1FFPLAY-ALL(1) FFPLAY-ALL(1)
2
6 ffplay - FFplay media player
7
9 ffplay [options] [input_url]
10
12 FFplay is a very simple and portable media player using the FFmpeg
13 libraries and the SDL library. It is mostly used as a testbed for the
14 various FFmpeg APIs.
15
17 All the numerical options, if not specified otherwise, accept a string
18 representing a number as input, which may be followed by one of the SI
19 unit prefixes, for example: 'K', 'M', or 'G'.
20 If 'i' is appended to the SI unit prefix, the complete prefix will be
22 interpreted as a unit prefix for binary multiples, which are based on
23 powers of 1024 instead of powers of 1000. Appending 'B' to the SI unit
24 prefix multiplies the value by 8. This allows using, for example: 'KB',
25 'MiB', 'G' and 'B' as number suffixes.
26 Options which do not take arguments are boolean options, and set the
28 corresponding value to true. They can be set to false by prefixing the
29 option name with "no". For example using "-nofoo" will set the boolean
30 option with name "foo" to false.
31 Stream specifiers
33 Some options are applied per-stream, e.g. bitrate or codec. Stream
34 specifiers are used to precisely specify which stream(s) a given option
35 belongs to.
36 A stream specifier is a string generally appended to the option name
38 and separated from it by a colon. E.g. "-codec:a:1 ac3" contains the
39 "a:1" stream specifier, which matches the second audio stream.
40 Therefore, it would select the ac3 codec for the second audio stream.
41 A stream specifier can match several streams, so that the option is
43 applied to all of them. E.g. the stream specifier in "-b:a 128k"
44 matches all audio streams.
45 An empty stream specifier matches all streams. For example, "-codec
47 copy" or "-codec: copy" would copy all the streams without reencoding.
48 Possible forms of stream specifiers are:
50 stream_index
52 Matches the stream with this index. E.g. "-threads:1 4" would set
53 the thread count for the second stream to 4. If stream_index is
54 used as an additional stream specifier (see below), then it selects
55 stream number stream_index from the matching streams. Stream
56 numbering is based on the order of the streams as detected by
57 libavformat except when a program ID is also specified. In this
58 case it is based on the ordering of the streams in the program.
59 stream_type[:additional_stream_specifier]
61 stream_type is one of following: 'v' or 'V' for video, 'a' for
62 audio, 's' for subtitle, 'd' for data, and 't' for attachments. 'v'
63 matches all video streams, 'V' only matches video streams which are
64 not attached pictures, video thumbnails or cover arts. If
65 additional_stream_specifier is used, then it matches streams which
66 both have this type and match the additional_stream_specifier.
67 Otherwise, it matches all streams of the specified type.
68 p:program_id[:additional_stream_specifier]
70 Matches streams which are in the program with the id program_id. If
71 additional_stream_specifier is used, then it matches streams which
72 both are part of the program and match the
73 additional_stream_specifier.
74 #stream_id or i:stream_id
76 Match the stream by stream id (e.g. PID in MPEG-TS container).
77 m:key[:value]
79 Matches streams with the metadata tag key having the specified
80 value. If value is not given, matches streams that contain the
81 given tag with any value.
82 u Matches streams with usable configuration, the codec must be
84 defined and the essential information such as video dimension or
85 audio sample rate must be present.
86 Note that in ffmpeg, matching by metadata will only work properly
88 for input files.
89 Generic options
91 These options are shared amongst the ff* tools.
92 -L Show license.
94 -h, -?, -help, --help [arg]
96 Show help. An optional parameter may be specified to print help
97 about a specific item. If no argument is specified, only basic (non
98 advanced) tool options are shown.
99 Possible values of arg are:
101 long
103 Print advanced tool options in addition to the basic tool
104 options.
105 full
107 Print complete list of options, including shared and private
108 options for encoders, decoders, demuxers, muxers, filters, etc.
109 decoder=decoder_name
111 Print detailed information about the decoder named
112 decoder_name. Use the -decoders option to get a list of all
113 decoders.
114 encoder=encoder_name
116 Print detailed information about the encoder named
117 encoder_name. Use the -encoders option to get a list of all
118 encoders.
119 demuxer=demuxer_name
121 Print detailed information about the demuxer named
122 demuxer_name. Use the -formats option to get a list of all
123 demuxers and muxers.
124 muxer=muxer_name
126 Print detailed information about the muxer named muxer_name.
127 Use the -formats option to get a list of all muxers and
128 demuxers.
129 filter=filter_name
131 Print detailed information about the filter named filter_name.
132 Use the -filters option to get a list of all filters.
133 bsf=bitstream_filter_name
135 Print detailed information about the bitstream filter named
136 bitstream_filter_name. Use the -bsfs option to get a list of
137 all bitstream filters.
138 protocol=protocol_name
140 Print detailed information about the protocol named
141 protocol_name. Use the -protocols option to get a list of all
142 protocols.
143 -version
145 Show version.
146 -buildconf
148 Show the build configuration, one option per line.
149 -formats
151 Show available formats (including devices).
152 -demuxers
154 Show available demuxers.
155 -muxers
157 Show available muxers.
158 -devices
160 Show available devices.
161 -codecs
163 Show all codecs known to libavcodec.
164 Note that the term 'codec' is used throughout this documentation as
166 a shortcut for what is more correctly called a media bitstream
167 format.
168 -decoders
170 Show available decoders.
171 -encoders
173 Show all available encoders.
174 -bsfs
176 Show available bitstream filters.
177 -protocols
179 Show available protocols.
180 -filters
182 Show available libavfilter filters.
183 -pix_fmts
185 Show available pixel formats.
186 -sample_fmts
188 Show available sample formats.
189 -layouts
191 Show channel names and standard channel layouts.
192 -dispositions
194 Show stream dispositions.
195 -colors
197 Show recognized color names.
198 -sources device[,opt1=val1[,opt2=val2]...]
200 Show autodetected sources of the input device. Some devices may
201 provide system-dependent source names that cannot be autodetected.
202 The returned list cannot be assumed to be always complete.
203 ffmpeg -sources pulse,server=192.168.0.4
205 -sinks device[,opt1=val1[,opt2=val2]...]
207 Show autodetected sinks of the output device. Some devices may
208 provide system-dependent sink names that cannot be autodetected.
209 The returned list cannot be assumed to be always complete.
210 ffmpeg -sinks pulse,server=192.168.0.4
212 -loglevel [flags+]loglevel | -v [flags+]loglevel
214 Set logging level and flags used by the library.
215 The optional flags prefix can consist of the following values:
217 repeat
219 Indicates that repeated log output should not be compressed to
220 the first line and the "Last message repeated n times" line
221 will be omitted.
222 level
224 Indicates that log output should add a "[level]" prefix to each
225 message line. This can be used as an alternative to log
226 coloring, e.g. when dumping the log to file.
227 Flags can also be used alone by adding a '+'/'-' prefix to
229 set/reset a single flag without affecting other flags or changing
230 loglevel. When setting both flags and loglevel, a '+' separator is
231 expected between the last flags value and before loglevel.
232 loglevel is a string or a number containing one of the following
234 values:
235 quiet, -8
237 Show nothing at all; be silent.
238 panic, 0
240 Only show fatal errors which could lead the process to crash,
241 such as an assertion failure. This is not currently used for
242 anything.
243 fatal, 8
245 Only show fatal errors. These are errors after which the
246 process absolutely cannot continue.
247 error, 16
249 Show all errors, including ones which can be recovered from.
250 warning, 24
252 Show all warnings and errors. Any message related to possibly
253 incorrect or unexpected events will be shown.
254 info, 32
256 Show informative messages during processing. This is in
257 addition to warnings and errors. This is the default value.
258 verbose, 40
260 Same as "info", except more verbose.
261 debug, 48
263 Show everything, including debugging information.
264 trace, 56
266 For example to enable repeated log output, add the "level" prefix,
268 and set loglevel to "verbose":
269 ffmpeg -loglevel repeat+level+verbose -i input output
271 Another example that enables repeated log output without affecting
273 current state of "level" prefix flag or loglevel:
274 ffmpeg [...] -loglevel +repeat
276 By default the program logs to stderr. If coloring is supported by
278 the terminal, colors are used to mark errors and warnings. Log
279 coloring can be disabled setting the environment variable
280 AV_LOG_FORCE_NOCOLOR, or can be forced setting the environment
281 variable AV_LOG_FORCE_COLOR.
282 -report
284 Dump full command line and log output to a file named
285 "program-YYYYMMDD-HHMMSS.log" in the current directory. This file
286 can be useful for bug reports. It also implies "-loglevel debug".
287 Setting the environment variable FFREPORT to any value has the same
289 effect. If the value is a ':'-separated key=value sequence, these
290 options will affect the report; option values must be escaped if
291 they contain special characters or the options delimiter ':' (see
292 the ``Quoting and escaping'' section in the ffmpeg-utils manual).
293 The following options are recognized:
295 file
297 set the file name to use for the report; %p is expanded to the
298 name of the program, %t is expanded to a timestamp, "%%" is
299 expanded to a plain "%"
300 level
302 set the log verbosity level using a numerical value (see
303 "-loglevel").
304 For example, to output a report to a file named ffreport.log using
306 a log level of 32 (alias for log level "info"):
307 FFREPORT=file=ffreport.log:level=32 ffmpeg -i input output
309 Errors in parsing the environment variable are not fatal, and will
311 not appear in the report.
312 -hide_banner
314 Suppress printing banner.
315 All FFmpeg tools will normally show a copyright notice, build
317 options and library versions. This option can be used to suppress
318 printing this information.
319 -cpuflags flags (global)
321 Allows setting and clearing cpu flags. This option is intended for
322 testing. Do not use it unless you know what you're doing.
323 ffmpeg -cpuflags -sse+mmx ...
325 ffmpeg -cpuflags mmx ...
326 ffmpeg -cpuflags 0 ...
327 Possible flags for this option are:
329 x86
331 mmx
332 mmxext
333 sse
334 sse2
335 sse2slow
336 sse3
337 sse3slow
338 ssse3
339 atom
340 sse4.1
341 sse4.2
342 avx
343 avx2
344 xop
345 fma3
346 fma4
347 3dnow
348 3dnowext
349 bmi1
350 bmi2
351 cmov
352 ARM
353 armv5te
354 armv6
355 armv6t2
356 vfp
357 vfpv3
358 neon
359 setend
360 AArch64
361 armv8
362 vfp
363 neon
364 PowerPC
365 altivec
366 Specific Processors
367 pentium2
368 pentium3
369 pentium4
370 k6
371 k62
372 athlon
373 athlonxp
374 k8
375 -cpucount count (global)
376 Override detection of CPU count. This option is intended for
377 testing. Do not use it unless you know what you're doing.
378 ffmpeg -cpucount 2
380 -max_alloc bytes
382 Set the maximum size limit for allocating a block on the heap by
383 ffmpeg's family of malloc functions. Exercise extreme caution when
384 using this option. Don't use if you do not understand the full
385 consequence of doing so. Default is INT_MAX.
386 AVOptions
388 These options are provided directly by the libavformat, libavdevice and
389 libavcodec libraries. To see the list of available AVOptions, use the
390 -help option. They are separated into two categories:
391 generic
393 These options can be set for any container, codec or device.
394 Generic options are listed under AVFormatContext options for
395 containers/devices and under AVCodecContext options for codecs.
396 private
398 These options are specific to the given container, device or codec.
399 Private options are listed under their corresponding
400 containers/devices/codecs.
401 For example to write an ID3v2.3 header instead of a default ID3v2.4 to
403 an MP3 file, use the id3v2_version private option of the MP3 muxer:
404 ffmpeg -i input.flac -id3v2_version 3 out.mp3
406 All codec AVOptions are per-stream, and thus a stream specifier should
408 be attached to them:
409 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
411 In the above example, a multichannel audio stream is mapped twice for
413 output. The first instance is encoded with codec ac3 and bitrate 640k.
414 The second instance is downmixed to 2 channels and encoded with codec
415 aac. A bitrate of 128k is specified for it using absolute index of the
416 output stream.
417 Note: the -nooption syntax cannot be used for boolean AVOptions, use
419 -option 0/-option 1.
420 Note: the old undocumented way of specifying per-stream AVOptions by
422 prepending v/a/s to the options name is now obsolete and will be
423 removed soon.
424 Main options
426 -x width
427 Force displayed width.
428 -y height
430 Force displayed height.
431 -fs Start in fullscreen mode.
433 -an Disable audio.
435 -vn Disable video.
437 -sn Disable subtitles.
439 -ss pos
441 Seek to pos. Note that in most formats it is not possible to seek
442 exactly, so ffplay will seek to the nearest seek point to pos.
443 pos must be a time duration specification, see the Time duration
445 section in the ffmpeg-utils(1) manual.
446 -t duration
448 Play duration seconds of audio/video.
449 duration must be a time duration specification, see the Time
451 duration section in the ffmpeg-utils(1) manual.
452 -bytes
454 Seek by bytes.
455 -seek_interval
457 Set custom interval, in seconds, for seeking using left/right keys.
458 Default is 10 seconds.
459 -nodisp
461 Disable graphical display.
462 -noborder
464 Borderless window.
465 -alwaysontop
467 Window always on top. Available on: X11 with SDL >= 2.0.5, Windows
468 SDL >= 2.0.6.
469 -volume
471 Set the startup volume. 0 means silence, 100 means no volume
472 reduction or amplification. Negative values are treated as 0,
473 values above 100 are treated as 100.
474 -f fmt
476 Force format.
477 -window_title title
479 Set window title (default is the input filename).
480 -left title
482 Set the x position for the left of the window (default is a
483 centered window).
484 -top title
486 Set the y position for the top of the window (default is a centered
487 window).
488 -loop number
490 Loops movie playback <number> times. 0 means forever.
491 -showmode mode
493 Set the show mode to use. Available values for mode are:
494 0, video
496 show video
497 1, waves
499 show audio waves
500 2, rdft
502 show audio frequency band using RDFT ((Inverse) Real Discrete
503 Fourier Transform)
504 Default value is "video", if video is not present or cannot be
506 played "rdft" is automatically selected.
507 You can interactively cycle through the available show modes by
509 pressing the key w.
510 -vf filtergraph
512 Create the filtergraph specified by filtergraph and use it to
513 filter the video stream.
514 filtergraph is a description of the filtergraph to apply to the
516 stream, and must have a single video input and a single video
517 output. In the filtergraph, the input is associated to the label
518 "in", and the output to the label "out". See the ffmpeg-filters
519 manual for more information about the filtergraph syntax.
520 You can specify this parameter multiple times and cycle through the
522 specified filtergraphs along with the show modes by pressing the
523 key w.
524 -af filtergraph
526 filtergraph is a description of the filtergraph to apply to the
527 input audio. Use the option "-filters" to show all the available
528 filters (including sources and sinks).
529 -i input_url
531 Read input_url.
532 Advanced options
534 -stats
535 Print several playback statistics, in particular show the stream
536 duration, the codec parameters, the current position in the stream
537 and the audio/video synchronisation drift. It is shown by default,
538 unless the log level is lower than "info". Its display can be
539 forced by manually specifying this option. To disable it, you need
540 to specify "-nostats".
541 -fast
543 Non-spec-compliant optimizations.
544 -genpts
546 Generate pts.
547 -sync type
549 Set the master clock to audio ("type=audio"), video ("type=video")
550 or external ("type=ext"). Default is audio. The master clock is
551 used to control audio-video synchronization. Most media players use
552 audio as master clock, but in some cases (streaming or high quality
553 broadcast) it is necessary to change that. This option is mainly
554 used for debugging purposes.
555 -ast audio_stream_specifier
557 Select the desired audio stream using the given stream specifier.
558 The stream specifiers are described in the Stream specifiers
559 chapter. If this option is not specified, the "best" audio stream
560 is selected in the program of the already selected video stream.
561 -vst video_stream_specifier
563 Select the desired video stream using the given stream specifier.
564 The stream specifiers are described in the Stream specifiers
565 chapter. If this option is not specified, the "best" video stream
566 is selected.
567 -sst subtitle_stream_specifier
569 Select the desired subtitle stream using the given stream
570 specifier. The stream specifiers are described in the Stream
571 specifiers chapter. If this option is not specified, the "best"
572 subtitle stream is selected in the program of the already selected
573 video or audio stream.
574 -autoexit
576 Exit when video is done playing.
577 -exitonkeydown
579 Exit if any key is pressed.
580 -exitonmousedown
582 Exit if any mouse button is pressed.
583 -codec:media_specifier codec_name
585 Force a specific decoder implementation for the stream identified
586 by media_specifier, which can assume the values "a" (audio), "v"
587 (video), and "s" subtitle.
588 -acodec codec_name
590 Force a specific audio decoder.
591 -vcodec codec_name
593 Force a specific video decoder.
594 -scodec codec_name
596 Force a specific subtitle decoder.
597 -autorotate
599 Automatically rotate the video according to file metadata. Enabled
600 by default, use -noautorotate to disable it.
601 -framedrop
603 Drop video frames if video is out of sync. Enabled by default if
604 the master clock is not set to video. Use this option to enable
605 frame dropping for all master clock sources, use -noframedrop to
606 disable it.
607 -infbuf
609 Do not limit the input buffer size, read as much data as possible
610 from the input as soon as possible. Enabled by default for realtime
611 streams, where data may be dropped if not read in time. Use this
612 option to enable infinite buffers for all inputs, use -noinfbuf to
613 disable it.
614 -filter_threads nb_threads
616 Defines how many threads are used to process a filter pipeline.
617 Each pipeline will produce a thread pool with this many threads
618 available for parallel processing. The default is 0 which means
619 that the thread count will be determined by the number of available
620 CPUs.
621 While playing
623 q, ESC
624 Quit.
625 f Toggle full screen.
627 p, SPC
629 Pause.
630 m Toggle mute.
632 9, 0
634 /, *
635 Decrease and increase volume respectively.
636 a Cycle audio channel in the current program.
638 v Cycle video channel.
640 t Cycle subtitle channel in the current program.
642 c Cycle program.
644 w Cycle video filters or show modes.
646 s Step to the next frame.
648 Pause if the stream is not already paused, step to the next video
650 frame, and pause.
651 left/right
653 Seek backward/forward 10 seconds.
654 down/up
656 Seek backward/forward 1 minute.
657 page down/page up
659 Seek to the previous/next chapter. or if there are no chapters
660 Seek backward/forward 10 minutes.
661 right mouse click
663 Seek to percentage in file corresponding to fraction of width.
664 left mouse double-click
666 Toggle full screen.
667
669 This section documents the syntax and formats employed by the FFmpeg
670 libraries and tools.
671 Quoting and escaping
673 FFmpeg adopts the following quoting and escaping mechanism, unless
674 explicitly specified. The following rules are applied:
675 • ' and \ are special characters (respectively used for quoting and
677 escaping). In addition to them, there might be other special
678 characters depending on the specific syntax where the escaping and
679 quoting are employed.
680 • A special character is escaped by prefixing it with a \.
682 • All characters enclosed between '' are included literally in the
684 parsed string. The quote character ' itself cannot be quoted, so
685 you may need to close the quote and escape it.
686 • Leading and trailing whitespaces, unless escaped or quoted, are
688 removed from the parsed string.
689 Note that you may need to add a second level of escaping when using the
691 command line or a script, which depends on the syntax of the adopted
692 shell language.
693 The function "av_get_token" defined in libavutil/avstring.h can be used
695 to parse a token quoted or escaped according to the rules defined
696 above.
697 The tool tools/ffescape in the FFmpeg source tree can be used to
699 automatically quote or escape a string in a script.
700 Examples
702 • Escape the string "Crime d'Amour" containing the "'" special
704 character:
705 Crime d\'Amour
707 • The string above contains a quote, so the "'" needs to be escaped
709 when quoting it:
710 'Crime d'\''Amour'
712 • Include leading or trailing whitespaces using quoting:
714 ' this string starts and ends with whitespaces '
716 • Escaping and quoting can be mixed together:
718 ' The string '\'string\'' is a string '
720 • To include a literal \ you can use either escaping or quoting:
722 'c:\foo' can be written as c:\\foo
724 Date
726 The accepted syntax is:
727 [(YYYY-MM-DD|YYYYMMDD)[T|t| ]]((HH:MM:SS[.m...]]])|(HHMMSS[.m...]]]))[Z]
729 now
730 If the value is "now" it takes the current time.
732 Time is local time unless Z is appended, in which case it is
734 interpreted as UTC. If the year-month-day part is not specified it
735 takes the current year-month-day.
736 Time duration
738 There are two accepted syntaxes for expressing time duration.
739 [-][<HH>:]<MM>:<SS>[.<m>...]
741 HH expresses the number of hours, MM the number of minutes for a
743 maximum of 2 digits, and SS the number of seconds for a maximum of 2
744 digits. The m at the end expresses decimal value for SS.
745 or
747 [-]<S>+[.<m>...][s|ms|us]
749 S expresses the number of seconds, with the optional decimal part m.
751 The optional literal suffixes s, ms or us indicate to interpret the
752 value as seconds, milliseconds or microseconds, respectively.
753 In both expressions, the optional - indicates negative duration.
755 Examples
757 The following examples are all valid time duration:
759 55 55 seconds
761 0.2 0.2 seconds
763 200ms
765 200 milliseconds, that's 0.2s
766 200000us
768 200000 microseconds, that's 0.2s
769 12:03:45
771 12 hours, 03 minutes and 45 seconds
772 23.189
774 23.189 seconds
775 Video size
777 Specify the size of the sourced video, it may be a string of the form
778 widthxheight, or the name of a size abbreviation.
779 The following abbreviations are recognized:
781 ntsc
783 720x480
784 pal 720x576
786 qntsc
788 352x240
789 qpal
791 352x288
792 sntsc
794 640x480
795 spal
797 768x576
798 film
800 352x240
801 ntsc-film
803 352x240
804 sqcif
806 128x96
807 qcif
809 176x144
810 cif 352x288
812 4cif
814 704x576
815 16cif
817 1408x1152
818 qqvga
820 160x120
821 qvga
823 320x240
824 vga 640x480
826 svga
828 800x600
829 xga 1024x768
831 uxga
833 1600x1200
834 qxga
836 2048x1536
837 sxga
839 1280x1024
840 qsxga
842 2560x2048
843 hsxga
845 5120x4096
846 wvga
848 852x480
849 wxga
851 1366x768
852 wsxga
854 1600x1024
855 wuxga
857 1920x1200
858 woxga
860 2560x1600
861 wqsxga
863 3200x2048
864 wquxga
866 3840x2400
867 whsxga
869 6400x4096
870 whuxga
872 7680x4800
873 cga 320x200
875 ega 640x350
877 hd480
879 852x480
880 hd720
882 1280x720
883 hd1080
885 1920x1080
886 2k 2048x1080
888 2kflat
890 1998x1080
891 2kscope
893 2048x858
894 4k 4096x2160
896 4kflat
898 3996x2160
899 4kscope
901 4096x1716
902 nhd 640x360
904 hqvga
906 240x160
907 wqvga
909 400x240
910 fwqvga
912 432x240
913 hvga
915 480x320
916 qhd 960x540
918 2kdci
920 2048x1080
921 4kdci
923 4096x2160
924 uhd2160
926 3840x2160
927 uhd4320
929 7680x4320
930 Video rate
932 Specify the frame rate of a video, expressed as the number of frames
933 generated per second. It has to be a string in the format
934 frame_rate_num/frame_rate_den, an integer number, a float number or a
935 valid video frame rate abbreviation.
936 The following abbreviations are recognized:
938 ntsc
940 30000/1001
941 pal 25/1
943 qntsc
945 30000/1001
946 qpal
948 25/1
949 sntsc
951 30000/1001
952 spal
954 25/1
955 film
957 24/1
958 ntsc-film
960 24000/1001
961 Ratio
963 A ratio can be expressed as an expression, or in the form
964 numerator:denominator.
965 Note that a ratio with infinite (1/0) or negative value is considered
967 valid, so you should check on the returned value if you want to exclude
968 those values.
969 The undefined value can be expressed using the "0:0" string.
971 Color
973 It can be the name of a color as defined below (case insensitive match)
974 or a "[0x|#]RRGGBB[AA]" sequence, possibly followed by @ and a string
975 representing the alpha component.
976 The alpha component may be a string composed by "0x" followed by an
978 hexadecimal number or a decimal number between 0.0 and 1.0, which
979 represents the opacity value (0x00 or 0.0 means completely transparent,
980 0xff or 1.0 completely opaque). If the alpha component is not specified
981 then 0xff is assumed.
982 The string random will result in a random color.
984 The following names of colors are recognized:
986 AliceBlue
988 0xF0F8FF
989 AntiqueWhite
991 0xFAEBD7
992 Aqua
994 0x00FFFF
995 Aquamarine
997 0x7FFFD4
998 Azure
1000 0xF0FFFF
1001 Beige
1003 0xF5F5DC
1004 Bisque
1006 0xFFE4C4
1007 Black
1009 0x000000
1010 BlanchedAlmond
1012 0xFFEBCD
1013 Blue
1015 0x0000FF
1016 BlueViolet
1018 0x8A2BE2
1019 Brown
1021 0xA52A2A
1022 BurlyWood
1024 0xDEB887
1025 CadetBlue
1027 0x5F9EA0
1028 Chartreuse
1030 0x7FFF00
1031 Chocolate
1033 0xD2691E
1034 Coral
1036 0xFF7F50
1037 CornflowerBlue
1039 0x6495ED
1040 Cornsilk
1042 0xFFF8DC
1043 Crimson
1045 0xDC143C
1046 Cyan
1048 0x00FFFF
1049 DarkBlue
1051 0x00008B
1052 DarkCyan
1054 0x008B8B
1055 DarkGoldenRod
1057 0xB8860B
1058 DarkGray
1060 0xA9A9A9
1061 DarkGreen
1063 0x006400
1064 DarkKhaki
1066 0xBDB76B
1067 DarkMagenta
1069 0x8B008B
1070 DarkOliveGreen
1072 0x556B2F
1073 Darkorange
1075 0xFF8C00
1076 DarkOrchid
1078 0x9932CC
1079 DarkRed
1081 0x8B0000
1082 DarkSalmon
1084 0xE9967A
1085 DarkSeaGreen
1087 0x8FBC8F
1088 DarkSlateBlue
1090 0x483D8B
1091 DarkSlateGray
1093 0x2F4F4F
1094 DarkTurquoise
1096 0x00CED1
1097 DarkViolet
1099 0x9400D3
1100 DeepPink
1102 0xFF1493
1103 DeepSkyBlue
1105 0x00BFFF
1106 DimGray
1108 0x696969
1109 DodgerBlue
1111 0x1E90FF
1112 FireBrick
1114 0xB22222
1115 FloralWhite
1117 0xFFFAF0
1118 ForestGreen
1120 0x228B22
1121 Fuchsia
1123 0xFF00FF
1124 Gainsboro
1126 0xDCDCDC
1127 GhostWhite
1129 0xF8F8FF
1130 Gold
1132 0xFFD700
1133 GoldenRod
1135 0xDAA520
1136 Gray
1138 0x808080
1139 Green
1141 0x008000
1142 GreenYellow
1144 0xADFF2F
1145 HoneyDew
1147 0xF0FFF0
1148 HotPink
1150 0xFF69B4
1151 IndianRed
1153 0xCD5C5C
1154 Indigo
1156 0x4B0082
1157 Ivory
1159 0xFFFFF0
1160 Khaki
1162 0xF0E68C
1163 Lavender
1165 0xE6E6FA
1166 LavenderBlush
1168 0xFFF0F5
1169 LawnGreen
1171 0x7CFC00
1172 LemonChiffon
1174 0xFFFACD
1175 LightBlue
1177 0xADD8E6
1178 LightCoral
1180 0xF08080
1181 LightCyan
1183 0xE0FFFF
1184 LightGoldenRodYellow
1186 0xFAFAD2
1187 LightGreen
1189 0x90EE90
1190 LightGrey
1192 0xD3D3D3
1193 LightPink
1195 0xFFB6C1
1196 LightSalmon
1198 0xFFA07A
1199 LightSeaGreen
1201 0x20B2AA
1202 LightSkyBlue
1204 0x87CEFA
1205 LightSlateGray
1207 0x778899
1208 LightSteelBlue
1210 0xB0C4DE
1211 LightYellow
1213 0xFFFFE0
1214 Lime
1216 0x00FF00
1217 LimeGreen
1219 0x32CD32
1220 Linen
1222 0xFAF0E6
1223 Magenta
1225 0xFF00FF
1226 Maroon
1228 0x800000
1229 MediumAquaMarine
1231 0x66CDAA
1232 MediumBlue
1234 0x0000CD
1235 MediumOrchid
1237 0xBA55D3
1238 MediumPurple
1240 0x9370D8
1241 MediumSeaGreen
1243 0x3CB371
1244 MediumSlateBlue
1246 0x7B68EE
1247 MediumSpringGreen
1249 0x00FA9A
1250 MediumTurquoise
1252 0x48D1CC
1253 MediumVioletRed
1255 0xC71585
1256 MidnightBlue
1258 0x191970
1259 MintCream
1261 0xF5FFFA
1262 MistyRose
1264 0xFFE4E1
1265 Moccasin
1267 0xFFE4B5
1268 NavajoWhite
1270 0xFFDEAD
1271 Navy
1273 0x000080
1274 OldLace
1276 0xFDF5E6
1277 Olive
1279 0x808000
1280 OliveDrab
1282 0x6B8E23
1283 Orange
1285 0xFFA500
1286 OrangeRed
1288 0xFF4500
1289 Orchid
1291 0xDA70D6
1292 PaleGoldenRod
1294 0xEEE8AA
1295 PaleGreen
1297 0x98FB98
1298 PaleTurquoise
1300 0xAFEEEE
1301 PaleVioletRed
1303 0xD87093
1304 PapayaWhip
1306 0xFFEFD5
1307 PeachPuff
1309 0xFFDAB9
1310 Peru
1312 0xCD853F
1313 Pink
1315 0xFFC0CB
1316 Plum
1318 0xDDA0DD
1319 PowderBlue
1321 0xB0E0E6
1322 Purple
1324 0x800080
1325 Red 0xFF0000
1327 RosyBrown
1329 0xBC8F8F
1330 RoyalBlue
1332 0x4169E1
1333 SaddleBrown
1335 0x8B4513
1336 Salmon
1338 0xFA8072
1339 SandyBrown
1341 0xF4A460
1342 SeaGreen
1344 0x2E8B57
1345 SeaShell
1347 0xFFF5EE
1348 Sienna
1350 0xA0522D
1351 Silver
1353 0xC0C0C0
1354 SkyBlue
1356 0x87CEEB
1357 SlateBlue
1359 0x6A5ACD
1360 SlateGray
1362 0x708090
1363 Snow
1365 0xFFFAFA
1366 SpringGreen
1368 0x00FF7F
1369 SteelBlue
1371 0x4682B4
1372 Tan 0xD2B48C
1374 Teal
1376 0x008080
1377 Thistle
1379 0xD8BFD8
1380 Tomato
1382 0xFF6347
1383 Turquoise
1385 0x40E0D0
1386 Violet
1388 0xEE82EE
1389 Wheat
1391 0xF5DEB3
1392 White
1394 0xFFFFFF
1395 WhiteSmoke
1397 0xF5F5F5
1398 Yellow
1400 0xFFFF00
1401 YellowGreen
1403 0x9ACD32
1404 Channel Layout
1406 A channel layout specifies the spatial disposition of the channels in a
1407 multi-channel audio stream. To specify a channel layout, FFmpeg makes
1408 use of a special syntax.
1409 Individual channels are identified by an id, as given by the table
1411 below:
1412 FL front left
1414 FR front right
1416 FC front center
1418 LFE low frequency
1420 BL back left
1422 BR back right
1424 FLC front left-of-center
1426 FRC front right-of-center
1428 BC back center
1430 SL side left
1432 SR side right
1434 TC top center
1436 TFL top front left
1438 TFC top front center
1440 TFR top front right
1442 TBL top back left
1444 TBC top back center
1446 TBR top back right
1448 DL downmix left
1450 DR downmix right
1452 WL wide left
1454 WR wide right
1456 SDL surround direct left
1458 SDR surround direct right
1460 LFE2
1462 low frequency 2
1463 Standard channel layout compositions can be specified by using the
1465 following identifiers:
1466 mono
1468 FC
1469 stereo
1471 FL+FR
1472 2.1 FL+FR+LFE
1474 3.0 FL+FR+FC
1476 3.0(back)
1478 FL+FR+BC
1479 4.0 FL+FR+FC+BC
1481 quad
1483 FL+FR+BL+BR
1484 quad(side)
1486 FL+FR+SL+SR
1487 3.1 FL+FR+FC+LFE
1489 5.0 FL+FR+FC+BL+BR
1491 5.0(side)
1493 FL+FR+FC+SL+SR
1494 4.1 FL+FR+FC+LFE+BC
1496 5.1 FL+FR+FC+LFE+BL+BR
1498 5.1(side)
1500 FL+FR+FC+LFE+SL+SR
1501 6.0 FL+FR+FC+BC+SL+SR
1503 6.0(front)
1505 FL+FR+FLC+FRC+SL+SR
1506 hexagonal
1508 FL+FR+FC+BL+BR+BC
1509 6.1 FL+FR+FC+LFE+BC+SL+SR
1511 6.1 FL+FR+FC+LFE+BL+BR+BC
1513 6.1(front)
1515 FL+FR+LFE+FLC+FRC+SL+SR
1516 7.0 FL+FR+FC+BL+BR+SL+SR
1518 7.0(front)
1520 FL+FR+FC+FLC+FRC+SL+SR
1521 7.1 FL+FR+FC+LFE+BL+BR+SL+SR
1523 7.1(wide)
1525 FL+FR+FC+LFE+BL+BR+FLC+FRC
1526 7.1(wide-side)
1528 FL+FR+FC+LFE+FLC+FRC+SL+SR
1529 7.1(top)
1531 FL+FR+FC+LFE+BL+BR+TFL+TFR
1532 octagonal
1534 FL+FR+FC+BL+BR+BC+SL+SR
1535 cube
1537 FL+FR+BL+BR+TFL+TFR+TBL+TBR
1538 hexadecagonal
1540 FL+FR+FC+BL+BR+BC+SL+SR+WL+WR+TBL+TBR+TBC+TFC+TFL+TFR
1541 downmix
1543 DL+DR
1544 22.2
1546 FL+FR+FC+LFE+BL+BR+FLC+FRC+BC+SL+SR+TC+TFL+TFC+TFR+TBL+TBC+TBR+LFE2+TSL+TSR+BFC+BFL+BFR
1547 A custom channel layout can be specified as a sequence of terms,
1549 separated by '+'. Each term can be:
1550 • the name of a single channel (e.g. FL, FR, FC, LFE, etc.), each
1552 optionally containing a custom name after a '@', (e.g. FL@Left,
1553 FR@Right, FC@Center, LFE@Low_Frequency, etc.)
1554 A standard channel layout can be specified by the following:
1556 • the name of a single channel (e.g. FL, FR, FC, LFE, etc.)
1558 • the name of a standard channel layout (e.g. mono, stereo, 4.0,
1560 quad, 5.0, etc.)
1561 • a number of channels, in decimal, followed by 'c', yielding the
1563 default channel layout for that number of channels (see the
1564 function "av_channel_layout_default"). Note that not all channel
1565 counts have a default layout.
1566 • a number of channels, in decimal, followed by 'C', yielding an
1568 unknown channel layout with the specified number of channels. Note
1569 that not all channel layout specification strings support unknown
1570 channel layouts.
1571 • a channel layout mask, in hexadecimal starting with "0x" (see the
1573 "AV_CH_*" macros in libavutil/channel_layout.h.
1574 Before libavutil version 53 the trailing character "c" to specify a
1576 number of channels was optional, but now it is required, while a
1577 channel layout mask can also be specified as a decimal number (if and
1578 only if not followed by "c" or "C").
1579 See also the function "av_channel_layout_from_string" defined in
1581 libavutil/channel_layout.h.
1582
1584 When evaluating an arithmetic expression, FFmpeg uses an internal
1585 formula evaluator, implemented through the libavutil/eval.h interface.
1586 An expression may contain unary, binary operators, constants, and
1588 functions.
1589 Two expressions expr1 and expr2 can be combined to form another
1591 expression "expr1;expr2". expr1 and expr2 are evaluated in turn, and
1592 the new expression evaluates to the value of expr2.
1593 The following binary operators are available: "+", "-", "*", "/", "^".
1595 The following unary operators are available: "+", "-".
1597 The following functions are available:
1599 abs(x)
1601 Compute absolute value of x.
1602 acos(x)
1604 Compute arccosine of x.
1605 asin(x)
1607 Compute arcsine of x.
1608 atan(x)
1610 Compute arctangent of x.
1611 atan2(x, y)
1613 Compute principal value of the arc tangent of y/x.
1614 between(x, min, max)
1616 Return 1 if x is greater than or equal to min and lesser than or
1617 equal to max, 0 otherwise.
1618 bitand(x, y)
1620 bitor(x, y)
1621 Compute bitwise and/or operation on x and y.
1622 The results of the evaluation of x and y are converted to integers
1624 before executing the bitwise operation.
1625 Note that both the conversion to integer and the conversion back to
1627 floating point can lose precision. Beware of unexpected results for
1628 large numbers (usually 2^53 and larger).
1629 ceil(expr)
1631 Round the value of expression expr upwards to the nearest integer.
1632 For example, "ceil(1.5)" is "2.0".
1633 clip(x, min, max)
1635 Return the value of x clipped between min and max.
1636 cos(x)
1638 Compute cosine of x.
1639 cosh(x)
1641 Compute hyperbolic cosine of x.
1642 eq(x, y)
1644 Return 1 if x and y are equivalent, 0 otherwise.
1645 exp(x)
1647 Compute exponential of x (with base "e", the Euler's number).
1648 floor(expr)
1650 Round the value of expression expr downwards to the nearest
1651 integer. For example, "floor(-1.5)" is "-2.0".
1652 gauss(x)
1654 Compute Gauss function of x, corresponding to "exp(-x*x/2) /
1655 sqrt(2*PI)".
1656 gcd(x, y)
1658 Return the greatest common divisor of x and y. If both x and y are
1659 0 or either or both are less than zero then behavior is undefined.
1660 gt(x, y)
1662 Return 1 if x is greater than y, 0 otherwise.
1663 gte(x, y)
1665 Return 1 if x is greater than or equal to y, 0 otherwise.
1666 hypot(x, y)
1668 This function is similar to the C function with the same name; it
1669 returns "sqrt(x*x + y*y)", the length of the hypotenuse of a right
1670 triangle with sides of length x and y, or the distance of the point
1671 (x, y) from the origin.
1672 if(x, y)
1674 Evaluate x, and if the result is non-zero return the result of the
1675 evaluation of y, return 0 otherwise.
1676 if(x, y, z)
1678 Evaluate x, and if the result is non-zero return the evaluation
1679 result of y, otherwise the evaluation result of z.
1680 ifnot(x, y)
1682 Evaluate x, and if the result is zero return the result of the
1683 evaluation of y, return 0 otherwise.
1684 ifnot(x, y, z)
1686 Evaluate x, and if the result is zero return the evaluation result
1687 of y, otherwise the evaluation result of z.
1688 isinf(x)
1690 Return 1.0 if x is +/-INFINITY, 0.0 otherwise.
1691 isnan(x)
1693 Return 1.0 if x is NAN, 0.0 otherwise.
1694 ld(var)
1696 Load the value of the internal variable with number var, which was
1697 previously stored with st(var, expr). The function returns the
1698 loaded value.
1699 lerp(x, y, z)
1701 Return linear interpolation between x and y by amount of z.
1702 log(x)
1704 Compute natural logarithm of x.
1705 lt(x, y)
1707 Return 1 if x is lesser than y, 0 otherwise.
1708 lte(x, y)
1710 Return 1 if x is lesser than or equal to y, 0 otherwise.
1711 max(x, y)
1713 Return the maximum between x and y.
1714 min(x, y)
1716 Return the minimum between x and y.
1717 mod(x, y)
1719 Compute the remainder of division of x by y.
1720 not(expr)
1722 Return 1.0 if expr is zero, 0.0 otherwise.
1723 pow(x, y)
1725 Compute the power of x elevated y, it is equivalent to "(x)^(y)".
1726 print(t)
1728 print(t, l)
1729 Print the value of expression t with loglevel l. If l is not
1730 specified then a default log level is used. Returns the value of
1731 the expression printed.
1732 Prints t with loglevel l
1734 random(x)
1736 Return a pseudo random value between 0.0 and 1.0. x is the index of
1737 the internal variable which will be used to save the seed/state.
1738 root(expr, max)
1740 Find an input value for which the function represented by expr with
1741 argument ld(0) is 0 in the interval 0..max.
1742 The expression in expr must denote a continuous function or the
1744 result is undefined.
1745 ld(0) is used to represent the function input value, which means
1747 that the given expression will be evaluated multiple times with
1748 various input values that the expression can access through ld(0).
1749 When the expression evaluates to 0 then the corresponding input
1750 value will be returned.
1751 round(expr)
1753 Round the value of expression expr to the nearest integer. For
1754 example, "round(1.5)" is "2.0".
1755 sgn(x)
1757 Compute sign of x.
1758 sin(x)
1760 Compute sine of x.
1761 sinh(x)
1763 Compute hyperbolic sine of x.
1764 sqrt(expr)
1766 Compute the square root of expr. This is equivalent to "(expr)^.5".
1767 squish(x)
1769 Compute expression "1/(1 + exp(4*x))".
1770 st(var, expr)
1772 Store the value of the expression expr in an internal variable. var
1773 specifies the number of the variable where to store the value, and
1774 it is a value ranging from 0 to 9. The function returns the value
1775 stored in the internal variable. Note, Variables are currently not
1776 shared between expressions.
1777 tan(x)
1779 Compute tangent of x.
1780 tanh(x)
1782 Compute hyperbolic tangent of x.
1783 taylor(expr, x)
1785 taylor(expr, x, id)
1786 Evaluate a Taylor series at x, given an expression representing the
1787 ld(id)-th derivative of a function at 0.
1788 When the series does not converge the result is undefined.
1790 ld(id) is used to represent the derivative order in expr, which
1792 means that the given expression will be evaluated multiple times
1793 with various input values that the expression can access through
1794 ld(id). If id is not specified then 0 is assumed.
1795 Note, when you have the derivatives at y instead of 0,
1797 "taylor(expr, x-y)" can be used.
1798 time(0)
1800 Return the current (wallclock) time in seconds.
1801 trunc(expr)
1803 Round the value of expression expr towards zero to the nearest
1804 integer. For example, "trunc(-1.5)" is "-1.0".
1805 while(cond, expr)
1807 Evaluate expression expr while the expression cond is non-zero, and
1808 returns the value of the last expr evaluation, or NAN if cond was
1809 always false.
1810 The following constants are available:
1812 PI area of the unit disc, approximately 3.14
1814 E exp(1) (Euler's number), approximately 2.718
1816 PHI golden ratio (1+sqrt(5))/2, approximately 1.618
1818 Assuming that an expression is considered "true" if it has a non-zero
1820 value, note that:
1821 "*" works like AND
1823 "+" works like OR
1825 For example the construct:
1827 if (A AND B) then C
1829 is equivalent to:
1831 if(A*B, C)
1833 In your C code, you can extend the list of unary and binary functions,
1835 and define recognized constants, so that they are available for your
1836 expressions.
1837 The evaluator also recognizes the International System unit prefixes.
1839 If 'i' is appended after the prefix, binary prefixes are used, which
1840 are based on powers of 1024 instead of powers of 1000. The 'B' postfix
1841 multiplies the value by 8, and can be appended after a unit prefix or
1842 used alone. This allows using for example 'KB', 'MiB', 'G' and 'B' as
1843 number postfix.
1844 The list of available International System prefixes follows, with
1846 indication of the corresponding powers of 10 and of 2.
1847 y 10^-24 / 2^-80
1849 z 10^-21 / 2^-70
1851 a 10^-18 / 2^-60
1853 f 10^-15 / 2^-50
1855 p 10^-12 / 2^-40
1857 n 10^-9 / 2^-30
1859 u 10^-6 / 2^-20
1861 m 10^-3 / 2^-10
1863 c 10^-2
1865 d 10^-1
1867 h 10^2
1869 k 10^3 / 2^10
1871 K 10^3 / 2^10
1873 M 10^6 / 2^20
1875 G 10^9 / 2^30
1877 T 10^12 / 2^40
1879 P 10^15 / 2^50
1881 E 10^18 / 2^60
1883 Z 10^21 / 2^70
1885 Y 10^24 / 2^80
1887
1889 libavcodec provides some generic global options, which can be set on
1890 all the encoders and decoders. In addition each codec may support so-
1891 called private options, which are specific for a given codec.
1892 Sometimes, a global option may only affect a specific kind of codec,
1894 and may be nonsensical or ignored by another, so you need to be aware
1895 of the meaning of the specified options. Also some options are meant
1896 only for decoding or encoding.
1897 Options may be set by specifying -option value in the FFmpeg tools, or
1899 by setting the value explicitly in the "AVCodecContext" options or
1900 using the libavutil/opt.h API for programmatic use.
1901 The list of supported options follow:
1903 b integer (encoding,audio,video)
1905 Set bitrate in bits/s. Default value is 200K.
1906 ab integer (encoding,audio)
1908 Set audio bitrate (in bits/s). Default value is 128K.
1909 bt integer (encoding,video)
1911 Set video bitrate tolerance (in bits/s). In 1-pass mode, bitrate
1912 tolerance specifies how far ratecontrol is willing to deviate from
1913 the target average bitrate value. This is not related to min/max
1914 bitrate. Lowering tolerance too much has an adverse effect on
1915 quality.
1916 flags flags (decoding/encoding,audio,video,subtitles)
1918 Set generic flags.
1919 Possible values:
1921 mv4 Use four motion vector by macroblock (mpeg4).
1923 qpel
1925 Use 1/4 pel motion compensation.
1926 loop
1928 Use loop filter.
1929 qscale
1931 Use fixed qscale.
1932 pass1
1934 Use internal 2pass ratecontrol in first pass mode.
1935 pass2
1937 Use internal 2pass ratecontrol in second pass mode.
1938 gray
1940 Only decode/encode grayscale.
1941 psnr
1943 Set error[?] variables during encoding.
1944 truncated
1946 Input bitstream might be randomly truncated.
1947 drop_changed
1949 Don't output frames whose parameters differ from first decoded
1950 frame in stream. Error AVERROR_INPUT_CHANGED is returned when
1951 a frame is dropped.
1952 ildct
1954 Use interlaced DCT.
1955 low_delay
1957 Force low delay.
1958 global_header
1960 Place global headers in extradata instead of every keyframe.
1961 bitexact
1963 Only write platform-, build- and time-independent data. (except
1964 (I)DCT). This ensures that file and data checksums are
1965 reproducible and match between platforms. Its primary use is
1966 for regression testing.
1967 aic Apply H263 advanced intra coding / mpeg4 ac prediction.
1969 ilme
1971 Apply interlaced motion estimation.
1972 cgop
1974 Use closed gop.
1975 output_corrupt
1977 Output even potentially corrupted frames.
1978 time_base rational number
1980 Set codec time base.
1981 It is the fundamental unit of time (in seconds) in terms of which
1983 frame timestamps are represented. For fixed-fps content, timebase
1984 should be "1 / frame_rate" and timestamp increments should be
1985 identically 1.
1986 g integer (encoding,video)
1988 Set the group of picture (GOP) size. Default value is 12.
1989 ar integer (decoding/encoding,audio)
1991 Set audio sampling rate (in Hz).
1992 ac integer (decoding/encoding,audio)
1994 Set number of audio channels.
1995 cutoff integer (encoding,audio)
1997 Set cutoff bandwidth. (Supported only by selected encoders, see
1998 their respective documentation sections.)
1999 frame_size integer (encoding,audio)
2001 Set audio frame size.
2002 Each submitted frame except the last must contain exactly
2004 frame_size samples per channel. May be 0 when the codec has
2005 CODEC_CAP_VARIABLE_FRAME_SIZE set, in that case the frame size is
2006 not restricted. It is set by some decoders to indicate constant
2007 frame size.
2008 frame_number integer
2010 Set the frame number.
2011 delay integer
2013 qcomp float (encoding,video)
2014 Set video quantizer scale compression (VBR). It is used as a
2015 constant in the ratecontrol equation. Recommended range for default
2016 rc_eq: 0.0-1.0.
2017 qblur float (encoding,video)
2019 Set video quantizer scale blur (VBR).
2020 qmin integer (encoding,video)
2022 Set min video quantizer scale (VBR). Must be included between -1
2023 and 69, default value is 2.
2024 qmax integer (encoding,video)
2026 Set max video quantizer scale (VBR). Must be included between -1
2027 and 1024, default value is 31.
2028 qdiff integer (encoding,video)
2030 Set max difference between the quantizer scale (VBR).
2031 bf integer (encoding,video)
2033 Set max number of B frames between non-B-frames.
2034 Must be an integer between -1 and 16. 0 means that B-frames are
2036 disabled. If a value of -1 is used, it will choose an automatic
2037 value depending on the encoder.
2038 Default value is 0.
2040 b_qfactor float (encoding,video)
2042 Set qp factor between P and B frames.
2043 codec_tag integer
2045 bug flags (decoding,video)
2046 Workaround not auto detected encoder bugs.
2047 Possible values:
2049 autodetect
2051 xvid_ilace
2052 Xvid interlacing bug (autodetected if fourcc==XVIX)
2053 ump4
2055 (autodetected if fourcc==UMP4)
2056 no_padding
2058 padding bug (autodetected)
2059 amv
2061 qpel_chroma
2062 std_qpel
2063 old standard qpel (autodetected per fourcc/version)
2064 qpel_chroma2
2066 direct_blocksize
2067 direct-qpel-blocksize bug (autodetected per fourcc/version)
2068 edge
2070 edge padding bug (autodetected per fourcc/version)
2071 hpel_chroma
2073 dc_clip
2074 ms Workaround various bugs in microsoft broken decoders.
2075 trunc
2077 trancated frames
2078 strict integer (decoding/encoding,audio,video)
2080 Specify how strictly to follow the standards.
2081 Possible values:
2083 very
2085 strictly conform to an older more strict version of the spec or
2086 reference software
2087 strict
2089 strictly conform to all the things in the spec no matter what
2090 consequences
2091 normal
2093 unofficial
2094 allow unofficial extensions
2095 experimental
2097 allow non standardized experimental things, experimental
2098 (unfinished/work in progress/not well tested) decoders and
2099 encoders. Note: experimental decoders can pose a security
2100 risk, do not use this for decoding untrusted input.
2101 b_qoffset float (encoding,video)
2103 Set QP offset between P and B frames.
2104 err_detect flags (decoding,audio,video)
2106 Set error detection flags.
2107 Possible values:
2109 crccheck
2111 verify embedded CRCs
2112 bitstream
2114 detect bitstream specification deviations
2115 buffer
2117 detect improper bitstream length
2118 explode
2120 abort decoding on minor error detection
2121 ignore_err
2123 ignore decoding errors, and continue decoding. This is useful
2124 if you want to analyze the content of a video and thus want
2125 everything to be decoded no matter what. This option will not
2126 result in a video that is pleasing to watch in case of errors.
2127 careful
2129 consider things that violate the spec and have not been seen in
2130 the wild as errors
2131 compliant
2133 consider all spec non compliancies as errors
2134 aggressive
2136 consider things that a sane encoder should not do as an error
2137 has_b_frames integer
2139 block_align integer
2140 rc_override_count integer
2141 maxrate integer (encoding,audio,video)
2142 Set max bitrate tolerance (in bits/s). Requires bufsize to be set.
2143 minrate integer (encoding,audio,video)
2145 Set min bitrate tolerance (in bits/s). Most useful in setting up a
2146 CBR encode. It is of little use elsewise.
2147 bufsize integer (encoding,audio,video)
2149 Set ratecontrol buffer size (in bits).
2150 i_qfactor float (encoding,video)
2152 Set QP factor between P and I frames.
2153 i_qoffset float (encoding,video)
2155 Set QP offset between P and I frames.
2156 dct integer (encoding,video)
2158 Set DCT algorithm.
2159 Possible values:
2161 auto
2163 autoselect a good one (default)
2164 fastint
2166 fast integer
2167 int accurate integer
2169 mmx
2171 altivec
2172 faan
2173 floating point AAN DCT
2174 lumi_mask float (encoding,video)
2176 Compress bright areas stronger than medium ones.
2177 tcplx_mask float (encoding,video)
2179 Set temporal complexity masking.
2180 scplx_mask float (encoding,video)
2182 Set spatial complexity masking.
2183 p_mask float (encoding,video)
2185 Set inter masking.
2186 dark_mask float (encoding,video)
2188 Compress dark areas stronger than medium ones.
2189 idct integer (decoding/encoding,video)
2191 Select IDCT implementation.
2192 Possible values:
2194 auto
2196 int
2197 simple
2198 simplemmx
2199 simpleauto
2200 Automatically pick a IDCT compatible with the simple one
2201 arm
2203 altivec
2204 sh4
2205 simplearm
2206 simplearmv5te
2207 simplearmv6
2208 simpleneon
2209 xvid
2210 faani
2211 floating point AAN IDCT
2212 slice_count integer
2214 ec flags (decoding,video)
2215 Set error concealment strategy.
2216 Possible values:
2218 guess_mvs
2220 iterative motion vector (MV) search (slow)
2221 deblock
2223 use strong deblock filter for damaged MBs
2224 favor_inter
2226 favor predicting from the previous frame instead of the current
2227 bits_per_coded_sample integer
2229 aspect rational number (encoding,video)
2230 Set sample aspect ratio.
2231 sar rational number (encoding,video)
2233 Set sample aspect ratio. Alias to aspect.
2234 debug flags (decoding/encoding,audio,video,subtitles)
2236 Print specific debug info.
2237 Possible values:
2239 pict
2241 picture info
2242 rc rate control
2244 bitstream
2246 mb_type
2247 macroblock (MB) type
2248 qp per-block quantization parameter (QP)
2250 dct_coeff
2252 green_metadata
2253 display complexity metadata for the upcoming frame, GoP or for
2254 a given duration.
