff as defined in ETSI
13227 300 468.
13228 digital_tv
13230 Digital TV service.
13231 digital_radio
13233 Digital Radio service.
13234 teletext
13236 Teletext service.
13237 advanced_codec_digital_radio
13239 Advanced Codec Digital Radio service.
13240 mpeg2_digital_hdtv
13242 MPEG2 Digital HDTV service.
13243 advanced_codec_digital_sdtv
13245 Advanced Codec Digital SDTV service.
13246 advanced_codec_digital_hdtv
13248 Advanced Codec Digital HDTV service.
13249 mpegts_pmt_start_pid integer
13251 Set the first PID for PMTs. Default is 0x1000, minimum is 0x0020,
13252 maximum is 0x1ffa. This option has no effect in m2ts mode where the
13253 PMT PID is fixed 0x0100.
13254 mpegts_start_pid integer
13256 Set the first PID for elementary streams. Default is 0x0100,
13257 minimum is 0x0020, maximum is 0x1ffa. This option has no effect in
13258 m2ts mode where the elementary stream PIDs are fixed.
13259 mpegts_m2ts_mode boolean
13261 Enable m2ts mode if set to 1. Default value is -1 which disables
13262 m2ts mode.
13263 muxrate integer
13265 Set a constant muxrate. Default is VBR.
13266 pes_payload_size integer
13268 Set minimum PES packet payload in bytes. Default is 2930.
13269 mpegts_flags flags
13271 Set mpegts flags. Accepts the following options:
13272 resend_headers
13274 Reemit PAT/PMT before writing the next packet.
13275 latm
13277 Use LATM packetization for AAC.
13278 pat_pmt_at_frames
13280 Reemit PAT and PMT at each video frame.
13281 system_b
13283 Conform to System B (DVB) instead of System A (ATSC).
13284 initial_discontinuity
13286 Mark the initial packet of each stream as discontinuity.
13287 nit Emit NIT table.
13289 omit_rai
13291 Disable writing of random access indicator.
13292 mpegts_copyts boolean
13294 Preserve original timestamps, if value is set to 1. Default value
13295 is -1, which results in shifting timestamps so that they start from
13296 0.
13297 omit_video_pes_length boolean
13299 Omit the PES packet length for video packets. Default is 1 (true).
13300 pcr_period integer
13302 Override the default PCR retransmission time in milliseconds.
13303 Default is -1 which means that the PCR interval will be determined
13304 automatically: 20 ms is used for CBR streams, the highest multiple
13305 of the frame duration which is less than 100 ms is used for VBR
13306 streams.
13307 pat_period duration
13309 Maximum time in seconds between PAT/PMT tables. Default is 0.1.
13310 sdt_period duration
13312 Maximum time in seconds between SDT tables. Default is 0.5.
13313 nit_period duration
13315 Maximum time in seconds between NIT tables. Default is 0.5.
13316 tables_version integer
13318 Set PAT, PMT, SDT and NIT version (default 0, valid values are from
13319 0 to 31, inclusively). This option allows updating stream
13320 structure so that standard consumer may detect the change. To do
13321 so, reopen output "AVFormatContext" (in case of API usage) or
13322 restart ffmpeg instance, cyclically changing tables_version value:
13323 ffmpeg -i source1.ts -codec copy -f mpegts -tables_version 0 udp://1.1.1.1:1111
13325 ffmpeg -i source2.ts -codec copy -f mpegts -tables_version 1 udp://1.1.1.1:1111
13326 ...
13327 ffmpeg -i source3.ts -codec copy -f mpegts -tables_version 31 udp://1.1.1.1:1111
13328 ffmpeg -i source1.ts -codec copy -f mpegts -tables_version 0 udp://1.1.1.1:1111
13329 ffmpeg -i source2.ts -codec copy -f mpegts -tables_version 1 udp://1.1.1.1:1111
13330 ...
13331 Example
13333 ffmpeg -i file.mpg -c copy \
13335 -mpegts_original_network_id 0x1122 \
13336 -mpegts_transport_stream_id 0x3344 \
13337 -mpegts_service_id 0x5566 \
13338 -mpegts_pmt_start_pid 0x1500 \
13339 -mpegts_start_pid 0x150 \
13340 -metadata service_provider="Some provider" \
13341 -metadata service_name="Some Channel" \
13342 out.ts
13343 mxf, mxf_d10, mxf_opatom
13345 MXF muxer.
13346 Options
13348 The muxer options are:
13350 store_user_comments bool
13352 Set if user comments should be stored if available or never. IRT
13353 D-10 does not allow user comments. The default is thus to write
13354 them for mxf and mxf_opatom but not for mxf_d10
13355 null
13357 Null muxer.
13358 This muxer does not generate any output file, it is mainly useful for
13360 testing or benchmarking purposes.
13361 For example to benchmark decoding with ffmpeg you can use the command:
13363 ffmpeg -benchmark -i INPUT -f null out.null
13365 Note that the above command does not read or write the out.null file,
13367 but specifying the output file is required by the ffmpeg syntax.
13368 Alternatively you can write the command as:
13370 ffmpeg -benchmark -i INPUT -f null -
13372 nut
13374 -syncpoints flags
13375 Change the syncpoint usage in nut:
13376 default use the normal low-overhead seeking aids.
13378 none do not use the syncpoints at all, reducing the overhead but
13379 making the stream non-seekable;
13380 Use of this option is not recommended, as the resulting files are very damage
13381 sensitive and seeking is not possible. Also in general the overhead from
13382 syncpoints is negligible. Note, -C<write_index> 0 can be used to disable
13383 all growing data tables, allowing to mux endless streams with limited memory
13384 and without these disadvantages.
13385 timestamped extend the syncpoint with a wallclock field.
13387 The none and timestamped flags are experimental.
13389 -write_index bool
13391 Write index at the end, the default is to write an index.
13392 ffmpeg -i INPUT -f_strict experimental -syncpoints none - | processor
13394 ogg
13396 Ogg container muxer.
13397 -page_duration duration
13399 Preferred page duration, in microseconds. The muxer will attempt to
13400 create pages that are approximately duration microseconds long.
13401 This allows the user to compromise between seek granularity and
13402 container overhead. The default is 1 second. A value of 0 will fill
13403 all segments, making pages as large as possible. A value of 1 will
13404 effectively use 1 packet-per-page in most situations, giving a
13405 small seek granularity at the cost of additional container
13406 overhead.
13407 -serial_offset value
13409 Serial value from which to set the streams serial number. Setting
13410 it to different and sufficiently large values ensures that the
13411 produced ogg files can be safely chained.
13412 raw muxers
13414 Raw muxers accept a single stream matching the designated codec. They
13415 do not store timestamps or metadata. The recognized extension is the
13416 same as the muxer name unless indicated otherwise.
13417 ac3
13419 Dolby Digital, also known as AC-3, audio.
13421 adx
13423 CRI Middleware ADX audio.
13425 This muxer will write out the total sample count near the start of the
13427 first packet when the output is seekable and the count can be stored in
13428 32 bits.
13429 aptx
13431 aptX (Audio Processing Technology for Bluetooth) audio.
13433 aptx_hd
13435 aptX HD (Audio Processing Technology for Bluetooth) audio.
13437 Extensions: aptxhd
13439 avs2
13441 AVS2-P2/IEEE1857.4 video.
13443 Extensions: avs, avs2
13445 cavsvideo
13447 Chinese AVS (Audio Video Standard) video.
13449 Extensions: cavs
13451 codec2raw
13453 Codec 2 audio.
13455 No extension is registered so format name has to be supplied e.g. with
13457 the ffmpeg CLI tool "-f codec2raw".
13458 data
13460 Data muxer accepts a single stream with any codec of any type. The
13462 input stream has to be selected using the "-map" option with the ffmpeg
13463 CLI tool.
13464 No extension is registered so format name has to be supplied e.g. with
13466 the ffmpeg CLI tool "-f data".
13467 dirac
13469 BBC Dirac video. The Dirac Pro codec is a subset and is standardized as
13471 SMPTE VC-2.
13472 Extensions: drc, vc2
13474 dnxhd
13476 Avid DNxHD video. It is standardized as SMPTE VC-3. Accepts DNxHR
13478 streams.
13479 Extensions: dnxhd, dnxhr
13481 dts
13483 DTS Coherent Acoustics (DCA) audio.
13485 eac3
13487 Dolby Digital Plus, also known as Enhanced AC-3, audio.
13489 g722
13491 ITU-T G.722 audio.
13493 g723_1
13495 ITU-T G.723.1 audio.
13497 Extensions: tco, rco
13499 g726
13501 ITU-T G.726 big-endian ("left-justified") audio.
13503 No extension is registered so format name has to be supplied e.g. with
13505 the ffmpeg CLI tool "-f g726".
13506 g726le
13508 ITU-T G.726 little-endian ("right-justified") audio.
13510 No extension is registered so format name has to be supplied e.g. with
13512 the ffmpeg CLI tool "-f g726le".
13513 gsm
13515 Global System for Mobile Communications audio.
13517 h261
13519 ITU-T H.261 video.
13521 h263
13523 ITU-T H.263 / H.263-1996, H.263+ / H.263-1998 / H.263 version 2 video.
13525 h264
13527 ITU-T H.264 / MPEG-4 Part 10 AVC video. Bitstream shall be converted to
13529 Annex B syntax if it's in length-prefixed mode.
13530 Extensions: h264, 264
13532 hevc
13534 ITU-T H.265 / MPEG-H Part 2 HEVC video. Bitstream shall be converted to
13536 Annex B syntax if it's in length-prefixed mode.
13537 Extensions: hevc, h265, 265
13539 m4v
13541 MPEG-4 Part 2 video.
13543 mjpeg
13545 Motion JPEG video.
13547 Extensions: mjpg, mjpeg
13549 mlp
13551 Meridian Lossless Packing, also known as Packed PCM, audio.
13553 mp2
13555 MPEG-1 Audio Layer II audio.
13557 Extensions: mp2, m2a, mpa
13559 mpeg1video
13561 MPEG-1 Part 2 video.
13563 Extensions: mpg, mpeg, m1v
13565 mpeg2video
13567 ITU-T H.262 / MPEG-2 Part 2 video.
13569 Extensions: m2v
13571 obu
13573 AV1 low overhead Open Bitstream Units muxer. Temporal delimiter OBUs
13575 will be inserted in all temporal units of the stream.
13576 rawvideo
13578 Raw uncompressed video.
13580 Extensions: yuv, rgb
13582 sbc
13584 Bluetooth SIG low-complexity subband codec audio.
13586 Extensions: sbc, msbc
13588 truehd
13590 Dolby TrueHD audio.
13592 Extensions: thd
13594 vc1
13596 SMPTE 421M / VC-1 video.
13598 segment, stream_segment, ssegment
13600 Basic stream segmenter.
13601 This muxer outputs streams to a number of separate files of nearly
13603 fixed duration. Output filename pattern can be set in a fashion similar
13604 to image2, or by using a "strftime" template if the strftime option is
13605 enabled.
13606 "stream_segment" is a variant of the muxer used to write to streaming
13608 output formats, i.e. which do not require global headers, and is
13609 recommended for outputting e.g. to MPEG transport stream segments.
13610 "ssegment" is a shorter alias for "stream_segment".
13611 Every segment starts with a keyframe of the selected reference stream,
13613 which is set through the reference_stream option.
13614 Note that if you want accurate splitting for a video file, you need to
13616 make the input key frames correspond to the exact splitting times
13617 expected by the segmenter, or the segment muxer will start the new
13618 segment with the key frame found next after the specified start time.
13619 The segment muxer works best with a single constant frame rate video.
13621 Optionally it can generate a list of the created segments, by setting
13623 the option segment_list. The list type is specified by the
13624 segment_list_type option. The entry filenames in the segment list are
13625 set by default to the basename of the corresponding segment files.
13626 See also the hls muxer, which provides a more specific implementation
13628 for HLS segmentation.
13629 Options
13631 The segment muxer supports the following options:
13633 increment_tc 1|0
13635 if set to 1, increment timecode between each segment If this is
13636 selected, the input need to have a timecode in the first video
13637 stream. Default value is 0.
13638 reference_stream specifier
13640 Set the reference stream, as specified by the string specifier. If
13641 specifier is set to "auto", the reference is chosen automatically.
13642 Otherwise it must be a stream specifier (see the ``Stream
13643 specifiers'' chapter in the ffmpeg manual) which specifies the
13644 reference stream. The default value is "auto".
13645 segment_format format
13647 Override the inner container format, by default it is guessed by
13648 the filename extension.
13649 segment_format_options options_list
13651 Set output format options using a :-separated list of key=value
13652 parameters. Values containing the ":" special character must be
13653 escaped.
13654 segment_list name
13656 Generate also a listfile named name. If not specified no listfile
13657 is generated.
13658 segment_list_flags flags
13660 Set flags affecting the segment list generation.
13661 It currently supports the following flags:
13663 cache
13665 Allow caching (only affects M3U8 list files).
13666 live
13668 Allow live-friendly file generation.
13669 segment_list_size size
13671 Update the list file so that it contains at most size segments. If
13672 0 the list file will contain all the segments. Default value is 0.
13673 segment_list_entry_prefix prefix
13675 Prepend prefix to each entry. Useful to generate absolute paths.
13676 By default no prefix is applied.
13677 segment_list_type type
13679 Select the listing format.
13680 The following values are recognized:
13682 flat
13684 Generate a flat list for the created segments, one segment per
13685 line.
13686 csv, ext
13688 Generate a list for the created segments, one segment per line,
13689 each line matching the format (comma-separated values):
13690 <segment_filename>,<segment_start_time>,<segment_end_time>
13692 segment_filename is the name of the output file generated by
13694 the muxer according to the provided pattern. CSV escaping
13695 (according to RFC4180) is applied if required.
13696 segment_start_time and segment_end_time specify the segment
13698 start and end time expressed in seconds.
13699 A list file with the suffix ".csv" or ".ext" will auto-select
13701 this format.
13702 ext is deprecated in favor or csv.
13704 ffconcat
13706 Generate an ffconcat file for the created segments. The
13707 resulting file can be read using the FFmpeg concat demuxer.
13708 A list file with the suffix ".ffcat" or ".ffconcat" will auto-
13710 select this format.
13711 m3u8
13713 Generate an extended M3U8 file, version 3, compliant with
13714 <http://tools.ietf.org/id/draft-pantos-http-live-streaming>.
13715 A list file with the suffix ".m3u8" will auto-select this
13717 format.
13718 If not specified the type is guessed from the list file name
13720 suffix.
13721 segment_time time
13723 Set segment duration to time, the value must be a duration
13724 specification. Default value is "2". See also the segment_times
13725 option.
13726 Note that splitting may not be accurate, unless you force the
13728 reference stream key-frames at the given time. See the introductory
13729 notice and the examples below.
13730 min_seg_duration time
13732 Set minimum segment duration to time, the value must be a duration
13733 specification. This prevents the muxer ending segments at a
13734 duration below this value. Only effective with "segment_time".
13735 Default value is "0".
13736 segment_atclocktime 1|0
13738 If set to "1" split at regular clock time intervals starting from
13739 00:00 o'clock. The time value specified in segment_time is used for
13740 setting the length of the splitting interval.
13741 For example with segment_time set to "900" this makes it possible
13743 to create files at 12:00 o'clock, 12:15, 12:30, etc.
13744 Default value is "0".
13746 segment_clocktime_offset duration
13748 Delay the segment splitting times with the specified duration when
13749 using segment_atclocktime.
13750 For example with segment_time set to "900" and
13752 segment_clocktime_offset set to "300" this makes it possible to
13753 create files at 12:05, 12:20, 12:35, etc.
13754 Default value is "0".
13756 segment_clocktime_wrap_duration duration
13758 Force the segmenter to only start a new segment if a packet reaches
13759 the muxer within the specified duration after the segmenting clock
13760 time. This way you can make the segmenter more resilient to
13761 backward local time jumps, such as leap seconds or transition to
13762 standard time from daylight savings time.
13763 Default is the maximum possible duration which means starting a new
13765 segment regardless of the elapsed time since the last clock time.
13766 segment_time_delta delta
13768 Specify the accuracy time when selecting the start time for a
13769 segment, expressed as a duration specification. Default value is
13770 "0".
13771 When delta is specified a key-frame will start a new segment if its
13773 PTS satisfies the relation:
13774 PTS >= start_time - time_delta
13776 This option is useful when splitting video content, which is always
13778 split at GOP boundaries, in case a key frame is found just before
13779 the specified split time.
13780 In particular may be used in combination with the ffmpeg option
13782 force_key_frames. The key frame times specified by force_key_frames
13783 may not be set accurately because of rounding issues, with the
13784 consequence that a key frame time may result set just before the
13785 specified time. For constant frame rate videos a value of
13786 1/(2*frame_rate) should address the worst case mismatch between the
13787 specified time and the time set by force_key_frames.
13788 segment_times times
13790 Specify a list of split points. times contains a list of comma
13791 separated duration specifications, in increasing order. See also
13792 the segment_time option.
13793 segment_frames frames
13795 Specify a list of split video frame numbers. frames contains a list
13796 of comma separated integer numbers, in increasing order.
13797 This option specifies to start a new segment whenever a reference
13799 stream key frame is found and the sequential number (starting from
13800 0) of the frame is greater or equal to the next value in the list.
13801 segment_wrap limit
13803 Wrap around segment index once it reaches limit.
13804 segment_start_number number
13806 Set the sequence number of the first segment. Defaults to 0.
13807 strftime 1|0
13809 Use the "strftime" function to define the name of the new segments
13810 to write. If this is selected, the output segment name must contain
13811 a "strftime" function template. Default value is 0.
13812 break_non_keyframes 1|0
13814 If enabled, allow segments to start on frames other than keyframes.
13815 This improves behavior on some players when the time between
13816 keyframes is inconsistent, but may make things worse on others, and
13817 can cause some oddities during seeking. Defaults to 0.
13818 reset_timestamps 1|0
13820 Reset timestamps at the beginning of each segment, so that each
13821 segment will start with near-zero timestamps. It is meant to ease
13822 the playback of the generated segments. May not work with some
13823 combinations of muxers/codecs. It is set to 0 by default.
13824 initial_offset offset
13826 Specify timestamp offset to apply to the output packet timestamps.
13827 The argument must be a time duration specification, and defaults to
13828 0.
13829 write_empty_segments 1|0
13831 If enabled, write an empty segment if there are no packets during
13832 the period a segment would usually span. Otherwise, the segment
13833 will be filled with the next packet written. Defaults to 0.
13834 Make sure to require a closed GOP when encoding and to set the GOP size
13836 to fit your segment time constraint.
13837 Examples
13839 • Remux the content of file in.mkv to a list of segments out-000.nut,
13841 out-001.nut, etc., and write the list of generated segments to
13842 out.list:
13843 ffmpeg -i in.mkv -codec hevc -flags +cgop -g 60 -map 0 -f segment -segment_list out.list out%03d.nut
13845 • Segment input and set output format options for the output
13847 segments:
13848 ffmpeg -i in.mkv -f segment -segment_time 10 -segment_format_options movflags=+faststart out%03d.mp4
13850 • Segment the input file according to the split points specified by
13852 the segment_times option:
13853 ffmpeg -i in.mkv -codec copy -map 0 -f segment -segment_list out.csv -segment_times 1,2,3,5,8,13,21 out%03d.nut
13855 • Use the ffmpeg force_key_frames option to force key frames in the
13857 input at the specified location, together with the segment option
13858 segment_time_delta to account for possible roundings operated when
13859 setting key frame times.
13860 ffmpeg -i in.mkv -force_key_frames 1,2,3,5,8,13,21 -codec:v mpeg4 -codec:a pcm_s16le -map 0 \
13862 -f segment -segment_list out.csv -segment_times 1,2,3,5,8,13,21 -segment_time_delta 0.05 out%03d.nut
13863 In order to force key frames on the input file, transcoding is
13865 required.
13866 • Segment the input file by splitting the input file according to the
13868 frame numbers sequence specified with the segment_frames option:
13869 ffmpeg -i in.mkv -codec copy -map 0 -f segment -segment_list out.csv -segment_frames 100,200,300,500,800 out%03d.nut
13871 • Convert the in.mkv to TS segments using the "libx264" and "aac"
13873 encoders:
13874 ffmpeg -i in.mkv -map 0 -codec:v libx264 -codec:a aac -f ssegment -segment_list out.list out%03d.ts
13876 • Segment the input file, and create an M3U8 live playlist (can be
13878 used as live HLS source):
13879 ffmpeg -re -i in.mkv -codec copy -map 0 -f segment -segment_list playlist.m3u8 \
13881 -segment_list_flags +live -segment_time 10 out%03d.mkv
13882 smoothstreaming
13884 Smooth Streaming muxer generates a set of files (Manifest, chunks)
13885 suitable for serving with conventional web server.
13886 window_size
13888 Specify the number of fragments kept in the manifest. Default 0
13889 (keep all).
13890 extra_window_size
13892 Specify the number of fragments kept outside of the manifest before
13893 removing from disk. Default 5.
13894 lookahead_count
13896 Specify the number of lookahead fragments. Default 2.
13897 min_frag_duration
13899 Specify the minimum fragment duration (in microseconds). Default
13900 5000000.
13901 remove_at_exit
13903 Specify whether to remove all fragments when finished. Default 0
13904 (do not remove).
13905 streamhash
13907 Per stream hash testing format.
13908 This muxer computes and prints a cryptographic hash of all the input
13910 frames, on a per-stream basis. This can be used for equality checks
13911 without having to do a complete binary comparison.
13912 By default audio frames are converted to signed 16-bit raw audio and
13914 video frames to raw video before computing the hash, but the output of
13915 explicit conversions to other codecs can also be used. Timestamps are
13916 ignored. It uses the SHA-256 cryptographic hash function by default,
13917 but supports several other algorithms.
13918 The output of the muxer consists of one line per stream of the form:
13920 streamindex,streamtype,algo=hash, where streamindex is the index of the
13921 mapped stream, streamtype is a single character indicating the type of
13922 stream, algo is a short string representing the hash function used, and
13923 hash is a hexadecimal number representing the computed hash.
13924 hash algorithm
13926 Use the cryptographic hash function specified by the string
13927 algorithm. Supported values include "MD5", "murmur3", "RIPEMD128",
13928 "RIPEMD160", "RIPEMD256", "RIPEMD320", "SHA160", "SHA224", "SHA256"
13929 (default), "SHA512/224", "SHA512/256", "SHA384", "SHA512", "CRC32"
13930 and "adler32".
13931 Examples
13933 To compute the SHA-256 hash of the input converted to raw audio and
13935 video, and store it in the file out.sha256:
13936 ffmpeg -i INPUT -f streamhash out.sha256
13938 To print an MD5 hash to stdout use the command:
13940 ffmpeg -i INPUT -f streamhash -hash md5 -
13942 See also the hash and framehash muxers.
13944 tee
13946 The tee muxer can be used to write the same data to several outputs,
13947 such as files or streams. It can be used, for example, to stream a
13948 video over a network and save it to disk at the same time.
13949 It is different from specifying several outputs to the ffmpeg command-
13951 line tool. With the tee muxer, the audio and video data will be encoded
13952 only once. With conventional multiple outputs, multiple encoding
13953 operations in parallel are initiated, which can be a very expensive
13954 process. The tee muxer is not useful when using the libavformat API
13955 directly because it is then possible to feed the same packets to
13956 several muxers directly.
13957 Since the tee muxer does not represent any particular output format,
13959 ffmpeg cannot auto-select output streams. So all streams intended for
13960 output must be specified using "-map". See the examples below.
13961 Some encoders may need different options depending on the output
13963 format; the auto-detection of this can not work with the tee muxer, so
13964 they need to be explicitly specified. The main example is the
13965 global_header flag.
13966 The slave outputs are specified in the file name given to the muxer,
13968 separated by '|'. If any of the slave name contains the '|' separator,
13969 leading or trailing spaces or any special character, those must be
13970 escaped (see the "Quoting and escaping" section in the ffmpeg-utils(1)
13971 manual).
13972 Options
13974 use_fifo bool
13976 If set to 1, slave outputs will be processed in separate threads
13977 using the fifo muxer. This allows to compensate for different
13978 speed/latency/reliability of outputs and setup transparent
13979 recovery. By default this feature is turned off.
13980 fifo_options
13982 Options to pass to fifo pseudo-muxer instances. See fifo.
13983 Muxer options can be specified for each slave by prepending them as a
13985 list of key=value pairs separated by ':', between square brackets. If
13986 the options values contain a special character or the ':' separator,
13987 they must be escaped; note that this is a second level escaping.
13988 The following special options are also recognized:
13990 f Specify the format name. Required if it cannot be guessed from the
13992 output URL.
13993 bsfs[/spec]
13995 Specify a list of bitstream filters to apply to the specified
13996 output.
13997 It is possible to specify to which streams a given bitstream filter
13999 applies, by appending a stream specifier to the option separated by
14000 "/". spec must be a stream specifier (see Format stream
14001 specifiers).
14002 If the stream specifier is not specified, the bitstream filters
14004 will be applied to all streams in the output. This will cause that
14005 output operation to fail if the output contains streams to which
14006 the bitstream filter cannot be applied e.g. "h264_mp4toannexb"
14007 being applied to an output containing an audio stream.
14008 Options for a bitstream filter must be specified in the form of
14010 "opt=value".
14011 Several bitstream filters can be specified, separated by ",".
14013 use_fifo bool
14015 This allows to override tee muxer use_fifo option for individual
14016 slave muxer.
14017 fifo_options
14019 This allows to override tee muxer fifo_options for individual slave
14020 muxer. See fifo.
14021 select
14023 Select the streams that should be mapped to the slave output,
14024 specified by a stream specifier. If not specified, this defaults to
14025 all the mapped streams. This will cause that output operation to
14026 fail if the output format does not accept all mapped streams.
14027 You may use multiple stream specifiers separated by commas (",")
14029 e.g.: "a:0,v"
14030 onfail
14032 Specify behaviour on output failure. This can be set to either
14033 "abort" (which is default) or "ignore". "abort" will cause whole
14034 process to fail in case of failure on this slave output. "ignore"
14035 will ignore failure on this output, so other outputs will continue
14036 without being affected.
14037 Examples
14039 • Encode something and both archive it in a WebM file and stream it
14041 as MPEG-TS over UDP:
14042 ffmpeg -i ... -c:v libx264 -c:a mp2 -f tee -map 0:v -map 0:a
14044 "archive-20121107.mkv|[f=mpegts]udp://10.0.1.255:1234/"
14045 • As above, but continue streaming even if output to local file fails
14047 (for example local drive fills up):
14048 ffmpeg -i ... -c:v libx264 -c:a mp2 -f tee -map 0:v -map 0:a
14050 "[onfail=ignore]archive-20121107.mkv|[f=mpegts]udp://10.0.1.255:1234/"
14051 • Use ffmpeg to encode the input, and send the output to three
14053 different destinations. The "dump_extra" bitstream filter is used
14054 to add extradata information to all the output video keyframes
14055 packets, as requested by the MPEG-TS format. The select option is
14056 applied to out.aac in order to make it contain only audio packets.
14057 ffmpeg -i ... -map 0 -flags +global_header -c:v libx264 -c:a aac
14059 -f tee "[bsfs/v=dump_extra=freq=keyframe]out.ts|[movflags=+faststart]out.mp4|[select=a]out.aac"
14060 • As above, but select only stream "a:1" for the audio output. Note
14062 that a second level escaping must be performed, as ":" is a special
14063 character used to separate options.
14064 ffmpeg -i ... -map 0 -flags +global_header -c:v libx264 -c:a aac
14066 -f tee "[bsfs/v=dump_extra=freq=keyframe]out.ts|[movflags=+faststart]out.mp4|[select=\'a:1\']out.aac"
14067 webm_chunk
14069 WebM Live Chunk Muxer.
14070 This muxer writes out WebM headers and chunks as separate files which
14072 can be consumed by clients that support WebM Live streams via DASH.
14073 Options
14075 This muxer supports the following options:
14077 chunk_start_index
14079 Index of the first chunk (defaults to 0).
14080 header
14082 Filename of the header where the initialization data will be
14083 written.
14084 audio_chunk_duration
14086 Duration of each audio chunk in milliseconds (defaults to 5000).
14087 Example
14089 ffmpeg -f v4l2 -i /dev/video0 \
14091 -f alsa -i hw:0 \
14092 -map 0:0 \
14093 -c:v libvpx-vp9 \
14094 -s 640x360 -keyint_min 30 -g 30 \
14095 -f webm_chunk \
14096 -header webm_live_video_360.hdr \
14097 -chunk_start_index 1 \
14098 webm_live_video_360_%d.chk \
14099 -map 1:0 \
14100 -c:a libvorbis \
14101 -b:a 128k \
14102 -f webm_chunk \
14103 -header webm_live_audio_128.hdr \
14104 -chunk_start_index 1 \
14105 -audio_chunk_duration 1000 \
14106 webm_live_audio_128_%d.chk
14107 webm_dash_manifest
14109 WebM DASH Manifest muxer.
14110 This muxer implements the WebM DASH Manifest specification to generate
14112 the DASH manifest XML. It also supports manifest generation for DASH
14113 live streams.
14114 For more information see:
14116 • WebM DASH Specification:
14118 <https://sites.google.com/a/webmproject.org/wiki/adaptive-streaming/webm-dash-specification>
14119 • ISO DASH Specification:
14121 <http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip>
14122 Options
14124 This muxer supports the following options:
14126 adaptation_sets
14128 This option has the following syntax: "id=x,streams=a,b,c
14129 id=y,streams=d,e" where x and y are the unique identifiers of the
14130 adaptation sets and a,b,c,d and e are the indices of the
14131 corresponding audio and video streams. Any number of adaptation
14132 sets can be added using this option.
14133 live
14135 Set this to 1 to create a live stream DASH Manifest. Default: 0.
14136 chunk_start_index
14138 Start index of the first chunk. This will go in the startNumber
14139 attribute of the SegmentTemplate element in the manifest. Default:
14140 0.
14141 chunk_duration_ms
14143 Duration of each chunk in milliseconds. This will go in the
14144 duration attribute of the SegmentTemplate element in the manifest.
14145 Default: 1000.
14146 utc_timing_url
14148 URL of the page that will return the UTC timestamp in ISO format.
14149 This will go in the value attribute of the UTCTiming element in the
14150 manifest. Default: None.
14151 time_shift_buffer_depth
14153 Smallest time (in seconds) shifting buffer for which any
14154 Representation is guaranteed to be available. This will go in the
14155 timeShiftBufferDepth attribute of the MPD element. Default: 60.
14156 minimum_update_period
14158 Minimum update period (in seconds) of the manifest. This will go in
14159 the minimumUpdatePeriod attribute of the MPD element. Default: 0.
14160 Example
14162 ffmpeg -f webm_dash_manifest -i video1.webm \
14164 -f webm_dash_manifest -i video2.webm \
14165 -f webm_dash_manifest -i audio1.webm \
14166 -f webm_dash_manifest -i audio2.webm \
14167 -map 0 -map 1 -map 2 -map 3 \
14168 -c copy \
14169 -f webm_dash_manifest \
14170 -adaptation_sets "id=0,streams=0,1 id=1,streams=2,3" \
14171 manifest.xml
14172
14174 FFmpeg is able to dump metadata from media files into a simple
14175 UTF-8-encoded INI-like text file and then load it back using the
14176 metadata muxer/demuxer.
14177 The file format is as follows:
14179 1. A file consists of a header and a number of metadata tags divided
14181 into sections, each on its own line.
14182 2. The header is a ;FFMETADATA string, followed by a version number
14184 (now 1).
14185 3. Metadata tags are of the form key=value
14187 4. Immediately after header follows global metadata
14189 5. After global metadata there may be sections with
14191 per-stream/per-chapter metadata.
14192 6. A section starts with the section name in uppercase (i.e. STREAM or
14194 CHAPTER) in brackets ([, ]) and ends with next section or end of
14195 file.
14196 7. At the beginning of a chapter section there may be an optional
14198 timebase to be used for start/end values. It must be in form
14199 TIMEBASE=num/den, where num and den are integers. If the timebase
14200 is missing then start/end times are assumed to be in nanoseconds.
14201 Next a chapter section must contain chapter start and end times in
14203 form START=num, END=num, where num is a positive integer.
14204 8. Empty lines and lines starting with ; or # are ignored.
14206 9. Metadata keys or values containing special characters (=, ;, #, \
14208 and a newline) must be escaped with a backslash \.
14209 10. Note that whitespace in metadata (e.g. foo = bar) is considered to
14211 be a part of the tag (in the example above key is foo , value is
14212 bar).
14213 A ffmetadata file might look like this:
14215 ;FFMETADATA1
14217 title=bike\\shed
14218 ;this is a comment
14219 artist=FFmpeg troll team
14220 [CHAPTER]
14222 TIMEBASE=1/1000
14223 START=0
14224 #chapter ends at 0:01:00
14225 END=60000
14226 title=chapter \#1
14227 [STREAM]
14228 title=multi\
14229 line
14230 By using the ffmetadata muxer and demuxer it is possible to extract
14232 metadata from an input file to an ffmetadata file, and then transcode
14233 the file into an output file with the edited ffmetadata file.
14234 Extracting an ffmetadata file with ffmpeg goes as follows:
14236 ffmpeg -i INPUT -f ffmetadata FFMETADATAFILE
14238 Reinserting edited metadata information from the FFMETADATAFILE file
14240 can be done as:
14241 ffmpeg -i INPUT -i FFMETADATAFILE -map_metadata 1 -codec copy OUTPUT
14243
14245 The libavformat library provides some generic global options, which can
14246 be set on all the protocols. In addition each protocol may support so-
14247 called private options, which are specific for that component.
14248 Options may be set by specifying -option value in the FFmpeg tools, or
14250 by setting the value explicitly in the "AVFormatContext" options or
14251 using the libavutil/opt.h API for programmatic use.
14252 The list of supported options follows:
14254 protocol_whitelist list (input)
14256 Set a ","-separated list of allowed protocols. "ALL" matches all
14257 protocols. Protocols prefixed by "-" are disabled. All protocols
14258 are allowed by default but protocols used by an another protocol
14259 (nested protocols) are restricted to a per protocol subset.
14260
14262 Protocols are configured elements in FFmpeg that enable access to
14263 resources that require specific protocols.
14264 When you configure your FFmpeg build, all the supported protocols are
14266 enabled by default. You can list all available ones using the configure
14267 option "--list-protocols".
14268 You can disable all the protocols using the configure option
14270 "--disable-protocols", and selectively enable a protocol using the
14271 option "--enable-protocol=PROTOCOL", or you can disable a particular
14272 protocol using the option "--disable-protocol=PROTOCOL".
14273 The option "-protocols" of the ff* tools will display the list of
14275 supported protocols.
14276 All protocols accept the following options:
14278 rw_timeout
14280 Maximum time to wait for (network) read/write operations to
14281 complete, in microseconds.
14282 A description of the currently available protocols follows.
14284 amqp
14286 Advanced Message Queueing Protocol (AMQP) version 0-9-1 is a broker
14287 based publish-subscribe communication protocol.
14288 FFmpeg must be compiled with --enable-librabbitmq to support AMQP. A
14290 separate AMQP broker must also be run. An example open-source AMQP
14291 broker is RabbitMQ.
14292 After starting the broker, an FFmpeg client may stream data to the
14294 broker using the command:
14295 ffmpeg -re -i input -f mpegts amqp://[[user]:[password]@]hostname[:port][/vhost]
14297 Where hostname and port (default is 5672) is the address of the broker.
14299 The client may also set a user/password for authentication. The default
14300 for both fields is "guest". Name of virtual host on broker can be set
14301 with vhost. The default value is "/".
14302 Muliple subscribers may stream from the broker using the command:
14304 ffplay amqp://[[user]:[password]@]hostname[:port][/vhost]
14306 In RabbitMQ all data published to the broker flows through a specific
14308 exchange, and each subscribing client has an assigned queue/buffer.
14309 When a packet arrives at an exchange, it may be copied to a client's
14310 queue depending on the exchange and routing_key fields.
14311 The following options are supported:
14313 exchange
14315 Sets the exchange to use on the broker. RabbitMQ has several
14316 predefined exchanges: "amq.direct" is the default exchange, where
14317 the publisher and subscriber must have a matching routing_key;
14318 "amq.fanout" is the same as a broadcast operation (i.e. the data is
14319 forwarded to all queues on the fanout exchange independent of the
14320 routing_key); and "amq.topic" is similar to "amq.direct", but
14321 allows for more complex pattern matching (refer to the RabbitMQ
14322 documentation).
14323 routing_key
14325 Sets the routing key. The default value is "amqp". The routing key
14326 is used on the "amq.direct" and "amq.topic" exchanges to decide
14327 whether packets are written to the queue of a subscriber.
14328 pkt_size
14330 Maximum size of each packet sent/received to the broker. Default is
14331 131072. Minimum is 4096 and max is any large value (representable
14332 by an int). When receiving packets, this sets an internal buffer
14333 size in FFmpeg. It should be equal to or greater than the size of
14334 the published packets to the broker. Otherwise the received message
14335 may be truncated causing decoding errors.
14336 connection_timeout
14338 The timeout in seconds during the initial connection to the broker.
14339 The default value is rw_timeout, or 5 seconds if rw_timeout is not
14340 set.
14341 delivery_mode mode
14343 Sets the delivery mode of each message sent to broker. The
14344 following values are accepted:
14345 persistent
14347 Delivery mode set to "persistent" (2). This is the default
14348 value. Messages may be written to the broker's disk depending
14349 on its setup.
14350 non-persistent
14352 Delivery mode set to "non-persistent" (1). Messages will stay
14353 in broker's memory unless the broker is under memory pressure.
14354 async
14356 Asynchronous data filling wrapper for input stream.
14357 Fill data in a background thread, to decouple I/O operation from demux
14359 thread.
14360 async:<URL>
14362 async:http://host/resource
14363 async:cache:http://host/resource
14364 bluray
14366 Read BluRay playlist.
14367 The accepted options are:
14369 angle
14371 BluRay angle
14372 chapter
14374 Start chapter (1...N)
14375 playlist
14377 Playlist to read (BDMV/PLAYLIST/?????.mpls)
14378 Examples:
14380 Read longest playlist from BluRay mounted to /mnt/bluray:
14382 bluray:/mnt/bluray
14384 Read angle 2 of playlist 4 from BluRay mounted to /mnt/bluray, start
14386 from chapter 2:
14387 -playlist 4 -angle 2 -chapter 2 bluray:/mnt/bluray
14389 cache
14391 Caching wrapper for input stream.
14392 Cache the input stream to temporary file. It brings seeking capability
14394 to live streams.
14395 The accepted options are:
14397 read_ahead_limit
14399 Amount in bytes that may be read ahead when seeking isn't
14400 supported. Range is -1 to INT_MAX. -1 for unlimited. Default is
14401 65536.
14402 URL Syntax is
14404 cache:<URL>
14406 concat
14408 Physical concatenation protocol.
14409 Read and seek from many resources in sequence as if they were a unique
14411 resource.
14412 A URL accepted by this protocol has the syntax:
14414 concat:<URL1>|<URL2>|...|<URLN>
14416 where URL1, URL2, ..., URLN are the urls of the resource to be
14418 concatenated, each one possibly specifying a distinct protocol.
14419 For example to read a sequence of files split1.mpeg, split2.mpeg,
14421 split3.mpeg with ffplay use the command:
14422 ffplay concat:split1.mpeg\|split2.mpeg\|split3.mpeg
14424 Note that you may need to escape the character "|" which is special for
14426 many shells.
14427 concatf
14429 Physical concatenation protocol using a line break delimited list of
14430 resources.
14431 Read and seek from many resources in sequence as if they were a unique
14433 resource.
14434 A URL accepted by this protocol has the syntax:
14436 concatf:<URL>
14438 where URL is the url containing a line break delimited list of
14440 resources to be concatenated, each one possibly specifying a distinct
14441 protocol. Special characters must be escaped with backslash or single
14442 quotes. See the "Quoting and escaping" section in the ffmpeg-utils(1)
14443 manual.
14444 For example to read a sequence of files split1.mpeg, split2.mpeg,
14446 split3.mpeg listed in separate lines within a file split.txt with
14447 ffplay use the command:
14448 ffplay concatf:split.txt
14450 Where split.txt contains the lines:
14452 split1.mpeg
14454 split2.mpeg
14455 split3.mpeg
14456 crypto
14458 AES-encrypted stream reading protocol.
14459 The accepted options are:
14461 key Set the AES decryption key binary block from given hexadecimal
14463 representation.
14464 iv Set the AES decryption initialization vector binary block from
14466 given hexadecimal representation.
14467 Accepted URL formats:
14469 crypto:<URL>
14471 crypto+<URL>
14472 data
14474 Data in-line in the URI. See
14475 <http://en.wikipedia.org/wiki/Data_URI_scheme>.
14476 For example, to convert a GIF file given inline with ffmpeg:
14478 ffmpeg -i "data:image/gif;base64,R0lGODdhCAAIAMIEAAAAAAAA//8AAP//AP///////////////ywAAAAACAAIAAADF0gEDLojDgdGiJdJqUX02iB4E8Q9jUMkADs=" smiley.png
14480 fd
14482 File descriptor access protocol.
14483 The accepted syntax is:
14485 fd: -fd <file_descriptor>
14487 If fd is not specified, by default the stdout file descriptor will be
14489 used for writing, stdin for reading. Unlike the pipe protocol, fd
14490 protocol has seek support if it corresponding to a regular file. fd
14491 protocol doesn't support pass file descriptor via URL for security.
14492 This protocol accepts the following options:
14494 blocksize
14496 Set I/O operation maximum block size, in bytes. Default value is
14497 "INT_MAX", which results in not limiting the requested block size.
14498 Setting this value reasonably low improves user termination request
14499 reaction time, which is valuable if data transmission is slow.
14500 fd Set file descriptor.
14502 file
14504 File access protocol.
14505 Read from or write to a file.
14507 A file URL can have the form:
14509 file:<filename>
14511 where filename is the path of the file to read.
14513 An URL that does not have a protocol prefix will be assumed to be a
14515 file URL. Depending on the build, an URL that looks like a Windows path
14516 with the drive letter at the beginning will also be assumed to be a
14517 file URL (usually not the case in builds for unix-like systems).
14518 For example to read from a file input.mpeg with ffmpeg use the command:
14520 ffmpeg -i file:input.mpeg output.mpeg
14522 This protocol accepts the following options:
14524 truncate
14526 Truncate existing files on write, if set to 1. A value of 0
14527 prevents truncating. Default value is 1.
14528 blocksize
14530 Set I/O operation maximum block size, in bytes. Default value is
14531 "INT_MAX", which results in not limiting the requested block size.
14532 Setting this value reasonably low improves user termination request
14533 reaction time, which is valuable for files on slow medium.
14534 follow
14536 If set to 1, the protocol will retry reading at the end of the
14537 file, allowing reading files that still are being written. In order
14538 for this to terminate, you either need to use the rw_timeout
14539 option, or use the interrupt callback (for API users).
14540 seekable
14542 Controls if seekability is advertised on the file. 0 means non-
14543 seekable, -1 means auto (seekable for normal files, non-seekable
14544 for named pipes).
14545 Many demuxers handle seekable and non-seekable resources
14547 differently, overriding this might speed up opening certain files
14548 at the cost of losing some features (e.g. accurate seeking).
14549 ftp
14551 FTP (File Transfer Protocol).
14552 Read from or write to remote resources using FTP protocol.
14554 Following syntax is required.
14556 ftp://[user[:password]@]server[:port]/path/to/remote/resource.mpeg
14558 This protocol accepts the following options.
14560 timeout
14562 Set timeout in microseconds of socket I/O operations used by the
14563 underlying low level operation. By default it is set to -1, which
14564 means that the timeout is not specified.
14565 ftp-user
14567 Set a user to be used for authenticating to the FTP server. This is
14568 overridden by the user in the FTP URL.
14569 ftp-password
14571 Set a password to be used for authenticating to the FTP server.
14572 This is overridden by the password in the FTP URL, or by ftp-
14573 anonymous-password if no user is set.
14574 ftp-anonymous-password
14576 Password used when login as anonymous user. Typically an e-mail
14577 address should be used.
14578 ftp-write-seekable
14580 Control seekability of connection during encoding. If set to 1 the
14581 resource is supposed to be seekable, if set to 0 it is assumed not
14582 to be seekable. Default value is 0.
14583 NOTE: Protocol can be used as output, but it is recommended to not do
14585 it, unless special care is taken (tests, customized server
14586 configuration etc.). Different FTP servers behave in different way
14587 during seek operation. ff* tools may produce incomplete content due to
14588 server limitations.
14589 gopher
14591 Gopher protocol.
14592 gophers
14594 Gophers protocol.
14595 The Gopher protocol with TLS encapsulation.
14597 hls
14599 Read Apple HTTP Live Streaming compliant segmented stream as a uniform
14600 one. The M3U8 playlists describing the segments can be remote HTTP
14601 resources or local files, accessed using the standard file protocol.
14602 The nested protocol is declared by specifying "+proto" after the hls
14603 URI scheme name, where proto is either "file" or "http".
14604 hls+http://host/path/to/remote/resource.m3u8
14606 hls+file://path/to/local/resource.m3u8
14607 Using this protocol is discouraged - the hls demuxer should work just
14609 as well (if not, please report the issues) and is more complete. To
14610 use the hls demuxer instead, simply use the direct URLs to the m3u8
14611 files.
14612 http
14614 HTTP (Hyper Text Transfer Protocol).
14615 This protocol accepts the following options:
14617 seekable
14619 Control seekability of connection. If set to 1 the resource is
14620 supposed to be seekable, if set to 0 it is assumed not to be
14621 seekable, if set to -1 it will try to autodetect if it is seekable.
14622 Default value is -1.
14623 chunked_post
14625 If set to 1 use chunked Transfer-Encoding for posts, default is 1.
14626 content_type
14628 Set a specific content type for the POST messages or for listen
14629 mode.
14630 http_proxy
14632 set HTTP proxy to tunnel through e.g. http://example.com:1234
14633 headers
14635 Set custom HTTP headers, can override built in default headers. The
14636 value must be a string encoding the headers.
14637 multiple_requests
14639 Use persistent connections if set to 1, default is 0.
14640 post_data
14642 Set custom HTTP post data.
14643 referer
14645 Set the Referer header. Include 'Referer: URL' header in HTTP
14646 request.
14647 user_agent
14649 Override the User-Agent header. If not specified the protocol will
14650 use a string describing the libavformat build. ("Lavf/<version>")
14651 reconnect_at_eof
14653 If set then eof is treated like an error and causes reconnection,
14654 this is useful for live / endless streams.
14655 reconnect_streamed
14657 If set then even streamed/non seekable streams will be reconnected
14658 on errors.
14659 reconnect_on_network_error
14661 Reconnect automatically in case of TCP/TLS errors during connect.
14662 reconnect_on_http_error
14664 A comma separated list of HTTP status codes to reconnect on. The
14665 list can include specific status codes (e.g. '503') or the strings
14666 '4xx' / '5xx'.
14667 reconnect_delay_max
14669 Sets the maximum delay in seconds after which to give up
14670 reconnecting
14671 mime_type
14673 Export the MIME type.
14674 http_version
14676 Exports the HTTP response version number. Usually "1.0" or "1.1".
14677 icy If set to 1 request ICY (SHOUTcast) metadata from the server. If
14679 the server supports this, the metadata has to be retrieved by the
14680 application by reading the icy_metadata_headers and
14681 icy_metadata_packet options. The default is 1.
14682 icy_metadata_headers
14684 If the server supports ICY metadata, this contains the ICY-specific
14685 HTTP reply headers, separated by newline characters.
14686 icy_metadata_packet
14688 If the server supports ICY metadata, and icy was set to 1, this
14689 contains the last non-empty metadata packet sent by the server. It
14690 should be polled in regular intervals by applications interested in
14691 mid-stream metadata updates.
14692 cookies
14694 Set the cookies to be sent in future requests. The format of each
14695 cookie is the same as the value of a Set-Cookie HTTP response
14696 field. Multiple cookies can be delimited by a newline character.
14697 offset
14699 Set initial byte offset.
14700 end_offset
14702 Try to limit the request to bytes preceding this offset.
14703 method
14705 When used as a client option it sets the HTTP method for the
14706 request.
14707 When used as a server option it sets the HTTP method that is going
14709 to be expected from the client(s). If the expected and the
14710 received HTTP method do not match the client will be given a Bad
14711 Request response. When unset the HTTP method is not checked for
14712 now. This will be replaced by autodetection in the future.
14713 listen
14715 If set to 1 enables experimental HTTP server. This can be used to
14716 send data when used as an output option, or read data from a client
14717 with HTTP POST when used as an input option. If set to 2 enables
14718 experimental multi-client HTTP server. This is not yet implemented
14719 in ffmpeg.c and thus must not be used as a command line option.
14720 # Server side (sending):
14722 ffmpeg -i somefile.ogg -c copy -listen 1 -f ogg http://<server>:<port>
14723 # Client side (receiving):
14725 ffmpeg -i http://<server>:<port> -c copy somefile.ogg
14726 # Client can also be done with wget:
14728 wget http://<server>:<port> -O somefile.ogg
14729 # Server side (receiving):
14731 ffmpeg -listen 1 -i http://<server>:<port> -c copy somefile.ogg
14732 # Client side (sending):
14734 ffmpeg -i somefile.ogg -chunked_post 0 -c copy -f ogg http://<server>:<port>
14735 # Client can also be done with wget:
14737 wget --post-file=somefile.ogg http://<server>:<port>
14738 send_expect_100
14740 Send an Expect: 100-continue header for POST. If set to 1 it will
14741 send, if set to 0 it won't, if set to -1 it will try to send if it
14742 is applicable. Default value is -1.
14743 auth_type
14745 Set HTTP authentication type. No option for Digest, since this
14746 method requires getting nonce parameters from the server first and
14747 can't be used straight away like Basic.
14748 none
14750 Choose the HTTP authentication type automatically. This is the
14751 default.
14752 basic
14754 Choose the HTTP basic authentication.
14755 Basic authentication sends a Base64-encoded string that
14757 contains a user name and password for the client. Base64 is not
14758 a form of encryption and should be considered the same as
14759 sending the user name and password in clear text (Base64 is a
14760 reversible encoding). If a resource needs to be protected,
14761 strongly consider using an authentication scheme other than
14762 basic authentication. HTTPS/TLS should be used with basic
14763 authentication. Without these additional security
14764 enhancements, basic authentication should not be used to
14765 protect sensitive or valuable information.
14766 HTTP Cookies
14768 Some HTTP requests will be denied unless cookie values are passed in
14770 with the request. The cookies option allows these cookies to be
14771 specified. At the very least, each cookie must specify a value along
14772 with a path and domain. HTTP requests that match both the domain and
14773 path will automatically include the cookie value in the HTTP Cookie
14774 header field. Multiple cookies can be delimited by a newline.
14775 The required syntax to play a stream specifying a cookie is:
14777 ffplay -cookies "nlqptid=nltid=tsn; path=/; domain=somedomain.com;" http://somedomain.com/somestream.m3u8
14779 Icecast
14781 Icecast protocol (stream to Icecast servers)
14782 This protocol accepts the following options:
14784 ice_genre
14786 Set the stream genre.
14787 ice_name
14789 Set the stream name.
14790 ice_description
14792 Set the stream description.
14793 ice_url
14795 Set the stream website URL.
14796 ice_public
14798 Set if the stream should be public. The default is 0 (not public).
14799 user_agent
14801 Override the User-Agent header. If not specified a string of the
14802 form "Lavf/<version>" will be used.
14803 password
14805 Set the Icecast mountpoint password.
14806 content_type
14808 Set the stream content type. This must be set if it is different
14809 from audio/mpeg.
14810 legacy_icecast
14812 This enables support for Icecast versions < 2.4.0, that do not
14813 support the HTTP PUT method but the SOURCE method.
14814 tls Establish a TLS (HTTPS) connection to Icecast.
14816 icecast://[<username>[:<password>]@]<server>:<port>/<mountpoint>
14818 ipfs
14820 InterPlanetary File System (IPFS) protocol support. One can access
14821 files stored on the IPFS network through so-called gateways. These are
14822 http(s) endpoints. This protocol wraps the IPFS native protocols
14823 (ipfs:// and ipns://) to be sent to such a gateway. Users can (and
14824 should) host their own node which means this protocol will use one's
14825 local gateway to access files on the IPFS network.
14826 This protocol accepts the following options:
14828 gateway
14830 Defines the gateway to use. When not set, the protocol will first
14831 try locating the local gateway by looking at $IPFS_GATEWAY,
14832 $IPFS_PATH and "$HOME/.ipfs/", in that order.
14833 One can use this protocol in 2 ways. Using IPFS:
14835 ffplay ipfs://<hash>
14837 Or the IPNS protocol (IPNS is mutable IPFS):
14839 ffplay ipns://<hash>
14841 mmst
14843 MMS (Microsoft Media Server) protocol over TCP.
14844 mmsh
14846 MMS (Microsoft Media Server) protocol over HTTP.
14847 The required syntax is:
14849 mmsh://<server>[:<port>][/<app>][/<playpath>]
14851 md5
14853 MD5 output protocol.
14854 Computes the MD5 hash of the data to be written, and on close writes
14856 this to the designated output or stdout if none is specified. It can be
14857 used to test muxers without writing an actual file.
14858 Some examples follow.
14860 # Write the MD5 hash of the encoded AVI file to the file output.avi.md5.
14862 ffmpeg -i input.flv -f avi -y md5:output.avi.md5
14863 # Write the MD5 hash of the encoded AVI file to stdout.
14865 ffmpeg -i input.flv -f avi -y md5:
14866 Note that some formats (typically MOV) require the output protocol to
14868 be seekable, so they will fail with the MD5 output protocol.
14869 pipe
14871 UNIX pipe access protocol.
14872 Read and write from UNIX pipes.
14874 The accepted syntax is:
14876 pipe:[<number>]
14878 If fd isn't specified, number is the number corresponding to the file
14880 descriptor of the pipe (e.g. 0 for stdin, 1 for stdout, 2 for stderr).
14881 If number is not specified, by default the stdout file descriptor will
14882 be used for writing, stdin for reading.
14883 For example to read from stdin with ffmpeg:
14885 cat test.wav | ffmpeg -i pipe:0
14887 # ...this is the same as...
14888 cat test.wav | ffmpeg -i pipe:
14889 For writing to stdout with ffmpeg:
14891 ffmpeg -i test.wav -f avi pipe:1 | cat > test.avi
14893 # ...this is the same as...
14894 ffmpeg -i test.wav -f avi pipe: | cat > test.avi
14895 This protocol accepts the following options:
14897 blocksize
14899 Set I/O operation maximum block size, in bytes. Default value is
14900 "INT_MAX", which results in not limiting the requested block size.
14901 Setting this value reasonably low improves user termination request
14902 reaction time, which is valuable if data transmission is slow.
14903 fd Set file descriptor.
14905 Note that some formats (typically MOV), require the output protocol to
14907 be seekable, so they will fail with the pipe output protocol.
14908 prompeg
14910 Pro-MPEG Code of Practice #3 Release 2 FEC protocol.
14911 The Pro-MPEG CoP#3 FEC is a 2D parity-check forward error correction
14913 mechanism for MPEG-2 Transport Streams sent over RTP.
14914 This protocol must be used in conjunction with the "rtp_mpegts" muxer
14916 and the "rtp" protocol.
14917 The required syntax is:
14919 -f rtp_mpegts -fec prompeg=<option>=<val>... rtp://<hostname>:<port>
14921 The destination UDP ports are "port + 2" for the column FEC stream and
14923 "port + 4" for the row FEC stream.
14924 This protocol accepts the following options:
14926 l=n The number of columns (4-20, LxD <= 100)
14928 d=n The number of rows (4-20, LxD <= 100)
14930 Example usage:
14932 -f rtp_mpegts -fec prompeg=l=8:d=4 rtp://<hostname>:<port>
14934 rist
14936 Reliable Internet Streaming Transport protocol
14937 The accepted options are:
14939 rist_profile
14941 Supported values:
14942 simple
14944 main
14945 This one is default.
14946 advanced
14948 buffer_size
14949 Set internal RIST buffer size in milliseconds for retransmission of
14950 data. Default value is 0 which means the librist default (1 sec).
14951 Maximum value is 30 seconds.
14952 fifo_size
14954 Size of the librist receiver output fifo in number of packets. This
14955 must be a power of 2. Defaults to 8192 (vs the librist default of
14956 1024).
14957 overrun_nonfatal=1|0
14959 Survive in case of librist fifo buffer overrun. Default value is 0.
14960 pkt_size
14962 Set maximum packet size for sending data. 1316 by default.
14963 log_level
14965 Set loglevel for RIST logging messages. You only need to set this
14966 if you explicitly want to enable debug level messages or packet
14967 loss simulation, otherwise the regular loglevel is respected.
14968 secret
14970 Set override of encryption secret, by default is unset.
14971 encryption
14973 Set encryption type, by default is disabled. Acceptable values are
14974 128 and 256.
14975 rtmp
14977 Real-Time Messaging Protocol.
14978 The Real-Time Messaging Protocol (RTMP) is used for streaming
14980 multimedia content across a TCP/IP network.
14981 The required syntax is:
14983 rtmp://[<username>:<password>@]<server>[:<port>][/<app>][/<instance>][/<playpath>]
14985 The accepted parameters are:
14987 username
14989 An optional username (mostly for publishing).
14990 password
14992 An optional password (mostly for publishing).
14993 server
14995 The address of the RTMP server.
14996 port
14998 The number of the TCP port to use (by default is 1935).
14999 app It is the name of the application to access. It usually corresponds
15001 to the path where the application is installed on the RTMP server
15002 (e.g. /ondemand/, /flash/live/, etc.). You can override the value
15003 parsed from the URI through the "rtmp_app" option, too.
15004 playpath
15006 It is the path or name of the resource to play with reference to
15007 the application specified in app, may be prefixed by "mp4:". You
15008 can override the value parsed from the URI through the
15009 "rtmp_playpath" option, too.
15010 listen
15012 Act as a server, listening for an incoming connection.
15013 timeout
15015 Maximum time to wait for the incoming connection. Implies listen.
15016 Additionally, the following parameters can be set via command line
15018 options (or in code via "AVOption"s):
15019 rtmp_app
15021 Name of application to connect on the RTMP server. This option
15022 overrides the parameter specified in the URI.
15023 rtmp_buffer
15025 Set the client buffer time in milliseconds. The default is 3000.
15026 rtmp_conn
15028 Extra arbitrary AMF connection parameters, parsed from a string,
15029 e.g. like "B:1 S:authMe O:1 NN:code:1.23 NS:flag:ok O:0". Each
15030 value is prefixed by a single character denoting the type, B for
15031 Boolean, N for number, S for string, O for object, or Z for null,
15032 followed by a colon. For Booleans the data must be either 0 or 1
15033 for FALSE or TRUE, respectively. Likewise for Objects the data
15034 must be 0 or 1 to end or begin an object, respectively. Data items
15035 in subobjects may be named, by prefixing the type with 'N' and
15036 specifying the name before the value (i.e. "NB:myFlag:1"). This
15037 option may be used multiple times to construct arbitrary AMF
15038 sequences.
15039 rtmp_flashver
15041 Version of the Flash plugin used to run the SWF player. The default
15042 is LNX 9,0,124,2. (When publishing, the default is FMLE/3.0
15043 (compatible; <libavformat version>).)
15044 rtmp_flush_interval
15046 Number of packets flushed in the same request (RTMPT only). The
15047 default is 10.
15048 rtmp_live
15050 Specify that the media is a live stream. No resuming or seeking in
15051 live streams is possible. The default value is "any", which means
15052 the subscriber first tries to play the live stream specified in the
15053 playpath. If a live stream of that name is not found, it plays the
15054 recorded stream. The other possible values are "live" and
15055 "recorded".
15056 rtmp_pageurl
15058 URL of the web page in which the media was embedded. By default no
15059 value will be sent.
15060 rtmp_playpath
15062 Stream identifier to play or to publish. This option overrides the
15063 parameter specified in the URI.
15064 rtmp_subscribe
15066 Name of live stream to subscribe to. By default no value will be
15067 sent. It is only sent if the option is specified or if rtmp_live
15068 is set to live.
15069 rtmp_swfhash
15071 SHA256 hash of the decompressed SWF file (32 bytes).
15072 rtmp_swfsize
15074 Size of the decompressed SWF file, required for SWFVerification.
15075 rtmp_swfurl
15077 URL of the SWF player for the media. By default no value will be
15078 sent.
15079 rtmp_swfverify
15081 URL to player swf file, compute hash/size automatically.
15082 rtmp_tcurl
15084 URL of the target stream. Defaults to proto://host[:port]/app.
15085 tcp_nodelay=1|0
15087 Set TCP_NODELAY to disable Nagle's algorithm. Default value is 0.
15088 Remark: Writing to the socket is currently not optimized to
15090 minimize system calls and reduces the efficiency / effect of
15091 TCP_NODELAY.
15092 For example to read with ffplay a multimedia resource named "sample"
15094 from the application "vod" from an RTMP server "myserver":
15095 ffplay rtmp://myserver/vod/sample
15097 To publish to a password protected server, passing the playpath and app
15099 names separately:
15100 ffmpeg -re -i <input> -f flv -rtmp_playpath some/long/path -rtmp_app long/app/name rtmp://username:password@myserver/
15102 rtmpe
15104 Encrypted Real-Time Messaging Protocol.
15105 The Encrypted Real-Time Messaging Protocol (RTMPE) is used for
15107 streaming multimedia content within standard cryptographic primitives,
15108 consisting of Diffie-Hellman key exchange and HMACSHA256, generating a
15109 pair of RC4 keys.
15110 rtmps
15112 Real-Time Messaging Protocol over a secure SSL connection.
15113 The Real-Time Messaging Protocol (RTMPS) is used for streaming
15115 multimedia content across an encrypted connection.
15116 rtmpt
15118 Real-Time Messaging Protocol tunneled through HTTP.
15119 The Real-Time Messaging Protocol tunneled through HTTP (RTMPT) is used
15121 for streaming multimedia content within HTTP requests to traverse
15122 firewalls.
15123 rtmpte
15125 Encrypted Real-Time Messaging Protocol tunneled through HTTP.
15126 The Encrypted Real-Time Messaging Protocol tunneled through HTTP
15128 (RTMPTE) is used for streaming multimedia content within HTTP requests
15129 to traverse firewalls.
15130 rtmpts
15132 Real-Time Messaging Protocol tunneled through HTTPS.
15133 The Real-Time Messaging Protocol tunneled through HTTPS (RTMPTS) is
15135 used for streaming multimedia content within HTTPS requests to traverse
15136 firewalls.
15137 libsmbclient
15139 libsmbclient permits one to manipulate CIFS/SMB network resources.
15140 Following syntax is required.
15142 smb://[[domain:]user[:password@]]server[/share[/path[/file]]]
15144 This protocol accepts the following options.
15146 timeout
15148 Set timeout in milliseconds of socket I/O operations used by the
15149 underlying low level operation. By default it is set to -1, which
15150 means that the timeout is not specified.
15151 truncate
15153 Truncate existing files on write, if set to 1. A value of 0
15154 prevents truncating. Default value is 1.
15155 workgroup
15157 Set the workgroup used for making connections. By default workgroup
15158 is not specified.
15159 For more information see: <http://www.samba.org/>.
15161 libssh
15163 Secure File Transfer Protocol via libssh
15164 Read from or write to remote resources using SFTP protocol.
15166 Following syntax is required.
15168 sftp://[user[:password]@]server[:port]/path/to/remote/resource.mpeg
15170 This protocol accepts the following options.
15172 timeout
15174 Set timeout of socket I/O operations used by the underlying low
15175 level operation. By default it is set to -1, which means that the
15176 timeout is not specified.
15177 truncate
15179 Truncate existing files on write, if set to 1. A value of 0
15180 prevents truncating. Default value is 1.
15181 private_key
15183 Specify the path of the file containing private key to use during
15184 authorization. By default libssh searches for keys in the ~/.ssh/
15185 directory.
15186 Example: Play a file stored on remote server.
15188 ffplay sftp://user:password@server_address:22/home/user/resource.mpeg
15190 librtmp rtmp, rtmpe, rtmps, rtmpt, rtmpte
15192 Real-Time Messaging Protocol and its variants supported through
15193 librtmp.
15194 Requires the presence of the librtmp headers and library during
15196 configuration. You need to explicitly configure the build with
15197 "--enable-librtmp". If enabled this will replace the native RTMP
15198 protocol.
15199 This protocol provides most client functions and a few server functions
15201 needed to support RTMP, RTMP tunneled in HTTP (RTMPT), encrypted RTMP
15202 (RTMPE), RTMP over SSL/TLS (RTMPS) and tunneled variants of these
15203 encrypted types (RTMPTE, RTMPTS).
15204 The required syntax is:
15206 <rtmp_proto>://<server>[:<port>][/<app>][/<playpath>] <options>
15208 where rtmp_proto is one of the strings "rtmp", "rtmpt", "rtmpe",
15210 "rtmps", "rtmpte", "rtmpts" corresponding to each RTMP variant, and
15211 server, port, app and playpath have the same meaning as specified for
15212 the RTMP native protocol. options contains a list of space-separated
15213 options of the form key=val.
15214 See the librtmp manual page (man 3 librtmp) for more information.
15216 For example, to stream a file in real-time to an RTMP server using
15218 ffmpeg:
15219 ffmpeg -re -i myfile -f flv rtmp://myserver/live/mystream
15221 To play the same stream using ffplay:
15223 ffplay "rtmp://myserver/live/mystream live=1"
15225 rtp
15227 Real-time Transport Protocol.
15228 The required syntax for an RTP URL is:
15230 rtp://hostname[:port][?option=val...]
15231 port specifies the RTP port to use.
15233 The following URL options are supported:
15235 ttl=n
15237 Set the TTL (Time-To-Live) value (for multicast only).
15238 rtcpport=n
15240 Set the remote RTCP port to n.
15241 localrtpport=n
15243 Set the local RTP port to n.
15244 localrtcpport=n'
15246 Set the local RTCP port to n.
15247 pkt_size=n
15249 Set max packet size (in bytes) to n.
15250 buffer_size=size
15252 Set the maximum UDP socket buffer size in bytes.
15253 connect=0|1
15255 Do a connect() on the UDP socket (if set to 1) or not (if set to
15256 0).
15257 sources=ip[,ip]
15259 List allowed source IP addresses.
15260 block=ip[,ip]
15262 List disallowed (blocked) source IP addresses.
15263 write_to_source=0|1
15265 Send packets to the source address of the latest received packet
15266 (if set to 1) or to a default remote address (if set to 0).
15267 localport=n
15269 Set the local RTP port to n.
15270 localaddr=addr
15272 Local IP address of a network interface used for sending packets or
15273 joining multicast groups.
15274 timeout=n
15276 Set timeout (in microseconds) of socket I/O operations to n.
15277 This is a deprecated option. Instead, localrtpport should be used.
15279 Important notes:
15281 1. If rtcpport is not set the RTCP port will be set to the RTP port
15283 value plus 1.
15284 2. If localrtpport (the local RTP port) is not set any available port
15286 will be used for the local RTP and RTCP ports.
15287 3. If localrtcpport (the local RTCP port) is not set it will be set to
15289 the local RTP port value plus 1.
15290 rtsp
15292 Real-Time Streaming Protocol.
15293 RTSP is not technically a protocol handler in libavformat, it is a
15295 demuxer and muxer. The demuxer supports both normal RTSP (with data
15296 transferred over RTP; this is used by e.g. Apple and Microsoft) and
15297 Real-RTSP (with data transferred over RDT).
15298 The muxer can be used to send a stream using RTSP ANNOUNCE to a server
15300 supporting it (currently Darwin Streaming Server and Mischa
15301 Spiegelmock's <https://github.com/revmischa/rtsp-server>).
15302 The required syntax for a RTSP url is:
15304 rtsp://<hostname>[:<port>]/<path>
15306 Options can be set on the ffmpeg/ffplay command line, or set in code
15308 via "AVOption"s or in "avformat_open_input".
15309 Muxer
15311 The following options are supported.
15313 rtsp_transport
15315 Set RTSP transport protocols.
15316 It accepts the following values:
15318 udp Use UDP as lower transport protocol.
15320 tcp Use TCP (interleaving within the RTSP control channel) as lower
15322 transport protocol.
15323 Default value is 0.
15325 rtsp_flags
15327 Set RTSP flags.
15328 The following values are accepted:
15330 latm
15332 Use MP4A-LATM packetization instead of MPEG4-GENERIC for AAC.
15333 rfc2190
15335 Use RFC 2190 packetization instead of RFC 4629 for H.263.
15336 skip_rtcp
15338 Don't send RTCP sender reports.
15339 h264_mode0
15341 Use mode 0 for H.264 in RTP.
15342 send_bye
15344 Send RTCP BYE packets when finishing.
15345 Default value is 0.
15347 min_port
15349 Set minimum local UDP port. Default value is 5000.
15350 max_port
15352 Set maximum local UDP port. Default value is 65000.
15353 buffer_size
15355 Set the maximum socket buffer size in bytes.
15356 pkt_size
15358 Set max send packet size (in bytes). Default value is 1472.
15359 Demuxer
15361 The following options are supported.
15363 initial_pause
15365 Do not start playing the stream immediately if set to 1. Default
15366 value is 0.
15367 rtsp_transport
15369 Set RTSP transport protocols.
15370 It accepts the following values:
15372 udp Use UDP as lower transport protocol.
15374 tcp Use TCP (interleaving within the RTSP control channel) as lower
15376 transport protocol.
15377 udp_multicast
15379 Use UDP multicast as lower transport protocol.
15380 http
15382 Use HTTP tunneling as lower transport protocol, which is useful
15383 for passing proxies.
15384 https
15386 Use HTTPs tunneling as lower transport protocol, which is
15387 useful for passing proxies and widely used for security
15388 consideration.
15389 Multiple lower transport protocols may be specified, in that case
15391 they are tried one at a time (if the setup of one fails, the next
15392 one is tried). For the muxer, only the tcp and udp options are
15393 supported.
15394 rtsp_flags
15396 Set RTSP flags.
15397 The following values are accepted:
15399 filter_src
15401 Accept packets only from negotiated peer address and port.
15402 listen
15404 Act as a server, listening for an incoming connection.
15405 prefer_tcp
15407 Try TCP for RTP transport first, if TCP is available as RTSP
15408 RTP transport.
15409 satip_raw
15411 Export raw MPEG-TS stream instead of demuxing. The flag will
15412 simply write out the raw stream, with the original PAT/PMT/PIDs
15413 intact.
15414 Default value is none.
15416 allowed_media_types
15418 Set media types to accept from the server.
15419 The following flags are accepted:
15421 video
15423 audio
15424 data
15425 subtitle
15426 By default it accepts all media types.
15428 min_port
15430 Set minimum local UDP port. Default value is 5000.
15431 max_port
15433 Set maximum local UDP port. Default value is 65000.
15434 listen_timeout
15436 Set maximum timeout (in seconds) to establish an initial
15437 connection. Setting listen_timeout > 0 sets rtsp_flags to listen.
15438 Default is -1 which means an infinite timeout when listen mode is
15439 set.
15440 reorder_queue_size
15442 Set number of packets to buffer for handling of reordered packets.
15443 timeout
15445 Set socket TCP I/O timeout in microseconds.
15446 user_agent
15448 Override User-Agent header. If not specified, it defaults to the
15449 libavformat identifier string.
15450 buffer_size
15452 Set the maximum socket buffer size in bytes.
15453 When receiving data over UDP, the demuxer tries to reorder received
15455 packets (since they may arrive out of order, or packets may get lost
15456 totally). This can be disabled by setting the maximum demuxing delay to
15457 zero (via the "max_delay" field of AVFormatContext).
15458 When watching multi-bitrate Real-RTSP streams with ffplay, the streams
15460 to display can be chosen with "-vst" n and "-ast" n for video and audio
15461 respectively, and can be switched on the fly by pressing "v" and "a".
15462 Examples
15464 The following examples all make use of the ffplay and ffmpeg tools.
15466 • Watch a stream over UDP, with a max reordering delay of 0.5
15468 seconds:
15469 ffplay -max_delay 500000 -rtsp_transport udp rtsp://server/video.mp4
15471 • Watch a stream tunneled over HTTP:
15473 ffplay -rtsp_transport http rtsp://server/video.mp4
15475 • Send a stream in realtime to a RTSP server, for others to watch:
15477 ffmpeg -re -i <input> -f rtsp -muxdelay 0.1 rtsp://server/live.sdp
15479 • Receive a stream in realtime:
15481 ffmpeg -rtsp_flags listen -i rtsp://ownaddress/live.sdp <output>
15483 sap
15485 Session Announcement Protocol (RFC 2974). This is not technically a
15486 protocol handler in libavformat, it is a muxer and demuxer. It is used
15487 for signalling of RTP streams, by announcing the SDP for the streams
15488 regularly on a separate port.
15489 Muxer
15491 The syntax for a SAP url given to the muxer is:
15493 sap://<destination>[:<port>][?<options>]
15495 The RTP packets are sent to destination on port port, or to port 5004
15497 if no port is specified. options is a "&"-separated list. The
15498 following options are supported:
15499 announce_addr=address
15501 Specify the destination IP address for sending the announcements
15502 to. If omitted, the announcements are sent to the commonly used
15503 SAP announcement multicast address 224.2.127.254 (sap.mcast.net),
15504 or ff0e::2:7ffe if destination is an IPv6 address.
15505 announce_port=port
15507 Specify the port to send the announcements on, defaults to 9875 if
15508 not specified.
15509 ttl=ttl
15511 Specify the time to live value for the announcements and RTP
15512 packets, defaults to 255.
15513 same_port=0|1
15515 If set to 1, send all RTP streams on the same port pair. If zero
15516 (the default), all streams are sent on unique ports, with each
15517 stream on a port 2 numbers higher than the previous. VLC/Live555
15518 requires this to be set to 1, to be able to receive the stream.
15519 The RTP stack in libavformat for receiving requires all streams to
15520 be sent on unique ports.
15521 Example command lines follow.
15523 To broadcast a stream on the local subnet, for watching in VLC:
15525 ffmpeg -re -i <input> -f sap sap://224.0.0.255?same_port=1
15527 Similarly, for watching in ffplay:
15529 ffmpeg -re -i <input> -f sap sap://224.0.0.255
15531 And for watching in ffplay, over IPv6:
15533 ffmpeg -re -i <input> -f sap sap://[ff0e::1:2:3:4]
15535 Demuxer
15537 The syntax for a SAP url given to the demuxer is:
15539 sap://[<address>][:<port>]
15541 address is the multicast address to listen for announcements on, if
15543 omitted, the default 224.2.127.254 (sap.mcast.net) is used. port is the
15544 port that is listened on, 9875 if omitted.
15545 The demuxers listens for announcements on the given address and port.
15547 Once an announcement is received, it tries to receive that particular
15548 stream.
15549 Example command lines follow.
15551 To play back the first stream announced on the normal SAP multicast
15553 address:
15554 ffplay sap://
15556 To play back the first stream announced on one the default IPv6 SAP
15558 multicast address:
15559 ffplay sap://[ff0e::2:7ffe]
15561 sctp
15563 Stream Control Transmission Protocol.
15564 The accepted URL syntax is:
15566 sctp://<host>:<port>[?<options>]
15568 The protocol accepts the following options:
15570 listen
15572 If set to any value, listen for an incoming connection. Outgoing
15573 connection is done by default.
15574 max_streams
15576 Set the maximum number of streams. By default no limit is set.
15577 srt
15579 Haivision Secure Reliable Transport Protocol via libsrt.
15580 The supported syntax for a SRT URL is:
15582 srt://<hostname>:<port>[?<options>]
15584 options contains a list of &-separated options of the form key=val.
15586 or
15588 <options> srt://<hostname>:<port>
15590 options contains a list of '-key val' options.
15592 This protocol accepts the following options.
15594 connect_timeout=milliseconds
15596 Connection timeout; SRT cannot connect for RTT > 1500 msec (2
15597 handshake exchanges) with the default connect timeout of 3 seconds.
15598 This option applies to the caller and rendezvous connection modes.
15599 The connect timeout is 10 times the value set for the rendezvous
15600 mode (which can be used as a workaround for this connection problem
15601 with earlier versions).
15602 ffs=bytes
15604 Flight Flag Size (Window Size), in bytes. FFS is actually an
15605 internal parameter and you should set it to not less than
15606 recv_buffer_size and mss. The default value is relatively large,
15607 therefore unless you set a very large receiver buffer, you do not
15608 need to change this option. Default value is 25600.
15609 inputbw=bytes/seconds
15611 Sender nominal input rate, in bytes per seconds. Used along with
15612 oheadbw, when maxbw is set to relative (0), to calculate maximum
15613 sending rate when recovery packets are sent along with the main
15614 media stream: inputbw * (100 + oheadbw) / 100 if inputbw is not set
15615 while maxbw is set to relative (0), the actual input rate is
15616 evaluated inside the library. Default value is 0.
15617 iptos=tos
15619 IP Type of Service. Applies to sender only. Default value is 0xB8.
15620 ipttl=ttl
15622 IP Time To Live. Applies to sender only. Default value is 64.
15623 latency=microseconds
15625 Timestamp-based Packet Delivery Delay. Used to absorb bursts of
15626 missed packet retransmissions. This flag sets both rcvlatency and
15627 peerlatency to the same value. Note that prior to version 1.3.0
15628 this is the only flag to set the latency, however this is
15629 effectively equivalent to setting peerlatency, when side is sender
15630 and rcvlatency when side is receiver, and the bidirectional stream
15631 sending is not supported.
15632 listen_timeout=microseconds
15634 Set socket listen timeout.
15635 maxbw=bytes/seconds
15637 Maximum sending bandwidth, in bytes per seconds. -1 infinite
15638 (CSRTCC limit is 30mbps) 0 relative to input rate (see inputbw) >0
15639 absolute limit value Default value is 0 (relative)
15640 mode=caller|listener|rendezvous
15642 Connection mode. caller opens client connection. listener starts
15643 server to listen for incoming connections. rendezvous use Rendez-
15644 Vous connection mode. Default value is caller.
15645 mss=bytes
15647 Maximum Segment Size, in bytes. Used for buffer allocation and rate
15648 calculation using a packet counter assuming fully filled packets.
15649 The smallest MSS between the peers is used. This is 1500 by default
15650 in the overall internet. This is the maximum size of the UDP
15651 packet and can be only decreased, unless you have some unusual
15652 dedicated network settings. Default value is 1500.
15653 nakreport=1|0
15655 If set to 1, Receiver will send `UMSG_LOSSREPORT` messages
15656 periodically until a lost packet is retransmitted or intentionally
15657 dropped. Default value is 1.
15658 oheadbw=percents
15660 Recovery bandwidth overhead above input rate, in percents. See
15661 inputbw. Default value is 25%.
15662 passphrase=string
15664 HaiCrypt Encryption/Decryption Passphrase string, length from 10 to
15665 79 characters. The passphrase is the shared secret between the
15666 sender and the receiver. It is used to generate the Key Encrypting
15667 Key using PBKDF2 (Password-Based Key Derivation Function). It is
15668 used only if pbkeylen is non-zero. It is used on the receiver only
15669 if the received data is encrypted. The configured passphrase
15670 cannot be recovered (write-only).
15671 enforced_encryption=1|0
15673 If true, both connection parties must have the same password set
15674 (including empty, that is, with no encryption). If the password
15675 doesn't match or only one side is unencrypted, the connection is
15676 rejected. Default is true.
15677 kmrefreshrate=packets
15679 The number of packets to be transmitted after which the encryption
15680 key is switched to a new key. Default is -1. -1 means auto
15681 (0x1000000 in srt library). The range for this option is integers
15682 in the 0 - "INT_MAX".
15683 kmpreannounce=packets
15685 The interval between when a new encryption key is sent and when
15686 switchover occurs. This value also applies to the subsequent
15687 interval between when switchover occurs and when the old encryption
15688 key is decommissioned. Default is -1. -1 means auto (0x1000 in srt
15689 library). The range for this option is integers in the 0 -
15690 "INT_MAX".
15691 snddropdelay=microseconds
15693 The sender's extra delay before dropping packets. This delay is
15694 added to the default drop delay time interval value.
15695 Special value -1: Do not drop packets on the sender at all.
15697 payload_size=bytes
15699 Sets the maximum declared size of a packet transferred during the
15700 single call to the sending function in Live mode. Use 0 if this
15701 value isn't used (which is default in file mode). Default is -1
15702 (automatic), which typically means MPEG-TS; if you are going to use
15703 SRT to send any different kind of payload, such as, for example,
15704 wrapping a live stream in very small frames, then you can use a
15705 bigger maximum frame size, though not greater than 1456 bytes.
15706 pkt_size=bytes
15708 Alias for payload_size.
15709 peerlatency=microseconds
15711 The latency value (as described in rcvlatency) that is set by the
15712 sender side as a minimum value for the receiver.
15713 pbkeylen=bytes
15715 Sender encryption key length, in bytes. Only can be set to 0, 16,
15716 24 and 32. Enable sender encryption if not 0. Not required on
15717 receiver (set to 0), key size obtained from sender in HaiCrypt
15718 handshake. Default value is 0.
15719 rcvlatency=microseconds
15721 The time that should elapse since the moment when the packet was
15722 sent and the moment when it's delivered to the receiver application
15723 in the receiving function. This time should be a buffer time large
15724 enough to cover the time spent for sending, unexpectedly extended
15725 RTT time, and the time needed to retransmit the lost UDP packet.
15726 The effective latency value will be the maximum of this options'
15727 value and the value of peerlatency set by the peer side. Before
15728 version 1.3.0 this option is only available as latency.
15729 recv_buffer_size=bytes
15731 Set UDP receive buffer size, expressed in bytes.
15732 send_buffer_size=bytes
15734 Set UDP send buffer size, expressed in bytes.
15735 timeout=microseconds
15737 Set raise error timeouts for read, write and connect operations.
15738 Note that the SRT library has internal timeouts which can be
15739 controlled separately, the value set here is only a cap on those.
15740 tlpktdrop=1|0
15742 Too-late Packet Drop. When enabled on receiver, it skips missing
15743 packets that have not been delivered in time and delivers the
15744 following packets to the application when their time-to-play has
15745 come. It also sends a fake ACK to the sender. When enabled on
15746 sender and enabled on the receiving peer, the sender drops the
15747 older packets that have no chance of being delivered in time. It
15748 was automatically enabled in the sender if the receiver supports
15749 it.
15750 sndbuf=bytes
15752 Set send buffer size, expressed in bytes.
15753 rcvbuf=bytes
15755 Set receive buffer size, expressed in bytes.
15756 Receive buffer must not be greater than ffs.
15758 lossmaxttl=packets
15760 The value up to which the Reorder Tolerance may grow. When Reorder
15761 Tolerance is > 0, then packet loss report is delayed until that
15762 number of packets come in. Reorder Tolerance increases every time a
15763 "belated" packet has come, but it wasn't due to retransmission
15764 (that is, when UDP packets tend to come out of order), with the
15765 difference between the latest sequence and this packet's sequence,
15766 and not more than the value of this option. By default it's 0,
15767 which means that this mechanism is turned off, and the loss report
15768 is always sent immediately upon experiencing a "gap" in sequences.
15769 minversion
15771 The minimum SRT version that is required from the peer. A
15772 connection to a peer that does not satisfy the minimum version
15773 requirement will be rejected.
15774 The version format in hex is 0xXXYYZZ for x.y.z in human readable
15776 form.
15777 streamid=string
15779 A string limited to 512 characters that can be set on the socket
15780 prior to connecting. This stream ID will be able to be retrieved by
15781 the listener side from the socket that is returned from srt_accept
15782 and was connected by a socket with that set stream ID. SRT does not
15783 enforce any special interpretation of the contents of this string.
15784 This option doesn’t make sense in Rendezvous connection; the result
15785 might be that simply one side will override the value from the
15786 other side and it’s the matter of luck which one would win
15787 srt_streamid=string
15789 Alias for streamid to avoid conflict with ffmpeg command line
15790 option.
15791 smoother=live|file
15793 The type of Smoother used for the transmission for that socket,
15794 which is responsible for the transmission and congestion control.
15795 The Smoother type must be exactly the same on both connecting
15796 parties, otherwise the connection is rejected.
15797 messageapi=1|0
15799 When set, this socket uses the Message API, otherwise it uses
15800 Buffer API. Note that in live mode (see transtype) there’s only
15801 message API available. In File mode you can chose to use one of two
15802 modes:
15803 Stream API (default, when this option is false). In this mode you
15805 may send as many data as you wish with one sending instruction, or
15806 even use dedicated functions that read directly from a file. The
15807 internal facility will take care of any speed and congestion
15808 control. When receiving, you can also receive as many data as
15809 desired, the data not extracted will be waiting for the next call.
15810 There is no boundary between data portions in the Stream mode.
15811 Message API. In this mode your single sending instruction passes
15813 exactly one piece of data that has boundaries (a message). Contrary
15814 to Live mode, this message may span across multiple UDP packets and
15815 the only size limitation is that it shall fit as a whole in the
15816 sending buffer. The receiver shall use as large buffer as necessary
15817 to receive the message, otherwise the message will not be given up.
15818 When the message is not complete (not all packets received or there
15819 was a packet loss) it will not be given up.
15820 transtype=live|file
15822 Sets the transmission type for the socket, in particular, setting
15823 this option sets multiple other parameters to their default values
15824 as required for a particular transmission type.
15825 live: Set options as for live transmission. In this mode, you
15827 should send by one sending instruction only so many data that fit
15828 in one UDP packet, and limited to the value defined first in
15829 payload_size (1316 is default in this mode). There is no speed
15830 control in this mode, only the bandwidth control, if configured, in
15831 order to not exceed the bandwidth with the overhead transmission
15832 (retransmitted and control packets).
15833 file: Set options as for non-live transmission. See messageapi for
15835 further explanations
15836 linger=seconds
15838 The number of seconds that the socket waits for unsent data when
15839 closing. Default is -1. -1 means auto (off with 0 seconds in live
15840 mode, on with 180 seconds in file mode). The range for this option
15841 is integers in the 0 - "INT_MAX".
15842 tsbpd=1|0
15844 When true, use Timestamp-based Packet Delivery mode. The default
15845 behavior depends on the transmission type: enabled in live mode,
15846 disabled in file mode.
15847 For more information see: <https://github.com/Haivision/srt>.
15849 srtp
15851 Secure Real-time Transport Protocol.
15852 The accepted options are:
15854 srtp_in_suite
15856 srtp_out_suite
15857 Select input and output encoding suites.
15858 Supported values:
15860 AES_CM_128_HMAC_SHA1_80
15862 SRTP_AES128_CM_HMAC_SHA1_80
15863 AES_CM_128_HMAC_SHA1_32
15864 SRTP_AES128_CM_HMAC_SHA1_32
15865 srtp_in_params
15866 srtp_out_params
15867 Set input and output encoding parameters, which are expressed by a
15868 base64-encoded representation of a binary block. The first 16 bytes
15869 of this binary block are used as master key, the following 14 bytes
15870 are used as master salt.
15871 subfile
15873 Virtually extract a segment of a file or another stream. The
15874 underlying stream must be seekable.
15875 Accepted options:
15877 start
15879 Start offset of the extracted segment, in bytes.
15880 end End offset of the extracted segment, in bytes. If set to 0,
15882 extract till end of file.
15883 Examples:
15885 Extract a chapter from a DVD VOB file (start and end sectors obtained
15887 externally and multiplied by 2048):
15888 subfile,,start,153391104,end,268142592,,:/media/dvd/VIDEO_TS/VTS_08_1.VOB
15890 Play an AVI file directly from a TAR archive:
15892 subfile,,start,183241728,end,366490624,,:archive.tar
15894 Play a MPEG-TS file from start offset till end:
15896 subfile,,start,32815239,end,0,,:video.ts
15898 tee
15900 Writes the output to multiple protocols. The individual outputs are
15901 separated by |
15902 tee:file://path/to/local/this.avi|file://path/to/local/that.avi
15904 tcp
15906 Transmission Control Protocol.
15907 The required syntax for a TCP url is:
15909 tcp://<hostname>:<port>[?<options>]
15911 options contains a list of &-separated options of the form key=val.
15913 The list of supported options follows.
15915 listen=2|1|0
15917 Listen for an incoming connection. 0 disables listen, 1 enables
15918 listen in single client mode, 2 enables listen in multi-client
15919 mode. Default value is 0.
15920 timeout=microseconds
15922 Set raise error timeout, expressed in microseconds.
15923 This option is only relevant in read mode: if no data arrived in
15925 more than this time interval, raise error.
15926 listen_timeout=milliseconds
15928 Set listen timeout, expressed in milliseconds.
15929 recv_buffer_size=bytes
15931 Set receive buffer size, expressed bytes.
15932 send_buffer_size=bytes
15934 Set send buffer size, expressed bytes.
15935 tcp_nodelay=1|0
15937 Set TCP_NODELAY to disable Nagle's algorithm. Default value is 0.
15938 Remark: Writing to the socket is currently not optimized to
15940 minimize system calls and reduces the efficiency / effect of
15941 TCP_NODELAY.
15942 tcp_mss=bytes
15944 Set maximum segment size for outgoing TCP packets, expressed in
15945 bytes.
15946 The following example shows how to setup a listening TCP connection
15948 with ffmpeg, which is then accessed with ffplay:
15949 ffmpeg -i <input> -f <format> tcp://<hostname>:<port>?listen
15951 ffplay tcp://<hostname>:<port>
15952 tls
15954 Transport Layer Security (TLS) / Secure Sockets Layer (SSL)
15955 The required syntax for a TLS/SSL url is:
15957 tls://<hostname>:<port>[?<options>]
15959 The following parameters can be set via command line options (or in
15961 code via "AVOption"s):
15962 ca_file, cafile=filename
15964 A file containing certificate authority (CA) root certificates to
15965 treat as trusted. If the linked TLS library contains a default this
15966 might not need to be specified for verification to work, but not
15967 all libraries and setups have defaults built in. The file must be
15968 in OpenSSL PEM format.
15969 tls_verify=1|0
15971 If enabled, try to verify the peer that we are communicating with.
15972 Note, if using OpenSSL, this currently only makes sure that the
15973 peer certificate is signed by one of the root certificates in the
15974 CA database, but it does not validate that the certificate actually
15975 matches the host name we are trying to connect to. (With other
15976 backends, the host name is validated as well.)
15977 This is disabled by default since it requires a CA database to be
15979 provided by the caller in many cases.
15980 cert_file, cert=filename
15982 A file containing a certificate to use in the handshake with the
15983 peer. (When operating as server, in listen mode, this is more
15984 often required by the peer, while client certificates only are
15985 mandated in certain setups.)
15986 key_file, key=filename
15988 A file containing the private key for the certificate.
15989 listen=1|0
15991 If enabled, listen for connections on the provided port, and assume
15992 the server role in the handshake instead of the client role.
15993 http_proxy
15995 The HTTP proxy to tunnel through, e.g. "http://example.com:1234".
15996 The proxy must support the CONNECT method.
15997 Example command lines:
15999 To create a TLS/SSL server that serves an input stream.
16001 ffmpeg -i <input> -f <format> tls://<hostname>:<port>?listen&cert=<server.crt>&key=<server.key>
16003 To play back a stream from the TLS/SSL server using ffplay:
16005 ffplay tls://<hostname>:<port>
16007 udp
16009 User Datagram Protocol.
16010 The required syntax for an UDP URL is:
16012 udp://<hostname>:<port>[?<options>]
16014 options contains a list of &-separated options of the form key=val.
16016 In case threading is enabled on the system, a circular buffer is used
16018 to store the incoming data, which allows one to reduce loss of data due
16019 to UDP socket buffer overruns. The fifo_size and overrun_nonfatal
16020 options are related to this buffer.
16021 The list of supported options follows.
16023 buffer_size=size
16025 Set the UDP maximum socket buffer size in bytes. This is used to
16026 set either the receive or send buffer size, depending on what the
16027 socket is used for. Default is 32 KB for output, 384 KB for input.
16028 See also fifo_size.
16029 bitrate=bitrate
16031 If set to nonzero, the output will have the specified constant
16032 bitrate if the input has enough packets to sustain it.
16033 burst_bits=bits
16035 When using bitrate this specifies the maximum number of bits in
16036 packet bursts.
16037 localport=port
16039 Override the local UDP port to bind with.
16040 localaddr=addr
16042 Local IP address of a network interface used for sending packets or
16043 joining multicast groups.
16044 pkt_size=size
16046 Set the size in bytes of UDP packets.
16047 reuse=1|0
16049 Explicitly allow or disallow reusing UDP sockets.
16050 ttl=ttl
16052 Set the time to live value (for multicast only).
16053 connect=1|0
16055 Initialize the UDP socket with connect(). In this case, the
16056 destination address can't be changed with ff_udp_set_remote_url
16057 later. If the destination address isn't known at the start, this
16058 option can be specified in ff_udp_set_remote_url, too. This allows
16059 finding out the source address for the packets with getsockname,
16060 and makes writes return with AVERROR(ECONNREFUSED) if "destination
16061 unreachable" is received. For receiving, this gives the benefit of
16062 only receiving packets from the specified peer address/port.
16063 sources=address[,address]
16065 Only receive packets sent from the specified addresses. In case of
16066 multicast, also subscribe to multicast traffic coming from these
16067 addresses only.
16068 block=address[,address]
16070 Ignore packets sent from the specified addresses. In case of
16071 multicast, also exclude the source addresses in the multicast
16072 subscription.
16073 fifo_size=units
16075 Set the UDP receiving circular buffer size, expressed as a number
16076 of packets with size of 188 bytes. If not specified defaults to
16077 7*4096.
16078 overrun_nonfatal=1|0
16080 Survive in case of UDP receiving circular buffer overrun. Default
16081 value is 0.
16082 timeout=microseconds
16084 Set raise error timeout, expressed in microseconds.
16085 This option is only relevant in read mode: if no data arrived in
16087 more than this time interval, raise error.
16088 broadcast=1|0
16090 Explicitly allow or disallow UDP broadcasting.
16091 Note that broadcasting may not work properly on networks having a
16093 broadcast storm protection.
16094 Examples
16096 • Use ffmpeg to stream over UDP to a remote endpoint:
16098 ffmpeg -i <input> -f <format> udp://<hostname>:<port>
16100 • Use ffmpeg to stream in mpegts format over UDP using 188 sized UDP
16102 packets, using a large input buffer:
16103 ffmpeg -i <input> -f mpegts udp://<hostname>:<port>?pkt_size=188&buffer_size=65535
16105 • Use ffmpeg to receive over UDP from a remote endpoint:
16107 ffmpeg -i udp://[<multicast-address>]:<port> ...
16109 unix
16111 Unix local socket
16112 The required syntax for a Unix socket URL is:
16114 unix://<filepath>
16116 The following parameters can be set via command line options (or in
16118 code via "AVOption"s):
16119 timeout
16121 Timeout in ms.
16122 listen
16124 Create the Unix socket in listening mode.
16125 zmq
16127 ZeroMQ asynchronous messaging using the libzmq library.
16128 This library supports unicast streaming to multiple clients without
16130 relying on an external server.
16131 The required syntax for streaming or connecting to a stream is:
16133 zmq:tcp://ip-address:port
16135 Example: Create a localhost stream on port 5555:
16137 ffmpeg -re -i input -f mpegts zmq:tcp://127.0.0.1:5555
16139 Multiple clients may connect to the stream using:
16141 ffplay zmq:tcp://127.0.0.1:5555
16143 Streaming to multiple clients is implemented using a ZeroMQ Pub-Sub
16145 pattern. The server side binds to a port and publishes data. Clients
16146 connect to the server (via IP address/port) and subscribe to the
16147 stream. The order in which the server and client start generally does
16148 not matter.
16149 ffmpeg must be compiled with the --enable-libzmq option to support this
16151 protocol.
16152 Options can be set on the ffmpeg/ffplay command line. The following
16154 options are supported:
16155 pkt_size
16157 Forces the maximum packet size for sending/receiving data. The
16158 default value is 131,072 bytes. On the server side, this sets the
16159 maximum size of sent packets via ZeroMQ. On the clients, it sets an
16160 internal buffer size for receiving packets. Note that pkt_size on
16161 the clients should be equal to or greater than pkt_size on the
16162 server. Otherwise the received message may be truncated causing
16163 decoding errors.
16164
16166 The libavdevice library provides the same interface as libavformat.
16167 Namely, an input device is considered like a demuxer, and an output
16168 device like a muxer, and the interface and generic device options are
16169 the same provided by libavformat (see the ffmpeg-formats manual).
16170 In addition each input or output device may support so-called private
16172 options, which are specific for that component.
16173 Options may be set by specifying -option value in the FFmpeg tools, or
16175 by setting the value explicitly in the device "AVFormatContext" options
16176 or using the libavutil/opt.h API for programmatic use.
16177
16179 Input devices are configured elements in FFmpeg which enable accessing
16180 the data coming from a multimedia device attached to your system.
16181 When you configure your FFmpeg build, all the supported input devices
16183 are enabled by default. You can list all available ones using the
16184 configure option "--list-indevs".
16185 You can disable all the input devices using the configure option
16187 "--disable-indevs", and selectively enable an input device using the
16188 option "--enable-indev=INDEV", or you can disable a particular input
16189 device using the option "--disable-indev=INDEV".
16190 The option "-devices" of the ff* tools will display the list of
16192 supported input devices.
16193 A description of the currently available input devices follows.
16195 alsa
16197 ALSA (Advanced Linux Sound Architecture) input device.
16198 To enable this input device during configuration you need libasound
16200 installed on your system.
16201 This device allows capturing from an ALSA device. The name of the
16203 device to capture has to be an ALSA card identifier.
16204 An ALSA identifier has the syntax:
16206 hw:<CARD>[,<DEV>[,<SUBDEV>]]
16208 where the DEV and SUBDEV components are optional.
16210 The three arguments (in order: CARD,DEV,SUBDEV) specify card number or
16212 identifier, device number and subdevice number (-1 means any).
16213 To see the list of cards currently recognized by your system check the
16215 files /proc/asound/cards and /proc/asound/devices.
16216 For example to capture with ffmpeg from an ALSA device with card id 0,
16218 you may run the command:
16219 ffmpeg -f alsa -i hw:0 alsaout.wav
16221 For more information see:
16223 <http://www.alsa-project.org/alsa-doc/alsa-lib/pcm.html>
16224 Options
16226 sample_rate
16228 Set the sample rate in Hz. Default is 48000.
16229 channels
16231 Set the number of channels. Default is 2.
16232 android_camera
16234 Android camera input device.
16235 This input devices uses the Android Camera2 NDK API which is available
16237 on devices with API level 24+. The availability of android_camera is
16238 autodetected during configuration.
16239 This device allows capturing from all cameras on an Android device,
16241 which are integrated into the Camera2 NDK API.
16242 The available cameras are enumerated internally and can be selected
16244 with the camera_index parameter. The input file string is discarded.
16245 Generally the back facing camera has index 0 while the front facing
16247 camera has index 1.
16248 Options
16250 video_size
16252 Set the video size given as a string such as 640x480 or hd720.
16253 Falls back to the first available configuration reported by Android
16254 if requested video size is not available or by default.
16255 framerate
16257 Set the video framerate. Falls back to the first available
16258 configuration reported by Android if requested framerate is not
16259 available or by default (-1).
16260 camera_index
16262 Set the index of the camera to use. Default is 0.
16263 input_queue_size
16265 Set the maximum number of frames to buffer. Default is 5.
16266 avfoundation
16268 AVFoundation input device.
16269 AVFoundation is the currently recommended framework by Apple for
16271 streamgrabbing on OSX >= 10.7 as well as on iOS.
16272 The input filename has to be given in the following syntax:
16274 -i "[[VIDEO]:[AUDIO]]"
16276 The first entry selects the video input while the latter selects the
16278 audio input. The stream has to be specified by the device name or the
16279 device index as shown by the device list. Alternatively, the video
16280 and/or audio input device can be chosen by index using the
16281 B<-video_device_index E<lt>INDEXE<gt>>
16283 and/or
16285 B<-audio_device_index E<lt>INDEXE<gt>>
16287 , overriding any device name or index given in the input filename.
16289 All available devices can be enumerated by using -list_devices true,
16291 listing all device names and corresponding indices.
16292 There are two device name aliases:
16294 "default"
16296 Select the AVFoundation default device of the corresponding type.
16297 "none"
16299 Do not record the corresponding media type. This is equivalent to
16300 specifying an empty device name or index.
16301 Options
16303 AVFoundation supports the following options:
16305 -list_devices <TRUE|FALSE>
16307 If set to true, a list of all available input devices is given
16308 showing all device names and indices.
16309 -video_device_index <INDEX>
16311 Specify the video device by its index. Overrides anything given in
16312 the input filename.
16313 -audio_device_index <INDEX>
16315 Specify the audio device by its index. Overrides anything given in
16316 the input filename.
16317 -pixel_format <FORMAT>
16319 Request the video device to use a specific pixel format. If the
16320 specified format is not supported, a list of available formats is
16321 given and the first one in this list is used instead. Available
16322 pixel formats are: "monob, rgb555be, rgb555le, rgb565be, rgb565le,
16323 rgb24, bgr24, 0rgb, bgr0, 0bgr, rgb0,
16324 bgr48be, uyvy422, yuva444p, yuva444p16le, yuv444p, yuv422p16,
16325 yuv422p10, yuv444p10,
16326 yuv420p, nv12, yuyv422, gray"
16327 -framerate
16329 Set the grabbing frame rate. Default is "ntsc", corresponding to a
16330 frame rate of "30000/1001".
16331 -video_size
16333 Set the video frame size.
16334 -capture_cursor
16336 Capture the mouse pointer. Default is 0.
16337 -capture_mouse_clicks
16339 Capture the screen mouse clicks. Default is 0.
16340 -capture_raw_data
16342 Capture the raw device data. Default is 0. Using this option may
16343 result in receiving the underlying data delivered to the
16344 AVFoundation framework. E.g. for muxed devices that sends raw DV
16345 data to the framework (like tape-based camcorders), setting this
16346 option to false results in extracted video frames captured in the
16347 designated pixel format only. Setting this option to true results
16348 in receiving the raw DV stream untouched.
16349 Examples
16351 • Print the list of AVFoundation supported devices and exit:
16353 $ ffmpeg -f avfoundation -list_devices true -i ""
16355 • Record video from video device 0 and audio from audio device 0 into
16357 out.avi:
16358 $ ffmpeg -f avfoundation -i "0:0" out.avi
16360 • Record video from video device 2 and audio from audio device 1 into
16362 out.avi:
16363 $ ffmpeg -f avfoundation -video_device_index 2 -i ":1" out.avi
16365 • Record video from the system default video device using the pixel
16367 format bgr0 and do not record any audio into out.avi:
16368 $ ffmpeg -f avfoundation -pixel_format bgr0 -i "default:none" out.avi
16370 • Record raw DV data from a suitable input device and write the
16372 output into out.dv:
16373 $ ffmpeg -f avfoundation -capture_raw_data true -i "zr100:none" out.dv
16375 bktr
16377 BSD video input device.
16378 Options
16380 framerate
16382 Set the frame rate.
16383 video_size
16385 Set the video frame size. Default is "vga".
16386 standard
16388 Available values are:
16389 pal
16391 ntsc
16392 secam
16393 paln
16394 palm
16395 ntscj
16396 decklink
16398 The decklink input device provides capture capabilities for Blackmagic
16399 DeckLink devices.
16400 To enable this input device, you need the Blackmagic DeckLink SDK and
16402 you need to configure with the appropriate "--extra-cflags" and
16403 "--extra-ldflags". On Windows, you need to run the IDL files through
16404 widl.
16405 DeckLink is very picky about the formats it supports. Pixel format of
16407 the input can be set with raw_format. Framerate and video size must be
16408 determined for your device with -list_formats 1. Audio sample rate is
16409 always 48 kHz and the number of channels can be 2, 8 or 16. Note that
16410 all audio channels are bundled in one single audio track.
16411 Options
16413 list_devices
16415 If set to true, print a list of devices and exit. Defaults to
16416 false. This option is deprecated, please use the "-sources" option
16417 of ffmpeg to list the available input devices.
16418 list_formats
16420 If set to true, print a list of supported formats and exit.
16421 Defaults to false.
16422 format_code <FourCC>
16424 This sets the input video format to the format given by the FourCC.
16425 To see the supported values of your device(s) use list_formats.
16426 Note that there is a FourCC 'pal ' that can also be used as pal (3
16427 letters). Default behavior is autodetection of the input video
16428 format, if the hardware supports it.
16429 raw_format
16431 Set the pixel format of the captured video. Available values are:
16432 auto
16434 This is the default which means 8-bit YUV 422 or 8-bit ARGB if
16435 format autodetection is used, 8-bit YUV 422 otherwise.
16436 uyvy422
16438 8-bit YUV 422.
16439 yuv422p10
16441 10-bit YUV 422.
16442 argb
16444 8-bit RGB.
16445 bgra
16447 8-bit RGB.
16448 rgb10
16450 10-bit RGB.
16451 teletext_lines
16453 If set to nonzero, an additional teletext stream will be captured
16454 from the vertical ancillary data. Both SD PAL (576i) and HD (1080i
16455 or 1080p) sources are supported. In case of HD sources, OP47
16456 packets are decoded.
16457 This option is a bitmask of the SD PAL VBI lines captured,
16459 specifically lines 6 to 22, and lines 318 to 335. Line 6 is the LSB
16460 in the mask. Selected lines which do not contain teletext
16461 information will be ignored. You can use the special all constant
16462 to select all possible lines, or standard to skip lines 6, 318 and
16463 319, which are not compatible with all receivers.
16464 For SD sources, ffmpeg needs to be compiled with
16466 "--enable-libzvbi". For HD sources, on older (pre-4K) DeckLink card
16467 models you have to capture in 10 bit mode.
16468 channels
16470 Defines number of audio channels to capture. Must be 2, 8 or 16.
16471 Defaults to 2.
16472 duplex_mode
16474 Sets the decklink device duplex/profile mode. Must be unset, half,
16475 full, one_sub_device_full, one_sub_device_half,
16476 two_sub_device_full, four_sub_device_half Defaults to unset.
16477 Note: DeckLink SDK 11.0 have replaced the duplex property by a
16479 profile property. For the DeckLink Duo 2 and DeckLink Quad 2, a
16480 profile is shared between any 2 sub-devices that utilize the same
16481 connectors. For the DeckLink 8K Pro, a profile is shared between
16482 all 4 sub-devices. So DeckLink 8K Pro support four profiles.
16483 Valid profile modes for DeckLink 8K Pro(with DeckLink SDK >= 11.0):
16485 one_sub_device_full, one_sub_device_half, two_sub_device_full,
16486 four_sub_device_half
16487 Valid profile modes for DeckLink Quad 2 and DeckLink Duo 2: half,
16489 full
16490 timecode_format
16492 Timecode type to include in the frame and video stream metadata.
16493 Must be none, rp188vitc, rp188vitc2, rp188ltc, rp188hfr, rp188any,
16494 vitc, vitc2, or serial. Defaults to none (not included).
16495 In order to properly support 50/60 fps timecodes, the ordering of
16497 the queried timecode types for rp188any is HFR, VITC1, VITC2 and
16498 LTC for >30 fps content. Note that this is slightly different to
16499 the ordering used by the DeckLink API, which is HFR, VITC1, LTC,
16500 VITC2.
16501 video_input
16503 Sets the video input source. Must be unset, sdi, hdmi, optical_sdi,
16504 component, composite or s_video. Defaults to unset.
16505 audio_input
16507 Sets the audio input source. Must be unset, embedded, aes_ebu,
16508 analog, analog_xlr, analog_rca or microphone. Defaults to unset.
16509 video_pts
16511 Sets the video packet timestamp source. Must be video, audio,
16512 reference, wallclock or abs_wallclock. Defaults to video.
16513 audio_pts
16515 Sets the audio packet timestamp source. Must be video, audio,
16516 reference, wallclock or abs_wallclock. Defaults to audio.
16517 draw_bars
16519 If set to true, color bars are drawn in the event of a signal loss.
16520 Defaults to true.
16521 queue_size
16523 Sets maximum input buffer size in bytes. If the buffering reaches
16524 this value, incoming frames will be dropped. Defaults to
16525 1073741824.
16526 audio_depth
16528 Sets the audio sample bit depth. Must be 16 or 32. Defaults to 16.
16529 decklink_copyts
16531 If set to true, timestamps are forwarded as they are without
16532 removing the initial offset. Defaults to false.
16533 timestamp_align
16535 Capture start time alignment in seconds. If set to nonzero, input
16536 frames are dropped till the system timestamp aligns with configured
16537 value. Alignment difference of up to one frame duration is
16538 tolerated. This is useful for maintaining input synchronization
16539 across N different hardware devices deployed for 'N-way'
16540 redundancy. The system time of different hardware devices should be
16541 synchronized with protocols such as NTP or PTP, before using this
16542 option. Note that this method is not foolproof. In some border
16543 cases input synchronization may not happen due to thread scheduling
16544 jitters in the OS. Either sync could go wrong by 1 frame or in a
16545 rarer case timestamp_align seconds. Defaults to 0.
16546 wait_for_tc (bool)
16548 Drop frames till a frame with timecode is received. Sometimes
16549 serial timecode isn't received with the first input frame. If that
16550 happens, the stored stream timecode will be inaccurate. If this
16551 option is set to true, input frames are dropped till a frame with
16552 timecode is received. Option timecode_format must be specified.
16553 Defaults to false.
16554 enable_klv(bool)
16556 If set to true, extracts KLV data from VANC and outputs KLV
16557 packets. KLV VANC packets are joined based on MID and PSC fields
16558 and aggregated into one KLV packet. Defaults to false.
16559 Examples
16561 • List input devices:
16563 ffmpeg -sources decklink
16565 • List supported formats:
16567 ffmpeg -f decklink -list_formats 1 -i 'Intensity Pro'
16569 • Capture video clip at 1080i50:
16571 ffmpeg -format_code Hi50 -f decklink -i 'Intensity Pro' -c:a copy -c:v copy output.avi
16573 • Capture video clip at 1080i50 10 bit:
16575 ffmpeg -raw_format yuv422p10 -format_code Hi50 -f decklink -i 'UltraStudio Mini Recorder' -c:a copy -c:v copy output.avi
16577 • Capture video clip at 1080i50 with 16 audio channels:
16579 ffmpeg -channels 16 -format_code Hi50 -f decklink -i 'UltraStudio Mini Recorder' -c:a copy -c:v copy output.avi
16581 dshow
16583 Windows DirectShow input device.
16584 DirectShow support is enabled when FFmpeg is built with the mingw-w64
16586 project. Currently only audio and video devices are supported.
16587 Multiple devices may be opened as separate inputs, but they may also be
16589 opened on the same input, which should improve synchronism between
16590 them.
16591 The input name should be in the format:
16593 <TYPE>=<NAME>[:<TYPE>=<NAME>]
16595 where TYPE can be either audio or video, and NAME is the device's name
16597 or alternative name..
16598 Options
16600 If no options are specified, the device's defaults are used. If the
16602 device does not support the requested options, it will fail to open.
16603 video_size
16605 Set the video size in the captured video.
16606 framerate
16608 Set the frame rate in the captured video.
16609 sample_rate
16611 Set the sample rate (in Hz) of the captured audio.
16612 sample_size
16614 Set the sample size (in bits) of the captured audio.
16615 channels
16617 Set the number of channels in the captured audio.
16618 list_devices
16620 If set to true, print a list of devices and exit.
16621 list_options
16623 If set to true, print a list of selected device's options and exit.
16624 video_device_number
16626 Set video device number for devices with the same name (starts at
16627 0, defaults to 0).
16628 audio_device_number
16630 Set audio device number for devices with the same name (starts at
16631 0, defaults to 0).
16632 pixel_format
16634 Select pixel format to be used by DirectShow. This may only be set
16635 when the video codec is not set or set to rawvideo.
16636 audio_buffer_size
16638 Set audio device buffer size in milliseconds (which can directly
16639 impact latency, depending on the device). Defaults to using the
16640 audio device's default buffer size (typically some multiple of
16641 500ms). Setting this value too low can degrade performance. See
16642 also
16643 <http://msdn.microsoft.com/en-us/library/windows/desktop/dd377582(v=vs.85).aspx>
16644 video_pin_name
16646 Select video capture pin to use by name or alternative name.
16647 audio_pin_name
16649 Select audio capture pin to use by name or alternative name.
16650 crossbar_video_input_pin_number
16652 Select video input pin number for crossbar device. This will be
16653 routed to the crossbar device's Video Decoder output pin. Note
16654 that changing this value can affect future invocations (sets a new
16655 default) until system reboot occurs.
16656 crossbar_audio_input_pin_number
16658 Select audio input pin number for crossbar device. This will be
16659 routed to the crossbar device's Audio Decoder output pin. Note
16660 that changing this value can affect future invocations (sets a new
16661 default) until system reboot occurs.
16662 show_video_device_dialog
16664 If set to true, before capture starts, popup a display dialog to
16665 the end user, allowing them to change video filter properties and
16666 configurations manually. Note that for crossbar devices, adjusting
16667 values in this dialog may be needed at times to toggle between PAL
16668 (25 fps) and NTSC (29.97) input frame rates, sizes, interlacing,
16669 etc. Changing these values can enable different scan rates/frame
16670 rates and avoiding green bars at the bottom, flickering scan lines,
16671 etc. Note that with some devices, changing these properties can
16672 also affect future invocations (sets new defaults) until system
16673 reboot occurs.
16674 show_audio_device_dialog
16676 If set to true, before capture starts, popup a display dialog to
16677 the end user, allowing them to change audio filter properties and
16678 configurations manually.
16679 show_video_crossbar_connection_dialog
16681 If set to true, before capture starts, popup a display dialog to
16682 the end user, allowing them to manually modify crossbar pin
16683 routings, when it opens a video device.
16684 show_audio_crossbar_connection_dialog
16686 If set to true, before capture starts, popup a display dialog to
16687 the end user, allowing them to manually modify crossbar pin
16688 routings, when it opens an audio device.
16689 show_analog_tv_tuner_dialog
16691 If set to true, before capture starts, popup a display dialog to
16692 the end user, allowing them to manually modify TV channels and
16693 frequencies.
16694 show_analog_tv_tuner_audio_dialog
16696 If set to true, before capture starts, popup a display dialog to
16697 the end user, allowing them to manually modify TV audio (like mono
16698 vs. stereo, Language A,B or C).
16699 audio_device_load
16701 Load an audio capture filter device from file instead of searching
16702 it by name. It may load additional parameters too, if the filter
16703 supports the serialization of its properties to. To use this an
16704 audio capture source has to be specified, but it can be anything
16705 even fake one.
16706 audio_device_save
16708 Save the currently used audio capture filter device and its
16709 parameters (if the filter supports it) to a file. If a file with
16710 the same name exists it will be overwritten.
16711 video_device_load
16713 Load a video capture filter device from file instead of searching
16714 it by name. It may load additional parameters too, if the filter
16715 supports the serialization of its properties to. To use this a
16716 video capture source has to be specified, but it can be anything
16717 even fake one.
16718 video_device_save
16720 Save the currently used video capture filter device and its
16721 parameters (if the filter supports it) to a file. If a file with
16722 the same name exists it will be overwritten.
16723 use_video_device_timestamps
16725 If set to false, the timestamp for video frames will be derived
16726 from the wallclock instead of the timestamp provided by the capture
16727 device. This allows working around devices that provide unreliable
16728 timestamps.
16729 Examples
16731 • Print the list of DirectShow supported devices and exit:
16733 $ ffmpeg -list_devices true -f dshow -i dummy
16735 • Open video device Camera:
16737 $ ffmpeg -f dshow -i video="Camera"
16739 • Open second video device with name Camera:
16741 $ ffmpeg -f dshow -video_device_number 1 -i video="Camera"
16743 • Open video device Camera and audio device Microphone:
16745 $ ffmpeg -f dshow -i video="Camera":audio="Microphone"
16747 • Print the list of supported options in selected device and exit:
16749 $ ffmpeg -list_options true -f dshow -i video="Camera"
16751 • Specify pin names to capture by name or alternative name, specify
16753 alternative device name:
16754 $ ffmpeg -f dshow -audio_pin_name "Audio Out" -video_pin_name 2 -i video=video="@device_pnp_\\?\pci#ven_1a0a&dev_6200&subsys_62021461&rev_01#4&e2c7dd6&0&00e1#{65e8773d-8f56-11d0-a3b9-00a0c9223196}\{ca465100-deb0-4d59-818f-8c477184adf6}":audio="Microphone"
16756 • Configure a crossbar device, specifying crossbar pins, allow user
16758 to adjust video capture properties at startup:
16759 $ ffmpeg -f dshow -show_video_device_dialog true -crossbar_video_input_pin_number 0
16761 -crossbar_audio_input_pin_number 3 -i video="AVerMedia BDA Analog Capture":audio="AVerMedia BDA Analog Capture"
16762 fbdev
16764 Linux framebuffer input device.
16765 The Linux framebuffer is a graphic hardware-independent abstraction
16767 layer to show graphics on a computer monitor, typically on the console.
16768 It is accessed through a file device node, usually /dev/fb0.
16769 For more detailed information read the file
16771 Documentation/fb/framebuffer.txt included in the Linux source tree.
16772 See also <http://linux-fbdev.sourceforge.net/>, and fbset(1).
16774 To record from the framebuffer device /dev/fb0 with ffmpeg:
16776 ffmpeg -f fbdev -framerate 10 -i /dev/fb0 out.avi
16778 You can take a single screenshot image with the command:
16780 ffmpeg -f fbdev -framerate 1 -i /dev/fb0 -frames:v 1 screenshot.jpeg
16782 Options
16784 framerate
16786 Set the frame rate. Default is 25.
16787 gdigrab
16789 Win32 GDI-based screen capture device.
16790 This device allows you to capture a region of the display on Windows.
16792 There are two options for the input filename:
16794 desktop
16796 or
16798 title=<window_title>
16800 The first option will capture the entire desktop, or a fixed region of
16802 the desktop. The second option will instead capture the contents of a
16803 single window, regardless of its position on the screen.
16804 For example, to grab the entire desktop using ffmpeg:
16806 ffmpeg -f gdigrab -framerate 6 -i desktop out.mpg
16808 Grab a 640x480 region at position "10,20":
16810 ffmpeg -f gdigrab -framerate 6 -offset_x 10 -offset_y 20 -video_size vga -i desktop out.mpg
16812 Grab the contents of the window named "Calculator"
16814 ffmpeg -f gdigrab -framerate 6 -i title=Calculator out.mpg
16816 Options
16818 draw_mouse
16820 Specify whether to draw the mouse pointer. Use the value 0 to not
16821 draw the pointer. Default value is 1.
16822 framerate
16824 Set the grabbing frame rate. Default value is "ntsc", corresponding
16825 to a frame rate of "30000/1001".
16826 show_region
16828 Show grabbed region on screen.
16829 If show_region is specified with 1, then the grabbing region will
16831 be indicated on screen. With this option, it is easy to know what
16832 is being grabbed if only a portion of the screen is grabbed.
16833 Note that show_region is incompatible with grabbing the contents of
16835 a single window.
16836 For example:
16838 ffmpeg -f gdigrab -show_region 1 -framerate 6 -video_size cif -offset_x 10 -offset_y 20 -i desktop out.mpg
16840 video_size
16842 Set the video frame size. The default is to capture the full screen
16843 if desktop is selected, or the full window size if
16844 title=window_title is selected.
16845 offset_x
16847 When capturing a region with video_size, set the distance from the
16848 left edge of the screen or desktop.
16849 Note that the offset calculation is from the top left corner of the
16851 primary monitor on Windows. If you have a monitor positioned to the
16852 left of your primary monitor, you will need to use a negative
16853 offset_x value to move the region to that monitor.
16854 offset_y
16856 When capturing a region with video_size, set the distance from the
16857 top edge of the screen or desktop.
16858 Note that the offset calculation is from the top left corner of the
16860 primary monitor on Windows. If you have a monitor positioned above
16861 your primary monitor, you will need to use a negative offset_y
16862 value to move the region to that monitor.
16863 iec61883
16865 FireWire DV/HDV input device using libiec61883.
16866 To enable this input device, you need libiec61883, libraw1394 and
16868 libavc1394 installed on your system. Use the configure option
16869 "--enable-libiec61883" to compile with the device enabled.
16870 The iec61883 capture device supports capturing from a video device
16872 connected via IEEE1394 (FireWire), using libiec61883 and the new Linux
16873 FireWire stack (juju). This is the default DV/HDV input method in Linux
16874 Kernel 2.6.37 and later, since the old FireWire stack was removed.
16875 Specify the FireWire port to be used as input file, or "auto" to choose
16877 the first port connected.
16878 Options
16880 dvtype
16882 Override autodetection of DV/HDV. This should only be used if auto
16883 detection does not work, or if usage of a different device type
16884 should be prohibited. Treating a DV device as HDV (or vice versa)
16885 will not work and result in undefined behavior. The values auto,
16886 dv and hdv are supported.
16887 dvbuffer
16889 Set maximum size of buffer for incoming data, in frames. For DV,
16890 this is an exact value. For HDV, it is not frame exact, since HDV
16891 does not have a fixed frame size.
16892 dvguid
16894 Select the capture device by specifying its GUID. Capturing will
16895 only be performed from the specified device and fails if no device
16896 with the given GUID is found. This is useful to select the input if
16897 multiple devices are connected at the same time. Look at
16898 /sys/bus/firewire/devices to find out the GUIDs.
16899 Examples
16901 • Grab and show the input of a FireWire DV/HDV device.
16903 ffplay -f iec61883 -i auto
16905 • Grab and record the input of a FireWire DV/HDV device, using a
16907 packet buffer of 100000 packets if the source is HDV.
16908 ffmpeg -f iec61883 -i auto -dvbuffer 100000 out.mpg
16910 jack
16912 JACK input device.
16913 To enable this input device during configuration you need libjack
16915 installed on your system.
16916 A JACK input device creates one or more JACK writable clients, one for
16918 each audio channel, with name client_name:input_N, where client_name is
16919 the name provided by the application, and N is a number which
16920 identifies the channel. Each writable client will send the acquired
16921 data to the FFmpeg input device.
16922 Once you have created one or more JACK readable clients, you need to
16924 connect them to one or more JACK writable clients.
16925 To connect or disconnect JACK clients you can use the jack_connect and
16927 jack_disconnect programs, or do it through a graphical interface, for
16928 example with qjackctl.
16929 To list the JACK clients and their properties you can invoke the
16931 command jack_lsp.
16932 Follows an example which shows how to capture a JACK readable client
16934 with ffmpeg.
16935 # Create a JACK writable client with name "ffmpeg".
16937 $ ffmpeg -f jack -i ffmpeg -y out.wav
16938 # Start the sample jack_metro readable client.
16940 $ jack_metro -b 120 -d 0.2 -f 4000
16941 # List the current JACK clients.
16943 $ jack_lsp -c
16944 system:capture_1
16945 system:capture_2
16946 system:playback_1
16947 system:playback_2
16948 ffmpeg:input_1
16949 metro:120_bpm
16950 # Connect metro to the ffmpeg writable client.
16952 $ jack_connect metro:120_bpm ffmpeg:input_1
16953 For more information read: <http://jackaudio.org/>
16955 Options
16957 channels
16959 Set the number of channels. Default is 2.
16960 kmsgrab
16962 KMS video input device.
16963 Captures the KMS scanout framebuffer associated with a specified CRTC
16965 or plane as a DRM object that can be passed to other hardware
16966 functions.
16967 Requires either DRM master or CAP_SYS_ADMIN to run.
16969 If you don't understand what all of that means, you probably don't want
16971 this. Look at x11grab instead.
16972 Options
16974 device
16976 DRM device to capture on. Defaults to /dev/dri/card0.
16977 format
16979 Pixel format of the framebuffer. This can be autodetected if you
16980 are running Linux 5.7 or later, but needs to be provided for
16981 earlier versions. Defaults to bgr0, which is the most common
16982 format used by the Linux console and Xorg X server.
16983 format_modifier
16985 Format modifier to signal on output frames. This is necessary to
16986 import correctly into some APIs. It can be autodetected if you are
16987 running Linux 5.7 or later, but will need to be provided explicitly
16988 when needed in earlier versions. See the libdrm documentation for
16989 possible values.
16990 crtc_id
16992 KMS CRTC ID to define the capture source. The first active plane
16993 on the given CRTC will be used.
16994 plane_id
16996 KMS plane ID to define the capture source. Defaults to the first
16997 active plane found if neither crtc_id nor plane_id are specified.
16998 framerate
17000 Framerate to capture at. This is not synchronised to any page
17001 flipping or framebuffer changes - it just defines the interval at
17002 which the framebuffer is sampled. Sampling faster than the
17003 framebuffer update rate will generate independent frames with the
17004 same content. Defaults to 30.
17005 Examples
17007 • Capture from the first active plane, download the result to normal
17009 frames and encode. This will only work if the framebuffer is both
17010 linear and mappable - if not, the result may be scrambled or fail
17011 to download.
17012 ffmpeg -f kmsgrab -i - -vf 'hwdownload,format=bgr0' output.mp4
17014 • Capture from CRTC ID 42 at 60fps, map the result to VAAPI, convert
17016 to NV12 and encode as H.264.
17017 ffmpeg -crtc_id 42 -framerate 60 -f kmsgrab -i - -vf 'hwmap=derive_device=vaapi,scale_vaapi=w=1920:h=1080:format=nv12' -c:v h264_vaapi output.mp4
17019 • To capture only part of a plane the output can be cropped - this
17021 can be used to capture a single window, as long as it has a known
17022 absolute position and size. For example, to capture and encode the
17023 middle quarter of a 1920x1080 plane:
17024 ffmpeg -f kmsgrab -i - -vf 'hwmap=derive_device=vaapi,crop=960:540:480:270,scale_vaapi=960:540:nv12' -c:v h264_vaapi output.mp4
17026 lavfi
17028 Libavfilter input virtual device.
17029 This input device reads data from the open output pads of a libavfilter
17031 filtergraph.
17032 For each filtergraph open output, the input device will create a
17034 corresponding stream which is mapped to the generated output. Currently
17035 only video data is supported. The filtergraph is specified through the
17036 option graph.
17037 Options
17039 graph
17041 Specify the filtergraph to use as input. Each video open output
17042 must be labelled by a unique string of the form "outN", where N is
17043 a number starting from 0 corresponding to the mapped input stream
17044 generated by the device. The first unlabelled output is
17045 automatically assigned to the "out0" label, but all the others need
17046 to be specified explicitly.
17047 The suffix "+subcc" can be appended to the output label to create
17049 an extra stream with the closed captions packets attached to that
17050 output (experimental; only for EIA-608 / CEA-708 for now). The
17051 subcc streams are created after all the normal streams, in the
17052 order of the corresponding stream. For example, if there is
17053 "out19+subcc", "out7+subcc" and up to "out42", the stream #43 is
17054 subcc for stream #7 and stream #44 is subcc for stream #19.
17055 If not specified defaults to the filename specified for the input
17057 device.
17058 graph_file
17060 Set the filename of the filtergraph to be read and sent to the
17061 other filters. Syntax of the filtergraph is the same as the one
17062 specified by the option graph.
17063 dumpgraph
17065 Dump graph to stderr.
17066 Examples
17068 • Create a color video stream and play it back with ffplay:
17070 ffplay -f lavfi -graph "color=c=pink [out0]" dummy
17072 • As the previous example, but use filename for specifying the graph
17074 description, and omit the "out0" label:
17075 ffplay -f lavfi color=c=pink
17077 • Create three different video test filtered sources and play them:
17079 ffplay -f lavfi -graph "testsrc [out0]; testsrc,hflip [out1]; testsrc,negate [out2]" test3
17081 • Read an audio stream from a file using the amovie source and play
17083 it back with ffplay:
17084 ffplay -f lavfi "amovie=test.wav"
17086 • Read an audio stream and a video stream and play it back with
17088 ffplay:
17089 ffplay -f lavfi "movie=test.avi[out0];amovie=test.wav[out1]"
17091 • Dump decoded frames to images and closed captions to a file
17093 (experimental):
17094 ffmpeg -f lavfi -i "movie=test.ts[out0+subcc]" -map v frame%08d.png -map s -c copy -f rawvideo subcc.bin
17096 libcdio
17098 Audio-CD input device based on libcdio.
17099 To enable this input device during configuration you need libcdio
17101 installed on your system. It requires the configure option
17102 "--enable-libcdio".
17103 This device allows playing and grabbing from an Audio-CD.
17105 For example to copy with ffmpeg the entire Audio-CD in /dev/sr0, you
17107 may run the command:
17108 ffmpeg -f libcdio -i /dev/sr0 cd.wav
17110 Options
17112 speed
17114 Set drive reading speed. Default value is 0.
17115 The speed is specified CD-ROM speed units. The speed is set through
17117 the libcdio "cdio_cddap_speed_set" function. On many CD-ROM drives,
17118 specifying a value too large will result in using the fastest
17119 speed.
17120 paranoia_mode
17122 Set paranoia recovery mode flags. It accepts one of the following
17123 values:
17124 disable
17126 verify
17127 overlap
17128 neverskip
17129 full
17130 Default value is disable.
17132 For more information about the available recovery modes, consult
17134 the paranoia project documentation.
17135 libdc1394
17137 IIDC1394 input device, based on libdc1394 and libraw1394.
17138 Requires the configure option "--enable-libdc1394".
17140 Options
17142 framerate
17144 Set the frame rate. Default is "ntsc", corresponding to a frame
17145 rate of "30000/1001".
17146 pixel_format
17148 Select the pixel format. Default is "uyvy422".
17149 video_size
17151 Set the video size given as a string such as "640x480" or "hd720".
17152 Default is "qvga".
17153 openal
17155 The OpenAL input device provides audio capture on all systems with a
17156 working OpenAL 1.1 implementation.
17157 To enable this input device during configuration, you need OpenAL
17159 headers and libraries installed on your system, and need to configure
17160 FFmpeg with "--enable-openal".
17161 OpenAL headers and libraries should be provided as part of your OpenAL
17163 implementation, or as an additional download (an SDK). Depending on
17164 your installation you may need to specify additional flags via the
17165 "--extra-cflags" and "--extra-ldflags" for allowing the build system to
17166 locate the OpenAL headers and libraries.
17167 An incomplete list of OpenAL implementations follows:
17169 Creative
17171 The official Windows implementation, providing hardware
17172 acceleration with supported devices and software fallback. See
17173 <http://openal.org/>.
17174 OpenAL Soft
17176 Portable, open source (LGPL) software implementation. Includes
17177 backends for the most common sound APIs on the Windows, Linux,
17178 Solaris, and BSD operating systems. See
17179 <http://kcat.strangesoft.net/openal.html>.
17180 Apple
17182 OpenAL is part of Core Audio, the official Mac OS X Audio
17183 interface. See
17184 <http://developer.apple.com/technologies/mac/audio-and-video.html>
17185 This device allows one to capture from an audio input device handled
17187 through OpenAL.
17188 You need to specify the name of the device to capture in the provided
17190 filename. If the empty string is provided, the device will
17191 automatically select the default device. You can get the list of the
17192 supported devices by using the option list_devices.
17193 Options
17195 channels
17197 Set the number of channels in the captured audio. Only the values 1
17198 (monaural) and 2 (stereo) are currently supported. Defaults to 2.
17199 sample_size
17201 Set the sample size (in bits) of the captured audio. Only the
17202 values 8 and 16 are currently supported. Defaults to 16.
17203 sample_rate
17205 Set the sample rate (in Hz) of the captured audio. Defaults to
17206 44.1k.
17207 list_devices
17209 If set to true, print a list of devices and exit. Defaults to
17210 false.
17211 Examples
17213 Print the list of OpenAL supported devices and exit:
17215 $ ffmpeg -list_devices true -f openal -i dummy out.ogg
17217 Capture from the OpenAL device DR-BT101 via PulseAudio:
17219 $ ffmpeg -f openal -i 'DR-BT101 via PulseAudio' out.ogg
17221 Capture from the default device (note the empty string '' as filename):
17223 $ ffmpeg -f openal -i '' out.ogg
17225 Capture from two devices simultaneously, writing to two different
17227 files, within the same ffmpeg command:
17228 $ ffmpeg -f openal -i 'DR-BT101 via PulseAudio' out1.ogg -f openal -i 'ALSA Default' out2.ogg
17230 Note: not all OpenAL implementations support multiple simultaneous
17232 capture - try the latest OpenAL Soft if the above does not work.
17233 oss
17235 Open Sound System input device.
17236 The filename to provide to the input device is the device node
17238 representing the OSS input device, and is usually set to /dev/dsp.
17239 For example to grab from /dev/dsp using ffmpeg use the command:
17241 ffmpeg -f oss -i /dev/dsp /tmp/oss.wav
17243 For more information about OSS see:
17245 <http://manuals.opensound.com/usersguide/dsp.html>
17246 Options
17248 sample_rate
17250 Set the sample rate in Hz. Default is 48000.
17251 channels
17253 Set the number of channels. Default is 2.
17254 pulse
17256 PulseAudio input device.
17257 To enable this output device you need to configure FFmpeg with
17259 "--enable-libpulse".
17260 The filename to provide to the input device is a source device or the
17262 string "default"
17263 To list the PulseAudio source devices and their properties you can
17265 invoke the command pactl list sources.
17266 More information about PulseAudio can be found on
17268 <http://www.pulseaudio.org>.
17269 Options
17271 server
17273 Connect to a specific PulseAudio server, specified by an IP
17274 address. Default server is used when not provided.
17275 name
17277 Specify the application name PulseAudio will use when showing
17278 active clients, by default it is the "LIBAVFORMAT_IDENT" string.
17279 stream_name
17281 Specify the stream name PulseAudio will use when showing active
17282 streams, by default it is "record".
17283 sample_rate
17285 Specify the samplerate in Hz, by default 48kHz is used.
17286 channels
17288 Specify the channels in use, by default 2 (stereo) is set.
17289 frame_size
17291 This option does nothing and is deprecated.
17292 fragment_size
17294 Specify the size in bytes of the minimal buffering fragment in
17295 PulseAudio, it will affect the audio latency. By default it is set
17296 to 50 ms amount of data.
17297 wallclock
17299 Set the initial PTS using the current time. Default is 1.
17300 Examples
17302 Record a stream from default device:
17304 ffmpeg -f pulse -i default /tmp/pulse.wav
17306 sndio
17308 sndio input device.
17309 To enable this input device during configuration you need libsndio
17311 installed on your system.
17312 The filename to provide to the input device is the device node
17314 representing the sndio input device, and is usually set to /dev/audio0.
17315 For example to grab from /dev/audio0 using ffmpeg use the command:
17317 ffmpeg -f sndio -i /dev/audio0 /tmp/oss.wav
17319 Options
17321 sample_rate
17323 Set the sample rate in Hz. Default is 48000.
17324 channels
17326 Set the number of channels. Default is 2.
17327 video4linux2, v4l2
17329 Video4Linux2 input video device.
17330 "v4l2" can be used as alias for "video4linux2".
17332 If FFmpeg is built with v4l-utils support (by using the
17334 "--enable-libv4l2" configure option), it is possible to use it with the
17335 "-use_libv4l2" input device option.
17336 The name of the device to grab is a file device node, usually Linux
17338 systems tend to automatically create such nodes when the device (e.g.
17339 an USB webcam) is plugged into the system, and has a name of the kind
17340 /dev/videoN, where N is a number associated to the device.
17341 Video4Linux2 devices usually support a limited set of widthxheight
17343 sizes and frame rates. You can check which are supported using
17344 -list_formats all for Video4Linux2 devices. Some devices, like TV
17345 cards, support one or more standards. It is possible to list all the
17346 supported standards using -list_standards all.
17347 The time base for the timestamps is 1 microsecond. Depending on the
17349 kernel version and configuration, the timestamps may be derived from
17350 the real time clock (origin at the Unix Epoch) or the monotonic clock
17351 (origin usually at boot time, unaffected by NTP or manual changes to
17352 the clock). The -timestamps abs or -ts abs option can be used to force
17353 conversion into the real time clock.
17354 Some usage examples of the video4linux2 device with ffmpeg and ffplay:
17356 • List supported formats for a video4linux2 device:
17358 ffplay -f video4linux2 -list_formats all /dev/video0
17360 • Grab and show the input of a video4linux2 device:
17362 ffplay -f video4linux2 -framerate 30 -video_size hd720 /dev/video0
17364 • Grab and record the input of a video4linux2 device, leave the frame
17366 rate and size as previously set:
17367 ffmpeg -f video4linux2 -input_format mjpeg -i /dev/video0 out.mpeg
17369 For more information about Video4Linux, check <http://linuxtv.org/>.
17371 Options
17373 standard
17375 Set the standard. Must be the name of a supported standard. To get
17376 a list of the supported standards, use the list_standards option.
17377 channel
17379 Set the input channel number. Default to -1, which means using the
17380 previously selected channel.
17381 video_size
17383 Set the video frame size. The argument must be a string in the form
17384 WIDTHxHEIGHT or a valid size abbreviation.
17385 pixel_format
17387 Select the pixel format (only valid for raw video input).
17388 input_format
17390 Set the preferred pixel format (for raw video) or a codec name.
17391 This option allows one to select the input format, when several are
17392 available.
17393 framerate
17395 Set the preferred video frame rate.
17396 list_formats
17398 List available formats (supported pixel formats, codecs, and frame
17399 sizes) and exit.
17400 Available values are:
17402 all Show all available (compressed and non-compressed) formats.
17404 raw Show only raw video (non-compressed) formats.
17406 compressed
17408 Show only compressed formats.
17409 list_standards
17411 List supported standards and exit.
17412 Available values are:
17414 all Show all supported standards.
17416 timestamps, ts
17418 Set type of timestamps for grabbed frames.
17419 Available values are:
17421 default
17423 Use timestamps from the kernel.
17424 abs Use absolute timestamps (wall clock).
17426 mono2abs
17428 Force conversion from monotonic to absolute timestamps.
17429 Default value is "default".
17431 use_libv4l2
17433 Use libv4l2 (v4l-utils) conversion functions. Default is 0.
17434 vfwcap
17436 VfW (Video for Windows) capture input device.
17437 The filename passed as input is the capture driver number, ranging from
17439 0 to 9. You may use "list" as filename to print a list of drivers. Any
17440 other filename will be interpreted as device number 0.
17441 Options
17443 video_size
17445 Set the video frame size.
17446 framerate
17448 Set the grabbing frame rate. Default value is "ntsc", corresponding
17449 to a frame rate of "30000/1001".
17450 x11grab
17452 X11 video input device.
17453 To enable this input device during configuration you need libxcb
17455 installed on your system. It will be automatically detected during
17456 configuration.
17457 This device allows one to capture a region of an X11 display.
17459 The filename passed as input has the syntax:
17461 [<hostname>]:<display_number>.<screen_number>[+<x_offset>,<y_offset>]
17463 hostname:display_number.screen_number specifies the X11 display name of
17465 the screen to grab from. hostname can be omitted, and defaults to
17466 "localhost". The environment variable DISPLAY contains the default
17467 display name.
17468 x_offset and y_offset specify the offsets of the grabbed area with
17470 respect to the top-left border of the X11 screen. They default to 0.
17471 Check the X11 documentation (e.g. man X) for more detailed information.
17473 Use the xdpyinfo program for getting basic information about the
17475 properties of your X11 display (e.g. grep for "name" or "dimensions").
17476 For example to grab from :0.0 using ffmpeg:
17478 ffmpeg -f x11grab -framerate 25 -video_size cif -i :0.0 out.mpg
17480 Grab at position "10,20":
17482 ffmpeg -f x11grab -framerate 25 -video_size cif -i :0.0+10,20 out.mpg
17484 Options
17486 select_region
17488 Specify whether to select the grabbing area graphically using the
17489 pointer. A value of 1 prompts the user to select the grabbing area
17490 graphically by clicking and dragging. A single click with no
17491 dragging will select the whole screen. A region with zero width or
17492 height will also select the whole screen. This option overwrites
17493 the video_size, grab_x, and grab_y options. Default value is 0.
17494 draw_mouse
17496 Specify whether to draw the mouse pointer. A value of 0 specifies
17497 not to draw the pointer. Default value is 1.
17498 follow_mouse
17500 Make the grabbed area follow the mouse. The argument can be
17501 "centered" or a number of pixels PIXELS.
17502 When it is specified with "centered", the grabbing region follows
17504 the mouse pointer and keeps the pointer at the center of region;
17505 otherwise, the region follows only when the mouse pointer reaches
17506 within PIXELS (greater than zero) to the edge of region.
17507 For example:
17509 ffmpeg -f x11grab -follow_mouse centered -framerate 25 -video_size cif -i :0.0 out.mpg
17511 To follow only when the mouse pointer reaches within 100 pixels to
17513 edge:
17514 ffmpeg -f x11grab -follow_mouse 100 -framerate 25 -video_size cif -i :0.0 out.mpg
17516 framerate
17518 Set the grabbing frame rate. Default value is "ntsc", corresponding
17519 to a frame rate of "30000/1001".
17520 show_region
17522 Show grabbed region on screen.
17523 If show_region is specified with 1, then the grabbing region will
17525 be indicated on screen. With this option, it is easy to know what
17526 is being grabbed if only a portion of the screen is grabbed.
17527 region_border
17529 Set the region border thickness if -show_region 1 is used. Range
17530 is 1 to 128 and default is 3 (XCB-based x11grab only).
17531 For example:
17533 ffmpeg -f x11grab -show_region 1 -framerate 25 -video_size cif -i :0.0+10,20 out.mpg
17535 With follow_mouse:
17537 ffmpeg -f x11grab -follow_mouse centered -show_region 1 -framerate 25 -video_size cif -i :0.0 out.mpg
17539 window_id
17541 Grab this window, instead of the whole screen. Default value is 0,
17542 which maps to the whole screen (root window).
17543 The id of a window can be found using the xwininfo program,
17545 possibly with options -tree and -root.
17546 If the window is later enlarged, the new area is not recorded.
17548 Video ends when the window is closed, unmapped (i.e., iconified) or
17549 shrunk beyond the video size (which defaults to the initial window
17550 size).
17551 This option disables options follow_mouse and select_region.
17553 video_size
17555 Set the video frame size. Default is the full desktop or window.
17556 grab_x
17558 grab_y
17559 Set the grabbing region coordinates. They are expressed as offset
17560 from the top left corner of the X11 window and correspond to the
17561 x_offset and y_offset parameters in the device name. The default
17562 value for both options is 0.
17563
17565 Output devices are configured elements in FFmpeg that can write
17566 multimedia data to an output device attached to your system.
17567 When you configure your FFmpeg build, all the supported output devices
17569 are enabled by default. You can list all available ones using the
17570 configure option "--list-outdevs".
17571 You can disable all the output devices using the configure option
17573 "--disable-outdevs", and selectively enable an output device using the
17574 option "--enable-outdev=OUTDEV", or you can disable a particular input
17575 device using the option "--disable-outdev=OUTDEV".
17576 The option "-devices" of the ff* tools will display the list of enabled
17578 output devices.
17579 A description of the currently available output devices follows.
17581 alsa
17583 ALSA (Advanced Linux Sound Architecture) output device.
17584 Examples
17586 • Play a file on default ALSA device:
17588 ffmpeg -i INPUT -f alsa default
17590 • Play a file on soundcard 1, audio device 7:
17592 ffmpeg -i INPUT -f alsa hw:1,7
17594 AudioToolbox
17596 AudioToolbox output device.
17597 Allows native output to CoreAudio devices on OSX.
17599 The output filename can be empty (or "-") to refer to the default
17601 system output device or a number that refers to the device index as
17602 shown using: "-list_devices true".
17603 Alternatively, the audio input device can be chosen by index using the
17605 B<-audio_device_index E<lt>INDEXE<gt>>
17607 , overriding any device name or index given in the input filename.
17609 All available devices can be enumerated by using -list_devices true,
17611 listing all device names, UIDs and corresponding indices.
17612 Options
17614 AudioToolbox supports the following options:
17616 -audio_device_index <INDEX>
17618 Specify the audio device by its index. Overrides anything given in
17619 the output filename.
17620 Examples
17622 • Print the list of supported devices and output a sine wave to the
17624 default device:
17625 $ ffmpeg -f lavfi -i sine=r=44100 -f audiotoolbox -list_devices true -
17627 • Output a sine wave to the device with the index 2, overriding any
17629 output filename:
17630 $ ffmpeg -f lavfi -i sine=r=44100 -f audiotoolbox -audio_device_index 2 -
17632 caca
17634 CACA output device.
17635 This output device allows one to show a video stream in CACA window.
17637 Only one CACA window is allowed per application, so you can have only
17638 one instance of this output device in an application.
17639 To enable this output device you need to configure FFmpeg with
17641 "--enable-libcaca". libcaca is a graphics library that outputs text
17642 instead of pixels.
17643 For more information about libcaca, check:
17645 <http://caca.zoy.org/wiki/libcaca>
17646 Options
17648 window_title
17650 Set the CACA window title, if not specified default to the filename
17651 specified for the output device.
17652 window_size
17654 Set the CACA window size, can be a string of the form widthxheight
17655 or a video size abbreviation. If not specified it defaults to the
17656 size of the input video.
17657 driver
17659 Set display driver.
17660 algorithm
17662 Set dithering algorithm. Dithering is necessary because the picture
17663 being rendered has usually far more colours than the available
17664 palette. The accepted values are listed with "-list_dither
17665 algorithms".
17666 antialias
17668 Set antialias method. Antialiasing smoothens the rendered image and
17669 avoids the commonly seen staircase effect. The accepted values are
17670 listed with "-list_dither antialiases".
17671 charset
17673 Set which characters are going to be used when rendering text. The
17674 accepted values are listed with "-list_dither charsets".
17675 color
17677 Set color to be used when rendering text. The accepted values are
17678 listed with "-list_dither colors".
17679 list_drivers
17681 If set to true, print a list of available drivers and exit.
17682 list_dither
17684 List available dither options related to the argument. The
17685 argument must be one of "algorithms", "antialiases", "charsets",
17686 "colors".
17687 Examples
17689 • The following command shows the ffmpeg output is an CACA window,
17691 forcing its size to 80x25:
17692 ffmpeg -i INPUT -c:v rawvideo -pix_fmt rgb24 -window_size 80x25 -f caca -
17694 • Show the list of available drivers and exit:
17696 ffmpeg -i INPUT -pix_fmt rgb24 -f caca -list_drivers true -
17698 • Show the list of available dither colors and exit:
17700 ffmpeg -i INPUT -pix_fmt rgb24 -f caca -list_dither colors -
17702 decklink
17704 The decklink output device provides playback capabilities for
17705 Blackmagic DeckLink devices.
17706 To enable this output device, you need the Blackmagic DeckLink SDK and
17708 you need to configure with the appropriate "--extra-cflags" and
17709 "--extra-ldflags". On Windows, you need to run the IDL files through
17710 widl.
17711 DeckLink is very picky about the formats it supports. Pixel format is
17713 always uyvy422, framerate, field order and video size must be
17714 determined for your device with -list_formats 1. Audio sample rate is
17715 always 48 kHz.
17716 Options
17718 list_devices
17720 If set to true, print a list of devices and exit. Defaults to
17721 false. This option is deprecated, please use the "-sinks" option of
17722 ffmpeg to list the available output devices.
17723 list_formats
17725 If set to true, print a list of supported formats and exit.
17726 Defaults to false.
17727 preroll
17729 Amount of time to preroll video in seconds. Defaults to 0.5.
17730 duplex_mode
17732 Sets the decklink device duplex/profile mode. Must be unset, half,
17733 full, one_sub_device_full, one_sub_device_half,
17734 two_sub_device_full, four_sub_device_half Defaults to unset.
17735 Note: DeckLink SDK 11.0 have replaced the duplex property by a
17737 profile property. For the DeckLink Duo 2 and DeckLink Quad 2, a
17738 profile is shared between any 2 sub-devices that utilize the same
17739 connectors. For the DeckLink 8K Pro, a profile is shared between
17740 all 4 sub-devices. So DeckLink 8K Pro support four profiles.
17741 Valid profile modes for DeckLink 8K Pro(with DeckLink SDK >= 11.0):
17743 one_sub_device_full, one_sub_device_half, two_sub_device_full,
17744 four_sub_device_half
17745 Valid profile modes for DeckLink Quad 2 and DeckLink Duo 2: half,
17747 full
17748 timing_offset
17750 Sets the genlock timing pixel offset on the used output. Defaults
17751 to unset.
17752 link
17754 Sets the SDI video link configuration on the used output. Must be
17755 unset, single link SDI, dual link SDI or quad link SDI. Defaults
17756 to unset.
17757 sqd Enable Square Division Quad Split mode for Quad-link SDI output.
17759 Must be unset, true or false. Defaults to unset.
17760 level_a
17762 Enable SMPTE Level A mode on the used output. Must be unset, true
17763 or false. Defaults to unset.
17764 Examples
17766 • List output devices:
17768 ffmpeg -sinks decklink
17770 • List supported formats:
17772 ffmpeg -i test.avi -f decklink -list_formats 1 'DeckLink Mini Monitor'
17774 • Play video clip:
17776 ffmpeg -i test.avi -f decklink -pix_fmt uyvy422 'DeckLink Mini Monitor'
17778 • Play video clip with non-standard framerate or video size:
17780 ffmpeg -i test.avi -f decklink -pix_fmt uyvy422 -s 720x486 -r 24000/1001 'DeckLink Mini Monitor'
17782 fbdev
17784 Linux framebuffer output device.
17785 The Linux framebuffer is a graphic hardware-independent abstraction
17787 layer to show graphics on a computer monitor, typically on the console.
17788 It is accessed through a file device node, usually /dev/fb0.
17789 For more detailed information read the file
17791 Documentation/fb/framebuffer.txt included in the Linux source tree.
17792 Options
17794 xoffset
17796 yoffset
17797 Set x/y coordinate of top left corner. Default is 0.
17798 Examples
17800 Play a file on framebuffer device /dev/fb0. Required pixel format
17802 depends on current framebuffer settings.
17803 ffmpeg -re -i INPUT -c:v rawvideo -pix_fmt bgra -f fbdev /dev/fb0
17805 See also <http://linux-fbdev.sourceforge.net/>, and fbset(1).
17807 opengl
17809 OpenGL output device.
17810 To enable this output device you need to configure FFmpeg with
17812 "--enable-opengl".
17813 This output device allows one to render to OpenGL context. Context may
17815 be provided by application or default SDL window is created.
17816 When device renders to external context, application must implement
17818 handlers for following messages: "AV_DEV_TO_APP_CREATE_WINDOW_BUFFER" -
17819 create OpenGL context on current thread.
17820 "AV_DEV_TO_APP_PREPARE_WINDOW_BUFFER" - make OpenGL context current.
17821 "AV_DEV_TO_APP_DISPLAY_WINDOW_BUFFER" - swap buffers.
17822 "AV_DEV_TO_APP_DESTROY_WINDOW_BUFFER" - destroy OpenGL context.
17823 Application is also required to inform a device about current
17824 resolution by sending "AV_APP_TO_DEV_WINDOW_SIZE" message.
17825 Options
17827 background
17829 Set background color. Black is a default.
17830 no_window
17832 Disables default SDL window when set to non-zero value.
17833 Application must provide OpenGL context and both "window_size_cb"
17834 and "window_swap_buffers_cb" callbacks when set.
17835 window_title
17837 Set the SDL window title, if not specified default to the filename
17838 specified for the output device. Ignored when no_window is set.
17839 window_size
17841 Set preferred window size, can be a string of the form widthxheight
17842 or a video size abbreviation. If not specified it defaults to the
17843 size of the input video, downscaled according to the aspect ratio.
17844 Mostly usable when no_window is not set.
17845 Examples
17847 Play a file on SDL window using OpenGL rendering:
17849 ffmpeg -i INPUT -f opengl "window title"
17851 oss
17853 OSS (Open Sound System) output device.
17854 pulse
17856 PulseAudio output device.
17857 To enable this output device you need to configure FFmpeg with
17859 "--enable-libpulse".
17860 More information about PulseAudio can be found on
17862 <http://www.pulseaudio.org>
17863 Options
17865 server
17867 Connect to a specific PulseAudio server, specified by an IP
17868 address. Default server is used when not provided.
17869 name
17871 Specify the application name PulseAudio will use when showing
17872 active clients, by default it is the "LIBAVFORMAT_IDENT" string.
17873 stream_name
17875 Specify the stream name PulseAudio will use when showing active
17876 streams, by default it is set to the specified output name.
17877 device
17879 Specify the device to use. Default device is used when not
17880 provided. List of output devices can be obtained with command
17881 pactl list sinks.
17882 buffer_size
17884 buffer_duration
17885 Control the size and duration of the PulseAudio buffer. A small
17886 buffer gives more control, but requires more frequent updates.
17887 buffer_size specifies size in bytes while buffer_duration specifies
17889 duration in milliseconds.
17890 When both options are provided then the highest value is used
17892 (duration is recalculated to bytes using stream parameters). If
17893 they are set to 0 (which is default), the device will use the
17894 default PulseAudio duration value. By default PulseAudio set buffer
17895 duration to around 2 seconds.
17896 prebuf
17898 Specify pre-buffering size in bytes. The server does not start with
17899 playback before at least prebuf bytes are available in the buffer.
17900 By default this option is initialized to the same value as
17901 buffer_size or buffer_duration (whichever is bigger).
17902 minreq
17904 Specify minimum request size in bytes. The server does not request
17905 less than minreq bytes from the client, instead waits until the
17906 buffer is free enough to request more bytes at once. It is
17907 recommended to not set this option, which will initialize this to a
17908 value that is deemed sensible by the server.
17909 Examples
17911 Play a file on default device on default server:
17913 ffmpeg -i INPUT -f pulse "stream name"
17915 sdl
17917 SDL (Simple DirectMedia Layer) output device.
17918 "sdl2" can be used as alias for "sdl".
17920 This output device allows one to show a video stream in an SDL window.
17922 Only one SDL window is allowed per application, so you can have only
17923 one instance of this output device in an application.
17924 To enable this output device you need libsdl installed on your system
17926 when configuring your build.
17927 For more information about SDL, check: <http://www.libsdl.org/>
17929 Options
17931 window_title
17933 Set the SDL window title, if not specified default to the filename
17934 specified for the output device.
17935 icon_title
17937 Set the name of the iconified SDL window, if not specified it is
17938 set to the same value of window_title.
17939 window_size
17941 Set the SDL window size, can be a string of the form widthxheight
17942 or a video size abbreviation. If not specified it defaults to the
17943 size of the input video, downscaled according to the aspect ratio.
17944 window_x
17946 window_y
17947 Set the position of the window on the screen.
17948 window_fullscreen
17950 Set fullscreen mode when non-zero value is provided. Default value
17951 is zero.
17952 window_enable_quit
17954 Enable quit action (using window button or keyboard key) when non-
17955 zero value is provided. Default value is 1 (enable quit action)
17956 Interactive commands
17958 The window created by the device can be controlled through the
17960 following interactive commands.
17961 q, ESC
17963 Quit the device immediately.
17964 Examples
17966 The following command shows the ffmpeg output is an SDL window, forcing
17968 its size to the qcif format:
17969 ffmpeg -i INPUT -c:v rawvideo -pix_fmt yuv420p -window_size qcif -f sdl "SDL output"
17971 sndio
17973 sndio audio output device.
17974 v4l2
17976 Video4Linux2 output device.
17977 xv
17979 XV (XVideo) output device.
17980 This output device allows one to show a video stream in a X Window
17982 System window.
17983 Options
17985 display_name
17987 Specify the hardware display name, which determines the display and
17988 communications domain to be used.
17989 The display name or DISPLAY environment variable can be a string in
17991 the format hostname[:number[.screen_number]].
17992 hostname specifies the name of the host machine on which the
17994 display is physically attached. number specifies the number of the
17995 display server on that host machine. screen_number specifies the
17996 screen to be used on that server.
17997 If unspecified, it defaults to the value of the DISPLAY environment
17999 variable.
18000 For example, "dual-headed:0.1" would specify screen 1 of display 0
18002 on the machine named ``dual-headed''.
18003 Check the X11 specification for more detailed information about the
18005 display name format.
18006 window_id
18008 When set to non-zero value then device doesn't create new window,
18009 but uses existing one with provided window_id. By default this
18010 options is set to zero and device creates its own window.
18011 window_size
18013 Set the created window size, can be a string of the form
18014 widthxheight or a video size abbreviation. If not specified it
18015 defaults to the size of the input video. Ignored when window_id is
18016 set.
18017 window_x
18019 window_y
18020 Set the X and Y window offsets for the created window. They are
18021 both set to 0 by default. The values may be ignored by the window
18022 manager. Ignored when window_id is set.
18023 window_title
18025 Set the window title, if not specified default to the filename
18026 specified for the output device. Ignored when window_id is set.
18027 For more information about XVideo see <http://www.x.org/>.
18029 Examples
18031 • Decode, display and encode video input with ffmpeg at the same
18033 time:
18034 ffmpeg -i INPUT OUTPUT -f xv display
18036 • Decode and display the input video to multiple X11 windows:
18038 ffmpeg -i INPUT -f xv normal -vf negate -f xv negated
18040
18042 The audio resampler supports the following named options.
18043 Options may be set by specifying -option value in the FFmpeg tools,
18045 option=value for the aresample filter, by setting the value explicitly
18046 in the "SwrContext" options or using the libavutil/opt.h API for
18047 programmatic use.
18048 uchl, used_chlayout
18050 Set used input channel layout. Default is unset. This option is
18051 only used for special remapping.
18052 isr, in_sample_rate
18054 Set the input sample rate. Default value is 0.
18055 osr, out_sample_rate
18057 Set the output sample rate. Default value is 0.
18058 isf, in_sample_fmt
18060 Specify the input sample format. It is set by default to "none".
18061 osf, out_sample_fmt
18063 Specify the output sample format. It is set by default to "none".
18064 tsf, internal_sample_fmt
18066 Set the internal sample format. Default value is "none". This will
18067 automatically be chosen when it is not explicitly set.
18068 ichl, in_chlayout
18070 ochl, out_chlayout
18071 Set the input/output channel layout.
18072 See the Channel Layout section in the ffmpeg-utils(1) manual for
18074 the required syntax.
18075 clev, center_mix_level
18077 Set the center mix level. It is a value expressed in deciBel, and
18078 must be in the interval [-32,32].
18079 slev, surround_mix_level
18081 Set the surround mix level. It is a value expressed in deciBel, and
18082 must be in the interval [-32,32].
18083 lfe_mix_level
18085 Set LFE mix into non LFE level. It is used when there is a LFE
18086 input but no LFE output. It is a value expressed in deciBel, and
18087 must be in the interval [-32,32].
18088 rmvol, rematrix_volume
18090 Set rematrix volume. Default value is 1.0.
18091 rematrix_maxval
18093 Set maximum output value for rematrixing. This can be used to
18094 prevent clipping vs. preventing volume reduction. A value of 1.0
18095 prevents clipping.
18096 flags, swr_flags
18098 Set flags used by the converter. Default value is 0.
18099 It supports the following individual flags:
18101 res force resampling, this flag forces resampling to be used even
18103 when the input and output sample rates match.
18104 dither_scale
18106 Set the dither scale. Default value is 1.
18107 dither_method
18109 Set dither method. Default value is 0.
18110 Supported values:
18112 rectangular
18114 select rectangular dither
18115 triangular
18117 select triangular dither
18118 triangular_hp
18120 select triangular dither with high pass
18121 lipshitz
18123 select Lipshitz noise shaping dither.
18124 shibata
18126 select Shibata noise shaping dither.
18127 low_shibata
18129 select low Shibata noise shaping dither.
18130 high_shibata
18132 select high Shibata noise shaping dither.
18133 f_weighted
18135 select f-weighted noise shaping dither
18136 modified_e_weighted
18138 select modified-e-weighted noise shaping dither
18139 improved_e_weighted
18141 select improved-e-weighted noise shaping dither
18142 resampler
18144 Set resampling engine. Default value is swr.
18145 Supported values:
18147 swr select the native SW Resampler; filter options precision and
18149 cheby are not applicable in this case.
18150 soxr
18152 select the SoX Resampler (where available); compensation, and
18153 filter options filter_size, phase_shift, exact_rational,
18154 filter_type & kaiser_beta, are not applicable in this case.
18155 filter_size
18157 For swr only, set resampling filter size, default value is 32.
18158 phase_shift
18160 For swr only, set resampling phase shift, default value is 10, and
18161 must be in the interval [0,30].
18162 linear_interp
18164 Use linear interpolation when enabled (the default). Disable it if
18165 you want to preserve speed instead of quality when exact_rational
18166 fails.
18167 exact_rational
18169 For swr only, when enabled, try to use exact phase_count based on
18170 input and output sample rate. However, if it is larger than "1 <<
18171 phase_shift", the phase_count will be "1 << phase_shift" as
18172 fallback. Default is enabled.
18173 cutoff
18175 Set cutoff frequency (swr: 6dB point; soxr: 0dB point) ratio; must
18176 be a float value between 0 and 1. Default value is 0.97 with swr,
18177 and 0.91 with soxr (which, with a sample-rate of 44100, preserves
18178 the entire audio band to 20kHz).
18179 precision
18181 For soxr only, the precision in bits to which the resampled signal
18182 will be calculated. The default value of 20 (which, with suitable
18183 dithering, is appropriate for a destination bit-depth of 16) gives
18184 SoX's 'High Quality'; a value of 28 gives SoX's 'Very High
18185 Quality'.
18186 cheby
18188 For soxr only, selects passband rolloff none (Chebyshev) & higher-
18189 precision approximation for 'irrational' ratios. Default value is
18190 0.
18191 async
18193 For swr only, simple 1 parameter audio sync to timestamps using
18194 stretching, squeezing, filling and trimming. Setting this to 1 will
18195 enable filling and trimming, larger values represent the maximum
18196 amount in samples that the data may be stretched or squeezed for
18197 each second. Default value is 0, thus no compensation is applied
18198 to make the samples match the audio timestamps.
18199 first_pts
18201 For swr only, assume the first pts should be this value. The time
18202 unit is 1 / sample rate. This allows for padding/trimming at the
18203 start of stream. By default, no assumption is made about the first
18204 frame's expected pts, so no padding or trimming is done. For
18205 example, this could be set to 0 to pad the beginning with silence
18206 if an audio stream starts after the video stream or to trim any
18207 samples with a negative pts due to encoder delay.
18208 min_comp
18210 For swr only, set the minimum difference between timestamps and
18211 audio data (in seconds) to trigger stretching/squeezing/filling or
18212 trimming of the data to make it match the timestamps. The default
18213 is that stretching/squeezing/filling and trimming is disabled
18214 (min_comp = "FLT_MAX").
18215 min_hard_comp
18217 For swr only, set the minimum difference between timestamps and
18218 audio data (in seconds) to trigger adding/dropping samples to make
18219 it match the timestamps. This option effectively is a threshold to
18220 select between hard (trim/fill) and soft (squeeze/stretch)
18221 compensation. Note that all compensation is by default disabled
18222 through min_comp. The default is 0.1.
18223 comp_duration
18225 For swr only, set duration (in seconds) over which data is
18226 stretched/squeezed to make it match the timestamps. Must be a non-
18227 negative double float value, default value is 1.0.
18228 max_soft_comp
18230 For swr only, set maximum factor by which data is
18231 stretched/squeezed to make it match the timestamps. Must be a non-
18232 negative double float value, default value is 0.
18233 matrix_encoding
18235 Select matrixed stereo encoding.
18236 It accepts the following values:
18238 none
18240 select none
18241 dolby
18243 select Dolby
18244 dplii
18246 select Dolby Pro Logic II
18247 Default value is "none".
18249 filter_type
18251 For swr only, select resampling filter type. This only affects
18252 resampling operations.
18253 It accepts the following values:
18255 cubic
18257 select cubic
18258 blackman_nuttall
18260 select Blackman Nuttall windowed sinc
18261 kaiser
18263 select Kaiser windowed sinc
18264 kaiser_beta
18266 For swr only, set Kaiser window beta value. Must be a double float
18267 value in the interval [2,16], default value is 9.
18268 output_sample_bits
18270 For swr only, set number of used output sample bits for dithering.
18271 Must be an integer in the interval [0,64], default value is 0,
18272 which means it's not used.
18273
18275 The video scaler supports the following named options.
18276 Options may be set by specifying -option value in the FFmpeg tools,
18278 with a few API-only exceptions noted below. For programmatic use, they
18279 can be set explicitly in the "SwsContext" options or through the
18280 libavutil/opt.h API.
18281 sws_flags
18283 Set the scaler flags. This is also used to set the scaling
18284 algorithm. Only a single algorithm should be selected. Default
18285 value is bicubic.
18286 It accepts the following values:
18288 fast_bilinear
18290 Select fast bilinear scaling algorithm.
18291 bilinear
18293 Select bilinear scaling algorithm.
18294 bicubic
18296 Select bicubic scaling algorithm.
18297 experimental
18299 Select experimental scaling algorithm.
18300 neighbor
18302 Select nearest neighbor rescaling algorithm.
18303 area
18305 Select averaging area rescaling algorithm.
18306 bicublin
18308 Select bicubic scaling algorithm for the luma component,
18309 bilinear for chroma components.
18310 gauss
18312 Select Gaussian rescaling algorithm.
18313 sinc
18315 Select sinc rescaling algorithm.
18316 lanczos
18318 Select Lanczos rescaling algorithm. The default width (alpha)
18319 is 3 and can be changed by setting "param0".
18320 spline
18322 Select natural bicubic spline rescaling algorithm.
18323 print_info
18325 Enable printing/debug logging.
18326 accurate_rnd
18328 Enable accurate rounding.
18329 full_chroma_int
18331 Enable full chroma interpolation.
18332 full_chroma_inp
18334 Select full chroma input.
18335 bitexact
18337 Enable bitexact output.
18338 srcw (API only)
18340 Set source width.
18341 srch (API only)
18343 Set source height.
18344 dstw (API only)
18346 Set destination width.
18347 dsth (API only)
18349 Set destination height.
18350 src_format (API only)
18352 Set source pixel format (must be expressed as an integer).
18353 dst_format (API only)
18355 Set destination pixel format (must be expressed as an integer).
18356 src_range (boolean)
18358 If value is set to 1, indicates source is full range. Default value
18359 is 0, which indicates source is limited range.
18360 dst_range (boolean)
18362 If value is set to 1, enable full range for destination. Default
18363 value is 0, which enables limited range.
18364 param0, param1
18366 Set scaling algorithm parameters. The specified values are specific
18367 of some scaling algorithms and ignored by others. The specified
18368 values are floating point number values.
18369 sws_dither
18371 Set the dithering algorithm. Accepts one of the following values.
18372 Default value is auto.
18373 auto
18375 automatic choice
18376 none
18378 no dithering
18379 bayer
18381 bayer dither
18382 ed error diffusion dither
18384 a_dither
18386 arithmetic dither, based using addition
18387 x_dither
18389 arithmetic dither, based using xor (more random/less apparent
18390 patterning that a_dither).
18391 alphablend
18393 Set the alpha blending to use when the input has alpha but the
18394 output does not. Default value is none.
18395 uniform_color
18397 Blend onto a uniform background color
18398 checkerboard
18400 Blend onto a checkerboard
18401 none
18403 No blending
18404
18406 Filtering in FFmpeg is enabled through the libavfilter library.
18407 In libavfilter, a filter can have multiple inputs and multiple outputs.
18409 To illustrate the sorts of things that are possible, we consider the
18410 following filtergraph.
18411 [main]
18413 input --> split ---------------------> overlay --> output
18414 | ^
18415 |[tmp] [flip]|
18416 +-----> crop --> vflip -------+
18417 This filtergraph splits the input stream in two streams, then sends one
18419 stream through the crop filter and the vflip filter, before merging it
18420 back with the other stream by overlaying it on top. You can use the
18421 following command to achieve this:
18422 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
18424 The result will be that the top half of the video is mirrored onto the
18426 bottom half of the output video.
18427 Filters in the same linear chain are separated by commas, and distinct
18429 linear chains of filters are separated by semicolons. In our example,
18430 crop,vflip are in one linear chain, split and overlay are separately in
18431 another. The points where the linear chains join are labelled by names
18432 enclosed in square brackets. In the example, the split filter generates
18433 two outputs that are associated to the labels [main] and [tmp].
18434 The stream sent to the second output of split, labelled as [tmp], is
18436 processed through the crop filter, which crops away the lower half part
18437 of the video, and then vertically flipped. The overlay filter takes in
18438 input the first unchanged output of the split filter (which was
18439 labelled as [main]), and overlay on its lower half the output generated
18440 by the crop,vflip filterchain.
18441 Some filters take in input a list of parameters: they are specified
18443 after the filter name and an equal sign, and are separated from each
18444 other by a colon.
18445 There exist so-called source filters that do not have an audio/video
18447 input, and sink filters that will not have audio/video output.
18448
18450 The graph2dot program included in the FFmpeg tools directory can be
18451 used to parse a filtergraph description and issue a corresponding
18452 textual representation in the dot language.
18453 Invoke the command:
18455 graph2dot -h
18457 to see how to use graph2dot.
18459 You can then pass the dot description to the dot program (from the
18461 graphviz suite of programs) and obtain a graphical representation of
18462 the filtergraph.
18463 For example the sequence of commands:
18465 echo <GRAPH_DESCRIPTION> | \
18467 tools/graph2dot -o graph.tmp && \
18468 dot -Tpng graph.tmp -o graph.png && \
18469 display graph.png
18470 can be used to create and display an image representing the graph
18472 described by the GRAPH_DESCRIPTION string. Note that this string must
18473 be a complete self-contained graph, with its inputs and outputs
18474 explicitly defined. For example if your command line is of the form:
18475 ffmpeg -i infile -vf scale=640:360 outfile
18477 your GRAPH_DESCRIPTION string will need to be of the form:
18479 nullsrc,scale=640:360,nullsink
18481 you may also need to set the nullsrc parameters and add a format filter
18483 in order to simulate a specific input file.
18484
18486 A filtergraph is a directed graph of connected filters. It can contain
18487 cycles, and there can be multiple links between a pair of filters. Each
18488 link has one input pad on one side connecting it to one filter from
18489 which it takes its input, and one output pad on the other side
18490 connecting it to one filter accepting its output.
18491 Each filter in a filtergraph is an instance of a filter class
18493 registered in the application, which defines the features and the
18494 number of input and output pads of the filter.
18495 A filter with no input pads is called a "source", and a filter with no
18497 output pads is called a "sink".
18498 Filtergraph syntax
18500 A filtergraph has a textual representation, which is recognized by the
18501 -filter/-vf/-af and -filter_complex options in ffmpeg and -vf/-af in
18502 ffplay, and by the avfilter_graph_parse_ptr() function defined in
18503 libavfilter/avfilter.h.
18504 A filterchain consists of a sequence of connected filters, each one
18506 connected to the previous one in the sequence. A filterchain is
18507 represented by a list of ","-separated filter descriptions.
18508 A filtergraph consists of a sequence of filterchains. A sequence of
18510 filterchains is represented by a list of ";"-separated filterchain
18511 descriptions.
18512 A filter is represented by a string of the form:
18514 [in_link_1]...[in_link_N]filter_name@id=arguments[out_link_1]...[out_link_M]
18515 filter_name is the name of the filter class of which the described
18517 filter is an instance of, and has to be the name of one of the filter
18518 classes registered in the program optionally followed by "@id". The
18519 name of the filter class is optionally followed by a string
18520 "=arguments".
18521 arguments is a string which contains the parameters used to initialize
18523 the filter instance. It may have one of two forms:
18524 • A ':'-separated list of key=value pairs.
18526 • A ':'-separated list of value. In this case, the keys are assumed
18528 to be the option names in the order they are declared. E.g. the
18529 "fade" filter declares three options in this order -- type,
18530 start_frame and nb_frames. Then the parameter list in:0:30 means
18531 that the value in is assigned to the option type, 0 to start_frame
18532 and 30 to nb_frames.
18533 • A ':'-separated list of mixed direct value and long key=value
18535 pairs. The direct value must precede the key=value pairs, and
18536 follow the same constraints order of the previous point. The
18537 following key=value pairs can be set in any preferred order.
18538 If the option value itself is a list of items (e.g. the "format" filter
18540 takes a list of pixel formats), the items in the list are usually
18541 separated by |.
18542 The list of arguments can be quoted using the character ' as initial
18544 and ending mark, and the character \ for escaping the characters within
18545 the quoted text; otherwise the argument string is considered terminated
18546 when the next special character (belonging to the set []=;,) is
18547 encountered.
18548 A special syntax implemented in the ffmpeg CLI tool allows loading
18550 option values from files. This is done be prepending a slash '/' to the
18551 option name, then the supplied value is interpreted as a path from
18552 which the actual value is loaded. E.g.
18553 ffmpeg -i <INPUT> -vf drawtext=/text=/tmp/some_text <OUTPUT>
18555 will load the text to be drawn from /tmp/some_text. API users wishing
18557 to implement a similar feature should use the
18558 "avfilter_graph_segment_*()" functions together with custom IO code.
18559 The name and arguments of the filter are optionally preceded and
18561 followed by a list of link labels. A link label allows one to name a
18562 link and associate it to a filter output or input pad. The preceding
18563 labels in_link_1 ... in_link_N, are associated to the filter input
18564 pads, the following labels out_link_1 ... out_link_M, are associated to
18565 the output pads.
18566 When two link labels with the same name are found in the filtergraph, a
18568 link between the corresponding input and output pad is created.
18569 If an output pad is not labelled, it is linked by default to the first
18571 unlabelled input pad of the next filter in the filterchain. For
18572 example in the filterchain
18573 nullsrc, split[L1], [L2]overlay, nullsink
18575 the split filter instance has two output pads, and the overlay filter
18577 instance two input pads. The first output pad of split is labelled
18578 "L1", the first input pad of overlay is labelled "L2", and the second
18579 output pad of split is linked to the second input pad of overlay, which
18580 are both unlabelled.
18581 In a filter description, if the input label of the first filter is not
18583 specified, "in" is assumed; if the output label of the last filter is
18584 not specified, "out" is assumed.
18585 In a complete filterchain all the unlabelled filter input and output
18587 pads must be connected. A filtergraph is considered valid if all the
18588 filter input and output pads of all the filterchains are connected.
18589 Libavfilter will automatically insert scale filters where format
18591 conversion is required. It is possible to specify swscale flags for
18592 those automatically inserted scalers by prepending "sws_flags=flags;"
18593 to the filtergraph description.
18594 Here is a BNF description of the filtergraph syntax:
18596 <NAME> ::= sequence of alphanumeric characters and '_'
18598 <FILTER_NAME> ::= <NAME>["@"<NAME>]
18599 <LINKLABEL> ::= "[" <NAME> "]"
18600 <LINKLABELS> ::= <LINKLABEL> [<LINKLABELS>]
18601 <FILTER_ARGUMENTS> ::= sequence of chars (possibly quoted)
18602 <FILTER> ::= [<LINKLABELS>] <FILTER_NAME> ["=" <FILTER_ARGUMENTS>] [<LINKLABELS>]
18603 <FILTERCHAIN> ::= <FILTER> [,<FILTERCHAIN>]
18604 <FILTERGRAPH> ::= [sws_flags=<flags>;] <FILTERCHAIN> [;<FILTERGRAPH>]
18605 Notes on filtergraph escaping
18607 Filtergraph description composition entails several levels of escaping.
18608 See the "Quoting and escaping" section in the ffmpeg-utils(1) manual
18609 for more information about the employed escaping procedure.
18610 A first level escaping affects the content of each filter option value,
18612 which may contain the special character ":" used to separate values, or
18613 one of the escaping characters "\'".
18614 A second level escaping affects the whole filter description, which may
18616 contain the escaping characters "\'" or the special characters "[],;"
18617 used by the filtergraph description.
18618 Finally, when you specify a filtergraph on a shell commandline, you
18620 need to perform a third level escaping for the shell special characters
18621 contained within it.
18622 For example, consider the following string to be embedded in the
18624 drawtext filter description text value:
18625 this is a 'string': may contain one, or more, special characters
18627 This string contains the "'" special escaping character, and the ":"
18629 special character, so it needs to be escaped in this way:
18630 text=this is a \'string\'\: may contain one, or more, special characters
18632 A second level of escaping is required when embedding the filter
18634 description in a filtergraph description, in order to escape all the
18635 filtergraph special characters. Thus the example above becomes:
18636 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
18638 (note that in addition to the "\'" escaping special characters, also
18640 "," needs to be escaped).
18641 Finally an additional level of escaping is needed when writing the
18643 filtergraph description in a shell command, which depends on the
18644 escaping rules of the adopted shell. For example, assuming that "\" is
18645 special and needs to be escaped with another "\", the previous string
18646 will finally result in:
18647 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
18649
18651 Some filters support a generic enable option. For the filters
18652 supporting timeline editing, this option can be set to an expression
18653 which is evaluated before sending a frame to the filter. If the
18654 evaluation is non-zero, the filter will be enabled, otherwise the frame
18655 will be sent unchanged to the next filter in the filtergraph.
18656 The expression accepts the following values:
18658 t timestamp expressed in seconds, NAN if the input timestamp is
18660 unknown
18661 n sequential number of the input frame, starting from 0
18663 pos the position in the file of the input frame, NAN if unknown
18665 w
18667 h width and height of the input frame if video
18668 Additionally, these filters support an enable command that can be used
18670 to re-define the expression.
18671 Like any other filtering option, the enable option follows the same
18673 rules.
18674 For example, to enable a blur filter (smartblur) from 10 seconds to 3
18676 minutes, and a curves filter starting at 3 seconds:
18677 smartblur = enable='between(t,10,3*60)',
18679 curves = enable='gte(t,3)' : preset=cross_process
18680 See "ffmpeg -filters" to view which filters have timeline support.
18682
18684 Some options can be changed during the operation of the filter using a
18685 command. These options are marked 'T' on the output of ffmpeg -h
18686 filter=<name of filter>. The name of the command is the name of the
18687 option and the argument is the new value.
18688
18690 Some filters with several inputs support a common set of options.
18691 These options can only be set by name, not with the short notation.
18692 eof_action
18694 The action to take when EOF is encountered on the secondary input;
18695 it accepts one of the following values:
18696 repeat
18698 Repeat the last frame (the default).
18699 endall
18701 End both streams.
18702 pass
18704 Pass the main input through.
18705 shortest
18707 If set to 1, force the output to terminate when the shortest input
18708 terminates. Default value is 0.
18709 repeatlast
18711 If set to 1, force the filter to extend the last frame of secondary
18712 streams until the end of the primary stream. A value of 0 disables
18713 this behavior. Default value is 1.
18714 ts_sync_mode
18716 How strictly to sync streams based on secondary input timestamps;
18717 it accepts one of the following values:
18718 default
18720 Frame from secondary input with the nearest lower or equal
18721 timestamp to the primary input frame.
18722 nearest
18724 Frame from secondary input with the absolute nearest timestamp
18725 to the primary input frame.
18726
18728 When you configure your FFmpeg build, you can disable any of the
18729 existing filters using "--disable-filters". The configure output will
18730 show the audio filters included in your build.
18731 Below is a description of the currently available audio filters.
18733 acompressor
18735 A compressor is mainly used to reduce the dynamic range of a signal.
18736 Especially modern music is mostly compressed at a high ratio to improve
18737 the overall loudness. It's done to get the highest attention of a
18738 listener, "fatten" the sound and bring more "power" to the track. If a
18739 signal is compressed too much it may sound dull or "dead" afterwards or
18740 it may start to "pump" (which could be a powerful effect but can also
18741 destroy a track completely). The right compression is the key to reach
18742 a professional sound and is the high art of mixing and mastering.
18743 Because of its complex settings it may take a long time to get the
18744 right feeling for this kind of effect.
18745 Compression is done by detecting the volume above a chosen level
18747 "threshold" and dividing it by the factor set with "ratio". So if you
18748 set the threshold to -12dB and your signal reaches -6dB a ratio of 2:1
18749 will result in a signal at -9dB. Because an exact manipulation of the
18750 signal would cause distortion of the waveform the reduction can be
18751 levelled over the time. This is done by setting "Attack" and "Release".
18752 "attack" determines how long the signal has to rise above the threshold
18753 before any reduction will occur and "release" sets the time the signal
18754 has to fall below the threshold to reduce the reduction again. Shorter
18755 signals than the chosen attack time will be left untouched. The
18756 overall reduction of the signal can be made up afterwards with the
18757 "makeup" setting. So compressing the peaks of a signal about 6dB and
18758 raising the makeup to this level results in a signal twice as loud than
18759 the source. To gain a softer entry in the compression the "knee"
18760 flattens the hard edge at the threshold in the range of the chosen
18761 decibels.
18762 The filter accepts the following options:
18764 level_in
18766 Set input gain. Default is 1. Range is between 0.015625 and 64.
18767 mode
18769 Set mode of compressor operation. Can be "upward" or "downward".
18770 Default is "downward".
18771 threshold
18773 If a signal of stream rises above this level it will affect the
18774 gain reduction. By default it is 0.125. Range is between
18775 0.00097563 and 1.
18776 ratio
18778 Set a ratio by which the signal is reduced. 1:2 means that if the
18779 level rose 4dB above the threshold, it will be only 2dB above after
18780 the reduction. Default is 2. Range is between 1 and 20.
18781 attack
18783 Amount of milliseconds the signal has to rise above the threshold
18784 before gain reduction starts. Default is 20. Range is between 0.01
18785 and 2000.
18786 release
18788 Amount of milliseconds the signal has to fall below the threshold
18789 before reduction is decreased again. Default is 250. Range is
18790 between 0.01 and 9000.
18791 makeup
18793 Set the amount by how much signal will be amplified after
18794 processing. Default is 1. Range is from 1 to 64.
18795 knee
18797 Curve the sharp knee around the threshold to enter gain reduction
18798 more softly. Default is 2.82843. Range is between 1 and 8.
18799 link
18801 Choose if the "average" level between all channels of input stream
18802 or the louder("maximum") channel of input stream affects the
18803 reduction. Default is "average".
18804 detection
18806 Should the exact signal be taken in case of "peak" or an RMS one in
18807 case of "rms". Default is "rms" which is mostly smoother.
18808 mix How much to use compressed signal in output. Default is 1. Range
18810 is between 0 and 1.
18811 Commands
18813 This filter supports the all above options as commands.
18815 acontrast
18817 Simple audio dynamic range compression/expansion filter.
18818 The filter accepts the following options:
18820 contrast
18822 Set contrast. Default is 33. Allowed range is between 0 and 100.
18823 acopy
18825 Copy the input audio source unchanged to the output. This is mainly
18826 useful for testing purposes.
18827 acrossfade
18829 Apply cross fade from one input audio stream to another input audio
18830 stream. The cross fade is applied for specified duration near the end
18831 of first stream.
18832 The filter accepts the following options:
18834 nb_samples, ns
18836 Specify the number of samples for which the cross fade effect has
18837 to last. At the end of the cross fade effect the first input audio
18838 will be completely silent. Default is 44100.
18839 duration, d
18841 Specify the duration of the cross fade effect. See the Time
18842 duration section in the ffmpeg-utils(1) manual for the accepted
18843 syntax. By default the duration is determined by nb_samples. If
18844 set this option is used instead of nb_samples.
18845 overlap, o
18847 Should first stream end overlap with second stream start. Default
18848 is enabled.
18849 curve1
18851 Set curve for cross fade transition for first stream.
18852 curve2
18854 Set curve for cross fade transition for second stream.
18855 For description of available curve types see afade filter
18857 description.
18858 Examples
18860 • Cross fade from one input to another:
18862 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
18864 • Cross fade from one input to another but without overlapping:
18866 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
18868 acrossover
18870 Split audio stream into several bands.
18871 This filter splits audio stream into two or more frequency ranges.
18873 Summing all streams back will give flat output.
18874 The filter accepts the following options:
18876 split
18878 Set split frequencies. Those must be positive and increasing.
18879 order
18881 Set filter order for each band split. This controls filter roll-off
18882 or steepness of filter transfer function. Available values are:
18883 2nd 12 dB per octave.
18885 4th 24 dB per octave.
18887 6th 36 dB per octave.
18889 8th 48 dB per octave.
18891 10th
18893 60 dB per octave.
18894 12th
18896 72 dB per octave.
18897 14th
18899 84 dB per octave.
18900 16th
18902 96 dB per octave.
18903 18th
18905 108 dB per octave.
18906 20th
18908 120 dB per octave.
18909 Default is 4th.
18911 level
18913 Set input gain level. Allowed range is from 0 to 1. Default value
18914 is 1.
18915 gains
18917 Set output gain for each band. Default value is 1 for all bands.
18918 precision
18920 Set which precision to use when processing samples.
18921 auto
18923 Auto pick internal sample format depending on other filters.
18924 float
18926 Always use single-floating point precision sample format.
18927 double
18929 Always use double-floating point precision sample format.
18930 Default value is "auto".
18932 Examples
18934 • Split input audio stream into two bands (low and high) with split
18936 frequency of 1500 Hz, each band will be in separate stream:
18937 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
18939 • Same as above, but with higher filter order:
18941 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
18943 • Same as above, but also with additional middle band (frequencies
18945 between 1500 and 8000):
18946 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
18948 acrusher
18950 Reduce audio bit resolution.
18951 This filter is bit crusher with enhanced functionality. A bit crusher
18953 is used to audibly reduce number of bits an audio signal is sampled
18954 with. This doesn't change the bit depth at all, it just produces the
18955 effect. Material reduced in bit depth sounds more harsh and "digital".
18956 This filter is able to even round to continuous values instead of
18957 discrete bit depths. Additionally it has a D/C offset which results in
18958 different crushing of the lower and the upper half of the signal. An
18959 Anti-Aliasing setting is able to produce "softer" crushing sounds.
18960 Another feature of this filter is the logarithmic mode. This setting
18962 switches from linear distances between bits to logarithmic ones. The
18963 result is a much more "natural" sounding crusher which doesn't gate low
18964 signals for example. The human ear has a logarithmic perception, so
18965 this kind of crushing is much more pleasant. Logarithmic crushing is
18966 also able to get anti-aliased.
18967 The filter accepts the following options:
18969 level_in
18971 Set level in.
18972 level_out
18974 Set level out.
18975 bits
18977 Set bit reduction.
18978 mix Set mixing amount.
18980 mode
18982 Can be linear: "lin" or logarithmic: "log".
18983 dc Set DC.
18985 aa Set anti-aliasing.
18987 samples
18989 Set sample reduction.
18990 lfo Enable LFO. By default disabled.
18992 lforange
18994 Set LFO range.
18995 lforate
18997 Set LFO rate.
18998 Commands
19000 This filter supports the all above options as commands.
19002 acue
19004 Delay audio filtering until a given wallclock timestamp. See the cue
19005 filter.
19006 adeclick
19008 Remove impulsive noise from input audio.
19009 Samples detected as impulsive noise are replaced by interpolated
19011 samples using autoregressive modelling.
19012 window, w
19014 Set window size, in milliseconds. Allowed range is from 10 to 100.
19015 Default value is 55 milliseconds. This sets size of window which
19016 will be processed at once.
19017 overlap, o
19019 Set window overlap, in percentage of window size. Allowed range is
19020 from 50 to 95. Default value is 75 percent. Setting this to a very
19021 high value increases impulsive noise removal but makes whole
19022 process much slower.
19023 arorder, a
19025 Set autoregression order, in percentage of window size. Allowed
19026 range is from 0 to 25. Default value is 2 percent. This option also
19027 controls quality of interpolated samples using neighbour good
19028 samples.
19029 threshold, t
19031 Set threshold value. Allowed range is from 1 to 100. Default value
19032 is 2. This controls the strength of impulsive noise which is going
19033 to be removed. The lower value, the more samples will be detected
19034 as impulsive noise.
19035 burst, b
19037 Set burst fusion, in percentage of window size. Allowed range is 0
19038 to 10. Default value is 2. If any two samples detected as noise
19039 are spaced less than this value then any sample between those two
19040 samples will be also detected as noise.
19041 method, m
19043 Set overlap method.
19044 It accepts the following values:
19046 add, a
19048 Select overlap-add method. Even not interpolated samples are
19049 slightly changed with this method.
19050 save, s
19052 Select overlap-save method. Not interpolated samples remain
19053 unchanged.
19054 Default value is "a".
19056 adeclip
19058 Remove clipped samples from input audio.
19059 Samples detected as clipped are replaced by interpolated samples using
19061 autoregressive modelling.
19062 window, w
19064 Set window size, in milliseconds. Allowed range is from 10 to 100.
19065 Default value is 55 milliseconds. This sets size of window which
19066 will be processed at once.
19067 overlap, o
19069 Set window overlap, in percentage of window size. Allowed range is
19070 from 50 to 95. Default value is 75 percent.
19071 arorder, a
19073 Set autoregression order, in percentage of window size. Allowed
19074 range is from 0 to 25. Default value is 8 percent. This option also
19075 controls quality of interpolated samples using neighbour good
19076 samples.
19077 threshold, t
19079 Set threshold value. Allowed range is from 1 to 100. Default value
19080 is 10. Higher values make clip detection less aggressive.
19081 hsize, n
19083 Set size of histogram used to detect clips. Allowed range is from
19084 100 to 9999. Default value is 1000. Higher values make clip
19085 detection less aggressive.
19086 method, m
19088 Set overlap method.
19089 It accepts the following values:
19091 add, a
19093 Select overlap-add method. Even not interpolated samples are
19094 slightly changed with this method.
19095 save, s
19097 Select overlap-save method. Not interpolated samples remain
19098 unchanged.
19099 Default value is "a".
19101 adecorrelate
19103 Apply decorrelation to input audio stream.
19104 The filter accepts the following options:
19106 stages
19108 Set decorrelation stages of filtering. Allowed range is from 1 to
19109 16. Default value is 6.
19110 seed
19112 Set random seed used for setting delay in samples across channels.
19113 adelay
19115 Delay one or more audio channels.
19116 Samples in delayed channel are filled with silence.
19118 The filter accepts the following option:
19120 delays
19122 Set list of delays in milliseconds for each channel separated by
19123 '|'. Unused delays will be silently ignored. If number of given
19124 delays is smaller than number of channels all remaining channels
19125 will not be delayed. If you want to delay exact number of samples,
19126 append 'S' to number. If you want instead to delay in seconds,
19127 append 's' to number.
19128 all Use last set delay for all remaining channels. By default is
19130 disabled. This option if enabled changes how option "delays" is
19131 interpreted.
19132 Examples
19134 • Delay first channel by 1.5 seconds, the third channel by 0.5
19136 seconds and leave the second channel (and any other channels that
19137 may be present) unchanged.
19138 adelay=1500|0|500
19140 • Delay second channel by 500 samples, the third channel by 700
19142 samples and leave the first channel (and any other channels that
19143 may be present) unchanged.
19144 adelay=0|500S|700S
19146 • Delay all channels by same number of samples:
19148 adelay=delays=64S:all=1
19150 adenorm
19152 Remedy denormals in audio by adding extremely low-level noise.
19153 This filter shall be placed before any filter that can produce
19155 denormals.
19156 A description of the accepted parameters follows.
19158 level
19160 Set level of added noise in dB. Default is -351. Allowed range is
19161 from -451 to -90.
19162 type
19164 Set type of added noise.
19165 dc Add DC signal.
19167 ac Add AC signal.
19169 square
19171 Add square signal.
19172 pulse
19174 Add pulse signal.
19175 Default is "dc".
19177 Commands
19179 This filter supports the all above options as commands.
19181 aderivative, aintegral
19183 Compute derivative/integral of audio stream.
19184 Applying both filters one after another produces original audio.
19186 adrc
19188 Apply spectral dynamic range controller filter to input audio stream.
19189 A description of the accepted options follows.
19191 transfer
19193 Set the transfer expression.
19194 The expression can contain the following constants:
19196 ch current channel number
19198 sn current sample number
19200 nb_channels
19202 number of channels
19203 t timestamp expressed in seconds
19205 sr sample rate
19207 p current frequency power value, in dB
19209 f current frequency in Hz
19211 Default value is "p".
19213 attack
19215 Set the attack in milliseconds. Default is 50 milliseconds.
19216 Allowed range is from 1 to 1000 milliseconds.
19217 release
19219 Set the release in milliseconds. Default is 100 milliseconds.
19220 Allowed range is from 5 to 2000 milliseconds.
19221 channels
19223 Set which channels to filter, by default "all" channels in audio
19224 stream are filtered.
19225 Commands
19227 This filter supports the all above options as commands.
19229 Examples
19231 • Apply spectral compression to all frequencies with threshold of -50
19233 dB and 1:6 ratio:
19234 adrc=transfer='if(gt(p,-50),-50+(p-(-50))/6,p)':attack=50:release=100
19236 • Similar to above but with 1:2 ratio and filtering only front center
19238 channel:
19239 adrc=transfer='if(gt(p,-50),-50+(p-(-50))/2,p)':attack=50:release=100:channels=FC
19241 • Apply spectral noise gate to all frequencies with threshold of -85
19243 dB and with short attack time and short release time:
19244 adrc=transfer='if(lte(p,-85),p-800,p)':attack=1:release=5
19246 • Apply spectral expansion to all frequencies with threshold of -10
19248 dB and 1:2 ratio:
19249 adrc=transfer='if(lt(p,-10),-10+(p-(-10))*2,p)':attack=50:release=100
19251 • Apply limiter to max -60 dB to all frequencies, with attack of 2 ms
19253 and release of 10 ms:
19254 adrc=transfer='min(p,-60)':attack=2:release=10
19256 adynamicequalizer
19258 Apply dynamic equalization to input audio stream.
19259 A description of the accepted options follows.
19261 threshold
19263 Set the detection threshold used to trigger equalization.
19264 Threshold detection is using bandpass filter. Default value is 0.
19265 Allowed range is from 0 to 100.
19266 dfrequency
19268 Set the detection frequency in Hz used for bandpass filter used to
19269 trigger equalization. Default value is 1000 Hz. Allowed range is
19270 between 2 and 1000000 Hz.
19271 dqfactor
19273 Set the detection resonance factor for bandpass filter used to
19274 trigger equalization. Default value is 1. Allowed range is from
19275 0.001 to 1000.
19276 tfrequency
19278 Set the target frequency of equalization filter. Default value is
19279 1000 Hz. Allowed range is between 2 and 1000000 Hz.
19280 tqfactor
19282 Set the target resonance factor for target equalization filter.
19283 Default value is 1. Allowed range is from 0.001 to 1000.
19284 attack
19286 Set the amount of milliseconds the signal from detection has to
19287 rise above the detection threshold before equalization starts.
19288 Default is 20. Allowed range is between 1 and 2000.
19289 release
19291 Set the amount of milliseconds the signal from detection has to
19292 fall below the detection threshold before equalization ends.
19293 Default is 200. Allowed range is between 1 and 2000.
19294 ratio
19296 Set the ratio by which the equalization gain is raised. Default is
19297 1. Allowed range is between 0 and 30.
19298 makeup
19300 Set the makeup offset by which the equalization gain is raised.
19301 Default is 0. Allowed range is between 0 and 100.
19302 range
19304 Set the max allowed cut/boost amount. Default is 50. Allowed range
19305 is from 1 to 200.
19306 mode
19308 Set the mode of filter operation, can be one of the following:
19309 listen
19311 Output only isolated bandpass signal.
19312 cut Cut frequencies above detection threshold.
19314 boost
19316 Boost frequencies bellow detection threshold.
19317 Default mode is cut.
19319 tftype
19321 Set the type of target filter, can be one of the following:
19322 bell
19324 lowshelf
19325 highshelf
19326 Default type is bell.
19328 direction
19330 Set processing direction relative to threshold.
19331 downward
19333 Boost/Cut if threshold is higher/lower than detected volume.
19334 upward
19336 Boost/Cut if threshold is lower/higher than detected volume.
19337 Default direction is downward.
19339 auto
19341 Automatically gather threshold from detection filter. By default is
19342 disabled. This option is useful to detect threshold in certain
19343 time frame of input audio stream, in such case option value is
19344 changed at runtime.
19345 Available values are:
19347 disabled
19349 Disable using automatically gathered threshold value.
19350 off Stop picking threshold value.
19352 on Start picking threshold value.
19354 Commands
19356 This filter supports the all above options as commands.
19358 adynamicsmooth
19360 Apply dynamic smoothing to input audio stream.
19361 A description of the accepted options follows.
19363 sensitivity
19365 Set an amount of sensitivity to frequency fluctations. Default is
19366 2. Allowed range is from 0 to 1e+06.
19367 basefreq
19369 Set a base frequency for smoothing. Default value is 22050.
19370 Allowed range is from 2 to 1e+06.
19371 Commands
19373 This filter supports the all above options as commands.
19375 aecho
19377 Apply echoing to the input audio.
19378 Echoes are reflected sound and can occur naturally amongst mountains
19380 (and sometimes large buildings) when talking or shouting; digital echo
19381 effects emulate this behaviour and are often used to help fill out the
19382 sound of a single instrument or vocal. The time difference between the
19383 original signal and the reflection is the "delay", and the loudness of
19384 the reflected signal is the "decay". Multiple echoes can have
19385 different delays and decays.
19386 A description of the accepted parameters follows.
19388 in_gain
19390 Set input gain of reflected signal. Default is 0.6.
19391 out_gain
19393 Set output gain of reflected signal. Default is 0.3.
19394 delays
19396 Set list of time intervals in milliseconds between original signal
19397 and reflections separated by '|'. Allowed range for each "delay" is
19398 "(0 - 90000.0]". Default is 1000.
19399 decays
19401 Set list of loudness of reflected signals separated by '|'.
19402 Allowed range for each "decay" is "(0 - 1.0]". Default is 0.5.
19403 Examples
19405 • Make it sound as if there are twice as many instruments as are
19407 actually playing:
19408 aecho=0.8:0.88:60:0.4
19410 • If delay is very short, then it sounds like a (metallic) robot
19412 playing music:
19413 aecho=0.8:0.88:6:0.4
19415 • A longer delay will sound like an open air concert in the
19417 mountains:
19418 aecho=0.8:0.9:1000:0.3
19420 • Same as above but with one more mountain:
19422 aecho=0.8:0.9:1000|1800:0.3|0.25
19424 aemphasis
19426 Audio emphasis filter creates or restores material directly taken from
19427 LPs or emphased CDs with different filter curves. E.g. to store music
19428 on vinyl the signal has to be altered by a filter first to even out the
19429 disadvantages of this recording medium. Once the material is played
19430 back the inverse filter has to be applied to restore the distortion of
19431 the frequency response.
19432 The filter accepts the following options:
19434 level_in
19436 Set input gain.
19437 level_out
19439 Set output gain.
19440 mode
19442 Set filter mode. For restoring material use "reproduction" mode,
19443 otherwise use "production" mode. Default is "reproduction" mode.
19444 type
19446 Set filter type. Selects medium. Can be one of the following:
19447 col select Columbia.
19449 emi select EMI.
19451 bsi select BSI (78RPM).
19453 riaa
19455 select RIAA.
19456 cd select Compact Disc (CD).
19458 50fm
19460 select 50µs (FM).
19461 75fm
19463 select 75µs (FM).
19464 50kf
19466 select 50µs (FM-KF).
19467 75kf
19469 select 75µs (FM-KF).
19470 Commands
19472 This filter supports the all above options as commands.
19474 aeval
19476 Modify an audio signal according to the specified expressions.
19477 This filter accepts one or more expressions (one for each channel),
19479 which are evaluated and used to modify a corresponding audio signal.
19480 It accepts the following parameters:
19482 exprs
19484 Set the '|'-separated expressions list for each separate channel.
19485 If the number of input channels is greater than the number of
19486 expressions, the last specified expression is used for the
19487 remaining output channels.
19488 channel_layout, c
19490 Set output channel layout. If not specified, the channel layout is
19491 specified by the number of expressions. If set to same, it will use
19492 by default the same input channel layout.
19493 Each expression in exprs can contain the following constants and
19495 functions:
19496 ch channel number of the current expression
19498 n number of the evaluated sample, starting from 0
19500 s sample rate
19502 t time of the evaluated sample expressed in seconds
19504 nb_in_channels
19506 nb_out_channels
19507 input and output number of channels
19508 val(CH)
19510 the value of input channel with number CH
19511 Note: this filter is slow. For faster processing you should use a
19513 dedicated filter.
19514 Examples
19516 • Half volume:
19518 aeval=val(ch)/2:c=same
19520 • Invert phase of the second channel:
19522 aeval=val(0)|-val(1)
19524 aexciter
19526 An exciter is used to produce high sound that is not present in the
19527 original signal. This is done by creating harmonic distortions of the
19528 signal which are restricted in range and added to the original signal.
19529 An Exciter raises the upper end of an audio signal without simply
19530 raising the higher frequencies like an equalizer would do to create a
19531 more "crisp" or "brilliant" sound.
19532 The filter accepts the following options:
19534 level_in
19536 Set input level prior processing of signal. Allowed range is from
19537 0 to 64. Default value is 1.
19538 level_out
19540 Set output level after processing of signal. Allowed range is from
19541 0 to 64. Default value is 1.
19542 amount
19544 Set the amount of harmonics added to original signal. Allowed
19545 range is from 0 to 64. Default value is 1.
19546 drive
19548 Set the amount of newly created harmonics. Allowed range is from
19549 0.1 to 10. Default value is 8.5.
19550 blend
19552 Set the octave of newly created harmonics. Allowed range is from
19553 -10 to 10. Default value is 0.
19554 freq
19556 Set the lower frequency limit of producing harmonics in Hz.
19557 Allowed range is from 2000 to 12000 Hz. Default is 7500 Hz.
19558 ceil
19560 Set the upper frequency limit of producing harmonics. Allowed
19561 range is from 9999 to 20000 Hz. If value is lower than 10000 Hz no
19562 limit is applied.
19563 listen
19565 Mute the original signal and output only added harmonics. By
19566 default is disabled.
19567 Commands
19569 This filter supports the all above options as commands.
19571 afade
19573 Apply fade-in/out effect to input audio.
19574 A description of the accepted parameters follows.
19576 type, t
19578 Specify the effect type, can be either "in" for fade-in, or "out"
19579 for a fade-out effect. Default is "in".
19580 start_sample, ss
19582 Specify the number of the start sample for starting to apply the
19583 fade effect. Default is 0.
19584 nb_samples, ns
19586 Specify the number of samples for which the fade effect has to
19587 last. At the end of the fade-in effect the output audio will have
19588 the same volume as the input audio, at the end of the fade-out
19589 transition the output audio will be silence. Default is 44100.
19590 start_time, st
19592 Specify the start time of the fade effect. Default is 0. The value
19593 must be specified as a time duration; see the Time duration section
19594 in the ffmpeg-utils(1) manual for the accepted syntax. If set this
19595 option is used instead of start_sample.
19596 duration, d
19598 Specify the duration of the fade effect. See the Time duration
19599 section in the ffmpeg-utils(1) manual for the accepted syntax. At
19600 the end of the fade-in effect the output audio will have the same
19601 volume as the input audio, at the end of the fade-out transition
19602 the output audio will be silence. By default the duration is
19603 determined by nb_samples. If set this option is used instead of
19604 nb_samples.
19605 curve
19607 Set curve for fade transition.
19608 It accepts the following values:
19610 tri select triangular, linear slope (default)
19612 qsin
19614 select quarter of sine wave
19615 hsin
19617 select half of sine wave
19618 esin
19620 select exponential sine wave
19621 log select logarithmic
19623 ipar
19625 select inverted parabola
19626 qua select quadratic
19628 cub select cubic
19630 squ select square root
19632 cbr select cubic root
19634 par select parabola
19636 exp select exponential
19638 iqsin
19640 select inverted quarter of sine wave
19641 ihsin
19643 select inverted half of sine wave
19644 dese
19646 select double-exponential seat
19647 desi
19649 select double-exponential sigmoid
19650 losi
19652 select logistic sigmoid
19653 sinc
19655 select sine cardinal function
19656 isinc
19658 select inverted sine cardinal function
19659 nofade
19661 no fade applied
19662 silence
19664 Set the initial gain for fade-in or final gain for fade-out.
19665 Default value is 0.0.
19666 unity
19668 Set the initial gain for fade-out or final gain for fade-in.
19669 Default value is 1.0.
19670 Commands
19672 This filter supports the all above options as commands.
19674 Examples
19676 • Fade in first 15 seconds of audio:
19678 afade=t=in:ss=0:d=15
19680 • Fade out last 25 seconds of a 900 seconds audio:
19682 afade=t=out:st=875:d=25
19684 afftdn
19686 Denoise audio samples with FFT.
19687 A description of the accepted parameters follows.
19689 noise_reduction, nr
19691 Set the noise reduction in dB, allowed range is 0.01 to 97.
19692 Default value is 12 dB.
19693 noise_floor, nf
19695 Set the noise floor in dB, allowed range is -80 to -20. Default
19696 value is -50 dB.
19697 noise_type, nt
19699 Set the noise type.
19700 It accepts the following values:
19702 white, w
19704 Select white noise.
19705 vinyl, v
19707 Select vinyl noise.
19708 shellac, s
19710 Select shellac noise.
19711 custom, c
19713 Select custom noise, defined in "bn" option.
19714 Default value is white noise.
19716 band_noise, bn
19718 Set custom band noise profile for every one of 15 bands. Bands are
19719 separated by ' ' or '|'.
19720 residual_floor, rf
19722 Set the residual floor in dB, allowed range is -80 to -20. Default
19723 value is -38 dB.
19724 track_noise, tn
19726 Enable noise floor tracking. By default is disabled. With this
19727 enabled, noise floor is automatically adjusted.
19728 track_residual, tr
19730 Enable residual tracking. By default is disabled.
19731 output_mode, om
19733 Set the output mode.
19734 It accepts the following values:
19736 input, i
19738 Pass input unchanged.
19739 output, o
19741 Pass noise filtered out.
19742 noise, n
19744 Pass only noise.
19745 Default value is output.
19747 adaptivity, ad
19749 Set the adaptivity factor, used how fast to adapt gains adjustments
19750 per each frequency bin. Value 0 enables instant adaptation, while
19751 higher values react much slower. Allowed range is from 0 to 1.
19752 Default value is 0.5.
19753 floor_offset, fo
19755 Set the noise floor offset factor. This option is used to adjust
19756 offset applied to measured noise floor. It is only effective when
19757 noise floor tracking is enabled. Allowed range is from -2.0 to
19758 2.0. Default value is 1.0.
19759 noise_link, nl
19761 Set the noise link used for multichannel audio.
19762 It accepts the following values:
19764 none
19766 Use unchanged channel's noise floor.
19767 min Use measured min noise floor of all channels.
19769 max Use measured max noise floor of all channels.
19771 average
19773 Use measured average noise floor of all channels.
19774 Default value is min.
19776 band_multiplier, bm
19778 Set the band multiplier factor, used how much to spread bands
19779 across frequency bins. Allowed range is from 0.2 to 5. Default
19780 value is 1.25.
19781 sample_noise, sn
19783 Toggle capturing and measurement of noise profile from input audio.
19784 It accepts the following values:
19786 start, begin
19788 Start sample noise capture.
19789 stop, end
19791 Stop sample noise capture and measure new noise band profile.
19792 Default value is "none".
19794 gain_smooth, gs
19796 Set gain smooth spatial radius, used to smooth gains applied to
19797 each frequency bin. Useful to reduce random music noise artefacts.
19798 Higher values increases smoothing of gains. Allowed range is from
19799 0 to 50. Default value is 0.
19800 Commands
19802 This filter supports the some above mentioned options as commands.
19804 Examples
19806 • Reduce white noise by 10dB, and use previously measured noise floor
19808 of -40dB:
19809 afftdn=nr=10:nf=-40
19811 • Reduce white noise by 10dB, also set initial noise floor to -80dB
19813 and enable automatic tracking of noise floor so noise floor will
19814 gradually change during processing:
19815 afftdn=nr=10:nf=-80:tn=1
19817 • Reduce noise by 20dB, using noise floor of -40dB and using commands
19819 to take noise profile of first 0.4 seconds of input audio:
19820 asendcmd=0.0 afftdn sn start,asendcmd=0.4 afftdn sn stop,afftdn=nr=20:nf=-40
19822 afftfilt
19824 Apply arbitrary expressions to samples in frequency domain.
19825 real
19827 Set frequency domain real expression for each separate channel
19828 separated by '|'. Default is "re". If the number of input channels
19829 is greater than the number of expressions, the last specified
19830 expression is used for the remaining output channels.
19831 imag
19833 Set frequency domain imaginary expression for each separate channel
19834 separated by '|'. Default is "im".
19835 Each expression in real and imag can contain the following
19837 constants and functions:
19838 sr sample rate
19840 b current frequency bin number
19842 nb number of available bins
19844 ch channel number of the current expression
19846 chs number of channels
19848 pts current frame pts
19850 re current real part of frequency bin of current channel
19852 im current imaginary part of frequency bin of current channel
19854 real(b, ch)
19856 Return the value of real part of frequency bin at location
19857 (bin,channel)
19858 imag(b, ch)
19860 Return the value of imaginary part of frequency bin at location
19861 (bin,channel)
19862 win_size
19864 Set window size. Allowed range is from 16 to 131072. Default is
19865 4096
19866 win_func
19868 Set window function.
19869 It accepts the following values:
19871 rect
19873 bartlett
19874 hann, hanning
19875 hamming
19876 blackman
19877 welch
19878 flattop
19879 bharris
19880 bnuttall
19881 bhann
19882 sine
19883 nuttall
19884 lanczos
19885 gauss
19886 tukey
19887 dolph
19888 cauchy
19889 parzen
19890 poisson
19891 bohman
19892 kaiser
19893 Default is "hann".
19895 overlap
19897 Set window overlap. If set to 1, the recommended overlap for
19898 selected window function will be picked. Default is 0.75.
19899 Examples
19901 • Leave almost only low frequencies in audio:
19903 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
19905 • Apply robotize effect:
19907 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
19909 • Apply whisper effect:
19911 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
19913 • Apply phase shift:
19915 afftfilt="real=re*cos(1)-im*sin(1):imag=re*sin(1)+im*cos(1)"
19917 afir
19919 Apply an arbitrary Finite Impulse Response filter.
19920 This filter is designed for applying long FIR filters, up to 60 seconds
19922 long.
19923 It can be used as component for digital crossover filters, room
19925 equalization, cross talk cancellation, wavefield synthesis,
19926 auralization, ambiophonics, ambisonics and spatialization.
19927 This filter uses the streams higher than first one as FIR coefficients.
19929 If the non-first stream holds a single channel, it will be used for all
19930 input channels in the first stream, otherwise the number of channels in
19931 the non-first stream must be same as the number of channels in the
19932 first stream.
19933 It accepts the following parameters:
19935 dry Set dry gain. This sets input gain.
19937 wet Set wet gain. This sets final output gain.
19939 length
19941 Set Impulse Response filter length. Default is 1, which means whole
19942 IR is processed.
19943 gtype
19945 Enable applying gain measured from power of IR.
19946 Set which approach to use for auto gain measurement.
19948 none
19950 Do not apply any gain.
19951 peak
19953 select peak gain, very conservative approach. This is default
19954 value.
19955 dc select DC gain, limited application.
19957 gn select gain to noise approach, this is most popular one.
19959 ac select AC gain.
19961 rms select RMS gain.
19963 irgain
19965 Set gain to be applied to IR coefficients before filtering.
19966 Allowed range is 0 to 1. This gain is applied after any gain
19967 applied with gtype option.
19968 irfmt
19970 Set format of IR stream. Can be "mono" or "input". Default is
19971 "input".
19972 maxir
19974 Set max allowed Impulse Response filter duration in seconds.
19975 Default is 30 seconds. Allowed range is 0.1 to 60 seconds.
19976 response
19978 Show IR frequency response, magnitude(magenta), phase(green) and
19979 group delay(yellow) in additional video stream. By default it is
19980 disabled.
19981 channel
19983 Set for which IR channel to display frequency response. By default
19984 is first channel displayed. This option is used only when response
19985 is enabled.
19986 size
19988 Set video stream size. This option is used only when response is
19989 enabled.
19990 rate
19992 Set video stream frame rate. This option is used only when response
19993 is enabled.
19994 minp
19996 Set minimal partition size used for convolution. Default is 8192.
19997 Allowed range is from 1 to 65536. Lower values decreases latency
19998 at cost of higher CPU usage.
19999 maxp
20001 Set maximal partition size used for convolution. Default is 8192.
20002 Allowed range is from 8 to 65536. Lower values may increase CPU
20003 usage.
20004 nbirs
20006 Set number of input impulse responses streams which will be
20007 switchable at runtime. Allowed range is from 1 to 32. Default is
20008 1.
20009 ir Set IR stream which will be used for convolution, starting from 0,
20011 should always be lower than supplied value by "nbirs" option.
20012 Default is 0. This option can be changed at runtime via commands.
20013 precision
20015 Set which precision to use when processing samples.
20016 auto
20018 Auto pick internal sample format depending on other filters.
20019 float
20021 Always use single-floating point precision sample format.
20022 double
20024 Always use double-floating point precision sample format.
20025 Default value is auto.
20027 Examples
20029 • Apply reverb to stream using mono IR file as second input, complete
20031 command using ffmpeg:
20032 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
20034 • Apply true stereo processing given input stereo stream, and two
20036 stereo impulse responses for left and right channel, the impulse
20037 response files are files with names l_ir.wav and r_ir.wav:
20038 "pan=4C|c0=FL|c1=FL|c2=FR|c3=FR[a];amovie=l_ir.wav[LIR];amovie=r_ir.wav[RIR];[LIR][RIR]amerge[ir];[a][ir]afir=irfmt=input:gtype=gn:irgain=-5dB,pan=stereo|FL<c0+c2|FR<c1+c3"
20040 aformat
20042 Set output format constraints for the input audio. The framework will
20043 negotiate the most appropriate format to minimize conversions.
20044 It accepts the following parameters:
20046 sample_fmts, f
20048 A '|'-separated list of requested sample formats.
20049 sample_rates, r
20051 A '|'-separated list of requested sample rates.
20052 channel_layouts, cl
20054 A '|'-separated list of requested channel layouts.
20055 See the Channel Layout section in the ffmpeg-utils(1) manual for
20057 the required syntax.
20058 If a parameter is omitted, all values are allowed.
20060 Force the output to either unsigned 8-bit or signed 16-bit stereo
20062 aformat=sample_fmts=u8|s16:channel_layouts=stereo
20064 afreqshift
20066 Apply frequency shift to input audio samples.
20067 The filter accepts the following options:
20069 shift
20071 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
20072 Default value is 0.0.
20073 level
20075 Set output gain applied to final output. Allowed range is from 0.0
20076 to 1.0. Default value is 1.0.
20077 order
20079 Set filter order used for filtering. Allowed range is from 1 to 16.
20080 Default value is 8.
20081 Commands
20083 This filter supports the all above options as commands.
20085 afwtdn
20087 Reduce broadband noise from input samples using Wavelets.
20088 A description of the accepted options follows.
20090 sigma
20092 Set the noise sigma, allowed range is from 0 to 1. Default value
20093 is 0. This option controls strength of denoising applied to input
20094 samples. Most useful way to set this option is via decibels, eg.
20095 -45dB.
20096 levels
20098 Set the number of wavelet levels of decomposition. Allowed range
20099 is from 1 to 12. Default value is 10. Setting this too low make
20100 denoising performance very poor.
20101 wavet
20103 Set wavelet type for decomposition of input frame. They are sorted
20104 by number of coefficients, from lowest to highest. More
20105 coefficients means worse filtering speed, but overall better
20106 quality. Available wavelets are:
20107 sym2
20109 sym4
20110 rbior68
20111 deb10
20112 sym10
20113 coif5
20114 bl3
20115 percent
20116 Set percent of full denoising. Allowed range is from 0 to 100
20117 percent. Default value is 85 percent or partial denoising.
20118 profile
20120 If enabled, first input frame will be used as noise profile. If
20121 first frame samples contain non-noise performance will be very
20122 poor.
20123 adaptive
20125 If enabled, input frames are analyzed for presence of noise. If
20126 noise is detected with high possibility then input frame profile
20127 will be used for processing following frames, until new noise frame
20128 is detected.
20129 samples
20131 Set size of single frame in number of samples. Allowed range is
20132 from 512 to 65536. Default frame size is 8192 samples.
20133 softness
20135 Set softness applied inside thresholding function. Allowed range is
20136 from 0 to 10. Default softness is 1.
20137 Commands
20139 This filter supports the all above options as commands.
20141 agate
20143 A gate is mainly used to reduce lower parts of a signal. This kind of
20144 signal processing reduces disturbing noise between useful signals.
20145 Gating is done by detecting the volume below a chosen level threshold
20147 and dividing it by the factor set with ratio. The bottom of the noise
20148 floor is set via range. Because an exact manipulation of the signal
20149 would cause distortion of the waveform the reduction can be levelled
20150 over time. This is done by setting attack and release.
20151 attack determines how long the signal has to fall below the threshold
20153 before any reduction will occur and release sets the time the signal
20154 has to rise above the threshold to reduce the reduction again. Shorter
20155 signals than the chosen attack time will be left untouched.
20156 level_in
20158 Set input level before filtering. Default is 1. Allowed range is
20159 from 0.015625 to 64.
20160 mode
20162 Set the mode of operation. Can be "upward" or "downward". Default
20163 is "downward". If set to "upward" mode, higher parts of signal will
20164 be amplified, expanding dynamic range in upward direction.
20165 Otherwise, in case of "downward" lower parts of signal will be
20166 reduced.
20167 range
20169 Set the level of gain reduction when the signal is below the
20170 threshold. Default is 0.06125. Allowed range is from 0 to 1.
20171 Setting this to 0 disables reduction and then filter behaves like
20172 expander.
20173 threshold
20175 If a signal rises above this level the gain reduction is released.
20176 Default is 0.125. Allowed range is from 0 to 1.
20177 ratio
20179 Set a ratio by which the signal is reduced. Default is 2. Allowed
20180 range is from 1 to 9000.
20181 attack
20183 Amount of milliseconds the signal has to rise above the threshold
20184 before gain reduction stops. Default is 20 milliseconds. Allowed
20185 range is from 0.01 to 9000.
20186 release
20188 Amount of milliseconds the signal has to fall below the threshold
20189 before the reduction is increased again. Default is 250
20190 milliseconds. Allowed range is from 0.01 to 9000.
20191 makeup
20193 Set amount of amplification of signal after processing. Default is
20194 1. Allowed range is from 1 to 64.
20195 knee
20197 Curve the sharp knee around the threshold to enter gain reduction
20198 more softly. Default is 2.828427125. Allowed range is from 1 to 8.
20199 detection
20201 Choose if exact signal should be taken for detection or an RMS like
20202 one. Default is "rms". Can be "peak" or "rms".
20203 link
20205 Choose if the average level between all channels or the louder
20206 channel affects the reduction. Default is "average". Can be
20207 "average" or "maximum".
20208 Commands
20210 This filter supports the all above options as commands.
20212 aiir
20214 Apply an arbitrary Infinite Impulse Response filter.
20215 It accepts the following parameters:
20217 zeros, z
20219 Set B/numerator/zeros/reflection coefficients.
20220 poles, p
20222 Set A/denominator/poles/ladder coefficients.
20223 gains, k
20225 Set channels gains.
20226 dry_gain
20228 Set input gain.
20229 wet_gain
20231 Set output gain.
20232 format, f
20234 Set coefficients format.
20235 ll lattice-ladder function
20237 sf analog transfer function
20239 tf digital transfer function
20241 zp Z-plane zeros/poles, cartesian (default)
20243 pr Z-plane zeros/poles, polar radians
20245 pd Z-plane zeros/poles, polar degrees
20247 sp S-plane zeros/poles
20249 process, r
20251 Set type of processing.
20252 d direct processing
20254 s serial processing
20256 p parallel processing
20258 precision, e
20260 Set filtering precision.
20261 dbl double-precision floating-point (default)
20263 flt single-precision floating-point
20265 i32 32-bit integers
20267 i16 16-bit integers
20269 normalize, n
20271 Normalize filter coefficients, by default is enabled. Enabling it
20272 will normalize magnitude response at DC to 0dB.
20273 mix How much to use filtered signal in output. Default is 1. Range is
20275 between 0 and 1.
20276 response
20278 Show IR frequency response, magnitude(magenta), phase(green) and
20279 group delay(yellow) in additional video stream. By default it is
20280 disabled.
20281 channel
20283 Set for which IR channel to display frequency response. By default
20284 is first channel displayed. This option is used only when response
20285 is enabled.
20286 size
20288 Set video stream size. This option is used only when response is
20289 enabled.
20290 Coefficients in "tf" and "sf" format are separated by spaces and are in
20292 ascending order.
20293 Coefficients in "zp" format are separated by spaces and order of
20295 coefficients doesn't matter. Coefficients in "zp" format are complex
20296 numbers with i imaginary unit.
20297 Different coefficients and gains can be provided for every channel, in
20299 such case use '|' to separate coefficients or gains. Last provided
20300 coefficients will be used for all remaining channels.
20301 Examples
20303 • Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample
20305 rate:
20306 aiir=k=1:z=7.957584807809675810E-1 -2.575128568908332300 3.674839853930788710 -2.57512875289799137 7.957586296317130880E-1:p=1 -2.86950072432325953 3.63022088054647218 -2.28075678147272232 6.361362326477423500E-1:f=tf:r=d
20308 • Same as above but in "zp" format:
20310 aiir=k=0.79575848078096756:z=0.80918701+0.58773007i 0.80918701-0.58773007i 0.80884700+0.58784055i 0.80884700-0.58784055i:p=0.63892345+0.59951235i 0.63892345-0.59951235i 0.79582691+0.44198673i 0.79582691-0.44198673i:f=zp:r=s
20312 • Apply 3-rd order analog normalized Butterworth low-pass filter,
20314 using analog transfer function format:
20315 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
20317 alimiter
20319 The limiter prevents an input signal from rising over a desired
20320 threshold. This limiter uses lookahead technology to prevent your
20321 signal from distorting. It means that there is a small delay after the
20322 signal is processed. Keep in mind that the delay it produces is the
20323 attack time you set.
20324 The filter accepts the following options:
20326 level_in
20328 Set input gain. Default is 1.
20329 level_out
20331 Set output gain. Default is 1.
20332 limit
20334 Don't let signals above this level pass the limiter. Default is 1.
20335 attack
20337 The limiter will reach its attenuation level in this amount of time
20338 in milliseconds. Default is 5 milliseconds.
20339 release
20341 Come back from limiting to attenuation 1.0 in this amount of
20342 milliseconds. Default is 50 milliseconds.
20343 asc When gain reduction is always needed ASC takes care of releasing to
20345 an average reduction level rather than reaching a reduction of 0 in
20346 the release time.
20347 asc_level
20349 Select how much the release time is affected by ASC, 0 means nearly
20350 no changes in release time while 1 produces higher release times.
20351 level
20353 Auto level output signal. Default is enabled. This normalizes
20354 audio back to 0dB if enabled.
20355 latency
20357 Compensate the delay introduced by using the lookahead buffer set
20358 with attack parameter. Also flush the valid audio data in the
20359 lookahead buffer when the stream hits EOF.
20360 Depending on picked setting it is recommended to upsample input 2x or
20362 4x times with aresample before applying this filter.
20363 allpass
20365 Apply a two-pole all-pass filter with central frequency (in Hz)
20366 frequency, and filter-width width. An all-pass filter changes the
20367 audio's frequency to phase relationship without changing its frequency
20368 to amplitude relationship.
20369 The filter accepts the following options:
20371 frequency, f
20373 Set frequency in Hz.
20374 width_type, t
20376 Set method to specify band-width of filter.
20377 h Hz
20379 q Q-Factor
20381 o octave
20383 s slope
20385 k kHz
20387 width, w
20389 Specify the band-width of a filter in width_type units.
20390 mix, m
20392 How much to use filtered signal in output. Default is 1. Range is
20393 between 0 and 1.
20394 channels, c
20396 Specify which channels to filter, by default all available are
20397 filtered.
20398 normalize, n
20400 Normalize biquad coefficients, by default is disabled. Enabling it
20401 will normalize magnitude response at DC to 0dB.
20402 order, o
20404 Set the filter order, can be 1 or 2. Default is 2.
20405 transform, a
20407 Set transform type of IIR filter.
20408 di
20410 dii
20411 tdi
20412 tdii
20413 latt
20414 svf
20415 zdf
20416 precision, r
20417 Set precison of filtering.
20418 auto
20420 Pick automatic sample format depending on surround filters.
20421 s16 Always use signed 16-bit.
20423 s32 Always use signed 32-bit.
20425 f32 Always use float 32-bit.
20427 f64 Always use float 64-bit.
20429 Commands
20431 This filter supports the following commands:
20433 frequency, f
20435 Change allpass frequency. Syntax for the command is : "frequency"
20436 width_type, t
20438 Change allpass width_type. Syntax for the command is :
20439 "width_type"
20440 width, w
20442 Change allpass width. Syntax for the command is : "width"
20443 mix, m
20445 Change allpass mix. Syntax for the command is : "mix"
20446 aloop
20448 Loop audio samples.
20449 The filter accepts the following options:
20451 loop
20453 Set the number of loops. Setting this value to -1 will result in
20454 infinite loops. Default is 0.
20455 size
20457 Set maximal number of samples. Default is 0.
20458 start
20460 Set first sample of loop. Default is 0.
20461 amerge
20463 Merge two or more audio streams into a single multi-channel stream.
20464 The filter accepts the following options:
20466 inputs
20468 Set the number of inputs. Default is 2.
20469 If the channel layouts of the inputs are disjoint, and therefore
20471 compatible, the channel layout of the output will be set accordingly
20472 and the channels will be reordered as necessary. If the channel layouts
20473 of the inputs are not disjoint, the output will have all the channels
20474 of the first input then all the channels of the second input, in that
20475 order, and the channel layout of the output will be the default value
20476 corresponding to the total number of channels.
20477 For example, if the first input is in 2.1 (FL+FR+LF) and the second
20479 input is FC+BL+BR, then the output will be in 5.1, with the channels in
20480 the following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of
20481 the first input, b1 is the first channel of the second input).
20482 On the other hand, if both input are in stereo, the output channels
20484 will be in the default order: a1, a2, b1, b2, and the channel layout
20485 will be arbitrarily set to 4.0, which may or may not be the expected
20486 value.
20487 All inputs must have the same sample rate, and format.
20489 If inputs do not have the same duration, the output will stop with the
20491 shortest.
20492 Examples
20494 • Merge two mono files into a stereo stream:
20496 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
20498 • Multiple merges assuming 1 video stream and 6 audio streams in
20500 input.mkv:
20501 ffmpeg -i input.mkv -filter_complex "[0:1][0:2][0:3][0:4][0:5][0:6] amerge=inputs=6" -c:a pcm_s16le output.mkv
20503 amix
20505 Mixes multiple audio inputs into a single output.
20506 Note that this filter only supports float samples (the amerge and pan
20508 audio filters support many formats). If the amix input has integer
20509 samples then aresample will be automatically inserted to perform the
20510 conversion to float samples.
20511 It accepts the following parameters:
20513 inputs
20515 The number of inputs. If unspecified, it defaults to 2.
20516 duration
20518 How to determine the end-of-stream.
20519 longest
20521 The duration of the longest input. (default)
20522 shortest
20524 The duration of the shortest input.
20525 first
20527 The duration of the first input.
20528 dropout_transition
20530 The transition time, in seconds, for volume renormalization when an
20531 input stream ends. The default value is 2 seconds.
20532 weights
20534 Specify weight of each input audio stream as a sequence of numbers
20535 separated by a space. If fewer weights are specified compared to
20536 number of inputs, the last weight is assigned to the remaining
20537 inputs. Default weight for each input is 1.
20538 normalize
20540 Always scale inputs instead of only doing summation of samples.
20541 Beware of heavy clipping if inputs are not normalized prior or
20542 after filtering by this filter if this option is disabled. By
20543 default is enabled.
20544 Examples
20546 • This will mix 3 input audio streams to a single output with the
20548 same duration as the first input and a dropout transition time of 3
20549 seconds:
20550 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
20552 • This will mix one vocal and one music input audio stream to a
20554 single output with the same duration as the longest input. The
20555 music will have quarter the weight as the vocals, and the inputs
20556 are not normalized:
20557 ffmpeg -i VOCALS -i MUSIC -filter_complex amix=inputs=2:duration=longest:dropout_transition=0:weights="1 0.25":normalize=0 OUTPUT
20559 Commands
20561 This filter supports the following commands:
20563 weights
20565 normalize
20566 Syntax is same as option with same name.
20567 amultiply
20569 Multiply first audio stream with second audio stream and store result
20570 in output audio stream. Multiplication is done by multiplying each
20571 sample from first stream with sample at same position from second
20572 stream.
20573 With this element-wise multiplication one can create amplitude fades
20575 and amplitude modulations.
20576 anequalizer
20578 High-order parametric multiband equalizer for each channel.
20579 It accepts the following parameters:
20581 params
20583 This option string is in format: "cchn f=cf w=w g=g t=f | ..."
20584 Each equalizer band is separated by '|'.
20585 chn Set channel number to which equalization will be applied. If
20587 input doesn't have that channel the entry is ignored.
20588 f Set central frequency for band. If input doesn't have that
20590 frequency the entry is ignored.
20591 w Set band width in Hertz.
20593 g Set band gain in dB.
20595 t Set filter type for band, optional, can be:
20597 0 Butterworth, this is default.
20599 1 Chebyshev type 1.
20601 2 Chebyshev type 2.
20603 curves
20605 With this option activated frequency response of anequalizer is
20606 displayed in video stream.
20607 size
20609 Set video stream size. Only useful if curves option is activated.
20610 mgain
20612 Set max gain that will be displayed. Only useful if curves option
20613 is activated. Setting this to a reasonable value makes it possible
20614 to display gain which is derived from neighbour bands which are too
20615 close to each other and thus produce higher gain when both are
20616 activated.
20617 fscale
20619 Set frequency scale used to draw frequency response in video
20620 output. Can be linear or logarithmic. Default is logarithmic.
20621 colors
20623 Set color for each channel curve which is going to be displayed in
20624 video stream. This is list of color names separated by space or by
20625 '|'. Unrecognised or missing colors will be replaced by white
20626 color.
20627 Examples
20629 • Lower gain by 10 of central frequency 200Hz and width 100 Hz for
20631 first 2 channels using Chebyshev type 1 filter:
20632 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
20634 Commands
20636 This filter supports the following commands:
20638 change
20640 Alter existing filter parameters. Syntax for the commands is :
20641 "fN|f=freq|w=width|g=gain"
20642 fN is existing filter number, starting from 0, if no such filter is
20644 available error is returned. freq set new frequency parameter.
20645 width set new width parameter in Hertz. gain set new gain
20646 parameter in dB.
20647 Full filter invocation with asendcmd may look like this:
20649 asendcmd=c='4.0 anequalizer change
20650 0|f=200|w=50|g=1',anequalizer=...
20651 anlmdn
20653 Reduce broadband noise in audio samples using Non-Local Means
20654 algorithm.
20655 Each sample is adjusted by looking for other samples with similar
20657 contexts. This context similarity is defined by comparing their
20658 surrounding patches of size p. Patches are searched in an area of r
20659 around the sample.
20660 The filter accepts the following options:
20662 strength, s
20664 Set denoising strength. Allowed range is from 0.00001 to 10000.
20665 Default value is 0.00001.
20666 patch, p
20668 Set patch radius duration. Allowed range is from 1 to 100
20669 milliseconds. Default value is 2 milliseconds.
20670 research, r
20672 Set research radius duration. Allowed range is from 2 to 300
20673 milliseconds. Default value is 6 milliseconds.
20674 output, o
20676 Set the output mode.
20677 It accepts the following values:
20679 i Pass input unchanged.
20681 o Pass noise filtered out.
20683 n Pass only noise.
20685 Default value is o.
20687 smooth, m
20689 Set smooth factor. Default value is 11. Allowed range is from 1 to
20690 1000.
20691 Commands
20693 This filter supports the all above options as commands.
20695 anlmf, anlms
20697 Apply Normalized Least-Mean-(Squares|Fourth) algorithm to the first
20698 audio stream using the second audio stream.
20699 This adaptive filter is used to mimic a desired filter by finding the
20701 filter coefficients that relate to producing the least mean square of
20702 the error signal (difference between the desired, 2nd input audio
20703 stream and the actual signal, the 1st input audio stream).
20704 A description of the accepted options follows.
20706 order
20708 Set filter order.
20709 mu Set filter mu.
20711 eps Set the filter eps.
20713 leakage
20715 Set the filter leakage.
20716 out_mode
20718 It accepts the following values:
20719 i Pass the 1st input.
20721 d Pass the 2nd input.
20723 o Pass filtered samples.
20725 n Pass difference between desired and filtered samples.
20727 Default value is o.
20729 Examples
20731 • One of many usages of this filter is noise reduction, input audio
20733 is filtered with same samples that are delayed by fixed amount, one
20734 such example for stereo audio is:
20735 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
20737 Commands
20739 This filter supports the same commands as options, excluding option
20741 "order".
20742 anull
20744 Pass the audio source unchanged to the output.
20745 apad
20747 Pad the end of an audio stream with silence.
20748 This can be used together with ffmpeg -shortest to extend audio streams
20750 to the same length as the video stream.
20751 A description of the accepted options follows.
20753 packet_size
20755 Set silence packet size. Default value is 4096.
20756 pad_len
20758 Set the number of samples of silence to add to the end. After the
20759 value is reached, the stream is terminated. This option is mutually
20760 exclusive with whole_len.
20761 whole_len
20763 Set the minimum total number of samples in the output audio stream.
20764 If the value is longer than the input audio length, silence is
20765 added to the end, until the value is reached. This option is
20766 mutually exclusive with pad_len.
20767 pad_dur
20769 Specify the duration of samples of silence to add. See the Time
20770 duration section in the ffmpeg-utils(1) manual for the accepted
20771 syntax. Used only if set to non-negative value.
20772 whole_dur
20774 Specify the minimum total duration in the output audio stream. See
20775 the Time duration section in the ffmpeg-utils(1) manual for the
20776 accepted syntax. Used only if set to non-negative value. If the
20777 value is longer than the input audio length, silence is added to
20778 the end, until the value is reached. This option is mutually
20779 exclusive with pad_dur
20780 If neither the pad_len nor the whole_len nor pad_dur nor whole_dur
20782 option is set, the filter will add silence to the end of the input
20783 stream indefinitely.
20784 Note that for ffmpeg 4.4 and earlier a zero pad_dur or whole_dur also
20786 caused the filter to add silence indefinitely.
20787 Examples
20789 • Add 1024 samples of silence to the end of the input:
20791 apad=pad_len=1024
20793 • Make sure the audio output will contain at least 10000 samples, pad
20795 the input with silence if required:
20796 apad=whole_len=10000
20798 • Use ffmpeg to pad the audio input with silence, so that the video
20800 stream will always result the shortest and will be converted until
20801 the end in the output file when using the shortest option:
20802 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
20804 aphaser
20806 Add a phasing effect to the input audio.
20807 A phaser filter creates series of peaks and troughs in the frequency
20809 spectrum. The position of the peaks and troughs are modulated so that
20810 they vary over time, creating a sweeping effect.
20811 A description of the accepted parameters follows.
20813 in_gain
20815 Set input gain. Default is 0.4.
20816 out_gain
20818 Set output gain. Default is 0.74
20819 delay
20821 Set delay in milliseconds. Default is 3.0.
20822 decay
20824 Set decay. Default is 0.4.
20825 speed
20827 Set modulation speed in Hz. Default is 0.5.
20828 type
20830 Set modulation type. Default is triangular.
20831 It accepts the following values:
20833 triangular, t
20835 sinusoidal, s
20836 aphaseshift
20838 Apply phase shift to input audio samples.
20839 The filter accepts the following options:
20841 shift
20843 Specify phase shift. Allowed range is from -1.0 to 1.0. Default
20844 value is 0.0.
20845 level
20847 Set output gain applied to final output. Allowed range is from 0.0
20848 to 1.0. Default value is 1.0.
20849 order
20851 Set filter order used for filtering. Allowed range is from 1 to 16.
20852 Default value is 8.
20853 Commands
20855 This filter supports the all above options as commands.
20857 apsyclip
20859 Apply Psychoacoustic clipper to input audio stream.
20860 The filter accepts the following options:
20862 level_in
20864 Set input gain. By default it is 1. Range is [0.015625 - 64].
20865 level_out
20867 Set output gain. By default it is 1. Range is [0.015625 - 64].
20868 clip
20870 Set the clipping start value. Default value is 0dBFS or 1.
20871 diff
20873 Output only difference samples, useful to hear introduced
20874 distortions. By default is disabled.
20875 adaptive
20877 Set strength of adaptive distortion applied. Default value is 0.5.
20878 Allowed range is from 0 to 1.
20879 iterations
20881 Set number of iterations of psychoacoustic clipper. Allowed range
20882 is from 1 to 20. Default value is 10.
20883 level
20885 Auto level output signal. Default is disabled. This normalizes
20886 audio back to 0dBFS if enabled.
20887 Commands
20889 This filter supports the all above options as commands.
20891 apulsator
20893 Audio pulsator is something between an autopanner and a tremolo. But
20894 it can produce funny stereo effects as well. Pulsator changes the
20895 volume of the left and right channel based on a LFO (low frequency
20896 oscillator) with different waveforms and shifted phases. This filter
20897 have the ability to define an offset between left and right channel. An
20898 offset of 0 means that both LFO shapes match each other. The left and
20899 right channel are altered equally - a conventional tremolo. An offset
20900 of 50% means that the shape of the right channel is exactly shifted in
20901 phase (or moved backwards about half of the frequency) - pulsator acts
20902 as an autopanner. At 1 both curves match again. Every setting in
20903 between moves the phase shift gapless between all stages and produces
20904 some "bypassing" sounds with sine and triangle waveforms. The more you
20905 set the offset near 1 (starting from the 0.5) the faster the signal
20906 passes from the left to the right speaker.
20907 The filter accepts the following options:
20909 level_in
20911 Set input gain. By default it is 1. Range is [0.015625 - 64].
20912 level_out
20914 Set output gain. By default it is 1. Range is [0.015625 - 64].
20915 mode
20917 Set waveform shape the LFO will use. Can be one of: sine, triangle,
20918 square, sawup or sawdown. Default is sine.
20919 amount
20921 Set modulation. Define how much of original signal is affected by
20922 the LFO.
20923 offset_l
20925 Set left channel offset. Default is 0. Allowed range is [0 - 1].
20926 offset_r
20928 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
20929 width
20931 Set pulse width. Default is 1. Allowed range is [0 - 2].
20932 timing
20934 Set possible timing mode. Can be one of: bpm, ms or hz. Default is
20935 hz.
20936 bpm Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if
20938 timing is set to bpm.
20939 ms Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if
20941 timing is set to ms.
20942 hz Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100].
20944 Only used if timing is set to hz.
20945 aresample
20947 Resample the input audio to the specified parameters, using the
20948 libswresample library. If none are specified then the filter will
20949 automatically convert between its input and output.
20950 This filter is also able to stretch/squeeze the audio data to make it
20952 match the timestamps or to inject silence / cut out audio to make it
20953 match the timestamps, do a combination of both or do neither.
20954 The filter accepts the syntax [sample_rate:]resampler_options, where
20956 sample_rate expresses a sample rate and resampler_options is a list of
20957 key=value pairs, separated by ":". See the "Resampler Options" section
20958 in the ffmpeg-resampler(1) manual for the complete list of supported
20959 options.
20960 Examples
20962 • Resample the input audio to 44100Hz:
20964 aresample=44100
20966 • Stretch/squeeze samples to the given timestamps, with a maximum of
20968 1000 samples per second compensation:
20969 aresample=async=1000
20971 areverse
20973 Reverse an audio clip.
20974 Warning: This filter requires memory to buffer the entire clip, so
20976 trimming is suggested.
20977 Examples
20979 • Take the first 5 seconds of a clip, and reverse it.
20981 atrim=end=5,areverse
20983 arnndn
20985 Reduce noise from speech using Recurrent Neural Networks.
20986 This filter accepts the following options:
20988 model, m
20990 Set train model file to load. This option is always required.
20991 mix Set how much to mix filtered samples into final output. Allowed
20993 range is from -1 to 1. Default value is 1. Negative values are
20994 special, they set how much to keep filtered noise in the final
20995 filter output. Set this option to -1 to hear actual noise removed
20996 from input signal.
20997 Commands
20999 This filter supports the all above options as commands.
21001 asdr
21003 Measure Audio Signal-to-Distortion Ratio.
21004 This filter takes two audio streams for input, and outputs first audio
21006 stream. Results are in dB per channel at end of either input.
21007 asetnsamples
21009 Set the number of samples per each output audio frame.
21010 The last output packet may contain a different number of samples, as
21012 the filter will flush all the remaining samples when the input audio
21013 signals its end.
21014 The filter accepts the following options:
21016 nb_out_samples, n
21018 Set the number of frames per each output audio frame. The number is
21019 intended as the number of samples per each channel. Default value
21020 is 1024.
21021 pad, p
21023 If set to 1, the filter will pad the last audio frame with zeroes,
21024 so that the last frame will contain the same number of samples as
21025 the previous ones. Default value is 1.
21026 For example, to set the number of per-frame samples to 1234 and disable
21028 padding for the last frame, use:
21029 asetnsamples=n=1234:p=0
21031 asetrate
21033 Set the sample rate without altering the PCM data. This will result in
21034 a change of speed and pitch.
21035 The filter accepts the following options:
21037 sample_rate, r
21039 Set the output sample rate. Default is 44100 Hz.
21040 ashowinfo
21042 Show a line containing various information for each input audio frame.
21043 The input audio is not modified.
21044 The shown line contains a sequence of key/value pairs of the form
21046 key:value.
21047 The following values are shown in the output:
21049 n The (sequential) number of the input frame, starting from 0.
21051 pts The presentation timestamp of the input frame, in time base units;
21053 the time base depends on the filter input pad, and is usually
21054 1/sample_rate.
21055 pts_time
21057 The presentation timestamp of the input frame in seconds.
21058 pos position of the frame in the input stream, -1 if this information
21060 in unavailable and/or meaningless (for example in case of synthetic
21061 audio)
21062 fmt The sample format.
21064 chlayout
21066 The channel layout.
21067 rate
21069 The sample rate for the audio frame.
21070 nb_samples
21072 The number of samples (per channel) in the frame.
21073 checksum
21075 The Adler-32 checksum (printed in hexadecimal) of the audio data.
21076 For planar audio, the data is treated as if all the planes were
21077 concatenated.
21078 plane_checksums
21080 A list of Adler-32 checksums for each data plane.
21081 asoftclip
21083 Apply audio soft clipping.
21084 Soft clipping is a type of distortion effect where the amplitude of a
21086 signal is saturated along a smooth curve, rather than the abrupt shape
21087 of hard-clipping.
21088 This filter accepts the following options:
21090 type
21092 Set type of soft-clipping.
21093 It accepts the following values:
21095 hard
21097 tanh
21098 atan
21099 cubic
21100 exp
21101 alg
21102 quintic
21103 sin
21104 erf
21105 threshold
21106 Set threshold from where to start clipping. Default value is 0dB or
21107 1.
21108 output
21110 Set gain applied to output. Default value is 0dB or 1.
21111 param
21113 Set additional parameter which controls sigmoid function.
21114 oversample
21116 Set oversampling factor.
21117 Commands
21119 This filter supports the all above options as commands.
21121 aspectralstats
21123 Display frequency domain statistical information about the audio
21124 channels. Statistics are calculated and stored as metadata for each
21125 audio channel and for each audio frame.
21126 It accepts the following option:
21128 win_size
21130 Set the window length in samples. Default value is 2048. Allowed
21131 range is from 32 to 65536.
21132 win_func
21134 Set window function.
21135 It accepts the following values:
21137 rect
21139 bartlett
21140 hann, hanning
21141 hamming
21142 blackman
21143 welch
21144 flattop
21145 bharris
21146 bnuttall
21147 bhann
21148 sine
21149 nuttall
21150 lanczos
21151 gauss
21152 tukey
21153 dolph
21154 cauchy
21155 parzen
21156 poisson
21157 bohman
21158 kaiser
21159 Default is "hann".
21161 overlap
21163 Set window overlap. Allowed range is from 0 to 1. Default value is
21164 0.5.
21165 measure
21167 Select the parameters which are measured. The metadata keys can be
21168 used as flags, default is all which measures everything. none
21169 disables all measurement.
21170 A list of each metadata key follows:
21172 mean
21174 variance
21175 centroid
21176 spread
21177 skewness
21178 kurtosis
21179 entropy
21180 flatness
21181 crest
21182 flux
21183 slope
21184 decrease
21185 rolloff
21186 asr
21188 Automatic Speech Recognition
21189 This filter uses PocketSphinx for speech recognition. To enable
21191 compilation of this filter, you need to configure FFmpeg with
21192 "--enable-pocketsphinx".
21193 It accepts the following options:
21195 rate
21197 Set sampling rate of input audio. Defaults is 16000. This need to
21198 match speech models, otherwise one will get poor results.
21199 hmm Set dictionary containing acoustic model files.
21201 dict
21203 Set pronunciation dictionary.
21204 lm Set language model file.