2255 skip
2257 startcode
2258 er error recognition
2259 mmco
2261 memory management control operations (H.264)
2262 bugs
2264 buffers
2265 picture buffer allocations
2266 thread_ops
2268 threading operations
2269 nomc
2271 skip motion compensation
2272 cmp integer (encoding,video)
2274 Set full pel me compare function.
2275 Possible values:
2277 sad sum of absolute differences, fast (default)
2279 sse sum of squared errors
2281 satd
2283 sum of absolute Hadamard transformed differences
2284 dct sum of absolute DCT transformed differences
2286 psnr
2288 sum of squared quantization errors (avoid, low quality)
2289 bit number of bits needed for the block
2291 rd rate distortion optimal, slow
2293 zero
2295 0
2296 vsad
2298 sum of absolute vertical differences
2299 vsse
2301 sum of squared vertical differences
2302 nsse
2304 noise preserving sum of squared differences
2305 w53 5/3 wavelet, only used in snow
2307 w97 9/7 wavelet, only used in snow
2309 dctmax
2311 chroma
2312 subcmp integer (encoding,video)
2313 Set sub pel me compare function.
2314 Possible values:
2316 sad sum of absolute differences, fast (default)
2318 sse sum of squared errors
2320 satd
2322 sum of absolute Hadamard transformed differences
2323 dct sum of absolute DCT transformed differences
2325 psnr
2327 sum of squared quantization errors (avoid, low quality)
2328 bit number of bits needed for the block
2330 rd rate distortion optimal, slow
2332 zero
2334 0
2335 vsad
2337 sum of absolute vertical differences
2338 vsse
2340 sum of squared vertical differences
2341 nsse
2343 noise preserving sum of squared differences
2344 w53 5/3 wavelet, only used in snow
2346 w97 9/7 wavelet, only used in snow
2348 dctmax
2350 chroma
2351 mbcmp integer (encoding,video)
2352 Set macroblock compare function.
2353 Possible values:
2355 sad sum of absolute differences, fast (default)
2357 sse sum of squared errors
2359 satd
2361 sum of absolute Hadamard transformed differences
2362 dct sum of absolute DCT transformed differences
2364 psnr
2366 sum of squared quantization errors (avoid, low quality)
2367 bit number of bits needed for the block
2369 rd rate distortion optimal, slow
2371 zero
2373 0
2374 vsad
2376 sum of absolute vertical differences
2377 vsse
2379 sum of squared vertical differences
2380 nsse
2382 noise preserving sum of squared differences
2383 w53 5/3 wavelet, only used in snow
2385 w97 9/7 wavelet, only used in snow
2387 dctmax
2389 chroma
2390 ildctcmp integer (encoding,video)
2391 Set interlaced dct compare function.
2392 Possible values:
2394 sad sum of absolute differences, fast (default)
2396 sse sum of squared errors
2398 satd
2400 sum of absolute Hadamard transformed differences
2401 dct sum of absolute DCT transformed differences
2403 psnr
2405 sum of squared quantization errors (avoid, low quality)
2406 bit number of bits needed for the block
2408 rd rate distortion optimal, slow
2410 zero
2412 0
2413 vsad
2415 sum of absolute vertical differences
2416 vsse
2418 sum of squared vertical differences
2419 nsse
2421 noise preserving sum of squared differences
2422 w53 5/3 wavelet, only used in snow
2424 w97 9/7 wavelet, only used in snow
2426 dctmax
2428 chroma
2429 dia_size integer (encoding,video)
2430 Set diamond type & size for motion estimation.
2431 (1024, INT_MAX)
2433 full motion estimation(slowest)
2434 (768, 1024]
2436 umh motion estimation
2437 (512, 768]
2439 hex motion estimation
2440 (256, 512]
2442 l2s diamond motion estimation
2443 [2,256]
2445 var diamond motion estimation
2446 (-1, 2)
2448 small diamond motion estimation
2449 -1 funny diamond motion estimation
2451 (INT_MIN, -1)
2453 sab diamond motion estimation
2454 last_pred integer (encoding,video)
2456 Set amount of motion predictors from the previous frame.
2457 precmp integer (encoding,video)
2459 Set pre motion estimation compare function.
2460 Possible values:
2462 sad sum of absolute differences, fast (default)
2464 sse sum of squared errors
2466 satd
2468 sum of absolute Hadamard transformed differences
2469 dct sum of absolute DCT transformed differences
2471 psnr
2473 sum of squared quantization errors (avoid, low quality)
2474 bit number of bits needed for the block
2476 rd rate distortion optimal, slow
2478 zero
2480 0
2481 vsad
2483 sum of absolute vertical differences
2484 vsse
2486 sum of squared vertical differences
2487 nsse
2489 noise preserving sum of squared differences
2490 w53 5/3 wavelet, only used in snow
2492 w97 9/7 wavelet, only used in snow
2494 dctmax
2496 chroma
2497 pre_dia_size integer (encoding,video)
2498 Set diamond type & size for motion estimation pre-pass.
2499 subq integer (encoding,video)
2501 Set sub pel motion estimation quality.
2502 me_range integer (encoding,video)
2504 Set limit motion vectors range (1023 for DivX player).
2505 global_quality integer (encoding,audio,video)
2507 slice_flags integer
2508 mbd integer (encoding,video)
2509 Set macroblock decision algorithm (high quality mode).
2510 Possible values:
2512 simple
2514 use mbcmp (default)
2515 bits
2517 use fewest bits
2518 rd use best rate distortion
2520 rc_init_occupancy integer (encoding,video)
2522 Set number of bits which should be loaded into the rc buffer before
2523 decoding starts.
2524 flags2 flags (decoding/encoding,audio,video,subtitles)
2526 Possible values:
2527 fast
2529 Allow non spec compliant speedup tricks.
2530 noout
2532 Skip bitstream encoding.
2533 ignorecrop
2535 Ignore cropping information from sps.
2536 local_header
2538 Place global headers at every keyframe instead of in extradata.
2539 chunks
2541 Frame data might be split into multiple chunks.
2542 showall
2544 Show all frames before the first keyframe.
2545 export_mvs
2547 Export motion vectors into frame side-data (see
2548 "AV_FRAME_DATA_MOTION_VECTORS") for codecs that support it. See
2549 also doc/examples/export_mvs.c.
2550 skip_manual
2552 Do not skip samples and export skip information as frame side
2553 data.
2554 ass_ro_flush_noop
2556 Do not reset ASS ReadOrder field on flush.
2557 icc_profiles
2559 Generate/parse embedded ICC profiles from/to colorimetry tags.
2560 export_side_data flags (decoding/encoding,audio,video,subtitles)
2562 Possible values:
2563 mvs Export motion vectors into frame side-data (see
2565 "AV_FRAME_DATA_MOTION_VECTORS") for codecs that support it. See
2566 also doc/examples/export_mvs.c.
2567 prft
2569 Export encoder Producer Reference Time into packet side-data
2570 (see "AV_PKT_DATA_PRFT") for codecs that support it.
2571 venc_params
2573 Export video encoding parameters through frame side data (see
2574 "AV_FRAME_DATA_VIDEO_ENC_PARAMS") for codecs that support it.
2575 At present, those are H.264 and VP9.
2576 film_grain
2578 Export film grain parameters through frame side data (see
2579 "AV_FRAME_DATA_FILM_GRAIN_PARAMS"). Supported at present by
2580 AV1 decoders.
2581 threads integer (decoding/encoding,video)
2583 Set the number of threads to be used, in case the selected codec
2584 implementation supports multi-threading.
2585 Possible values:
2587 auto, 0
2589 automatically select the number of threads to set
2590 Default value is auto.
2592 dc integer (encoding,video)
2594 Set intra_dc_precision.
2595 nssew integer (encoding,video)
2597 Set nsse weight.
2598 skip_top integer (decoding,video)
2600 Set number of macroblock rows at the top which are skipped.
2601 skip_bottom integer (decoding,video)
2603 Set number of macroblock rows at the bottom which are skipped.
2604 profile integer (encoding,audio,video)
2606 Set encoder codec profile. Default value is unknown. Encoder
2607 specific profiles are documented in the relevant encoder
2608 documentation.
2609 level integer (encoding,audio,video)
2611 Possible values:
2612 unknown
2614 lowres integer (decoding,audio,video)
2615 Decode at 1= 1/2, 2=1/4, 3=1/8 resolutions.
2616 mblmin integer (encoding,video)
2618 Set min macroblock lagrange factor (VBR).
2619 mblmax integer (encoding,video)
2621 Set max macroblock lagrange factor (VBR).
2622 skip_loop_filter integer (decoding,video)
2624 skip_idct integer (decoding,video)
2625 skip_frame integer (decoding,video)
2626 Make decoder discard processing depending on the frame type
2627 selected by the option value.
2628 skip_loop_filter skips frame loop filtering, skip_idct skips frame
2630 IDCT/dequantization, skip_frame skips decoding.
2631 Possible values:
2633 none
2635 Discard no frame.
2636 default
2638 Discard useless frames like 0-sized frames.
2639 noref
2641 Discard all non-reference frames.
2642 bidir
2644 Discard all bidirectional frames.
2645 nokey
2647 Discard all frames excepts keyframes.
2648 nointra
2650 Discard all frames except I frames.
2651 all Discard all frames.
2653 Default value is default.
2655 bidir_refine integer (encoding,video)
2657 Refine the two motion vectors used in bidirectional macroblocks.
2658 keyint_min integer (encoding,video)
2660 Set minimum interval between IDR-frames.
2661 refs integer (encoding,video)
2663 Set reference frames to consider for motion compensation.
2664 trellis integer (encoding,audio,video)
2666 Set rate-distortion optimal quantization.
2667 mv0_threshold integer (encoding,video)
2669 compression_level integer (encoding,audio,video)
2670 bits_per_raw_sample integer
2671 channel_layout integer (decoding/encoding,audio)
2672 Possible values:
2673 request_channel_layout integer (decoding,audio)
2675 Possible values:
2676 rc_max_vbv_use float (encoding,video)
2678 rc_min_vbv_use float (encoding,video)
2679 ticks_per_frame integer (decoding/encoding,audio,video)
2680 color_primaries integer (decoding/encoding,video)
2681 Possible values:
2682 bt709
2684 BT.709
2685 bt470m
2687 BT.470 M
2688 bt470bg
2690 BT.470 BG
2691 smpte170m
2693 SMPTE 170 M
2694 smpte240m
2696 SMPTE 240 M
2697 film
2699 Film
2700 bt2020
2702 BT.2020
2703 smpte428
2705 smpte428_1
2706 SMPTE ST 428-1
2707 smpte431
2709 SMPTE 431-2
2710 smpte432
2712 SMPTE 432-1
2713 jedec-p22
2715 JEDEC P22
2716 color_trc integer (decoding/encoding,video)
2718 Possible values:
2719 bt709
2721 BT.709
2722 gamma22
2724 BT.470 M
2725 gamma28
2727 BT.470 BG
2728 smpte170m
2730 SMPTE 170 M
2731 smpte240m
2733 SMPTE 240 M
2734 linear
2736 Linear
2737 log
2739 log100
2740 Log
2741 log_sqrt
2743 log316
2744 Log square root
2745 iec61966_2_4
2747 iec61966-2-4
2748 IEC 61966-2-4
2749 bt1361
2751 bt1361e
2752 BT.1361
2753 iec61966_2_1
2755 iec61966-2-1
2756 IEC 61966-2-1
2757 bt2020_10
2759 bt2020_10bit
2760 BT.2020 - 10 bit
2761 bt2020_12
2763 bt2020_12bit
2764 BT.2020 - 12 bit
2765 smpte2084
2767 SMPTE ST 2084
2768 smpte428
2770 smpte428_1
2771 SMPTE ST 428-1
2772 arib-std-b67
2774 ARIB STD-B67
2775 colorspace integer (decoding/encoding,video)
2777 Possible values:
2778 rgb RGB
2780 bt709
2782 BT.709
2783 fcc FCC
2785 bt470bg
2787 BT.470 BG
2788 smpte170m
2790 SMPTE 170 M
2791 smpte240m
2793 SMPTE 240 M
2794 ycocg
2796 YCOCG
2797 bt2020nc
2799 bt2020_ncl
2800 BT.2020 NCL
2801 bt2020c
2803 bt2020_cl
2804 BT.2020 CL
2805 smpte2085
2807 SMPTE 2085
2808 chroma-derived-nc
2810 Chroma-derived NCL
2811 chroma-derived-c
2813 Chroma-derived CL
2814 ictcp
2816 ICtCp
2817 color_range integer (decoding/encoding,video)
2819 If used as input parameter, it serves as a hint to the decoder,
2820 which color_range the input has. Possible values:
2821 tv
2823 mpeg
2824 MPEG (219*2^(n-8))
2825 pc
2827 jpeg
2828 JPEG (2^n-1)
2829 chroma_sample_location integer (decoding/encoding,video)
2831 Possible values:
2832 left
2834 center
2835 topleft
2836 top
2837 bottomleft
2838 bottom
2839 log_level_offset integer
2840 Set the log level offset.
2841 slices integer (encoding,video)
2843 Number of slices, used in parallelized encoding.
2844 thread_type flags (decoding/encoding,video)
2846 Select which multithreading methods to use.
2847 Use of frame will increase decoding delay by one frame per thread,
2849 so clients which cannot provide future frames should not use it.
2850 Possible values:
2852 slice
2854 Decode more than one part of a single frame at once.
2855 Multithreading using slices works only when the video was
2857 encoded with slices.
2858 frame
2860 Decode more than one frame at once.
2861 Default value is slice+frame.
2863 audio_service_type integer (encoding,audio)
2865 Set audio service type.
2866 Possible values:
2868 ma Main Audio Service
2870 ef Effects
2872 vi Visually Impaired
2874 hi Hearing Impaired
2876 di Dialogue
2878 co Commentary
2880 em Emergency
2882 vo Voice Over
2884 ka Karaoke
2886 request_sample_fmt sample_fmt (decoding,audio)
2888 Set sample format audio decoders should prefer. Default value is
2889 "none".
2890 pkt_timebase rational number
2892 sub_charenc encoding (decoding,subtitles)
2893 Set the input subtitles character encoding.
2894 field_order field_order (video)
2896 Set/override the field order of the video. Possible values:
2897 progressive
2899 Progressive video
2900 tt Interlaced video, top field coded and displayed first
2902 bb Interlaced video, bottom field coded and displayed first
2904 tb Interlaced video, top coded first, bottom displayed first
2906 bt Interlaced video, bottom coded first, top displayed first
2908 skip_alpha bool (decoding,video)
2910 Set to 1 to disable processing alpha (transparency). This works
2911 like the gray flag in the flags option which skips chroma
2912 information instead of alpha. Default is 0.
2913 codec_whitelist list (input)
2915 "," separated list of allowed decoders. By default all are allowed.
2916 dump_separator string (input)
2918 Separator used to separate the fields printed on the command line
2919 about the Stream parameters. For example, to separate the fields
2920 with newlines and indentation:
2921 ffprobe -dump_separator "
2923 " -i ~/videos/matrixbench_mpeg2.mpg
2924 max_pixels integer (decoding/encoding,video)
2926 Maximum number of pixels per image. This value can be used to avoid
2927 out of memory failures due to large images.
2928 apply_cropping bool (decoding,video)
2930 Enable cropping if cropping parameters are multiples of the
2931 required alignment for the left and top parameters. If the
2932 alignment is not met the cropping will be partially applied to
2933 maintain alignment. Default is 1 (enabled). Note: The required
2934 alignment depends on if "AV_CODEC_FLAG_UNALIGNED" is set and the
2935 CPU. "AV_CODEC_FLAG_UNALIGNED" cannot be changed from the command
2936 line. Also hardware decoders will not apply left/top Cropping.
2937
2939 Decoders are configured elements in FFmpeg which allow the decoding of
2940 multimedia streams.
2941 When you configure your FFmpeg build, all the supported native decoders
2943 are enabled by default. Decoders requiring an external library must be
2944 enabled manually via the corresponding "--enable-lib" option. You can
2945 list all available decoders using the configure option
2946 "--list-decoders".
2947 You can disable all the decoders with the configure option
2949 "--disable-decoders" and selectively enable / disable single decoders
2950 with the options "--enable-decoder=DECODER" /
2951 "--disable-decoder=DECODER".
2952 The option "-decoders" of the ff* tools will display the list of
2954 enabled decoders.
2955
2957 A description of some of the currently available video decoders
2958 follows.
2959 av1
2961 AOMedia Video 1 (AV1) decoder.
2962 Options
2964 operating_point
2966 Select an operating point of a scalable AV1 bitstream (0 - 31).
2967 Default is 0.
2968 rawvideo
2970 Raw video decoder.
2971 This decoder decodes rawvideo streams.
2973 Options
2975 top top_field_first
2977 Specify the assumed field type of the input video.
2978 -1 the video is assumed to be progressive (default)
2980 0 bottom-field-first is assumed
2982 1 top-field-first is assumed
2984 libdav1d
2986 dav1d AV1 decoder.
2987 libdav1d allows libavcodec to decode the AOMedia Video 1 (AV1) codec.
2989 Requires the presence of the libdav1d headers and library during
2990 configuration. You need to explicitly configure the build with
2991 "--enable-libdav1d".
2992 Options
2994 The following options are supported by the libdav1d wrapper.
2996 framethreads
2998 Set amount of frame threads to use during decoding. The default
2999 value is 0 (autodetect). This option is deprecated for libdav1d >=
3000 1.0 and will be removed in the future. Use the option
3001 "max_frame_delay" and the global option "threads" instead.
3002 tilethreads
3004 Set amount of tile threads to use during decoding. The default
3005 value is 0 (autodetect). This option is deprecated for libdav1d >=
3006 1.0 and will be removed in the future. Use the global option
3007 "threads" instead.
3008 max_frame_delay
3010 Set max amount of frames the decoder may buffer internally. The
3011 default value is 0 (autodetect).
3012 filmgrain
3014 Apply film grain to the decoded video if present in the bitstream.
3015 Defaults to the internal default of the library. This option is
3016 deprecated and will be removed in the future. See the global option
3017 "export_side_data" to export Film Grain parameters instead of
3018 applying it.
3019 oppoint
3021 Select an operating point of a scalable AV1 bitstream (0 - 31).
3022 Defaults to the internal default of the library.
3023 alllayers
3025 Output all spatial layers of a scalable AV1 bitstream. The default
3026 value is false.
3027 libdavs2
3029 AVS2-P2/IEEE1857.4 video decoder wrapper.
3030 This decoder allows libavcodec to decode AVS2 streams with davs2
3032 library.
3033 libuavs3d
3035 AVS3-P2/IEEE1857.10 video decoder.
3036 libuavs3d allows libavcodec to decode AVS3 streams. Requires the
3038 presence of the libuavs3d headers and library during configuration.
3039 You need to explicitly configure the build with "--enable-libuavs3d".
3040 Options
3042 The following option is supported by the libuavs3d wrapper.
3044 frame_threads
3046 Set amount of frame threads to use during decoding. The default
3047 value is 0 (autodetect).
3048 QSV Decoders
3050 The family of Intel QuickSync Video decoders (VC1, MPEG-2, H.264, HEVC,
3051 JPEG/MJPEG, VP8, VP9, AV1).
3052 Common Options
3054 The following options are supported by all qsv decoders.
3056 async_depth
3058 Internal parallelization depth, the higher the value the higher the
3059 latency.
3060 gpu_copy
3062 A GPU-accelerated copy between video and system memory
3063 default
3065 on
3066 off
3067 HEVC Options
3069 Extra options for hevc_qsv.
3071 load_plugin
3073 A user plugin to load in an internal session
3074 none
3076 hevc_sw
3077 hevc_hw
3078 load_plugins
3079 A :-separate list of hexadecimal plugin UIDs to load in an internal
3080 session
3081 v210
3083 Uncompressed 4:2:2 10-bit decoder.
3084 Options
3086 custom_stride
3088 Set the line size of the v210 data in bytes. The default value is 0
3089 (autodetect). You can use the special -1 value for a strideless
3090 v210 as seen in BOXX files.
3091
3093 A description of some of the currently available audio decoders
3094 follows.
3095 ac3
3097 AC-3 audio decoder.
3098 This decoder implements part of ATSC A/52:2010 and ETSI TS 102 366, as
3100 well as the undocumented RealAudio 3 (a.k.a. dnet).
3101 AC-3 Decoder Options
3103 -drc_scale value
3105 Dynamic Range Scale Factor. The factor to apply to dynamic range
3106 values from the AC-3 stream. This factor is applied exponentially.
3107 The default value is 1. There are 3 notable scale factor ranges:
3108 drc_scale == 0
3110 DRC disabled. Produces full range audio.
3111 0 < drc_scale <= 1
3113 DRC enabled. Applies a fraction of the stream DRC value.
3114 Audio reproduction is between full range and full compression.
3115 drc_scale > 1
3117 DRC enabled. Applies drc_scale asymmetrically. Loud sounds are
3118 fully compressed. Soft sounds are enhanced.
3119 flac
3121 FLAC audio decoder.
3122 This decoder aims to implement the complete FLAC specification from
3124 Xiph.
3125 FLAC Decoder options
3127 -use_buggy_lpc
3129 The lavc FLAC encoder used to produce buggy streams with high lpc
3130 values (like the default value). This option makes it possible to
3131 decode such streams correctly by using lavc's old buggy lpc logic
3132 for decoding.
3133 ffwavesynth
3135 Internal wave synthesizer.
3136 This decoder generates wave patterns according to predefined sequences.
3138 Its use is purely internal and the format of the data it accepts is not
3139 publicly documented.
3140 libcelt
3142 libcelt decoder wrapper.
3143 libcelt allows libavcodec to decode the Xiph CELT ultra-low delay audio
3145 codec. Requires the presence of the libcelt headers and library during
3146 configuration. You need to explicitly configure the build with
3147 "--enable-libcelt".
3148 libgsm
3150 libgsm decoder wrapper.
3151 libgsm allows libavcodec to decode the GSM full rate audio codec.
3153 Requires the presence of the libgsm headers and library during
3154 configuration. You need to explicitly configure the build with
3155 "--enable-libgsm".
3156 This decoder supports both the ordinary GSM and the Microsoft variant.
3158 libilbc
3160 libilbc decoder wrapper.
3161 libilbc allows libavcodec to decode the Internet Low Bitrate Codec
3163 (iLBC) audio codec. Requires the presence of the libilbc headers and
3164 library during configuration. You need to explicitly configure the
3165 build with "--enable-libilbc".
3166 Options
3168 The following option is supported by the libilbc wrapper.
3170 enhance
3172 Enable the enhancement of the decoded audio when set to 1. The
3173 default value is 0 (disabled).
3174 libopencore-amrnb
3176 libopencore-amrnb decoder wrapper.
3177 libopencore-amrnb allows libavcodec to decode the Adaptive Multi-Rate
3179 Narrowband audio codec. Using it requires the presence of the
3180 libopencore-amrnb headers and library during configuration. You need to
3181 explicitly configure the build with "--enable-libopencore-amrnb".
3182 An FFmpeg native decoder for AMR-NB exists, so users can decode AMR-NB
3184 without this library.
3185 libopencore-amrwb
3187 libopencore-amrwb decoder wrapper.
3188 libopencore-amrwb allows libavcodec to decode the Adaptive Multi-Rate
3190 Wideband audio codec. Using it requires the presence of the
3191 libopencore-amrwb headers and library during configuration. You need to
3192 explicitly configure the build with "--enable-libopencore-amrwb".
3193 An FFmpeg native decoder for AMR-WB exists, so users can decode AMR-WB
3195 without this library.
3196 libopus
3198 libopus decoder wrapper.
3199 libopus allows libavcodec to decode the Opus Interactive Audio Codec.
3201 Requires the presence of the libopus headers and library during
3202 configuration. You need to explicitly configure the build with
3203 "--enable-libopus".
3204 An FFmpeg native decoder for Opus exists, so users can decode Opus
3206 without this library.
3207
3209 libaribb24
3210 ARIB STD-B24 caption decoder.
3211 Implements profiles A and C of the ARIB STD-B24 standard.
3213 libaribb24 Decoder Options
3215 -aribb24-base-path path
3217 Sets the base path for the libaribb24 library. This is utilized for
3218 reading of configuration files (for custom unicode conversions),
3219 and for dumping of non-text symbols as images under that location.
3220 Unset by default.
3222 -aribb24-skip-ruby-text boolean
3224 Tells the decoder wrapper to skip text blocks that contain half-
3225 height ruby text.
3226 Enabled by default.
3228 dvbsub
3230 Options
3231 compute_clut
3233 -2 Compute clut once if no matching CLUT is in the stream.
3234 -1 Compute clut if no matching CLUT is in the stream.
3236 0 Never compute CLUT
3238 1 Always compute CLUT and override the one provided in the
3240 stream.
3241 dvb_substream
3243 Selects the dvb substream, or all substreams if -1 which is
3244 default.
3245 dvdsub
3247 This codec decodes the bitmap subtitles used in DVDs; the same
3248 subtitles can also be found in VobSub file pairs and in some Matroska
3249 files.
3250 Options
3252 palette
3254 Specify the global palette used by the bitmaps. When stored in
3255 VobSub, the palette is normally specified in the index file; in
3256 Matroska, the palette is stored in the codec extra-data in the same
3257 format as in VobSub. In DVDs, the palette is stored in the IFO
3258 file, and therefore not available when reading from dumped VOB
3259 files.
3260 The format for this option is a string containing 16 24-bits
3262 hexadecimal numbers (without 0x prefix) separated by commas, for
3263 example "0d00ee, ee450d, 101010, eaeaea, 0ce60b, ec14ed, ebff0b,
3264 0d617a, 7b7b7b, d1d1d1, 7b2a0e, 0d950c, 0f007b, cf0dec, cfa80c,
3265 7c127b".
3266 ifo_palette
3268 Specify the IFO file from which the global palette is obtained.
3269 (experimental)
3270 forced_subs_only
3272 Only decode subtitle entries marked as forced. Some titles have
3273 forced and non-forced subtitles in the same track. Setting this
3274 flag to 1 will only keep the forced subtitles. Default value is 0.
3275 libzvbi-teletext
3277 Libzvbi allows libavcodec to decode DVB teletext pages and DVB teletext
3278 subtitles. Requires the presence of the libzvbi headers and library
3279 during configuration. You need to explicitly configure the build with
3280 "--enable-libzvbi".
3281 Options
3283 txt_page
3285 List of teletext page numbers to decode. Pages that do not match
3286 the specified list are dropped. You may use the special "*" string
3287 to match all pages, or "subtitle" to match all subtitle pages.
3288 Default value is *.
3289 txt_default_region
3291 Set default character set used for decoding, a value between 0 and
3292 87 (see ETS 300 706, Section 15, Table 32). Default value is -1,
3293 which does not override the libzvbi default. This option is needed
3294 for some legacy level 1.0 transmissions which cannot signal the
3295 proper charset.
3296 txt_chop_top
3298 Discards the top teletext line. Default value is 1.
3299 txt_format
3301 Specifies the format of the decoded subtitles.
3302 bitmap
3304 The default format, you should use this for teletext pages,
3305 because certain graphics and colors cannot be expressed in
3306 simple text or even ASS.
3307 text
3309 Simple text based output without formatting.
3310 ass Formatted ASS output, subtitle pages and teletext pages are
3312 returned in different styles, subtitle pages are stripped down
3313 to text, but an effort is made to keep the text alignment and
3314 the formatting.
3315 txt_left
3317 X offset of generated bitmaps, default is 0.
3318 txt_top
3320 Y offset of generated bitmaps, default is 0.
3321 txt_chop_spaces
3323 Chops leading and trailing spaces and removes empty lines from the
3324 generated text. This option is useful for teletext based subtitles
3325 where empty spaces may be present at the start or at the end of the
3326 lines or empty lines may be present between the subtitle lines
3327 because of double-sized teletext characters. Default value is 1.
3328 txt_duration
3330 Sets the display duration of the decoded teletext pages or
3331 subtitles in milliseconds. Default value is -1 which means infinity
3332 or until the next subtitle event comes.
3333 txt_transparent
3335 Force transparent background of the generated teletext bitmaps.
3336 Default value is 0 which means an opaque background.
3337 txt_opacity
3339 Sets the opacity (0-255) of the teletext background. If
3340 txt_transparent is not set, it only affects characters between a
3341 start box and an end box, typically subtitles. Default value is 0
3342 if txt_transparent is set, 255 otherwise.
3343
3345 When you configure your FFmpeg build, all the supported bitstream
3346 filters are enabled by default. You can list all available ones using
3347 the configure option "--list-bsfs".
3348 You can disable all the bitstream filters using the configure option
3350 "--disable-bsfs", and selectively enable any bitstream filter using the
3351 option "--enable-bsf=BSF", or you can disable a particular bitstream
3352 filter using the option "--disable-bsf=BSF".
3353 The option "-bsfs" of the ff* tools will display the list of all the
3355 supported bitstream filters included in your build.
3356 The ff* tools have a -bsf option applied per stream, taking a comma-
3358 separated list of filters, whose parameters follow the filter name
3359 after a '='.
3360 ffmpeg -i INPUT -c:v copy -bsf:v filter1[=opt1=str1:opt2=str2][,filter2] OUTPUT
3362 Below is a description of the currently available bitstream filters,
3364 with their parameters, if any.
3365 aac_adtstoasc
3367 Convert MPEG-2/4 AAC ADTS to an MPEG-4 Audio Specific Configuration
3368 bitstream.
3369 This filter creates an MPEG-4 AudioSpecificConfig from an MPEG-2/4 ADTS
3371 header and removes the ADTS header.
3372 This filter is required for example when copying an AAC stream from a
3374 raw ADTS AAC or an MPEG-TS container to MP4A-LATM, to an FLV file, or
3375 to MOV/MP4 files and related formats such as 3GP or M4A. Please note
3376 that it is auto-inserted for MP4A-LATM and MOV/MP4 and related formats.
3377 av1_metadata
3379 Modify metadata embedded in an AV1 stream.
3380 td Insert or remove temporal delimiter OBUs in all temporal units of
3382 the stream.
3383 insert
3385 Insert a TD at the beginning of every TU which does not already
3386 have one.
3387 remove
3389 Remove the TD from the beginning of every TU which has one.
3390 color_primaries
3392 transfer_characteristics
3393 matrix_coefficients
3394 Set the color description fields in the stream (see AV1 section
3395 6.4.2).
3396 color_range
3398 Set the color range in the stream (see AV1 section 6.4.2; note that
3399 this cannot be set for streams using BT.709 primaries, sRGB
3400 transfer characteristic and identity (RGB) matrix coefficients).
3401 tv Limited range.
3403 pc Full range.
3405 chroma_sample_position
3407 Set the chroma sample location in the stream (see AV1 section
3408 6.4.2). This can only be set for 4:2:0 streams.
3409 vertical
3411 Left position (matching the default in MPEG-2 and H.264).
3412 colocated
3414 Top-left position.
3415 tick_rate
3417 Set the tick rate (time_scale / num_units_in_display_tick) in the
3418 timing info in the sequence header.
3419 num_ticks_per_picture
3421 Set the number of ticks in each picture, to indicate that the
3422 stream has a fixed framerate. Ignored if tick_rate is not also
3423 set.
3424 delete_padding
3426 Deletes Padding OBUs.
3427 chomp
3429 Remove zero padding at the end of a packet.
3430 dca_core
3432 Extract the core from a DCA/DTS stream, dropping extensions such as
3433 DTS-HD.
3434 dump_extra
3436 Add extradata to the beginning of the filtered packets except when said
3437 packets already exactly begin with the extradata that is intended to be
3438 added.
3439 freq
3441 The additional argument specifies which packets should be filtered.
3442 It accepts the values:
3443 k
3445 keyframe
3446 add extradata to all key packets
3447 e
3449 all add extradata to all packets
3450 If not specified it is assumed k.
3452 For example the following ffmpeg command forces a global header (thus
3454 disabling individual packet headers) in the H.264 packets generated by
3455 the "libx264" encoder, but corrects them by adding the header stored in
3456 extradata to the key packets:
3457 ffmpeg -i INPUT -map 0 -flags:v +global_header -c:v libx264 -bsf:v dump_extra out.ts
3459 dv_error_marker
3461 Blocks in DV which are marked as damaged are replaced by blocks of the
3462 specified color.
3463 color
3465 The color to replace damaged blocks by
3466 sta A 16 bit mask which specifies which of the 16 possible error status
3468 values are to be replaced by colored blocks. 0xFFFE is the default
3469 which replaces all non 0 error status values.
3470 ok No error, no concealment
3472 err Error, No concealment
3474 res Reserved
3476 notok
3478 Error or concealment
3479 notres
3481 Not reserved
3482 Aa, Ba, Ca, Ab, Bb, Cb, A, B, C, a, b, erri, erru
3484 The specific error status code
3485 see page 44-46 or section 5.5 of
3487 <http://web.archive.org/web/20060927044735/http://www.smpte.org/smpte_store/standards/pdf/s314m.pdf>
3488 eac3_core
3490 Extract the core from a E-AC-3 stream, dropping extra channels.
3491 extract_extradata
3493 Extract the in-band extradata.
3494 Certain codecs allow the long-term headers (e.g. MPEG-2 sequence
3496 headers, or H.264/HEVC (VPS/)SPS/PPS) to be transmitted either "in-
3497 band" (i.e. as a part of the bitstream containing the coded frames) or
3498 "out of band" (e.g. on the container level). This latter form is called
3499 "extradata" in FFmpeg terminology.
3500 This bitstream filter detects the in-band headers and makes them
3502 available as extradata.
3503 remove
3505 When this option is enabled, the long-term headers are removed from
3506 the bitstream after extraction.
3507 filter_units
3509 Remove units with types in or not in a given set from the stream.
3510 pass_types
3512 List of unit types or ranges of unit types to pass through while
3513 removing all others. This is specified as a '|'-separated list of
3514 unit type values or ranges of values with '-'.
3515 remove_types
3517 Identical to pass_types, except the units in the given set removed
3518 and all others passed through.
3519 Extradata is unchanged by this transformation, but note that if the
3521 stream contains inline parameter sets then the output may be unusable
3522 if they are removed.
3523 For example, to remove all non-VCL NAL units from an H.264 stream:
3525 ffmpeg -i INPUT -c:v copy -bsf:v 'filter_units=pass_types=1-5' OUTPUT
3527 To remove all AUDs, SEI and filler from an H.265 stream:
3529 ffmpeg -i INPUT -c:v copy -bsf:v 'filter_units=remove_types=35|38-40' OUTPUT
3531 hapqa_extract
3533 Extract Rgb or Alpha part of an HAPQA file, without recompression, in
3534 order to create an HAPQ or an HAPAlphaOnly file.
3535 texture
3537 Specifies the texture to keep.
3538 color
3540 alpha
3541 Convert HAPQA to HAPQ
3543 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
3545 Convert HAPQA to HAPAlphaOnly
3547 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
3549 h264_metadata
3551 Modify metadata embedded in an H.264 stream.
3552 aud Insert or remove AUD NAL units in all access units of the stream.
3554 pass
3556 insert
3557 remove
3558 Default is pass.
3560 sample_aspect_ratio
3562 Set the sample aspect ratio of the stream in the VUI parameters.
3563 See H.264 table E-1.
3564 overscan_appropriate_flag
3566 Set whether the stream is suitable for display using overscan or
3567 not (see H.264 section E.2.1).
3568 video_format
3570 video_full_range_flag
3571 Set the video format in the stream (see H.264 section E.2.1 and
3572 table E-2).
3573 colour_primaries
3575 transfer_characteristics
3576 matrix_coefficients
3577 Set the colour description in the stream (see H.264 section E.2.1
3578 and tables E-3, E-4 and E-5).
3579 chroma_sample_loc_type
3581 Set the chroma sample location in the stream (see H.264 section
3582 E.2.1 and figure E-1).
3583 tick_rate
3585 Set the tick rate (time_scale / num_units_in_tick) in the VUI
3586 parameters. This is the smallest time unit representable in the
3587 stream, and in many cases represents the field rate of the stream
3588 (double the frame rate).
3589 fixed_frame_rate_flag
3591 Set whether the stream has fixed framerate - typically this
3592 indicates that the framerate is exactly half the tick rate, but the
3593 exact meaning is dependent on interlacing and the picture structure
3594 (see H.264 section E.2.1 and table E-6).
3595 zero_new_constraint_set_flags
3597 Zero constraint_set4_flag and constraint_set5_flag in the SPS.
3598 These bits were reserved in a previous version of the H.264 spec,
3599 and thus some hardware decoders require these to be zero. The
3600 result of zeroing this is still a valid bitstream.
3601 crop_left
3603 crop_right
3604 crop_top
3605 crop_bottom
3606 Set the frame cropping offsets in the SPS. These values will
3607 replace the current ones if the stream is already cropped.
3608 These fields are set in pixels. Note that some sizes may not be
3610 representable if the chroma is subsampled or the stream is
3611 interlaced (see H.264 section 7.4.2.1.1).
3612 sei_user_data
3614 Insert a string as SEI unregistered user data. The argument must
3615 be of the form UUID+string, where the UUID is as hex digits
3616 possibly separated by hyphens, and the string can be anything.
3617 For example, 086f3693-b7b3-4f2c-9653-21492feee5b8+hello will insert
3619 the string ``hello'' associated with the given UUID.
3620 delete_filler
3622 Deletes both filler NAL units and filler SEI messages.
3623 display_orientation
3625 Insert, extract or remove Display orientation SEI messages. See
3626 H.264 section D.1.27 and D.2.27 for syntax and semantics.
3627 pass
3629 insert
3630 remove
3631 extract
3632 Default is pass.
3634 Insert mode works in conjunction with "rotate" and "flip" options.
3636 Any pre-existing Display orientation messages will be removed in
3637 insert or remove mode. Extract mode attaches the display matrix to
3638 the packet as side data.
3639 rotate
3641 Set rotation in display orientation SEI (anticlockwise angle in
3642 degrees). Range is -360 to +360. Default is NaN.
3643 flip
3645 Set flip in display orientation SEI.
3646 horizontal
3648 vertical
3649 Default is unset.
3651 level
3653 Set the level in the SPS. Refer to H.264 section A.3 and tables
3654 A-1 to A-5.
3655 The argument must be the name of a level (for example, 4.2), a
3657 level_idc value (for example, 42), or the special name auto
3658 indicating that the filter should attempt to guess the level from
3659 the input stream properties.
3660 h264_mp4toannexb
3662 Convert an H.264 bitstream from length prefixed mode to start code
3663 prefixed mode (as defined in the Annex B of the ITU-T H.264
3664 specification).
3665 This is required by some streaming formats, typically the MPEG-2
3667 transport stream format (muxer "mpegts").
3668 For example to remux an MP4 file containing an H.264 stream to mpegts
3670 format with ffmpeg, you can use the command:
3671 ffmpeg -i INPUT.mp4 -codec copy -bsf:v h264_mp4toannexb OUTPUT.ts
3673 Please note that this filter is auto-inserted for MPEG-TS (muxer
3675 "mpegts") and raw H.264 (muxer "h264") output formats.
3676 h264_redundant_pps
3678 This applies a specific fixup to some Blu-ray streams which contain
3679 redundant PPSs modifying irrelevant parameters of the stream which
3680 confuse other transformations which require correct extradata.
3681 hevc_metadata
3683 Modify metadata embedded in an HEVC stream.
3684 aud Insert or remove AUD NAL units in all access units of the stream.
3686 insert
3688 remove
3689 sample_aspect_ratio
3690 Set the sample aspect ratio in the stream in the VUI parameters.
3691 video_format
3693 video_full_range_flag
3694 Set the video format in the stream (see H.265 section E.3.1 and
3695 table E.2).
3696 colour_primaries
3698 transfer_characteristics
3699 matrix_coefficients
3700 Set the colour description in the stream (see H.265 section E.3.1
3701 and tables E.3, E.4 and E.5).
3702 chroma_sample_loc_type
3704 Set the chroma sample location in the stream (see H.265 section
3705 E.3.1 and figure E.1).
3706 tick_rate
3708 Set the tick rate in the VPS and VUI parameters (time_scale /
3709 num_units_in_tick). Combined with num_ticks_poc_diff_one, this can
3710 set a constant framerate in the stream. Note that it is likely to
3711 be overridden by container parameters when the stream is in a
3712 container.
3713 num_ticks_poc_diff_one
3715 Set poc_proportional_to_timing_flag in VPS and VUI and use this
3716 value to set num_ticks_poc_diff_one_minus1 (see H.265 sections
3717 7.4.3.1 and E.3.1). Ignored if tick_rate is not also set.
3718 crop_left
3720 crop_right
3721 crop_top
3722 crop_bottom
3723 Set the conformance window cropping offsets in the SPS. These
3724 values will replace the current ones if the stream is already
3725 cropped.
3726 These fields are set in pixels. Note that some sizes may not be
3728 representable if the chroma is subsampled (H.265 section
3729 7.4.3.2.1).
3730 level
3732 Set the level in the VPS and SPS. See H.265 section A.4 and tables
3733 A.6 and A.7.
3734 The argument must be the name of a level (for example, 5.1), a
3736 general_level_idc value (for example, 153 for level 5.1), or the
3737 special name auto indicating that the filter should attempt to
3738 guess the level from the input stream properties.
3739 hevc_mp4toannexb
3741 Convert an HEVC/H.265 bitstream from length prefixed mode to start code
3742 prefixed mode (as defined in the Annex B of the ITU-T H.265
3743 specification).
3744 This is required by some streaming formats, typically the MPEG-2
3746 transport stream format (muxer "mpegts").
3747 For example to remux an MP4 file containing an HEVC stream to mpegts
3749 format with ffmpeg, you can use the command:
3750 ffmpeg -i INPUT.mp4 -codec copy -bsf:v hevc_mp4toannexb OUTPUT.ts
3752 Please note that this filter is auto-inserted for MPEG-TS (muxer
3754 "mpegts") and raw HEVC/H.265 (muxer "h265" or "hevc") output formats.
3755 imxdump
3757 Modifies the bitstream to fit in MOV and to be usable by the Final Cut
3758 Pro decoder. This filter only applies to the mpeg2video codec, and is
3759 likely not needed for Final Cut Pro 7 and newer with the appropriate
3760 -tag:v.
3761 For example, to remux 30 MB/sec NTSC IMX to MOV:
3763 ffmpeg -i input.mxf -c copy -bsf:v imxdump -tag:v mx3n output.mov
3765 mjpeg2jpeg
3767 Convert MJPEG/AVI1 packets to full JPEG/JFIF packets.
3768 MJPEG is a video codec wherein each video frame is essentially a JPEG
3770 image. The individual frames can be extracted without loss, e.g. by
3771 ffmpeg -i ../some_mjpeg.avi -c:v copy frames_%d.jpg
3773 Unfortunately, these chunks are incomplete JPEG images, because they
3775 lack the DHT segment required for decoding. Quoting from
3776 <http://www.digitalpreservation.gov/formats/fdd/fdd000063.shtml>:
3777 Avery Lee, writing in the rec.video.desktop newsgroup in 2001,
3779 commented that "MJPEG, or at least the MJPEG in AVIs having the MJPG
3780 fourcc, is restricted JPEG with a fixed -- and *omitted* -- Huffman
3781 table. The JPEG must be YCbCr colorspace, it must be 4:2:2, and it must
3782 use basic Huffman encoding, not arithmetic or progressive. . . . You
3783 can indeed extract the MJPEG frames and decode them with a regular JPEG
3784 decoder, but you have to prepend the DHT segment to them, or else the
3785 decoder won't have any idea how to decompress the data. The exact table
3786 necessary is given in the OpenDML spec."
3787 This bitstream filter patches the header of frames extracted from an
3789 MJPEG stream (carrying the AVI1 header ID and lacking a DHT segment) to
3790 produce fully qualified JPEG images.
3791 ffmpeg -i mjpeg-movie.avi -c:v copy -bsf:v mjpeg2jpeg frame_%d.jpg
3793 exiftran -i -9 frame*.jpg
3794 ffmpeg -i frame_%d.jpg -c:v copy rotated.avi
3795 mjpegadump
3797 Add an MJPEG A header to the bitstream, to enable decoding by
3798 Quicktime.
3799 mov2textsub
3801 Extract a representable text file from MOV subtitles, stripping the
3802 metadata header from each subtitle packet.
3803 See also the text2movsub filter.
3805 mp3decomp
3807 Decompress non-standard compressed MP3 audio headers.
3808 mpeg2_metadata
3810 Modify metadata embedded in an MPEG-2 stream.
3811 display_aspect_ratio
3813 Set the display aspect ratio in the stream.
3814 The following fixed values are supported:
3816 4/3
3818 16/9
3819 221/100
3820 Any other value will result in square pixels being signalled
3822 instead (see H.262 section 6.3.3 and table 6-3).
3823 frame_rate
3825 Set the frame rate in the stream. This is constructed from a table
3826 of known values combined with a small multiplier and divisor - if
3827 the supplied value is not exactly representable, the nearest
3828 representable value will be used instead (see H.262 section 6.3.3
3829 and table 6-4).
3830 video_format
3832 Set the video format in the stream (see H.262 section 6.3.6 and
3833 table 6-6).
3834 colour_primaries
3836 transfer_characteristics
3837 matrix_coefficients
3838 Set the colour description in the stream (see H.262 section 6.3.6
3839 and tables 6-7, 6-8 and 6-9).
3840 mpeg4_unpack_bframes
3842 Unpack DivX-style packed B-frames.
3843 DivX-style packed B-frames are not valid MPEG-4 and were only a
3845 workaround for the broken Video for Windows subsystem. They use more
3846 space, can cause minor AV sync issues, require more CPU power to decode
3847 (unless the player has some decoded picture queue to compensate the
3848 2,0,2,0 frame per packet style) and cause trouble if copied into a
3849 standard container like mp4 or mpeg-ps/ts, because MPEG-4 decoders may
3850 not be able to decode them, since they are not valid MPEG-4.
3851 For example to fix an AVI file containing an MPEG-4 stream with DivX-
3853 style packed B-frames using ffmpeg, you can use the command:
3854 ffmpeg -i INPUT.avi -codec copy -bsf:v mpeg4_unpack_bframes OUTPUT.avi
3856 noise
3858 Damages the contents of packets or simply drops them without damaging
3859 the container. Can be used for fuzzing or testing error
3860 resilience/concealment.
3861 Parameters:
3863 amount
3865 Accepts an expression whose evaluation per-packet determines how
3866 often bytes in that packet will be modified. A value below 0 will
3867 result in a variable frequency. Default is 0 which results in no
3868 modification. However, if neither amount nor drop is specified,
3869 amount will be set to -1. See below for accepted variables.
3870 drop
3872 Accepts an expression evaluated per-packet whose value determines
3873 whether that packet is dropped. Evaluation to a positive value
3874 results in the packet being dropped. Evaluation to a negative value
3875 results in a variable chance of it being dropped, roughly inverse
3876 in proportion to the magnitude of the value. Default is 0 which
3877 results in no drops. See below for accepted variables.
3878 dropamount
3880 Accepts a non-negative integer, which assigns a variable chance of
3881 it being dropped, roughly inverse in proportion to the value.
3882 Default is 0 which results in no drops. This option is kept for
3883 backwards compatibility and is equivalent to setting drop to a
3884 negative value with the same magnitude i.e. "dropamount=4" is the
3885 same as "drop=-4". Ignored if drop is also specified.
3886 Both "amount" and "drop" accept expressions containing the following
3888 variables:
3889 n The index of the packet, starting from zero.
3891 tb The timebase for packet timestamps.
3893 pts Packet presentation timestamp.
3895 dts Packet decoding timestamp.
3897 nopts
3899 Constant representing AV_NOPTS_VALUE.
3900 startpts
3902 First non-AV_NOPTS_VALUE PTS seen in the stream.
3903 startdts
3905 First non-AV_NOPTS_VALUE DTS seen in the stream.
3906 duration
3908 d Packet duration, in timebase units.
3909 pos Packet position in input; may be -1 when unknown or not set.
3911 size
3913 Packet size, in bytes.
3914 key Whether packet is marked as a keyframe.
3916 state
3918 A pseudo random integer, primarily derived from the content of
3919 packet payload.
3920 Examples
3922 Apply modification to every byte but don't drop any packets.
3924 ffmpeg -i INPUT -c copy -bsf noise=1 output.mkv
3926 Drop every video packet not marked as a keyframe after timestamp 30s
3928 but do not modify any of the remaining packets.
3929 ffmpeg -i INPUT -c copy -bsf:v noise=drop='gt(t\,30)*not(key)' output.mkv
3931 Drop one second of audio every 10 seconds and add some random noise to
3933 the rest.
3934 ffmpeg -i INPUT -c copy -bsf:a noise=amount=-1:drop='between(mod(t\,10)\,9\,10)' output.mkv
3936 null
3938 This bitstream filter passes the packets through unchanged.
3939 pcm_rechunk
3941 Repacketize PCM audio to a fixed number of samples per packet or a
3942 fixed packet rate per second. This is similar to the asetnsamples audio
3943 filter but works on audio packets instead of audio frames.
3944 nb_out_samples, n
3946 Set the number of samples per each output audio packet. The number
3947 is intended as the number of samples per each channel. Default
3948 value is 1024.
3949 pad, p
3951 If set to 1, the filter will pad the last audio packet with
3952 silence, so that it will contain the same number of samples (or
3953 roughly the same number of samples, see frame_rate) as the previous
3954 ones. Default value is 1.
3955 frame_rate, r
3957 This option makes the filter output a fixed number of packets per
3958 second instead of a fixed number of samples per packet. If the
3959 audio sample rate is not divisible by the frame rate then the
3960 number of samples will not be constant but will vary slightly so
3961 that each packet will start as close to the frame boundary as
3962 possible. Using this option has precedence over nb_out_samples.
3963 You can generate the well known 1602-1601-1602-1601-1602 pattern of
3965 48kHz audio for NTSC frame rate using the frame_rate option.
3966 ffmpeg -f lavfi -i sine=r=48000:d=1 -c pcm_s16le -bsf pcm_rechunk=r=30000/1001 -f framecrc -
3968 pgs_frame_merge
3970 Merge a sequence of PGS Subtitle segments ending with an "end of
3971 display set" segment into a single packet.
3972 This is required by some containers that support PGS subtitles (muxer
3974 "matroska").
3975 prores_metadata
3977 Modify color property metadata embedded in prores stream.
3978 color_primaries
3980 Set the color primaries. Available values are:
3981 auto
3983 Keep the same color primaries property (default).
3984 unknown
3986 bt709
3987 bt470bg
3988 BT601 625
3989 smpte170m
3991 BT601 525
3992 bt2020
3994 smpte431
3995 DCI P3
3996 smpte432
3998 P3 D65
3999 transfer_characteristics
4001 Set the color transfer. Available values are:
4002 auto
4004 Keep the same transfer characteristics property (default).
4005 unknown
4007 bt709
4008 BT 601, BT 709, BT 2020
4009 smpte2084
4011 SMPTE ST 2084
4012 arib-std-b67
4014 ARIB STD-B67
4015 matrix_coefficients
4017 Set the matrix coefficient. Available values are:
4018 auto
4020 Keep the same colorspace property (default).
4021 unknown
4023 bt709
4024 smpte170m
4025 BT 601
4026 bt2020nc
4028 Set Rec709 colorspace for each frame of the file
4030 ffmpeg -i INPUT -c copy -bsf:v prores_metadata=color_primaries=bt709:color_trc=bt709:colorspace=bt709 output.mov
4032 Set Hybrid Log-Gamma parameters for each frame of the file
4034 ffmpeg -i INPUT -c copy -bsf:v prores_metadata=color_primaries=bt2020:color_trc=arib-std-b67:colorspace=bt2020nc output.mov
4036 remove_extra
4038 Remove extradata from packets.
4039 It accepts the following parameter:
4041 freq
4043 Set which frame types to remove extradata from.
4044 k Remove extradata from non-keyframes only.
4046 keyframe
4048 Remove extradata from keyframes only.
4049 e, all
4051 Remove extradata from all frames.
4052 setts
4054 Set PTS and DTS in packets.
4055 It accepts the following parameters:
4057 ts
4059 pts
4060 dts Set expressions for PTS, DTS or both.
4061 duration
4063 Set expression for duration.
4064 time_base
4066 Set output time base.
4067 The expressions are evaluated through the eval API and can contain the
4069 following constants:
4070 N The count of the input packet. Starting from 0.
4072 TS The demux timestamp in input in case of "ts" or "dts" option or
4074 presentation timestamp in case of "pts" option.
4075 POS The original position in the file of the packet, or undefined if
4077 undefined for the current packet
4078 DTS The demux timestamp in input.
4080 PTS The presentation timestamp in input.
4082 DURATION
4084 The duration in input.
4085 STARTDTS
4087 The DTS of the first packet.
4088 STARTPTS
4090 The PTS of the first packet.
4091 PREV_INDTS
4093 The previous input DTS.
4094 PREV_INPTS
4096 The previous input PTS.
4097 PREV_INDURATION
4099 The previous input duration.
4100 PREV_OUTDTS
4102 The previous output DTS.
4103 PREV_OUTPTS
4105 The previous output PTS.
4106 PREV_OUTDURATION
4108 The previous output duration.
4109 NEXT_DTS
4111 The next input DTS.
4112 NEXT_PTS
4114 The next input PTS.
4115 NEXT_DURATION
4117 The next input duration.
4118 TB The timebase of stream packet belongs.
4120 TB_OUT
4122 The output timebase.
4123 SR The sample rate of stream packet belongs.
4125 NOPTS
4127 The AV_NOPTS_VALUE constant.
4128 text2movsub
4130 Convert text subtitles to MOV subtitles (as used by the "mov_text"
4131 codec) with metadata headers.
4132 See also the mov2textsub filter.
4134 trace_headers
4136 Log trace output containing all syntax elements in the coded stream
4137 headers (everything above the level of individual coded blocks). This
4138 can be useful for debugging low-level stream issues.
4139 Supports AV1, H.264, H.265, (M)JPEG, MPEG-2 and VP9, but depending on
4141 the build only a subset of these may be available.
4142 truehd_core
4144 Extract the core from a TrueHD stream, dropping ATMOS data.
4145 vp9_metadata
4147 Modify metadata embedded in a VP9 stream.
4148 color_space
4150 Set the color space value in the frame header. Note that any frame
4151 set to RGB will be implicitly set to PC range and that RGB is
4152 incompatible with profiles 0 and 2.