21206 lmctl
21208 Set language model set.
21209 lmname
21211 Set which language model to use.
21212 logfn
21214 Set output for log messages.
21215 The filter exports recognized speech as the frame metadata
21217 "lavfi.asr.text".
21218 astats
21220 Display time domain statistical information about the audio channels.
21221 Statistics are calculated and displayed for each audio channel and,
21222 where applicable, an overall figure is also given.
21223 It accepts the following option:
21225 length
21227 Short window length in seconds, used for peak and trough RMS
21228 measurement. Default is 0.05 (50 milliseconds). Allowed range is
21229 "[0 - 10]".
21230 metadata
21232 Set metadata injection. All the metadata keys are prefixed with
21233 "lavfi.astats.X", where "X" is channel number starting from 1 or
21234 string "Overall". Default is disabled.
21235 Available keys for each channel are: Bit_depth Crest_factor
21237 DC_offset Dynamic_range Entropy Flat_factor Max_difference
21238 Max_level Mean_difference Min_difference Min_level Noise_floor
21239 Noise_floor_count Number_of_Infs Number_of_NaNs Number_of_denormals
21240 Peak_count Peak_level RMS_difference RMS_peak RMS_trough
21241 Zero_crossings Zero_crossings_rate
21242 and for "Overall": Bit_depth DC_offset Entropy Flat_factor
21244 Max_difference Max_level Mean_difference Min_difference Min_level
21245 Noise_floor Noise_floor_count Number_of_Infs Number_of_NaNs
21246 Number_of_denormals Number_of_samples Peak_count Peak_level
21247 RMS_difference RMS_level RMS_peak RMS_trough
21248 For example, a full key looks like "lavfi.astats.1.DC_offset" or
21250 "lavfi.astats.Overall.Peak_count".
21251 Read below for the description of the keys.
21253 reset
21255 Set the number of frames over which cumulative stats are calculated
21256 before being reset. Default is disabled.
21257 measure_perchannel
21259 Select the parameters which are measured per channel. The metadata
21260 keys can be used as flags, default is all which measures
21261 everything. none disables all per channel measurement.
21262 measure_overall
21264 Select the parameters which are measured overall. The metadata keys
21265 can be used as flags, default is all which measures everything.
21266 none disables all overall measurement.
21267 A description of the measure keys follow:
21269 none
21271 no measures
21272 all all measures
21274 Bit_depth
21276 overall bit depth of audio, i.e. number of bits used for each
21277 sample
21278 Crest_factor
21280 standard ratio of peak to RMS level (note: not in dB)
21281 DC_offset
21283 mean amplitude displacement from zero
21284 Dynamic_range
21286 measured dynamic range of audio in dB
21287 Entropy
21289 entropy measured across whole audio, entropy of value near 1.0 is
21290 typically measured for white noise
21291 Flat_factor
21293 flatness (i.e. consecutive samples with the same value) of the
21294 signal at its peak levels (i.e. either Min_level or Max_level)
21295 Max_difference
21297 maximal difference between two consecutive samples
21298 Max_level
21300 maximal sample level
21301 Mean_difference
21303 mean difference between two consecutive samples, i.e. the average
21304 of each difference between two consecutive samples
21305 Min_difference
21307 minimal difference between two consecutive samples
21308 Min_level
21310 minimal sample level
21311 Noise_floor
21313 minimum local peak measured in dBFS over a short window
21314 Noise_floor_count
21316 number of occasions (not the number of samples) that the signal
21317 attained Noise floor
21318 Number_of_Infs
21320 number of samples with an infinite value
21321 Number_of_NaNs
21323 number of samples with a NaN (not a number) value
21324 Number_of_denormals
21326 number of samples with a subnormal value
21327 Number_of_samples
21329 number of samples
21330 Peak_count
21332 number of occasions (not the number of samples) that the signal
21333 attained either Min_level or Max_level
21334 Peak_level
21336 standard peak level measured in dBFS
21337 RMS_difference
21339 Root Mean Square difference between two consecutive samples
21340 RMS_level
21342 standard RMS level measured in dBFS
21343 RMS_peak
21345 RMS_trough
21346 peak and trough values for RMS level measured over a short window,
21347 measured in dBFS.
21348 Zero crossings
21350 number of points where the waveform crosses the zero level axis
21351 Zero crossings rate
21353 rate of Zero crossings and number of audio samples
21354 asubboost
21356 Boost subwoofer frequencies.
21357 The filter accepts the following options:
21359 dry Set dry gain, how much of original signal is kept. Allowed range is
21361 from 0 to 1. Default value is 1.0.
21362 wet Set wet gain, how much of filtered signal is kept. Allowed range is
21364 from 0 to 1. Default value is 1.0.
21365 boost
21367 Set max boost factor. Allowed range is from 1 to 12. Default value
21368 is 2.
21369 decay
21371 Set delay line decay gain value. Allowed range is from 0 to 1.
21372 Default value is 0.0.
21373 feedback
21375 Set delay line feedback gain value. Allowed range is from 0 to 1.
21376 Default value is 0.9.
21377 cutoff
21379 Set cutoff frequency in Hertz. Allowed range is 50 to 900. Default
21380 value is 100.
21381 slope
21383 Set slope amount for cutoff frequency. Allowed range is 0.0001 to
21384 1. Default value is 0.5.
21385 delay
21387 Set delay. Allowed range is from 1 to 100. Default value is 20.
21388 channels
21390 Set the channels to process. Default value is all available.
21391 Commands
21393 This filter supports the all above options as commands.
21395 asubcut
21397 Cut subwoofer frequencies.
21398 This filter allows to set custom, steeper roll off than highpass
21400 filter, and thus is able to more attenuate frequency content in stop-
21401 band.
21402 The filter accepts the following options:
21404 cutoff
21406 Set cutoff frequency in Hertz. Allowed range is 2 to 200. Default
21407 value is 20.
21408 order
21410 Set filter order. Available values are from 3 to 20. Default value
21411 is 10.
21412 level
21414 Set input gain level. Allowed range is from 0 to 1. Default value
21415 is 1.
21416 Commands
21418 This filter supports the all above options as commands.
21420 asupercut
21422 Cut super frequencies.
21423 The filter accepts the following options:
21425 cutoff
21427 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
21428 Default value is 20000.
21429 order
21431 Set filter order. Available values are from 3 to 20. Default value
21432 is 10.
21433 level
21435 Set input gain level. Allowed range is from 0 to 1. Default value
21436 is 1.
21437 Commands
21439 This filter supports the all above options as commands.
21441 asuperpass
21443 Apply high order Butterworth band-pass filter.
21444 The filter accepts the following options:
21446 centerf
21448 Set center frequency in Hertz. Allowed range is 2 to 999999.
21449 Default value is 1000.
21450 order
21452 Set filter order. Available values are from 4 to 20. Default value
21453 is 4.
21454 qfactor
21456 Set Q-factor. Allowed range is from 0.01 to 100. Default value is
21457 1.
21458 level
21460 Set input gain level. Allowed range is from 0 to 2. Default value
21461 is 1.
21462 Commands
21464 This filter supports the all above options as commands.
21466 asuperstop
21468 Apply high order Butterworth band-stop filter.
21469 The filter accepts the following options:
21471 centerf
21473 Set center frequency in Hertz. Allowed range is 2 to 999999.
21474 Default value is 1000.
21475 order
21477 Set filter order. Available values are from 4 to 20. Default value
21478 is 4.
21479 qfactor
21481 Set Q-factor. Allowed range is from 0.01 to 100. Default value is
21482 1.
21483 level
21485 Set input gain level. Allowed range is from 0 to 2. Default value
21486 is 1.
21487 Commands
21489 This filter supports the all above options as commands.
21491 atempo
21493 Adjust audio tempo.
21494 The filter accepts exactly one parameter, the audio tempo. If not
21496 specified then the filter will assume nominal 1.0 tempo. Tempo must be
21497 in the [0.5, 100.0] range.
21498 Note that tempo greater than 2 will skip some samples rather than blend
21500 them in. If for any reason this is a concern it is always possible to
21501 daisy-chain several instances of atempo to achieve the desired product
21502 tempo.
21503 Examples
21505 • Slow down audio to 80% tempo:
21507 atempo=0.8
21509 • To speed up audio to 300% tempo:
21511 atempo=3
21513 • To speed up audio to 300% tempo by daisy-chaining two atempo
21515 instances:
21516 atempo=sqrt(3),atempo=sqrt(3)
21518 Commands
21520 This filter supports the following commands:
21522 tempo
21524 Change filter tempo scale factor. Syntax for the command is :
21525 "tempo"
21526 atilt
21528 Apply spectral tilt filter to audio stream.
21529 This filter apply any spectral roll-off slope over any specified
21531 frequency band.
21532 The filter accepts the following options:
21534 freq
21536 Set central frequency of tilt in Hz. Default is 10000 Hz.
21537 slope
21539 Set slope direction of tilt. Default is 0. Allowed range is from -1
21540 to 1.
21541 width
21543 Set width of tilt. Default is 1000. Allowed range is from 100 to
21544 10000.
21545 order
21547 Set order of tilt filter.
21548 level
21550 Set input volume level. Allowed range is from 0 to 4. Defalt is 1.
21551 Commands
21553 This filter supports the all above options as commands.
21555 atrim
21557 Trim the input so that the output contains one continuous subpart of
21558 the input.
21559 It accepts the following parameters:
21561 start
21563 Timestamp (in seconds) of the start of the section to keep. I.e.
21564 the audio sample with the timestamp start will be the first sample
21565 in the output.
21566 end Specify time of the first audio sample that will be dropped, i.e.
21568 the audio sample immediately preceding the one with the timestamp
21569 end will be the last sample in the output.
21570 start_pts
21572 Same as start, except this option sets the start timestamp in
21573 samples instead of seconds.
21574 end_pts
21576 Same as end, except this option sets the end timestamp in samples
21577 instead of seconds.
21578 duration
21580 The maximum duration of the output in seconds.
21581 start_sample
21583 The number of the first sample that should be output.
21584 end_sample
21586 The number of the first sample that should be dropped.
21587 start, end, and duration are expressed as time duration specifications;
21589 see the Time duration section in the ffmpeg-utils(1) manual.
21590 Note that the first two sets of the start/end options and the duration
21592 option look at the frame timestamp, while the _sample options simply
21593 count the samples that pass through the filter. So start/end_pts and
21594 start/end_sample will give different results when the timestamps are
21595 wrong, inexact or do not start at zero. Also note that this filter does
21596 not modify the timestamps. If you wish to have the output timestamps
21597 start at zero, insert the asetpts filter after the atrim filter.
21598 If multiple start or end options are set, this filter tries to be
21600 greedy and keep all samples that match at least one of the specified
21601 constraints. To keep only the part that matches all the constraints at
21602 once, chain multiple atrim filters.
21603 The defaults are such that all the input is kept. So it is possible to
21605 set e.g. just the end values to keep everything before the specified
21606 time.
21607 Examples:
21609 • Drop everything except the second minute of input:
21611 ffmpeg -i INPUT -af atrim=60:120
21613 • Keep only the first 1000 samples:
21615 ffmpeg -i INPUT -af atrim=end_sample=1000
21617 axcorrelate
21619 Calculate normalized windowed cross-correlation between two input audio
21620 streams.
21621 Resulted samples are always between -1 and 1 inclusive. If result is 1
21623 it means two input samples are highly correlated in that selected
21624 segment. Result 0 means they are not correlated at all. If result is
21625 -1 it means two input samples are out of phase, which means they cancel
21626 each other.
21627 The filter accepts the following options:
21629 size
21631 Set size of segment over which cross-correlation is calculated.
21632 Default is 256. Allowed range is from 2 to 131072.
21633 algo
21635 Set algorithm for cross-correlation. Can be "slow" or "fast".
21636 Default is "slow". Fast algorithm assumes mean values over any
21637 given segment are always zero and thus need much less calculations
21638 to make. This is generally not true, but is valid for typical
21639 audio streams.
21640 Examples
21642 • Calculate correlation between channels in stereo audio stream:
21644 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
21646 bandpass
21648 Apply a two-pole Butterworth band-pass filter with central frequency
21649 frequency, and (3dB-point) band-width width. The csg option selects a
21650 constant skirt gain (peak gain = Q) instead of the default: constant
21651 0dB peak gain. The filter roll off at 6dB per octave (20dB per
21652 decade).
21653 The filter accepts the following options:
21655 frequency, f
21657 Set the filter's central frequency. Default is 3000.
21658 csg Constant skirt gain if set to 1. Defaults to 0.
21660 width_type, t
21662 Set method to specify band-width of filter.
21663 h Hz
21665 q Q-Factor
21667 o octave
21669 s slope
21671 k kHz
21673 width, w
21675 Specify the band-width of a filter in width_type units.
21676 mix, m
21678 How much to use filtered signal in output. Default is 1. Range is
21679 between 0 and 1.
21680 channels, c
21682 Specify which channels to filter, by default all available are
21683 filtered.
21684 normalize, n
21686 Normalize biquad coefficients, by default is disabled. Enabling it
21687 will normalize magnitude response at DC to 0dB.
21688 transform, a
21690 Set transform type of IIR filter.
21691 di
21693 dii
21694 tdi
21695 tdii
21696 latt
21697 svf
21698 zdf
21699 precision, r
21700 Set precison of filtering.
21701 auto
21703 Pick automatic sample format depending on surround filters.
21704 s16 Always use signed 16-bit.
21706 s32 Always use signed 32-bit.
21708 f32 Always use float 32-bit.
21710 f64 Always use float 64-bit.
21712 block_size, b
21714 Set block size used for reverse IIR processing. If this value is
21715 set to high enough value (higher than impulse response length
21716 truncated when reaches near zero values) filtering will become
21717 linear phase otherwise if not big enough it will just produce nasty
21718 artifacts.
21719 Note that filter delay will be exactly this many samples when set
21721 to non-zero value.
21722 Commands
21724 This filter supports the following commands:
21726 frequency, f
21728 Change bandpass frequency. Syntax for the command is : "frequency"
21729 width_type, t
21731 Change bandpass width_type. Syntax for the command is :
21732 "width_type"
21733 width, w
21735 Change bandpass width. Syntax for the command is : "width"
21736 mix, m
21738 Change bandpass mix. Syntax for the command is : "mix"
21739 bandreject
21741 Apply a two-pole Butterworth band-reject filter with central frequency
21742 frequency, and (3dB-point) band-width width. The filter roll off at
21743 6dB per octave (20dB per decade).
21744 The filter accepts the following options:
21746 frequency, f
21748 Set the filter's central frequency. Default is 3000.
21749 width_type, t
21751 Set method to specify band-width of filter.
21752 h Hz
21754 q Q-Factor
21756 o octave
21758 s slope
21760 k kHz
21762 width, w
21764 Specify the band-width of a filter in width_type units.
21765 mix, m
21767 How much to use filtered signal in output. Default is 1. Range is
21768 between 0 and 1.
21769 channels, c
21771 Specify which channels to filter, by default all available are
21772 filtered.
21773 normalize, n
21775 Normalize biquad coefficients, by default is disabled. Enabling it
21776 will normalize magnitude response at DC to 0dB.
21777 transform, a
21779 Set transform type of IIR filter.
21780 di
21782 dii
21783 tdi
21784 tdii
21785 latt
21786 svf
21787 zdf
21788 precision, r
21789 Set precison of filtering.
21790 auto
21792 Pick automatic sample format depending on surround filters.
21793 s16 Always use signed 16-bit.
21795 s32 Always use signed 32-bit.
21797 f32 Always use float 32-bit.
21799 f64 Always use float 64-bit.
21801 block_size, b
21803 Set block size used for reverse IIR processing. If this value is
21804 set to high enough value (higher than impulse response length
21805 truncated when reaches near zero values) filtering will become
21806 linear phase otherwise if not big enough it will just produce nasty
21807 artifacts.
21808 Note that filter delay will be exactly this many samples when set
21810 to non-zero value.
21811 Commands
21813 This filter supports the following commands:
21815 frequency, f
21817 Change bandreject frequency. Syntax for the command is :
21818 "frequency"
21819 width_type, t
21821 Change bandreject width_type. Syntax for the command is :
21822 "width_type"
21823 width, w
21825 Change bandreject width. Syntax for the command is : "width"
21826 mix, m
21828 Change bandreject mix. Syntax for the command is : "mix"
21829 bass, lowshelf
21831 Boost or cut the bass (lower) frequencies of the audio using a two-pole
21832 shelving filter with a response similar to that of a standard hi-fi's
21833 tone-controls. This is also known as shelving equalisation (EQ).
21834 The filter accepts the following options:
21836 gain, g
21838 Give the gain at 0 Hz. Its useful range is about -20 (for a large
21839 cut) to +20 (for a large boost). Beware of clipping when using a
21840 positive gain.
21841 frequency, f
21843 Set the filter's central frequency and so can be used to extend or
21844 reduce the frequency range to be boosted or cut. The default value
21845 is 100 Hz.
21846 width_type, t
21848 Set method to specify band-width of filter.
21849 h Hz
21851 q Q-Factor
21853 o octave
21855 s slope
21857 k kHz
21859 width, w
21861 Determine how steep is the filter's shelf transition.
21862 poles, p
21864 Set number of poles. Default is 2.
21865 mix, m
21867 How much to use filtered signal in output. Default is 1. Range is
21868 between 0 and 1.
21869 channels, c
21871 Specify which channels to filter, by default all available are
21872 filtered.
21873 normalize, n
21875 Normalize biquad coefficients, by default is disabled. Enabling it
21876 will normalize magnitude response at DC to 0dB.
21877 transform, a
21879 Set transform type of IIR filter.
21880 di
21882 dii
21883 tdi
21884 tdii
21885 latt
21886 svf
21887 zdf
21888 precision, r
21889 Set precison of filtering.
21890 auto
21892 Pick automatic sample format depending on surround filters.
21893 s16 Always use signed 16-bit.
21895 s32 Always use signed 32-bit.
21897 f32 Always use float 32-bit.
21899 f64 Always use float 64-bit.
21901 block_size, b
21903 Set block size used for reverse IIR processing. If this value is
21904 set to high enough value (higher than impulse response length
21905 truncated when reaches near zero values) filtering will become
21906 linear phase otherwise if not big enough it will just produce nasty
21907 artifacts.
21908 Note that filter delay will be exactly this many samples when set
21910 to non-zero value.
21911 Commands
21913 This filter supports the following commands:
21915 frequency, f
21917 Change bass frequency. Syntax for the command is : "frequency"
21918 width_type, t
21920 Change bass width_type. Syntax for the command is : "width_type"
21921 width, w
21923 Change bass width. Syntax for the command is : "width"
21924 gain, g
21926 Change bass gain. Syntax for the command is : "gain"
21927 mix, m
21929 Change bass mix. Syntax for the command is : "mix"
21930 biquad
21932 Apply a biquad IIR filter with the given coefficients. Where b0, b1,
21933 b2 and a0, a1, a2 are the numerator and denominator coefficients
21934 respectively. and channels, c specify which channels to filter, by
21935 default all available are filtered.
21936 Commands
21938 This filter supports the following commands:
21940 a0
21942 a1
21943 a2
21944 b0
21945 b1
21946 b2 Change biquad parameter. Syntax for the command is : "value"
21947 mix, m
21949 How much to use filtered signal in output. Default is 1. Range is
21950 between 0 and 1.
21951 channels, c
21953 Specify which channels to filter, by default all available are
21954 filtered.
21955 normalize, n
21957 Normalize biquad coefficients, by default is disabled. Enabling it
21958 will normalize magnitude response at DC to 0dB.
21959 transform, a
21961 Set transform type of IIR filter.
21962 di
21964 dii
21965 tdi
21966 tdii
21967 latt
21968 svf
21969 zdf
21970 precision, r
21971 Set precison of filtering.
21972 auto
21974 Pick automatic sample format depending on surround filters.
21975 s16 Always use signed 16-bit.
21977 s32 Always use signed 32-bit.
21979 f32 Always use float 32-bit.
21981 f64 Always use float 64-bit.
21983 block_size, b
21985 Set block size used for reverse IIR processing. If this value is
21986 set to high enough value (higher than impulse response length
21987 truncated when reaches near zero values) filtering will become
21988 linear phase otherwise if not big enough it will just produce nasty
21989 artifacts.
21990 Note that filter delay will be exactly this many samples when set
21992 to non-zero value.
21993 bs2b
21995 Bauer stereo to binaural transformation, which improves headphone
21996 listening of stereo audio records.
21997 To enable compilation of this filter you need to configure FFmpeg with
21999 "--enable-libbs2b".
22000 It accepts the following parameters:
22002 profile
22004 Pre-defined crossfeed level.
22005 default
22007 Default level (fcut=700, feed=50).
22008 cmoy
22010 Chu Moy circuit (fcut=700, feed=60).
22011 jmeier
22013 Jan Meier circuit (fcut=650, feed=95).
22014 fcut
22016 Cut frequency (in Hz).
22017 feed
22019 Feed level (in Hz).
22020 channelmap
22022 Remap input channels to new locations.
22023 It accepts the following parameters:
22025 map Map channels from input to output. The argument is a '|'-separated
22027 list of mappings, each in the "in_channel-out_channel" or
22028 in_channel form. in_channel can be either the name of the input
22029 channel (e.g. FL for front left) or its index in the input channel
22030 layout. out_channel is the name of the output channel or its index
22031 in the output channel layout. If out_channel is not given then it
22032 is implicitly an index, starting with zero and increasing by one
22033 for each mapping.
22034 channel_layout
22036 The channel layout of the output stream.
22037 If no mapping is present, the filter will implicitly map input channels
22039 to output channels, preserving indices.
22040 Examples
22042 • For example, assuming a 5.1+downmix input MOV file,
22044 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
22046 will create an output WAV file tagged as stereo from the downmix
22048 channels of the input.
22049 • To fix a 5.1 WAV improperly encoded in AAC's native channel order
22051 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
22053 channelsplit
22055 Split each channel from an input audio stream into a separate output
22056 stream.
22057 It accepts the following parameters:
22059 channel_layout
22061 The channel layout of the input stream. The default is "stereo".
22062 channels
22064 A channel layout describing the channels to be extracted as
22065 separate output streams or "all" to extract each input channel as a
22066 separate stream. The default is "all".
22067 Choosing channels not present in channel layout in the input will
22069 result in an error.
22070 Examples
22072 • For example, assuming a stereo input MP3 file,
22074 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
22076 will create an output Matroska file with two audio streams, one
22078 containing only the left channel and the other the right channel.
22079 • Split a 5.1 WAV file into per-channel files:
22081 ffmpeg -i in.wav -filter_complex
22083 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
22084 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
22085 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
22086 side_right.wav
22087 • Extract only LFE from a 5.1 WAV file:
22089 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
22091 -map '[LFE]' lfe.wav
22092 chorus
22094 Add a chorus effect to the audio.
22095 Can make a single vocal sound like a chorus, but can also be applied to
22097 instrumentation.
22098 Chorus resembles an echo effect with a short delay, but whereas with
22100 echo the delay is constant, with chorus, it is varied using using
22101 sinusoidal or triangular modulation. The modulation depth defines the
22102 range the modulated delay is played before or after the delay. Hence
22103 the delayed sound will sound slower or faster, that is the delayed
22104 sound tuned around the original one, like in a chorus where some vocals
22105 are slightly off key.
22106 It accepts the following parameters:
22108 in_gain
22110 Set input gain. Default is 0.4.
22111 out_gain
22113 Set output gain. Default is 0.4.
22114 delays
22116 Set delays. A typical delay is around 40ms to 60ms.
22117 decays
22119 Set decays.
22120 speeds
22122 Set speeds.
22123 depths
22125 Set depths.
22126 Examples
22128 • A single delay:
22130 chorus=0.7:0.9:55:0.4:0.25:2
22132 • Two delays:
22134 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
22136 • Fuller sounding chorus with three delays:
22138 chorus=0.5:0.9:50|60|40:0.4|0.32|0.3:0.25|0.4|0.3:2|2.3|1.3
22140 compand
22142 Compress or expand the audio's dynamic range.
22143 It accepts the following parameters:
22145 attacks
22147 decays
22148 A list of times in seconds for each channel over which the
22149 instantaneous level of the input signal is averaged to determine
22150 its volume. attacks refers to increase of volume and decays refers
22151 to decrease of volume. For most situations, the attack time
22152 (response to the audio getting louder) should be shorter than the
22153 decay time, because the human ear is more sensitive to sudden loud
22154 audio than sudden soft audio. A typical value for attack is 0.3
22155 seconds and a typical value for decay is 0.8 seconds. If specified
22156 number of attacks & decays is lower than number of channels, the
22157 last set attack/decay will be used for all remaining channels.
22158 points
22160 A list of points for the transfer function, specified in dB
22161 relative to the maximum possible signal amplitude. Each key points
22162 list must be defined using the following syntax:
22163 "x0/y0|x1/y1|x2/y2|...." or "x0/y0 x1/y1 x2/y2 ...."
22164 The input values must be in strictly increasing order but the
22166 transfer function does not have to be monotonically rising. The
22167 point "0/0" is assumed but may be overridden (by "0/out-dBn").
22168 Typical values for the transfer function are "-70/-70|-60/-20|1/0".
22169 soft-knee
22171 Set the curve radius in dB for all joints. It defaults to 0.01.
22172 gain
22174 Set the additional gain in dB to be applied at all points on the
22175 transfer function. This allows for easy adjustment of the overall
22176 gain. It defaults to 0.
22177 volume
22179 Set an initial volume, in dB, to be assumed for each channel when
22180 filtering starts. This permits the user to supply a nominal level
22181 initially, so that, for example, a very large gain is not applied
22182 to initial signal levels before the companding has begun to
22183 operate. A typical value for audio which is initially quiet is -90
22184 dB. It defaults to 0.
22185 delay
22187 Set a delay, in seconds. The input audio is analyzed immediately,
22188 but audio is delayed before being fed to the volume adjuster.
22189 Specifying a delay approximately equal to the attack/decay times
22190 allows the filter to effectively operate in predictive rather than
22191 reactive mode. It defaults to 0.
22192 Examples
22194 • Make music with both quiet and loud passages suitable for listening
22196 to in a noisy environment:
22197 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
22199 Another example for audio with whisper and explosion parts:
22201 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
22203 • A noise gate for when the noise is at a lower level than the
22205 signal:
22206 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
22208 • Here is another noise gate, this time for when the noise is at a
22210 higher level than the signal (making it, in some ways, similar to
22211 squelch):
22212 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
22214 • 2:1 compression starting at -6dB:
22216 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
22218 • 2:1 compression starting at -9dB:
22220 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
22222 • 2:1 compression starting at -12dB:
22224 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
22226 • 2:1 compression starting at -18dB:
22228 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
22230 • 3:1 compression starting at -15dB:
22232 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
22234 • Compressor/Gate:
22236 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
22238 • Expander:
22240 compand=attacks=0:points=-80/-169|-54/-80|-49.5/-64.6|-41.1/-41.1|-25.8/-15|-10.8/-4.5|0/0|20/8.3
22242 • Hard limiter at -6dB:
22244 compand=attacks=0:points=-80/-80|-6/-6|20/-6
22246 • Hard limiter at -12dB:
22248 compand=attacks=0:points=-80/-80|-12/-12|20/-12
22250 • Hard noise gate at -35 dB:
22252 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
22254 • Soft limiter:
22256 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
22258 compensationdelay
22260 Compensation Delay Line is a metric based delay to compensate differing
22261 positions of microphones or speakers.
22262 For example, you have recorded guitar with two microphones placed in
22264 different locations. Because the front of sound wave has fixed speed in
22265 normal conditions, the phasing of microphones can vary and depends on
22266 their location and interposition. The best sound mix can be achieved
22267 when these microphones are in phase (synchronized). Note that a
22268 distance of ~30 cm between microphones makes one microphone capture the
22269 signal in antiphase to the other microphone. That makes the final mix
22270 sound moody. This filter helps to solve phasing problems by adding
22271 different delays to each microphone track and make them synchronized.
22272 The best result can be reached when you take one track as base and
22274 synchronize other tracks one by one with it. Remember that
22275 synchronization/delay tolerance depends on sample rate, too. Higher
22276 sample rates will give more tolerance.
22277 The filter accepts the following parameters:
22279 mm Set millimeters distance. This is compensation distance for fine
22281 tuning. Default is 0.
22282 cm Set cm distance. This is compensation distance for tightening
22284 distance setup. Default is 0.
22285 m Set meters distance. This is compensation distance for hard
22287 distance setup. Default is 0.
22288 dry Set dry amount. Amount of unprocessed (dry) signal. Default is 0.
22290 wet Set wet amount. Amount of processed (wet) signal. Default is 1.
22292 temp
22294 Set temperature in degrees Celsius. This is the temperature of the
22295 environment. Default is 20.
22296 Commands
22298 This filter supports the all above options as commands.
22300 crossfeed
22302 Apply headphone crossfeed filter.
22303 Crossfeed is the process of blending the left and right channels of
22305 stereo audio recording. It is mainly used to reduce extreme stereo
22306 separation of low frequencies.
22307 The intent is to produce more speaker like sound to the listener.
22309 The filter accepts the following options:
22311 strength
22313 Set strength of crossfeed. Default is 0.2. Allowed range is from 0
22314 to 1. This sets gain of low shelf filter for side part of stereo
22315 image. Default is -6dB. Max allowed is -30db when strength is set
22316 to 1.
22317 range
22319 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to
22320 1. This sets cut off frequency of low shelf filter. Default is cut
22321 off near 1550 Hz. With range set to 1 cut off frequency is set to
22322 2100 Hz.
22323 slope
22325 Set curve slope of low shelf filter. Default is 0.5. Allowed range
22326 is from 0.01 to 1.
22327 level_in
22329 Set input gain. Default is 0.9.
22330 level_out
22332 Set output gain. Default is 1.
22333 block_size
22335 Set block size used for reverse IIR processing. If this value is
22336 set to high enough value (higher than impulse response length
22337 truncated when reaches near zero values) filtering will become
22338 linear phase otherwise if not big enough it will just produce nasty
22339 artifacts.
22340 Note that filter delay will be exactly this many samples when set
22342 to non-zero value.
22343 Commands
22345 This filter supports the all above options as commands.
22347 crystalizer
22349 Simple algorithm for audio noise sharpening.
22350 This filter linearly increases differences betweeen each audio sample.
22352 The filter accepts the following options:
22354 i Sets the intensity of effect (default: 2.0). Must be in range
22356 between -10.0 to 0 (unchanged sound) to 10.0 (maximum effect). To
22357 inverse filtering use negative value.
22358 c Enable clipping. By default is enabled.
22360 Commands
22362 This filter supports the all above options as commands.
22364 dcshift
22366 Apply a DC shift to the audio.
22367 This can be useful to remove a DC offset (caused perhaps by a hardware
22369 problem in the recording chain) from the audio. The effect of a DC
22370 offset is reduced headroom and hence volume. The astats filter can be
22371 used to determine if a signal has a DC offset.
22372 shift
22374 Set the DC shift, allowed range is [-1, 1]. It indicates the amount
22375 to shift the audio.
22376 limitergain
22378 Optional. It should have a value much less than 1 (e.g. 0.05 or
22379 0.02) and is used to prevent clipping.
22380 deesser
22382 Apply de-essing to the audio samples.
22383 i Set intensity for triggering de-essing. Allowed range is from 0 to
22385 1. Default is 0.
22386 m Set amount of ducking on treble part of sound. Allowed range is
22388 from 0 to 1. Default is 0.5.
22389 f How much of original frequency content to keep when de-essing.
22391 Allowed range is from 0 to 1. Default is 0.5.
22392 s Set the output mode.
22394 It accepts the following values:
22396 i Pass input unchanged.
22398 o Pass ess filtered out.
22400 e Pass only ess.
22402 Default value is o.
22404 dialoguenhance
22406 Enhance dialogue in stereo audio.
22407 This filter accepts stereo input and produce surround (3.0) channels
22409 output. The newly produced front center channel have enhanced speech
22410 dialogue originally available in both stereo channels. This filter
22411 outputs front left and front right channels same as available in stereo
22412 input.
22413 The filter accepts the following options:
22415 original
22417 Set the original center factor to keep in front center channel
22418 output. Allowed range is from 0 to 1. Default value is 1.
22419 enhance
22421 Set the dialogue enhance factor to put in front center channel
22422 output. Allowed range is from 0 to 3. Default value is 1.
22423 voice
22425 Set the voice detection factor. Allowed range is from 2 to 32.
22426 Default value is 2.
22427 Commands
22429 This filter supports the all above options as commands.
22431 drmeter
22433 Measure audio dynamic range.
22434 DR values of 14 and higher is found in very dynamic material. DR of 8
22436 to 13 is found in transition material. And anything less that 8 have
22437 very poor dynamics and is very compressed.
22438 The filter accepts the following options:
22440 length
22442 Set window length in seconds used to split audio into segments of
22443 equal length. Default is 3 seconds.
22444 dynaudnorm
22446 Dynamic Audio Normalizer.
22447 This filter applies a certain amount of gain to the input audio in
22449 order to bring its peak magnitude to a target level (e.g. 0 dBFS).
22450 However, in contrast to more "simple" normalization algorithms, the
22451 Dynamic Audio Normalizer *dynamically* re-adjusts the gain factor to
22452 the input audio. This allows for applying extra gain to the "quiet"
22453 sections of the audio while avoiding distortions or clipping the "loud"
22454 sections. In other words: The Dynamic Audio Normalizer will "even out"
22455 the volume of quiet and loud sections, in the sense that the volume of
22456 each section is brought to the same target level. Note, however, that
22457 the Dynamic Audio Normalizer achieves this goal *without* applying
22458 "dynamic range compressing". It will retain 100% of the dynamic range
22459 *within* each section of the audio file.
22460 framelen, f
22462 Set the frame length in milliseconds. In range from 10 to 8000
22463 milliseconds. Default is 500 milliseconds. The Dynamic Audio
22464 Normalizer processes the input audio in small chunks, referred to
22465 as frames. This is required, because a peak magnitude has no
22466 meaning for just a single sample value. Instead, we need to
22467 determine the peak magnitude for a contiguous sequence of sample
22468 values. While a "standard" normalizer would simply use the peak
22469 magnitude of the complete file, the Dynamic Audio Normalizer
22470 determines the peak magnitude individually for each frame. The
22471 length of a frame is specified in milliseconds. By default, the
22472 Dynamic Audio Normalizer uses a frame length of 500 milliseconds,
22473 which has been found to give good results with most files. Note
22474 that the exact frame length, in number of samples, will be
22475 determined automatically, based on the sampling rate of the
22476 individual input audio file.
22477 gausssize, g
22479 Set the Gaussian filter window size. In range from 3 to 301, must
22480 be odd number. Default is 31. Probably the most important
22481 parameter of the Dynamic Audio Normalizer is the "window size" of
22482 the Gaussian smoothing filter. The filter's window size is
22483 specified in frames, centered around the current frame. For the
22484 sake of simplicity, this must be an odd number. Consequently, the
22485 default value of 31 takes into account the current frame, as well
22486 as the 15 preceding frames and the 15 subsequent frames. Using a
22487 larger window results in a stronger smoothing effect and thus in
22488 less gain variation, i.e. slower gain adaptation. Conversely, using
22489 a smaller window results in a weaker smoothing effect and thus in
22490 more gain variation, i.e. faster gain adaptation. In other words,
22491 the more you increase this value, the more the Dynamic Audio
22492 Normalizer will behave like a "traditional" normalization filter.
22493 On the contrary, the more you decrease this value, the more the
22494 Dynamic Audio Normalizer will behave like a dynamic range
22495 compressor.
22496 peak, p
22498 Set the target peak value. This specifies the highest permissible
22499 magnitude level for the normalized audio input. This filter will
22500 try to approach the target peak magnitude as closely as possible,
22501 but at the same time it also makes sure that the normalized signal
22502 will never exceed the peak magnitude. A frame's maximum local gain
22503 factor is imposed directly by the target peak magnitude. The
22504 default value is 0.95 and thus leaves a headroom of 5%*. It is not
22505 recommended to go above this value.
22506 maxgain, m
22508 Set the maximum gain factor. In range from 1.0 to 100.0. Default is
22509 10.0. The Dynamic Audio Normalizer determines the maximum possible
22510 (local) gain factor for each input frame, i.e. the maximum gain
22511 factor that does not result in clipping or distortion. The maximum
22512 gain factor is determined by the frame's highest magnitude sample.
22513 However, the Dynamic Audio Normalizer additionally bounds the
22514 frame's maximum gain factor by a predetermined (global) maximum
22515 gain factor. This is done in order to avoid excessive gain factors
22516 in "silent" or almost silent frames. By default, the maximum gain
22517 factor is 10.0, For most inputs the default value should be
22518 sufficient and it usually is not recommended to increase this
22519 value. Though, for input with an extremely low overall volume
22520 level, it may be necessary to allow even higher gain factors. Note,
22521 however, that the Dynamic Audio Normalizer does not simply apply a
22522 "hard" threshold (i.e. cut off values above the threshold).
22523 Instead, a "sigmoid" threshold function will be applied. This way,
22524 the gain factors will smoothly approach the threshold value, but
22525 never exceed that value.
22526 targetrms, r
22528 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 -
22529 disabled. By default, the Dynamic Audio Normalizer performs "peak"
22530 normalization. This means that the maximum local gain factor for
22531 each frame is defined (only) by the frame's highest magnitude
22532 sample. This way, the samples can be amplified as much as possible
22533 without exceeding the maximum signal level, i.e. without clipping.
22534 Optionally, however, the Dynamic Audio Normalizer can also take
22535 into account the frame's root mean square, abbreviated RMS. In
22536 electrical engineering, the RMS is commonly used to determine the
22537 power of a time-varying signal. It is therefore considered that the
22538 RMS is a better approximation of the "perceived loudness" than just
22539 looking at the signal's peak magnitude. Consequently, by adjusting
22540 all frames to a constant RMS value, a uniform "perceived loudness"
22541 can be established. If a target RMS value has been specified, a
22542 frame's local gain factor is defined as the factor that would
22543 result in exactly that RMS value. Note, however, that the maximum
22544 local gain factor is still restricted by the frame's highest
22545 magnitude sample, in order to prevent clipping.
22546 coupling, n
22548 Enable channels coupling. By default is enabled. By default, the
22549 Dynamic Audio Normalizer will amplify all channels by the same
22550 amount. This means the same gain factor will be applied to all
22551 channels, i.e. the maximum possible gain factor is determined by
22552 the "loudest" channel. However, in some recordings, it may happen
22553 that the volume of the different channels is uneven, e.g. one
22554 channel may be "quieter" than the other one(s). In this case, this
22555 option can be used to disable the channel coupling. This way, the
22556 gain factor will be determined independently for each channel,
22557 depending only on the individual channel's highest magnitude
22558 sample. This allows for harmonizing the volume of the different
22559 channels.
22560 correctdc, c
22562 Enable DC bias correction. By default is disabled. An audio signal
22563 (in the time domain) is a sequence of sample values. In the
22564 Dynamic Audio Normalizer these sample values are represented in the
22565 -1.0 to 1.0 range, regardless of the original input format.
22566 Normally, the audio signal, or "waveform", should be centered
22567 around the zero point. That means if we calculate the mean value
22568 of all samples in a file, or in a single frame, then the result
22569 should be 0.0 or at least very close to that value. If, however,
22570 there is a significant deviation of the mean value from 0.0, in
22571 either positive or negative direction, this is referred to as a DC
22572 bias or DC offset. Since a DC bias is clearly undesirable, the
22573 Dynamic Audio Normalizer provides optional DC bias correction.
22574 With DC bias correction enabled, the Dynamic Audio Normalizer will
22575 determine the mean value, or "DC correction" offset, of each input
22576 frame and subtract that value from all of the frame's sample values
22577 which ensures those samples are centered around 0.0 again. Also, in
22578 order to avoid "gaps" at the frame boundaries, the DC correction
22579 offset values will be interpolated smoothly between neighbouring
22580 frames.
22581 altboundary, b
22583 Enable alternative boundary mode. By default is disabled. The
22584 Dynamic Audio Normalizer takes into account a certain neighbourhood
22585 around each frame. This includes the preceding frames as well as
22586 the subsequent frames. However, for the "boundary" frames, located
22587 at the very beginning and at the very end of the audio file, not
22588 all neighbouring frames are available. In particular, for the first
22589 few frames in the audio file, the preceding frames are not known.
22590 And, similarly, for the last few frames in the audio file, the
22591 subsequent frames are not known. Thus, the question arises which
22592 gain factors should be assumed for the missing frames in the
22593 "boundary" region. The Dynamic Audio Normalizer implements two
22594 modes to deal with this situation. The default boundary mode
22595 assumes a gain factor of exactly 1.0 for the missing frames,
22596 resulting in a smooth "fade in" and "fade out" at the beginning and
22597 at the end of the input, respectively.
22598 compress, s
22600 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
22601 By default, the Dynamic Audio Normalizer does not apply
22602 "traditional" compression. This means that signal peaks will not be
22603 pruned and thus the full dynamic range will be retained within each
22604 local neighbourhood. However, in some cases it may be desirable to
22605 combine the Dynamic Audio Normalizer's normalization algorithm with
22606 a more "traditional" compression. For this purpose, the Dynamic
22607 Audio Normalizer provides an optional compression (thresholding)
22608 function. If (and only if) the compression feature is enabled, all
22609 input frames will be processed by a soft knee thresholding function
22610 prior to the actual normalization process. Put simply, the
22611 thresholding function is going to prune all samples whose magnitude
22612 exceeds a certain threshold value. However, the Dynamic Audio
22613 Normalizer does not simply apply a fixed threshold value. Instead,
22614 the threshold value will be adjusted for each individual frame. In
22615 general, smaller parameters result in stronger compression, and
22616 vice versa. Values below 3.0 are not recommended, because audible
22617 distortion may appear.
22618 threshold, t
22620 Set the target threshold value. This specifies the lowest
22621 permissible magnitude level for the audio input which will be
22622 normalized. If input frame volume is above this value frame will
22623 be normalized. Otherwise frame may not be normalized at all. The
22624 default value is set to 0, which means all input frames will be
22625 normalized. This option is mostly useful if digital noise is not
22626 wanted to be amplified.
22627 channels, h
22629 Specify which channels to filter, by default all available channels
22630 are filtered.
22631 overlap, o
22633 Specify overlap for frames. If set to 0 (default) no frame
22634 overlapping is done. Using >0 and <1 values will make less
22635 conservative gain adjustments, like when framelen option is set to
22636 smaller value, if framelen option value is compensated for non-zero
22637 overlap then gain adjustments will be smoother across time compared
22638 to zero overlap case.
22639 curve, v
22641 Specify the peak mapping curve expression which is going to be used
22642 when calculating gain applied to frames. The max output frame gain
22643 will still be limited by other options mentioned previously for
22644 this filter.
22645 The expression can contain the following constants:
22647 ch current channel number
22649 sn current sample number
22651 nb_channels
22653 number of channels
22654 t timestamp expressed in seconds
22656 sr sample rate
22658 p current frame peak value
22660 Commands
22662 This filter supports the all above options as commands.
22664 earwax
22666 Make audio easier to listen to on headphones.
22667 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
22669 so that when listened to on headphones the stereo image is moved from
22670 inside your head (standard for headphones) to outside and in front of
22671 the listener (standard for speakers).
22672 Ported from SoX.
22674 equalizer
22676 Apply a two-pole peaking equalisation (EQ) filter. With this filter,
22677 the signal-level at and around a selected frequency can be increased or
22678 decreased, whilst (unlike bandpass and bandreject filters) that at all
22679 other frequencies is unchanged.
22680 In order to produce complex equalisation curves, this filter can be
22682 given several times, each with a different central frequency.
22683 The filter accepts the following options:
22685 frequency, f
22687 Set the filter's central frequency in Hz.
22688 width_type, t
22690 Set method to specify band-width of filter.
22691 h Hz
22693 q Q-Factor
22695 o octave
22697 s slope
22699 k kHz
22701 width, w
22703 Specify the band-width of a filter in width_type units.
22704 gain, g
22706 Set the required gain or attenuation in dB. Beware of clipping
22707 when using a positive gain.
22708 mix, m
22710 How much to use filtered signal in output. Default is 1. Range is
22711 between 0 and 1.
22712 channels, c
22714 Specify which channels to filter, by default all available are
22715 filtered.
22716 normalize, n
22718 Normalize biquad coefficients, by default is disabled. Enabling it
22719 will normalize magnitude response at DC to 0dB.
22720 transform, a
22722 Set transform type of IIR filter.
22723 di
22725 dii
22726 tdi
22727 tdii
22728 latt
22729 svf
22730 zdf
22731 precision, r
22732 Set precison of filtering.
22733 auto
22735 Pick automatic sample format depending on surround filters.
22736 s16 Always use signed 16-bit.
22738 s32 Always use signed 32-bit.
22740 f32 Always use float 32-bit.
22742 f64 Always use float 64-bit.
22744 block_size, b
22746 Set block size used for reverse IIR processing. If this value is
22747 set to high enough value (higher than impulse response length
22748 truncated when reaches near zero values) filtering will become
22749 linear phase otherwise if not big enough it will just produce nasty
22750 artifacts.
22751 Note that filter delay will be exactly this many samples when set
22753 to non-zero value.
22754 Examples
22756 • Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
22758 equalizer=f=1000:t=h:width=200:g=-10
22760 • Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz
22762 with Q 2:
22763 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
22765 Commands
22767 This filter supports the following commands:
22769 frequency, f
22771 Change equalizer frequency. Syntax for the command is :
22772 "frequency"
22773 width_type, t
22775 Change equalizer width_type. Syntax for the command is :
22776 "width_type"
22777 width, w
22779 Change equalizer width. Syntax for the command is : "width"
22780 gain, g
22782 Change equalizer gain. Syntax for the command is : "gain"
22783 mix, m
22785 Change equalizer mix. Syntax for the command is : "mix"
22786 extrastereo
22788 Linearly increases the difference between left and right channels which
22789 adds some sort of "live" effect to playback.
22790 The filter accepts the following options:
22792 m Sets the difference coefficient (default: 2.5). 0.0 means mono
22794 sound (average of both channels), with 1.0 sound will be unchanged,
22795 with -1.0 left and right channels will be swapped.
22796 c Enable clipping. By default is enabled.
22798 Commands
22800 This filter supports the all above options as commands.
22802 firequalizer
22804 Apply FIR Equalization using arbitrary frequency response.
22805 The filter accepts the following option:
22807 gain
22809 Set gain curve equation (in dB). The expression can contain
22810 variables:
22811 f the evaluated frequency
22813 sr sample rate
22815 ch channel number, set to 0 when multichannels evaluation is
22817 disabled
22818 chid
22820 channel id, see libavutil/channel_layout.h, set to the first
22821 channel id when multichannels evaluation is disabled
22822 chs number of channels
22824 chlayout
22826 channel_layout, see libavutil/channel_layout.h
22827 and functions:
22829 gain_interpolate(f)
22831 interpolate gain on frequency f based on gain_entry
22832 cubic_interpolate(f)
22834 same as gain_interpolate, but smoother
22835 This option is also available as command. Default is
22837 gain_interpolate(f).
22838 gain_entry
22840 Set gain entry for gain_interpolate function. The expression can
22841 contain functions:
22842 entry(f, g)
22844 store gain entry at frequency f with value g
22845 This option is also available as command.
22847 delay
22849 Set filter delay in seconds. Higher value means more accurate.
22850 Default is 0.01.
22851 accuracy
22853 Set filter accuracy in Hz. Lower value means more accurate.
22854 Default is 5.
22855 wfunc
22857 Set window function. Acceptable values are:
22858 rectangular
22860 rectangular window, useful when gain curve is already smooth
22861 hann
22863 hann window (default)
22864 hamming
22866 hamming window
22867 blackman
22869 blackman window
22870 nuttall3
22872 3-terms continuous 1st derivative nuttall window
22873 mnuttall3
22875 minimum 3-terms discontinuous nuttall window
22876 nuttall
22878 4-terms continuous 1st derivative nuttall window
22879 bnuttall
22881 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
22882 bharris
22884 blackman-harris window
22885 tukey
22887 tukey window
22888 fixed
22890 If enabled, use fixed number of audio samples. This improves speed
22891 when filtering with large delay. Default is disabled.
22892 multi
22894 Enable multichannels evaluation on gain. Default is disabled.
22895 zero_phase
22897 Enable zero phase mode by subtracting timestamp to compensate
22898 delay. Default is disabled.
22899 scale
22901 Set scale used by gain. Acceptable values are:
22902 linlin
22904 linear frequency, linear gain
22905 linlog
22907 linear frequency, logarithmic (in dB) gain (default)
22908 loglin
22910 logarithmic (in octave scale where 20 Hz is 0) frequency,
22911 linear gain
22912 loglog
22914 logarithmic frequency, logarithmic gain
22915 dumpfile
22917 Set file for dumping, suitable for gnuplot.
22918 dumpscale
22920 Set scale for dumpfile. Acceptable values are same with scale
22921 option. Default is linlog.
22922 fft2
22924 Enable 2-channel convolution using complex FFT. This improves speed
22925 significantly. Default is disabled.
22926 min_phase
22928 Enable minimum phase impulse response. Default is disabled.
22929 Examples
22931 • lowpass at 1000 Hz:
22933 firequalizer=gain='if(lt(f,1000), 0, -INF)'
22935 • lowpass at 1000 Hz with gain_entry:
22937 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
22939 • custom equalization:
22941 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
22943 • higher delay with zero phase to compensate delay:
22945 firequalizer=delay=0.1:fixed=on:zero_phase=on
22947 • lowpass on left channel, highpass on right channel:
22949 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
22951 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
22952 flanger
22954 Apply a flanging effect to the audio.
22955 The filter accepts the following options:
22957 delay
22959 Set base delay in milliseconds. Range from 0 to 30. Default value
22960 is 0.
22961 depth
22963 Set added sweep delay in milliseconds. Range from 0 to 10. Default
22964 value is 2.
22965 regen
22967 Set percentage regeneration (delayed signal feedback). Range from
22968 -95 to 95. Default value is 0.
22969 width
22971 Set percentage of delayed signal mixed with original. Range from 0
22972 to 100. Default value is 71.
22973 speed
22975 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is
22976 0.5.
22977 shape
22979 Set swept wave shape, can be triangular or sinusoidal. Default
22980 value is sinusoidal.
22981 phase
22983 Set swept wave percentage-shift for multi channel. Range from 0 to
22984 100. Default value is 25.
22985 interp
22987 Set delay-line interpolation, linear or quadratic. Default is
22988 linear.
22989 haas
22991 Apply Haas effect to audio.
22992 Note that this makes most sense to apply on mono signals. With this
22994 filter applied to mono signals it give some directionality and
22995 stretches its stereo image.
22996 The filter accepts the following options:
22998 level_in
23000 Set input level. By default is 1, or 0dB
23001 level_out
23003 Set output level. By default is 1, or 0dB.
23004 side_gain
23006 Set gain applied to side part of signal. By default is 1.
23007 middle_source
23009 Set kind of middle source. Can be one of the following:
23010 left
23012 Pick left channel.
23013 right
23015 Pick right channel.
23016 mid Pick middle part signal of stereo image.
23018 side
23020 Pick side part signal of stereo image.
23021 middle_phase
23023 Change middle phase. By default is disabled.
23024 left_delay
23026 Set left channel delay. By default is 2.05 milliseconds.
23027 left_balance
23029 Set left channel balance. By default is -1.
23030 left_gain
23032 Set left channel gain. By default is 1.
23033 left_phase
23035 Change left phase. By default is disabled.
23036 right_delay
23038 Set right channel delay. By defaults is 2.12 milliseconds.
23039 right_balance
23041 Set right channel balance. By default is 1.
23042 right_gain
23044 Set right channel gain. By default is 1.
23045 right_phase
23047 Change right phase. By default is enabled.
23048 hdcd
23050 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM
23051 stream with embedded HDCD codes is expanded into a 20-bit PCM stream.
23052 The filter supports the Peak Extend and Low-level Gain Adjustment
23054 features of HDCD, and detects the Transient Filter flag.
23055 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
23057 When using the filter with wav, note the default encoding for wav is
23059 16-bit, so the resulting 20-bit stream will be truncated back to
23060 16-bit. Use something like -acodec pcm_s24le after the filter to get
23061 24-bit PCM output.
23062 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
23064 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
23065 The filter accepts the following options:
23067 disable_autoconvert
23069 Disable any automatic format conversion or resampling in the filter
23070 graph.
23071 process_stereo
23073 Process the stereo channels together. If target_gain does not match
23074 between channels, consider it invalid and use the last valid
23075 target_gain.
23076 cdt_ms
23078 Set the code detect timer period in ms.
23079 force_pe
23081 Always extend peaks above -3dBFS even if PE isn't signaled.
23082 analyze_mode
23084 Replace audio with a solid tone and adjust the amplitude to signal
23085 some specific aspect of the decoding process. The output file can
23086 be loaded in an audio editor alongside the original to aid
23087 analysis.
23088 "analyze_mode=pe:force_pe=true" can be used to see all samples
23090 above the PE level.
23091 Modes are:
23093 0, off
23095 Disabled
23096 1, lle
23098 Gain adjustment level at each sample
23099 2, pe
23101 Samples where peak extend occurs
23102 3, cdt
23104 Samples where the code detect timer is active
23105 4, tgm
23107 Samples where the target gain does not match between channels
23108 headphone
23110 Apply head-related transfer functions (HRTFs) to create virtual
23111 loudspeakers around the user for binaural listening via headphones.
23112 The HRIRs are provided via additional streams, for each channel one
23113 stereo input stream is needed.
23114 The filter accepts the following options:
23116 map Set mapping of input streams for convolution. The argument is a
23118 '|'-separated list of channel names in order as they are given as
23119 additional stream inputs for filter. This also specify number of
23120 input streams. Number of input streams must be not less than number
23121 of channels in first stream plus one.
23122 gain
23124 Set gain applied to audio. Value is in dB. Default is 0.
23125 type
23127 Set processing type. Can be time or freq. time is processing audio
23128 in time domain which is slow. freq is processing audio in
23129 frequency domain which is fast. Default is freq.
23130 lfe Set custom gain for LFE channels. Value is in dB. Default is 0.
23132 size
23134 Set size of frame in number of samples which will be processed at
23135 once. Default value is 1024. Allowed range is from 1024 to 96000.
23136 hrir
23138 Set format of hrir stream. Default value is stereo. Alternative
23139 value is multich. If value is set to stereo, number of additional
23140 streams should be greater or equal to number of input channels in
23141 first input stream. Also each additional stream should have stereo
23142 number of channels. If value is set to multich, number of
23143 additional streams should be exactly one. Also number of input
23144 channels of additional stream should be equal or greater than twice
23145 number of channels of first input stream.
23146 Examples
23148 • Full example using wav files as coefficients with amovie filters
23150 for 7.1 downmix, each amovie filter use stereo file with IR
23151 coefficients as input. The files give coefficients for each
23152 position of virtual loudspeaker:
23153 ffmpeg -i input.wav
23155 -filter_complex "amovie=azi_270_ele_0_DFC.wav[sr];amovie=azi_90_ele_0_DFC.wav[sl];amovie=azi_225_ele_0_DFC.wav[br];amovie=azi_135_ele_0_DFC.wav[bl];amovie=azi_0_ele_0_DFC.wav,asplit[fc][lfe];amovie=azi_35_ele_0_DFC.wav[fl];amovie=azi_325_ele_0_DFC.wav[fr];[0:a][fl][fr][fc][lfe][bl][br][sl][sr]headphone=FL|FR|FC|LFE|BL|BR|SL|SR"
23156 output.wav
23157 • Full example using wav files as coefficients with amovie filters
23159 for 7.1 downmix, but now in multich hrir format.
23160 ffmpeg -i input.wav -filter_complex "amovie=minp.wav[hrirs];[0:a][hrirs]headphone=map=FL|FR|FC|LFE|BL|BR|SL|SR:hrir=multich"
23162 output.wav
23163 highpass
23165 Apply a high-pass filter with 3dB point frequency. The filter can be
23166 either single-pole, or double-pole (the default). The filter roll off
23167 at 6dB per pole per octave (20dB per pole per decade).
23168 The filter accepts the following options:
23170 frequency, f
23172 Set frequency in Hz. Default is 3000.
23173 poles, p
23175 Set number of poles. Default is 2.
23176 width_type, t
23178 Set method to specify band-width of filter.
23179 h Hz
23181 q Q-Factor
23183 o octave
23185 s slope
23187 k kHz
23189 width, w
23191 Specify the band-width of a filter in width_type units. Applies
23192 only to double-pole filter. The default is 0.707q and gives a
23193 Butterworth response.
23194 mix, m
23196 How much to use filtered signal in output. Default is 1. Range is
23197 between 0 and 1.
23198 channels, c
23200 Specify which channels to filter, by default all available are
23201 filtered.
23202 normalize, n
23204 Normalize biquad coefficients, by default is disabled. Enabling it
23205 will normalize magnitude response at DC to 0dB.
23206 transform, a
23208 Set transform type of IIR filter.
23209 di
23211 dii
23212 tdi
23213 tdii
23214 latt
23215 svf
23216 zdf
23217 precision, r
23218 Set precison of filtering.
23219 auto
23221 Pick automatic sample format depending on surround filters.
23222 s16 Always use signed 16-bit.
23224 s32 Always use signed 32-bit.
23226 f32 Always use float 32-bit.
23228 f64 Always use float 64-bit.
23230 block_size, b
23232 Set block size used for reverse IIR processing. If this value is
23233 set to high enough value (higher than impulse response length
23234 truncated when reaches near zero values) filtering will become
23235 linear phase otherwise if not big enough it will just produce nasty
23236 artifacts.
23237 Note that filter delay will be exactly this many samples when set
23239 to non-zero value.
23240 Commands
23242 This filter supports the following commands:
23244 frequency, f
23246 Change highpass frequency. Syntax for the command is : "frequency"
23247 width_type, t
23249 Change highpass width_type. Syntax for the command is :
23250 "width_type"
23251 width, w
23253 Change highpass width. Syntax for the command is : "width"
23254 mix, m
23256 Change highpass mix. Syntax for the command is : "mix"
23257 join
23259 Join multiple input streams into one multi-channel stream.
23260 It accepts the following parameters:
23262 inputs
23264 The number of input streams. It defaults to 2.
23265 channel_layout
23267 The desired output channel layout. It defaults to stereo.
23268 map Map channels from inputs to output. The argument is a '|'-separated
23270 list of mappings, each in the "input_idx.in_channel-out_channel"
23271 form. input_idx is the 0-based index of the input stream.
23272 in_channel can be either the name of the input channel (e.g. FL for
23273 front left) or its index in the specified input stream. out_channel
23274 is the name of the output channel.
23275 The filter will attempt to guess the mappings when they are not
23277 specified explicitly. It does so by first trying to find an unused
23278 matching input channel and if that fails it picks the first unused
23279 input channel.
23280 Join 3 inputs (with properly set channel layouts):
23282 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
23284 Build a 5.1 output from 6 single-channel streams:
23286 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
23288 'join=inputs=6:channel_layout=5.1:map=0.0-FL|1.0-FR|2.0-FC|3.0-SL|4.0-SR|5.0-LFE'
23289 out
23290 ladspa
23292 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
23293 To enable compilation of this filter you need to configure FFmpeg with
23295 "--enable-ladspa".
23296 file, f
23298 Specifies the name of LADSPA plugin library to load. If the
23299 environment variable LADSPA_PATH is defined, the LADSPA plugin is
23300 searched in each one of the directories specified by the colon
23301 separated list in LADSPA_PATH, otherwise in the standard LADSPA
23302 paths, which are in this order: HOME/.ladspa/lib/,
23303 /usr/local/lib/ladspa/, /usr/lib/ladspa/.
23304 plugin, p
23306 Specifies the plugin within the library. Some libraries contain
23307 only one plugin, but others contain many of them. If this is not
23308 set filter will list all available plugins within the specified
23309 library.
23310 controls, c
23312 Set the '|' separated list of controls which are zero or more
23313 floating point values that determine the behavior of the loaded
23314 plugin (for example delay, threshold or gain). Controls need to be
23315 defined using the following syntax:
23316 c0=value0|c1=value1|c2=value2|..., where valuei is the value set on
23317 the i-th control. Alternatively they can be also defined using the
23318 following syntax: value0|value1|value2|..., where valuei is the
23319 value set on the i-th control. If controls is set to "help", all
23320 available controls and their valid ranges are printed.
23321 sample_rate, s
23323 Specify the sample rate, default to 44100. Only used if plugin have
23324 zero inputs.
23325 nb_samples, n
23327 Set the number of samples per channel per each output frame,
23328 default is 1024. Only used if plugin have zero inputs.
23329 duration, d
23331 Set the minimum duration of the sourced audio. See the Time
23332 duration section in the ffmpeg-utils(1) manual for the accepted
23333 syntax. Note that the resulting duration may be greater than the
23334 specified duration, as the generated audio is always cut at the end
23335 of a complete frame. If not specified, or the expressed duration
23336 is negative, the audio is supposed to be generated forever. Only
23337 used if plugin have zero inputs.
23338 latency, l
23340 Enable latency compensation, by default is disabled. Only used if
23341 plugin have inputs.
23342 Examples
23344 • List all available plugins within amp (LADSPA example plugin)
23346 library:
23347 ladspa=file=amp
23349 • List all available controls and their valid ranges for "vcf_notch"
23351 plugin from "VCF" library:
23352 ladspa=f=vcf:p=vcf_notch:c=help
23354 • Simulate low quality audio equipment using "Computer Music Toolkit"
23356 (CMT) plugin library:
23357 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
23359 • Add reverberation to the audio using TAP-plugins (Tom's Audio
23361 Processing plugins):
23362 ladspa=file=tap_reverb:tap_reverb
23364 • Generate white noise, with 0.2 amplitude:
23366 ladspa=file=cmt:noise_source_white:c=c0=.2
23368 • Generate 20 bpm clicks using plugin "C* Click - Metronome" from the
23370 "C* Audio Plugin Suite" (CAPS) library:
23371 ladspa=file=caps:Click:c=c1=20'
23373 • Apply "C* Eq10X2 - Stereo 10-band equaliser" effect:
23375 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
23377 • Increase volume by 20dB using fast lookahead limiter from Steve
23379 Harris "SWH Plugins" collection:
23380 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
23382 • Attenuate low frequencies using Multiband EQ from Steve Harris "SWH
23384 Plugins" collection:
23385 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
23387 • Reduce stereo image using "Narrower" from the "C* Audio Plugin
23389 Suite" (CAPS) library:
23390 ladspa=caps:Narrower
23392 • Another white noise, now using "C* Audio Plugin Suite" (CAPS)
23394 library:
23395 ladspa=caps:White:.2
23397 • Some fractal noise, using "C* Audio Plugin Suite" (CAPS) library:
23399 ladspa=caps:Fractal:c=c1=1
23401 • Dynamic volume normalization using "VLevel" plugin:
23403 ladspa=vlevel-ladspa:vlevel_mono
23405 Commands
23407 This filter supports the following commands:
23409 cN Modify the N-th control value.
23411 If the specified value is not valid, it is ignored and prior one is
23413 kept.
23414 loudnorm
23416 EBU R128 loudness normalization. Includes both dynamic and linear
23417 normalization modes. Support for both single pass (livestreams, files)
23418 and double pass (files) modes. This algorithm can target IL, LRA, and
23419 maximum true peak. In dynamic mode, to accurately detect true peaks,
23420 the audio stream will be upsampled to 192 kHz. Use the "-ar" option or
23421 "aresample" filter to explicitly set an output sample rate.
23422 The filter accepts the following options:
23424 I, i
23426 Set integrated loudness target. Range is -70.0 - -5.0. Default
23427 value is -24.0.
23428 LRA, lra
23430 Set loudness range target. Range is 1.0 - 50.0. Default value is
23431 7.0.
23432 TP, tp
23434 Set maximum true peak. Range is -9.0 - +0.0. Default value is
23435 -2.0.
23436 measured_I, measured_i
23438 Measured IL of input file. Range is -99.0 - +0.0.
23439 measured_LRA, measured_lra
23441 Measured LRA of input file. Range is 0.0 - 99.0.
23442 measured_TP, measured_tp
23444 Measured true peak of input file. Range is -99.0 - +99.0.
23445 measured_thresh
23447 Measured threshold of input file. Range is -99.0 - +0.0.
23448 offset
23450 Set offset gain. Gain is applied before the true-peak limiter.
23451 Range is -99.0 - +99.0. Default is +0.0.
23452 linear
23454 Normalize by linearly scaling the source audio. "measured_I",
23455 "measured_LRA", "measured_TP", and "measured_thresh" must all be
23456 specified. Target LRA shouldn't be lower than source LRA and the
23457 change in integrated loudness shouldn't result in a true peak which
23458 exceeds the target TP. If any of these conditions aren't met,
23459 normalization mode will revert to dynamic. Options are "true" or
23460 "false". Default is "true".
23461 dual_mono
23463 Treat mono input files as "dual-mono". If a mono file is intended
23464 for playback on a stereo system, its EBU R128 measurement will be
23465 perceptually incorrect. If set to "true", this option will
23466 compensate for this effect. Multi-channel input files are not
23467 affected by this option. Options are true or false. Default is
23468 false.
23469 print_format
23471 Set print format for stats. Options are summary, json, or none.
23472 Default value is none.
23473 lowpass
23475 Apply a low-pass filter with 3dB point frequency. The filter can be
23476 either single-pole or double-pole (the default). The filter roll off
23477 at 6dB per pole per octave (20dB per pole per decade).
23478 The filter accepts the following options:
23480 frequency, f
23482 Set frequency in Hz. Default is 500.
23483 poles, p
23485 Set number of poles. Default is 2.
23486 width_type, t
23488 Set method to specify band-width of filter.
23489 h Hz
23491 q Q-Factor
23493 o octave
23495 s slope
23497 k kHz
23499 width, w
23501 Specify the band-width of a filter in width_type units. Applies
23502 only to double-pole filter. The default is 0.707q and gives a
23503 Butterworth response.
23504 mix, m
23506 How much to use filtered signal in output. Default is 1. Range is
23507 between 0 and 1.
23508 channels, c
23510 Specify which channels to filter, by default all available are
23511 filtered.
23512 normalize, n
23514 Normalize biquad coefficients, by default is disabled. Enabling it
23515 will normalize magnitude response at DC to 0dB.
23516 transform, a
23518 Set transform type of IIR filter.
23519 di
23521 dii
23522 tdi
23523 tdii
23524 latt
23525 svf
23526 zdf
23527 precision, r
23528 Set precison of filtering.
23529 auto
23531 Pick automatic sample format depending on surround filters.
23532 s16 Always use signed 16-bit.
23534 s32 Always use signed 32-bit.
23536 f32 Always use float 32-bit.
23538 f64 Always use float 64-bit.
23540 block_size, b
23542 Set block size used for reverse IIR processing. If this value is
23543 set to high enough value (higher than impulse response length
23544 truncated when reaches near zero values) filtering will become
23545 linear phase otherwise if not big enough it will just produce nasty
23546 artifacts.
23547 Note that filter delay will be exactly this many samples when set
23549 to non-zero value.
23550 Examples
23552 • Lowpass only LFE channel, it LFE is not present it does nothing:
23554 lowpass=c=LFE
23556 Commands
23558 This filter supports the following commands:
23560 frequency, f
23562 Change lowpass frequency. Syntax for the command is : "frequency"
23563 width_type, t
23565 Change lowpass width_type. Syntax for the command is :
23566 "width_type"
23567 width, w
23569 Change lowpass width. Syntax for the command is : "width"
23570 mix, m
23572 Change lowpass mix. Syntax for the command is : "mix"
23573 lv2
23575 Load a LV2 (LADSPA Version 2) plugin.
23576 To enable compilation of this filter you need to configure FFmpeg with
23578 "--enable-lv2".
23579 plugin, p
23581 Specifies the plugin URI. You may need to escape ':'.
23582 controls, c
23584 Set the '|' separated list of controls which are zero or more
23585 floating point values that determine the behavior of the loaded
23586 plugin (for example delay, threshold or gain). If controls is set
23587 to "help", all available controls and their valid ranges are
23588 printed.
23589 sample_rate, s
23591 Specify the sample rate, default to 44100. Only used if plugin have
23592 zero inputs.
23593 nb_samples, n
23595 Set the number of samples per channel per each output frame,
23596 default is 1024. Only used if plugin have zero inputs.
23597 duration, d
23599 Set the minimum duration of the sourced audio. See the Time
23600 duration section in the ffmpeg-utils(1) manual for the accepted
23601 syntax. Note that the resulting duration may be greater than the
23602 specified duration, as the generated audio is always cut at the end
23603 of a complete frame. If not specified, or the expressed duration
23604 is negative, the audio is supposed to be generated forever. Only
23605 used if plugin have zero inputs.
23606 Examples
23608 • Apply bass enhancer plugin from Calf:
23610 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
23612 • Apply vinyl plugin from Calf:
23614 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
23616 • Apply bit crusher plugin from ArtyFX:
23618 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
23620 Commands
23622 This filter supports all options that are exported by plugin as
23624 commands.
23625 mcompand
23627 Multiband Compress or expand the audio's dynamic range.
23628 The input audio is divided into bands using 4th order Linkwitz-Riley
23630 IIRs. This is akin to the crossover of a loudspeaker, and results in
23631 flat frequency response when absent compander action.
23632 It accepts the following parameters:
23634 args
23636 This option syntax is: attack,decay,[attack,decay..] soft-knee
23637 points crossover_frequency [delay [initial_volume [gain]]] |
23638 attack,decay ... For explanation of each item refer to compand
23639 filter documentation.
23640 pan
23642 Mix channels with specific gain levels. The filter accepts the output
23643 channel layout followed by a set of channels definitions.
23644 This filter is also designed to efficiently remap the channels of an
23646 audio stream.
23647 The filter accepts parameters of the form: "l|outdef|outdef|..."
23649 l output channel layout or number of channels
23651 outdef
23653 output channel specification, of the form:
23654 "out_name=[gain*]in_name[(+-)[gain*]in_name...]"
23655 out_name
23657 output channel to define, either a channel name (FL, FR, etc.) or a
23658 channel number (c0, c1, etc.)
23659 gain
23661 multiplicative coefficient for the channel, 1 leaving the volume
23662 unchanged
23663 in_name
23665 input channel to use, see out_name for details; it is not possible
23666 to mix named and numbered input channels
23667 If the `=' in a channel specification is replaced by `<', then the
23669 gains for that specification will be renormalized so that the total is
23670 1, thus avoiding clipping noise.
23671 Mixing examples
23673 For example, if you want to down-mix from stereo to mono, but with a
23675 bigger factor for the left channel:
23676 pan=1c|c0=0.9*c0+0.1*c1
23678 A customized down-mix to stereo that works automatically for 3-, 4-, 5-
23680 and 7-channels surround:
23681 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
23683 Note that ffmpeg integrates a default down-mix (and up-mix) system that
23685 should be preferred (see "-ac" option) unless you have very specific
23686 needs.
23687 Remapping examples
23689 The channel remapping will be effective if, and only if:
23691 *<gain coefficients are zeroes or ones,>
23693 *<only one input per channel output,>
23694 If all these conditions are satisfied, the filter will notify the user
23696 ("Pure channel mapping detected"), and use an optimized and lossless
23697 method to do the remapping.
23698 For example, if you have a 5.1 source and want a stereo audio stream by
23700 dropping the extra channels:
23701 pan="stereo| c0=FL | c1=FR"
23703 Given the same source, you can also switch front left and front right
23705 channels and keep the input channel layout:
23706 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
23708 If the input is a stereo audio stream, you can mute the front left
23710 channel (and still keep the stereo channel layout) with:
23711 pan="stereo|c1=c1"
23713 Still with a stereo audio stream input, you can copy the right channel
23715 in both front left and right:
23716 pan="stereo| c0=FR | c1=FR"
23718 replaygain
23720 ReplayGain scanner filter. This filter takes an audio stream as an
23721 input and outputs it unchanged. At end of filtering it displays
23722 "track_gain" and "track_peak".
23723 resample
23725 Convert the audio sample format, sample rate and channel layout. It is
23726 not meant to be used directly.
23727 rubberband
23729 Apply time-stretching and pitch-shifting with librubberband.
23730 To enable compilation of this filter, you need to configure FFmpeg with
23732 "--enable-librubberband".
23733 The filter accepts the following options:
23735 tempo
23737 Set tempo scale factor.
23738 pitch
23740 Set pitch scale factor.
23741 transients
23743 Set transients detector. Possible values are:
23744 crisp
23746 mixed
23747 smooth
23748 detector
23749 Set detector. Possible values are:
23750 compound
23752 percussive
23753 soft
23754 phase
23755 Set phase. Possible values are:
23756 laminar
23758 independent
23759 window
23760 Set processing window size. Possible values are:
23761 standard
23763 short
23764 long
23765 smoothing
23766 Set smoothing. Possible values are:
23767 off
23769 on
23770 formant
23771 Enable formant preservation when shift pitching. Possible values
23772 are:
23773 shifted
23775 preserved
23776 pitchq
23777 Set pitch quality. Possible values are:
23778 quality
23780 speed
23781 consistency
23782 channels
23783 Set channels. Possible values are:
23784 apart
23786 together
23787 Commands
23789 This filter supports the following commands:
23791 tempo
23793 Change filter tempo scale factor. Syntax for the command is :
23794 "tempo"
23795 pitch
23797 Change filter pitch scale factor. Syntax for the command is :
23798 "pitch"
23799 sidechaincompress
23801 This filter acts like normal compressor but has the ability to compress
23802 detected signal using second input signal. It needs two input streams
23803 and returns one output stream. First input stream will be processed
23804 depending on second stream signal. The filtered signal then can be
23805 filtered with other filters in later stages of processing. See pan and
23806 amerge filter.
23807 The filter accepts the following options:
23809 level_in
23811 Set input gain. Default is 1. Range is between 0.015625 and 64.
23812 mode
23814 Set mode of compressor operation. Can be "upward" or "downward".
23815 Default is "downward".
23816 threshold
23818 If a signal of second stream raises above this level it will affect
23819 the gain reduction of first stream. By default is 0.125. Range is
23820 between 0.00097563 and 1.
23821 ratio
23823 Set a ratio about which the signal is reduced. 1:2 means that if
23824 the level raised 4dB above the threshold, it will be only 2dB above
23825 after the reduction. Default is 2. Range is between 1 and 20.
23826 attack
23828 Amount of milliseconds the signal has to rise above the threshold
23829 before gain reduction starts. Default is 20. Range is between 0.01
23830 and 2000.
23831 release
23833 Amount of milliseconds the signal has to fall below the threshold
23834 before reduction is decreased again. Default is 250. Range is
23835 between 0.01 and 9000.
23836 makeup
23838 Set the amount by how much signal will be amplified after
23839 processing. Default is 1. Range is from 1 to 64.
23840 knee
23842 Curve the sharp knee around the threshold to enter gain reduction
23843 more softly. Default is 2.82843. Range is between 1 and 8.
23844 link
23846 Choose if the "average" level between all channels of side-chain
23847 stream or the louder("maximum") channel of side-chain stream
23848 affects the reduction. Default is "average".
23849 detection
23851 Should the exact signal be taken in case of "peak" or an RMS one in
23852 case of "rms". Default is "rms" which is mainly smoother.
23853 level_sc
23855 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
23856 mix How much to use compressed signal in output. Default is 1. Range
23858 is between 0 and 1.
23859 Commands
23861 This filter supports the all above options as commands.
23863 Examples
23865 • Full ffmpeg example taking 2 audio inputs, 1st input to be
23867 compressed depending on the signal of 2nd input and later
23868 compressed signal to be merged with 2nd input:
23869 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
23871 sidechaingate
23873 A sidechain gate acts like a normal (wideband) gate but has the ability
23874 to filter the detected signal before sending it to the gain reduction
23875 stage. Normally a gate uses the full range signal to detect a level
23876 above the threshold. For example: If you cut all lower frequencies
23877 from your sidechain signal the gate will decrease the volume of your
23878 track only if not enough highs appear. With this technique you are able
23879 to reduce the resonation of a natural drum or remove "rumbling" of
23880 muted strokes from a heavily distorted guitar. It needs two input
23881 streams and returns one output stream. First input stream will be
23882 processed depending on second stream signal.
23883 The filter accepts the following options:
23885 level_in
23887 Set input level before filtering. Default is 1. Allowed range is
23888 from 0.015625 to 64.
23889 mode
23891 Set the mode of operation. Can be "upward" or "downward". Default
23892 is "downward". If set to "upward" mode, higher parts of signal will
23893 be amplified, expanding dynamic range in upward direction.
23894 Otherwise, in case of "downward" lower parts of signal will be
23895 reduced.
23896 range
23898 Set the level of gain reduction when the signal is below the
23899 threshold. Default is 0.06125. Allowed range is from 0 to 1.
23900 Setting this to 0 disables reduction and then filter behaves like
23901 expander.
23902 threshold
23904 If a signal rises above this level the gain reduction is released.
23905 Default is 0.125. Allowed range is from 0 to 1.
23906 ratio
23908 Set a ratio about which the signal is reduced. Default is 2.
23909 Allowed range is from 1 to 9000.
23910 attack
23912 Amount of milliseconds the signal has to rise above the threshold
23913 before gain reduction stops. Default is 20 milliseconds. Allowed
23914 range is from 0.01 to 9000.
23915 release
23917 Amount of milliseconds the signal has to fall below the threshold
23918 before the reduction is increased again. Default is 250
23919 milliseconds. Allowed range is from 0.01 to 9000.
23920 makeup
23922 Set amount of amplification of signal after processing. Default is
23923 1. Allowed range is from 1 to 64.
23924 knee
23926 Curve the sharp knee around the threshold to enter gain reduction
23927 more softly. Default is 2.828427125. Allowed range is from 1 to 8.
23928 detection
23930 Choose if exact signal should be taken for detection or an RMS like
23931 one. Default is rms. Can be peak or rms.
23932 link
23934 Choose if the average level between all channels or the louder
23935 channel affects the reduction. Default is average. Can be average
23936 or maximum.
23937 level_sc
23939 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
23940 Commands
23942 This filter supports the all above options as commands.
23944 silencedetect
23946 Detect silence in an audio stream.
23947 This filter logs a message when it detects that the input audio volume
23949 is less or equal to a noise tolerance value for a duration greater or
23950 equal to the minimum detected noise duration.
23951 The printed times and duration are expressed in seconds. The
23953 "lavfi.silence_start" or "lavfi.silence_start.X" metadata key is set on
23954 the first frame whose timestamp equals or exceeds the detection
23955 duration and it contains the timestamp of the first frame of the
23956 silence.
23957 The "lavfi.silence_duration" or "lavfi.silence_duration.X" and
23959 "lavfi.silence_end" or "lavfi.silence_end.X" metadata keys are set on
23960 the first frame after the silence. If mono is enabled, and each channel
23961 is evaluated separately, the ".X" suffixed keys are used, and "X"
23962 corresponds to the channel number.
23963 The filter accepts the following options:
23965 noise, n
23967 Set noise tolerance. Can be specified in dB (in case "dB" is
23968 appended to the specified value) or amplitude ratio. Default is
23969 -60dB, or 0.001.
23970 duration, d
23972 Set silence duration until notification (default is 2 seconds). See
23973 the Time duration section in the ffmpeg-utils(1) manual for the
23974 accepted syntax.
23975 mono, m
23977 Process each channel separately, instead of combined. By default is
23978 disabled.
23979 Examples
23981 • Detect 5 seconds of silence with -50dB noise tolerance:
23983 silencedetect=n=-50dB:d=5
23985 • Complete example with ffmpeg to detect silence with 0.0001 noise
23987 tolerance in silence.mp3:
23988 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
23990 silenceremove
23992 Remove silence from the beginning, middle or end of the audio.
23993 The filter accepts the following options:
23995 start_periods
23997 This value is used to indicate if audio should be trimmed at
23998 beginning of the audio. A value of zero indicates no silence should
23999 be trimmed from the beginning. When specifying a non-zero value, it
24000 trims audio up until it finds non-silence. Normally, when trimming
24001 silence from beginning of audio the start_periods will be 1 but it
24002 can be increased to higher values to trim all audio up to specific
24003 count of non-silence periods. Default value is 0.
24004 start_duration
24006 Specify the amount of time that non-silence must be detected before
24007 it stops trimming audio. By increasing the duration, bursts of
24008 noises can be treated as silence and trimmed off. Default value is
24009 0.
24010 start_threshold
24012 This indicates what sample value should be treated as silence. For
24013 digital audio, a value of 0 may be fine but for audio recorded from
24014 analog, you may wish to increase the value to account for
24015 background noise. Can be specified in dB (in case "dB" is appended
24016 to the specified value) or amplitude ratio. Default value is 0.
24017 start_silence
24019 Specify max duration of silence at beginning that will be kept
24020 after trimming. Default is 0, which is equal to trimming all
24021 samples detected as silence.
24022 start_mode
24024 Specify mode of detection of silence end in start of multi-channel
24025 audio. Can be any or all. Default is any. With any, any sample
24026 that is detected as non-silence will cause stopped trimming of
24027 silence. With all, only if all channels are detected as non-
24028 silence will cause stopped trimming of silence.
24029 stop_periods
24031 Set the count for trimming silence from the end of audio. To
24032 remove silence from the middle of a file, specify a stop_periods
24033 that is negative. This value is then treated as a positive value
24034 and is used to indicate the effect should restart processing as
24035 specified by start_periods, making it suitable for removing periods
24036 of silence in the middle of the audio. Default value is 0.
24037 stop_duration
24039 Specify a duration of silence that must exist before audio is not
24040 copied any more. By specifying a higher duration, silence that is
24041 wanted can be left in the audio. Default value is 0.
24042 stop_threshold
24044 This is the same as start_threshold but for trimming silence from
24045 the end of audio. Can be specified in dB (in case "dB" is appended
24046 to the specified value) or amplitude ratio. Default value is 0.
24047 stop_silence
24049 Specify max duration of silence at end that will be kept after
24050 trimming. Default is 0, which is equal to trimming all samples
24051 detected as silence.
24052 stop_mode
24054 Specify mode of detection of silence start in end of multi-channel
24055 audio. Can be any or all. Default is any. With any, any sample
24056 that is detected as non-silence will cause stopped trimming of
24057 silence. With all, only if all channels are detected as non-
24058 silence will cause stopped trimming of silence.
24059 detection
24061 Set how is silence detected. Can be "rms" or "peak". Second is
24062 faster and works better with digital silence which is exactly 0.
24063 Default value is "rms".
24064 window
24066 Set duration in number of seconds used to calculate size of window
24067 in number of samples for detecting silence. Default value is 0.02.
24068 Allowed range is from 0 to 10.
24069 Examples
24071 • The following example shows how this filter can be used to start a
24073 recording that does not contain the delay at the start which
24074 usually occurs between pressing the record button and the start of
24075 the performance:
24076 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
24078 • Trim all silence encountered from beginning to end where there is
24080 more than 1 second of silence in audio:
24081 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
24083 • Trim all digital silence samples, using peak detection, from
24085 beginning to end where there is more than 0 samples of digital
24086 silence in audio and digital silence is detected in all channels at
24087 same positions in stream:
24088 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
24090 sofalizer
24092 SOFAlizer uses head-related transfer functions (HRTFs) to create
24093 virtual loudspeakers around the user for binaural listening via
24094 headphones (audio formats up to 9 channels supported). The HRTFs are
24095 stored in SOFA files (see <http://www.sofacoustics.org/> for a
24096 database). SOFAlizer is developed at the Acoustics Research Institute
24097 (ARI) of the Austrian Academy of Sciences.
24098 To enable compilation of this filter you need to configure FFmpeg with
24100 "--enable-libmysofa".
24101 The filter accepts the following options:
24103 sofa
24105 Set the SOFA file used for rendering.
24106 gain
24108 Set gain applied to audio. Value is in dB. Default is 0.
24109 rotation
24111 Set rotation of virtual loudspeakers in deg. Default is 0.
24112 elevation
24114 Set elevation of virtual speakers in deg. Default is 0.
24115 radius
24117 Set distance in meters between loudspeakers and the listener with
24118 near-field HRTFs. Default is 1.
24119 type
24121 Set processing type. Can be time or freq. time is processing audio
24122 in time domain which is slow. freq is processing audio in
24123 frequency domain which is fast. Default is freq.
24124 speakers
24126 Set custom positions of virtual loudspeakers. Syntax for this
24127 option is: <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...]. Each
24128 virtual loudspeaker is described with short channel name following
24129 with azimuth and elevation in degrees. Each virtual loudspeaker
24130 description is separated by '|'. For example to override front
24131 left and front right channel positions use: 'speakers=FL 45 15|FR
24132 345 15'. Descriptions with unrecognised channel names are ignored.
24133 lfegain
24135 Set custom gain for LFE channels. Value is in dB. Default is 0.
24136 framesize
24138 Set custom frame size in number of samples. Default is 1024.
24139 Allowed range is from 1024 to 96000. Only used if option type is
24140 set to freq.
24141 normalize
24143 Should all IRs be normalized upon importing SOFA file. By default
24144 is enabled.
24145 interpolate
24147 Should nearest IRs be interpolated with neighbor IRs if exact
24148 position does not match. By default is disabled.
24149 minphase
24151 Minphase all IRs upon loading of SOFA file. By default is disabled.
24152 anglestep
24154 Set neighbor search angle step. Only used if option interpolate is
24155 enabled.
24156 radstep
24158 Set neighbor search radius step. Only used if option interpolate is
24159 enabled.
24160 Examples
24162 • Using ClubFritz6 sofa file:
24164 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
24166 • Using ClubFritz12 sofa file and bigger radius with small rotation:
24168 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
24170 • Similar as above but with custom speaker positions for front left,
24172 front right, back left and back right and also with custom gain:
24173 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
24175 speechnorm
24177 Speech Normalizer.
24178 This filter expands or compresses each half-cycle of audio samples
24180 (local set of samples all above or all below zero and between two
24181 nearest zero crossings) depending on threshold value, so audio reaches
24182 target peak value under conditions controlled by below options.
24183 The filter accepts the following options:
24185 peak, p
24187 Set the expansion target peak value. This specifies the highest
24188 allowed absolute amplitude level for the normalized audio input.
24189 Default value is 0.95. Allowed range is from 0.0 to 1.0.
24190 expansion, e
24192 Set the maximum expansion factor. Allowed range is from 1.0 to
24193 50.0. Default value is 2.0. This option controls maximum local
24194 half-cycle of samples expansion. The maximum expansion would be
24195 such that local peak value reaches target peak value but never to
24196 surpass it and that ratio between new and previous peak value does
24197 not surpass this option value.
24198 compression, c
24200 Set the maximum compression factor. Allowed range is from 1.0 to
24201 50.0. Default value is 2.0. This option controls maximum local
24202 half-cycle of samples compression. This option is used only if
24203 threshold option is set to value greater than 0.0, then in such
24204 cases when local peak is lower or same as value set by threshold
24205 all samples belonging to that peak's half-cycle will be compressed
24206 by current compression factor.
24207 threshold, t
24209 Set the threshold value. Default value is 0.0. Allowed range is
24210 from 0.0 to 1.0. This option specifies which half-cycles of
24211 samples will be compressed and which will be expanded. Any half-
24212 cycle samples with their local peak value below or same as this
24213 option value will be compressed by current compression factor,
24214 otherwise, if greater than threshold value they will be expanded
24215 with expansion factor so that it could reach peak target value but
24216 never surpass it.
24217 raise, r
24219 Set the expansion raising amount per each half-cycle of samples.
24220 Default value is 0.001. Allowed range is from 0.0 to 1.0. This
24221 controls how fast expansion factor is raised per each new half-
24222 cycle until it reaches expansion value. Setting this options too
24223 high may lead to distortions.
24224 fall, f
24226 Set the compression raising amount per each half-cycle of samples.
24227 Default value is 0.001. Allowed range is from 0.0 to 1.0. This
24228 controls how fast compression factor is raised per each new half-
24229 cycle until it reaches compression value.
24230 channels, h
24232 Specify which channels to filter, by default all available channels
24233 are filtered.
24234 invert, i
24236 Enable inverted filtering, by default is disabled. This inverts
24237 interpretation of threshold option. When enabled any half-cycle of
24238 samples with their local peak value below or same as threshold
24239 option will be expanded otherwise it will be compressed.
24240 link, l
24242 Link channels when calculating gain applied to each filtered
24243 channel sample, by default is disabled. When disabled each
24244 filtered channel gain calculation is independent, otherwise when
24245 this option is enabled the minimum of all possible gains for each
24246 filtered channel is used.
24247 rms, m
24249 Set the expansion target RMS value. This specifies the highest
24250 allowed RMS level for the normalized audio input. Default value is
24251 0.0, thus disabled. Allowed range is from 0.0 to 1.0.
24252 Commands
24254 This filter supports the all above options as commands.
24256 Examples
24258 • Weak and slow amplification:
24260 speechnorm=e=3:r=0.00001:l=1
24262 • Moderate and slow amplification:
24264 speechnorm=e=6.25:r=0.00001:l=1
24266 • Strong and fast amplification:
24268 speechnorm=e=12.5:r=0.0001:l=1
24270 • Very strong and fast amplification:
24272 speechnorm=e=25:r=0.0001:l=1
24274 • Extreme and fast amplification:
24276 speechnorm=e=50:r=0.0001:l=1
24278 stereotools
24280 This filter has some handy utilities to manage stereo signals, for
24281 converting M/S stereo recordings to L/R signal while having control
24282 over the parameters or spreading the stereo image of master track.
24283 The filter accepts the following options:
24285 level_in
24287 Set input level before filtering for both channels. Defaults is 1.
24288 Allowed range is from 0.015625 to 64.
24289 level_out
24291 Set output level after filtering for both channels. Defaults is 1.
24292 Allowed range is from 0.015625 to 64.
24293 balance_in
24295 Set input balance between both channels. Default is 0. Allowed
24296 range is from -1 to 1.
24297 balance_out
24299 Set output balance between both channels. Default is 0. Allowed
24300 range is from -1 to 1.
24301 softclip
24303 Enable softclipping. Results in analog distortion instead of harsh
24304 digital 0dB clipping. Disabled by default.
24305 mutel
24307 Mute the left channel. Disabled by default.
24308 muter
24310 Mute the right channel. Disabled by default.
24311 phasel
24313 Change the phase of the left channel. Disabled by default.
24314 phaser
24316 Change the phase of the right channel. Disabled by default.
24317 mode
24319 Set stereo mode. Available values are:
24320 lr>lr
24322 Left/Right to Left/Right, this is default.
24323 lr>ms
24325 Left/Right to Mid/Side.
24326 ms>lr
24328 Mid/Side to Left/Right.
24329 lr>ll
24331 Left/Right to Left/Left.
24332 lr>rr
24334 Left/Right to Right/Right.
24335 lr>l+r
24337 Left/Right to Left + Right.
24338 lr>rl
24340 Left/Right to Right/Left.
24341 ms>ll
24343 Mid/Side to Left/Left.
24344 ms>rr
24346 Mid/Side to Right/Right.
24347 ms>rl
24349 Mid/Side to Right/Left.
24350 lr>l-r
24352 Left/Right to Left - Right.
24353 slev
24355 Set level of side signal. Default is 1. Allowed range is from
24356 0.015625 to 64.
24357 sbal
24359 Set balance of side signal. Default is 0. Allowed range is from -1
24360 to 1.
24361 mlev
24363 Set level of the middle signal. Default is 1. Allowed range is
24364 from 0.015625 to 64.
24365 mpan
24367 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
24368 base
24370 Set stereo base between mono and inversed channels. Default is 0.
24371 Allowed range is from -1 to 1.
24372 delay
24374 Set delay in milliseconds how much to delay left from right channel
24375 and vice versa. Default is 0. Allowed range is from -20 to 20.
24376 sclevel
24378 Set S/C level. Default is 1. Allowed range is from 1 to 100.
24379 phase
24381 Set the stereo phase in degrees. Default is 0. Allowed range is
24382 from 0 to 360.
24383 bmode_in, bmode_out
24385 Set balance mode for balance_in/balance_out option.
24386 Can be one of the following:
24388 balance
24390 Classic balance mode. Attenuate one channel at time. Gain is
24391 raised up to 1.
24392 amplitude
24394 Similar as classic mode above but gain is raised up to 2.
24395 power
24397 Equal power distribution, from -6dB to +6dB range.
24398 Commands
24400 This filter supports the all above options as commands.
24402 Examples
24404 • Apply karaoke like effect:
24406 stereotools=mlev=0.015625
24408 • Convert M/S signal to L/R:
24410 "stereotools=mode=ms>lr"
24412 stereowiden
24414 This filter enhance the stereo effect by suppressing signal common to
24415 both channels and by delaying the signal of left into right and vice
24416 versa, thereby widening the stereo effect.
24417 The filter accepts the following options:
24419 delay
24421 Time in milliseconds of the delay of left signal into right and
24422 vice versa. Default is 20 milliseconds.
24423 feedback
24425 Amount of gain in delayed signal into right and vice versa. Gives a
24426 delay effect of left signal in right output and vice versa which
24427 gives widening effect. Default is 0.3.
24428 crossfeed
24430 Cross feed of left into right with inverted phase. This helps in
24431 suppressing the mono. If the value is 1 it will cancel all the
24432 signal common to both channels. Default is 0.3.
24433 drymix
24435 Set level of input signal of original channel. Default is 0.8.
24436 Commands
24438 This filter supports the all above options except "delay" as commands.
24440 superequalizer
24442 Apply 18 band equalizer.
24443 The filter accepts the following options:
24445 1b Set 65Hz band gain.
24447 2b Set 92Hz band gain.
24449 3b Set 131Hz band gain.
24451 4b Set 185Hz band gain.
24453 5b Set 262Hz band gain.
24455 6b Set 370Hz band gain.
24457 7b Set 523Hz band gain.
24459 8b Set 740Hz band gain.
24461 9b Set 1047Hz band gain.
24463 10b Set 1480Hz band gain.
24465 11b Set 2093Hz band gain.
24467 12b Set 2960Hz band gain.
24469 13b Set 4186Hz band gain.
24471 14b Set 5920Hz band gain.
24473 15b Set 8372Hz band gain.
24475 16b Set 11840Hz band gain.
24477 17b Set 16744Hz band gain.
24479 18b Set 20000Hz band gain.
24481 surround
24483 Apply audio surround upmix filter.
24484 This filter allows to produce multichannel output from audio stream.
24486 The filter accepts the following options:
24488 chl_out
24490 Set output channel layout. By default, this is 5.1.
24491 See the Channel Layout section in the ffmpeg-utils(1) manual for
24493 the required syntax.
24494 chl_in
24496 Set input channel layout. By default, this is stereo.
24497 See the Channel Layout section in the ffmpeg-utils(1) manual for
24499 the required syntax.
24500 level_in
24502 Set input volume level. By default, this is 1.
24503 level_out
24505 Set output volume level. By default, this is 1.
24506 lfe Enable LFE channel output if output channel layout has it. By
24508 default, this is enabled.
24509 lfe_low
24511 Set LFE low cut off frequency. By default, this is 128 Hz.
24512 lfe_high
24514 Set LFE high cut off frequency. By default, this is 256 Hz.
24515 lfe_mode
24517 Set LFE mode, can be add or sub. Default is add. In add mode, LFE
24518 channel is created from input audio and added to output. In sub
24519 mode, LFE channel is created from input audio and added to output
24520 but also all non-LFE output channels are subtracted with output LFE
24521 channel.
24522 smooth
24524 Set temporal smoothness strength, used to gradually change factors
24525 when transforming stereo sound in time. Allowed range is from 0.0
24526 to 1.0. Useful to improve output quality with focus option values
24527 greater than 0.0. Default is 0.0. Only values inside this range
24528 and without edges are effective.
24529 angle
24531 Set angle of stereo surround transform, Allowed range is from 0 to
24532 360. Default is 90.
24533 focus
24535 Set focus of stereo surround transform, Allowed range is from -1 to
24536 1. Default is 0.
24537 fc_in
24539 Set front center input volume. By default, this is 1.
24540 fc_out
24542 Set front center output volume. By default, this is 1.
24543 fl_in
24545 Set front left input volume. By default, this is 1.
24546 fl_out
24548 Set front left output volume. By default, this is 1.
24549 fr_in
24551 Set front right input volume. By default, this is 1.
24552 fr_out
24554 Set front right output volume. By default, this is 1.
24555 sl_in
24557 Set side left input volume. By default, this is 1.
24558 sl_out
24560 Set side left output volume. By default, this is 1.
24561 sr_in
24563 Set side right input volume. By default, this is 1.
24564 sr_out
24566 Set side right output volume. By default, this is 1.
24567 bl_in
24569 Set back left input volume. By default, this is 1.
24570 bl_out
24572 Set back left output volume. By default, this is 1.
24573 br_in
24575 Set back right input volume. By default, this is 1.
24576 br_out
24578 Set back right output volume. By default, this is 1.
24579 bc_in
24581 Set back center input volume. By default, this is 1.
24582 bc_out
24584 Set back center output volume. By default, this is 1.
24585 lfe_in
24587 Set LFE input volume. By default, this is 1.
24588 lfe_out
24590 Set LFE output volume. By default, this is 1.
24591 allx
24593 Set spread usage of stereo image across X axis for all channels.
24594 Allowed range is from -1 to 15. By default this value is negative
24595 -1, and thus unused.
24596 ally
24598 Set spread usage of stereo image across Y axis for all channels.
24599 Allowed range is from -1 to 15. By default this value is negative
24600 -1, and thus unused.
24601 fcx, flx, frx, blx, brx, slx, srx, bcx
24603 Set spread usage of stereo image across X axis for each channel.
24604 Allowed range is from 0.06 to 15. By default this value is 0.5.
24605 fcy, fly, fry, bly, bry, sly, sry, bcy
24607 Set spread usage of stereo image across Y axis for each channel.
24608 Allowed range is from 0.06 to 15. By default this value is 0.5.
24609 win_size
24611 Set window size. Allowed range is from 1024 to 65536. Default size
24612 is 4096.
24613 win_func
24615 Set window function.
24616 It accepts the following values:
24618 rect
24620 bartlett
24621 hann, hanning
24622 hamming
24623 blackman
24624 welch
24625 flattop
24626 bharris
24627 bnuttall
24628 bhann
24629 sine
24630 nuttall
24631 lanczos
24632 gauss
24633 tukey
24634 dolph
24635 cauchy
24636 parzen
24637 poisson
24638 bohman
24639 kaiser
24640 Default is "hann".
24642 overlap
24644 Set window overlap. If set to 1, the recommended overlap for
24645 selected window function will be picked. Default is 0.5.
24646 tiltshelf
24648 Boost or cut the lower frequencies and cut or boost higher frequencies
24649 of the audio using a two-pole shelving filter with a response similar
24650 to that of a standard hi-fi's tone-controls. This is also known as
24651 shelving equalisation (EQ).
24652 The filter accepts the following options:
24654 gain, g
24656 Give the gain at 0 Hz. Its useful range is about -20 (for a large
24657 cut) to +20 (for a large boost). Beware of clipping when using a
24658 positive gain.
24659 frequency, f
24661 Set the filter's central frequency and so can be used to extend or
24662 reduce the frequency range to be boosted or cut. The default value
24663 is 3000 Hz.
24664 width_type, t
24666 Set method to specify band-width of filter.
24667 h Hz
24669 q Q-Factor
24671 o octave
24673 s slope
24675 k kHz
24677 width, w
24679 Determine how steep is the filter's shelf transition.
24680 poles, p
24682 Set number of poles. Default is 2.
24683 mix, m
24685 How much to use filtered signal in output. Default is 1. Range is
24686 between 0 and 1.
24687 channels, c
24689 Specify which channels to filter, by default all available are
24690 filtered.
24691 normalize, n
24693 Normalize biquad coefficients, by default is disabled. Enabling it
24694 will normalize magnitude response at DC to 0dB.
24695 transform, a
24697 Set transform type of IIR filter.
24698 di
24700 dii
24701 tdi
24702 tdii
24703 latt
24704 svf
24705 zdf
24706 precision, r
24707 Set precison of filtering.
24708 auto
24710 Pick automatic sample format depending on surround filters.
24711 s16 Always use signed 16-bit.
24713 s32 Always use signed 32-bit.
24715 f32 Always use float 32-bit.
24717 f64 Always use float 64-bit.
24719 block_size, b
24721 Set block size used for reverse IIR processing. If this value is
24722 set to high enough value (higher than impulse response length
24723 truncated when reaches near zero values) filtering will become
24724 linear phase otherwise if not big enough it will just produce nasty
24725 artifacts.
24726 Note that filter delay will be exactly this many samples when set
24728 to non-zero value.
24729 Commands
24731 This filter supports some options as commands.
24733 treble, highshelf
24735 Boost or cut treble (upper) frequencies of the audio using a two-pole
24736 shelving filter with a response similar to that of a standard hi-fi's
24737 tone-controls. This is also known as shelving equalisation (EQ).
24738 The filter accepts the following options:
24740 gain, g
24742 Give the gain at whichever is the lower of ~22 kHz and the Nyquist
24743 frequency. Its useful range is about -20 (for a large cut) to +20
24744 (for a large boost). Beware of clipping when using a positive gain.
24745 frequency, f
24747 Set the filter's central frequency and so can be used to extend or
24748 reduce the frequency range to be boosted or cut. The default value
24749 is 3000 Hz.
24750 width_type, t
24752 Set method to specify band-width of filter.
24753 h Hz
24755 q Q-Factor
24757 o octave
24759 s slope
24761 k kHz
24763 width, w
24765 Determine how steep is the filter's shelf transition.
24766 poles, p
24768 Set number of poles. Default is 2.
24769 mix, m
24771 How much to use filtered signal in output. Default is 1. Range is
24772 between 0 and 1.
24773 channels, c
24775 Specify which channels to filter, by default all available are
24776 filtered.
24777 normalize, n
24779 Normalize biquad coefficients, by default is disabled. Enabling it
24780 will normalize magnitude response at DC to 0dB.
24781 transform, a
24783 Set transform type of IIR filter.
24784 di
24786 dii
24787 tdi
24788 tdii
24789 latt
24790 svf
24791 zdf
24792 precision, r
24793 Set precison of filtering.
24794 auto
24796 Pick automatic sample format depending on surround filters.
24797 s16 Always use signed 16-bit.
24799 s32 Always use signed 32-bit.
24801 f32 Always use float 32-bit.
24803 f64 Always use float 64-bit.
24805 block_size, b
24807 Set block size used for reverse IIR processing. If this value is
24808 set to high enough value (higher than impulse response length
24809 truncated when reaches near zero values) filtering will become
24810 linear phase otherwise if not big enough it will just produce nasty
24811 artifacts.
24812 Note that filter delay will be exactly this many samples when set
24814 to non-zero value.
24815 Commands
24817 This filter supports the following commands:
24819 frequency, f
24821 Change treble frequency. Syntax for the command is : "frequency"
24822 width_type, t
24824 Change treble width_type. Syntax for the command is : "width_type"
24825 width, w
24827 Change treble width. Syntax for the command is : "width"
24828 gain, g
24830 Change treble gain. Syntax for the command is : "gain"
24831 mix, m
24833 Change treble mix. Syntax for the command is : "mix"
24834 tremolo
24836 Sinusoidal amplitude modulation.
24837 The filter accepts the following options:
24839 f Modulation frequency in Hertz. Modulation frequencies in the
24841 subharmonic range (20 Hz or lower) will result in a tremolo effect.
24842 This filter may also be used as a ring modulator by specifying a
24843 modulation frequency higher than 20 Hz. Range is 0.1 - 20000.0.
24844 Default value is 5.0 Hz.
24845 d Depth of modulation as a percentage. Range is 0.0 - 1.0. Default
24847 value is 0.5.
24848 vibrato
24850 Sinusoidal phase modulation.
24851 The filter accepts the following options:
24853 f Modulation frequency in Hertz. Range is 0.1 - 20000.0. Default
24855 value is 5.0 Hz.
24856 d Depth of modulation as a percentage. Range is 0.0 - 1.0. Default
24858 value is 0.5.
24859 virtualbass
24861 Apply audio Virtual Bass filter.
24862 This filter accepts stereo input and produce stereo with LFE (2.1)
24864 channels output. The newly produced LFE channel have enhanced virtual
24865 bass originally obtained from both stereo channels. This filter
24866 outputs front left and front right channels unchanged as available in
24867 stereo input.
24868 The filter accepts the following options:
24870 cutoff
24872 Set the virtual bass cutoff frequency. Default value is 250 Hz.
24873 Allowed range is from 100 to 500 Hz.
24874 strength
24876 Set the virtual bass strength. Allowed range is from 0.5 to 3.
24877 Default value is 3.
24878 volume
24880 Adjust the input audio volume.
24881 It accepts the following parameters:
24883 volume
24885 Set audio volume expression.
24886 Output values are clipped to the maximum value.
24888 The output audio volume is given by the relation:
24890 <output_volume> = <volume> * <input_volume>
24892 The default value for volume is "1.0".
24894 precision
24896 This parameter represents the mathematical precision.
24897 It determines which input sample formats will be allowed, which
24899 affects the precision of the volume scaling.
24900 fixed
24902 8-bit fixed-point; this limits input sample format to U8, S16,
24903 and S32.
24904 float
24906 32-bit floating-point; this limits input sample format to FLT.
24907 (default)
24908 double
24910 64-bit floating-point; this limits input sample format to DBL.
24911 replaygain
24913 Choose the behaviour on encountering ReplayGain side data in input
24914 frames.
24915 drop
24917 Remove ReplayGain side data, ignoring its contents (the
24918 default).
24919 ignore
24921 Ignore ReplayGain side data, but leave it in the frame.
24922 track
24924 Prefer the track gain, if present.
24925 album
24927 Prefer the album gain, if present.
24928 replaygain_preamp
24930 Pre-amplification gain in dB to apply to the selected replaygain
24931 gain.
24932 Default value for replaygain_preamp is 0.0.
24934 replaygain_noclip
24936 Prevent clipping by limiting the gain applied.
24937 Default value for replaygain_noclip is 1.
24939 eval
24941 Set when the volume expression is evaluated.
24942 It accepts the following values:
24944 once
24946 only evaluate expression once during the filter initialization,
24947 or when the volume command is sent
24948 frame
24950 evaluate expression for each incoming frame
24951 Default value is once.
24953 The volume expression can contain the following parameters.
24955 n frame number (starting at zero)
24957 nb_channels
24959 number of channels
24960 nb_consumed_samples
24962 number of samples consumed by the filter
24963 nb_samples
24965 number of samples in the current frame
24966 pos original frame position in the file
24968 pts frame PTS
24970 sample_rate
24972 sample rate
24973 startpts
24975 PTS at start of stream
24976 startt
24978 time at start of stream
24979 t frame time
24981 tb timestamp timebase
24983 volume
24985 last set volume value
24986 Note that when eval is set to once only the sample_rate and tb
24988 variables are available, all other variables will evaluate to NAN.
24989 Commands
24991 This filter supports the following commands:
24993 volume
24995 Modify the volume expression. The command accepts the same syntax
24996 of the corresponding option.
24997 If the specified expression is not valid, it is kept at its current
24999 value.
25000 Examples
25002 • Halve the input audio volume:
25004 volume=volume=0.5
25006 volume=volume=1/2
25007 volume=volume=-6.0206dB
25008 In all the above example the named key for volume can be omitted,
25010 for example like in:
25011 volume=0.5
25013 • Increase input audio power by 6 decibels using fixed-point
25015 precision:
25016 volume=volume=6dB:precision=fixed
25018 • Fade volume after time 10 with an annihilation period of 5 seconds:
25020 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
25022 volumedetect
25024 Detect the volume of the input video.
25025 The filter has no parameters. It supports only 16-bit signed integer
25027 samples, so the input will be converted when needed. Statistics about
25028 the volume will be printed in the log when the input stream end is
25029 reached.
25030 In particular it will show the mean volume (root mean square), maximum
25032 volume (on a per-sample basis), and the beginning of a histogram of the
25033 registered volume values (from the maximum value to a cumulated 1/1000
25034 of the samples).
25035 All volumes are in decibels relative to the maximum PCM value.
25037 Examples
25039 Here is an excerpt of the output:
25041 [Parsed_volumedetect_0 0xa23120] mean_volume: -27 dB
25043 [Parsed_volumedetect_0 0xa23120] max_volume: -4 dB
25044 [Parsed_volumedetect_0 0xa23120] histogram_4db: 6
25045 [Parsed_volumedetect_0 0xa23120] histogram_5db: 62
25046 [Parsed_volumedetect_0 0xa23120] histogram_6db: 286
25047 [Parsed_volumedetect_0 0xa23120] histogram_7db: 1042
25048 [Parsed_volumedetect_0 0xa23120] histogram_8db: 2551
25049 [Parsed_volumedetect_0 0xa23120] histogram_9db: 4609
25050 [Parsed_volumedetect_0 0xa23120] histogram_10db: 8409
25051 It means that:
25053 • The mean square energy is approximately -27 dB, or 10^-2.7.
25055 • The largest sample is at -4 dB, or more precisely between -4 dB and
25057 -5 dB.
25058 • There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
25060 In other words, raising the volume by +4 dB does not cause any
25062 clipping, raising it by +5 dB causes clipping for 6 samples, etc.
25063
25065 Below is a description of the currently available audio sources.
25066 abuffer
25068 Buffer audio frames, and make them available to the filter chain.
25069 This source is mainly intended for a programmatic use, in particular
25071 through the interface defined in libavfilter/buffersrc.h.
25072 It accepts the following parameters:
25074 time_base
25076 The timebase which will be used for timestamps of submitted frames.
25077 It must be either a floating-point number or in
25078 numerator/denominator form.
25079 sample_rate
25081 The sample rate of the incoming audio buffers.
25082 sample_fmt
25084 The sample format of the incoming audio buffers. Either a sample
25085 format name or its corresponding integer representation from the
25086 enum AVSampleFormat in libavutil/samplefmt.h
25087 channel_layout
25089 The channel layout of the incoming audio buffers. Either a channel
25090 layout name from channel_layout_map in libavutil/channel_layout.c
25091 or its corresponding integer representation from the AV_CH_LAYOUT_*
25092 macros in libavutil/channel_layout.h
25093 channels
25095 The number of channels of the incoming audio buffers. If both
25096 channels and channel_layout are specified, then they must be
25097 consistent.
25098 Examples
25100 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
25102 will instruct the source to accept planar 16bit signed stereo at
25104 44100Hz. Since the sample format with name "s16p" corresponds to the
25105 number 6 and the "stereo" channel layout corresponds to the value 0x3,
25106 this is equivalent to:
25107 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
25109 aevalsrc
25111 Generate an audio signal specified by an expression.
25112 This source accepts in input one or more expressions (one for each
25114 channel), which are evaluated and used to generate a corresponding
25115 audio signal.
25116 This source accepts the following options:
25118 exprs
25120 Set the '|'-separated expressions list for each separate channel.
25121 In case the channel_layout option is not specified, the selected
25122 channel layout depends on the number of provided expressions.
25123 Otherwise the last specified expression is applied to the remaining
25124 output channels.
25125 channel_layout, c
25127 Set the channel layout. The number of channels in the specified
25128 layout must be equal to the number of specified expressions.
25129 duration, d
25131 Set the minimum duration of the sourced audio. See the Time
25132 duration section in the ffmpeg-utils(1) manual for the accepted
25133 syntax. Note that the resulting duration may be greater than the
25134 specified duration, as the generated audio is always cut at the end
25135 of a complete frame.
25136 If not specified, or the expressed duration is negative, the audio
25138 is supposed to be generated forever.
25139 nb_samples, n
25141 Set the number of samples per channel per each output frame,
25142 default to 1024.
25143 sample_rate, s
25145 Specify the sample rate, default to 44100.
25146 Each expression in exprs can contain the following constants:
25148 n number of the evaluated sample, starting from 0
25150 t time of the evaluated sample expressed in seconds, starting from 0
25152 s sample rate
25154 Examples
25156 • Generate silence:
25158 aevalsrc=0
25160 • Generate a sin signal with frequency of 440 Hz, set sample rate to
25162 8000 Hz:
25163 aevalsrc="sin(440*2*PI*t):s=8000"
25165 • Generate a two channels signal, specify the channel layout (Front
25167 Center + Back Center) explicitly:
25168 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
25170 • Generate white noise:
25172 aevalsrc="-2+random(0)"
25174 • Generate an amplitude modulated signal:
25176 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
25178 • Generate 2.5 Hz binaural beats on a 360 Hz carrier:
25180 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
25182 afdelaysrc
25184 Generate a fractional delay FIR coefficients.
25185 The resulting stream can be used with afir filter for filtering the
25187 audio signal.
25188 The filter accepts the following options:
25190 delay, d
25192 Set the fractional delay. Default is 0.
25193 sample_rate, r
25195 Set the sample rate, default is 44100.
25196 nb_samples, n
25198 Set the number of samples per each frame. Default is 1024.
25199 taps, t
25201 Set the number of filter coefficents in output audio stream.
25202 Default value is 0.
25203 channel_layout, c
25205 Specifies the channel layout, and can be a string representing a
25206 channel layout. The default value of channel_layout is "stereo".
25207 afirsrc
25209 Generate a FIR coefficients using frequency sampling method.
25210 The resulting stream can be used with afir filter for filtering the
25212 audio signal.
25213 The filter accepts the following options:
25215 taps, t
25217 Set number of filter coefficents in output audio stream. Default
25218 value is 1025.
25219 frequency, f
25221 Set frequency points from where magnitude and phase are set. This
25222 must be in non decreasing order, and first element must be 0, while
25223 last element must be 1. Elements are separated by white spaces.
25224 magnitude, m
25226 Set magnitude value for every frequency point set by frequency.
25227 Number of values must be same as number of frequency points.
25228 Values are separated by white spaces.
25229 phase, p
25231 Set phase value for every frequency point set by frequency. Number
25232 of values must be same as number of frequency points. Values are
25233 separated by white spaces.
25234 sample_rate, r
25236 Set sample rate, default is 44100.
25237 nb_samples, n
25239 Set number of samples per each frame. Default is 1024.
25240 win_func, w
25242 Set window function. Default is blackman.
25243 anullsrc
25245 The null audio source, return unprocessed audio frames. It is mainly
25246 useful as a template and to be employed in analysis / debugging tools,
25247 or as the source for filters which ignore the input data (for example
25248 the sox synth filter).
25249 This source accepts the following options:
25251 channel_layout, cl
25253 Specifies the channel layout, and can be either an integer or a
25254 string representing a channel layout. The default value of
25255 channel_layout is "stereo".
25256 Check the channel_layout_map definition in
25258 libavutil/channel_layout.c for the mapping between strings and
25259 channel layout values.
25260 sample_rate, r
25262 Specifies the sample rate, and defaults to 44100.
25263 nb_samples, n
25265 Set the number of samples per requested frames.
25266 duration, d
25268 Set the duration of the sourced audio. See the Time duration
25269 section in the ffmpeg-utils(1) manual for the accepted syntax.
25270 If not specified, or the expressed duration is negative, the audio
25272 is supposed to be generated forever.
25273 Examples
25275 • Set the sample rate to 48000 Hz and the channel layout to
25277 AV_CH_LAYOUT_MONO.
25278 anullsrc=r=48000:cl=4
25280 • Do the same operation with a more obvious syntax:
25282 anullsrc=r=48000:cl=mono
25284 All the parameters need to be explicitly defined.
25286 flite
25288 Synthesize a voice utterance using the libflite library.
25289 To enable compilation of this filter you need to configure FFmpeg with
25291 "--enable-libflite".
25292 Note that versions of the flite library prior to 2.0 are not thread-
25294 safe.
25295 The filter accepts the following options:
25297 list_voices
25299 If set to 1, list the names of the available voices and exit
25300 immediately. Default value is 0.
25301 nb_samples, n
25303 Set the maximum number of samples per frame. Default value is 512.
25304 textfile
25306 Set the filename containing the text to speak.
25307 text
25309 Set the text to speak.
25310 voice, v
25312 Set the voice to use for the speech synthesis. Default value is
25313 "kal". See also the list_voices option.
25314 Examples
25316 • Read from file speech.txt, and synthesize the text using the
25318 standard flite voice:
25319 flite=textfile=speech.txt
25321 • Read the specified text selecting the "slt" voice:
25323 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
25325 • Input text to ffmpeg:
25327 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
25329 • Make ffplay speak the specified text, using "flite" and the "lavfi"
25331 device:
25332 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
25334 For more information about libflite, check:
25336 <http://www.festvox.org/flite/>
25337 anoisesrc
25339 Generate a noise audio signal.
25340 The filter accepts the following options:
25342 sample_rate, r
25344 Specify the sample rate. Default value is 48000 Hz.
25345 amplitude, a
25347 Specify the amplitude (0.0 - 1.0) of the generated audio stream.
25348 Default value is 1.0.
25349 duration, d
25351 Specify the duration of the generated audio stream. Not specifying
25352 this option results in noise with an infinite length.
25353 color, colour, c
25355 Specify the color of noise. Available noise colors are white, pink,
25356 brown, blue, violet and velvet. Default color is white.
25357 seed, s
25359 Specify a value used to seed the PRNG.
25360 nb_samples, n
25362 Set the number of samples per each output frame, default is 1024.
25363 Examples
25365 • Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate
25367 and an amplitude of 0.5:
25368 anoisesrc=d=60:c=pink:r=44100:a=0.5
25370 hilbert
25372 Generate odd-tap Hilbert transform FIR coefficients.
25373 The resulting stream can be used with afir filter for phase-shifting
25375 the signal by 90 degrees.
25376 This is used in many matrix coding schemes and for analytic signal
25378 generation. The process is often written as a multiplication by i (or
25379 j), the imaginary unit.
25380 The filter accepts the following options:
25382 sample_rate, s
25384 Set sample rate, default is 44100.
25385 taps, t
25387 Set length of FIR filter, default is 22051.
25388 nb_samples, n
25390 Set number of samples per each frame.
25391 win_func, w
25393 Set window function to be used when generating FIR coefficients.
25394 sinc
25396 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or
25397 band-reject FIR coefficients.
25398 The resulting stream can be used with afir filter for filtering the
25400 audio signal.
25401 The filter accepts the following options:
25403 sample_rate, r
25405 Set sample rate, default is 44100.
25406 nb_samples, n
25408 Set number of samples per each frame. Default is 1024.
25409 hp Set high-pass frequency. Default is 0.
25411 lp Set low-pass frequency. Default is 0. If high-pass frequency is
25413 lower than low-pass frequency and low-pass frequency is higher than
25414 0 then filter will create band-pass filter coefficients, otherwise
25415 band-reject filter coefficients.
25416 phase
25418 Set filter phase response. Default is 50. Allowed range is from 0
25419 to 100.
25420 beta
25422 Set Kaiser window beta.
25423 att Set stop-band attenuation. Default is 120dB, allowed range is from
25425 40 to 180 dB.
25426 round
25428 Enable rounding, by default is disabled.
25429 hptaps
25431 Set number of taps for high-pass filter.
25432 lptaps
25434 Set number of taps for low-pass filter.
25435 sine
25437 Generate an audio signal made of a sine wave with amplitude 1/8.
25438 The audio signal is bit-exact.
25440 The filter accepts the following options:
25442 frequency, f
25444 Set the carrier frequency. Default is 440 Hz.
25445 beep_factor, b
25447 Enable a periodic beep every second with frequency beep_factor
25448 times the carrier frequency. Default is 0, meaning the beep is
25449 disabled.
25450 sample_rate, r
25452 Specify the sample rate, default is 44100.
25453 duration, d
25455 Specify the duration of the generated audio stream.
25456 samples_per_frame
25458 Set the number of samples per output frame.
25459 The expression can contain the following constants:
25461 n The (sequential) number of the output audio frame, starting
25463 from 0.
25464 pts The PTS (Presentation TimeStamp) of the output audio frame,
25466 expressed in TB units.
25467 t The PTS of the output audio frame, expressed in seconds.
25469 TB The timebase of the output audio frames.
25471 Default is 1024.
25473 Examples
25475 • Generate a simple 440 Hz sine wave:
25477 sine
25479 • Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5
25481 seconds:
25482 sine=220:4:d=5
25484 sine=f=220:b=4:d=5
25485 sine=frequency=220:beep_factor=4:duration=5
25486 • Generate a 1 kHz sine wave following "1602,1601,1602,1601,1602"
25488 NTSC pattern:
25489 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
25491
25493 Below is a description of the currently available audio sinks.
25494 abuffersink
25496 Buffer audio frames, and make them available to the end of filter
25497 chain.
25498 This sink is mainly intended for programmatic use, in particular
25500 through the interface defined in libavfilter/buffersink.h or the
25501 options system.
25502 It accepts a pointer to an AVABufferSinkContext structure, which
25504 defines the incoming buffers' formats, to be passed as the opaque
25505 parameter to "avfilter_init_filter" for initialization.
25506 anullsink
25508 Null audio sink; do absolutely nothing with the input audio. It is
25509 mainly useful as a template and for use in analysis / debugging tools.
25510
25512 When you configure your FFmpeg build, you can disable any of the
25513 existing filters using "--disable-filters". The configure output will
25514 show the video filters included in your build.
25515 Below is a description of the currently available video filters.
25517 addroi
25519 Mark a region of interest in a video frame.
25520 The frame data is passed through unchanged, but metadata is attached to
25522 the frame indicating regions of interest which can affect the behaviour
25523 of later encoding. Multiple regions can be marked by applying the
25524 filter multiple times.
25525 x Region distance in pixels from the left edge of the frame.
25527 y Region distance in pixels from the top edge of the frame.
25529 w Region width in pixels.
25531 h Region height in pixels.
25533 The parameters x, y, w and h are expressions, and may contain the
25535 following variables:
25536 iw Width of the input frame.
25538 ih Height of the input frame.
25540 qoffset
25542 Quantisation offset to apply within the region.
25543 This must be a real value in the range -1 to +1. A value of zero
25545 indicates no quality change. A negative value asks for better
25546 quality (less quantisation), while a positive value asks for worse
25547 quality (greater quantisation).
25548 The range is calibrated so that the extreme values indicate the
25550 largest possible offset - if the rest of the frame is encoded with
25551 the worst possible quality, an offset of -1 indicates that this
25552 region should be encoded with the best possible quality anyway.
25553 Intermediate values are then interpolated in some codec-dependent
25554 way.
25555 For example, in 10-bit H.264 the quantisation parameter varies
25557 between -12 and 51. A typical qoffset value of -1/10 therefore
25558 indicates that this region should be encoded with a QP around one-
25559 tenth of the full range better than the rest of the frame. So, if
25560 most of the frame were to be encoded with a QP of around 30, this
25561 region would get a QP of around 24 (an offset of approximately
25562 -1/10 * (51 - -12) = -6.3). An extreme value of -1 would indicate
25563 that this region should be encoded with the best possible quality
25564 regardless of the treatment of the rest of the frame - that is,
25565 should be encoded at a QP of -12.
25566 clear
25568 If set to true, remove any existing regions of interest marked on
25569 the frame before adding the new one.
25570 Examples
25572 • Mark the centre quarter of the frame as interesting.
25574 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
25576 • Mark the 100-pixel-wide region on the left edge of the frame as
25578 very uninteresting (to be encoded at much lower quality than the
25579 rest of the frame).
25580 addroi=0:0:100:ih:+1/5
25582 alphaextract
25584 Extract the alpha component from the input as a grayscale video. This
25585 is especially useful with the alphamerge filter.
25586 alphamerge
25588 Add or replace the alpha component of the primary input with the
25589 grayscale value of a second input. This is intended for use with
25590 alphaextract to allow the transmission or storage of frame sequences
25591 that have alpha in a format that doesn't support an alpha channel.
25592 For example, to reconstruct full frames from a normal YUV-encoded video
25594 and a separate video created with alphaextract, you might use:
25595 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
25597 amplify
25599 Amplify differences between current pixel and pixels of adjacent frames
25600 in same pixel location.
25601 This filter accepts the following options:
25603 radius
25605 Set frame radius. Default is 2. Allowed range is from 1 to 63. For
25606 example radius of 3 will instruct filter to calculate average of 7
25607 frames.
25608 factor
25610 Set factor to amplify difference. Default is 2. Allowed range is
25611 from 0 to 65535.
25612 threshold
25614 Set threshold for difference amplification. Any difference greater
25615 or equal to this value will not alter source pixel. Default is 10.
25616 Allowed range is from 0 to 65535.
25617 tolerance
25619 Set tolerance for difference amplification. Any difference lower to
25620 this value will not alter source pixel. Default is 0. Allowed
25621 range is from 0 to 65535.
25622 low Set lower limit for changing source pixel. Default is 65535.
25624 Allowed range is from 0 to 65535. This option controls maximum
25625 possible value that will decrease source pixel value.
25626 high
25628 Set high limit for changing source pixel. Default is 65535. Allowed
25629 range is from 0 to 65535. This option controls maximum possible
25630 value that will increase source pixel value.
25631 planes
25633 Set which planes to filter. Default is all. Allowed range is from 0
25634 to 15.
25635 Commands
25637 This filter supports the following commands that corresponds to option
25639 of same name:
25640 factor
25642 threshold
25643 tolerance
25644 low
25645 high
25646 planes
25647 ass
25649 Same as the subtitles filter, except that it doesn't require libavcodec
25650 and libavformat to work. On the other hand, it is limited to ASS
25651 (Advanced Substation Alpha) subtitles files.
25652 This filter accepts the following option in addition to the common
25654 options from the subtitles filter:
25655 shaping
25657 Set the shaping engine
25658 Available values are:
25660 auto
25662 The default libass shaping engine, which is the best available.
25663 simple
25665 Fast, font-agnostic shaper that can do only substitutions
25666 complex
25668 Slower shaper using OpenType for substitutions and positioning
25669 The default is "auto".
25671 atadenoise
25673 Apply an Adaptive Temporal Averaging Denoiser to the video input.
25674 The filter accepts the following options:
25676 0a Set threshold A for 1st plane. Default is 0.02. Valid range is 0
25678 to 0.3.
25679 0b Set threshold B for 1st plane. Default is 0.04. Valid range is 0
25681 to 5.
25682 1a Set threshold A for 2nd plane. Default is 0.02. Valid range is 0
25684 to 0.3.
25685 1b Set threshold B for 2nd plane. Default is 0.04. Valid range is 0
25687 to 5.
25688 2a Set threshold A for 3rd plane. Default is 0.02. Valid range is 0
25690 to 0.3.
25691 2b Set threshold B for 3rd plane. Default is 0.04. Valid range is 0
25693 to 5.
25694 Threshold A is designed to react on abrupt changes in the input
25696 signal and threshold B is designed to react on continuous changes
25697 in the input signal.
25698 s Set number of frames filter will use for averaging. Default is 9.
25700 Must be odd number in range [5, 129].
25701 p Set what planes of frame filter will use for averaging. Default is
25703 all.
25704 a Set what variant of algorithm filter will use for averaging.
25706 Default is "p" parallel. Alternatively can be set to "s" serial.
25707 Parallel can be faster then serial, while other way around is never
25709 true. Parallel will abort early on first change being greater then
25710 thresholds, while serial will continue processing other side of
25711 frames if they are equal or below thresholds.
25712 0s
25714 1s
25715 2s Set sigma for 1st plane, 2nd plane or 3rd plane. Default is 32767.
25716 Valid range is from 0 to 32767. This options controls weight for
25717 each pixel in radius defined by size. Default value means every
25718 pixel have same weight. Setting this option to 0 effectively
25719 disables filtering.
25720 Commands
25722 This filter supports same commands as options except option "s". The
25724 command accepts the same syntax of the corresponding option.
25725 avgblur
25727 Apply average blur filter.
25728 The filter accepts the following options:
25730 sizeX
25732 Set horizontal radius size.
25733 planes
25735 Set which planes to filter. By default all planes are filtered.
25736 sizeY
25738 Set vertical radius size, if zero it will be same as "sizeX".
25739 Default is 0.
25740 Commands
25742 This filter supports same commands as options. The command accepts the
25744 same syntax of the corresponding option.
25745 If the specified expression is not valid, it is kept at its current
25747 value.
25748 backgroundkey
25750 Turns a static background into transparency.
25751 The filter accepts the following option:
25753 threshold
25755 Threshold for scene change detection.
25756 similarity
25758 Similarity percentage with the background.
25759 blend
25761 Set the blend amount for pixels that are not similar.
25762 Commands
25764 This filter supports the all above options as commands.
25766 bbox
25768 Compute the bounding box for the non-black pixels in the input frame
25769 luminance plane.
25770 This filter computes the bounding box containing all the pixels with a
25772 luminance value greater than the minimum allowed value. The parameters
25773 describing the bounding box are printed on the filter log.
25774 The filter accepts the following option:
25776 min_val
25778 Set the minimal luminance value. Default is 16.
25779 Commands
25781 This filter supports the all above options as commands.
25783 bilateral
25785 Apply bilateral filter, spatial smoothing while preserving edges.
25786 The filter accepts the following options:
25788 sigmaS
25790 Set sigma of gaussian function to calculate spatial weight.
25791 Allowed range is 0 to 512. Default is 0.1.
25792 sigmaR
25794 Set sigma of gaussian function to calculate range weight. Allowed
25795 range is 0 to 1. Default is 0.1.
25796 planes
25798 Set planes to filter. Default is first only.
25799 Commands
25801 This filter supports the all above options as commands.
25803 bilateral_cuda
25805 CUDA accelerated bilateral filter, an edge preserving filter. This
25806 filter is mathematically accurate thanks to the use of GPU
25807 acceleration. For best output quality, use one to one chroma
25808 subsampling, i.e. yuv444p format.
25809 The filter accepts the following options:
25811 sigmaS
25813 Set sigma of gaussian function to calculate spatial weight, also
25814 called sigma space. Allowed range is 0.1 to 512. Default is 0.1.
25815 sigmaR
25817 Set sigma of gaussian function to calculate color range weight,
25818 also called sigma color. Allowed range is 0.1 to 512. Default is
25819 0.1.
25820 window_size
25822 Set window size of the bilateral function to determine the number
25823 of neighbours to loop on. If the number entered is even, one will
25824 be added automatically. Allowed range is 1 to 255. Default is 1.
25825 Examples
25827 • Apply the bilateral filter on a video.
25829 ./ffmpeg -v verbose \
25831 -hwaccel cuda -hwaccel_output_format cuda -i input.mp4 \
25832 -init_hw_device cuda \
25833 -filter_complex \
25834 " \
25835 [0:v]scale_cuda=format=yuv444p[scaled_video];
25836 [scaled_video]bilateral_cuda=window_size=9:sigmaS=3.0:sigmaR=50.0" \
25837 -an -sn -c:v h264_nvenc -cq 20 out.mp4
25838 bitplanenoise
25840 Show and measure bit plane noise.
25841 The filter accepts the following options:
25843 bitplane
25845 Set which plane to analyze. Default is 1.
25846 filter
25848 Filter out noisy pixels from "bitplane" set above. Default is
25849 disabled.
25850 blackdetect
25852 Detect video intervals that are (almost) completely black. Can be
25853 useful to detect chapter transitions, commercials, or invalid
25854 recordings.
25855 The filter outputs its detection analysis to both the log as well as
25857 frame metadata. If a black segment of at least the specified minimum
25858 duration is found, a line with the start and end timestamps as well as
25859 duration is printed to the log with level "info". In addition, a log
25860 line with level "debug" is printed per frame showing the black amount
25861 detected for that frame.
25862 The filter also attaches metadata to the first frame of a black segment
25864 with key "lavfi.black_start" and to the first frame after the black
25865 segment ends with key "lavfi.black_end". The value is the frame's
25866 timestamp. This metadata is added regardless of the minimum duration
25867 specified.
25868 The filter accepts the following options:
25870 black_min_duration, d
25872 Set the minimum detected black duration expressed in seconds. It
25873 must be a non-negative floating point number.
25874 Default value is 2.0.
25876 picture_black_ratio_th, pic_th
25878 Set the threshold for considering a picture "black". Express the
25879 minimum value for the ratio:
25880 <nb_black_pixels> / <nb_pixels>
25882 for which a picture is considered black. Default value is 0.98.
25884 pixel_black_th, pix_th
25886 Set the threshold for considering a pixel "black".
25887 The threshold expresses the maximum pixel luminance value for which
25889 a pixel is considered "black". The provided value is scaled
25890 according to the following equation:
25891 <absolute_threshold> = <luminance_minimum_value> + <pixel_black_th> * <luminance_range_size>
25893 luminance_range_size and luminance_minimum_value depend on the
25895 input video format, the range is [0-255] for YUV full-range formats
25896 and [16-235] for YUV non full-range formats.
25897 Default value is 0.10.
25899 The following example sets the maximum pixel threshold to the minimum
25901 value, and detects only black intervals of 2 or more seconds:
25902 blackdetect=d=2:pix_th=0.00
25904 blackframe
25906 Detect frames that are (almost) completely black. Can be useful to
25907 detect chapter transitions or commercials. Output lines consist of the
25908 frame number of the detected frame, the percentage of blackness, the
25909 position in the file if known or -1 and the timestamp in seconds.
25910 In order to display the output lines, you need to set the loglevel at
25912 least to the AV_LOG_INFO value.
25913 This filter exports frame metadata "lavfi.blackframe.pblack". The
25915 value represents the percentage of pixels in the picture that are below
25916 the threshold value.
25917 It accepts the following parameters:
25919 amount
25921 The percentage of the pixels that have to be below the threshold;
25922 it defaults to 98.
25923 threshold, thresh
25925 The threshold below which a pixel value is considered black; it
25926 defaults to 32.
25927 blend
25929 Blend two video frames into each other.
25930 The "blend" filter takes two input streams and outputs one stream, the
25932 first input is the "top" layer and second input is "bottom" layer. By
25933 default, the output terminates when the longest input terminates.
25934 The "tblend" (time blend) filter takes two consecutive frames from one
25936 single stream, and outputs the result obtained by blending the new
25937 frame on top of the old frame.
25938 A description of the accepted options follows.
25940 c0_mode
25942 c1_mode
25943 c2_mode
25944 c3_mode
25945 all_mode
25946 Set blend mode for specific pixel component or all pixel components
25947 in case of all_mode. Default value is "normal".
25948 Available values for component modes are:
25950 addition
25952 and
25953 average
25954 bleach
25955 burn
25956 darken
25957 difference
25958 divide
25959 dodge
25960 exclusion
25961 extremity
25962 freeze
25963 geometric
25964 glow
25965 grainextract
25966 grainmerge
25967 hardlight
25968 hardmix
25969 hardoverlay
25970 harmonic
25971 heat
25972 interpolate
25973 lighten
25974 linearlight
25975 multiply
25976 multiply128
25977 negation
25978 normal
25979 or
25980 overlay
25981 phoenix
25982 pinlight
25983 reflect
25984 screen
25985 softdifference
25986 softlight
25987 stain
25988 subtract
25989 vividlight
25990 xor
25991 c0_opacity
25992 c1_opacity
25993 c2_opacity
25994 c3_opacity
25995 all_opacity
25996 Set blend opacity for specific pixel component or all pixel
25997 components in case of all_opacity. Only used in combination with
25998 pixel component blend modes.
25999 c0_expr
26001 c1_expr
26002 c2_expr
26003 c3_expr
26004 all_expr
26005 Set blend expression for specific pixel component or all pixel
26006 components in case of all_expr. Note that related mode options will
26007 be ignored if those are set.
26008 The expressions can use the following variables:
26010 N The sequential number of the filtered frame, starting from 0.
26012 X
26014 Y the coordinates of the current sample
26015 W
26017 H the width and height of currently filtered plane
26018 SW
26020 SH Width and height scale for the plane being filtered. It is the
26021 ratio between the dimensions of the current plane to the luma
26022 plane, e.g. for a "yuv420p" frame, the values are "1,1" for the
26023 luma plane and "0.5,0.5" for the chroma planes.
26024 T Time of the current frame, expressed in seconds.
26026 TOP, A
26028 Value of pixel component at current location for first video
26029 frame (top layer).
26030 BOTTOM, B
26032 Value of pixel component at current location for second video
26033 frame (bottom layer).
26034 The "blend" filter also supports the framesync options.
26036 Examples
26038 • Apply transition from bottom layer to top layer in first 10
26040 seconds:
26041 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
26043 • Apply linear horizontal transition from top layer to bottom layer:
26045 blend=all_expr='A*(X/W)+B*(1-X/W)'
26047 • Apply 1x1 checkerboard effect:
26049 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
26051 • Apply uncover left effect:
26053 blend=all_expr='if(gte(N*SW+X,W),A,B)'
26055 • Apply uncover down effect:
26057 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
26059 • Apply uncover up-left effect:
26061 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
26063 • Split diagonally video and shows top and bottom layer on each side:
26065 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
26067 • Display differences between the current and the previous frame:
26069 tblend=all_mode=grainextract
26071 Commands
26073 This filter supports same commands as options.
26075 blockdetect
26077 Determines blockiness of frames without altering the input frames.
26078 Based on Remco Muijs and Ihor Kirenko: "A no-reference blocking
26080 artifact measure for adaptive video processing." 2005 13th European
26081 signal processing conference.
26082 The filter accepts the following options:
26084 period_min
26086 period_max
26087 Set minimum and maximum values for determining pixel grids
26088 (periods). Default values are [3,24].
26089 planes
26091 Set planes to filter. Default is first only.
26092 Examples
26094 • Determine blockiness for the first plane and search for periods
26096 within [8,32]:
26097 blockdetect=period_min=8:period_max=32:planes=1
26099 blurdetect
26101 Determines blurriness of frames without altering the input frames.
26102 Based on Marziliano, Pina, et al. "A no-reference perceptual blur
26104 metric." Allows for a block-based abbreviation.
26105 The filter accepts the following options:
26107 low
26109 high
26110 Set low and high threshold values used by the Canny thresholding
26111 algorithm.
26112 The high threshold selects the "strong" edge pixels, which are then
26114 connected through 8-connectivity with the "weak" edge pixels
26115 selected by the low threshold.
26116 low and high threshold values must be chosen in the range [0,1],
26118 and low should be lesser or equal to high.
26119 Default value for low is "20/255", and default value for high is
26121 "50/255".
26122 radius
26124 Define the radius to search around an edge pixel for local maxima.
26125 block_pct
26127 Determine blurriness only for the most significant blocks, given in
26128 percentage.
26129 block_width
26131 Determine blurriness for blocks of width block_width. If set to any
26132 value smaller 1, no blocks are used and the whole image is
26133 processed as one no matter of block_height.
26134 block_height
26136 Determine blurriness for blocks of height block_height. If set to
26137 any value smaller 1, no blocks are used and the whole image is
26138 processed as one no matter of block_width.
26139 planes
26141 Set planes to filter. Default is first only.
26142 Examples
26144 • Determine blur for 80% of most significant 32x32 blocks:
26146 blurdetect=block_width=32:block_height=32:block_pct=80
26148 bm3d
26150 Denoise frames using Block-Matching 3D algorithm.
26151 The filter accepts the following options.
26153 sigma
26155 Set denoising strength. Default value is 1. Allowed range is from
26156 0 to 999.9. The denoising algorithm is very sensitive to sigma, so
26157 adjust it according to the source.
26158 block
26160 Set local patch size. This sets dimensions in 2D.
26161 bstep
26163 Set sliding step for processing blocks. Default value is 4.
26164 Allowed range is from 1 to 64. Smaller values allows processing
26165 more reference blocks and is slower.
26166 group
26168 Set maximal number of similar blocks for 3rd dimension. Default
26169 value is 1. When set to 1, no block matching is done. Larger
26170 values allows more blocks in single group. Allowed range is from 1
26171 to 256.
26172 range
26174 Set radius for search block matching. Default is 9. Allowed range
26175 is from 1 to INT32_MAX.
26176 mstep
26178 Set step between two search locations for block matching. Default
26179 is 1. Allowed range is from 1 to 64. Smaller is slower.
26180 thmse
26182 Set threshold of mean square error for block matching. Valid range
26183 is 0 to INT32_MAX.
26184 hdthr
26186 Set thresholding parameter for hard thresholding in 3D transformed
26187 domain. Larger values results in stronger hard-thresholding
26188 filtering in frequency domain.
26189 estim
26191 Set filtering estimation mode. Can be "basic" or "final". Default
26192 is "basic".
26193 ref If enabled, filter will use 2nd stream for block matching. Default
26195 is disabled for "basic" value of estim option, and always enabled
26196 if value of estim is "final".
26197 planes
26199 Set planes to filter. Default is all available except alpha.
26200 Examples
26202 • Basic filtering with bm3d:
26204 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
26206 • Same as above, but filtering only luma:
26208 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
26210 • Same as above, but with both estimation modes:
26212 split[a][b],[a]bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic[a],[b][a]bm3d=sigma=3:block=4:bstep=2:group=16:estim=final:ref=1
26214 • Same as above, but prefilter with nlmeans filter instead:
26216 split[a][b],[a]nlmeans=s=3:r=7:p=3[a],[b][a]bm3d=sigma=3:block=4:bstep=2:group=16:estim=final:ref=1
26218 boxblur
26220 Apply a boxblur algorithm to the input video.
26221 It accepts the following parameters:
26223 luma_radius, lr
26225 luma_power, lp
26226 chroma_radius, cr
26227 chroma_power, cp
26228 alpha_radius, ar
26229 alpha_power, ap
26230 A description of the accepted options follows.
26232 luma_radius, lr
26234 chroma_radius, cr
26235 alpha_radius, ar
26236 Set an expression for the box radius in pixels used for blurring
26237 the corresponding input plane.
26238 The radius value must be a non-negative number, and must not be
26240 greater than the value of the expression "min(w,h)/2" for the luma
26241 and alpha planes, and of "min(cw,ch)/2" for the chroma planes.
26242 Default value for luma_radius is "2". If not specified,
26244 chroma_radius and alpha_radius default to the corresponding value
26245 set for luma_radius.
26246 The expressions can contain the following constants:
26248 w
26250 h The input width and height in pixels.
26251 cw
26253 ch The input chroma image width and height in pixels.
26254 hsub
26256 vsub
26257 The horizontal and vertical chroma subsample values. For
26258 example, for the pixel format "yuv422p", hsub is 2 and vsub is
26259 1.
26260 luma_power, lp
26262 chroma_power, cp
26263 alpha_power, ap
26264 Specify how many times the boxblur filter is applied to the
26265 corresponding plane.
26266 Default value for luma_power is 2. If not specified, chroma_power
26268 and alpha_power default to the corresponding value set for
26269 luma_power.
26270 A value of 0 will disable the effect.
26272 Examples
26274 • Apply a boxblur filter with the luma, chroma, and alpha radii set
26276 to 2:
26277 boxblur=luma_radius=2:luma_power=1
26279 boxblur=2:1
26280 • Set the luma radius to 2, and alpha and chroma radius to 0:
26282 boxblur=2:1:cr=0:ar=0
26284 • Set the luma and chroma radii to a fraction of the video dimension:
26286 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
26288 bwdif
26290 Deinterlace the input video ("bwdif" stands for "Bob Weaver
26291 Deinterlacing Filter").
26292 Motion adaptive deinterlacing based on yadif with the use of w3fdif and
26294 cubic interpolation algorithms. It accepts the following parameters:
26295 mode
26297 The interlacing mode to adopt. It accepts one of the following
26298 values:
26299 0, send_frame
26301 Output one frame for each frame.
26302 1, send_field
26304 Output one frame for each field.
26305 The default value is "send_field".
26307 parity
26309 The picture field parity assumed for the input interlaced video. It
26310 accepts one of the following values:
26311 0, tff
26313 Assume the top field is first.
26314 1, bff
26316 Assume the bottom field is first.
26317 -1, auto
26319 Enable automatic detection of field parity.
26320 The default value is "auto". If the interlacing is unknown or the
26322 decoder does not export this information, top field first will be
26323 assumed.
26324 deint
26326 Specify which frames to deinterlace. Accepts one of the following
26327 values:
26328 0, all
26330 Deinterlace all frames.
26331 1, interlaced
26333 Only deinterlace frames marked as interlaced.
26334 The default value is "all".
26336 cas
26338 Apply Contrast Adaptive Sharpen filter to video stream.
26339 The filter accepts the following options:
26341 strength
26343 Set the sharpening strength. Default value is 0.
26344 planes
26346 Set planes to filter. Default value is to filter all planes except
26347 alpha plane.
26348 Commands
26350 This filter supports same commands as options.
26352 chromahold
26354 Remove all color information for all colors except for certain one.
26355 The filter accepts the following options:
26357 color
26359 The color which will not be replaced with neutral chroma.
26360 similarity
26362 Similarity percentage with the above color. 0.01 matches only the
26363 exact key color, while 1.0 matches everything.
26364 blend
26366 Blend percentage. 0.0 makes pixels either fully gray, or not gray
26367 at all. Higher values result in more preserved color.
26368 yuv Signals that the color passed is already in YUV instead of RGB.
26370 Literal colors like "green" or "red" don't make sense with this
26372 enabled anymore. This can be used to pass exact YUV values as
26373 hexadecimal numbers.
26374 Commands
26376 This filter supports same commands as options. The command accepts the
26378 same syntax of the corresponding option.
26379 If the specified expression is not valid, it is kept at its current
26381 value.
26382 chromakey
26384 YUV colorspace color/chroma keying.
26385 The filter accepts the following options:
26387 color
26389 The color which will be replaced with transparency.
26390 similarity
26392 Similarity percentage with the key color.
26393 0.01 matches only the exact key color, while 1.0 matches
26395 everything.
26396 blend
26398 Blend percentage.
26399 0.0 makes pixels either fully transparent, or not transparent at
26401 all.
26402 Higher values result in semi-transparent pixels, with a higher
26404 transparency the more similar the pixels color is to the key color.
26405 yuv Signals that the color passed is already in YUV instead of RGB.
26407 Literal colors like "green" or "red" don't make sense with this
26409 enabled anymore. This can be used to pass exact YUV values as
26410 hexadecimal numbers.
26411 Commands
26413 This filter supports same commands as options. The command accepts the
26415 same syntax of the corresponding option.
26416 If the specified expression is not valid, it is kept at its current
26418 value.
26419 Examples
26421 • Make every green pixel in the input image transparent:
26423 ffmpeg -i input.png -vf chromakey=green out.png
26425 • Overlay a greenscreen-video on top of a static black background.
26427 ffmpeg -f lavfi -i color=c=black:s=1280x720 -i video.mp4 -shortest -filter_complex "[1:v]chromakey=0x70de77:0.1:0.2[ckout];[0:v][ckout]overlay[out]" -map "[out]" output.mkv
26429 chromakey_cuda
26431 CUDA accelerated YUV colorspace color/chroma keying.
26432 This filter works like normal chromakey filter but operates on CUDA
26434 frames. for more details and parameters see chromakey.
26435 Examples
26437 • Make all the green pixels in the input video transparent and use it
26439 as an overlay for another video:
26440 ./ffmpeg \
26442 -hwaccel cuda -hwaccel_output_format cuda -i input_green.mp4 \
26443 -hwaccel cuda -hwaccel_output_format cuda -i base_video.mp4 \
26444 -init_hw_device cuda \
26445 -filter_complex \
26446 " \
26447 [0:v]chromakey_cuda=0x25302D:0.1:0.12:1[overlay_video]; \
26448 [1:v]scale_cuda=format=yuv420p[base]; \
26449 [base][overlay_video]overlay_cuda" \
26450 -an -sn -c:v h264_nvenc -cq 20 output.mp4
26451 • Process two software sources, explicitly uploading the frames:
26453 ./ffmpeg -init_hw_device cuda=cuda -filter_hw_device cuda \
26455 -f lavfi -i color=size=800x600:color=white,format=yuv420p \
26456 -f lavfi -i yuvtestsrc=size=200x200,format=yuv420p \
26457 -filter_complex \
26458 " \
26459 [0]hwupload[under]; \
26460 [1]hwupload,chromakey_cuda=green:0.1:0.12[over]; \
26461 [under][over]overlay_cuda" \
26462 -c:v hevc_nvenc -cq 18 -preset slow output.mp4
26463 chromanr
26465 Reduce chrominance noise.
26466 The filter accepts the following options:
26468 thres
26470 Set threshold for averaging chrominance values. Sum of absolute
26471 difference of Y, U and V pixel components of current pixel and
26472 neighbour pixels lower than this threshold will be used in
26473 averaging. Luma component is left unchanged and is copied to
26474 output. Default value is 30. Allowed range is from 1 to 200.
26475 sizew
26477 Set horizontal radius of rectangle used for averaging. Allowed
26478 range is from 1 to 100. Default value is 5.
26479 sizeh
26481 Set vertical radius of rectangle used for averaging. Allowed range
26482 is from 1 to 100. Default value is 5.
26483 stepw
26485 Set horizontal step when averaging. Default value is 1. Allowed
26486 range is from 1 to 50. Mostly useful to speed-up filtering.
26487 steph
26489 Set vertical step when averaging. Default value is 1. Allowed
26490 range is from 1 to 50. Mostly useful to speed-up filtering.
26491 threy
26493 Set Y threshold for averaging chrominance values. Set finer
26494 control for max allowed difference between Y components of current
26495 pixel and neigbour pixels. Default value is 200. Allowed range is
26496 from 1 to 200.
26497 threu
26499 Set U threshold for averaging chrominance values. Set finer
26500 control for max allowed difference between U components of current
26501 pixel and neigbour pixels. Default value is 200. Allowed range is
26502 from 1 to 200.
26503 threv
26505 Set V threshold for averaging chrominance values. Set finer
26506 control for max allowed difference between V components of current
26507 pixel and neigbour pixels. Default value is 200. Allowed range is
26508 from 1 to 200.
26509 distance
26511 Set distance type used in calculations.
26512 manhattan
26514 Absolute difference.
26515 euclidean
26517 Difference squared.
26518 Default distance type is manhattan.
26520 Commands
26522 This filter supports same commands as options. The command accepts the
26524 same syntax of the corresponding option.
26525 chromashift
26527 Shift chroma pixels horizontally and/or vertically.
26528 The filter accepts the following options:
26530 cbh Set amount to shift chroma-blue horizontally.
26532 cbv Set amount to shift chroma-blue vertically.
26534 crh Set amount to shift chroma-red horizontally.
26536 crv Set amount to shift chroma-red vertically.
26538 edge
26540 Set edge mode, can be smear, default, or warp.
26541 Commands
26543 This filter supports the all above options as commands.
26545 ciescope
26547 Display CIE color diagram with pixels overlaid onto it.
26548 The filter accepts the following options:
26550 system
26552 Set color system.
26553 ntsc, 470m
26555 ebu, 470bg
26556 smpte
26557 240m
26558 apple
26559 widergb
26560 cie1931
26561 rec709, hdtv
26562 uhdtv, rec2020
26563 dcip3
26564 cie Set CIE system.
26565 xyy
26567 ucs
26568 luv
26569 gamuts
26570 Set what gamuts to draw.
26571 See "system" option for available values.
26573 size, s
26575 Set ciescope size, by default set to 512.
26576 intensity, i
26578 Set intensity used to map input pixel values to CIE diagram.
26579 contrast
26581 Set contrast used to draw tongue colors that are out of active
26582 color system gamut.
26583 corrgamma
26585 Correct gamma displayed on scope, by default enabled.
26586 showwhite
26588 Show white point on CIE diagram, by default disabled.
26589 gamma
26591 Set input gamma. Used only with XYZ input color space.
26592 fill
26594 Fill with CIE colors. By default is enabled.
26595 codecview
26597 Visualize information exported by some codecs.
26598 Some codecs can export information through frames using side-data or
26600 other means. For example, some MPEG based codecs export motion vectors
26601 through the export_mvs flag in the codec flags2 option.
26602 The filter accepts the following option:
26604 block
26606 Display block partition structure using the luma plane.
26607 mv Set motion vectors to visualize.
26609 Available flags for mv are:
26611 pf forward predicted MVs of P-frames
26613 bf forward predicted MVs of B-frames
26615 bb backward predicted MVs of B-frames
26617 qp Display quantization parameters using the chroma planes.
26619 mv_type, mvt
26621 Set motion vectors type to visualize. Includes MVs from all frames
26622 unless specified by frame_type option.
26623 Available flags for mv_type are:
26625 fp forward predicted MVs
26627 bp backward predicted MVs
26629 frame_type, ft
26631 Set frame type to visualize motion vectors of.
26632 Available flags for frame_type are:
26634 if intra-coded frames (I-frames)
26636 pf predicted frames (P-frames)
26638 bf bi-directionally predicted frames (B-frames)
26640 Examples
26642 • Visualize forward predicted MVs of all frames using ffplay:
26644 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
26646 • Visualize multi-directionals MVs of P and B-Frames using ffplay:
26648 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
26650 colorbalance
26652 Modify intensity of primary colors (red, green and blue) of input
26653 frames.
26654 The filter allows an input frame to be adjusted in the shadows,
26656 midtones or highlights regions for the red-cyan, green-magenta or blue-
26657 yellow balance.
26658 A positive adjustment value shifts the balance towards the primary
26660 color, a negative value towards the complementary color.
26661 The filter accepts the following options:
26663 rs
26665 gs
26666 bs Adjust red, green and blue shadows (darkest pixels).
26667 rm
26669 gm
26670 bm Adjust red, green and blue midtones (medium pixels).
26671 rh
26673 gh
26674 bh Adjust red, green and blue highlights (brightest pixels).
26675 Allowed ranges for options are "[-1.0, 1.0]". Defaults are 0.
26677 pl Preserve lightness when changing color balance. Default is
26679 disabled.
26680 Examples
26682 • Add red color cast to shadows:
26684 colorbalance=rs=.3
26686 Commands
26688 This filter supports the all above options as commands.
26690 colorcontrast
26692 Adjust color contrast between RGB components.
26693 The filter accepts the following options:
26695 rc Set the red-cyan contrast. Defaults is 0.0. Allowed range is from
26697 -1.0 to 1.0.
26698 gm Set the green-magenta contrast. Defaults is 0.0. Allowed range is
26700 from -1.0 to 1.0.
26701 by Set the blue-yellow contrast. Defaults is 0.0. Allowed range is
26703 from -1.0 to 1.0.
26704 rcw
26706 gmw
26707 byw Set the weight of each "rc", "gm", "by" option value. Default value
26708 is 0.0. Allowed range is from 0.0 to 1.0. If all weights are 0.0
26709 filtering is disabled.
26710 pl Set the amount of preserving lightness. Default value is 0.0.
26712 Allowed range is from 0.0 to 1.0.
26713 Commands
26715 This filter supports the all above options as commands.
26717 colorcorrect
26719 Adjust color white balance selectively for blacks and whites. This
26720 filter operates in YUV colorspace.
26721 The filter accepts the following options:
26723 rl Set the red shadow spot. Allowed range is from -1.0 to 1.0.
26725 Default value is 0.
26726 bl Set the blue shadow spot. Allowed range is from -1.0 to 1.0.
26728 Default value is 0.
26729 rh Set the red highlight spot. Allowed range is from -1.0 to 1.0.
26731 Default value is 0.
26732 bh Set the red highlight spot. Allowed range is from -1.0 to 1.0.
26734 Default value is 0.
26735 saturation
26737 Set the amount of saturation. Allowed range is from -3.0 to 3.0.
26738 Default value is 1.
26739 analyze
26741 If set to anything other than "manual" it will analyze every frame
26742 and use derived parameters for filtering output frame.
26743 Possible values are:
26745 manual
26747 average
26748 minmax
26749 median
26750 Default value is "manual".
26752 Commands
26754 This filter supports the all above options as commands.
26756 colorchannelmixer
26758 Adjust video input frames by re-mixing color channels.
26759 This filter modifies a color channel by adding the values associated to
26761 the other channels of the same pixels. For example if the value to
26762 modify is red, the output value will be:
26763 <red>=<red>*<rr> + <blue>*<rb> + <green>*<rg> + <alpha>*<ra>
26765 The filter accepts the following options:
26767 rr
26769 rg
26770 rb
26771 ra Adjust contribution of input red, green, blue and alpha channels
26772 for output red channel. Default is 1 for rr, and 0 for rg, rb and
26773 ra.
26774 gr
26776 gg
26777 gb
26778 ga Adjust contribution of input red, green, blue and alpha channels
26779 for output green channel. Default is 1 for gg, and 0 for gr, gb
26780 and ga.
26781 br
26783 bg
26784 bb
26785 ba Adjust contribution of input red, green, blue and alpha channels
26786 for output blue channel. Default is 1 for bb, and 0 for br, bg and
26787 ba.
26788 ar
26790 ag
26791 ab
26792 aa Adjust contribution of input red, green, blue and alpha channels
26793 for output alpha channel. Default is 1 for aa, and 0 for ar, ag
26794 and ab.
26795 Allowed ranges for options are "[-2.0, 2.0]".
26797 pc Set preserve color mode. The accepted values are:
26799 none
26801 Disable color preserving, this is default.
26802 lum Preserve luminance.
26804 max Preserve max value of RGB triplet.
26806 avg Preserve average value of RGB triplet.
26808 sum Preserve sum value of RGB triplet.
26810 nrm Preserve normalized value of RGB triplet.
26812 pwr Preserve power value of RGB triplet.
26814 pa Set the preserve color amount when changing colors. Allowed range
26816 is from "[0.0, 1.0]". Default is 0.0, thus disabled.
26817 Examples
26819 • Convert source to grayscale:
26821 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
26823 • Simulate sepia tones:
26825 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
26827 Commands
26829 This filter supports the all above options as commands.
26831 colorize
26833 Overlay a solid color on the video stream.
26834 The filter accepts the following options:
26836 hue Set the color hue. Allowed range is from 0 to 360. Default value
26838 is 0.
26839 saturation
26841 Set the color saturation. Allowed range is from 0 to 1. Default
26842 value is 0.5.
26843 lightness
26845 Set the color lightness. Allowed range is from 0 to 1. Default
26846 value is 0.5.
26847 mix Set the mix of source lightness. By default is set to 1.0. Allowed
26849 range is from 0.0 to 1.0.
26850 Commands
26852 This filter supports the all above options as commands.
26854 colorkey
26856 RGB colorspace color keying. This filter operates on 8-bit RGB format
26857 frames by setting the alpha component of each pixel which falls within
26858 the similarity radius of the key color to 0. The alpha value for pixels
26859 outside the similarity radius depends on the value of the blend option.
26860 The filter accepts the following options:
26862 color
26864 Set the color for which alpha will be set to 0 (full transparency).
26865 See "Color" section in the ffmpeg-utils manual. Default is
26866 "black".
26867 similarity
26869 Set the radius from the key color within which other colors also
26870 have full transparency. The computed distance is related to the
26871 unit fractional distance in 3D space between the RGB values of the
26872 key color and the pixel's color. Range is 0.01 to 1.0. 0.01 matches
26873 within a very small radius around the exact key color, while 1.0
26874 matches everything. Default is 0.01.
26875 blend
26877 Set how the alpha value for pixels that fall outside the similarity
26878 radius is computed. 0.0 makes pixels either fully transparent or
26879 fully opaque. Higher values result in semi-transparent pixels,
26880 with greater transparency the more similar the pixel color is to
26881 the key color. Range is 0.0 to 1.0. Default is 0.0.
26882 Examples
26884 • Make every green pixel in the input image transparent:
26886 ffmpeg -i input.png -vf colorkey=green out.png
26888 • Overlay a greenscreen-video on top of a static background image.
26890 ffmpeg -i background.png -i video.mp4 -filter_complex "[1:v]colorkey=0x3BBD1E:0.3:0.2[ckout];[0:v][ckout]overlay[out]" -map "[out]" output.flv
26892 Commands
26894 This filter supports same commands as options. The command accepts the
26896 same syntax of the corresponding option.
26897 If the specified expression is not valid, it is kept at its current
26899 value.
26900 colorhold
26902 Remove all color information for all RGB colors except for certain one.
26903 The filter accepts the following options:
26905 color
26907 The color which will not be replaced with neutral gray.
26908 similarity
26910 Similarity percentage with the above color. 0.01 matches only the
26911 exact key color, while 1.0 matches everything.
26912 blend
26914 Blend percentage. 0.0 makes pixels fully gray. Higher values
26915 result in more preserved color.
26916 Commands
26918 This filter supports same commands as options. The command accepts the
26920 same syntax of the corresponding option.
26921 If the specified expression is not valid, it is kept at its current
26923 value.
26924 colorlevels
26926 Adjust video input frames using levels.
26927 The filter accepts the following options:
26929 rimin
26931 gimin
26932 bimin
26933 aimin
26934 Adjust red, green, blue and alpha input black point. Allowed
26935 ranges for options are "[-1.0, 1.0]". Defaults are 0.
26936 rimax
26938 gimax
26939 bimax
26940 aimax
26941 Adjust red, green, blue and alpha input white point. Allowed
26942 ranges for options are "[-1.0, 1.0]". Defaults are 1.
26943 Input levels are used to lighten highlights (bright tones), darken
26945 shadows (dark tones), change the balance of bright and dark tones.
26946 romin
26948 gomin
26949 bomin
26950 aomin
26951 Adjust red, green, blue and alpha output black point. Allowed
26952 ranges for options are "[0, 1.0]". Defaults are 0.
26953 romax
26955 gomax
26956 bomax
26957 aomax
26958 Adjust red, green, blue and alpha output white point. Allowed
26959 ranges for options are "[0, 1.0]". Defaults are 1.
26960 Output levels allows manual selection of a constrained output level
26962 range.
26963 preserve
26965 Set preserve color mode. The accepted values are:
26966 none
26968 Disable color preserving, this is default.
26969 lum Preserve luminance.
26971 max Preserve max value of RGB triplet.
26973 avg Preserve average value of RGB triplet.
26975 sum Preserve sum value of RGB triplet.
26977 nrm Preserve normalized value of RGB triplet.
26979 pwr Preserve power value of RGB triplet.
26981 Examples
26983 • Make video output darker:
26985 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
26987 • Increase contrast:
26989 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
26991 • Make video output lighter:
26993 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
26995 • Increase brightness:
26997 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
26999 Commands
27001 This filter supports the all above options as commands.
27003 colormap
27005 Apply custom color maps to video stream.
27006 This filter needs three input video streams. First stream is video
27008 stream that is going to be filtered out. Second and third video stream
27009 specify color patches for source color to target color mapping.
27010 The filter accepts the following options:
27012 patch_size
27014 Set the source and target video stream patch size in pixels.
27015 nb_patches
27017 Set the max number of used patches from source and target video
27018 stream. Default value is number of patches available in additional
27019 video streams. Max allowed number of patches is 64.
27020 type
27022 Set the adjustments used for target colors. Can be "relative" or
27023 "absolute". Defaults is "absolute".
27024 kernel
27026 Set the kernel used to measure color differences between mapped
27027 colors.
27028 The accepted values are:
27030 euclidean
27032 weuclidean
27033 Default is "euclidean".
27035 colormatrix
27037 Convert color matrix.
27038 The filter accepts the following options:
27040 src
27042 dst Specify the source and destination color matrix. Both values must
27043 be specified.
27044 The accepted values are:
27046 bt709
27048 BT.709
27049 fcc FCC
27051 bt601
27053 BT.601
27054 bt470
27056 BT.470
27057 bt470bg
27059 BT.470BG
27060 smpte170m
27062 SMPTE-170M
27063 smpte240m
27065 SMPTE-240M
27066 bt2020
27068 BT.2020
27069 For example to convert from BT.601 to SMPTE-240M, use the command:
27071 colormatrix=bt601:smpte240m
27073 colorspace
27075 Convert colorspace, transfer characteristics or color primaries. Input
27076 video needs to have an even size.
27077 The filter accepts the following options:
27079 all Specify all color properties at once.
27081 The accepted values are:
27083 bt470m
27085 BT.470M
27086 bt470bg
27088 BT.470BG
27089 bt601-6-525
27091 BT.601-6 525
27092 bt601-6-625
27094 BT.601-6 625
27095 bt709
27097 BT.709
27098 smpte170m
27100 SMPTE-170M
27101 smpte240m
27103 SMPTE-240M
27104 bt2020
27106 BT.2020
27107 space
27109 Specify output colorspace.
27110 The accepted values are:
27112 bt709
27114 BT.709
27115 fcc FCC
27117 bt470bg
27119 BT.470BG or BT.601-6 625
27120 smpte170m
27122 SMPTE-170M or BT.601-6 525
27123 smpte240m
27125 SMPTE-240M
27126 ycgco
27128 YCgCo
27129 bt2020ncl
27131 BT.2020 with non-constant luminance
27132 trc Specify output transfer characteristics.
27134 The accepted values are:
27136 bt709
27138 BT.709
27139 bt470m
27141 BT.470M
27142 bt470bg
27144 BT.470BG
27145 gamma22
27147 Constant gamma of 2.2
27148 gamma28
27150 Constant gamma of 2.8
27151 smpte170m
27153 SMPTE-170M, BT.601-6 625 or BT.601-6 525
27154 smpte240m
27156 SMPTE-240M
27157 srgb
27159 SRGB
27160 iec61966-2-1
27162 iec61966-2-1
27163 iec61966-2-4
27165 iec61966-2-4
27166 xvycc
27168 xvycc
27169 bt2020-10
27171 BT.2020 for 10-bits content
27172 bt2020-12
27174 BT.2020 for 12-bits content
27175 primaries
27177 Specify output color primaries.
27178 The accepted values are:
27180 bt709
27182 BT.709
27183 bt470m
27185 BT.470M
27186 bt470bg
27188 BT.470BG or BT.601-6 625
27189 smpte170m
27191 SMPTE-170M or BT.601-6 525
27192 smpte240m
27194 SMPTE-240M
27195 film
27197 film
27198 smpte431
27200 SMPTE-431
27201 smpte432
27203 SMPTE-432
27204 bt2020
27206 BT.2020
27207 jedec-p22
27209 JEDEC P22 phosphors
27210 range
27212 Specify output color range.
27213 The accepted values are:
27215 tv TV (restricted) range
27217 mpeg
27219 MPEG (restricted) range
27220 pc PC (full) range
27222 jpeg
27224 JPEG (full) range
27225 format
27227 Specify output color format.
27228 The accepted values are:
27230 yuv420p
27232 YUV 4:2:0 planar 8-bits
27233 yuv420p10
27235 YUV 4:2:0 planar 10-bits
27236 yuv420p12
27238 YUV 4:2:0 planar 12-bits
27239 yuv422p
27241 YUV 4:2:2 planar 8-bits
27242 yuv422p10
27244 YUV 4:2:2 planar 10-bits
27245 yuv422p12
27247 YUV 4:2:2 planar 12-bits
27248 yuv444p
27250 YUV 4:4:4 planar 8-bits
27251 yuv444p10
27253 YUV 4:4:4 planar 10-bits
27254 yuv444p12
27256 YUV 4:4:4 planar 12-bits
27257 fast
27259 Do a fast conversion, which skips gamma/primary correction. This
27260 will take significantly less CPU, but will be mathematically
27261 incorrect. To get output compatible with that produced by the
27262 colormatrix filter, use fast=1.
27263 dither
27265 Specify dithering mode.
27266 The accepted values are:
27268 none
27270 No dithering
27271 fsb Floyd-Steinberg dithering
27273 wpadapt
27275 Whitepoint adaptation mode.
27276 The accepted values are:
27278 bradford
27280 Bradford whitepoint adaptation
27281 vonkries
27283 von Kries whitepoint adaptation
27284 identity
27286 identity whitepoint adaptation (i.e. no whitepoint adaptation)
27287 iall
27289 Override all input properties at once. Same accepted values as all.
27290 ispace
27292 Override input colorspace. Same accepted values as space.
27293 iprimaries
27295 Override input color primaries. Same accepted values as primaries.
27296 itrc
27298 Override input transfer characteristics. Same accepted values as
27299 trc.
27300 irange
27302 Override input color range. Same accepted values as range.
27303 The filter converts the transfer characteristics, color space and color
27305 primaries to the specified user values. The output value, if not
27306 specified, is set to a default value based on the "all" property. If
27307 that property is also not specified, the filter will log an error. The
27308 output color range and format default to the same value as the input
27309 color range and format. The input transfer characteristics, color
27310 space, color primaries and color range should be set on the input data.
27311 If any of these are missing, the filter will log an error and no
27312 conversion will take place.
27313 For example to convert the input to SMPTE-240M, use the command:
27315 colorspace=smpte240m
27317 colorspace_cuda
27319 CUDA accelerated implementation of the colorspace filter.
27320 It is by no means feature complete compared to the software colorspace
27322 filter, and at the current time only supports color range conversion
27323 between jpeg/full and mpeg/limited range.
27324 The filter accepts the following options:
27326 range
27328 Specify output color range.
27329 The accepted values are:
27331 tv TV (restricted) range
27333 mpeg
27335 MPEG (restricted) range
27336 pc PC (full) range
27338 jpeg
27340 JPEG (full) range
27341 colortemperature
27343 Adjust color temperature in video to simulate variations in ambient
27344 color temperature.
27345 The filter accepts the following options:
27347 temperature
27349 Set the temperature in Kelvin. Allowed range is from 1000 to 40000.
27350 Default value is 6500 K.
27351 mix Set mixing with filtered output. Allowed range is from 0 to 1.
27353 Default value is 1.
27354 pl Set the amount of preserving lightness. Allowed range is from 0 to
27356 1. Default value is 0.
27357 Commands
27359 This filter supports same commands as options.
27361 convolution
27363 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49
27364 elements.
27365 The filter accepts the following options:
27367 0m
27369 1m
27370 2m
27371 3m Set matrix for each plane. Matrix is sequence of 9, 25 or 49
27372 signed integers in square mode, and from 1 to 49 odd number of
27373 signed integers in row mode.
27374 0rdiv
27376 1rdiv
27377 2rdiv
27378 3rdiv
27379 Set multiplier for calculated value for each plane. If unset or 0,
27380 it will be sum of all matrix elements.
27381 0bias
27383 1bias
27384 2bias
27385 3bias
27386 Set bias for each plane. This value is added to the result of the
27387 multiplication. Useful for making the overall image brighter or
27388 darker. Default is 0.0.
27389 0mode
27391 1mode
27392 2mode
27393 3mode
27394 Set matrix mode for each plane. Can be square, row or column.
27395 Default is square.
27396 Commands
27398 This filter supports the all above options as commands.
27400 Examples
27402 • Apply sharpen:
27404 convolution="0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0"
27406 • Apply blur:
27408 convolution="1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1/9:1/9:1/9:1/9"
27410 • Apply edge enhance:
27412 convolution="0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:5:1:1:1:0:128:128:128"
27414 • Apply edge detect:
27416 convolution="0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:5:5:5:1:0:128:128:128"
27418 • Apply laplacian edge detector which includes diagonals:
27420 convolution="1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:5:5:5:1:0:128:128:0"
27422 • Apply emboss:
27424 convolution="-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2"
27426 convolve
27428 Apply 2D convolution of video stream in frequency domain using second
27429 stream as impulse.
27430 The filter accepts the following options:
27432 planes
27434 Set which planes to process.
27435 impulse
27437 Set which impulse video frames will be processed, can be first or
27438 all. Default is all.
27439 The "convolve" filter also supports the framesync options.
27441 copy
27443 Copy the input video source unchanged to the output. This is mainly
27444 useful for testing purposes.
27445 coreimage
27447 Video filtering on GPU using Apple's CoreImage API on OSX.
27448 Hardware acceleration is based on an OpenGL context. Usually, this
27450 means it is processed by video hardware. However, software-based OpenGL
27451 implementations exist which means there is no guarantee for hardware
27452 processing. It depends on the respective OSX.
27453 There are many filters and image generators provided by Apple that come
27455 with a large variety of options. The filter has to be referenced by its
27456 name along with its options.
27457 The coreimage filter accepts the following options:
27459 list_filters
27461 List all available filters and generators along with all their
27462 respective options as well as possible minimum and maximum values
27463 along with the default values.
27464 list_filters=true
27466 filter
27468 Specify all filters by their respective name and options. Use
27469 list_filters to determine all valid filter names and options.
27470 Numerical options are specified by a float value and are
27471 automatically clamped to their respective value range. Vector and
27472 color options have to be specified by a list of space separated
27473 float values. Character escaping has to be done. A special option
27474 name "default" is available to use default options for a filter.
27475 It is required to specify either "default" or at least one of the
27477 filter options. All omitted options are used with their default
27478 values. The syntax of the filter string is as follows:
27479 filter=<NAME>@<OPTION>=<VALUE>[@<OPTION>=<VALUE>][@...][#<NAME>@<OPTION>=<VALUE>[@<OPTION>=<VALUE>][@...]][#...]
27481 output_rect
27483 Specify a rectangle where the output of the filter chain is copied
27484 into the input image. It is given by a list of space separated
27485 float values:
27486 output_rect=x\ y\ width\ height
27488 If not given, the output rectangle equals the dimensions of the
27490 input image. The output rectangle is automatically cropped at the
27491 borders of the input image. Negative values are valid for each
27492 component.
27493 output_rect=25\ 25\ 100\ 100
27495 Several filters can be chained for successive processing without GPU-
27497 HOST transfers allowing for fast processing of complex filter chains.
27498 Currently, only filters with zero (generators) or exactly one (filters)
27499 input image and one output image are supported. Also, transition
27500 filters are not yet usable as intended.
27501 Some filters generate output images with additional padding depending
27503 on the respective filter kernel. The padding is automatically removed
27504 to ensure the filter output has the same size as the input image.
27505 For image generators, the size of the output image is determined by the
27507 previous output image of the filter chain or the input image of the
27508 whole filterchain, respectively. The generators do not use the pixel
27509 information of this image to generate their output. However, the
27510 generated output is blended onto this image, resulting in partial or
27511 complete coverage of the output image.
27512 The coreimagesrc video source can be used for generating input images
27514 which are directly fed into the filter chain. By using it, providing
27515 input images by another video source or an input video is not required.
27516 Examples
27518 • List all filters available:
27520 coreimage=list_filters=true
27522 • Use the CIBoxBlur filter with default options to blur an image:
27524 coreimage=filter=CIBoxBlur@default
27526 • Use a filter chain with CISepiaTone at default values and
27528 CIVignetteEffect with its center at 100x100 and a radius of 50
27529 pixels:
27530 coreimage=filter=CIBoxBlur@default#CIVignetteEffect@inputCenter=100\ 100@inputRadius=50
27532 • Use nullsrc and CIQRCodeGenerator to create a QR code for the
27534 FFmpeg homepage, given as complete and escaped command-line for
27535 Apple's standard bash shell:
27536 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@inputMessage=https\\\\\://FFmpeg.org/@inputCorrectionLevel=H -frames:v 1 QRCode.png
27538 corr
27540 Obtain the correlation between two input videos.
27541 This filter takes two input videos.
27543 Both input videos must have the same resolution and pixel format for
27545 this filter to work correctly. Also it assumes that both inputs have
27546 the same number of frames, which are compared one by one.
27547 The obtained per component, average, min and max correlation is printed
27549 through the logging system.
27550 The filter stores the calculated correlation of each frame in frame
27552 metadata.
27553 This filter also supports the framesync options.
27555 In the below example the input file main.mpg being processed is
27557 compared with the reference file ref.mpg.
27558 ffmpeg -i main.mpg -i ref.mpg -lavfi corr -f null -
27560 cover_rect
27562 Cover a rectangular object
27563 It accepts the following options:
27565 cover
27567 Filepath of the optional cover image, needs to be in yuv420.
27568 mode
27570 Set covering mode.
27571 It accepts the following values:
27573 cover
27575 cover it by the supplied image
27576 blur
27578 cover it by interpolating the surrounding pixels
27579 Default value is blur.
27581 Examples
27583 • Cover a rectangular object by the supplied image of a given video
27585 using ffmpeg:
27586 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
27588 crop
27590 Crop the input video to given dimensions.
27591 It accepts the following parameters:
27593 w, out_w
27595 The width of the output video. It defaults to "iw". This
27596 expression is evaluated only once during the filter configuration,
27597 or when the w or out_w command is sent.
27598 h, out_h
27600 The height of the output video. It defaults to "ih". This
27601 expression is evaluated only once during the filter configuration,
27602 or when the h or out_h command is sent.
27603 x The horizontal position, in the input video, of the left edge of
27605 the output video. It defaults to "(in_w-out_w)/2". This expression
27606 is evaluated per-frame.
27607 y The vertical position, in the input video, of the top edge of the
27609 output video. It defaults to "(in_h-out_h)/2". This expression is
27610 evaluated per-frame.
27611 keep_aspect
27613 If set to 1 will force the output display aspect ratio to be the
27614 same of the input, by changing the output sample aspect ratio. It
27615 defaults to 0.
27616 exact
27618 Enable exact cropping. If enabled, subsampled videos will be
27619 cropped at exact width/height/x/y as specified and will not be
27620 rounded to nearest smaller value. It defaults to 0.
27621 The out_w, out_h, x, y parameters are expressions containing the
27623 following constants:
27624 x
27626 y The computed values for x and y. They are evaluated for each new
27627 frame.
27628 in_w
27630 in_h
27631 The input width and height.
27632 iw
27634 ih These are the same as in_w and in_h.
27635 out_w
27637 out_h
27638 The output (cropped) width and height.
27639 ow
27641 oh These are the same as out_w and out_h.
27642 a same as iw / ih
27644 sar input sample aspect ratio
27646 dar input display aspect ratio, it is the same as (iw / ih) * sar
27648 hsub
27650 vsub
27651 horizontal and vertical chroma subsample values. For example for
27652 the pixel format "yuv422p" hsub is 2 and vsub is 1.
27653 n The number of the input frame, starting from 0.
27655 pos the position in the file of the input frame, NAN if unknown
27657 t The timestamp expressed in seconds. It's NAN if the input timestamp
27659 is unknown.
27660 The expression for out_w may depend on the value of out_h, and the
27662 expression for out_h may depend on out_w, but they cannot depend on x
27663 and y, as x and y are evaluated after out_w and out_h.
27664 The x and y parameters specify the expressions for the position of the
27666 top-left corner of the output (non-cropped) area. They are evaluated
27667 for each frame. If the evaluated value is not valid, it is approximated
27668 to the nearest valid value.
27669 The expression for x may depend on y, and the expression for y may
27671 depend on x.
27672 Examples
27674 • Crop area with size 100x100 at position (12,34).
27676 crop=100:100:12:34
27678 Using named options, the example above becomes:
27680 crop=w=100:h=100:x=12:y=34
27682 • Crop the central input area with size 100x100:
27684 crop=100:100
27686 • Crop the central input area with size 2/3 of the input video:
27688 crop=2/3*in_w:2/3*in_h
27690 • Crop the input video central square:
27692 crop=out_w=in_h
27694 crop=in_h
27695 • Delimit the rectangle with the top-left corner placed at position
27697 100:100 and the right-bottom corner corresponding to the right-
27698 bottom corner of the input image.
27699 crop=in_w-100:in_h-100:100:100
27701 • Crop 10 pixels from the left and right borders, and 20 pixels from
27703 the top and bottom borders
27704 crop=in_w-2*10:in_h-2*20
27706 • Keep only the bottom right quarter of the input image:
27708 crop=in_w/2:in_h/2:in_w/2:in_h/2
27710 • Crop height for getting Greek harmony:
27712 crop=in_w:1/PHI*in_w
27714 • Apply trembling effect:
27716 crop=in_w/2:in_h/2:(in_w-out_w)/2+((in_w-out_w)/2)*sin(n/10):(in_h-out_h)/2 +((in_h-out_h)/2)*sin(n/7)
27718 • Apply erratic camera effect depending on timestamp:
27720 crop=in_w/2:in_h/2:(in_w-out_w)/2+((in_w-out_w)/2)*sin(t*10):(in_h-out_h)/2 +((in_h-out_h)/2)*sin(t*13)"
27722 • Set x depending on the value of y:
27724 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
27726 Commands
27728 This filter supports the following commands:
27730 w, out_w
27732 h, out_h
27733 x
27734 y Set width/height of the output video and the horizontal/vertical
27735 position in the input video. The command accepts the same syntax
27736 of the corresponding option.
27737 If the specified expression is not valid, it is kept at its current
27739 value.
27740 cropdetect
27742 Auto-detect the crop size.
27743 It calculates the necessary cropping parameters and prints the
27745 recommended parameters via the logging system. The detected dimensions
27746 correspond to the non-black or video area of the input video according
27747 to mode.
27748 It accepts the following parameters:
27750 mode
27752 Depending on mode crop detection is based on either the mere black
27753 value of surrounding pixels or a combination of motion vectors and
27754 edge pixels.
27755 black
27757 Detect black pixels surrounding the playing video. For fine
27758 control use option limit.
27759 mvedges
27761 Detect the playing video by the motion vectors inside the video
27762 and scanning for edge pixels typically forming the border of a
27763 playing video.
27764 limit
27766 Set higher black value threshold, which can be optionally specified
27767 from nothing (0) to everything (255 for 8-bit based formats). An
27768 intensity value greater to the set value is considered non-black.
27769 It defaults to 24. You can also specify a value between 0.0 and
27770 1.0 which will be scaled depending on the bitdepth of the pixel
27771 format.
27772 round
27774 The value which the width/height should be divisible by. It
27775 defaults to 16. The offset is automatically adjusted to center the
27776 video. Use 2 to get only even dimensions (needed for 4:2:2 video).
27777 16 is best when encoding to most video codecs.
27778 skip
27780 Set the number of initial frames for which evaluation is skipped.
27781 Default is 2. Range is 0 to INT_MAX.
27782 reset_count, reset
27784 Set the counter that determines after how many frames cropdetect
27785 will reset the previously detected largest video area and start
27786 over to detect the current optimal crop area. Default value is 0.
27787 This can be useful when channel logos distort the video area. 0
27789 indicates 'never reset', and returns the largest area encountered
27790 during playback.
27791 mv_threshold
27793 Set motion in pixel units as threshold for motion detection. It
27794 defaults to 8.
27795 low
27797 high
27798 Set low and high threshold values used by the Canny thresholding
27799 algorithm.
27800 The high threshold selects the "strong" edge pixels, which are then
27802 connected through 8-connectivity with the "weak" edge pixels
27803 selected by the low threshold.
27804 low and high threshold values must be chosen in the range [0,1],
27806 and low should be lesser or equal to high.
27807 Default value for low is "5/255", and default value for high is
27809 "15/255".
27810 Examples
27812 • Find video area surrounded by black borders:
27814 ffmpeg -i file.mp4 -vf cropdetect,metadata=mode=print -f null -
27816 • Find an embedded video area, generate motion vectors beforehand:
27818 ffmpeg -i file.mp4 -vf mestimate,cropdetect=mode=mvedges,metadata=mode=print -f null -
27820 • Find an embedded video area, use motion vectors from decoder:
27822 ffmpeg -flags2 +export_mvs -i file.mp4 -vf cropdetect=mode=mvedges,metadata=mode=print -f null -
27824 Commands
27826 This filter supports the following commands:
27828 limit
27830 The command accepts the same syntax of the corresponding option.
27831 If the specified expression is not valid, it is kept at its current
27832 value.
27833 cue
27835 Delay video filtering until a given wallclock timestamp. The filter
27836 first passes on preroll amount of frames, then it buffers at most
27837 buffer amount of frames and waits for the cue. After reaching the cue
27838 it forwards the buffered frames and also any subsequent frames coming
27839 in its input.
27840 The filter can be used synchronize the output of multiple ffmpeg
27842 processes for realtime output devices like decklink. By putting the
27843 delay in the filtering chain and pre-buffering frames the process can
27844 pass on data to output almost immediately after the target wallclock
27845 timestamp is reached.
27846 Perfect frame accuracy cannot be guaranteed, but the result is good
27848 enough for some use cases.
27849 cue The cue timestamp expressed in a UNIX timestamp in microseconds.
27851 Default is 0.
27852 preroll
27854 The duration of content to pass on as preroll expressed in seconds.
27855 Default is 0.
27856 buffer
27858 The maximum duration of content to buffer before waiting for the
27859 cue expressed in seconds. Default is 0.
27860 curves
27862 Apply color adjustments using curves.
27863 This filter is similar to the Adobe Photoshop and GIMP curves tools.
27865 Each component (red, green and blue) has its values defined by N key
27866 points tied from each other using a smooth curve. The x-axis represents
27867 the pixel values from the input frame, and the y-axis the new pixel
27868 values to be set for the output frame.
27869 By default, a component curve is defined by the two points (0;0) and
27871 (1;1). This creates a straight line where each original pixel value is
27872 "adjusted" to its own value, which means no change to the image.
27873 The filter allows you to redefine these two points and add some more. A
27875 new curve will be define to pass smoothly through all these new
27876 coordinates. The new defined points needs to be strictly increasing
27877 over the x-axis, and their x and y values must be in the [0;1]
27878 interval. The curve is formed by using a natural or monotonic cubic
27879 spline interpolation, depending on the interp option (default:
27880 "natural"). The "natural" spline produces a smoother curve in general
27881 while the monotonic ("pchip") spline guarantees the transitions between
27882 the specified points to be monotonic. If the computed curves happened
27883 to go outside the vector spaces, the values will be clipped
27884 accordingly.
27885 The filter accepts the following options:
27887 preset
27889 Select one of the available color presets. This option can be used
27890 in addition to the r, g, b parameters; in this case, the later
27891 options takes priority on the preset values. Available presets
27892 are:
27893 none
27895 color_negative
27896 cross_process
27897 darker
27898 increase_contrast
27899 lighter
27900 linear_contrast
27901 medium_contrast
27902 negative
27903 strong_contrast
27904 vintage
27905 Default is "none".
27907 master, m
27909 Set the master key points. These points will define a second pass
27910 mapping. It is sometimes called a "luminance" or "value" mapping.
27911 It can be used with r, g, b or all since it acts like a post-
27912 processing LUT.
27913 red, r
27915 Set the key points for the red component.
27916 green, g
27918 Set the key points for the green component.
27919 blue, b
27921 Set the key points for the blue component.
27922 all Set the key points for all components (not including master). Can
27924 be used in addition to the other key points component options. In
27925 this case, the unset component(s) will fallback on this all
27926 setting.
27927 psfile
27929 Specify a Photoshop curves file (".acv") to import the settings
27930 from.
27931 plot
27933 Save Gnuplot script of the curves in specified file.
27934 interp
27936 Specify the kind of interpolation. Available algorithms are:
27937 natural
27939 Natural cubic spline using a piece-wise cubic polynomial that
27940 is twice continuously differentiable.
27941 pchip
27943 Monotonic cubic spline using a piecewise cubic Hermite
27944 interpolating polynomial (PCHIP).
27945 To avoid some filtergraph syntax conflicts, each key points list need
27947 to be defined using the following syntax: "x0/y0 x1/y1 x2/y2 ...".
27948 Commands
27950 This filter supports same commands as options.
27952 Examples
27954 • Increase slightly the middle level of blue:
27956 curves=blue='0/0 0.5/0.58 1/1'
27958 • Vintage effect:
27960 curves=r='0/0.11 .42/.51 1/0.95':g='0/0 0.50/0.48 1/1':b='0/0.22 .49/.44 1/0.8'
27962 Here we obtain the following coordinates for each components:
27964 red "(0;0.11) (0.42;0.51) (1;0.95)"
27966 green
27968 "(0;0) (0.50;0.48) (1;1)"
27969 blue
27971 "(0;0.22) (0.49;0.44) (1;0.80)"
27972 • The previous example can also be achieved with the associated
27974 built-in preset:
27975 curves=preset=vintage
27977 • Or simply:
27979 curves=vintage
27981 • Use a Photoshop preset and redefine the points of the green
27983 component:
27984 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
27986 • Check out the curves of the "cross_process" profile using ffmpeg
27988 and gnuplot:
27989 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
27991 gnuplot -p /tmp/curves.plt
27992 datascope
27994 Video data analysis filter.
27995 This filter shows hexadecimal pixel values of part of video.
27997 The filter accepts the following options:
27999 size, s
28001 Set output video size.
28002 x Set x offset from where to pick pixels.
28004 y Set y offset from where to pick pixels.
28006 mode
28008 Set scope mode, can be one of the following:
28009 mono
28011 Draw hexadecimal pixel values with white color on black
28012 background.
28013 color
28015 Draw hexadecimal pixel values with input video pixel color on
28016 black background.
28017 color2
28019 Draw hexadecimal pixel values on color background picked from
28020 input video, the text color is picked in such way so its always
28021 visible.
28022 axis
28024 Draw rows and columns numbers on left and top of video.
28025 opacity
28027 Set background opacity.
28028 format
28030 Set display number format. Can be "hex", or "dec". Default is
28031 "hex".
28032 components
28034 Set pixel components to display. By default all pixel components
28035 are displayed.
28036 Commands
28038 This filter supports same commands as options excluding "size" option.
28040 dblur
28042 Apply Directional blur filter.
28043 The filter accepts the following options:
28045 angle
28047 Set angle of directional blur. Default is 45.
28048 radius
28050 Set radius of directional blur. Default is 5.
28051 planes
28053 Set which planes to filter. By default all planes are filtered.
28054 Commands
28056 This filter supports same commands as options. The command accepts the
28058 same syntax of the corresponding option.
28059 If the specified expression is not valid, it is kept at its current
28061 value.
28062 dctdnoiz
28064 Denoise frames using 2D DCT (frequency domain filtering).
28065 This filter is not designed for real time.
28067 The filter accepts the following options:
28069 sigma, s
28071 Set the noise sigma constant.
28072 This sigma defines a hard threshold of "3 * sigma"; every DCT
28074 coefficient (absolute value) below this threshold with be dropped.
28075 If you need a more advanced filtering, see expr.
28077 Default is 0.
28079 overlap
28081 Set number overlapping pixels for each block. Since the filter can
28082 be slow, you may want to reduce this value, at the cost of a less
28083 effective filter and the risk of various artefacts.
28084 If the overlapping value doesn't permit processing the whole input
28086 width or height, a warning will be displayed and according borders
28087 won't be denoised.
28088 Default value is blocksize-1, which is the best possible setting.
28090 expr, e
28092 Set the coefficient factor expression.
28093 For each coefficient of a DCT block, this expression will be
28095 evaluated as a multiplier value for the coefficient.
28096 If this is option is set, the sigma option will be ignored.
28098 The absolute value of the coefficient can be accessed through the c
28100 variable.
28101 n Set the blocksize using the number of bits. "1<<n" defines the
28103 blocksize, which is the width and height of the processed blocks.
28104 The default value is 3 (8x8) and can be raised to 4 for a blocksize
28106 of 16x16. Note that changing this setting has huge consequences on
28107 the speed processing. Also, a larger block size does not
28108 necessarily means a better de-noising.
28109 Examples
28111 Apply a denoise with a sigma of 4.5:
28113 dctdnoiz=4.5
28115 The same operation can be achieved using the expression system:
28117 dctdnoiz=e='gte(c, 4.5*3)'
28119 Violent denoise using a block size of "16x16":
28121 dctdnoiz=15:n=4
28123 deband
28125 Remove banding artifacts from input video. It works by replacing
28126 banded pixels with average value of referenced pixels.
28127 The filter accepts the following options:
28129 1thr
28131 2thr
28132 3thr
28133 4thr
28134 Set banding detection threshold for each plane. Default is 0.02.
28135 Valid range is 0.00003 to 0.5. If difference between current pixel
28136 and reference pixel is less than threshold, it will be considered
28137 as banded.
28138 range, r
28140 Banding detection range in pixels. Default is 16. If positive,
28141 random number in range 0 to set value will be used. If negative,
28142 exact absolute value will be used. The range defines square of
28143 four pixels around current pixel.
28144 direction, d
28146 Set direction in radians from which four pixel will be compared. If
28147 positive, random direction from 0 to set direction will be picked.
28148 If negative, exact of absolute value will be picked. For example
28149 direction 0, -PI or -2*PI radians will pick only pixels on same row
28150 and -PI/2 will pick only pixels on same column.
28151 blur, b
28153 If enabled, current pixel is compared with average value of all
28154 four surrounding pixels. The default is enabled. If disabled
28155 current pixel is compared with all four surrounding pixels. The
28156 pixel is considered banded if only all four differences with
28157 surrounding pixels are less than threshold.
28158 coupling, c
28160 If enabled, current pixel is changed if and only if all pixel
28161 components are banded, e.g. banding detection threshold is
28162 triggered for all color components. The default is disabled.
28163 Commands
28165 This filter supports the all above options as commands.
28167 deblock
28169 Remove blocking artifacts from input video.
28170 The filter accepts the following options:
28172 filter
28174 Set filter type, can be weak or strong. Default is strong. This
28175 controls what kind of deblocking is applied.
28176 block
28178 Set size of block, allowed range is from 4 to 512. Default is 8.
28179 alpha
28181 beta
28182 gamma
28183 delta
28184 Set blocking detection thresholds. Allowed range is 0 to 1.
28185 Defaults are: 0.098 for alpha and 0.05 for the rest. Using higher
28186 threshold gives more deblocking strength. Setting alpha controls
28187 threshold detection at exact edge of block. Remaining options
28188 controls threshold detection near the edge. Each one for
28189 below/above or left/right. Setting any of those to 0 disables
28190 deblocking.
28191 planes
28193 Set planes to filter. Default is to filter all available planes.
28194 Examples
28196 • Deblock using weak filter and block size of 4 pixels.
28198 deblock=filter=weak:block=4
28200 • Deblock using strong filter, block size of 4 pixels and custom
28202 thresholds for deblocking more edges.
28203 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
28205 • Similar as above, but filter only first plane.
28207 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
28209 • Similar as above, but filter only second and third plane.
28211 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
28213 Commands
28215 This filter supports the all above options as commands.
28217 decimate
28219 Drop duplicated frames at regular intervals.
28220 The filter accepts the following options:
28222 cycle
28224 Set the number of frames from which one will be dropped. Setting
28225 this to N means one frame in every batch of N frames will be
28226 dropped. Default is 5.
28227 dupthresh
28229 Set the threshold for duplicate detection. If the difference metric
28230 for a frame is less than or equal to this value, then it is
28231 declared as duplicate. Default is 1.1
28232 scthresh
28234 Set scene change threshold. Default is 15.
28235 blockx
28237 blocky
28238 Set the size of the x and y-axis blocks used during metric
28239 calculations. Larger blocks give better noise suppression, but
28240 also give worse detection of small movements. Must be a power of
28241 two. Default is 32.
28242 ppsrc
28244 Mark main input as a pre-processed input and activate clean source
28245 input stream. This allows the input to be pre-processed with
28246 various filters to help the metrics calculation while keeping the
28247 frame selection lossless. When set to 1, the first stream is for
28248 the pre-processed input, and the second stream is the clean source
28249 from where the kept frames are chosen. Default is 0.
28250 chroma
28252 Set whether or not chroma is considered in the metric calculations.
28253 Default is 1.
28254 mixed
28256 Set whether or not the input only partially contains content to be
28257 decimated. Default is "false". If enabled video output stream
28258 will be in variable frame rate.
28259 deconvolve
28261 Apply 2D deconvolution of video stream in frequency domain using second
28262 stream as impulse.
28263 The filter accepts the following options:
28265 planes
28267 Set which planes to process.
28268 impulse
28270 Set which impulse video frames will be processed, can be first or
28271 all. Default is all.
28272 noise
28274 Set noise when doing divisions. Default is 0.0000001. Useful when
28275 width and height are not same and not power of 2 or if stream prior
28276 to convolving had noise.
28277 The "deconvolve" filter also supports the framesync options.
28279 dedot
28281 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from
28282 video.
28283 It accepts the following options:
28285 m Set mode of operation. Can be combination of dotcrawl for cross-
28287 luminance reduction and/or rainbows for cross-color reduction.
28288 lt Set spatial luma threshold. Lower values increases reduction of
28290 cross-luminance.
28291 tl Set tolerance for temporal luma. Higher values increases reduction
28293 of cross-luminance.
28294 tc Set tolerance for chroma temporal variation. Higher values
28296 increases reduction of cross-color.
28297 ct Set temporal chroma threshold. Lower values increases reduction of
28299 cross-color.
28300 deflate
28302 Apply deflate effect to the video.
28303 This filter replaces the pixel by the local(3x3) average by taking into
28305 account only values lower than the pixel.
28306 It accepts the following options:
28308 threshold0
28310 threshold1
28311 threshold2
28312 threshold3
28313 Limit the maximum change for each plane, default is 65535. If 0,
28314 plane will remain unchanged.
28315 Commands
28317 This filter supports the all above options as commands.
28319 deflicker
28321 Remove temporal frame luminance variations.
28322 It accepts the following options:
28324 size, s
28326 Set moving-average filter size in frames. Default is 5. Allowed
28327 range is 2 - 129.
28328 mode, m
28330 Set averaging mode to smooth temporal luminance variations.
28331 Available values are:
28333 am Arithmetic mean
28335 gm Geometric mean
28337 hm Harmonic mean
28339 qm Quadratic mean
28341 cm Cubic mean
28343 pm Power mean
28345 median
28347 Median
28348 bypass
28350 Do not actually modify frame. Useful when one only wants metadata.
28351 dejudder
28353 Remove judder produced by partially interlaced telecined content.
28354 Judder can be introduced, for instance, by pullup filter. If the
28356 original source was partially telecined content then the output of
28357 "pullup,dejudder" will have a variable frame rate. May change the
28358 recorded frame rate of the container. Aside from that change, this
28359 filter will not affect constant frame rate video.
28360 The option available in this filter is:
28362 cycle
28364 Specify the length of the window over which the judder repeats.
28365 Accepts any integer greater than 1. Useful values are:
28367 4 If the original was telecined from 24 to 30 fps (Film to NTSC).
28369 5 If the original was telecined from 25 to 30 fps (PAL to NTSC).
28371 20 If a mixture of the two.
28373 The default is 4.
28375 delogo
28377 Suppress a TV station logo by a simple interpolation of the surrounding
28378 pixels. Just set a rectangle covering the logo and watch it disappear
28379 (and sometimes something even uglier appear - your mileage may vary).
28380 It accepts the following parameters:
28382 x
28384 y Specify the top left corner coordinates of the logo. They must be
28385 specified.
28386 w
28388 h Specify the width and height of the logo to clear. They must be
28389 specified.
28390 show
28392 When set to 1, a green rectangle is drawn on the screen to simplify
28393 finding the right x, y, w, and h parameters. The default value is
28394 0.
28395 The rectangle is drawn on the outermost pixels which will be
28397 (partly) replaced with interpolated values. The values of the next
28398 pixels immediately outside this rectangle in each direction will be
28399 used to compute the interpolated pixel values inside the rectangle.
28400 Examples
28402 • Set a rectangle covering the area with top left corner coordinates
28404 0,0 and size 100x77:
28405 delogo=x=0:y=0:w=100:h=77
28407 derain
28409 Remove the rain in the input image/video by applying the derain methods
28410 based on convolutional neural networks. Supported models:
28411 • Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
28413 See
28414 <http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf>.
28415 Training as well as model generation scripts are provided in the
28417 repository at <https://github.com/XueweiMeng/derain_filter.git>.
28418 Native model files (.model) can be generated from TensorFlow model
28420 files (.pb) by using tools/python/convert.py
28421 The filter accepts the following options:
28423 filter_type
28425 Specify which filter to use. This option accepts the following
28426 values:
28427 derain
28429 Derain filter. To conduct derain filter, you need to use a
28430 derain model.
28431 dehaze
28433 Dehaze filter. To conduct dehaze filter, you need to use a
28434 dehaze model.
28435 Default value is derain.
28437 dnn_backend
28439 Specify which DNN backend to use for model loading and execution.
28440 This option accepts the following values:
28441 native
28443 Native implementation of DNN loading and execution.
28444 tensorflow
28446 TensorFlow backend. To enable this backend you need to install
28447 the TensorFlow for C library (see
28448 <https://www.tensorflow.org/install/lang_c>) and configure
28449 FFmpeg with "--enable-libtensorflow"
28450 Default value is native.
28452 model
28454 Set path to model file specifying network architecture and its
28455 parameters. Note that different backends use different file
28456 formats. TensorFlow and native backend can load files for only its
28457 format.
28458 To get full functionality (such as async execution), please use the
28460 dnn_processing filter.
28461 deshake
28463 Attempt to fix small changes in horizontal and/or vertical shift. This
28464 filter helps remove camera shake from hand-holding a camera, bumping a
28465 tripod, moving on a vehicle, etc.
28466 The filter accepts the following options:
28468 x
28470 y
28471 w
28472 h Specify a rectangular area where to limit the search for motion
28473 vectors. If desired the search for motion vectors can be limited
28474 to a rectangular area of the frame defined by its top left corner,
28475 width and height. These parameters have the same meaning as the
28476 drawbox filter which can be used to visualise the position of the
28477 bounding box.
28478 This is useful when simultaneous movement of subjects within the
28480 frame might be confused for camera motion by the motion vector
28481 search.
28482 If any or all of x, y, w and h are set to -1 then the full frame is
28484 used. This allows later options to be set without specifying the
28485 bounding box for the motion vector search.
28486 Default - search the whole frame.
28488 rx
28490 ry Specify the maximum extent of movement in x and y directions in the
28491 range 0-64 pixels. Default 16.
28492 edge
28494 Specify how to generate pixels to fill blanks at the edge of the
28495 frame. Available values are:
28496 blank, 0
28498 Fill zeroes at blank locations
28499 original, 1
28501 Original image at blank locations
28502 clamp, 2
28504 Extruded edge value at blank locations
28505 mirror, 3
28507 Mirrored edge at blank locations
28508 Default value is mirror.
28510 blocksize
28512 Specify the blocksize to use for motion search. Range 4-128 pixels,
28513 default 8.
28514 contrast
28516 Specify the contrast threshold for blocks. Only blocks with more
28517 than the specified contrast (difference between darkest and
28518 lightest pixels) will be considered. Range 1-255, default 125.
28519 search
28521 Specify the search strategy. Available values are:
28522 exhaustive, 0
28524 Set exhaustive search
28525 less, 1
28527 Set less exhaustive search.
28528 Default value is exhaustive.
28530 filename
28532 If set then a detailed log of the motion search is written to the
28533 specified file.
28534 despill
28536 Remove unwanted contamination of foreground colors, caused by reflected
28537 color of greenscreen or bluescreen.
28538 This filter accepts the following options:
28540 type
28542 Set what type of despill to use.
28543 mix Set how spillmap will be generated.
28545 expand
28547 Set how much to get rid of still remaining spill.
28548 red Controls amount of red in spill area.
28550 green
28552 Controls amount of green in spill area. Should be -1 for
28553 greenscreen.
28554 blue
28556 Controls amount of blue in spill area. Should be -1 for
28557 bluescreen.
28558 brightness
28560 Controls brightness of spill area, preserving colors.
28561 alpha
28563 Modify alpha from generated spillmap.
28564 Commands
28566 This filter supports the all above options as commands.
28568 detelecine
28570 Apply an exact inverse of the telecine operation. It requires a
28571 predefined pattern specified using the pattern option which must be the
28572 same as that passed to the telecine filter.
28573 This filter accepts the following options:
28575 first_field
28577 top, t
28578 top field first
28579 bottom, b
28581 bottom field first The default value is "top".
28582 pattern
28584 A string of numbers representing the pulldown pattern you wish to
28585 apply. The default value is 23.
28586 start_frame
28588 A number representing position of the first frame with respect to
28589 the telecine pattern. This is to be used if the stream is cut. The
28590 default value is 0.
28591 dilation
28593 Apply dilation effect to the video.
28594 This filter replaces the pixel by the local(3x3) maximum.
28596 It accepts the following options:
28598 threshold0
28600 threshold1
28601 threshold2
28602 threshold3
28603 Limit the maximum change for each plane, default is 65535. If 0,
28604 plane will remain unchanged.
28605 coordinates
28607 Flag which specifies the pixel to refer to. Default is 255 i.e. all
28608 eight pixels are used.
28609 Flags to local 3x3 coordinates maps like this:
28611 1 2 3
28613 4 5
28614 6 7 8
28615 Commands
28617 This filter supports the all above options as commands.
28619 displace
28621 Displace pixels as indicated by second and third input stream.
28622 It takes three input streams and outputs one stream, the first input is
28624 the source, and second and third input are displacement maps.
28625 The second input specifies how much to displace pixels along the
28627 x-axis, while the third input specifies how much to displace pixels
28628 along the y-axis. If one of displacement map streams terminates, last
28629 frame from that displacement map will be used.
28630 Note that once generated, displacements maps can be reused over and
28632 over again.
28633 A description of the accepted options follows.
28635 edge
28637 Set displace behavior for pixels that are out of range.
28638 Available values are:
28640 blank
28642 Missing pixels are replaced by black pixels.
28643 smear
28645 Adjacent pixels will spread out to replace missing pixels.
28646 wrap
28648 Out of range pixels are wrapped so they point to pixels of
28649 other side.
28650 mirror
28652 Out of range pixels will be replaced with mirrored pixels.
28653 Default is smear.
28655 Examples
28657 • Add ripple effect to rgb input of video size hd720:
28659 ffmpeg -i INPUT -f lavfi -i nullsrc=s=hd720,lutrgb=128:128:128 -f lavfi -i nullsrc=s=hd720,geq='r=128+30*sin(2*PI*X/400+T):g=128+30*sin(2*PI*X/400+T):b=128+30*sin(2*PI*X/400+T)' -lavfi '[0][1][2]displace' OUTPUT
28661 • Add wave effect to rgb input of video size hd720:
28663 ffmpeg -i INPUT -f lavfi -i nullsrc=hd720,geq='r=128+80*(sin(sqrt((X-W/2)*(X-W/2)+(Y-H/2)*(Y-H/2))/220*2*PI+T)):g=128+80*(sin(sqrt((X-W/2)*(X-W/2)+(Y-H/2)*(Y-H/2))/220*2*PI+T)):b=128+80*(sin(sqrt((X-W/2)*(X-W/2)+(Y-H/2)*(Y-H/2))/220*2*PI+T))' -lavfi '[1]split[x][y],[0][x][y]displace' OUTPUT
28665 dnn_classify
28667 Do classification with deep neural networks based on bounding boxes.
28668 The filter accepts the following options:
28670 dnn_backend
28672 Specify which DNN backend to use for model loading and execution.
28673 This option accepts only openvino now, tensorflow backends will be
28674 added.
28675 model
28677 Set path to model file specifying network architecture and its
28678 parameters. Note that different backends use different file
28679 formats.
28680 input
28682 Set the input name of the dnn network.
28683 output
28685 Set the output name of the dnn network.
28686 confidence
28688 Set the confidence threshold (default: 0.5).
28689 labels
28691 Set path to label file specifying the mapping between label id and
28692 name. Each label name is written in one line, tailing spaces and
28693 empty lines are skipped. The first line is the name of label id 0,
28694 and the second line is the name of label id 1, etc. The label id
28695 is considered as name if the label file is not provided.
28696 backend_configs
28698 Set the configs to be passed into backend
28699 For tensorflow backend, you can set its configs with sess_config
28701 options, please use tools/python/tf_sess_config.py to get the
28702 configs for your system.
28703 dnn_detect
28705 Do object detection with deep neural networks.
28706 The filter accepts the following options:
28708 dnn_backend
28710 Specify which DNN backend to use for model loading and execution.
28711 This option accepts only openvino now, tensorflow backends will be
28712 added.
28713 model
28715 Set path to model file specifying network architecture and its
28716 parameters. Note that different backends use different file
28717 formats.
28718 input
28720 Set the input name of the dnn network.
28721 output
28723 Set the output name of the dnn network.
28724 confidence
28726 Set the confidence threshold (default: 0.5).
28727 labels
28729 Set path to label file specifying the mapping between label id and
28730 name. Each label name is written in one line, tailing spaces and
28731 empty lines are skipped. The first line is the name of label id 0
28732 (usually it is 'background'), and the second line is the name of
28733 label id 1, etc. The label id is considered as name if the label
28734 file is not provided.
28735 backend_configs
28737 Set the configs to be passed into backend. To use async execution,
28738 set async (default: set). Roll back to sync execution if the
28739 backend does not support async.
28740 dnn_processing
28742 Do image processing with deep neural networks. It works together with
28743 another filter which converts the pixel format of the Frame to what the
28744 dnn network requires.
28745 The filter accepts the following options:
28747 dnn_backend
28749 Specify which DNN backend to use for model loading and execution.
28750 This option accepts the following values:
28751 native
28753 Native implementation of DNN loading and execution.
28754 tensorflow
28756 TensorFlow backend. To enable this backend you need to install
28757 the TensorFlow for C library (see
28758 <https://www.tensorflow.org/install/lang_c>) and configure
28759 FFmpeg with "--enable-libtensorflow"
28760 openvino
28762 OpenVINO backend. To enable this backend you need to build and
28763 install the OpenVINO for C library (see
28764 <https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md>)
28765 and configure FFmpeg with "--enable-libopenvino"
28766 (--extra-cflags=-I... --extra-ldflags=-L... might be needed if
28767 the header files and libraries are not installed into system
28768 path)
28769 Default value is native.
28771 model
28773 Set path to model file specifying network architecture and its
28774 parameters. Note that different backends use different file
28775 formats. TensorFlow, OpenVINO and native backend can load files for
28776 only its format.
28777 Native model file (.model) can be generated from TensorFlow model
28779 file (.pb) by using tools/python/convert.py
28780 input
28782 Set the input name of the dnn network.
28783 output
28785 Set the output name of the dnn network.
28786 backend_configs
28788 Set the configs to be passed into backend. To use async execution,
28789 set async (default: set). Roll back to sync execution if the
28790 backend does not support async.
28791 For tensorflow backend, you can set its configs with sess_config
28793 options, please use tools/python/tf_sess_config.py to get the
28794 configs of TensorFlow backend for your system.
28795 Examples
28797 • Remove rain in rgb24 frame with can.pb (see derain filter):
28799 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
28801 • Halve the pixel value of the frame with format gray32f:
28803 ffmpeg -i input.jpg -vf format=grayf32,dnn_processing=model=halve_gray_float.model:input=dnn_in:output=dnn_out:dnn_backend=native -y out.native.png
28805 • Handle the Y channel with srcnn.pb (see sr filter) for frame with
28807 yuv420p (planar YUV formats supported):
28808 ./ffmpeg -i 480p.jpg -vf format=yuv420p,scale=w=iw*2:h=ih*2,dnn_processing=dnn_backend=tensorflow:model=srcnn.pb:input=x:output=y -y srcnn.jpg
28810 • Handle the Y channel with espcn.pb (see sr filter), which changes
28812 frame size, for format yuv420p (planar YUV formats supported),
28813 please use tools/python/tf_sess_config.py to get the configs of
28814 TensorFlow backend for your system.
28815 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y:backend_configs=sess_config=0x10022805320e09cdccccccccccec3f20012a01303801 -y tmp.espcn.jpg
28817 drawbox
28819 Draw a colored box on the input image.
28820 It accepts the following parameters:
28822 x
28824 y The expressions which specify the top left corner coordinates of
28825 the box. It defaults to 0.
28826 width, w
28828 height, h
28829 The expressions which specify the width and height of the box; if 0
28830 they are interpreted as the input width and height. It defaults to
28831 0.
28832 color, c
28834 Specify the color of the box to write. For the general syntax of
28835 this option, check the "Color" section in the ffmpeg-utils manual.
28836 If the special value "invert" is used, the box edge color is the
28837 same as the video with inverted luma.
28838 thickness, t
28840 The expression which sets the thickness of the box edge. A value
28841 of "fill" will create a filled box. Default value is 3.
28842 See below for the list of accepted constants.
28844 replace
28846 Applicable if the input has alpha. With value 1, the pixels of the
28847 painted box will overwrite the video's color and alpha pixels.
28848 Default is 0, which composites the box onto the input, leaving the
28849 video's alpha intact.
28850 The parameters for x, y, w and h and t are expressions containing the
28852 following constants:
28853 dar The input display aspect ratio, it is the same as (w / h) * sar.
28855 hsub
28857 vsub
28858 horizontal and vertical chroma subsample values. For example for
28859 the pixel format "yuv422p" hsub is 2 and vsub is 1.
28860 in_h, ih
28862 in_w, iw
28863 The input width and height.
28864 sar The input sample aspect ratio.
28866 x
28868 y The x and y offset coordinates where the box is drawn.
28869 w
28871 h The width and height of the drawn box.
28872 box_source
28874 Box source can be set as side_data_detection_bboxes if you want to
28875 use box data in detection bboxes of side data.
28876 If box_source is set, the x, y, width and height will be ignored
28878 and still use box data in detection bboxes of side data. So please
28879 do not use this parameter if you were not sure about the box
28880 source.
28881 t The thickness of the drawn box.
28883 These constants allow the x, y, w, h and t expressions to refer to
28885 each other, so you may for example specify "y=x/dar" or "h=w/dar".
28886 Examples
28888 • Draw a black box around the edge of the input image:
28890 drawbox
28892 • Draw a box with color red and an opacity of 50%:
28894 drawbox=10:20:200:60:red@0.5
28896 The previous example can be specified as:
28898 drawbox=x=10:y=20:w=200:h=60:color=red@0.5
28900 • Fill the box with pink color:
28902 drawbox=x=10:y=10:w=100:h=100:color=pink@0.5:t=fill
28904 • Draw a 2-pixel red 2.40:1 mask:
28906 drawbox=x=-t:y=0.5*(ih-iw/2.4)-t:w=iw+t*2:h=iw/2.4+t*2:t=2:c=red
28908 Commands
28910 This filter supports same commands as options. The command accepts the
28912 same syntax of the corresponding option.
28913 If the specified expression is not valid, it is kept at its current
28915 value.
28916 drawgraph
28918 Draw a graph using input video metadata.
28919 It accepts the following parameters:
28921 m1 Set 1st frame metadata key from which metadata values will be used
28923 to draw a graph.
28924 fg1 Set 1st foreground color expression.
28926 m2 Set 2nd frame metadata key from which metadata values will be used
28928 to draw a graph.
28929 fg2 Set 2nd foreground color expression.
28931 m3 Set 3rd frame metadata key from which metadata values will be used
28933 to draw a graph.
28934 fg3 Set 3rd foreground color expression.
28936 m4 Set 4th frame metadata key from which metadata values will be used
28938 to draw a graph.
28939 fg4 Set 4th foreground color expression.
28941 min Set minimal value of metadata value.
28943 max Set maximal value of metadata value.
28945 bg Set graph background color. Default is white.
28947 mode
28949 Set graph mode.
28950 Available values for mode is:
28952 bar
28954 dot
28955 line
28956 Default is "line".
28958 slide
28960 Set slide mode.
28961 Available values for slide is:
28963 frame
28965 Draw new frame when right border is reached.
28966 replace
28968 Replace old columns with new ones.
28969 scroll
28971 Scroll from right to left.
28972 rscroll
28974 Scroll from left to right.
28975 picture
28977 Draw single picture.
28978 Default is "frame".
28980 size
28982 Set size of graph video. For the syntax of this option, check the
28983 "Video size" section in the ffmpeg-utils manual. The default value
28984 is "900x256".
28985 rate, r
28987 Set the output frame rate. Default value is 25.
28988 The foreground color expressions can use the following variables:
28990 MIN Minimal value of metadata value.
28992 MAX Maximal value of metadata value.
28994 VAL Current metadata key value.
28996 The color is defined as 0xAABBGGRR.
28998 Example using metadata from signalstats filter:
29000 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
29002 Example using metadata from ebur128 filter:
29004 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
29006 drawgrid
29008 Draw a grid on the input image.
29009 It accepts the following parameters:
29011 x
29013 y The expressions which specify the coordinates of some point of grid
29014 intersection (meant to configure offset). Both default to 0.
29015 width, w
29017 height, h
29018 The expressions which specify the width and height of the grid
29019 cell, if 0 they are interpreted as the input width and height,
29020 respectively, minus "thickness", so image gets framed. Default to
29021 0.
29022 color, c
29024 Specify the color of the grid. For the general syntax of this
29025 option, check the "Color" section in the ffmpeg-utils manual. If
29026 the special value "invert" is used, the grid color is the same as
29027 the video with inverted luma.
29028 thickness, t
29030 The expression which sets the thickness of the grid line. Default
29031 value is 1.
29032 See below for the list of accepted constants.
29034 replace
29036 Applicable if the input has alpha. With 1 the pixels of the painted
29037 grid will overwrite the video's color and alpha pixels. Default is
29038 0, which composites the grid onto the input, leaving the video's
29039 alpha intact.
29040 The parameters for x, y, w and h and t are expressions containing the
29042 following constants:
29043 dar The input display aspect ratio, it is the same as (w / h) * sar.
29045 hsub
29047 vsub
29048 horizontal and vertical chroma subsample values. For example for
29049 the pixel format "yuv422p" hsub is 2 and vsub is 1.
29050 in_h, ih
29052 in_w, iw
29053 The input grid cell width and height.
29054 sar The input sample aspect ratio.
29056 x
29058 y The x and y coordinates of some point of grid intersection (meant
29059 to configure offset).
29060 w
29062 h The width and height of the drawn cell.
29063 t The thickness of the drawn cell.
29065 These constants allow the x, y, w, h and t expressions to refer to
29067 each other, so you may for example specify "y=x/dar" or "h=w/dar".
29068 Examples
29070 • Draw a grid with cell 100x100 pixels, thickness 2 pixels, with
29072 color red and an opacity of 50%:
29073 drawgrid=width=100:height=100:thickness=2:color=red@0.5
29075 • Draw a white 3x3 grid with an opacity of 50%:
29077 drawgrid=w=iw/3:h=ih/3:t=2:c=white@0.5
29079 Commands
29081 This filter supports same commands as options. The command accepts the
29083 same syntax of the corresponding option.
29084 If the specified expression is not valid, it is kept at its current
29086 value.
29087 drawtext
29089 Draw a text string or text from a specified file on top of a video,
29090 using the libfreetype library.
29091 To enable compilation of this filter, you need to configure FFmpeg with
29093 "--enable-libfreetype". To enable default font fallback and the font
29094 option you need to configure FFmpeg with "--enable-libfontconfig". To
29095 enable the text_shaping option, you need to configure FFmpeg with
29096 "--enable-libfribidi".
29097 Syntax
29099 It accepts the following parameters:
29101 box Used to draw a box around text using the background color. The
29103 value must be either 1 (enable) or 0 (disable). The default value
29104 of box is 0.
29105 boxborderw
29107 Set the width of the border to be drawn around the box using
29108 boxcolor. The default value of boxborderw is 0.
29109 boxcolor
29111 The color to be used for drawing box around text. For the syntax of
29112 this option, check the "Color" section in the ffmpeg-utils manual.
29113 The default value of boxcolor is "white".
29115 line_spacing
29117 Set the line spacing in pixels of the border to be drawn around the
29118 box using box. The default value of line_spacing is 0.
29119 borderw
29121 Set the width of the border to be drawn around the text using
29122 bordercolor. The default value of borderw is 0.
29123 bordercolor
29125 Set the color to be used for drawing border around text. For the
29126 syntax of this option, check the "Color" section in the ffmpeg-
29127 utils manual.
29128 The default value of bordercolor is "black".
29130 expansion
29132 Select how the text is expanded. Can be either "none", "strftime"
29133 (deprecated) or "normal" (default). See the drawtext_expansion,
29134 Text expansion section below for details.
29135 basetime
29137 Set a start time for the count. Value is in microseconds. Only
29138 applied in the deprecated strftime expansion mode. To emulate in
29139 normal expansion mode use the "pts" function, supplying the start
29140 time (in seconds) as the second argument.
29141 fix_bounds
29143 If true, check and fix text coords to avoid clipping.
29144 fontcolor
29146 The color to be used for drawing fonts. For the syntax of this
29147 option, check the "Color" section in the ffmpeg-utils manual.
29148 The default value of fontcolor is "black".
29150 fontcolor_expr
29152 String which is expanded the same way as text to obtain dynamic
29153 fontcolor value. By default this option has empty value and is not
29154 processed. When this option is set, it overrides fontcolor option.
29155 font
29157 The font family to be used for drawing text. By default Sans.
29158 fontfile
29160 The font file to be used for drawing text. The path must be
29161 included. This parameter is mandatory if the fontconfig support is
29162 disabled.
29163 alpha
29165 Draw the text applying alpha blending. The value can be a number
29166 between 0.0 and 1.0. The expression accepts the same variables x,
29167 y as well. The default value is 1. Please see fontcolor_expr.
29168 fontsize
29170 The font size to be used for drawing text. The default value of
29171 fontsize is 16.
29172 text_shaping
29174 If set to 1, attempt to shape the text (for example, reverse the
29175 order of right-to-left text and join Arabic characters) before
29176 drawing it. Otherwise, just draw the text exactly as given. By
29177 default 1 (if supported).
29178 ft_load_flags
29180 The flags to be used for loading the fonts.
29181 The flags map the corresponding flags supported by libfreetype, and
29183 are a combination of the following values:
29184 default
29186 no_scale
29187 no_hinting
29188 render
29189 no_bitmap
29190 vertical_layout
29191 force_autohint
29192 crop_bitmap
29193 pedantic
29194 ignore_global_advance_width
29195 no_recurse
29196 ignore_transform
29197 monochrome
29198 linear_design
29199 no_autohint
29200 Default value is "default".
29202 For more information consult the documentation for the FT_LOAD_*
29204 libfreetype flags.
29205 shadowcolor
29207 The color to be used for drawing a shadow behind the drawn text.
29208 For the syntax of this option, check the "Color" section in the
29209 ffmpeg-utils manual.
29210 The default value of shadowcolor is "black".
29212 shadowx
29214 shadowy
29215 The x and y offsets for the text shadow position with respect to
29216 the position of the text. They can be either positive or negative
29217 values. The default value for both is "0".
29218 start_number
29220 The starting frame number for the n/frame_num variable. The default
29221 value is "0".
29222 tabsize
29224 The size in number of spaces to use for rendering the tab. Default
29225 value is 4.
29226 timecode
29228 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
29229 format. It can be used with or without text parameter.
29230 timecode_rate option must be specified.
29231 timecode_rate, rate, r
29233 Set the timecode frame rate (timecode only). Value will be rounded
29234 to nearest integer. Minimum value is "1". Drop-frame timecode is
29235 supported for frame rates 30 & 60.
29236 tc24hmax
29238 If set to 1, the output of the timecode option will wrap around at
29239 24 hours. Default is 0 (disabled).
29240 text
29242 The text string to be drawn. The text must be a sequence of UTF-8
29243 encoded characters. This parameter is mandatory if no file is
29244 specified with the parameter textfile.
29245 textfile
29247 A text file containing text to be drawn. The text must be a
29248 sequence of UTF-8 encoded characters.
29249 This parameter is mandatory if no text string is specified with the
29251 parameter text.
29252 If both text and textfile are specified, an error is thrown.
29254 text_source
29256 Text source should be set as side_data_detection_bboxes if you want
29257 to use text data in detection bboxes of side data.
29258 If text source is set, text and textfile will be ignored and still
29260 use text data in detection bboxes of side data. So please do not
29261 use this parameter if you are not sure about the text source.
29262 reload
29264 The textfile will be reloaded at specified frame interval. Be sure
29265 to update textfile atomically, or it may be read partially, or even
29266 fail. Range is 0 to INT_MAX. Default is 0.
29267 x
29269 y The expressions which specify the offsets where text will be drawn
29270 within the video frame. They are relative to the top/left border of
29271 the output image.
29272 The default value of x and y is "0".
29274 See below for the list of accepted constants and functions.
29276 The parameters for x and y are expressions containing the following
29278 constants and functions:
29279 dar input display aspect ratio, it is the same as (w / h) * sar
29281 hsub
29283 vsub
29284 horizontal and vertical chroma subsample values. For example for
29285 the pixel format "yuv422p" hsub is 2 and vsub is 1.
29286 line_h, lh
29288 the height of each text line
29289 main_h, h, H
29291 the input height
29292 main_w, w, W
29294 the input width
29295 max_glyph_a, ascent
29297 the maximum distance from the baseline to the highest/upper grid
29298 coordinate used to place a glyph outline point, for all the
29299 rendered glyphs. It is a positive value, due to the grid's
29300 orientation with the Y axis upwards.
29301 max_glyph_d, descent
29303 the maximum distance from the baseline to the lowest grid
29304 coordinate used to place a glyph outline point, for all the
29305 rendered glyphs. This is a negative value, due to the grid's
29306 orientation, with the Y axis upwards.
29307 max_glyph_h
29309 maximum glyph height, that is the maximum height for all the glyphs
29310 contained in the rendered text, it is equivalent to ascent -
29311 descent.
29312 max_glyph_w
29314 maximum glyph width, that is the maximum width for all the glyphs
29315 contained in the rendered text
29316 n the number of input frame, starting from 0
29318 rand(min, max)
29320 return a random number included between min and max
29321 sar The input sample aspect ratio.
29323 t timestamp expressed in seconds, NAN if the input timestamp is
29325 unknown
29326 text_h, th
29328 the height of the rendered text
29329 text_w, tw
29331 the width of the rendered text
29332 x
29334 y the x and y offset coordinates where the text is drawn.
29335 These parameters allow the x and y expressions to refer to each
29337 other, so you can for example specify "y=x/dar".
29338 pict_type
29340 A one character description of the current frame's picture type.
29341 pkt_pos
29343 The current packet's position in the input file or stream (in
29344 bytes, from the start of the input). A value of -1 indicates this
29345 info is not available.
29346 duration
29348 The current packet's duration, in seconds.
29349 pkt_size
29351 The current packet's size (in bytes).
29352 Text expansion
29354 If expansion is set to "strftime", the filter recognizes strftime()
29356 sequences in the provided text and expands them accordingly. Check the
29357 documentation of strftime(). This feature is deprecated.
29358 If expansion is set to "none", the text is printed verbatim.
29360 If expansion is set to "normal" (which is the default), the following
29362 expansion mechanism is used.
29363 The backslash character \, followed by any character, always expands to
29365 the second character.
29366 Sequences of the form "%{...}" are expanded. The text between the
29368 braces is a function name, possibly followed by arguments separated by
29369 ':'. If the arguments contain special characters or delimiters (':' or
29370 '}'), they should be escaped.
29371 Note that they probably must also be escaped as the value for the text
29373 option in the filter argument string and as the filter argument in the
29374 filtergraph description, and possibly also for the shell, that makes up
29375 to four levels of escaping; using a text file avoids these problems.
29376 The following functions are available:
29378 expr, e
29380 The expression evaluation result.
29381 It must take one argument specifying the expression to be
29383 evaluated, which accepts the same constants and functions as the x
29384 and y values. Note that not all constants should be used, for
29385 example the text size is not known when evaluating the expression,
29386 so the constants text_w and text_h will have an undefined value.
29387 expr_int_format, eif
29389 Evaluate the expression's value and output as formatted integer.
29390 The first argument is the expression to be evaluated, just as for
29392 the expr function. The second argument specifies the output
29393 format. Allowed values are x, X, d and u. They are treated exactly
29394 as in the "printf" function. The third parameter is optional and
29395 sets the number of positions taken by the output. It can be used
29396 to add padding with zeros from the left.
29397 gmtime
29399 The time at which the filter is running, expressed in UTC. It can
29400 accept an argument: a strftime() format string. The format string
29401 is extended to support the variable %[1-6]N which prints fractions
29402 of the second with optionally specified number of digits.
29403 localtime
29405 The time at which the filter is running, expressed in the local
29406 time zone. It can accept an argument: a strftime() format string.
29407 The format string is extended to support the variable %[1-6]N which
29408 prints fractions of the second with optionally specified number of
29409 digits.
29410 metadata
29412 Frame metadata. Takes one or two arguments.
29413 The first argument is mandatory and specifies the metadata key.
29415 The second argument is optional and specifies a default value, used
29417 when the metadata key is not found or empty.
29418 Available metadata can be identified by inspecting entries starting
29420 with TAG included within each frame section printed by running
29421 "ffprobe -show_frames".
29422 String metadata generated in filters leading to the drawtext filter
29424 are also available.
29425 n, frame_num
29427 The frame number, starting from 0.
29428 pict_type
29430 A one character description of the current picture type.
29431 pts The timestamp of the current frame. It can take up to three
29433 arguments.
29434 The first argument is the format of the timestamp; it defaults to
29436 "flt" for seconds as a decimal number with microsecond accuracy;
29437 "hms" stands for a formatted [-]HH:MM:SS.mmm timestamp with
29438 millisecond accuracy. "gmtime" stands for the timestamp of the
29439 frame formatted as UTC time; "localtime" stands for the timestamp
29440 of the frame formatted as local time zone time.
29441 The second argument is an offset added to the timestamp.
29443 If the format is set to "hms", a third argument "24HH" may be
29445 supplied to present the hour part of the formatted timestamp in 24h
29446 format (00-23).
29447 If the format is set to "localtime" or "gmtime", a third argument
29449 may be supplied: a strftime() format string. By default, YYYY-MM-
29450 DD HH:MM:SS format will be used.
29451 Commands
29453 This filter supports altering parameters via commands:
29455 reinit
29457 Alter existing filter parameters.
29458 Syntax for the argument is the same as for filter invocation, e.g.
29460 fontsize=56:fontcolor=green:text='Hello World'
29462 Full filter invocation with sendcmd would look like this:
29464 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
29466 If the entire argument can't be parsed or applied as valid values then
29468 the filter will continue with its existing parameters.
29469 Examples
29471 • Draw "Test Text" with font FreeSerif, using the default values for
29473 the optional parameters.
29474 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
29476 • Draw 'Test Text' with font FreeSerif of size 24 at position x=100
29478 and y=50 (counting from the top-left corner of the screen), text is
29479 yellow with a red box around it. Both the text and the box have an
29480 opacity of 20%.
29481 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
29483 x=100: y=50: fontsize=24: fontcolor=yellow@0.2: box=1: boxcolor=red@0.2"
29484 Note that the double quotes are not necessary if spaces are not
29486 used within the parameter list.
29487 • Show the text at the center of the video frame:
29489 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
29491 • Show the text at a random position, switching to a new position
29493 every 30 seconds:
29494 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=if(eq(mod(t\,30)\,0)\,rand(0\,(w-text_w))\,x):y=if(eq(mod(t\,30)\,0)\,rand(0\,(h-text_h))\,y)"
29496 • Show a text line sliding from right to left in the last row of the
29498 video frame. The file LONG_LINE is assumed to contain a single line
29499 with no newlines.
29500 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
29502 • Show the content of file CREDITS off the bottom of the frame and
29504 scroll up.
29505 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
29507 • Draw a single green letter "g", at the center of the input video.
29509 The glyph baseline is placed at half screen height.
29510 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
29512 • Show text for 1 second every 3 seconds:
29514 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
29516 • Use fontconfig to set the font. Note that the colons need to be
29518 escaped.
29519 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
29521 • Draw "Test Text" with font size dependent on height of the video.
29523 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
29525 • Print the date of a real-time encoding (see strftime(3)):
29527 drawtext='fontfile=FreeSans.ttf:text=%{localtime\:%a %b %d %Y}'
29529 • Show text fading in and out (appearing/disappearing):
29531 #!/bin/sh
29533 DS=1.0 # display start
29534 DE=10.0 # display end
29535 FID=1.5 # fade in duration
29536 FOD=5 # fade out duration
29537 ffplay -f lavfi "color,drawtext=text=TEST:fontsize=50:fontfile=FreeSerif.ttf:fontcolor_expr=ff0000%{eif\\\\: clip(255*(1*between(t\\, $DS + $FID\\, $DE - $FOD) + ((t - $DS)/$FID)*between(t\\, $DS\\, $DS + $FID) + (-(t - $DE)/$FOD)*between(t\\, $DE - $FOD\\, $DE) )\\, 0\\, 255) \\\\: x\\\\: 2 }"
29538 • Horizontally align multiple separate texts. Note that max_glyph_a
29540 and the fontsize value are included in the y offset.
29541 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
29543 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
29544 • Plot special lavf.image2dec.source_basename metadata onto each
29546 frame if such metadata exists. Otherwise, plot the string "NA".
29547 Note that image2 demuxer must have option -export_path_metadata 1
29548 for the special metadata fields to be available for filters.
29549 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%{metadata\:lavf.image2dec.source_basename\:NA}':x=10:y=10"
29551 For more information about libfreetype, check:
29553 <http://www.freetype.org/>.
29554 For more information about fontconfig, check:
29556 <http://freedesktop.org/software/fontconfig/fontconfig-user.html>.
29557 For more information about libfribidi, check: <http://fribidi.org/>.
29559 edgedetect
29561 Detect and draw edges. The filter uses the Canny Edge Detection
29562 algorithm.
29563 The filter accepts the following options:
29565 low
29567 high
29568 Set low and high threshold values used by the Canny thresholding
29569 algorithm.
29570 The high threshold selects the "strong" edge pixels, which are then
29572 connected through 8-connectivity with the "weak" edge pixels
29573 selected by the low threshold.
29574 low and high threshold values must be chosen in the range [0,1],
29576 and low should be lesser or equal to high.
29577 Default value for low is "20/255", and default value for high is
29579 "50/255".
29580 mode
29582 Define the drawing mode.
29583 wires
29585 Draw white/gray wires on black background.
29586 colormix
29588 Mix the colors to create a paint/cartoon effect.
29589 canny
29591 Apply Canny edge detector on all selected planes.
29592 Default value is wires.
29594 planes
29596 Select planes for filtering. By default all available planes are
29597 filtered.
29598 Examples
29600 • Standard edge detection with custom values for the hysteresis
29602 thresholding:
29603 edgedetect=low=0.1:high=0.4
29605 • Painting effect without thresholding:
29607 edgedetect=mode=colormix:high=0
29609 elbg
29611 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
29612 For each input image, the filter will compute the optimal mapping from
29614 the input to the output given the codebook length, that is the number
29615 of distinct output colors.
29616 This filter accepts the following options.
29618 codebook_length, l
29620 Set codebook length. The value must be a positive integer, and
29621 represents the number of distinct output colors. Default value is
29622 256.
29623 nb_steps, n
29625 Set the maximum number of iterations to apply for computing the
29626 optimal mapping. The higher the value the better the result and the
29627 higher the computation time. Default value is 1.
29628 seed, s
29630 Set a random seed, must be an integer included between 0 and
29631 UINT32_MAX. If not specified, or if explicitly set to -1, the
29632 filter will try to use a good random seed on a best effort basis.
29633 pal8
29635 Set pal8 output pixel format. This option does not work with
29636 codebook length greater than 256. Default is disabled.
29637 use_alpha
29639 Include alpha values in the quantization calculation. Allows
29640 creating palettized output images (e.g. PNG8) with multiple alpha
29641 smooth blending.
29642 entropy
29644 Measure graylevel entropy in histogram of color channels of video
29645 frames.
29646 It accepts the following parameters:
29648 mode
29650 Can be either normal or diff. Default is normal.
29651 diff mode measures entropy of histogram delta values, absolute
29653 differences between neighbour histogram values.
29654 epx
29656 Apply the EPX magnification filter which is designed for pixel art.
29657 It accepts the following option:
29659 n Set the scaling dimension: 2 for "2xEPX", 3 for "3xEPX". Default
29661 is 3.
29662 eq
29664 Set brightness, contrast, saturation and approximate gamma adjustment.
29665 The filter accepts the following options:
29667 contrast
29669 Set the contrast expression. The value must be a float value in
29670 range -1000.0 to 1000.0. The default value is "1".
29671 brightness
29673 Set the brightness expression. The value must be a float value in
29674 range -1.0 to 1.0. The default value is "0".
29675 saturation
29677 Set the saturation expression. The value must be a float in range
29678 0.0 to 3.0. The default value is "1".
29679 gamma
29681 Set the gamma expression. The value must be a float in range 0.1 to
29682 10.0. The default value is "1".
29683 gamma_r
29685 Set the gamma expression for red. The value must be a float in
29686 range 0.1 to 10.0. The default value is "1".
29687 gamma_g
29689 Set the gamma expression for green. The value must be a float in
29690 range 0.1 to 10.0. The default value is "1".
29691 gamma_b
29693 Set the gamma expression for blue. The value must be a float in
29694 range 0.1 to 10.0. The default value is "1".
29695 gamma_weight
29697 Set the gamma weight expression. It can be used to reduce the
29698 effect of a high gamma value on bright image areas, e.g. keep them
29699 from getting overamplified and just plain white. The value must be
29700 a float in range 0.0 to 1.0. A value of 0.0 turns the gamma
29701 correction all the way down while 1.0 leaves it at its full
29702 strength. Default is "1".
29703 eval
29705 Set when the expressions for brightness, contrast, saturation and
29706 gamma expressions are evaluated.
29707 It accepts the following values:
29709 init
29711 only evaluate expressions once during the filter initialization
29712 or when a command is processed
29713 frame
29715 evaluate expressions for each incoming frame
29716 Default value is init.
29718 The expressions accept the following parameters:
29720 n frame count of the input frame starting from 0
29722 pos byte position of the corresponding packet in the input file, NAN if
29724 unspecified
29725 r frame rate of the input video, NAN if the input frame rate is
29727 unknown
29728 t timestamp expressed in seconds, NAN if the input timestamp is
29730 unknown
29731 Commands
29733 The filter supports the following commands:
29735 contrast
29737 Set the contrast expression.
29738 brightness
29740 Set the brightness expression.
29741 saturation
29743 Set the saturation expression.
29744 gamma
29746 Set the gamma expression.
29747 gamma_r
29749 Set the gamma_r expression.
29750 gamma_g
29752 Set gamma_g expression.
29753 gamma_b
29755 Set gamma_b expression.
29756 gamma_weight
29758 Set gamma_weight expression.
29759 The command accepts the same syntax of the corresponding option.
29761 If the specified expression is not valid, it is kept at its current
29763 value.
29764 erosion
29766 Apply erosion effect to the video.
29767 This filter replaces the pixel by the local(3x3) minimum.
29769 It accepts the following options:
29771 threshold0
29773 threshold1
29774 threshold2
29775 threshold3
29776 Limit the maximum change for each plane, default is 65535. If 0,
29777 plane will remain unchanged.
29778 coordinates
29780 Flag which specifies the pixel to refer to. Default is 255 i.e. all
29781 eight pixels are used.
29782 Flags to local 3x3 coordinates maps like this:
29784 1 2 3
29786 4 5
29787 6 7 8
29788 Commands
29790 This filter supports the all above options as commands.
29792 estdif
29794 Deinterlace the input video ("estdif" stands for "Edge Slope Tracing
29795 Deinterlacing Filter").
29796 Spatial only filter that uses edge slope tracing algorithm to
29798 interpolate missing lines. It accepts the following parameters:
29799 mode
29801 The interlacing mode to adopt. It accepts one of the following
29802 values:
29803 frame
29805 Output one frame for each frame.
29806 field
29808 Output one frame for each field.
29809 The default value is "field".
29811 parity
29813 The picture field parity assumed for the input interlaced video. It
29814 accepts one of the following values:
29815 tff Assume the top field is first.
29817 bff Assume the bottom field is first.
29819 auto
29821 Enable automatic detection of field parity.
29822 The default value is "auto". If the interlacing is unknown or the
29824 decoder does not export this information, top field first will be
29825 assumed.
29826 deint
29828 Specify which frames to deinterlace. Accepts one of the following
29829 values:
29830 all Deinterlace all frames.
29832 interlaced
29834 Only deinterlace frames marked as interlaced.
29835 The default value is "all".
29837 rslope
29839 Specify the search radius for edge slope tracing. Default value is
29840 1. Allowed range is from 1 to 15.
29841 redge
29843 Specify the search radius for best edge matching. Default value is
29844 2. Allowed range is from 0 to 15.
29845 ecost
29847 Specify the edge cost for edge matching. Default value is 1.0.
29848 Allowed range is from 0 to 9.
29849 mcost
29851 Specify the middle cost for edge matching. Default value is 0.5.
29852 Allowed range is from 0 to 1.
29853 dcost
29855 Specify the distance cost for edge matching. Default value is 0.5.
29856 Allowed range is from 0 to 1.
29857 interp
29859 Specify the interpolation used. Default is 4-point interpolation.
29860 It accepts one of the following values:
29861 2p Two-point interpolation.
29863 4p Four-point interpolation.
29865 6p Six-point interpolation.
29867 Commands
29869 This filter supports same commands as options.
29871 exposure
29873 Adjust exposure of the video stream.
29874 The filter accepts the following options:
29876 exposure
29878 Set the exposure correction in EV. Allowed range is from -3.0 to
29879 3.0 EV Default value is 0 EV.
29880 black
29882 Set the black level correction. Allowed range is from -1.0 to 1.0.
29883 Default value is 0.
29884 Commands
29886 This filter supports same commands as options.
29888 extractplanes
29890 Extract color channel components from input video stream into separate
29891 grayscale video streams.
29892 The filter accepts the following option:
29894 planes
29896 Set plane(s) to extract.
29897 Available values for planes are:
29899 y
29901 u
29902 v
29903 a
29904 r
29905 g
29906 b
29907 Choosing planes not available in the input will result in an error.
29909 That means you cannot select "r", "g", "b" planes with "y", "u",
29910 "v" planes at same time.
29911 Examples
29913 • Extract luma, u and v color channel component from input video
29915 frame into 3 grayscale outputs:
29916 ffmpeg -i video.avi -filter_complex 'extractplanes=y+u+v[y][u][v]' -map '[y]' y.avi -map '[u]' u.avi -map '[v]' v.avi
29918 fade
29920 Apply a fade-in/out effect to the input video.
29921 It accepts the following parameters:
29923 type, t
29925 The effect type can be either "in" for a fade-in, or "out" for a
29926 fade-out effect. Default is "in".
29927 start_frame, s
29929 Specify the number of the frame to start applying the fade effect
29930 at. Default is 0.
29931 nb_frames, n
29933 The number of frames that the fade effect lasts. At the end of the
29934 fade-in effect, the output video will have the same intensity as
29935 the input video. At the end of the fade-out transition, the output
29936 video will be filled with the selected color. Default is 25.
29937 alpha
29939 If set to 1, fade only alpha channel, if one exists on the input.
29940 Default value is 0.
29941 start_time, st
29943 Specify the timestamp (in seconds) of the frame to start to apply
29944 the fade effect. If both start_frame and start_time are specified,
29945 the fade will start at whichever comes last. Default is 0.
29946 duration, d
29948 The number of seconds for which the fade effect has to last. At the
29949 end of the fade-in effect the output video will have the same
29950 intensity as the input video, at the end of the fade-out transition
29951 the output video will be filled with the selected color. If both
29952 duration and nb_frames are specified, duration is used. Default is
29953 0 (nb_frames is used by default).
29954 color, c
29956 Specify the color of the fade. Default is "black".
29957 Examples
29959 • Fade in the first 30 frames of video:
29961 fade=in:0:30
29963 The command above is equivalent to:
29965 fade=t=in:s=0:n=30
29967 • Fade out the last 45 frames of a 200-frame video:
29969 fade=out:155:45
29971 fade=type=out:start_frame=155:nb_frames=45
29972 • Fade in the first 25 frames and fade out the last 25 frames of a
29974 1000-frame video:
29975 fade=in:0:25, fade=out:975:25
29977 • Make the first 5 frames yellow, then fade in from frame 5-24:
29979 fade=in:5:20:color=yellow
29981 • Fade in alpha over first 25 frames of video:
29983 fade=in:0:25:alpha=1
29985 • Make the first 5.5 seconds black, then fade in for 0.5 seconds:
29987 fade=t=in:st=5.5:d=0.5
29989 feedback
29991 Apply feedback video filter.
29992 This filter pass cropped input frames to 2nd output. From there it can
29994 be filtered with other video filters. After filter receives frame from
29995 2nd input, that frame is combined on top of original frame from 1st
29996 input and passed to 1st output.
29997 The typical usage is filter only part of frame.
29999 The filter accepts the following options:
30001 x
30003 y Set the top left crop position.
30004 w
30006 h Set the crop size.
30007 Examples
30009 • Blur only top left rectangular part of video frame size 100x100
30011 with gblur filter.
30012 [in][blurin]feedback=x=0:y=0:w=100:h=100[out][blurout];[blurout]gblur=8[blurin]
30014 • Draw black box on top left part of video frame of size 100x100 with
30016 drawbox filter.
30017 [in][blurin]feedback=x=0:y=0:w=100:h=100[out][blurout];[blurout]drawbox=x=0:y=0:w=100:h=100:t=100[blurin]
30019 fftdnoiz
30021 Denoise frames using 3D FFT (frequency domain filtering).
30022 The filter accepts the following options:
30024 sigma
30026 Set the noise sigma constant. This sets denoising strength.
30027 Default value is 1. Allowed range is from 0 to 30. Using very high
30028 sigma with low overlap may give blocking artifacts.
30029 amount
30031 Set amount of denoising. By default all detected noise is reduced.
30032 Default value is 1. Allowed range is from 0 to 1.
30033 block
30035 Set size of block in pixels, Default is 32, can be 8 to 256.
30036 overlap
30038 Set block overlap. Default is 0.5. Allowed range is from 0.2 to
30039 0.8.
30040 method
30042 Set denoising method. Default is "wiener", can also be "hard".
30043 prev
30045 Set number of previous frames to use for denoising. By default is
30046 set to 0.
30047 next
30049 Set number of next frames to to use for denoising. By default is
30050 set to 0.
30051 planes
30053 Set planes which will be filtered, by default are all available
30054 filtered except alpha.
30055 fftfilt
30057 Apply arbitrary expressions to samples in frequency domain
30058 dc_Y
30060 Adjust the dc value (gain) of the luma plane of the image. The
30061 filter accepts an integer value in range 0 to 1000. The default
30062 value is set to 0.
30063 dc_U
30065 Adjust the dc value (gain) of the 1st chroma plane of the image.
30066 The filter accepts an integer value in range 0 to 1000. The default
30067 value is set to 0.
30068 dc_V
30070 Adjust the dc value (gain) of the 2nd chroma plane of the image.
30071 The filter accepts an integer value in range 0 to 1000. The default
30072 value is set to 0.