4153 unknown
4155 bt601
4156 bt709
4157 smpte170
4158 smpte240
4159 bt2020
4160 rgb
4161 color_range
4162 Set the color range value in the frame header. Note that any value
4163 imposed by the color space will take precedence over this value.
4164 tv
4166 pc
4167 vp9_superframe
4169 Merge VP9 invisible (alt-ref) frames back into VP9 superframes. This
4170 fixes merging of split/segmented VP9 streams where the alt-ref frame
4171 was split from its visible counterpart.
4172 vp9_superframe_split
4174 Split VP9 superframes into single frames.
4175 vp9_raw_reorder
4177 Given a VP9 stream with correct timestamps but possibly out of order,
4178 insert additional show-existing-frame packets to correct the ordering.
4179
4181 The libavformat library provides some generic global options, which can
4182 be set on all the muxers and demuxers. In addition each muxer or
4183 demuxer may support so-called private options, which are specific for
4184 that component.
4185 Options may be set by specifying -option value in the FFmpeg tools, or
4187 by setting the value explicitly in the "AVFormatContext" options or
4188 using the libavutil/opt.h API for programmatic use.
4189 The list of supported options follows:
4191 avioflags flags (input/output)
4193 Possible values:
4194 direct
4196 Reduce buffering.
4197 probesize integer (input)
4199 Set probing size in bytes, i.e. the size of the data to analyze to
4200 get stream information. A higher value will enable detecting more
4201 information in case it is dispersed into the stream, but will
4202 increase latency. Must be an integer not lesser than 32. It is
4203 5000000 by default.
4204 max_probe_packets integer (input)
4206 Set the maximum number of buffered packets when probing a codec.
4207 Default is 2500 packets.
4208 packetsize integer (output)
4210 Set packet size.
4211 fflags flags
4213 Set format flags. Some are implemented for a limited number of
4214 formats.
4215 Possible values for input files:
4217 discardcorrupt
4219 Discard corrupted packets.
4220 fastseek
4222 Enable fast, but inaccurate seeks for some formats.
4223 genpts
4225 Generate missing PTS if DTS is present.
4226 igndts
4228 Ignore DTS if PTS is set. Inert when nofillin is set.
4229 ignidx
4231 Ignore index.
4232 nobuffer
4234 Reduce the latency introduced by buffering during initial input
4235 streams analysis.
4236 nofillin
4238 Do not fill in missing values in packet fields that can be
4239 exactly calculated.
4240 noparse
4242 Disable AVParsers, this needs "+nofillin" too.
4243 sortdts
4245 Try to interleave output packets by DTS. At present, available
4246 only for AVIs with an index.
4247 Possible values for output files:
4249 autobsf
4251 Automatically apply bitstream filters as required by the output
4252 format. Enabled by default.
4253 bitexact
4255 Only write platform-, build- and time-independent data. This
4256 ensures that file and data checksums are reproducible and match
4257 between platforms. Its primary use is for regression testing.
4258 flush_packets
4260 Write out packets immediately.
4261 shortest
4263 Stop muxing at the end of the shortest stream. It may be
4264 needed to increase max_interleave_delta to avoid flushing the
4265 longer streams before EOF.
4266 seek2any integer (input)
4268 Allow seeking to non-keyframes on demuxer level when supported if
4269 set to 1. Default is 0.
4270 analyzeduration integer (input)
4272 Specify how many microseconds are analyzed to probe the input. A
4273 higher value will enable detecting more accurate information, but
4274 will increase latency. It defaults to 5,000,000 microseconds = 5
4275 seconds.
4276 cryptokey hexadecimal string (input)
4278 Set decryption key.
4279 indexmem integer (input)
4281 Set max memory used for timestamp index (per stream).
4282 rtbufsize integer (input)
4284 Set max memory used for buffering real-time frames.
4285 fdebug flags (input/output)
4287 Print specific debug info.
4288 Possible values:
4290 ts
4292 max_delay integer (input/output)
4293 Set maximum muxing or demuxing delay in microseconds.
4294 fpsprobesize integer (input)
4296 Set number of frames used to probe fps.
4297 audio_preload integer (output)
4299 Set microseconds by which audio packets should be interleaved
4300 earlier.
4301 chunk_duration integer (output)
4303 Set microseconds for each chunk.
4304 chunk_size integer (output)
4306 Set size in bytes for each chunk.
4307 err_detect, f_err_detect flags (input)
4309 Set error detection flags. "f_err_detect" is deprecated and should
4310 be used only via the ffmpeg tool.
4311 Possible values:
4313 crccheck
4315 Verify embedded CRCs.
4316 bitstream
4318 Detect bitstream specification deviations.
4319 buffer
4321 Detect improper bitstream length.
4322 explode
4324 Abort decoding on minor error detection.
4325 careful
4327 Consider things that violate the spec and have not been seen in
4328 the wild as errors.
4329 compliant
4331 Consider all spec non compliancies as errors.
4332 aggressive
4334 Consider things that a sane encoder should not do as an error.
4335 max_interleave_delta integer (output)
4337 Set maximum buffering duration for interleaving. The duration is
4338 expressed in microseconds, and defaults to 10000000 (10 seconds).
4339 To ensure all the streams are interleaved correctly, libavformat
4341 will wait until it has at least one packet for each stream before
4342 actually writing any packets to the output file. When some streams
4343 are "sparse" (i.e. there are large gaps between successive
4344 packets), this can result in excessive buffering.
4345 This field specifies the maximum difference between the timestamps
4347 of the first and the last packet in the muxing queue, above which
4348 libavformat will output a packet regardless of whether it has
4349 queued a packet for all the streams.
4350 If set to 0, libavformat will continue buffering packets until it
4352 has a packet for each stream, regardless of the maximum timestamp
4353 difference between the buffered packets.
4354 use_wallclock_as_timestamps integer (input)
4356 Use wallclock as timestamps if set to 1. Default is 0.
4357 avoid_negative_ts integer (output)
4359 Possible values:
4360 make_non_negative
4362 Shift timestamps to make them non-negative. Also note that
4363 this affects only leading negative timestamps, and not non-
4364 monotonic negative timestamps.
4365 make_zero
4367 Shift timestamps so that the first timestamp is 0.
4368 auto (default)
4370 Enables shifting when required by the target format.
4371 disabled
4373 Disables shifting of timestamp.
4374 When shifting is enabled, all output timestamps are shifted by the
4376 same amount. Audio, video, and subtitles desynching and relative
4377 timestamp differences are preserved compared to how they would have
4378 been without shifting.
4379 skip_initial_bytes integer (input)
4381 Set number of bytes to skip before reading header and frames if set
4382 to 1. Default is 0.
4383 correct_ts_overflow integer (input)
4385 Correct single timestamp overflows if set to 1. Default is 1.
4386 flush_packets integer (output)
4388 Flush the underlying I/O stream after each packet. Default is -1
4389 (auto), which means that the underlying protocol will decide, 1
4390 enables it, and has the effect of reducing the latency, 0 disables
4391 it and may increase IO throughput in some cases.
4392 output_ts_offset offset (output)
4394 Set the output time offset.
4395 offset must be a time duration specification, see the Time duration
4397 section in the ffmpeg-utils(1) manual.
4398 The offset is added by the muxer to the output timestamps.
4400 Specifying a positive offset means that the corresponding streams
4402 are delayed bt the time duration specified in offset. Default value
4403 is 0 (meaning that no offset is applied).
4404 format_whitelist list (input)
4406 "," separated list of allowed demuxers. By default all are allowed.
4407 dump_separator string (input)
4409 Separator used to separate the fields printed on the command line
4410 about the Stream parameters. For example, to separate the fields
4411 with newlines and indentation:
4412 ffprobe -dump_separator "
4414 " -i ~/videos/matrixbench_mpeg2.mpg
4415 max_streams integer (input)
4417 Specifies the maximum number of streams. This can be used to reject
4418 files that would require too many resources due to a large number
4419 of streams.
4420 skip_estimate_duration_from_pts bool (input)
4422 Skip estimation of input duration when calculated using PTS. At
4423 present, applicable for MPEG-PS and MPEG-TS.
4424 strict, f_strict integer (input/output)
4426 Specify how strictly to follow the standards. "f_strict" is
4427 deprecated and should be used only via the ffmpeg tool.
4428 Possible values:
4430 very
4432 strictly conform to an older more strict version of the spec or
4433 reference software
4434 strict
4436 strictly conform to all the things in the spec no matter what
4437 consequences
4438 normal
4440 unofficial
4441 allow unofficial extensions
4442 experimental
4444 allow non standardized experimental things, experimental
4445 (unfinished/work in progress/not well tested) decoders and
4446 encoders. Note: experimental decoders can pose a security
4447 risk, do not use this for decoding untrusted input.
4448 Format stream specifiers
4450 Format stream specifiers allow selection of one or more streams that
4451 match specific properties.
4452 The exact semantics of stream specifiers is defined by the
4454 avformat_match_stream_specifier() function declared in the
4455 libavformat/avformat.h header and documented in the Stream specifiers
4456 section in the ffmpeg(1) manual.
4457
4459 Demuxers are configured elements in FFmpeg that can read the multimedia
4460 streams from a particular type of file.
4461 When you configure your FFmpeg build, all the supported demuxers are
4463 enabled by default. You can list all available ones using the configure
4464 option "--list-demuxers".
4465 You can disable all the demuxers using the configure option
4467 "--disable-demuxers", and selectively enable a single demuxer with the
4468 option "--enable-demuxer=DEMUXER", or disable it with the option
4469 "--disable-demuxer=DEMUXER".
4470 The option "-demuxers" of the ff* tools will display the list of
4472 enabled demuxers. Use "-formats" to view a combined list of enabled
4473 demuxers and muxers.
4474 The description of some of the currently available demuxers follows.
4476 aa
4478 Audible Format 2, 3, and 4 demuxer.
4479 This demuxer is used to demux Audible Format 2, 3, and 4 (.aa) files.
4481 aac
4483 Raw Audio Data Transport Stream AAC demuxer.
4484 This demuxer is used to demux an ADTS input containing a single AAC
4486 stream alongwith any ID3v1/2 or APE tags in it.
4487 apng
4489 Animated Portable Network Graphics demuxer.
4490 This demuxer is used to demux APNG files. All headers, but the PNG
4492 signature, up to (but not including) the first fcTL chunk are
4493 transmitted as extradata. Frames are then split as being all the
4494 chunks between two fcTL ones, or between the last fcTL and IEND chunks.
4495 -ignore_loop bool
4497 Ignore the loop variable in the file if set. Default is enabled.
4498 -max_fps int
4500 Maximum framerate in frames per second. Default of 0 imposes no
4501 limit.
4502 -default_fps int
4504 Default framerate in frames per second when none is specified in
4505 the file (0 meaning as fast as possible). Default is 15.
4506 asf
4508 Advanced Systems Format demuxer.
4509 This demuxer is used to demux ASF files and MMS network streams.
4511 -no_resync_search bool
4513 Do not try to resynchronize by looking for a certain optional start
4514 code.
4515 concat
4517 Virtual concatenation script demuxer.
4518 This demuxer reads a list of files and other directives from a text
4520 file and demuxes them one after the other, as if all their packets had
4521 been muxed together.
4522 The timestamps in the files are adjusted so that the first file starts
4524 at 0 and each next file starts where the previous one finishes. Note
4525 that it is done globally and may cause gaps if all streams do not have
4526 exactly the same length.
4527 All files must have the same streams (same codecs, same time base,
4529 etc.).
4530 The duration of each file is used to adjust the timestamps of the next
4532 file: if the duration is incorrect (because it was computed using the
4533 bit-rate or because the file is truncated, for example), it can cause
4534 artifacts. The "duration" directive can be used to override the
4535 duration stored in each file.
4536 Syntax
4538 The script is a text file in extended-ASCII, with one directive per
4540 line. Empty lines, leading spaces and lines starting with '#' are
4541 ignored. The following directive is recognized:
4542 "file path"
4544 Path to a file to read; special characters and spaces must be
4545 escaped with backslash or single quotes.
4546 All subsequent file-related directives apply to that file.
4548 "ffconcat version 1.0"
4550 Identify the script type and version.
4551 To make FFmpeg recognize the format automatically, this directive
4553 must appear exactly as is (no extra space or byte-order-mark) on
4554 the very first line of the script.
4555 "duration dur"
4557 Duration of the file. This information can be specified from the
4558 file; specifying it here may be more efficient or help if the
4559 information from the file is not available or accurate.
4560 If the duration is set for all files, then it is possible to seek
4562 in the whole concatenated video.
4563 "inpoint timestamp"
4565 In point of the file. When the demuxer opens the file it instantly
4566 seeks to the specified timestamp. Seeking is done so that all
4567 streams can be presented successfully at In point.
4568 This directive works best with intra frame codecs, because for non-
4570 intra frame ones you will usually get extra packets before the
4571 actual In point and the decoded content will most likely contain
4572 frames before In point too.
4573 For each file, packets before the file In point will have
4575 timestamps less than the calculated start timestamp of the file
4576 (negative in case of the first file), and the duration of the files
4577 (if not specified by the "duration" directive) will be reduced
4578 based on their specified In point.
4579 Because of potential packets before the specified In point, packet
4581 timestamps may overlap between two concatenated files.
4582 "outpoint timestamp"
4584 Out point of the file. When the demuxer reaches the specified
4585 decoding timestamp in any of the streams, it handles it as an end
4586 of file condition and skips the current and all the remaining
4587 packets from all streams.
4588 Out point is exclusive, which means that the demuxer will not
4590 output packets with a decoding timestamp greater or equal to Out
4591 point.
4592 This directive works best with intra frame codecs and formats where
4594 all streams are tightly interleaved. For non-intra frame codecs you
4595 will usually get additional packets with presentation timestamp
4596 after Out point therefore the decoded content will most likely
4597 contain frames after Out point too. If your streams are not tightly
4598 interleaved you may not get all the packets from all streams before
4599 Out point and you may only will be able to decode the earliest
4600 stream until Out point.
4601 The duration of the files (if not specified by the "duration"
4603 directive) will be reduced based on their specified Out point.
4604 "file_packet_metadata key=value"
4606 Metadata of the packets of the file. The specified metadata will be
4607 set for each file packet. You can specify this directive multiple
4608 times to add multiple metadata entries. This directive is
4609 deprecated, use "file_packet_meta" instead.
4610 "file_packet_meta key value"
4612 Metadata of the packets of the file. The specified metadata will be
4613 set for each file packet. You can specify this directive multiple
4614 times to add multiple metadata entries.
4615 "option key value"
4617 Option to access, open and probe the file. Can be present multiple
4618 times.
4619 "stream"
4621 Introduce a stream in the virtual file. All subsequent stream-
4622 related directives apply to the last introduced stream. Some
4623 streams properties must be set in order to allow identifying the
4624 matching streams in the subfiles. If no streams are defined in the
4625 script, the streams from the first file are copied.
4626 "exact_stream_id id"
4628 Set the id of the stream. If this directive is given, the string
4629 with the corresponding id in the subfiles will be used. This is
4630 especially useful for MPEG-PS (VOB) files, where the order of the
4631 streams is not reliable.
4632 "stream_meta key value"
4634 Metadata for the stream. Can be present multiple times.
4635 "stream_codec value"
4637 Codec for the stream.
4638 "stream_extradata hex_string"
4640 Extradata for the string, encoded in hexadecimal.
4641 "chapter id start end"
4643 Add a chapter. id is an unique identifier, possibly small and
4644 consecutive.
4645 Options
4647 This demuxer accepts the following option:
4649 safe
4651 If set to 1, reject unsafe file paths and directives. A file path
4652 is considered safe if it does not contain a protocol specification
4653 and is relative and all components only contain characters from the
4654 portable character set (letters, digits, period, underscore and
4655 hyphen) and have no period at the beginning of a component.
4656 If set to 0, any file name is accepted.
4658 The default is 1.
4660 auto_convert
4662 If set to 1, try to perform automatic conversions on packet data to
4663 make the streams concatenable. The default is 1.
4664 Currently, the only conversion is adding the h264_mp4toannexb
4666 bitstream filter to H.264 streams in MP4 format. This is necessary
4667 in particular if there are resolution changes.
4668 segment_time_metadata
4670 If set to 1, every packet will contain the lavf.concat.start_time
4671 and the lavf.concat.duration packet metadata values which are the
4672 start_time and the duration of the respective file segments in the
4673 concatenated output expressed in microseconds. The duration
4674 metadata is only set if it is known based on the concat file. The
4675 default is 0.
4676 Examples
4678 • Use absolute filenames and include some comments:
4680 # my first filename
4682 file /mnt/share/file-1.wav
4683 # my second filename including whitespace
4684 file '/mnt/share/file 2.wav'
4685 # my third filename including whitespace plus single quote
4686 file '/mnt/share/file 3'\''.wav'
4687 • Allow for input format auto-probing, use safe filenames and set the
4689 duration of the first file:
4690 ffconcat version 1.0
4692 file file-1.wav
4694 duration 20.0
4695 file subdir/file-2.wav
4697 dash
4699 Dynamic Adaptive Streaming over HTTP demuxer.
4700 This demuxer presents all AVStreams found in the manifest. By setting
4702 the discard flags on AVStreams the caller can decide which streams to
4703 actually receive. Each stream mirrors the "id" and "bandwidth"
4704 properties from the "<Representation>" as metadata keys named "id" and
4705 "variant_bitrate" respectively.
4706 Options
4708 This demuxer accepts the following option:
4710 cenc_decryption_key
4712 16-byte key, in hex, to decrypt files encrypted using ISO Common
4713 Encryption (CENC/AES-128 CTR; ISO/IEC 23001-7).
4714 ea
4716 Electronic Arts Multimedia format demuxer.
4717 This format is used by various Electronic Arts games.
4719 Options
4721 merge_alpha bool
4723 Normally the VP6 alpha channel (if exists) is returned as a
4724 secondary video stream, by setting this option you can make the
4725 demuxer return a single video stream which contains the alpha
4726 channel in addition to the ordinary video.
4727 imf
4729 Interoperable Master Format demuxer.
4730 This demuxer presents audio and video streams found in an IMF
4732 Composition.
4733 flv, live_flv, kux
4735 Adobe Flash Video Format demuxer.
4736 This demuxer is used to demux FLV files and RTMP network streams. In
4738 case of live network streams, if you force format, you may use live_flv
4739 option instead of flv to survive timestamp discontinuities. KUX is a
4740 flv variant used on the Youku platform.
4741 ffmpeg -f flv -i myfile.flv ...
4743 ffmpeg -f live_flv -i rtmp://<any.server>/anything/key ....
4744 -flv_metadata bool
4746 Allocate the streams according to the onMetaData array content.
4747 -flv_ignore_prevtag bool
4749 Ignore the size of previous tag value.
4750 -flv_full_metadata bool
4752 Output all context of the onMetadata.
4753 gif
4755 Animated GIF demuxer.
4756 It accepts the following options:
4758 min_delay
4760 Set the minimum valid delay between frames in hundredths of
4761 seconds. Range is 0 to 6000. Default value is 2.
4762 max_gif_delay
4764 Set the maximum valid delay between frames in hundredth of seconds.
4765 Range is 0 to 65535. Default value is 65535 (nearly eleven
4766 minutes), the maximum value allowed by the specification.
4767 default_delay
4769 Set the default delay between frames in hundredths of seconds.
4770 Range is 0 to 6000. Default value is 10.
4771 ignore_loop
4773 GIF files can contain information to loop a certain number of times
4774 (or infinitely). If ignore_loop is set to 1, then the loop setting
4775 from the input will be ignored and looping will not occur. If set
4776 to 0, then looping will occur and will cycle the number of times
4777 according to the GIF. Default value is 1.
4778 For example, with the overlay filter, place an infinitely looping GIF
4780 over another video:
4781 ffmpeg -i input.mp4 -ignore_loop 0 -i input.gif -filter_complex overlay=shortest=1 out.mkv
4783 Note that in the above example the shortest option for overlay filter
4785 is used to end the output video at the length of the shortest input
4786 file, which in this case is input.mp4 as the GIF in this example loops
4787 infinitely.
4788 hls
4790 HLS demuxer
4791 Apple HTTP Live Streaming demuxer.
4793 This demuxer presents all AVStreams from all variant streams. The id
4795 field is set to the bitrate variant index number. By setting the
4796 discard flags on AVStreams (by pressing 'a' or 'v' in ffplay), the
4797 caller can decide which variant streams to actually receive. The total
4798 bitrate of the variant that the stream belongs to is available in a
4799 metadata key named "variant_bitrate".
4800 It accepts the following options:
4802 live_start_index
4804 segment index to start live streams at (negative values are from
4805 the end).
4806 prefer_x_start
4808 prefer to use #EXT-X-START if it's in playlist instead of
4809 live_start_index.
4810 allowed_extensions
4812 ',' separated list of file extensions that hls is allowed to
4813 access.
4814 max_reload
4816 Maximum number of times a insufficient list is attempted to be
4817 reloaded. Default value is 1000.
4818 m3u8_hold_counters
4820 The maximum number of times to load m3u8 when it refreshes without
4821 new segments. Default value is 1000.
4822 http_persistent
4824 Use persistent HTTP connections. Applicable only for HTTP streams.
4825 Enabled by default.
4826 http_multiple
4828 Use multiple HTTP connections for downloading HTTP segments.
4829 Enabled by default for HTTP/1.1 servers.
4830 http_seekable
4832 Use HTTP partial requests for downloading HTTP segments. 0 =
4833 disable, 1 = enable, -1 = auto, Default is auto.
4834 seg_format_options
4836 Set options for the demuxer of media segments using a list of
4837 key=value pairs separated by ":".
4838 seg_max_retry
4840 Maximum number of times to reload a segment on error, useful when
4841 segment skip on network error is not desired. Default value is 0.
4842 image2
4844 Image file demuxer.
4845 This demuxer reads from a list of image files specified by a pattern.
4847 The syntax and meaning of the pattern is specified by the option
4848 pattern_type.
4849 The pattern may contain a suffix which is used to automatically
4851 determine the format of the images contained in the files.
4852 The size, the pixel format, and the format of each image must be the
4854 same for all the files in the sequence.
4855 This demuxer accepts the following options:
4857 framerate
4859 Set the frame rate for the video stream. It defaults to 25.
4860 loop
4862 If set to 1, loop over the input. Default value is 0.
4863 pattern_type
4865 Select the pattern type used to interpret the provided filename.
4866 pattern_type accepts one of the following values.
4868 none
4870 Disable pattern matching, therefore the video will only contain
4871 the specified image. You should use this option if you do not
4872 want to create sequences from multiple images and your
4873 filenames may contain special pattern characters.
4874 sequence
4876 Select a sequence pattern type, used to specify a sequence of
4877 files indexed by sequential numbers.
4878 A sequence pattern may contain the string "%d" or "%0Nd", which
4880 specifies the position of the characters representing a
4881 sequential number in each filename matched by the pattern. If
4882 the form "%d0Nd" is used, the string representing the number in
4883 each filename is 0-padded and N is the total number of 0-padded
4884 digits representing the number. The literal character '%' can
4885 be specified in the pattern with the string "%%".
4886 If the sequence pattern contains "%d" or "%0Nd", the first
4888 filename of the file list specified by the pattern must contain
4889 a number inclusively contained between start_number and
4890 start_number+start_number_range-1, and all the following
4891 numbers must be sequential.
4892 For example the pattern "img-%03d.bmp" will match a sequence of
4894 filenames of the form img-001.bmp, img-002.bmp, ...,
4895 img-010.bmp, etc.; the pattern "i%%m%%g-%d.jpg" will match a
4896 sequence of filenames of the form i%m%g-1.jpg, i%m%g-2.jpg,
4897 ..., i%m%g-10.jpg, etc.
4898 Note that the pattern must not necessarily contain "%d" or
4900 "%0Nd", for example to convert a single image file img.jpeg you
4901 can employ the command:
4902 ffmpeg -i img.jpeg img.png
4904 glob
4906 Select a glob wildcard pattern type.
4907 The pattern is interpreted like a glob() pattern. This is only
4909 selectable if libavformat was compiled with globbing support.
4910 glob_sequence (deprecated, will be removed)
4912 Select a mixed glob wildcard/sequence pattern.
4913 If your version of libavformat was compiled with globbing
4915 support, and the provided pattern contains at least one glob
4916 meta character among "%*?[]{}" that is preceded by an unescaped
4917 "%", the pattern is interpreted like a glob() pattern,
4918 otherwise it is interpreted like a sequence pattern.
4919 All glob special characters "%*?[]{}" must be prefixed with
4921 "%". To escape a literal "%" you shall use "%%".
4922 For example the pattern "foo-%*.jpeg" will match all the
4924 filenames prefixed by "foo-" and terminating with ".jpeg", and
4925 "foo-%?%?%?.jpeg" will match all the filenames prefixed with
4926 "foo-", followed by a sequence of three characters, and
4927 terminating with ".jpeg".
4928 This pattern type is deprecated in favor of glob and sequence.
4930 Default value is glob_sequence.
4932 pixel_format
4934 Set the pixel format of the images to read. If not specified the
4935 pixel format is guessed from the first image file in the sequence.
4936 start_number
4938 Set the index of the file matched by the image file pattern to
4939 start to read from. Default value is 0.
4940 start_number_range
4942 Set the index interval range to check when looking for the first
4943 image file in the sequence, starting from start_number. Default
4944 value is 5.
4945 ts_from_file
4947 If set to 1, will set frame timestamp to modification time of image
4948 file. Note that monotonity of timestamps is not provided: images go
4949 in the same order as without this option. Default value is 0. If
4950 set to 2, will set frame timestamp to the modification time of the
4951 image file in nanosecond precision.
4952 video_size
4954 Set the video size of the images to read. If not specified the
4955 video size is guessed from the first image file in the sequence.
4956 export_path_metadata
4958 If set to 1, will add two extra fields to the metadata found in
4959 input, making them also available for other filters (see drawtext
4960 filter for examples). Default value is 0. The extra fields are
4961 described below:
4962 lavf.image2dec.source_path
4964 Corresponds to the full path to the input file being read.
4965 lavf.image2dec.source_basename
4967 Corresponds to the name of the file being read.
4968 Examples
4970 • Use ffmpeg for creating a video from the images in the file
4972 sequence img-001.jpeg, img-002.jpeg, ..., assuming an input frame
4973 rate of 10 frames per second:
4974 ffmpeg -framerate 10 -i 'img-%03d.jpeg' out.mkv
4976 • As above, but start by reading from a file with index 100 in the
4978 sequence:
4979 ffmpeg -framerate 10 -start_number 100 -i 'img-%03d.jpeg' out.mkv
4981 • Read images matching the "*.png" glob pattern , that is all the
4983 files terminating with the ".png" suffix:
4984 ffmpeg -framerate 10 -pattern_type glob -i "*.png" out.mkv
4986 libgme
4988 The Game Music Emu library is a collection of video game music file
4989 emulators.
4990 See <https://bitbucket.org/mpyne/game-music-emu/overview> for more
4992 information.
4993 It accepts the following options:
4995 track_index
4997 Set the index of which track to demux. The demuxer can only export
4998 one track. Track indexes start at 0. Default is to pick the first
4999 track. Number of tracks is exported as tracks metadata entry.
5000 sample_rate
5002 Set the sampling rate of the exported track. Range is 1000 to
5003 999999. Default is 44100.
5004 max_size (bytes)
5006 The demuxer buffers the entire file into memory. Adjust this value
5007 to set the maximum buffer size, which in turn, acts as a ceiling
5008 for the size of files that can be read. Default is 50 MiB.
5009 libmodplug
5011 ModPlug based module demuxer
5012 See <https://github.com/Konstanty/libmodplug>
5014 It will export one 2-channel 16-bit 44.1 kHz audio stream. Optionally,
5016 a "pal8" 16-color video stream can be exported with or without printed
5017 metadata.
5018 It accepts the following options:
5020 noise_reduction
5022 Apply a simple low-pass filter. Can be 1 (on) or 0 (off). Default
5023 is 0.
5024 reverb_depth
5026 Set amount of reverb. Range 0-100. Default is 0.
5027 reverb_delay
5029 Set delay in ms, clamped to 40-250 ms. Default is 0.
5030 bass_amount
5032 Apply bass expansion a.k.a. XBass or megabass. Range is 0 (quiet)
5033 to 100 (loud). Default is 0.
5034 bass_range
5036 Set cutoff i.e. upper-bound for bass frequencies. Range is 10-100
5037 Hz. Default is 0.
5038 surround_depth
5040 Apply a Dolby Pro-Logic surround effect. Range is 0 (quiet) to 100
5041 (heavy). Default is 0.
5042 surround_delay
5044 Set surround delay in ms, clamped to 5-40 ms. Default is 0.
5045 max_size
5047 The demuxer buffers the entire file into memory. Adjust this value
5048 to set the maximum buffer size, which in turn, acts as a ceiling
5049 for the size of files that can be read. Range is 0 to 100 MiB. 0
5050 removes buffer size limit (not recommended). Default is 5 MiB.
5051 video_stream_expr
5053 String which is evaluated using the eval API to assign colors to
5054 the generated video stream. Variables which can be used are "x",
5055 "y", "w", "h", "t", "speed", "tempo", "order", "pattern" and "row".
5056 video_stream
5058 Generate video stream. Can be 1 (on) or 0 (off). Default is 0.
5059 video_stream_w
5061 Set video frame width in 'chars' where one char indicates 8 pixels.
5062 Range is 20-512. Default is 30.
5063 video_stream_h
5065 Set video frame height in 'chars' where one char indicates 8
5066 pixels. Range is 20-512. Default is 30.
5067 video_stream_ptxt
5069 Print metadata on video stream. Includes "speed", "tempo", "order",
5070 "pattern", "row" and "ts" (time in ms). Can be 1 (on) or 0 (off).
5071 Default is 1.
5072 libopenmpt
5074 libopenmpt based module demuxer
5075 See <https://lib.openmpt.org/libopenmpt/> for more information.
5077 Some files have multiple subsongs (tracks) this can be set with the
5079 subsong option.
5080 It accepts the following options:
5082 subsong
5084 Set the subsong index. This can be either 'all', 'auto', or the
5085 index of the subsong. Subsong indexes start at 0. The default is
5086 'auto'.
5087 The default value is to let libopenmpt choose.
5089 layout
5091 Set the channel layout. Valid values are 1, 2, and 4 channel
5092 layouts. The default value is STEREO.
5093 sample_rate
5095 Set the sample rate for libopenmpt to output. Range is from 1000
5096 to INT_MAX. The value default is 48000.
5097 mov/mp4/3gp
5099 Demuxer for Quicktime File Format & ISO/IEC Base Media File Format
5100 (ISO/IEC 14496-12 or MPEG-4 Part 12, ISO/IEC 15444-12 or JPEG 2000 Part
5101 12).
5102 Registered extensions: mov, mp4, m4a, 3gp, 3g2, mj2, psp, m4b, ism,
5104 ismv, isma, f4v
5105 Options
5107 This demuxer accepts the following options:
5109 enable_drefs
5111 Enable loading of external tracks, disabled by default. Enabling
5112 this can theoretically leak information in some use cases.
5113 use_absolute_path
5115 Allows loading of external tracks via absolute paths, disabled by
5116 default. Enabling this poses a security risk. It should only be
5117 enabled if the source is known to be non-malicious.
5118 seek_streams_individually
5120 When seeking, identify the closest point in each stream
5121 individually and demux packets in that stream from identified
5122 point. This can lead to a different sequence of packets compared to
5123 demuxing linearly from the beginning. Default is true.
5124 ignore_editlist
5126 Ignore any edit list atoms. The demuxer, by default, modifies the
5127 stream index to reflect the timeline described by the edit list.
5128 Default is false.
5129 advanced_editlist
5131 Modify the stream index to reflect the timeline described by the
5132 edit list. "ignore_editlist" must be set to false for this option
5133 to be effective. If both "ignore_editlist" and this option are set
5134 to false, then only the start of the stream index is modified to
5135 reflect initial dwell time or starting timestamp described by the
5136 edit list. Default is true.
5137 ignore_chapters
5139 Don't parse chapters. This includes GoPro 'HiLight' tags/moments.
5140 Note that chapters are only parsed when input is seekable. Default
5141 is false.
5142 use_mfra_for
5144 For seekable fragmented input, set fragment's starting timestamp
5145 from media fragment random access box, if present.
5146 Following options are available:
5148 auto
5150 Auto-detect whether to set mfra timestamps as PTS or DTS
5151 (default)
5152 dts Set mfra timestamps as DTS
5154 pts Set mfra timestamps as PTS
5156 0 Don't use mfra box to set timestamps
5158 use_tfdt
5160 For fragmented input, set fragment's starting timestamp to
5161 "baseMediaDecodeTime" from the "tfdt" box. Default is enabled,
5162 which will prefer to use the "tfdt" box to set DTS. Disable to use
5163 the "earliest_presentation_time" from the "sidx" box. In either
5164 case, the timestamp from the "mfra" box will be used if it's
5165 available and "use_mfra_for" is set to pts or dts.
5166 export_all
5168 Export unrecognized boxes within the udta box as metadata entries.
5169 The first four characters of the box type are set as the key.
5170 Default is false.
5171 export_xmp
5173 Export entire contents of XMP_ box and uuid box as a string with
5174 key "xmp". Note that if "export_all" is set and this option isn't,
5175 the contents of XMP_ box are still exported but with key "XMP_".
5176 Default is false.
5177 activation_bytes
5179 4-byte key required to decrypt Audible AAX and AAX+ files. See
5180 Audible AAX subsection below.
5181 audible_fixed_key
5183 Fixed key used for handling Audible AAX/AAX+ files. It has been
5184 pre-set so should not be necessary to specify.
5185 decryption_key
5187 16-byte key, in hex, to decrypt files encrypted using ISO Common
5188 Encryption (CENC/AES-128 CTR; ISO/IEC 23001-7).
5189 max_stts_delta
5191 Very high sample deltas written in a trak's stts box may
5192 occasionally be intended but usually they are written in error or
5193 used to store a negative value for dts correction when treated as
5194 signed 32-bit integers. This option lets the user set an upper
5195 limit, beyond which the delta is clamped to 1. Values greater than
5196 the limit if negative when cast to int32 are used to adjust onward
5197 dts.
5198 Unit is the track time scale. Range is 0 to UINT_MAX. Default is
5200 "UINT_MAX - 48000*10" which allows upto a 10 second dts correction
5201 for 48 kHz audio streams while accommodating 99.9% of "uint32"
5202 range.
5203 Audible AAX
5205 Audible AAX files are encrypted M4B files, and they can be decrypted by
5207 specifying a 4 byte activation secret.
5208 ffmpeg -activation_bytes 1CEB00DA -i test.aax -vn -c:a copy output.mp4
5210 mpegts
5212 MPEG-2 transport stream demuxer.
5213 This demuxer accepts the following options:
5215 resync_size
5217 Set size limit for looking up a new synchronization. Default value
5218 is 65536.
5219 skip_unknown_pmt
5221 Skip PMTs for programs not defined in the PAT. Default value is 0.
5222 fix_teletext_pts
5224 Override teletext packet PTS and DTS values with the timestamps
5225 calculated from the PCR of the first program which the teletext
5226 stream is part of and is not discarded. Default value is 1, set
5227 this option to 0 if you want your teletext packet PTS and DTS
5228 values untouched.
5229 ts_packetsize
5231 Output option carrying the raw packet size in bytes. Show the
5232 detected raw packet size, cannot be set by the user.
5233 scan_all_pmts
5235 Scan and combine all PMTs. The value is an integer with value from
5236 -1 to 1 (-1 means automatic setting, 1 means enabled, 0 means
5237 disabled). Default value is -1.
5238 merge_pmt_versions
5240 Re-use existing streams when a PMT's version is updated and
5241 elementary streams move to different PIDs. Default value is 0.
5242 max_packet_size
5244 Set maximum size, in bytes, of packet emitted by the demuxer.
5245 Payloads above this size are split across multiple packets. Range
5246 is 1 to INT_MAX/2. Default is 204800 bytes.
5247 mpjpeg
5249 MJPEG encapsulated in multi-part MIME demuxer.
5250 This demuxer allows reading of MJPEG, where each frame is represented
5252 as a part of multipart/x-mixed-replace stream.
5253 strict_mime_boundary
5255 Default implementation applies a relaxed standard to multi-part
5256 MIME boundary detection, to prevent regression with numerous
5257 existing endpoints not generating a proper MIME MJPEG stream.
5258 Turning this option on by setting it to 1 will result in a stricter
5259 check of the boundary value.
5260 rawvideo
5262 Raw video demuxer.
5263 This demuxer allows one to read raw video data. Since there is no
5265 header specifying the assumed video parameters, the user must specify
5266 them in order to be able to decode the data correctly.
5267 This demuxer accepts the following options:
5269 framerate
5271 Set input video frame rate. Default value is 25.
5272 pixel_format
5274 Set the input video pixel format. Default value is "yuv420p".
5275 video_size
5277 Set the input video size. This value must be specified explicitly.
5278 For example to read a rawvideo file input.raw with ffplay, assuming a
5280 pixel format of "rgb24", a video size of "320x240", and a frame rate of
5281 10 images per second, use the command:
5282 ffplay -f rawvideo -pixel_format rgb24 -video_size 320x240 -framerate 10 input.raw
5284 sbg
5286 SBaGen script demuxer.
5287 This demuxer reads the script language used by SBaGen
5289 <http://uazu.net/sbagen/> to generate binaural beats sessions. A SBG
5290 script looks like that:
5291 -SE
5293 a: 300-2.5/3 440+4.5/0
5294 b: 300-2.5/0 440+4.5/3
5295 off: -
5296 NOW == a
5297 +0:07:00 == b
5298 +0:14:00 == a
5299 +0:21:00 == b
5300 +0:30:00 off
5301 A SBG script can mix absolute and relative timestamps. If the script
5303 uses either only absolute timestamps (including the script start time)
5304 or only relative ones, then its layout is fixed, and the conversion is
5305 straightforward. On the other hand, if the script mixes both kind of
5306 timestamps, then the NOW reference for relative timestamps will be
5307 taken from the current time of day at the time the script is read, and
5308 the script layout will be frozen according to that reference. That
5309 means that if the script is directly played, the actual times will
5310 match the absolute timestamps up to the sound controller's clock
5311 accuracy, but if the user somehow pauses the playback or seeks, all
5312 times will be shifted accordingly.
5313 tedcaptions
5315 JSON captions used for <http://www.ted.com/>.
5316 TED does not provide links to the captions, but they can be guessed
5318 from the page. The file tools/bookmarklets.html from the FFmpeg source
5319 tree contains a bookmarklet to expose them.
5320 This demuxer accepts the following option:
5322 start_time
5324 Set the start time of the TED talk, in milliseconds. The default is
5325 15000 (15s). It is used to sync the captions with the downloadable
5326 videos, because they include a 15s intro.
5327 Example: convert the captions to a format most players understand:
5329 ffmpeg -i http://www.ted.com/talks/subtitles/id/1/lang/en talk1-en.srt
5331 vapoursynth
5333 Vapoursynth wrapper.
5334 Due to security concerns, Vapoursynth scripts will not be autodetected
5336 so the input format has to be forced. For ff* CLI tools, add "-f
5337 vapoursynth" before the input "-i yourscript.vpy".
5338 This demuxer accepts the following option:
5340 max_script_size
5342 The demuxer buffers the entire script into memory. Adjust this
5343 value to set the maximum buffer size, which in turn, acts as a
5344 ceiling for the size of scripts that can be read. Default is 1
5345 MiB.
5346
5348 FFmpeg is able to dump metadata from media files into a simple
5349 UTF-8-encoded INI-like text file and then load it back using the
5350 metadata muxer/demuxer.
5351 The file format is as follows:
5353 1. A file consists of a header and a number of metadata tags divided
5355 into sections, each on its own line.
5356 2. The header is a ;FFMETADATA string, followed by a version number
5358 (now 1).
5359 3. Metadata tags are of the form key=value
5361 4. Immediately after header follows global metadata
5363 5. After global metadata there may be sections with
5365 per-stream/per-chapter metadata.
5366 6. A section starts with the section name in uppercase (i.e. STREAM or
5368 CHAPTER) in brackets ([, ]) and ends with next section or end of
5369 file.
5370 7. At the beginning of a chapter section there may be an optional
5372 timebase to be used for start/end values. It must be in form
5373 TIMEBASE=num/den, where num and den are integers. If the timebase
5374 is missing then start/end times are assumed to be in nanoseconds.
5375 Next a chapter section must contain chapter start and end times in
5377 form START=num, END=num, where num is a positive integer.
5378 8. Empty lines and lines starting with ; or # are ignored.
5380 9. Metadata keys or values containing special characters (=, ;, #, \
5382 and a newline) must be escaped with a backslash \.
5383 10. Note that whitespace in metadata (e.g. foo = bar) is considered to
5385 be a part of the tag (in the example above key is foo , value is
5386 bar).
5387 A ffmetadata file might look like this:
5389 ;FFMETADATA1
5391 title=bike\\shed
5392 ;this is a comment
5393 artist=FFmpeg troll team
5394 [CHAPTER]
5396 TIMEBASE=1/1000
5397 START=0
5398 #chapter ends at 0:01:00
5399 END=60000
5400 title=chapter \#1
5401 [STREAM]
5402 title=multi\
5403 line
5404 By using the ffmetadata muxer and demuxer it is possible to extract
5406 metadata from an input file to an ffmetadata file, and then transcode
5407 the file into an output file with the edited ffmetadata file.
5408 Extracting an ffmetadata file with ffmpeg goes as follows:
5410 ffmpeg -i INPUT -f ffmetadata FFMETADATAFILE
5412 Reinserting edited metadata information from the FFMETADATAFILE file
5414 can be done as:
5415 ffmpeg -i INPUT -i FFMETADATAFILE -map_metadata 1 -codec copy OUTPUT
5417
5419 The libavformat library provides some generic global options, which can
5420 be set on all the protocols. In addition each protocol may support so-
5421 called private options, which are specific for that component.
5422 Options may be set by specifying -option value in the FFmpeg tools, or
5424 by setting the value explicitly in the "AVFormatContext" options or
5425 using the libavutil/opt.h API for programmatic use.
5426 The list of supported options follows:
5428 protocol_whitelist list (input)
5430 Set a ","-separated list of allowed protocols. "ALL" matches all
5431 protocols. Protocols prefixed by "-" are disabled. All protocols
5432 are allowed by default but protocols used by an another protocol
5433 (nested protocols) are restricted to a per protocol subset.
5434
5436 Protocols are configured elements in FFmpeg that enable access to
5437 resources that require specific protocols.
5438 When you configure your FFmpeg build, all the supported protocols are
5440 enabled by default. You can list all available ones using the configure
5441 option "--list-protocols".
5442 You can disable all the protocols using the configure option
5444 "--disable-protocols", and selectively enable a protocol using the
5445 option "--enable-protocol=PROTOCOL", or you can disable a particular
5446 protocol using the option "--disable-protocol=PROTOCOL".
5447 The option "-protocols" of the ff* tools will display the list of
5449 supported protocols.
5450 All protocols accept the following options:
5452 rw_timeout
5454 Maximum time to wait for (network) read/write operations to
5455 complete, in microseconds.
5456 A description of the currently available protocols follows.
5458 amqp
5460 Advanced Message Queueing Protocol (AMQP) version 0-9-1 is a broker
5461 based publish-subscribe communication protocol.
5462 FFmpeg must be compiled with --enable-librabbitmq to support AMQP. A
5464 separate AMQP broker must also be run. An example open-source AMQP
5465 broker is RabbitMQ.
5466 After starting the broker, an FFmpeg client may stream data to the
5468 broker using the command:
5469 ffmpeg -re -i input -f mpegts amqp://[[user]:[password]@]hostname[:port][/vhost]
5471 Where hostname and port (default is 5672) is the address of the broker.
5473 The client may also set a user/password for authentication. The default
5474 for both fields is "guest". Name of virtual host on broker can be set
5475 with vhost. The default value is "/".
5476 Muliple subscribers may stream from the broker using the command:
5478 ffplay amqp://[[user]:[password]@]hostname[:port][/vhost]
5480 In RabbitMQ all data published to the broker flows through a specific
5482 exchange, and each subscribing client has an assigned queue/buffer.
5483 When a packet arrives at an exchange, it may be copied to a client's
5484 queue depending on the exchange and routing_key fields.
5485 The following options are supported:
5487 exchange
5489 Sets the exchange to use on the broker. RabbitMQ has several
5490 predefined exchanges: "amq.direct" is the default exchange, where
5491 the publisher and subscriber must have a matching routing_key;
5492 "amq.fanout" is the same as a broadcast operation (i.e. the data is
5493 forwarded to all queues on the fanout exchange independent of the
5494 routing_key); and "amq.topic" is similar to "amq.direct", but
5495 allows for more complex pattern matching (refer to the RabbitMQ
5496 documentation).
5497 routing_key
5499 Sets the routing key. The default value is "amqp". The routing key
5500 is used on the "amq.direct" and "amq.topic" exchanges to decide
5501 whether packets are written to the queue of a subscriber.
5502 pkt_size
5504 Maximum size of each packet sent/received to the broker. Default is
5505 131072. Minimum is 4096 and max is any large value (representable
5506 by an int). When receiving packets, this sets an internal buffer
5507 size in FFmpeg. It should be equal to or greater than the size of
5508 the published packets to the broker. Otherwise the received message
5509 may be truncated causing decoding errors.
5510 connection_timeout
5512 The timeout in seconds during the initial connection to the broker.
5513 The default value is rw_timeout, or 5 seconds if rw_timeout is not
5514 set.
5515 delivery_mode mode
5517 Sets the delivery mode of each message sent to broker. The
5518 following values are accepted:
5519 persistent
5521 Delivery mode set to "persistent" (2). This is the default
5522 value. Messages may be written to the broker's disk depending
5523 on its setup.
5524 non-persistent
5526 Delivery mode set to "non-persistent" (1). Messages will stay
5527 in broker's memory unless the broker is under memory pressure.
5528 async
5530 Asynchronous data filling wrapper for input stream.
5531 Fill data in a background thread, to decouple I/O operation from demux
5533 thread.
5534 async:<URL>
5536 async:http://host/resource
5537 async:cache:http://host/resource
5538 bluray
5540 Read BluRay playlist.
5541 The accepted options are:
5543 angle
5545 BluRay angle
5546 chapter
5548 Start chapter (1...N)
5549 playlist
5551 Playlist to read (BDMV/PLAYLIST/?????.mpls)
5552 Examples:
5554 Read longest playlist from BluRay mounted to /mnt/bluray:
5556 bluray:/mnt/bluray
5558 Read angle 2 of playlist 4 from BluRay mounted to /mnt/bluray, start
5560 from chapter 2:
5561 -playlist 4 -angle 2 -chapter 2 bluray:/mnt/bluray
5563 cache
5565 Caching wrapper for input stream.
5566 Cache the input stream to temporary file. It brings seeking capability
5568 to live streams.
5569 The accepted options are:
5571 read_ahead_limit
5573 Amount in bytes that may be read ahead when seeking isn't
5574 supported. Range is -1 to INT_MAX. -1 for unlimited. Default is
5575 65536.
5576 URL Syntax is
5578 cache:<URL>
5580 concat
5582 Physical concatenation protocol.
5583 Read and seek from many resources in sequence as if they were a unique
5585 resource.
5586 A URL accepted by this protocol has the syntax:
5588 concat:<URL1>|<URL2>|...|<URLN>
5590 where URL1, URL2, ..., URLN are the urls of the resource to be
5592 concatenated, each one possibly specifying a distinct protocol.
5593 For example to read a sequence of files split1.mpeg, split2.mpeg,
5595 split3.mpeg with ffplay use the command:
5596 ffplay concat:split1.mpeg\|split2.mpeg\|split3.mpeg
5598 Note that you may need to escape the character "|" which is special for
5600 many shells.
5601 concatf
5603 Physical concatenation protocol using a line break delimited list of
5604 resources.
5605 Read and seek from many resources in sequence as if they were a unique
5607 resource.
5608 A URL accepted by this protocol has the syntax:
5610 concatf:<URL>
5612 where URL is the url containing a line break delimited list of
5614 resources to be concatenated, each one possibly specifying a distinct
5615 protocol. Special characters must be escaped with backslash or single
5616 quotes. See the "Quoting and escaping" section in the ffmpeg-utils(1)
5617 manual.
5618 For example to read a sequence of files split1.mpeg, split2.mpeg,
5620 split3.mpeg listed in separate lines within a file split.txt with
5621 ffplay use the command:
5622 ffplay concatf:split.txt
5624 Where split.txt contains the lines:
5626 split1.mpeg
5628 split2.mpeg
5629 split3.mpeg
5630 crypto
5632 AES-encrypted stream reading protocol.
5633 The accepted options are:
5635 key Set the AES decryption key binary block from given hexadecimal
5637 representation.
5638 iv Set the AES decryption initialization vector binary block from
5640 given hexadecimal representation.
5641 Accepted URL formats:
5643 crypto:<URL>
5645 crypto+<URL>
5646 data
5648 Data in-line in the URI. See
5649 <http://en.wikipedia.org/wiki/Data_URI_scheme>.
5650 For example, to convert a GIF file given inline with ffmpeg:
5652 ffmpeg -i "data:image/gif;base64,R0lGODdhCAAIAMIEAAAAAAAA//8AAP//AP///////////////ywAAAAACAAIAAADF0gEDLojDgdGiJdJqUX02iB4E8Q9jUMkADs=" smiley.png
5654 fd
5656 File descriptor access protocol.
5657 The accepted syntax is:
5659 fd: -fd <file_descriptor>
5661 If fd is not specified, by default the stdout file descriptor will be
5663 used for writing, stdin for reading. Unlike the pipe protocol, fd
5664 protocol has seek support if it corresponding to a regular file. fd
5665 protocol doesn't support pass file descriptor via URL for security.
5666 This protocol accepts the following options:
5668 blocksize
5670 Set I/O operation maximum block size, in bytes. Default value is
5671 "INT_MAX", which results in not limiting the requested block size.
5672 Setting this value reasonably low improves user termination request
5673 reaction time, which is valuable if data transmission is slow.
5674 fd Set file descriptor.
5676 file
5678 File access protocol.
5679 Read from or write to a file.
5681 A file URL can have the form:
5683 file:<filename>
5685 where filename is the path of the file to read.
5687 An URL that does not have a protocol prefix will be assumed to be a
5689 file URL. Depending on the build, an URL that looks like a Windows path
5690 with the drive letter at the beginning will also be assumed to be a
5691 file URL (usually not the case in builds for unix-like systems).
5692 For example to read from a file input.mpeg with ffmpeg use the command:
5694 ffmpeg -i file:input.mpeg output.mpeg
5696 This protocol accepts the following options:
5698 truncate
5700 Truncate existing files on write, if set to 1. A value of 0
5701 prevents truncating. Default value is 1.
5702 blocksize
5704 Set I/O operation maximum block size, in bytes. Default value is
5705 "INT_MAX", which results in not limiting the requested block size.
5706 Setting this value reasonably low improves user termination request
5707 reaction time, which is valuable for files on slow medium.
5708 follow
5710 If set to 1, the protocol will retry reading at the end of the
5711 file, allowing reading files that still are being written. In order
5712 for this to terminate, you either need to use the rw_timeout
5713 option, or use the interrupt callback (for API users).
5714 seekable
5716 Controls if seekability is advertised on the file. 0 means non-
5717 seekable, -1 means auto (seekable for normal files, non-seekable
5718 for named pipes).
5719 Many demuxers handle seekable and non-seekable resources
5721 differently, overriding this might speed up opening certain files
5722 at the cost of losing some features (e.g. accurate seeking).
5723 ftp
5725 FTP (File Transfer Protocol).
5726 Read from or write to remote resources using FTP protocol.
5728 Following syntax is required.
5730 ftp://[user[:password]@]server[:port]/path/to/remote/resource.mpeg
5732 This protocol accepts the following options.
5734 timeout
5736 Set timeout in microseconds of socket I/O operations used by the
5737 underlying low level operation. By default it is set to -1, which
5738 means that the timeout is not specified.
5739 ftp-user
5741 Set a user to be used for authenticating to the FTP server. This is
5742 overridden by the user in the FTP URL.
5743 ftp-password
5745 Set a password to be used for authenticating to the FTP server.
5746 This is overridden by the password in the FTP URL, or by ftp-
5747 anonymous-password if no user is set.
5748 ftp-anonymous-password
5750 Password used when login as anonymous user. Typically an e-mail
5751 address should be used.
5752 ftp-write-seekable
5754 Control seekability of connection during encoding. If set to 1 the
5755 resource is supposed to be seekable, if set to 0 it is assumed not
5756 to be seekable. Default value is 0.
5757 NOTE: Protocol can be used as output, but it is recommended to not do
5759 it, unless special care is taken (tests, customized server
5760 configuration etc.). Different FTP servers behave in different way
5761 during seek operation. ff* tools may produce incomplete content due to
5762 server limitations.
5763 gopher
5765 Gopher protocol.
5766 gophers
5768 Gophers protocol.
5769 The Gopher protocol with TLS encapsulation.
5771 hls
5773 Read Apple HTTP Live Streaming compliant segmented stream as a uniform
5774 one. The M3U8 playlists describing the segments can be remote HTTP
5775 resources or local files, accessed using the standard file protocol.
5776 The nested protocol is declared by specifying "+proto" after the hls
5777 URI scheme name, where proto is either "file" or "http".
5778 hls+http://host/path/to/remote/resource.m3u8
5780 hls+file://path/to/local/resource.m3u8
5781 Using this protocol is discouraged - the hls demuxer should work just
5783 as well (if not, please report the issues) and is more complete. To
5784 use the hls demuxer instead, simply use the direct URLs to the m3u8
5785 files.
5786 http
5788 HTTP (Hyper Text Transfer Protocol).
5789 This protocol accepts the following options:
5791 seekable
5793 Control seekability of connection. If set to 1 the resource is
5794 supposed to be seekable, if set to 0 it is assumed not to be
5795 seekable, if set to -1 it will try to autodetect if it is seekable.
5796 Default value is -1.
5797 chunked_post
5799 If set to 1 use chunked Transfer-Encoding for posts, default is 1.
5800 content_type
5802 Set a specific content type for the POST messages or for listen
5803 mode.
5804 http_proxy
5806 set HTTP proxy to tunnel through e.g. http://example.com:1234
5807 headers
5809 Set custom HTTP headers, can override built in default headers. The
5810 value must be a string encoding the headers.