30073 weight_Y
30075 Set the frequency domain weight expression for the luma plane.
30076 weight_U
30078 Set the frequency domain weight expression for the 1st chroma
30079 plane.
30080 weight_V
30082 Set the frequency domain weight expression for the 2nd chroma
30083 plane.
30084 eval
30086 Set when the expressions are evaluated.
30087 It accepts the following values:
30089 init
30091 Only evaluate expressions once during the filter
30092 initialization.
30093 frame
30095 Evaluate expressions for each incoming frame.
30096 Default value is init.
30098 The filter accepts the following variables:
30100 X
30102 Y The coordinates of the current sample.
30103 W
30105 H The width and height of the image.
30106 N The number of input frame, starting from 0.
30108 WS
30110 HS The size of FFT array for horizontal and vertical processing.
30111 Examples
30113 • High-pass:
30115 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
30117 • Low-pass:
30119 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
30121 • Sharpen:
30123 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
30125 • Blur:
30127 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
30129 field
30131 Extract a single field from an interlaced image using stride arithmetic
30132 to avoid wasting CPU time. The output frames are marked as non-
30133 interlaced.
30134 The filter accepts the following options:
30136 type
30138 Specify whether to extract the top (if the value is 0 or "top") or
30139 the bottom field (if the value is 1 or "bottom").
30140 fieldhint
30142 Create new frames by copying the top and bottom fields from surrounding
30143 frames supplied as numbers by the hint file.
30144 hint
30146 Set file containing hints: absolute/relative frame numbers.
30147 There must be one line for each frame in a clip. Each line must
30149 contain two numbers separated by the comma, optionally followed by
30150 "-" or "+". Numbers supplied on each line of file can not be out
30151 of [N-1,N+1] where N is current frame number for "absolute" mode or
30152 out of [-1, 1] range for "relative" mode. First number tells from
30153 which frame to pick up top field and second number tells from which
30154 frame to pick up bottom field.
30155 If optionally followed by "+" output frame will be marked as
30157 interlaced, else if followed by "-" output frame will be marked as
30158 progressive, else it will be marked same as input frame. If
30159 optionally followed by "t" output frame will use only top field, or
30160 in case of "b" it will use only bottom field. If line starts with
30161 "#" or ";" that line is skipped.
30162 mode
30164 Can be item "absolute" or "relative" or "pattern". Default is
30165 "absolute". The "pattern" mode is same as "relative" mode, except
30166 at last entry of file if there are more frames to process than
30167 "hint" file is seek back to start.
30168 Example of first several lines of "hint" file for "relative" mode:
30170 0,0 - # first frame
30172 1,0 - # second frame, use third's frame top field and second's frame bottom field
30173 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
30174 1,0 -
30175 0,0 -
30176 0,0 -
30177 1,0 -
30178 1,0 -
30179 1,0 -
30180 0,0 -
30181 0,0 -
30182 1,0 -
30183 1,0 -
30184 1,0 -
30185 0,0 -
30186 fieldmatch
30188 Field matching filter for inverse telecine. It is meant to reconstruct
30189 the progressive frames from a telecined stream. The filter does not
30190 drop duplicated frames, so to achieve a complete inverse telecine
30191 "fieldmatch" needs to be followed by a decimation filter such as
30192 decimate in the filtergraph.
30193 The separation of the field matching and the decimation is notably
30195 motivated by the possibility of inserting a de-interlacing filter
30196 fallback between the two. If the source has mixed telecined and real
30197 interlaced content, "fieldmatch" will not be able to match fields for
30198 the interlaced parts. But these remaining combed frames will be marked
30199 as interlaced, and thus can be de-interlaced by a later filter such as
30200 yadif before decimation.
30201 In addition to the various configuration options, "fieldmatch" can take
30203 an optional second stream, activated through the ppsrc option. If
30204 enabled, the frames reconstruction will be based on the fields and
30205 frames from this second stream. This allows the first input to be pre-
30206 processed in order to help the various algorithms of the filter, while
30207 keeping the output lossless (assuming the fields are matched properly).
30208 Typically, a field-aware denoiser, or brightness/contrast adjustments
30209 can help.
30210 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth
30212 project) and VIVTC/VFM (VapourSynth project). The later is a light
30213 clone of TFM from which "fieldmatch" is based on. While the semantic
30214 and usage are very close, some behaviour and options names can differ.
30215 The decimate filter currently only works for constant frame rate input.
30217 If your input has mixed telecined (30fps) and progressive content with
30218 a lower framerate like 24fps use the following filterchain to produce
30219 the necessary cfr stream:
30220 "dejudder,fps=30000/1001,fieldmatch,decimate".
30221 The filter accepts the following options:
30223 order
30225 Specify the assumed field order of the input stream. Available
30226 values are:
30227 auto
30229 Auto detect parity (use FFmpeg's internal parity value).
30230 bff Assume bottom field first.
30232 tff Assume top field first.
30234 Note that it is sometimes recommended not to trust the parity
30236 announced by the stream.
30237 Default value is auto.
30239 mode
30241 Set the matching mode or strategy to use. pc mode is the safest in
30242 the sense that it won't risk creating jerkiness due to duplicate
30243 frames when possible, but if there are bad edits or blended fields
30244 it will end up outputting combed frames when a good match might
30245 actually exist. On the other hand, pcn_ub mode is the most risky in
30246 terms of creating jerkiness, but will almost always find a good
30247 frame if there is one. The other values are all somewhere in
30248 between pc and pcn_ub in terms of risking jerkiness and creating
30249 duplicate frames versus finding good matches in sections with bad
30250 edits, orphaned fields, blended fields, etc.
30251 More details about p/c/n/u/b are available in p/c/n/u/b meaning
30253 section.
30254 Available values are:
30256 pc 2-way matching (p/c)
30258 pc_n
30260 2-way matching, and trying 3rd match if still combed (p/c + n)
30261 pc_u
30263 2-way matching, and trying 3rd match (same order) if still
30264 combed (p/c + u)
30265 pc_n_ub
30267 2-way matching, trying 3rd match if still combed, and trying
30268 4th/5th matches if still combed (p/c + n + u/b)
30269 pcn 3-way matching (p/c/n)
30271 pcn_ub
30273 3-way matching, and trying 4th/5th matches if all 3 of the
30274 original matches are detected as combed (p/c/n + u/b)
30275 The parenthesis at the end indicate the matches that would be used
30277 for that mode assuming order=tff (and field on auto or top).
30278 In terms of speed pc mode is by far the fastest and pcn_ub is the
30280 slowest.
30281 Default value is pc_n.
30283 ppsrc
30285 Mark the main input stream as a pre-processed input, and enable the
30286 secondary input stream as the clean source to pick the fields from.
30287 See the filter introduction for more details. It is similar to the
30288 clip2 feature from VFM/TFM.
30289 Default value is 0 (disabled).
30291 field
30293 Set the field to match from. It is recommended to set this to the
30294 same value as order unless you experience matching failures with
30295 that setting. In certain circumstances changing the field that is
30296 used to match from can have a large impact on matching performance.
30297 Available values are:
30298 auto
30300 Automatic (same value as order).
30301 bottom
30303 Match from the bottom field.
30304 top Match from the top field.
30306 Default value is auto.
30308 mchroma
30310 Set whether or not chroma is included during the match comparisons.
30311 In most cases it is recommended to leave this enabled. You should
30312 set this to 0 only if your clip has bad chroma problems such as
30313 heavy rainbowing or other artifacts. Setting this to 0 could also
30314 be used to speed things up at the cost of some accuracy.
30315 Default value is 1.
30317 y0
30319 y1 These define an exclusion band which excludes the lines between y0
30320 and y1 from being included in the field matching decision. An
30321 exclusion band can be used to ignore subtitles, a logo, or other
30322 things that may interfere with the matching. y0 sets the starting
30323 scan line and y1 sets the ending line; all lines in between y0 and
30324 y1 (including y0 and y1) will be ignored. Setting y0 and y1 to the
30325 same value will disable the feature. y0 and y1 defaults to 0.
30326 scthresh
30328 Set the scene change detection threshold as a percentage of maximum
30329 change on the luma plane. Good values are in the "[8.0, 14.0]"
30330 range. Scene change detection is only relevant in case
30331 combmatch=sc. The range for scthresh is "[0.0, 100.0]".
30332 Default value is 12.0.
30334 combmatch
30336 When combatch is not none, "fieldmatch" will take into account the
30337 combed scores of matches when deciding what match to use as the
30338 final match. Available values are:
30339 none
30341 No final matching based on combed scores.
30342 sc Combed scores are only used when a scene change is detected.
30344 full
30346 Use combed scores all the time.
30347 Default is sc.
30349 combdbg
30351 Force "fieldmatch" to calculate the combed metrics for certain
30352 matches and print them. This setting is known as micout in TFM/VFM
30353 vocabulary. Available values are:
30354 none
30356 No forced calculation.
30357 pcn Force p/c/n calculations.
30359 pcnub
30361 Force p/c/n/u/b calculations.
30362 Default value is none.
30364 cthresh
30366 This is the area combing threshold used for combed frame detection.
30367 This essentially controls how "strong" or "visible" combing must be
30368 to be detected. Larger values mean combing must be more visible
30369 and smaller values mean combing can be less visible or strong and
30370 still be detected. Valid settings are from -1 (every pixel will be
30371 detected as combed) to 255 (no pixel will be detected as combed).
30372 This is basically a pixel difference value. A good range is "[8,
30373 12]".
30374 Default value is 9.
30376 chroma
30378 Sets whether or not chroma is considered in the combed frame
30379 decision. Only disable this if your source has chroma problems
30380 (rainbowing, etc.) that are causing problems for the combed frame
30381 detection with chroma enabled. Actually, using chroma=0 is usually
30382 more reliable, except for the case where there is chroma only
30383 combing in the source.
30384 Default value is 0.
30386 blockx
30388 blocky
30389 Respectively set the x-axis and y-axis size of the window used
30390 during combed frame detection. This has to do with the size of the
30391 area in which combpel pixels are required to be detected as combed
30392 for a frame to be declared combed. See the combpel parameter
30393 description for more info. Possible values are any number that is
30394 a power of 2 starting at 4 and going up to 512.
30395 Default value is 16.
30397 combpel
30399 The number of combed pixels inside any of the blocky by blockx size
30400 blocks on the frame for the frame to be detected as combed. While
30401 cthresh controls how "visible" the combing must be, this setting
30402 controls "how much" combing there must be in any localized area (a
30403 window defined by the blockx and blocky settings) on the frame.
30404 Minimum value is 0 and maximum is "blocky x blockx" (at which point
30405 no frames will ever be detected as combed). This setting is known
30406 as MI in TFM/VFM vocabulary.
30407 Default value is 80.
30409 p/c/n/u/b meaning
30411 p/c/n
30413 We assume the following telecined stream:
30415 Top fields: 1 2 2 3 4
30417 Bottom fields: 1 2 3 4 4
30418 The numbers correspond to the progressive frame the fields relate to.
30420 Here, the first two frames are progressive, the 3rd and 4th are combed,
30421 and so on.
30422 When "fieldmatch" is configured to run a matching from bottom
30424 (field=bottom) this is how this input stream get transformed:
30425 Input stream:
30427 T 1 2 2 3 4
30428 B 1 2 3 4 4 <-- matching reference
30429 Matches: c c n n c
30431 Output stream:
30433 T 1 2 3 4 4
30434 B 1 2 3 4 4
30435 As a result of the field matching, we can see that some frames get
30437 duplicated. To perform a complete inverse telecine, you need to rely
30438 on a decimation filter after this operation. See for instance the
30439 decimate filter.
30440 The same operation now matching from top fields (field=top) looks like
30442 this:
30443 Input stream:
30445 T 1 2 2 3 4 <-- matching reference
30446 B 1 2 3 4 4
30447 Matches: c c p p c
30449 Output stream:
30451 T 1 2 2 3 4
30452 B 1 2 2 3 4
30453 In these examples, we can see what p, c and n mean; basically, they
30455 refer to the frame and field of the opposite parity:
30456 *<p matches the field of the opposite parity in the previous frame>
30458 *<c matches the field of the opposite parity in the current frame>
30459 *<n matches the field of the opposite parity in the next frame>
30460 u/b
30462 The u and b matching are a bit special in the sense that they match
30464 from the opposite parity flag. In the following examples, we assume
30465 that we are currently matching the 2nd frame (Top:2, bottom:2).
30466 According to the match, a 'x' is placed above and below each matched
30467 fields.
30468 With bottom matching (field=bottom):
30470 Match: c p n b u
30472 x x x x x
30474 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
30475 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
30476 x x x x x
30477 Output frames:
30479 2 1 2 2 2
30480 2 2 2 1 3
30481 With top matching (field=top):
30483 Match: c p n b u
30485 x x x x x
30487 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
30488 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
30489 x x x x x
30490 Output frames:
30492 2 2 2 1 2
30493 2 1 3 2 2
30494 Examples
30496 Simple IVTC of a top field first telecined stream:
30498 fieldmatch=order=tff:combmatch=none, decimate
30500 Advanced IVTC, with fallback on yadif for still combed frames:
30502 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
30504 fieldorder
30506 Transform the field order of the input video.
30507 It accepts the following parameters:
30509 order
30511 The output field order. Valid values are tff for top field first or
30512 bff for bottom field first.
30513 The default value is tff.
30515 The transformation is done by shifting the picture content up or down
30517 by one line, and filling the remaining line with appropriate picture
30518 content. This method is consistent with most broadcast field order
30519 converters.
30520 If the input video is not flagged as being interlaced, or it is already
30522 flagged as being of the required output field order, then this filter
30523 does not alter the incoming video.
30524 It is very useful when converting to or from PAL DV material, which is
30526 bottom field first.
30527 For example:
30529 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
30531 fifo, afifo
30533 Buffer input images and send them when they are requested.
30534 It is mainly useful when auto-inserted by the libavfilter framework.
30536 It does not take parameters.
30538 fillborders
30540 Fill borders of the input video, without changing video stream
30541 dimensions. Sometimes video can have garbage at the four edges and you
30542 may not want to crop video input to keep size multiple of some number.
30543 This filter accepts the following options:
30545 left
30547 Number of pixels to fill from left border.
30548 right
30550 Number of pixels to fill from right border.
30551 top Number of pixels to fill from top border.
30553 bottom
30555 Number of pixels to fill from bottom border.
30556 mode
30558 Set fill mode.
30559 It accepts the following values:
30561 smear
30563 fill pixels using outermost pixels
30564 mirror
30566 fill pixels using mirroring (half sample symmetric)
30567 fixed
30569 fill pixels with constant value
30570 reflect
30572 fill pixels using reflecting (whole sample symmetric)
30573 wrap
30575 fill pixels using wrapping
30576 fade
30578 fade pixels to constant value
30579 margins
30581 fill pixels at top and bottom with weighted averages pixels
30582 near borders
30583 Default is smear.
30585 color
30587 Set color for pixels in fixed or fade mode. Default is black.
30588 Commands
30590 This filter supports same commands as options. The command accepts the
30592 same syntax of the corresponding option.
30593 If the specified expression is not valid, it is kept at its current
30595 value.
30596 find_rect
30598 Find a rectangular object
30599 It accepts the following options:
30601 object
30603 Filepath of the object image, needs to be in gray8.
30604 threshold
30606 Detection threshold, default is 0.5.
30607 mipmaps
30609 Number of mipmaps, default is 3.
30610 xmin, ymin, xmax, ymax
30612 Specifies the rectangle in which to search.
30613 discard
30615 Discard frames where object is not detected. Default is disabled.
30616 Examples
30618 • Cover a rectangular object by the supplied image of a given video
30620 using ffmpeg:
30621 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
30623 floodfill
30625 Flood area with values of same pixel components with another values.
30626 It accepts the following options:
30628 x Set pixel x coordinate.
30630 y Set pixel y coordinate.
30632 s0 Set source #0 component value.
30634 s1 Set source #1 component value.
30636 s2 Set source #2 component value.
30638 s3 Set source #3 component value.
30640 d0 Set destination #0 component value.
30642 d1 Set destination #1 component value.
30644 d2 Set destination #2 component value.
30646 d3 Set destination #3 component value.
30648 format
30650 Convert the input video to one of the specified pixel formats.
30651 Libavfilter will try to pick one that is suitable as input to the next
30652 filter.
30653 It accepts the following parameters:
30655 pix_fmts
30657 A '|'-separated list of pixel format names, such as
30658 "pix_fmts=yuv420p|monow|rgb24".
30659 Examples
30661 • Convert the input video to the yuv420p format
30663 format=pix_fmts=yuv420p
30665 Convert the input video to any of the formats in the list
30667 format=pix_fmts=yuv420p|yuv444p|yuv410p
30669 fps
30671 Convert the video to specified constant frame rate by duplicating or
30672 dropping frames as necessary.
30673 It accepts the following parameters:
30675 fps The desired output frame rate. It accepts expressions containing
30677 the following constants:
30678 source_fps
30680 The input's frame rate
30681 ntsc
30683 NTSC frame rate of "30000/1001"
30684 pal PAL frame rate of 25.0
30686 film
30688 Film frame rate of 24.0
30689 ntsc_film
30691 NTSC-film frame rate of "24000/1001"
30692 The default is 25.
30694 start_time
30696 Assume the first PTS should be the given value, in seconds. This
30697 allows for padding/trimming at the start of stream. By default, no
30698 assumption is made about the first frame's expected PTS, so no
30699 padding or trimming is done. For example, this could be set to 0
30700 to pad the beginning with duplicates of the first frame if a video
30701 stream starts after the audio stream or to trim any frames with a
30702 negative PTS.
30703 round
30705 Timestamp (PTS) rounding method.
30706 Possible values are:
30708 zero
30710 round towards 0
30711 inf round away from 0
30713 down
30715 round towards -infinity
30716 up round towards +infinity
30718 near
30720 round to nearest
30721 The default is "near".
30723 eof_action
30725 Action performed when reading the last frame.
30726 Possible values are:
30728 round
30730 Use same timestamp rounding method as used for other frames.
30731 pass
30733 Pass through last frame if input duration has not been reached
30734 yet.
30735 The default is "round".
30737 Alternatively, the options can be specified as a flat string:
30739 fps[:start_time[:round]].
30740 See also the setpts filter.
30742 Examples
30744 • A typical usage in order to set the fps to 25:
30746 fps=fps=25
30748 • Sets the fps to 24, using abbreviation and rounding method to round
30750 to nearest:
30751 fps=fps=film:round=near
30753 framepack
30755 Pack two different video streams into a stereoscopic video, setting
30756 proper metadata on supported codecs. The two views should have the same
30757 size and framerate and processing will stop when the shorter video
30758 ends. Please note that you may conveniently adjust view properties with
30759 the scale and fps filters.
30760 It accepts the following parameters:
30762 format
30764 The desired packing format. Supported values are:
30765 sbs The views are next to each other (default).
30767 tab The views are on top of each other.
30769 lines
30771 The views are packed by line.
30772 columns
30774 The views are packed by column.
30775 frameseq
30777 The views are temporally interleaved.
30778 Some examples:
30780 # Convert left and right views into a frame-sequential video
30782 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
30783 # Convert views into a side-by-side video with the same output resolution as the input
30785 ffmpeg -i LEFT -i RIGHT -filter_complex [0:v]scale=w=iw/2[left],[1:v]scale=w=iw/2[right],[left][right]framepack=sbs OUTPUT
30786 framerate
30788 Change the frame rate by interpolating new video output frames from the
30789 source frames.
30790 This filter is not designed to function correctly with interlaced
30792 media. If you wish to change the frame rate of interlaced media then
30793 you are required to deinterlace before this filter and re-interlace
30794 after this filter.
30795 A description of the accepted options follows.
30797 fps Specify the output frames per second. This option can also be
30799 specified as a value alone. The default is 50.
30800 interp_start
30802 Specify the start of a range where the output frame will be created
30803 as a linear interpolation of two frames. The range is [0-255], the
30804 default is 15.
30805 interp_end
30807 Specify the end of a range where the output frame will be created
30808 as a linear interpolation of two frames. The range is [0-255], the
30809 default is 240.
30810 scene
30812 Specify the level at which a scene change is detected as a value
30813 between 0 and 100 to indicate a new scene; a low value reflects a
30814 low probability for the current frame to introduce a new scene,
30815 while a higher value means the current frame is more likely to be
30816 one. The default is 8.2.
30817 flags
30819 Specify flags influencing the filter process.
30820 Available value for flags is:
30822 scene_change_detect, scd
30824 Enable scene change detection using the value of the option
30825 scene. This flag is enabled by default.
30826 framestep
30828 Select one frame every N-th frame.
30829 This filter accepts the following option:
30831 step
30833 Select frame after every "step" frames. Allowed values are
30834 positive integers higher than 0. Default value is 1.
30835 freezedetect
30837 Detect frozen video.
30838 This filter logs a message and sets frame metadata when it detects that
30840 the input video has no significant change in content during a specified
30841 duration. Video freeze detection calculates the mean average absolute
30842 difference of all the components of video frames and compares it to a
30843 noise floor.
30844 The printed times and duration are expressed in seconds. The
30846 "lavfi.freezedetect.freeze_start" metadata key is set on the first
30847 frame whose timestamp equals or exceeds the detection duration and it
30848 contains the timestamp of the first frame of the freeze. The
30849 "lavfi.freezedetect.freeze_duration" and
30850 "lavfi.freezedetect.freeze_end" metadata keys are set on the first
30851 frame after the freeze.
30852 The filter accepts the following options:
30854 noise, n
30856 Set noise tolerance. Can be specified in dB (in case "dB" is
30857 appended to the specified value) or as a difference ratio between 0
30858 and 1. Default is -60dB, or 0.001.
30859 duration, d
30861 Set freeze duration until notification (default is 2 seconds).
30862 freezeframes
30864 Freeze video frames.
30865 This filter freezes video frames using frame from 2nd input.
30867 The filter accepts the following options:
30869 first
30871 Set number of first frame from which to start freeze.
30872 last
30874 Set number of last frame from which to end freeze.
30875 replace
30877 Set number of frame from 2nd input which will be used instead of
30878 replaced frames.
30879 frei0r
30881 Apply a frei0r effect to the input video.
30882 To enable the compilation of this filter, you need to install the
30884 frei0r header and configure FFmpeg with "--enable-frei0r".
30885 It accepts the following parameters:
30887 filter_name
30889 The name of the frei0r effect to load. If the environment variable
30890 FREI0R_PATH is defined, the frei0r effect is searched for in each
30891 of the directories specified by the colon-separated list in
30892 FREI0R_PATH. Otherwise, the standard frei0r paths are searched, in
30893 this order: HOME/.frei0r-1/lib/, /usr/local/lib/frei0r-1/,
30894 /usr/lib/frei0r-1/.
30895 filter_params
30897 A '|'-separated list of parameters to pass to the frei0r effect.
30898 A frei0r effect parameter can be a boolean (its value is either "y" or
30900 "n"), a double, a color (specified as R/G/B, where R, G, and B are
30901 floating point numbers between 0.0 and 1.0, inclusive) or a color
30902 description as specified in the "Color" section in the ffmpeg-utils
30903 manual, a position (specified as X/Y, where X and Y are floating point
30904 numbers) and/or a string.
30905 The number and types of parameters depend on the loaded effect. If an
30907 effect parameter is not specified, the default value is set.
30908 Examples
30910 • Apply the distort0r effect, setting the first two double
30912 parameters:
30913 frei0r=filter_name=distort0r:filter_params=0.5|0.01
30915 • Apply the colordistance effect, taking a color as the first
30917 parameter:
30918 frei0r=colordistance:0.2/0.3/0.4
30920 frei0r=colordistance:violet
30921 frei0r=colordistance:0x112233
30922 • Apply the perspective effect, specifying the top left and top right
30924 image positions:
30925 frei0r=perspective:0.2/0.2|0.8/0.2
30927 For more information, see <http://frei0r.dyne.org>
30929 Commands
30931 This filter supports the filter_params option as commands.
30933 fspp
30935 Apply fast and simple postprocessing. It is a faster version of spp.
30936 It splits (I)DCT into horizontal/vertical passes. Unlike the simple
30938 post- processing filter, one of them is performed once per block, not
30939 per pixel. This allows for much higher speed.
30940 The filter accepts the following options:
30942 quality
30944 Set quality. This option defines the number of levels for
30945 averaging. It accepts an integer in the range 4-5. Default value is
30946 4.
30947 qp Force a constant quantization parameter. It accepts an integer in
30949 range 0-63. If not set, the filter will use the QP from the video
30950 stream (if available).
30951 strength
30953 Set filter strength. It accepts an integer in range -15 to 32.
30954 Lower values mean more details but also more artifacts, while
30955 higher values make the image smoother but also blurrier. Default
30956 value is 0 − PSNR optimal.
30957 use_bframe_qp
30959 Enable the use of the QP from the B-Frames if set to 1. Using this
30960 option may cause flicker since the B-Frames have often larger QP.
30961 Default is 0 (not enabled).
30962 gblur
30964 Apply Gaussian blur filter.
30965 The filter accepts the following options:
30967 sigma
30969 Set horizontal sigma, standard deviation of Gaussian blur. Default
30970 is 0.5.
30971 steps
30973 Set number of steps for Gaussian approximation. Default is 1.
30974 planes
30976 Set which planes to filter. By default all planes are filtered.
30977 sigmaV
30979 Set vertical sigma, if negative it will be same as "sigma".
30980 Default is -1.
30981 Commands
30983 This filter supports same commands as options. The command accepts the
30985 same syntax of the corresponding option.
30986 If the specified expression is not valid, it is kept at its current
30988 value.
30989 geq
30991 Apply generic equation to each pixel.
30992 The filter accepts the following options:
30994 lum_expr, lum
30996 Set the luminance expression.
30997 cb_expr, cb
30999 Set the chrominance blue expression.
31000 cr_expr, cr
31002 Set the chrominance red expression.
31003 alpha_expr, a
31005 Set the alpha expression.
31006 red_expr, r
31008 Set the red expression.
31009 green_expr, g
31011 Set the green expression.
31012 blue_expr, b
31014 Set the blue expression.
31015 The colorspace is selected according to the specified options. If one
31017 of the lum_expr, cb_expr, or cr_expr options is specified, the filter
31018 will automatically select a YCbCr colorspace. If one of the red_expr,
31019 green_expr, or blue_expr options is specified, it will select an RGB
31020 colorspace.
31021 If one of the chrominance expression is not defined, it falls back on
31023 the other one. If no alpha expression is specified it will evaluate to
31024 opaque value. If none of chrominance expressions are specified, they
31025 will evaluate to the luminance expression.
31026 The expressions can use the following variables and functions:
31028 N The sequential number of the filtered frame, starting from 0.
31030 X
31032 Y The coordinates of the current sample.
31033 W
31035 H The width and height of the image.
31036 SW
31038 SH Width and height scale depending on the currently filtered plane.
31039 It is the ratio between the corresponding luma plane number of
31040 pixels and the current plane ones. E.g. for YUV4:2:0 the values are
31041 "1,1" for the luma plane, and "0.5,0.5" for chroma planes.
31042 T Time of the current frame, expressed in seconds.
31044 p(x, y)
31046 Return the value of the pixel at location (x,y) of the current
31047 plane.
31048 lum(x, y)
31050 Return the value of the pixel at location (x,y) of the luminance
31051 plane.
31052 cb(x, y)
31054 Return the value of the pixel at location (x,y) of the blue-
31055 difference chroma plane. Return 0 if there is no such plane.
31056 cr(x, y)
31058 Return the value of the pixel at location (x,y) of the red-
31059 difference chroma plane. Return 0 if there is no such plane.
31060 r(x, y)
31062 g(x, y)
31063 b(x, y)
31064 Return the value of the pixel at location (x,y) of the
31065 red/green/blue component. Return 0 if there is no such component.
31066 alpha(x, y)
31068 Return the value of the pixel at location (x,y) of the alpha plane.
31069 Return 0 if there is no such plane.
31070 psum(x,y), lumsum(x, y), cbsum(x,y), crsum(x,y), rsum(x,y), gsum(x,y),
31072 bsum(x,y), alphasum(x,y)
31073 Sum of sample values in the rectangle from (0,0) to (x,y), this
31074 allows obtaining sums of samples within a rectangle. See the
31075 functions without the sum postfix.
31076 interpolation
31078 Set one of interpolation methods:
31079 nearest, n
31081 bilinear, b
31082 Default is bilinear.
31084 For functions, if x and y are outside the area, the value will be
31086 automatically clipped to the closer edge.
31087 Please note that this filter can use multiple threads in which case
31089 each slice will have its own expression state. If you want to use only
31090 a single expression state because your expressions depend on previous
31091 state then you should limit the number of filter threads to 1.
31092 Examples
31094 • Flip the image horizontally:
31096 geq=p(W-X\,Y)
31098 • Generate a bidimensional sine wave, with angle "PI/3" and a
31100 wavelength of 100 pixels:
31101 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
31103 • Generate a fancy enigmatic moving light:
31105 nullsrc=s=256x256,geq=random(1)/hypot(X-cos(N*0.07)*W/2-W/2\,Y-sin(N*0.09)*H/2-H/2)^2*1000000*sin(N*0.02):128:128
31107 • Generate a quick emboss effect:
31109 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
31111 • Modify RGB components depending on pixel position:
31113 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
31115 • Create a radial gradient that is the same size as the input (also
31117 see the vignette filter):
31118 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
31120 gradfun
31122 Fix the banding artifacts that are sometimes introduced into nearly
31123 flat regions by truncation to 8-bit color depth. Interpolate the
31124 gradients that should go where the bands are, and dither them.
31125 It is designed for playback only. Do not use it prior to lossy
31127 compression, because compression tends to lose the dither and bring
31128 back the bands.
31129 It accepts the following parameters:
31131 strength
31133 The maximum amount by which the filter will change any one pixel.
31134 This is also the threshold for detecting nearly flat regions.
31135 Acceptable values range from .51 to 64; the default value is 1.2.
31136 Out-of-range values will be clipped to the valid range.
31137 radius
31139 The neighborhood to fit the gradient to. A larger radius makes for
31140 smoother gradients, but also prevents the filter from modifying the
31141 pixels near detailed regions. Acceptable values are 8-32; the
31142 default value is 16. Out-of-range values will be clipped to the
31143 valid range.
31144 Alternatively, the options can be specified as a flat string:
31146 strength[:radius]
31147 Examples
31149 • Apply the filter with a 3.5 strength and radius of 8:
31151 gradfun=3.5:8
31153 • Specify radius, omitting the strength (which will fall-back to the
31155 default value):
31156 gradfun=radius=8
31158 graphmonitor
31160 Show various filtergraph stats.
31161 With this filter one can debug complete filtergraph. Especially issues
31163 with links filling with queued frames.
31164 The filter accepts the following options:
31166 size, s
31168 Set video output size. Default is hd720.
31169 opacity, o
31171 Set video opacity. Default is 0.9. Allowed range is from 0 to 1.
31172 mode, m
31174 Set output mode, can be fulll or compact. In compact mode only
31175 filters with some queued frames have displayed stats.
31176 flags, f
31178 Set flags which enable which stats are shown in video.
31179 Available values for flags are:
31181 queue
31183 Display number of queued frames in each link.
31184 frame_count_in
31186 Display number of frames taken from filter.
31187 frame_count_out
31189 Display number of frames given out from filter.
31190 frame_count_delta
31192 Display delta number of frames between above two values.
31193 pts Display current filtered frame pts.
31195 pts_delta
31197 Display pts delta between current and previous frame.
31198 time
31200 Display current filtered frame time.
31201 time_delta
31203 Display time delta between current and previous frame.
31204 timebase
31206 Display time base for filter link.
31207 format
31209 Display used format for filter link.
31210 size
31212 Display video size or number of audio channels in case of audio
31213 used by filter link.
31214 rate
31216 Display video frame rate or sample rate in case of audio used
31217 by filter link.
31218 eof Display link output status.
31220 sample_count_in
31222 Display number of samples taken from filter.
31223 sample_count_out
31225 Display number of samples given out from filter.
31226 sample_count_delta
31228 Display delta number of samples between above two values.
31229 rate, r
31231 Set upper limit for video rate of output stream, Default value is
31232 25. This guarantee that output video frame rate will not be higher
31233 than this value.
31234 grayworld
31236 A color constancy filter that applies color correction based on the
31237 grayworld assumption
31238 See:
31240 <https://www.researchgate.net/publication/275213614_A_New_Color_Correction_Method_for_Underwater_Imaging>
31241 The algorithm uses linear light, so input data should be linearized
31243 beforehand (and possibly correctly tagged).
31244 ffmpeg -i INPUT -vf zscale=transfer=linear,grayworld,zscale=transfer=bt709,format=yuv420p OUTPUT
31246 greyedge
31248 A color constancy variation filter which estimates scene illumination
31249 via grey edge algorithm and corrects the scene colors accordingly.
31250 See: <https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf>
31252 The filter accepts the following options:
31254 difford
31256 The order of differentiation to be applied on the scene. Must be
31257 chosen in the range [0,2] and default value is 1.
31258 minknorm
31260 The Minkowski parameter to be used for calculating the Minkowski
31261 distance. Must be chosen in the range [0,20] and default value is
31262 1. Set to 0 for getting max value instead of calculating Minkowski
31263 distance.
31264 sigma
31266 The standard deviation of Gaussian blur to be applied on the scene.
31267 Must be chosen in the range [0,1024.0] and default value = 1.
31268 floor( sigma * break_off_sigma(3) ) can't be equal to 0 if difford
31269 is greater than 0.
31270 Examples
31272 • Grey Edge:
31274 greyedge=difford=1:minknorm=5:sigma=2
31276 • Max Edge:
31278 greyedge=difford=1:minknorm=0:sigma=2
31280 guided
31282 Apply guided filter for edge-preserving smoothing, dehazing and so on.
31283 The filter accepts the following options:
31285 radius
31287 Set the box radius in pixels. Allowed range is 1 to 20. Default is
31288 3.
31289 eps Set regularization parameter (with square). Allowed range is 0 to
31291 1. Default is 0.01.
31292 mode
31294 Set filter mode. Can be "basic" or "fast". Default is "basic".
31295 sub Set subsampling ratio for "fast" mode. Range is 2 to 64. Default
31297 is 4. No subsampling occurs in "basic" mode.
31298 guidance
31300 Set guidance mode. Can be "off" or "on". Default is "off". If
31301 "off", single input is required. If "on", two inputs of the same
31302 resolution and pixel format are required. The second input serves
31303 as the guidance.
31304 planes
31306 Set planes to filter. Default is first only.
31307 Commands
31309 This filter supports the all above options as commands.
31311 Examples
31313 • Edge-preserving smoothing with guided filter:
31315 ffmpeg -i in.png -vf guided out.png
31317 • Dehazing, structure-transferring filtering, detail enhancement with
31319 guided filter. For the generation of guidance image, refer to
31320 paper "Guided Image Filtering". See:
31321 <http://kaiminghe.com/publications/pami12guidedfilter.pdf>.
31322 ffmpeg -i in.png -i guidance.png -filter_complex guided=guidance=on out.png
31324 haldclut
31326 Apply a Hald CLUT to a video stream.
31327 First input is the video stream to process, and second one is the Hald
31329 CLUT. The Hald CLUT input can be a simple picture or a complete video
31330 stream.
31331 The filter accepts the following options:
31333 clut
31335 Set which CLUT video frames will be processed from second input
31336 stream, can be first or all. Default is all.
31337 shortest
31339 Force termination when the shortest input terminates. Default is 0.
31340 repeatlast
31342 Continue applying the last CLUT after the end of the stream. A
31343 value of 0 disable the filter after the last frame of the CLUT is
31344 reached. Default is 1.
31345 "haldclut" also has the same interpolation options as lut3d (both
31347 filters share the same internals).
31348 This filter also supports the framesync options.
31350 More information about the Hald CLUT can be found on Eskil Steenberg's
31352 website (Hald CLUT author) at
31353 <http://www.quelsolaar.com/technology/clut.html>.
31354 Commands
31356 This filter supports the "interp" option as commands.
31358 Workflow examples
31360 Hald CLUT video stream
31362 Generate an identity Hald CLUT stream altered with various effects:
31364 ffmpeg -f lavfi -i B<haldclutsrc>=8 -vf "hue=H=2*PI*t:s=sin(2*PI*t)+1, curves=cross_process" -t 10 -c:v ffv1 clut.nut
31366 Note: make sure you use a lossless codec.
31368 Then use it with "haldclut" to apply it on some random stream:
31370 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
31372 The Hald CLUT will be applied to the 10 first seconds (duration of
31374 clut.nut), then the latest picture of that CLUT stream will be applied
31375 to the remaining frames of the "mandelbrot" stream.
31376 Hald CLUT with preview
31378 A Hald CLUT is supposed to be a squared image of "Level*Level*Level" by
31380 "Level*Level*Level" pixels. For a given Hald CLUT, FFmpeg will select
31381 the biggest possible square starting at the top left of the picture.
31382 The remaining padding pixels (bottom or right) will be ignored. This
31383 area can be used to add a preview of the Hald CLUT.
31384 Typically, the following generated Hald CLUT will be supported by the
31386 "haldclut" filter:
31387 ffmpeg -f lavfi -i B<haldclutsrc>=8 -vf "
31389 pad=iw+320 [padded_clut];
31390 smptebars=s=320x256, split [a][b];
31391 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
31392 [main][b] overlay=W-320" -frames:v 1 clut.png
31393 It contains the original and a preview of the effect of the CLUT: SMPTE
31395 color bars are displayed on the right-top, and below the same color
31396 bars processed by the color changes.
31397 Then, the effect of this Hald CLUT can be visualized with:
31399 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
31401 hflip
31403 Flip the input video horizontally.
31404 For example, to horizontally flip the input video with ffmpeg:
31406 ffmpeg -i in.avi -vf "hflip" out.avi
31408 histeq
31410 This filter applies a global color histogram equalization on a per-
31411 frame basis.
31412 It can be used to correct video that has a compressed range of pixel
31414 intensities. The filter redistributes the pixel intensities to
31415 equalize their distribution across the intensity range. It may be
31416 viewed as an "automatically adjusting contrast filter". This filter is
31417 useful only for correcting degraded or poorly captured source video.
31418 The filter accepts the following options:
31420 strength
31422 Determine the amount of equalization to be applied. As the
31423 strength is reduced, the distribution of pixel intensities more-
31424 and-more approaches that of the input frame. The value must be a
31425 float number in the range [0,1] and defaults to 0.200.
31426 intensity
31428 Set the maximum intensity that can generated and scale the output
31429 values appropriately. The strength should be set as desired and
31430 then the intensity can be limited if needed to avoid washing-out.
31431 The value must be a float number in the range [0,1] and defaults to
31432 0.210.
31433 antibanding
31435 Set the antibanding level. If enabled the filter will randomly vary
31436 the luminance of output pixels by a small amount to avoid banding
31437 of the histogram. Possible values are "none", "weak" or "strong".
31438 It defaults to "none".
31439 histogram
31441 Compute and draw a color distribution histogram for the input video.
31442 The computed histogram is a representation of the color component
31444 distribution in an image.
31445 Standard histogram displays the color components distribution in an
31447 image. Displays color graph for each color component. Shows
31448 distribution of the Y, U, V, A or R, G, B components, depending on
31449 input format, in the current frame. Below each graph a color component
31450 scale meter is shown.
31451 The filter accepts the following options:
31453 level_height
31455 Set height of level. Default value is 200. Allowed range is [50,
31456 2048].
31457 scale_height
31459 Set height of color scale. Default value is 12. Allowed range is
31460 [0, 40].
31461 display_mode
31463 Set display mode. It accepts the following values:
31464 stack
31466 Per color component graphs are placed below each other.
31467 parade
31469 Per color component graphs are placed side by side.
31470 overlay
31472 Presents information identical to that in the "parade", except
31473 that the graphs representing color components are superimposed
31474 directly over one another.
31475 Default is "stack".
31477 levels_mode
31479 Set mode. Can be either "linear", or "logarithmic". Default is
31480 "linear".
31481 components
31483 Set what color components to display. Default is 7.
31484 fgopacity
31486 Set foreground opacity. Default is 0.7.
31487 bgopacity
31489 Set background opacity. Default is 0.5.
31490 colors_mode
31492 Set colors mode. It accepts the following values:
31493 whiteonblack
31495 blackonwhite
31496 whiteongray
31497 blackongray
31498 coloronblack
31499 coloronwhite
31500 colorongray
31501 blackoncolor
31502 whiteoncolor
31503 grayoncolor
31504 Default is "whiteonblack".
31506 Examples
31508 • Calculate and draw histogram:
31510 ffplay -i input -vf histogram
31512 hqdn3d
31514 This is a high precision/quality 3d denoise filter. It aims to reduce
31515 image noise, producing smooth images and making still images really
31516 still. It should enhance compressibility.
31517 It accepts the following optional parameters:
31519 luma_spatial
31521 A non-negative floating point number which specifies spatial luma
31522 strength. It defaults to 4.0.
31523 chroma_spatial
31525 A non-negative floating point number which specifies spatial chroma
31526 strength. It defaults to 3.0*luma_spatial/4.0.
31527 luma_tmp
31529 A floating point number which specifies luma temporal strength. It
31530 defaults to 6.0*luma_spatial/4.0.
31531 chroma_tmp
31533 A floating point number which specifies chroma temporal strength.
31534 It defaults to luma_tmp*chroma_spatial/luma_spatial.
31535 Commands
31537 This filter supports same commands as options. The command accepts the
31539 same syntax of the corresponding option.
31540 If the specified expression is not valid, it is kept at its current
31542 value.
31543 hwdownload
31545 Download hardware frames to system memory.
31546 The input must be in hardware frames, and the output a non-hardware
31548 format. Not all formats will be supported on the output - it may be
31549 necessary to insert an additional format filter immediately following
31550 in the graph to get the output in a supported format.
31551 hwmap
31553 Map hardware frames to system memory or to another device.
31554 This filter has several different modes of operation; which one is used
31556 depends on the input and output formats:
31557 • Hardware frame input, normal frame output
31559 Map the input frames to system memory and pass them to the output.
31561 If the original hardware frame is later required (for example,
31562 after overlaying something else on part of it), the hwmap filter
31563 can be used again in the next mode to retrieve it.
31564 • Normal frame input, hardware frame output
31566 If the input is actually a software-mapped hardware frame, then
31568 unmap it - that is, return the original hardware frame.
31569 Otherwise, a device must be provided. Create new hardware surfaces
31571 on that device for the output, then map them back to the software
31572 format at the input and give those frames to the preceding filter.
31573 This will then act like the hwupload filter, but may be able to
31574 avoid an additional copy when the input is already in a compatible
31575 format.
31576 • Hardware frame input and output
31578 A device must be supplied for the output, either directly or with
31580 the derive_device option. The input and output devices must be of
31581 different types and compatible - the exact meaning of this is
31582 system-dependent, but typically it means that they must refer to
31583 the same underlying hardware context (for example, refer to the
31584 same graphics card).
31585 If the input frames were originally created on the output device,
31587 then unmap to retrieve the original frames.
31588 Otherwise, map the frames to the output device - create new
31590 hardware frames on the output corresponding to the frames on the
31591 input.
31592 The following additional parameters are accepted:
31594 mode
31596 Set the frame mapping mode. Some combination of:
31597 read
31599 The mapped frame should be readable.
31600 write
31602 The mapped frame should be writeable.
31603 overwrite
31605 The mapping will always overwrite the entire frame.
31606 This may improve performance in some cases, as the original
31608 contents of the frame need not be loaded.
31609 direct
31611 The mapping must not involve any copying.
31612 Indirect mappings to copies of frames are created in some cases
31614 where either direct mapping is not possible or it would have
31615 unexpected properties. Setting this flag ensures that the
31616 mapping is direct and will fail if that is not possible.
31617 Defaults to read+write if not specified.
31619 derive_device type
31621 Rather than using the device supplied at initialisation, instead
31622 derive a new device of type type from the device the input frames
31623 exist on.
31624 reverse
31626 In a hardware to hardware mapping, map in reverse - create frames
31627 in the sink and map them back to the source. This may be necessary
31628 in some cases where a mapping in one direction is required but only
31629 the opposite direction is supported by the devices being used.
31630 This option is dangerous - it may break the preceding filter in
31632 undefined ways if there are any additional constraints on that
31633 filter's output. Do not use it without fully understanding the
31634 implications of its use.
31635 hwupload
31637 Upload system memory frames to hardware surfaces.
31638 The device to upload to must be supplied when the filter is
31640 initialised. If using ffmpeg, select the appropriate device with the
31641 -filter_hw_device option or with the derive_device option. The input
31642 and output devices must be of different types and compatible - the
31643 exact meaning of this is system-dependent, but typically it means that
31644 they must refer to the same underlying hardware context (for example,
31645 refer to the same graphics card).
31646 The following additional parameters are accepted:
31648 derive_device type
31650 Rather than using the device supplied at initialisation, instead
31651 derive a new device of type type from the device the input frames
31652 exist on.
31653 hwupload_cuda
31655 Upload system memory frames to a CUDA device.
31656 It accepts the following optional parameters:
31658 device
31660 The number of the CUDA device to use
31661 hqx
31663 Apply a high-quality magnification filter designed for pixel art. This
31664 filter was originally created by Maxim Stepin.
31665 It accepts the following option:
31667 n Set the scaling dimension: 2 for "hq2x", 3 for "hq3x" and 4 for
31669 "hq4x". Default is 3.
31670 hstack
31672 Stack input videos horizontally.
31673 All streams must be of same pixel format and of same height.
31675 Note that this filter is faster than using overlay and pad filter to
31677 create same output.
31678 The filter accepts the following option:
31680 inputs
31682 Set number of input streams. Default is 2.
31683 shortest
31685 If set to 1, force the output to terminate when the shortest input
31686 terminates. Default value is 0.
31687 hsvhold
31689 Turns a certain HSV range into gray values.
31690 This filter measures color difference between set HSV color in options
31692 and ones measured in video stream. Depending on options, output colors
31693 can be changed to be gray or not.
31694 The filter accepts the following options:
31696 hue Set the hue value which will be used in color difference
31698 calculation. Allowed range is from -360 to 360. Default value is
31699 0.
31700 sat Set the saturation value which will be used in color difference
31702 calculation. Allowed range is from -1 to 1. Default value is 0.
31703 val Set the value which will be used in color difference calculation.
31705 Allowed range is from -1 to 1. Default value is 0.
31706 similarity
31708 Set similarity percentage with the key color. Allowed range is
31709 from 0 to 1. Default value is 0.01.
31710 0.00001 matches only the exact key color, while 1.0 matches
31712 everything.
31713 blend
31715 Blend percentage. Allowed range is from 0 to 1. Default value is
31716 0.
31717 0.0 makes pixels either fully gray, or not gray at all.
31719 Higher values result in more gray pixels, with a higher gray pixel
31721 the more similar the pixels color is to the key color.
31722 hsvkey
31724 Turns a certain HSV range into transparency.
31725 This filter measures color difference between set HSV color in options
31727 and ones measured in video stream. Depending on options, output colors
31728 can be changed to transparent by adding alpha channel.
31729 The filter accepts the following options:
31731 hue Set the hue value which will be used in color difference
31733 calculation. Allowed range is from -360 to 360. Default value is
31734 0.
31735 sat Set the saturation value which will be used in color difference
31737 calculation. Allowed range is from -1 to 1. Default value is 0.
31738 val Set the value which will be used in color difference calculation.
31740 Allowed range is from -1 to 1. Default value is 0.
31741 similarity
31743 Set similarity percentage with the key color. Allowed range is
31744 from 0 to 1. Default value is 0.01.
31745 0.00001 matches only the exact key color, while 1.0 matches
31747 everything.
31748 blend
31750 Blend percentage. Allowed range is from 0 to 1. Default value is
31751 0.
31752 0.0 makes pixels either fully transparent, or not transparent at
31754 all.
31755 Higher values result in semi-transparent pixels, with a higher
31757 transparency the more similar the pixels color is to the key color.
31758 hue
31760 Modify the hue and/or the saturation of the input.
31761 It accepts the following parameters:
31763 h Specify the hue angle as a number of degrees. It accepts an
31765 expression, and defaults to "0".
31766 s Specify the saturation in the [-10,10] range. It accepts an
31768 expression and defaults to "1".
31769 H Specify the hue angle as a number of radians. It accepts an
31771 expression, and defaults to "0".
31772 b Specify the brightness in the [-10,10] range. It accepts an
31774 expression and defaults to "0".
31775 h and H are mutually exclusive, and can't be specified at the same
31777 time.
31778 The b, h, H and s option values are expressions containing the
31780 following constants:
31781 n frame count of the input frame starting from 0
31783 pts presentation timestamp of the input frame expressed in time base
31785 units
31786 r frame rate of the input video, NAN if the input frame rate is
31788 unknown
31789 t timestamp expressed in seconds, NAN if the input timestamp is
31791 unknown
31792 tb time base of the input video
31794 Examples
31796 • Set the hue to 90 degrees and the saturation to 1.0:
31798 hue=h=90:s=1
31800 • Same command but expressing the hue in radians:
31802 hue=H=PI/2:s=1
31804 • Rotate hue and make the saturation swing between 0 and 2 over a
31806 period of 1 second:
31807 hue="H=2*PI*t: s=sin(2*PI*t)+1"
31809 • Apply a 3 seconds saturation fade-in effect starting at 0:
31811 hue="s=min(t/3\,1)"
31813 The general fade-in expression can be written as:
31815 hue="s=min(0\, max((t-START)/DURATION\, 1))"
31817 • Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
31819 hue="s=max(0\, min(1\, (8-t)/3))"
31821 The general fade-out expression can be written as:
31823 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
31825 Commands
31827 This filter supports the following commands:
31829 b
31831 s
31832 h
31833 H Modify the hue and/or the saturation and/or brightness of the input
31834 video. The command accepts the same syntax of the corresponding
31835 option.
31836 If the specified expression is not valid, it is kept at its current
31838 value.
31839 huesaturation
31841 Apply hue-saturation-intensity adjustments to input video stream.
31842 This filter operates in RGB colorspace.
31844 This filter accepts the following options:
31846 hue Set the hue shift in degrees to apply. Default is 0. Allowed range
31848 is from -180 to 180.
31849 saturation
31851 Set the saturation shift. Default is 0. Allowed range is from -1
31852 to 1.
31853 intensity
31855 Set the intensity shift. Default is 0. Allowed range is from -1 to
31856 1.
31857 colors
31859 Set which primary and complementary colors are going to be
31860 adjusted. This options is set by providing one or multiple values.
31861 This can select multiple colors at once. By default all colors are
31862 selected.
31863 r Adjust reds.
31865 y Adjust yellows.
31867 g Adjust greens.
31869 c Adjust cyans.
31871 b Adjust blues.
31873 m Adjust magentas.
31875 a Adjust all colors.
31877 strength
31879 Set strength of filtering. Allowed range is from 0 to 100. Default
31880 value is 1.
31881 rw, gw, bw
31883 Set weight for each RGB component. Allowed range is from 0 to 1.
31884 By default is set to 0.333, 0.334, 0.333. Those options are used
31885 in saturation and lightess processing.
31886 lightness
31888 Set preserving lightness, by default is disabled. Adjusting hues
31889 can change lightness from original RGB triplet, with this option
31890 enabled lightness is kept at same value.
31891 hysteresis
31893 Grow first stream into second stream by connecting components. This
31894 makes it possible to build more robust edge masks.
31895 This filter accepts the following options:
31897 planes
31899 Set which planes will be processed as bitmap, unprocessed planes
31900 will be copied from first stream. By default value 0xf, all planes
31901 will be processed.
31902 threshold
31904 Set threshold which is used in filtering. If pixel component value
31905 is higher than this value filter algorithm for connecting
31906 components is activated. By default value is 0.
31907 The "hysteresis" filter also supports the framesync options.
31909 iccdetect
31911 Detect the colorspace from an embedded ICC profile (if present), and
31912 update the frame's tags accordingly.
31913 This filter accepts the following options:
31915 force
31917 If true, the frame's existing colorspace tags will always be
31918 overridden by values detected from an ICC profile. Otherwise, they
31919 will only be assigned if they contain "unknown". Enabled by
31920 default.
31921 iccgen
31923 Generate ICC profiles and attach them to frames.
31924 This filter accepts the following options:
31926 color_primaries
31928 color_trc
31929 Configure the colorspace that the ICC profile will be generated
31930 for. The default value of "auto" infers the value from the input
31931 frame's metadata, defaulting to BT.709/sRGB as appropriate.
31932 See the setparams filter for a list of possible values, but note
31934 that "unknown" are not valid values for this filter.
31935 force
31937 If true, an ICC profile will be generated even if it would
31938 overwrite an already existing ICC profile. Disabled by default.
31939 identity
31941 Obtain the identity score between two input videos.
31942 This filter takes two input videos.
31944 Both input videos must have the same resolution and pixel format for
31946 this filter to work correctly. Also it assumes that both inputs have
31947 the same number of frames, which are compared one by one.
31948 The obtained per component, average, min and max identity score is
31950 printed through the logging system.
31951 The filter stores the calculated identity scores of each frame in frame
31953 metadata.
31954 This filter also supports the framesync options.
31956 In the below example the input file main.mpg being processed is
31958 compared with the reference file ref.mpg.
31959 ffmpeg -i main.mpg -i ref.mpg -lavfi identity -f null -
31961 idet
31963 Detect video interlacing type.
31964 This filter tries to detect if the input frames are interlaced,
31966 progressive, top or bottom field first. It will also try to detect
31967 fields that are repeated between adjacent frames (a sign of telecine).
31968 Single frame detection considers only immediately adjacent frames when
31970 classifying each frame. Multiple frame detection incorporates the
31971 classification history of previous frames.
31972 The filter will log these metadata values:
31974 single.current_frame
31976 Detected type of current frame using single-frame detection. One
31977 of: ``tff'' (top field first), ``bff'' (bottom field first),
31978 ``progressive'', or ``undetermined''
31979 single.tff
31981 Cumulative number of frames detected as top field first using
31982 single-frame detection.
31983 multiple.tff
31985 Cumulative number of frames detected as top field first using
31986 multiple-frame detection.
31987 single.bff
31989 Cumulative number of frames detected as bottom field first using
31990 single-frame detection.
31991 multiple.current_frame
31993 Detected type of current frame using multiple-frame detection. One
31994 of: ``tff'' (top field first), ``bff'' (bottom field first),
31995 ``progressive'', or ``undetermined''
31996 multiple.bff
31998 Cumulative number of frames detected as bottom field first using
31999 multiple-frame detection.
32000 single.progressive
32002 Cumulative number of frames detected as progressive using single-
32003 frame detection.
32004 multiple.progressive
32006 Cumulative number of frames detected as progressive using multiple-
32007 frame detection.
32008 single.undetermined
32010 Cumulative number of frames that could not be classified using
32011 single-frame detection.
32012 multiple.undetermined
32014 Cumulative number of frames that could not be classified using
32015 multiple-frame detection.
32016 repeated.current_frame
32018 Which field in the current frame is repeated from the last. One of
32019 ``neither'', ``top'', or ``bottom''.
32020 repeated.neither
32022 Cumulative number of frames with no repeated field.
32023 repeated.top
32025 Cumulative number of frames with the top field repeated from the
32026 previous frame's top field.
32027 repeated.bottom
32029 Cumulative number of frames with the bottom field repeated from the
32030 previous frame's bottom field.
32031 The filter accepts the following options:
32033 intl_thres
32035 Set interlacing threshold.
32036 prog_thres
32038 Set progressive threshold.
32039 rep_thres
32041 Threshold for repeated field detection.
32042 half_life
32044 Number of frames after which a given frame's contribution to the
32045 statistics is halved (i.e., it contributes only 0.5 to its
32046 classification). The default of 0 means that all frames seen are
32047 given full weight of 1.0 forever.
32048 analyze_interlaced_flag
32050 When this is not 0 then idet will use the specified number of
32051 frames to determine if the interlaced flag is accurate, it will not
32052 count undetermined frames. If the flag is found to be accurate it
32053 will be used without any further computations, if it is found to be
32054 inaccurate it will be cleared without any further computations.
32055 This allows inserting the idet filter as a low computational method
32056 to clean up the interlaced flag
32057 il
32059 Deinterleave or interleave fields.
32060 This filter allows one to process interlaced images fields without
32062 deinterlacing them. Deinterleaving splits the input frame into 2 fields
32063 (so called half pictures). Odd lines are moved to the top half of the
32064 output image, even lines to the bottom half. You can process (filter)
32065 them independently and then re-interleave them.
32066 The filter accepts the following options:
32068 luma_mode, l
32070 chroma_mode, c
32071 alpha_mode, a
32072 Available values for luma_mode, chroma_mode and alpha_mode are:
32073 none
32075 Do nothing.
32076 deinterleave, d
32078 Deinterleave fields, placing one above the other.
32079 interleave, i
32081 Interleave fields. Reverse the effect of deinterleaving.
32082 Default value is "none".
32084 luma_swap, ls
32086 chroma_swap, cs
32087 alpha_swap, as
32088 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default
32089 value is 0.
32090 Commands
32092 This filter supports the all above options as commands.
32094 inflate
32096 Apply inflate effect to the video.
32097 This filter replaces the pixel by the local(3x3) average by taking into
32099 account only values higher than the pixel.
32100 It accepts the following options:
32102 threshold0
32104 threshold1
32105 threshold2
32106 threshold3
32107 Limit the maximum change for each plane, default is 65535. If 0,
32108 plane will remain unchanged.
32109 Commands
32111 This filter supports the all above options as commands.
32113 interlace
32115 Simple interlacing filter from progressive contents. This interleaves
32116 upper (or lower) lines from odd frames with lower (or upper) lines from
32117 even frames, halving the frame rate and preserving image height.
32118 Original Original New Frame
32120 Frame 'j' Frame 'j+1' (tff)
32121 ========== =========== ==================
32122 Line 0 --------------------> Frame 'j' Line 0
32123 Line 1 Line 1 ----> Frame 'j+1' Line 1
32124 Line 2 ---------------------> Frame 'j' Line 2
32125 Line 3 Line 3 ----> Frame 'j+1' Line 3
32126 ... ... ...
32127 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
32128 It accepts the following optional parameters:
32130 scan
32132 This determines whether the interlaced frame is taken from the even
32133 (tff - default) or odd (bff) lines of the progressive frame.
32134 lowpass
32136 Vertical lowpass filter to avoid twitter interlacing and reduce
32137 moire patterns.
32138 0, off
32140 Disable vertical lowpass filter
32141 1, linear
32143 Enable linear filter (default)
32144 2, complex
32146 Enable complex filter. This will slightly less reduce twitter
32147 and moire but better retain detail and subjective sharpness
32148 impression.
32149 kerndeint
32151 Deinterlace input video by applying Donald Graft's adaptive kernel
32152 deinterling. Work on interlaced parts of a video to produce progressive
32153 frames.
32154 The description of the accepted parameters follows.
32156 thresh
32158 Set the threshold which affects the filter's tolerance when
32159 determining if a pixel line must be processed. It must be an
32160 integer in the range [0,255] and defaults to 10. A value of 0 will
32161 result in applying the process on every pixels.
32162 map Paint pixels exceeding the threshold value to white if set to 1.
32164 Default is 0.
32165 order
32167 Set the fields order. Swap fields if set to 1, leave fields alone
32168 if 0. Default is 0.
32169 sharp
32171 Enable additional sharpening if set to 1. Default is 0.
32172 twoway
32174 Enable twoway sharpening if set to 1. Default is 0.
32175 Examples
32177 • Apply default values:
32179 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
32181 • Enable additional sharpening:
32183 kerndeint=sharp=1
32185 • Paint processed pixels in white:
32187 kerndeint=map=1
32189 kirsch
32191 Apply kirsch operator to input video stream.
32192 The filter accepts the following option:
32194 planes
32196 Set which planes will be processed, unprocessed planes will be
32197 copied. By default value 0xf, all planes will be processed.
32198 scale
32200 Set value which will be multiplied with filtered result.
32201 delta
32203 Set value which will be added to filtered result.
32204 Commands
32206 This filter supports the all above options as commands.
32208 lagfun
32210 Slowly update darker pixels.
32211 This filter makes short flashes of light appear longer. This filter
32213 accepts the following options:
32214 decay
32216 Set factor for decaying. Default is .95. Allowed range is from 0 to
32217 1.
32218 planes
32220 Set which planes to filter. Default is all. Allowed range is from 0
32221 to 15.
32222 Commands
32224 This filter supports the all above options as commands.
32226 lenscorrection
32228 Correct radial lens distortion
32229 This filter can be used to correct for radial distortion as can result
32231 from the use of wide angle lenses, and thereby re-rectify the image. To
32232 find the right parameters one can use tools available for example as
32233 part of opencv or simply trial-and-error. To use opencv use the
32234 calibration sample (under samples/cpp) from the opencv sources and
32235 extract the k1 and k2 coefficients from the resulting matrix.
32236 Note that effectively the same filter is available in the open-source
32238 tools Krita and Digikam from the KDE project.
32239 In contrast to the vignette filter, which can also be used to
32241 compensate lens errors, this filter corrects the distortion of the
32242 image, whereas vignette corrects the brightness distribution, so you
32243 may want to use both filters together in certain cases, though you will
32244 have to take care of ordering, i.e. whether vignetting should be
32245 applied before or after lens correction.
32246 Options
32248 The filter accepts the following options:
32250 cx Relative x-coordinate of the focal point of the image, and thereby
32252 the center of the distortion. This value has a range [0,1] and is
32253 expressed as fractions of the image width. Default is 0.5.
32254 cy Relative y-coordinate of the focal point of the image, and thereby
32256 the center of the distortion. This value has a range [0,1] and is
32257 expressed as fractions of the image height. Default is 0.5.
32258 k1 Coefficient of the quadratic correction term. This value has a
32260 range [-1,1]. 0 means no correction. Default is 0.
32261 k2 Coefficient of the double quadratic correction term. This value has
32263 a range [-1,1]. 0 means no correction. Default is 0.
32264 i Set interpolation type. Can be "nearest" or "bilinear". Default is
32266 "nearest".
32267 fc Specify the color of the unmapped pixels. For the syntax of this
32269 option, check the "Color" section in the ffmpeg-utils manual.
32270 Default color is "black@0".
32271 The formula that generates the correction is:
32273 r_src = r_tgt * (1 + k1 * (r_tgt / r_0)^2 + k2 * (r_tgt / r_0)^4)
32275 where r_0 is halve of the image diagonal and r_src and r_tgt are the
32277 distances from the focal point in the source and target images,
32278 respectively.
32279 Commands
32281 This filter supports the all above options as commands.
32283 lensfun
32285 Apply lens correction via the lensfun library
32286 (<http://lensfun.sourceforge.net/>).
32287 The "lensfun" filter requires the camera make, camera model, and lens
32289 model to apply the lens correction. The filter will load the lensfun
32290 database and query it to find the corresponding camera and lens entries
32291 in the database. As long as these entries can be found with the given
32292 options, the filter can perform corrections on frames. Note that
32293 incomplete strings will result in the filter choosing the best match
32294 with the given options, and the filter will output the chosen camera
32295 and lens models (logged with level "info"). You must provide the make,
32296 camera model, and lens model as they are required.
32297 To obtain a list of available makes and models, leave out one or both
32299 of "make" and "model" options. The filter will send the full list to
32300 the log with level "INFO". The first column is the make and the second
32301 column is the model. To obtain a list of available lenses, set any
32302 values for make and model and leave out the "lens_model" option. The
32303 filter will send the full list of lenses in the log with level "INFO".
32304 The ffmpeg tool will exit after the list is printed.
32305 The filter accepts the following options:
32307 make
32309 The make of the camera (for example, "Canon"). This option is
32310 required.
32311 model
32313 The model of the camera (for example, "Canon EOS 100D"). This
32314 option is required.
32315 lens_model
32317 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6
32318 IS STM"). This option is required.
32319 db_path
32321 The full path to the lens database folder. If not set, the filter
32322 will attempt to load the database from the install path when the
32323 library was built. Default is unset.
32324 mode
32326 The type of correction to apply. The following values are valid
32327 options:
32328 vignetting
32330 Enables fixing lens vignetting.
32331 geometry
32333 Enables fixing lens geometry. This is the default.
32334 subpixel
32336 Enables fixing chromatic aberrations.
32337 vig_geo
32339 Enables fixing lens vignetting and lens geometry.
32340 vig_subpixel
32342 Enables fixing lens vignetting and chromatic aberrations.
32343 distortion
32345 Enables fixing both lens geometry and chromatic aberrations.
32346 all Enables all possible corrections.
32348 focal_length
32350 The focal length of the image/video (zoom; expected constant for
32351 video). For example, a 18--55mm lens has focal length range of
32352 [18--55], so a value in that range should be chosen when using that
32353 lens. Default 18.
32354 aperture
32356 The aperture of the image/video (expected constant for video). Note
32357 that aperture is only used for vignetting correction. Default 3.5.
32358 focus_distance
32360 The focus distance of the image/video (expected constant for
32361 video). Note that focus distance is only used for vignetting and
32362 only slightly affects the vignetting correction process. If
32363 unknown, leave it at the default value (which is 1000).
32364 scale
32366 The scale factor which is applied after transformation. After
32367 correction the video is no longer necessarily rectangular. This
32368 parameter controls how much of the resulting image is visible. The
32369 value 0 means that a value will be chosen automatically such that
32370 there is little or no unmapped area in the output image. 1.0 means
32371 that no additional scaling is done. Lower values may result in more
32372 of the corrected image being visible, while higher values may avoid
32373 unmapped areas in the output.
32374 target_geometry
32376 The target geometry of the output image/video. The following values
32377 are valid options:
32378 rectilinear (default)
32380 fisheye
32381 panoramic
32382 equirectangular
32383 fisheye_orthographic
32384 fisheye_stereographic
32385 fisheye_equisolid
32386 fisheye_thoby
32387 reverse
32388 Apply the reverse of image correction (instead of correcting
32389 distortion, apply it).
32390 interpolation
32392 The type of interpolation used when correcting distortion. The
32393 following values are valid options:
32394 nearest
32396 linear (default)
32397 lanczos
32398 Examples
32400 • Apply lens correction with make "Canon", camera model "Canon EOS
32402 100D", and lens model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with
32403 focal length of "18" and aperture of "8.0".
32404 ffmpeg -i input.mov -vf lensfun=make=Canon:model="Canon EOS 100D":lens_model="Canon EF-S 18-55mm f/3.5-5.6 IS STM":focal_length=18:aperture=8 -c:v h264 -b:v 8000k output.mov
32406 • Apply the same as before, but only for the first 5 seconds of
32408 video.
32409 ffmpeg -i input.mov -vf lensfun=make=Canon:model="Canon EOS 100D":lens_model="Canon EF-S 18-55mm f/3.5-5.6 IS STM":focal_length=18:aperture=8:enable='lte(t\,5)' -c:v h264 -b:v 8000k output.mov
32411 libplacebo
32413 Flexible GPU-accelerated processing filter based on libplacebo
32414 (<https://code.videolan.org/videolan/libplacebo>). Note that this
32415 filter currently only accepts Vulkan input frames.
32416 Options
32418 The options for this filter are divided into the following sections:
32420 Output mode
32422 These options control the overall output mode. By default, libplacebo
32424 will try to preserve the source colorimetry and size as best as it can,
32425 but it will apply any embedded film grain, dolby vision metadata or
32426 anamorphic SAR present in source frames.
32427 w
32429 h Set the output video dimension expression. Default value is the
32430 input dimension.
32431 Allows for the same expressions as the scale filter.
32433 format
32435 Set the output format override. If unset (the default), frames will
32436 be output in the same format as the respective input frames.
32437 Otherwise, format conversion will be performed.
32438 force_original_aspect_ratio
32440 force_divisible_by
32441 Work the same as the identical scale filter options.
32442 normalize_sar
32444 If enabled, output frames will always have a pixel aspect ratio of
32445 1:1. This will introduce padding/cropping as necessary. If disabled
32446 (the default), any aspect ratio mismatches, including those from
32447 e.g. anamorphic video sources, are forwarded to the output pixel
32448 aspect ratio.
32449 pad_crop_ratio
32451 Specifies a ratio (between 0.0 and 1.0) between padding and
32452 cropping when the input aspect ratio does not match the output
32453 aspect ratio and normalize_sar is in effect. The default of 0.0
32454 always pads the content with black borders, while a value of 1.0
32455 always crops off parts of the content. Intermediate values are
32456 possible, leading to a mix of the two approaches.
32457 colorspace
32459 color_primaries
32460 color_trc
32461 range
32462 Configure the colorspace that output frames will be delivered in.
32463 The default value of "auto" outputs frames in the same format as
32464 the input frames, leading to no change. For any other value,
32465 conversion will be performed.
32466 See the setparams filter for a list of possible values.
32468 apply_filmgrain
32470 Apply film grain (e.g. AV1 or H.274) if present in source frames,
32471 and strip it from the output. Enabled by default.
32472 apply_dolbyvision
32474 Apply Dolby Vision RPU metadata if present in source frames, and
32475 strip it from the output. Enabled by default. Note that Dolby
32476 Vision will always output BT.2020+PQ, overriding the usual input
32477 frame metadata. These will also be picked as the values of "auto"
32478 for the respective frame output options.
32479 Scaling
32481 The options in this section control how libplacebo performs upscaling
32483 and (if necessary) downscaling. Note that libplacebo will always
32484 internally operate on 4:4:4 content, so any sub-sampled chroma formats
32485 such as "yuv420p" will necessarily be upsampled and downsampled as part
32486 of the rendering process. That means scaling might be in effect even if
32487 the source and destination resolution are the same.
32488 upscaler
32490 downscaler
32491 Configure the filter kernel used for upscaling and downscaling. The
32492 respective defaults are "spline36" and "mitchell". For a full list
32493 of possible values, pass "help" to these options. The most
32494 important values are:
32495 none
32497 Forces the use of built-in GPU texture sampling (typically
32498 bilinear). Extremely fast but poor quality, especially when
32499 downscaling.
32500 bilinear
32502 Bilinear interpolation. Can generally be done for free on GPUs,
32503 except when doing so would lead to aliasing. Fast and low
32504 quality.
32505 nearest
32507 Nearest-neighbour interpolation. Sharp but highly aliasing.
32508 oversample
32510 Algorithm that looks visually similar to nearest-neighbour
32511 interpolation but tries to preserve pixel aspect ratio. Good
32512 for pixel art, since it results in minimal distortion of the
32513 artistic appearance.
32514 lanczos
32516 Standard sinc-sinc interpolation kernel.
32517 spline36
32519 Cubic spline approximation of lanczos. No difference in
32520 performance, but has very slightly less ringing.
32521 ewa_lanczos
32523 Elliptically weighted average version of lanczos, based on a
32524 jinc-sinc kernel. This is also popularly referred to as just
32525 "Jinc scaling". Slow but very high quality.
32526 gaussian
32528 Gaussian kernel. Has certain ideal mathematical properties, but
32529 subjectively very blurry.
32530 mitchell
32532 Cubic BC spline with parameters recommended by Mitchell and
32533 Netravali. Very little ringing.
32534 lut_entries
32536 Configures the size of scaler LUTs, ranging from 1 to 256. The
32537 default of 0 will pick libplacebo's internal default, typically 64.
32538 antiringing
32540 Enables anti-ringing (for non-EWA filters). The value (between 0.0
32541 and 1.0) configures the strength of the anti-ringing algorithm. May
32542 increase aliasing if set too high. Disabled by default.
32543 sigmoid
32545 Enable sigmoidal compression during upscaling. Reduces ringing
32546 slightly. Enabled by default.
32547 Debanding
32549 Libplacebo comes with a built-in debanding filter that is good at
32551 counteracting many common sources of banding and blocking. Turning this
32552 on is highly recommended whenever quality is desired.
32553 deband
32555 Enable (fast) debanding algorithm. Disabled by default.
32556 deband_iterations
32558 Number of deband iterations of the debanding algorithm. Each
32559 iteration is performed with progressively increased radius (and
32560 diminished threshold). Recommended values are in the range 1 to 4.
32561 Defaults to 1.
32562 deband_threshold
32564 Debanding filter strength. Higher numbers lead to more aggressive
32565 debanding. Defaults to 4.0.
32566 deband_radius
32568 Debanding filter radius. A higher radius is better for slow
32569 gradients, while a lower radius is better for steep gradients.
32570 Defaults to 16.0.
32571 deband_grain
32573 Amount of extra output grain to add. Helps hide imperfections.
32574 Defaults to 6.0.
32575 Color adjustment
32577 A collection of subjective color controls. Not very rigorous, so the
32579 exact effect will vary somewhat depending on the input primaries and
32580 colorspace.
32581 brightness
32583 Brightness boost, between -1.0 and 1.0. Defaults to 0.0.
32584 contrast
32586 Contrast gain, between 0.0 and 16.0. Defaults to 1.0.
32587 saturation
32589 Saturation gain, between 0.0 and 16.0. Defaults to 1.0.
32590 hue Hue shift in radians, between -3.14 and 3.14. Defaults to 0.0. This
32592 will rotate the UV subvector, defaulting to BT.709 coefficients for
32593 RGB inputs.
32594 gamma
32596 Gamma adjustment, between 0.0 and 16.0. Defaults to 1.0.
32597 cones
32599 Cone model to use for color blindness simulation. Accepts any
32600 combination of "l", "m" and "s". Here are some examples:
32601 m Deuteranomaly / deuteranopia (affecting 3%-4% of the
32603 population)
32604 l Protanomaly / protanopia (affecting 1%-2% of the population)
32606 l+m Monochromacy (very rare)
32608 l+m+s
32610 Achromatopsy (complete loss of daytime vision, extremely rare)
32611 cone-strength
32613 Gain factor for the cones specified by "cones", between 0.0 and
32614 10.0. A value of 1.0 results in no change to color vision. A value
32615 of 0.0 (the default) simulates complete loss of those cones. Values
32616 above 1.0 result in exaggerating the differences between cones,
32617 which may help compensate for reduced color vision.
32618 Peak detection
32620 To help deal with sources that only have static HDR10 metadata (or no
32622 tagging whatsoever), libplacebo uses its own internal frame analysis
32623 compute shader to analyze source frames and adapt the tone mapping
32624 function in realtime. If this is too slow, or if exactly reproducible
32625 frame-perfect results are needed, it's recommended to turn this feature
32626 off.
32627 peak_detect
32629 Enable HDR peak detection. Ignores static MaxCLL/MaxFALL values in
32630 favor of dynamic detection from the input. Note that the detected
32631 values do not get written back to the output frames, they merely
32632 guide the internal tone mapping process. Enabled by default.
32633 smoothing_period
32635 Peak detection smoothing period, between 0.0 and 1000.0. Higher
32636 values result in peak detection becoming less responsive to changes
32637 in the input. Defaults to 100.0.
32638 minimum_peak
32640 Lower bound on the detected peak (relative to SDR white), between
32641 0.0 and 100.0. Defaults to 1.0.
32642 scene_threshold_low
32644 scene_threshold_high
32645 Lower and upper thresholds for scene change detection. Expressed in
32646 a logarithmic scale between 0.0 and 100.0. Default to 5.5 and 10.0,
32647 respectively. Setting either to a negative value disables this
32648 functionality.
32649 overshoot
32651 Peak smoothing overshoot margin, between 0.0 and 1.0. Provides a
32652 safety margin to prevent clipping as a result of peak smoothing.
32653 Defaults to 0.05, corresponding to a margin of 5%.
32654 Tone mapping
32656 The options in this section control how libplacebo performs tone-
32658 mapping and gamut-mapping when dealing with mismatches between wide-
32659 gamut or HDR content. In general, libplacebo relies on accurate source
32660 tagging and mastering display gamut information to produce the best
32661 results.
32662 intent
32664 Rendering intent to use when adapting between different primary
32665 color gamuts (after tone-mapping).
32666 perceptual
32668 Perceptual gamut mapping. Currently equivalent to relative
32669 colorimetric.
32670 relative
32672 Relative colorimetric. This is the default.
32673 absolute
32675 Absolute colorimetric.
32676 saturation
32678 Saturation mapping. Forcibly stretches the source gamut to the
32679 target gamut.
32680 gamut_mode
32682 How to handle out-of-gamut colors that can occur as a result of
32683 colorimetric gamut mapping.
32684 clip
32686 Do nothing, simply clip out-of-range colors to the RGB volume.
32687 This is the default.
32688 warn
32690 Highlight out-of-gamut pixels (by coloring them pink).
32691 darken
32693 Linearly reduces content brightness to preserves saturated
32694 details, followed by clipping the remaining out-of-gamut
32695 colors. As the name implies, this makes everything darker, but
32696 provides a good balance between preserving details and colors.
32697 desaturate
32699 Hard-desaturates out-of-gamut colors towards white, while
32700 preserving the luminance. Has a tendency to shift colors.
32701 tonemapping
32703 Tone-mapping algorithm to use. Available values are:
32704 auto
32706 Automatic selection based on internal heuristics. This is the
32707 default.
32708 clip
32710 Performs no tone-mapping, just clips out-of-range colors.
32711 Retains perfect color accuracy for in-range colors but
32712 completely destroys out-of-range information. Does not perform
32713 any black point adaptation. Not configurable.
32714 st2094-40
32716 EETF from SMPTE ST 2094-40 Annex B, which applies the Bezier
32717 curves from HDR10+ dynamic metadata based on Bezier curves to
32718 perform tone-mapping. The OOTF used is adjusted based on the
32719 ratio between the targeted and actual display peak luminances.
32720 st2094-10
32722 EETF from SMPTE ST 2094-10 Annex B.2, which takes into account
32723 the input signal average luminance in addition to the
32724 maximum/minimum. The configurable contrast parameter influences
32725 the slope of the linear output segment, defaulting to 1.0 for
32726 no increase/decrease in contrast. Note that this does not
32727 currently include the subjective gain/offset/gamma controls
32728 defined in Annex B.3.
32729 bt.2390
32731 EETF from the ITU-R Report BT.2390, a hermite spline roll-off
32732 with linear segment. The knee point offset is configurable.
32733 Note that this parameter defaults to 1.0, rather than the value
32734 of 0.5 from the ITU-R spec.
32735 bt.2446a
32737 EETF from ITU-R Report BT.2446, method A. Designed for well-
32738 mastered HDR sources. Can be used for both forward and inverse
32739 tone mapping. Not configurable.
32740 spline
32742 Simple spline consisting of two polynomials, joined by a single
32743 pivot point. The parameter gives the pivot point (in PQ
32744 space), defaulting to 0.30. Can be used for both forward and
32745 inverse tone mapping.
32746 reinhard
32748 Simple non-linear, global tone mapping algorithm. The parameter
32749 specifies the local contrast coefficient at the display peak.
32750 Essentially, a parameter of 0.5 implies that the reference
32751 white will be about half as bright as when clipping. Defaults
32752 to 0.5, which results in the simplest formulation of this
32753 function.
32754 mobius
32756 Generalization of the reinhard tone mapping algorithm to
32757 support an additional linear slope near black. The tone mapping
32758 parameter indicates the trade-off between the linear section
32759 and the non-linear section. Essentially, for a given parameter
32760 x, every color value below x will be mapped linearly, while
32761 higher values get non-linearly tone-mapped. Values near 1.0
32762 make this curve behave like "clip", while values near 0.0 make
32763 this curve behave like "reinhard". The default value is 0.3,
32764 which provides a good balance between colorimetric accuracy and
32765 preserving out-of-gamut details.
32766 hable
32768 Piece-wise, filmic tone-mapping algorithm developed by John
32769 Hable for use in Uncharted 2, inspired by a similar tone-
32770 mapping algorithm used by Kodak. Popularized by its use in
32771 video games with HDR rendering. Preserves both dark and bright
32772 details very well, but comes with the drawback of changing the
32773 average brightness quite significantly. This is sort of similar
32774 to "reinhard" with parameter 0.24.
32775 gamma
32777 Fits a gamma (power) function to transfer between the source
32778 and target color spaces, effectively resulting in a perceptual
32779 hard-knee joining two roughly linear sections. This preserves
32780 details at all scales fairly accurately, but can result in an
32781 image with a muted or dull appearance. The parameter is used as
32782 the cutoff point, defaulting to 0.5.
32783 linear
32785 Linearly stretches the input range to the output range, in PQ
32786 space. This will preserve all details accurately, but results
32787 in a significantly different average brightness. Can be used
32788 for inverse tone-mapping in addition to regular tone-mapping.
32789 The parameter can be used as an additional linear gain
32790 coefficient (defaulting to 1.0).
32791 tonemapping_param
32793 For tunable tone mapping functions, this parameter can be used to
32794 fine-tune the curve behavior. Refer to the documentation of
32795 "tonemapping". The default value of 0.0 is replaced by the curve's
32796 preferred default setting.
32797 tonemapping_mode
32799 This option determines how the tone mapping function specified by
32800 "tonemapping" is applied to the colors in a scene. Possible values
32801 are:
32802 auto
32804 Automatic selection based on internal heuristics. This is the
32805 default.
32806 rgb Apply the function per-channel in the RGB colorspace. Per-
32808 channel tone-mapping in RGB. Guarantees no clipping and heavily
32809 desaturates the output, but distorts the colors quite
32810 significantly. Very similar to the "Hollywood" look and feel.
32811 max Tone-mapping is performed on the brightest component found in
32813 the signal. Good at preserving details in highlights, but has a
32814 tendency to crush blacks.
32815 hybrid
32817 Tone-map per-channel for highlights and linearly (luma-based)
32818 for midtones/shadows, based on a fixed gamma 2.4 coefficient
32819 curve.
32820 luma
32822 Tone-map linearly on the luma component (CIE Y), and adjust
32823 (desaturate) the chromaticities to compensate using a simple
32824 constant factor. This is essentially the mode used in ITU-R
32825 BT.2446 method A.
32826 inverse_tonemapping
32828 If enabled, this filter will also attempt stretching SDR signals to
32829 fill HDR output color volumes. Disabled by default.
32830 tonemapping_crosstalk
32832 Extra tone-mapping crosstalk factor, between 0.0 and 0.3. This can
32833 help reduce issues tone-mapping certain bright spectral colors.
32834 Defaults to 0.04.
32835 tonemapping_lut_size
32837 Size of the tone-mapping LUT, between 2 and 1024. Defaults to 256.
32838 Note that this figure is squared when combined with "peak_detect".
32839 Dithering
32841 By default, libplacebo will dither whenever necessary, which includes
32843 rendering to any integer format below 16-bit precision. It's
32844 recommended to always leave this on, since not doing so may result in
32845 visible banding in the output, even if the "debanding" filter is
32846 enabled. If maximum performance is needed, use "ordered_fixed" instead
32847 of disabling dithering.
32848 dithering
32850 Dithering method to use. Accepts the following values:
32851 none
32853 Disables dithering completely. May result in visible banding.
32854 blue
32856 Dither with pseudo-blue noise. This is the default.
32857 ordered
32859 Tunable ordered dither pattern.
32860 ordered_fixed
32862 Faster ordered dither with a fixed size of 6. Texture-less.
32863 white
32865 Dither with white noise. Texture-less.
32866 dither_lut_size
32868 Dither LUT size, as log base2 between 1 and 8. Defaults to 6,
32869 corresponding to a LUT size of "64x64".
32870 dither_temporal
32872 Enables temporal dithering. Disabled by default.
32873 Custom shaders
32875 libplacebo supports a number of custom shaders based on the mpv .hook
32877 GLSL syntax. A collection of such shaders can be found here:
32878 <https://github.com/mpv-player/mpv/wiki/User-Scripts#user-shaders>
32879 A full description of the mpv shader format is beyond the scope of this
32881 section, but a summary can be found here:
32882 <https://mpv.io/manual/master/#options-glsl-shader>
32883 custom_shader_path
32885 Specifies a path to a custom shader file to load at runtime.
32886 custom_shader_bin
32888 Specifies a complete custom shader as a raw string.
32889 Debugging / performance
32891 All of the options in this section default off. They may be of
32893 assistance when attempting to squeeze the maximum performance at the
32894 cost of quality.
32895 skip_aa
32897 Disable anti-aliasing when downscaling.
32898 polar_cutoff
32900 Truncate polar (EWA) scaler kernels below this absolute magnitude,
32901 between 0.0 and 1.0.
32902 disable_linear
32904 Disable linear light scaling.
32905 disable_builtin
32907 Disable built-in GPU sampling (forces LUT).
32908 disable_fbos
32910 Forcibly disable FBOs, resulting in loss of almost all
32911 functionality, but offering the maximum possible speed.
32912 Commands
32914 This filter supports almost all of the above options as commands.
32916 Examples
32918 • Complete example for how to initialize the Vulkan device, upload
32920 frames to the GPU, perform filter conversion to yuv420p, and
32921 download frames back to the CPU for output. Note that in specific
32922 cases you can get around the need to perform format conversion by
32923 specifying the correct "format" filter option corresponding to the
32924 input frames.
32925 ffmpeg -i $INPUT -init_hw_device vulkan -vf hwupload,libplacebo=format=yuv420p,hwdownload,format=yuv420p $OUTPUT
32927 • Tone-map input to standard gamut BT.709 output:
32929 libplacebo=colorspace=bt709:color_primaries=bt709:color_trc=bt709:range=tv
32931 • Rescale input to fit into standard 1080p, with high quality
32933 scaling:
32934 libplacebo=w=1920:h=1080:force_original_aspect_ratio=decrease:normalize_sar=true:upscaler=ewa_lanczos:downscaler=ewa_lanczos
32936 • Convert input to standard sRGB JPEG:
32938 libplacebo=format=yuv420p:colorspace=bt470bg:color_primaries=bt709:color_trc=iec61966-2-1:range=pc
32940 • Use higher quality debanding settings:
32942 libplacebo=deband=true:deband_iterations=3:deband_radius=8:deband_threshold=6
32944 • Run this filter on the CPU, on systems with Mesa installed (and
32946 with the most expensive options disabled):
32947 ffmpeg ... -init_hw_device vulkan:llvmpipe ... -vf libplacebo=upscaler=none:downscaler=none:peak_detect=false
32949 • Suppress CPU-based AV1/H.274 film grain application in the decoder,
32951 in favor of doing it with this filter. Note that this is only a
32952 gain if the frames are either already on the GPU, or if you're
32953 using libplacebo for other purposes, since otherwise the VRAM
32954 roundtrip will more than offset any expected speedup.
32955 ffmpeg -export_side_data +film_grain ... -vf libplacebo=apply_filmgrain=true
32957 libvmaf
32959 Calulate the VMAF (Video Multi-Method Assessment Fusion) score for a
32960 reference/distorted pair of input videos.
32961 The first input is the distorted video, and the second input is the
32963 reference video.
32964 The obtained VMAF score is printed through the logging system.
32966 It requires Netflix's vmaf library (libvmaf) as a pre-requisite. After
32968 installing the library it can be enabled using: "./configure
32969 --enable-libvmaf".
32970 The filter has following options:
32972 model
32974 A `|` delimited list of vmaf models. Each model can be configured
32975 with a number of parameters. Default value: "version=vmaf_v0.6.1"
32976 model_path
32978 Deprecated, use model='path=...'.
32979 enable_transform
32981 Deprecated, use model='enable_transform=true'.
32982 phone_model
32984 Deprecated, use model='enable_transform=true'.
32985 enable_conf_interval
32987 Deprecated, use model='enable_conf_interval=true'.
32988 feature
32990 A `|` delimited list of features. Each feature can be configured
32991 with a number of parameters.
32992 psnr
32994 Deprecated, use feature='name=psnr'.
32995 ssim
32997 Deprecated, use feature='name=ssim'.
32998 ms_ssim
33000 Deprecated, use feature='name=ms_ssim'.
33001 log_path
33003 Set the file path to be used to store log files.
33004 log_fmt
33006 Set the format of the log file (xml, json, csv, or sub).
33007 n_threads
33009 Set number of threads to be used when initializing libvmaf.
33010 Default value: 0, no threads.
33011 n_subsample
33013 Set frame subsampling interval to be used.
33014 This filter also supports the framesync options.
33016 Examples
33018 • In the examples below, a distorted video distorted.mpg is compared
33020 with a reference file reference.mpg.
33021 • Basic usage:
33023 ffmpeg -i distorted.mpg -i reference.mpg -lavfi libvmaf=log_path=output.xml -f null -
33025 • Example with multiple models:
33027 ffmpeg -i distorted.mpg -i reference.mpg -lavfi libvmaf='model=version=vmaf_v0.6.1\\:name=vmaf|version=vmaf_v0.6.1neg\\:name=vmaf_neg' -f null -
33029 • Example with multiple addtional features:
33031 ffmpeg -i distorted.mpg -i reference.mpg -lavfi libvmaf='feature=name=psnr|name=ciede' -f null -
33033 • Example with options and different containers:
33035 ffmpeg -i distorted.mpg -i reference.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]libvmaf=log_fmt=json:log_path=output.json" -f null -
33037 limitdiff
33039 Apply limited difference filter using second and optionally third video
33040 stream.
33041 The filter accepts the following options:
33043 threshold
33045 Set the threshold to use when allowing certain differences between
33046 video streams. Any absolute difference value lower or exact than
33047 this threshold will pick pixel components from first video stream.
33048 elasticity
33050 Set the elasticity of soft thresholding when processing video
33051 streams. This value multiplied with first one sets second
33052 threshold. Any absolute difference value greater or exact than
33053 second threshold will pick pixel components from second video
33054 stream. For values between those two threshold linear interpolation
33055 between first and second video stream will be used.
33056 reference
33058 Enable the reference (third) video stream processing. By default is
33059 disabled. If set, this video stream will be used for calculating
33060 absolute difference with first video stream.
33061 planes
33063 Specify which planes will be processed. Defaults to all available.
33064 Commands
33066 This filter supports the all above options as commands except option
33068 reference.
33069 limiter
33071 Limits the pixel components values to the specified range [min, max].
33072 The filter accepts the following options:
33074 min Lower bound. Defaults to the lowest allowed value for the input.
33076 max Upper bound. Defaults to the highest allowed value for the input.
33078 planes
33080 Specify which planes will be processed. Defaults to all available.
33081 Commands
33083 This filter supports the all above options as commands.
33085 loop
33087 Loop video frames.
33088 The filter accepts the following options:
33090 loop
33092 Set the number of loops. Setting this value to -1 will result in
33093 infinite loops. Default is 0.
33094 size
33096 Set maximal size in number of frames. Default is 0.
33097 start
33099 Set first frame of loop. Default is 0.
33100 Examples
33102 • Loop single first frame infinitely:
33104 loop=loop=-1:size=1:start=0
33106 • Loop single first frame 10 times:
33108 loop=loop=10:size=1:start=0
33110 • Loop 10 first frames 5 times:
33112 loop=loop=5:size=10:start=0
33114 lut1d
33116 Apply a 1D LUT to an input video.
33117 The filter accepts the following options:
33119 file
33121 Set the 1D LUT file name.
33122 Currently supported formats:
33124 cube
33126 Iridas
33127 csp cineSpace
33129 interp
33131 Select interpolation mode.
33132 Available values are:
33134 nearest
33136 Use values from the nearest defined point.
33137 linear
33139 Interpolate values using the linear interpolation.
33140 cosine
33142 Interpolate values using the cosine interpolation.
33143 cubic
33145 Interpolate values using the cubic interpolation.
33146 spline
33148 Interpolate values using the spline interpolation.
33149 Commands
33151 This filter supports the all above options as commands.
33153 lut3d
33155 Apply a 3D LUT to an input video.
33156 The filter accepts the following options:
33158 file
33160 Set the 3D LUT file name.
33161 Currently supported formats:
33163 3dl AfterEffects
33165 cube
33167 Iridas
33168 dat DaVinci
33170 m3d Pandora
33172 csp cineSpace
33174 interp
33176 Select interpolation mode.
33177 Available values are:
33179 nearest
33181 Use values from the nearest defined point.
33182 trilinear
33184 Interpolate values using the 8 points defining a cube.
33185 tetrahedral
33187 Interpolate values using a tetrahedron.
33188 pyramid
33190 Interpolate values using a pyramid.
33191 prism
33193 Interpolate values using a prism.
33194 Commands
33196 This filter supports the "interp" option as commands.
33198 lumakey
33200 Turn certain luma values into transparency.
33201 The filter accepts the following options:
33203 threshold
33205 Set the luma which will be used as base for transparency. Default
33206 value is 0.
33207 tolerance
33209 Set the range of luma values to be keyed out. Default value is
33210 0.01.
33211 softness
33213 Set the range of softness. Default value is 0. Use this to control
33214 gradual transition from zero to full transparency.
33215 Commands
33217 This filter supports same commands as options. The command accepts the
33219 same syntax of the corresponding option.
33220 If the specified expression is not valid, it is kept at its current
33222 value.
33223 lut, lutrgb, lutyuv
33225 Compute a look-up table for binding each pixel component input value to
33226 an output value, and apply it to the input video.
33227 lutyuv applies a lookup table to a YUV input video, lutrgb to an RGB
33229 input video.
33230 These filters accept the following parameters:
33232 c0 set first pixel component expression
33234 c1 set second pixel component expression
33236 c2 set third pixel component expression
33238 c3 set fourth pixel component expression, corresponds to the alpha
33240 component
33241 r set red component expression
33243 g set green component expression
33245 b set blue component expression
33247 a alpha component expression
33249 y set Y/luminance component expression
33251 u set U/Cb component expression
33253 v set V/Cr component expression
33255 Each of them specifies the expression to use for computing the lookup
33257 table for the corresponding pixel component values.
33258 The exact component associated to each of the c* options depends on the
33260 format in input.
33261 The lut filter requires either YUV or RGB pixel formats in input,
33263 lutrgb requires RGB pixel formats in input, and lutyuv requires YUV.
33264 The expressions can contain the following constants and functions:
33266 w
33268 h The input width and height.
33269 val The input value for the pixel component.
33271 clipval
33273 The input value, clipped to the minval-maxval range.
33274 maxval
33276 The maximum value for the pixel component.
33277 minval
33279 The minimum value for the pixel component.
33280 negval
33282 The negated value for the pixel component value, clipped to the
33283 minval-maxval range; it corresponds to the expression
33284 "maxval-clipval+minval".
33285 clip(val)
33287 The computed value in val, clipped to the minval-maxval range.
33288 gammaval(gamma)
33290 The computed gamma correction value of the pixel component value,
33291 clipped to the minval-maxval range. It corresponds to the
33292 expression
33293 "pow((clipval-minval)/(maxval-minval)\,gamma)*(maxval-minval)+minval"
33294 All expressions default to "clipval".
33296 Commands
33298 This filter supports same commands as options.
33300 Examples
33302 • Negate input video:
33304 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
33306 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
33307 The above is the same as:
33309 lutrgb="r=negval:g=negval:b=negval"
33311 lutyuv="y=negval:u=negval:v=negval"
33312 • Negate luminance:
33314 lutyuv=y=negval
33316 • Remove chroma components, turning the video into a graytone image:
33318 lutyuv="u=128:v=128"
33320 • Apply a luma burning effect:
33322 lutyuv="y=2*val"
33324 • Remove green and blue components:
33326 lutrgb="g=0:b=0"
33328 • Set a constant alpha channel value on input:
33330 format=rgba,lutrgb=a="maxval-minval/2"
33332 • Correct luminance gamma by a factor of 0.5:
33334 lutyuv=y=gammaval(0.5)
33336 • Discard least significant bits of luma:
33338 lutyuv=y='bitand(val, 128+64+32)'
33340 • Technicolor like effect:
33342 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
33344 lut2, tlut2
33346 The "lut2" filter takes two input streams and outputs one stream.
33347 The "tlut2" (time lut2) filter takes two consecutive frames from one
33349 single stream.
33350 This filter accepts the following parameters:
33352 c0 set first pixel component expression
33354 c1 set second pixel component expression
33356 c2 set third pixel component expression
33358 c3 set fourth pixel component expression, corresponds to the alpha
33360 component
33361 d set output bit depth, only available for "lut2" filter. By default
33363 is 0, which means bit depth is automatically picked from first
33364 input format.
33365 The "lut2" filter also supports the framesync options.
33367 Each of them specifies the expression to use for computing the lookup
33369 table for the corresponding pixel component values.
33370 The exact component associated to each of the c* options depends on the
33372 format in inputs.
33373 The expressions can contain the following constants:
33375 w
33377 h The input width and height.
33378 x The first input value for the pixel component.
33380 y The second input value for the pixel component.
33382 bdx The first input video bit depth.
33384 bdy The second input video bit depth.
33386 All expressions default to "x".
33388 Commands
33390 This filter supports the all above options as commands except option
33392 "d".
33393 Examples
33395 • Highlight differences between two RGB video streams:
33397 lut2='ifnot(x-y,0,pow(2,bdx)-1):ifnot(x-y,0,pow(2,bdx)-1):ifnot(x-y,0,pow(2,bdx)-1)'
33399 • Highlight differences between two YUV video streams:
33401 lut2='ifnot(x-y,0,pow(2,bdx)-1):ifnot(x-y,pow(2,bdx-1),pow(2,bdx)-1):ifnot(x-y,pow(2,bdx-1),pow(2,bdx)-1)'
33403 • Show max difference between two video streams:
33405 lut2='if(lt(x,y),0,if(gt(x,y),pow(2,bdx)-1,pow(2,bdx-1))):if(lt(x,y),0,if(gt(x,y),pow(2,bdx)-1,pow(2,bdx-1))):if(lt(x,y),0,if(gt(x,y),pow(2,bdx)-1,pow(2,bdx-1)))'
33407 maskedclamp
33409 Clamp the first input stream with the second input and third input
33410 stream.
33411 Returns the value of first stream to be between second input stream -
33413 "undershoot" and third input stream + "overshoot".
33414 This filter accepts the following options:
33416 undershoot
33418 Default value is 0.
33419 overshoot
33421 Default value is 0.
33422 planes
33424 Set which planes will be processed as bitmap, unprocessed planes
33425 will be copied from first stream. By default value 0xf, all planes
33426 will be processed.
33427 Commands
33429 This filter supports the all above options as commands.
33431 maskedmax
33433 Merge the second and third input stream into output stream using
33434 absolute differences between second input stream and first input stream
33435 and absolute difference between third input stream and first input
33436 stream. The picked value will be from second input stream if second
33437 absolute difference is greater than first one or from third input
33438 stream otherwise.
33439 This filter accepts the following options:
33441 planes
33443 Set which planes will be processed as bitmap, unprocessed planes
33444 will be copied from first stream. By default value 0xf, all planes
33445 will be processed.
33446 Commands
33448 This filter supports the all above options as commands.
33450 maskedmerge
33452 Merge the first input stream with the second input stream using per
33453 pixel weights in the third input stream.
33454 A value of 0 in the third stream pixel component means that pixel
33456 component from first stream is returned unchanged, while maximum value
33457 (eg. 255 for 8-bit videos) means that pixel component from second
33458 stream is returned unchanged. Intermediate values define the amount of
33459 merging between both input stream's pixel components.
33460 This filter accepts the following options:
33462 planes
33464 Set which planes will be processed as bitmap, unprocessed planes
33465 will be copied from first stream. By default value 0xf, all planes
33466 will be processed.
33467 Commands
33469 This filter supports the all above options as commands.
33471 maskedmin
33473 Merge the second and third input stream into output stream using
33474 absolute differences between second input stream and first input stream
33475 and absolute difference between third input stream and first input
33476 stream. The picked value will be from second input stream if second
33477 absolute difference is less than first one or from third input stream
33478 otherwise.
33479 This filter accepts the following options:
33481 planes
33483 Set which planes will be processed as bitmap, unprocessed planes
33484 will be copied from first stream. By default value 0xf, all planes
33485 will be processed.
33486 Commands
33488 This filter supports the all above options as commands.
33490 maskedthreshold
33492 Pick pixels comparing absolute difference of two video streams with
33493 fixed threshold.
33494 If absolute difference between pixel component of first and second
33496 video stream is equal or lower than user supplied threshold than pixel
33497 component from first video stream is picked, otherwise pixel component
33498 from second video stream is picked.
33499 This filter accepts the following options:
33501 threshold
33503 Set threshold used when picking pixels from absolute difference
33504 from two input video streams.
33505 planes
33507 Set which planes will be processed as bitmap, unprocessed planes
33508 will be copied from second stream. By default value 0xf, all
33509 planes will be processed.
33510 mode
33512 Set mode of filter operation. Can be "abs" or "diff". Default is
33513 "abs".
33514 Commands
33516 This filter supports the all above options as commands.
33518 maskfun
33520 Create mask from input video.
33521 For example it is useful to create motion masks after "tblend" filter.
33523 This filter accepts the following options:
33525 low Set low threshold. Any pixel component lower or exact than this
33527 value will be set to 0.
33528 high
33530 Set high threshold. Any pixel component higher than this value will
33531 be set to max value allowed for current pixel format.
33532 planes
33534 Set planes to filter, by default all available planes are filtered.
33535 fill
33537 Fill all frame pixels with this value.
33538 sum Set max average pixel value for frame. If sum of all pixel
33540 components is higher that this average, output frame will be
33541 completely filled with value set by fill option. Typically useful
33542 for scene changes when used in combination with "tblend" filter.
33543 Commands
33545 This filter supports the all above options as commands.
33547 mcdeint
33549 Apply motion-compensation deinterlacing.
33550 It needs one field per frame as input and must thus be used together
33552 with yadif=1/3 or equivalent.
33553 This filter is only available in ffmpeg version 4.4 or earlier.
33555 This filter accepts the following options:
33557 mode
33559 Set the deinterlacing mode.
33560 It accepts one of the following values:
33562 fast
33564 medium
33565 slow
33566 use iterative motion estimation
33567 extra_slow
33569 like slow, but use multiple reference frames.
33570 Default value is fast.
33572 parity
33574 Set the picture field parity assumed for the input video. It must
33575 be one of the following values:
33576 0, tff
33578 assume top field first
33579 1, bff
33581 assume bottom field first
33582 Default value is bff.
33584 qp Set per-block quantization parameter (QP) used by the internal
33586 encoder.
33587 Higher values should result in a smoother motion vector field but
33589 less optimal individual vectors. Default value is 1.
33590 median
33592 Pick median pixel from certain rectangle defined by radius.
33593 This filter accepts the following options:
33595 radius
33597 Set horizontal radius size. Default value is 1. Allowed range is
33598 integer from 1 to 127.
33599 planes
33601 Set which planes to process. Default is 15, which is all available
33602 planes.
33603 radiusV
33605 Set vertical radius size. Default value is 0. Allowed range is
33606 integer from 0 to 127. If it is 0, value will be picked from
33607 horizontal "radius" option.
33608 percentile
33610 Set median percentile. Default value is 0.5. Default value of 0.5
33611 will pick always median values, while 0 will pick minimum values,
33612 and 1 maximum values.
33613 Commands
33615 This filter supports same commands as options. The command accepts the
33617 same syntax of the corresponding option.
33618 If the specified expression is not valid, it is kept at its current
33620 value.
33621 mergeplanes
33623 Merge color channel components from several video streams.
33624 The filter accepts up to 4 input streams, and merge selected input
33626 planes to the output video.
33627 This filter accepts the following options:
33629 mapping
33631 Set input to output plane mapping. Default is 0.
33632 The mappings is specified as a bitmap. It should be specified as a
33634 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
33635 mapping for the first plane of the output stream. 'A' sets the
33636 number of the input stream to use (from 0 to 3), and 'a' the plane
33637 number of the corresponding input to use (from 0 to 3). The rest of
33638 the mappings is similar, 'Bb' describes the mapping for the output
33639 stream second plane, 'Cc' describes the mapping for the output
33640 stream third plane and 'Dd' describes the mapping for the output
33641 stream fourth plane.
33642 format
33644 Set output pixel format. Default is "yuva444p".
33645 map0s
33647 map1s
33648 map2s
33649 map3s
33650 Set input to output stream mapping for output Nth plane. Default is
33651 0.
33652 map0p
33654 map1p
33655 map2p
33656 map3p
33657 Set input to output plane mapping for output Nth plane. Default is
33658 0.
33659 Examples
33661 • Merge three gray video streams of same width and height into single
33663 video stream:
33664 [a0][a1][a2]mergeplanes=0x001020:yuv444p
33666 • Merge 1st yuv444p stream and 2nd gray video stream into yuva444p
33668 video stream:
33669 [a0][a1]mergeplanes=0x00010210:yuva444p
33671 • Swap Y and A plane in yuva444p stream:
33673 format=yuva444p,mergeplanes=0x03010200:yuva444p
33675 • Swap U and V plane in yuv420p stream:
33677 format=yuv420p,mergeplanes=0x000201:yuv420p
33679 • Cast a rgb24 clip to yuv444p:
33681 format=rgb24,mergeplanes=0x000102:yuv444p
33683 mestimate
33685 Estimate and export motion vectors using block matching algorithms.
33686 Motion vectors are stored in frame side data to be used by other
33687 filters.
33688 This filter accepts the following options:
33690 method
33692 Specify the motion estimation method. Accepts one of the following
33693 values:
33694 esa Exhaustive search algorithm.
33696 tss Three step search algorithm.
33698 tdls
33700 Two dimensional logarithmic search algorithm.
33701 ntss
33703 New three step search algorithm.
33704 fss Four step search algorithm.
33706 ds Diamond search algorithm.
33708 hexbs
33710 Hexagon-based search algorithm.
33711 epzs
33713 Enhanced predictive zonal search algorithm.
33714 umh Uneven multi-hexagon search algorithm.
33716 Default value is esa.
33718 mb_size
33720 Macroblock size. Default 16.
33721 search_param
33723 Search parameter. Default 7.
33724 midequalizer
33726 Apply Midway Image Equalization effect using two video streams.
33727 Midway Image Equalization adjusts a pair of images to have the same
33729 histogram, while maintaining their dynamics as much as possible. It's
33730 useful for e.g. matching exposures from a pair of stereo cameras.
33731 This filter has two inputs and one output, which must be of same pixel
33733 format, but may be of different sizes. The output of filter is first
33734 input adjusted with midway histogram of both inputs.
33735 This filter accepts the following option:
33737 planes
33739 Set which planes to process. Default is 15, which is all available
33740 planes.
33741 minterpolate
33743 Convert the video to specified frame rate using motion interpolation.
33744 This filter accepts the following options:
33746 fps Specify the output frame rate. This can be rational e.g.
33748 "60000/1001". Frames are dropped if fps is lower than source fps.
33749 Default 60.
33750 mi_mode
33752 Motion interpolation mode. Following values are accepted:
33753 dup Duplicate previous or next frame for interpolating new ones.
33755 blend
33757 Blend source frames. Interpolated frame is mean of previous and
33758 next frames.
33759 mci Motion compensated interpolation. Following options are
33761 effective when this mode is selected:
33762 mc_mode
33764 Motion compensation mode. Following values are accepted:
33765 obmc
33767 Overlapped block motion compensation.
33768 aobmc
33770 Adaptive overlapped block motion compensation. Window
33771 weighting coefficients are controlled adaptively
33772 according to the reliabilities of the neighboring
33773 motion vectors to reduce oversmoothing.
33774 Default mode is obmc.
33776 me_mode
33778 Motion estimation mode. Following values are accepted:
33779 bidir
33781 Bidirectional motion estimation. Motion vectors are
33782 estimated for each source frame in both forward and
33783 backward directions.
33784 bilat
33786 Bilateral motion estimation. Motion vectors are
33787 estimated directly for interpolated frame.
33788 Default mode is bilat.
33790 me The algorithm to be used for motion estimation. Following
33792 values are accepted:
33793 esa Exhaustive search algorithm.
33795 tss Three step search algorithm.
33797 tdls
33799 Two dimensional logarithmic search algorithm.
33800 ntss
33802 New three step search algorithm.
33803 fss Four step search algorithm.
33805 ds Diamond search algorithm.
33807 hexbs
33809 Hexagon-based search algorithm.
33810 epzs
33812 Enhanced predictive zonal search algorithm.
33813 umh Uneven multi-hexagon search algorithm.
33815 Default algorithm is epzs.
33817 mb_size
33819 Macroblock size. Default 16.
33820 search_param
33822 Motion estimation search parameter. Default 32.
33823 vsbmc
33825 Enable variable-size block motion compensation. Motion
33826 estimation is applied with smaller block sizes at object
33827 boundaries in order to make the them less blur. Default is
33828 0 (disabled).
33829 scd Scene change detection method. Scene change leads motion vectors to
33831 be in random direction. Scene change detection replace interpolated
33832 frames by duplicate ones. May not be needed for other modes.
33833 Following values are accepted:
33834 none
33836 Disable scene change detection.
33837 fdiff
33839 Frame difference. Corresponding pixel values are compared and
33840 if it satisfies scd_threshold scene change is detected.
33841 Default method is fdiff.
33843 scd_threshold
33845 Scene change detection threshold. Default is 10..
33846 mix
33848 Mix several video input streams into one video stream.
33849 A description of the accepted options follows.
33851 inputs
33853 The number of inputs. If unspecified, it defaults to 2.
33854 weights
33856 Specify weight of each input video stream as sequence. Each weight
33857 is separated by space. If number of weights is smaller than number
33858 of frames last specified weight will be used for all remaining
33859 unset weights.
33860 scale
33862 Specify scale, if it is set it will be multiplied with sum of each
33863 weight multiplied with pixel values to give final destination pixel
33864 value. By default scale is auto scaled to sum of weights.
33865 planes
33867 Set which planes to filter. Default is all. Allowed range is from 0
33868 to 15.
33869 duration
33871 Specify how end of stream is determined.
33872 longest
33874 The duration of the longest input. (default)
33875 shortest
33877 The duration of the shortest input.
33878 first
33880 The duration of the first input.
33881 Commands
33883 This filter supports the following commands:
33885 weights
33887 scale
33888 planes
33889 Syntax is same as option with same name.
33890 monochrome
33892 Convert video to gray using custom color filter.
33893 A description of the accepted options follows.
33895 cb Set the chroma blue spot. Allowed range is from -1 to 1. Default
33897 value is 0.
33898 cr Set the chroma red spot. Allowed range is from -1 to 1. Default
33900 value is 0.
33901 size
33903 Set the color filter size. Allowed range is from .1 to 10. Default
33904 value is 1.
33905 high
33907 Set the highlights strength. Allowed range is from 0 to 1. Default
33908 value is 0.
33909 Commands
33911 This filter supports the all above options as commands.
33913 morpho
33915 This filter allows to apply main morphological grayscale transforms,
33916 erode and dilate with arbitrary structures set in second input stream.
33917 Unlike naive implementation and much slower performance in erosion and
33919 dilation filters, when speed is critical "morpho" filter should be used
33920 instead.
33921 A description of accepted options follows,
33923 mode
33925 Set morphological transform to apply, can be:
33926 erode
33928 dilate
33929 open
33930 close
33931 gradient
33932 tophat
33933 blackhat
33934 Default is "erode".
33936 planes
33938 Set planes to filter, by default all planes except alpha are
33939 filtered.
33940 structure
33942 Set which structure video frames will be processed from second
33943 input stream, can be first or all. Default is all.
33944 The "morpho" filter also supports the framesync options.
33946 Commands
33948 This filter supports same commands as options.
33950 mpdecimate
33952 Drop frames that do not differ greatly from the previous frame in order
33953 to reduce frame rate.
33954 The main use of this filter is for very-low-bitrate encoding (e.g.
33956 streaming over dialup modem), but it could in theory be used for fixing
33957 movies that were inverse-telecined incorrectly.
33958 A description of the accepted options follows.
33960 max Set the maximum number of consecutive frames which can be dropped
33962 (if positive), or the minimum interval between dropped frames (if
33963 negative). If the value is 0, the frame is dropped disregarding the
33964 number of previous sequentially dropped frames.
33965 Default value is 0.
33967 hi
33969 lo
33970 frac
33971 Set the dropping threshold values.
33972 Values for hi and lo are for 8x8 pixel blocks and represent actual
33974 pixel value differences, so a threshold of 64 corresponds to 1 unit
33975 of difference for each pixel, or the same spread out differently
33976 over the block.
33977 A frame is a candidate for dropping if no 8x8 blocks differ by more
33979 than a threshold of hi, and if no more than frac blocks (1 meaning
33980 the whole image) differ by more than a threshold of lo.
33981 Default value for hi is 64*12, default value for lo is 64*5, and
33983 default value for frac is 0.33.
33984 msad
33986 Obtain the MSAD (Mean Sum of Absolute Differences) between two input
33987 videos.
33988 This filter takes two input videos.
33990 Both input videos must have the same resolution and pixel format for
33992 this filter to work correctly. Also it assumes that both inputs have
33993 the same number of frames, which are compared one by one.
33994 The obtained per component, average, min and max MSAD is printed
33996 through the logging system.
33997 The filter stores the calculated MSAD of each frame in frame metadata.
33999 This filter also supports the framesync options.
34001 In the below example the input file main.mpg being processed is
34003 compared with the reference file ref.mpg.
34004 ffmpeg -i main.mpg -i ref.mpg -lavfi msad -f null -
34006 multiply
34008 Multiply first video stream pixels values with second video stream
34009 pixels values.
34010 The filter accepts the following options:
34012 scale
34014 Set the scale applied to second video stream. By default is 1.
34015 Allowed range is from 0 to 9.
34016 offset
34018 Set the offset applied to second video stream. By default is 0.5.
34019 Allowed range is from -1 to 1.
34020 planes
34022 Specify planes from input video stream that will be processed. By
34023 default all planes are processed.
34024 Commands
34026 This filter supports same commands as options.
34028 negate
34030 Negate (invert) the input video.
34031 It accepts the following option:
34033 components
34035 Set components to negate.
34036 Available values for components are:
34038 y
34040 u
34041 v
34042 a
34043 r
34044 g
34045 b
34046 negate_alpha
34047 With value 1, it negates the alpha component, if present. Default
34048 value is 0.
34049 Commands
34051 This filter supports same commands as options.
34053 nlmeans
34055 Denoise frames using Non-Local Means algorithm.
34056 Each pixel is adjusted by looking for other pixels with similar
34058 contexts. This context similarity is defined by comparing their
34059 surrounding patches of size pxp. Patches are searched in an area of rxr
34060 around the pixel.
34061 Note that the research area defines centers for patches, which means
34063 some patches will be made of pixels outside that research area.
34064 The filter accepts the following options.
34066 s Set denoising strength. Default is 1.0. Must be in range [1.0,
34068 30.0].
34069 p Set patch size. Default is 7. Must be odd number in range [0, 99].
34071 pc Same as p but for chroma planes.
34073 The default value is 0 and means automatic.
34075 r Set research size. Default is 15. Must be odd number in range [0,
34077 99].
34078 rc Same as r but for chroma planes.
34080 The default value is 0 and means automatic.
34082 nnedi
34084 Deinterlace video using neural network edge directed interpolation.
34085 This filter accepts the following options:
34087 weights
34089 Mandatory option, without binary file filter can not work.
34090 Currently file can be found here:
34091 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
34092 deint
34094 Set which frames to deinterlace, by default it is "all". Can be
34095 "all" or "interlaced".
34096 field
34098 Set mode of operation.
34099 Can be one of the following:
34101 af Use frame flags, both fields.
34103 a Use frame flags, single field.
34105 t Use top field only.
34107 b Use bottom field only.
34109 tf Use both fields, top first.
34111 bf Use both fields, bottom first.
34113 planes
34115 Set which planes to process, by default filter process all frames.
34116 nsize
34118 Set size of local neighborhood around each pixel, used by the
34119 predictor neural network.
34120 Can be one of the following:
34122 s8x6
34124 s16x6
34125 s32x6
34126 s48x6
34127 s8x4
34128 s16x4
34129 s32x4
34130 nns Set the number of neurons in predictor neural network. Can be one
34131 of the following:
34132 n16
34134 n32
34135 n64
34136 n128
34137 n256
34138 qual
34139 Controls the number of different neural network predictions that
34140 are blended together to compute the final output value. Can be
34141 "fast", default or "slow".
34142 etype
34144 Set which set of weights to use in the predictor. Can be one of
34145 the following:
34146 a, abs
34148 weights trained to minimize absolute error
34149 s, mse
34151 weights trained to minimize squared error
34152 pscrn
34154 Controls whether or not the prescreener neural network is used to
34155 decide which pixels should be processed by the predictor neural
34156 network and which can be handled by simple cubic interpolation.
34157 The prescreener is trained to know whether cubic interpolation will
34158 be sufficient for a pixel or whether it should be predicted by the
34159 predictor nn. The computational complexity of the prescreener nn
34160 is much less than that of the predictor nn. Since most pixels can
34161 be handled by cubic interpolation, using the prescreener generally
34162 results in much faster processing. The prescreener is pretty
34163 accurate, so the difference between using it and not using it is
34164 almost always unnoticeable.
34165 Can be one of the following:
34167 none
34169 original
34170 new
34171 new2
34172 new3
34173 Default is "new".
34175 Commands
34177 This filter supports same commands as options, excluding weights
34179 option.
34180 noformat
34182 Force libavfilter not to use any of the specified pixel formats for the
34183 input to the next filter.
34184 It accepts the following parameters:
34186 pix_fmts
34188 A '|'-separated list of pixel format names, such as
34189 pix_fmts=yuv420p|monow|rgb24".
34190 Examples
34192 • Force libavfilter to use a format different from yuv420p for the
34194 input to the vflip filter:
34195 noformat=pix_fmts=yuv420p,vflip
34197 • Convert the input video to any of the formats not contained in the
34199 list:
34200 noformat=yuv420p|yuv444p|yuv410p
34202 noise
34204 Add noise on video input frame.
34205 The filter accepts the following options:
34207 all_seed
34209 c0_seed
34210 c1_seed
34211 c2_seed
34212 c3_seed
34213 Set noise seed for specific pixel component or all pixel components
34214 in case of all_seed. Default value is 123457.
34215 all_strength, alls
34217 c0_strength, c0s
34218 c1_strength, c1s
34219 c2_strength, c2s
34220 c3_strength, c3s
34221 Set noise strength for specific pixel component or all pixel
34222 components in case all_strength. Default value is 0. Allowed range
34223 is [0, 100].
34224 all_flags, allf
34226 c0_flags, c0f
34227 c1_flags, c1f
34228 c2_flags, c2f
34229 c3_flags, c3f
34230 Set pixel component flags or set flags for all components if
34231 all_flags. Available values for component flags are:
34232 a averaged temporal noise (smoother)
34234 p mix random noise with a (semi)regular pattern
34236 t temporal noise (noise pattern changes between frames)
34238 u uniform noise (gaussian otherwise)
34240 Examples
34242 Add temporal and uniform noise to input video:
34244 noise=alls=20:allf=t+u
34246 normalize
34248 Normalize RGB video (aka histogram stretching, contrast stretching).
34249 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
34250 For each channel of each frame, the filter computes the input range and
34252 maps it linearly to the user-specified output range. The output range
34253 defaults to the full dynamic range from pure black to pure white.
34254 Temporal smoothing can be used on the input range to reduce flickering
34256 (rapid changes in brightness) caused when small dark or bright objects
34257 enter or leave the scene. This is similar to the auto-exposure
34258 (automatic gain control) on a video camera, and, like a video camera,
34259 it may cause a period of over- or under-exposure of the video.
34260 The R,G,B channels can be normalized independently, which may cause
34262 some color shifting, or linked together as a single channel, which
34263 prevents color shifting. Linked normalization preserves hue.
34264 Independent normalization does not, so it can be used to remove some
34265 color casts. Independent and linked normalization can be combined in
34266 any ratio.
34267 The normalize filter accepts the following options:
34269 blackpt
34271 whitept
34272 Colors which define the output range. The minimum input value is
34273 mapped to the blackpt. The maximum input value is mapped to the
34274 whitept. The defaults are black and white respectively. Specifying
34275 white for blackpt and black for whitept will give color-inverted,
34276 normalized video. Shades of grey can be used to reduce the dynamic
34277 range (contrast). Specifying saturated colors here can create some
34278 interesting effects.
34279 smoothing
34281 The number of previous frames to use for temporal smoothing. The
34282 input range of each channel is smoothed using a rolling average
34283 over the current frame and the smoothing previous frames. The
34284 default is 0 (no temporal smoothing).
34285 independence
34287 Controls the ratio of independent (color shifting) channel
34288 normalization to linked (color preserving) normalization. 0.0 is
34289 fully linked, 1.0 is fully independent. Defaults to 1.0 (fully
34290 independent).
34291 strength
34293 Overall strength of the filter. 1.0 is full strength. 0.0 is a
34294 rather expensive no-op. Defaults to 1.0 (full strength).
34295 Commands
34297 This filter supports same commands as options, excluding smoothing
34299 option. The command accepts the same syntax of the corresponding
34300 option.
34301 If the specified expression is not valid, it is kept at its current
34303 value.
34304 Examples
34306 Stretch video contrast to use the full dynamic range, with no temporal
34308 smoothing; may flicker depending on the source content:
34309 normalize=blackpt=black:whitept=white:smoothing=0
34311 As above, but with 50 frames of temporal smoothing; flicker should be
34313 reduced, depending on the source content:
34314 normalize=blackpt=black:whitept=white:smoothing=50
34316 As above, but with hue-preserving linked channel normalization:
34318 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
34320 As above, but with half strength:
34322 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
34324 Map the darkest input color to red, the brightest input color to cyan:
34326 normalize=blackpt=red:whitept=cyan
34328 null
34330 Pass the video source unchanged to the output.
34331 ocr
34333 Optical Character Recognition
34334 This filter uses Tesseract for optical character recognition. To enable
34336 compilation of this filter, you need to configure FFmpeg with
34337 "--enable-libtesseract".
34338 It accepts the following options:
34340 datapath
34342 Set datapath to tesseract data. Default is to use whatever was set
34343 at installation.
34344 language
34346 Set language, default is "eng".
34347 whitelist
34349 Set character whitelist.
34350 blacklist
34352 Set character blacklist.
34353 The filter exports recognized text as the frame metadata
34355 "lavfi.ocr.text". The filter exports confidence of recognized words as
34356 the frame metadata "lavfi.ocr.confidence".
34357 ocv
34359 Apply a video transform using libopencv.
34360 To enable this filter, install the libopencv library and headers and
34362 configure FFmpeg with "--enable-libopencv".
34363 It accepts the following parameters:
34365 filter_name
34367 The name of the libopencv filter to apply.
34368 filter_params
34370 The parameters to pass to the libopencv filter. If not specified,
34371 the default values are assumed.
34372 Refer to the official libopencv documentation for more precise
34374 information:
34375 <http://docs.opencv.org/master/modules/imgproc/doc/filtering.html>
34376 Several libopencv filters are supported; see the following subsections.
34378 dilate
34380 Dilate an image by using a specific structuring element. It
34382 corresponds to the libopencv function "cvDilate".
34383 It accepts the parameters: struct_el|nb_iterations.
34385 struct_el represents a structuring element, and has the syntax:
34387 colsxrows+anchor_xxanchor_y/shape
34388 cols and rows represent the number of columns and rows of the
34390 structuring element, anchor_x and anchor_y the anchor point, and shape
34391 the shape for the structuring element. shape must be "rect", "cross",
34392 "ellipse", or "custom".
34393 If the value for shape is "custom", it must be followed by a string of
34395 the form "=filename". The file with name filename is assumed to
34396 represent a binary image, with each printable character corresponding
34397 to a bright pixel. When a custom shape is used, cols and rows are
34398 ignored, the number or columns and rows of the read file are assumed
34399 instead.
34400 The default value for struct_el is "3x3+0x0/rect".
34402 nb_iterations specifies the number of times the transform is applied to
34404 the image, and defaults to 1.
34405 Some examples:
34407 # Use the default values
34409 ocv=dilate
34410 # Dilate using a structuring element with a 5x5 cross, iterating two times
34412 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
34413 # Read the shape from the file diamond.shape, iterating two times.
34415 # The file diamond.shape may contain a pattern of characters like this
34416 # *
34417 # ***
34418 # *****
34419 # ***
34420 # *
34421 # The specified columns and rows are ignored
34422 # but the anchor point coordinates are not
34423 ocv=dilate:0x0+2x2/custom=diamond.shape|2
34424 erode
34426 Erode an image by using a specific structuring element. It corresponds
34428 to the libopencv function "cvErode".
34429 It accepts the parameters: struct_el:nb_iterations, with the same
34431 syntax and semantics as the dilate filter.
34432 smooth
34434 Smooth the input video.
34436 The filter takes the following parameters:
34438 type|param1|param2|param3|param4.
34439 type is the type of smooth filter to apply, and must be one of the
34441 following values: "blur", "blur_no_scale", "median", "gaussian", or
34442 "bilateral". The default value is "gaussian".
34443 The meaning of param1, param2, param3, and param4 depends on the smooth
34445 type. param1 and param2 accept integer positive values or 0. param3 and
34446 param4 accept floating point values.
34447 The default value for param1 is 3. The default value for the other
34449 parameters is 0.
34450 These parameters correspond to the parameters assigned to the libopencv
34452 function "cvSmooth".
34453 oscilloscope
34455 2D Video Oscilloscope.
34456 Useful to measure spatial impulse, step responses, chroma delays, etc.
34458 It accepts the following parameters:
34460 x Set scope center x position.
34462 y Set scope center y position.
34464 s Set scope size, relative to frame diagonal.
34466 t Set scope tilt/rotation.
34468 o Set trace opacity.
34470 tx Set trace center x position.
34472 ty Set trace center y position.
34474 tw Set trace width, relative to width of frame.
34476 th Set trace height, relative to height of frame.
34478 c Set which components to trace. By default it traces first three
34480 components.
34481 g Draw trace grid. By default is enabled.
34483 st Draw some statistics. By default is enabled.
34485 sc Draw scope. By default is enabled.
34487 Commands
34489 This filter supports same commands as options. The command accepts the
34491 same syntax of the corresponding option.
34492 If the specified expression is not valid, it is kept at its current
34494 value.
34495 Examples
34497 • Inspect full first row of video frame.
34499 oscilloscope=x=0.5:y=0:s=1
34501 • Inspect full last row of video frame.
34503 oscilloscope=x=0.5:y=1:s=1
34505 • Inspect full 5th line of video frame of height 1080.
34507 oscilloscope=x=0.5:y=5/1080:s=1
34509 • Inspect full last column of video frame.
34511 oscilloscope=x=1:y=0.5:s=1:t=1
34513 overlay
34515 Overlay one video on top of another.
34516 It takes two inputs and has one output. The first input is the "main"
34518 video on which the second input is overlaid.
34519 It accepts the following parameters:
34521 A description of the accepted options follows.
34523 x
34525 y Set the expression for the x and y coordinates of the overlaid
34526 video on the main video. Default value is "0" for both expressions.
34527 In case the expression is invalid, it is set to a huge value
34528 (meaning that the overlay will not be displayed within the output
34529 visible area).
34530 eof_action
34532 See framesync.
34533 eval
34535 Set when the expressions for x, and y are evaluated.
34536 It accepts the following values:
34538 init
34540 only evaluate expressions once during the filter initialization
34541 or when a command is processed
34542 frame
34544 evaluate expressions for each incoming frame
34545 Default value is frame.
34547 shortest
34549 See framesync.
34550 format
34552 Set the format for the output video.
34553 It accepts the following values:
34555 yuv420
34557 force YUV420 output
34558 yuv420p10
34560 force YUV420p10 output
34561 yuv422
34563 force YUV422 output
34564 yuv422p10
34566 force YUV422p10 output
34567 yuv444
34569 force YUV444 output
34570 rgb force packed RGB output
34572 gbrp
34574 force planar RGB output
34575 auto
34577 automatically pick format
34578 Default value is yuv420.
34580 repeatlast
34582 See framesync.
34583 alpha
34585 Set format of alpha of the overlaid video, it can be straight or
34586 premultiplied. Default is straight.
34587 The x, and y expressions can contain the following parameters.
34589 main_w, W
34591 main_h, H
34592 The main input width and height.
34593 overlay_w, w
34595 overlay_h, h
34596 The overlay input width and height.
34597 x
34599 y The computed values for x and y. They are evaluated for each new
34600 frame.
34601 hsub
34603 vsub
34604 horizontal and vertical chroma subsample values of the output
34605 format. For example for the pixel format "yuv422p" hsub is 2 and
34606 vsub is 1.
34607 n the number of input frame, starting from 0
34609 pos the position in the file of the input frame, NAN if unknown
34611 t The timestamp, expressed in seconds. It's NAN if the input
34613 timestamp is unknown.
34614 This filter also supports the framesync options.
34616 Note that the n, pos, t variables are available only when evaluation is
34618 done per frame, and will evaluate to NAN when eval is set to init.
34619 Be aware that frames are taken from each input video in timestamp
34621 order, hence, if their initial timestamps differ, it is a good idea to
34622 pass the two inputs through a setpts=PTS-STARTPTS filter to have them
34623 begin in the same zero timestamp, as the example for the movie filter
34624 does.
34625 You can chain together more overlays but you should test the efficiency
34627 of such approach.
34628 Commands
34630 This filter supports the following commands:
34632 x
34634 y Modify the x and y of the overlay input. The command accepts the
34635 same syntax of the corresponding option.
34636 If the specified expression is not valid, it is kept at its current
34638 value.
34639 Examples
34641 • Draw the overlay at 10 pixels from the bottom right corner of the
34643 main video:
34644 overlay=main_w-overlay_w-10:main_h-overlay_h-10
34646 Using named options the example above becomes:
34648 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
34650 • Insert a transparent PNG logo in the bottom left corner of the
34652 input, using the ffmpeg tool with the "-filter_complex" option:
34653 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
34655 • Insert 2 different transparent PNG logos (second logo on bottom
34657 right corner) using the ffmpeg tool:
34658 ffmpeg -i input -i logo1 -i logo2 -filter_complex 'overlay=x=10:y=H-h-10,overlay=x=W-w-10:y=H-h-10' output
34660 • Add a transparent color layer on top of the main video; "WxH" must
34662 specify the size of the main input to the overlay filter:
34663 color=color=red@.3:size=WxH [over]; [in][over] overlay [out]
34665 • Play an original video and a filtered version (here with the
34667 deshake filter) side by side using the ffplay tool:
34668 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
34670 The above command is the same as:
34672 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
34674 • Make a sliding overlay appearing from the left to the right top
34676 part of the screen starting since time 2:
34677 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
34679 • Compose output by putting two input videos side to side:
34681 ffmpeg -i left.avi -i right.avi -filter_complex "
34683 nullsrc=size=200x100 [background];
34684 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
34685 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
34686 [background][left] overlay=shortest=1 [background+left];
34687 [background+left][right] overlay=shortest=1:x=100 [left+right]
34688 "
34689 • Mask 10-20 seconds of a video by applying the delogo filter to a
34691 section
34692 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
34694 -vf '[in]split[split_main][split_delogo];[split_delogo]trim=start=360:end=371,delogo=0:0:640:480[delogoed];[split_main][delogoed]overlay=eof_action=pass[out]'
34695 masked.avi
34696 • Chain several overlays in cascade:
34698 nullsrc=s=200x200 [bg];
34700 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
34701 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
34702 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
34703 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
34704 [in3] null, [mid2] overlay=100:100 [out0]
34705 overlay_cuda
34707 Overlay one video on top of another.
34708 This is the CUDA variant of the overlay filter. It only accepts CUDA
34710 frames. The underlying input pixel formats have to match.
34711 It takes two inputs and has one output. The first input is the "main"
34713 video on which the second input is overlaid.
34714 It accepts the following parameters:
34716 x
34718 y Set expressions for the x and y coordinates of the overlaid video
34719 on the main video.
34720 They can contain the following parameters:
34722 main_w, W
34724 main_h, H
34725 The main input width and height.
34726 overlay_w, w
34728 overlay_h, h
34729 The overlay input width and height.
34730 x
34732 y The computed values for x and y. They are evaluated for each
34733 new frame.
34734 n The ordinal index of the main input frame, starting from 0.
34736 pos The byte offset position in the file of the main input frame,
34738 NAN if unknown.
34739 t The timestamp of the main input frame, expressed in seconds,
34741 NAN if unknown.
34742 Default value is "0" for both expressions.
34744 eval
34746 Set when the expressions for x and y are evaluated.
34747 It accepts the following values:
34749 init
34751 Evaluate expressions once during filter initialization or when
34752 a command is processed.
34753 frame
34755 Evaluate expressions for each incoming frame
34756 Default value is frame.
34758 eof_action
34760 See framesync.
34761 shortest
34763 See framesync.
34764 repeatlast
34766 See framesync.
34767 This filter also supports the framesync options.
34769 owdenoise
34771 Apply Overcomplete Wavelet denoiser.
34772 The filter accepts the following options:
34774 depth
34776 Set depth.
34777 Larger depth values will denoise lower frequency components more,
34779 but slow down filtering.
34780 Must be an int in the range 8-16, default is 8.
34782 luma_strength, ls
34784 Set luma strength.
34785 Must be a double value in the range 0-1000, default is 1.0.
34787 chroma_strength, cs
34789 Set chroma strength.
34790 Must be a double value in the range 0-1000, default is 1.0.
34792 pad
34794 Add paddings to the input image, and place the original input at the
34795 provided x, y coordinates.
34796 It accepts the following parameters:
34798 width, w
34800 height, h
34801 Specify an expression for the size of the output image with the
34802 paddings added. If the value for width or height is 0, the
34803 corresponding input size is used for the output.
34804 The width expression can reference the value set by the height
34806 expression, and vice versa.
34807 The default value of width and height is 0.
34809 x
34811 y Specify the offsets to place the input image at within the padded
34812 area, with respect to the top/left border of the output image.
34813 The x expression can reference the value set by the y expression,
34815 and vice versa.
34816 The default value of x and y is 0.
34818 If x or y evaluate to a negative number, they'll be changed so the
34820 input image is centered on the padded area.
34821 color
34823 Specify the color of the padded area. For the syntax of this
34824 option, check the "Color" section in the ffmpeg-utils manual.
34825 The default value of color is "black".
34827 eval
34829 Specify when to evaluate width, height, x and y expression.
34830 It accepts the following values:
34832 init
34834 Only evaluate expressions once during the filter initialization
34835 or when a command is processed.
34836 frame
34838 Evaluate expressions for each incoming frame.
34839 Default value is init.
34841 aspect
34843 Pad to aspect instead to a resolution.
34844 The value for the width, height, x, and y options are expressions
34846 containing the following constants:
34847 in_w
34849 in_h
34850 The input video width and height.
34851 iw
34853 ih These are the same as in_w and in_h.
34854 out_w
34856 out_h
34857 The output width and height (the size of the padded area), as
34858 specified by the width and height expressions.
34859 ow
34861 oh These are the same as out_w and out_h.
34862 x
34864 y The x and y offsets as specified by the x and y expressions, or NAN
34865 if not yet specified.
34866 a same as iw / ih
34868 sar input sample aspect ratio
34870 dar input display aspect ratio, it is the same as (iw / ih) * sar
34872 hsub
34874 vsub
34875 The horizontal and vertical chroma subsample values. For example
34876 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
34877 Examples
34879 • Add paddings with the color "violet" to the input video. The output
34881 video size is 640x480, and the top-left corner of the input video
34882 is placed at column 0, row 40
34883 pad=640:480:0:40:violet
34885 The example above is equivalent to the following command:
34887 pad=width=640:height=480:x=0:y=40:color=violet
34889 • Pad the input to get an output with dimensions increased by 3/2,
34891 and put the input video at the center of the padded area:
34892 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
34894 • Pad the input to get a squared output with size equal to the
34896 maximum value between the input width and height, and put the input
34897 video at the center of the padded area:
34898 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
34900 • Pad the input to get a final w/h ratio of 16:9:
34902 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
34904 • In case of anamorphic video, in order to set the output display
34906 aspect correctly, it is necessary to use sar in the expression,
34907 according to the relation:
34908 (ih * X / ih) * sar = output_dar
34910 X = output_dar / sar
34911 Thus the previous example needs to be modified to:
34913 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
34915 • Double the output size and put the input video in the bottom-right
34917 corner of the output padded area:
34918 pad="2*iw:2*ih:ow-iw:oh-ih"
34920 palettegen
34922 Generate one palette for a whole video stream.
34923 It accepts the following options:
34925 max_colors
34927 Set the maximum number of colors to quantize in the palette. Note:
34928 the palette will still contain 256 colors; the unused palette
34929 entries will be black.
34930 reserve_transparent
34932 Create a palette of 255 colors maximum and reserve the last one for
34933 transparency. Reserving the transparency color is useful for GIF
34934 optimization. If not set, the maximum of colors in the palette
34935 will be 256. You probably want to disable this option for a
34936 standalone image. Set by default.
34937 transparency_color
34939 Set the color that will be used as background for transparency.
34940 stats_mode
34942 Set statistics mode.
34943 It accepts the following values:
34945 full
34947 Compute full frame histograms.
34948 diff
34950 Compute histograms only for the part that differs from previous
34951 frame. This might be relevant to give more importance to the
34952 moving part of your input if the background is static.
34953 single
34955 Compute new histogram for each frame.
34956 Default value is full.
34958 The filter also exports the frame metadata "lavfi.color_quant_ratio"
34960 ("nb_color_in / nb_color_out") which you can use to evaluate the degree
34961 of color quantization of the palette. This information is also visible
34962 at info logging level.
34963 Examples
34965 • Generate a representative palette of a given video using ffmpeg:
34967 ffmpeg -i input.mkv -vf palettegen palette.png
34969 paletteuse
34971 Use a palette to downsample an input video stream.
34972 The filter takes two inputs: one video stream and a palette. The
34974 palette must be a 256 pixels image.
34975 It accepts the following options:
34977 dither
34979 Select dithering mode. Available algorithms are:
34980 bayer
34982 Ordered 8x8 bayer dithering (deterministic)
34983 heckbert
34985 Dithering as defined by Paul Heckbert in 1982 (simple error
34986 diffusion). Note: this dithering is sometimes considered
34987 "wrong" and is included as a reference.
34988 floyd_steinberg
34990 Floyd and Steingberg dithering (error diffusion)
34991 sierra2
34993 Frankie Sierra dithering v2 (error diffusion)
34994 sierra2_4a
34996 Frankie Sierra dithering v2 "Lite" (error diffusion)
34997 sierra3
34999 Frankie Sierra dithering v3 (error diffusion)
35000 burkes
35002 Burkes dithering (error diffusion)
35003 atkinson
35005 Atkinson dithering by Bill Atkinson at Apple Computer (error
35006 diffusion)
35007 Default is sierra2_4a.
35009 bayer_scale
35011 When bayer dithering is selected, this option defines the scale of
35012 the pattern (how much the crosshatch pattern is visible). A low
35013 value means more visible pattern for less banding, and higher value
35014 means less visible pattern at the cost of more banding.
35015 The option must be an integer value in the range [0,5]. Default is
35017 2.
35018 diff_mode
35020 If set, define the zone to process
35021 rectangle
35023 Only the changing rectangle will be reprocessed. This is
35024 similar to GIF cropping/offsetting compression mechanism. This
35025 option can be useful for speed if only a part of the image is
35026 changing, and has use cases such as limiting the scope of the
35027 error diffusal dither to the rectangle that bounds the moving
35028 scene (it leads to more deterministic output if the scene
35029 doesn't change much, and as a result less moving noise and
35030 better GIF compression).
35031 Default is none.
35033 new Take new palette for each output frame.
35035 alpha_threshold
35037 Sets the alpha threshold for transparency. Alpha values above this
35038 threshold will be treated as completely opaque, and values below
35039 this threshold will be treated as completely transparent.
35040 The option must be an integer value in the range [0,255]. Default
35042 is 128.
35043 Examples
35045 • Use a palette (generated for example with palettegen) to encode a
35047 GIF using ffmpeg:
35048 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
35050 perspective
35052 Correct perspective of video not recorded perpendicular to the screen.
35053 A description of the accepted parameters follows.
35055 x0
35057 y0
35058 x1
35059 y1
35060 x2
35061 y2
35062 x3
35063 y3 Set coordinates expression for top left, top right, bottom left and
35064 bottom right corners. Default values are "0:0:W:0:0:H:W:H" with
35065 which perspective will remain unchanged. If the "sense" option is
35066 set to "source", then the specified points will be sent to the
35067 corners of the destination. If the "sense" option is set to
35068 "destination", then the corners of the source will be sent to the
35069 specified coordinates.
35070 The expressions can use the following variables:
35072 W
35074 H the width and height of video frame.
35075 in Input frame count.
35077 on Output frame count.
35079 interpolation
35081 Set interpolation for perspective correction.
35082 It accepts the following values:
35084 linear
35086 cubic
35087 Default value is linear.
35089 sense
35091 Set interpretation of coordinate options.
35092 It accepts the following values:
35094 0, source
35096 Send point in the source specified by the given coordinates to
35097 the corners of the destination.
35098 1, destination
35100 Send the corners of the source to the point in the destination
35101 specified by the given coordinates.
35102 Default value is source.
35104 eval
35106 Set when the expressions for coordinates x0,y0,...x3,y3 are
35107 evaluated.
35108 It accepts the following values:
35110 init
35112 only evaluate expressions once during the filter initialization
35113 or when a command is processed
35114 frame
35116 evaluate expressions for each incoming frame
35117 Default value is init.
35119 phase
35121 Delay interlaced video by one field time so that the field order
35122 changes.
35123 The intended use is to fix PAL movies that have been captured with the
35125 opposite field order to the film-to-video transfer.
35126 A description of the accepted parameters follows.
35128 mode
35130 Set phase mode.
35131 It accepts the following values:
35133 t Capture field order top-first, transfer bottom-first. Filter
35135 will delay the bottom field.
35136 b Capture field order bottom-first, transfer top-first. Filter
35138 will delay the top field.
35139 p Capture and transfer with the same field order. This mode only
35141 exists for the documentation of the other options to refer to,
35142 but if you actually select it, the filter will faithfully do
35143 nothing.
35144 a Capture field order determined automatically by field flags,
35146 transfer opposite. Filter selects among t and b modes on a
35147 frame by frame basis using field flags. If no field information
35148 is available, then this works just like u.
35149 u Capture unknown or varying, transfer opposite. Filter selects
35151 among t and b on a frame by frame basis by analyzing the images
35152 and selecting the alternative that produces best match between
35153 the fields.
35154 T Capture top-first, transfer unknown or varying. Filter selects
35156 among t and p using image analysis.
35157 B Capture bottom-first, transfer unknown or varying. Filter
35159 selects among b and p using image analysis.
35160 A Capture determined by field flags, transfer unknown or varying.
35162 Filter selects among t, b and p using field flags and image
35163 analysis. If no field information is available, then this works
35164 just like U. This is the default mode.
35165 U Both capture and transfer unknown or varying. Filter selects
35167 among t, b and p using image analysis only.
35168 Commands
35170 This filter supports the all above options as commands.
35172 photosensitivity
35174 Reduce various flashes in video, so to help users with epilepsy.
35175 It accepts the following options:
35177 frames, f
35179 Set how many frames to use when filtering. Default is 30.
35180 threshold, t
35182 Set detection threshold factor. Default is 1. Lower is stricter.
35183 skip
35185 Set how many pixels to skip when sampling frames. Default is 1.
35186 Allowed range is from 1 to 1024.
35187 bypass
35189 Leave frames unchanged. Default is disabled.
35190 pixdesctest
35192 Pixel format descriptor test filter, mainly useful for internal
35193 testing. The output video should be equal to the input video.
35194 For example:
35196 format=monow, pixdesctest
35198 can be used to test the monowhite pixel format descriptor definition.
35200 pixelize
35202 Apply pixelization to video stream.
35203 The filter accepts the following options:
35205 width, w
35207 height, h
35208 Set block dimensions that will be used for pixelization. Default
35209 value is 16.
35210 mode, m
35212 Set the mode of pixelization used.
35213 Possible values are:
35215 avg
35217 min
35218 max
35219 Default value is "avg".
35221 planes, p
35223 Set what planes to filter. Default is to filter all planes.
35224 Commands
35226 This filter supports all options as commands.
35228 pixscope
35230 Display sample values of color channels. Mainly useful for checking
35231 color and levels. Minimum supported resolution is 640x480.
35232 The filters accept the following options:
35234 x Set scope X position, relative offset on X axis.
35236 y Set scope Y position, relative offset on Y axis.
35238 w Set scope width.
35240 h Set scope height.
35242 o Set window opacity. This window also holds statistics about pixel
35244 area.
35245 wx Set window X position, relative offset on X axis.
35247 wy Set window Y position, relative offset on Y axis.
35249 Commands
35251 This filter supports same commands as options.
35253 pp
35255 Enable the specified chain of postprocessing subfilters using
35256 libpostproc. This library should be automatically selected with a GPL
35257 build ("--enable-gpl"). Subfilters must be separated by '/' and can be
35258 disabled by prepending a '-'. Each subfilter and some options have a
35259 short and a long name that can be used interchangeably, i.e. dr/dering
35260 are the same.
35261 The filters accept the following options:
35263 subfilters
35265 Set postprocessing subfilters string.
35266 All subfilters share common options to determine their scope:
35268 a/autoq
35270 Honor the quality commands for this subfilter.
35271 c/chrom
35273 Do chrominance filtering, too (default).
35274 y/nochrom
35276 Do luminance filtering only (no chrominance).
35277 n/noluma
35279 Do chrominance filtering only (no luminance).
35280 These options can be appended after the subfilter name, separated by a
35282 '|'.
35283 Available subfilters are:
35285 hb/hdeblock[|difference[|flatness]]
35287 Horizontal deblocking filter
35288 difference
35290 Difference factor where higher values mean more deblocking
35291 (default: 32).
35292 flatness
35294 Flatness threshold where lower values mean more deblocking
35295 (default: 39).
35296 vb/vdeblock[|difference[|flatness]]
35298 Vertical deblocking filter
35299 difference
35301 Difference factor where higher values mean more deblocking
35302 (default: 32).
35303 flatness
35305 Flatness threshold where lower values mean more deblocking
35306 (default: 39).
35307 ha/hadeblock[|difference[|flatness]]
35309 Accurate horizontal deblocking filter
35310 difference
35312 Difference factor where higher values mean more deblocking
35313 (default: 32).
35314 flatness
35316 Flatness threshold where lower values mean more deblocking
35317 (default: 39).
35318 va/vadeblock[|difference[|flatness]]
35320 Accurate vertical deblocking filter
35321 difference
35323 Difference factor where higher values mean more deblocking
35324 (default: 32).
35325 flatness
35327 Flatness threshold where lower values mean more deblocking
35328 (default: 39).
35329 The horizontal and vertical deblocking filters share the difference and
35331 flatness values so you cannot set different horizontal and vertical
35332 thresholds.
35333 h1/x1hdeblock
35335 Experimental horizontal deblocking filter
35336 v1/x1vdeblock
35338 Experimental vertical deblocking filter
35339 dr/dering
35341 Deringing filter
35342 tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise
35344 reducer
35345 threshold1
35346 larger -> stronger filtering
35347 threshold2
35349 larger -> stronger filtering
35350 threshold3
35352 larger -> stronger filtering
35353 al/autolevels[:f/fullyrange], automatic brightness / contrast
35355 correction
35356 f/fullyrange
35357 Stretch luminance to "0-255".
35358 lb/linblenddeint
35360 Linear blend deinterlacing filter that deinterlaces the given block
35361 by filtering all lines with a "(1 2 1)" filter.
35362 li/linipoldeint
35364 Linear interpolating deinterlacing filter that deinterlaces the
35365 given block by linearly interpolating every second line.
35366 ci/cubicipoldeint
35368 Cubic interpolating deinterlacing filter deinterlaces the given
35369 block by cubically interpolating every second line.
35370 md/mediandeint
35372 Median deinterlacing filter that deinterlaces the given block by
35373 applying a median filter to every second line.
35374 fd/ffmpegdeint
35376 FFmpeg deinterlacing filter that deinterlaces the given block by
35377 filtering every second line with a "(-1 4 2 4 -1)" filter.
35378 l5/lowpass5
35380 Vertically applied FIR lowpass deinterlacing filter that
35381 deinterlaces the given block by filtering all lines with a "(-1 2 6
35382 2 -1)" filter.
35383 fq/forceQuant[|quantizer]
35385 Overrides the quantizer table from the input with the constant
35386 quantizer you specify.
35387 quantizer
35389 Quantizer to use
35390 de/default
35392 Default pp filter combination ("hb|a,vb|a,dr|a")
35393 fa/fast
35395 Fast pp filter combination ("h1|a,v1|a,dr|a")
35396 ac High quality pp filter combination ("ha|a|128|7,va|a,dr|a")
35398 Examples
35400 • Apply horizontal and vertical deblocking, deringing and automatic
35402 brightness/contrast:
35403 pp=hb/vb/dr/al
35405 • Apply default filters without brightness/contrast correction:
35407 pp=de/-al
35409 • Apply default filters and temporal denoiser:
35411 pp=default/tmpnoise|1|2|3
35413 • Apply deblocking on luminance only, and switch vertical deblocking
35415 on or off automatically depending on available CPU time:
35416 pp=hb|y/vb|a
35418 pp7
35420 Apply Postprocessing filter 7. It is variant of the spp filter, similar
35421 to spp = 6 with 7 point DCT, where only the center sample is used after
35422 IDCT.
35423 The filter accepts the following options:
35425 qp Force a constant quantization parameter. It accepts an integer in
35427 range 0 to 63. If not set, the filter will use the QP from the
35428 video stream (if available).
35429 mode
35431 Set thresholding mode. Available modes are:
35432 hard
35434 Set hard thresholding.
35435 soft
35437 Set soft thresholding (better de-ringing effect, but likely
35438 blurrier).
35439 medium
35441 Set medium thresholding (good results, default).
35442 premultiply
35444 Apply alpha premultiply effect to input video stream using first plane
35445 of second stream as alpha.
35446 Both streams must have same dimensions and same pixel format.
35448 The filter accepts the following option:
35450 planes
35452 Set which planes will be processed, unprocessed planes will be
35453 copied. By default value 0xf, all planes will be processed.
35454 inplace
35456 Do not require 2nd input for processing, instead use alpha plane
35457 from input stream.
35458 prewitt
35460 Apply prewitt operator to input video stream.
35461 The filter accepts the following option:
35463 planes
35465 Set which planes will be processed, unprocessed planes will be
35466 copied. By default value 0xf, all planes will be processed.
35467 scale
35469 Set value which will be multiplied with filtered result.
35470 delta
35472 Set value which will be added to filtered result.
35473 Commands
35475 This filter supports the all above options as commands.
35477 pseudocolor
35479 Alter frame colors in video with pseudocolors.
35480 This filter accepts the following options:
35482 c0 set pixel first component expression
35484 c1 set pixel second component expression
35486 c2 set pixel third component expression
35488 c3 set pixel fourth component expression, corresponds to the alpha
35490 component
35491 index, i
35493 set component to use as base for altering colors
35494 preset, p
35496 Pick one of built-in LUTs. By default is set to none.
35497 Available LUTs:
35499 magma
35501 inferno
35502 plasma
35503 viridis
35504 turbo
35505 cividis
35506 range1
35507 range2
35508 shadows
35509 highlights
35510 solar
35511 nominal
35512 preferred
35513 total
35514 spectral
35515 opacity
35516 Set opacity of output colors. Allowed range is from 0 to 1.
35517 Default value is set to 1.
35518 Each of the expression options specifies the expression to use for
35520 computing the lookup table for the corresponding pixel component
35521 values.
35522 The expressions can contain the following constants and functions:
35524 w
35526 h The input width and height.
35527 val The input value for the pixel component.
35529 ymin, umin, vmin, amin
35531 The minimum allowed component value.
35532 ymax, umax, vmax, amax
35534 The maximum allowed component value.
35535 All expressions default to "val".
35537 Commands
35539 This filter supports the all above options as commands.
35541 Examples
35543 • Change too high luma values to gradient:
35545 pseudocolor="'if(between(val,ymax,amax),lerp(ymin,ymax,(val-ymax)/(amax-ymax)),-1):if(between(val,ymax,amax),lerp(umax,umin,(val-ymax)/(amax-ymax)),-1):if(between(val,ymax,amax),lerp(vmin,vmax,(val-ymax)/(amax-ymax)),-1):-1'"
35547 psnr
35549 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
35550 Ratio) between two input videos.
35551 This filter takes in input two input videos, the first input is
35553 considered the "main" source and is passed unchanged to the output. The
35554 second input is used as a "reference" video for computing the PSNR.
35555 Both video inputs must have the same resolution and pixel format for
35557 this filter to work correctly. Also it assumes that both inputs have
35558 the same number of frames, which are compared one by one.
35559 The obtained average PSNR is printed through the logging system.
35561 The filter stores the accumulated MSE (mean squared error) of each
35563 frame, and at the end of the processing it is averaged across all
35564 frames equally, and the following formula is applied to obtain the
35565 PSNR:
35566 PSNR = 10*log10(MAX^2/MSE)
35568 Where MAX is the average of the maximum values of each component of the
35570 image.
35571 The description of the accepted parameters follows.
35573 stats_file, f
35575 If specified the filter will use the named file to save the PSNR of
35576 each individual frame. When filename equals "-" the data is sent to
35577 standard output.
35578 stats_version
35580 Specifies which version of the stats file format to use. Details of
35581 each format are written below. Default value is 1.
35582 stats_add_max
35584 Determines whether the max value is output to the stats log.
35585 Default value is 0. Requires stats_version >= 2. If this is set
35586 and stats_version < 2, the filter will return an error.
35587 This filter also supports the framesync options.
35589 The file printed if stats_file is selected, contains a sequence of
35591 key/value pairs of the form key:value for each compared couple of
35592 frames.
35593 If a stats_version greater than 1 is specified, a header line precedes
35595 the list of per-frame-pair stats, with key value pairs following the
35596 frame format with the following parameters:
35597 psnr_log_version
35599 The version of the log file format. Will match stats_version.
35600 fields
35602 A comma separated list of the per-frame-pair parameters included in
35603 the log.
35604 A description of each shown per-frame-pair parameter follows:
35606 n sequential number of the input frame, starting from 1
35608 mse_avg
35610 Mean Square Error pixel-by-pixel average difference of the compared
35611 frames, averaged over all the image components.
35612 mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
35614 Mean Square Error pixel-by-pixel average difference of the compared
35615 frames for the component specified by the suffix.
35616 psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
35618 Peak Signal to Noise ratio of the compared frames for the component
35619 specified by the suffix.
35620 max_avg, max_y, max_u, max_v
35622 Maximum allowed value for each channel, and average over all
35623 channels.
35624 Examples
35626 • For example:
35628 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
35630 [main][ref] psnr="stats_file=stats.log" [out]
35631 On this example the input file being processed is compared with the
35633 reference file ref_movie.mpg. The PSNR of each individual frame is
35634 stored in stats.log.
35635 • Another example with different containers:
35637 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]psnr" -f null -
35639 pullup
35641 Pulldown reversal (inverse telecine) filter, capable of handling mixed
35642 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps
35643 progressive content.
35644 The pullup filter is designed to take advantage of future context in
35646 making its decisions. This filter is stateless in the sense that it
35647 does not lock onto a pattern to follow, but it instead looks forward to
35648 the following fields in order to identify matches and rebuild
35649 progressive frames.
35650 To produce content with an even framerate, insert the fps filter after
35652 pullup, use "fps=24000/1001" if the input frame rate is 29.97fps,
35653 "fps=24" for 30fps and the (rare) telecined 25fps input.
35654 The filter accepts the following options:
35656 jl
35658 jr
35659 jt
35660 jb These options set the amount of "junk" to ignore at the left,
35661 right, top, and bottom of the image, respectively. Left and right
35662 are in units of 8 pixels, while top and bottom are in units of 2
35663 lines. The default is 8 pixels on each side.
35664 sb Set the strict breaks. Setting this option to 1 will reduce the
35666 chances of filter generating an occasional mismatched frame, but it
35667 may also cause an excessive number of frames to be dropped during
35668 high motion sequences. Conversely, setting it to -1 will make
35669 filter match fields more easily. This may help processing of video
35670 where there is slight blurring between the fields, but may also
35671 cause there to be interlaced frames in the output. Default value
35672 is 0.
35673 mp Set the metric plane to use. It accepts the following values:
35675 l Use luma plane.
35677 u Use chroma blue plane.
35679 v Use chroma red plane.
35681 This option may be set to use chroma plane instead of the default
35683 luma plane for doing filter's computations. This may improve
35684 accuracy on very clean source material, but more likely will
35685 decrease accuracy, especially if there is chroma noise (rainbow
35686 effect) or any grayscale video. The main purpose of setting mp to
35687 a chroma plane is to reduce CPU load and make pullup usable in
35688 realtime on slow machines.
35689 For best results (without duplicated frames in the output file) it is
35691 necessary to change the output frame rate. For example, to inverse
35692 telecine NTSC input:
35693 ffmpeg -i input -vf pullup -r 24000/1001 ...
35695 qp
35697 Change video quantization parameters (QP).
35698 The filter accepts the following option:
35700 qp Set expression for quantization parameter.
35702 The expression is evaluated through the eval API and can contain, among
35704 others, the following constants:
35705 known
35707 1 if index is not 129, 0 otherwise.
35708 qp Sequential index starting from -129 to 128.
35710 Examples
35712 • Some equation like:
35714 qp=2+2*sin(PI*qp)
35716 random
35718 Flush video frames from internal cache of frames into a random order.
35719 No frame is discarded. Inspired by frei0r nervous filter.
35720 frames
35722 Set size in number of frames of internal cache, in range from 2 to
35723 512. Default is 30.
35724 seed
35726 Set seed for random number generator, must be an integer included
35727 between 0 and "UINT32_MAX". If not specified, or if explicitly set
35728 to less than 0, the filter will try to use a good random seed on a
35729 best effort basis.
35730 readeia608
35732 Read closed captioning (EIA-608) information from the top lines of a
35733 video frame.
35734 This filter adds frame metadata for "lavfi.readeia608.X.cc" and
35736 "lavfi.readeia608.X.line", where "X" is the number of the identified
35737 line with EIA-608 data (starting from 0). A description of each
35738 metadata value follows:
35739 lavfi.readeia608.X.cc
35741 The two bytes stored as EIA-608 data (printed in hexadecimal).
35742 lavfi.readeia608.X.line
35744 The number of the line on which the EIA-608 data was identified and
35745 read.
35746 This filter accepts the following options:
35748 scan_min
35750 Set the line to start scanning for EIA-608 data. Default is 0.
35751 scan_max
35753 Set the line to end scanning for EIA-608 data. Default is 29.
35754 spw Set the ratio of width reserved for sync code detection. Default
35756 is 0.27. Allowed range is "[0.1 - 0.7]".
35757 chp Enable checking the parity bit. In the event of a parity error, the
35759 filter will output 0x00 for that character. Default is false.
35760 lp Lowpass lines prior to further processing. Default is enabled.
35762 Commands
35764 This filter supports the all above options as commands.
35766 Examples
35768 • Output a csv with presentation time and the first two lines of
35770 identified EIA-608 captioning data.
35771 ffprobe -f lavfi -i movie=captioned_video.mov,readeia608 -show_entries frame=pts_time:frame_tags=lavfi.readeia608.0.cc,lavfi.readeia608.1.cc -of csv
35773 readvitc
35775 Read vertical interval timecode (VITC) information from the top lines
35776 of a video frame.
35777 The filter adds frame metadata key "lavfi.readvitc.tc_str" with the
35779 timecode value, if a valid timecode has been detected. Further metadata
35780 key "lavfi.readvitc.found" is set to 0/1 depending on whether timecode
35781 data has been found or not.
35782 This filter accepts the following options:
35784 scan_max
35786 Set the maximum number of lines to scan for VITC data. If the value
35787 is set to -1 the full video frame is scanned. Default is 45.
35788 thr_b
35790 Set the luma threshold for black. Accepts float numbers in the
35791 range [0.0,1.0], default value is 0.2. The value must be equal or
35792 less than "thr_w".
35793 thr_w
35795 Set the luma threshold for white. Accepts float numbers in the
35796 range [0.0,1.0], default value is 0.6. The value must be equal or
35797 greater than "thr_b".
35798 Examples
35800 • Detect and draw VITC data onto the video frame; if no valid VITC is
35802 detected, draw "--:--:--:--" as a placeholder:
35803 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--}:x=(w-tw)/2:y=400-ascent'
35805 remap
35807 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
35808 Destination pixel at position (X, Y) will be picked from source (x, y)
35810 position where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are
35811 out of range, zero value for pixel will be used for destination pixel.
35812 Xmap and Ymap input video streams must be of same dimensions. Output
35814 video stream will have Xmap/Ymap video stream dimensions. Xmap and
35815 Ymap input video streams are 16bit depth, single channel.
35816 format
35818 Specify pixel format of output from this filter. Can be "color" or
35819 "gray". Default is "color".
35820 fill
35822 Specify the color of the unmapped pixels. For the syntax of this
35823 option, check the "Color" section in the ffmpeg-utils manual.
35824 Default color is "black".
35825 removegrain
35827 The removegrain filter is a spatial denoiser for progressive video.
35828 m0 Set mode for the first plane.
35830 m1 Set mode for the second plane.
35832 m2 Set mode for the third plane.
35834 m3 Set mode for the fourth plane.
35836 Range of mode is from 0 to 24. Description of each mode follows:
35838 0 Leave input plane unchanged. Default.
35840 1 Clips the pixel with the minimum and maximum of the 8 neighbour
35842 pixels.
35843 2 Clips the pixel with the second minimum and maximum of the 8
35845 neighbour pixels.
35846 3 Clips the pixel with the third minimum and maximum of the 8
35848 neighbour pixels.
35849 4 Clips the pixel with the fourth minimum and maximum of the 8
35851 neighbour pixels. This is equivalent to a median filter.
35852 5 Line-sensitive clipping giving the minimal change.
35854 6 Line-sensitive clipping, intermediate.
35856 7 Line-sensitive clipping, intermediate.
35858 8 Line-sensitive clipping, intermediate.
35860 9 Line-sensitive clipping on a line where the neighbours pixels are
35862 the closest.
35863 10 Replaces the target pixel with the closest neighbour.
35865 11 [1 2 1] horizontal and vertical kernel blur.
35867 12 Same as mode 11.
35869 13 Bob mode, interpolates top field from the line where the neighbours
35871 pixels are the closest.
35872 14 Bob mode, interpolates bottom field from the line where the
35874 neighbours pixels are the closest.
35875 15 Bob mode, interpolates top field. Same as 13 but with a more
35877 complicated interpolation formula.
35878 16 Bob mode, interpolates bottom field. Same as 14 but with a more
35880 complicated interpolation formula.
35881 17 Clips the pixel with the minimum and maximum of respectively the
35883 maximum and minimum of each pair of opposite neighbour pixels.
35884 18 Line-sensitive clipping using opposite neighbours whose greatest
35886 distance from the current pixel is minimal.
35887 19 Replaces the pixel with the average of its 8 neighbours.
35889 20 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
35891 21 Clips pixels using the averages of opposite neighbour.
35893 22 Same as mode 21 but simpler and faster.
35895 23 Small edge and halo removal, but reputed useless.
35897 24 Similar as 23.
35899 removelogo
35901 Suppress a TV station logo, using an image file to determine which
35902 pixels comprise the logo. It works by filling in the pixels that
35903 comprise the logo with neighboring pixels.
35904 The filter accepts the following options:
35906 filename, f
35908 Set the filter bitmap file, which can be any image format supported
35909 by libavformat. The width and height of the image file must match
35910 those of the video stream being processed.
35911 Pixels in the provided bitmap image with a value of zero are not
35913 considered part of the logo, non-zero pixels are considered part of the
35914 logo. If you use white (255) for the logo and black (0) for the rest,
35915 you will be safe. For making the filter bitmap, it is recommended to
35916 take a screen capture of a black frame with the logo visible, and then
35917 using a threshold filter followed by the erode filter once or twice.
35918 If needed, little splotches can be fixed manually. Remember that if
35920 logo pixels are not covered, the filter quality will be much reduced.
35921 Marking too many pixels as part of the logo does not hurt as much, but
35922 it will increase the amount of blurring needed to cover over the image
35923 and will destroy more information than necessary, and extra pixels will
35924 slow things down on a large logo.
35925 repeatfields
35927 This filter uses the repeat_field flag from the Video ES headers and
35928 hard repeats fields based on its value.
35929 reverse
35931 Reverse a video clip.
35932 Warning: This filter requires memory to buffer the entire clip, so
35934 trimming is suggested.
35935 Examples
35937 • Take the first 5 seconds of a clip, and reverse it.
35939 trim=end=5,reverse
35941 rgbashift
35943 Shift R/G/B/A pixels horizontally and/or vertically.
35944 The filter accepts the following options:
35946 rh Set amount to shift red horizontally.
35948 rv Set amount to shift red vertically.
35950 gh Set amount to shift green horizontally.
35952 gv Set amount to shift green vertically.
35954 bh Set amount to shift blue horizontally.
35956 bv Set amount to shift blue vertically.
35958 ah Set amount to shift alpha horizontally.
35960 av Set amount to shift alpha vertically.
35962 edge
35964 Set edge mode, can be smear, default, or warp.
35965 Commands
35967 This filter supports the all above options as commands.
35969 roberts
35971 Apply roberts cross operator to input video stream.
35972 The filter accepts the following option:
35974 planes
35976 Set which planes will be processed, unprocessed planes will be
35977 copied. By default value 0xf, all planes will be processed.
35978 scale
35980 Set value which will be multiplied with filtered result.
35981 delta
35983 Set value which will be added to filtered result.
35984 Commands
35986 This filter supports the all above options as commands.
35988 rotate
35990 Rotate video by an arbitrary angle expressed in radians.
35991 The filter accepts the following options:
35993 A description of the optional parameters follows.
35995 angle, a
35997 Set an expression for the angle by which to rotate the input video
35998 clockwise, expressed as a number of radians. A negative value will
35999 result in a counter-clockwise rotation. By default it is set to
36000 "0".
36001 This expression is evaluated for each frame.
36003 out_w, ow
36005 Set the output width expression, default value is "iw". This
36006 expression is evaluated just once during configuration.
36007 out_h, oh
36009 Set the output height expression, default value is "ih". This
36010 expression is evaluated just once during configuration.
36011 bilinear
36013 Enable bilinear interpolation if set to 1, a value of 0 disables
36014 it. Default value is 1.
36015 fillcolor, c
36017 Set the color used to fill the output area not covered by the
36018 rotated image. For the general syntax of this option, check the
36019 "Color" section in the ffmpeg-utils manual. If the special value
36020 "none" is selected then no background is printed (useful for
36021 example if the background is never shown).
36022 Default value is "black".
36024 The expressions for the angle and the output size can contain the
36026 following constants and functions:
36027 n sequential number of the input frame, starting from 0. It is always
36029 NAN before the first frame is filtered.
36030 t time in seconds of the input frame, it is set to 0 when the filter
36032 is configured. It is always NAN before the first frame is filtered.
36033 hsub
36035 vsub
36036 horizontal and vertical chroma subsample values. For example for
36037 the pixel format "yuv422p" hsub is 2 and vsub is 1.
36038 in_w, iw
36040 in_h, ih
36041 the input video width and height
36042 out_w, ow
36044 out_h, oh
36045 the output width and height, that is the size of the padded area as
36046 specified by the width and height expressions
36047 rotw(a)
36049 roth(a)
36050 the minimal width/height required for completely containing the
36051 input video rotated by a radians.
36052 These are only available when computing the out_w and out_h
36054 expressions.
36055 Examples
36057 • Rotate the input by PI/6 radians clockwise:
36059 rotate=PI/6
36061 • Rotate the input by PI/6 radians counter-clockwise:
36063 rotate=-PI/6
36065 • Rotate the input by 45 degrees clockwise:
36067 rotate=45*PI/180
36069 • Apply a constant rotation with period T, starting from an angle of
36071 PI/3:
36072 rotate=PI/3+2*PI*t/T
36074 • Make the input video rotation oscillating with a period of T
36076 seconds and an amplitude of A radians:
36077 rotate=A*sin(2*PI/T*t)
36079 • Rotate the video, output size is chosen so that the whole rotating
36081 input video is always completely contained in the output:
36082 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
36084 • Rotate the video, reduce the output size so that no background is
36086 ever shown:
36087 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
36089 Commands
36091 The filter supports the following commands:
36093 a, angle
36095 Set the angle expression. The command accepts the same syntax of
36096 the corresponding option.
36097 If the specified expression is not valid, it is kept at its current
36099 value.
36100 sab
36102 Apply Shape Adaptive Blur.
36103 The filter accepts the following options:
36105 luma_radius, lr
36107 Set luma blur filter strength, must be a value in range 0.1-4.0,
36108 default value is 1.0. A greater value will result in a more blurred
36109 image, and in slower processing.
36110 luma_pre_filter_radius, lpfr
36112 Set luma pre-filter radius, must be a value in the 0.1-2.0 range,
36113 default value is 1.0.
36114 luma_strength, ls
36116 Set luma maximum difference between pixels to still be considered,
36117 must be a value in the 0.1-100.0 range, default value is 1.0.
36118 chroma_radius, cr
36120 Set chroma blur filter strength, must be a value in range -0.9-4.0.
36121 A greater value will result in a more blurred image, and in slower
36122 processing.
36123 chroma_pre_filter_radius, cpfr
36125 Set chroma pre-filter radius, must be a value in the -0.9-2.0
36126 range.
36127 chroma_strength, cs
36129 Set chroma maximum difference between pixels to still be
36130 considered, must be a value in the -0.9-100.0 range.
36131 Each chroma option value, if not explicitly specified, is set to the
36133 corresponding luma option value.
36134 scale
36136 Scale (resize) the input video, using the libswscale library.
36137 The scale filter forces the output display aspect ratio to be the same
36139 of the input, by changing the output sample aspect ratio.
36140 If the input image format is different from the format requested by the
36142 next filter, the scale filter will convert the input to the requested
36143 format.
36144 Options
36146 The filter accepts the following options, or any of the options
36148 supported by the libswscale scaler.
36149 See the ffmpeg-scaler manual for the complete list of scaler options.
36151 width, w
36153 height, h
36154 Set the output video dimension expression. Default value is the
36155 input dimension.
36156 If the width or w value is 0, the input width is used for the
36158 output. If the height or h value is 0, the input height is used for
36159 the output.
36160 If one and only one of the values is -n with n >= 1, the scale
36162 filter will use a value that maintains the aspect ratio of the
36163 input image, calculated from the other specified dimension. After
36164 that it will, however, make sure that the calculated dimension is
36165 divisible by n and adjust the value if necessary.
36166 If both values are -n with n >= 1, the behavior will be identical
36168 to both values being set to 0 as previously detailed.
36169 See below for the list of accepted constants for use in the
36171 dimension expression.
36172 eval
36174 Specify when to evaluate width and height expression. It accepts
36175 the following values:
36176 init
36178 Only evaluate expressions once during the filter initialization
36179 or when a command is processed.
36180 frame
36182 Evaluate expressions for each incoming frame.
36183 Default value is init.
36185 interl
36187 Set the interlacing mode. It accepts the following values:
36188 1 Force interlaced aware scaling.
36190 0 Do not apply interlaced scaling.
36192 -1 Select interlaced aware scaling depending on whether the source
36194 frames are flagged as interlaced or not.
36195 Default value is 0.
36197 flags
36199 Set libswscale scaling flags. See the ffmpeg-scaler manual for the
36200 complete list of values. If not explicitly specified the filter
36201 applies the default flags.
36202 param0, param1
36204 Set libswscale input parameters for scaling algorithms that need
36205 them. See the ffmpeg-scaler manual for the complete documentation.
36206 If not explicitly specified the filter applies empty parameters.
36207 size, s
36209 Set the video size. For the syntax of this option, check the "Video
36210 size" section in the ffmpeg-utils manual.
36211 in_color_matrix
36213 out_color_matrix
36214 Set in/output YCbCr color space type.
36215 This allows the autodetected value to be overridden as well as
36217 allows forcing a specific value used for the output and encoder.
36218 If not specified, the color space type depends on the pixel format.
36220 Possible values:
36222 auto
36224 Choose automatically.
36225 bt709
36227 Format conforming to International Telecommunication Union
36228 (ITU) Recommendation BT.709.
36229 fcc Set color space conforming to the United States Federal
36231 Communications Commission (FCC) Code of Federal Regulations
36232 (CFR) Title 47 (2003) 73.682 (a).
36233 bt601
36235 bt470
36236 smpte170m
36237 Set color space conforming to:
36238 • ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
36240 • ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
36242 • Society of Motion Picture and Television Engineers (SMPTE)
36244 ST 170:2004
36245 smpte240m
36247 Set color space conforming to SMPTE ST 240:1999.
36248 bt2020
36250 Set color space conforming to ITU-R BT.2020 non-constant
36251 luminance system.
36252 in_range
36254 out_range
36255 Set in/output YCbCr sample range.
36256 This allows the autodetected value to be overridden as well as
36258 allows forcing a specific value used for the output and encoder. If
36259 not specified, the range depends on the pixel format. Possible
36260 values:
36261 auto/unknown
36263 Choose automatically.
36264 jpeg/full/pc
36266 Set full range (0-255 in case of 8-bit luma).
36267 mpeg/limited/tv
36269 Set "MPEG" range (16-235 in case of 8-bit luma).
36270 force_original_aspect_ratio
36272 Enable decreasing or increasing output video width or height if
36273 necessary to keep the original aspect ratio. Possible values:
36274 disable
36276 Scale the video as specified and disable this feature.
36277 decrease
36279 The output video dimensions will automatically be decreased if
36280 needed.
36281 increase
36283 The output video dimensions will automatically be increased if
36284 needed.
36285 One useful instance of this option is that when you know a specific
36287 device's maximum allowed resolution, you can use this to limit the
36288 output video to that, while retaining the aspect ratio. For
36289 example, device A allows 1280x720 playback, and your video is
36290 1920x800. Using this option (set it to decrease) and specifying
36291 1280x720 to the command line makes the output 1280x533.
36292 Please note that this is a different thing than specifying -1 for w
36294 or h, you still need to specify the output resolution for this
36295 option to work.
36296 force_divisible_by
36298 Ensures that both the output dimensions, width and height, are
36299 divisible by the given integer when used together with
36300 force_original_aspect_ratio. This works similar to using "-n" in
36301 the w and h options.
36302 This option respects the value set for force_original_aspect_ratio,
36304 increasing or decreasing the resolution accordingly. The video's
36305 aspect ratio may be slightly modified.
36306 This option can be handy if you need to have a video fit within or
36308 exceed a defined resolution using force_original_aspect_ratio but
36309 also have encoder restrictions on width or height divisibility.
36310 The values of the w and h options are expressions containing the
36312 following constants:
36313 in_w
36315 in_h
36316 The input width and height
36317 iw
36319 ih These are the same as in_w and in_h.
36320 out_w
36322 out_h
36323 The output (scaled) width and height
36324 ow
36326 oh These are the same as out_w and out_h
36327 a The same as iw / ih
36329 sar input sample aspect ratio
36331 dar The input display aspect ratio. Calculated from "(iw / ih) * sar".
36333 hsub
36335 vsub
36336 horizontal and vertical input chroma subsample values. For example
36337 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
36338 ohsub
36340 ovsub
36341 horizontal and vertical output chroma subsample values. For example
36342 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
36343 n The (sequential) number of the input frame, starting from 0. Only
36345 available with "eval=frame".
36346 t The presentation timestamp of the input frame, expressed as a
36348 number of seconds. Only available with "eval=frame".
36349 pos The position (byte offset) of the frame in the input stream, or NaN
36351 if this information is unavailable and/or meaningless (for example
36352 in case of synthetic video). Only available with "eval=frame".
36353 Examples
36355 • Scale the input video to a size of 200x100
36357 scale=w=200:h=100
36359 This is equivalent to:
36361 scale=200:100
36363 or:
36365 scale=200x100
36367 • Specify a size abbreviation for the output size:
36369 scale=qcif
36371 which can also be written as:
36373 scale=size=qcif
36375 • Scale the input to 2x:
36377 scale=w=2*iw:h=2*ih
36379 • The above is the same as:
36381 scale=2*in_w:2*in_h
36383 • Scale the input to 2x with forced interlaced scaling:
36385 scale=2*iw:2*ih:interl=1
36387 • Scale the input to half size:
36389 scale=w=iw/2:h=ih/2
36391 • Increase the width, and set the height to the same size:
36393 scale=3/2*iw:ow
36395 • Seek Greek harmony:
36397 scale=iw:1/PHI*iw
36399 scale=ih*PHI:ih
36400 • Increase the height, and set the width to 3/2 of the height:
36402 scale=w=3/2*oh:h=3/5*ih
36404 • Increase the size, making the size a multiple of the chroma
36406 subsample values:
36407 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
36409 • Increase the width to a maximum of 500 pixels, keeping the same
36411 aspect ratio as the input:
36412 scale=w='min(500\, iw*3/2):h=-1'
36414 • Make pixels square by combining scale and setsar:
36416 scale='trunc(ih*dar):ih',setsar=1/1
36418 • Make pixels square by combining scale and setsar, making sure the
36420 resulting resolution is even (required by some codecs):
36421 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
36423 Commands
36425 This filter supports the following commands:
36427 width, w
36429 height, h
36430 Set the output video dimension expression. The command accepts the
36431 same syntax of the corresponding option.
36432 If the specified expression is not valid, it is kept at its current
36434 value.
36435 scale_cuda
36437 Scale (resize) and convert (pixel format) the input video, using
36438 accelerated CUDA kernels. Setting the output width and height works in
36439 the same way as for the scale filter.
36440 The filter accepts the following options:
36442 w
36444 h Set the output video dimension expression. Default value is the
36445 input dimension.
36446 Allows for the same expressions as the scale filter.
36448 interp_algo
36450 Sets the algorithm used for scaling:
36451 nearest
36453 Nearest neighbour
36454 Used by default if input parameters match the desired output.
36456 bilinear
36458 Bilinear
36459 bicubic
36461 Bicubic
36462 This is the default.
36464 lanczos
36466 Lanczos
36467 format
36469 Controls the output pixel format. By default, or if none is
36470 specified, the input pixel format is used.
36471 The filter does not support converting between YUV and RGB pixel
36473 formats.
36474 passthrough
36476 If set to 0, every frame is processed, even if no conversion is
36477 neccesary. This mode can be useful to use the filter as a buffer
36478 for a downstream frame-consumer that exhausts the limited decoder
36479 frame pool.
36480 If set to 1, frames are passed through as-is if they match the
36482 desired output parameters. This is the default behaviour.
36483 param
36485 Algorithm-Specific parameter.
36486 Affects the curves of the bicubic algorithm.
36488 force_original_aspect_ratio
36490 force_divisible_by
36491 Work the same as the identical scale filter options.
36492 Examples
36494 • Scale input to 720p, keeping aspect ratio and ensuring the output
36496 is yuv420p.
36497 scale_cuda=-2:720:format=yuv420p
36499 • Upscale to 4K using nearest neighbour algorithm.
36501 scale_cuda=4096:2160:interp_algo=nearest
36503 • Don't do any conversion or scaling, but copy all input frames into
36505 newly allocated ones. This can be useful to deal with a filter and
36506 encode chain that otherwise exhausts the decoders frame pool.
36507 scale_cuda=passthrough=0
36509 scale_npp
36511 Use the NVIDIA Performance Primitives (libnpp) to perform scaling
36512 and/or pixel format conversion on CUDA video frames. Setting the output
36513 width and height works in the same way as for the scale filter.
36514 The following additional options are accepted:
36516 format
36518 The pixel format of the output CUDA frames. If set to the string
36519 "same" (the default), the input format will be kept. Note that
36520 automatic format negotiation and conversion is not yet supported
36521 for hardware frames
36522 interp_algo
36524 The interpolation algorithm used for resizing. One of the
36525 following:
36526 nn Nearest neighbour.
36528 linear
36530 cubic
36531 cubic2p_bspline
36532 2-parameter cubic (B=1, C=0)
36533 cubic2p_catmullrom
36535 2-parameter cubic (B=0, C=1/2)
36536 cubic2p_b05c03
36538 2-parameter cubic (B=1/2, C=3/10)
36539 super
36541 Supersampling
36542 lanczos
36544 force_original_aspect_ratio
36545 Enable decreasing or increasing output video width or height if
36546 necessary to keep the original aspect ratio. Possible values:
36547 disable
36549 Scale the video as specified and disable this feature.
36550 decrease
36552 The output video dimensions will automatically be decreased if
36553 needed.
36554 increase
36556 The output video dimensions will automatically be increased if
36557 needed.
36558 One useful instance of this option is that when you know a specific
36560 device's maximum allowed resolution, you can use this to limit the
36561 output video to that, while retaining the aspect ratio. For
36562 example, device A allows 1280x720 playback, and your video is
36563 1920x800. Using this option (set it to decrease) and specifying
36564 1280x720 to the command line makes the output 1280x533.
36565 Please note that this is a different thing than specifying -1 for w
36567 or h, you still need to specify the output resolution for this
36568 option to work.
36569 force_divisible_by
36571 Ensures that both the output dimensions, width and height, are
36572 divisible by the given integer when used together with
36573 force_original_aspect_ratio. This works similar to using "-n" in
36574 the w and h options.
36575 This option respects the value set for force_original_aspect_ratio,
36577 increasing or decreasing the resolution accordingly. The video's
36578 aspect ratio may be slightly modified.
36579 This option can be handy if you need to have a video fit within or
36581 exceed a defined resolution using force_original_aspect_ratio but
36582 also have encoder restrictions on width or height divisibility.
36583 eval
36585 Specify when to evaluate width and height expression. It accepts
36586 the following values:
36587 init
36589 Only evaluate expressions once during the filter initialization
36590 or when a command is processed.
36591 frame
36593 Evaluate expressions for each incoming frame.
36594 The values of the w and h options are expressions containing the
36596 following constants:
36597 in_w
36599 in_h
36600 The input width and height
36601 iw
36603 ih These are the same as in_w and in_h.
36604 out_w
36606 out_h
36607 The output (scaled) width and height
36608 ow
36610 oh These are the same as out_w and out_h
36611 a The same as iw / ih
36613 sar input sample aspect ratio
36615 dar The input display aspect ratio. Calculated from "(iw / ih) * sar".
36617 n The (sequential) number of the input frame, starting from 0. Only
36619 available with "eval=frame".
36620 t The presentation timestamp of the input frame, expressed as a
36622 number of seconds. Only available with "eval=frame".
36623 pos The position (byte offset) of the frame in the input stream, or NaN
36625 if this information is unavailable and/or meaningless (for example
36626 in case of synthetic video). Only available with "eval=frame".
36627 scale2ref
36629 Scale (resize) the input video, based on a reference video.
36630 See the scale filter for available options, scale2ref supports the same
36632 but uses the reference video instead of the main input as basis.
36633 scale2ref also supports the following additional constants for the w
36634 and h options:
36635 main_w
36637 main_h
36638 The main input video's width and height
36639 main_a
36641 The same as main_w / main_h
36642 main_sar
36644 The main input video's sample aspect ratio
36645 main_dar, mdar
36647 The main input video's display aspect ratio. Calculated from
36648 "(main_w / main_h) * main_sar".
36649 main_hsub
36651 main_vsub
36652 The main input video's horizontal and vertical chroma subsample
36653 values. For example for the pixel format "yuv422p" hsub is 2 and
36654 vsub is 1.
36655 main_n
36657 The (sequential) number of the main input frame, starting from 0.
36658 Only available with "eval=frame".
36659 main_t
36661 The presentation timestamp of the main input frame, expressed as a
36662 number of seconds. Only available with "eval=frame".
36663 main_pos
36665 The position (byte offset) of the frame in the main input stream,
36666 or NaN if this information is unavailable and/or meaningless (for
36667 example in case of synthetic video). Only available with
36668 "eval=frame".
36669 Examples
36671 • Scale a subtitle stream (b) to match the main video (a) in size
36673 before overlaying
36674 'scale2ref[b][a];[a][b]overlay'
36676 • Scale a logo to 1/10th the height of a video, while preserving its
36678 display aspect ratio.
36679 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
36681 Commands
36683 This filter supports the following commands:
36685 width, w
36687 height, h
36688 Set the output video dimension expression. The command accepts the
36689 same syntax of the corresponding option.
36690 If the specified expression is not valid, it is kept at its current
36692 value.
36693 scale2ref_npp
36695 Use the NVIDIA Performance Primitives (libnpp) to scale (resize) the
36696 input video, based on a reference video.
36697 See the scale_npp filter for available options, scale2ref_npp supports
36699 the same but uses the reference video instead of the main input as
36700 basis. scale2ref_npp also supports the following additional constants
36701 for the w and h options:
36702 main_w
36704 main_h
36705 The main input video's width and height
36706 main_a
36708 The same as main_w / main_h
36709 main_sar
36711 The main input video's sample aspect ratio
36712 main_dar, mdar
36714 The main input video's display aspect ratio. Calculated from
36715 "(main_w / main_h) * main_sar".
36716 main_n
36718 The (sequential) number of the main input frame, starting from 0.
36719 Only available with "eval=frame".
36720 main_t
36722 The presentation timestamp of the main input frame, expressed as a
36723 number of seconds. Only available with "eval=frame".
36724 main_pos
36726 The position (byte offset) of the frame in the main input stream,
36727 or NaN if this information is unavailable and/or meaningless (for
36728 example in case of synthetic video). Only available with
36729 "eval=frame".
36730 Examples
36732 • Scale a subtitle stream (b) to match the main video (a) in size
36734 before overlaying
36735 'scale2ref_npp[b][a];[a][b]overlay_cuda'
36737 • Scale a logo to 1/10th the height of a video, while preserving its
36739 display aspect ratio.
36740 [logo-in][video-in]scale2ref_npp=w=oh*mdar:h=ih/10[logo-out][video-out]
36742 scharr
36744 Apply scharr operator to input video stream.
36745 The filter accepts the following option:
36747 planes
36749 Set which planes will be processed, unprocessed planes will be
36750 copied. By default value 0xf, all planes will be processed.
36751 scale
36753 Set value which will be multiplied with filtered result.
36754 delta
36756 Set value which will be added to filtered result.
36757 Commands
36759 This filter supports the all above options as commands.
36761 scroll
36763 Scroll input video horizontally and/or vertically by constant speed.
36764 The filter accepts the following options:
36766 horizontal, h
36768 Set the horizontal scrolling speed. Default is 0. Allowed range is
36769 from -1 to 1. Negative values changes scrolling direction.
36770 vertical, v
36772 Set the vertical scrolling speed. Default is 0. Allowed range is
36773 from -1 to 1. Negative values changes scrolling direction.
36774 hpos
36776 Set the initial horizontal scrolling position. Default is 0.
36777 Allowed range is from 0 to 1.
36778 vpos
36780 Set the initial vertical scrolling position. Default is 0. Allowed
36781 range is from 0 to 1.
36782 Commands
36784 This filter supports the following commands:
36786 horizontal, h
36788 Set the horizontal scrolling speed.
36789 vertical, v
36791 Set the vertical scrolling speed.
36792 scdet
36794 Detect video scene change.
36795 This filter sets frame metadata with mafd between frame, the scene
36797 score, and forward the frame to the next filter, so they can use these
36798 metadata to detect scene change or others.
36799 In addition, this filter logs a message and sets frame metadata when it
36801 detects a scene change by threshold.
36802 "lavfi.scd.mafd" metadata keys are set with mafd for every frame.
36804 "lavfi.scd.score" metadata keys are set with scene change score for
36806 every frame to detect scene change.
36807 "lavfi.scd.time" metadata keys are set with current filtered frame time
36809 which detect scene change with threshold.
36810 The filter accepts the following options:
36812 threshold, t
36814 Set the scene change detection threshold as a percentage of maximum
36815 change. Good values are in the "[8.0, 14.0]" range. The range for
36816 threshold is "[0., 100.]".
36817 Default value is 10..
36819 sc_pass, s
36821 Set the flag to pass scene change frames to the next filter.
36822 Default value is 0 You can enable it if you want to get snapshot of
36823 scene change frames only.
36824 selectivecolor
36826 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of
36827 colors (such as "reds", "yellows", "greens", "cyans", ...). The
36828 adjustment range is defined by the "purity" of the color (that is, how
36829 saturated it already is).
36830 This filter is similar to the Adobe Photoshop Selective Color tool.
36832 The filter accepts the following options:
36834 correction_method
36836 Select color correction method.
36837 Available values are:
36839 absolute
36841 Specified adjustments are applied "as-is" (added/subtracted to
36842 original pixel component value).
36843 relative
36845 Specified adjustments are relative to the original component
36846 value.
36847 Default is "absolute".
36849 reds
36851 Adjustments for red pixels (pixels where the red component is the
36852 maximum)
36853 yellows
36855 Adjustments for yellow pixels (pixels where the blue component is
36856 the minimum)
36857 greens
36859 Adjustments for green pixels (pixels where the green component is
36860 the maximum)
36861 cyans
36863 Adjustments for cyan pixels (pixels where the red component is the
36864 minimum)
36865 blues
36867 Adjustments for blue pixels (pixels where the blue component is the
36868 maximum)
36869 magentas
36871 Adjustments for magenta pixels (pixels where the green component is
36872 the minimum)
36873 whites
36875 Adjustments for white pixels (pixels where all components are
36876 greater than 128)
36877 neutrals
36879 Adjustments for all pixels except pure black and pure white
36880 blacks
36882 Adjustments for black pixels (pixels where all components are
36883 lesser than 128)
36884 psfile
36886 Specify a Photoshop selective color file (".asv") to import the
36887 settings from.
36888 All the adjustment settings (reds, yellows, ...) accept up to 4 space
36890 separated floating point adjustment values in the [-1,1] range,
36891 respectively to adjust the amount of cyan, magenta, yellow and black
36892 for the pixels of its range.
36893 Examples
36895 • Increase cyan by 50% and reduce yellow by 33% in every green areas,
36897 and increase magenta by 27% in blue areas:
36898 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
36900 • Use a Photoshop selective color preset:
36902 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
36904 separatefields
36906 The "separatefields" takes a frame-based video input and splits each
36907 frame into its components fields, producing a new half height clip with
36908 twice the frame rate and twice the frame count.
36909 This filter use field-dominance information in frame to decide which of
36911 each pair of fields to place first in the output. If it gets it wrong
36912 use setfield filter before "separatefields" filter.
36913 setdar, setsar
36915 The "setdar" filter sets the Display Aspect Ratio for the filter output
36916 video.
36917 This is done by changing the specified Sample (aka Pixel) Aspect Ratio,
36919 according to the following equation:
36920 <DAR> = <HORIZONTAL_RESOLUTION> / <VERTICAL_RESOLUTION> * <SAR>
36922 Keep in mind that the "setdar" filter does not modify the pixel
36924 dimensions of the video frame. Also, the display aspect ratio set by
36925 this filter may be changed by later filters in the filterchain, e.g. in
36926 case of scaling or if another "setdar" or a "setsar" filter is applied.
36927 The "setsar" filter sets the Sample (aka Pixel) Aspect Ratio for the
36929 filter output video.
36930 Note that as a consequence of the application of this filter, the
36932 output display aspect ratio will change according to the equation
36933 above.
36934 Keep in mind that the sample aspect ratio set by the "setsar" filter
36936 may be changed by later filters in the filterchain, e.g. if another
36937 "setsar" or a "setdar" filter is applied.
36938 It accepts the following parameters:
36940 r, ratio, dar ("setdar" only), sar ("setsar" only)
36942 Set the aspect ratio used by the filter.
36943 The parameter can be a floating point number string, an expression,
36945 or a string of the form num:den, where num and den are the
36946 numerator and denominator of the aspect ratio. If the parameter is
36947 not specified, it is assumed the value "0". In case the form
36948 "num:den" is used, the ":" character should be escaped.
36949 max Set the maximum integer value to use for expressing numerator and
36951 denominator when reducing the expressed aspect ratio to a rational.
36952 Default value is 100.
36953 The parameter sar is an expression containing the following constants:
36955 E, PI, PHI
36957 These are approximated values for the mathematical constants e
36958 (Euler's number), pi (Greek pi), and phi (the golden ratio).
36959 w, h
36961 The input width and height.
36962 a These are the same as w / h.
36964 sar The input sample aspect ratio.
36966 dar The input display aspect ratio. It is the same as (w / h) * sar.
36968 hsub, vsub
36970 Horizontal and vertical chroma subsample values. For example, for
36971 the pixel format "yuv422p" hsub is 2 and vsub is 1.
36972 Examples
36974 • To change the display aspect ratio to 16:9, specify one of the
36976 following:
36977 setdar=dar=1.77777
36979 setdar=dar=16/9
36980 • To change the sample aspect ratio to 10:11, specify:
36982 setsar=sar=10/11
36984 • To set a display aspect ratio of 16:9, and specify a maximum
36986 integer value of 1000 in the aspect ratio reduction, use the
36987 command:
36988 setdar=ratio=16/9:max=1000
36990 setfield
36992 Force field for the output video frame.
36993 The "setfield" filter marks the interlace type field for the output
36995 frames. It does not change the input frame, but only sets the
36996 corresponding property, which affects how the frame is treated by
36997 following filters (e.g. "fieldorder" or "yadif").
36998 The filter accepts the following options:
37000 mode
37002 Available values are:
37003 auto
37005 Keep the same field property.
37006 bff Mark the frame as bottom-field-first.
37008 tff Mark the frame as top-field-first.
37010 prog
37012 Mark the frame as progressive.
37013 setparams
37015 Force frame parameter for the output video frame.
37016 The "setparams" filter marks interlace and color range for the output
37018 frames. It does not change the input frame, but only sets the
37019 corresponding property, which affects how the frame is treated by
37020 filters/encoders.
37021 field_mode
37023 Available values are:
37024 auto
37026 Keep the same field property (default).
37027 bff Mark the frame as bottom-field-first.
37029 tff Mark the frame as top-field-first.
37031 prog
37033 Mark the frame as progressive.
37034 range
37036 Available values are:
37037 auto
37039 Keep the same color range property (default).
37040 unspecified, unknown
37042 Mark the frame as unspecified color range.
37043 limited, tv, mpeg
37045 Mark the frame as limited range.
37046 full, pc, jpeg
37048 Mark the frame as full range.
37049 color_primaries
37051 Set the color primaries. Available values are:
37052 auto
37054 Keep the same color primaries property (default).
37055 bt709
37057 unknown
37058 bt470m
37059 bt470bg
37060 smpte170m
37061 smpte240m
37062 film
37063 bt2020
37064 smpte428
37065 smpte431
37066 smpte432
37067 jedec-p22
37068 color_trc
37069 Set the color transfer. Available values are:
37070 auto
37072 Keep the same color trc property (default).
37073 bt709
37075 unknown
37076 bt470m
37077 bt470bg
37078 smpte170m
37079 smpte240m
37080 linear
37081 log100
37082 log316
37083 iec61966-2-4
37084 bt1361e
37085 iec61966-2-1
37086 bt2020-10
37087 bt2020-12
37088 smpte2084
37089 smpte428
37090 arib-std-b67
37091 colorspace
37092 Set the colorspace. Available values are:
37093 auto
37095 Keep the same colorspace property (default).
37096 gbr
37098 bt709
37099 unknown
37100 fcc
37101 bt470bg
37102 smpte170m
37103 smpte240m
37104 ycgco
37105 bt2020nc
37106 bt2020c
37107 smpte2085
37108 chroma-derived-nc
37109 chroma-derived-c
37110 ictcp
37111 sharpen_npp
37113 Use the NVIDIA Performance Primitives (libnpp) to perform image
37114 sharpening with border control.
37115 The following additional options are accepted:
37117 border_type
37119 Type of sampling to be used ad frame borders. One of the following:
37120 replicate
37122 Replicate pixel values.
37123 shear
37125 Apply shear transform to input video.
37126 This filter supports the following options:
37128 shx Shear factor in X-direction. Default value is 0. Allowed range is
37130 from -2 to 2.
37131 shy Shear factor in Y-direction. Default value is 0. Allowed range is
37133 from -2 to 2.
37134 fillcolor, c
37136 Set the color used to fill the output area not covered by the
37137 transformed video. For the general syntax of this option, check the
37138 "Color" section in the ffmpeg-utils manual. If the special value
37139 "none" is selected then no background is printed (useful for
37140 example if the background is never shown).
37141 Default value is "black".
37143 interp
37145 Set interpolation type. Can be "bilinear" or "nearest". Default is
37146 "bilinear".
37147 Commands
37149 This filter supports the all above options as commands.
37151 showinfo
37153 Show a line containing various information for each input video frame.
37154 The input video is not modified.
37155 This filter supports the following options:
37157 checksum
37159 Calculate checksums of each plane. By default enabled.
37160 The shown line contains a sequence of key/value pairs of the form
37162 key:value.
37163 The following values are shown in the output:
37165 n The (sequential) number of the input frame, starting from 0.
37167 pts The Presentation TimeStamp of the input frame, expressed as a
37169 number of time base units. The time base unit depends on the filter
37170 input pad.
37171 pts_time
37173 The Presentation TimeStamp of the input frame, expressed as a
37174 number of seconds.
37175 pos The position of the frame in the input stream, or -1 if this
37177 information is unavailable and/or meaningless (for example in case
37178 of synthetic video).
37179 fmt The pixel format name.
37181 sar The sample aspect ratio of the input frame, expressed in the form
37183 num/den.
37184 s The size of the input frame. For the syntax of this option, check
37186 the "Video size" section in the ffmpeg-utils manual.
37187 i The type of interlaced mode ("P" for "progressive", "T" for top
37189 field first, "B" for bottom field first).
37190 iskey
37192 This is 1 if the frame is a key frame, 0 otherwise.
37193 type
37195 The picture type of the input frame ("I" for an I-frame, "P" for a
37196 P-frame, "B" for a B-frame, or "?" for an unknown type). Also
37197 refer to the documentation of the "AVPictureType" enum and of the
37198 "av_get_picture_type_char" function defined in libavutil/avutil.h.
37199 checksum
37201 The Adler-32 checksum (printed in hexadecimal) of all the planes of
37202 the input frame.
37203 plane_checksum
37205 The Adler-32 checksum (printed in hexadecimal) of each plane of the
37206 input frame, expressed in the form "[c0 c1 c2 c3]".
37207 mean
37209 The mean value of pixels in each plane of the input frame,
37210 expressed in the form "[mean0 mean1 mean2 mean3]".
37211 stdev
37213 The standard deviation of pixel values in each plane of the input
37214 frame, expressed in the form "[stdev0 stdev1 stdev2 stdev3]".
37215 showpalette
37217 Displays the 256 colors palette of each frame. This filter is only
37218 relevant for pal8 pixel format frames.
37219 It accepts the following option:
37221 s Set the size of the box used to represent one palette color entry.
37223 Default is 30 (for a "30x30" pixel box).
37224 shuffleframes
37226 Reorder and/or duplicate and/or drop video frames.
37227 It accepts the following parameters:
37229 mapping
37231 Set the destination indexes of input frames. This is space or '|'
37232 separated list of indexes that maps input frames to output frames.
37233 Number of indexes also sets maximal value that each index may have.
37234 '-1' index have special meaning and that is to drop frame.
37235 The first frame has the index 0. The default is to keep the input
37237 unchanged.
37238 Examples
37240 • Swap second and third frame of every three frames of the input:
37242 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
37244 • Swap 10th and 1st frame of every ten frames of the input:
37246 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
37248 shufflepixels
37250 Reorder pixels in video frames.
37251 This filter accepts the following options:
37253 direction, d
37255 Set shuffle direction. Can be forward or inverse direction.
37256 Default direction is forward.
37257 mode, m
37259 Set shuffle mode. Can be horizontal, vertical or block mode.
37260 width, w
37262 height, h
37263 Set shuffle block_size. In case of horizontal shuffle mode only
37264 width part of size is used, and in case of vertical shuffle mode
37265 only height part of size is used.
37266 seed, s
37268 Set random seed used with shuffling pixels. Mainly useful to set to
37269 be able to reverse filtering process to get original input. For
37270 example, to reverse forward shuffle you need to use same parameters
37271 and exact same seed and to set direction to inverse.
37272 shuffleplanes
37274 Reorder and/or duplicate video planes.
37275 It accepts the following parameters:
37277 map0
37279 The index of the input plane to be used as the first output plane.
37280 map1
37282 The index of the input plane to be used as the second output plane.
37283 map2
37285 The index of the input plane to be used as the third output plane.
37286 map3
37288 The index of the input plane to be used as the fourth output plane.
37289 The first plane has the index 0. The default is to keep the input
37291 unchanged.
37292 Examples
37294 • Swap the second and third planes of the input:
37296 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
37298 signalstats
37300 Evaluate various visual metrics that assist in determining issues
37301 associated with the digitization of analog video media.
37302 By default the filter will log these metadata values:
37304 YMIN
37306 Display the minimal Y value contained within the input frame.
37307 Expressed in range of [0-255].
37308 YLOW
37310 Display the Y value at the 10% percentile within the input frame.
37311 Expressed in range of [0-255].
37312 YAVG
37314 Display the average Y value within the input frame. Expressed in
37315 range of [0-255].
37316 YHIGH
37318 Display the Y value at the 90% percentile within the input frame.
37319 Expressed in range of [0-255].
37320 YMAX
37322 Display the maximum Y value contained within the input frame.
37323 Expressed in range of [0-255].
37324 UMIN
37326 Display the minimal U value contained within the input frame.
37327 Expressed in range of [0-255].
37328 ULOW
37330 Display the U value at the 10% percentile within the input frame.
37331 Expressed in range of [0-255].
37332 UAVG
37334 Display the average U value within the input frame. Expressed in
37335 range of [0-255].
37336 UHIGH
37338 Display the U value at the 90% percentile within the input frame.
37339 Expressed in range of [0-255].
37340 UMAX
37342 Display the maximum U value contained within the input frame.
37343 Expressed in range of [0-255].
37344 VMIN
37346 Display the minimal V value contained within the input frame.
37347 Expressed in range of [0-255].
37348 VLOW
37350 Display the V value at the 10% percentile within the input frame.
37351 Expressed in range of [0-255].
37352 VAVG
37354 Display the average V value within the input frame. Expressed in
37355 range of [0-255].
37356 VHIGH
37358 Display the V value at the 90% percentile within the input frame.
37359 Expressed in range of [0-255].
37360 VMAX
37362 Display the maximum V value contained within the input frame.
37363 Expressed in range of [0-255].
37364 SATMIN
37366 Display the minimal saturation value contained within the input
37367 frame. Expressed in range of [0-~181.02].
37368 SATLOW
37370 Display the saturation value at the 10% percentile within the input
37371 frame. Expressed in range of [0-~181.02].
37372 SATAVG
37374 Display the average saturation value within the input frame.
37375 Expressed in range of [0-~181.02].
37376 SATHIGH
37378 Display the saturation value at the 90% percentile within the input
37379 frame. Expressed in range of [0-~181.02].
37380 SATMAX
37382 Display the maximum saturation value contained within the input
37383 frame. Expressed in range of [0-~181.02].
37384 HUEMED
37386 Display the median value for hue within the input frame. Expressed
37387 in range of [0-360].
37388 HUEAVG
37390 Display the average value for hue within the input frame. Expressed
37391 in range of [0-360].
37392 YDIF
37394 Display the average of sample value difference between all values
37395 of the Y plane in the current frame and corresponding values of the
37396 previous input frame. Expressed in range of [0-255].
37397 UDIF
37399 Display the average of sample value difference between all values
37400 of the U plane in the current frame and corresponding values of the
37401 previous input frame. Expressed in range of [0-255].
37402 VDIF
37404 Display the average of sample value difference between all values
37405 of the V plane in the current frame and corresponding values of the
37406 previous input frame. Expressed in range of [0-255].
37407 YBITDEPTH
37409 Display bit depth of Y plane in current frame. Expressed in range
37410 of [0-16].
37411 UBITDEPTH
37413 Display bit depth of U plane in current frame. Expressed in range
37414 of [0-16].
37415 VBITDEPTH
37417 Display bit depth of V plane in current frame. Expressed in range
37418 of [0-16].
37419 The filter accepts the following options:
37421 stat
37423 out stat specify an additional form of image analysis. out output
37424 video with the specified type of pixel highlighted.
37425 Both options accept the following values:
37427 tout
37429 Identify temporal outliers pixels. A temporal outlier is a
37430 pixel unlike the neighboring pixels of the same field. Examples
37431 of temporal outliers include the results of video dropouts,
37432 head clogs, or tape tracking issues.
37433 vrep
37435 Identify vertical line repetition. Vertical line repetition
37436 includes similar rows of pixels within a frame. In born-digital
37437 video vertical line repetition is common, but this pattern is
37438 uncommon in video digitized from an analog source. When it
37439 occurs in video that results from the digitization of an analog
37440 source it can indicate concealment from a dropout compensator.
37441 brng
37443 Identify pixels that fall outside of legal broadcast range.
37444 color, c
37446 Set the highlight color for the out option. The default color is
37447 yellow.
37448 Examples
37450 • Output data of various video metrics:
37452 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
37454 • Output specific data about the minimum and maximum values of the Y
37456 plane per frame:
37457 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
37459 • Playback video while highlighting pixels that are outside of
37461 broadcast range in red.
37462 ffplay example.mov -vf signalstats="out=brng:color=red"
37464 • Playback video with signalstats metadata drawn over the frame.
37466 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
37468 The contents of signalstat_drawtext.txt used in the command are:
37470 time %{pts:hms}
37472 Y (%{metadata:lavfi.signalstats.YMIN}-%{metadata:lavfi.signalstats.YMAX})
37473 U (%{metadata:lavfi.signalstats.UMIN}-%{metadata:lavfi.signalstats.UMAX})
37474 V (%{metadata:lavfi.signalstats.VMIN}-%{metadata:lavfi.signalstats.VMAX})
37475 saturation maximum: %{metadata:lavfi.signalstats.SATMAX}
37476 signature
37478 Calculates the MPEG-7 Video Signature. The filter can handle more than
37479 one input. In this case the matching between the inputs can be
37480 calculated additionally. The filter always passes through the first
37481 input. The signature of each stream can be written into a file.
37482 It accepts the following options:
37484 detectmode
37486 Enable or disable the matching process.
37487 Available values are:
37489 off Disable the calculation of a matching (default).
37491 full
37493 Calculate the matching for the whole video and output whether
37494 the whole video matches or only parts.
37495 fast
37497 Calculate only until a matching is found or the video ends.
37498 Should be faster in some cases.
37499 nb_inputs
37501 Set the number of inputs. The option value must be a non negative
37502 integer. Default value is 1.
37503 filename
37505 Set the path to which the output is written. If there is more than
37506 one input, the path must be a prototype, i.e. must contain %d or
37507 %0nd (where n is a positive integer), that will be replaced with
37508 the input number. If no filename is specified, no output will be
37509 written. This is the default.
37510 format
37512 Choose the output format.
37513 Available values are:
37515 binary
37517 Use the specified binary representation (default).
37518 xml Use the specified xml representation.
37520 th_d
37522 Set threshold to detect one word as similar. The option value must
37523 be an integer greater than zero. The default value is 9000.
37524 th_dc
37526 Set threshold to detect all words as similar. The option value must
37527 be an integer greater than zero. The default value is 60000.
37528 th_xh
37530 Set threshold to detect frames as similar. The option value must be
37531 an integer greater than zero. The default value is 116.
37532 th_di
37534 Set the minimum length of a sequence in frames to recognize it as
37535 matching sequence. The option value must be a non negative integer
37536 value. The default value is 0.
37537 th_it
37539 Set the minimum relation, that matching frames to all frames must
37540 have. The option value must be a double value between 0 and 1. The
37541 default value is 0.5.
37542 Examples
37544 • To calculate the signature of an input video and store it in
37546 signature.bin:
37547 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
37549 • To detect whether two videos match and store the signatures in XML
37551 format in signature0.xml and signature1.xml:
37552 ffmpeg -i input1.mkv -i input2.mkv -filter_complex "[0:v][1:v] signature=nb_inputs=2:detectmode=full:format=xml:filename=signature%d.xml" -map :v -f null -
37554 siti
37556 Calculate Spatial Info (SI) and Temporal Info (TI) scores for a video,
37557 as defined in ITU-T P.910: Subjective video quality assessment methods
37558 for multimedia applications. Available PDF at
37559 <https://www.itu.int/rec/T-REC-P.910-199909-S/en >.
37560 It accepts the following option:
37562 print_summary
37564 If set to 1, Summary statistics will be printed to the console.
37565 Default 0.
37566 Examples
37568 • To calculate SI/TI metrics and print summary:
37570 ffmpeg -i input.mp4 -vf siti=print_summary=1 -f null -
37572 smartblur
37574 Blur the input video without impacting the outlines.
37575 It accepts the following options:
37577 luma_radius, lr
37579 Set the luma radius. The option value must be a float number in the
37580 range [0.1,5.0] that specifies the variance of the gaussian filter
37581 used to blur the image (slower if larger). Default value is 1.0.
37582 luma_strength, ls
37584 Set the luma strength. The option value must be a float number in
37585 the range [-1.0,1.0] that configures the blurring. A value included
37586 in [0.0,1.0] will blur the image whereas a value included in
37587 [-1.0,0.0] will sharpen the image. Default value is 1.0.
37588 luma_threshold, lt
37590 Set the luma threshold used as a coefficient to determine whether a
37591 pixel should be blurred or not. The option value must be an integer
37592 in the range [-30,30]. A value of 0 will filter all the image, a
37593 value included in [0,30] will filter flat areas and a value
37594 included in [-30,0] will filter edges. Default value is 0.
37595 chroma_radius, cr
37597 Set the chroma radius. The option value must be a float number in
37598 the range [0.1,5.0] that specifies the variance of the gaussian
37599 filter used to blur the image (slower if larger). Default value is
37600 luma_radius.
37601 chroma_strength, cs
37603 Set the chroma strength. The option value must be a float number in
37604 the range [-1.0,1.0] that configures the blurring. A value included
37605 in [0.0,1.0] will blur the image whereas a value included in
37606 [-1.0,0.0] will sharpen the image. Default value is luma_strength.
37607 chroma_threshold, ct
37609 Set the chroma threshold used as a coefficient to determine whether
37610 a pixel should be blurred or not. The option value must be an
37611 integer in the range [-30,30]. A value of 0 will filter all the
37612 image, a value included in [0,30] will filter flat areas and a
37613 value included in [-30,0] will filter edges. Default value is
37614 luma_threshold.
37615 If a chroma option is not explicitly set, the corresponding luma value
37617 is set.
37618 sobel
37620 Apply sobel operator to input video stream.
37621 The filter accepts the following option:
37623 planes
37625 Set which planes will be processed, unprocessed planes will be
37626 copied. By default value 0xf, all planes will be processed.
37627 scale
37629 Set value which will be multiplied with filtered result.
37630 delta
37632 Set value which will be added to filtered result.
37633 Commands
37635 This filter supports the all above options as commands.
37637 spp
37639 Apply a simple postprocessing filter that compresses and decompresses
37640 the image at several (or - in the case of quality level 6 - all) shifts
37641 and average the results.
37642 The filter accepts the following options:
37644 quality
37646 Set quality. This option defines the number of levels for
37647 averaging. It accepts an integer in the range 0-6. If set to 0, the
37648 filter will have no effect. A value of 6 means the higher quality.
37649 For each increment of that value the speed drops by a factor of
37650 approximately 2. Default value is 3.
37651 qp Force a constant quantization parameter. If not set, the filter
37653 will use the QP from the video stream (if available).
37654 mode
37656 Set thresholding mode. Available modes are:
37657 hard
37659 Set hard thresholding (default).
37660 soft
37662 Set soft thresholding (better de-ringing effect, but likely
37663 blurrier).
37664 use_bframe_qp
37666 Enable the use of the QP from the B-Frames if set to 1. Using this
37667 option may cause flicker since the B-Frames have often larger QP.
37668 Default is 0 (not enabled).
37669 Commands
37671 This filter supports the following commands:
37673 quality, level
37675 Set quality level. The value "max" can be used to set the maximum
37676 level, currently 6.
37677 sr
37679 Scale the input by applying one of the super-resolution methods based
37680 on convolutional neural networks. Supported models:
37681 • Super-Resolution Convolutional Neural Network model (SRCNN). See
37683 <https://arxiv.org/abs/1501.00092>.
37684 • Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
37686 See <https://arxiv.org/abs/1609.05158>.
37687 Training scripts as well as scripts for model file (.pb) saving can be
37689 found at <https://github.com/XueweiMeng/sr/tree/sr_dnn_native>.
37690 Original repository is at
37691 <https://github.com/HighVoltageRocknRoll/sr.git>.
37692 Native model files (.model) can be generated from TensorFlow model
37694 files (.pb) by using tools/python/convert.py
37695 The filter accepts the following options:
37697 dnn_backend
37699 Specify which DNN backend to use for model loading and execution.
37700 This option accepts the following values:
37701 native
37703 Native implementation of DNN loading and execution.
37704 tensorflow
37706 TensorFlow backend. To enable this backend you need to install
37707 the TensorFlow for C library (see
37708 <https://www.tensorflow.org/install/lang_c>) and configure
37709 FFmpeg with "--enable-libtensorflow"
37710 Default value is native.
37712 model
37714 Set path to model file specifying network architecture and its
37715 parameters. Note that different backends use different file
37716 formats. TensorFlow backend can load files for both formats, while
37717 native backend can load files for only its format.
37718 scale_factor
37720 Set scale factor for SRCNN model. Allowed values are 2, 3 and 4.
37721 Default value is 2. Scale factor is necessary for SRCNN model,
37722 because it accepts input upscaled using bicubic upscaling with
37723 proper scale factor.
37724 To get full functionality (such as async execution), please use the
37726 dnn_processing filter.
37727 ssim
37729 Obtain the SSIM (Structural SImilarity Metric) between two input
37730 videos.
37731 This filter takes in input two input videos, the first input is
37733 considered the "main" source and is passed unchanged to the output. The
37734 second input is used as a "reference" video for computing the SSIM.
37735 Both video inputs must have the same resolution and pixel format for
37737 this filter to work correctly. Also it assumes that both inputs have
37738 the same number of frames, which are compared one by one.
37739 The filter stores the calculated SSIM of each frame.
37741 The description of the accepted parameters follows.
37743 stats_file, f
37745 If specified the filter will use the named file to save the SSIM of
37746 each individual frame. When filename equals "-" the data is sent to
37747 standard output.
37748 The file printed if stats_file is selected, contains a sequence of
37750 key/value pairs of the form key:value for each compared couple of
37751 frames.
37752 A description of each shown parameter follows:
37754 n sequential number of the input frame, starting from 1
37756 Y, U, V, R, G, B
37758 SSIM of the compared frames for the component specified by the
37759 suffix.
37760 All SSIM of the compared frames for the whole frame.
37762 dB Same as above but in dB representation.
37764 This filter also supports the framesync options.
37766 Examples
37768 • For example:
37770 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
37772 [main][ref] ssim="stats_file=stats.log" [out]
37773 On this example the input file being processed is compared with the
37775 reference file ref_movie.mpg. The SSIM of each individual frame is
37776 stored in stats.log.
37777 • Another example with both psnr and ssim at same time:
37779 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
37781 • Another example with different containers:
37783 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]ssim" -f null -
37785 stereo3d
37787 Convert between different stereoscopic image formats.
37788 The filters accept the following options:
37790 in Set stereoscopic image format of input.
37792 Available values for input image formats are:
37794 sbsl
37796 side by side parallel (left eye left, right eye right)
37797 sbsr
37799 side by side crosseye (right eye left, left eye right)
37800 sbs2l
37802 side by side parallel with half width resolution (left eye
37803 left, right eye right)
37804 sbs2r
37806 side by side crosseye with half width resolution (right eye
37807 left, left eye right)
37808 abl
37810 tbl above-below (left eye above, right eye below)
37811 abr
37813 tbr above-below (right eye above, left eye below)
37814 ab2l
37816 tb2l
37817 above-below with half height resolution (left eye above, right
37818 eye below)
37819 ab2r
37821 tb2r
37822 above-below with half height resolution (right eye above, left
37823 eye below)
37824 al alternating frames (left eye first, right eye second)
37826 ar alternating frames (right eye first, left eye second)
37828 irl interleaved rows (left eye has top row, right eye starts on
37830 next row)
37831 irr interleaved rows (right eye has top row, left eye starts on
37833 next row)
37834 icl interleaved columns, left eye first
37836 icr interleaved columns, right eye first
37838 Default value is sbsl.
37840 out Set stereoscopic image format of output.
37842 sbsl
37844 side by side parallel (left eye left, right eye right)
37845 sbsr
37847 side by side crosseye (right eye left, left eye right)
37848 sbs2l
37850 side by side parallel with half width resolution (left eye
37851 left, right eye right)
37852 sbs2r
37854 side by side crosseye with half width resolution (right eye
37855 left, left eye right)
37856 abl
37858 tbl above-below (left eye above, right eye below)
37859 abr
37861 tbr above-below (right eye above, left eye below)
37862 ab2l
37864 tb2l
37865 above-below with half height resolution (left eye above, right
37866 eye below)
37867 ab2r
37869 tb2r
37870 above-below with half height resolution (right eye above, left
37871 eye below)
37872 al alternating frames (left eye first, right eye second)
37874 ar alternating frames (right eye first, left eye second)
37876 irl interleaved rows (left eye has top row, right eye starts on
37878 next row)
37879 irr interleaved rows (right eye has top row, left eye starts on
37881 next row)
37882 arbg
37884 anaglyph red/blue gray (red filter on left eye, blue filter on
37885 right eye)
37886 argg
37888 anaglyph red/green gray (red filter on left eye, green filter
37889 on right eye)
37890 arcg
37892 anaglyph red/cyan gray (red filter on left eye, cyan filter on
37893 right eye)
37894 arch
37896 anaglyph red/cyan half colored (red filter on left eye, cyan
37897 filter on right eye)
37898 arcc
37900 anaglyph red/cyan color (red filter on left eye, cyan filter on
37901 right eye)
37902 arcd
37904 anaglyph red/cyan color optimized with the least squares
37905 projection of dubois (red filter on left eye, cyan filter on
37906 right eye)
37907 agmg
37909 anaglyph green/magenta gray (green filter on left eye, magenta
37910 filter on right eye)
37911 agmh
37913 anaglyph green/magenta half colored (green filter on left eye,
37914 magenta filter on right eye)
37915 agmc
37917 anaglyph green/magenta colored (green filter on left eye,
37918 magenta filter on right eye)
37919 agmd
37921 anaglyph green/magenta color optimized with the least squares
37922 projection of dubois (green filter on left eye, magenta filter
37923 on right eye)
37924 aybg
37926 anaglyph yellow/blue gray (yellow filter on left eye, blue
37927 filter on right eye)
37928 aybh
37930 anaglyph yellow/blue half colored (yellow filter on left eye,
37931 blue filter on right eye)
37932 aybc
37934 anaglyph yellow/blue colored (yellow filter on left eye, blue
37935 filter on right eye)
37936 aybd
37938 anaglyph yellow/blue color optimized with the least squares
37939 projection of dubois (yellow filter on left eye, blue filter on
37940 right eye)
37941 ml mono output (left eye only)
37943 mr mono output (right eye only)
37945 chl checkerboard, left eye first
37947 chr checkerboard, right eye first
37949 icl interleaved columns, left eye first
37951 icr interleaved columns, right eye first
37953 hdmi
37955 HDMI frame pack
37956 Default value is arcd.
37958 Examples
37960 • Convert input video from side by side parallel to anaglyph
37962 yellow/blue dubois:
37963 stereo3d=sbsl:aybd
37965 • Convert input video from above below (left eye above, right eye
37967 below) to side by side crosseye.
37968 stereo3d=abl:sbsr
37970 streamselect, astreamselect
37972 Select video or audio streams.
37973 The filter accepts the following options:
37975 inputs
37977 Set number of inputs. Default is 2.
37978 map Set input indexes to remap to outputs.
37980 Commands
37982 The "streamselect" and "astreamselect" filter supports the following
37984 commands:
37985 map Set input indexes to remap to outputs.
37987 Examples
37989 • Select first 5 seconds 1st stream and rest of time 2nd stream:
37991 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
37993 • Same as above, but for audio:
37995 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
37997 subtitles
37999 Draw subtitles on top of input video using the libass library.
38000 To enable compilation of this filter you need to configure FFmpeg with
38002 "--enable-libass". This filter also requires a build with libavcodec
38003 and libavformat to convert the passed subtitles file to ASS (Advanced
38004 Substation Alpha) subtitles format.
38005 The filter accepts the following options:
38007 filename, f
38009 Set the filename of the subtitle file to read. It must be
38010 specified.
38011 original_size
38013 Specify the size of the original video, the video for which the ASS
38014 file was composed. For the syntax of this option, check the "Video
38015 size" section in the ffmpeg-utils manual. Due to a misdesign in
38016 ASS aspect ratio arithmetic, this is necessary to correctly scale
38017 the fonts if the aspect ratio has been changed.
38018 fontsdir
38020 Set a directory path containing fonts that can be used by the
38021 filter. These fonts will be used in addition to whatever the font
38022 provider uses.
38023 alpha
38025 Process alpha channel, by default alpha channel is untouched.
38026 charenc
38028 Set subtitles input character encoding. "subtitles" filter only.
38029 Only useful if not UTF-8.
38030 stream_index, si
38032 Set subtitles stream index. "subtitles" filter only.
38033 force_style
38035 Override default style or script info parameters of the subtitles.
38036 It accepts a string containing ASS style format "KEY=VALUE" couples
38037 separated by ",".
38038 If the first key is not specified, it is assumed that the first value
38040 specifies the filename.
38041 For example, to render the file sub.srt on top of the input video, use
38043 the command:
38044 subtitles=sub.srt
38046 which is equivalent to:
38048 subtitles=filename=sub.srt
38050 To render the default subtitles stream from file video.mkv, use:
38052 subtitles=video.mkv
38054 To render the second subtitles stream from that file, use:
38056 subtitles=video.mkv:si=1
38058 To make the subtitles stream from sub.srt appear in 80% transparent
38060 blue "DejaVu Serif", use:
38061 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
38063 super2xsai
38065 Scale the input by 2x and smooth using the Super2xSaI (Scale and
38066 Interpolate) pixel art scaling algorithm.
38067 Useful for enlarging pixel art images without reducing sharpness.
38069 swaprect
38071 Swap two rectangular objects in video.
38072 This filter accepts the following options:
38074 w Set object width.
38076 h Set object height.
38078 x1 Set 1st rect x coordinate.
38080 y1 Set 1st rect y coordinate.
38082 x2 Set 2nd rect x coordinate.
38084 y2 Set 2nd rect y coordinate.
38086 All expressions are evaluated once for each frame.
38088 The all options are expressions containing the following constants:
38090 w
38092 h The input width and height.
38093 a same as w / h
38095 sar input sample aspect ratio
38097 dar input display aspect ratio, it is the same as (w / h) * sar
38099 n The number of the input frame, starting from 0.
38101 t The timestamp expressed in seconds. It's NAN if the input timestamp
38103 is unknown.
38104 pos the position in the file of the input frame, NAN if unknown
38106 Commands
38108 This filter supports the all above options as commands.
38110 swapuv
38112 Swap U & V plane.
38113 tblend
38115 Blend successive video frames.
38116 See blend
38118 telecine
38120 Apply telecine process to the video.
38121 This filter accepts the following options:
38123 first_field
38125 top, t
38126 top field first
38127 bottom, b
38129 bottom field first The default value is "top".
38130 pattern
38132 A string of numbers representing the pulldown pattern you wish to
38133 apply. The default value is 23.
38134 Some typical patterns:
38136 NTSC output (30i):
38138 27.5p: 32222
38139 24p: 23 (classic)
38140 24p: 2332 (preferred)
38141 20p: 33
38142 18p: 334
38143 16p: 3444
38144 PAL output (25i):
38146 27.5p: 12222
38147 24p: 222222222223 ("Euro pulldown")
38148 16.67p: 33
38149 16p: 33333334
38150 thistogram
38152 Compute and draw a color distribution histogram for the input video
38153 across time.
38154 Unlike histogram video filter which only shows histogram of single
38156 input frame at certain time, this filter shows also past histograms of
38157 number of frames defined by "width" option.
38158 The computed histogram is a representation of the color component
38160 distribution in an image.
38161 The filter accepts the following options:
38163 width, w
38165 Set width of single color component output. Default value is 0.
38166 Value of 0 means width will be picked from input video. This also
38167 set number of passed histograms to keep. Allowed range is [0,
38168 8192].
38169 display_mode, d
38171 Set display mode. It accepts the following values:
38172 stack
38174 Per color component graphs are placed below each other.
38175 parade
38177 Per color component graphs are placed side by side.
38178 overlay
38180 Presents information identical to that in the "parade", except
38181 that the graphs representing color components are superimposed
38182 directly over one another.
38183 Default is "stack".
38185 levels_mode, m
38187 Set mode. Can be either "linear", or "logarithmic". Default is
38188 "linear".
38189 components, c
38191 Set what color components to display. Default is 7.
38192 bgopacity, b
38194 Set background opacity. Default is 0.9.
38195 envelope, e
38197 Show envelope. Default is disabled.
38198 ecolor, ec
38200 Set envelope color. Default is "gold".
38201 slide
38203 Set slide mode.
38204 Available values for slide is:
38206 frame
38208 Draw new frame when right border is reached.
38209 replace
38211 Replace old columns with new ones.
38212 scroll
38214 Scroll from right to left.
38215 rscroll
38217 Scroll from left to right.
38218 picture
38220 Draw single picture.
38221 Default is "replace".
38223 threshold
38225 Apply threshold effect to video stream.
38226 This filter needs four video streams to perform thresholding. First
38228 stream is stream we are filtering. Second stream is holding threshold
38229 values, third stream is holding min values, and last, fourth stream is
38230 holding max values.
38231 The filter accepts the following option:
38233 planes
38235 Set which planes will be processed, unprocessed planes will be
38236 copied. By default value 0xf, all planes will be processed.
38237 For example if first stream pixel's component value is less then
38239 threshold value of pixel component from 2nd threshold stream, third
38240 stream value will picked, otherwise fourth stream pixel component value
38241 will be picked.
38242 Using color source filter one can perform various types of
38244 thresholding:
38245 Commands
38247 This filter supports the all options as commands.
38249 Examples
38251 • Binary threshold, using gray color as threshold:
38253 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
38255 • Inverted binary threshold, using gray color as threshold:
38257 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
38259 • Truncate binary threshold, using gray color as threshold:
38261 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
38263 • Threshold to zero, using gray color as threshold:
38265 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
38267 • Inverted threshold to zero, using gray color as threshold:
38269 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
38271 thumbnail
38273 Select the most representative frame in a given sequence of consecutive
38274 frames.
38275 The filter accepts the following options:
38277 n Set the frames batch size to analyze; in a set of n frames, the
38279 filter will pick one of them, and then handle the next batch of n
38280 frames until the end. Default is 100.
38281 log Set the log level to display picked frame stats. Default is
38283 "info".
38284 Since the filter keeps track of the whole frames sequence, a bigger n
38286 value will result in a higher memory usage, so a high value is not
38287 recommended.
38288 Examples
38290 • Extract one picture each 50 frames:
38292 thumbnail=50
38294 • Complete example of a thumbnail creation with ffmpeg:
38296 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
38298 tile
38300 Tile several successive frames together.
38301 The untile filter can do the reverse.
38303 The filter accepts the following options:
38305 layout
38307 Set the grid size in the form "COLUMNSxROWS". Range is upto
38308 UINT_MAX cells. Default is "6x5".
38309 nb_frames
38311 Set the maximum number of frames to render in the given area. It
38312 must be less than or equal to wxh. The default value is 0, meaning
38313 all the area will be used.
38314 margin
38316 Set the outer border margin in pixels. Range is 0 to 1024. Default
38317 is 0.
38318 padding
38320 Set the inner border thickness (i.e. the number of pixels between
38321 frames). For more advanced padding options (such as having
38322 different values for the edges), refer to the pad video filter.
38323 Range is 0 to 1024. Default is 0.
38324 color
38326 Specify the color of the unused area. For the syntax of this
38327 option, check the "Color" section in the ffmpeg-utils manual. The
38328 default value of color is "black".
38329 overlap
38331 Set the number of frames to overlap when tiling several successive
38332 frames together. The value must be between 0 and nb_frames - 1.
38333 Default is 0.
38334 init_padding
38336 Set the number of frames to initially be empty before displaying
38337 first output frame. This controls how soon will one get first
38338 output frame. The value must be between 0 and nb_frames - 1.
38339 Default is 0.
38340 Examples
38342 • Produce 8x8 PNG tiles of all keyframes (-skip_frame nokey) in a
38344 movie:
38345 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
38347 The -vsync 0 is necessary to prevent ffmpeg from duplicating each
38349 output frame to accommodate the originally detected frame rate.
38350 • Display 5 pictures in an area of "3x2" frames, with 7 pixels
38352 between them, and 2 pixels of initial margin, using mixed flat and
38353 named options:
38354 tile=3x2:nb_frames=5:padding=7:margin=2
38356 tinterlace
38358 Perform various types of temporal field interlacing.
38359 Frames are counted starting from 1, so the first input frame is
38361 considered odd.
38362 The filter accepts the following options:
38364 mode
38366 Specify the mode of the interlacing. This option can also be
38367 specified as a value alone. See below for a list of values for this
38368 option.
38369 Available values are:
38371 merge, 0
38373 Move odd frames into the upper field, even into the lower
38374 field, generating a double height frame at half frame rate.
38375 ------> time
38377 Input:
38378 Frame 1 Frame 2 Frame 3 Frame 4
38379 11111 22222 33333 44444
38381 11111 22222 33333 44444
38382 11111 22222 33333 44444
38383 11111 22222 33333 44444
38384 Output:
38386 11111 33333
38387 22222 44444
38388 11111 33333
38389 22222 44444
38390 11111 33333
38391 22222 44444
38392 11111 33333
38393 22222 44444
38394 drop_even, 1
38396 Only output odd frames, even frames are dropped, generating a
38397 frame with unchanged height at half frame rate.
38398 ------> time
38400 Input:
38401 Frame 1 Frame 2 Frame 3 Frame 4
38402 11111 22222 33333 44444
38404 11111 22222 33333 44444
38405 11111 22222 33333 44444
38406 11111 22222 33333 44444
38407 Output:
38409 11111 33333
38410 11111 33333
38411 11111 33333
38412 11111 33333
38413 drop_odd, 2
38415 Only output even frames, odd frames are dropped, generating a
38416 frame with unchanged height at half frame rate.
38417 ------> time
38419 Input:
38420 Frame 1 Frame 2 Frame 3 Frame 4
38421 11111 22222 33333 44444
38423 11111 22222 33333 44444
38424 11111 22222 33333 44444
38425 11111 22222 33333 44444
38426 Output:
38428 22222 44444
38429 22222 44444
38430 22222 44444
38431 22222 44444
38432 pad, 3
38434 Expand each frame to full height, but pad alternate lines with
38435 black, generating a frame with double height at the same input
38436 frame rate.
38437 ------> time
38439 Input:
38440 Frame 1 Frame 2 Frame 3 Frame 4
38441 11111 22222 33333 44444
38443 11111 22222 33333 44444
38444 11111 22222 33333 44444
38445 11111 22222 33333 44444
38446 Output:
38448 11111 ..... 33333 .....
38449 ..... 22222 ..... 44444
38450 11111 ..... 33333 .....
38451 ..... 22222 ..... 44444
38452 11111 ..... 33333 .....
38453 ..... 22222 ..... 44444
38454 11111 ..... 33333 .....
38455 ..... 22222 ..... 44444
38456 interleave_top, 4
38458 Interleave the upper field from odd frames with the lower field
38459 from even frames, generating a frame with unchanged height at
38460 half frame rate.
38461 ------> time
38463 Input:
38464 Frame 1 Frame 2 Frame 3 Frame 4
38465 11111<- 22222 33333<- 44444
38467 11111 22222<- 33333 44444<-
38468 11111<- 22222 33333<- 44444
38469 11111 22222<- 33333 44444<-
38470 Output:
38472 11111 33333
38473 22222 44444
38474 11111 33333
38475 22222 44444
38476 interleave_bottom, 5
38478 Interleave the lower field from odd frames with the upper field
38479 from even frames, generating a frame with unchanged height at
38480 half frame rate.
38481 ------> time
38483 Input:
38484 Frame 1 Frame 2 Frame 3 Frame 4
38485 11111 22222<- 33333 44444<-
38487 11111<- 22222 33333<- 44444
38488 11111 22222<- 33333 44444<-
38489 11111<- 22222 33333<- 44444
38490 Output:
38492 22222 44444
38493 11111 33333
38494 22222 44444
38495 11111 33333
38496 interlacex2, 6
38498 Double frame rate with unchanged height. Frames are inserted
38499 each containing the second temporal field from the previous
38500 input frame and the first temporal field from the next input
38501 frame. This mode relies on the top_field_first flag. Useful for
38502 interlaced video displays with no field synchronisation.
38503 ------> time
38505 Input:
38506 Frame 1 Frame 2 Frame 3 Frame 4
38507 11111 22222 33333 44444
38509 11111 22222 33333 44444
38510 11111 22222 33333 44444
38511 11111 22222 33333 44444
38512 Output:
38514 11111 22222 22222 33333 33333 44444 44444
38515 11111 11111 22222 22222 33333 33333 44444
38516 11111 22222 22222 33333 33333 44444 44444
38517 11111 11111 22222 22222 33333 33333 44444
38518 mergex2, 7
38520 Move odd frames into the upper field, even into the lower
38521 field, generating a double height frame at same frame rate.
38522 ------> time
38524 Input:
38525 Frame 1 Frame 2 Frame 3 Frame 4
38526 11111 22222 33333 44444
38528 11111 22222 33333 44444
38529 11111 22222 33333 44444
38530 11111 22222 33333 44444
38531 Output:
38533 11111 33333 33333 55555
38534 22222 22222 44444 44444
38535 11111 33333 33333 55555
38536 22222 22222 44444 44444
38537 11111 33333 33333 55555
38538 22222 22222 44444 44444
38539 11111 33333 33333 55555
38540 22222 22222 44444 44444
38541 Numeric values are deprecated but are accepted for backward
38543 compatibility reasons.
38544 Default mode is "merge".
38546 flags
38548 Specify flags influencing the filter process.
38549 Available value for flags is:
38551 low_pass_filter, vlpf
38553 Enable linear vertical low-pass filtering in the filter.
38554 Vertical low-pass filtering is required when creating an
38555 interlaced destination from a progressive source which contains
38556 high-frequency vertical detail. Filtering will reduce interlace
38557 'twitter' and Moire patterning.
38558 complex_filter, cvlpf
38560 Enable complex vertical low-pass filtering. This will slightly
38561 less reduce interlace 'twitter' and Moire patterning but better
38562 retain detail and subjective sharpness impression.
38563 bypass_il
38565 Bypass already interlaced frames, only adjust the frame rate.
38566 Vertical low-pass filtering and bypassing already interlaced frames
38568 can only be enabled for mode interleave_top and interleave_bottom.
38569 tmedian
38571 Pick median pixels from several successive input video frames.
38572 The filter accepts the following options:
38574 radius
38576 Set radius of median filter. Default is 1. Allowed range is from 1
38577 to 127.
38578 planes
38580 Set which planes to filter. Default value is 15, by which all
38581 planes are processed.
38582 percentile
38584 Set median percentile. Default value is 0.5. Default value of 0.5
38585 will pick always median values, while 0 will pick minimum values,
38586 and 1 maximum values.
38587 Commands
38589 This filter supports all above options as commands, excluding option
38591 "radius".
38592 tmidequalizer
38594 Apply Temporal Midway Video Equalization effect.
38595 Midway Video Equalization adjusts a sequence of video frames to have
38597 the same histograms, while maintaining their dynamics as much as
38598 possible. It's useful for e.g. matching exposures from a video frames
38599 sequence.
38600 This filter accepts the following option:
38602 radius
38604 Set filtering radius. Default is 5. Allowed range is from 1 to 127.
38605 sigma
38607 Set filtering sigma. Default is 0.5. This controls strength of
38608 filtering. Setting this option to 0 effectively does nothing.
38609 planes
38611 Set which planes to process. Default is 15, which is all available
38612 planes.
38613 tmix
38615 Mix successive video frames.
38616 A description of the accepted options follows.
38618 frames
38620 The number of successive frames to mix. If unspecified, it defaults
38621 to 3.
38622 weights
38624 Specify weight of each input video frame. Each weight is separated
38625 by space. If number of weights is smaller than number of frames
38626 last specified weight will be used for all remaining unset weights.
38627 scale
38629 Specify scale, if it is set it will be multiplied with sum of each
38630 weight multiplied with pixel values to give final destination pixel
38631 value. By default scale is auto scaled to sum of weights.
38632 planes
38634 Set which planes to filter. Default is all. Allowed range is from 0
38635 to 15.
38636 Examples
38638 • Average 7 successive frames:
38640 tmix=frames=7:weights="1 1 1 1 1 1 1"
38642 • Apply simple temporal convolution:
38644 tmix=frames=3:weights="-1 3 -1"
38646 • Similar as above but only showing temporal differences:
38648 tmix=frames=3:weights="-1 2 -1":scale=1
38650 Commands
38652 This filter supports the following commands:
38654 weights
38656 scale
38657 planes
38658 Syntax is same as option with same name.
38659 tonemap
38661 Tone map colors from different dynamic ranges.
38662 This filter expects data in single precision floating point, as it
38664 needs to operate on (and can output) out-of-range values. Another
38665 filter, such as zscale, is needed to convert the resulting frame to a
38666 usable format.
38667 The tonemapping algorithms implemented only work on linear light, so
38669 input data should be linearized beforehand (and possibly correctly
38670 tagged).
38671 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
38673 Options
38675 The filter accepts the following options.
38677 tonemap
38679 Set the tone map algorithm to use.
38680 Possible values are:
38682 none
38684 Do not apply any tone map, only desaturate overbright pixels.
38685 clip
38687 Hard-clip any out-of-range values. Use it for perfect color
38688 accuracy for in-range values, while distorting out-of-range
38689 values.
38690 linear
38692 Stretch the entire reference gamut to a linear multiple of the
38693 display.
38694 gamma
38696 Fit a logarithmic transfer between the tone curves.
38697 reinhard
38699 Preserve overall image brightness with a simple curve, using
38700 nonlinear contrast, which results in flattening details and
38701 degrading color accuracy.
38702 hable
38704 Preserve both dark and bright details better than reinhard, at
38705 the cost of slightly darkening everything. Use it when detail
38706 preservation is more important than color and brightness
38707 accuracy.
38708 mobius
38710 Smoothly map out-of-range values, while retaining contrast and
38711 colors for in-range material as much as possible. Use it when
38712 color accuracy is more important than detail preservation.
38713 Default is none.
38715 param
38717 Tune the tone mapping algorithm.
38718 This affects the following algorithms:
38720 none
38722 Ignored.
38723 linear
38725 Specifies the scale factor to use while stretching. Default to
38726 1.0.
38727 gamma
38729 Specifies the exponent of the function. Default to 1.8.
38730 clip
38732 Specify an extra linear coefficient to multiply into the signal
38733 before clipping. Default to 1.0.
38734 reinhard
38736 Specify the local contrast coefficient at the display peak.
38737 Default to 0.5, which means that in-gamut values will be about
38738 half as bright as when clipping.
38739 hable
38741 Ignored.
38742 mobius
38744 Specify the transition point from linear to mobius transform.
38745 Every value below this point is guaranteed to be mapped 1:1.
38746 The higher the value, the more accurate the result will be, at
38747 the cost of losing bright details. Default to 0.3, which due
38748 to the steep initial slope still preserves in-range colors
38749 fairly accurately.
38750 desat
38752 Apply desaturation for highlights that exceed this level of
38753 brightness. The higher the parameter, the more color information
38754 will be preserved. This setting helps prevent unnaturally blown-out
38755 colors for super-highlights, by (smoothly) turning into white
38756 instead. This makes images feel more natural, at the cost of
38757 reducing information about out-of-range colors.
38758 The default of 2.0 is somewhat conservative and will mostly just
38760 apply to skies or directly sunlit surfaces. A setting of 0.0
38761 disables this option.
38762 This option works only if the input frame has a supported color
38764 tag.
38765 peak
38767 Override signal/nominal/reference peak with this value. Useful when
38768 the embedded peak information in display metadata is not reliable
38769 or when tone mapping from a lower range to a higher range.
38770 tpad
38772 Temporarily pad video frames.
38773 The filter accepts the following options:
38775 start
38777 Specify number of delay frames before input video stream. Default
38778 is 0.
38779 stop
38781 Specify number of padding frames after input video stream. Set to
38782 -1 to pad indefinitely. Default is 0.
38783 start_mode
38785 Set kind of frames added to beginning of stream. Can be either add
38786 or clone. With add frames of solid-color are added. With clone
38787 frames are clones of first frame. Default is add.
38788 stop_mode
38790 Set kind of frames added to end of stream. Can be either add or
38791 clone. With add frames of solid-color are added. With clone
38792 frames are clones of last frame. Default is add.
38793 start_duration, stop_duration
38795 Specify the duration of the start/stop delay. See the Time duration
38796 section in the ffmpeg-utils(1) manual for the accepted syntax.
38797 These options override start and stop. Default is 0.
38798 color
38800 Specify the color of the padded area. For the syntax of this
38801 option, check the "Color" section in the ffmpeg-utils manual.
38802 The default value of color is "black".
38804 transpose
38806 Transpose rows with columns in the input video and optionally flip it.
38807 It accepts the following parameters:
38809 dir Specify the transposition direction.
38811 Can assume the following values:
38813 0, 4, cclock_flip
38815 Rotate by 90 degrees counterclockwise and vertically flip
38816 (default), that is:
38817 L.R L.l
38819 . . -> . .
38820 l.r R.r
38821 1, 5, clock
38823 Rotate by 90 degrees clockwise, that is:
38824 L.R l.L
38826 . . -> . .
38827 l.r r.R
38828 2, 6, cclock
38830 Rotate by 90 degrees counterclockwise, that is:
38831 L.R R.r
38833 . . -> . .
38834 l.r L.l
38835 3, 7, clock_flip
38837 Rotate by 90 degrees clockwise and vertically flip, that is:
38838 L.R r.R
38840 . . -> . .
38841 l.r l.L
38842 For values between 4-7, the transposition is only done if the input
38844 video geometry is portrait and not landscape. These values are
38845 deprecated, the "passthrough" option should be used instead.
38846 Numerical values are deprecated, and should be dropped in favor of
38848 symbolic constants.
38849 passthrough
38851 Do not apply the transposition if the input geometry matches the
38852 one specified by the specified value. It accepts the following
38853 values:
38854 none
38856 Always apply transposition.
38857 portrait
38859 Preserve portrait geometry (when height >= width).
38860 landscape
38862 Preserve landscape geometry (when width >= height).
38863 Default value is "none".
38865 For example to rotate by 90 degrees clockwise and preserve portrait
38867 layout:
38868 transpose=dir=1:passthrough=portrait
38870 The command above can also be specified as:
38872 transpose=1:portrait
38874 transpose_npp
38876 Transpose rows with columns in the input video and optionally flip it.
38877 For more in depth examples see the transpose video filter, which shares
38878 mostly the same options.
38879 It accepts the following parameters:
38881 dir Specify the transposition direction.
38883 Can assume the following values:
38885 cclock_flip
38887 Rotate by 90 degrees counterclockwise and vertically flip.
38888 (default)
38889 clock
38891 Rotate by 90 degrees clockwise.
38892 cclock
38894 Rotate by 90 degrees counterclockwise.
38895 clock_flip
38897 Rotate by 90 degrees clockwise and vertically flip.
38898 passthrough
38900 Do not apply the transposition if the input geometry matches the
38901 one specified by the specified value. It accepts the following
38902 values:
38903 none
38905 Always apply transposition. (default)
38906 portrait
38908 Preserve portrait geometry (when height >= width).
38909 landscape
38911 Preserve landscape geometry (when width >= height).
38912 trim
38914 Trim the input so that the output contains one continuous subpart of
38915 the input.
38916 It accepts the following parameters:
38918 start
38920 Specify the time of the start of the kept section, i.e. the frame
38921 with the timestamp start will be the first frame in the output.
38922 end Specify the time of the first frame that will be dropped, i.e. the
38924 frame immediately preceding the one with the timestamp end will be
38925 the last frame in the output.
38926 start_pts
38928 This is the same as start, except this option sets the start
38929 timestamp in timebase units instead of seconds.
38930 end_pts
38932 This is the same as end, except this option sets the end timestamp
38933 in timebase units instead of seconds.
38934 duration
38936 The maximum duration of the output in seconds.
38937 start_frame
38939 The number of the first frame that should be passed to the output.
38940 end_frame
38942 The number of the first frame that should be dropped.
38943 start, end, and duration are expressed as time duration specifications;
38945 see the Time duration section in the ffmpeg-utils(1) manual for the
38946 accepted syntax.
38947 Note that the first two sets of the start/end options and the duration
38949 option look at the frame timestamp, while the _frame variants simply
38950 count the frames that pass through the filter. Also note that this
38951 filter does not modify the timestamps. If you wish for the output
38952 timestamps to start at zero, insert a setpts filter after the trim
38953 filter.
38954 If multiple start or end options are set, this filter tries to be
38956 greedy and keep all the frames that match at least one of the specified
38957 constraints. To keep only the part that matches all the constraints at
38958 once, chain multiple trim filters.
38959 The defaults are such that all the input is kept. So it is possible to
38961 set e.g. just the end values to keep everything before the specified
38962 time.