5811 multiple_requests
5813 Use persistent connections if set to 1, default is 0.
5814 post_data
5816 Set custom HTTP post data.
5817 referer
5819 Set the Referer header. Include 'Referer: URL' header in HTTP
5820 request.
5821 user_agent
5823 Override the User-Agent header. If not specified the protocol will
5824 use a string describing the libavformat build. ("Lavf/<version>")
5825 reconnect_at_eof
5827 If set then eof is treated like an error and causes reconnection,
5828 this is useful for live / endless streams.
5829 reconnect_streamed
5831 If set then even streamed/non seekable streams will be reconnected
5832 on errors.
5833 reconnect_on_network_error
5835 Reconnect automatically in case of TCP/TLS errors during connect.
5836 reconnect_on_http_error
5838 A comma separated list of HTTP status codes to reconnect on. The
5839 list can include specific status codes (e.g. '503') or the strings
5840 '4xx' / '5xx'.
5841 reconnect_delay_max
5843 Sets the maximum delay in seconds after which to give up
5844 reconnecting
5845 mime_type
5847 Export the MIME type.
5848 http_version
5850 Exports the HTTP response version number. Usually "1.0" or "1.1".
5851 icy If set to 1 request ICY (SHOUTcast) metadata from the server. If
5853 the server supports this, the metadata has to be retrieved by the
5854 application by reading the icy_metadata_headers and
5855 icy_metadata_packet options. The default is 1.
5856 icy_metadata_headers
5858 If the server supports ICY metadata, this contains the ICY-specific
5859 HTTP reply headers, separated by newline characters.
5860 icy_metadata_packet
5862 If the server supports ICY metadata, and icy was set to 1, this
5863 contains the last non-empty metadata packet sent by the server. It
5864 should be polled in regular intervals by applications interested in
5865 mid-stream metadata updates.
5866 cookies
5868 Set the cookies to be sent in future requests. The format of each
5869 cookie is the same as the value of a Set-Cookie HTTP response
5870 field. Multiple cookies can be delimited by a newline character.
5871 offset
5873 Set initial byte offset.
5874 end_offset
5876 Try to limit the request to bytes preceding this offset.
5877 method
5879 When used as a client option it sets the HTTP method for the
5880 request.
5881 When used as a server option it sets the HTTP method that is going
5883 to be expected from the client(s). If the expected and the
5884 received HTTP method do not match the client will be given a Bad
5885 Request response. When unset the HTTP method is not checked for
5886 now. This will be replaced by autodetection in the future.
5887 listen
5889 If set to 1 enables experimental HTTP server. This can be used to
5890 send data when used as an output option, or read data from a client
5891 with HTTP POST when used as an input option. If set to 2 enables
5892 experimental multi-client HTTP server. This is not yet implemented
5893 in ffmpeg.c and thus must not be used as a command line option.
5894 # Server side (sending):
5896 ffmpeg -i somefile.ogg -c copy -listen 1 -f ogg http://<server>:<port>
5897 # Client side (receiving):
5899 ffmpeg -i http://<server>:<port> -c copy somefile.ogg
5900 # Client can also be done with wget:
5902 wget http://<server>:<port> -O somefile.ogg
5903 # Server side (receiving):
5905 ffmpeg -listen 1 -i http://<server>:<port> -c copy somefile.ogg
5906 # Client side (sending):
5908 ffmpeg -i somefile.ogg -chunked_post 0 -c copy -f ogg http://<server>:<port>
5909 # Client can also be done with wget:
5911 wget --post-file=somefile.ogg http://<server>:<port>
5912 send_expect_100
5914 Send an Expect: 100-continue header for POST. If set to 1 it will
5915 send, if set to 0 it won't, if set to -1 it will try to send if it
5916 is applicable. Default value is -1.
5917 auth_type
5919 Set HTTP authentication type. No option for Digest, since this
5920 method requires getting nonce parameters from the server first and
5921 can't be used straight away like Basic.
5922 none
5924 Choose the HTTP authentication type automatically. This is the
5925 default.
5926 basic
5928 Choose the HTTP basic authentication.
5929 Basic authentication sends a Base64-encoded string that
5931 contains a user name and password for the client. Base64 is not
5932 a form of encryption and should be considered the same as
5933 sending the user name and password in clear text (Base64 is a
5934 reversible encoding). If a resource needs to be protected,
5935 strongly consider using an authentication scheme other than
5936 basic authentication. HTTPS/TLS should be used with basic
5937 authentication. Without these additional security
5938 enhancements, basic authentication should not be used to
5939 protect sensitive or valuable information.
5940 HTTP Cookies
5942 Some HTTP requests will be denied unless cookie values are passed in
5944 with the request. The cookies option allows these cookies to be
5945 specified. At the very least, each cookie must specify a value along
5946 with a path and domain. HTTP requests that match both the domain and
5947 path will automatically include the cookie value in the HTTP Cookie
5948 header field. Multiple cookies can be delimited by a newline.
5949 The required syntax to play a stream specifying a cookie is:
5951 ffplay -cookies "nlqptid=nltid=tsn; path=/; domain=somedomain.com;" http://somedomain.com/somestream.m3u8
5953 Icecast
5955 Icecast protocol (stream to Icecast servers)
5956 This protocol accepts the following options:
5958 ice_genre
5960 Set the stream genre.
5961 ice_name
5963 Set the stream name.
5964 ice_description
5966 Set the stream description.
5967 ice_url
5969 Set the stream website URL.
5970 ice_public
5972 Set if the stream should be public. The default is 0 (not public).
5973 user_agent
5975 Override the User-Agent header. If not specified a string of the
5976 form "Lavf/<version>" will be used.
5977 password
5979 Set the Icecast mountpoint password.
5980 content_type
5982 Set the stream content type. This must be set if it is different
5983 from audio/mpeg.
5984 legacy_icecast
5986 This enables support for Icecast versions < 2.4.0, that do not
5987 support the HTTP PUT method but the SOURCE method.
5988 tls Establish a TLS (HTTPS) connection to Icecast.
5990 icecast://[<username>[:<password>]@]<server>:<port>/<mountpoint>
5992 ipfs
5994 InterPlanetary File System (IPFS) protocol support. One can access
5995 files stored on the IPFS network through so-called gateways. These are
5996 http(s) endpoints. This protocol wraps the IPFS native protocols
5997 (ipfs:// and ipns://) to be sent to such a gateway. Users can (and
5998 should) host their own node which means this protocol will use one's
5999 local gateway to access files on the IPFS network.
6000 This protocol accepts the following options:
6002 gateway
6004 Defines the gateway to use. When not set, the protocol will first
6005 try locating the local gateway by looking at $IPFS_GATEWAY,
6006 $IPFS_PATH and "$HOME/.ipfs/", in that order.
6007 One can use this protocol in 2 ways. Using IPFS:
6009 ffplay ipfs://<hash>
6011 Or the IPNS protocol (IPNS is mutable IPFS):
6013 ffplay ipns://<hash>
6015 mmst
6017 MMS (Microsoft Media Server) protocol over TCP.
6018 mmsh
6020 MMS (Microsoft Media Server) protocol over HTTP.
6021 The required syntax is:
6023 mmsh://<server>[:<port>][/<app>][/<playpath>]
6025 md5
6027 MD5 output protocol.
6028 Computes the MD5 hash of the data to be written, and on close writes
6030 this to the designated output or stdout if none is specified. It can be
6031 used to test muxers without writing an actual file.
6032 Some examples follow.
6034 # Write the MD5 hash of the encoded AVI file to the file output.avi.md5.
6036 ffmpeg -i input.flv -f avi -y md5:output.avi.md5
6037 # Write the MD5 hash of the encoded AVI file to stdout.
6039 ffmpeg -i input.flv -f avi -y md5:
6040 Note that some formats (typically MOV) require the output protocol to
6042 be seekable, so they will fail with the MD5 output protocol.
6043 pipe
6045 UNIX pipe access protocol.
6046 Read and write from UNIX pipes.
6048 The accepted syntax is:
6050 pipe:[<number>]
6052 If fd isn't specified, number is the number corresponding to the file
6054 descriptor of the pipe (e.g. 0 for stdin, 1 for stdout, 2 for stderr).
6055 If number is not specified, by default the stdout file descriptor will
6056 be used for writing, stdin for reading.
6057 For example to read from stdin with ffmpeg:
6059 cat test.wav | ffmpeg -i pipe:0
6061 # ...this is the same as...
6062 cat test.wav | ffmpeg -i pipe:
6063 For writing to stdout with ffmpeg:
6065 ffmpeg -i test.wav -f avi pipe:1 | cat > test.avi
6067 # ...this is the same as...
6068 ffmpeg -i test.wav -f avi pipe: | cat > test.avi
6069 This protocol accepts the following options:
6071 blocksize
6073 Set I/O operation maximum block size, in bytes. Default value is
6074 "INT_MAX", which results in not limiting the requested block size.
6075 Setting this value reasonably low improves user termination request
6076 reaction time, which is valuable if data transmission is slow.
6077 fd Set file descriptor.
6079 Note that some formats (typically MOV), require the output protocol to
6081 be seekable, so they will fail with the pipe output protocol.
6082 prompeg
6084 Pro-MPEG Code of Practice #3 Release 2 FEC protocol.
6085 The Pro-MPEG CoP#3 FEC is a 2D parity-check forward error correction
6087 mechanism for MPEG-2 Transport Streams sent over RTP.
6088 This protocol must be used in conjunction with the "rtp_mpegts" muxer
6090 and the "rtp" protocol.
6091 The required syntax is:
6093 -f rtp_mpegts -fec prompeg=<option>=<val>... rtp://<hostname>:<port>
6095 The destination UDP ports are "port + 2" for the column FEC stream and
6097 "port + 4" for the row FEC stream.
6098 This protocol accepts the following options:
6100 l=n The number of columns (4-20, LxD <= 100)
6102 d=n The number of rows (4-20, LxD <= 100)
6104 Example usage:
6106 -f rtp_mpegts -fec prompeg=l=8:d=4 rtp://<hostname>:<port>
6108 rist
6110 Reliable Internet Streaming Transport protocol
6111 The accepted options are:
6113 rist_profile
6115 Supported values:
6116 simple
6118 main
6119 This one is default.
6120 advanced
6122 buffer_size
6123 Set internal RIST buffer size in milliseconds for retransmission of
6124 data. Default value is 0 which means the librist default (1 sec).
6125 Maximum value is 30 seconds.
6126 fifo_size
6128 Size of the librist receiver output fifo in number of packets. This
6129 must be a power of 2. Defaults to 8192 (vs the librist default of
6130 1024).
6131 overrun_nonfatal=1|0
6133 Survive in case of librist fifo buffer overrun. Default value is 0.
6134 pkt_size
6136 Set maximum packet size for sending data. 1316 by default.
6137 log_level
6139 Set loglevel for RIST logging messages. You only need to set this
6140 if you explicitly want to enable debug level messages or packet
6141 loss simulation, otherwise the regular loglevel is respected.
6142 secret
6144 Set override of encryption secret, by default is unset.
6145 encryption
6147 Set encryption type, by default is disabled. Acceptable values are
6148 128 and 256.
6149 rtmp
6151 Real-Time Messaging Protocol.
6152 The Real-Time Messaging Protocol (RTMP) is used for streaming
6154 multimedia content across a TCP/IP network.
6155 The required syntax is:
6157 rtmp://[<username>:<password>@]<server>[:<port>][/<app>][/<instance>][/<playpath>]
6159 The accepted parameters are:
6161 username
6163 An optional username (mostly for publishing).
6164 password
6166 An optional password (mostly for publishing).
6167 server
6169 The address of the RTMP server.
6170 port
6172 The number of the TCP port to use (by default is 1935).
6173 app It is the name of the application to access. It usually corresponds
6175 to the path where the application is installed on the RTMP server
6176 (e.g. /ondemand/, /flash/live/, etc.). You can override the value
6177 parsed from the URI through the "rtmp_app" option, too.
6178 playpath
6180 It is the path or name of the resource to play with reference to
6181 the application specified in app, may be prefixed by "mp4:". You
6182 can override the value parsed from the URI through the
6183 "rtmp_playpath" option, too.
6184 listen
6186 Act as a server, listening for an incoming connection.
6187 timeout
6189 Maximum time to wait for the incoming connection. Implies listen.
6190 Additionally, the following parameters can be set via command line
6192 options (or in code via "AVOption"s):
6193 rtmp_app
6195 Name of application to connect on the RTMP server. This option
6196 overrides the parameter specified in the URI.
6197 rtmp_buffer
6199 Set the client buffer time in milliseconds. The default is 3000.
6200 rtmp_conn
6202 Extra arbitrary AMF connection parameters, parsed from a string,
6203 e.g. like "B:1 S:authMe O:1 NN:code:1.23 NS:flag:ok O:0". Each
6204 value is prefixed by a single character denoting the type, B for
6205 Boolean, N for number, S for string, O for object, or Z for null,
6206 followed by a colon. For Booleans the data must be either 0 or 1
6207 for FALSE or TRUE, respectively. Likewise for Objects the data
6208 must be 0 or 1 to end or begin an object, respectively. Data items
6209 in subobjects may be named, by prefixing the type with 'N' and
6210 specifying the name before the value (i.e. "NB:myFlag:1"). This
6211 option may be used multiple times to construct arbitrary AMF
6212 sequences.
6213 rtmp_flashver
6215 Version of the Flash plugin used to run the SWF player. The default
6216 is LNX 9,0,124,2. (When publishing, the default is FMLE/3.0
6217 (compatible; <libavformat version>).)
6218 rtmp_flush_interval
6220 Number of packets flushed in the same request (RTMPT only). The
6221 default is 10.
6222 rtmp_live
6224 Specify that the media is a live stream. No resuming or seeking in
6225 live streams is possible. The default value is "any", which means
6226 the subscriber first tries to play the live stream specified in the
6227 playpath. If a live stream of that name is not found, it plays the
6228 recorded stream. The other possible values are "live" and
6229 "recorded".
6230 rtmp_pageurl
6232 URL of the web page in which the media was embedded. By default no
6233 value will be sent.
6234 rtmp_playpath
6236 Stream identifier to play or to publish. This option overrides the
6237 parameter specified in the URI.
6238 rtmp_subscribe
6240 Name of live stream to subscribe to. By default no value will be
6241 sent. It is only sent if the option is specified or if rtmp_live
6242 is set to live.
6243 rtmp_swfhash
6245 SHA256 hash of the decompressed SWF file (32 bytes).
6246 rtmp_swfsize
6248 Size of the decompressed SWF file, required for SWFVerification.
6249 rtmp_swfurl
6251 URL of the SWF player for the media. By default no value will be
6252 sent.
6253 rtmp_swfverify
6255 URL to player swf file, compute hash/size automatically.
6256 rtmp_tcurl
6258 URL of the target stream. Defaults to proto://host[:port]/app.
6259 tcp_nodelay=1|0
6261 Set TCP_NODELAY to disable Nagle's algorithm. Default value is 0.
6262 Remark: Writing to the socket is currently not optimized to
6264 minimize system calls and reduces the efficiency / effect of
6265 TCP_NODELAY.
6266 For example to read with ffplay a multimedia resource named "sample"
6268 from the application "vod" from an RTMP server "myserver":
6269 ffplay rtmp://myserver/vod/sample
6271 To publish to a password protected server, passing the playpath and app
6273 names separately:
6274 ffmpeg -re -i <input> -f flv -rtmp_playpath some/long/path -rtmp_app long/app/name rtmp://username:password@myserver/
6276 rtmpe
6278 Encrypted Real-Time Messaging Protocol.
6279 The Encrypted Real-Time Messaging Protocol (RTMPE) is used for
6281 streaming multimedia content within standard cryptographic primitives,
6282 consisting of Diffie-Hellman key exchange and HMACSHA256, generating a
6283 pair of RC4 keys.
6284 rtmps
6286 Real-Time Messaging Protocol over a secure SSL connection.
6287 The Real-Time Messaging Protocol (RTMPS) is used for streaming
6289 multimedia content across an encrypted connection.
6290 rtmpt
6292 Real-Time Messaging Protocol tunneled through HTTP.
6293 The Real-Time Messaging Protocol tunneled through HTTP (RTMPT) is used
6295 for streaming multimedia content within HTTP requests to traverse
6296 firewalls.
6297 rtmpte
6299 Encrypted Real-Time Messaging Protocol tunneled through HTTP.
6300 The Encrypted Real-Time Messaging Protocol tunneled through HTTP
6302 (RTMPTE) is used for streaming multimedia content within HTTP requests
6303 to traverse firewalls.
6304 rtmpts
6306 Real-Time Messaging Protocol tunneled through HTTPS.
6307 The Real-Time Messaging Protocol tunneled through HTTPS (RTMPTS) is
6309 used for streaming multimedia content within HTTPS requests to traverse
6310 firewalls.
6311 libsmbclient
6313 libsmbclient permits one to manipulate CIFS/SMB network resources.
6314 Following syntax is required.
6316 smb://[[domain:]user[:password@]]server[/share[/path[/file]]]
6318 This protocol accepts the following options.
6320 timeout
6322 Set timeout in milliseconds of socket I/O operations used by the
6323 underlying low level operation. By default it is set to -1, which
6324 means that the timeout is not specified.
6325 truncate
6327 Truncate existing files on write, if set to 1. A value of 0
6328 prevents truncating. Default value is 1.
6329 workgroup
6331 Set the workgroup used for making connections. By default workgroup
6332 is not specified.
6333 For more information see: <http://www.samba.org/>.
6335 libssh
6337 Secure File Transfer Protocol via libssh
6338 Read from or write to remote resources using SFTP protocol.
6340 Following syntax is required.
6342 sftp://[user[:password]@]server[:port]/path/to/remote/resource.mpeg
6344 This protocol accepts the following options.
6346 timeout
6348 Set timeout of socket I/O operations used by the underlying low
6349 level operation. By default it is set to -1, which means that the
6350 timeout is not specified.
6351 truncate
6353 Truncate existing files on write, if set to 1. A value of 0
6354 prevents truncating. Default value is 1.
6355 private_key
6357 Specify the path of the file containing private key to use during
6358 authorization. By default libssh searches for keys in the ~/.ssh/
6359 directory.
6360 Example: Play a file stored on remote server.
6362 ffplay sftp://user:password@server_address:22/home/user/resource.mpeg
6364 librtmp rtmp, rtmpe, rtmps, rtmpt, rtmpte
6366 Real-Time Messaging Protocol and its variants supported through
6367 librtmp.
6368 Requires the presence of the librtmp headers and library during
6370 configuration. You need to explicitly configure the build with
6371 "--enable-librtmp". If enabled this will replace the native RTMP
6372 protocol.
6373 This protocol provides most client functions and a few server functions
6375 needed to support RTMP, RTMP tunneled in HTTP (RTMPT), encrypted RTMP
6376 (RTMPE), RTMP over SSL/TLS (RTMPS) and tunneled variants of these
6377 encrypted types (RTMPTE, RTMPTS).
6378 The required syntax is:
6380 <rtmp_proto>://<server>[:<port>][/<app>][/<playpath>] <options>
6382 where rtmp_proto is one of the strings "rtmp", "rtmpt", "rtmpe",
6384 "rtmps", "rtmpte", "rtmpts" corresponding to each RTMP variant, and
6385 server, port, app and playpath have the same meaning as specified for
6386 the RTMP native protocol. options contains a list of space-separated
6387 options of the form key=val.
6388 See the librtmp manual page (man 3 librtmp) for more information.
6390 For example, to stream a file in real-time to an RTMP server using
6392 ffmpeg:
6393 ffmpeg -re -i myfile -f flv rtmp://myserver/live/mystream
6395 To play the same stream using ffplay:
6397 ffplay "rtmp://myserver/live/mystream live=1"
6399 rtp
6401 Real-time Transport Protocol.
6402 The required syntax for an RTP URL is:
6404 rtp://hostname[:port][?option=val...]
6405 port specifies the RTP port to use.
6407 The following URL options are supported:
6409 ttl=n
6411 Set the TTL (Time-To-Live) value (for multicast only).
6412 rtcpport=n
6414 Set the remote RTCP port to n.
6415 localrtpport=n
6417 Set the local RTP port to n.
6418 localrtcpport=n'
6420 Set the local RTCP port to n.
6421 pkt_size=n
6423 Set max packet size (in bytes) to n.
6424 buffer_size=size
6426 Set the maximum UDP socket buffer size in bytes.
6427 connect=0|1
6429 Do a connect() on the UDP socket (if set to 1) or not (if set to
6430 0).
6431 sources=ip[,ip]
6433 List allowed source IP addresses.
6434 block=ip[,ip]
6436 List disallowed (blocked) source IP addresses.
6437 write_to_source=0|1
6439 Send packets to the source address of the latest received packet
6440 (if set to 1) or to a default remote address (if set to 0).
6441 localport=n
6443 Set the local RTP port to n.
6444 localaddr=addr
6446 Local IP address of a network interface used for sending packets or
6447 joining multicast groups.
6448 timeout=n
6450 Set timeout (in microseconds) of socket I/O operations to n.
6451 This is a deprecated option. Instead, localrtpport should be used.
6453 Important notes:
6455 1. If rtcpport is not set the RTCP port will be set to the RTP port
6457 value plus 1.
6458 2. If localrtpport (the local RTP port) is not set any available port
6460 will be used for the local RTP and RTCP ports.
6461 3. If localrtcpport (the local RTCP port) is not set it will be set to
6463 the local RTP port value plus 1.
6464 rtsp
6466 Real-Time Streaming Protocol.
6467 RTSP is not technically a protocol handler in libavformat, it is a
6469 demuxer and muxer. The demuxer supports both normal RTSP (with data
6470 transferred over RTP; this is used by e.g. Apple and Microsoft) and
6471 Real-RTSP (with data transferred over RDT).
6472 The muxer can be used to send a stream using RTSP ANNOUNCE to a server
6474 supporting it (currently Darwin Streaming Server and Mischa
6475 Spiegelmock's <https://github.com/revmischa/rtsp-server>).
6476 The required syntax for a RTSP url is:
6478 rtsp://<hostname>[:<port>]/<path>
6480 Options can be set on the ffmpeg/ffplay command line, or set in code
6482 via "AVOption"s or in "avformat_open_input".
6483 Muxer
6485 The following options are supported.
6487 rtsp_transport
6489 Set RTSP transport protocols.
6490 It accepts the following values:
6492 udp Use UDP as lower transport protocol.
6494 tcp Use TCP (interleaving within the RTSP control channel) as lower
6496 transport protocol.
6497 Default value is 0.
6499 rtsp_flags
6501 Set RTSP flags.
6502 The following values are accepted:
6504 latm
6506 Use MP4A-LATM packetization instead of MPEG4-GENERIC for AAC.
6507 rfc2190
6509 Use RFC 2190 packetization instead of RFC 4629 for H.263.
6510 skip_rtcp
6512 Don't send RTCP sender reports.
6513 h264_mode0
6515 Use mode 0 for H.264 in RTP.
6516 send_bye
6518 Send RTCP BYE packets when finishing.
6519 Default value is 0.
6521 min_port
6523 Set minimum local UDP port. Default value is 5000.
6524 max_port
6526 Set maximum local UDP port. Default value is 65000.
6527 buffer_size
6529 Set the maximum socket buffer size in bytes.
6530 pkt_size
6532 Set max send packet size (in bytes). Default value is 1472.
6533 Demuxer
6535 The following options are supported.
6537 initial_pause
6539 Do not start playing the stream immediately if set to 1. Default
6540 value is 0.
6541 rtsp_transport
6543 Set RTSP transport protocols.
6544 It accepts the following values:
6546 udp Use UDP as lower transport protocol.
6548 tcp Use TCP (interleaving within the RTSP control channel) as lower
6550 transport protocol.
6551 udp_multicast
6553 Use UDP multicast as lower transport protocol.
6554 http
6556 Use HTTP tunneling as lower transport protocol, which is useful
6557 for passing proxies.
6558 https
6560 Use HTTPs tunneling as lower transport protocol, which is
6561 useful for passing proxies and widely used for security
6562 consideration.
6563 Multiple lower transport protocols may be specified, in that case
6565 they are tried one at a time (if the setup of one fails, the next
6566 one is tried). For the muxer, only the tcp and udp options are
6567 supported.
6568 rtsp_flags
6570 Set RTSP flags.
6571 The following values are accepted:
6573 filter_src
6575 Accept packets only from negotiated peer address and port.
6576 listen
6578 Act as a server, listening for an incoming connection.
6579 prefer_tcp
6581 Try TCP for RTP transport first, if TCP is available as RTSP
6582 RTP transport.
6583 satip_raw
6585 Export raw MPEG-TS stream instead of demuxing. The flag will
6586 simply write out the raw stream, with the original PAT/PMT/PIDs
6587 intact.
6588 Default value is none.
6590 allowed_media_types
6592 Set media types to accept from the server.
6593 The following flags are accepted:
6595 video
6597 audio
6598 data
6599 subtitle
6600 By default it accepts all media types.
6602 min_port
6604 Set minimum local UDP port. Default value is 5000.
6605 max_port
6607 Set maximum local UDP port. Default value is 65000.
6608 listen_timeout
6610 Set maximum timeout (in seconds) to establish an initial
6611 connection. Setting listen_timeout > 0 sets rtsp_flags to listen.
6612 Default is -1 which means an infinite timeout when listen mode is
6613 set.
6614 reorder_queue_size
6616 Set number of packets to buffer for handling of reordered packets.
6617 timeout
6619 Set socket TCP I/O timeout in microseconds.
6620 user_agent
6622 Override User-Agent header. If not specified, it defaults to the
6623 libavformat identifier string.
6624 buffer_size
6626 Set the maximum socket buffer size in bytes.
6627 When receiving data over UDP, the demuxer tries to reorder received
6629 packets (since they may arrive out of order, or packets may get lost
6630 totally). This can be disabled by setting the maximum demuxing delay to
6631 zero (via the "max_delay" field of AVFormatContext).
6632 When watching multi-bitrate Real-RTSP streams with ffplay, the streams
6634 to display can be chosen with "-vst" n and "-ast" n for video and audio
6635 respectively, and can be switched on the fly by pressing "v" and "a".
6636 Examples
6638 The following examples all make use of the ffplay and ffmpeg tools.
6640 • Watch a stream over UDP, with a max reordering delay of 0.5
6642 seconds:
6643 ffplay -max_delay 500000 -rtsp_transport udp rtsp://server/video.mp4
6645 • Watch a stream tunneled over HTTP:
6647 ffplay -rtsp_transport http rtsp://server/video.mp4
6649 • Send a stream in realtime to a RTSP server, for others to watch:
6651 ffmpeg -re -i <input> -f rtsp -muxdelay 0.1 rtsp://server/live.sdp
6653 • Receive a stream in realtime:
6655 ffmpeg -rtsp_flags listen -i rtsp://ownaddress/live.sdp <output>
6657 sap
6659 Session Announcement Protocol (RFC 2974). This is not technically a
6660 protocol handler in libavformat, it is a muxer and demuxer. It is used
6661 for signalling of RTP streams, by announcing the SDP for the streams
6662 regularly on a separate port.
6663 Muxer
6665 The syntax for a SAP url given to the muxer is:
6667 sap://<destination>[:<port>][?<options>]
6669 The RTP packets are sent to destination on port port, or to port 5004
6671 if no port is specified. options is a "&"-separated list. The
6672 following options are supported:
6673 announce_addr=address
6675 Specify the destination IP address for sending the announcements
6676 to. If omitted, the announcements are sent to the commonly used
6677 SAP announcement multicast address 224.2.127.254 (sap.mcast.net),
6678 or ff0e::2:7ffe if destination is an IPv6 address.
6679 announce_port=port
6681 Specify the port to send the announcements on, defaults to 9875 if
6682 not specified.
6683 ttl=ttl
6685 Specify the time to live value for the announcements and RTP
6686 packets, defaults to 255.
6687 same_port=0|1
6689 If set to 1, send all RTP streams on the same port pair. If zero
6690 (the default), all streams are sent on unique ports, with each
6691 stream on a port 2 numbers higher than the previous. VLC/Live555
6692 requires this to be set to 1, to be able to receive the stream.
6693 The RTP stack in libavformat for receiving requires all streams to
6694 be sent on unique ports.
6695 Example command lines follow.
6697 To broadcast a stream on the local subnet, for watching in VLC:
6699 ffmpeg -re -i <input> -f sap sap://224.0.0.255?same_port=1
6701 Similarly, for watching in ffplay:
6703 ffmpeg -re -i <input> -f sap sap://224.0.0.255
6705 And for watching in ffplay, over IPv6:
6707 ffmpeg -re -i <input> -f sap sap://[ff0e::1:2:3:4]
6709 Demuxer
6711 The syntax for a SAP url given to the demuxer is:
6713 sap://[<address>][:<port>]
6715 address is the multicast address to listen for announcements on, if
6717 omitted, the default 224.2.127.254 (sap.mcast.net) is used. port is the
6718 port that is listened on, 9875 if omitted.
6719 The demuxers listens for announcements on the given address and port.
6721 Once an announcement is received, it tries to receive that particular
6722 stream.
6723 Example command lines follow.
6725 To play back the first stream announced on the normal SAP multicast
6727 address:
6728 ffplay sap://
6730 To play back the first stream announced on one the default IPv6 SAP
6732 multicast address:
6733 ffplay sap://[ff0e::2:7ffe]
6735 sctp
6737 Stream Control Transmission Protocol.
6738 The accepted URL syntax is:
6740 sctp://<host>:<port>[?<options>]
6742 The protocol accepts the following options:
6744 listen
6746 If set to any value, listen for an incoming connection. Outgoing
6747 connection is done by default.
6748 max_streams
6750 Set the maximum number of streams. By default no limit is set.
6751 srt
6753 Haivision Secure Reliable Transport Protocol via libsrt.
6754 The supported syntax for a SRT URL is:
6756 srt://<hostname>:<port>[?<options>]
6758 options contains a list of &-separated options of the form key=val.
6760 or
6762 <options> srt://<hostname>:<port>
6764 options contains a list of '-key val' options.
6766 This protocol accepts the following options.
6768 connect_timeout=milliseconds
6770 Connection timeout; SRT cannot connect for RTT > 1500 msec (2
6771 handshake exchanges) with the default connect timeout of 3 seconds.
6772 This option applies to the caller and rendezvous connection modes.
6773 The connect timeout is 10 times the value set for the rendezvous
6774 mode (which can be used as a workaround for this connection problem
6775 with earlier versions).
6776 ffs=bytes
6778 Flight Flag Size (Window Size), in bytes. FFS is actually an
6779 internal parameter and you should set it to not less than
6780 recv_buffer_size and mss. The default value is relatively large,
6781 therefore unless you set a very large receiver buffer, you do not
6782 need to change this option. Default value is 25600.
6783 inputbw=bytes/seconds
6785 Sender nominal input rate, in bytes per seconds. Used along with
6786 oheadbw, when maxbw is set to relative (0), to calculate maximum
6787 sending rate when recovery packets are sent along with the main
6788 media stream: inputbw * (100 + oheadbw) / 100 if inputbw is not set
6789 while maxbw is set to relative (0), the actual input rate is
6790 evaluated inside the library. Default value is 0.
6791 iptos=tos
6793 IP Type of Service. Applies to sender only. Default value is 0xB8.
6794 ipttl=ttl
6796 IP Time To Live. Applies to sender only. Default value is 64.
6797 latency=microseconds
6799 Timestamp-based Packet Delivery Delay. Used to absorb bursts of
6800 missed packet retransmissions. This flag sets both rcvlatency and
6801 peerlatency to the same value. Note that prior to version 1.3.0
6802 this is the only flag to set the latency, however this is
6803 effectively equivalent to setting peerlatency, when side is sender
6804 and rcvlatency when side is receiver, and the bidirectional stream
6805 sending is not supported.
6806 listen_timeout=microseconds
6808 Set socket listen timeout.
6809 maxbw=bytes/seconds
6811 Maximum sending bandwidth, in bytes per seconds. -1 infinite
6812 (CSRTCC limit is 30mbps) 0 relative to input rate (see inputbw) >0
6813 absolute limit value Default value is 0 (relative)
6814 mode=caller|listener|rendezvous
6816 Connection mode. caller opens client connection. listener starts
6817 server to listen for incoming connections. rendezvous use Rendez-
6818 Vous connection mode. Default value is caller.
6819 mss=bytes
6821 Maximum Segment Size, in bytes. Used for buffer allocation and rate
6822 calculation using a packet counter assuming fully filled packets.
6823 The smallest MSS between the peers is used. This is 1500 by default
6824 in the overall internet. This is the maximum size of the UDP
6825 packet and can be only decreased, unless you have some unusual
6826 dedicated network settings. Default value is 1500.
6827 nakreport=1|0
6829 If set to 1, Receiver will send `UMSG_LOSSREPORT` messages
6830 periodically until a lost packet is retransmitted or intentionally
6831 dropped. Default value is 1.
6832 oheadbw=percents
6834 Recovery bandwidth overhead above input rate, in percents. See
6835 inputbw. Default value is 25%.
6836 passphrase=string
6838 HaiCrypt Encryption/Decryption Passphrase string, length from 10 to
6839 79 characters. The passphrase is the shared secret between the
6840 sender and the receiver. It is used to generate the Key Encrypting
6841 Key using PBKDF2 (Password-Based Key Derivation Function). It is
6842 used only if pbkeylen is non-zero. It is used on the receiver only
6843 if the received data is encrypted. The configured passphrase
6844 cannot be recovered (write-only).
6845 enforced_encryption=1|0
6847 If true, both connection parties must have the same password set
6848 (including empty, that is, with no encryption). If the password
6849 doesn't match or only one side is unencrypted, the connection is
6850 rejected. Default is true.
6851 kmrefreshrate=packets
6853 The number of packets to be transmitted after which the encryption
6854 key is switched to a new key. Default is -1. -1 means auto
6855 (0x1000000 in srt library). The range for this option is integers
6856 in the 0 - "INT_MAX".
6857 kmpreannounce=packets
6859 The interval between when a new encryption key is sent and when
6860 switchover occurs. This value also applies to the subsequent
6861 interval between when switchover occurs and when the old encryption
6862 key is decommissioned. Default is -1. -1 means auto (0x1000 in srt
6863 library). The range for this option is integers in the 0 -
6864 "INT_MAX".
6865 snddropdelay=microseconds
6867 The sender's extra delay before dropping packets. This delay is
6868 added to the default drop delay time interval value.
6869 Special value -1: Do not drop packets on the sender at all.
6871 payload_size=bytes
6873 Sets the maximum declared size of a packet transferred during the
6874 single call to the sending function in Live mode. Use 0 if this
6875 value isn't used (which is default in file mode). Default is -1
6876 (automatic), which typically means MPEG-TS; if you are going to use
6877 SRT to send any different kind of payload, such as, for example,
6878 wrapping a live stream in very small frames, then you can use a
6879 bigger maximum frame size, though not greater than 1456 bytes.
6880 pkt_size=bytes
6882 Alias for payload_size.
6883 peerlatency=microseconds
6885 The latency value (as described in rcvlatency) that is set by the
6886 sender side as a minimum value for the receiver.
6887 pbkeylen=bytes
6889 Sender encryption key length, in bytes. Only can be set to 0, 16,
6890 24 and 32. Enable sender encryption if not 0. Not required on
6891 receiver (set to 0), key size obtained from sender in HaiCrypt
6892 handshake. Default value is 0.
6893 rcvlatency=microseconds
6895 The time that should elapse since the moment when the packet was
6896 sent and the moment when it's delivered to the receiver application
6897 in the receiving function. This time should be a buffer time large
6898 enough to cover the time spent for sending, unexpectedly extended
6899 RTT time, and the time needed to retransmit the lost UDP packet.
6900 The effective latency value will be the maximum of this options'
6901 value and the value of peerlatency set by the peer side. Before
6902 version 1.3.0 this option is only available as latency.
6903 recv_buffer_size=bytes
6905 Set UDP receive buffer size, expressed in bytes.
6906 send_buffer_size=bytes
6908 Set UDP send buffer size, expressed in bytes.
6909 timeout=microseconds
6911 Set raise error timeouts for read, write and connect operations.
6912 Note that the SRT library has internal timeouts which can be
6913 controlled separately, the value set here is only a cap on those.
6914 tlpktdrop=1|0
6916 Too-late Packet Drop. When enabled on receiver, it skips missing
6917 packets that have not been delivered in time and delivers the
6918 following packets to the application when their time-to-play has
6919 come. It also sends a fake ACK to the sender. When enabled on
6920 sender and enabled on the receiving peer, the sender drops the
6921 older packets that have no chance of being delivered in time. It
6922 was automatically enabled in the sender if the receiver supports
6923 it.
6924 sndbuf=bytes
6926 Set send buffer size, expressed in bytes.
6927 rcvbuf=bytes
6929 Set receive buffer size, expressed in bytes.
6930 Receive buffer must not be greater than ffs.
6932 lossmaxttl=packets
6934 The value up to which the Reorder Tolerance may grow. When Reorder
6935 Tolerance is > 0, then packet loss report is delayed until that
6936 number of packets come in. Reorder Tolerance increases every time a
6937 "belated" packet has come, but it wasn't due to retransmission
6938 (that is, when UDP packets tend to come out of order), with the
6939 difference between the latest sequence and this packet's sequence,
6940 and not more than the value of this option. By default it's 0,
6941 which means that this mechanism is turned off, and the loss report
6942 is always sent immediately upon experiencing a "gap" in sequences.
6943 minversion
6945 The minimum SRT version that is required from the peer. A
6946 connection to a peer that does not satisfy the minimum version
6947 requirement will be rejected.
6948 The version format in hex is 0xXXYYZZ for x.y.z in human readable
6950 form.
6951 streamid=string
6953 A string limited to 512 characters that can be set on the socket
6954 prior to connecting. This stream ID will be able to be retrieved by
6955 the listener side from the socket that is returned from srt_accept
6956 and was connected by a socket with that set stream ID. SRT does not
6957 enforce any special interpretation of the contents of this string.
6958 This option doesn’t make sense in Rendezvous connection; the result
6959 might be that simply one side will override the value from the
6960 other side and it’s the matter of luck which one would win
6961 srt_streamid=string
6963 Alias for streamid to avoid conflict with ffmpeg command line
6964 option.
6965 smoother=live|file
6967 The type of Smoother used for the transmission for that socket,
6968 which is responsible for the transmission and congestion control.
6969 The Smoother type must be exactly the same on both connecting
6970 parties, otherwise the connection is rejected.
6971 messageapi=1|0
6973 When set, this socket uses the Message API, otherwise it uses
6974 Buffer API. Note that in live mode (see transtype) there’s only
6975 message API available. In File mode you can chose to use one of two
6976 modes:
6977 Stream API (default, when this option is false). In this mode you
6979 may send as many data as you wish with one sending instruction, or
6980 even use dedicated functions that read directly from a file. The
6981 internal facility will take care of any speed and congestion
6982 control. When receiving, you can also receive as many data as
6983 desired, the data not extracted will be waiting for the next call.
6984 There is no boundary between data portions in the Stream mode.
6985 Message API. In this mode your single sending instruction passes
6987 exactly one piece of data that has boundaries (a message). Contrary
6988 to Live mode, this message may span across multiple UDP packets and
6989 the only size limitation is that it shall fit as a whole in the
6990 sending buffer. The receiver shall use as large buffer as necessary
6991 to receive the message, otherwise the message will not be given up.
6992 When the message is not complete (not all packets received or there
6993 was a packet loss) it will not be given up.
6994 transtype=live|file
6996 Sets the transmission type for the socket, in particular, setting
6997 this option sets multiple other parameters to their default values
6998 as required for a particular transmission type.
6999 live: Set options as for live transmission. In this mode, you
7001 should send by one sending instruction only so many data that fit
7002 in one UDP packet, and limited to the value defined first in
7003 payload_size (1316 is default in this mode). There is no speed
7004 control in this mode, only the bandwidth control, if configured, in
7005 order to not exceed the bandwidth with the overhead transmission
7006 (retransmitted and control packets).
7007 file: Set options as for non-live transmission. See messageapi for
7009 further explanations
7010 linger=seconds
7012 The number of seconds that the socket waits for unsent data when
7013 closing. Default is -1. -1 means auto (off with 0 seconds in live
7014 mode, on with 180 seconds in file mode). The range for this option
7015 is integers in the 0 - "INT_MAX".
7016 tsbpd=1|0
7018 When true, use Timestamp-based Packet Delivery mode. The default
7019 behavior depends on the transmission type: enabled in live mode,
7020 disabled in file mode.
7021 For more information see: <https://github.com/Haivision/srt>.
7023 srtp
7025 Secure Real-time Transport Protocol.
7026 The accepted options are:
7028 srtp_in_suite
7030 srtp_out_suite
7031 Select input and output encoding suites.
7032 Supported values:
7034 AES_CM_128_HMAC_SHA1_80
7036 SRTP_AES128_CM_HMAC_SHA1_80
7037 AES_CM_128_HMAC_SHA1_32
7038 SRTP_AES128_CM_HMAC_SHA1_32
7039 srtp_in_params
7040 srtp_out_params
7041 Set input and output encoding parameters, which are expressed by a
7042 base64-encoded representation of a binary block. The first 16 bytes
7043 of this binary block are used as master key, the following 14 bytes
7044 are used as master salt.
7045 subfile
7047 Virtually extract a segment of a file or another stream. The
7048 underlying stream must be seekable.
7049 Accepted options:
7051 start
7053 Start offset of the extracted segment, in bytes.
7054 end End offset of the extracted segment, in bytes. If set to 0,
7056 extract till end of file.
7057 Examples:
7059 Extract a chapter from a DVD VOB file (start and end sectors obtained
7061 externally and multiplied by 2048):
7062 subfile,,start,153391104,end,268142592,,:/media/dvd/VIDEO_TS/VTS_08_1.VOB
7064 Play an AVI file directly from a TAR archive:
7066 subfile,,start,183241728,end,366490624,,:archive.tar
7068 Play a MPEG-TS file from start offset till end:
7070 subfile,,start,32815239,end,0,,:video.ts
7072 tee
7074 Writes the output to multiple protocols. The individual outputs are
7075 separated by |
7076 tee:file://path/to/local/this.avi|file://path/to/local/that.avi
7078 tcp
7080 Transmission Control Protocol.
7081 The required syntax for a TCP url is:
7083 tcp://<hostname>:<port>[?<options>]
7085 options contains a list of &-separated options of the form key=val.
7087 The list of supported options follows.
7089 listen=2|1|0
7091 Listen for an incoming connection. 0 disables listen, 1 enables
7092 listen in single client mode, 2 enables listen in multi-client
7093 mode. Default value is 0.
7094 timeout=microseconds
7096 Set raise error timeout, expressed in microseconds.
7097 This option is only relevant in read mode: if no data arrived in
7099 more than this time interval, raise error.
7100 listen_timeout=milliseconds
7102 Set listen timeout, expressed in milliseconds.
7103 recv_buffer_size=bytes
7105 Set receive buffer size, expressed bytes.
7106 send_buffer_size=bytes
7108 Set send buffer size, expressed bytes.
7109 tcp_nodelay=1|0
7111 Set TCP_NODELAY to disable Nagle's algorithm. Default value is 0.
7112 Remark: Writing to the socket is currently not optimized to
7114 minimize system calls and reduces the efficiency / effect of
7115 TCP_NODELAY.
7116 tcp_mss=bytes
7118 Set maximum segment size for outgoing TCP packets, expressed in
7119 bytes.
7120 The following example shows how to setup a listening TCP connection
7122 with ffmpeg, which is then accessed with ffplay:
7123 ffmpeg -i <input> -f <format> tcp://<hostname>:<port>?listen
7125 ffplay tcp://<hostname>:<port>
7126 tls
7128 Transport Layer Security (TLS) / Secure Sockets Layer (SSL)
7129 The required syntax for a TLS/SSL url is:
7131 tls://<hostname>:<port>[?<options>]
7133 The following parameters can be set via command line options (or in
7135 code via "AVOption"s):
7136 ca_file, cafile=filename
7138 A file containing certificate authority (CA) root certificates to
7139 treat as trusted. If the linked TLS library contains a default this
7140 might not need to be specified for verification to work, but not
7141 all libraries and setups have defaults built in. The file must be
7142 in OpenSSL PEM format.
7143 tls_verify=1|0
7145 If enabled, try to verify the peer that we are communicating with.
7146 Note, if using OpenSSL, this currently only makes sure that the
7147 peer certificate is signed by one of the root certificates in the
7148 CA database, but it does not validate that the certificate actually
7149 matches the host name we are trying to connect to. (With other
7150 backends, the host name is validated as well.)
7151 This is disabled by default since it requires a CA database to be
7153 provided by the caller in many cases.
7154 cert_file, cert=filename
7156 A file containing a certificate to use in the handshake with the
7157 peer. (When operating as server, in listen mode, this is more
7158 often required by the peer, while client certificates only are
7159 mandated in certain setups.)
7160 key_file, key=filename
7162 A file containing the private key for the certificate.
7163 listen=1|0
7165 If enabled, listen for connections on the provided port, and assume
7166 the server role in the handshake instead of the client role.
7167 http_proxy
7169 The HTTP proxy to tunnel through, e.g. "http://example.com:1234".
7170 The proxy must support the CONNECT method.
7171 Example command lines:
7173 To create a TLS/SSL server that serves an input stream.
7175 ffmpeg -i <input> -f <format> tls://<hostname>:<port>?listen&cert=<server.crt>&key=<server.key>
7177 To play back a stream from the TLS/SSL server using ffplay:
7179 ffplay tls://<hostname>:<port>
7181 udp
7183 User Datagram Protocol.
7184 The required syntax for an UDP URL is:
7186 udp://<hostname>:<port>[?<options>]
7188 options contains a list of &-separated options of the form key=val.
7190 In case threading is enabled on the system, a circular buffer is used
7192 to store the incoming data, which allows one to reduce loss of data due
7193 to UDP socket buffer overruns. The fifo_size and overrun_nonfatal
7194 options are related to this buffer.
7195 The list of supported options follows.
7197 buffer_size=size
7199 Set the UDP maximum socket buffer size in bytes. This is used to
7200 set either the receive or send buffer size, depending on what the
7201 socket is used for. Default is 32 KB for output, 384 KB for input.
7202 See also fifo_size.
7203 bitrate=bitrate
7205 If set to nonzero, the output will have the specified constant
7206 bitrate if the input has enough packets to sustain it.
7207 burst_bits=bits
7209 When using bitrate this specifies the maximum number of bits in
7210 packet bursts.
7211 localport=port
7213 Override the local UDP port to bind with.
7214 localaddr=addr
7216 Local IP address of a network interface used for sending packets or
7217 joining multicast groups.
7218 pkt_size=size
7220 Set the size in bytes of UDP packets.
7221 reuse=1|0
7223 Explicitly allow or disallow reusing UDP sockets.
7224 ttl=ttl
7226 Set the time to live value (for multicast only).
7227 connect=1|0
7229 Initialize the UDP socket with connect(). In this case, the
7230 destination address can't be changed with ff_udp_set_remote_url
7231 later. If the destination address isn't known at the start, this
7232 option can be specified in ff_udp_set_remote_url, too. This allows
7233 finding out the source address for the packets with getsockname,
7234 and makes writes return with AVERROR(ECONNREFUSED) if "destination
7235 unreachable" is received. For receiving, this gives the benefit of
7236 only receiving packets from the specified peer address/port.
7237 sources=address[,address]
7239 Only receive packets sent from the specified addresses. In case of
7240 multicast, also subscribe to multicast traffic coming from these
7241 addresses only.
7242 block=address[,address]
7244 Ignore packets sent from the specified addresses. In case of
7245 multicast, also exclude the source addresses in the multicast
7246 subscription.
7247 fifo_size=units
7249 Set the UDP receiving circular buffer size, expressed as a number
7250 of packets with size of 188 bytes. If not specified defaults to
7251 7*4096.
7252 overrun_nonfatal=1|0
7254 Survive in case of UDP receiving circular buffer overrun. Default
7255 value is 0.
7256 timeout=microseconds
7258 Set raise error timeout, expressed in microseconds.
7259 This option is only relevant in read mode: if no data arrived in
7261 more than this time interval, raise error.
7262 broadcast=1|0
7264 Explicitly allow or disallow UDP broadcasting.
7265 Note that broadcasting may not work properly on networks having a
7267 broadcast storm protection.
7268 Examples
7270 • Use ffmpeg to stream over UDP to a remote endpoint:
7272 ffmpeg -i <input> -f <format> udp://<hostname>:<port>
7274 • Use ffmpeg to stream in mpegts format over UDP using 188 sized UDP
7276 packets, using a large input buffer:
7277 ffmpeg -i <input> -f mpegts udp://<hostname>:<port>?pkt_size=188&buffer_size=65535
7279 • Use ffmpeg to receive over UDP from a remote endpoint:
7281 ffmpeg -i udp://[<multicast-address>]:<port> ...
7283 unix
7285 Unix local socket
7286 The required syntax for a Unix socket URL is:
7288 unix://<filepath>
7290 The following parameters can be set via command line options (or in
7292 code via "AVOption"s):
7293 timeout
7295 Timeout in ms.
7296 listen
7298 Create the Unix socket in listening mode.
7299 zmq
7301 ZeroMQ asynchronous messaging using the libzmq library.
7302 This library supports unicast streaming to multiple clients without
7304 relying on an external server.
7305 The required syntax for streaming or connecting to a stream is:
7307 zmq:tcp://ip-address:port
7309 Example: Create a localhost stream on port 5555:
7311 ffmpeg -re -i input -f mpegts zmq:tcp://127.0.0.1:5555
7313 Multiple clients may connect to the stream using:
7315 ffplay zmq:tcp://127.0.0.1:5555
7317 Streaming to multiple clients is implemented using a ZeroMQ Pub-Sub
7319 pattern. The server side binds to a port and publishes data. Clients
7320 connect to the server (via IP address/port) and subscribe to the
7321 stream. The order in which the server and client start generally does
7322 not matter.
7323 ffmpeg must be compiled with the --enable-libzmq option to support this
7325 protocol.
7326 Options can be set on the ffmpeg/ffplay command line. The following
7328 options are supported:
7329 pkt_size
7331 Forces the maximum packet size for sending/receiving data. The
7332 default value is 131,072 bytes. On the server side, this sets the
7333 maximum size of sent packets via ZeroMQ. On the clients, it sets an
7334 internal buffer size for receiving packets. Note that pkt_size on
7335 the clients should be equal to or greater than pkt_size on the
7336 server. Otherwise the received message may be truncated causing
7337 decoding errors.
7338
7340 The libavdevice library provides the same interface as libavformat.
7341 Namely, an input device is considered like a demuxer, and an output
7342 device like a muxer, and the interface and generic device options are
7343 the same provided by libavformat (see the ffmpeg-formats manual).
7344 In addition each input or output device may support so-called private
7346 options, which are specific for that component.
7347 Options may be set by specifying -option value in the FFmpeg tools, or
7349 by setting the value explicitly in the device "AVFormatContext" options
7350 or using the libavutil/opt.h API for programmatic use.
7351
7353 Input devices are configured elements in FFmpeg which enable accessing
7354 the data coming from a multimedia device attached to your system.
7355 When you configure your FFmpeg build, all the supported input devices
7357 are enabled by default. You can list all available ones using the
7358 configure option "--list-indevs".
7359 You can disable all the input devices using the configure option
7361 "--disable-indevs", and selectively enable an input device using the
7362 option "--enable-indev=INDEV", or you can disable a particular input
7363 device using the option "--disable-indev=INDEV".
7364 The option "-devices" of the ff* tools will display the list of
7366 supported input devices.
7367 A description of the currently available input devices follows.
7369 alsa
7371 ALSA (Advanced Linux Sound Architecture) input device.
7372 To enable this input device during configuration you need libasound
7374 installed on your system.
7375 This device allows capturing from an ALSA device. The name of the
7377 device to capture has to be an ALSA card identifier.
7378 An ALSA identifier has the syntax:
7380 hw:<CARD>[,<DEV>[,<SUBDEV>]]
7382 where the DEV and SUBDEV components are optional.
7384 The three arguments (in order: CARD,DEV,SUBDEV) specify card number or
7386 identifier, device number and subdevice number (-1 means any).
7387 To see the list of cards currently recognized by your system check the
7389 files /proc/asound/cards and /proc/asound/devices.
7390 For example to capture with ffmpeg from an ALSA device with card id 0,
7392 you may run the command:
7393 ffmpeg -f alsa -i hw:0 alsaout.wav
7395 For more information see:
7397 <http://www.alsa-project.org/alsa-doc/alsa-lib/pcm.html>
7398 Options
7400 sample_rate
7402 Set the sample rate in Hz. Default is 48000.
7403 channels
7405 Set the number of channels. Default is 2.
7406 android_camera
7408 Android camera input device.
7409 This input devices uses the Android Camera2 NDK API which is available
7411 on devices with API level 24+. The availability of android_camera is
7412 autodetected during configuration.
7413 This device allows capturing from all cameras on an Android device,
7415 which are integrated into the Camera2 NDK API.
7416 The available cameras are enumerated internally and can be selected
7418 with the camera_index parameter. The input file string is discarded.
7419 Generally the back facing camera has index 0 while the front facing
7421 camera has index 1.
7422 Options
7424 video_size
7426 Set the video size given as a string such as 640x480 or hd720.
7427 Falls back to the first available configuration reported by Android
7428 if requested video size is not available or by default.
7429 framerate
7431 Set the video framerate. Falls back to the first available
7432 configuration reported by Android if requested framerate is not
7433 available or by default (-1).
7434 camera_index
7436 Set the index of the camera to use. Default is 0.
7437 input_queue_size
7439 Set the maximum number of frames to buffer. Default is 5.
7440 avfoundation
7442 AVFoundation input device.
7443 AVFoundation is the currently recommended framework by Apple for
7445 streamgrabbing on OSX >= 10.7 as well as on iOS.
7446 The input filename has to be given in the following syntax:
7448 -i "[[VIDEO]:[AUDIO]]"
7450 The first entry selects the video input while the latter selects the
7452 audio input. The stream has to be specified by the device name or the
7453 device index as shown by the device list. Alternatively, the video
7454 and/or audio input device can be chosen by index using the
7455 B<-video_device_index E<lt>INDEXE<gt>>
7457 and/or
7459 B<-audio_device_index E<lt>INDEXE<gt>>
7461 , overriding any device name or index given in the input filename.