38963 Examples:
38965 • Drop everything except the second minute of input:
38967 ffmpeg -i INPUT -vf trim=60:120
38969 • Keep only the first second:
38971 ffmpeg -i INPUT -vf trim=duration=1
38973 unpremultiply
38975 Apply alpha unpremultiply effect to input video stream using first
38976 plane of second stream as alpha.
38977 Both streams must have same dimensions and same pixel format.
38979 The filter accepts the following option:
38981 planes
38983 Set which planes will be processed, unprocessed planes will be
38984 copied. By default value 0xf, all planes will be processed.
38985 If the format has 1 or 2 components, then luma is bit 0. If the
38987 format has 3 or 4 components: for RGB formats bit 0 is green, bit 1
38988 is blue and bit 2 is red; for YUV formats bit 0 is luma, bit 1 is
38989 chroma-U and bit 2 is chroma-V. If present, the alpha channel is
38990 always the last bit.
38991 inplace
38993 Do not require 2nd input for processing, instead use alpha plane
38994 from input stream.
38995 unsharp
38997 Sharpen or blur the input video.
38998 It accepts the following parameters:
39000 luma_msize_x, lx
39002 Set the luma matrix horizontal size. It must be an odd integer
39003 between 3 and 23. The default value is 5.
39004 luma_msize_y, ly
39006 Set the luma matrix vertical size. It must be an odd integer
39007 between 3 and 23. The default value is 5.
39008 luma_amount, la
39010 Set the luma effect strength. It must be a floating point number,
39011 reasonable values lay between -1.5 and 1.5.
39012 Negative values will blur the input video, while positive values
39014 will sharpen it, a value of zero will disable the effect.
39015 Default value is 1.0.
39017 chroma_msize_x, cx
39019 Set the chroma matrix horizontal size. It must be an odd integer
39020 between 3 and 23. The default value is 5.
39021 chroma_msize_y, cy
39023 Set the chroma matrix vertical size. It must be an odd integer
39024 between 3 and 23. The default value is 5.
39025 chroma_amount, ca
39027 Set the chroma effect strength. It must be a floating point number,
39028 reasonable values lay between -1.5 and 1.5.
39029 Negative values will blur the input video, while positive values
39031 will sharpen it, a value of zero will disable the effect.
39032 Default value is 0.0.
39034 alpha_msize_x, ax
39036 Set the alpha matrix horizontal size. It must be an odd integer
39037 between 3 and 23. The default value is 5.
39038 alpha_msize_y, ay
39040 Set the alpha matrix vertical size. It must be an odd integer
39041 between 3 and 23. The default value is 5.
39042 alpha_amount, aa
39044 Set the alpha effect strength. It must be a floating point number,
39045 reasonable values lay between -1.5 and 1.5.
39046 Negative values will blur the input video, while positive values
39048 will sharpen it, a value of zero will disable the effect.
39049 Default value is 0.0.
39051 All parameters are optional and default to the equivalent of the string
39053 '5:5:1.0:5:5:0.0'.
39054 Examples
39056 • Apply strong luma sharpen effect:
39058 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
39060 • Apply a strong blur of both luma and chroma parameters:
39062 unsharp=7:7:-2:7:7:-2
39064 untile
39066 Decompose a video made of tiled images into the individual images.
39067 The frame rate of the output video is the frame rate of the input video
39069 multiplied by the number of tiles.
39070 This filter does the reverse of tile.
39072 The filter accepts the following options:
39074 layout
39076 Set the grid size (i.e. the number of lines and columns). For the
39077 syntax of this option, check the "Video size" section in the
39078 ffmpeg-utils manual.
39079 Examples
39081 • Produce a 1-second video from a still image file made of 25 frames
39083 stacked vertically, like an analogic film reel:
39084 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
39086 uspp
39088 Apply ultra slow/simple postprocessing filter that compresses and
39089 decompresses the image at several (or - in the case of quality level 8
39090 - all) shifts and average the results.
39091 The way this differs from the behavior of spp is that uspp actually
39093 encodes & decodes each case with libavcodec Snow, whereas spp uses a
39094 simplified intra only 8x8 DCT similar to MJPEG.
39095 This filter is only available in ffmpeg version 4.4 or earlier.
39097 The filter accepts the following options:
39099 quality
39101 Set quality. This option defines the number of levels for
39102 averaging. It accepts an integer in the range 0-8. If set to 0, the
39103 filter will have no effect. A value of 8 means the higher quality.
39104 For each increment of that value the speed drops by a factor of
39105 approximately 2. Default value is 3.
39106 qp Force a constant quantization parameter. If not set, the filter
39108 will use the QP from the video stream (if available).
39109 v360
39111 Convert 360 videos between various formats.
39112 The filter accepts the following options:
39114 input
39116 output
39117 Set format of the input/output video.
39118 Available formats:
39120 e
39122 equirect
39123 Equirectangular projection.
39124 c3x2
39126 c6x1
39127 c1x6
39128 Cubemap with 3x2/6x1/1x6 layout.
39129 Format specific options:
39131 in_pad
39133 out_pad
39134 Set padding proportion for the input/output cubemap. Values
39135 in decimals.
39136 Example values:
39138 0 No padding.
39140 0.01
39142 1% of face is padding. For example, with 1920x1280
39143 resolution face size would be 640x640 and padding would
39144 be 3 pixels from each side. (640 * 0.01 = 6 pixels)
39145 Default value is @samp{0}. Maximum value is @samp{0.1}.
39147 fin_pad
39149 fout_pad
39150 Set fixed padding for the input/output cubemap. Values in
39151 pixels.
39152 Default value is @samp{0}. If greater than zero it
39154 overrides other padding options.
39155 in_forder
39157 out_forder
39158 Set order of faces for the input/output cubemap. Choose one
39159 direction for each position.
39160 Designation of directions:
39162 r right
39164 l left
39166 u up
39168 d down
39170 f forward
39172 b back
39174 Default value is @samp{rludfb}.
39176 in_frot
39178 out_frot
39179 Set rotation of faces for the input/output cubemap. Choose
39180 one angle for each position.
39181 Designation of angles:
39183 0 0 degrees clockwise
39185 1 90 degrees clockwise
39187 2 180 degrees clockwise
39189 3 270 degrees clockwise
39191 Default value is @samp{000000}.
39193 eac Equi-Angular Cubemap.
39195 flat
39197 gnomonic
39198 rectilinear
39199 Regular video.
39200 Format specific options:
39202 h_fov
39204 v_fov
39205 d_fov
39206 Set output horizontal/vertical/diagonal field of view.
39207 Values in degrees.
39208 If diagonal field of view is set it overrides horizontal
39210 and vertical field of view.
39211 ih_fov
39213 iv_fov
39214 id_fov
39215 Set input horizontal/vertical/diagonal field of view.
39216 Values in degrees.
39217 If diagonal field of view is set it overrides horizontal
39219 and vertical field of view.
39220 dfisheye
39222 Dual fisheye.
39223 Format specific options:
39225 h_fov
39227 v_fov
39228 d_fov
39229 Set output horizontal/vertical/diagonal field of view.
39230 Values in degrees.
39231 If diagonal field of view is set it overrides horizontal
39233 and vertical field of view.
39234 ih_fov
39236 iv_fov
39237 id_fov
39238 Set input horizontal/vertical/diagonal field of view.
39239 Values in degrees.
39240 If diagonal field of view is set it overrides horizontal
39242 and vertical field of view.
39243 barrel
39245 fb
39246 barrelsplit
39247 Facebook's 360 formats.
39248 sg Stereographic format.
39250 Format specific options:
39252 h_fov
39254 v_fov
39255 d_fov
39256 Set output horizontal/vertical/diagonal field of view.
39257 Values in degrees.
39258 If diagonal field of view is set it overrides horizontal
39260 and vertical field of view.
39261 ih_fov
39263 iv_fov
39264 id_fov
39265 Set input horizontal/vertical/diagonal field of view.
39266 Values in degrees.
39267 If diagonal field of view is set it overrides horizontal
39269 and vertical field of view.
39270 mercator
39272 Mercator format.
39273 ball
39275 Ball format, gives significant distortion toward the back.
39276 hammer
39278 Hammer-Aitoff map projection format.
39279 sinusoidal
39281 Sinusoidal map projection format.
39282 fisheye
39284 Fisheye projection.
39285 Format specific options:
39287 h_fov
39289 v_fov
39290 d_fov
39291 Set output horizontal/vertical/diagonal field of view.
39292 Values in degrees.
39293 If diagonal field of view is set it overrides horizontal
39295 and vertical field of view.
39296 ih_fov
39298 iv_fov
39299 id_fov
39300 Set input horizontal/vertical/diagonal field of view.
39301 Values in degrees.
39302 If diagonal field of view is set it overrides horizontal
39304 and vertical field of view.
39305 pannini
39307 Pannini projection.
39308 Format specific options:
39310 h_fov
39312 Set output pannini parameter.
39313 ih_fov
39315 Set input pannini parameter.
39316 cylindrical
39318 Cylindrical projection.
39319 Format specific options:
39321 h_fov
39323 v_fov
39324 d_fov
39325 Set output horizontal/vertical/diagonal field of view.
39326 Values in degrees.
39327 If diagonal field of view is set it overrides horizontal
39329 and vertical field of view.
39330 ih_fov
39332 iv_fov
39333 id_fov
39334 Set input horizontal/vertical/diagonal field of view.
39335 Values in degrees.
39336 If diagonal field of view is set it overrides horizontal
39338 and vertical field of view.
39339 perspective
39341 Perspective projection. (output only)
39342 Format specific options:
39344 v_fov
39346 Set perspective parameter.
39347 tetrahedron
39349 Tetrahedron projection.
39350 tsp Truncated square pyramid projection.
39352 he
39354 hequirect
39355 Half equirectangular projection.
39356 equisolid
39358 Equisolid format.
39359 Format specific options:
39361 h_fov
39363 v_fov
39364 d_fov
39365 Set output horizontal/vertical/diagonal field of view.
39366 Values in degrees.
39367 If diagonal field of view is set it overrides horizontal
39369 and vertical field of view.
39370 ih_fov
39372 iv_fov
39373 id_fov
39374 Set input horizontal/vertical/diagonal field of view.
39375 Values in degrees.
39376 If diagonal field of view is set it overrides horizontal
39378 and vertical field of view.
39379 og Orthographic format.
39381 Format specific options:
39383 h_fov
39385 v_fov
39386 d_fov
39387 Set output horizontal/vertical/diagonal field of view.
39388 Values in degrees.
39389 If diagonal field of view is set it overrides horizontal
39391 and vertical field of view.
39392 ih_fov
39394 iv_fov
39395 id_fov
39396 Set input horizontal/vertical/diagonal field of view.
39397 Values in degrees.
39398 If diagonal field of view is set it overrides horizontal
39400 and vertical field of view.
39401 octahedron
39403 Octahedron projection.
39404 cylindricalea
39406 Cylindrical Equal Area projection.
39407 interp
39409 Set interpolation method.Note: more complex interpolation methods
39410 require much more memory to run.
39411 Available methods:
39413 near
39415 nearest
39416 Nearest neighbour.
39417 line
39419 linear
39420 Bilinear interpolation.
39421 lagrange9
39423 Lagrange9 interpolation.
39424 cube
39426 cubic
39427 Bicubic interpolation.
39428 lanc
39430 lanczos
39431 Lanczos interpolation.
39432 sp16
39434 spline16
39435 Spline16 interpolation.
39436 gauss
39438 gaussian
39439 Gaussian interpolation.
39440 mitchell
39442 Mitchell interpolation.
39443 Default value is @samp{line}.
39445 w
39447 h Set the output video resolution.
39448 Default resolution depends on formats.
39450 in_stereo
39452 out_stereo
39453 Set the input/output stereo format.
39454 2d 2D mono
39456 sbs Side by side
39458 tb Top bottom
39460 Default value is @samp{2d} for input and output format.
39462 yaw
39464 pitch
39465 roll
39466 Set rotation for the output video. Values in degrees.
39467 rorder
39469 Set rotation order for the output video. Choose one item for each
39470 position.
39471 y, Y
39473 yaw
39474 p, P
39476 pitch
39477 r, R
39479 roll
39480 Default value is @samp{ypr}.
39482 h_flip
39484 v_flip
39485 d_flip
39486 Flip the output video horizontally(swaps
39487 left-right)/vertically(swaps up-down)/in-depth(swaps back-forward).
39488 Boolean values.
39489 ih_flip
39491 iv_flip
39492 Set if input video is flipped horizontally/vertically. Boolean
39493 values.
39494 in_trans
39496 Set if input video is transposed. Boolean value, by default
39497 disabled.
39498 out_trans
39500 Set if output video needs to be transposed. Boolean value, by
39501 default disabled.
39502 h_offset
39504 v_offset
39505 Set output horizontal/vertical off-axis offset. Default is set to
39506 0. Allowed range is from -1 to 1.
39507 alpha_mask
39509 Build mask in alpha plane for all unmapped pixels by marking them
39510 fully transparent. Boolean value, by default disabled.
39511 reset_rot
39513 Reset rotation of output video. Boolean value, by default disabled.
39514 Examples
39516 • Convert equirectangular video to cubemap with 3x2 layout and 1%
39518 padding using bicubic interpolation:
39519 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
39521 • Extract back view of Equi-Angular Cubemap:
39523 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
39525 • Convert transposed and horizontally flipped Equi-Angular Cubemap in
39527 side-by-side stereo format to equirectangular top-bottom stereo
39528 format:
39529 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
39531 Commands
39533 This filter supports subset of above options as commands.
39535 vaguedenoiser
39537 Apply a wavelet based denoiser.
39538 It transforms each frame from the video input into the wavelet domain,
39540 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
39541 the obtained coefficients. It does an inverse wavelet transform after.
39542 Due to wavelet properties, it should give a nice smoothed result, and
39543 reduced noise, without blurring picture features.
39544 This filter accepts the following options:
39546 threshold
39548 The filtering strength. The higher, the more filtered the video
39549 will be. Hard thresholding can use a higher threshold than soft
39550 thresholding before the video looks overfiltered. Default value is
39551 2.
39552 method
39554 The filtering method the filter will use.
39555 It accepts the following values:
39557 hard
39559 All values under the threshold will be zeroed.
39560 soft
39562 All values under the threshold will be zeroed. All values above
39563 will be reduced by the threshold.
39564 garrote
39566 Scales or nullifies coefficients - intermediary between (more)
39567 soft and (less) hard thresholding.
39568 Default is garrote.
39570 nsteps
39572 Number of times, the wavelet will decompose the picture. Picture
39573 can't be decomposed beyond a particular point (typically, 8 for a
39574 640x480 frame - as 2^9 = 512 > 480). Valid values are integers
39575 between 1 and 32. Default value is 6.
39576 percent
39578 Partial of full denoising (limited coefficients shrinking), from 0
39579 to 100. Default value is 85.
39580 planes
39582 A list of the planes to process. By default all planes are
39583 processed.
39584 type
39586 The threshold type the filter will use.
39587 It accepts the following values:
39589 universal
39591 Threshold used is same for all decompositions.
39592 bayes
39594 Threshold used depends also on each decomposition coefficients.
39595 Default is universal.
39597 varblur
39599 Apply variable blur filter by using 2nd video stream to set blur
39600 radius. The 2nd stream must have the same dimensions.
39601 This filter accepts the following options:
39603 min_r
39605 Set min allowed radius. Allowed range is from 0 to 254. Default is
39606 0.
39607 max_r
39609 Set max allowed radius. Allowed range is from 1 to 255. Default is
39610 8.
39611 planes
39613 Set which planes to process. By default, all are used.
39614 The "varblur" filter also supports the framesync options.
39616 Commands
39618 This filter supports all the above options as commands.
39620 vectorscope
39622 Display 2 color component values in the two dimensional graph (which is
39623 called a vectorscope).
39624 This filter accepts the following options:
39626 mode, m
39628 Set vectorscope mode.
39629 It accepts the following values:
39631 gray
39633 tint
39634 Gray values are displayed on graph, higher brightness means
39635 more pixels have same component color value on location in
39636 graph. This is the default mode.
39637 color
39639 Gray values are displayed on graph. Surrounding pixels values
39640 which are not present in video frame are drawn in gradient of 2
39641 color components which are set by option "x" and "y". The 3rd
39642 color component is static.
39643 color2
39645 Actual color components values present in video frame are
39646 displayed on graph.
39647 color3
39649 Similar as color2 but higher frequency of same values "x" and
39650 "y" on graph increases value of another color component, which
39651 is luminance by default values of "x" and "y".
39652 color4
39654 Actual colors present in video frame are displayed on graph. If
39655 two different colors map to same position on graph then color
39656 with higher value of component not present in graph is picked.
39657 color5
39659 Gray values are displayed on graph. Similar to "color" but with
39660 3rd color component picked from radial gradient.
39661 x Set which color component will be represented on X-axis. Default is
39663 1.
39664 y Set which color component will be represented on Y-axis. Default is
39666 2.
39667 intensity, i
39669 Set intensity, used by modes: gray, color, color3 and color5 for
39670 increasing brightness of color component which represents frequency
39671 of (X, Y) location in graph.
39672 envelope, e
39674 none
39675 No envelope, this is default.
39676 instant
39678 Instant envelope, even darkest single pixel will be clearly
39679 highlighted.
39680 peak
39682 Hold maximum and minimum values presented in graph over time.
39683 This way you can still spot out of range values without
39684 constantly looking at vectorscope.
39685 peak+instant
39687 Peak and instant envelope combined together.
39688 graticule, g
39690 Set what kind of graticule to draw.
39691 none
39693 green
39694 color
39695 invert
39696 opacity, o
39697 Set graticule opacity.
39698 flags, f
39700 Set graticule flags.
39701 white
39703 Draw graticule for white point.
39704 black
39706 Draw graticule for black point.
39707 name
39709 Draw color points short names.
39710 bgopacity, b
39712 Set background opacity.
39713 lthreshold, l
39715 Set low threshold for color component not represented on X or Y
39716 axis. Values lower than this value will be ignored. Default is 0.
39717 Note this value is multiplied with actual max possible value one
39718 pixel component can have. So for 8-bit input and low threshold
39719 value of 0.1 actual threshold is 0.1 * 255 = 25.
39720 hthreshold, h
39722 Set high threshold for color component not represented on X or Y
39723 axis. Values higher than this value will be ignored. Default is 1.
39724 Note this value is multiplied with actual max possible value one
39725 pixel component can have. So for 8-bit input and high threshold
39726 value of 0.9 actual threshold is 0.9 * 255 = 230.
39727 colorspace, c
39729 Set what kind of colorspace to use when drawing graticule.
39730 auto
39732 601
39733 709
39734 Default is auto.
39736 tint0, t0
39738 tint1, t1
39739 Set color tint for gray/tint vectorscope mode. By default both
39740 options are zero. This means no tint, and output will remain gray.
39741 vidstabdetect
39743 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
39744 vidstabtransform for pass 2.
39745 This filter generates a file with relative translation and rotation
39747 transform information about subsequent frames, which is then used by
39748 the vidstabtransform filter.
39749 To enable compilation of this filter you need to configure FFmpeg with
39751 "--enable-libvidstab".
39752 This filter accepts the following options:
39754 result
39756 Set the path to the file used to write the transforms information.
39757 Default value is transforms.trf.
39758 shakiness
39760 Set how shaky the video is and how quick the camera is. It accepts
39761 an integer in the range 1-10, a value of 1 means little shakiness,
39762 a value of 10 means strong shakiness. Default value is 5.
39763 accuracy
39765 Set the accuracy of the detection process. It must be a value in
39766 the range 1-15. A value of 1 means low accuracy, a value of 15
39767 means high accuracy. Default value is 15.
39768 stepsize
39770 Set stepsize of the search process. The region around minimum is
39771 scanned with 1 pixel resolution. Default value is 6.
39772 mincontrast
39774 Set minimum contrast. Below this value a local measurement field is
39775 discarded. Must be a floating point value in the range 0-1. Default
39776 value is 0.3.
39777 tripod
39779 Set reference frame number for tripod mode.
39780 If enabled, the motion of the frames is compared to a reference
39782 frame in the filtered stream, identified by the specified number.
39783 The idea is to compensate all movements in a more-or-less static
39784 scene and keep the camera view absolutely still.
39785 If set to 0, it is disabled. The frames are counted starting from
39787 1.
39788 show
39790 Show fields and transforms in the resulting frames. It accepts an
39791 integer in the range 0-2. Default value is 0, which disables any
39792 visualization.
39793 Examples
39795 • Use default values:
39797 vidstabdetect
39799 • Analyze strongly shaky movie and put the results in file
39801 mytransforms.trf:
39802 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
39804 • Visualize the result of internal transformations in the resulting
39806 video:
39807 vidstabdetect=show=1
39809 • Analyze a video with medium shakiness using ffmpeg:
39811 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
39813 vidstabtransform
39815 Video stabilization/deshaking: pass 2 of 2, see vidstabdetect for pass
39816 1.
39817 Read a file with transform information for each frame and
39819 apply/compensate them. Together with the vidstabdetect filter this can
39820 be used to deshake videos. See also
39821 <http://public.hronopik.de/vid.stab>. It is important to also use the
39822 unsharp filter, see below.
39823 To enable compilation of this filter you need to configure FFmpeg with
39825 "--enable-libvidstab".
39826 Options
39828 input
39830 Set path to the file used to read the transforms. Default value is
39831 transforms.trf.
39832 smoothing
39834 Set the number of frames (value*2 + 1) used for lowpass filtering
39835 the camera movements. Default value is 10.
39836 For example a number of 10 means that 21 frames are used (10 in the
39838 past and 10 in the future) to smoothen the motion in the video. A
39839 larger value leads to a smoother video, but limits the acceleration
39840 of the camera (pan/tilt movements). 0 is a special case where a
39841 static camera is simulated.
39842 optalgo
39844 Set the camera path optimization algorithm.
39845 Accepted values are:
39847 gauss
39849 gaussian kernel low-pass filter on camera motion (default)
39850 avg averaging on transformations
39852 maxshift
39854 Set maximal number of pixels to translate frames. Default value is
39855 -1, meaning no limit.
39856 maxangle
39858 Set maximal angle in radians (degree*PI/180) to rotate frames.
39859 Default value is -1, meaning no limit.
39860 crop
39862 Specify how to deal with borders that may be visible due to
39863 movement compensation.
39864 Available values are:
39866 keep
39868 keep image information from previous frame (default)
39869 black
39871 fill the border black
39872 invert
39874 Invert transforms if set to 1. Default value is 0.
39875 relative
39877 Consider transforms as relative to previous frame if set to 1,
39878 absolute if set to 0. Default value is 0.
39879 zoom
39881 Set percentage to zoom. A positive value will result in a zoom-in
39882 effect, a negative value in a zoom-out effect. Default value is 0
39883 (no zoom).
39884 optzoom
39886 Set optimal zooming to avoid borders.
39887 Accepted values are:
39889 0 disabled
39891 1 optimal static zoom value is determined (only very strong
39893 movements will lead to visible borders) (default)
39894 2 optimal adaptive zoom value is determined (no borders will be
39896 visible), see zoomspeed
39897 Note that the value given at zoom is added to the one calculated
39899 here.
39900 zoomspeed
39902 Set percent to zoom maximally each frame (enabled when optzoom is
39903 set to 2). Range is from 0 to 5, default value is 0.25.
39904 interpol
39906 Specify type of interpolation.
39907 Available values are:
39909 no no interpolation
39911 linear
39913 linear only horizontal
39914 bilinear
39916 linear in both directions (default)
39917 bicubic
39919 cubic in both directions (slow)
39920 tripod
39922 Enable virtual tripod mode if set to 1, which is equivalent to
39923 "relative=0:smoothing=0". Default value is 0.
39924 Use also "tripod" option of vidstabdetect.
39926 debug
39928 Increase log verbosity if set to 1. Also the detected global
39929 motions are written to the temporary file global_motions.trf.
39930 Default value is 0.
39931 Examples
39933 • Use ffmpeg for a typical stabilization with default values:
39935 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
39937 Note the use of the unsharp filter which is always recommended.
39939 • Zoom in a bit more and load transform data from a given file:
39941 vidstabtransform=zoom=5:input="mytransforms.trf"
39943 • Smoothen the video even more:
39945 vidstabtransform=smoothing=30
39947 vflip
39949 Flip the input video vertically.
39950 For example, to vertically flip a video with ffmpeg:
39952 ffmpeg -i in.avi -vf "vflip" out.avi
39954 vfrdet
39956 Detect variable frame rate video.
39957 This filter tries to detect if the input is variable or constant frame
39959 rate.
39960 At end it will output number of frames detected as having variable
39962 delta pts, and ones with constant delta pts. If there was frames with
39963 variable delta, than it will also show min, max and average delta
39964 encountered.
39965 vibrance
39967 Boost or alter saturation.
39968 The filter accepts the following options:
39970 intensity
39972 Set strength of boost if positive value or strength of alter if
39973 negative value. Default is 0. Allowed range is from -2 to 2.
39974 rbal
39976 Set the red balance. Default is 1. Allowed range is from -10 to 10.
39977 gbal
39979 Set the green balance. Default is 1. Allowed range is from -10 to
39980 10.
39981 bbal
39983 Set the blue balance. Default is 1. Allowed range is from -10 to
39984 10.
39985 rlum
39987 Set the red luma coefficient.
39988 glum
39990 Set the green luma coefficient.
39991 blum
39993 Set the blue luma coefficient.
39994 alternate
39996 If "intensity" is negative and this is set to 1, colors will
39997 change, otherwise colors will be less saturated, more towards gray.
39998 Commands
40000 This filter supports the all above options as commands.
40002 vif
40004 Obtain the average VIF (Visual Information Fidelity) between two input
40005 videos.
40006 This filter takes two input videos.
40008 Both input videos must have the same resolution and pixel format for
40010 this filter to work correctly. Also it assumes that both inputs have
40011 the same number of frames, which are compared one by one.
40012 The obtained average VIF score is printed through the logging system.
40014 The filter stores the calculated VIF score of each frame.
40016 This filter also supports the framesync options.
40018 In the below example the input file main.mpg being processed is
40020 compared with the reference file ref.mpg.
40021 ffmpeg -i main.mpg -i ref.mpg -lavfi vif -f null -
40023 vignette
40025 Make or reverse a natural vignetting effect.
40026 The filter accepts the following options:
40028 angle, a
40030 Set lens angle expression as a number of radians.
40031 The value is clipped in the "[0,PI/2]" range.
40033 Default value: "PI/5"
40035 x0
40037 y0 Set center coordinates expressions. Respectively "w/2" and "h/2" by
40038 default.
40039 mode
40041 Set forward/backward mode.
40042 Available modes are:
40044 forward
40046 The larger the distance from the central point, the darker the
40047 image becomes.
40048 backward
40050 The larger the distance from the central point, the brighter
40051 the image becomes. This can be used to reverse a vignette
40052 effect, though there is no automatic detection to extract the
40053 lens angle and other settings (yet). It can also be used to
40054 create a burning effect.
40055 Default value is forward.
40057 eval
40059 Set evaluation mode for the expressions (angle, x0, y0).
40060 It accepts the following values:
40062 init
40064 Evaluate expressions only once during the filter
40065 initialization.
40066 frame
40068 Evaluate expressions for each incoming frame. This is way
40069 slower than the init mode since it requires all the scalers to
40070 be re-computed, but it allows advanced dynamic expressions.
40071 Default value is init.
40073 dither
40075 Set dithering to reduce the circular banding effects. Default is 1
40076 (enabled).
40077 aspect
40079 Set vignette aspect. This setting allows one to adjust the shape of
40080 the vignette. Setting this value to the SAR of the input will make
40081 a rectangular vignetting following the dimensions of the video.
40082 Default is "1/1".
40084 Expressions
40086 The alpha, x0 and y0 expressions can contain the following parameters.
40088 w
40090 h input width and height
40091 n the number of input frame, starting from 0
40093 pts the PTS (Presentation TimeStamp) time of the filtered video frame,
40095 expressed in TB units, NAN if undefined
40096 r frame rate of the input video, NAN if the input frame rate is
40098 unknown
40099 t the PTS (Presentation TimeStamp) of the filtered video frame,
40101 expressed in seconds, NAN if undefined
40102 tb time base of the input video
40104 Examples
40106 • Apply simple strong vignetting effect:
40108 vignette=PI/4
40110 • Make a flickering vignetting:
40112 vignette='PI/4+random(1)*PI/50':eval=frame
40114 vmafmotion
40116 Obtain the average VMAF motion score of a video. It is one of the
40117 component metrics of VMAF.
40118 The obtained average motion score is printed through the logging
40120 system.
40121 The filter accepts the following options:
40123 stats_file
40125 If specified, the filter will use the named file to save the motion
40126 score of each frame with respect to the previous frame. When
40127 filename equals "-" the data is sent to standard output.
40128 Example:
40130 ffmpeg -i ref.mpg -vf vmafmotion -f null -
40132 vstack
40134 Stack input videos vertically.
40135 All streams must be of same pixel format and of same width.
40137 Note that this filter is faster than using overlay and pad filter to
40139 create same output.
40140 The filter accepts the following options:
40142 inputs
40144 Set number of input streams. Default is 2.
40145 shortest
40147 If set to 1, force the output to terminate when the shortest input
40148 terminates. Default value is 0.
40149 w3fdif
40151 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
40152 Deinterlacing Filter").
40153 Based on the process described by Martin Weston for BBC R&D, and
40155 implemented based on the de-interlace algorithm written by Jim
40156 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter uses
40157 filter coefficients calculated by BBC R&D.
40158 This filter uses field-dominance information in frame to decide which
40160 of each pair of fields to place first in the output. If it gets it
40161 wrong use setfield filter before "w3fdif" filter.
40162 There are two sets of filter coefficients, so called "simple" and
40164 "complex". Which set of filter coefficients is used can be set by
40165 passing an optional parameter:
40166 filter
40168 Set the interlacing filter coefficients. Accepts one of the
40169 following values:
40170 simple
40172 Simple filter coefficient set.
40173 complex
40175 More-complex filter coefficient set.
40176 Default value is complex.
40178 mode
40180 The interlacing mode to adopt. It accepts one of the following
40181 values:
40182 frame
40184 Output one frame for each frame.
40185 field
40187 Output one frame for each field.
40188 The default value is "field".
40190 parity
40192 The picture field parity assumed for the input interlaced video. It
40193 accepts one of the following values:
40194 tff Assume the top field is first.
40196 bff Assume the bottom field is first.
40198 auto
40200 Enable automatic detection of field parity.
40201 The default value is "auto". If the interlacing is unknown or the
40203 decoder does not export this information, top field first will be
40204 assumed.
40205 deint
40207 Specify which frames to deinterlace. Accepts one of the following
40208 values:
40209 all Deinterlace all frames,
40211 interlaced
40213 Only deinterlace frames marked as interlaced.
40214 Default value is all.
40216 Commands
40218 This filter supports same commands as options.
40220 waveform
40222 Video waveform monitor.
40223 The waveform monitor plots color component intensity. By default
40225 luminance only. Each column of the waveform corresponds to a column of
40226 pixels in the source video.
40227 It accepts the following options:
40229 mode, m
40231 Can be either "row", or "column". Default is "column". In row
40232 mode, the graph on the left side represents color component value 0
40233 and the right side represents value = 255. In column mode, the top
40234 side represents color component value = 0 and bottom side
40235 represents value = 255.
40236 intensity, i
40238 Set intensity. Smaller values are useful to find out how many
40239 values of the same luminance are distributed across input
40240 rows/columns. Default value is 0.04. Allowed range is [0, 1].
40241 mirror, r
40243 Set mirroring mode. 0 means unmirrored, 1 means mirrored. In
40244 mirrored mode, higher values will be represented on the left side
40245 for "row" mode and at the top for "column" mode. Default is 1
40246 (mirrored).
40247 display, d
40249 Set display mode. It accepts the following values:
40250 overlay
40252 Presents information identical to that in the "parade", except
40253 that the graphs representing color components are superimposed
40254 directly over one another.
40255 This display mode makes it easier to spot relative differences
40257 or similarities in overlapping areas of the color components
40258 that are supposed to be identical, such as neutral whites,
40259 grays, or blacks.
40260 stack
40262 Display separate graph for the color components side by side in
40263 "row" mode or one below the other in "column" mode.
40264 parade
40266 Display separate graph for the color components side by side in
40267 "column" mode or one below the other in "row" mode.
40268 Using this display mode makes it easy to spot color casts in
40270 the highlights and shadows of an image, by comparing the
40271 contours of the top and the bottom graphs of each waveform.
40272 Since whites, grays, and blacks are characterized by exactly
40273 equal amounts of red, green, and blue, neutral areas of the
40274 picture should display three waveforms of roughly equal
40275 width/height. If not, the correction is easy to perform by
40276 making level adjustments the three waveforms.
40277 Default is "stack".
40279 components, c
40281 Set which color components to display. Default is 1, which means
40282 only luminance or red color component if input is in RGB
40283 colorspace. If is set for example to 7 it will display all 3 (if)
40284 available color components.
40285 envelope, e
40287 none
40288 No envelope, this is default.
40289 instant
40291 Instant envelope, minimum and maximum values presented in graph
40292 will be easily visible even with small "step" value.
40293 peak
40295 Hold minimum and maximum values presented in graph across time.
40296 This way you can still spot out of range values without
40297 constantly looking at waveforms.
40298 peak+instant
40300 Peak and instant envelope combined together.
40301 filter, f
40303 lowpass
40304 No filtering, this is default.
40305 flat
40307 Luma and chroma combined together.
40308 aflat
40310 Similar as above, but shows difference between blue and red
40311 chroma.
40312 xflat
40314 Similar as above, but use different colors.
40315 yflat
40317 Similar as above, but again with different colors.
40318 chroma
40320 Displays only chroma.
40321 color
40323 Displays actual color value on waveform.
40324 acolor
40326 Similar as above, but with luma showing frequency of chroma
40327 values.
40328 graticule, g
40330 Set which graticule to display.
40331 none
40333 Do not display graticule.
40334 green
40336 Display green graticule showing legal broadcast ranges.
40337 orange
40339 Display orange graticule showing legal broadcast ranges.
40340 invert
40342 Display invert graticule showing legal broadcast ranges.
40343 opacity, o
40345 Set graticule opacity.
40346 flags, fl
40348 Set graticule flags.
40349 numbers
40351 Draw numbers above lines. By default enabled.
40352 dots
40354 Draw dots instead of lines.
40355 scale, s
40357 Set scale used for displaying graticule.
40358 digital
40360 millivolts
40361 ire
40362 Default is digital.
40364 bgopacity, b
40366 Set background opacity.
40367 tint0, t0
40369 tint1, t1
40370 Set tint for output. Only used with lowpass filter and when
40371 display is not overlay and input pixel formats are not RGB.
40372 fitmode, fm
40374 Set sample aspect ratio of video output frames. Can be used to
40375 configure waveform so it is not streched too much in one of
40376 directions.
40377 none
40379 Set sample aspect ration to 1/1.
40380 size
40382 Set sample aspect ratio to match input size of video
40383 Default is none.
40385 weave, doubleweave
40387 The "weave" takes a field-based video input and join each two
40388 sequential fields into single frame, producing a new double height clip
40389 with half the frame rate and half the frame count.
40390 The "doubleweave" works same as "weave" but without halving frame rate
40392 and frame count.
40393 It accepts the following option:
40395 first_field
40397 Set first field. Available values are:
40398 top, t
40400 Set the frame as top-field-first.
40401 bottom, b
40403 Set the frame as bottom-field-first.
40404 Examples
40406 • Interlace video using select and separatefields filter:
40408 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
40410 xbr
40412 Apply the xBR high-quality magnification filter which is designed for
40413 pixel art. It follows a set of edge-detection rules, see
40414 <https://forums.libretro.com/t/xbr-algorithm-tutorial/123>.
40415 It accepts the following option:
40417 n Set the scaling dimension: 2 for "2xBR", 3 for "3xBR" and 4 for
40419 "4xBR". Default is 3.
40420 xcorrelate
40422 Apply normalized cross-correlation between first and second input video
40423 stream.
40424 Second input video stream dimensions must be lower than first input
40426 video stream.
40427 The filter accepts the following options:
40429 planes
40431 Set which planes to process.
40432 secondary
40434 Set which secondary video frames will be processed from second
40435 input video stream, can be first or all. Default is all.
40436 The "xcorrelate" filter also supports the framesync options.
40438 xfade
40440 Apply cross fade from one input video stream to another input video
40441 stream. The cross fade is applied for specified duration.
40442 Both inputs must be constant frame-rate and have the same resolution,
40444 pixel format, frame rate and timebase.
40445 The filter accepts the following options:
40447 transition
40449 Set one of available transition effects:
40450 custom
40452 fade
40453 wipeleft
40454 wiperight
40455 wipeup
40456 wipedown
40457 slideleft
40458 slideright
40459 slideup
40460 slidedown
40461 circlecrop
40462 rectcrop
40463 distance
40464 fadeblack
40465 fadewhite
40466 radial
40467 smoothleft
40468 smoothright
40469 smoothup
40470 smoothdown
40471 circleopen
40472 circleclose
40473 vertopen
40474 vertclose
40475 horzopen
40476 horzclose
40477 dissolve
40478 pixelize
40479 diagtl
40480 diagtr
40481 diagbl
40482 diagbr
40483 hlslice
40484 hrslice
40485 vuslice
40486 vdslice
40487 hblur
40488 fadegrays
40489 wipetl
40490 wipetr
40491 wipebl
40492 wipebr
40493 squeezeh
40494 squeezev
40495 zoomin
40496 fadefast
40497 fadeslow
40498 Default transition effect is fade.
40500 duration
40502 Set cross fade duration in seconds. Range is 0 to 60 seconds.
40503 Default duration is 1 second.
40504 offset
40506 Set cross fade start relative to first input stream in seconds.
40507 Default offset is 0.
40508 expr
40510 Set expression for custom transition effect.
40511 The expressions can use the following variables and functions:
40513 X
40515 Y The coordinates of the current sample.
40516 W
40518 H The width and height of the image.
40519 P Progress of transition effect.
40521 PLANE
40523 Currently processed plane.
40524 A Return value of first input at current location and plane.
40526 B Return value of second input at current location and plane.
40528 a0(x, y)
40530 a1(x, y)
40531 a2(x, y)
40532 a3(x, y)
40533 Return the value of the pixel at location (x,y) of the
40534 first/second/third/fourth component of first input.
40535 b0(x, y)
40537 b1(x, y)
40538 b2(x, y)
40539 b3(x, y)
40540 Return the value of the pixel at location (x,y) of the
40541 first/second/third/fourth component of second input.
40542 Examples
40544 • Cross fade from one input video to another input video, with fade
40546 transition and duration of transition of 2 seconds starting at
40547 offset of 5 seconds:
40548 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
40550 xmedian
40552 Pick median pixels from several input videos.
40553 The filter accepts the following options:
40555 inputs
40557 Set number of inputs. Default is 3. Allowed range is from 3 to
40558 255. If number of inputs is even number, than result will be mean
40559 value between two median values.
40560 planes
40562 Set which planes to filter. Default value is 15, by which all
40563 planes are processed.
40564 percentile
40566 Set median percentile. Default value is 0.5. Default value of 0.5
40567 will pick always median values, while 0 will pick minimum values,
40568 and 1 maximum values.
40569 Commands
40571 This filter supports all above options as commands, excluding option
40573 "inputs".
40574 xstack
40576 Stack video inputs into custom layout.
40577 All streams must be of same pixel format.
40579 The filter accepts the following options:
40581 inputs
40583 Set number of input streams. Default is 2.
40584 layout
40586 Specify layout of inputs. This option requires the desired layout
40587 configuration to be explicitly set by the user. This sets position
40588 of each video input in output. Each input is separated by '|'. The
40589 first number represents the column, and the second number
40590 represents the row. Numbers start at 0 and are separated by '_'.
40591 Optionally one can use wX and hX, where X is video input from which
40592 to take width or height. Multiple values can be used when
40593 separated by '+'. In such case values are summed together.
40594 Note that if inputs are of different sizes gaps may appear, as not
40596 all of the output video frame will be filled. Similarly, videos can
40597 overlap each other if their position doesn't leave enough space for
40598 the full frame of adjoining videos.
40599 For 2 inputs, a default layout of "0_0|w0_0" (equivalent to
40601 "grid=2x1") is set. In all other cases, a layout or a grid must be
40602 set by the user. Either "grid" or "layout" can be specified at a
40603 time. Specifying both will result in an error.
40604 grid
40606 Specify a fixed size grid of inputs. This option is used to create
40607 a fixed size grid of the input streams. Set the grid size in the
40608 form "COLUMNSxROWS". There must be "ROWS * COLUMNS" input streams
40609 and they will be arranged as a grid with "ROWS" rows and "COLUMNS"
40610 columns. When using this option, each input stream within a row
40611 must have the same height and all the rows must have the same
40612 width.
40613 If "grid" is set, then "inputs" option is ignored and is implicitly
40615 set to "ROWS * COLUMNS".
40616 For 2 inputs, a default grid of "2x1" (equivalent to
40618 "layout=0_0|w0_0") is set. In all other cases, a layout or a grid
40619 must be set by the user. Either "grid" or "layout" can be specified
40620 at a time. Specifying both will result in an error.
40621 shortest
40623 If set to 1, force the output to terminate when the shortest input
40624 terminates. Default value is 0.
40625 fill
40627 If set to valid color, all unused pixels will be filled with that
40628 color. By default fill is set to none, so it is disabled.
40629 Examples
40631 • Display 4 inputs into 2x2 grid.
40633 Layout:
40635 input1(0, 0) | input3(w0, 0)
40637 input2(0, h0) | input4(w0, h0)
40638 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
40642 Note that if inputs are of different sizes, gaps or overlaps may
40644 occur.
40645 • Display 4 inputs into 1x4 grid.
40647 Layout:
40649 input1(0, 0)
40651 input2(0, h0)
40652 input3(0, h0+h1)
40653 input4(0, h0+h1+h2)
40654 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
40658 Note that if inputs are of different widths, unused space will
40660 appear.
40661 • Display 9 inputs into 3x3 grid.
40663 Layout:
40665 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
40667 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
40668 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
40669 xstack=inputs=9:layout=0_0|0_h0|0_h0+h1|w0_0|w0_h0|w0_h0+h1|w0+w3_0|w0+w3_h0|w0+w3_h0+h1
40673 Note that if inputs are of different sizes, gaps or overlaps may
40675 occur.
40676 • Display 16 inputs into 4x4 grid.
40678 Layout:
40680 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
40682 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
40683 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
40684 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
40685 xstack=inputs=16:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2|w0_0|w0_h0|w0_h0+h1|w0_h0+h1+h2|w0+w4_0|
40689 w0+w4_h0|w0+w4_h0+h1|w0+w4_h0+h1+h2|w0+w4+w8_0|w0+w4+w8_h0|w0+w4+w8_h0+h1|w0+w4+w8_h0+h1+h2
40690 Note that if inputs are of different sizes, gaps or overlaps may
40692 occur.
40693 yadif
40695 Deinterlace the input video ("yadif" means "yet another deinterlacing
40696 filter").
40697 It accepts the following parameters:
40699 mode
40701 The interlacing mode to adopt. It accepts one of the following
40702 values:
40703 0, send_frame
40705 Output one frame for each frame.
40706 1, send_field
40708 Output one frame for each field.
40709 2, send_frame_nospatial
40711 Like "send_frame", but it skips the spatial interlacing check.
40712 3, send_field_nospatial
40714 Like "send_field", but it skips the spatial interlacing check.
40715 The default value is "send_frame".
40717 parity
40719 The picture field parity assumed for the input interlaced video. It
40720 accepts one of the following values:
40721 0, tff
40723 Assume the top field is first.
40724 1, bff
40726 Assume the bottom field is first.
40727 -1, auto
40729 Enable automatic detection of field parity.
40730 The default value is "auto". If the interlacing is unknown or the
40732 decoder does not export this information, top field first will be
40733 assumed.
40734 deint
40736 Specify which frames to deinterlace. Accepts one of the following
40737 values:
40738 0, all
40740 Deinterlace all frames.
40741 1, interlaced
40743 Only deinterlace frames marked as interlaced.
40744 The default value is "all".
40746 yadif_cuda
40748 Deinterlace the input video using the yadif algorithm, but implemented
40749 in CUDA so that it can work as part of a GPU accelerated pipeline with
40750 nvdec and/or nvenc.
40751 It accepts the following parameters:
40753 mode
40755 The interlacing mode to adopt. It accepts one of the following
40756 values:
40757 0, send_frame
40759 Output one frame for each frame.
40760 1, send_field
40762 Output one frame for each field.
40763 2, send_frame_nospatial
40765 Like "send_frame", but it skips the spatial interlacing check.
40766 3, send_field_nospatial
40768 Like "send_field", but it skips the spatial interlacing check.
40769 The default value is "send_frame".
40771 parity
40773 The picture field parity assumed for the input interlaced video. It
40774 accepts one of the following values:
40775 0, tff
40777 Assume the top field is first.
40778 1, bff
40780 Assume the bottom field is first.
40781 -1, auto
40783 Enable automatic detection of field parity.
40784 The default value is "auto". If the interlacing is unknown or the
40786 decoder does not export this information, top field first will be
40787 assumed.
40788 deint
40790 Specify which frames to deinterlace. Accepts one of the following
40791 values:
40792 0, all
40794 Deinterlace all frames.
40795 1, interlaced
40797 Only deinterlace frames marked as interlaced.
40798 The default value is "all".
40800 yaepblur
40802 Apply blur filter while preserving edges ("yaepblur" means "yet another
40803 edge preserving blur filter"). The algorithm is described in "J. S.
40804 Lee, Digital image enhancement and noise filtering by use of local
40805 statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
40806 It accepts the following parameters:
40808 radius, r
40810 Set the window radius. Default value is 3.
40811 planes, p
40813 Set which planes to filter. Default is only the first plane.
40814 sigma, s
40816 Set blur strength. Default value is 128.
40817 Commands
40819 This filter supports same commands as options.
40821 zoompan
40823 Apply Zoom & Pan effect.
40824 This filter accepts the following options:
40826 zoom, z
40828 Set the zoom expression. Range is 1-10. Default is 1.
40829 x
40831 y Set the x and y expression. Default is 0.
40832 d Set the duration expression in number of frames. This sets for how
40834 many number of frames effect will last for single input image.
40835 Default is 90.
40836 s Set the output image size, default is 'hd720'.
40838 fps Set the output frame rate, default is '25'.
40840 Each expression can contain the following constants:
40842 in_w, iw
40844 Input width.
40845 in_h, ih
40847 Input height.
40848 out_w, ow
40850 Output width.
40851 out_h, oh
40853 Output height.
40854 in Input frame count.
40856 on Output frame count.
40858 in_time, it
40860 The input timestamp expressed in seconds. It's NAN if the input
40861 timestamp is unknown.
40862 out_time, time, ot
40864 The output timestamp expressed in seconds.
40865 x
40867 y Last calculated 'x' and 'y' position from 'x' and 'y' expression
40868 for current input frame.
40869 px
40871 py 'x' and 'y' of last output frame of previous input frame or 0 when
40872 there was not yet such frame (first input frame).
40873 zoom
40875 Last calculated zoom from 'z' expression for current input frame.
40876 pzoom
40878 Last calculated zoom of last output frame of previous input frame.
40879 duration
40881 Number of output frames for current input frame. Calculated from
40882 'd' expression for each input frame.
40883 pduration
40885 number of output frames created for previous input frame
40886 a Rational number: input width / input height
40888 sar sample aspect ratio
40890 dar display aspect ratio
40892 Examples
40894 • Zoom in up to 1.5x and pan at same time to some spot near center of
40896 picture:
40897 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='if(gte(zoom,1.5),x,x+1/a)':y='if(gte(zoom,1.5),y,y+1)':s=640x360
40899 • Zoom in up to 1.5x and pan always at center of picture:
40901 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
40903 • Same as above but without pausing:
40905 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
40907 • Zoom in 2x into center of picture only for the first second of the
40909 input video:
40910 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
40912 zscale
40914 Scale (resize) the input video, using the z.lib library:
40915 <https://github.com/sekrit-twc/zimg>. To enable compilation of this
40916 filter, you need to configure FFmpeg with "--enable-libzimg".
40917 The zscale filter forces the output display aspect ratio to be the same
40919 as the input, by changing the output sample aspect ratio.
40920 If the input image format is different from the format requested by the
40922 next filter, the zscale filter will convert the input to the requested
40923 format.
40924 Options
40926 The filter accepts the following options.
40928 width, w
40930 height, h
40931 Set the output video dimension expression. Default value is the
40932 input dimension.
40933 If the width or w value is 0, the input width is used for the
40935 output. If the height or h value is 0, the input height is used for
40936 the output.
40937 If one and only one of the values is -n with n >= 1, the zscale
40939 filter will use a value that maintains the aspect ratio of the
40940 input image, calculated from the other specified dimension. After
40941 that it will, however, make sure that the calculated dimension is
40942 divisible by n and adjust the value if necessary.
40943 If both values are -n with n >= 1, the behavior will be identical
40945 to both values being set to 0 as previously detailed.
40946 See below for the list of accepted constants for use in the
40948 dimension expression.
40949 size, s
40951 Set the video size. For the syntax of this option, check the "Video
40952 size" section in the ffmpeg-utils manual.
40953 dither, d
40955 Set the dither type.
40956 Possible values are:
40958 none
40960 ordered
40961 random
40962 error_diffusion
40963 Default is none.
40965 filter, f
40967 Set the resize filter type.
40968 Possible values are:
40970 point
40972 bilinear
40973 bicubic
40974 spline16
40975 spline36
40976 lanczos
40977 Default is bilinear.
40979 range, r
40981 Set the color range.
40982 Possible values are:
40984 input
40986 limited
40987 full
40988 Default is same as input.
40990 primaries, p
40992 Set the color primaries.
40993 Possible values are:
40995 input
40997 709
40998 unspecified
40999 170m
41000 240m
41001 2020
41002 Default is same as input.
41004 transfer, t
41006 Set the transfer characteristics.
41007 Possible values are:
41009 input
41011 709
41012 unspecified
41013 601
41014 linear
41015 2020_10
41016 2020_12
41017 smpte2084
41018 iec61966-2-1
41019 arib-std-b67
41020 Default is same as input.
41022 matrix, m
41024 Set the colorspace matrix.
41025 Possible value are:
41027 input
41029 709
41030 unspecified
41031 470bg
41032 170m
41033 2020_ncl
41034 2020_cl
41035 Default is same as input.
41037 rangein, rin
41039 Set the input color range.
41040 Possible values are:
41042 input
41044 limited
41045 full
41046 Default is same as input.
41048 primariesin, pin
41050 Set the input color primaries.
41051 Possible values are:
41053 input
41055 709
41056 unspecified
41057 170m
41058 240m
41059 2020
41060 Default is same as input.
41062 transferin, tin
41064 Set the input transfer characteristics.
41065 Possible values are:
41067 input
41069 709
41070 unspecified
41071 601
41072 linear
41073 2020_10
41074 2020_12
41075 Default is same as input.
41077 matrixin, min
41079 Set the input colorspace matrix.
41080 Possible value are:
41082 input
41084 709
41085 unspecified
41086 470bg
41087 170m
41088 2020_ncl
41089 2020_cl
41090 chromal, c
41091 Set the output chroma location.
41092 Possible values are:
41094 input
41096 left
41097 center
41098 topleft
41099 top
41100 bottomleft
41101 bottom
41102 chromalin, cin
41103 Set the input chroma location.
41104 Possible values are:
41106 input
41108 left
41109 center
41110 topleft
41111 top
41112 bottomleft
41113 bottom
41114 npl Set the nominal peak luminance.
41115 param_a
41117 Parameter A for scaling filters. Parameter "b" for bicubic, and the
41118 number of filter taps for lanczos.
41119 param_b
41121 Parameter B for scaling filters. Parameter "c" for bicubic.
41122 The values of the w and h options are expressions containing the
41124 following constants:
41125 in_w
41127 in_h
41128 The input width and height
41129 iw
41131 ih These are the same as in_w and in_h.
41132 out_w
41134 out_h
41135 The output (scaled) width and height
41136 ow
41138 oh These are the same as out_w and out_h
41139 a The same as iw / ih
41141 sar input sample aspect ratio
41143 dar The input display aspect ratio. Calculated from "(iw / ih) * sar".
41145 hsub
41147 vsub
41148 horizontal and vertical input chroma subsample values. For example
41149 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
41150 ohsub
41152 ovsub
41153 horizontal and vertical output chroma subsample values. For example
41154 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
41155 Commands
41157 This filter supports the following commands:
41159 width, w
41161 height, h
41162 Set the output video dimension expression. The command accepts the
41163 same syntax of the corresponding option.
41164 If the specified expression is not valid, it is kept at its current
41166 value.
41167
41169 Below is a description of the currently available OpenCL video filters.
41170 To enable compilation of these filters you need to configure FFmpeg
41172 with "--enable-opencl".
41173 Running OpenCL filters requires you to initialize a hardware device and
41175 to pass that device to all filters in any filter graph.
41176 -init_hw_device opencl[=name][:device[,key=value...]]
41178 Initialise a new hardware device of type opencl called name, using
41179 the given device parameters.
41180 -filter_hw_device name
41182 Pass the hardware device called name to all filters in any filter
41183 graph.
41184 For more detailed information see
41186 <https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options>
41187 • Example of choosing the first device on the second platform and
41189 running avgblur_opencl filter with default parameters on it.
41190 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
41192 Since OpenCL filters are not able to access frame data in normal
41194 memory, all frame data needs to be uploaded(hwupload) to hardware
41195 surfaces connected to the appropriate device before being used and then
41196 downloaded(hwdownload) back to normal memory. Note that hwupload will
41197 upload to a surface with the same layout as the software frame, so it
41198 may be necessary to add a format filter immediately before to get the
41199 input into the right format and hwdownload does not support all formats
41200 on the output - it may be necessary to insert an additional format
41201 filter immediately following in the graph to get the output in a
41202 supported format.
41203 avgblur_opencl
41205 Apply average blur filter.
41206 The filter accepts the following options:
41208 sizeX
41210 Set horizontal radius size. Range is "[1, 1024]" and default value
41211 is 1.
41212 planes
41214 Set which planes to filter. Default value is 0xf, by which all
41215 planes are processed.
41216 sizeY
41218 Set vertical radius size. Range is "[1, 1024]" and default value is
41219 0. If zero, "sizeX" value will be used.
41220 Example
41222 • Apply average blur filter with horizontal and vertical size of 3,
41224 setting each pixel of the output to the average value of the 7x7
41225 region centered on it in the input. For pixels on the edges of the
41226 image, the region does not extend beyond the image boundaries, and
41227 so out-of-range coordinates are not used in the calculations.
41228 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
41230 boxblur_opencl
41232 Apply a boxblur algorithm to the input video.
41233 It accepts the following parameters:
41235 luma_radius, lr
41237 luma_power, lp
41238 chroma_radius, cr
41239 chroma_power, cp
41240 alpha_radius, ar
41241 alpha_power, ap
41242 A description of the accepted options follows.
41244 luma_radius, lr
41246 chroma_radius, cr
41247 alpha_radius, ar
41248 Set an expression for the box radius in pixels used for blurring
41249 the corresponding input plane.
41250 The radius value must be a non-negative number, and must not be
41252 greater than the value of the expression "min(w,h)/2" for the luma
41253 and alpha planes, and of "min(cw,ch)/2" for the chroma planes.
41254 Default value for luma_radius is "2". If not specified,
41256 chroma_radius and alpha_radius default to the corresponding value
41257 set for luma_radius.
41258 The expressions can contain the following constants:
41260 w
41262 h The input width and height in pixels.
41263 cw
41265 ch The input chroma image width and height in pixels.
41266 hsub
41268 vsub
41269 The horizontal and vertical chroma subsample values. For
41270 example, for the pixel format "yuv422p", hsub is 2 and vsub is
41271 1.
41272 luma_power, lp
41274 chroma_power, cp
41275 alpha_power, ap
41276 Specify how many times the boxblur filter is applied to the
41277 corresponding plane.
41278 Default value for luma_power is 2. If not specified, chroma_power
41280 and alpha_power default to the corresponding value set for
41281 luma_power.
41282 A value of 0 will disable the effect.
41284 Examples
41286 Apply boxblur filter, setting each pixel of the output to the average
41288 value of box-radiuses luma_radius, chroma_radius, alpha_radius for each
41289 plane respectively. The filter will apply luma_power, chroma_power,
41290 alpha_power times onto the corresponding plane. For pixels on the edges
41291 of the image, the radius does not extend beyond the image boundaries,
41292 and so out-of-range coordinates are not used in the calculations.
41293 • Apply a boxblur filter with the luma, chroma, and alpha radius set
41295 to 2 and luma, chroma, and alpha power set to 3. The filter will
41296 run 3 times with box-radius set to 2 for every plane of the image.
41297 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
41299 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
41300 • Apply a boxblur filter with luma radius set to 2, luma_power to 1,
41302 chroma_radius to 4, chroma_power to 5, alpha_radius to 3 and
41303 alpha_power to 7.
41304 For the luma plane, a 2x2 box radius will be run once.
41306 For the chroma plane, a 4x4 box radius will be run 5 times.
41308 For the alpha plane, a 3x3 box radius will be run 7 times.
41310 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
41312 colorkey_opencl
41314 RGB colorspace color keying.
41315 The filter accepts the following options:
41317 color
41319 The color which will be replaced with transparency.
41320 similarity
41322 Similarity percentage with the key color.
41323 0.01 matches only the exact key color, while 1.0 matches
41325 everything.
41326 blend
41328 Blend percentage.
41329 0.0 makes pixels either fully transparent, or not transparent at
41331 all.
41332 Higher values result in semi-transparent pixels, with a higher
41334 transparency the more similar the pixels color is to the key color.
41335 Examples
41337 • Make every semi-green pixel in the input transparent with some
41339 slight blending:
41340 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
41342 convolution_opencl
41344 Apply convolution of 3x3, 5x5, 7x7 matrix.
41345 The filter accepts the following options:
41347 0m
41349 1m
41350 2m
41351 3m Set matrix for each plane. Matrix is sequence of 9, 25 or 49
41352 signed numbers. Default value for each plane is "0 0 0 0 1 0 0 0
41353 0".
41354 0rdiv
41356 1rdiv
41357 2rdiv
41358 3rdiv
41359 Set multiplier for calculated value for each plane. If unset or 0,
41360 it will be sum of all matrix elements. The option value must be a
41361 float number greater or equal to 0.0. Default value is 1.0.
41362 0bias
41364 1bias
41365 2bias
41366 3bias
41367 Set bias for each plane. This value is added to the result of the
41368 multiplication. Useful for making the overall image brighter or
41369 darker. The option value must be a float number greater or equal
41370 to 0.0. Default value is 0.0.
41371 Examples
41373 • Apply sharpen:
41375 -i INPUT -vf "hwupload, convolution_opencl=0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0, hwdownload" OUTPUT
41377 • Apply blur:
41379 -i INPUT -vf "hwupload, convolution_opencl=1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1/9:1/9:1/9:1/9, hwdownload" OUTPUT
41381 • Apply edge enhance:
41383 -i INPUT -vf "hwupload, convolution_opencl=0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:5:1:1:1:0:128:128:128, hwdownload" OUTPUT
41385 • Apply edge detect:
41387 -i INPUT -vf "hwupload, convolution_opencl=0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:5:5:5:1:0:128:128:128, hwdownload" OUTPUT
41389 • Apply laplacian edge detector which includes diagonals:
41391 -i INPUT -vf "hwupload, convolution_opencl=1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:5:5:5:1:0:128:128:0, hwdownload" OUTPUT
41393 • Apply emboss:
41395 -i INPUT -vf "hwupload, convolution_opencl=-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2, hwdownload" OUTPUT
41397 erosion_opencl
41399 Apply erosion effect to the video.
41400 This filter replaces the pixel by the local(3x3) minimum.
41402 It accepts the following options:
41404 threshold0
41406 threshold1
41407 threshold2
41408 threshold3
41409 Limit the maximum change for each plane. Range is "[0, 65535]" and
41410 default value is 65535. If 0, plane will remain unchanged.
41411 coordinates
41413 Flag which specifies the pixel to refer to. Range is "[0, 255]"
41414 and default value is 255, i.e. all eight pixels are used.
41415 Flags to local 3x3 coordinates region centered on "x":
41417 1 2 3
41419 4 x 5
41421 6 7 8
41423 Example
41425 • Apply erosion filter with threshold0 set to 30, threshold1 set 40,
41427 threshold2 set to 50 and coordinates set to 231, setting each pixel
41428 of the output to the local minimum between pixels: 1, 2, 3, 6, 7, 8
41429 of the 3x3 region centered on it in the input. If the difference
41430 between input pixel and local minimum is more then threshold of the
41431 corresponding plane, output pixel will be set to input pixel -
41432 threshold of corresponding plane.
41433 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
41435 deshake_opencl
41437 Feature-point based video stabilization filter.
41438 The filter accepts the following options:
41440 tripod
41442 Simulates a tripod by preventing any camera movement whatsoever
41443 from the original frame. Defaults to 0.
41444 debug
41446 Whether or not additional debug info should be displayed, both in
41447 the processed output and in the console.
41448 Note that in order to see console debug output you will also need
41450 to pass "-v verbose" to ffmpeg.
41451 Viewing point matches in the output video is only supported for RGB
41453 input.
41454 Defaults to 0.
41456 adaptive_crop
41458 Whether or not to do a tiny bit of cropping at the borders to cut
41459 down on the amount of mirrored pixels.
41460 Defaults to 1.
41462 refine_features
41464 Whether or not feature points should be refined at a sub-pixel
41465 level.
41466 This can be turned off for a slight performance gain at the cost of
41468 precision.
41469 Defaults to 1.
41471 smooth_strength
41473 The strength of the smoothing applied to the camera path from 0.0
41474 to 1.0.
41475 1.0 is the maximum smoothing strength while values less than that
41477 result in less smoothing.
41478 0.0 causes the filter to adaptively choose a smoothing strength on
41480 a per-frame basis.
41481 Defaults to 0.0.
41483 smooth_window_multiplier
41485 Controls the size of the smoothing window (the number of frames
41486 buffered to determine motion information from).
41487 The size of the smoothing window is determined by multiplying the
41489 framerate of the video by this number.
41490 Acceptable values range from 0.1 to 10.0.
41492 Larger values increase the amount of motion data available for
41494 determining how to smooth the camera path, potentially improving
41495 smoothness, but also increase latency and memory usage.
41496 Defaults to 2.0.
41498 Examples
41500 • Stabilize a video with a fixed, medium smoothing strength:
41502 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
41504 • Stabilize a video with debugging (both in console and in rendered
41506 video):
41507 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
41509 dilation_opencl
41511 Apply dilation effect to the video.
41512 This filter replaces the pixel by the local(3x3) maximum.
41514 It accepts the following options:
41516 threshold0
41518 threshold1
41519 threshold2
41520 threshold3
41521 Limit the maximum change for each plane. Range is "[0, 65535]" and
41522 default value is 65535. If 0, plane will remain unchanged.
41523 coordinates
41525 Flag which specifies the pixel to refer to. Range is "[0, 255]"
41526 and default value is 255, i.e. all eight pixels are used.
41527 Flags to local 3x3 coordinates region centered on "x":
41529 1 2 3
41531 4 x 5
41533 6 7 8
41535 Example
41537 • Apply dilation filter with threshold0 set to 30, threshold1 set 40,
41539 threshold2 set to 50 and coordinates set to 231, setting each pixel
41540 of the output to the local maximum between pixels: 1, 2, 3, 6, 7, 8
41541 of the 3x3 region centered on it in the input. If the difference
41542 between input pixel and local maximum is more then threshold of the
41543 corresponding plane, output pixel will be set to input pixel +
41544 threshold of corresponding plane.
41545 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
41547 nlmeans_opencl
41549 Non-local Means denoise filter through OpenCL, this filter accepts same
41550 options as nlmeans.
41551 overlay_opencl
41553 Overlay one video on top of another.
41554 It takes two inputs and has one output. The first input is the "main"
41556 video on which the second input is overlaid. This filter requires same
41557 memory layout for all the inputs. So, format conversion may be needed.
41558 The filter accepts the following options:
41560 x Set the x coordinate of the overlaid video on the main video.
41562 Default value is 0.
41563 y Set the y coordinate of the overlaid video on the main video.
41565 Default value is 0.
41566 Examples
41568 • Overlay an image LOGO at the top-left corner of the INPUT video.
41570 Both inputs are yuv420p format.
41571 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
41573 • The inputs have same memory layout for color channels , the overlay
41575 has additional alpha plane, like INPUT is yuv420p, and the LOGO is
41576 yuva420p.
41577 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
41579 pad_opencl
41581 Add paddings to the input image, and place the original input at the
41582 provided x, y coordinates.
41583 It accepts the following options:
41585 width, w
41587 height, h
41588 Specify an expression for the size of the output image with the
41589 paddings added. If the value for width or height is 0, the
41590 corresponding input size is used for the output.
41591 The width expression can reference the value set by the height
41593 expression, and vice versa.
41594 The default value of width and height is 0.
41596 x
41598 y Specify the offsets to place the input image at within the padded
41599 area, with respect to the top/left border of the output image.
41600 The x expression can reference the value set by the y expression,
41602 and vice versa.
41603 The default value of x and y is 0.
41605 If x or y evaluate to a negative number, they'll be changed so the
41607 input image is centered on the padded area.
41608 color
41610 Specify the color of the padded area. For the syntax of this
41611 option, check the "Color" section in the ffmpeg-utils manual.
41612 aspect
41614 Pad to an aspect instead to a resolution.
41615 The value for the width, height, x, and y options are expressions
41617 containing the following constants:
41618 in_w
41620 in_h
41621 The input video width and height.
41622 iw
41624 ih These are the same as in_w and in_h.
41625 out_w
41627 out_h
41628 The output width and height (the size of the padded area), as
41629 specified by the width and height expressions.
41630 ow
41632 oh These are the same as out_w and out_h.
41633 x
41635 y The x and y offsets as specified by the x and y expressions, or NAN
41636 if not yet specified.
41637 a same as iw / ih
41639 sar input sample aspect ratio
41641 dar input display aspect ratio, it is the same as (iw / ih) * sar
41643 prewitt_opencl
41645 Apply the Prewitt operator
41646 (<https://en.wikipedia.org/wiki/Prewitt_operator>) to input video
41647 stream.
41648 The filter accepts the following option:
41650 planes
41652 Set which planes to filter. Default value is 0xf, by which all
41653 planes are processed.
41654 scale
41656 Set value which will be multiplied with filtered result. Range is
41657 "[0.0, 65535]" and default value is 1.0.
41658 delta
41660 Set value which will be added to filtered result. Range is
41661 "[-65535, 65535]" and default value is 0.0.
41662 Example
41664 • Apply the Prewitt operator with scale set to 2 and delta set to 10.
41666 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
41668 program_opencl
41670 Filter video using an OpenCL program.
41671 source
41673 OpenCL program source file.
41674 kernel
41676 Kernel name in program.
41677 inputs
41679 Number of inputs to the filter. Defaults to 1.
41680 size, s
41682 Size of output frames. Defaults to the same as the first input.
41683 The "program_opencl" filter also supports the framesync options.
41685 The program source file must contain a kernel function with the given
41687 name, which will be run once for each plane of the output. Each run on
41688 a plane gets enqueued as a separate 2D global NDRange with one work-
41689 item for each pixel to be generated. The global ID offset for each
41690 work-item is therefore the coordinates of a pixel in the destination
41691 image.
41692 The kernel function needs to take the following arguments:
41694 • Destination image, __write_only image2d_t.
41696 This image will become the output; the kernel should write all of
41698 it.
41699 • Frame index, unsigned int.
41701 This is a counter starting from zero and increasing by one for each
41703 frame.
41704 • Source images, __read_only image2d_t.
41706 These are the most recent images on each input. The kernel may
41708 read from them to generate the output, but they can't be written
41709 to.
41710 Example programs:
41712 • Copy the input to the output (output must be the same size as the
41714 input).
41715 __kernel void copy(__write_only image2d_t destination,
41717 unsigned int index,
41718 __read_only image2d_t source)
41719 {
41720 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
41721 int2 location = (int2)(get_global_id(0), get_global_id(1));
41723 float4 value = read_imagef(source, sampler, location);
41725 write_imagef(destination, location, value);
41727 }
41728 • Apply a simple transformation, rotating the input by an amount
41730 increasing with the index counter. Pixel values are linearly
41731 interpolated by the sampler, and the output need not have the same
41732 dimensions as the input.
41733 __kernel void rotate_image(__write_only image2d_t dst,
41735 unsigned int index,
41736 __read_only image2d_t src)
41737 {
41738 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
41739 CLK_FILTER_LINEAR);
41740 float angle = (float)index / 100.0f;
41742 float2 dst_dim = convert_float2(get_image_dim(dst));
41744 float2 src_dim = convert_float2(get_image_dim(src));
41745 float2 dst_cen = dst_dim / 2.0f;
41747 float2 src_cen = src_dim / 2.0f;
41748 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
41750 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
41752 float2 src_pos = {
41753 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
41754 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
41755 };
41756 src_pos = src_pos * src_dim / dst_dim;
41757 float2 src_loc = src_pos + src_cen;
41759 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
41761 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
41762 write_imagef(dst, dst_loc, 0.5f);
41763 else
41764 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
41765 }
41766 • Blend two inputs together, with the amount of each input used
41768 varying with the index counter.
41769 __kernel void blend_images(__write_only image2d_t dst,
41771 unsigned int index,
41772 __read_only image2d_t src1,
41773 __read_only image2d_t src2)
41774 {
41775 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
41776 CLK_FILTER_LINEAR);
41777 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
41779 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
41781 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
41782 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
41783 float4 val1 = read_imagef(src1, sampler, src1_loc);
41785 float4 val2 = read_imagef(src2, sampler, src2_loc);
41786 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
41788 }
41789 remap_opencl
41791 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
41792 Destination pixel at position (X, Y) will be picked from source (x, y)
41794 position where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are
41795 out of range, zero value for pixel will be used for destination pixel.
41796 Xmap and Ymap input video streams must be of same dimensions. Output
41798 video stream will have Xmap/Ymap video stream dimensions. Xmap and
41799 Ymap input video streams are 32bit float pixel format, single channel.
41800 interp
41802 Specify interpolation used for remapping of pixels. Allowed values
41803 are "near" and "linear". Default value is "linear".
41804 fill
41806 Specify the color of the unmapped pixels. For the syntax of this
41807 option, check the "Color" section in the ffmpeg-utils manual.
41808 Default color is "black".
41809 roberts_opencl
41811 Apply the Roberts cross operator
41812 (<https://en.wikipedia.org/wiki/Roberts_cross>) to input video stream.
41813 The filter accepts the following option:
41815 planes
41817 Set which planes to filter. Default value is 0xf, by which all
41818 planes are processed.
41819 scale
41821 Set value which will be multiplied with filtered result. Range is
41822 "[0.0, 65535]" and default value is 1.0.
41823 delta
41825 Set value which will be added to filtered result. Range is
41826 "[-65535, 65535]" and default value is 0.0.
41827 Example
41829 • Apply the Roberts cross operator with scale set to 2 and delta set
41831 to 10
41832 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
41834 sobel_opencl
41836 Apply the Sobel operator
41837 (<https://en.wikipedia.org/wiki/Sobel_operator>) to input video stream.
41838 The filter accepts the following option:
41840 planes
41842 Set which planes to filter. Default value is 0xf, by which all
41843 planes are processed.
41844 scale
41846 Set value which will be multiplied with filtered result. Range is
41847 "[0.0, 65535]" and default value is 1.0.
41848 delta
41850 Set value which will be added to filtered result. Range is
41851 "[-65535, 65535]" and default value is 0.0.
41852 Example
41854 • Apply sobel operator with scale set to 2 and delta set to 10
41856 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
41858 tonemap_opencl
41860 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
41861 It accepts the following parameters:
41863 tonemap
41865 Specify the tone-mapping operator to be used. Same as tonemap
41866 option in tonemap.
41867 param
41869 Tune the tone mapping algorithm. same as param option in tonemap.
41870 desat
41872 Apply desaturation for highlights that exceed this level of
41873 brightness. The higher the parameter, the more color information
41874 will be preserved. This setting helps prevent unnaturally blown-out
41875 colors for super-highlights, by (smoothly) turning into white
41876 instead. This makes images feel more natural, at the cost of
41877 reducing information about out-of-range colors.
41878 The default value is 0.5, and the algorithm here is a little
41880 different from the cpu version tonemap currently. A setting of 0.0
41881 disables this option.
41882 threshold
41884 The tonemapping algorithm parameters is fine-tuned per each scene.
41885 And a threshold is used to detect whether the scene has changed or
41886 not. If the distance between the current frame average brightness
41887 and the current running average exceeds a threshold value, we would
41888 re-calculate scene average and peak brightness. The default value
41889 is 0.2.
41890 format
41892 Specify the output pixel format.
41893 Currently supported formats are:
41895 p010
41897 nv12
41898 range, r
41899 Set the output color range.
41900 Possible values are:
41902 tv/mpeg
41904 pc/jpeg
41905 Default is same as input.
41907 primaries, p
41909 Set the output color primaries.
41910 Possible values are:
41912 bt709
41914 bt2020
41915 Default is same as input.
41917 transfer, t
41919 Set the output transfer characteristics.
41920 Possible values are:
41922 bt709
41924 bt2020
41925 Default is bt709.
41927 matrix, m
41929 Set the output colorspace matrix.
41930 Possible value are:
41932 bt709
41934 bt2020
41935 Default is same as input.
41937 Example
41939 • Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010
41941 format using linear operator.
41942 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
41944 unsharp_opencl
41946 Sharpen or blur the input video.
41947 It accepts the following parameters:
41949 luma_msize_x, lx
41951 Set the luma matrix horizontal size. Range is "[1, 23]" and
41952 default value is 5.
41953 luma_msize_y, ly
41955 Set the luma matrix vertical size. Range is "[1, 23]" and default
41956 value is 5.
41957 luma_amount, la
41959 Set the luma effect strength. Range is "[-10, 10]" and default
41960 value is 1.0.
41961 Negative values will blur the input video, while positive values
41963 will sharpen it, a value of zero will disable the effect.
41964 chroma_msize_x, cx
41966 Set the chroma matrix horizontal size. Range is "[1, 23]" and
41967 default value is 5.
41968 chroma_msize_y, cy
41970 Set the chroma matrix vertical size. Range is "[1, 23]" and
41971 default value is 5.
41972 chroma_amount, ca
41974 Set the chroma effect strength. Range is "[-10, 10]" and default
41975 value is 0.0.
41976 Negative values will blur the input video, while positive values
41978 will sharpen it, a value of zero will disable the effect.
41979 All parameters are optional and default to the equivalent of the string
41981 '5:5:1.0:5:5:0.0'.
41982 Examples
41984 • Apply strong luma sharpen effect:
41986 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
41988 • Apply a strong blur of both luma and chroma parameters:
41990 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
41992 xfade_opencl
41994 Cross fade two videos with custom transition effect by using OpenCL.
41995 It accepts the following options:
41997 transition
41999 Set one of possible transition effects.
42000 custom
42002 Select custom transition effect, the actual transition
42003 description will be picked from source and kernel options.
42004 fade
42006 wipeleft
42007 wiperight
42008 wipeup
42009 wipedown
42010 slideleft
42011 slideright
42012 slideup
42013 slidedown
42014 Default transition is fade.
42015 source
42017 OpenCL program source file for custom transition.
42018 kernel
42020 Set name of kernel to use for custom transition from program source
42021 file.
42022 duration
42024 Set duration of video transition.
42025 offset
42027 Set time of start of transition relative to first video.
42028 The program source file must contain a kernel function with the given
42030 name, which will be run once for each plane of the output. Each run on
42031 a plane gets enqueued as a separate 2D global NDRange with one work-
42032 item for each pixel to be generated. The global ID offset for each
42033 work-item is therefore the coordinates of a pixel in the destination
42034 image.
42035 The kernel function needs to take the following arguments:
42037 • Destination image, __write_only image2d_t.
42039 This image will become the output; the kernel should write all of
42041 it.
42042 • First Source image, __read_only image2d_t. Second Source image,
42044 __read_only image2d_t.
42045 These are the most recent images on each input. The kernel may
42047 read from them to generate the output, but they can't be written
42048 to.
42049 • Transition progress, float. This value is always between 0 and 1
42051 inclusive.
42052 Example programs:
42054 • Apply dots curtain transition effect:
42056 __kernel void blend_images(__write_only image2d_t dst,
42058 __read_only image2d_t src1,
42059 __read_only image2d_t src2,
42060 float progress)
42061 {
42062 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
42063 CLK_FILTER_LINEAR);
42064 int2 p = (int2)(get_global_id(0), get_global_id(1));
42065 float2 rp = (float2)(get_global_id(0), get_global_id(1));
42066 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
42067 rp = rp / dim;
42068 float2 dots = (float2)(20.0, 20.0);
42070 float2 center = (float2)(0,0);
42071 float2 unused;
42072 float4 val1 = read_imagef(src1, sampler, p);
42074 float4 val2 = read_imagef(src2, sampler, p);
42075 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
42076 write_imagef(dst, p, next ? val1 : val2);
42078 }
42079
42081 VAAPI Video filters are usually used with VAAPI decoder and VAAPI
42082 encoder. Below is a description of VAAPI video filters.
42083 To enable compilation of these filters you need to configure FFmpeg
42085 with "--enable-vaapi".
42086 To use vaapi filters, you need to setup the vaapi device correctly. For
42088 more information, please read
42089 <https://trac.ffmpeg.org/wiki/Hardware/VAAPI>
42090 overlay_vaapi
42092 Overlay one video on the top of another.
42093 It takes two inputs and has one output. The first input is the "main"
42095 video on which the second input is overlaid.
42096 The filter accepts the following options:
42098 x
42100 y Set expressions for the x and y coordinates of the overlaid video
42101 on the main video.
42102 Default value is "0" for both expressions.
42104 w
42106 h Set expressions for the width and height the overlaid video on the
42107 main video.
42108 Default values are 'overlay_iw' for 'w' and
42110 'overlay_ih*w/overlay_iw' for 'h'.
42111 The expressions can contain the following parameters:
42113 main_w, W
42115 main_h, H
42116 The main input width and height.
42117 overlay_iw
42119 overlay_ih
42120 The overlay input width and height.
42121 overlay_w, w
42123 overlay_h, h
42124 The overlay output width and height.
42125 overlay_x, x
42127 overlay_y, y
42128 Position of the overlay layer inside of main
42129 alpha
42131 Set transparency of overlaid video. Allowed range is 0.0 to 1.0.
42132 Higher value means lower transparency. Default value is 1.0.
42133 eof_action
42135 See framesync.
42136 shortest
42138 See framesync.
42139 repeatlast
42141 See framesync.
42142 This filter also supports the framesync options.
42144 Examples
42146 • Overlay an image LOGO at the top-left corner of the INPUT video.
42148 Both inputs for this filter are yuv420p format.
42149 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_vaapi" OUTPUT
42151 • Overlay an image LOGO at the offset (200, 100) from the top-left
42153 corner of the INPUT video. The inputs have same memory layout for
42154 color channels, the overlay has additional alpha plane, like INPUT
42155 is yuv420p, and the LOGO is yuva420p.
42156 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_vaapi=x=200:y=100:w=400:h=300:alpha=1.0, hwdownload, format=nv12" OUTPUT
42158 tonemap_vaapi
42160 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range)
42161 conversion with tone-mapping. It maps the dynamic range of HDR10
42162 content to the SDR content. It currently only accepts HDR10 as input.
42163 It accepts the following parameters:
42165 format
42167 Specify the output pixel format.
42168 Currently supported formats are:
42170 p010
42172 nv12
42173 Default is nv12.
42175 primaries, p
42177 Set the output color primaries.
42178 Default is same as input.
42180 transfer, t
42182 Set the output transfer characteristics.
42183 Default is bt709.
42185 matrix, m
42187 Set the output colorspace matrix.
42188 Default is same as input.
42190 Example
42192 • Convert HDR(HDR10) video to bt2020-transfer-characteristic p010
42194 format
42195 tonemap_vaapi=format=p010:t=bt2020-10
42197 hstack_vaapi
42199 Stack input videos horizontally.
42200 This is the VA-API variant of the hstack filter, each input stream may
42202 have different height, this filter will scale down/up each input stream
42203 while keeping the orignal aspect.
42204 It accepts the following options:
42206 inputs
42208 See hstack.
42209 shortest
42211 See hstack.
42212 height
42214 Set height of output. If set to 0, this filter will set height of
42215 output to height of the first input stream. Default value is 0.
42216 vstack_vaapi
42218 Stack input videos vertically.
42219 This is the VA-API variant of the vstack filter, each input stream may
42221 have different width, this filter will scale down/up each input stream
42222 while keeping the orignal aspect.
42223 It accepts the following options:
42225 inputs
42227 See vstack.
42228 shortest
42230 See vstack.
42231 width
42233 Set width of output. If set to 0, this filter will set width of
42234 output to width of the first input stream. Default value is 0.
42235 xstack_vaapi
42237 Stack video inputs into custom layout.
42238 This is the VA-API variant of the xstack filter, each input stream may
42240 have different size, this filter will scale down/up each input stream
42241 to the given output size, or the size of the first input stream.
42242 It accepts the following options:
42244 inputs
42246 See xstack.
42247 shortest
42249 See xstack.
42250 layout
42252 See xstack. Moreover, this permits the user to supply output size
42253 for each input stream.
42254 xstack_vaapi=inputs=4:layout=0_0_1920x1080|0_h0_1920x1080|w0_0_1920x1080|w0_h0_1920x1080
42256 grid
42258 See xstack.
42259 grid_tile_size
42261 Set output size for each input stream when grid is set. If this
42262 option is not set, this filter will set output size by default to
42263 the size of the first input stream. For the syntax of this option,
42264 check the "Video size" section in the ffmpeg-utils manual.
42265 fill
42267 See xstack.
42268
42270 Below is a description of the currently available QSV video filters.
42271 To enable compilation of these filters you need to configure FFmpeg
42273 with "--enable-libmfx" or "--enable-libvpl".
42274 To use QSV filters, you need to setup the QSV device correctly. For
42276 more information, please read
42277 <https://trac.ffmpeg.org/wiki/Hardware/QuickSync>
42278 hstack_qsv
42280 Stack input videos horizontally.
42281 This is the QSV variant of the hstack filter, each input stream may
42283 have different height, this filter will scale down/up each input stream
42284 while keeping the orignal aspect.
42285 It accepts the following options:
42287 inputs
42289 See hstack.
42290 shortest
42292 See hstack.
42293 height
42295 Set height of output. If set to 0, this filter will set height of
42296 output to height of the first input stream. Default value is 0.
42297 vstack_qsv
42299 Stack input videos vertically.
42300 This is the QSV variant of the vstack filter, each input stream may
42302 have different width, this filter will scale down/up each input stream
42303 while keeping the orignal aspect.
42304 It accepts the following options:
42306 inputs
42308 See vstack.
42309 shortest
42311 See vstack.
42312 width
42314 Set width of output. If set to 0, this filter will set width of
42315 output to width of the first input stream. Default value is 0.
42316 xstack_qsv
42318 Stack video inputs into custom layout.
42319 This is the QSV variant of the xstack filter.
42321 It accepts the following options:
42323 inputs
42325 See xstack.
42326 shortest
42328 See xstack.
42329 layout
42331 See xstack. Moreover, this permits the user to supply output size
42332 for each input stream.
42333 xstack_qsv=inputs=4:layout=0_0_1920x1080|0_h0_1920x1080|w0_0_1920x1080|w0_h0_1920x1080
42335 grid
42337 See xstack.
42338 grid_tile_size
42340 Set output size for each input stream when grid is set. If this
42341 option is not set, this filter will set output size by default to
42342 the size of the first input stream. For the syntax of this option,
42343 check the "Video size" section in the ffmpeg-utils manual.
42344 fill
42346 See xstack.
42347
42349 Below is a description of the currently available video sources.
42350 buffer
42352 Buffer video frames, and make them available to the filter chain.
42353 This source is mainly intended for a programmatic use, in particular
42355 through the interface defined in libavfilter/buffersrc.h.
42356 It accepts the following parameters:
42358 video_size
42360 Specify the size (width and height) of the buffered video frames.
42361 For the syntax of this option, check the "Video size" section in
42362 the ffmpeg-utils manual.
42363 width
42365 The input video width.
42366 height
42368 The input video height.
42369 pix_fmt
42371 A string representing the pixel format of the buffered video
42372 frames. It may be a number corresponding to a pixel format, or a
42373 pixel format name.
42374 time_base
42376 Specify the timebase assumed by the timestamps of the buffered
42377 frames.
42378 frame_rate
42380 Specify the frame rate expected for the video stream.
42381 pixel_aspect, sar
42383 The sample (pixel) aspect ratio of the input video.
42384 hw_frames_ctx
42386 When using a hardware pixel format, this should be a reference to
42387 an AVHWFramesContext describing input frames.
42388 For example:
42390 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
42392 will instruct the source to accept video frames with size 320x240 and
42394 with format "yuv410p", assuming 1/24 as the timestamps timebase and
42395 square pixels (1:1 sample aspect ratio). Since the pixel format with
42396 name "yuv410p" corresponds to the number 6 (check the enum
42397 AVPixelFormat definition in libavutil/pixfmt.h), this example
42398 corresponds to:
42399 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
42401 Alternatively, the options can be specified as a flat string, but this
42403 syntax is deprecated:
42404 width:height:pix_fmt:time_base.num:time_base.den:pixel_aspect.num:pixel_aspect.den
42406 cellauto
42408 Create a pattern generated by an elementary cellular automaton.
42409 The initial state of the cellular automaton can be defined through the
42411 filename and pattern options. If such options are not specified an
42412 initial state is created randomly.
42413 At each new frame a new row in the video is filled with the result of
42415 the cellular automaton next generation. The behavior when the whole
42416 frame is filled is defined by the scroll option.
42417 This source accepts the following options:
42419 filename, f
42421 Read the initial cellular automaton state, i.e. the starting row,
42422 from the specified file. In the file, each non-whitespace
42423 character is considered an alive cell, a newline will terminate the
42424 row, and further characters in the file will be ignored.
42425 pattern, p
42427 Read the initial cellular automaton state, i.e. the starting row,
42428 from the specified string.
42429 Each non-whitespace character in the string is considered an alive
42431 cell, a newline will terminate the row, and further characters in
42432 the string will be ignored.
42433 rate, r
42435 Set the video rate, that is the number of frames generated per
42436 second. Default is 25.
42437 random_fill_ratio, ratio
42439 Set the random fill ratio for the initial cellular automaton row.
42440 It is a floating point number value ranging from 0 to 1, defaults
42441 to 1/PHI.
42442 This option is ignored when a file or a pattern is specified.
42444 random_seed, seed
42446 Set the seed for filling randomly the initial row, must be an
42447 integer included between 0 and UINT32_MAX. If not specified, or if
42448 explicitly set to -1, the filter will try to use a good random seed
42449 on a best effort basis.
42450 rule
42452 Set the cellular automaton rule, it is a number ranging from 0 to
42453 255. Default value is 110.
42454 size, s
42456 Set the size of the output video. For the syntax of this option,
42457 check the "Video size" section in the ffmpeg-utils manual.
42458 If filename or pattern is specified, the size is set by default to
42460 the width of the specified initial state row, and the height is set
42461 to width * PHI.
42462 If size is set, it must contain the width of the specified pattern
42464 string, and the specified pattern will be centered in the larger
42465 row.
42466 If a filename or a pattern string is not specified, the size value
42468 defaults to "320x518" (used for a randomly generated initial
42469 state).
42470 scroll
42472 If set to 1, scroll the output upward when all the rows in the
42473 output have been already filled. If set to 0, the new generated row
42474 will be written over the top row just after the bottom row is
42475 filled. Defaults to 1.
42476 start_full, full
42478 If set to 1, completely fill the output with generated rows before
42479 outputting the first frame. This is the default behavior, for
42480 disabling set the value to 0.
42481 stitch
42483 If set to 1, stitch the left and right row edges together. This is
42484 the default behavior, for disabling set the value to 0.
42485 Examples
42487 • Read the initial state from pattern, and specify an output of size
42489 200x400.
42490 cellauto=f=pattern:s=200x400
42492 • Generate a random initial row with a width of 200 cells, with a
42494 fill ratio of 2/3:
42495 cellauto=ratio=2/3:s=200x200
42497 • Create a pattern generated by rule 18 starting by a single alive
42499 cell centered on an initial row with width 100:
42500 cellauto=p=@s=100x400:full=0:rule=18
42502 • Specify a more elaborated initial pattern:
42504 cellauto=p='@@ @ @@':s=100x400:full=0:rule=18
42506 coreimagesrc
42508 Video source generated on GPU using Apple's CoreImage API on OSX.
42509 This video source is a specialized version of the coreimage video
42511 filter. Use a core image generator at the beginning of the applied
42512 filterchain to generate the content.
42513 The coreimagesrc video source accepts the following options:
42515 list_generators
42517 List all available generators along with all their respective
42518 options as well as possible minimum and maximum values along with
42519 the default values.
42520 list_generators=true
42522 size, s
42524 Specify the size of the sourced video. For the syntax of this
42525 option, check the "Video size" section in the ffmpeg-utils manual.
42526 The default value is "320x240".
42527 rate, r
42529 Specify the frame rate of the sourced video, as the number of
42530 frames generated per second. It has to be a string in the format
42531 frame_rate_num/frame_rate_den, an integer number, a floating point
42532 number or a valid video frame rate abbreviation. The default value
42533 is "25".
42534 sar Set the sample aspect ratio of the sourced video.
42536 duration, d
42538 Set the duration of the sourced video. See the Time duration
42539 section in the ffmpeg-utils(1) manual for the accepted syntax.
42540 If not specified, or the expressed duration is negative, the video
42542 is supposed to be generated forever.
42543 Additionally, all options of the coreimage video filter are accepted.
42545 A complete filterchain can be used for further processing of the
42546 generated input without CPU-HOST transfer. See coreimage documentation
42547 and examples for details.
42548 Examples
42550 • Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
42552 given as complete and escaped command-line for Apple's standard
42553 bash shell:
42554 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@inputMessage=https\\\\\://FFmpeg.org/@inputCorrectionLevel=H -frames:v 1 QRCode.png
42556 This example is equivalent to the QRCode example of coreimage
42558 without the need for a nullsrc video source.
42559 ddagrab
42561 Captures the Windows Desktop via Desktop Duplication API.
42562 The filter exclusively returns D3D11 Hardware Frames, for on-gpu
42564 encoding or processing. So an explicit hwdownload is needed for any
42565 kind of software processing.
42566 It accepts the following options:
42568 output_idx
42570 DXGI Output Index to capture.
42571 Usually corresponds to the index Windows has given the screen minus
42573 one, so it's starting at 0.
42574 Defaults to output 0.
42576 draw_mouse
42578 Whether to draw the mouse cursor.
42579 Defaults to true.
42581 Only affects hardware cursors. If a game or application renders its
42583 own cursor, it'll always be captured.
42584 framerate
42586 Framerate at which the desktop will be captured.
42587 Defaults to 30 FPS.
42589 video_size
42591 Specify the size of the captured video.
42592 Defaults to the full size of the screen.
42594 Cropped from the bottom/right if smaller than screen size.
42596 offset_x
42598 Horizontal offset of the captured video.
42599 offset_y
42601 Vertical offset of the captured video.
42602 output_fmt
42604 Desired filter output format. Defaults to 8 Bit BGRA.
42605 It accepts the following values:
42607 auto
42609 Passes all supported output formats to DDA and returns what DDA
42610 decides to use.
42611 8bit
42613 bgra
42614 8 Bit formats always work, and DDA will convert to them if
42615 neccesary.
42616 10bit
42618 x2bgr10
42619 Filter initialization will fail if 10 bit format is requested
42620 but unavailable.
42621 Examples
42623 Capture primary screen and encode using nvenc:
42625 ffmpeg -f lavfi -i ddagrab -c:v h264_nvenc -cq 18 output.mp4
42627 You can also skip the lavfi device and directly use the filter. Also
42629 demonstrates downloading the frame and encoding with libx264. Explicit
42630 output format specification is required in this case:
42631 ffmpeg -filter_complex ddagrab=output_idx=1:framerate=60,hwdownload,format=bgra -c:v libx264 -crf 18 output.mp4
42633 If you want to capture only a subsection of the desktop, this can be
42635 achieved by specifying a smaller size and its offsets into the screen:
42636 ddagrab=video_size=800x600:offset_x=100:offset_y=100
42638 gradients
42640 Generate several gradients.
42641 size, s
42643 Set frame size. For the syntax of this option, check the "Video
42644 size" section in the ffmpeg-utils manual. Default value is
42645 "640x480".
42646 rate, r
42648 Set frame rate, expressed as number of frames per second. Default
42649 value is "25".
42650 c0, c1, c2, c3, c4, c5, c6, c7
42652 Set 8 colors. Default values for colors is to pick random one.
42653 x0, y0, y0, y1
42655 Set gradient line source and destination points. If negative or out
42656 of range, random ones are picked.
42657 nb_colors, n
42659 Set number of colors to use at once. Allowed range is from 2 to 8.
42660 Default value is 2.
42661 seed
42663 Set seed for picking gradient line points.
42664 duration, d
42666 Set the duration of the sourced video. See the Time duration
42667 section in the ffmpeg-utils(1) manual for the accepted syntax.
42668 If not specified, or the expressed duration is negative, the video
42670 is supposed to be generated forever.
42671 speed
42673 Set speed of gradients rotation.
42674 type, t
42676 Set type of gradients, can be "linear" or "radial" or "circular" or
42677 "spiral".
42678 mandelbrot
42680 Generate a Mandelbrot set fractal, and progressively zoom towards the
42681 point specified with start_x and start_y.
42682 This source accepts the following options:
42684 end_pts
42686 Set the terminal pts value. Default value is 400.
42687 end_scale
42689 Set the terminal scale value. Must be a floating point value.
42690 Default value is 0.3.
42691 inner
42693 Set the inner coloring mode, that is the algorithm used to draw the
42694 Mandelbrot fractal internal region.
42695 It shall assume one of the following values:
42697 black
42699 Set black mode.
42700 convergence
42702 Show time until convergence.
42703 mincol
42705 Set color based on point closest to the origin of the
42706 iterations.
42707 period
42709 Set period mode.
42710 Default value is mincol.
42712 bailout
42714 Set the bailout value. Default value is 10.0.
42715 maxiter
42717 Set the maximum of iterations performed by the rendering algorithm.
42718 Default value is 7189.
42719 outer
42721 Set outer coloring mode. It shall assume one of following values:
42722 iteration_count
42724 Set iteration count mode.
42725 normalized_iteration_count
42727 set normalized iteration count mode.
42728 Default value is normalized_iteration_count.
42730 rate, r
42732 Set frame rate, expressed as number of frames per second. Default
42733 value is "25".
42734 size, s
42736 Set frame size. For the syntax of this option, check the "Video
42737 size" section in the ffmpeg-utils manual. Default value is
42738 "640x480".
42739 start_scale
42741 Set the initial scale value. Default value is 3.0.
42742 start_x
42744 Set the initial x position. Must be a floating point value between
42745 -100 and 100. Default value is
42746 -0.743643887037158704752191506114774.
42747 start_y
42749 Set the initial y position. Must be a floating point value between
42750 -100 and 100. Default value is
42751 -0.131825904205311970493132056385139.
42752 mptestsrc
42754 Generate various test patterns, as generated by the MPlayer test
42755 filter.
42756 The size of the generated video is fixed, and is 256x256. This source
42758 is useful in particular for testing encoding features.
42759 This source accepts the following options:
42761 rate, r
42763 Specify the frame rate of the sourced video, as the number of
42764 frames generated per second. It has to be a string in the format
42765 frame_rate_num/frame_rate_den, an integer number, a floating point
42766 number or a valid video frame rate abbreviation. The default value
42767 is "25".
42768 duration, d
42770 Set the duration of the sourced video. See the Time duration
42771 section in the ffmpeg-utils(1) manual for the accepted syntax.
42772 If not specified, or the expressed duration is negative, the video
42774 is supposed to be generated forever.
42775 test, t
42777 Set the number or the name of the test to perform. Supported tests
42778 are:
42779 dc_luma
42781 dc_chroma
42782 freq_luma
42783 freq_chroma
42784 amp_luma
42785 amp_chroma
42786 cbp
42787 mv
42788 ring1
42789 ring2
42790 all
42791 max_frames, m
42792 Set the maximum number of frames generated for each test,
42793 default value is 30.
42794 Default value is "all", which will cycle through the list of all
42796 tests.
42797 Some examples:
42799 mptestsrc=t=dc_luma
42801 will generate a "dc_luma" test pattern.
42803 frei0r_src
42805 Provide a frei0r source.
42806 To enable compilation of this filter you need to install the frei0r
42808 header and configure FFmpeg with "--enable-frei0r".
42809 This source accepts the following parameters:
42811 size
42813 The size of the video to generate. For the syntax of this option,
42814 check the "Video size" section in the ffmpeg-utils manual.
42815 framerate
42817 The framerate of the generated video. It may be a string of the
42818 form num/den or a frame rate abbreviation.
42819 filter_name
42821 The name to the frei0r source to load. For more information
42822 regarding frei0r and how to set the parameters, read the frei0r
42823 section in the video filters documentation.
42824 filter_params
42826 A '|'-separated list of parameters to pass to the frei0r source.
42827 For example, to generate a frei0r partik0l source with size 200x200 and
42829 frame rate 10 which is overlaid on the overlay filter main input:
42830 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
42832 life
42834 Generate a life pattern.
42835 This source is based on a generalization of John Conway's life game.
42837 The sourced input represents a life grid, each pixel represents a cell
42839 which can be in one of two possible states, alive or dead. Every cell
42840 interacts with its eight neighbours, which are the cells that are
42841 horizontally, vertically, or diagonally adjacent.
42842 At each interaction the grid evolves according to the adopted rule,
42844 which specifies the number of neighbor alive cells which will make a
42845 cell stay alive or born. The rule option allows one to specify the rule
42846 to adopt.
42847 This source accepts the following options:
42849 filename, f
42851 Set the file from which to read the initial grid state. In the
42852 file, each non-whitespace character is considered an alive cell,
42853 and newline is used to delimit the end of each row.
42854 If this option is not specified, the initial grid is generated
42856 randomly.
42857 rate, r
42859 Set the video rate, that is the number of frames generated per
42860 second. Default is 25.
42861 random_fill_ratio, ratio
42863 Set the random fill ratio for the initial random grid. It is a
42864 floating point number value ranging from 0 to 1, defaults to 1/PHI.
42865 It is ignored when a file is specified.
42866 random_seed, seed
42868 Set the seed for filling the initial random grid, must be an
42869 integer included between 0 and UINT32_MAX. If not specified, or if
42870 explicitly set to -1, the filter will try to use a good random seed
42871 on a best effort basis.
42872 rule
42874 Set the life rule.
42875 A rule can be specified with a code of the kind "SNS/BNB", where NS
42877 and NB are sequences of numbers in the range 0-8, NS specifies the
42878 number of alive neighbor cells which make a live cell stay alive,
42879 and NB the number of alive neighbor cells which make a dead cell to
42880 become alive (i.e. to "born"). "s" and "b" can be used in place of
42881 "S" and "B", respectively.
42882 Alternatively a rule can be specified by an 18-bits integer. The 9
42884 high order bits are used to encode the next cell state if it is
42885 alive for each number of neighbor alive cells, the low order bits
42886 specify the rule for "borning" new cells. Higher order bits encode
42887 for an higher number of neighbor cells. For example the number
42888 6153 = "(12<<9)+9" specifies a stay alive rule of 12 and a born
42889 rule of 9, which corresponds to "S23/B03".
42890 Default value is "S23/B3", which is the original Conway's game of
42892 life rule, and will keep a cell alive if it has 2 or 3 neighbor
42893 alive cells, and will born a new cell if there are three alive
42894 cells around a dead cell.
42895 size, s
42897 Set the size of the output video. For the syntax of this option,
42898 check the "Video size" section in the ffmpeg-utils manual.
42899 If filename is specified, the size is set by default to the same
42901 size of the input file. If size is set, it must contain the size
42902 specified in the input file, and the initial grid defined in that
42903 file is centered in the larger resulting area.
42904 If a filename is not specified, the size value defaults to
42906 "320x240" (used for a randomly generated initial grid).
42907 stitch
42909 If set to 1, stitch the left and right grid edges together, and the
42910 top and bottom edges also. Defaults to 1.
42911 mold
42913 Set cell mold speed. If set, a dead cell will go from death_color
42914 to mold_color with a step of mold. mold can have a value from 0 to
42915 255.
42916 life_color
42918 Set the color of living (or new born) cells.
42919 death_color
42921 Set the color of dead cells. If mold is set, this is the first
42922 color used to represent a dead cell.
42923 mold_color
42925 Set mold color, for definitely dead and moldy cells.
42926 For the syntax of these 3 color options, check the "Color" section
42928 in the ffmpeg-utils manual.
42929 Examples
42931 • Read a grid from pattern, and center it on a grid of size 300x300
42933 pixels:
42934 life=f=pattern:s=300x300
42936 • Generate a random grid of size 200x200, with a fill ratio of 2/3:
42938 life=ratio=2/3:s=200x200
42940 • Specify a custom rule for evolving a randomly generated grid:
42942 life=rule=S14/B34
42944 • Full example with slow death effect (mold) using ffplay:
42946 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
42948 allrgb, allyuv, color, colorchart, colorspectrum, haldclutsrc, nullsrc,
42950 pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc,
42951 testsrc2, yuvtestsrc
42952 The "allrgb" source returns frames of size 4096x4096 of all rgb colors.
42953 The "allyuv" source returns frames of size 4096x4096 of all yuv colors.
42955 The "color" source provides an uniformly colored input.
42957 The "colorchart" source provides a colors checker chart.
42959 The "colorspectrum" source provides a color spectrum input.
42961 The "haldclutsrc" source provides an identity Hald CLUT. See also
42963 haldclut filter.
42964 The "nullsrc" source returns unprocessed video frames. It is mainly
42966 useful to be employed in analysis / debugging tools, or as the source
42967 for filters which ignore the input data.
42968 The "pal75bars" source generates a color bars pattern, based on EBU PAL
42970 recommendations with 75% color levels.
42971 The "pal100bars" source generates a color bars pattern, based on EBU
42973 PAL recommendations with 100% color levels.
42974 The "rgbtestsrc" source generates an RGB test pattern useful for
42976 detecting RGB vs BGR issues. You should see a red, green and blue
42977 stripe from top to bottom.
42978 The "smptebars" source generates a color bars pattern, based on the
42980 SMPTE Engineering Guideline EG 1-1990.
42981 The "smptehdbars" source generates a color bars pattern, based on the
42983 SMPTE RP 219-2002.
42984 The "testsrc" source generates a test video pattern, showing a color
42986 pattern, a scrolling gradient and a timestamp. This is mainly intended
42987 for testing purposes.
42988 The "testsrc2" source is similar to testsrc, but supports more pixel
42990 formats instead of just "rgb24". This allows using it as an input for
42991 other tests without requiring a format conversion.
42992 The "yuvtestsrc" source generates an YUV test pattern. You should see a
42994 y, cb and cr stripe from top to bottom.
42995 The sources accept the following parameters:
42997 level
42999 Specify the level of the Hald CLUT, only available in the
43000 "haldclutsrc" source. A level of "N" generates a picture of "N*N*N"
43001 by "N*N*N" pixels to be used as identity matrix for 3D lookup
43002 tables. Each component is coded on a "1/(N*N)" scale.
43003 color, c
43005 Specify the color of the source, only available in the "color"
43006 source. For the syntax of this option, check the "Color" section in
43007 the ffmpeg-utils manual.
43008 size, s
43010 Specify the size of the sourced video. For the syntax of this
43011 option, check the "Video size" section in the ffmpeg-utils manual.
43012 The default value is "320x240".
43013 This option is not available with the "allrgb", "allyuv", and
43015 "haldclutsrc" filters.
43016 rate, r
43018 Specify the frame rate of the sourced video, as the number of
43019 frames generated per second. It has to be a string in the format
43020 frame_rate_num/frame_rate_den, an integer number, a floating point
43021 number or a valid video frame rate abbreviation. The default value
43022 is "25".
43023 duration, d
43025 Set the duration of the sourced video. See the Time duration
43026 section in the ffmpeg-utils(1) manual for the accepted syntax.
43027 If not specified, or the expressed duration is negative, the video
43029 is supposed to be generated forever.
43030 Since the frame rate is used as time base, all frames including the
43032 last one will have their full duration. If the specified duration
43033 is not a multiple of the frame duration, it will be rounded up.
43034 sar Set the sample aspect ratio of the sourced video.
43036 alpha
43038 Specify the alpha (opacity) of the background, only available in
43039 the "testsrc2" source. The value must be between 0 (fully
43040 transparent) and 255 (fully opaque, the default).
43041 decimals, n
43043 Set the number of decimals to show in the timestamp, only available
43044 in the "testsrc" source.
43045 The displayed timestamp value will correspond to the original
43047 timestamp value multiplied by the power of 10 of the specified
43048 value. Default value is 0.
43049 type
43051 Set the type of the color spectrum, only available in the
43052 "colorspectrum" source. Can be one of the following:
43053 black
43055 white
43056 all
43057 patch_size
43058 Set patch size of single color patch, only available in the
43059 "colorchart" source. Default is "64x64".
43060 preset
43062 Set colorchecker colors preset, only available in the "colorchart"
43063 source.
43064 Available values are:
43066 reference
43068 skintones
43069 Default value is "reference".
43071 Examples
43073 • Generate a video with a duration of 5.3 seconds, with size 176x144
43075 and a frame rate of 10 frames per second:
43076 testsrc=duration=5.3:size=qcif:rate=10
43078 • The following graph description will generate a red source with an
43080 opacity of 0.2, with size "qcif" and a frame rate of 10 frames per
43081 second:
43082 color=c=red@0.2:s=qcif:r=10
43084 • If the input content is to be ignored, "nullsrc" can be used. The
43086 following command generates noise in the luminance plane by
43087 employing the "geq" filter:
43088 nullsrc=s=256x256, geq=random(1)*255:128:128
43090 Commands
43092 The "color" source supports the following commands:
43094 c, color
43096 Set the color of the created image. Accepts the same syntax of the
43097 corresponding color option.
43098 openclsrc
43100 Generate video using an OpenCL program.
43101 source
43103 OpenCL program source file.
43104 kernel
43106 Kernel name in program.
43107 size, s
43109 Size of frames to generate. This must be set.
43110 format
43112 Pixel format to use for the generated frames. This must be set.
43113 rate, r
43115 Number of frames generated every second. Default value is '25'.
43116 For details of how the program loading works, see the program_opencl
43118 filter.
43119 Example programs:
43121 • Generate a colour ramp by setting pixel values from the position of
43123 the pixel in the output image. (Note that this will work with all
43124 pixel formats, but the generated output will not be the same.)
43125 __kernel void ramp(__write_only image2d_t dst,
43127 unsigned int index)
43128 {
43129 int2 loc = (int2)(get_global_id(0), get_global_id(1));
43130 float4 val;
43132 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
43133 write_imagef(dst, loc, val);
43135 }
43136 • Generate a Sierpinski carpet pattern, panning by a single pixel
43138 each frame.
43139 __kernel void sierpinski_carpet(__write_only image2d_t dst,
43141 unsigned int index)
43142 {
43143 int2 loc = (int2)(get_global_id(0), get_global_id(1));
43144 float4 value = 0.0f;
43146 int x = loc.x + index;
43147 int y = loc.y + index;
43148 while (x > 0 || y > 0) {
43149 if (x % 3 == 1 && y % 3 == 1) {
43150 value = 1.0f;
43151 break;
43152 }
43153 x /= 3;
43154 y /= 3;
43155 }
43156 write_imagef(dst, loc, value);
43158 }
43159 sierpinski
43161 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
43162 This source accepts the following options:
43164 size, s
43166 Set frame size. For the syntax of this option, check the "Video
43167 size" section in the ffmpeg-utils manual. Default value is
43168 "640x480".
43169 rate, r
43171 Set frame rate, expressed as number of frames per second. Default
43172 value is "25".
43173 seed
43175 Set seed which is used for random panning.
43176 jump
43178 Set max jump for single pan destination. Allowed range is from 1 to
43179 10000.
43180 type
43182 Set fractal type, can be default "carpet" or "triangle".
43183
43185 Below is a description of the currently available video sinks.
43186 buffersink
43188 Buffer video frames, and make them available to the end of the filter
43189 graph.
43190 This sink is mainly intended for programmatic use, in particular
43192 through the interface defined in libavfilter/buffersink.h or the
43193 options system.
43194 It accepts a pointer to an AVBufferSinkContext structure, which defines
43196 the incoming buffers' formats, to be passed as the opaque parameter to
43197 "avfilter_init_filter" for initialization.
43198 nullsink
43200 Null video sink: do absolutely nothing with the input video. It is
43201 mainly useful as a template and for use in analysis / debugging tools.
43202
43204 Below is a description of the currently available multimedia filters.
43205 a3dscope
43207 Convert input audio to 3d scope video output.
43208 The filter accepts the following options:
43210 rate, r
43212 Set frame rate, expressed as number of frames per second. Default
43213 value is "25".
43214 size, s
43216 Specify the video size for the output. For the syntax of this
43217 option, check the "Video size" section in the ffmpeg-utils manual.
43218 Default value is "hd720".
43219 fov Set the camera field of view. Default is 90 degrees. Allowed range
43221 is from 40 to 150.
43222 roll
43224 Set the camera roll.
43225 pitch
43227 Set the camera pitch.
43228 yaw Set the camera yaw.
43230 xzoom
43232 Set the camera zoom on X-axis.
43233 yzoom
43235 Set the camera zoom on Y-axis.
43236 zzoom
43238 Set the camera zoom on Z-axis.
43239 xpos
43241 Set the camera position on X-axis.
43242 ypos
43244 Set the camera position on Y-axis.
43245 zpos
43247 Set the camera position on Z-axis.
43248 length
43250 Set the length of displayed audio waves in number of frames.
43251 Commands
43253 Filter supports the some above options as commands.
43255 abitscope
43257 Convert input audio to a video output, displaying the audio bit scope.
43258 The filter accepts the following options:
43260 rate, r
43262 Set frame rate, expressed as number of frames per second. Default
43263 value is "25".
43264 size, s
43266 Specify the video size for the output. For the syntax of this
43267 option, check the "Video size" section in the ffmpeg-utils manual.
43268 Default value is "1024x256".
43269 colors
43271 Specify list of colors separated by space or by '|' which will be
43272 used to draw channels. Unrecognized or missing colors will be
43273 replaced by white color.
43274 mode, m
43276 Set output mode. Can be "bars" or "trace". Default is "bars".
43277 adrawgraph
43279 Draw a graph using input audio metadata.
43280 See drawgraph
43282 agraphmonitor
43284 See graphmonitor.
43285 ahistogram
43287 Convert input audio to a video output, displaying the volume histogram.
43288 The filter accepts the following options:
43290 dmode
43292 Specify how histogram is calculated.
43293 It accepts the following values:
43295 single
43297 Use single histogram for all channels.
43298 separate
43300 Use separate histogram for each channel.
43301 Default is "single".
43303 rate, r
43305 Set frame rate, expressed as number of frames per second. Default
43306 value is "25".
43307 size, s
43309 Specify the video size for the output. For the syntax of this
43310 option, check the "Video size" section in the ffmpeg-utils manual.
43311 Default value is "hd720".
43312 scale
43314 Set display scale.
43315 It accepts the following values:
43317 log logarithmic
43319 sqrt
43321 square root
43322 cbrt
43324 cubic root
43325 lin linear
43327 rlog
43329 reverse logarithmic
43330 Default is "log".
43332 ascale
43334 Set amplitude scale.
43335 It accepts the following values:
43337 log logarithmic
43339 lin linear
43341 Default is "log".
43343 acount
43345 Set how much frames to accumulate in histogram. Default is 1.
43346 Setting this to -1 accumulates all frames.
43347 rheight
43349 Set histogram ratio of window height.
43350 slide
43352 Set sonogram sliding.
43353 It accepts the following values:
43355 replace
43357 replace old rows with new ones.
43358 scroll
43360 scroll from top to bottom.
43361 Default is "replace".
43363 hmode
43365 Set histogram mode.
43366 It accepts the following values:
43368 abs Use absolute values of samples.
43370 sign
43372 Use untouched values of samples.
43373 Default is "abs".
43375 aphasemeter
43377 Measures phase of input audio, which is exported as metadata
43378 "lavfi.aphasemeter.phase", representing mean phase of current audio
43379 frame. A video output can also be produced and is enabled by default.
43380 The audio is passed through as first output.
43381 Audio will be rematrixed to stereo if it has a different channel
43383 layout. Phase value is in range "[-1, 1]" where -1 means left and right
43384 channels are completely out of phase and 1 means channels are in phase.
43385 The filter accepts the following options, all related to its video
43387 output:
43388 rate, r
43390 Set the output frame rate. Default value is 25.
43391 size, s
43393 Set the video size for the output. For the syntax of this option,
43394 check the "Video size" section in the ffmpeg-utils manual. Default
43395 value is "800x400".
43396 rc
43398 gc
43399 bc Specify the red, green, blue contrast. Default values are 2, 7 and
43400 1. Allowed range is "[0, 255]".
43401 mpc Set color which will be used for drawing median phase. If color is
43403 "none" which is default, no median phase value will be drawn.
43404 video
43406 Enable video output. Default is enabled.
43407 phasing detection
43409 The filter also detects out of phase and mono sequences in stereo
43411 streams. It logs the sequence start, end and duration when it lasts
43412 longer or as long as the minimum set.
43413 The filter accepts the following options for this detection:
43415 phasing
43417 Enable mono and out of phase detection. Default is disabled.
43418 tolerance, t
43420 Set phase tolerance for mono detection, in amplitude ratio. Default
43421 is 0. Allowed range is "[0, 1]".
43422 angle, a
43424 Set angle threshold for out of phase detection, in degree. Default
43425 is 170. Allowed range is "[90, 180]".
43426 duration, d
43428 Set mono or out of phase duration until notification, expressed in
43429 seconds. Default is 2.
43430 Examples
43432 • Complete example with ffmpeg to detect 1 second of mono with 0.001
43434 phase tolerance:
43435 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
43437 avectorscope
43439 Convert input audio to a video output, representing the audio vector
43440 scope.
43441 The filter is used to measure the difference between channels of stereo
43443 audio stream. A monaural signal, consisting of identical left and right
43444 signal, results in straight vertical line. Any stereo separation is
43445 visible as a deviation from this line, creating a Lissajous figure. If
43446 the straight (or deviation from it) but horizontal line appears this
43447 indicates that the left and right channels are out of phase.
43448 The filter accepts the following options:
43450 mode, m
43452 Set the vectorscope mode.
43453 Available values are:
43455 lissajous
43457 Lissajous rotated by 45 degrees.
43458 lissajous_xy
43460 Same as above but not rotated.
43461 polar
43463 Shape resembling half of circle.
43464 Default value is lissajous.
43466 size, s
43468 Set the video size for the output. For the syntax of this option,
43469 check the "Video size" section in the ffmpeg-utils manual. Default
43470 value is "400x400".
43471 rate, r
43473 Set the output frame rate. Default value is 25.
43474 rc
43476 gc
43477 bc
43478 ac Specify the red, green, blue and alpha contrast. Default values are
43479 40, 160, 80 and 255. Allowed range is "[0, 255]".
43480 rf
43482 gf
43483 bf
43484 af Specify the red, green, blue and alpha fade. Default values are 15,
43485 10, 5 and 5. Allowed range is "[0, 255]".
43486 zoom
43488 Set the zoom factor. Default value is 1. Allowed range is "[0,
43489 10]". Values lower than 1 will auto adjust zoom factor to maximal
43490 possible value.
43491 draw
43493 Set the vectorscope drawing mode.
43494 Available values are:
43496 dot Draw dot for each sample.
43498 line
43500 Draw line between previous and current sample.
43501 aaline
43503 Draw anti-aliased line between previous and current sample.
43504 Default value is dot.
43506 scale
43508 Specify amplitude scale of audio samples.
43509 Available values are:
43511 lin Linear.
43513 sqrt
43515 Square root.
43516 cbrt
43518 Cubic root.
43519 log Logarithmic.
43521 swap
43523 Swap left channel axis with right channel axis.
43524 mirror
43526 Mirror axis.
43527 none
43529 No mirror.
43530 x Mirror only x axis.
43532 y Mirror only y axis.
43534 xy Mirror both axis.
43536 Examples
43538 • Complete example using ffplay:
43540 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
43542 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
43543 Commands
43545 This filter supports the all above options as commands except options
43547 "size" and "rate".
43548 bench, abench
43550 Benchmark part of a filtergraph.
43551 The filter accepts the following options:
43553 action
43555 Start or stop a timer.
43556 Available values are:
43558 start
43560 Get the current time, set it as frame metadata (using the key
43561 "lavfi.bench.start_time"), and forward the frame to the next
43562 filter.
43563 stop
43565 Get the current time and fetch the "lavfi.bench.start_time"
43566 metadata from the input frame metadata to get the time
43567 difference. Time difference, average, maximum and minimum time
43568 (respectively "t", "avg", "max" and "min") are then printed.
43569 The timestamps are expressed in seconds.
43570 Examples
43572 • Benchmark selectivecolor filter:
43574 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
43576 concat
43578 Concatenate audio and video streams, joining them together one after
43579 the other.
43580 The filter works on segments of synchronized video and audio streams.
43582 All segments must have the same number of streams of each type, and
43583 that will also be the number of streams at output.
43584 The filter accepts the following options:
43586 n Set the number of segments. Default is 2.
43588 v Set the number of output video streams, that is also the number of
43590 video streams in each segment. Default is 1.
43591 a Set the number of output audio streams, that is also the number of
43593 audio streams in each segment. Default is 0.
43594 unsafe
43596 Activate unsafe mode: do not fail if segments have a different
43597 format.
43598 The filter has v+a outputs: first v video outputs, then a audio
43600 outputs.
43601 There are nx(v+a) inputs: first the inputs for the first segment, in
43603 the same order as the outputs, then the inputs for the second segment,
43604 etc.
43605 Related streams do not always have exactly the same duration, for
43607 various reasons including codec frame size or sloppy authoring. For
43608 that reason, related synchronized streams (e.g. a video and its audio
43609 track) should be concatenated at once. The concat filter will use the
43610 duration of the longest stream in each segment (except the last one),
43611 and if necessary pad shorter audio streams with silence.
43612 For this filter to work correctly, all segments must start at timestamp
43614 0.
43615 All corresponding streams must have the same parameters in all
43617 segments; the filtering system will automatically select a common pixel
43618 format for video streams, and a common sample format, sample rate and
43619 channel layout for audio streams, but other settings, such as
43620 resolution, must be converted explicitly by the user.
43621 Different frame rates are acceptable but will result in variable frame
43623 rate at output; be sure to configure the output file to handle it.
43624 Examples
43626 • Concatenate an opening, an episode and an ending, all in bilingual
43628 version (video in stream 0, audio in streams 1 and 2):
43629 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
43631 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
43632 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
43633 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
43634 • Concatenate two parts, handling audio and video separately, using
43636 the (a)movie sources, and adjusting the resolution:
43637 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
43639 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
43640 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
43641 Note that a desync will happen at the stitch if the audio and video
43643 streams do not have exactly the same duration in the first file.
43644 Commands
43646 This filter supports the following commands:
43648 next
43650 Close the current segment and step to the next one
43651 ebur128
43653 EBU R128 scanner filter. This filter takes an audio stream and analyzes
43654 its loudness level. By default, it logs a message at a frequency of
43655 10Hz with the Momentary loudness (identified by "M"), Short-term
43656 loudness ("S"), Integrated loudness ("I") and Loudness Range ("LRA").
43657 The filter can only analyze streams which have sample format is double-
43659 precision floating point. The input stream will be converted to this
43660 specification, if needed. Users may need to insert aformat and/or
43661 aresample filters after this filter to obtain the original parameters.
43662 The filter also has a video output (see the video option) with a real
43664 time graph to observe the loudness evolution. The graphic contains the
43665 logged message mentioned above, so it is not printed anymore when this
43666 option is set, unless the verbose logging is set. The main graphing
43667 area contains the short-term loudness (3 seconds of analysis), and the
43668 gauge on the right is for the momentary loudness (400 milliseconds),
43669 but can optionally be configured to instead display short-term loudness
43670 (see gauge).
43671 The green area marks a +/- 1LU target range around the target loudness
43673 (-23LUFS by default, unless modified through target).
43674 More information about the Loudness Recommendation EBU R128 on
43676 <http://tech.ebu.ch/loudness>.
43677 The filter accepts the following options:
43679 video
43681 Activate the video output. The audio stream is passed unchanged
43682 whether this option is set or no. The video stream will be the
43683 first output stream if activated. Default is 0.
43684 size
43686 Set the video size. This option is for video only. For the syntax
43687 of this option, check the "Video size" section in the ffmpeg-utils
43688 manual. Default and minimum resolution is "640x480".
43689 meter
43691 Set the EBU scale meter. Default is 9. Common values are 9 and 18,
43692 respectively for EBU scale meter +9 and EBU scale meter +18. Any
43693 other integer value between this range is allowed.
43694 metadata
43696 Set metadata injection. If set to 1, the audio input will be
43697 segmented into 100ms output frames, each of them containing various
43698 loudness information in metadata. All the metadata keys are
43699 prefixed with "lavfi.r128.".
43700 Default is 0.
43702 framelog
43704 Force the frame logging level.
43705 Available values are:
43707 quiet
43709 logging disabled
43710 info
43712 information logging level
43713 verbose
43715 verbose logging level
43716 By default, the logging level is set to info. If the video or the
43718 metadata options are set, it switches to verbose.
43719 peak
43721 Set peak mode(s).
43722 Available modes can be cumulated (the option is a "flag" type).
43724 Possible values are:
43725 none
43727 Disable any peak mode (default).
43728 sample
43730 Enable sample-peak mode.
43731 Simple peak mode looking for the higher sample value. It logs a
43733 message for sample-peak (identified by "SPK").
43734 true
43736 Enable true-peak mode.
43737 If enabled, the peak lookup is done on an over-sampled version
43739 of the input stream for better peak accuracy. It logs a message
43740 for true-peak. (identified by "TPK") and true-peak per frame
43741 (identified by "FTPK"). This mode requires a build with
43742 "libswresample".
43743 dualmono
43745 Treat mono input files as "dual mono". If a mono file is intended
43746 for playback on a stereo system, its EBU R128 measurement will be
43747 perceptually incorrect. If set to "true", this option will
43748 compensate for this effect. Multi-channel input files are not
43749 affected by this option.
43750 panlaw
43752 Set a specific pan law to be used for the measurement of dual mono
43753 files. This parameter is optional, and has a default value of
43754 -3.01dB.
43755 target
43757 Set a specific target level (in LUFS) used as relative zero in the
43758 visualization. This parameter is optional and has a default value
43759 of -23LUFS as specified by EBU R128. However, material published
43760 online may prefer a level of -16LUFS (e.g. for use with podcasts or
43761 video platforms).
43762 gauge
43764 Set the value displayed by the gauge. Valid values are "momentary"
43765 and s "shortterm". By default the momentary value will be used, but
43766 in certain scenarios it may be more useful to observe the short
43767 term value instead (e.g. live mixing).
43768 scale
43770 Sets the display scale for the loudness. Valid parameters are
43771 "absolute" (in LUFS) or "relative" (LU) relative to the target.
43772 This only affects the video output, not the summary or continuous
43773 log output.
43774 Examples
43776 • Real-time graph using ffplay, with a EBU scale meter +18:
43778 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
43780 • Run an analysis with ffmpeg:
43782 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
43784 interleave, ainterleave
43786 Temporally interleave frames from several inputs.
43787 "interleave" works with video inputs, "ainterleave" with audio.
43789 These filters read frames from several inputs and send the oldest
43791 queued frame to the output.
43792 Input streams must have well defined, monotonically increasing frame
43794 timestamp values.
43795 In order to submit one frame to output, these filters need to enqueue
43797 at least one frame for each input, so they cannot work in case one
43798 input is not yet terminated and will not receive incoming frames.
43799 For example consider the case when one input is a "select" filter which
43801 always drops input frames. The "interleave" filter will keep reading
43802 from that input, but it will never be able to send new frames to output
43803 until the input sends an end-of-stream signal.
43804 Also, depending on inputs synchronization, the filters will drop frames
43806 in case one input receives more frames than the other ones, and the
43807 queue is already filled.
43808 These filters accept the following options:
43810 nb_inputs, n
43812 Set the number of different inputs, it is 2 by default.
43813 duration
43815 How to determine the end-of-stream.
43816 longest
43818 The duration of the longest input. (default)
43819 shortest
43821 The duration of the shortest input.
43822 first
43824 The duration of the first input.
43825 Examples
43827 • Interleave frames belonging to different streams using ffmpeg:
43829 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
43831 • Add flickering blur effect:
43833 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
43835 latency, alatency
43837 Measure filtering latency.
43838 Report previous filter filtering latency, delay in number of audio
43840 samples for audio filters or number of video frames for video filters.
43841 On end of input stream, filter will report min and max measured latency
43843 for previous running filter in filtergraph.
43844 metadata, ametadata
43846 Manipulate frame metadata.
43847 This filter accepts the following options:
43849 mode
43851 Set mode of operation of the filter.
43852 Can be one of the following:
43854 select
43856 If both "value" and "key" is set, select frames which have such
43857 metadata. If only "key" is set, select every frame that has
43858 such key in metadata.
43859 add Add new metadata "key" and "value". If key is already available
43861 do nothing.
43862 modify
43864 Modify value of already present key.
43865 delete
43867 If "value" is set, delete only keys that have such value.
43868 Otherwise, delete key. If "key" is not set, delete all metadata
43869 values in the frame.
43870 print
43872 Print key and its value if metadata was found. If "key" is not
43873 set print all metadata values available in frame.
43874 key Set key used with all modes. Must be set for all modes except
43876 "print" and "delete".
43877 value
43879 Set metadata value which will be used. This option is mandatory for
43880 "modify" and "add" mode.
43881 function
43883 Which function to use when comparing metadata value and "value".
43884 Can be one of following:
43886 same_str
43888 Values are interpreted as strings, returns true if metadata
43889 value is same as "value".
43890 starts_with
43892 Values are interpreted as strings, returns true if metadata
43893 value starts with the "value" option string.
43894 less
43896 Values are interpreted as floats, returns true if metadata
43897 value is less than "value".
43898 equal
43900 Values are interpreted as floats, returns true if "value" is
43901 equal with metadata value.
43902 greater
43904 Values are interpreted as floats, returns true if metadata
43905 value is greater than "value".
43906 expr
43908 Values are interpreted as floats, returns true if expression
43909 from option "expr" evaluates to true.
43910 ends_with
43912 Values are interpreted as strings, returns true if metadata
43913 value ends with the "value" option string.
43914 expr
43916 Set expression which is used when "function" is set to "expr". The
43917 expression is evaluated through the eval API and can contain the
43918 following constants:
43919 VALUE1, FRAMEVAL
43921 Float representation of "value" from metadata key.
43922 VALUE2, USERVAL
43924 Float representation of "value" as supplied by user in "value"
43925 option.
43926 file
43928 If specified in "print" mode, output is written to the named file.
43929 Instead of plain filename any writable url can be specified.
43930 Filename ``-'' is a shorthand for standard output. If "file" option
43931 is not set, output is written to the log with AV_LOG_INFO loglevel.
43932 direct
43934 Reduces buffering in print mode when output is written to a URL set
43935 using file.
43936 Examples
43938 • Print all metadata values for frames with key
43940 "lavfi.signalstats.YDIF" with values between 0 and 1.
43941 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
43943 • Print silencedetect output to file metadata.txt.
43945 silencedetect,ametadata=mode=print:file=metadata.txt
43947 • Direct all metadata to a pipe with file descriptor 4.
43949 metadata=mode=print:file='pipe\:4'
43951 perms, aperms
43953 Set read/write permissions for the output frames.
43954 These filters are mainly aimed at developers to test direct path in the
43956 following filter in the filtergraph.
43957 The filters accept the following options:
43959 mode
43961 Select the permissions mode.
43962 It accepts the following values:
43964 none
43966 Do nothing. This is the default.
43967 ro Set all the output frames read-only.
43969 rw Set all the output frames directly writable.
43971 toggle
43973 Make the frame read-only if writable, and writable if read-
43974 only.
43975 random
43977 Set each output frame read-only or writable randomly.
43978 seed
43980 Set the seed for the random mode, must be an integer included
43981 between 0 and "UINT32_MAX". If not specified, or if explicitly set
43982 to -1, the filter will try to use a good random seed on a best
43983 effort basis.
43984 Note: in case of auto-inserted filter between the permission filter and
43986 the following one, the permission might not be received as expected in
43987 that following filter. Inserting a format or aformat filter before the
43988 perms/aperms filter can avoid this problem.
43989 realtime, arealtime
43991 Slow down filtering to match real time approximately.
43992 These filters will pause the filtering for a variable amount of time to
43994 match the output rate with the input timestamps. They are similar to
43995 the re option to "ffmpeg".
43996 They accept the following options:
43998 limit
44000 Time limit for the pauses. Any pause longer than that will be
44001 considered a timestamp discontinuity and reset the timer. Default
44002 is 2 seconds.
44003 speed
44005 Speed factor for processing. The value must be a float larger than
44006 zero. Values larger than 1.0 will result in faster than realtime
44007 processing, smaller will slow processing down. The limit is
44008 automatically adapted accordingly. Default is 1.0.
44009 A processing speed faster than what is possible without these
44011 filters cannot be achieved.
44012 Commands
44014 Both filters supports the all above options as commands.
44016 segment, asegment
44018 Split single input stream into multiple streams.
44019 This filter does opposite of concat filters.
44021 "segment" works on video frames, "asegment" on audio samples.
44023 This filter accepts the following options:
44025 timestamps
44027 Timestamps of output segments separated by '|'. The first segment
44028 will run from the beginning of the input stream. The last segment
44029 will run until the end of the input stream
44030 frames, samples
44032 Exact frame/sample count to split the segments.
44033 In all cases, prefixing an each segment with '+' will make it relative
44035 to the previous segment.
44036 Examples
44038 • Split input audio stream into three output audio streams, starting
44040 at start of input audio stream and storing that in 1st output audio
44041 stream, then following at 60th second and storing than in 2nd
44042 output audio stream, and last after 150th second of input audio
44043 stream store in 3rd output audio stream:
44044 asegment=timestamps="60|150"
44046 select, aselect
44048 Select frames to pass in output.
44049 This filter accepts the following options:
44051 expr, e
44053 Set expression, which is evaluated for each input frame.
44054 If the expression is evaluated to zero, the frame is discarded.
44056 If the evaluation result is negative or NaN, the frame is sent to
44058 the first output; otherwise it is sent to the output with index
44059 "ceil(val)-1", assuming that the input index starts from 0.
44060 For example a value of 1.2 corresponds to the output with index
44062 "ceil(1.2)-1 = 2-1 = 1", that is the second output.
44063 outputs, n
44065 Set the number of outputs. The output to which to send the selected
44066 frame is based on the result of the evaluation. Default value is 1.
44067 The expression can contain the following constants:
44069 n The (sequential) number of the filtered frame, starting from 0.
44071 selected_n
44073 The (sequential) number of the selected frame, starting from 0.
44074 prev_selected_n
44076 The sequential number of the last selected frame. It's NAN if
44077 undefined.
44078 TB The timebase of the input timestamps.
44080 pts The PTS (Presentation TimeStamp) of the filtered frame, expressed
44082 in TB units. It's NAN if undefined.
44083 t The PTS of the filtered frame, expressed in seconds. It's NAN if
44085 undefined.
44086 prev_pts
44088 The PTS of the previously filtered frame. It's NAN if undefined.
44089 prev_selected_pts
44091 The PTS of the last previously filtered frame. It's NAN if
44092 undefined.
44093 prev_selected_t
44095 The PTS of the last previously selected frame, expressed in
44096 seconds. It's NAN if undefined.
44097 start_pts
44099 The first PTS in the stream which is not NAN. It remains NAN if not
44100 found.
44101 start_t
44103 The first PTS, in seconds, in the stream which is not NAN. It
44104 remains NAN if not found.
44105 pict_type (video only)
44107 The type of the filtered frame. It can assume one of the following
44108 values:
44109 I
44111 P
44112 B
44113 S
44114 SI
44115 SP
44116 BI
44117 interlace_type (video only)
44118 The frame interlace type. It can assume one of the following
44119 values:
44120 PROGRESSIVE
44122 The frame is progressive (not interlaced).
44123 TOPFIRST
44125 The frame is top-field-first.
44126 BOTTOMFIRST
44128 The frame is bottom-field-first.
44129 consumed_sample_n (audio only)
44131 the number of selected samples before the current frame
44132 samples_n (audio only)
44134 the number of samples in the current frame
44135 sample_rate (audio only)
44137 the input sample rate
44138 key This is 1 if the filtered frame is a key-frame, 0 otherwise.
44140 pos the position in the file of the filtered frame, -1 if the
44142 information is not available (e.g. for synthetic video)
44143 scene (video only)
44145 value between 0 and 1 to indicate a new scene; a low value reflects
44146 a low probability for the current frame to introduce a new scene,
44147 while a higher value means the current frame is more likely to be
44148 one (see the example below)
44149 concatdec_select
44151 The concat demuxer can select only part of a concat input file by
44152 setting an inpoint and an outpoint, but the output packets may not
44153 be entirely contained in the selected interval. By using this
44154 variable, it is possible to skip frames generated by the concat
44155 demuxer which are not exactly contained in the selected interval.
44156 This works by comparing the frame pts against the
44158 lavf.concat.start_time and the lavf.concat.duration packet metadata
44159 values which are also present in the decoded frames.
44160 The concatdec_select variable is -1 if the frame pts is at least
44162 start_time and either the duration metadata is missing or the frame
44163 pts is less than start_time + duration, 0 otherwise, and NaN if the
44164 start_time metadata is missing.
44165 That basically means that an input frame is selected if its pts is
44167 within the interval set by the concat demuxer.
44168 The default value of the select expression is "1".
44170 Examples
44172 • Select all frames in input:
44174 select
44176 The example above is the same as:
44178 select=1
44180 • Skip all frames:
44182 select=0
44184 • Select only I-frames:
44186 select='eq(pict_type\,I)'
44188 • Select one frame every 100:
44190 select='not(mod(n\,100))'
44192 • Select only frames contained in the 10-20 time interval:
44194 select=between(t\,10\,20)
44196 • Select only I-frames contained in the 10-20 time interval:
44198 select=between(t\,10\,20)*eq(pict_type\,I)
44200 • Select frames with a minimum distance of 10 seconds:
44202 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
44204 • Use aselect to select only audio frames with samples number > 100:
44206 aselect='gt(samples_n\,100)'
44208 • Create a mosaic of the first scenes:
44210 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
44212 Comparing scene against a value between 0.3 and 0.5 is generally a
44214 sane choice.
44215 • Send even and odd frames to separate outputs, and compose them:
44217 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
44219 • Select useful frames from an ffconcat file which is using inpoints
44221 and outpoints but where the source files are not intra frame only.
44222 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
44224 sendcmd, asendcmd
44226 Send commands to filters in the filtergraph.
44227 These filters read commands to be sent to other filters in the
44229 filtergraph.
44230 "sendcmd" must be inserted between two video filters, "asendcmd" must
44232 be inserted between two audio filters, but apart from that they act the
44233 same way.
44234 The specification of commands can be provided in the filter arguments
44236 with the commands option, or in a file specified by the filename
44237 option.
44238 These filters accept the following options:
44240 commands, c
44242 Set the commands to be read and sent to the other filters.
44243 filename, f
44245 Set the filename of the commands to be read and sent to the other
44246 filters.
44247 Commands syntax
44249 A commands description consists of a sequence of interval
44251 specifications, comprising a list of commands to be executed when a
44252 particular event related to that interval occurs. The occurring event
44253 is typically the current frame time entering or leaving a given time
44254 interval.
44255 An interval is specified by the following syntax:
44257 <START>[-<END>] <COMMANDS>;
44259 The time interval is specified by the START and END times. END is
44261 optional and defaults to the maximum time.
44262 The current frame time is considered within the specified interval if
44264 it is included in the interval [START, END), that is when the time is
44265 greater or equal to START and is lesser than END.
44266 COMMANDS consists of a sequence of one or more command specifications,
44268 separated by ",", relating to that interval. The syntax of a command
44269 specification is given by:
44270 [<FLAGS>] <TARGET> <COMMAND> <ARG>
44272 FLAGS is optional and specifies the type of events relating to the time
44274 interval which enable sending the specified command, and must be a non-
44275 null sequence of identifier flags separated by "+" or "|" and enclosed
44276 between "[" and "]".
44277 The following flags are recognized:
44279 enter
44281 The command is sent when the current frame timestamp enters the
44282 specified interval. In other words, the command is sent when the
44283 previous frame timestamp was not in the given interval, and the
44284 current is.
44285 leave
44287 The command is sent when the current frame timestamp leaves the
44288 specified interval. In other words, the command is sent when the
44289 previous frame timestamp was in the given interval, and the current
44290 is not.
44291 expr
44293 The command ARG is interpreted as expression and result of
44294 expression is passed as ARG.
44295 The expression is evaluated through the eval API and can contain
44297 the following constants:
44298 POS Original position in the file of the frame, or undefined if
44300 undefined for the current frame.
44301 PTS The presentation timestamp in input.
44303 N The count of the input frame for video or audio, starting from
44305 0.
44306 T The time in seconds of the current frame.
44308 TS The start time in seconds of the current command interval.
44310 TE The end time in seconds of the current command interval.
44312 TI The interpolated time of the current command interval, TI = (T
44314 - TS) / (TE - TS).
44315 W The video frame width.
44317 H The video frame height.
44319 If FLAGS is not specified, a default value of "[enter]" is assumed.
44321 TARGET specifies the target of the command, usually the name of the
44323 filter class or a specific filter instance name.
44324 COMMAND specifies the name of the command for the target filter.
44326 ARG is optional and specifies the optional list of argument for the
44328 given COMMAND.
44329 Between one interval specification and another, whitespaces, or
44331 sequences of characters starting with "#" until the end of line, are
44332 ignored and can be used to annotate comments.
44333 A simplified BNF description of the commands specification syntax
44335 follows:
44336 <COMMAND_FLAG> ::= "enter" | "leave"
44338 <COMMAND_FLAGS> ::= <COMMAND_FLAG> [(+|"|")<COMMAND_FLAG>]
44339 <COMMAND> ::= ["[" <COMMAND_FLAGS> "]"] <TARGET> <COMMAND> [<ARG>]
44340 <COMMANDS> ::= <COMMAND> [,<COMMANDS>]
44341 <INTERVAL> ::= <START>[-<END>] <COMMANDS>
44342 <INTERVALS> ::= <INTERVAL>[;<INTERVALS>]
44343 Examples
44345 • Specify audio tempo change at second 4:
44347 asendcmd=c='4.0 atempo tempo 1.5',atempo
44349 • Target a specific filter instance:
44351 asendcmd=c='4.0 atempo@my tempo 1.5',atempo@my
44353 • Specify a list of drawtext and hue commands in a file.
44355 # show text in the interval 5-10
44357 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
44358 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
44359 # desaturate the image in the interval 15-20
44361 15.0-20.0 [enter] hue s 0,
44362 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
44363 [leave] hue s 1,
44364 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
44365 # apply an exponential saturation fade-out effect, starting from time 25
44367 25 [enter] hue s exp(25-t)
44368 A filtergraph allowing to read and process the above command list
44370 stored in a file test.cmd, can be specified with:
44371 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
44373 setpts, asetpts
44375 Change the PTS (presentation timestamp) of the input frames.
44376 "setpts" works on video frames, "asetpts" on audio frames.
44378 This filter accepts the following options:
44380 expr
44382 The expression which is evaluated for each frame to construct its
44383 timestamp.
44384 The expression is evaluated through the eval API and can contain the
44386 following constants:
44387 FRAME_RATE, FR
44389 frame rate, only defined for constant frame-rate video
44390 PTS The presentation timestamp in input
44392 N The count of the input frame for video or the number of consumed
44394 samples, not including the current frame for audio, starting from
44395 0.
44396 NB_CONSUMED_SAMPLES
44398 The number of consumed samples, not including the current frame
44399 (only audio)
44400 NB_SAMPLES, S
44402 The number of samples in the current frame (only audio)
44403 SAMPLE_RATE, SR
44405 The audio sample rate.
44406 STARTPTS
44408 The PTS of the first frame.
44409 STARTT
44411 the time in seconds of the first frame
44412 INTERLACED
44414 State whether the current frame is interlaced.
44415 T the time in seconds of the current frame
44417 POS original position in the file of the frame, or undefined if
44419 undefined for the current frame
44420 PREV_INPTS
44422 The previous input PTS.
44423 PREV_INT
44425 previous input time in seconds
44426 PREV_OUTPTS
44428 The previous output PTS.
44429 PREV_OUTT
44431 previous output time in seconds
44432 RTCTIME
44434 The wallclock (RTC) time in microseconds. This is deprecated, use
44435 time(0) instead.
44436 RTCSTART
44438 The wallclock (RTC) time at the start of the movie in microseconds.
44439 TB The timebase of the input timestamps.
44441 Examples
44443 • Start counting PTS from zero
44445 setpts=PTS-STARTPTS
44447 • Apply fast motion effect:
44449 setpts=0.5*PTS
44451 • Apply slow motion effect:
44453 setpts=2.0*PTS
44455 • Set fixed rate of 25 frames per second:
44457 setpts=N/(25*TB)
44459 • Set fixed rate 25 fps with some jitter:
44461 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
44463 • Apply an offset of 10 seconds to the input PTS:
44465 setpts=PTS+10/TB
44467 • Generate timestamps from a "live source" and rebase onto the
44469 current timebase:
44470 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
44472 • Generate timestamps by counting samples:
44474 asetpts=N/SR/TB
44476 setrange
44478 Force color range for the output video frame.
44479 The "setrange" filter marks the color range property for the output
44481 frames. It does not change the input frame, but only sets the
44482 corresponding property, which affects how the frame is treated by
44483 following filters.
44484 The filter accepts the following options:
44486 range
44488 Available values are:
44489 auto
44491 Keep the same color range property.
44492 unspecified, unknown
44494 Set the color range as unspecified.
44495 limited, tv, mpeg
44497 Set the color range as limited.
44498 full, pc, jpeg
44500 Set the color range as full.
44501 settb, asettb
44503 Set the timebase to use for the output frames timestamps. It is mainly
44504 useful for testing timebase configuration.
44505 It accepts the following parameters:
44507 expr, tb
44509 The expression which is evaluated into the output timebase.
44510 The value for tb is an arithmetic expression representing a rational.
44512 The expression can contain the constants "AVTB" (the default timebase),
44513 "intb" (the input timebase) and "sr" (the sample rate, audio only).
44514 Default value is "intb".
44515 Examples
44517 • Set the timebase to 1/25:
44519 settb=expr=1/25
44521 • Set the timebase to 1/10:
44523 settb=expr=0.1
44525 • Set the timebase to 1001/1000:
44527 settb=1+0.001
44529 • Set the timebase to 2*intb:
44531 settb=2*intb
44533 • Set the default timebase value:
44535 settb=AVTB
44537 showcqt
44539 Convert input audio to a video output representing frequency spectrum
44540 logarithmically using Brown-Puckette constant Q transform algorithm
44541 with direct frequency domain coefficient calculation (but the transform
44542 itself is not really constant Q, instead the Q factor is actually
44543 variable/clamped), with musical tone scale, from E0 to D#10.
44544 The filter accepts the following options:
44546 size, s
44548 Specify the video size for the output. It must be even. For the
44549 syntax of this option, check the "Video size" section in the
44550 ffmpeg-utils manual. Default value is "1920x1080".
44551 fps, rate, r
44553 Set the output frame rate. Default value is 25.
44554 bar_h
44556 Set the bargraph height. It must be even. Default value is -1 which
44557 computes the bargraph height automatically.
44558 axis_h
44560 Set the axis height. It must be even. Default value is -1 which
44561 computes the axis height automatically.
44562 sono_h
44564 Set the sonogram height. It must be even. Default value is -1 which
44565 computes the sonogram height automatically.
44566 fullhd
44568 Set the fullhd resolution. This option is deprecated, use size, s
44569 instead. Default value is 1.
44570 sono_v, volume
44572 Specify the sonogram volume expression. It can contain variables:
44573 bar_v
44575 the bar_v evaluated expression
44576 frequency, freq, f
44578 the frequency where it is evaluated
44579 timeclamp, tc
44581 the value of timeclamp option
44582 and functions:
44584 a_weighting(f)
44586 A-weighting of equal loudness
44587 b_weighting(f)
44589 B-weighting of equal loudness
44590 c_weighting(f)
44592 C-weighting of equal loudness.
44593 Default value is 16.
44595 bar_v, volume2
44597 Specify the bargraph volume expression. It can contain variables:
44598 sono_v
44600 the sono_v evaluated expression
44601 frequency, freq, f
44603 the frequency where it is evaluated
44604 timeclamp, tc
44606 the value of timeclamp option
44607 and functions:
44609 a_weighting(f)
44611 A-weighting of equal loudness
44612 b_weighting(f)
44614 B-weighting of equal loudness
44615 c_weighting(f)
44617 C-weighting of equal loudness.
44618 Default value is "sono_v".
44620 sono_g, gamma
44622 Specify the sonogram gamma. Lower gamma makes the spectrum more
44623 contrast, higher gamma makes the spectrum having more range.
44624 Default value is 3. Acceptable range is "[1, 7]".
44625 bar_g, gamma2
44627 Specify the bargraph gamma. Default value is 1. Acceptable range is
44628 "[1, 7]".
44629 bar_t
44631 Specify the bargraph transparency level. Lower value makes the
44632 bargraph sharper. Default value is 1. Acceptable range is "[0,
44633 1]".
44634 timeclamp, tc
44636 Specify the transform timeclamp. At low frequency, there is trade-
44637 off between accuracy in time domain and frequency domain. If
44638 timeclamp is lower, event in time domain is represented more
44639 accurately (such as fast bass drum), otherwise event in frequency
44640 domain is represented more accurately (such as bass guitar).
44641 Acceptable range is "[0.002, 1]". Default value is 0.17.
44642 attack
44644 Set attack time in seconds. The default is 0 (disabled). Otherwise,
44645 it limits future samples by applying asymmetric windowing in time
44646 domain, useful when low latency is required. Accepted range is "[0,
44647 1]".
44648 basefreq
44650 Specify the transform base frequency. Default value is
44651 20.01523126408007475, which is frequency 50 cents below E0.
44652 Acceptable range is "[10, 100000]".
44653 endfreq
44655 Specify the transform end frequency. Default value is
44656 20495.59681441799654, which is frequency 50 cents above D#10.
44657 Acceptable range is "[10, 100000]".
44658 coeffclamp
44660 This option is deprecated and ignored.
44661 tlength
44663 Specify the transform length in time domain. Use this option to
44664 control accuracy trade-off between time domain and frequency domain
44665 at every frequency sample. It can contain variables:
44666 frequency, freq, f
44668 the frequency where it is evaluated
44669 timeclamp, tc
44671 the value of timeclamp option.
44672 Default value is "384*tc/(384+tc*f)".
44674 count
44676 Specify the transform count for every video frame. Default value is
44677 6. Acceptable range is "[1, 30]".
44678 fcount
44680 Specify the transform count for every single pixel. Default value
44681 is 0, which makes it computed automatically. Acceptable range is
44682 "[0, 10]".
44683 fontfile
44685 Specify font file for use with freetype to draw the axis. If not
44686 specified, use embedded font. Note that drawing with font file or
44687 embedded font is not implemented with custom basefreq and endfreq,
44688 use axisfile option instead.
44689 font
44691 Specify fontconfig pattern. This has lower priority than fontfile.
44692 The ":" in the pattern may be replaced by "|" to avoid unnecessary
44693 escaping.
44694 fontcolor
44696 Specify font color expression. This is arithmetic expression that
44697 should return integer value 0xRRGGBB. It can contain variables:
44698 frequency, freq, f
44700 the frequency where it is evaluated
44701 timeclamp, tc
44703 the value of timeclamp option
44704 and functions:
44706 midi(f)
44708 midi number of frequency f, some midi numbers: E0(16), C1(24),
44709 C2(36), A4(69)
44710 r(x), g(x), b(x)
44712 red, green, and blue value of intensity x.
44713 Default value is "st(0, (midi(f)-59.5)/12); st(1,
44715 if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0)); r(1-ld(1)) +
44716 b(ld(1))".
44717 axisfile
44719 Specify image file to draw the axis. This option override fontfile
44720 and fontcolor option.
44721 axis, text
44723 Enable/disable drawing text to the axis. If it is set to 0, drawing
44724 to the axis is disabled, ignoring fontfile and axisfile option.
44725 Default value is 1.
44726 csp Set colorspace. The accepted values are:
44728 unspecified
44730 Unspecified (default)
44731 bt709
44733 BT.709
44734 fcc FCC
44736 bt470bg
44738 BT.470BG or BT.601-6 625
44739 smpte170m
44741 SMPTE-170M or BT.601-6 525
44742 smpte240m
44744 SMPTE-240M
44745 bt2020ncl
44747 BT.2020 with non-constant luminance
44748 cscheme
44750 Set spectrogram color scheme. This is list of floating point values
44751 with format "left_r|left_g|left_b|right_r|right_g|right_b". The
44752 default is "1|0.5|0|0|0.5|1".
44753 Examples
44755 • Playing audio while showing the spectrum:
44757 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
44759 • Same as above, but with frame rate 30 fps:
44761 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
44763 • Playing at 1280x720:
44765 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
44767 • Disable sonogram display:
44769 sono_h=0
44771 • A1 and its harmonics: A1, A2, (near)E3, A3:
44773 ffplay -f lavfi 'aevalsrc=0.1*sin(2*PI*55*t)+0.1*sin(4*PI*55*t)+0.1*sin(6*PI*55*t)+0.1*sin(8*PI*55*t),
44775 asplit[a][out1]; [a] showcqt [out0]'
44776 • Same as above, but with more accuracy in frequency domain:
44778 ffplay -f lavfi 'aevalsrc=0.1*sin(2*PI*55*t)+0.1*sin(4*PI*55*t)+0.1*sin(6*PI*55*t)+0.1*sin(8*PI*55*t),
44780 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
44781 • Custom volume:
44783 bar_v=10:sono_v=bar_v*a_weighting(f)
44785 • Custom gamma, now spectrum is linear to the amplitude.
44787 bar_g=2:sono_g=2
44789 • Custom tlength equation:
44791 tc=0.33:tlength='st(0,0.17); 384*tc / (384 / ld(0) + tc*f /(1-ld(0))) + 384*tc / (tc*f / ld(0) + 384 /(1-ld(0)))'
44793 • Custom fontcolor and fontfile, C-note is colored green, others are
44795 colored blue:
44796 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
44798 • Custom font using fontconfig:
44800 font='Courier New,Monospace,mono|bold'
44802 • Custom frequency range with custom axis using image file:
44804 axisfile=myaxis.png:basefreq=40:endfreq=10000
44806 showcwt
44808 Convert input audio to video output representing frequency spectrum
44809 using Continuous Wavelet Transform and Morlet wavelet.
44810 The filter accepts the following options:
44812 size, s
44814 Specify the video size for the output. For the syntax of this
44815 option, check the "Video size" section in the ffmpeg-utils manual.
44816 Default value is "640x512".
44817 rate, r
44819 Set the output frame rate. Default value is 25.
44820 scale
44822 Set the frequency scale used. Allowed values are:
44823 linear
44825 log2
44826 bark
44827 mel
44828 erbs
44829 Default value is "linear".
44831 min Set the minimum frequency that will be used in output. Default is
44833 20 Hz.
44834 max Set the maximum frequency that will be used in output. Default is
44836 20000 Hz. The real frequency upper limit depends on input audio's
44837 sample rate and such will be enforced on this value when it is set
44838 to value greater than Nyquist frequency.
44839 logb
44841 Set the logarithmic basis for brightness strength when mapping
44842 calculated magnitude values to pixel values. Allowed range is from
44843 0 to 1. Default value is 0.0001.
44844 deviation
44846 Set the frequency deviation. Lower values than 1 are more
44847 frequency oriented, while higher values than 1 are more time
44848 oriented. Allowed range is from 0 to 10. Default value is 1.
44849 pps Set the number of pixel output per each second in one row. Allowed
44851 range is from 1 to 1024. Default value is 64.
44852 mode
44854 Set the output visual mode. Allowed values are:
44855 magnitude
44857 Show magnitude.
44858 phase
44860 Show only phase.
44861 magphase
44863 Show combination of magnitude and phase. Magnitude is mapped
44864 to brightness and phase to color.
44865 channel
44867 Show unique color per channel magnitude.
44868 stereo
44870 Show unique color per stereo difference.
44871 Default value is "magnitude".
44873 slide
44875 Set the output slide method. Allowed values are:
44876 replace
44878 scroll
44879 frame
44880 direction
44881 Set the direction method for output slide method. Allowed values
44882 are:
44883 lr Direction from left to right.
44885 rl Direction from right to left.
44887 ud Direction from up to down.
44889 du Direction from down to up.
44891 showfreqs
44893 Convert input audio to video output representing the audio power
44894 spectrum. Audio amplitude is on Y-axis while frequency is on X-axis.
44895 The filter accepts the following options:
44897 size, s
44899 Specify size of video. For the syntax of this option, check the
44900 "Video size" section in the ffmpeg-utils manual. Default is
44901 "1024x512".
44902 rate, r
44904 Set video rate. Default is 25.
44905 mode
44907 Set display mode. This set how each frequency bin will be
44908 represented.
44909 It accepts the following values:
44911 line
44913 bar
44914 dot
44915 Default is "bar".
44917 ascale
44919 Set amplitude scale.
44920 It accepts the following values:
44922 lin Linear scale.
44924 sqrt
44926 Square root scale.
44927 cbrt
44929 Cubic root scale.
44930 log Logarithmic scale.
44932 Default is "log".
44934 fscale
44936 Set frequency scale.
44937 It accepts the following values:
44939 lin Linear scale.
44941 log Logarithmic scale.
44943 rlog
44945 Reverse logarithmic scale.
44946 Default is "lin".
44948 win_size
44950 Set window size. Allowed range is from 16 to 65536.
44951 Default is 2048
44953 win_func
44955 Set windowing function.
44956 It accepts the following values:
44958 rect
44960 bartlett
44961 hanning
44962 hamming
44963 blackman
44964 welch
44965 flattop
44966 bharris
44967 bnuttall
44968 bhann
44969 sine
44970 nuttall
44971 lanczos
44972 gauss
44973 tukey
44974 dolph
44975 cauchy
44976 parzen
44977 poisson
44978 bohman
44979 kaiser
44980 Default is "hanning".
44982 overlap
44984 Set window overlap. In range "[0, 1]". Default is 1, which means
44985 optimal overlap for selected window function will be picked.
44986 averaging
44988 Set time averaging. Setting this to 0 will display current maximal
44989 peaks. Default is 1, which means time averaging is disabled.
44990 colors
44992 Specify list of colors separated by space or by '|' which will be
44993 used to draw channel frequencies. Unrecognized or missing colors
44994 will be replaced by white color.
44995 cmode
44997 Set channel display mode.
44998 It accepts the following values:
45000 combined
45002 separate
45003 Default is "combined".
45005 minamp
45007 Set minimum amplitude used in "log" amplitude scaler.
45008 data
45010 Set data display mode.
45011 It accepts the following values:
45013 magnitude
45015 phase
45016 delay
45017 Default is "magnitude".
45019 channels
45021 Set channels to use when processing audio. By default all are
45022 processed.
45023 showspatial
45025 Convert stereo input audio to a video output, representing the spatial
45026 relationship between two channels.
45027 The filter accepts the following options:
45029 size, s
45031 Specify the video size for the output. For the syntax of this
45032 option, check the "Video size" section in the ffmpeg-utils manual.
45033 Default value is "512x512".
45034 win_size
45036 Set window size. Allowed range is from 1024 to 65536. Default size
45037 is 4096.
45038 win_func
45040 Set window function.
45041 It accepts the following values:
45043 rect
45045 bartlett
45046 hann
45047 hanning
45048 hamming
45049 blackman
45050 welch
45051 flattop
45052 bharris
45053 bnuttall
45054 bhann
45055 sine
45056 nuttall
45057 lanczos
45058 gauss
45059 tukey
45060 dolph
45061 cauchy
45062 parzen
45063 poisson
45064 bohman
45065 kaiser
45066 Default value is "hann".
45068 rate, r
45070 Set output framerate.
45071 showspectrum
45073 Convert input audio to a video output, representing the audio frequency
45074 spectrum.
45075 The filter accepts the following options:
45077 size, s
45079 Specify the video size for the output. For the syntax of this
45080 option, check the "Video size" section in the ffmpeg-utils manual.
45081 Default value is "640x512".
45082 slide
45084 Specify how the spectrum should slide along the window.
45085 It accepts the following values:
45087 replace
45089 the samples start again on the left when they reach the right
45090 scroll
45092 the samples scroll from right to left
45093 fullframe
45095 frames are only produced when the samples reach the right
45096 rscroll
45098 the samples scroll from left to right
45099 lreplace
45101 the samples start again on the right when they reach the left
45102 Default value is "replace".
45104 mode
45106 Specify display mode.
45107 It accepts the following values:
45109 combined
45111 all channels are displayed in the same row
45112 separate
45114 all channels are displayed in separate rows
45115 Default value is combined.
45117 color
45119 Specify display color mode.
45120 It accepts the following values:
45122 channel
45124 each channel is displayed in a separate color
45125 intensity
45127 each channel is displayed using the same color scheme
45128 rainbow
45130 each channel is displayed using the rainbow color scheme
45131 moreland
45133 each channel is displayed using the moreland color scheme
45134 nebulae
45136 each channel is displayed using the nebulae color scheme
45137 fire
45139 each channel is displayed using the fire color scheme
45140 fiery
45142 each channel is displayed using the fiery color scheme
45143 fruit
45145 each channel is displayed using the fruit color scheme
45146 cool
45148 each channel is displayed using the cool color scheme
45149 magma
45151 each channel is displayed using the magma color scheme
45152 green
45154 each channel is displayed using the green color scheme
45155 viridis
45157 each channel is displayed using the viridis color scheme
45158 plasma
45160 each channel is displayed using the plasma color scheme
45161 cividis
45163 each channel is displayed using the cividis color scheme
45164 terrain
45166 each channel is displayed using the terrain color scheme
45167 Default value is channel.
45169 scale
45171 Specify scale used for calculating intensity color values.
45172 It accepts the following values:
45174 lin linear
45176 sqrt
45178 square root, default
45179 cbrt
45181 cubic root
45182 log logarithmic
45184 4thrt
45186 4th root
45187 5thrt
45189 5th root
45190 Default value is sqrt.
45192 fscale
45194 Specify frequency scale.
45195 It accepts the following values:
45197 lin linear
45199 log logarithmic
45201 Default value is lin.
45203 saturation
45205 Set saturation modifier for displayed colors. Negative values
45206 provide alternative color scheme. 0 is no saturation at all.
45207 Saturation must be in [-10.0, 10.0] range. Default value is 1.
45208 win_func
45210 Set window function.
45211 It accepts the following values:
45213 rect
45215 bartlett
45216 hann
45217 hanning
45218 hamming
45219 blackman
45220 welch
45221 flattop
45222 bharris
45223 bnuttall
45224 bhann
45225 sine
45226 nuttall
45227 lanczos
45228 gauss
45229 tukey
45230 dolph
45231 cauchy
45232 parzen
45233 poisson
45234 bohman
45235 kaiser
45236 Default value is "hann".
45238 orientation
45240 Set orientation of time vs frequency axis. Can be "vertical" or
45241 "horizontal". Default is "vertical".
45242 overlap
45244 Set ratio of overlap window. Default value is 0. When value is 1
45245 overlap is set to recommended size for specific window function
45246 currently used.
45247 gain
45249 Set scale gain for calculating intensity color values. Default
45250 value is 1.
45251 data
45253 Set which data to display. Can be "magnitude", default or "phase",
45254 or unwrapped phase: "uphase".
45255 rotation
45257 Set color rotation, must be in [-1.0, 1.0] range. Default value is
45258 0.
45259 start
45261 Set start frequency from which to display spectrogram. Default is
45262 0.
45263 stop
45265 Set stop frequency to which to display spectrogram. Default is 0.
45266 fps Set upper frame rate limit. Default is "auto", unlimited.
45268 legend
45270 Draw time and frequency axes and legends. Default is disabled.
45271 drange
45273 Set dynamic range used to calculate intensity color values. Default
45274 is 120 dBFS. Allowed range is from 10 to 200.
45275 limit
45277 Set upper limit of input audio samples volume in dBFS. Default is 0
45278 dBFS. Allowed range is from -100 to 100.
45279 opacity
45281 Set opacity strength when using pixel format output with alpha
45282 component.
45283 The usage is very similar to the showwaves filter; see the examples in
45285 that section.
45286 Examples
45288 • Large window with logarithmic color scaling:
45290 showspectrum=s=1280x480:scale=log
45292 • Complete example for a colored and sliding spectrum per channel
45294 using ffplay:
45295 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
45297 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
45298 showspectrumpic
45300 Convert input audio to a single video frame, representing the audio
45301 frequency spectrum.
45302 The filter accepts the following options:
45304 size, s
45306 Specify the video size for the output. For the syntax of this
45307 option, check the "Video size" section in the ffmpeg-utils manual.
45308 Default value is "4096x2048".
45309 mode
45311 Specify display mode.
45312 It accepts the following values:
45314 combined
45316 all channels are displayed in the same row
45317 separate
45319 all channels are displayed in separate rows
45320 Default value is combined.
45322 color
45324 Specify display color mode.
45325 It accepts the following values:
45327 channel
45329 each channel is displayed in a separate color
45330 intensity
45332 each channel is displayed using the same color scheme
45333 rainbow
45335 each channel is displayed using the rainbow color scheme
45336 moreland
45338 each channel is displayed using the moreland color scheme
45339 nebulae
45341 each channel is displayed using the nebulae color scheme
45342 fire
45344 each channel is displayed using the fire color scheme
45345 fiery
45347 each channel is displayed using the fiery color scheme
45348 fruit
45350 each channel is displayed using the fruit color scheme
45351 cool
45353 each channel is displayed using the cool color scheme
45354 magma
45356 each channel is displayed using the magma color scheme
45357 green
45359 each channel is displayed using the green color scheme
45360 viridis
45362 each channel is displayed using the viridis color scheme
45363 plasma
45365 each channel is displayed using the plasma color scheme
45366 cividis
45368 each channel is displayed using the cividis color scheme
45369 terrain
45371 each channel is displayed using the terrain color scheme
45372 Default value is intensity.
45374 scale
45376 Specify scale used for calculating intensity color values.
45377 It accepts the following values:
45379 lin linear
45381 sqrt
45383 square root, default
45384 cbrt
45386 cubic root
45387 log logarithmic
45389 4thrt
45391 4th root
45392 5thrt
45394 5th root
45395 Default value is log.
45397 fscale
45399 Specify frequency scale.
45400 It accepts the following values:
45402 lin linear
45404 log logarithmic
45406 Default value is lin.
45408 saturation
45410 Set saturation modifier for displayed colors. Negative values
45411 provide alternative color scheme. 0 is no saturation at all.
45412 Saturation must be in [-10.0, 10.0] range. Default value is 1.
45413 win_func
45415 Set window function.
45416 It accepts the following values:
45418 rect
45420 bartlett
45421 hann
45422 hanning
45423 hamming
45424 blackman
45425 welch
45426 flattop
45427 bharris
45428 bnuttall
45429 bhann
45430 sine
45431 nuttall
45432 lanczos
45433 gauss
45434 tukey
45435 dolph
45436 cauchy
45437 parzen
45438 poisson
45439 bohman
45440 kaiser
45441 Default value is "hann".
45443 orientation
45445 Set orientation of time vs frequency axis. Can be "vertical" or
45446 "horizontal". Default is "vertical".
45447 gain
45449 Set scale gain for calculating intensity color values. Default
45450 value is 1.
45451 legend
45453 Draw time and frequency axes and legends. Default is enabled.
45454 rotation
45456 Set color rotation, must be in [-1.0, 1.0] range. Default value is
45457 0.
45458 start
45460 Set start frequency from which to display spectrogram. Default is
45461 0.
45462 stop
45464 Set stop frequency to which to display spectrogram. Default is 0.
45465 drange
45467 Set dynamic range used to calculate intensity color values. Default
45468 is 120 dBFS. Allowed range is from 10 to 200.
45469 limit
45471 Set upper limit of input audio samples volume in dBFS. Default is 0
45472 dBFS. Allowed range is from -100 to 100.
45473 opacity
45475 Set opacity strength when using pixel format output with alpha
45476 component.
45477 Examples
45479 • Extract an audio spectrogram of a whole audio track in a 1024x1024
45481 picture using ffmpeg:
45482 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
45484 showvolume
45486 Convert input audio volume to a video output.
45487 The filter accepts the following options:
45489 rate, r
45491 Set video rate.
45492 b Set border width, allowed range is [0, 5]. Default is 1.
45494 w Set channel width, allowed range is [80, 8192]. Default is 400.
45496 h Set channel height, allowed range is [1, 900]. Default is 20.
45498 f Set fade, allowed range is [0, 1]. Default is 0.95.
45500 c Set volume color expression.
45502 The expression can use the following variables:
45504 VOLUME
45506 Current max volume of channel in dB.
45507 PEAK
45509 Current peak.
45510 CHANNEL
45512 Current channel number, starting from 0.
45513 t If set, displays channel names. Default is enabled.
45515 v If set, displays volume values. Default is enabled.
45517 o Set orientation, can be horizontal: "h" or vertical: "v", default
45519 is "h".
45520 s Set step size, allowed range is [0, 5]. Default is 0, which means
45522 step is disabled.
45523 p Set background opacity, allowed range is [0, 1]. Default is 0.
45525 m Set metering mode, can be peak: "p" or rms: "r", default is "p".
45527 ds Set display scale, can be linear: "lin" or log: "log", default is
45529 "lin".
45530 dm In second. If set to > 0., display a line for the max level in the
45532 previous seconds. default is disabled: 0.
45533 dmc The color of the max line. Use when "dm" option is set to > 0.
45535 default is: "orange"
45536 showwaves
45538 Convert input audio to a video output, representing the samples waves.
45539 The filter accepts the following options:
45541 size, s
45543 Specify the video size for the output. For the syntax of this
45544 option, check the "Video size" section in the ffmpeg-utils manual.
45545 Default value is "600x240".
45546 mode
45548 Set display mode.
45549 Available values are:
45551 point
45553 Draw a point for each sample.
45554 line
45556 Draw a vertical line for each sample.
45557 p2p Draw a point for each sample and a line between them.
45559 cline
45561 Draw a centered vertical line for each sample.
45562 Default value is "point".
45564 n Set the number of samples which are printed on the same column. A
45566 larger value will decrease the frame rate. Must be a positive
45567 integer. This option can be set only if the value for rate is not
45568 explicitly specified.
45569 rate, r
45571 Set the (approximate) output frame rate. This is done by setting
45572 the option n. Default value is "25".
45573 split_channels
45575 Set if channels should be drawn separately or overlap. Default
45576 value is 0.
45577 colors
45579 Set colors separated by '|' which are going to be used for drawing
45580 of each channel.
45581 scale
45583 Set amplitude scale.
45584 Available values are:
45586 lin Linear.
45588 log Logarithmic.
45590 sqrt
45592 Square root.
45593 cbrt
45595 Cubic root.
45596 Default is linear.
45598 draw
45600 Set the draw mode. This is mostly useful to set for high n.
45601 Available values are:
45603 scale
45605 Scale pixel values for each drawn sample.
45606 full
45608 Draw every sample directly.
45609 Default value is "scale".
45611 Examples
45613 • Output the input file audio and the corresponding video
45615 representation at the same time:
45616 amovie=a.mp3,asplit[out0],showwaves[out1]
45618 • Create a synthetic signal and show it with showwaves, forcing a
45620 frame rate of 30 frames per second:
45621 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
45623 showwavespic
45625 Convert input audio to a single video frame, representing the samples
45626 waves.
45627 The filter accepts the following options:
45629 size, s
45631 Specify the video size for the output. For the syntax of this
45632 option, check the "Video size" section in the ffmpeg-utils manual.
45633 Default value is "600x240".
45634 split_channels
45636 Set if channels should be drawn separately or overlap. Default
45637 value is 0.
45638 colors
45640 Set colors separated by '|' which are going to be used for drawing
45641 of each channel.
45642 scale
45644 Set amplitude scale.
45645 Available values are:
45647 lin Linear.
45649 log Logarithmic.
45651 sqrt
45653 Square root.
45654 cbrt
45656 Cubic root.
45657 Default is linear.
45659 draw
45661 Set the draw mode.
45662 Available values are:
45664 scale
45666 Scale pixel values for each drawn sample.
45667 full
45669 Draw every sample directly.
45670 Default value is "scale".
45672 filter
45674 Set the filter mode.
45675 Available values are:
45677 average
45679 Use average samples values for each drawn sample.
45680 peak
45682 Use peak samples values for each drawn sample.
45683 Default value is "average".
45685 Examples
45687 • Extract a channel split representation of the wave form of a whole
45689 audio track in a 1024x800 picture using ffmpeg:
45690 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
45692 sidedata, asidedata
45694 Delete frame side data, or select frames based on it.
45695 This filter accepts the following options:
45697 mode
45699 Set mode of operation of the filter.
45700 Can be one of the following:
45702 select
45704 Select every frame with side data of "type".
45705 delete
45707 Delete side data of "type". If "type" is not set, delete all
45708 side data in the frame.
45709 type
45711 Set side data type used with all modes. Must be set for "select"
45712 mode. For the list of frame side data types, refer to the
45713 "AVFrameSideDataType" enum in libavutil/frame.h. For example, to
45714 choose "AV_FRAME_DATA_PANSCAN" side data, you must specify
45715 "PANSCAN".
45716 spectrumsynth
45718 Synthesize audio from 2 input video spectrums, first input stream
45719 represents magnitude across time and second represents phase across
45720 time. The filter will transform from frequency domain as displayed in
45721 videos back to time domain as presented in audio output.
45722 This filter is primarily created for reversing processed showspectrum
45724 filter outputs, but can synthesize sound from other spectrograms too.
45725 But in such case results are going to be poor if the phase data is not
45726 available, because in such cases phase data need to be recreated,
45727 usually it's just recreated from random noise. For best results use
45728 gray only output ("channel" color mode in showspectrum filter) and
45729 "log" scale for magnitude video and "lin" scale for phase video. To
45730 produce phase, for 2nd video, use "data" option. Inputs videos should
45731 generally use "fullframe" slide mode as that saves resources needed for
45732 decoding video.
45733 The filter accepts the following options:
45735 sample_rate
45737 Specify sample rate of output audio, the sample rate of audio from
45738 which spectrum was generated may differ.
45739 channels
45741 Set number of channels represented in input video spectrums.
45742 scale
45744 Set scale which was used when generating magnitude input spectrum.
45745 Can be "lin" or "log". Default is "log".
45746 slide
45748 Set slide which was used when generating inputs spectrums. Can be
45749 "replace", "scroll", "fullframe" or "rscroll". Default is
45750 "fullframe".
45751 win_func
45753 Set window function used for resynthesis.
45754 overlap
45756 Set window overlap. In range "[0, 1]". Default is 1, which means
45757 optimal overlap for selected window function will be picked.
45758 orientation
45760 Set orientation of input videos. Can be "vertical" or "horizontal".
45761 Default is "vertical".
45762 Examples
45764 • First create magnitude and phase videos from audio, assuming audio
45766 is stereo with 44100 sample rate, then resynthesize videos back to
45767 audio with spectrumsynth:
45768 ffmpeg -i input.flac -lavfi showspectrum=mode=separate:scale=log:overlap=0.875:color=channel:slide=fullframe:data=magnitude -an -c:v rawvideo magnitude.nut
45770 ffmpeg -i input.flac -lavfi showspectrum=mode=separate:scale=lin:overlap=0.875:color=channel:slide=fullframe:data=phase -an -c:v rawvideo phase.nut
45771 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
45772 split, asplit
45774 Split input into several identical outputs.
45775 "asplit" works with audio input, "split" with video.
45777 The filter accepts a single parameter which specifies the number of
45779 outputs. If unspecified, it defaults to 2.
45780 Examples
45782 • Create two separate outputs from the same input:
45784 [in] split [out0][out1]
45786 • To create 3 or more outputs, you need to specify the number of
45788 outputs, like in:
45789 [in] asplit=3 [out0][out1][out2]
45791 • Create two separate outputs from the same input, one cropped and
45793 one padded:
45794 [in] split [splitout1][splitout2];
45796 [splitout1] crop=100:100:0:0 [cropout];
45797 [splitout2] pad=200:200:100:100 [padout];
45798 • Create 5 copies of the input audio with ffmpeg:
45800 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
45802 zmq, azmq
45804 Receive commands sent through a libzmq client, and forward them to
45805 filters in the filtergraph.
45806 "zmq" and "azmq" work as a pass-through filters. "zmq" must be inserted
45808 between two video filters, "azmq" between two audio filters. Both are
45809 capable to send messages to any filter type.
45810 To enable these filters you need to install the libzmq library and
45812 headers and configure FFmpeg with "--enable-libzmq".
45813 For more information about libzmq see: <http://www.zeromq.org/>
45815 The "zmq" and "azmq" filters work as a libzmq server, which receives
45817 messages sent through a network interface defined by the bind_address
45818 (or the abbreviation "b") option. Default value of this option is
45819 tcp://localhost:5555. You may want to alter this value to your needs,
45820 but do not forget to escape any ':' signs (see filtergraph escaping).
45821 The received message must be in the form:
45823 <TARGET> <COMMAND> [<ARG>]
45825 TARGET specifies the target of the command, usually the name of the
45827 filter class or a specific filter instance name. The default filter
45828 instance name uses the pattern Parsed_<filter_name>_<index>, but you
45829 can override this by using the filter_name@id syntax (see Filtergraph
45830 syntax).
45831 COMMAND specifies the name of the command for the target filter.
45833 ARG is optional and specifies the optional argument list for the given
45835 COMMAND.
45836 Upon reception, the message is processed and the corresponding command
45838 is injected into the filtergraph. Depending on the result, the filter
45839 will send a reply to the client, adopting the format:
45840 <ERROR_CODE> <ERROR_REASON>
45842 <MESSAGE>
45843 MESSAGE is optional.
45845 Examples
45847 Look at tools/zmqsend for an example of a zmq client which can be used
45849 to send commands processed by these filters.
45850 Consider the following filtergraph generated by ffplay. In this
45852 example the last overlay filter has an instance name. All other filters
45853 will have default instance names.
45854 ffplay -dumpgraph 1 -f lavfi "
45856 color=s=100x100:c=red [l];
45857 color=s=100x100:c=blue [r];
45858 nullsrc=s=200x100, zmq [bg];
45859 [bg][l] overlay [bg+l];
45860 [bg+l][r] overlay@my=x=100 "
45861 To change the color of the left side of the video, the following
45863 command can be used:
45864 echo Parsed_color_0 c yellow | tools/zmqsend
45866 To change the right side:
45868 echo Parsed_color_1 c pink | tools/zmqsend
45870 To change the position of the right side:
45872 echo overlay@my x 150 | tools/zmqsend
45874
45876 Below is a description of the currently available multimedia sources.
45877 amovie
45879 This is the same as movie source, except it selects an audio stream by
45880 default.
45881 avsynctest
45883 Generate an Audio/Video Sync Test.
45884 Generated stream periodically shows flash video frame and emits beep in
45886 audio. Useful to inspect A/V sync issues.
45887 It accepts the following options:
45889 size, s
45891 Set output video size. Default value is "hd720".
45892 framerate, fr
45894 Set output video frame rate. Default value is 30.
45895 samplerate, sr
45897 Set output audio sample rate. Default value is 44100.
45898 amplitude, a
45900 Set output audio beep amplitude. Default value is 0.7.
45901 period, p
45903 Set output audio beep period in seconds. Default value is 3.
45904 delay, dl
45906 Set output video flash delay in number of frames. Default value is
45907 0.
45908 cycle, c
45910 Enable cycling of video delays, by default is disabled.
45911 duration, d
45913 Set stream output duration. By default duration is unlimited.
45914 fg, bg, ag
45916 Set foreground/background/additional color.
45917 movie
45919 Read audio and/or video stream(s) from a movie container.
45920 It accepts the following parameters:
45922 filename
45924 The name of the resource to read (not necessarily a file; it can
45925 also be a device or a stream accessed through some protocol).
45926 format_name, f
45928 Specifies the format assumed for the movie to read, and can be
45929 either the name of a container or an input device. If not
45930 specified, the format is guessed from movie_name or by probing.
45931 seek_point, sp
45933 Specifies the seek point in seconds. The frames will be output
45934 starting from this seek point. The parameter is evaluated with
45935 "av_strtod", so the numerical value may be suffixed by an IS
45936 postfix. The default value is "0".
45937 streams, s
45939 Specifies the streams to read. Several streams can be specified,
45940 separated by "+". The source will then have as many outputs, in the
45941 same order. The syntax is explained in the "Stream specifiers"
45942 section in the ffmpeg manual. Two special names, "dv" and "da"
45943 specify respectively the default (best suited) video and audio
45944 stream. Default is "dv", or "da" if the filter is called as
45945 "amovie".
45946 stream_index, si
45948 Specifies the index of the video stream to read. If the value is
45949 -1, the most suitable video stream will be automatically selected.
45950 The default value is "-1". Deprecated. If the filter is called
45951 "amovie", it will select audio instead of video.
45952 loop
45954 Specifies how many times to read the stream in sequence. If the
45955 value is 0, the stream will be looped infinitely. Default value is
45956 "1".
45957 Note that when the movie is looped the source timestamps are not
45959 changed, so it will generate non monotonically increasing
45960 timestamps.
45961 discontinuity
45963 Specifies the time difference between frames above which the point
45964 is considered a timestamp discontinuity which is removed by
45965 adjusting the later timestamps.
45966 dec_threads
45968 Specifies the number of threads for decoding
45969 format_opts
45971 Specify format options for the opened file. Format options can be
45972 specified as a list of key=value pairs separated by ':'. The
45973 following example shows how to add protocol_whitelist and
45974 protocol_blacklist options:
45975 ffplay -f lavfi
45977 "movie=filename='1.sdp':format_opts='protocol_whitelist=file,rtp,udp\:protocol_blacklist=http'"
45978 It allows overlaying a second video on top of the main input of a
45980 filtergraph, as shown in this graph:
45981 input -----------> deltapts0 --> overlay --> output
45983 ^
45984 |
45985 movie --> scale--> deltapts1 -------+
45986 Examples
45988 • Skip 3.2 seconds from the start of the AVI file in.avi, and overlay
45990 it on top of the input labelled "in":
45991 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
45993 [in] setpts=PTS-STARTPTS [main];
45994 [main][over] overlay=16:16 [out]
45995 • Read from a video4linux2 device, and overlay it on top of the input
45997 labelled "in":
45998 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
46000 [in] setpts=PTS-STARTPTS [main];
46001 [main][over] overlay=16:16 [out]
46002 • Read the first video stream and the audio stream with id 0x81 from
46004 dvd.vob; the video is connected to the pad named "video" and the
46005 audio is connected to the pad named "audio":
46006 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
46008 Commands
46010 Both movie and amovie support the following commands:
46012 seek
46014 Perform seek using "av_seek_frame". The syntax is: seek
46015 stream_index|timestamp|flags
46016 • stream_index: If stream_index is -1, a default stream is
46018 selected, and timestamp is automatically converted from
46019 AV_TIME_BASE units to the stream specific time_base.
46020 • timestamp: Timestamp in AVStream.time_base units or, if no
46022 stream is specified, in AV_TIME_BASE units.
46023 • flags: Flags which select direction and seeking mode.
46025 get_duration
46027 Get movie duration in AV_TIME_BASE units.
46028
46030 FFmpeg can be hooked up with a number of external libraries to add
46031 support for more formats. None of them are used by default, their use
46032 has to be explicitly requested by passing the appropriate flags to
46033 ./configure.
46034 Alliance for Open Media (AOM)
46036 FFmpeg can make use of the AOM library for AV1 decoding and encoding.
46037 Go to <http://aomedia.org/> and follow the instructions for installing
46039 the library. Then pass "--enable-libaom" to configure to enable it.
46040 AMD AMF/VCE
46042 FFmpeg can use the AMD Advanced Media Framework library for accelerated
46043 H.264 and HEVC(only windows) encoding on hardware with Video Coding
46044 Engine (VCE).
46045 To enable support you must obtain the AMF framework header
46047 files(version 1.4.9+) from
46048 <https://github.com/GPUOpen-LibrariesAndSDKs/AMF.git>.
46049 Create an "AMF/" directory in the system include path. Copy the
46051 contents of "AMF/amf/public/include/" into that directory. Then
46052 configure FFmpeg with "--enable-amf".
46053 Initialization of amf encoder occurs in this order: 1) trying to
46055 initialize through dx11(only windows) 2) trying to initialize through
46056 dx9(only windows) 3) trying to initialize through vulkan
46057 To use h.264(AMD VCE) encoder on linux amdgru-pro version 19.20+ and
46059 amf-amdgpu-pro package(amdgru-pro contains, but does not install
46060 automatically) are required.
46061 This driver can be installed using amdgpu-pro-install script in
46063 official amd driver archive.
46064 AviSynth
46066 FFmpeg can read AviSynth scripts as input. To enable support, pass
46067 "--enable-avisynth" to configure after installing the headers provided
46068 by <https://github.com/AviSynth/AviSynthPlus>. AviSynth+ can be
46069 configured to install only the headers by either passing
46070 "-DHEADERS_ONLY:bool=on" to the normal CMake-based build system, or by
46071 using the supplied "GNUmakefile".
46072 For Windows, supported AviSynth variants are <http://avisynth.nl> for
46074 32-bit builds and <http://avisynth.nl/index.php/AviSynth+> for 32-bit
46075 and 64-bit builds.
46076 For Linux, macOS, and BSD, the only supported AviSynth variant is
46078 <https://github.com/AviSynth/AviSynthPlus>, starting with version 3.5.
46079 In 2016, AviSynth+ added support for building with GCC. However,
46081 due to the eccentricities of Windows' calling conventions, 32-bit
46082 GCC builds of AviSynth+ are not compatible with typical 32-bit
46083 builds of FFmpeg.
46084 By default, FFmpeg assumes compatibility with 32-bit MSVC builds of
46086 AviSynth+ since that is the most widely-used and entrenched build
46087 configuration. Users can override this and enable support for
46088 32-bit GCC builds of AviSynth+ by passing "-DAVSC_WIN32_GCC32" to
46089 "--extra-cflags" when configuring FFmpeg.
46090 64-bit builds of FFmpeg are not affected, and can use either MSVC
46092 or GCC builds of AviSynth+ without any special flags.
46093 AviSynth(+) is loaded dynamically. Distributors can build FFmpeg
46095 with "--enable-avisynth", and the binaries will work regardless of
46096 the end user having AviSynth installed. If/when an end user would
46097 like to use AviSynth scripts, then they can install AviSynth(+) and
46098 FFmpeg will be able to find and use it to open scripts.
46099 Chromaprint
46101 FFmpeg can make use of the Chromaprint library for generating audio
46102 fingerprints. Pass "--enable-chromaprint" to configure to enable it.
46103 See <https://acoustid.org/chromaprint>.
46104 codec2
46106 FFmpeg can make use of the codec2 library for codec2 decoding and
46107 encoding. There is currently no native decoder, so libcodec2 must be
46108 used for decoding.
46109 Go to <http://freedv.org/>, download "Codec 2 source archive". Build
46111 and install using CMake. Debian users can install the libcodec2-dev
46112 package instead. Once libcodec2 is installed you can pass
46113 "--enable-libcodec2" to configure to enable it.
46114 The easiest way to use codec2 is with .c2 files, since they contain the
46116 mode information required for decoding. To encode such a file, use a
46117 .c2 file extension and give the libcodec2 encoder the -mode option:
46118 "ffmpeg -i input.wav -mode 700C output.c2". Playback is as simple as
46119 "ffplay output.c2". For a list of supported modes, run "ffmpeg -h
46120 encoder=libcodec2". Raw codec2 files are also supported. To make
46121 sense of them the mode in use needs to be specified as a format option:
46122 "ffmpeg -f codec2raw -mode 1300 -i input.raw output.wav".
46123 dav1d
46125 FFmpeg can make use of the dav1d library for AV1 video decoding.
46126 Go to <https://code.videolan.org/videolan/dav1d> and follow the
46128 instructions for installing the library. Then pass "--enable-libdav1d"
46129 to configure to enable it.
46130 davs2
46132 FFmpeg can make use of the davs2 library for AVS2-P2/IEEE1857.4 video
46133 decoding.
46134 Go to <https://github.com/pkuvcl/davs2> and follow the instructions for
46136 installing the library. Then pass "--enable-libdavs2" to configure to
46137 enable it.
46138 libdavs2 is under the GNU Public License Version 2 or later (see
46140 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
46141 details), you must upgrade FFmpeg's license to GPL in order to use
46142 it.
46143 uavs3d
46145 FFmpeg can make use of the uavs3d library for AVS3-P2/IEEE1857.10 video
46146 decoding.
46147 Go to <https://github.com/uavs3/uavs3d> and follow the instructions for
46149 installing the library. Then pass "--enable-libuavs3d" to configure to
46150 enable it.
46151 Game Music Emu
46153 FFmpeg can make use of the Game Music Emu library to read audio from
46154 supported video game music file formats. Pass "--enable-libgme" to
46155 configure to enable it. See
46156 <https://bitbucket.org/mpyne/game-music-emu/overview>.
46157 Intel QuickSync Video
46159 FFmpeg can use Intel QuickSync Video (QSV) for accelerated decoding and
46160 encoding of multiple codecs. To use QSV, FFmpeg must be linked against
46161 the "libmfx" dispatcher, which loads the actual decoding libraries.
46162 The dispatcher is open source and can be downloaded from
46164 <https://github.com/lu-zero/mfx_dispatch.git>. FFmpeg needs to be
46165 configured with the "--enable-libmfx" option and "pkg-config" needs to
46166 be able to locate the dispatcher's ".pc" files.
46167 Kvazaar
46169 FFmpeg can make use of the Kvazaar library for HEVC encoding.
46170 Go to <https://github.com/ultravideo/kvazaar> and follow the
46172 instructions for installing the library. Then pass
46173 "--enable-libkvazaar" to configure to enable it.
46174 LAME
46176 FFmpeg can make use of the LAME library for MP3 encoding.
46177 Go to <http://lame.sourceforge.net/> and follow the instructions for
46179 installing the library. Then pass "--enable-libmp3lame" to configure
46180 to enable it.
46181 libilbc
46183 iLBC is a narrowband speech codec that has been made freely available
46184 by Google as part of the WebRTC project. libilbc is a packaging
46185 friendly copy of the iLBC codec. FFmpeg can make use of the libilbc
46186 library for iLBC decoding and encoding.
46187 Go to <https://github.com/TimothyGu/libilbc> and follow the
46189 instructions for installing the library. Then pass "--enable-libilbc"
46190 to configure to enable it.
46191 libjxl
46193 JPEG XL is an image format intended to fully replace legacy JPEG for an
46194 extended period of life. See <https://jpegxl.info/> for more
46195 information, and see <https://github.com/libjxl/libjxl> for the library
46196 source. You can pass "--enable-libjxl" to configure in order enable the
46197 libjxl wrapper.
46198 libvpx
46200 FFmpeg can make use of the libvpx library for VP8/VP9 decoding and
46201 encoding.
46202 Go to <http://www.webmproject.org/> and follow the instructions for
46204 installing the library. Then pass "--enable-libvpx" to configure to
46205 enable it.
46206 ModPlug
46208 FFmpeg can make use of this library, originating in Modplug-XMMS, to
46209 read from MOD-like music files. See
46210 <https://github.com/Konstanty/libmodplug>. Pass "--enable-libmodplug"
46211 to configure to enable it.
46212 OpenCORE, VisualOn, and Fraunhofer libraries
46214 Spun off Google Android sources, OpenCore, VisualOn and Fraunhofer
46215 libraries provide encoders for a number of audio codecs.
46216 OpenCORE and VisualOn libraries are under the Apache License 2.0
46218 (see <http://www.apache.org/licenses/LICENSE-2.0> for details),
46219 which is incompatible to the LGPL version 2.1 and GPL version 2.
46220 You have to upgrade FFmpeg's license to LGPL version 3 (or if you
46221 have enabled GPL components, GPL version 3) by passing
46222 "--enable-version3" to configure in order to use it.
46223 The license of the Fraunhofer AAC library is incompatible with the
46225 GPL. Therefore, for GPL builds, you have to pass
46226 "--enable-nonfree" to configure in order to use it. To the best of
46227 our knowledge, it is compatible with the LGPL.
46228 OpenCORE AMR
46230 FFmpeg can make use of the OpenCORE libraries for AMR-NB
46232 decoding/encoding and AMR-WB decoding.
46233 Go to <http://sourceforge.net/projects/opencore-amr/> and follow the
46235 instructions for installing the libraries. Then pass
46236 "--enable-libopencore-amrnb" and/or "--enable-libopencore-amrwb" to
46237 configure to enable them.
46238 VisualOn AMR-WB encoder library
46240 FFmpeg can make use of the VisualOn AMR-WBenc library for AMR-WB
46242 encoding.
46243 Go to <http://sourceforge.net/projects/opencore-amr/> and follow the
46245 instructions for installing the library. Then pass
46246 "--enable-libvo-amrwbenc" to configure to enable it.
46247 Fraunhofer AAC library
46249 FFmpeg can make use of the Fraunhofer AAC library for AAC decoding &
46251 encoding.
46252 Go to <http://sourceforge.net/projects/opencore-amr/> and follow the
46254 instructions for installing the library. Then pass
46255 "--enable-libfdk-aac" to configure to enable it.
46256 OpenH264
46258 FFmpeg can make use of the OpenH264 library for H.264 decoding and
46259 encoding.
46260 Go to <http://www.openh264.org/> and follow the instructions for
46262 installing the library. Then pass "--enable-libopenh264" to configure
46263 to enable it.
46264 For decoding, this library is much more limited than the built-in
46266 decoder in libavcodec; currently, this library lacks support for
46267 decoding B-frames and some other main/high profile features. (It
46268 currently only supports constrained baseline profile and CABAC.) Using
46269 it is mostly useful for testing and for taking advantage of Cisco's
46270 patent portfolio license
46271 (<http://www.openh264.org/BINARY_LICENSE.txt>).
46272 OpenJPEG
46274 FFmpeg can use the OpenJPEG libraries for decoding/encoding J2K videos.
46275 Go to <http://www.openjpeg.org/> to get the libraries and follow the
46276 installation instructions. To enable using OpenJPEG in FFmpeg, pass
46277 "--enable-libopenjpeg" to ./configure.
46278 rav1e
46280 FFmpeg can make use of rav1e (Rust AV1 Encoder) via its C bindings to
46281 encode videos. Go to <https://github.com/xiph/rav1e/> and follow the
46282 instructions to build the C library. To enable using rav1e in FFmpeg,
46283 pass "--enable-librav1e" to ./configure.
46284 SVT-AV1
46286 FFmpeg can make use of the Scalable Video Technology for AV1 library
46287 for AV1 encoding.
46288 Go to <https://gitlab.com/AOMediaCodec/SVT-AV1/> and follow the
46290 instructions for installing the library. Then pass "--enable-libsvtav1"
46291 to configure to enable it.
46292 TwoLAME
46294 FFmpeg can make use of the TwoLAME library for MP2 encoding.
46295 Go to <http://www.twolame.org/> and follow the instructions for
46297 installing the library. Then pass "--enable-libtwolame" to configure
46298 to enable it.
46299 VapourSynth
46301 FFmpeg can read VapourSynth scripts as input. To enable support, pass
46302 "--enable-vapoursynth" to configure. Vapoursynth is detected via
46303 "pkg-config". Versions 42 or greater supported. See
46304 <http://www.vapoursynth.com/>.
46305 Due to security concerns, Vapoursynth scripts will not be autodetected
46307 so the input format has to be forced. For ff* CLI tools, add "-f
46308 vapoursynth" before the input "-i yourscript.vpy".
46309 x264
46311 FFmpeg can make use of the x264 library for H.264 encoding.
46312 Go to <http://www.videolan.org/developers/x264.html> and follow the
46314 instructions for installing the library. Then pass "--enable-libx264"
46315 to configure to enable it.
46316 x264 is under the GNU Public License Version 2 or later (see
46318 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
46319 details), you must upgrade FFmpeg's license to GPL in order to use
46320 it.
46321 x265
46323 FFmpeg can make use of the x265 library for HEVC encoding.
46324 Go to <http://x265.org/developers.html> and follow the instructions for
46326 installing the library. Then pass "--enable-libx265" to configure to
46327 enable it.
46328 x265 is under the GNU Public License Version 2 or later (see
46330 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
46331 details), you must upgrade FFmpeg's license to GPL in order to use
46332 it.
46333 xavs
46335 FFmpeg can make use of the xavs library for AVS encoding.
46336 Go to <http://xavs.sf.net/> and follow the instructions for installing
46338 the library. Then pass "--enable-libxavs" to configure to enable it.
46339 xavs2
46341 FFmpeg can make use of the xavs2 library for AVS2-P2/IEEE1857.4 video
46342 encoding.
46343 Go to <https://github.com/pkuvcl/xavs2> and follow the instructions for
46345 installing the library. Then pass "--enable-libxavs2" to configure to
46346 enable it.
46347 libxavs2 is under the GNU Public License Version 2 or later (see
46349 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
46350 details), you must upgrade FFmpeg's license to GPL in order to use
46351 it.
46352 ZVBI
46354 ZVBI is a VBI decoding library which can be used by FFmpeg to decode
46355 DVB teletext pages and DVB teletext subtitles.
46356 Go to <http://sourceforge.net/projects/zapping/> and follow the
46358 instructions for installing the library. Then pass "--enable-libzvbi"
46359 to configure to enable it.
46360
46362 You can use the "-formats" and "-codecs" options to have an exhaustive
46363 list.
46364 File Formats
46366 FFmpeg supports the following file formats through the "libavformat"
46367 library:
46368 Name : Encoding @tab Decoding @tab Comments
46370 3dostr : @tab X
46371 4xm : @tab X
46372 @tab 4X Technologies format, used in some games.
46373 8088flex TMV : @tab X
46375 AAX : @tab X
46376 @tab Audible Enhanced Audio format, used in audiobooks.
46377 AA : @tab X
46379 @tab Audible Format 2, 3, and 4, used in audiobooks.
46380 ACT Voice : @tab X
46382 @tab contains G.729 audio
46383 Adobe Filmstrip : X @tab X
46385 Audio IFF (AIFF) : X @tab X
46386 American Laser Games MM : @tab X
46387 @tab Multimedia format used in games like Mad Dog McCree.
46388 3GPP AMR : X @tab X
46390 Amazing Studio Packed Animation File : @tab X
46391 @tab Multimedia format used in game Heart Of Darkness.
46392 Apple HTTP Live Streaming : @tab X
46394 Artworx Data Format : @tab X
46395 Interplay ACM : @tab X
46396 @tab Audio only format used in some Interplay games.
46397 ADP : @tab X
46399 @tab Audio format used on the Nintendo Gamecube.
46400 AFC : @tab X
46402 @tab Audio format used on the Nintendo Gamecube.
46403 ADS/SS2 : @tab X
46405 @tab Audio format used on the PS2.
46406 APNG : X @tab X
46408 ASF : X @tab X
46409 @tab Advanced / Active Streaming Format.
46410 AST : X @tab X
46412 @tab Audio format used on the Nintendo Wii.
46413 AVI : X @tab X
46415 AviSynth : @tab X
46416 AVR : @tab X
46417 @tab Audio format used on Mac.
46418 AVS : @tab X
46420 @tab Multimedia format used by the Creature Shock game.
46421 Beam Software SIFF : @tab X
46423 @tab Audio and video format used in some games by Beam Software.
46424 Bethesda Softworks VID : @tab X
46426 @tab Used in some games from Bethesda Softworks.
46427 Binary text : @tab X
46429 Bink : @tab X
46430 @tab Multimedia format used by many games.
46431 Bink Audio : @tab X
46433 @tab Audio only multimedia format used by some games.
46434 Bitmap Brothers JV : @tab X
46436 @tab Used in Z and Z95 games.
46437 BRP : @tab X
46439 @tab Argonaut Games format.
46440 Brute Force & Ignorance : @tab X
46442 @tab Used in the game Flash Traffic: City of Angels.
46443 BFSTM : @tab X
46445 @tab Audio format used on the Nintendo WiiU (based on BRSTM).
46446 BRSTM : @tab X
46448 @tab Audio format used on the Nintendo Wii.
46449 BW64 : @tab X
46451 @tab Broadcast Wave 64bit.
46452 BWF : X @tab X
46454 codec2 (raw) : X @tab X
46455 @tab Must be given -mode format option to decode correctly.
46456 codec2 (.c2 files) : X @tab X
46458 @tab Contains header with version and mode info, simplifying playback.
46459 CRI ADX : X @tab X
46461 @tab Audio-only format used in console video games.
46462 CRI AIX : @tab X
46464 CRI HCA : @tab X
46465 @tab Audio-only format used in console video games.
46466 Discworld II BMV : @tab X
46468 Interplay C93 : @tab X
46469 @tab Used in the game Cyberia from Interplay.
46470 Delphine Software International CIN : @tab X
46472 @tab Multimedia format used by Delphine Software games.
46473 Digital Speech Standard (DSS) : @tab X
46475 CD+G : @tab X
46476 @tab Video format used by CD+G karaoke disks
46477 Phantom Cine : @tab X
46479 Commodore CDXL : @tab X
46480 @tab Amiga CD video format
46481 Core Audio Format : X @tab X
46483 @tab Apple Core Audio Format
46484 CRC testing format : X @tab
46486 Creative Voice : X @tab X
46487 @tab Created for the Sound Blaster Pro.
46488 CRYO APC : @tab X
46490 @tab Audio format used in some games by CRYO Interactive Entertainment.
46491 D-Cinema audio : X @tab X
46493 Deluxe Paint Animation : @tab X
46494 DCSTR : @tab X
46495 DFA : @tab X
46496 @tab This format is used in Chronomaster game
46497 DirectDraw Surface : @tab X
46499 DSD Stream File (DSF) : @tab X
46500 DV video : X @tab X
46501 DXA : @tab X
46502 @tab This format is used in the non-Windows version of the Feeble Files
46503 game and different game cutscenes repacked for use with ScummVM.
46504 Electronic Arts cdata : @tab X
46506 Electronic Arts Multimedia : @tab X
46507 @tab Used in various EA games; files have extensions like WVE and UV2.
46508 Ensoniq Paris Audio File : @tab X
46510 FFM (FFserver live feed) : X @tab X
46511 Flash (SWF) : X @tab X
46512 Flash 9 (AVM2) : X @tab X
46513 @tab Only embedded audio is decoded.
46514 FLI/FLC/FLX animation : @tab X
46516 @tab .fli/.flc files
46517 Flash Video (FLV) : X @tab X
46519 @tab Macromedia Flash video files
46520 framecrc testing format : X @tab
46522 FunCom ISS : @tab X
46523 @tab Audio format used in various games from FunCom like The Longest Journey.
46524 G.723.1 : X @tab X
46526 G.726 : @tab X @tab Both left- and right-
46527 justified.
46528 G.729 BIT : X @tab X
46529 G.729 raw : @tab X
46530 GENH : @tab X
46531 @tab Audio format for various games.
46532 GIF Animation : X @tab X
46534 GXF : X @tab X
46535 @tab General eXchange Format SMPTE 360M, used by Thomson Grass Valley
46536 playout servers.
46537 HNM : @tab X
46539 @tab Only version 4 supported, used in some games from Cryo Interactive
46540 iCEDraw File : @tab X
46542 ICO : X @tab X
46543 @tab Microsoft Windows ICO
46544 id Quake II CIN video : @tab X
46546 id RoQ : X @tab X
46547 @tab Used in Quake III, Jedi Knight 2 and other computer games.
46548 IEC61937 encapsulation : X @tab X
46550 IFF : @tab X
46551 @tab Interchange File Format
46552 IFV : @tab X
46554 @tab A format used by some old CCTV DVRs.
46555 iLBC : X @tab X
46557 Interplay MVE : @tab X
46558 @tab Format used in various Interplay computer games.
46559 Iterated Systems ClearVideo : @tab X
46561 @tab I-frames only
46562 IV8 : @tab X
46564 @tab A format generated by IndigoVision 8000 video server.
46565 IVF (On2) : X @tab X
46567 @tab A format used by libvpx
46568 Internet Video Recording : @tab X
46570 IRCAM : X @tab X
46571 LAF : @tab X
46572 @tab Limitless Audio Format
46573 LATM : X @tab X
46575 LMLM4 : @tab X
46576 @tab Used by Linux Media Labs MPEG-4 PCI boards
46577 LOAS : @tab X
46579 @tab contains LATM multiplexed AAC audio
46580 LRC : X @tab X
46582 LVF : @tab X
46583 LXF : @tab X
46584 @tab VR native stream format, used by Leitch/Harris' video servers.
46585 Magic Lantern Video (MLV) : @tab X
46587 Matroska : X @tab X
46588 Matroska audio : X @tab
46589 FFmpeg metadata : X @tab X
46590 @tab Metadata in text format.
46591 MAXIS XA : @tab X
46593 @tab Used in Sim City 3000; file extension .xa.
46594 MCA : @tab X
46596 @tab Used in some games from Capcom; file extension .mca.
46597 MD Studio : @tab X
46599 Metal Gear Solid: The Twin Snakes : @tab X
46600 Megalux Frame : @tab X
46601 @tab Used by Megalux Ultimate Paint
46602 MobiClip MODS : @tab X
46604 MobiClip MOFLEX : @tab X
46605 Mobotix .mxg : @tab X
46606 Monkey's Audio : @tab X
46607 Motion Pixels MVI : @tab X
46608 MOV/QuickTime/MP4 : X @tab X
46609 @tab 3GP, 3GP2, PSP, iPod variants supported
46610 MP2 : X @tab X
46612 MP3 : X @tab X
46613 MPEG-1 System : X @tab X
46614 @tab muxed audio and video, VCD format supported
46615 MPEG-PS (program stream) : X @tab X
46617 @tab also known as C<VOB> file, SVCD and DVD format supported
46618 MPEG-TS (transport stream) : X @tab X
46620 @tab also known as DVB Transport Stream
46621 MPEG-4 : X @tab X
46623 @tab MPEG-4 is a variant of QuickTime.
46624 MSF : @tab X
46626 @tab Audio format used on the PS3.
46627 Mirillis FIC video : @tab X
46629 @tab No cursor rendering.
46630 MIDI Sample Dump Standard : @tab X
46632 MIME multipart JPEG : X @tab
46633 MSN TCP webcam : @tab X
46634 @tab Used by MSN Messenger webcam streams.
46635 MTV : @tab X
46637 Musepack : @tab X
46638 Musepack SV8 : @tab X
46639 Material eXchange Format (MXF) : X @tab X
46640 @tab SMPTE 377M, used by D-Cinema, broadcast industry.
46641 Material eXchange Format (MXF), D-10 Mapping : X @tab X
46643 @tab SMPTE 386M, D-10/IMX Mapping.
46644 NC camera feed : @tab X
46646 @tab NC (AVIP NC4600) camera streams
46647 NIST SPeech HEader REsources : @tab X
46649 Computerized Speech Lab NSP : @tab X
46650 NTT TwinVQ (VQF) : @tab X
46651 @tab Nippon Telegraph and Telephone Corporation TwinVQ.
46652 Nullsoft Streaming Video : @tab X
46654 NuppelVideo : @tab X
46655 NUT : X @tab X
46656 @tab NUT Open Container Format
46657 Ogg : X @tab X
46659 Playstation Portable PMP : @tab X
46660 Portable Voice Format : @tab X
46661 RK Audio (RKA) : @tab X
46662 TechnoTrend PVA : @tab X
46663 @tab Used by TechnoTrend DVB PCI boards.
46664 QCP : @tab X
46666 raw ADTS (AAC) : X @tab X
46667 raw AC-3 : X @tab X
46668 raw AMR-NB : @tab X
46669 raw AMR-WB : @tab X
46670 raw APAC : @tab X
46671 raw aptX : X @tab X
46672 raw aptX HD : X @tab X
46673 raw Bonk : @tab X
46674 raw Chinese AVS video : X @tab X
46675 raw DFPWM : X @tab X
46676 raw Dirac : X @tab X
46677 raw DNxHD : X @tab X
46678 raw DTS : X @tab X
46679 raw DTS-HD : @tab X
46680 raw E-AC-3 : X @tab X
46681 raw FLAC : X @tab X
46682 raw GSM : @tab X
46683 raw H.261 : X @tab X
46684 raw H.263 : X @tab X
46685 raw H.264 : X @tab X
46686 raw HEVC : X @tab X
46687 raw Ingenient MJPEG : @tab X
46688 raw MJPEG : X @tab X
46689 raw MLP : @tab X
46690 raw MPEG : @tab X
46691 raw MPEG-1 : @tab X
46692 raw MPEG-2 : @tab X
46693 raw MPEG-4 : X @tab X
46694 raw NULL : X @tab
46695 raw video : X @tab X
46696 raw id RoQ : X @tab
46697 raw OBU : X @tab X
46698 raw SBC : X @tab X
46699 raw Shorten : @tab X
46700 raw TAK : @tab X
46701 raw TrueHD : X @tab X
46702 raw VC-1 : X @tab X
46703 raw PCM A-law : X @tab X
46704 raw PCM mu-law : X @tab X
46705 raw PCM Archimedes VIDC : X @tab X
46706 raw PCM signed 8 bit : X @tab X
46707 raw PCM signed 16 bit big-endian : X @tab X
46708 raw PCM signed 16 bit little-endian : X @tab X
46709 raw PCM signed 24 bit big-endian : X @tab X
46710 raw PCM signed 24 bit little-endian : X @tab X
46711 raw PCM signed 32 bit big-endian : X @tab X
46712 raw PCM signed 32 bit little-endian : X @tab X
46713 raw PCM signed 64 bit big-endian : X @tab X
46714 raw PCM signed 64 bit little-endian : X @tab X
46715 raw PCM unsigned 8 bit : X @tab X
46716 raw PCM unsigned 16 bit big-endian : X @tab X
46717 raw PCM unsigned 16 bit little-endian : X @tab X
46718 raw PCM unsigned 24 bit big-endian : X @tab X
46719 raw PCM unsigned 24 bit little-endian : X @tab X
46720 raw PCM unsigned 32 bit big-endian : X @tab X
46721 raw PCM unsigned 32 bit little-endian : X @tab X
46722 raw PCM 16.8 floating point little-endian : @tab X
46723 raw PCM 24.0 floating point little-endian : @tab X
46724 raw PCM floating-point 32 bit big-endian : X @tab X
46725 raw PCM floating-point 32 bit little-endian : X @tab X
46726 raw PCM floating-point 64 bit big-endian : X @tab X
46727 raw PCM floating-point 64 bit little-endian : X @tab X
46728 RDT : @tab X
46729 REDCODE R3D : @tab X
46730 @tab File format used by RED Digital cameras, contains JPEG 2000 frames and PCM audio.
46731 RealMedia : X @tab X
46733 Redirector : @tab X
46734 RedSpark : @tab X
46735 Renderware TeXture Dictionary : @tab X
46736 Resolume DXV : @tab X
46737 RF64 : @tab X
46738 RL2 : @tab X
46739 @tab Audio and video format used in some games by Entertainment Software Partners.
46740 RPL/ARMovie : @tab X
46742 Lego Mindstorms RSO : X @tab X
46743 RSD : @tab X
46744 RTMP : X @tab X
46745 @tab Output is performed by publishing stream to RTMP server
46746 RTP : X @tab X
46748 RTSP : X @tab X
46749 Sample Dump eXchange : @tab X
46750 SAP : X @tab X
46751 SBG : @tab X
46752 SDNS : @tab X
46753 SDP : @tab X
46754 SER : @tab X
46755 Digital Pictures SGA : @tab X
46756 Sega FILM/CPK : X @tab X
46757 @tab Used in many Sega Saturn console games.
46758 Silicon Graphics Movie : @tab X
46760 Sierra SOL : @tab X
46761 @tab .sol files used in Sierra Online games.
46762 Sierra VMD : @tab X
46764 @tab Used in Sierra CD-ROM games.
46765 Smacker : @tab X
46767 @tab Multimedia format used by many games.
46768 SMJPEG : X @tab X
46770 @tab Used in certain Loki game ports.
46771 SMPTE 337M encapsulation : @tab X
46773 Smush : @tab X
46774 @tab Multimedia format used in some LucasArts games.
46775 Sony OpenMG (OMA) : X @tab X
46777 @tab Audio format used in Sony Sonic Stage and Sony Vegas.
46778 Sony PlayStation STR : @tab X
46780 Sony Wave64 (W64) : X @tab X
46781 SoX native format : X @tab X
46782 SUN AU format : X @tab X
46783 SUP raw PGS subtitles : X @tab X
46784 SVAG : @tab X
46785 @tab Audio format used in Konami PS2 games.
46786 TDSC : @tab X
46788 Text files : @tab X
46789 THP : @tab X
46790 @tab Used on the Nintendo GameCube.
46791 Tiertex Limited SEQ : @tab X
46793 @tab Tiertex .seq files used in the DOS CD-ROM version of the game Flashback.
46794 True Audio : X @tab X
46796 VAG : @tab X
46797 @tab Audio format used in many Sony PS2 games.
46798 VC-1 test bitstream : X @tab X
46800 Vidvox Hap : X @tab X
46801 Vivo : @tab X
46802 VPK : @tab X
46803 @tab Audio format used in Sony PS games.
46804 Marble WADY : @tab X
46806 WAV : X @tab X
46807 Waveform Archiver : @tab X
46808 WavPack : X @tab X
46809 WebM : X @tab X
46810 Windows Televison (WTV) : X @tab X
46811 Wing Commander III movie : @tab X
46812 @tab Multimedia format used in Origin's Wing Commander III computer game.
46813 Westwood Studios audio : X @tab X
46815 @tab Multimedia format used in Westwood Studios games.
46816 Westwood Studios VQA : @tab X
46818 @tab Multimedia format used in Westwood Studios games.
46819 Wideband Single-bit Data (WSD) : @tab X
46821 WVE : @tab X
46822 Konami XMD : @tab X
46823 XMV : @tab X
46824 @tab Microsoft video container used in Xbox games.
46825 XVAG : @tab X
46827 @tab Audio format used on the PS3.
46828 xWMA : @tab X
46830 @tab Microsoft audio container used by XAudio 2.
46831 eXtended BINary text (XBIN) : @tab X
46833 YUV4MPEG pipe : X @tab X
46834 Psygnosis YOP : @tab X
46835 "X" means that the feature in that column (encoding / decoding) is
46837 supported.
46838 Image Formats
46840 FFmpeg can read and write images for each frame of a video sequence.
46841 The following image formats are supported:
46842 Name : Encoding @tab Decoding @tab Comments
46844 .Y.U.V : X @tab X
46845 @tab one raw file per component
46846 Alias PIX : X @tab X
46848 @tab Alias/Wavefront PIX image format
46849 animated GIF : X @tab X
46851 APNG : X @tab X
46852 @tab Animated Portable Network Graphics
46853 BMP : X @tab X
46855 @tab Microsoft BMP image
46856 BRender PIX : @tab X
46858 @tab Argonaut BRender 3D engine image format.
46859 CRI : @tab X
46861 @tab Cintel RAW
46862 DPX : X @tab X
46864 @tab Digital Picture Exchange
46865 EXR : @tab X
46867 @tab OpenEXR
46868 FITS : X @tab X
46870 @tab Flexible Image Transport System
46871 HDR : X @tab X
46873 @tab Radiance HDR RGBE Image format
46874 IMG : @tab X
46876 @tab GEM Raster image
46877 JPEG : X @tab X
46879 @tab Progressive JPEG is not supported.
46880 JPEG 2000 : X @tab X
46882 JPEG-LS : X @tab X
46883 LJPEG : X @tab
46884 @tab Lossless JPEG
46885 Media 100 : @tab X
46887 MSP : @tab X
46888 @tab Microsoft Paint image
46889 PAM : X @tab X
46891 @tab PAM is a PNM extension with alpha support.
46892 PBM : X @tab X
46894 @tab Portable BitMap image
46895 PCD : @tab X
46897 @tab PhotoCD
46898 PCX : X @tab X
46900 @tab PC Paintbrush
46901 PFM : X @tab X
46903 @tab Portable FloatMap image
46904 PGM : X @tab X
46906 @tab Portable GrayMap image
46907 PGMYUV : X @tab X
46909 @tab PGM with U and V components in YUV 4:2:0
46910 PGX : @tab X
46912 @tab PGX file decoder
46913 PHM : X @tab X
46915 @tab Portable HalfFloatMap image
46916 PIC : @tab X
46918 @tab Pictor/PC Paint
46919 PNG : X @tab X
46921 @tab Portable Network Graphics image
46922 PPM : X @tab X
46924 @tab Portable PixelMap image
46925 PSD : @tab X
46927 @tab Photoshop
46928 PTX : @tab X
46930 @tab V.Flash PTX format
46931 QOI : X @tab X
46933 @tab Quite OK Image format
46934 SGI : X @tab X
46936 @tab SGI RGB image format
46937 Sun Rasterfile : X @tab X
46939 @tab Sun RAS image format
46940 TIFF : X @tab X
46942 @tab YUV, JPEG and some extension is not supported yet.
46943 Truevision Targa : X @tab X
46945 @tab Targa (.TGA) image format
46946 VBN : X @tab X
46948 @tab Vizrt Binary Image format
46949 WBMP : X @tab X
46951 @tab Wireless Application Protocol Bitmap image format
46952 WebP : E @tab X
46954 @tab WebP image format, encoding supported through external library libwebp
46955 XBM : X @tab X
46957 @tab X BitMap image format
46958 XFace : X @tab X
46960 @tab X-Face image format
46961 XPM : @tab X
46963 @tab X PixMap image format
46964 XWD : X @tab X
46966 @tab X Window Dump image format
46967 "X" means that the feature in that column (encoding / decoding) is
46969 supported.
46970 "E" means that support is provided through an external library.
46972 Video Codecs
46974 Name : Encoding @tab Decoding @tab Comments
46975 4X Movie : @tab X
46976 @tab Used in certain computer games.
46977 8088flex TMV : @tab X
46979 A64 multicolor : X @tab
46980 @tab Creates video suitable to be played on a commodore 64 (multicolor mode).
46981 Amazing Studio PAF Video : @tab X
46983 American Laser Games MM : @tab X
46984 @tab Used in games like Mad Dog McCree.
46985 Amuse Graphics Movie : @tab X
46987 AMV Video : X @tab X
46988 @tab Used in Chinese MP3 players.
46989 ANSI/ASCII art : @tab X
46991 Apple Intermediate Codec : @tab X
46992 Apple MJPEG-B : @tab X
46993 Apple Pixlet : @tab X
46994 Apple ProRes : X @tab X
46995 @tab fourcc: apch,apcn,apcs,apco,ap4h,ap4x
46996 Apple QuickDraw : @tab X
46998 @tab fourcc: qdrw
46999 Argonaut Video : @tab X
47001 @tab Used in some Argonaut games.
47002 Asus v1 : X @tab X
47004 @tab fourcc: ASV1
47005 Asus v2 : X @tab X
47007 @tab fourcc: ASV2
47008 ATI VCR1 : @tab X
47010 @tab fourcc: VCR1
47011 ATI VCR2 : @tab X
47013 @tab fourcc: VCR2
47014 Auravision Aura : @tab X
47016 Auravision Aura 2 : @tab X
47017 Autodesk Animator Flic video : @tab X
47018 Autodesk RLE : @tab X
47019 @tab fourcc: AASC
47020 AV1 : E @tab E
47022 @tab Supported through external libraries libaom, libdav1d, librav1e and libsvtav1
47023 Avid 1:1 10-bit RGB Packer : X @tab X
47025 @tab fourcc: AVrp
47026 AVS (Audio Video Standard) video : @tab X
47028 @tab Video encoding used by the Creature Shock game.
47029 AVS2-P2/IEEE1857.4 : E @tab E
47031 @tab Supported through external libraries libxavs2 and libdavs2
47032 AVS3-P2/IEEE1857.10 : @tab E
47034 @tab Supported through external library libuavs3d
47035 AYUV : X @tab X
47037 @tab Microsoft uncompressed packed 4:4:4:4
47038 Beam Software VB : @tab X
47040 Bethesda VID video : @tab X
47041 @tab Used in some games from Bethesda Softworks.
47042 Bink Video : @tab X
47044 BitJazz SheerVideo : @tab X
47045 Bitmap Brothers JV video : @tab X
47046 y41p Brooktree uncompressed 4:1:1 12-bit : X @tab X
47047 Brooktree ProSumer Video : @tab X
47048 @tab fourcc: BT20
47049 Brute Force & Ignorance : @tab X
47051 @tab Used in the game Flash Traffic: City of Angels.
47052 C93 video : @tab X
47054 @tab Codec used in Cyberia game.
47055 CamStudio : @tab X
47057 @tab fourcc: CSCD
47058 CD+G : @tab X
47060 @tab Video codec for CD+G karaoke disks
47061 CDXL : @tab X
47063 @tab Amiga CD video codec
47064 Chinese AVS video : E @tab X
47066 @tab AVS1-P2, JiZhun profile, encoding through external library libxavs
47067 Delphine Software International CIN video : @tab X
47069 @tab Codec used in Delphine Software International games.
47070 Discworld II BMV Video : @tab X
47072 CineForm HD : X @tab X
47073 Canopus HQ : @tab X
47074 Canopus HQA : @tab X
47075 Canopus HQX : @tab X
47076 Canopus Lossless Codec : @tab X
47077 CDToons : @tab X
47078 @tab Codec used in various Broderbund games.
47079 Cinepak : @tab X
47081 Cirrus Logic AccuPak : X @tab X
47082 @tab fourcc: CLJR
47083 CPiA Video Format : @tab X
47085 Creative YUV (CYUV) : @tab X
47086 DFA : @tab X
47087 @tab Codec used in Chronomaster game.
47088 Dirac : E @tab X
47090 @tab supported though the native vc2 (Dirac Pro) encoder
47091 Deluxe Paint Animation : @tab X
47093 DNxHD : X @tab X
47094 @tab aka SMPTE VC3
47095 Duck TrueMotion 1.0 : @tab X
47097 @tab fourcc: DUCK
47098 Duck TrueMotion 2.0 : @tab X
47100 @tab fourcc: TM20
47101 Duck TrueMotion 2.0 RT : @tab X
47103 @tab fourcc: TR20
47104 DV (Digital Video) : X @tab X
47106 Dxtory capture format : @tab X
47107 Feeble Files/ScummVM DXA : @tab X
47108 @tab Codec originally used in Feeble Files game.
47109 Electronic Arts CMV video : @tab X
47111 @tab Used in NHL 95 game.
47112 Electronic Arts Madcow video : @tab X
47114 Electronic Arts TGV video : @tab X
47115 Electronic Arts TGQ video : @tab X
47116 Electronic Arts TQI video : @tab X
47117 Escape 124 : @tab X
47118 Escape 130 : @tab X
47119 FFmpeg video codec #1 : X @tab X
47120 @tab lossless codec (fourcc: FFV1)
47121 Flash Screen Video v1 : X @tab X
47123 @tab fourcc: FSV1
47124 Flash Screen Video v2 : X @tab X
47126 Flash Video (FLV) : X @tab X
47127 @tab Sorenson H.263 used in Flash
47128 FM Screen Capture Codec : @tab X
47130 Forward Uncompressed : @tab X
47131 Fraps : @tab X
47132 Go2Meeting : @tab X
47133 @tab fourcc: G2M2, G2M3
47134 Go2Webinar : @tab X
47136 @tab fourcc: G2M4
47137 Gremlin Digital Video : @tab X
47139 H.261 : X @tab X
47140 H.263 / H.263-1996 : X @tab X
47141 H.263+ / H.263-1998 / H.263 version 2 : X @tab X
47142 H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10 : E @tab X
47143 @tab encoding supported through external library libx264 and OpenH264
47144 HEVC : X @tab X
47146 @tab encoding supported through external library libx265 and libkvazaar
47147 HNM version 4 : @tab X
47149 HuffYUV : X @tab X
47150 HuffYUV FFmpeg variant : X @tab X
47151 IBM Ultimotion : @tab X
47152 @tab fourcc: ULTI
47153 id Cinematic video : @tab X
47155 @tab Used in Quake II.
47156 id RoQ video : X @tab X
47158 @tab Used in Quake III, Jedi Knight 2, other computer games.
47159 IFF ILBM : @tab X
47161 @tab IFF interleaved bitmap
47162 IFF ByteRun1 : @tab X
47164 @tab IFF run length encoded bitmap
47165 Infinity IMM4 : @tab X
47167 Intel H.263 : @tab X
47168 Intel Indeo 2 : @tab X
47169 Intel Indeo 3 : @tab X
47170 Intel Indeo 4 : @tab X
47171 Intel Indeo 5 : @tab X
47172 Interplay C93 : @tab X
47173 @tab Used in the game Cyberia from Interplay.
47174 Interplay MVE video : @tab X
47176 @tab Used in Interplay .MVE files.
47177 J2K : X @tab X
47179 Karl Morton's video codec : @tab X
47180 @tab Codec used in Worms games.
47181 Kega Game Video (KGV1) : @tab X
47183 @tab Kega emulator screen capture codec.
47184 Lagarith : @tab X
47186 LCL (LossLess Codec Library) MSZH : @tab X
47187 LCL (LossLess Codec Library) ZLIB : E @tab E
47188 LOCO : @tab X
47189 LucasArts SANM/Smush : @tab X
47190 @tab Used in LucasArts games / SMUSH animations.
47191 lossless MJPEG : X @tab X
47193 MagicYUV Video : X @tab X
47194 Mandsoft Screen Capture Codec : @tab X
47195 Microsoft ATC Screen : @tab X
47196 @tab Also known as Microsoft Screen 3.
47197 Microsoft Expression Encoder Screen : @tab X
47199 @tab Also known as Microsoft Titanium Screen 2.
47200 Microsoft RLE : @tab X
47202 Microsoft Screen 1 : @tab X
47203 @tab Also known as Windows Media Video V7 Screen.
47204 Microsoft Screen 2 : @tab X
47206 @tab Also known as Windows Media Video V9 Screen.
47207 Microsoft Video 1 : @tab X
47209 Mimic : @tab X
47210 @tab Used in MSN Messenger Webcam streams.
47211 Miro VideoXL : @tab X
47213 @tab fourcc: VIXL
47214 MJPEG (Motion JPEG) : X @tab X
47216 Mobotix MxPEG video : @tab X
47217 Motion Pixels video : @tab X
47218 MPEG-1 video : X @tab X
47219 MPEG-2 video : X @tab X
47220 MPEG-4 part 2 : X @tab X
47221 @tab libxvidcore can be used alternatively for encoding.
47222 MPEG-4 part 2 Microsoft variant version 1 : @tab X
47224 MPEG-4 part 2 Microsoft variant version 2 : X @tab X
47225 MPEG-4 part 2 Microsoft variant version 3 : X @tab X
47226 Newtek SpeedHQ : X @tab X
47227 Nintendo Gamecube THP video : @tab X
47228 NotchLC : @tab X
47229 NuppelVideo/RTjpeg : @tab X
47230 @tab Video encoding used in NuppelVideo files.
47231 On2 VP3 : @tab X
47233 @tab still experimental
47234 On2 VP4 : @tab X
47236 @tab fourcc: VP40
47237 On2 VP5 : @tab X
47239 @tab fourcc: VP50
47240 On2 VP6 : @tab X
47242 @tab fourcc: VP60,VP61,VP62
47243 On2 VP7 : @tab X
47245 @tab fourcc: VP70,VP71
47246 VP8 : E @tab X
47248 @tab fourcc: VP80, encoding supported through external library libvpx
47249 VP9 : E @tab X
47251 @tab encoding supported through external library libvpx
47252 Pinnacle TARGA CineWave YUV16 : @tab X
47254 @tab fourcc: Y216
47255 Q-team QPEG : @tab X
47257 @tab fourccs: QPEG, Q1.0, Q1.1
47258 QuickTime 8BPS video : @tab X
47260 QuickTime Animation (RLE) video : X @tab X
47261 @tab fourcc: 'rle '
47262 QuickTime Graphics (SMC) : X @tab X
47264 @tab fourcc: 'smc '
47265 QuickTime video (RPZA) : X @tab X
47267 @tab fourcc: rpza
47268 R10K AJA Kona 10-bit RGB Codec : X @tab X
47270 R210 Quicktime Uncompressed RGB 10-bit : X @tab X
47271 Raw Video : X @tab X
47272 RealVideo 1.0 : X @tab X
47273 RealVideo 2.0 : X @tab X
47274 RealVideo 3.0 : @tab X
47275 @tab still far from ideal
47276 RealVideo 4.0 : @tab X
47278 Renderware TXD (TeXture Dictionary) : @tab X
47279 @tab Texture dictionaries used by the Renderware Engine.
47280 RL2 video : @tab X
47282 @tab used in some games by Entertainment Software Partners
47283 ScreenPressor : @tab X
47285 Screenpresso : @tab X
47286 Screen Recorder Gold Codec : @tab X
47287 Sierra VMD video : @tab X
47288 @tab Used in Sierra VMD files.
47289 Silicon Graphics Motion Video Compressor 1 (MVC1) : @tab X
47291 Silicon Graphics Motion Video Compressor 2 (MVC2) : @tab X
47292 Silicon Graphics RLE 8-bit video : @tab X
47293 Smacker video : @tab X
47294 @tab Video encoding used in Smacker.
47295 SMPTE VC-1 : @tab X
47297 Snow : X @tab X
47298 @tab experimental wavelet codec (fourcc: SNOW)
47299 Sony PlayStation MDEC (Motion DECoder) : @tab X
47301 Sorenson Vector Quantizer 1 : X @tab X
47302 @tab fourcc: SVQ1
47303 Sorenson Vector Quantizer 3 : @tab X
47305 @tab fourcc: SVQ3
47306 Sunplus JPEG (SP5X) : @tab X
47308 @tab fourcc: SP5X
47309 TechSmith Screen Capture Codec : @tab X
47311 @tab fourcc: TSCC
47312 TechSmith Screen Capture Codec 2 : @tab X
47314 @tab fourcc: TSC2
47315 Theora : E @tab X
47317 @tab encoding supported through external library libtheora
47318 Tiertex Limited SEQ video : @tab X
47320 @tab Codec used in DOS CD-ROM FlashBack game.
47321 Ut Video : X @tab X
47323 v210 QuickTime uncompressed 4:2:2 10-bit : X @tab X
47324 v308 QuickTime uncompressed 4:4:4 : X @tab X
47325 v408 QuickTime uncompressed 4:4:4:4 : X @tab X
47326 v410 QuickTime uncompressed 4:4:4 10-bit : X @tab X
47327 VBLE Lossless Codec : @tab X
47328 VMware Screen Codec / VMware Video : @tab X
47329 @tab Codec used in videos captured by VMware.
47330 Westwood Studios VQA (Vector Quantized Animation) video : @tab
47332 X
47333 Windows Media Image : @tab X
47334 Windows Media Video 7 : X @tab X
47335 Windows Media Video 8 : X @tab X
47336 Windows Media Video 9 : @tab X
47337 @tab not completely working
47338 Wing Commander III / Xan : @tab X
47340 @tab Used in Wing Commander III .MVE files.
47341 Wing Commander IV / Xan : @tab X
47343 @tab Used in Wing Commander IV.
47344 Winnov WNV1 : @tab X
47346 WMV7 : X @tab X
47347 YAMAHA SMAF : X @tab X
47348 Psygnosis YOP Video : @tab X
47349 yuv4 : X @tab X
47350 @tab libquicktime uncompressed packed 4:2:0
47351 ZeroCodec Lossless Video : @tab X
47353 ZLIB : X @tab X
47354 @tab part of LCL, encoder experimental
47355 Zip Motion Blocks Video : X @tab X
47357 @tab Encoder works only in PAL8.
47358 "X" means that the feature in that column (encoding / decoding) is
47360 supported.
47361 "E" means that support is provided through an external library.
47363 Audio Codecs
47365 Name : Encoding @tab Decoding @tab Comments
47366 8SVX exponential : @tab X
47367 8SVX fibonacci : @tab X
47368 AAC : EX @tab X
47369 @tab encoding supported through internal encoder and external library libfdk-aac
47370 AAC+ : E @tab IX
47372 @tab encoding supported through external library libfdk-aac
47373 AC-3 : IX @tab IX
47375 ACELP.KELVIN : @tab X
47376 ADPCM 4X Movie : @tab X
47377 ADPCM Yamaha AICA : @tab X
47378 ADPCM AmuseGraphics Movie : @tab X
47379 ADPCM Argonaut Games : X @tab X
47380 ADPCM CDROM XA : @tab X
47381 ADPCM Creative Technology : @tab X
47382 @tab 16 -E<gt> 4, 8 -E<gt> 4, 8 -E<gt> 3, 8 -E<gt> 2
47383 ADPCM Electronic Arts : @tab X
47385 @tab Used in various EA titles.
47386 ADPCM Electronic Arts Maxis CDROM XS : @tab X
47388 @tab Used in Sim City 3000.
47389 ADPCM Electronic Arts R1 : @tab X
47391 ADPCM Electronic Arts R2 : @tab X
47392 ADPCM Electronic Arts R3 : @tab X
47393 ADPCM Electronic Arts XAS : @tab X
47394 ADPCM G.722 : X @tab X
47395 ADPCM G.726 : X @tab X
47396 ADPCM IMA Acorn Replay : @tab X
47397 ADPCM IMA AMV : X @tab X
47398 @tab Used in AMV files
47399 ADPCM IMA Cunning Developments : @tab X
47401 ADPCM IMA Electronic Arts EACS : @tab X
47402 ADPCM IMA Electronic Arts SEAD : @tab X
47403 ADPCM IMA Funcom : @tab X
47404 ADPCM IMA High Voltage Software ALP : X @tab X
47405 ADPCM IMA Mobiclip MOFLEX : @tab X
47406 ADPCM IMA QuickTime : X @tab X
47407 ADPCM IMA Simon & Schuster Interactive : X @tab X
47408 ADPCM IMA Ubisoft APM : X @tab X
47409 ADPCM IMA Loki SDL MJPEG : @tab X
47410 ADPCM IMA WAV : X @tab X
47411 ADPCM IMA Westwood : @tab X
47412 ADPCM ISS IMA : @tab X
47413 @tab Used in FunCom games.
47414 ADPCM IMA Dialogic : @tab X
47416 ADPCM IMA Duck DK3 : @tab X
47417 @tab Used in some Sega Saturn console games.
47418 ADPCM IMA Duck DK4 : @tab X
47420 @tab Used in some Sega Saturn console games.
47421 ADPCM IMA Radical : @tab X
47423 ADPCM Microsoft : X @tab X
47424 ADPCM MS IMA : X @tab X
47425 ADPCM Nintendo Gamecube AFC : @tab X
47426 ADPCM Nintendo Gamecube DTK : @tab X
47427 ADPCM Nintendo THP : @tab X
47428 ADPCM Playstation : @tab X
47429 ADPCM QT IMA : X @tab X
47430 ADPCM SEGA CRI ADX : X @tab X
47431 @tab Used in Sega Dreamcast games.
47432 ADPCM Shockwave Flash : X @tab X
47434 ADPCM Sound Blaster Pro 2-bit : @tab X
47435 ADPCM Sound Blaster Pro 2.6-bit : @tab X
47436 ADPCM Sound Blaster Pro 4-bit : @tab X
47437 ADPCM VIMA : @tab X
47438 @tab Used in LucasArts SMUSH animations.
47439 ADPCM Konami XMD : @tab X
47441 ADPCM Westwood Studios IMA : X @tab X
47442 @tab Used in Westwood Studios games like Command and Conquer.
47443 ADPCM Yamaha : X @tab X
47445 ADPCM Zork : @tab X
47446 AMR-NB : E @tab X
47447 @tab encoding supported through external library libopencore-amrnb
47448 AMR-WB : E @tab X
47450 @tab encoding supported through external library libvo-amrwbenc
47451 Amazing Studio PAF Audio : @tab X
47453 Apple lossless audio : X @tab X
47454 @tab QuickTime fourcc 'alac'
47455 aptX : X @tab X
47457 @tab Used in Bluetooth A2DP
47458 aptX HD : X @tab X
47460 @tab Used in Bluetooth A2DP
47461 ATRAC1 : @tab X
47463 ATRAC3 : @tab X
47464 ATRAC3+ : @tab X
47465 ATRAC9 : @tab X
47466 Bink Audio : @tab X
47467 @tab Used in Bink and Smacker files in many games.
47468 Bonk audio : @tab X
47470 CELT : @tab E
47471 @tab decoding supported through external library libcelt
47472 codec2 : E @tab E
47474 @tab en/decoding supported through external library libcodec2
47475 CRI HCA : @tab X
47477 Delphine Software International CIN audio : @tab X
47478 @tab Codec used in Delphine Software International games.
47479 DFPWM : X @tab X
47481 Digital Speech Standard - Standard Play mode (DSS SP) : @tab X
47482 Discworld II BMV Audio : @tab X
47483 COOK : @tab X
47484 @tab All versions except 5.1 are supported.
47485 DCA (DTS Coherent Acoustics) : X @tab X
47487 @tab supported extensions: XCh, XXCH, X96, XBR, XLL, LBR (partially)
47488 Dolby E : @tab X
47490 DPCM Cuberoot-Delta-Exact : @tab X
47491 @tab Used in few games.
47492 DPCM Gremlin : @tab X
47494 DPCM id RoQ : X @tab X
47495 @tab Used in Quake III, Jedi Knight 2 and other computer games.
47496 DPCM Marble WADY : @tab X
47498 DPCM Interplay : @tab X
47499 @tab Used in various Interplay computer games.
47500 DPCM Squareroot-Delta-Exact : @tab X
47502 @tab Used in various games.
47503 DPCM Sierra Online : @tab X
47505 @tab Used in Sierra Online game audio files.
47506 DPCM Sol : @tab X
47508 DPCM Xan : @tab X
47509 @tab Used in Origin's Wing Commander IV AVI files.
47510 DPCM Xilam DERF : @tab X
47512 DSD (Direct Stream Digital), least significant bit first : @tab X
47513 DSD (Direct Stream Digital), most significant bit first : @tab X
47514 DSD (Direct Stream Digital), least significant bit first, planar :
47515 @tab X
47516 DSD (Direct Stream Digital), most significant bit first, planar :
47517 @tab X
47518 DSP Group TrueSpeech : @tab X
47519 DST (Direct Stream Transfer) : @tab X
47520 DV audio : @tab X
47521 Enhanced AC-3 : X @tab X
47522 EVRC (Enhanced Variable Rate Codec) : @tab X
47523 FLAC (Free Lossless Audio Codec) : X @tab IX
47524 FTR Voice : @tab X
47525 G.723.1 : X @tab X
47526 G.729 : @tab X
47527 GSM : E @tab X
47528 @tab encoding supported through external library libgsm
47529 GSM Microsoft variant : E @tab X
47531 @tab encoding supported through external library libgsm
47532 IAC (Indeo Audio Coder) : @tab X
47534 iLBC (Internet Low Bitrate Codec) : E @tab EX
47535 @tab encoding and decoding supported through external library libilbc
47536 IMC (Intel Music Coder) : @tab X
47538 Interplay ACM : @tab X
47539 MACE (Macintosh Audio Compression/Expansion) 3:1 : @tab X
47540 MACE (Macintosh Audio Compression/Expansion) 6:1 : @tab X
47541 Marian's A-pac audio : @tab X
47542 MI-SC4 (Micronas SC-4 Audio) : @tab X
47543 MLP (Meridian Lossless Packing) : X @tab X
47544 @tab Used in DVD-Audio discs.
47545 Monkey's Audio : @tab X
47547 MP1 (MPEG audio layer 1) : @tab IX
47548 MP2 (MPEG audio layer 2) : IX @tab IX
47549 @tab encoding supported also through external library TwoLAME
47550 MP3 (MPEG audio layer 3) : E @tab IX
47552 @tab encoding supported through external library LAME, ADU MP3 and MP3onMP4 also supported
47553 MPEG-4 Audio Lossless Coding (ALS) : @tab X
47555 MobiClip FastAudio : @tab X
47556 Musepack SV7 : @tab X
47557 Musepack SV8 : @tab X
47558 Nellymoser Asao : X @tab X
47559 On2 AVC (Audio for Video Codec) : @tab X
47560 Opus : E @tab X
47561 @tab encoding supported through external library libopus
47562 PCM A-law : X @tab X
47564 PCM mu-law : X @tab X
47565 PCM Archimedes VIDC : X @tab X
47566 PCM signed 8-bit planar : X @tab X
47567 PCM signed 16-bit big-endian planar : X @tab X
47568 PCM signed 16-bit little-endian planar : X @tab X
47569 PCM signed 24-bit little-endian planar : X @tab X
47570 PCM signed 32-bit little-endian planar : X @tab X
47571 PCM 32-bit floating point big-endian : X @tab X
47572 PCM 32-bit floating point little-endian : X @tab X
47573 PCM 64-bit floating point big-endian : X @tab X
47574 PCM 64-bit floating point little-endian : X @tab X
47575 PCM D-Cinema audio signed 24-bit : X @tab X
47576 PCM signed 8-bit : X @tab X
47577 PCM signed 16-bit big-endian : X @tab X
47578 PCM signed 16-bit little-endian : X @tab X
47579 PCM signed 24-bit big-endian : X @tab X
47580 PCM signed 24-bit little-endian : X @tab X
47581 PCM signed 32-bit big-endian : X @tab X
47582 PCM signed 32-bit little-endian : X @tab X
47583 PCM signed 16/20/24-bit big-endian in MPEG-TS : @tab X
47584 PCM unsigned 8-bit : X @tab X
47585 PCM unsigned 16-bit big-endian : X @tab X
47586 PCM unsigned 16-bit little-endian : X @tab X
47587 PCM unsigned 24-bit big-endian : X @tab X
47588 PCM unsigned 24-bit little-endian : X @tab X
47589 PCM unsigned 32-bit big-endian : X @tab X
47590 PCM unsigned 32-bit little-endian : X @tab X
47591 PCM SGA : @tab X
47592 QCELP / PureVoice : @tab X
47593 QDesign Music Codec 1 : @tab X
47594 QDesign Music Codec 2 : @tab X
47595 @tab There are still some distortions.
47596 RealAudio 1.0 (14.4K) : X @tab X
47598 @tab Real 14400 bit/s codec
47599 RealAudio 2.0 (28.8K) : @tab X
47601 @tab Real 28800 bit/s codec
47602 RealAudio 3.0 (dnet) : IX @tab X
47604 @tab Real low bitrate AC-3 codec
47605 RealAudio Lossless : @tab X
47607 RealAudio SIPR / ACELP.NET : @tab X
47608 RK Audio (RKA) : @tab X
47609 SBC (low-complexity subband codec) : X @tab X
47610 @tab Used in Bluetooth A2DP
47611 Shorten : @tab X
47613 Sierra VMD audio : @tab X
47614 @tab Used in Sierra VMD files.
47615 Smacker audio : @tab X
47617 SMPTE 302M AES3 audio : X @tab X
47618 Sonic : X @tab X
47619 @tab experimental codec
47620 Sonic lossless : X @tab X
47622 @tab experimental codec
47623 Speex : E @tab EX
47625 @tab supported through external library libspeex
47626 TAK (Tom's lossless Audio Kompressor) : @tab X
47628 True Audio (TTA) : X @tab X
47629 TrueHD : X @tab X
47630 @tab Used in HD-DVD and Blu-Ray discs.
47631 TwinVQ (VQF flavor) : @tab X
47633 VIMA : @tab X
47634 @tab Used in LucasArts SMUSH animations.
47635 ViewQuest VQC : @tab X
47637 Vorbis : E @tab X
47638 @tab A native but very primitive encoder exists.
47639 Voxware MetaSound : @tab X
47641 Waveform Archiver : @tab X
47642 WavPack : X @tab X
47643 Westwood Audio (SND1) : @tab X
47644 Windows Media Audio 1 : X @tab X
47645 Windows Media Audio 2 : X @tab X
47646 Windows Media Audio Lossless : @tab X
47647 Windows Media Audio Pro : @tab X
47648 Windows Media Audio Voice : @tab X
47649 Xbox Media Audio 1 : @tab X
47650 Xbox Media Audio 2 : @tab X
47651 "X" means that the feature in that column (encoding / decoding) is
47653 supported.
47654 "E" means that support is provided through an external library.
47656 "I" means that an integer-only version is available, too (ensures high
47658 performance on systems without hardware floating point support).
47659 Subtitle Formats
47661 Name : Muxing @tab Demuxing @tab Encoding @tab Decoding
47662 3GPP Timed Text : @tab @tab X @tab X
47663 AQTitle : @tab X @tab @tab X
47664 DVB : X @tab X @tab X @tab X
47665 DVB teletext : @tab X @tab @tab E
47666 DVD : X @tab X @tab X @tab X
47667 JACOsub : X @tab X @tab @tab X
47668 MicroDVD : X @tab X @tab @tab X
47669 MPL2 : @tab X @tab @tab X
47670 MPsub (MPlayer) : @tab X @tab @tab X
47671 PGS : @tab @tab @tab X
47672 PJS (Phoenix) : @tab X @tab @tab X
47673 RealText : @tab X @tab @tab X
47674 SAMI : @tab X @tab @tab X
47675 Spruce format (STL) : @tab X @tab @tab X
47676 SSA/ASS : X @tab X @tab X @tab X
47677 SubRip (SRT) : X @tab X @tab X @tab X
47678 SubViewer v1 : @tab X @tab @tab X
47679 SubViewer : @tab X @tab @tab X
47680 TED Talks captions : @tab X @tab @tab X
47681 TTML : X @tab @tab X @tab
47682 VobSub (IDX+SUB) : @tab X @tab @tab X
47683 VPlayer : @tab X @tab @tab X
47684 WebVTT : X @tab X @tab X @tab X
47685 XSUB : @tab @tab X @tab X
47686 "X" means that the feature is supported.
47688 "E" means that support is provided through an external library.
47690 Network Protocols
47692 Name : Support
47693 AMQP : E
47694 file : X
47695 FTP : X
47696 Gopher : X
47697 Gophers : X
47698 HLS : X
47699 HTTP : X
47700 HTTPS : X
47701 Icecast : X
47702 MMSH : X
47703 MMST : X
47704 pipe : X
47705 Pro-MPEG FEC : X
47706 RTMP : X
47707 RTMPE : X
47708 RTMPS : X
47709 RTMPT : X
47710 RTMPTE : X
47711 RTMPTS : X
47712 RTP : X
47713 SAMBA : E
47714 SCTP : X
47715 SFTP : E
47716 TCP : X
47717 TLS : X
47718 UDP : X
47719 ZMQ : E
47720 "X" means that the protocol is supported.
47722 "E" means that support is provided through an external library.
47724 Input/Output Devices
47726 Name : Input @tab Output
47727 ALSA : X @tab X
47728 BKTR : X @tab
47729 caca : @tab X
47730 DV1394 : X @tab
47731 Lavfi virtual device : X @tab
47732 Linux framebuffer : X @tab X
47733 JACK : X @tab
47734 LIBCDIO : X
47735 LIBDC1394 : X @tab
47736 OpenAL : X
47737 OpenGL : @tab X
47738 OSS : X @tab X
47739 PulseAudio : X @tab X
47740 SDL : @tab X
47741 Video4Linux2 : X @tab X
47742 VfW capture : X @tab
47743 X11 grabbing : X @tab
47744 Win32 grabbing : X @tab
47745 "X" means that input/output is supported.
47747 Timecode
47749 Codec/format : Read @tab Write
47750 AVI : X @tab X
47751 DV : X @tab X
47752 GXF : X @tab X
47753 MOV : X @tab X
47754 MPEG1/2 : X @tab X
47755 MXF : X @tab X
47756
47758 ffmpeg(1), ffplay(1), ffprobe(1), ffmpeg-utils(1), ffmpeg-scaler(1),
47759 ffmpeg-resampler(1), ffmpeg-codecs(1), ffmpeg-bitstream-filters(1),
47760 ffmpeg-formats(1), ffmpeg-devices(1), ffmpeg-protocols(1),
47761 ffmpeg-filters(1)
47762
47764 The FFmpeg developers.
47765 For details about the authorship, see the Git history of the project
47767 (https://git.ffmpeg.org/ffmpeg), e.g. by typing the command git log in
47768 the FFmpeg source directory, or browsing the online repository at
47769 <https://git.ffmpeg.org/ffmpeg>.
47770 Maintainers for the specific components are listed in the file
47772 MAINTAINERS in the source code tree.
47773 FFMPEG-ALL(1)