7463 All available devices can be enumerated by using -list_devices true,
7465 listing all device names and corresponding indices.
7466 There are two device name aliases:
7468 "default"
7470 Select the AVFoundation default device of the corresponding type.
7471 "none"
7473 Do not record the corresponding media type. This is equivalent to
7474 specifying an empty device name or index.
7475 Options
7477 AVFoundation supports the following options:
7479 -list_devices <TRUE|FALSE>
7481 If set to true, a list of all available input devices is given
7482 showing all device names and indices.
7483 -video_device_index <INDEX>
7485 Specify the video device by its index. Overrides anything given in
7486 the input filename.
7487 -audio_device_index <INDEX>
7489 Specify the audio device by its index. Overrides anything given in
7490 the input filename.
7491 -pixel_format <FORMAT>
7493 Request the video device to use a specific pixel format. If the
7494 specified format is not supported, a list of available formats is
7495 given and the first one in this list is used instead. Available
7496 pixel formats are: "monob, rgb555be, rgb555le, rgb565be, rgb565le,
7497 rgb24, bgr24, 0rgb, bgr0, 0bgr, rgb0,
7498 bgr48be, uyvy422, yuva444p, yuva444p16le, yuv444p, yuv422p16,
7499 yuv422p10, yuv444p10,
7500 yuv420p, nv12, yuyv422, gray"
7501 -framerate
7503 Set the grabbing frame rate. Default is "ntsc", corresponding to a
7504 frame rate of "30000/1001".
7505 -video_size
7507 Set the video frame size.
7508 -capture_cursor
7510 Capture the mouse pointer. Default is 0.
7511 -capture_mouse_clicks
7513 Capture the screen mouse clicks. Default is 0.
7514 -capture_raw_data
7516 Capture the raw device data. Default is 0. Using this option may
7517 result in receiving the underlying data delivered to the
7518 AVFoundation framework. E.g. for muxed devices that sends raw DV
7519 data to the framework (like tape-based camcorders), setting this
7520 option to false results in extracted video frames captured in the
7521 designated pixel format only. Setting this option to true results
7522 in receiving the raw DV stream untouched.
7523 Examples
7525 • Print the list of AVFoundation supported devices and exit:
7527 $ ffmpeg -f avfoundation -list_devices true -i ""
7529 • Record video from video device 0 and audio from audio device 0 into
7531 out.avi:
7532 $ ffmpeg -f avfoundation -i "0:0" out.avi
7534 • Record video from video device 2 and audio from audio device 1 into
7536 out.avi:
7537 $ ffmpeg -f avfoundation -video_device_index 2 -i ":1" out.avi
7539 • Record video from the system default video device using the pixel
7541 format bgr0 and do not record any audio into out.avi:
7542 $ ffmpeg -f avfoundation -pixel_format bgr0 -i "default:none" out.avi
7544 • Record raw DV data from a suitable input device and write the
7546 output into out.dv:
7547 $ ffmpeg -f avfoundation -capture_raw_data true -i "zr100:none" out.dv
7549 bktr
7551 BSD video input device.
7552 Options
7554 framerate
7556 Set the frame rate.
7557 video_size
7559 Set the video frame size. Default is "vga".
7560 standard
7562 Available values are:
7563 pal
7565 ntsc
7566 secam
7567 paln
7568 palm
7569 ntscj
7570 decklink
7572 The decklink input device provides capture capabilities for Blackmagic
7573 DeckLink devices.
7574 To enable this input device, you need the Blackmagic DeckLink SDK and
7576 you need to configure with the appropriate "--extra-cflags" and
7577 "--extra-ldflags". On Windows, you need to run the IDL files through
7578 widl.
7579 DeckLink is very picky about the formats it supports. Pixel format of
7581 the input can be set with raw_format. Framerate and video size must be
7582 determined for your device with -list_formats 1. Audio sample rate is
7583 always 48 kHz and the number of channels can be 2, 8 or 16. Note that
7584 all audio channels are bundled in one single audio track.
7585 Options
7587 list_devices
7589 If set to true, print a list of devices and exit. Defaults to
7590 false. This option is deprecated, please use the "-sources" option
7591 of ffmpeg to list the available input devices.
7592 list_formats
7594 If set to true, print a list of supported formats and exit.
7595 Defaults to false.
7596 format_code <FourCC>
7598 This sets the input video format to the format given by the FourCC.
7599 To see the supported values of your device(s) use list_formats.
7600 Note that there is a FourCC 'pal ' that can also be used as pal (3
7601 letters). Default behavior is autodetection of the input video
7602 format, if the hardware supports it.
7603 raw_format
7605 Set the pixel format of the captured video. Available values are:
7606 auto
7608 This is the default which means 8-bit YUV 422 or 8-bit ARGB if
7609 format autodetection is used, 8-bit YUV 422 otherwise.
7610 uyvy422
7612 8-bit YUV 422.
7613 yuv422p10
7615 10-bit YUV 422.
7616 argb
7618 8-bit RGB.
7619 bgra
7621 8-bit RGB.
7622 rgb10
7624 10-bit RGB.
7625 teletext_lines
7627 If set to nonzero, an additional teletext stream will be captured
7628 from the vertical ancillary data. Both SD PAL (576i) and HD (1080i
7629 or 1080p) sources are supported. In case of HD sources, OP47
7630 packets are decoded.
7631 This option is a bitmask of the SD PAL VBI lines captured,
7633 specifically lines 6 to 22, and lines 318 to 335. Line 6 is the LSB
7634 in the mask. Selected lines which do not contain teletext
7635 information will be ignored. You can use the special all constant
7636 to select all possible lines, or standard to skip lines 6, 318 and
7637 319, which are not compatible with all receivers.
7638 For SD sources, ffmpeg needs to be compiled with
7640 "--enable-libzvbi". For HD sources, on older (pre-4K) DeckLink card
7641 models you have to capture in 10 bit mode.
7642 channels
7644 Defines number of audio channels to capture. Must be 2, 8 or 16.
7645 Defaults to 2.
7646 duplex_mode
7648 Sets the decklink device duplex/profile mode. Must be unset, half,
7649 full, one_sub_device_full, one_sub_device_half,
7650 two_sub_device_full, four_sub_device_half Defaults to unset.
7651 Note: DeckLink SDK 11.0 have replaced the duplex property by a
7653 profile property. For the DeckLink Duo 2 and DeckLink Quad 2, a
7654 profile is shared between any 2 sub-devices that utilize the same
7655 connectors. For the DeckLink 8K Pro, a profile is shared between
7656 all 4 sub-devices. So DeckLink 8K Pro support four profiles.
7657 Valid profile modes for DeckLink 8K Pro(with DeckLink SDK >= 11.0):
7659 one_sub_device_full, one_sub_device_half, two_sub_device_full,
7660 four_sub_device_half
7661 Valid profile modes for DeckLink Quad 2 and DeckLink Duo 2: half,
7663 full
7664 timecode_format
7666 Timecode type to include in the frame and video stream metadata.
7667 Must be none, rp188vitc, rp188vitc2, rp188ltc, rp188hfr, rp188any,
7668 vitc, vitc2, or serial. Defaults to none (not included).
7669 In order to properly support 50/60 fps timecodes, the ordering of
7671 the queried timecode types for rp188any is HFR, VITC1, VITC2 and
7672 LTC for >30 fps content. Note that this is slightly different to
7673 the ordering used by the DeckLink API, which is HFR, VITC1, LTC,
7674 VITC2.
7675 video_input
7677 Sets the video input source. Must be unset, sdi, hdmi, optical_sdi,
7678 component, composite or s_video. Defaults to unset.
7679 audio_input
7681 Sets the audio input source. Must be unset, embedded, aes_ebu,
7682 analog, analog_xlr, analog_rca or microphone. Defaults to unset.
7683 video_pts
7685 Sets the video packet timestamp source. Must be video, audio,
7686 reference, wallclock or abs_wallclock. Defaults to video.
7687 audio_pts
7689 Sets the audio packet timestamp source. Must be video, audio,
7690 reference, wallclock or abs_wallclock. Defaults to audio.
7691 draw_bars
7693 If set to true, color bars are drawn in the event of a signal loss.
7694 Defaults to true.
7695 queue_size
7697 Sets maximum input buffer size in bytes. If the buffering reaches
7698 this value, incoming frames will be dropped. Defaults to
7699 1073741824.
7700 audio_depth
7702 Sets the audio sample bit depth. Must be 16 or 32. Defaults to 16.
7703 decklink_copyts
7705 If set to true, timestamps are forwarded as they are without
7706 removing the initial offset. Defaults to false.
7707 timestamp_align
7709 Capture start time alignment in seconds. If set to nonzero, input
7710 frames are dropped till the system timestamp aligns with configured
7711 value. Alignment difference of up to one frame duration is
7712 tolerated. This is useful for maintaining input synchronization
7713 across N different hardware devices deployed for 'N-way'
7714 redundancy. The system time of different hardware devices should be
7715 synchronized with protocols such as NTP or PTP, before using this
7716 option. Note that this method is not foolproof. In some border
7717 cases input synchronization may not happen due to thread scheduling
7718 jitters in the OS. Either sync could go wrong by 1 frame or in a
7719 rarer case timestamp_align seconds. Defaults to 0.
7720 wait_for_tc (bool)
7722 Drop frames till a frame with timecode is received. Sometimes
7723 serial timecode isn't received with the first input frame. If that
7724 happens, the stored stream timecode will be inaccurate. If this
7725 option is set to true, input frames are dropped till a frame with
7726 timecode is received. Option timecode_format must be specified.
7727 Defaults to false.
7728 enable_klv(bool)
7730 If set to true, extracts KLV data from VANC and outputs KLV
7731 packets. KLV VANC packets are joined based on MID and PSC fields
7732 and aggregated into one KLV packet. Defaults to false.
7733 Examples
7735 • List input devices:
7737 ffmpeg -sources decklink
7739 • List supported formats:
7741 ffmpeg -f decklink -list_formats 1 -i 'Intensity Pro'
7743 • Capture video clip at 1080i50:
7745 ffmpeg -format_code Hi50 -f decklink -i 'Intensity Pro' -c:a copy -c:v copy output.avi
7747 • Capture video clip at 1080i50 10 bit:
7749 ffmpeg -raw_format yuv422p10 -format_code Hi50 -f decklink -i 'UltraStudio Mini Recorder' -c:a copy -c:v copy output.avi
7751 • Capture video clip at 1080i50 with 16 audio channels:
7753 ffmpeg -channels 16 -format_code Hi50 -f decklink -i 'UltraStudio Mini Recorder' -c:a copy -c:v copy output.avi
7755 dshow
7757 Windows DirectShow input device.
7758 DirectShow support is enabled when FFmpeg is built with the mingw-w64
7760 project. Currently only audio and video devices are supported.
7761 Multiple devices may be opened as separate inputs, but they may also be
7763 opened on the same input, which should improve synchronism between
7764 them.
7765 The input name should be in the format:
7767 <TYPE>=<NAME>[:<TYPE>=<NAME>]
7769 where TYPE can be either audio or video, and NAME is the device's name
7771 or alternative name..
7772 Options
7774 If no options are specified, the device's defaults are used. If the
7776 device does not support the requested options, it will fail to open.
7777 video_size
7779 Set the video size in the captured video.
7780 framerate
7782 Set the frame rate in the captured video.
7783 sample_rate
7785 Set the sample rate (in Hz) of the captured audio.
7786 sample_size
7788 Set the sample size (in bits) of the captured audio.
7789 channels
7791 Set the number of channels in the captured audio.
7792 list_devices
7794 If set to true, print a list of devices and exit.
7795 list_options
7797 If set to true, print a list of selected device's options and exit.
7798 video_device_number
7800 Set video device number for devices with the same name (starts at
7801 0, defaults to 0).
7802 audio_device_number
7804 Set audio device number for devices with the same name (starts at
7805 0, defaults to 0).
7806 pixel_format
7808 Select pixel format to be used by DirectShow. This may only be set
7809 when the video codec is not set or set to rawvideo.
7810 audio_buffer_size
7812 Set audio device buffer size in milliseconds (which can directly
7813 impact latency, depending on the device). Defaults to using the
7814 audio device's default buffer size (typically some multiple of
7815 500ms). Setting this value too low can degrade performance. See
7816 also
7817 <http://msdn.microsoft.com/en-us/library/windows/desktop/dd377582(v=vs.85).aspx>
7818 video_pin_name
7820 Select video capture pin to use by name or alternative name.
7821 audio_pin_name
7823 Select audio capture pin to use by name or alternative name.
7824 crossbar_video_input_pin_number
7826 Select video input pin number for crossbar device. This will be
7827 routed to the crossbar device's Video Decoder output pin. Note
7828 that changing this value can affect future invocations (sets a new
7829 default) until system reboot occurs.
7830 crossbar_audio_input_pin_number
7832 Select audio input pin number for crossbar device. This will be
7833 routed to the crossbar device's Audio Decoder output pin. Note
7834 that changing this value can affect future invocations (sets a new
7835 default) until system reboot occurs.
7836 show_video_device_dialog
7838 If set to true, before capture starts, popup a display dialog to
7839 the end user, allowing them to change video filter properties and
7840 configurations manually. Note that for crossbar devices, adjusting
7841 values in this dialog may be needed at times to toggle between PAL
7842 (25 fps) and NTSC (29.97) input frame rates, sizes, interlacing,
7843 etc. Changing these values can enable different scan rates/frame
7844 rates and avoiding green bars at the bottom, flickering scan lines,
7845 etc. Note that with some devices, changing these properties can
7846 also affect future invocations (sets new defaults) until system
7847 reboot occurs.
7848 show_audio_device_dialog
7850 If set to true, before capture starts, popup a display dialog to
7851 the end user, allowing them to change audio filter properties and
7852 configurations manually.
7853 show_video_crossbar_connection_dialog
7855 If set to true, before capture starts, popup a display dialog to
7856 the end user, allowing them to manually modify crossbar pin
7857 routings, when it opens a video device.
7858 show_audio_crossbar_connection_dialog
7860 If set to true, before capture starts, popup a display dialog to
7861 the end user, allowing them to manually modify crossbar pin
7862 routings, when it opens an audio device.
7863 show_analog_tv_tuner_dialog
7865 If set to true, before capture starts, popup a display dialog to
7866 the end user, allowing them to manually modify TV channels and
7867 frequencies.
7868 show_analog_tv_tuner_audio_dialog
7870 If set to true, before capture starts, popup a display dialog to
7871 the end user, allowing them to manually modify TV audio (like mono
7872 vs. stereo, Language A,B or C).
7873 audio_device_load
7875 Load an audio capture filter device from file instead of searching
7876 it by name. It may load additional parameters too, if the filter
7877 supports the serialization of its properties to. To use this an
7878 audio capture source has to be specified, but it can be anything
7879 even fake one.
7880 audio_device_save
7882 Save the currently used audio capture filter device and its
7883 parameters (if the filter supports it) to a file. If a file with
7884 the same name exists it will be overwritten.
7885 video_device_load
7887 Load a video capture filter device from file instead of searching
7888 it by name. It may load additional parameters too, if the filter
7889 supports the serialization of its properties to. To use this a
7890 video capture source has to be specified, but it can be anything
7891 even fake one.
7892 video_device_save
7894 Save the currently used video capture filter device and its
7895 parameters (if the filter supports it) to a file. If a file with
7896 the same name exists it will be overwritten.
7897 use_video_device_timestamps
7899 If set to false, the timestamp for video frames will be derived
7900 from the wallclock instead of the timestamp provided by the capture
7901 device. This allows working around devices that provide unreliable
7902 timestamps.
7903 Examples
7905 • Print the list of DirectShow supported devices and exit:
7907 $ ffmpeg -list_devices true -f dshow -i dummy
7909 • Open video device Camera:
7911 $ ffmpeg -f dshow -i video="Camera"
7913 • Open second video device with name Camera:
7915 $ ffmpeg -f dshow -video_device_number 1 -i video="Camera"
7917 • Open video device Camera and audio device Microphone:
7919 $ ffmpeg -f dshow -i video="Camera":audio="Microphone"
7921 • Print the list of supported options in selected device and exit:
7923 $ ffmpeg -list_options true -f dshow -i video="Camera"
7925 • Specify pin names to capture by name or alternative name, specify
7927 alternative device name:
7928 $ 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"
7930 • Configure a crossbar device, specifying crossbar pins, allow user
7932 to adjust video capture properties at startup:
7933 $ ffmpeg -f dshow -show_video_device_dialog true -crossbar_video_input_pin_number 0
7935 -crossbar_audio_input_pin_number 3 -i video="AVerMedia BDA Analog Capture":audio="AVerMedia BDA Analog Capture"
7936 fbdev
7938 Linux framebuffer input device.
7939 The Linux framebuffer is a graphic hardware-independent abstraction
7941 layer to show graphics on a computer monitor, typically on the console.
7942 It is accessed through a file device node, usually /dev/fb0.
7943 For more detailed information read the file
7945 Documentation/fb/framebuffer.txt included in the Linux source tree.
7946 See also <http://linux-fbdev.sourceforge.net/>, and fbset(1).
7948 To record from the framebuffer device /dev/fb0 with ffmpeg:
7950 ffmpeg -f fbdev -framerate 10 -i /dev/fb0 out.avi
7952 You can take a single screenshot image with the command:
7954 ffmpeg -f fbdev -framerate 1 -i /dev/fb0 -frames:v 1 screenshot.jpeg
7956 Options
7958 framerate
7960 Set the frame rate. Default is 25.
7961 gdigrab
7963 Win32 GDI-based screen capture device.
7964 This device allows you to capture a region of the display on Windows.
7966 There are two options for the input filename:
7968 desktop
7970 or
7972 title=<window_title>
7974 The first option will capture the entire desktop, or a fixed region of
7976 the desktop. The second option will instead capture the contents of a
7977 single window, regardless of its position on the screen.
7978 For example, to grab the entire desktop using ffmpeg:
7980 ffmpeg -f gdigrab -framerate 6 -i desktop out.mpg
7982 Grab a 640x480 region at position "10,20":
7984 ffmpeg -f gdigrab -framerate 6 -offset_x 10 -offset_y 20 -video_size vga -i desktop out.mpg
7986 Grab the contents of the window named "Calculator"
7988 ffmpeg -f gdigrab -framerate 6 -i title=Calculator out.mpg
7990 Options
7992 draw_mouse
7994 Specify whether to draw the mouse pointer. Use the value 0 to not
7995 draw the pointer. Default value is 1.
7996 framerate
7998 Set the grabbing frame rate. Default value is "ntsc", corresponding
7999 to a frame rate of "30000/1001".
8000 show_region
8002 Show grabbed region on screen.
8003 If show_region is specified with 1, then the grabbing region will
8005 be indicated on screen. With this option, it is easy to know what
8006 is being grabbed if only a portion of the screen is grabbed.
8007 Note that show_region is incompatible with grabbing the contents of
8009 a single window.
8010 For example:
8012 ffmpeg -f gdigrab -show_region 1 -framerate 6 -video_size cif -offset_x 10 -offset_y 20 -i desktop out.mpg
8014 video_size
8016 Set the video frame size. The default is to capture the full screen
8017 if desktop is selected, or the full window size if
8018 title=window_title is selected.
8019 offset_x
8021 When capturing a region with video_size, set the distance from the
8022 left edge of the screen or desktop.
8023 Note that the offset calculation is from the top left corner of the
8025 primary monitor on Windows. If you have a monitor positioned to the
8026 left of your primary monitor, you will need to use a negative
8027 offset_x value to move the region to that monitor.
8028 offset_y
8030 When capturing a region with video_size, set the distance from the
8031 top edge of the screen or desktop.
8032 Note that the offset calculation is from the top left corner of the
8034 primary monitor on Windows. If you have a monitor positioned above
8035 your primary monitor, you will need to use a negative offset_y
8036 value to move the region to that monitor.
8037 iec61883
8039 FireWire DV/HDV input device using libiec61883.
8040 To enable this input device, you need libiec61883, libraw1394 and
8042 libavc1394 installed on your system. Use the configure option
8043 "--enable-libiec61883" to compile with the device enabled.
8044 The iec61883 capture device supports capturing from a video device
8046 connected via IEEE1394 (FireWire), using libiec61883 and the new Linux
8047 FireWire stack (juju). This is the default DV/HDV input method in Linux
8048 Kernel 2.6.37 and later, since the old FireWire stack was removed.
8049 Specify the FireWire port to be used as input file, or "auto" to choose
8051 the first port connected.
8052 Options
8054 dvtype
8056 Override autodetection of DV/HDV. This should only be used if auto
8057 detection does not work, or if usage of a different device type
8058 should be prohibited. Treating a DV device as HDV (or vice versa)
8059 will not work and result in undefined behavior. The values auto,
8060 dv and hdv are supported.
8061 dvbuffer
8063 Set maximum size of buffer for incoming data, in frames. For DV,
8064 this is an exact value. For HDV, it is not frame exact, since HDV
8065 does not have a fixed frame size.
8066 dvguid
8068 Select the capture device by specifying its GUID. Capturing will
8069 only be performed from the specified device and fails if no device
8070 with the given GUID is found. This is useful to select the input if
8071 multiple devices are connected at the same time. Look at
8072 /sys/bus/firewire/devices to find out the GUIDs.
8073 Examples
8075 • Grab and show the input of a FireWire DV/HDV device.
8077 ffplay -f iec61883 -i auto
8079 • Grab and record the input of a FireWire DV/HDV device, using a
8081 packet buffer of 100000 packets if the source is HDV.
8082 ffmpeg -f iec61883 -i auto -dvbuffer 100000 out.mpg
8084 jack
8086 JACK input device.
8087 To enable this input device during configuration you need libjack
8089 installed on your system.
8090 A JACK input device creates one or more JACK writable clients, one for
8092 each audio channel, with name client_name:input_N, where client_name is
8093 the name provided by the application, and N is a number which
8094 identifies the channel. Each writable client will send the acquired
8095 data to the FFmpeg input device.
8096 Once you have created one or more JACK readable clients, you need to
8098 connect them to one or more JACK writable clients.
8099 To connect or disconnect JACK clients you can use the jack_connect and
8101 jack_disconnect programs, or do it through a graphical interface, for
8102 example with qjackctl.
8103 To list the JACK clients and their properties you can invoke the
8105 command jack_lsp.
8106 Follows an example which shows how to capture a JACK readable client
8108 with ffmpeg.
8109 # Create a JACK writable client with name "ffmpeg".
8111 $ ffmpeg -f jack -i ffmpeg -y out.wav
8112 # Start the sample jack_metro readable client.
8114 $ jack_metro -b 120 -d 0.2 -f 4000
8115 # List the current JACK clients.
8117 $ jack_lsp -c
8118 system:capture_1
8119 system:capture_2
8120 system:playback_1
8121 system:playback_2
8122 ffmpeg:input_1
8123 metro:120_bpm
8124 # Connect metro to the ffmpeg writable client.
8126 $ jack_connect metro:120_bpm ffmpeg:input_1
8127 For more information read: <http://jackaudio.org/>
8129 Options
8131 channels
8133 Set the number of channels. Default is 2.
8134 kmsgrab
8136 KMS video input device.
8137 Captures the KMS scanout framebuffer associated with a specified CRTC
8139 or plane as a DRM object that can be passed to other hardware
8140 functions.
8141 Requires either DRM master or CAP_SYS_ADMIN to run.
8143 If you don't understand what all of that means, you probably don't want
8145 this. Look at x11grab instead.
8146 Options
8148 device
8150 DRM device to capture on. Defaults to /dev/dri/card0.
8151 format
8153 Pixel format of the framebuffer. This can be autodetected if you
8154 are running Linux 5.7 or later, but needs to be provided for
8155 earlier versions. Defaults to bgr0, which is the most common
8156 format used by the Linux console and Xorg X server.
8157 format_modifier
8159 Format modifier to signal on output frames. This is necessary to
8160 import correctly into some APIs. It can be autodetected if you are
8161 running Linux 5.7 or later, but will need to be provided explicitly
8162 when needed in earlier versions. See the libdrm documentation for
8163 possible values.
8164 crtc_id
8166 KMS CRTC ID to define the capture source. The first active plane
8167 on the given CRTC will be used.
8168 plane_id
8170 KMS plane ID to define the capture source. Defaults to the first
8171 active plane found if neither crtc_id nor plane_id are specified.
8172 framerate
8174 Framerate to capture at. This is not synchronised to any page
8175 flipping or framebuffer changes - it just defines the interval at
8176 which the framebuffer is sampled. Sampling faster than the
8177 framebuffer update rate will generate independent frames with the
8178 same content. Defaults to 30.
8179 Examples
8181 • Capture from the first active plane, download the result to normal
8183 frames and encode. This will only work if the framebuffer is both
8184 linear and mappable - if not, the result may be scrambled or fail
8185 to download.
8186 ffmpeg -f kmsgrab -i - -vf 'hwdownload,format=bgr0' output.mp4
8188 • Capture from CRTC ID 42 at 60fps, map the result to VAAPI, convert
8190 to NV12 and encode as H.264.
8191 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
8193 • To capture only part of a plane the output can be cropped - this
8195 can be used to capture a single window, as long as it has a known
8196 absolute position and size. For example, to capture and encode the
8197 middle quarter of a 1920x1080 plane:
8198 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
8200 lavfi
8202 Libavfilter input virtual device.
8203 This input device reads data from the open output pads of a libavfilter
8205 filtergraph.
8206 For each filtergraph open output, the input device will create a
8208 corresponding stream which is mapped to the generated output. Currently
8209 only video data is supported. The filtergraph is specified through the
8210 option graph.
8211 Options
8213 graph
8215 Specify the filtergraph to use as input. Each video open output
8216 must be labelled by a unique string of the form "outN", where N is
8217 a number starting from 0 corresponding to the mapped input stream
8218 generated by the device. The first unlabelled output is
8219 automatically assigned to the "out0" label, but all the others need
8220 to be specified explicitly.
8221 The suffix "+subcc" can be appended to the output label to create
8223 an extra stream with the closed captions packets attached to that
8224 output (experimental; only for EIA-608 / CEA-708 for now). The
8225 subcc streams are created after all the normal streams, in the
8226 order of the corresponding stream. For example, if there is
8227 "out19+subcc", "out7+subcc" and up to "out42", the stream #43 is
8228 subcc for stream #7 and stream #44 is subcc for stream #19.
8229 If not specified defaults to the filename specified for the input
8231 device.
8232 graph_file
8234 Set the filename of the filtergraph to be read and sent to the
8235 other filters. Syntax of the filtergraph is the same as the one
8236 specified by the option graph.
8237 dumpgraph
8239 Dump graph to stderr.
8240 Examples
8242 • Create a color video stream and play it back with ffplay:
8244 ffplay -f lavfi -graph "color=c=pink [out0]" dummy
8246 • As the previous example, but use filename for specifying the graph
8248 description, and omit the "out0" label:
8249 ffplay -f lavfi color=c=pink
8251 • Create three different video test filtered sources and play them:
8253 ffplay -f lavfi -graph "testsrc [out0]; testsrc,hflip [out1]; testsrc,negate [out2]" test3
8255 • Read an audio stream from a file using the amovie source and play
8257 it back with ffplay:
8258 ffplay -f lavfi "amovie=test.wav"
8260 • Read an audio stream and a video stream and play it back with
8262 ffplay:
8263 ffplay -f lavfi "movie=test.avi[out0];amovie=test.wav[out1]"
8265 • Dump decoded frames to images and closed captions to a file
8267 (experimental):
8268 ffmpeg -f lavfi -i "movie=test.ts[out0+subcc]" -map v frame%08d.png -map s -c copy -f rawvideo subcc.bin
8270 libcdio
8272 Audio-CD input device based on libcdio.
8273 To enable this input device during configuration you need libcdio
8275 installed on your system. It requires the configure option
8276 "--enable-libcdio".
8277 This device allows playing and grabbing from an Audio-CD.
8279 For example to copy with ffmpeg the entire Audio-CD in /dev/sr0, you
8281 may run the command:
8282 ffmpeg -f libcdio -i /dev/sr0 cd.wav
8284 Options
8286 speed
8288 Set drive reading speed. Default value is 0.
8289 The speed is specified CD-ROM speed units. The speed is set through
8291 the libcdio "cdio_cddap_speed_set" function. On many CD-ROM drives,
8292 specifying a value too large will result in using the fastest
8293 speed.
8294 paranoia_mode
8296 Set paranoia recovery mode flags. It accepts one of the following
8297 values:
8298 disable
8300 verify
8301 overlap
8302 neverskip
8303 full
8304 Default value is disable.
8306 For more information about the available recovery modes, consult
8308 the paranoia project documentation.
8309 libdc1394
8311 IIDC1394 input device, based on libdc1394 and libraw1394.
8312 Requires the configure option "--enable-libdc1394".
8314 Options
8316 framerate
8318 Set the frame rate. Default is "ntsc", corresponding to a frame
8319 rate of "30000/1001".
8320 pixel_format
8322 Select the pixel format. Default is "uyvy422".
8323 video_size
8325 Set the video size given as a string such as "640x480" or "hd720".
8326 Default is "qvga".
8327 openal
8329 The OpenAL input device provides audio capture on all systems with a
8330 working OpenAL 1.1 implementation.
8331 To enable this input device during configuration, you need OpenAL
8333 headers and libraries installed on your system, and need to configure
8334 FFmpeg with "--enable-openal".
8335 OpenAL headers and libraries should be provided as part of your OpenAL
8337 implementation, or as an additional download (an SDK). Depending on
8338 your installation you may need to specify additional flags via the
8339 "--extra-cflags" and "--extra-ldflags" for allowing the build system to
8340 locate the OpenAL headers and libraries.
8341 An incomplete list of OpenAL implementations follows:
8343 Creative
8345 The official Windows implementation, providing hardware
8346 acceleration with supported devices and software fallback. See
8347 <http://openal.org/>.
8348 OpenAL Soft
8350 Portable, open source (LGPL) software implementation. Includes
8351 backends for the most common sound APIs on the Windows, Linux,
8352 Solaris, and BSD operating systems. See
8353 <http://kcat.strangesoft.net/openal.html>.
8354 Apple
8356 OpenAL is part of Core Audio, the official Mac OS X Audio
8357 interface. See
8358 <http://developer.apple.com/technologies/mac/audio-and-video.html>
8359 This device allows one to capture from an audio input device handled
8361 through OpenAL.
8362 You need to specify the name of the device to capture in the provided
8364 filename. If the empty string is provided, the device will
8365 automatically select the default device. You can get the list of the
8366 supported devices by using the option list_devices.
8367 Options
8369 channels
8371 Set the number of channels in the captured audio. Only the values 1
8372 (monaural) and 2 (stereo) are currently supported. Defaults to 2.
8373 sample_size
8375 Set the sample size (in bits) of the captured audio. Only the
8376 values 8 and 16 are currently supported. Defaults to 16.
8377 sample_rate
8379 Set the sample rate (in Hz) of the captured audio. Defaults to
8380 44.1k.
8381 list_devices
8383 If set to true, print a list of devices and exit. Defaults to
8384 false.
8385 Examples
8387 Print the list of OpenAL supported devices and exit:
8389 $ ffmpeg -list_devices true -f openal -i dummy out.ogg
8391 Capture from the OpenAL device DR-BT101 via PulseAudio:
8393 $ ffmpeg -f openal -i 'DR-BT101 via PulseAudio' out.ogg
8395 Capture from the default device (note the empty string '' as filename):
8397 $ ffmpeg -f openal -i '' out.ogg
8399 Capture from two devices simultaneously, writing to two different
8401 files, within the same ffmpeg command:
8402 $ ffmpeg -f openal -i 'DR-BT101 via PulseAudio' out1.ogg -f openal -i 'ALSA Default' out2.ogg
8404 Note: not all OpenAL implementations support multiple simultaneous
8406 capture - try the latest OpenAL Soft if the above does not work.
8407 oss
8409 Open Sound System input device.
8410 The filename to provide to the input device is the device node
8412 representing the OSS input device, and is usually set to /dev/dsp.
8413 For example to grab from /dev/dsp using ffmpeg use the command:
8415 ffmpeg -f oss -i /dev/dsp /tmp/oss.wav
8417 For more information about OSS see:
8419 <http://manuals.opensound.com/usersguide/dsp.html>
8420 Options
8422 sample_rate
8424 Set the sample rate in Hz. Default is 48000.
8425 channels
8427 Set the number of channels. Default is 2.
8428 pulse
8430 PulseAudio input device.
8431 To enable this output device you need to configure FFmpeg with
8433 "--enable-libpulse".
8434 The filename to provide to the input device is a source device or the
8436 string "default"
8437 To list the PulseAudio source devices and their properties you can
8439 invoke the command pactl list sources.
8440 More information about PulseAudio can be found on
8442 <http://www.pulseaudio.org>.
8443 Options
8445 server
8447 Connect to a specific PulseAudio server, specified by an IP
8448 address. Default server is used when not provided.
8449 name
8451 Specify the application name PulseAudio will use when showing
8452 active clients, by default it is the "LIBAVFORMAT_IDENT" string.
8453 stream_name
8455 Specify the stream name PulseAudio will use when showing active
8456 streams, by default it is "record".
8457 sample_rate
8459 Specify the samplerate in Hz, by default 48kHz is used.
8460 channels
8462 Specify the channels in use, by default 2 (stereo) is set.
8463 frame_size
8465 This option does nothing and is deprecated.
8466 fragment_size
8468 Specify the size in bytes of the minimal buffering fragment in
8469 PulseAudio, it will affect the audio latency. By default it is set
8470 to 50 ms amount of data.
8471 wallclock
8473 Set the initial PTS using the current time. Default is 1.
8474 Examples
8476 Record a stream from default device:
8478 ffmpeg -f pulse -i default /tmp/pulse.wav
8480 sndio
8482 sndio input device.
8483 To enable this input device during configuration you need libsndio
8485 installed on your system.
8486 The filename to provide to the input device is the device node
8488 representing the sndio input device, and is usually set to /dev/audio0.
8489 For example to grab from /dev/audio0 using ffmpeg use the command:
8491 ffmpeg -f sndio -i /dev/audio0 /tmp/oss.wav
8493 Options
8495 sample_rate
8497 Set the sample rate in Hz. Default is 48000.
8498 channels
8500 Set the number of channels. Default is 2.
8501 video4linux2, v4l2
8503 Video4Linux2 input video device.
8504 "v4l2" can be used as alias for "video4linux2".
8506 If FFmpeg is built with v4l-utils support (by using the
8508 "--enable-libv4l2" configure option), it is possible to use it with the
8509 "-use_libv4l2" input device option.
8510 The name of the device to grab is a file device node, usually Linux
8512 systems tend to automatically create such nodes when the device (e.g.
8513 an USB webcam) is plugged into the system, and has a name of the kind
8514 /dev/videoN, where N is a number associated to the device.
8515 Video4Linux2 devices usually support a limited set of widthxheight
8517 sizes and frame rates. You can check which are supported using
8518 -list_formats all for Video4Linux2 devices. Some devices, like TV
8519 cards, support one or more standards. It is possible to list all the
8520 supported standards using -list_standards all.
8521 The time base for the timestamps is 1 microsecond. Depending on the
8523 kernel version and configuration, the timestamps may be derived from
8524 the real time clock (origin at the Unix Epoch) or the monotonic clock
8525 (origin usually at boot time, unaffected by NTP or manual changes to
8526 the clock). The -timestamps abs or -ts abs option can be used to force
8527 conversion into the real time clock.
8528 Some usage examples of the video4linux2 device with ffmpeg and ffplay:
8530 • List supported formats for a video4linux2 device:
8532 ffplay -f video4linux2 -list_formats all /dev/video0
8534 • Grab and show the input of a video4linux2 device:
8536 ffplay -f video4linux2 -framerate 30 -video_size hd720 /dev/video0
8538 • Grab and record the input of a video4linux2 device, leave the frame
8540 rate and size as previously set:
8541 ffmpeg -f video4linux2 -input_format mjpeg -i /dev/video0 out.mpeg
8543 For more information about Video4Linux, check <http://linuxtv.org/>.
8545 Options
8547 standard
8549 Set the standard. Must be the name of a supported standard. To get
8550 a list of the supported standards, use the list_standards option.
8551 channel
8553 Set the input channel number. Default to -1, which means using the
8554 previously selected channel.
8555 video_size
8557 Set the video frame size. The argument must be a string in the form
8558 WIDTHxHEIGHT or a valid size abbreviation.
8559 pixel_format
8561 Select the pixel format (only valid for raw video input).
8562 input_format
8564 Set the preferred pixel format (for raw video) or a codec name.
8565 This option allows one to select the input format, when several are
8566 available.
8567 framerate
8569 Set the preferred video frame rate.
8570 list_formats
8572 List available formats (supported pixel formats, codecs, and frame
8573 sizes) and exit.
8574 Available values are:
8576 all Show all available (compressed and non-compressed) formats.
8578 raw Show only raw video (non-compressed) formats.
8580 compressed
8582 Show only compressed formats.
8583 list_standards
8585 List supported standards and exit.
8586 Available values are:
8588 all Show all supported standards.
8590 timestamps, ts
8592 Set type of timestamps for grabbed frames.
8593 Available values are:
8595 default
8597 Use timestamps from the kernel.
8598 abs Use absolute timestamps (wall clock).
8600 mono2abs
8602 Force conversion from monotonic to absolute timestamps.
8603 Default value is "default".
8605 use_libv4l2
8607 Use libv4l2 (v4l-utils) conversion functions. Default is 0.
8608 vfwcap
8610 VfW (Video for Windows) capture input device.
8611 The filename passed as input is the capture driver number, ranging from
8613 0 to 9. You may use "list" as filename to print a list of drivers. Any
8614 other filename will be interpreted as device number 0.
8615 Options
8617 video_size
8619 Set the video frame size.
8620 framerate
8622 Set the grabbing frame rate. Default value is "ntsc", corresponding
8623 to a frame rate of "30000/1001".
8624 x11grab
8626 X11 video input device.
8627 To enable this input device during configuration you need libxcb
8629 installed on your system. It will be automatically detected during
8630 configuration.
8631 This device allows one to capture a region of an X11 display.
8633 The filename passed as input has the syntax:
8635 [<hostname>]:<display_number>.<screen_number>[+<x_offset>,<y_offset>]
8637 hostname:display_number.screen_number specifies the X11 display name of
8639 the screen to grab from. hostname can be omitted, and defaults to
8640 "localhost". The environment variable DISPLAY contains the default
8641 display name.
8642 x_offset and y_offset specify the offsets of the grabbed area with
8644 respect to the top-left border of the X11 screen. They default to 0.
8645 Check the X11 documentation (e.g. man X) for more detailed information.
8647 Use the xdpyinfo program for getting basic information about the
8649 properties of your X11 display (e.g. grep for "name" or "dimensions").
8650 For example to grab from :0.0 using ffmpeg:
8652 ffmpeg -f x11grab -framerate 25 -video_size cif -i :0.0 out.mpg
8654 Grab at position "10,20":
8656 ffmpeg -f x11grab -framerate 25 -video_size cif -i :0.0+10,20 out.mpg
8658 Options
8660 select_region
8662 Specify whether to select the grabbing area graphically using the
8663 pointer. A value of 1 prompts the user to select the grabbing area
8664 graphically by clicking and dragging. A single click with no
8665 dragging will select the whole screen. A region with zero width or
8666 height will also select the whole screen. This option overwrites
8667 the video_size, grab_x, and grab_y options. Default value is 0.
8668 draw_mouse
8670 Specify whether to draw the mouse pointer. A value of 0 specifies
8671 not to draw the pointer. Default value is 1.
8672 follow_mouse
8674 Make the grabbed area follow the mouse. The argument can be
8675 "centered" or a number of pixels PIXELS.
8676 When it is specified with "centered", the grabbing region follows
8678 the mouse pointer and keeps the pointer at the center of region;
8679 otherwise, the region follows only when the mouse pointer reaches
8680 within PIXELS (greater than zero) to the edge of region.
8681 For example:
8683 ffmpeg -f x11grab -follow_mouse centered -framerate 25 -video_size cif -i :0.0 out.mpg
8685 To follow only when the mouse pointer reaches within 100 pixels to
8687 edge:
8688 ffmpeg -f x11grab -follow_mouse 100 -framerate 25 -video_size cif -i :0.0 out.mpg
8690 framerate
8692 Set the grabbing frame rate. Default value is "ntsc", corresponding
8693 to a frame rate of "30000/1001".
8694 show_region
8696 Show grabbed region on screen.
8697 If show_region is specified with 1, then the grabbing region will
8699 be indicated on screen. With this option, it is easy to know what
8700 is being grabbed if only a portion of the screen is grabbed.
8701 region_border
8703 Set the region border thickness if -show_region 1 is used. Range
8704 is 1 to 128 and default is 3 (XCB-based x11grab only).
8705 For example:
8707 ffmpeg -f x11grab -show_region 1 -framerate 25 -video_size cif -i :0.0+10,20 out.mpg
8709 With follow_mouse:
8711 ffmpeg -f x11grab -follow_mouse centered -show_region 1 -framerate 25 -video_size cif -i :0.0 out.mpg
8713 window_id
8715 Grab this window, instead of the whole screen. Default value is 0,
8716 which maps to the whole screen (root window).
8717 The id of a window can be found using the xwininfo program,
8719 possibly with options -tree and -root.
8720 If the window is later enlarged, the new area is not recorded.
8722 Video ends when the window is closed, unmapped (i.e., iconified) or
8723 shrunk beyond the video size (which defaults to the initial window
8724 size).
8725 This option disables options follow_mouse and select_region.
8727 video_size
8729 Set the video frame size. Default is the full desktop or window.
8730 grab_x
8732 grab_y
8733 Set the grabbing region coordinates. They are expressed as offset
8734 from the top left corner of the X11 window and correspond to the
8735 x_offset and y_offset parameters in the device name. The default
8736 value for both options is 0.
8737
8739 The audio resampler supports the following named options.
8740 Options may be set by specifying -option value in the FFmpeg tools,
8742 option=value for the aresample filter, by setting the value explicitly
8743 in the "SwrContext" options or using the libavutil/opt.h API for
8744 programmatic use.
8745 uchl, used_chlayout
8747 Set used input channel layout. Default is unset. This option is
8748 only used for special remapping.
8749 isr, in_sample_rate
8751 Set the input sample rate. Default value is 0.
8752 osr, out_sample_rate
8754 Set the output sample rate. Default value is 0.
8755 isf, in_sample_fmt
8757 Specify the input sample format. It is set by default to "none".
8758 osf, out_sample_fmt
8760 Specify the output sample format. It is set by default to "none".
8761 tsf, internal_sample_fmt
8763 Set the internal sample format. Default value is "none". This will
8764 automatically be chosen when it is not explicitly set.
8765 ichl, in_chlayout
8767 ochl, out_chlayout
8768 Set the input/output channel layout.
8769 See the Channel Layout section in the ffmpeg-utils(1) manual for
8771 the required syntax.
8772 clev, center_mix_level
8774 Set the center mix level. It is a value expressed in deciBel, and
8775 must be in the interval [-32,32].
8776 slev, surround_mix_level
8778 Set the surround mix level. It is a value expressed in deciBel, and
8779 must be in the interval [-32,32].
8780 lfe_mix_level
8782 Set LFE mix into non LFE level. It is used when there is a LFE
8783 input but no LFE output. It is a value expressed in deciBel, and
8784 must be in the interval [-32,32].
8785 rmvol, rematrix_volume
8787 Set rematrix volume. Default value is 1.0.
8788 rematrix_maxval
8790 Set maximum output value for rematrixing. This can be used to
8791 prevent clipping vs. preventing volume reduction. A value of 1.0
8792 prevents clipping.
8793 flags, swr_flags
8795 Set flags used by the converter. Default value is 0.
8796 It supports the following individual flags:
8798 res force resampling, this flag forces resampling to be used even
8800 when the input and output sample rates match.
8801 dither_scale
8803 Set the dither scale. Default value is 1.
8804 dither_method
8806 Set dither method. Default value is 0.
8807 Supported values:
8809 rectangular
8811 select rectangular dither
8812 triangular
8814 select triangular dither
8815 triangular_hp
8817 select triangular dither with high pass
8818 lipshitz
8820 select Lipshitz noise shaping dither.
8821 shibata
8823 select Shibata noise shaping dither.
8824 low_shibata
8826 select low Shibata noise shaping dither.
8827 high_shibata
8829 select high Shibata noise shaping dither.
8830 f_weighted
8832 select f-weighted noise shaping dither
8833 modified_e_weighted
8835 select modified-e-weighted noise shaping dither
8836 improved_e_weighted
8838 select improved-e-weighted noise shaping dither
8839 resampler
8841 Set resampling engine. Default value is swr.
8842 Supported values:
8844 swr select the native SW Resampler; filter options precision and
8846 cheby are not applicable in this case.
8847 soxr
8849 select the SoX Resampler (where available); compensation, and
8850 filter options filter_size, phase_shift, exact_rational,
8851 filter_type & kaiser_beta, are not applicable in this case.
8852 filter_size
8854 For swr only, set resampling filter size, default value is 32.
8855 phase_shift
8857 For swr only, set resampling phase shift, default value is 10, and
8858 must be in the interval [0,30].
8859 linear_interp
8861 Use linear interpolation when enabled (the default). Disable it if
8862 you want to preserve speed instead of quality when exact_rational
8863 fails.
8864 exact_rational
8866 For swr only, when enabled, try to use exact phase_count based on
8867 input and output sample rate. However, if it is larger than "1 <<
8868 phase_shift", the phase_count will be "1 << phase_shift" as
8869 fallback. Default is enabled.
8870 cutoff
8872 Set cutoff frequency (swr: 6dB point; soxr: 0dB point) ratio; must
8873 be a float value between 0 and 1. Default value is 0.97 with swr,
8874 and 0.91 with soxr (which, with a sample-rate of 44100, preserves
8875 the entire audio band to 20kHz).
8876 precision
8878 For soxr only, the precision in bits to which the resampled signal
8879 will be calculated. The default value of 20 (which, with suitable
8880 dithering, is appropriate for a destination bit-depth of 16) gives
8881 SoX's 'High Quality'; a value of 28 gives SoX's 'Very High
8882 Quality'.
8883 cheby
8885 For soxr only, selects passband rolloff none (Chebyshev) & higher-
8886 precision approximation for 'irrational' ratios. Default value is
8887 0.
8888 async
8890 For swr only, simple 1 parameter audio sync to timestamps using
8891 stretching, squeezing, filling and trimming. Setting this to 1 will
8892 enable filling and trimming, larger values represent the maximum
8893 amount in samples that the data may be stretched or squeezed for
8894 each second. Default value is 0, thus no compensation is applied
8895 to make the samples match the audio timestamps.
8896 first_pts
8898 For swr only, assume the first pts should be this value. The time
8899 unit is 1 / sample rate. This allows for padding/trimming at the
8900 start of stream. By default, no assumption is made about the first
8901 frame's expected pts, so no padding or trimming is done. For
8902 example, this could be set to 0 to pad the beginning with silence
8903 if an audio stream starts after the video stream or to trim any
8904 samples with a negative pts due to encoder delay.
8905 min_comp
8907 For swr only, set the minimum difference between timestamps and
8908 audio data (in seconds) to trigger stretching/squeezing/filling or
8909 trimming of the data to make it match the timestamps. The default
8910 is that stretching/squeezing/filling and trimming is disabled
8911 (min_comp = "FLT_MAX").
8912 min_hard_comp
8914 For swr only, set the minimum difference between timestamps and
8915 audio data (in seconds) to trigger adding/dropping samples to make
8916 it match the timestamps. This option effectively is a threshold to
8917 select between hard (trim/fill) and soft (squeeze/stretch)
8918 compensation. Note that all compensation is by default disabled
8919 through min_comp. The default is 0.1.
8920 comp_duration
8922 For swr only, set duration (in seconds) over which data is
8923 stretched/squeezed to make it match the timestamps. Must be a non-
8924 negative double float value, default value is 1.0.
8925 max_soft_comp
8927 For swr only, set maximum factor by which data is
8928 stretched/squeezed to make it match the timestamps. Must be a non-
8929 negative double float value, default value is 0.
8930 matrix_encoding
8932 Select matrixed stereo encoding.
8933 It accepts the following values:
8935 none
8937 select none
8938 dolby
8940 select Dolby
8941 dplii
8943 select Dolby Pro Logic II
8944 Default value is "none".
8946 filter_type
8948 For swr only, select resampling filter type. This only affects
8949 resampling operations.
8950 It accepts the following values:
8952 cubic
8954 select cubic
8955 blackman_nuttall
8957 select Blackman Nuttall windowed sinc
8958 kaiser
8960 select Kaiser windowed sinc
8961 kaiser_beta
8963 For swr only, set Kaiser window beta value. Must be a double float
8964 value in the interval [2,16], default value is 9.
8965 output_sample_bits
8967 For swr only, set number of used output sample bits for dithering.
8968 Must be an integer in the interval [0,64], default value is 0,
8969 which means it's not used.
8970
8972 The video scaler supports the following named options.
8973 Options may be set by specifying -option value in the FFmpeg tools,
8975 with a few API-only exceptions noted below. For programmatic use, they
8976 can be set explicitly in the "SwsContext" options or through the
8977 libavutil/opt.h API.
8978 sws_flags
8980 Set the scaler flags. This is also used to set the scaling
8981 algorithm. Only a single algorithm should be selected. Default
8982 value is bicubic.
8983 It accepts the following values:
8985 fast_bilinear
8987 Select fast bilinear scaling algorithm.
8988 bilinear
8990 Select bilinear scaling algorithm.
8991 bicubic
8993 Select bicubic scaling algorithm.
8994 experimental
8996 Select experimental scaling algorithm.
8997 neighbor
8999 Select nearest neighbor rescaling algorithm.
9000 area
9002 Select averaging area rescaling algorithm.
9003 bicublin
9005 Select bicubic scaling algorithm for the luma component,
9006 bilinear for chroma components.
9007 gauss
9009 Select Gaussian rescaling algorithm.
9010 sinc
9012 Select sinc rescaling algorithm.
9013 lanczos
9015 Select Lanczos rescaling algorithm. The default width (alpha)
9016 is 3 and can be changed by setting "param0".
9017 spline
9019 Select natural bicubic spline rescaling algorithm.
9020 print_info
9022 Enable printing/debug logging.
9023 accurate_rnd
9025 Enable accurate rounding.
9026 full_chroma_int
9028 Enable full chroma interpolation.
9029 full_chroma_inp
9031 Select full chroma input.
9032 bitexact
9034 Enable bitexact output.
9035 srcw (API only)
9037 Set source width.
9038 srch (API only)
9040 Set source height.
9041 dstw (API only)
9043 Set destination width.
9044 dsth (API only)
9046 Set destination height.
9047 src_format (API only)
9049 Set source pixel format (must be expressed as an integer).
9050 dst_format (API only)
9052 Set destination pixel format (must be expressed as an integer).
9053 src_range (boolean)
9055 If value is set to 1, indicates source is full range. Default value
9056 is 0, which indicates source is limited range.
9057 dst_range (boolean)
9059 If value is set to 1, enable full range for destination. Default
9060 value is 0, which enables limited range.
9061 param0, param1
9063 Set scaling algorithm parameters. The specified values are specific
9064 of some scaling algorithms and ignored by others. The specified
9065 values are floating point number values.
9066 sws_dither
9068 Set the dithering algorithm. Accepts one of the following values.
9069 Default value is auto.
9070 auto
9072 automatic choice
9073 none
9075 no dithering
9076 bayer
9078 bayer dither
9079 ed error diffusion dither
9081 a_dither
9083 arithmetic dither, based using addition
9084 x_dither
9086 arithmetic dither, based using xor (more random/less apparent
9087 patterning that a_dither).
9088 alphablend
9090 Set the alpha blending to use when the input has alpha but the
9091 output does not. Default value is none.
9092 uniform_color
9094 Blend onto a uniform background color
9095 checkerboard
9097 Blend onto a checkerboard
9098 none
9100 No blending
9101
9103 Filtering in FFmpeg is enabled through the libavfilter library.
9104 In libavfilter, a filter can have multiple inputs and multiple outputs.
9106 To illustrate the sorts of things that are possible, we consider the
9107 following filtergraph.
9108 [main]
9110 input --> split ---------------------> overlay --> output
9111 | ^
9112 |[tmp] [flip]|
9113 +-----> crop --> vflip -------+
9114 This filtergraph splits the input stream in two streams, then sends one
9116 stream through the crop filter and the vflip filter, before merging it
9117 back with the other stream by overlaying it on top. You can use the
9118 following command to achieve this:
9119 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
9121 The result will be that the top half of the video is mirrored onto the
9123 bottom half of the output video.
9124 Filters in the same linear chain are separated by commas, and distinct
9126 linear chains of filters are separated by semicolons. In our example,
9127 crop,vflip are in one linear chain, split and overlay are separately in
9128 another. The points where the linear chains join are labelled by names
9129 enclosed in square brackets. In the example, the split filter generates
9130 two outputs that are associated to the labels [main] and [tmp].
9131 The stream sent to the second output of split, labelled as [tmp], is
9133 processed through the crop filter, which crops away the lower half part
9134 of the video, and then vertically flipped. The overlay filter takes in
9135 input the first unchanged output of the split filter (which was
9136 labelled as [main]), and overlay on its lower half the output generated
9137 by the crop,vflip filterchain.
9138 Some filters take in input a list of parameters: they are specified
9140 after the filter name and an equal sign, and are separated from each
9141 other by a colon.
9142 There exist so-called source filters that do not have an audio/video
9144 input, and sink filters that will not have audio/video output.
9145
9147 The graph2dot program included in the FFmpeg tools directory can be
9148 used to parse a filtergraph description and issue a corresponding
9149 textual representation in the dot language.
9150 Invoke the command:
9152 graph2dot -h
9154 to see how to use graph2dot.
9156 You can then pass the dot description to the dot program (from the
9158 graphviz suite of programs) and obtain a graphical representation of
9159 the filtergraph.
9160 For example the sequence of commands:
9162 echo <GRAPH_DESCRIPTION> | \
9164 tools/graph2dot -o graph.tmp && \
9165 dot -Tpng graph.tmp -o graph.png && \
9166 display graph.png
9167 can be used to create and display an image representing the graph
9169 described by the GRAPH_DESCRIPTION string. Note that this string must
9170 be a complete self-contained graph, with its inputs and outputs
9171 explicitly defined. For example if your command line is of the form:
9172 ffmpeg -i infile -vf scale=640:360 outfile
9174 your GRAPH_DESCRIPTION string will need to be of the form:
9176 nullsrc,scale=640:360,nullsink
9178 you may also need to set the nullsrc parameters and add a format filter
9180 in order to simulate a specific input file.
9181
9183 A filtergraph is a directed graph of connected filters. It can contain
9184 cycles, and there can be multiple links between a pair of filters. Each
9185 link has one input pad on one side connecting it to one filter from
9186 which it takes its input, and one output pad on the other side
9187 connecting it to one filter accepting its output.
9188 Each filter in a filtergraph is an instance of a filter class
9190 registered in the application, which defines the features and the
9191 number of input and output pads of the filter.
9192 A filter with no input pads is called a "source", and a filter with no
9194 output pads is called a "sink".
9195 Filtergraph syntax
9197 A filtergraph has a textual representation, which is recognized by the
9198 -filter/-vf/-af and -filter_complex options in ffmpeg and -vf/-af in
9199 ffplay, and by the avfilter_graph_parse_ptr() function defined in
9200 libavfilter/avfilter.h.
9201 A filterchain consists of a sequence of connected filters, each one
9203 connected to the previous one in the sequence. A filterchain is
9204 represented by a list of ","-separated filter descriptions.
9205 A filtergraph consists of a sequence of filterchains. A sequence of
9207 filterchains is represented by a list of ";"-separated filterchain
9208 descriptions.
9209 A filter is represented by a string of the form:
9211 [in_link_1]...[in_link_N]filter_name@id=arguments[out_link_1]...[out_link_M]
9212 filter_name is the name of the filter class of which the described
9214 filter is an instance of, and has to be the name of one of the filter
9215 classes registered in the program optionally followed by "@id". The
9216 name of the filter class is optionally followed by a string
9217 "=arguments".
9218 arguments is a string which contains the parameters used to initialize
9220 the filter instance. It may have one of two forms:
9221 • A ':'-separated list of key=value pairs.
9223 • A ':'-separated list of value. In this case, the keys are assumed
9225 to be the option names in the order they are declared. E.g. the
9226 "fade" filter declares three options in this order -- type,
9227 start_frame and nb_frames. Then the parameter list in:0:30 means
9228 that the value in is assigned to the option type, 0 to start_frame
9229 and 30 to nb_frames.
9230 • A ':'-separated list of mixed direct value and long key=value
9232 pairs. The direct value must precede the key=value pairs, and
9233 follow the same constraints order of the previous point. The
9234 following key=value pairs can be set in any preferred order.
9235 If the option value itself is a list of items (e.g. the "format" filter
9237 takes a list of pixel formats), the items in the list are usually
9238 separated by |.
9239 The list of arguments can be quoted using the character ' as initial
9241 and ending mark, and the character \ for escaping the characters within
9242 the quoted text; otherwise the argument string is considered terminated
9243 when the next special character (belonging to the set []=;,) is
9244 encountered.
9245 A special syntax implemented in the ffmpeg CLI tool allows loading
9247 option values from files. This is done be prepending a slash '/' to the
9248 option name, then the supplied value is interpreted as a path from
9249 which the actual value is loaded. E.g.
9250 ffmpeg -i <INPUT> -vf drawtext=/text=/tmp/some_text <OUTPUT>
9252 will load the text to be drawn from /tmp/some_text. API users wishing
9254 to implement a similar feature should use the
9255 "avfilter_graph_segment_*()" functions together with custom IO code.
9256 The name and arguments of the filter are optionally preceded and
9258 followed by a list of link labels. A link label allows one to name a
9259 link and associate it to a filter output or input pad. The preceding
9260 labels in_link_1 ... in_link_N, are associated to the filter input
9261 pads, the following labels out_link_1 ... out_link_M, are associated to
9262 the output pads.
9263 When two link labels with the same name are found in the filtergraph, a
9265 link between the corresponding input and output pad is created.
9266 If an output pad is not labelled, it is linked by default to the first
9268 unlabelled input pad of the next filter in the filterchain. For
9269 example in the filterchain
9270 nullsrc, split[L1], [L2]overlay, nullsink
9272 the split filter instance has two output pads, and the overlay filter
9274 instance two input pads. The first output pad of split is labelled
9275 "L1", the first input pad of overlay is labelled "L2", and the second
9276 output pad of split is linked to the second input pad of overlay, which
9277 are both unlabelled.
9278 In a filter description, if the input label of the first filter is not
9280 specified, "in" is assumed; if the output label of the last filter is
9281 not specified, "out" is assumed.
9282 In a complete filterchain all the unlabelled filter input and output
9284 pads must be connected. A filtergraph is considered valid if all the
9285 filter input and output pads of all the filterchains are connected.
9286 Libavfilter will automatically insert scale filters where format
9288 conversion is required. It is possible to specify swscale flags for
9289 those automatically inserted scalers by prepending "sws_flags=flags;"
9290 to the filtergraph description.
9291 Here is a BNF description of the filtergraph syntax:
9293 <NAME> ::= sequence of alphanumeric characters and '_'
9295 <FILTER_NAME> ::= <NAME>["@"<NAME>]
9296 <LINKLABEL> ::= "[" <NAME> "]"
9297 <LINKLABELS> ::= <LINKLABEL> [<LINKLABELS>]
9298 <FILTER_ARGUMENTS> ::= sequence of chars (possibly quoted)
9299 <FILTER> ::= [<LINKLABELS>] <FILTER_NAME> ["=" <FILTER_ARGUMENTS>] [<LINKLABELS>]
9300 <FILTERCHAIN> ::= <FILTER> [,<FILTERCHAIN>]
9301 <FILTERGRAPH> ::= [sws_flags=<flags>;] <FILTERCHAIN> [;<FILTERGRAPH>]
9302 Notes on filtergraph escaping
9304 Filtergraph description composition entails several levels of escaping.
9305 See the "Quoting and escaping" section in the ffmpeg-utils(1) manual
9306 for more information about the employed escaping procedure.
9307 A first level escaping affects the content of each filter option value,
9309 which may contain the special character ":" used to separate values, or
9310 one of the escaping characters "\'".
9311 A second level escaping affects the whole filter description, which may
9313 contain the escaping characters "\'" or the special characters "[],;"
9314 used by the filtergraph description.
9315 Finally, when you specify a filtergraph on a shell commandline, you
9317 need to perform a third level escaping for the shell special characters
9318 contained within it.
9319 For example, consider the following string to be embedded in the
9321 drawtext filter description text value:
9322 this is a 'string': may contain one, or more, special characters
9324 This string contains the "'" special escaping character, and the ":"
9326 special character, so it needs to be escaped in this way:
9327 text=this is a \'string\'\: may contain one, or more, special characters
9329 A second level of escaping is required when embedding the filter
9331 description in a filtergraph description, in order to escape all the
9332 filtergraph special characters. Thus the example above becomes:
9333 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
9335 (note that in addition to the "\'" escaping special characters, also
9337 "," needs to be escaped).
9338 Finally an additional level of escaping is needed when writing the
9340 filtergraph description in a shell command, which depends on the
9341 escaping rules of the adopted shell. For example, assuming that "\" is
9342 special and needs to be escaped with another "\", the previous string
9343 will finally result in:
9344 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
9346
9348 Some filters support a generic enable option. For the filters
9349 supporting timeline editing, this option can be set to an expression
9350 which is evaluated before sending a frame to the filter. If the
9351 evaluation is non-zero, the filter will be enabled, otherwise the frame
9352 will be sent unchanged to the next filter in the filtergraph.
9353 The expression accepts the following values:
9355 t timestamp expressed in seconds, NAN if the input timestamp is
9357 unknown
9358 n sequential number of the input frame, starting from 0
9360 pos the position in the file of the input frame, NAN if unknown
9362 w
9364 h width and height of the input frame if video
9365 Additionally, these filters support an enable command that can be used
9367 to re-define the expression.
9368 Like any other filtering option, the enable option follows the same
9370 rules.
9371 For example, to enable a blur filter (smartblur) from 10 seconds to 3
9373 minutes, and a curves filter starting at 3 seconds:
9374 smartblur = enable='between(t,10,3*60)',
9376 curves = enable='gte(t,3)' : preset=cross_process
9377 See "ffmpeg -filters" to view which filters have timeline support.
9379
9381 Some options can be changed during the operation of the filter using a
9382 command. These options are marked 'T' on the output of ffmpeg -h
9383 filter=<name of filter>. The name of the command is the name of the
9384 option and the argument is the new value.
9385
9387 Some filters with several inputs support a common set of options.
9388 These options can only be set by name, not with the short notation.
9389 eof_action
9391 The action to take when EOF is encountered on the secondary input;
9392 it accepts one of the following values:
9393 repeat
9395 Repeat the last frame (the default).
9396 endall
9398 End both streams.
9399 pass
9401 Pass the main input through.
9402 shortest
9404 If set to 1, force the output to terminate when the shortest input
9405 terminates. Default value is 0.
9406 repeatlast
9408 If set to 1, force the filter to extend the last frame of secondary
9409 streams until the end of the primary stream. A value of 0 disables
9410 this behavior. Default value is 1.
9411 ts_sync_mode
9413 How strictly to sync streams based on secondary input timestamps;
9414 it accepts one of the following values:
9415 default
9417 Frame from secondary input with the nearest lower or equal
9418 timestamp to the primary input frame.
9419 nearest
9421 Frame from secondary input with the absolute nearest timestamp
9422 to the primary input frame.
9423
9425 When you configure your FFmpeg build, you can disable any of the
9426 existing filters using "--disable-filters". The configure output will
9427 show the audio filters included in your build.
9428 Below is a description of the currently available audio filters.
9430 acompressor
9432 A compressor is mainly used to reduce the dynamic range of a signal.
9433 Especially modern music is mostly compressed at a high ratio to improve
9434 the overall loudness. It's done to get the highest attention of a
9435 listener, "fatten" the sound and bring more "power" to the track. If a
9436 signal is compressed too much it may sound dull or "dead" afterwards or
9437 it may start to "pump" (which could be a powerful effect but can also
9438 destroy a track completely). The right compression is the key to reach
9439 a professional sound and is the high art of mixing and mastering.
9440 Because of its complex settings it may take a long time to get the
9441 right feeling for this kind of effect.
9442 Compression is done by detecting the volume above a chosen level
9444 "threshold" and dividing it by the factor set with "ratio". So if you
9445 set the threshold to -12dB and your signal reaches -6dB a ratio of 2:1
9446 will result in a signal at -9dB. Because an exact manipulation of the
9447 signal would cause distortion of the waveform the reduction can be
9448 levelled over the time. This is done by setting "Attack" and "Release".
9449 "attack" determines how long the signal has to rise above the threshold
9450 before any reduction will occur and "release" sets the time the signal
9451 has to fall below the threshold to reduce the reduction again. Shorter
9452 signals than the chosen attack time will be left untouched. The
9453 overall reduction of the signal can be made up afterwards with the
9454 "makeup" setting. So compressing the peaks of a signal about 6dB and
9455 raising the makeup to this level results in a signal twice as loud than
9456 the source. To gain a softer entry in the compression the "knee"
9457 flattens the hard edge at the threshold in the range of the chosen
9458 decibels.
9459 The filter accepts the following options:
9461 level_in
9463 Set input gain. Default is 1. Range is between 0.015625 and 64.
9464 mode
9466 Set mode of compressor operation. Can be "upward" or "downward".
9467 Default is "downward".
9468 threshold
9470 If a signal of stream rises above this level it will affect the
9471 gain reduction. By default it is 0.125. Range is between
9472 0.00097563 and 1.
9473 ratio
9475 Set a ratio by which the signal is reduced. 1:2 means that if the
9476 level rose 4dB above the threshold, it will be only 2dB above after
9477 the reduction. Default is 2. Range is between 1 and 20.
9478 attack
9480 Amount of milliseconds the signal has to rise above the threshold
9481 before gain reduction starts. Default is 20. Range is between 0.01
9482 and 2000.
9483 release
9485 Amount of milliseconds the signal has to fall below the threshold
9486 before reduction is decreased again. Default is 250. Range is
9487 between 0.01 and 9000.
9488 makeup
9490 Set the amount by how much signal will be amplified after
9491 processing. Default is 1. Range is from 1 to 64.
9492 knee
9494 Curve the sharp knee around the threshold to enter gain reduction
9495 more softly. Default is 2.82843. Range is between 1 and 8.
9496 link
9498 Choose if the "average" level between all channels of input stream
9499 or the louder("maximum") channel of input stream affects the
9500 reduction. Default is "average".
9501 detection
9503 Should the exact signal be taken in case of "peak" or an RMS one in
9504 case of "rms". Default is "rms" which is mostly smoother.
9505 mix How much to use compressed signal in output. Default is 1. Range
9507 is between 0 and 1.
9508 Commands
9510 This filter supports the all above options as commands.
9512 acontrast
9514 Simple audio dynamic range compression/expansion filter.
9515 The filter accepts the following options:
9517 contrast
9519 Set contrast. Default is 33. Allowed range is between 0 and 100.
9520 acopy
9522 Copy the input audio source unchanged to the output. This is mainly
9523 useful for testing purposes.
9524 acrossfade
9526 Apply cross fade from one input audio stream to another input audio
9527 stream. The cross fade is applied for specified duration near the end
9528 of first stream.
9529 The filter accepts the following options:
9531 nb_samples, ns
9533 Specify the number of samples for which the cross fade effect has
9534 to last. At the end of the cross fade effect the first input audio
9535 will be completely silent. Default is 44100.
9536 duration, d
9538 Specify the duration of the cross fade effect. See the Time
9539 duration section in the ffmpeg-utils(1) manual for the accepted
9540 syntax. By default the duration is determined by nb_samples. If
9541 set this option is used instead of nb_samples.
9542 overlap, o
9544 Should first stream end overlap with second stream start. Default
9545 is enabled.
9546 curve1
9548 Set curve for cross fade transition for first stream.
9549 curve2
9551 Set curve for cross fade transition for second stream.
9552 For description of available curve types see afade filter
9554 description.
9555 Examples
9557 • Cross fade from one input to another:
9559 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
9561 • Cross fade from one input to another but without overlapping:
9563 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
9565 acrossover
9567 Split audio stream into several bands.
9568 This filter splits audio stream into two or more frequency ranges.
9570 Summing all streams back will give flat output.
9571 The filter accepts the following options:
9573 split
9575 Set split frequencies. Those must be positive and increasing.
9576 order
9578 Set filter order for each band split. This controls filter roll-off
9579 or steepness of filter transfer function. Available values are:
9580 2nd 12 dB per octave.
9582 4th 24 dB per octave.
9584 6th 36 dB per octave.
9586 8th 48 dB per octave.
9588 10th
9590 60 dB per octave.
9591 12th
9593 72 dB per octave.
9594 14th
9596 84 dB per octave.
9597 16th
9599 96 dB per octave.
9600 18th
9602 108 dB per octave.
9603 20th
9605 120 dB per octave.
9606 Default is 4th.
9608 level
9610 Set input gain level. Allowed range is from 0 to 1. Default value
9611 is 1.
9612 gains
9614 Set output gain for each band. Default value is 1 for all bands.
9615 precision
9617 Set which precision to use when processing samples.
9618 auto
9620 Auto pick internal sample format depending on other filters.
9621 float
9623 Always use single-floating point precision sample format.
9624 double
9626 Always use double-floating point precision sample format.
9627 Default value is "auto".
9629 Examples
9631 • Split input audio stream into two bands (low and high) with split
9633 frequency of 1500 Hz, each band will be in separate stream:
9634 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
9636 • Same as above, but with higher filter order:
9638 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
9640 • Same as above, but also with additional middle band (frequencies
9642 between 1500 and 8000):
9643 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
9645 acrusher
9647 Reduce audio bit resolution.
9648 This filter is bit crusher with enhanced functionality. A bit crusher
9650 is used to audibly reduce number of bits an audio signal is sampled
9651 with. This doesn't change the bit depth at all, it just produces the
9652 effect. Material reduced in bit depth sounds more harsh and "digital".
9653 This filter is able to even round to continuous values instead of
9654 discrete bit depths. Additionally it has a D/C offset which results in
9655 different crushing of the lower and the upper half of the signal. An
9656 Anti-Aliasing setting is able to produce "softer" crushing sounds.
9657 Another feature of this filter is the logarithmic mode. This setting
9659 switches from linear distances between bits to logarithmic ones. The
9660 result is a much more "natural" sounding crusher which doesn't gate low
9661 signals for example. The human ear has a logarithmic perception, so
9662 this kind of crushing is much more pleasant. Logarithmic crushing is
9663 also able to get anti-aliased.
9664 The filter accepts the following options:
9666 level_in
9668 Set level in.
9669 level_out
9671 Set level out.
9672 bits
9674 Set bit reduction.
9675 mix Set mixing amount.
9677 mode
9679 Can be linear: "lin" or logarithmic: "log".
9680 dc Set DC.
9682 aa Set anti-aliasing.
9684 samples
9686 Set sample reduction.
9687 lfo Enable LFO. By default disabled.
9689 lforange
9691 Set LFO range.
9692 lforate
9694 Set LFO rate.
9695 Commands
9697 This filter supports the all above options as commands.
9699 acue
9701 Delay audio filtering until a given wallclock timestamp. See the cue
9702 filter.
9703 adeclick
9705 Remove impulsive noise from input audio.
9706 Samples detected as impulsive noise are replaced by interpolated
9708 samples using autoregressive modelling.
9709 window, w
9711 Set window size, in milliseconds. Allowed range is from 10 to 100.
9712 Default value is 55 milliseconds. This sets size of window which
9713 will be processed at once.
9714 overlap, o
9716 Set window overlap, in percentage of window size. Allowed range is
9717 from 50 to 95. Default value is 75 percent. Setting this to a very
9718 high value increases impulsive noise removal but makes whole
9719 process much slower.
9720 arorder, a
9722 Set autoregression order, in percentage of window size. Allowed
9723 range is from 0 to 25. Default value is 2 percent. This option also
9724 controls quality of interpolated samples using neighbour good
9725 samples.
9726 threshold, t
9728 Set threshold value. Allowed range is from 1 to 100. Default value
9729 is 2. This controls the strength of impulsive noise which is going
9730 to be removed. The lower value, the more samples will be detected
9731 as impulsive noise.
9732 burst, b
9734 Set burst fusion, in percentage of window size. Allowed range is 0
9735 to 10. Default value is 2. If any two samples detected as noise
9736 are spaced less than this value then any sample between those two
9737 samples will be also detected as noise.
9738 method, m
9740 Set overlap method.
9741 It accepts the following values:
9743 add, a
9745 Select overlap-add method. Even not interpolated samples are
9746 slightly changed with this method.
9747 save, s
9749 Select overlap-save method. Not interpolated samples remain
9750 unchanged.
9751 Default value is "a".
9753 adeclip
9755 Remove clipped samples from input audio.
9756 Samples detected as clipped are replaced by interpolated samples using
9758 autoregressive modelling.
9759 window, w
9761 Set window size, in milliseconds. Allowed range is from 10 to 100.
9762 Default value is 55 milliseconds. This sets size of window which
9763 will be processed at once.
9764 overlap, o
9766 Set window overlap, in percentage of window size. Allowed range is
9767 from 50 to 95. Default value is 75 percent.
9768 arorder, a
9770 Set autoregression order, in percentage of window size. Allowed
9771 range is from 0 to 25. Default value is 8 percent. This option also
9772 controls quality of interpolated samples using neighbour good
9773 samples.
9774 threshold, t
9776 Set threshold value. Allowed range is from 1 to 100. Default value
9777 is 10. Higher values make clip detection less aggressive.
9778 hsize, n
9780 Set size of histogram used to detect clips. Allowed range is from
9781 100 to 9999. Default value is 1000. Higher values make clip
9782 detection less aggressive.
9783 method, m
9785 Set overlap method.
9786 It accepts the following values:
9788 add, a
9790 Select overlap-add method. Even not interpolated samples are
9791 slightly changed with this method.
9792 save, s
9794 Select overlap-save method. Not interpolated samples remain
9795 unchanged.
9796 Default value is "a".
9798 adecorrelate
9800 Apply decorrelation to input audio stream.
9801 The filter accepts the following options:
9803 stages
9805 Set decorrelation stages of filtering. Allowed range is from 1 to
9806 16. Default value is 6.
9807 seed
9809 Set random seed used for setting delay in samples across channels.
9810 adelay
9812 Delay one or more audio channels.
9813 Samples in delayed channel are filled with silence.
9815 The filter accepts the following option:
9817 delays
9819 Set list of delays in milliseconds for each channel separated by
9820 '|'. Unused delays will be silently ignored. If number of given
9821 delays is smaller than number of channels all remaining channels
9822 will not be delayed. If you want to delay exact number of samples,
9823 append 'S' to number. If you want instead to delay in seconds,
9824 append 's' to number.
9825 all Use last set delay for all remaining channels. By default is
9827 disabled. This option if enabled changes how option "delays" is
9828 interpreted.
9829 Examples
9831 • Delay first channel by 1.5 seconds, the third channel by 0.5
9833 seconds and leave the second channel (and any other channels that
9834 may be present) unchanged.
9835 adelay=1500|0|500
9837 • Delay second channel by 500 samples, the third channel by 700
9839 samples and leave the first channel (and any other channels that
9840 may be present) unchanged.
9841 adelay=0|500S|700S
9843 • Delay all channels by same number of samples:
9845 adelay=delays=64S:all=1
9847 adenorm
9849 Remedy denormals in audio by adding extremely low-level noise.
9850 This filter shall be placed before any filter that can produce
9852 denormals.
9853 A description of the accepted parameters follows.
9855 level
9857 Set level of added noise in dB. Default is -351. Allowed range is
9858 from -451 to -90.
9859 type
9861 Set type of added noise.
9862 dc Add DC signal.
9864 ac Add AC signal.
9866 square
9868 Add square signal.
9869 pulse
9871 Add pulse signal.
9872 Default is "dc".
9874 Commands
9876 This filter supports the all above options as commands.
9878 aderivative, aintegral
9880 Compute derivative/integral of audio stream.
9881 Applying both filters one after another produces original audio.
9883 adrc
9885 Apply spectral dynamic range controller filter to input audio stream.
9886 A description of the accepted options follows.
9888 transfer
9890 Set the transfer expression.
9891 The expression can contain the following constants:
9893 ch current channel number
9895 sn current sample number
9897 nb_channels
9899 number of channels
9900 t timestamp expressed in seconds
9902 sr sample rate
9904 p current frequency power value, in dB
9906 f current frequency in Hz
9908 Default value is "p".
9910 attack
9912 Set the attack in milliseconds. Default is 50 milliseconds.
9913 Allowed range is from 1 to 1000 milliseconds.
9914 release
9916 Set the release in milliseconds. Default is 100 milliseconds.
9917 Allowed range is from 5 to 2000 milliseconds.
9918 channels
9920 Set which channels to filter, by default "all" channels in audio
9921 stream are filtered.
9922 Commands
9924 This filter supports the all above options as commands.
9926 Examples
9928 • Apply spectral compression to all frequencies with threshold of -50
9930 dB and 1:6 ratio:
9931 adrc=transfer='if(gt(p,-50),-50+(p-(-50))/6,p)':attack=50:release=100
9933 • Similar to above but with 1:2 ratio and filtering only front center
9935 channel:
9936 adrc=transfer='if(gt(p,-50),-50+(p-(-50))/2,p)':attack=50:release=100:channels=FC
9938 • Apply spectral noise gate to all frequencies with threshold of -85
9940 dB and with short attack time and short release time:
9941 adrc=transfer='if(lte(p,-85),p-800,p)':attack=1:release=5
9943 • Apply spectral expansion to all frequencies with threshold of -10
9945 dB and 1:2 ratio:
9946 adrc=transfer='if(lt(p,-10),-10+(p-(-10))*2,p)':attack=50:release=100
9948 • Apply limiter to max -60 dB to all frequencies, with attack of 2 ms
9950 and release of 10 ms:
9951 adrc=transfer='min(p,-60)':attack=2:release=10
9953 adynamicequalizer
9955 Apply dynamic equalization to input audio stream.
9956 A description of the accepted options follows.
9958 threshold
9960 Set the detection threshold used to trigger equalization.
9961 Threshold detection is using bandpass filter. Default value is 0.
9962 Allowed range is from 0 to 100.
9963 dfrequency
9965 Set the detection frequency in Hz used for bandpass filter used to
9966 trigger equalization. Default value is 1000 Hz. Allowed range is
9967 between 2 and 1000000 Hz.
9968 dqfactor
9970 Set the detection resonance factor for bandpass filter used to
9971 trigger equalization. Default value is 1. Allowed range is from
9972 0.001 to 1000.
9973 tfrequency
9975 Set the target frequency of equalization filter. Default value is
9976 1000 Hz. Allowed range is between 2 and 1000000 Hz.
9977 tqfactor
9979 Set the target resonance factor for target equalization filter.
9980 Default value is 1. Allowed range is from 0.001 to 1000.
9981 attack
9983 Set the amount of milliseconds the signal from detection has to
9984 rise above the detection threshold before equalization starts.
9985 Default is 20. Allowed range is between 1 and 2000.
9986 release
9988 Set the amount of milliseconds the signal from detection has to
9989 fall below the detection threshold before equalization ends.
9990 Default is 200. Allowed range is between 1 and 2000.
9991 ratio
9993 Set the ratio by which the equalization gain is raised. Default is
9994 1. Allowed range is between 0 and 30.
9995 makeup
9997 Set the makeup offset by which the equalization gain is raised.
9998 Default is 0. Allowed range is between 0 and 100.
9999 range
10001 Set the max allowed cut/boost amount. Default is 50. Allowed range
10002 is from 1 to 200.
10003 mode
10005 Set the mode of filter operation, can be one of the following:
10006 listen
10008 Output only isolated bandpass signal.
10009 cut Cut frequencies above detection threshold.
10011 boost
10013 Boost frequencies bellow detection threshold.
10014 Default mode is cut.
10016 tftype
10018 Set the type of target filter, can be one of the following:
10019 bell
10021 lowshelf
10022 highshelf
10023 Default type is bell.
10025 direction
10027 Set processing direction relative to threshold.
10028 downward
10030 Boost/Cut if threshold is higher/lower than detected volume.
10031 upward
10033 Boost/Cut if threshold is lower/higher than detected volume.
10034 Default direction is downward.
10036 auto
10038 Automatically gather threshold from detection filter. By default is
10039 disabled. This option is useful to detect threshold in certain
10040 time frame of input audio stream, in such case option value is
10041 changed at runtime.
10042 Available values are:
10044 disabled
10046 Disable using automatically gathered threshold value.
10047 off Stop picking threshold value.
10049 on Start picking threshold value.
10051 Commands
10053 This filter supports the all above options as commands.
10055 adynamicsmooth
10057 Apply dynamic smoothing to input audio stream.
10058 A description of the accepted options follows.
10060 sensitivity
10062 Set an amount of sensitivity to frequency fluctations. Default is
10063 2. Allowed range is from 0 to 1e+06.
10064 basefreq
10066 Set a base frequency for smoothing. Default value is 22050.
10067 Allowed range is from 2 to 1e+06.
10068 Commands
10070 This filter supports the all above options as commands.
10072 aecho
10074 Apply echoing to the input audio.
10075 Echoes are reflected sound and can occur naturally amongst mountains
10077 (and sometimes large buildings) when talking or shouting; digital echo
10078 effects emulate this behaviour and are often used to help fill out the
10079 sound of a single instrument or vocal. The time difference between the
10080 original signal and the reflection is the "delay", and the loudness of
10081 the reflected signal is the "decay". Multiple echoes can have
10082 different delays and decays.
10083 A description of the accepted parameters follows.
10085 in_gain
10087 Set input gain of reflected signal. Default is 0.6.
10088 out_gain
10090 Set output gain of reflected signal. Default is 0.3.
10091 delays
10093 Set list of time intervals in milliseconds between original signal
10094 and reflections separated by '|'. Allowed range for each "delay" is
10095 "(0 - 90000.0]". Default is 1000.
10096 decays
10098 Set list of loudness of reflected signals separated by '|'.
10099 Allowed range for each "decay" is "(0 - 1.0]". Default is 0.5.
10100 Examples
10102 • Make it sound as if there are twice as many instruments as are
10104 actually playing:
10105 aecho=0.8:0.88:60:0.4
10107 • If delay is very short, then it sounds like a (metallic) robot
10109 playing music:
10110 aecho=0.8:0.88:6:0.4
10112 • A longer delay will sound like an open air concert in the
10114 mountains:
10115 aecho=0.8:0.9:1000:0.3
10117 • Same as above but with one more mountain:
10119 aecho=0.8:0.9:1000|1800:0.3|0.25
10121 aemphasis
10123 Audio emphasis filter creates or restores material directly taken from
10124 LPs or emphased CDs with different filter curves. E.g. to store music
10125 on vinyl the signal has to be altered by a filter first to even out the
10126 disadvantages of this recording medium. Once the material is played
10127 back the inverse filter has to be applied to restore the distortion of
10128 the frequency response.
10129 The filter accepts the following options:
10131 level_in
10133 Set input gain.
10134 level_out
10136 Set output gain.
10137 mode
10139 Set filter mode. For restoring material use "reproduction" mode,
10140 otherwise use "production" mode. Default is "reproduction" mode.
10141 type
10143 Set filter type. Selects medium. Can be one of the following:
10144 col select Columbia.
10146 emi select EMI.
10148 bsi select BSI (78RPM).
10150 riaa
10152 select RIAA.
10153 cd select Compact Disc (CD).
10155 50fm
10157 select 50µs (FM).
10158 75fm
10160 select 75µs (FM).
10161 50kf
10163 select 50µs (FM-KF).
10164 75kf
10166 select 75µs (FM-KF).
10167 Commands
10169 This filter supports the all above options as commands.
10171 aeval
10173 Modify an audio signal according to the specified expressions.
10174 This filter accepts one or more expressions (one for each channel),
10176 which are evaluated and used to modify a corresponding audio signal.
10177 It accepts the following parameters:
10179 exprs
10181 Set the '|'-separated expressions list for each separate channel.
10182 If the number of input channels is greater than the number of
10183 expressions, the last specified expression is used for the
10184 remaining output channels.
10185 channel_layout, c
10187 Set output channel layout. If not specified, the channel layout is
10188 specified by the number of expressions. If set to same, it will use
10189 by default the same input channel layout.
10190 Each expression in exprs can contain the following constants and
10192 functions:
10193 ch channel number of the current expression
10195 n number of the evaluated sample, starting from 0
10197 s sample rate
10199 t time of the evaluated sample expressed in seconds
10201 nb_in_channels
10203 nb_out_channels
10204 input and output number of channels
10205 val(CH)
10207 the value of input channel with number CH
10208 Note: this filter is slow. For faster processing you should use a
10210 dedicated filter.
10211 Examples
10213 • Half volume:
10215 aeval=val(ch)/2:c=same
10217 • Invert phase of the second channel:
10219 aeval=val(0)|-val(1)
10221 aexciter
10223 An exciter is used to produce high sound that is not present in the
10224 original signal. This is done by creating harmonic distortions of the
10225 signal which are restricted in range and added to the original signal.
10226 An Exciter raises the upper end of an audio signal without simply
10227 raising the higher frequencies like an equalizer would do to create a
10228 more "crisp" or "brilliant" sound.
10229 The filter accepts the following options:
10231 level_in
10233 Set input level prior processing of signal. Allowed range is from
10234 0 to 64. Default value is 1.
10235 level_out
10237 Set output level after processing of signal. Allowed range is from
10238 0 to 64. Default value is 1.
10239 amount
10241 Set the amount of harmonics added to original signal. Allowed
10242 range is from 0 to 64. Default value is 1.
10243 drive
10245 Set the amount of newly created harmonics. Allowed range is from
10246 0.1 to 10. Default value is 8.5.
10247 blend
10249 Set the octave of newly created harmonics. Allowed range is from
10250 -10 to 10. Default value is 0.
10251 freq
10253 Set the lower frequency limit of producing harmonics in Hz.
10254 Allowed range is from 2000 to 12000 Hz. Default is 7500 Hz.
10255 ceil
10257 Set the upper frequency limit of producing harmonics. Allowed
10258 range is from 9999 to 20000 Hz. If value is lower than 10000 Hz no
10259 limit is applied.
10260 listen
10262 Mute the original signal and output only added harmonics. By
10263 default is disabled.
10264 Commands
10266 This filter supports the all above options as commands.
10268 afade
10270 Apply fade-in/out effect to input audio.
10271 A description of the accepted parameters follows.
10273 type, t
10275 Specify the effect type, can be either "in" for fade-in, or "out"
10276 for a fade-out effect. Default is "in".
10277 start_sample, ss
10279 Specify the number of the start sample for starting to apply the
10280 fade effect. Default is 0.
10281 nb_samples, ns
10283 Specify the number of samples for which the fade effect has to
10284 last. At the end of the fade-in effect the output audio will have
10285 the same volume as the input audio, at the end of the fade-out
10286 transition the output audio will be silence. Default is 44100.
10287 start_time, st
10289 Specify the start time of the fade effect. Default is 0. The value
10290 must be specified as a time duration; see the Time duration section
10291 in the ffmpeg-utils(1) manual for the accepted syntax. If set this
10292 option is used instead of start_sample.
10293 duration, d
10295 Specify the duration of the fade effect. See the Time duration
10296 section in the ffmpeg-utils(1) manual for the accepted syntax. At
10297 the end of the fade-in effect the output audio will have the same
10298 volume as the input audio, at the end of the fade-out transition
10299 the output audio will be silence. By default the duration is
10300 determined by nb_samples. If set this option is used instead of
10301 nb_samples.
10302 curve
10304 Set curve for fade transition.
10305 It accepts the following values:
10307 tri select triangular, linear slope (default)
10309 qsin
10311 select quarter of sine wave
10312 hsin
10314 select half of sine wave
10315 esin
10317 select exponential sine wave
10318 log select logarithmic
10320 ipar
10322 select inverted parabola
10323 qua select quadratic
10325 cub select cubic
10327 squ select square root
10329 cbr select cubic root
10331 par select parabola
10333 exp select exponential
10335 iqsin
10337 select inverted quarter of sine wave
10338 ihsin
10340 select inverted half of sine wave
10341 dese
10343 select double-exponential seat
10344 desi
10346 select double-exponential sigmoid
10347 losi
10349 select logistic sigmoid
10350 sinc
10352 select sine cardinal function
10353 isinc
10355 select inverted sine cardinal function
10356 nofade
10358 no fade applied
10359 silence
10361 Set the initial gain for fade-in or final gain for fade-out.
10362 Default value is 0.0.
10363 unity
10365 Set the initial gain for fade-out or final gain for fade-in.
10366 Default value is 1.0.
10367 Commands
10369 This filter supports the all above options as commands.
10371 Examples
10373 • Fade in first 15 seconds of audio:
10375 afade=t=in:ss=0:d=15
10377 • Fade out last 25 seconds of a 900 seconds audio:
10379 afade=t=out:st=875:d=25
10381 afftdn
10383 Denoise audio samples with FFT.
10384 A description of the accepted parameters follows.
10386 noise_reduction, nr
10388 Set the noise reduction in dB, allowed range is 0.01 to 97.
10389 Default value is 12 dB.
10390 noise_floor, nf
10392 Set the noise floor in dB, allowed range is -80 to -20. Default
10393 value is -50 dB.
10394 noise_type, nt
10396 Set the noise type.
10397 It accepts the following values:
10399 white, w
10401 Select white noise.
10402 vinyl, v
10404 Select vinyl noise.
10405 shellac, s
10407 Select shellac noise.
10408 custom, c
10410 Select custom noise, defined in "bn" option.
10411 Default value is white noise.
10413 band_noise, bn
10415 Set custom band noise profile for every one of 15 bands. Bands are
10416 separated by ' ' or '|'.
10417 residual_floor, rf
10419 Set the residual floor in dB, allowed range is -80 to -20. Default
10420 value is -38 dB.
10421 track_noise, tn
10423 Enable noise floor tracking. By default is disabled. With this
10424 enabled, noise floor is automatically adjusted.
10425 track_residual, tr
10427 Enable residual tracking. By default is disabled.
10428 output_mode, om
10430 Set the output mode.
10431 It accepts the following values:
10433 input, i
10435 Pass input unchanged.
10436 output, o
10438 Pass noise filtered out.
10439 noise, n
10441 Pass only noise.
10442 Default value is output.
10444 adaptivity, ad
10446 Set the adaptivity factor, used how fast to adapt gains adjustments
10447 per each frequency bin. Value 0 enables instant adaptation, while
10448 higher values react much slower. Allowed range is from 0 to 1.
10449 Default value is 0.5.
10450 floor_offset, fo
10452 Set the noise floor offset factor. This option is used to adjust
10453 offset applied to measured noise floor. It is only effective when
10454 noise floor tracking is enabled. Allowed range is from -2.0 to
10455 2.0. Default value is 1.0.
10456 noise_link, nl
10458 Set the noise link used for multichannel audio.
10459 It accepts the following values:
10461 none
10463 Use unchanged channel's noise floor.
10464 min Use measured min noise floor of all channels.
10466 max Use measured max noise floor of all channels.
10468 average
10470 Use measured average noise floor of all channels.
10471 Default value is min.
10473 band_multiplier, bm
10475 Set the band multiplier factor, used how much to spread bands
10476 across frequency bins. Allowed range is from 0.2 to 5. Default
10477 value is 1.25.
10478 sample_noise, sn
10480 Toggle capturing and measurement of noise profile from input audio.
10481 It accepts the following values:
10483 start, begin
10485 Start sample noise capture.
10486 stop, end
10488 Stop sample noise capture and measure new noise band profile.
10489 Default value is "none".
10491 gain_smooth, gs
10493 Set gain smooth spatial radius, used to smooth gains applied to
10494 each frequency bin. Useful to reduce random music noise artefacts.
10495 Higher values increases smoothing of gains. Allowed range is from
10496 0 to 50. Default value is 0.
10497 Commands
10499 This filter supports the some above mentioned options as commands.
10501 Examples
10503 • Reduce white noise by 10dB, and use previously measured noise floor
10505 of -40dB:
10506 afftdn=nr=10:nf=-40
10508 • Reduce white noise by 10dB, also set initial noise floor to -80dB
10510 and enable automatic tracking of noise floor so noise floor will
10511 gradually change during processing:
10512 afftdn=nr=10:nf=-80:tn=1
10514 • Reduce noise by 20dB, using noise floor of -40dB and using commands
10516 to take noise profile of first 0.4 seconds of input audio:
10517 asendcmd=0.0 afftdn sn start,asendcmd=0.4 afftdn sn stop,afftdn=nr=20:nf=-40
10519 afftfilt
10521 Apply arbitrary expressions to samples in frequency domain.
10522 real
10524 Set frequency domain real expression for each separate channel
10525 separated by '|'. Default is "re". If the number of input channels
10526 is greater than the number of expressions, the last specified
10527 expression is used for the remaining output channels.
10528 imag
10530 Set frequency domain imaginary expression for each separate channel
10531 separated by '|'. Default is "im".
10532 Each expression in real and imag can contain the following
10534 constants and functions:
10535 sr sample rate
10537 b current frequency bin number
10539 nb number of available bins
10541 ch channel number of the current expression
10543 chs number of channels
10545 pts current frame pts
10547 re current real part of frequency bin of current channel
10549 im current imaginary part of frequency bin of current channel
10551 real(b, ch)
10553 Return the value of real part of frequency bin at location
10554 (bin,channel)
10555 imag(b, ch)
10557 Return the value of imaginary part of frequency bin at location
10558 (bin,channel)
10559 win_size
10561 Set window size. Allowed range is from 16 to 131072. Default is
10562 4096
10563 win_func
10565 Set window function.
10566 It accepts the following values:
10568 rect
10570 bartlett
10571 hann, hanning
10572 hamming
10573 blackman
10574 welch
10575 flattop
10576 bharris
10577 bnuttall
10578 bhann
10579 sine
10580 nuttall
10581 lanczos
10582 gauss
10583 tukey
10584 dolph
10585 cauchy
10586 parzen
10587 poisson
10588 bohman
10589 kaiser
10590 Default is "hann".
10592 overlap
10594 Set window overlap. If set to 1, the recommended overlap for
10595 selected window function will be picked. Default is 0.75.
10596 Examples
10598 • Leave almost only low frequencies in audio:
10600 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
10602 • Apply robotize effect:
10604 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
10606 • Apply whisper effect:
10608 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"
10610 • Apply phase shift:
10612 afftfilt="real=re*cos(1)-im*sin(1):imag=re*sin(1)+im*cos(1)"
10614 afir
10616 Apply an arbitrary Finite Impulse Response filter.
10617 This filter is designed for applying long FIR filters, up to 60 seconds
10619 long.
10620 It can be used as component for digital crossover filters, room
10622 equalization, cross talk cancellation, wavefield synthesis,
10623 auralization, ambiophonics, ambisonics and spatialization.
10624 This filter uses the streams higher than first one as FIR coefficients.
10626 If the non-first stream holds a single channel, it will be used for all
10627 input channels in the first stream, otherwise the number of channels in
10628 the non-first stream must be same as the number of channels in the
10629 first stream.
10630 It accepts the following parameters:
10632 dry Set dry gain. This sets input gain.
10634 wet Set wet gain. This sets final output gain.
10636 length
10638 Set Impulse Response filter length. Default is 1, which means whole
10639 IR is processed.
10640 gtype
10642 Enable applying gain measured from power of IR.
10643 Set which approach to use for auto gain measurement.
10645 none
10647 Do not apply any gain.
10648 peak
10650 select peak gain, very conservative approach. This is default
10651 value.
10652 dc select DC gain, limited application.
10654 gn select gain to noise approach, this is most popular one.
10656 ac select AC gain.
10658 rms select RMS gain.
10660 irgain
10662 Set gain to be applied to IR coefficients before filtering.
10663 Allowed range is 0 to 1. This gain is applied after any gain
10664 applied with gtype option.
10665 irfmt
10667 Set format of IR stream. Can be "mono" or "input". Default is
10668 "input".
10669 maxir
10671 Set max allowed Impulse Response filter duration in seconds.
10672 Default is 30 seconds. Allowed range is 0.1 to 60 seconds.
10673 response
10675 Show IR frequency response, magnitude(magenta), phase(green) and
10676 group delay(yellow) in additional video stream. By default it is
10677 disabled.
10678 channel
10680 Set for which IR channel to display frequency response. By default
10681 is first channel displayed. This option is used only when response
10682 is enabled.
10683 size
10685 Set video stream size. This option is used only when response is
10686 enabled.
10687 rate
10689 Set video stream frame rate. This option is used only when response
10690 is enabled.
10691 minp
10693 Set minimal partition size used for convolution. Default is 8192.
10694 Allowed range is from 1 to 65536. Lower values decreases latency
10695 at cost of higher CPU usage.
10696 maxp
10698 Set maximal partition size used for convolution. Default is 8192.
10699 Allowed range is from 8 to 65536. Lower values may increase CPU
10700 usage.
10701 nbirs
10703 Set number of input impulse responses streams which will be
10704 switchable at runtime. Allowed range is from 1 to 32. Default is
10705 1.
10706 ir Set IR stream which will be used for convolution, starting from 0,
10708 should always be lower than supplied value by "nbirs" option.
10709 Default is 0. This option can be changed at runtime via commands.
10710 precision
10712 Set which precision to use when processing samples.
10713 auto
10715 Auto pick internal sample format depending on other filters.
10716 float
10718 Always use single-floating point precision sample format.
10719 double
10721 Always use double-floating point precision sample format.
10722 Default value is auto.
10724 Examples
10726 • Apply reverb to stream using mono IR file as second input, complete
10728 command using ffmpeg:
10729 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
10731 • Apply true stereo processing given input stereo stream, and two
10733 stereo impulse responses for left and right channel, the impulse
10734 response files are files with names l_ir.wav and r_ir.wav:
10735 "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"
10737 aformat
10739 Set output format constraints for the input audio. The framework will
10740 negotiate the most appropriate format to minimize conversions.
10741 It accepts the following parameters:
10743 sample_fmts, f
10745 A '|'-separated list of requested sample formats.
10746 sample_rates, r
10748 A '|'-separated list of requested sample rates.
10749 channel_layouts, cl
10751 A '|'-separated list of requested channel layouts.
10752 See the Channel Layout section in the ffmpeg-utils(1) manual for
10754 the required syntax.
10755 If a parameter is omitted, all values are allowed.
10757 Force the output to either unsigned 8-bit or signed 16-bit stereo
10759 aformat=sample_fmts=u8|s16:channel_layouts=stereo
10761 afreqshift
10763 Apply frequency shift to input audio samples.
10764 The filter accepts the following options:
10766 shift
10768 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
10769 Default value is 0.0.
10770 level
10772 Set output gain applied to final output. Allowed range is from 0.0
10773 to 1.0. Default value is 1.0.
10774 order
10776 Set filter order used for filtering. Allowed range is from 1 to 16.
10777 Default value is 8.
10778 Commands
10780 This filter supports the all above options as commands.
10782 afwtdn
10784 Reduce broadband noise from input samples using Wavelets.
10785 A description of the accepted options follows.
10787 sigma
10789 Set the noise sigma, allowed range is from 0 to 1. Default value
10790 is 0. This option controls strength of denoising applied to input
10791 samples. Most useful way to set this option is via decibels, eg.
10792 -45dB.
10793 levels
10795 Set the number of wavelet levels of decomposition. Allowed range
10796 is from 1 to 12. Default value is 10. Setting this too low make
10797 denoising performance very poor.
10798 wavet
10800 Set wavelet type for decomposition of input frame. They are sorted
10801 by number of coefficients, from lowest to highest. More
10802 coefficients means worse filtering speed, but overall better
10803 quality. Available wavelets are:
10804 sym2
10806 sym4
10807 rbior68
10808 deb10
10809 sym10
10810 coif5
10811 bl3
10812 percent
10813 Set percent of full denoising. Allowed range is from 0 to 100
10814 percent. Default value is 85 percent or partial denoising.
10815 profile
10817 If enabled, first input frame will be used as noise profile. If
10818 first frame samples contain non-noise performance will be very
10819 poor.
10820 adaptive
10822 If enabled, input frames are analyzed for presence of noise. If
10823 noise is detected with high possibility then input frame profile
10824 will be used for processing following frames, until new noise frame
10825 is detected.
10826 samples
10828 Set size of single frame in number of samples. Allowed range is
10829 from 512 to 65536. Default frame size is 8192 samples.
10830 softness
10832 Set softness applied inside thresholding function. Allowed range is
10833 from 0 to 10. Default softness is 1.
10834 Commands
10836 This filter supports the all above options as commands.
10838 agate
10840 A gate is mainly used to reduce lower parts of a signal. This kind of
10841 signal processing reduces disturbing noise between useful signals.
10842 Gating is done by detecting the volume below a chosen level threshold
10844 and dividing it by the factor set with ratio. The bottom of the noise
10845 floor is set via range. Because an exact manipulation of the signal
10846 would cause distortion of the waveform the reduction can be levelled
10847 over time. This is done by setting attack and release.
10848 attack determines how long the signal has to fall below the threshold
10850 before any reduction will occur and release sets the time the signal
10851 has to rise above the threshold to reduce the reduction again. Shorter
10852 signals than the chosen attack time will be left untouched.
10853 level_in
10855 Set input level before filtering. Default is 1. Allowed range is
10856 from 0.015625 to 64.
10857 mode
10859 Set the mode of operation. Can be "upward" or "downward". Default
10860 is "downward". If set to "upward" mode, higher parts of signal will
10861 be amplified, expanding dynamic range in upward direction.
10862 Otherwise, in case of "downward" lower parts of signal will be
10863 reduced.
10864 range
10866 Set the level of gain reduction when the signal is below the
10867 threshold. Default is 0.06125. Allowed range is from 0 to 1.
10868 Setting this to 0 disables reduction and then filter behaves like
10869 expander.
10870 threshold
10872 If a signal rises above this level the gain reduction is released.
10873 Default is 0.125. Allowed range is from 0 to 1.
10874 ratio
10876 Set a ratio by which the signal is reduced. Default is 2. Allowed
10877 range is from 1 to 9000.
10878 attack
10880 Amount of milliseconds the signal has to rise above the threshold
10881 before gain reduction stops. Default is 20 milliseconds. Allowed
10882 range is from 0.01 to 9000.
10883 release
10885 Amount of milliseconds the signal has to fall below the threshold
10886 before the reduction is increased again. Default is 250
10887 milliseconds. Allowed range is from 0.01 to 9000.
10888 makeup
10890 Set amount of amplification of signal after processing. Default is
10891 1. Allowed range is from 1 to 64.
10892 knee
10894 Curve the sharp knee around the threshold to enter gain reduction
10895 more softly. Default is 2.828427125. Allowed range is from 1 to 8.
10896 detection
10898 Choose if exact signal should be taken for detection or an RMS like
10899 one. Default is "rms". Can be "peak" or "rms".
10900 link
10902 Choose if the average level between all channels or the louder
10903 channel affects the reduction. Default is "average". Can be
10904 "average" or "maximum".
10905 Commands
10907 This filter supports the all above options as commands.
10909 aiir
10911 Apply an arbitrary Infinite Impulse Response filter.
10912 It accepts the following parameters:
10914 zeros, z
10916 Set B/numerator/zeros/reflection coefficients.
10917 poles, p
10919 Set A/denominator/poles/ladder coefficients.
10920 gains, k
10922 Set channels gains.
10923 dry_gain
10925 Set input gain.
10926 wet_gain
10928 Set output gain.
10929 format, f
10931 Set coefficients format.
10932 ll lattice-ladder function
10934 sf analog transfer function
10936 tf digital transfer function
10938 zp Z-plane zeros/poles, cartesian (default)
10940 pr Z-plane zeros/poles, polar radians
10942 pd Z-plane zeros/poles, polar degrees
10944 sp S-plane zeros/poles
10946 process, r
10948 Set type of processing.
10949 d direct processing
10951 s serial processing
10953 p parallel processing
10955 precision, e
10957 Set filtering precision.
10958 dbl double-precision floating-point (default)
10960 flt single-precision floating-point
10962 i32 32-bit integers
10964 i16 16-bit integers
10966 normalize, n
10968 Normalize filter coefficients, by default is enabled. Enabling it
10969 will normalize magnitude response at DC to 0dB.
10970 mix How much to use filtered signal in output. Default is 1. Range is
10972 between 0 and 1.
10973 response
10975 Show IR frequency response, magnitude(magenta), phase(green) and
10976 group delay(yellow) in additional video stream. By default it is
10977 disabled.
10978 channel
10980 Set for which IR channel to display frequency response. By default
10981 is first channel displayed. This option is used only when response
10982 is enabled.
10983 size
10985 Set video stream size. This option is used only when response is
10986 enabled.
10987 Coefficients in "tf" and "sf" format are separated by spaces and are in
10989 ascending order.
10990 Coefficients in "zp" format are separated by spaces and order of
10992 coefficients doesn't matter. Coefficients in "zp" format are complex
10993 numbers with i imaginary unit.
10994 Different coefficients and gains can be provided for every channel, in
10996 such case use '|' to separate coefficients or gains. Last provided
10997 coefficients will be used for all remaining channels.
10998 Examples
11000 • Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample
11002 rate:
11003 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
11005 • Same as above but in "zp" format:
11007 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
11009 • Apply 3-rd order analog normalized Butterworth low-pass filter,
11011 using analog transfer function format:
11012 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
11014 alimiter
11016 The limiter prevents an input signal from rising over a desired
11017 threshold. This limiter uses lookahead technology to prevent your
11018 signal from distorting. It means that there is a small delay after the
11019 signal is processed. Keep in mind that the delay it produces is the
11020 attack time you set.
11021 The filter accepts the following options:
11023 level_in
11025 Set input gain. Default is 1.
11026 level_out
11028 Set output gain. Default is 1.
11029 limit
11031 Don't let signals above this level pass the limiter. Default is 1.
11032 attack
11034 The limiter will reach its attenuation level in this amount of time
11035 in milliseconds. Default is 5 milliseconds.
11036 release
11038 Come back from limiting to attenuation 1.0 in this amount of
11039 milliseconds. Default is 50 milliseconds.
11040 asc When gain reduction is always needed ASC takes care of releasing to
11042 an average reduction level rather than reaching a reduction of 0 in
11043 the release time.
11044 asc_level
11046 Select how much the release time is affected by ASC, 0 means nearly
11047 no changes in release time while 1 produces higher release times.
11048 level
11050 Auto level output signal. Default is enabled. This normalizes
11051 audio back to 0dB if enabled.
11052 latency
11054 Compensate the delay introduced by using the lookahead buffer set
11055 with attack parameter. Also flush the valid audio data in the
11056 lookahead buffer when the stream hits EOF.
11057 Depending on picked setting it is recommended to upsample input 2x or
11059 4x times with aresample before applying this filter.
11060 allpass
11062 Apply a two-pole all-pass filter with central frequency (in Hz)
11063 frequency, and filter-width width. An all-pass filter changes the
11064 audio's frequency to phase relationship without changing its frequency
11065 to amplitude relationship.
11066 The filter accepts the following options:
11068 frequency, f
11070 Set frequency in Hz.
11071 width_type, t
11073 Set method to specify band-width of filter.
11074 h Hz
11076 q Q-Factor
11078 o octave
11080 s slope
11082 k kHz
11084 width, w
11086 Specify the band-width of a filter in width_type units.
11087 mix, m
11089 How much to use filtered signal in output. Default is 1. Range is
11090 between 0 and 1.
11091 channels, c
11093 Specify which channels to filter, by default all available are
11094 filtered.
11095 normalize, n
11097 Normalize biquad coefficients, by default is disabled. Enabling it
11098 will normalize magnitude response at DC to 0dB.
11099 order, o
11101 Set the filter order, can be 1 or 2. Default is 2.
11102 transform, a
11104 Set transform type of IIR filter.
11105 di
11107 dii
11108 tdi
11109 tdii
11110 latt
11111 svf
11112 zdf
11113 precision, r
11114 Set precison of filtering.
11115 auto
11117 Pick automatic sample format depending on surround filters.
11118 s16 Always use signed 16-bit.
11120 s32 Always use signed 32-bit.
11122 f32 Always use float 32-bit.
11124 f64 Always use float 64-bit.
11126 Commands
11128 This filter supports the following commands:
11130 frequency, f
11132 Change allpass frequency. Syntax for the command is : "frequency"
11133 width_type, t
11135 Change allpass width_type. Syntax for the command is :
11136 "width_type"
11137 width, w
11139 Change allpass width. Syntax for the command is : "width"
11140 mix, m
11142 Change allpass mix. Syntax for the command is : "mix"
11143 aloop
11145 Loop audio samples.
11146 The filter accepts the following options:
11148 loop
11150 Set the number of loops. Setting this value to -1 will result in
11151 infinite loops. Default is 0.
11152 size
11154 Set maximal number of samples. Default is 0.
11155 start
11157 Set first sample of loop. Default is 0.
11158 amerge
11160 Merge two or more audio streams into a single multi-channel stream.
11161 The filter accepts the following options:
11163 inputs
11165 Set the number of inputs. Default is 2.
11166 If the channel layouts of the inputs are disjoint, and therefore
11168 compatible, the channel layout of the output will be set accordingly
11169 and the channels will be reordered as necessary. If the channel layouts
11170 of the inputs are not disjoint, the output will have all the channels
11171 of the first input then all the channels of the second input, in that
11172 order, and the channel layout of the output will be the default value
11173 corresponding to the total number of channels.
11174 For example, if the first input is in 2.1 (FL+FR+LF) and the second
11176 input is FC+BL+BR, then the output will be in 5.1, with the channels in
11177 the following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of
11178 the first input, b1 is the first channel of the second input).
11179 On the other hand, if both input are in stereo, the output channels
11181 will be in the default order: a1, a2, b1, b2, and the channel layout
11182 will be arbitrarily set to 4.0, which may or may not be the expected
11183 value.
11184 All inputs must have the same sample rate, and format.
11186 If inputs do not have the same duration, the output will stop with the
11188 shortest.
11189 Examples
11191 • Merge two mono files into a stereo stream:
11193 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
11195 • Multiple merges assuming 1 video stream and 6 audio streams in
11197 input.mkv:
11198 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
11200 amix
11202 Mixes multiple audio inputs into a single output.
11203 Note that this filter only supports float samples (the amerge and pan
11205 audio filters support many formats). If the amix input has integer
11206 samples then aresample will be automatically inserted to perform the
11207 conversion to float samples.
11208 It accepts the following parameters:
11210 inputs
11212 The number of inputs. If unspecified, it defaults to 2.
11213 duration
11215 How to determine the end-of-stream.
11216 longest
11218 The duration of the longest input. (default)
11219 shortest
11221 The duration of the shortest input.
11222 first
11224 The duration of the first input.
11225 dropout_transition
11227 The transition time, in seconds, for volume renormalization when an
11228 input stream ends. The default value is 2 seconds.
11229 weights
11231 Specify weight of each input audio stream as a sequence of numbers
11232 separated by a space. If fewer weights are specified compared to
11233 number of inputs, the last weight is assigned to the remaining
11234 inputs. Default weight for each input is 1.
11235 normalize
11237 Always scale inputs instead of only doing summation of samples.
11238 Beware of heavy clipping if inputs are not normalized prior or
11239 after filtering by this filter if this option is disabled. By
11240 default is enabled.
11241 Examples
11243 • This will mix 3 input audio streams to a single output with the
11245 same duration as the first input and a dropout transition time of 3
11246 seconds:
11247 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
11249 • This will mix one vocal and one music input audio stream to a
11251 single output with the same duration as the longest input. The
11252 music will have quarter the weight as the vocals, and the inputs
11253 are not normalized:
11254 ffmpeg -i VOCALS -i MUSIC -filter_complex amix=inputs=2:duration=longest:dropout_transition=0:weights="1 0.25":normalize=0 OUTPUT
11256 Commands
11258 This filter supports the following commands:
11260 weights
11262 normalize
11263 Syntax is same as option with same name.
11264 amultiply
11266 Multiply first audio stream with second audio stream and store result
11267 in output audio stream. Multiplication is done by multiplying each
11268 sample from first stream with sample at same position from second
11269 stream.
11270 With this element-wise multiplication one can create amplitude fades
11272 and amplitude modulations.
11273 anequalizer
11275 High-order parametric multiband equalizer for each channel.
11276 It accepts the following parameters:
11278 params
11280 This option string is in format: "cchn f=cf w=w g=g t=f | ..."
11281 Each equalizer band is separated by '|'.
11282 chn Set channel number to which equalization will be applied. If
11284 input doesn't have that channel the entry is ignored.
11285 f Set central frequency for band. If input doesn't have that
11287 frequency the entry is ignored.
11288 w Set band width in Hertz.
11290 g Set band gain in dB.
11292 t Set filter type for band, optional, can be:
11294 0 Butterworth, this is default.
11296 1 Chebyshev type 1.
11298 2 Chebyshev type 2.
11300 curves
11302 With this option activated frequency response of anequalizer is
11303 displayed in video stream.
11304 size
11306 Set video stream size. Only useful if curves option is activated.
11307 mgain
11309 Set max gain that will be displayed. Only useful if curves option
11310 is activated. Setting this to a reasonable value makes it possible
11311 to display gain which is derived from neighbour bands which are too
11312 close to each other and thus produce higher gain when both are
11313 activated.
11314 fscale
11316 Set frequency scale used to draw frequency response in video
11317 output. Can be linear or logarithmic. Default is logarithmic.
11318 colors
11320 Set color for each channel curve which is going to be displayed in
11321 video stream. This is list of color names separated by space or by
11322 '|'. Unrecognised or missing colors will be replaced by white
11323 color.
11324 Examples
11326 • Lower gain by 10 of central frequency 200Hz and width 100 Hz for
11328 first 2 channels using Chebyshev type 1 filter:
11329 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
11331 Commands
11333 This filter supports the following commands:
11335 change
11337 Alter existing filter parameters. Syntax for the commands is :
11338 "fN|f=freq|w=width|g=gain"
11339 fN is existing filter number, starting from 0, if no such filter is
11341 available error is returned. freq set new frequency parameter.
11342 width set new width parameter in Hertz. gain set new gain
11343 parameter in dB.
11344 Full filter invocation with asendcmd may look like this:
11346 asendcmd=c='4.0 anequalizer change
11347 0|f=200|w=50|g=1',anequalizer=...
11348 anlmdn
11350 Reduce broadband noise in audio samples using Non-Local Means
11351 algorithm.
11352 Each sample is adjusted by looking for other samples with similar
11354 contexts. This context similarity is defined by comparing their
11355 surrounding patches of size p. Patches are searched in an area of r
11356 around the sample.
11357 The filter accepts the following options:
11359 strength, s
11361 Set denoising strength. Allowed range is from 0.00001 to 10000.
11362 Default value is 0.00001.
11363 patch, p
11365 Set patch radius duration. Allowed range is from 1 to 100
11366 milliseconds. Default value is 2 milliseconds.
11367 research, r
11369 Set research radius duration. Allowed range is from 2 to 300
11370 milliseconds. Default value is 6 milliseconds.
11371 output, o
11373 Set the output mode.
11374 It accepts the following values:
11376 i Pass input unchanged.
11378 o Pass noise filtered out.
11380 n Pass only noise.
11382 Default value is o.
11384 smooth, m
11386 Set smooth factor. Default value is 11. Allowed range is from 1 to
11387 1000.
11388 Commands
11390 This filter supports the all above options as commands.
11392 anlmf, anlms
11394 Apply Normalized Least-Mean-(Squares|Fourth) algorithm to the first
11395 audio stream using the second audio stream.
11396 This adaptive filter is used to mimic a desired filter by finding the
11398 filter coefficients that relate to producing the least mean square of
11399 the error signal (difference between the desired, 2nd input audio
11400 stream and the actual signal, the 1st input audio stream).
11401 A description of the accepted options follows.
11403 order
11405 Set filter order.
11406 mu Set filter mu.
11408 eps Set the filter eps.
11410 leakage
11412 Set the filter leakage.
11413 out_mode
11415 It accepts the following values:
11416 i Pass the 1st input.
11418 d Pass the 2nd input.
11420 o Pass filtered samples.
11422 n Pass difference between desired and filtered samples.
11424 Default value is o.
11426 Examples
11428 • One of many usages of this filter is noise reduction, input audio
11430 is filtered with same samples that are delayed by fixed amount, one
11431 such example for stereo audio is:
11432 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
11434 Commands
11436 This filter supports the same commands as options, excluding option
11438 "order".
11439 anull
11441 Pass the audio source unchanged to the output.
11442 apad
11444 Pad the end of an audio stream with silence.
11445 This can be used together with ffmpeg -shortest to extend audio streams
11447 to the same length as the video stream.
11448 A description of the accepted options follows.
11450 packet_size
11452 Set silence packet size. Default value is 4096.
11453 pad_len
11455 Set the number of samples of silence to add to the end. After the
11456 value is reached, the stream is terminated. This option is mutually
11457 exclusive with whole_len.
11458 whole_len
11460 Set the minimum total number of samples in the output audio stream.
11461 If the value is longer than the input audio length, silence is
11462 added to the end, until the value is reached. This option is
11463 mutually exclusive with pad_len.
11464 pad_dur
11466 Specify the duration of samples of silence to add. See the Time
11467 duration section in the ffmpeg-utils(1) manual for the accepted
11468 syntax. Used only if set to non-negative value.
11469 whole_dur
11471 Specify the minimum total duration in the output audio stream. See
11472 the Time duration section in the ffmpeg-utils(1) manual for the
11473 accepted syntax. Used only if set to non-negative value. If the
11474 value is longer than the input audio length, silence is added to
11475 the end, until the value is reached. This option is mutually
11476 exclusive with pad_dur
11477 If neither the pad_len nor the whole_len nor pad_dur nor whole_dur
11479 option is set, the filter will add silence to the end of the input
11480 stream indefinitely.
11481 Note that for ffmpeg 4.4 and earlier a zero pad_dur or whole_dur also
11483 caused the filter to add silence indefinitely.
11484 Examples
11486 • Add 1024 samples of silence to the end of the input:
11488 apad=pad_len=1024
11490 • Make sure the audio output will contain at least 10000 samples, pad
11492 the input with silence if required:
11493 apad=whole_len=10000
11495 • Use ffmpeg to pad the audio input with silence, so that the video
11497 stream will always result the shortest and will be converted until
11498 the end in the output file when using the shortest option:
11499 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
11501 aphaser
11503 Add a phasing effect to the input audio.
11504 A phaser filter creates series of peaks and troughs in the frequency
11506 spectrum. The position of the peaks and troughs are modulated so that
11507 they vary over time, creating a sweeping effect.
11508 A description of the accepted parameters follows.
11510 in_gain
11512 Set input gain. Default is 0.4.
11513 out_gain
11515 Set output gain. Default is 0.74
11516 delay
11518 Set delay in milliseconds. Default is 3.0.
11519 decay
11521 Set decay. Default is 0.4.
11522 speed
11524 Set modulation speed in Hz. Default is 0.5.
11525 type
11527 Set modulation type. Default is triangular.
11528 It accepts the following values:
11530 triangular, t
11532 sinusoidal, s
11533 aphaseshift
11535 Apply phase shift to input audio samples.
11536 The filter accepts the following options:
11538 shift
11540 Specify phase shift. Allowed range is from -1.0 to 1.0. Default
11541 value is 0.0.
11542 level
11544 Set output gain applied to final output. Allowed range is from 0.0
11545 to 1.0. Default value is 1.0.
11546 order
11548 Set filter order used for filtering. Allowed range is from 1 to 16.
11549 Default value is 8.
11550 Commands
11552 This filter supports the all above options as commands.
11554 apsyclip
11556 Apply Psychoacoustic clipper to input audio stream.
11557 The filter accepts the following options:
11559 level_in
11561 Set input gain. By default it is 1. Range is [0.015625 - 64].
11562 level_out
11564 Set output gain. By default it is 1. Range is [0.015625 - 64].
11565 clip
11567 Set the clipping start value. Default value is 0dBFS or 1.
11568 diff
11570 Output only difference samples, useful to hear introduced
11571 distortions. By default is disabled.
11572 adaptive
11574 Set strength of adaptive distortion applied. Default value is 0.5.
11575 Allowed range is from 0 to 1.
11576 iterations
11578 Set number of iterations of psychoacoustic clipper. Allowed range
11579 is from 1 to 20. Default value is 10.
11580 level
11582 Auto level output signal. Default is disabled. This normalizes
11583 audio back to 0dBFS if enabled.
11584 Commands
11586 This filter supports the all above options as commands.
11588 apulsator
11590 Audio pulsator is something between an autopanner and a tremolo. But
11591 it can produce funny stereo effects as well. Pulsator changes the
11592 volume of the left and right channel based on a LFO (low frequency
11593 oscillator) with different waveforms and shifted phases. This filter
11594 have the ability to define an offset between left and right channel. An
11595 offset of 0 means that both LFO shapes match each other. The left and
11596 right channel are altered equally - a conventional tremolo. An offset
11597 of 50% means that the shape of the right channel is exactly shifted in
11598 phase (or moved backwards about half of the frequency) - pulsator acts
11599 as an autopanner. At 1 both curves match again. Every setting in
11600 between moves the phase shift gapless between all stages and produces
11601 some "bypassing" sounds with sine and triangle waveforms. The more you
11602 set the offset near 1 (starting from the 0.5) the faster the signal
11603 passes from the left to the right speaker.
11604 The filter accepts the following options:
11606 level_in
11608 Set input gain. By default it is 1. Range is [0.015625 - 64].
11609 level_out
11611 Set output gain. By default it is 1. Range is [0.015625 - 64].
11612 mode
11614 Set waveform shape the LFO will use. Can be one of: sine, triangle,
11615 square, sawup or sawdown. Default is sine.
11616 amount
11618 Set modulation. Define how much of original signal is affected by
11619 the LFO.
11620 offset_l
11622 Set left channel offset. Default is 0. Allowed range is [0 - 1].
11623 offset_r
11625 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
11626 width
11628 Set pulse width. Default is 1. Allowed range is [0 - 2].
11629 timing
11631 Set possible timing mode. Can be one of: bpm, ms or hz. Default is
11632 hz.
11633 bpm Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if
11635 timing is set to bpm.
11636 ms Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if
11638 timing is set to ms.
11639 hz Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100].
11641 Only used if timing is set to hz.
11642 aresample
11644 Resample the input audio to the specified parameters, using the
11645 libswresample library. If none are specified then the filter will
11646 automatically convert between its input and output.
11647 This filter is also able to stretch/squeeze the audio data to make it
11649 match the timestamps or to inject silence / cut out audio to make it
11650 match the timestamps, do a combination of both or do neither.
11651 The filter accepts the syntax [sample_rate:]resampler_options, where
11653 sample_rate expresses a sample rate and resampler_options is a list of
11654 key=value pairs, separated by ":". See the "Resampler Options" section
11655 in the ffmpeg-resampler(1) manual for the complete list of supported
11656 options.
11657 Examples
11659 • Resample the input audio to 44100Hz:
11661 aresample=44100
11663 • Stretch/squeeze samples to the given timestamps, with a maximum of
11665 1000 samples per second compensation:
11666 aresample=async=1000
11668 areverse
11670 Reverse an audio clip.
11671 Warning: This filter requires memory to buffer the entire clip, so
11673 trimming is suggested.
11674 Examples
11676 • Take the first 5 seconds of a clip, and reverse it.
11678 atrim=end=5,areverse
11680 arnndn
11682 Reduce noise from speech using Recurrent Neural Networks.
11683 This filter accepts the following options:
11685 model, m
11687 Set train model file to load. This option is always required.
11688 mix Set how much to mix filtered samples into final output. Allowed
11690 range is from -1 to 1. Default value is 1. Negative values are
11691 special, they set how much to keep filtered noise in the final
11692 filter output. Set this option to -1 to hear actual noise removed
11693 from input signal.
11694 Commands
11696 This filter supports the all above options as commands.
11698 asdr
11700 Measure Audio Signal-to-Distortion Ratio.
11701 This filter takes two audio streams for input, and outputs first audio
11703 stream. Results are in dB per channel at end of either input.
11704 asetnsamples
11706 Set the number of samples per each output audio frame.
11707 The last output packet may contain a different number of samples, as
11709 the filter will flush all the remaining samples when the input audio
11710 signals its end.
11711 The filter accepts the following options:
11713 nb_out_samples, n
11715 Set the number of frames per each output audio frame. The number is
11716 intended as the number of samples per each channel. Default value
11717 is 1024.
11718 pad, p
11720 If set to 1, the filter will pad the last audio frame with zeroes,
11721 so that the last frame will contain the same number of samples as
11722 the previous ones. Default value is 1.
11723 For example, to set the number of per-frame samples to 1234 and disable
11725 padding for the last frame, use:
11726 asetnsamples=n=1234:p=0
11728 asetrate
11730 Set the sample rate without altering the PCM data. This will result in
11731 a change of speed and pitch.
11732 The filter accepts the following options:
11734 sample_rate, r
11736 Set the output sample rate. Default is 44100 Hz.
11737 ashowinfo
11739 Show a line containing various information for each input audio frame.
11740 The input audio is not modified.
11741 The shown line contains a sequence of key/value pairs of the form
11743 key:value.
11744 The following values are shown in the output:
11746 n The (sequential) number of the input frame, starting from 0.
11748 pts The presentation timestamp of the input frame, in time base units;
11750 the time base depends on the filter input pad, and is usually
11751 1/sample_rate.
11752 pts_time
11754 The presentation timestamp of the input frame in seconds.
11755 pos position of the frame in the input stream, -1 if this information
11757 in unavailable and/or meaningless (for example in case of synthetic
11758 audio)
11759 fmt The sample format.
11761 chlayout
11763 The channel layout.
11764 rate
11766 The sample rate for the audio frame.
11767 nb_samples
11769 The number of samples (per channel) in the frame.
11770 checksum
11772 The Adler-32 checksum (printed in hexadecimal) of the audio data.
11773 For planar audio, the data is treated as if all the planes were
11774 concatenated.
11775 plane_checksums
11777 A list of Adler-32 checksums for each data plane.
11778 asoftclip
11780 Apply audio soft clipping.
11781 Soft clipping is a type of distortion effect where the amplitude of a
11783 signal is saturated along a smooth curve, rather than the abrupt shape
11784 of hard-clipping.
11785 This filter accepts the following options:
11787 type
11789 Set type of soft-clipping.
11790 It accepts the following values:
11792 hard
11794 tanh
11795 atan
11796 cubic
11797 exp
11798 alg
11799 quintic
11800 sin
11801 erf
11802 threshold
11803 Set threshold from where to start clipping. Default value is 0dB or
11804 1.
11805 output
11807 Set gain applied to output. Default value is 0dB or 1.
11808 param
11810 Set additional parameter which controls sigmoid function.
11811 oversample
11813 Set oversampling factor.
11814 Commands
11816 This filter supports the all above options as commands.
11818 aspectralstats
11820 Display frequency domain statistical information about the audio
11821 channels. Statistics are calculated and stored as metadata for each
11822 audio channel and for each audio frame.
11823 It accepts the following option:
11825 win_size
11827 Set the window length in samples. Default value is 2048. Allowed
11828 range is from 32 to 65536.
11829 win_func
11831 Set window function.
11832 It accepts the following values:
11834 rect
11836 bartlett
11837 hann, hanning
11838 hamming
11839 blackman
11840 welch
11841 flattop
11842 bharris
11843 bnuttall
11844 bhann
11845 sine
11846 nuttall
11847 lanczos
11848 gauss
11849 tukey
11850 dolph
11851 cauchy
11852 parzen
11853 poisson
11854 bohman
11855 kaiser
11856 Default is "hann".
11858 overlap
11860 Set window overlap. Allowed range is from 0 to 1. Default value is
11861 0.5.
11862 measure
11864 Select the parameters which are measured. The metadata keys can be
11865 used as flags, default is all which measures everything. none
11866 disables all measurement.
11867 A list of each metadata key follows:
11869 mean
11871 variance
11872 centroid
11873 spread
11874 skewness
11875 kurtosis
11876 entropy
11877 flatness
11878 crest
11879 flux
11880 slope
11881 decrease
11882 rolloff
11883 asr
11885 Automatic Speech Recognition
11886 This filter uses PocketSphinx for speech recognition. To enable
11888 compilation of this filter, you need to configure FFmpeg with
11889 "--enable-pocketsphinx".
11890 It accepts the following options:
11892 rate
11894 Set sampling rate of input audio. Defaults is 16000. This need to
11895 match speech models, otherwise one will get poor results.
11896 hmm Set dictionary containing acoustic model files.
11898 dict
11900 Set pronunciation dictionary.
11901 lm Set language model file.
11903 lmctl
11905 Set language model set.
11906 lmname
11908 Set which language model to use.
11909 logfn
11911 Set output for log messages.
11912 The filter exports recognized speech as the frame metadata
11914 "lavfi.asr.text".
11915 astats
11917 Display time domain statistical information about the audio channels.
11918 Statistics are calculated and displayed for each audio channel and,
11919 where applicable, an overall figure is also given.
11920 It accepts the following option:
11922 length
11924 Short window length in seconds, used for peak and trough RMS
11925 measurement. Default is 0.05 (50 milliseconds). Allowed range is
11926 "[0 - 10]".
11927 metadata
11929 Set metadata injection. All the metadata keys are prefixed with
11930 "lavfi.astats.X", where "X" is channel number starting from 1 or
11931 string "Overall". Default is disabled.
11932 Available keys for each channel are: Bit_depth Crest_factor
11934 DC_offset Dynamic_range Entropy Flat_factor Max_difference
11935 Max_level Mean_difference Min_difference Min_level Noise_floor
11936 Noise_floor_count Number_of_Infs Number_of_NaNs Number_of_denormals
11937 Peak_count Peak_level RMS_difference RMS_peak RMS_trough
11938 Zero_crossings Zero_crossings_rate
11939 and for "Overall": Bit_depth DC_offset Entropy Flat_factor
11941 Max_difference Max_level Mean_difference Min_difference Min_level
11942 Noise_floor Noise_floor_count Number_of_Infs Number_of_NaNs
11943 Number_of_denormals Number_of_samples Peak_count Peak_level
11944 RMS_difference RMS_level RMS_peak RMS_trough
11945 For example, a full key looks like "lavfi.astats.1.DC_offset" or
11947 "lavfi.astats.Overall.Peak_count".
11948 Read below for the description of the keys.
11950 reset
11952 Set the number of frames over which cumulative stats are calculated
11953 before being reset. Default is disabled.
11954 measure_perchannel
11956 Select the parameters which are measured per channel. The metadata
11957 keys can be used as flags, default is all which measures
11958 everything. none disables all per channel measurement.
11959 measure_overall
11961 Select the parameters which are measured overall. The metadata keys
11962 can be used as flags, default is all which measures everything.
11963 none disables all overall measurement.
11964 A description of the measure keys follow:
11966 none
11968 no measures
11969 all all measures
11971 Bit_depth
11973 overall bit depth of audio, i.e. number of bits used for each
11974 sample
11975 Crest_factor
11977 standard ratio of peak to RMS level (note: not in dB)
11978 DC_offset
11980 mean amplitude displacement from zero
11981 Dynamic_range
11983 measured dynamic range of audio in dB
11984 Entropy
11986 entropy measured across whole audio, entropy of value near 1.0 is
11987 typically measured for white noise
11988 Flat_factor
11990 flatness (i.e. consecutive samples with the same value) of the
11991 signal at its peak levels (i.e. either Min_level or Max_level)
11992 Max_difference
11994 maximal difference between two consecutive samples
11995 Max_level
11997 maximal sample level
11998 Mean_difference
12000 mean difference between two consecutive samples, i.e. the average
12001 of each difference between two consecutive samples
12002 Min_difference
12004 minimal difference between two consecutive samples
12005 Min_level
12007 minimal sample level
12008 Noise_floor
12010 minimum local peak measured in dBFS over a short window
12011 Noise_floor_count
12013 number of occasions (not the number of samples) that the signal
12014 attained Noise floor
12015 Number_of_Infs
12017 number of samples with an infinite value
12018 Number_of_NaNs
12020 number of samples with a NaN (not a number) value
12021 Number_of_denormals
12023 number of samples with a subnormal value
12024 Number_of_samples
12026 number of samples
12027 Peak_count
12029 number of occasions (not the number of samples) that the signal
12030 attained either Min_level or Max_level
12031 Peak_level
12033 standard peak level measured in dBFS
12034 RMS_difference
12036 Root Mean Square difference between two consecutive samples
12037 RMS_level
12039 standard RMS level measured in dBFS
12040 RMS_peak
12042 RMS_trough
12043 peak and trough values for RMS level measured over a short window,
12044 measured in dBFS.
12045 Zero crossings
12047 number of points where the waveform crosses the zero level axis
12048 Zero crossings rate
12050 rate of Zero crossings and number of audio samples
12051 asubboost
12053 Boost subwoofer frequencies.
12054 The filter accepts the following options:
12056 dry Set dry gain, how much of original signal is kept. Allowed range is
12058 from 0 to 1. Default value is 1.0.
12059 wet Set wet gain, how much of filtered signal is kept. Allowed range is
12061 from 0 to 1. Default value is 1.0.
12062 boost
12064 Set max boost factor. Allowed range is from 1 to 12. Default value
12065 is 2.
12066 decay
12068 Set delay line decay gain value. Allowed range is from 0 to 1.
12069 Default value is 0.0.
12070 feedback
12072 Set delay line feedback gain value. Allowed range is from 0 to 1.
12073 Default value is 0.9.
12074 cutoff
12076 Set cutoff frequency in Hertz. Allowed range is 50 to 900. Default
12077 value is 100.
12078 slope
12080 Set slope amount for cutoff frequency. Allowed range is 0.0001 to
12081 1. Default value is 0.5.
12082 delay
12084 Set delay. Allowed range is from 1 to 100. Default value is 20.
12085 channels
12087 Set the channels to process. Default value is all available.
12088 Commands
12090 This filter supports the all above options as commands.
12092 asubcut
12094 Cut subwoofer frequencies.
12095 This filter allows to set custom, steeper roll off than highpass
12097 filter, and thus is able to more attenuate frequency content in stop-
12098 band.
12099 The filter accepts the following options:
12101 cutoff
12103 Set cutoff frequency in Hertz. Allowed range is 2 to 200. Default
12104 value is 20.
12105 order
12107 Set filter order. Available values are from 3 to 20. Default value
12108 is 10.
12109 level
12111 Set input gain level. Allowed range is from 0 to 1. Default value
12112 is 1.
12113 Commands
12115 This filter supports the all above options as commands.
12117 asupercut
12119 Cut super frequencies.
12120 The filter accepts the following options:
12122 cutoff
12124 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
12125 Default value is 20000.
12126 order
12128 Set filter order. Available values are from 3 to 20. Default value
12129 is 10.
12130 level
12132 Set input gain level. Allowed range is from 0 to 1. Default value
12133 is 1.
12134 Commands
12136 This filter supports the all above options as commands.
12138 asuperpass
12140 Apply high order Butterworth band-pass filter.
12141 The filter accepts the following options:
12143 centerf
12145 Set center frequency in Hertz. Allowed range is 2 to 999999.
12146 Default value is 1000.
12147 order
12149 Set filter order. Available values are from 4 to 20. Default value
12150 is 4.
12151 qfactor
12153 Set Q-factor. Allowed range is from 0.01 to 100. Default value is
12154 1.
12155 level
12157 Set input gain level. Allowed range is from 0 to 2. Default value
12158 is 1.
12159 Commands
12161 This filter supports the all above options as commands.
12163 asuperstop
12165 Apply high order Butterworth band-stop filter.
12166 The filter accepts the following options:
12168 centerf
12170 Set center frequency in Hertz. Allowed range is 2 to 999999.
12171 Default value is 1000.
12172 order
12174 Set filter order. Available values are from 4 to 20. Default value
12175 is 4.
12176 qfactor
12178 Set Q-factor. Allowed range is from 0.01 to 100. Default value is
12179 1.
12180 level
12182 Set input gain level. Allowed range is from 0 to 2. Default value
12183 is 1.
12184 Commands
12186 This filter supports the all above options as commands.
12188 atempo
12190 Adjust audio tempo.
12191 The filter accepts exactly one parameter, the audio tempo. If not
12193 specified then the filter will assume nominal 1.0 tempo. Tempo must be
12194 in the [0.5, 100.0] range.
12195 Note that tempo greater than 2 will skip some samples rather than blend
12197 them in. If for any reason this is a concern it is always possible to
12198 daisy-chain several instances of atempo to achieve the desired product
12199 tempo.
12200 Examples
12202 • Slow down audio to 80% tempo:
12204 atempo=0.8
12206 • To speed up audio to 300% tempo:
12208 atempo=3
12210 • To speed up audio to 300% tempo by daisy-chaining two atempo
12212 instances:
12213 atempo=sqrt(3),atempo=sqrt(3)
12215 Commands
12217 This filter supports the following commands:
12219 tempo
12221 Change filter tempo scale factor. Syntax for the command is :
12222 "tempo"
12223 atilt
12225 Apply spectral tilt filter to audio stream.
12226 This filter apply any spectral roll-off slope over any specified
12228 frequency band.
12229 The filter accepts the following options:
12231 freq
12233 Set central frequency of tilt in Hz. Default is 10000 Hz.
12234 slope
12236 Set slope direction of tilt. Default is 0. Allowed range is from -1
12237 to 1.
12238 width
12240 Set width of tilt. Default is 1000. Allowed range is from 100 to
12241 10000.
12242 order
12244 Set order of tilt filter.
12245 level
12247 Set input volume level. Allowed range is from 0 to 4. Defalt is 1.
12248 Commands
12250 This filter supports the all above options as commands.
12252 atrim
12254 Trim the input so that the output contains one continuous subpart of
12255 the input.
12256 It accepts the following parameters:
12258 start
12260 Timestamp (in seconds) of the start of the section to keep. I.e.
12261 the audio sample with the timestamp start will be the first sample
12262 in the output.
12263 end Specify time of the first audio sample that will be dropped, i.e.
12265 the audio sample immediately preceding the one with the timestamp
12266 end will be the last sample in the output.
12267 start_pts
12269 Same as start, except this option sets the start timestamp in
12270 samples instead of seconds.
12271 end_pts
12273 Same as end, except this option sets the end timestamp in samples
12274 instead of seconds.
12275 duration
12277 The maximum duration of the output in seconds.
12278 start_sample
12280 The number of the first sample that should be output.
12281 end_sample
12283 The number of the first sample that should be dropped.
12284 start, end, and duration are expressed as time duration specifications;
12286 see the Time duration section in the ffmpeg-utils(1) manual.
12287 Note that the first two sets of the start/end options and the duration
12289 option look at the frame timestamp, while the _sample options simply
12290 count the samples that pass through the filter. So start/end_pts and
12291 start/end_sample will give different results when the timestamps are
12292 wrong, inexact or do not start at zero. Also note that this filter does
12293 not modify the timestamps. If you wish to have the output timestamps
12294 start at zero, insert the asetpts filter after the atrim filter.
12295 If multiple start or end options are set, this filter tries to be
12297 greedy and keep all samples that match at least one of the specified
12298 constraints. To keep only the part that matches all the constraints at
12299 once, chain multiple atrim filters.
12300 The defaults are such that all the input is kept. So it is possible to
12302 set e.g. just the end values to keep everything before the specified
12303 time.
12304 Examples:
12306 • Drop everything except the second minute of input:
12308 ffmpeg -i INPUT -af atrim=60:120
12310 • Keep only the first 1000 samples:
12312 ffmpeg -i INPUT -af atrim=end_sample=1000
12314 axcorrelate
12316 Calculate normalized windowed cross-correlation between two input audio
12317 streams.
12318 Resulted samples are always between -1 and 1 inclusive. If result is 1
12320 it means two input samples are highly correlated in that selected
12321 segment. Result 0 means they are not correlated at all. If result is
12322 -1 it means two input samples are out of phase, which means they cancel
12323 each other.
12324 The filter accepts the following options:
12326 size
12328 Set size of segment over which cross-correlation is calculated.
12329 Default is 256. Allowed range is from 2 to 131072.
12330 algo
12332 Set algorithm for cross-correlation. Can be "slow" or "fast".
12333 Default is "slow". Fast algorithm assumes mean values over any
12334 given segment are always zero and thus need much less calculations
12335 to make. This is generally not true, but is valid for typical
12336 audio streams.
12337 Examples
12339 • Calculate correlation between channels in stereo audio stream:
12341 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
12343 bandpass
12345 Apply a two-pole Butterworth band-pass filter with central frequency
12346 frequency, and (3dB-point) band-width width. The csg option selects a
12347 constant skirt gain (peak gain = Q) instead of the default: constant
12348 0dB peak gain. The filter roll off at 6dB per octave (20dB per
12349 decade).
12350 The filter accepts the following options:
12352 frequency, f
12354 Set the filter's central frequency. Default is 3000.
12355 csg Constant skirt gain if set to 1. Defaults to 0.
12357 width_type, t
12359 Set method to specify band-width of filter.
12360 h Hz
12362 q Q-Factor
12364 o octave
12366 s slope
12368 k kHz
12370 width, w
12372 Specify the band-width of a filter in width_type units.
12373 mix, m
12375 How much to use filtered signal in output. Default is 1. Range is
12376 between 0 and 1.
12377 channels, c
12379 Specify which channels to filter, by default all available are
12380 filtered.
12381 normalize, n
12383 Normalize biquad coefficients, by default is disabled. Enabling it
12384 will normalize magnitude response at DC to 0dB.
12385 transform, a
12387 Set transform type of IIR filter.
12388 di
12390 dii
12391 tdi
12392 tdii
12393 latt
12394 svf
12395 zdf
12396 precision, r
12397 Set precison of filtering.
12398 auto
12400 Pick automatic sample format depending on surround filters.
12401 s16 Always use signed 16-bit.
12403 s32 Always use signed 32-bit.
12405 f32 Always use float 32-bit.
12407 f64 Always use float 64-bit.
12409 block_size, b
12411 Set block size used for reverse IIR processing. If this value is
12412 set to high enough value (higher than impulse response length
12413 truncated when reaches near zero values) filtering will become
12414 linear phase otherwise if not big enough it will just produce nasty
12415 artifacts.
12416 Note that filter delay will be exactly this many samples when set
12418 to non-zero value.
12419 Commands
12421 This filter supports the following commands:
12423 frequency, f
12425 Change bandpass frequency. Syntax for the command is : "frequency"
12426 width_type, t
12428 Change bandpass width_type. Syntax for the command is :
12429 "width_type"
12430 width, w
12432 Change bandpass width. Syntax for the command is : "width"
12433 mix, m
12435 Change bandpass mix. Syntax for the command is : "mix"
12436 bandreject
12438 Apply a two-pole Butterworth band-reject filter with central frequency
12439 frequency, and (3dB-point) band-width width. The filter roll off at
12440 6dB per octave (20dB per decade).
12441 The filter accepts the following options:
12443 frequency, f
12445 Set the filter's central frequency. Default is 3000.
12446 width_type, t
12448 Set method to specify band-width of filter.
12449 h Hz
12451 q Q-Factor
12453 o octave
12455 s slope
12457 k kHz
12459 width, w
12461 Specify the band-width of a filter in width_type units.
12462 mix, m
12464 How much to use filtered signal in output. Default is 1. Range is
12465 between 0 and 1.
12466 channels, c
12468 Specify which channels to filter, by default all available are
12469 filtered.
12470 normalize, n
12472 Normalize biquad coefficients, by default is disabled. Enabling it
12473 will normalize magnitude response at DC to 0dB.
12474 transform, a
12476 Set transform type of IIR filter.
12477 di
12479 dii
12480 tdi
12481 tdii
12482 latt
12483 svf
12484 zdf
12485 precision, r
12486 Set precison of filtering.
12487 auto
12489 Pick automatic sample format depending on surround filters.
12490 s16 Always use signed 16-bit.
12492 s32 Always use signed 32-bit.
12494 f32 Always use float 32-bit.
12496 f64 Always use float 64-bit.
12498 block_size, b
12500 Set block size used for reverse IIR processing. If this value is
12501 set to high enough value (higher than impulse response length
12502 truncated when reaches near zero values) filtering will become
12503 linear phase otherwise if not big enough it will just produce nasty
12504 artifacts.
12505 Note that filter delay will be exactly this many samples when set
12507 to non-zero value.
12508 Commands
12510 This filter supports the following commands:
12512 frequency, f
12514 Change bandreject frequency. Syntax for the command is :
12515 "frequency"
12516 width_type, t
12518 Change bandreject width_type. Syntax for the command is :
12519 "width_type"
12520 width, w
12522 Change bandreject width. Syntax for the command is : "width"
12523 mix, m
12525 Change bandreject mix. Syntax for the command is : "mix"
12526 bass, lowshelf
12528 Boost or cut the bass (lower) frequencies of the audio using a two-pole
12529 shelving filter with a response similar to that of a standard hi-fi's
12530 tone-controls. This is also known as shelving equalisation (EQ).
12531 The filter accepts the following options:
12533 gain, g
12535 Give the gain at 0 Hz. Its useful range is about -20 (for a large
12536 cut) to +20 (for a large boost). Beware of clipping when using a
12537 positive gain.
12538 frequency, f
12540 Set the filter's central frequency and so can be used to extend or
12541 reduce the frequency range to be boosted or cut. The default value
12542 is 100 Hz.
12543 width_type, t
12545 Set method to specify band-width of filter.
12546 h Hz
12548 q Q-Factor
12550 o octave
12552 s slope
12554 k kHz
12556 width, w
12558 Determine how steep is the filter's shelf transition.
12559 poles, p
12561 Set number of poles. Default is 2.
12562 mix, m
12564 How much to use filtered signal in output. Default is 1. Range is
12565 between 0 and 1.
12566 channels, c
12568 Specify which channels to filter, by default all available are
12569 filtered.
12570 normalize, n
12572 Normalize biquad coefficients, by default is disabled. Enabling it
12573 will normalize magnitude response at DC to 0dB.
12574 transform, a
12576 Set transform type of IIR filter.
12577 di
12579 dii
12580 tdi
12581 tdii
12582 latt
12583 svf
12584 zdf
12585 precision, r
12586 Set precison of filtering.
12587 auto
12589 Pick automatic sample format depending on surround filters.
12590 s16 Always use signed 16-bit.
12592 s32 Always use signed 32-bit.
12594 f32 Always use float 32-bit.
12596 f64 Always use float 64-bit.
12598 block_size, b
12600 Set block size used for reverse IIR processing. If this value is
12601 set to high enough value (higher than impulse response length
12602 truncated when reaches near zero values) filtering will become
12603 linear phase otherwise if not big enough it will just produce nasty
12604 artifacts.
12605 Note that filter delay will be exactly this many samples when set
12607 to non-zero value.
12608 Commands
12610 This filter supports the following commands:
12612 frequency, f
12614 Change bass frequency. Syntax for the command is : "frequency"
12615 width_type, t
12617 Change bass width_type. Syntax for the command is : "width_type"
12618 width, w
12620 Change bass width. Syntax for the command is : "width"
12621 gain, g
12623 Change bass gain. Syntax for the command is : "gain"
12624 mix, m
12626 Change bass mix. Syntax for the command is : "mix"
12627 biquad
12629 Apply a biquad IIR filter with the given coefficients. Where b0, b1,
12630 b2 and a0, a1, a2 are the numerator and denominator coefficients
12631 respectively. and channels, c specify which channels to filter, by
12632 default all available are filtered.
12633 Commands
12635 This filter supports the following commands:
12637 a0
12639 a1
12640 a2
12641 b0
12642 b1
12643 b2 Change biquad parameter. Syntax for the command is : "value"
12644 mix, m
12646 How much to use filtered signal in output. Default is 1. Range is
12647 between 0 and 1.
12648 channels, c
12650 Specify which channels to filter, by default all available are
12651 filtered.
12652 normalize, n
12654 Normalize biquad coefficients, by default is disabled. Enabling it
12655 will normalize magnitude response at DC to 0dB.
12656 transform, a
12658 Set transform type of IIR filter.
12659 di
12661 dii
12662 tdi
12663 tdii
12664 latt
12665 svf
12666 zdf
12667 precision, r
12668 Set precison of filtering.
12669 auto
12671 Pick automatic sample format depending on surround filters.
12672 s16 Always use signed 16-bit.
12674 s32 Always use signed 32-bit.
12676 f32 Always use float 32-bit.
12678 f64 Always use float 64-bit.
12680 block_size, b
12682 Set block size used for reverse IIR processing. If this value is
12683 set to high enough value (higher than impulse response length
12684 truncated when reaches near zero values) filtering will become
12685 linear phase otherwise if not big enough it will just produce nasty
12686 artifacts.
12687 Note that filter delay will be exactly this many samples when set
12689 to non-zero value.
12690 bs2b
12692 Bauer stereo to binaural transformation, which improves headphone
12693 listening of stereo audio records.
12694 To enable compilation of this filter you need to configure FFmpeg with
12696 "--enable-libbs2b".
12697 It accepts the following parameters:
12699 profile
12701 Pre-defined crossfeed level.
12702 default
12704 Default level (fcut=700, feed=50).
12705 cmoy
12707 Chu Moy circuit (fcut=700, feed=60).
12708 jmeier
12710 Jan Meier circuit (fcut=650, feed=95).
12711 fcut
12713 Cut frequency (in Hz).
12714 feed
12716 Feed level (in Hz).
12717 channelmap
12719 Remap input channels to new locations.
12720 It accepts the following parameters:
12722 map Map channels from input to output. The argument is a '|'-separated
12724 list of mappings, each in the "in_channel-out_channel" or
12725 in_channel form. in_channel can be either the name of the input
12726 channel (e.g. FL for front left) or its index in the input channel
12727 layout. out_channel is the name of the output channel or its index
12728 in the output channel layout. If out_channel is not given then it
12729 is implicitly an index, starting with zero and increasing by one
12730 for each mapping.
12731 channel_layout
12733 The channel layout of the output stream.
12734 If no mapping is present, the filter will implicitly map input channels
12736 to output channels, preserving indices.
12737 Examples
12739 • For example, assuming a 5.1+downmix input MOV file,
12741 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
12743 will create an output WAV file tagged as stereo from the downmix
12745 channels of the input.
12746 • To fix a 5.1 WAV improperly encoded in AAC's native channel order
12748 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
12750 channelsplit
12752 Split each channel from an input audio stream into a separate output
12753 stream.
12754 It accepts the following parameters:
12756 channel_layout
12758 The channel layout of the input stream. The default is "stereo".
12759 channels
12761 A channel layout describing the channels to be extracted as
12762 separate output streams or "all" to extract each input channel as a
12763 separate stream. The default is "all".
12764 Choosing channels not present in channel layout in the input will
12766 result in an error.
12767 Examples
12769 • For example, assuming a stereo input MP3 file,
12771 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
12773 will create an output Matroska file with two audio streams, one
12775 containing only the left channel and the other the right channel.
12776 • Split a 5.1 WAV file into per-channel files:
12778 ffmpeg -i in.wav -filter_complex
12780 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
12781 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
12782 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
12783 side_right.wav
12784 • Extract only LFE from a 5.1 WAV file:
12786 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
12788 -map '[LFE]' lfe.wav
12789 chorus
12791 Add a chorus effect to the audio.
12792 Can make a single vocal sound like a chorus, but can also be applied to
12794 instrumentation.
12795 Chorus resembles an echo effect with a short delay, but whereas with
12797 echo the delay is constant, with chorus, it is varied using using
12798 sinusoidal or triangular modulation. The modulation depth defines the
12799 range the modulated delay is played before or after the delay. Hence
12800 the delayed sound will sound slower or faster, that is the delayed
12801 sound tuned around the original one, like in a chorus where some vocals
12802 are slightly off key.
12803 It accepts the following parameters:
12805 in_gain
12807 Set input gain. Default is 0.4.
12808 out_gain
12810 Set output gain. Default is 0.4.
12811 delays
12813 Set delays. A typical delay is around 40ms to 60ms.
12814 decays
12816 Set decays.
12817 speeds
12819 Set speeds.
12820 depths
12822 Set depths.
12823 Examples
12825 • A single delay:
12827 chorus=0.7:0.9:55:0.4:0.25:2
12829 • Two delays:
12831 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
12833 • Fuller sounding chorus with three delays:
12835 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
12837 compand
12839 Compress or expand the audio's dynamic range.
12840 It accepts the following parameters:
12842 attacks
12844 decays
12845 A list of times in seconds for each channel over which the
12846 instantaneous level of the input signal is averaged to determine
12847 its volume. attacks refers to increase of volume and decays refers
12848 to decrease of volume. For most situations, the attack time
12849 (response to the audio getting louder) should be shorter than the
12850 decay time, because the human ear is more sensitive to sudden loud
12851 audio than sudden soft audio. A typical value for attack is 0.3
12852 seconds and a typical value for decay is 0.8 seconds. If specified
12853 number of attacks & decays is lower than number of channels, the
12854 last set attack/decay will be used for all remaining channels.
12855 points
12857 A list of points for the transfer function, specified in dB
12858 relative to the maximum possible signal amplitude. Each key points
12859 list must be defined using the following syntax:
12860 "x0/y0|x1/y1|x2/y2|...." or "x0/y0 x1/y1 x2/y2 ...."
12861 The input values must be in strictly increasing order but the
12863 transfer function does not have to be monotonically rising. The
12864 point "0/0" is assumed but may be overridden (by "0/out-dBn").
12865 Typical values for the transfer function are "-70/-70|-60/-20|1/0".
12866 soft-knee
12868 Set the curve radius in dB for all joints. It defaults to 0.01.
12869 gain
12871 Set the additional gain in dB to be applied at all points on the
12872 transfer function. This allows for easy adjustment of the overall
12873 gain. It defaults to 0.
12874 volume
12876 Set an initial volume, in dB, to be assumed for each channel when
12877 filtering starts. This permits the user to supply a nominal level
12878 initially, so that, for example, a very large gain is not applied
12879 to initial signal levels before the companding has begun to
12880 operate. A typical value for audio which is initially quiet is -90
12881 dB. It defaults to 0.
12882 delay
12884 Set a delay, in seconds. The input audio is analyzed immediately,
12885 but audio is delayed before being fed to the volume adjuster.
12886 Specifying a delay approximately equal to the attack/decay times
12887 allows the filter to effectively operate in predictive rather than
12888 reactive mode. It defaults to 0.
12889 Examples
12891 • Make music with both quiet and loud passages suitable for listening
12893 to in a noisy environment:
12894 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
12896 Another example for audio with whisper and explosion parts:
12898 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
12900 • A noise gate for when the noise is at a lower level than the
12902 signal:
12903 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
12905 • Here is another noise gate, this time for when the noise is at a
12907 higher level than the signal (making it, in some ways, similar to
12908 squelch):
12909 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
12911 • 2:1 compression starting at -6dB:
12913 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
12915 • 2:1 compression starting at -9dB:
12917 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
12919 • 2:1 compression starting at -12dB:
12921 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
12923 • 2:1 compression starting at -18dB:
12925 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
12927 • 3:1 compression starting at -15dB:
12929 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
12931 • Compressor/Gate:
12933 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
12935 • Expander:
12937 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
12939 • Hard limiter at -6dB:
12941 compand=attacks=0:points=-80/-80|-6/-6|20/-6
12943 • Hard limiter at -12dB:
12945 compand=attacks=0:points=-80/-80|-12/-12|20/-12
12947 • Hard noise gate at -35 dB:
12949 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
12951 • Soft limiter:
12953 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
12955 compensationdelay
12957 Compensation Delay Line is a metric based delay to compensate differing
12958 positions of microphones or speakers.
12959 For example, you have recorded guitar with two microphones placed in
12961 different locations. Because the front of sound wave has fixed speed in
12962 normal conditions, the phasing of microphones can vary and depends on
12963 their location and interposition. The best sound mix can be achieved
12964 when these microphones are in phase (synchronized). Note that a
12965 distance of ~30 cm between microphones makes one microphone capture the
12966 signal in antiphase to the other microphone. That makes the final mix
12967 sound moody. This filter helps to solve phasing problems by adding
12968 different delays to each microphone track and make them synchronized.
12969 The best result can be reached when you take one track as base and
12971 synchronize other tracks one by one with it. Remember that
12972 synchronization/delay tolerance depends on sample rate, too. Higher
12973 sample rates will give more tolerance.
12974 The filter accepts the following parameters:
12976 mm Set millimeters distance. This is compensation distance for fine
12978 tuning. Default is 0.
12979 cm Set cm distance. This is compensation distance for tightening
12981 distance setup. Default is 0.
12982 m Set meters distance. This is compensation distance for hard
12984 distance setup. Default is 0.
12985 dry Set dry amount. Amount of unprocessed (dry) signal. Default is 0.
12987 wet Set wet amount. Amount of processed (wet) signal. Default is 1.
12989 temp
12991 Set temperature in degrees Celsius. This is the temperature of the
12992 environment. Default is 20.
12993 Commands
12995 This filter supports the all above options as commands.
12997 crossfeed
12999 Apply headphone crossfeed filter.
13000 Crossfeed is the process of blending the left and right channels of
13002 stereo audio recording. It is mainly used to reduce extreme stereo
13003 separation of low frequencies.
13004 The intent is to produce more speaker like sound to the listener.
13006 The filter accepts the following options:
13008 strength
13010 Set strength of crossfeed. Default is 0.2. Allowed range is from 0
13011 to 1. This sets gain of low shelf filter for side part of stereo
13012 image. Default is -6dB. Max allowed is -30db when strength is set
13013 to 1.
13014 range
13016 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to
13017 1. This sets cut off frequency of low shelf filter. Default is cut
13018 off near 1550 Hz. With range set to 1 cut off frequency is set to
13019 2100 Hz.
13020 slope
13022 Set curve slope of low shelf filter. Default is 0.5. Allowed range
13023 is from 0.01 to 1.
13024 level_in
13026 Set input gain. Default is 0.9.
13027 level_out
13029 Set output gain. Default is 1.
13030 block_size
13032 Set block size used for reverse IIR processing. If this value is
13033 set to high enough value (higher than impulse response length
13034 truncated when reaches near zero values) filtering will become
13035 linear phase otherwise if not big enough it will just produce nasty
13036 artifacts.
13037 Note that filter delay will be exactly this many samples when set
13039 to non-zero value.
13040 Commands
13042 This filter supports the all above options as commands.
13044 crystalizer
13046 Simple algorithm for audio noise sharpening.
13047 This filter linearly increases differences betweeen each audio sample.
13049 The filter accepts the following options:
13051 i Sets the intensity of effect (default: 2.0). Must be in range
13053 between -10.0 to 0 (unchanged sound) to 10.0 (maximum effect). To
13054 inverse filtering use negative value.
13055 c Enable clipping. By default is enabled.
13057 Commands
13059 This filter supports the all above options as commands.
13061 dcshift
13063 Apply a DC shift to the audio.
13064 This can be useful to remove a DC offset (caused perhaps by a hardware
13066 problem in the recording chain) from the audio. The effect of a DC
13067 offset is reduced headroom and hence volume. The astats filter can be
13068 used to determine if a signal has a DC offset.
13069 shift
13071 Set the DC shift, allowed range is [-1, 1]. It indicates the amount
13072 to shift the audio.
13073 limitergain
13075 Optional. It should have a value much less than 1 (e.g. 0.05 or
13076 0.02) and is used to prevent clipping.
13077 deesser
13079 Apply de-essing to the audio samples.
13080 i Set intensity for triggering de-essing. Allowed range is from 0 to
13082 1. Default is 0.
13083 m Set amount of ducking on treble part of sound. Allowed range is
13085 from 0 to 1. Default is 0.5.
13086 f How much of original frequency content to keep when de-essing.
13088 Allowed range is from 0 to 1. Default is 0.5.
13089 s Set the output mode.
13091 It accepts the following values:
13093 i Pass input unchanged.
13095 o Pass ess filtered out.
13097 e Pass only ess.
13099 Default value is o.
13101 dialoguenhance
13103 Enhance dialogue in stereo audio.
13104 This filter accepts stereo input and produce surround (3.0) channels
13106 output. The newly produced front center channel have enhanced speech
13107 dialogue originally available in both stereo channels. This filter
13108 outputs front left and front right channels same as available in stereo
13109 input.
13110 The filter accepts the following options:
13112 original
13114 Set the original center factor to keep in front center channel
13115 output. Allowed range is from 0 to 1. Default value is 1.
13116 enhance
13118 Set the dialogue enhance factor to put in front center channel
13119 output. Allowed range is from 0 to 3. Default value is 1.
13120 voice
13122 Set the voice detection factor. Allowed range is from 2 to 32.
13123 Default value is 2.
13124 Commands
13126 This filter supports the all above options as commands.
13128 drmeter
13130 Measure audio dynamic range.
13131 DR values of 14 and higher is found in very dynamic material. DR of 8
13133 to 13 is found in transition material. And anything less that 8 have
13134 very poor dynamics and is very compressed.
13135 The filter accepts the following options:
13137 length
13139 Set window length in seconds used to split audio into segments of
13140 equal length. Default is 3 seconds.
13141 dynaudnorm
13143 Dynamic Audio Normalizer.
13144 This filter applies a certain amount of gain to the input audio in
13146 order to bring its peak magnitude to a target level (e.g. 0 dBFS).
13147 However, in contrast to more "simple" normalization algorithms, the
13148 Dynamic Audio Normalizer *dynamically* re-adjusts the gain factor to
13149 the input audio. This allows for applying extra gain to the "quiet"
13150 sections of the audio while avoiding distortions or clipping the "loud"
13151 sections. In other words: The Dynamic Audio Normalizer will "even out"
13152 the volume of quiet and loud sections, in the sense that the volume of
13153 each section is brought to the same target level. Note, however, that
13154 the Dynamic Audio Normalizer achieves this goal *without* applying
13155 "dynamic range compressing". It will retain 100% of the dynamic range
13156 *within* each section of the audio file.
13157 framelen, f
13159 Set the frame length in milliseconds. In range from 10 to 8000
13160 milliseconds. Default is 500 milliseconds. The Dynamic Audio
13161 Normalizer processes the input audio in small chunks, referred to
13162 as frames. This is required, because a peak magnitude has no
13163 meaning for just a single sample value. Instead, we need to
13164 determine the peak magnitude for a contiguous sequence of sample
13165 values. While a "standard" normalizer would simply use the peak
13166 magnitude of the complete file, the Dynamic Audio Normalizer
13167 determines the peak magnitude individually for each frame. The
13168 length of a frame is specified in milliseconds. By default, the
13169 Dynamic Audio Normalizer uses a frame length of 500 milliseconds,
13170 which has been found to give good results with most files. Note
13171 that the exact frame length, in number of samples, will be
13172 determined automatically, based on the sampling rate of the
13173 individual input audio file.
13174 gausssize, g
13176 Set the Gaussian filter window size. In range from 3 to 301, must
13177 be odd number. Default is 31. Probably the most important
13178 parameter of the Dynamic Audio Normalizer is the "window size" of
13179 the Gaussian smoothing filter. The filter's window size is
13180 specified in frames, centered around the current frame. For the
13181 sake of simplicity, this must be an odd number. Consequently, the
13182 default value of 31 takes into account the current frame, as well
13183 as the 15 preceding frames and the 15 subsequent frames. Using a
13184 larger window results in a stronger smoothing effect and thus in
13185 less gain variation, i.e. slower gain adaptation. Conversely, using
13186 a smaller window results in a weaker smoothing effect and thus in
13187 more gain variation, i.e. faster gain adaptation. In other words,
13188 the more you increase this value, the more the Dynamic Audio
13189 Normalizer will behave like a "traditional" normalization filter.
13190 On the contrary, the more you decrease this value, the more the
13191 Dynamic Audio Normalizer will behave like a dynamic range
13192 compressor.
13193 peak, p
13195 Set the target peak value. This specifies the highest permissible
13196 magnitude level for the normalized audio input. This filter will
13197 try to approach the target peak magnitude as closely as possible,
13198 but at the same time it also makes sure that the normalized signal
13199 will never exceed the peak magnitude. A frame's maximum local gain
13200 factor is imposed directly by the target peak magnitude. The
13201 default value is 0.95 and thus leaves a headroom of 5%*. It is not
13202 recommended to go above this value.
13203 maxgain, m
13205 Set the maximum gain factor. In range from 1.0 to 100.0. Default is
13206 10.0. The Dynamic Audio Normalizer determines the maximum possible
13207 (local) gain factor for each input frame, i.e. the maximum gain
13208 factor that does not result in clipping or distortion. The maximum
13209 gain factor is determined by the frame's highest magnitude sample.
13210 However, the Dynamic Audio Normalizer additionally bounds the
13211 frame's maximum gain factor by a predetermined (global) maximum
13212 gain factor. This is done in order to avoid excessive gain factors
13213 in "silent" or almost silent frames. By default, the maximum gain
13214 factor is 10.0, For most inputs the default value should be
13215 sufficient and it usually is not recommended to increase this
13216 value. Though, for input with an extremely low overall volume
13217 level, it may be necessary to allow even higher gain factors. Note,
13218 however, that the Dynamic Audio Normalizer does not simply apply a
13219 "hard" threshold (i.e. cut off values above the threshold).
13220 Instead, a "sigmoid" threshold function will be applied. This way,
13221 the gain factors will smoothly approach the threshold value, but
13222 never exceed that value.
13223 targetrms, r
13225 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 -
13226 disabled. By default, the Dynamic Audio Normalizer performs "peak"
13227 normalization. This means that the maximum local gain factor for
13228 each frame is defined (only) by the frame's highest magnitude
13229 sample. This way, the samples can be amplified as much as possible
13230 without exceeding the maximum signal level, i.e. without clipping.
13231 Optionally, however, the Dynamic Audio Normalizer can also take
13232 into account the frame's root mean square, abbreviated RMS. In
13233 electrical engineering, the RMS is commonly used to determine the
13234 power of a time-varying signal. It is therefore considered that the
13235 RMS is a better approximation of the "perceived loudness" than just
13236 looking at the signal's peak magnitude. Consequently, by adjusting
13237 all frames to a constant RMS value, a uniform "perceived loudness"
13238 can be established. If a target RMS value has been specified, a
13239 frame's local gain factor is defined as the factor that would
13240 result in exactly that RMS value. Note, however, that the maximum
13241 local gain factor is still restricted by the frame's highest
13242 magnitude sample, in order to prevent clipping.
13243 coupling, n
13245 Enable channels coupling. By default is enabled. By default, the
13246 Dynamic Audio Normalizer will amplify all channels by the same
13247 amount. This means the same gain factor will be applied to all
13248 channels, i.e. the maximum possible gain factor is determined by
13249 the "loudest" channel. However, in some recordings, it may happen
13250 that the volume of the different channels is uneven, e.g. one
13251 channel may be "quieter" than the other one(s). In this case, this
13252 option can be used to disable the channel coupling. This way, the
13253 gain factor will be determined independently for each channel,
13254 depending only on the individual channel's highest magnitude
13255 sample. This allows for harmonizing the volume of the different
13256 channels.
13257 correctdc, c
13259 Enable DC bias correction. By default is disabled. An audio signal
13260 (in the time domain) is a sequence of sample values. In the
13261 Dynamic Audio Normalizer these sample values are represented in the
13262 -1.0 to 1.0 range, regardless of the original input format.
13263 Normally, the audio signal, or "waveform", should be centered
13264 around the zero point. That means if we calculate the mean value
13265 of all samples in a file, or in a single frame, then the result
13266 should be 0.0 or at least very close to that value. If, however,
13267 there is a significant deviation of the mean value from 0.0, in
13268 either positive or negative direction, this is referred to as a DC
13269 bias or DC offset. Since a DC bias is clearly undesirable, the
13270 Dynamic Audio Normalizer provides optional DC bias correction.
13271 With DC bias correction enabled, the Dynamic Audio Normalizer will
13272 determine the mean value, or "DC correction" offset, of each input
13273 frame and subtract that value from all of the frame's sample values
13274 which ensures those samples are centered around 0.0 again. Also, in
13275 order to avoid "gaps" at the frame boundaries, the DC correction
13276 offset values will be interpolated smoothly between neighbouring
13277 frames.
13278 altboundary, b
13280 Enable alternative boundary mode. By default is disabled. The
13281 Dynamic Audio Normalizer takes into account a certain neighbourhood
13282 around each frame. This includes the preceding frames as well as
13283 the subsequent frames. However, for the "boundary" frames, located
13284 at the very beginning and at the very end of the audio file, not
13285 all neighbouring frames are available. In particular, for the first
13286 few frames in the audio file, the preceding frames are not known.
13287 And, similarly, for the last few frames in the audio file, the
13288 subsequent frames are not known. Thus, the question arises which
13289 gain factors should be assumed for the missing frames in the
13290 "boundary" region. The Dynamic Audio Normalizer implements two
13291 modes to deal with this situation. The default boundary mode
13292 assumes a gain factor of exactly 1.0 for the missing frames,
13293 resulting in a smooth "fade in" and "fade out" at the beginning and
13294 at the end of the input, respectively.
13295 compress, s
13297 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
13298 By default, the Dynamic Audio Normalizer does not apply
13299 "traditional" compression. This means that signal peaks will not be
13300 pruned and thus the full dynamic range will be retained within each
13301 local neighbourhood. However, in some cases it may be desirable to
13302 combine the Dynamic Audio Normalizer's normalization algorithm with
13303 a more "traditional" compression. For this purpose, the Dynamic
13304 Audio Normalizer provides an optional compression (thresholding)
13305 function. If (and only if) the compression feature is enabled, all
13306 input frames will be processed by a soft knee thresholding function
13307 prior to the actual normalization process. Put simply, the
13308 thresholding function is going to prune all samples whose magnitude
13309 exceeds a certain threshold value. However, the Dynamic Audio
13310 Normalizer does not simply apply a fixed threshold value. Instead,
13311 the threshold value will be adjusted for each individual frame. In
13312 general, smaller parameters result in stronger compression, and
13313 vice versa. Values below 3.0 are not recommended, because audible
13314 distortion may appear.
13315 threshold, t
13317 Set the target threshold value. This specifies the lowest
13318 permissible magnitude level for the audio input which will be
13319 normalized. If input frame volume is above this value frame will
13320 be normalized. Otherwise frame may not be normalized at all. The
13321 default value is set to 0, which means all input frames will be
13322 normalized. This option is mostly useful if digital noise is not
13323 wanted to be amplified.
13324 channels, h
13326 Specify which channels to filter, by default all available channels
13327 are filtered.
13328 overlap, o
13330 Specify overlap for frames. If set to 0 (default) no frame
13331 overlapping is done. Using >0 and <1 values will make less
13332 conservative gain adjustments, like when framelen option is set to
13333 smaller value, if framelen option value is compensated for non-zero
13334 overlap then gain adjustments will be smoother across time compared
13335 to zero overlap case.
13336 curve, v
13338 Specify the peak mapping curve expression which is going to be used
13339 when calculating gain applied to frames. The max output frame gain
13340 will still be limited by other options mentioned previously for
13341 this filter.
13342 The expression can contain the following constants:
13344 ch current channel number
13346 sn current sample number
13348 nb_channels
13350 number of channels
13351 t timestamp expressed in seconds
13353 sr sample rate
13355 p current frame peak value
13357 Commands
13359 This filter supports the all above options as commands.
13361 earwax
13363 Make audio easier to listen to on headphones.
13364 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
13366 so that when listened to on headphones the stereo image is moved from
13367 inside your head (standard for headphones) to outside and in front of
13368 the listener (standard for speakers).
13369 Ported from SoX.
13371 equalizer
13373 Apply a two-pole peaking equalisation (EQ) filter. With this filter,
13374 the signal-level at and around a selected frequency can be increased or
13375 decreased, whilst (unlike bandpass and bandreject filters) that at all
13376 other frequencies is unchanged.
13377 In order to produce complex equalisation curves, this filter can be
13379 given several times, each with a different central frequency.
13380 The filter accepts the following options:
13382 frequency, f
13384 Set the filter's central frequency in Hz.
13385 width_type, t
13387 Set method to specify band-width of filter.
13388 h Hz
13390 q Q-Factor
13392 o octave
13394 s slope
13396 k kHz
13398 width, w
13400 Specify the band-width of a filter in width_type units.
13401 gain, g
13403 Set the required gain or attenuation in dB. Beware of clipping
13404 when using a positive gain.
13405 mix, m
13407 How much to use filtered signal in output. Default is 1. Range is
13408 between 0 and 1.
13409 channels, c
13411 Specify which channels to filter, by default all available are
13412 filtered.
13413 normalize, n
13415 Normalize biquad coefficients, by default is disabled. Enabling it
13416 will normalize magnitude response at DC to 0dB.
13417 transform, a
13419 Set transform type of IIR filter.
13420 di
13422 dii
13423 tdi
13424 tdii
13425 latt
13426 svf
13427 zdf
13428 precision, r
13429 Set precison of filtering.
13430 auto
13432 Pick automatic sample format depending on surround filters.
13433 s16 Always use signed 16-bit.
13435 s32 Always use signed 32-bit.
13437 f32 Always use float 32-bit.
13439 f64 Always use float 64-bit.
13441 block_size, b
13443 Set block size used for reverse IIR processing. If this value is
13444 set to high enough value (higher than impulse response length
13445 truncated when reaches near zero values) filtering will become
13446 linear phase otherwise if not big enough it will just produce nasty
13447 artifacts.
13448 Note that filter delay will be exactly this many samples when set
13450 to non-zero value.
13451 Examples
13453 • Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
13455 equalizer=f=1000:t=h:width=200:g=-10
13457 • Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz
13459 with Q 2:
13460 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
13462 Commands
13464 This filter supports the following commands:
13466 frequency, f
13468 Change equalizer frequency. Syntax for the command is :
13469 "frequency"
13470 width_type, t
13472 Change equalizer width_type. Syntax for the command is :
13473 "width_type"
13474 width, w
13476 Change equalizer width. Syntax for the command is : "width"
13477 gain, g
13479 Change equalizer gain. Syntax for the command is : "gain"
13480 mix, m
13482 Change equalizer mix. Syntax for the command is : "mix"
13483 extrastereo
13485 Linearly increases the difference between left and right channels which
13486 adds some sort of "live" effect to playback.
13487 The filter accepts the following options:
13489 m Sets the difference coefficient (default: 2.5). 0.0 means mono
13491 sound (average of both channels), with 1.0 sound will be unchanged,
13492 with -1.0 left and right channels will be swapped.
13493 c Enable clipping. By default is enabled.
13495 Commands
13497 This filter supports the all above options as commands.
13499 firequalizer
13501 Apply FIR Equalization using arbitrary frequency response.
13502 The filter accepts the following option:
13504 gain
13506 Set gain curve equation (in dB). The expression can contain
13507 variables:
13508 f the evaluated frequency
13510 sr sample rate
13512 ch channel number, set to 0 when multichannels evaluation is
13514 disabled
13515 chid
13517 channel id, see libavutil/channel_layout.h, set to the first
13518 channel id when multichannels evaluation is disabled
13519 chs number of channels
13521 chlayout
13523 channel_layout, see libavutil/channel_layout.h
13524 and functions:
13526 gain_interpolate(f)
13528 interpolate gain on frequency f based on gain_entry
13529 cubic_interpolate(f)
13531 same as gain_interpolate, but smoother
13532 This option is also available as command. Default is
13534 gain_interpolate(f).
13535 gain_entry
13537 Set gain entry for gain_interpolate function. The expression can
13538 contain functions:
13539 entry(f, g)
13541 store gain entry at frequency f with value g
13542 This option is also available as command.
13544 delay
13546 Set filter delay in seconds. Higher value means more accurate.
13547 Default is 0.01.
13548 accuracy
13550 Set filter accuracy in Hz. Lower value means more accurate.
13551 Default is 5.
13552 wfunc
13554 Set window function. Acceptable values are:
13555 rectangular
13557 rectangular window, useful when gain curve is already smooth
13558 hann
13560 hann window (default)
13561 hamming
13563 hamming window
13564 blackman
13566 blackman window
13567 nuttall3
13569 3-terms continuous 1st derivative nuttall window
13570 mnuttall3
13572 minimum 3-terms discontinuous nuttall window
13573 nuttall
13575 4-terms continuous 1st derivative nuttall window
13576 bnuttall
13578 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
13579 bharris
13581 blackman-harris window
13582 tukey
13584 tukey window
13585 fixed
13587 If enabled, use fixed number of audio samples. This improves speed
13588 when filtering with large delay. Default is disabled.
13589 multi
13591 Enable multichannels evaluation on gain. Default is disabled.
13592 zero_phase
13594 Enable zero phase mode by subtracting timestamp to compensate
13595 delay. Default is disabled.
13596 scale
13598 Set scale used by gain. Acceptable values are:
13599 linlin
13601 linear frequency, linear gain
13602 linlog
13604 linear frequency, logarithmic (in dB) gain (default)
13605 loglin
13607 logarithmic (in octave scale where 20 Hz is 0) frequency,
13608 linear gain
13609 loglog
13611 logarithmic frequency, logarithmic gain
13612 dumpfile
13614 Set file for dumping, suitable for gnuplot.
13615 dumpscale
13617 Set scale for dumpfile. Acceptable values are same with scale
13618 option. Default is linlog.
13619 fft2
13621 Enable 2-channel convolution using complex FFT. This improves speed
13622 significantly. Default is disabled.
13623 min_phase
13625 Enable minimum phase impulse response. Default is disabled.
13626 Examples
13628 • lowpass at 1000 Hz:
13630 firequalizer=gain='if(lt(f,1000), 0, -INF)'
13632 • lowpass at 1000 Hz with gain_entry:
13634 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
13636 • custom equalization:
13638 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
13640 • higher delay with zero phase to compensate delay:
13642 firequalizer=delay=0.1:fixed=on:zero_phase=on
13644 • lowpass on left channel, highpass on right channel:
13646 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
13648 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
13649 flanger
13651 Apply a flanging effect to the audio.
13652 The filter accepts the following options:
13654 delay
13656 Set base delay in milliseconds. Range from 0 to 30. Default value
13657 is 0.
13658 depth
13660 Set added sweep delay in milliseconds. Range from 0 to 10. Default
13661 value is 2.
13662 regen
13664 Set percentage regeneration (delayed signal feedback). Range from
13665 -95 to 95. Default value is 0.
13666 width
13668 Set percentage of delayed signal mixed with original. Range from 0
13669 to 100. Default value is 71.
13670 speed
13672 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is
13673 0.5.
13674 shape
13676 Set swept wave shape, can be triangular or sinusoidal. Default
13677 value is sinusoidal.
13678 phase
13680 Set swept wave percentage-shift for multi channel. Range from 0 to
13681 100. Default value is 25.
13682 interp
13684 Set delay-line interpolation, linear or quadratic. Default is
13685 linear.
13686 haas
13688 Apply Haas effect to audio.
13689 Note that this makes most sense to apply on mono signals. With this
13691 filter applied to mono signals it give some directionality and
13692 stretches its stereo image.
13693 The filter accepts the following options:
13695 level_in
13697 Set input level. By default is 1, or 0dB
13698 level_out
13700 Set output level. By default is 1, or 0dB.
13701 side_gain
13703 Set gain applied to side part of signal. By default is 1.
13704 middle_source
13706 Set kind of middle source. Can be one of the following:
13707 left
13709 Pick left channel.
13710 right
13712 Pick right channel.
13713 mid Pick middle part signal of stereo image.
13715 side
13717 Pick side part signal of stereo image.
13718 middle_phase
13720 Change middle phase. By default is disabled.
13721 left_delay
13723 Set left channel delay. By default is 2.05 milliseconds.
13724 left_balance
13726 Set left channel balance. By default is -1.
13727 left_gain
13729 Set left channel gain. By default is 1.
13730 left_phase
13732 Change left phase. By default is disabled.
13733 right_delay
13735 Set right channel delay. By defaults is 2.12 milliseconds.
13736 right_balance
13738 Set right channel balance. By default is 1.
13739 right_gain
13741 Set right channel gain. By default is 1.
13742 right_phase
13744 Change right phase. By default is enabled.
13745 hdcd
13747 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM
13748 stream with embedded HDCD codes is expanded into a 20-bit PCM stream.
13749 The filter supports the Peak Extend and Low-level Gain Adjustment
13751 features of HDCD, and detects the Transient Filter flag.
13752 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
13754 When using the filter with wav, note the default encoding for wav is
13756 16-bit, so the resulting 20-bit stream will be truncated back to
13757 16-bit. Use something like -acodec pcm_s24le after the filter to get
13758 24-bit PCM output.
13759 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
13761 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
13762 The filter accepts the following options:
13764 disable_autoconvert
13766 Disable any automatic format conversion or resampling in the filter
13767 graph.
13768 process_stereo
13770 Process the stereo channels together. If target_gain does not match
13771 between channels, consider it invalid and use the last valid
13772 target_gain.
13773 cdt_ms
13775 Set the code detect timer period in ms.
13776 force_pe
13778 Always extend peaks above -3dBFS even if PE isn't signaled.
13779 analyze_mode
13781 Replace audio with a solid tone and adjust the amplitude to signal
13782 some specific aspect of the decoding process. The output file can
13783 be loaded in an audio editor alongside the original to aid
13784 analysis.
13785 "analyze_mode=pe:force_pe=true" can be used to see all samples
13787 above the PE level.
13788 Modes are:
13790 0, off
13792 Disabled
13793 1, lle
13795 Gain adjustment level at each sample
13796 2, pe
13798 Samples where peak extend occurs
13799 3, cdt
13801 Samples where the code detect timer is active
13802 4, tgm
13804 Samples where the target gain does not match between channels
13805 headphone
13807 Apply head-related transfer functions (HRTFs) to create virtual
13808 loudspeakers around the user for binaural listening via headphones.
13809 The HRIRs are provided via additional streams, for each channel one
13810 stereo input stream is needed.
13811 The filter accepts the following options:
13813 map Set mapping of input streams for convolution. The argument is a
13815 '|'-separated list of channel names in order as they are given as
13816 additional stream inputs for filter. This also specify number of
13817 input streams. Number of input streams must be not less than number
13818 of channels in first stream plus one.
13819 gain
13821 Set gain applied to audio. Value is in dB. Default is 0.
13822 type
13824 Set processing type. Can be time or freq. time is processing audio
13825 in time domain which is slow. freq is processing audio in
13826 frequency domain which is fast. Default is freq.
13827 lfe Set custom gain for LFE channels. Value is in dB. Default is 0.
13829 size
13831 Set size of frame in number of samples which will be processed at
13832 once. Default value is 1024. Allowed range is from 1024 to 96000.
13833 hrir
13835 Set format of hrir stream. Default value is stereo. Alternative
13836 value is multich. If value is set to stereo, number of additional
13837 streams should be greater or equal to number of input channels in
13838 first input stream. Also each additional stream should have stereo
13839 number of channels. If value is set to multich, number of
13840 additional streams should be exactly one. Also number of input
13841 channels of additional stream should be equal or greater than twice
13842 number of channels of first input stream.
13843 Examples
13845 • Full example using wav files as coefficients with amovie filters
13847 for 7.1 downmix, each amovie filter use stereo file with IR
13848 coefficients as input. The files give coefficients for each
13849 position of virtual loudspeaker:
13850 ffmpeg -i input.wav
13852 -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"
13853 output.wav
13854 • Full example using wav files as coefficients with amovie filters
13856 for 7.1 downmix, but now in multich hrir format.
13857 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"
13859 output.wav
13860 highpass
13862 Apply a high-pass filter with 3dB point frequency. The filter can be
13863 either single-pole, or double-pole (the default). The filter roll off
13864 at 6dB per pole per octave (20dB per pole per decade).
13865 The filter accepts the following options:
13867 frequency, f
13869 Set frequency in Hz. Default is 3000.
13870 poles, p
13872 Set number of poles. Default is 2.
13873 width_type, t
13875 Set method to specify band-width of filter.
13876 h Hz
13878 q Q-Factor
13880 o octave
13882 s slope
13884 k kHz
13886 width, w
13888 Specify the band-width of a filter in width_type units. Applies
13889 only to double-pole filter. The default is 0.707q and gives a
13890 Butterworth response.
13891 mix, m
13893 How much to use filtered signal in output. Default is 1. Range is
13894 between 0 and 1.
13895 channels, c
13897 Specify which channels to filter, by default all available are
13898 filtered.
13899 normalize, n
13901 Normalize biquad coefficients, by default is disabled. Enabling it
13902 will normalize magnitude response at DC to 0dB.
13903 transform, a
13905 Set transform type of IIR filter.
13906 di
13908 dii
13909 tdi
13910 tdii
13911 latt
13912 svf
13913 zdf
13914 precision, r
13915 Set precison of filtering.
13916 auto
13918 Pick automatic sample format depending on surround filters.
13919 s16 Always use signed 16-bit.
13921 s32 Always use signed 32-bit.
13923 f32 Always use float 32-bit.
13925 f64 Always use float 64-bit.
13927 block_size, b
13929 Set block size used for reverse IIR processing. If this value is
13930 set to high enough value (higher than impulse response length
13931 truncated when reaches near zero values) filtering will become
13932 linear phase otherwise if not big enough it will just produce nasty
13933 artifacts.
13934 Note that filter delay will be exactly this many samples when set
13936 to non-zero value.
13937 Commands
13939 This filter supports the following commands:
13941 frequency, f
13943 Change highpass frequency. Syntax for the command is : "frequency"
13944 width_type, t
13946 Change highpass width_type. Syntax for the command is :
13947 "width_type"
13948 width, w
13950 Change highpass width. Syntax for the command is : "width"
13951 mix, m
13953 Change highpass mix. Syntax for the command is : "mix"
13954 join
13956 Join multiple input streams into one multi-channel stream.
13957 It accepts the following parameters:
13959 inputs
13961 The number of input streams. It defaults to 2.
13962 channel_layout
13964 The desired output channel layout. It defaults to stereo.
13965 map Map channels from inputs to output. The argument is a '|'-separated
13967 list of mappings, each in the "input_idx.in_channel-out_channel"
13968 form. input_idx is the 0-based index of the input stream.
13969 in_channel can be either the name of the input channel (e.g. FL for
13970 front left) or its index in the specified input stream. out_channel
13971 is the name of the output channel.
13972 The filter will attempt to guess the mappings when they are not
13974 specified explicitly. It does so by first trying to find an unused
13975 matching input channel and if that fails it picks the first unused
13976 input channel.
13977 Join 3 inputs (with properly set channel layouts):
13979 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
13981 Build a 5.1 output from 6 single-channel streams:
13983 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
13985 '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'
13986 out
13987 ladspa
13989 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
13990 To enable compilation of this filter you need to configure FFmpeg with
13992 "--enable-ladspa".
13993 file, f
13995 Specifies the name of LADSPA plugin library to load. If the
13996 environment variable LADSPA_PATH is defined, the LADSPA plugin is
13997 searched in each one of the directories specified by the colon
13998 separated list in LADSPA_PATH, otherwise in the standard LADSPA
13999 paths, which are in this order: HOME/.ladspa/lib/,
14000 /usr/local/lib/ladspa/, /usr/lib/ladspa/.
14001 plugin, p
14003 Specifies the plugin within the library. Some libraries contain
14004 only one plugin, but others contain many of them. If this is not
14005 set filter will list all available plugins within the specified
14006 library.
14007 controls, c
14009 Set the '|' separated list of controls which are zero or more
14010 floating point values that determine the behavior of the loaded
14011 plugin (for example delay, threshold or gain). Controls need to be
14012 defined using the following syntax:
14013 c0=value0|c1=value1|c2=value2|..., where valuei is the value set on
14014 the i-th control. Alternatively they can be also defined using the
14015 following syntax: value0|value1|value2|..., where valuei is the
14016 value set on the i-th control. If controls is set to "help", all
14017 available controls and their valid ranges are printed.
14018 sample_rate, s
14020 Specify the sample rate, default to 44100. Only used if plugin have
14021 zero inputs.
14022 nb_samples, n
14024 Set the number of samples per channel per each output frame,
14025 default is 1024. Only used if plugin have zero inputs.
14026 duration, d
14028 Set the minimum duration of the sourced audio. See the Time
14029 duration section in the ffmpeg-utils(1) manual for the accepted
14030 syntax. Note that the resulting duration may be greater than the
14031 specified duration, as the generated audio is always cut at the end
14032 of a complete frame. If not specified, or the expressed duration
14033 is negative, the audio is supposed to be generated forever. Only
14034 used if plugin have zero inputs.
14035 latency, l
14037 Enable latency compensation, by default is disabled. Only used if
14038 plugin have inputs.
14039 Examples
14041 • List all available plugins within amp (LADSPA example plugin)
14043 library:
14044 ladspa=file=amp
14046 • List all available controls and their valid ranges for "vcf_notch"
14048 plugin from "VCF" library:
14049 ladspa=f=vcf:p=vcf_notch:c=help
14051 • Simulate low quality audio equipment using "Computer Music Toolkit"
14053 (CMT) plugin library:
14054 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
14056 • Add reverberation to the audio using TAP-plugins (Tom's Audio
14058 Processing plugins):
14059 ladspa=file=tap_reverb:tap_reverb
14061 • Generate white noise, with 0.2 amplitude:
14063 ladspa=file=cmt:noise_source_white:c=c0=.2
14065 • Generate 20 bpm clicks using plugin "C* Click - Metronome" from the
14067 "C* Audio Plugin Suite" (CAPS) library:
14068 ladspa=file=caps:Click:c=c1=20'
14070 • Apply "C* Eq10X2 - Stereo 10-band equaliser" effect:
14072 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
14074 • Increase volume by 20dB using fast lookahead limiter from Steve
14076 Harris "SWH Plugins" collection:
14077 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
14079 • Attenuate low frequencies using Multiband EQ from Steve Harris "SWH
14081 Plugins" collection:
14082 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
14084 • Reduce stereo image using "Narrower" from the "C* Audio Plugin
14086 Suite" (CAPS) library:
14087 ladspa=caps:Narrower
14089 • Another white noise, now using "C* Audio Plugin Suite" (CAPS)
14091 library:
14092 ladspa=caps:White:.2
14094 • Some fractal noise, using "C* Audio Plugin Suite" (CAPS) library:
14096 ladspa=caps:Fractal:c=c1=1
14098 • Dynamic volume normalization using "VLevel" plugin:
14100 ladspa=vlevel-ladspa:vlevel_mono
14102 Commands
14104 This filter supports the following commands:
14106 cN Modify the N-th control value.
14108 If the specified value is not valid, it is ignored and prior one is
14110 kept.
14111 loudnorm
14113 EBU R128 loudness normalization. Includes both dynamic and linear
14114 normalization modes. Support for both single pass (livestreams, files)
14115 and double pass (files) modes. This algorithm can target IL, LRA, and
14116 maximum true peak. In dynamic mode, to accurately detect true peaks,
14117 the audio stream will be upsampled to 192 kHz. Use the "-ar" option or
14118 "aresample" filter to explicitly set an output sample rate.
14119 The filter accepts the following options:
14121 I, i
14123 Set integrated loudness target. Range is -70.0 - -5.0. Default
14124 value is -24.0.
14125 LRA, lra
14127 Set loudness range target. Range is 1.0 - 50.0. Default value is
14128 7.0.
14129 TP, tp
14131 Set maximum true peak. Range is -9.0 - +0.0. Default value is
14132 -2.0.
14133 measured_I, measured_i
14135 Measured IL of input file. Range is -99.0 - +0.0.
14136 measured_LRA, measured_lra
14138 Measured LRA of input file. Range is 0.0 - 99.0.
14139 measured_TP, measured_tp
14141 Measured true peak of input file. Range is -99.0 - +99.0.
14142 measured_thresh
14144 Measured threshold of input file. Range is -99.0 - +0.0.
14145 offset
14147 Set offset gain. Gain is applied before the true-peak limiter.
14148 Range is -99.0 - +99.0. Default is +0.0.
14149 linear
14151 Normalize by linearly scaling the source audio. "measured_I",
14152 "measured_LRA", "measured_TP", and "measured_thresh" must all be
14153 specified. Target LRA shouldn't be lower than source LRA and the
14154 change in integrated loudness shouldn't result in a true peak which
14155 exceeds the target TP. If any of these conditions aren't met,
14156 normalization mode will revert to dynamic. Options are "true" or
14157 "false". Default is "true".
14158 dual_mono
14160 Treat mono input files as "dual-mono". If a mono file is intended
14161 for playback on a stereo system, its EBU R128 measurement will be
14162 perceptually incorrect. If set to "true", this option will
14163 compensate for this effect. Multi-channel input files are not
14164 affected by this option. Options are true or false. Default is
14165 false.
14166 print_format
14168 Set print format for stats. Options are summary, json, or none.
14169 Default value is none.
14170 lowpass
14172 Apply a low-pass filter with 3dB point frequency. The filter can be
14173 either single-pole or double-pole (the default). The filter roll off
14174 at 6dB per pole per octave (20dB per pole per decade).
14175 The filter accepts the following options:
14177 frequency, f
14179 Set frequency in Hz. Default is 500.
14180 poles, p
14182 Set number of poles. Default is 2.
14183 width_type, t
14185 Set method to specify band-width of filter.
14186 h Hz
14188 q Q-Factor
14190 o octave
14192 s slope
14194 k kHz
14196 width, w
14198 Specify the band-width of a filter in width_type units. Applies
14199 only to double-pole filter. The default is 0.707q and gives a
14200 Butterworth response.
14201 mix, m
14203 How much to use filtered signal in output. Default is 1. Range is
14204 between 0 and 1.
14205 channels, c
14207 Specify which channels to filter, by default all available are
14208 filtered.
14209 normalize, n
14211 Normalize biquad coefficients, by default is disabled. Enabling it
14212 will normalize magnitude response at DC to 0dB.
14213 transform, a
14215 Set transform type of IIR filter.
14216 di
14218 dii
14219 tdi
14220 tdii
14221 latt
14222 svf
14223 zdf
14224 precision, r
14225 Set precison of filtering.
14226 auto
14228 Pick automatic sample format depending on surround filters.
14229 s16 Always use signed 16-bit.
14231 s32 Always use signed 32-bit.
14233 f32 Always use float 32-bit.
14235 f64 Always use float 64-bit.
14237 block_size, b
14239 Set block size used for reverse IIR processing. If this value is
14240 set to high enough value (higher than impulse response length
14241 truncated when reaches near zero values) filtering will become
14242 linear phase otherwise if not big enough it will just produce nasty
14243 artifacts.
14244 Note that filter delay will be exactly this many samples when set
14246 to non-zero value.
14247 Examples
14249 • Lowpass only LFE channel, it LFE is not present it does nothing:
14251 lowpass=c=LFE
14253 Commands
14255 This filter supports the following commands:
14257 frequency, f
14259 Change lowpass frequency. Syntax for the command is : "frequency"
14260 width_type, t
14262 Change lowpass width_type. Syntax for the command is :
14263 "width_type"
14264 width, w
14266 Change lowpass width. Syntax for the command is : "width"
14267 mix, m
14269 Change lowpass mix. Syntax for the command is : "mix"
14270 lv2
14272 Load a LV2 (LADSPA Version 2) plugin.
14273 To enable compilation of this filter you need to configure FFmpeg with
14275 "--enable-lv2".
14276 plugin, p
14278 Specifies the plugin URI. You may need to escape ':'.
14279 controls, c
14281 Set the '|' separated list of controls which are zero or more
14282 floating point values that determine the behavior of the loaded
14283 plugin (for example delay, threshold or gain). If controls is set
14284 to "help", all available controls and their valid ranges are
14285 printed.
14286 sample_rate, s
14288 Specify the sample rate, default to 44100. Only used if plugin have
14289 zero inputs.
14290 nb_samples, n
14292 Set the number of samples per channel per each output frame,
14293 default is 1024. Only used if plugin have zero inputs.
14294 duration, d
14296 Set the minimum duration of the sourced audio. See the Time
14297 duration section in the ffmpeg-utils(1) manual for the accepted
14298 syntax. Note that the resulting duration may be greater than the
14299 specified duration, as the generated audio is always cut at the end
14300 of a complete frame. If not specified, or the expressed duration
14301 is negative, the audio is supposed to be generated forever. Only
14302 used if plugin have zero inputs.
14303 Examples
14305 • Apply bass enhancer plugin from Calf:
14307 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
14309 • Apply vinyl plugin from Calf:
14311 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
14313 • Apply bit crusher plugin from ArtyFX:
14315 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
14317 Commands
14319 This filter supports all options that are exported by plugin as
14321 commands.
14322 mcompand
14324 Multiband Compress or expand the audio's dynamic range.
14325 The input audio is divided into bands using 4th order Linkwitz-Riley
14327 IIRs. This is akin to the crossover of a loudspeaker, and results in
14328 flat frequency response when absent compander action.
14329 It accepts the following parameters:
14331 args
14333 This option syntax is: attack,decay,[attack,decay..] soft-knee
14334 points crossover_frequency [delay [initial_volume [gain]]] |
14335 attack,decay ... For explanation of each item refer to compand
14336 filter documentation.
14337 pan
14339 Mix channels with specific gain levels. The filter accepts the output
14340 channel layout followed by a set of channels definitions.
14341 This filter is also designed to efficiently remap the channels of an
14343 audio stream.
14344 The filter accepts parameters of the form: "l|outdef|outdef|..."
14346 l output channel layout or number of channels
14348 outdef
14350 output channel specification, of the form:
14351 "out_name=[gain*]in_name[(+-)[gain*]in_name...]"
14352 out_name
14354 output channel to define, either a channel name (FL, FR, etc.) or a
14355 channel number (c0, c1, etc.)
14356 gain
14358 multiplicative coefficient for the channel, 1 leaving the volume
14359 unchanged
14360 in_name
14362 input channel to use, see out_name for details; it is not possible
14363 to mix named and numbered input channels
14364 If the `=' in a channel specification is replaced by `<', then the
14366 gains for that specification will be renormalized so that the total is
14367 1, thus avoiding clipping noise.
14368 Mixing examples
14370 For example, if you want to down-mix from stereo to mono, but with a
14372 bigger factor for the left channel:
14373 pan=1c|c0=0.9*c0+0.1*c1
14375 A customized down-mix to stereo that works automatically for 3-, 4-, 5-
14377 and 7-channels surround:
14378 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
14380 Note that ffmpeg integrates a default down-mix (and up-mix) system that
14382 should be preferred (see "-ac" option) unless you have very specific
14383 needs.
14384 Remapping examples
14386 The channel remapping will be effective if, and only if:
14388 *<gain coefficients are zeroes or ones,>
14390 *<only one input per channel output,>
14391 If all these conditions are satisfied, the filter will notify the user
14393 ("Pure channel mapping detected"), and use an optimized and lossless
14394 method to do the remapping.
14395 For example, if you have a 5.1 source and want a stereo audio stream by
14397 dropping the extra channels:
14398 pan="stereo| c0=FL | c1=FR"
14400 Given the same source, you can also switch front left and front right
14402 channels and keep the input channel layout:
14403 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
14405 If the input is a stereo audio stream, you can mute the front left
14407 channel (and still keep the stereo channel layout) with:
14408 pan="stereo|c1=c1"
14410 Still with a stereo audio stream input, you can copy the right channel
14412 in both front left and right:
14413 pan="stereo| c0=FR | c1=FR"
14415 replaygain
14417 ReplayGain scanner filter. This filter takes an audio stream as an
14418 input and outputs it unchanged. At end of filtering it displays
14419 "track_gain" and "track_peak".
14420 resample
14422 Convert the audio sample format, sample rate and channel layout. It is
14423 not meant to be used directly.
14424 rubberband
14426 Apply time-stretching and pitch-shifting with librubberband.
14427 To enable compilation of this filter, you need to configure FFmpeg with
14429 "--enable-librubberband".
14430 The filter accepts the following options:
14432 tempo
14434 Set tempo scale factor.
14435 pitch
14437 Set pitch scale factor.
14438 transients
14440 Set transients detector. Possible values are:
14441 crisp
14443 mixed
14444 smooth
14445 detector
14446 Set detector. Possible values are:
14447 compound
14449 percussive
14450 soft
14451 phase
14452 Set phase. Possible values are:
14453 laminar
14455 independent
14456 window
14457 Set processing window size. Possible values are:
14458 standard
14460 short
14461 long
14462 smoothing
14463 Set smoothing. Possible values are:
14464 off
14466 on
14467 formant
14468 Enable formant preservation when shift pitching. Possible values
14469 are:
14470 shifted
14472 preserved
14473 pitchq
14474 Set pitch quality. Possible values are:
14475 quality
14477 speed
14478 consistency
14479 channels
14480 Set channels. Possible values are:
14481 apart
14483 together
14484 Commands
14486 This filter supports the following commands:
14488 tempo
14490 Change filter tempo scale factor. Syntax for the command is :
14491 "tempo"
14492 pitch
14494 Change filter pitch scale factor. Syntax for the command is :
14495 "pitch"
14496 sidechaincompress
14498 This filter acts like normal compressor but has the ability to compress
14499 detected signal using second input signal. It needs two input streams
14500 and returns one output stream. First input stream will be processed
14501 depending on second stream signal. The filtered signal then can be
14502 filtered with other filters in later stages of processing. See pan and
14503 amerge filter.
14504 The filter accepts the following options:
14506 level_in
14508 Set input gain. Default is 1. Range is between 0.015625 and 64.
14509 mode
14511 Set mode of compressor operation. Can be "upward" or "downward".
14512 Default is "downward".
14513 threshold
14515 If a signal of second stream raises above this level it will affect
14516 the gain reduction of first stream. By default is 0.125. Range is
14517 between 0.00097563 and 1.
14518 ratio
14520 Set a ratio about which the signal is reduced. 1:2 means that if
14521 the level raised 4dB above the threshold, it will be only 2dB above
14522 after the reduction. Default is 2. Range is between 1 and 20.
14523 attack
14525 Amount of milliseconds the signal has to rise above the threshold
14526 before gain reduction starts. Default is 20. Range is between 0.01
14527 and 2000.
14528 release
14530 Amount of milliseconds the signal has to fall below the threshold
14531 before reduction is decreased again. Default is 250. Range is
14532 between 0.01 and 9000.
14533 makeup
14535 Set the amount by how much signal will be amplified after
14536 processing. Default is 1. Range is from 1 to 64.
14537 knee
14539 Curve the sharp knee around the threshold to enter gain reduction
14540 more softly. Default is 2.82843. Range is between 1 and 8.
14541 link
14543 Choose if the "average" level between all channels of side-chain
14544 stream or the louder("maximum") channel of side-chain stream
14545 affects the reduction. Default is "average".
14546 detection
14548 Should the exact signal be taken in case of "peak" or an RMS one in
14549 case of "rms". Default is "rms" which is mainly smoother.
14550 level_sc
14552 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
14553 mix How much to use compressed signal in output. Default is 1. Range
14555 is between 0 and 1.
14556 Commands
14558 This filter supports the all above options as commands.
14560 Examples
14562 • Full ffmpeg example taking 2 audio inputs, 1st input to be
14564 compressed depending on the signal of 2nd input and later
14565 compressed signal to be merged with 2nd input:
14566 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
14568 sidechaingate
14570 A sidechain gate acts like a normal (wideband) gate but has the ability
14571 to filter the detected signal before sending it to the gain reduction
14572 stage. Normally a gate uses the full range signal to detect a level
14573 above the threshold. For example: If you cut all lower frequencies
14574 from your sidechain signal the gate will decrease the volume of your
14575 track only if not enough highs appear. With this technique you are able
14576 to reduce the resonation of a natural drum or remove "rumbling" of
14577 muted strokes from a heavily distorted guitar. It needs two input
14578 streams and returns one output stream. First input stream will be
14579 processed depending on second stream signal.
14580 The filter accepts the following options:
14582 level_in
14584 Set input level before filtering. Default is 1. Allowed range is
14585 from 0.015625 to 64.
14586 mode
14588 Set the mode of operation. Can be "upward" or "downward". Default
14589 is "downward". If set to "upward" mode, higher parts of signal will
14590 be amplified, expanding dynamic range in upward direction.
14591 Otherwise, in case of "downward" lower parts of signal will be
14592 reduced.
14593 range
14595 Set the level of gain reduction when the signal is below the
14596 threshold. Default is 0.06125. Allowed range is from 0 to 1.
14597 Setting this to 0 disables reduction and then filter behaves like
14598 expander.
14599 threshold
14601 If a signal rises above this level the gain reduction is released.
14602 Default is 0.125. Allowed range is from 0 to 1.
14603 ratio
14605 Set a ratio about which the signal is reduced. Default is 2.
14606 Allowed range is from 1 to 9000.
14607 attack
14609 Amount of milliseconds the signal has to rise above the threshold
14610 before gain reduction stops. Default is 20 milliseconds. Allowed
14611 range is from 0.01 to 9000.
14612 release
14614 Amount of milliseconds the signal has to fall below the threshold
14615 before the reduction is increased again. Default is 250
14616 milliseconds. Allowed range is from 0.01 to 9000.
14617 makeup
14619 Set amount of amplification of signal after processing. Default is
14620 1. Allowed range is from 1 to 64.
14621 knee
14623 Curve the sharp knee around the threshold to enter gain reduction
14624 more softly. Default is 2.828427125. Allowed range is from 1 to 8.
14625 detection
14627 Choose if exact signal should be taken for detection or an RMS like
14628 one. Default is rms. Can be peak or rms.
14629 link
14631 Choose if the average level between all channels or the louder
14632 channel affects the reduction. Default is average. Can be average
14633 or maximum.
14634 level_sc
14636 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
14637 Commands
14639 This filter supports the all above options as commands.
14641 silencedetect
14643 Detect silence in an audio stream.
14644 This filter logs a message when it detects that the input audio volume
14646 is less or equal to a noise tolerance value for a duration greater or
14647 equal to the minimum detected noise duration.
14648 The printed times and duration are expressed in seconds. The
14650 "lavfi.silence_start" or "lavfi.silence_start.X" metadata key is set on
14651 the first frame whose timestamp equals or exceeds the detection
14652 duration and it contains the timestamp of the first frame of the
14653 silence.
14654 The "lavfi.silence_duration" or "lavfi.silence_duration.X" and
14656 "lavfi.silence_end" or "lavfi.silence_end.X" metadata keys are set on
14657 the first frame after the silence. If mono is enabled, and each channel
14658 is evaluated separately, the ".X" suffixed keys are used, and "X"
14659 corresponds to the channel number.
14660 The filter accepts the following options:
14662 noise, n
14664 Set noise tolerance. Can be specified in dB (in case "dB" is
14665 appended to the specified value) or amplitude ratio. Default is
14666 -60dB, or 0.001.
14667 duration, d
14669 Set silence duration until notification (default is 2 seconds). See
14670 the Time duration section in the ffmpeg-utils(1) manual for the
14671 accepted syntax.
14672 mono, m
14674 Process each channel separately, instead of combined. By default is
14675 disabled.
14676 Examples
14678 • Detect 5 seconds of silence with -50dB noise tolerance:
14680 silencedetect=n=-50dB:d=5
14682 • Complete example with ffmpeg to detect silence with 0.0001 noise
14684 tolerance in silence.mp3:
14685 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
14687 silenceremove
14689 Remove silence from the beginning, middle or end of the audio.
14690 The filter accepts the following options:
14692 start_periods
14694 This value is used to indicate if audio should be trimmed at
14695 beginning of the audio. A value of zero indicates no silence should
14696 be trimmed from the beginning. When specifying a non-zero value, it
14697 trims audio up until it finds non-silence. Normally, when trimming
14698 silence from beginning of audio the start_periods will be 1 but it
14699 can be increased to higher values to trim all audio up to specific
14700 count of non-silence periods. Default value is 0.
14701 start_duration
14703 Specify the amount of time that non-silence must be detected before
14704 it stops trimming audio. By increasing the duration, bursts of
14705 noises can be treated as silence and trimmed off. Default value is
14706 0.
14707 start_threshold
14709 This indicates what sample value should be treated as silence. For
14710 digital audio, a value of 0 may be fine but for audio recorded from
14711 analog, you may wish to increase the value to account for
14712 background noise. Can be specified in dB (in case "dB" is appended
14713 to the specified value) or amplitude ratio. Default value is 0.
14714 start_silence
14716 Specify max duration of silence at beginning that will be kept
14717 after trimming. Default is 0, which is equal to trimming all
14718 samples detected as silence.
14719 start_mode
14721 Specify mode of detection of silence end in start of multi-channel
14722 audio. Can be any or all. Default is any. With any, any sample
14723 that is detected as non-silence will cause stopped trimming of
14724 silence. With all, only if all channels are detected as non-
14725 silence will cause stopped trimming of silence.
14726 stop_periods
14728 Set the count for trimming silence from the end of audio. To
14729 remove silence from the middle of a file, specify a stop_periods
14730 that is negative. This value is then treated as a positive value
14731 and is used to indicate the effect should restart processing as
14732 specified by start_periods, making it suitable for removing periods
14733 of silence in the middle of the audio. Default value is 0.
14734 stop_duration
14736 Specify a duration of silence that must exist before audio is not
14737 copied any more. By specifying a higher duration, silence that is
14738 wanted can be left in the audio. Default value is 0.
14739 stop_threshold
14741 This is the same as start_threshold but for trimming silence from
14742 the end of audio. Can be specified in dB (in case "dB" is appended
14743 to the specified value) or amplitude ratio. Default value is 0.
14744 stop_silence
14746 Specify max duration of silence at end that will be kept after
14747 trimming. Default is 0, which is equal to trimming all samples
14748 detected as silence.
14749 stop_mode
14751 Specify mode of detection of silence start in end of multi-channel
14752 audio. Can be any or all. Default is any. With any, any sample
14753 that is detected as non-silence will cause stopped trimming of
14754 silence. With all, only if all channels are detected as non-
14755 silence will cause stopped trimming of silence.
14756 detection
14758 Set how is silence detected. Can be "rms" or "peak". Second is
14759 faster and works better with digital silence which is exactly 0.
14760 Default value is "rms".
14761 window
14763 Set duration in number of seconds used to calculate size of window
14764 in number of samples for detecting silence. Default value is 0.02.
14765 Allowed range is from 0 to 10.
14766 Examples
14768 • The following example shows how this filter can be used to start a
14770 recording that does not contain the delay at the start which
14771 usually occurs between pressing the record button and the start of
14772 the performance:
14773 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
14775 • Trim all silence encountered from beginning to end where there is
14777 more than 1 second of silence in audio:
14778 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
14780 • Trim all digital silence samples, using peak detection, from
14782 beginning to end where there is more than 0 samples of digital
14783 silence in audio and digital silence is detected in all channels at
14784 same positions in stream:
14785 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
14787 sofalizer
14789 SOFAlizer uses head-related transfer functions (HRTFs) to create
14790 virtual loudspeakers around the user for binaural listening via
14791 headphones (audio formats up to 9 channels supported). The HRTFs are
14792 stored in SOFA files (see <http://www.sofacoustics.org/> for a
14793 database). SOFAlizer is developed at the Acoustics Research Institute
14794 (ARI) of the Austrian Academy of Sciences.
14795 To enable compilation of this filter you need to configure FFmpeg with
14797 "--enable-libmysofa".
14798 The filter accepts the following options:
14800 sofa
14802 Set the SOFA file used for rendering.
14803 gain
14805 Set gain applied to audio. Value is in dB. Default is 0.
14806 rotation
14808 Set rotation of virtual loudspeakers in deg. Default is 0.
14809 elevation
14811 Set elevation of virtual speakers in deg. Default is 0.
14812 radius
14814 Set distance in meters between loudspeakers and the listener with
14815 near-field HRTFs. Default is 1.
14816 type
14818 Set processing type. Can be time or freq. time is processing audio
14819 in time domain which is slow. freq is processing audio in
14820 frequency domain which is fast. Default is freq.
14821 speakers
14823 Set custom positions of virtual loudspeakers. Syntax for this
14824 option is: <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...]. Each
14825 virtual loudspeaker is described with short channel name following
14826 with azimuth and elevation in degrees. Each virtual loudspeaker
14827 description is separated by '|'. For example to override front
14828 left and front right channel positions use: 'speakers=FL 45 15|FR
14829 345 15'. Descriptions with unrecognised channel names are ignored.
14830 lfegain
14832 Set custom gain for LFE channels. Value is in dB. Default is 0.
14833 framesize
14835 Set custom frame size in number of samples. Default is 1024.
14836 Allowed range is from 1024 to 96000. Only used if option type is
14837 set to freq.
14838 normalize
14840 Should all IRs be normalized upon importing SOFA file. By default
14841 is enabled.
14842 interpolate
14844 Should nearest IRs be interpolated with neighbor IRs if exact
14845 position does not match. By default is disabled.
14846 minphase
14848 Minphase all IRs upon loading of SOFA file. By default is disabled.
14849 anglestep
14851 Set neighbor search angle step. Only used if option interpolate is
14852 enabled.
14853 radstep
14855 Set neighbor search radius step. Only used if option interpolate is
14856 enabled.
14857 Examples
14859 • Using ClubFritz6 sofa file:
14861 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
14863 • Using ClubFritz12 sofa file and bigger radius with small rotation:
14865 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
14867 • Similar as above but with custom speaker positions for front left,
14869 front right, back left and back right and also with custom gain:
14870 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
14872 speechnorm
14874 Speech Normalizer.
14875 This filter expands or compresses each half-cycle of audio samples
14877 (local set of samples all above or all below zero and between two
14878 nearest zero crossings) depending on threshold value, so audio reaches
14879 target peak value under conditions controlled by below options.
14880 The filter accepts the following options:
14882 peak, p
14884 Set the expansion target peak value. This specifies the highest
14885 allowed absolute amplitude level for the normalized audio input.
14886 Default value is 0.95. Allowed range is from 0.0 to 1.0.
14887 expansion, e
14889 Set the maximum expansion factor. Allowed range is from 1.0 to
14890 50.0. Default value is 2.0. This option controls maximum local
14891 half-cycle of samples expansion. The maximum expansion would be
14892 such that local peak value reaches target peak value but never to
14893 surpass it and that ratio between new and previous peak value does
14894 not surpass this option value.
14895 compression, c
14897 Set the maximum compression factor. Allowed range is from 1.0 to
14898 50.0. Default value is 2.0. This option controls maximum local
14899 half-cycle of samples compression. This option is used only if
14900 threshold option is set to value greater than 0.0, then in such
14901 cases when local peak is lower or same as value set by threshold
14902 all samples belonging to that peak's half-cycle will be compressed
14903 by current compression factor.
14904 threshold, t
14906 Set the threshold value. Default value is 0.0. Allowed range is
14907 from 0.0 to 1.0. This option specifies which half-cycles of
14908 samples will be compressed and which will be expanded. Any half-
14909 cycle samples with their local peak value below or same as this
14910 option value will be compressed by current compression factor,
14911 otherwise, if greater than threshold value they will be expanded
14912 with expansion factor so that it could reach peak target value but
14913 never surpass it.
14914 raise, r
14916 Set the expansion raising amount per each half-cycle of samples.
14917 Default value is 0.001. Allowed range is from 0.0 to 1.0. This
14918 controls how fast expansion factor is raised per each new half-
14919 cycle until it reaches expansion value. Setting this options too
14920 high may lead to distortions.
14921 fall, f
14923 Set the compression raising amount per each half-cycle of samples.
14924 Default value is 0.001. Allowed range is from 0.0 to 1.0. This
14925 controls how fast compression factor is raised per each new half-
14926 cycle until it reaches compression value.
14927 channels, h
14929 Specify which channels to filter, by default all available channels
14930 are filtered.
14931 invert, i
14933 Enable inverted filtering, by default is disabled. This inverts
14934 interpretation of threshold option. When enabled any half-cycle of
14935 samples with their local peak value below or same as threshold
14936 option will be expanded otherwise it will be compressed.
14937 link, l
14939 Link channels when calculating gain applied to each filtered
14940 channel sample, by default is disabled. When disabled each
14941 filtered channel gain calculation is independent, otherwise when
14942 this option is enabled the minimum of all possible gains for each
14943 filtered channel is used.
14944 rms, m
14946 Set the expansion target RMS value. This specifies the highest
14947 allowed RMS level for the normalized audio input. Default value is
14948 0.0, thus disabled. Allowed range is from 0.0 to 1.0.
14949 Commands
14951 This filter supports the all above options as commands.
14953 Examples
14955 • Weak and slow amplification:
14957 speechnorm=e=3:r=0.00001:l=1
14959 • Moderate and slow amplification:
14961 speechnorm=e=6.25:r=0.00001:l=1
14963 • Strong and fast amplification:
14965 speechnorm=e=12.5:r=0.0001:l=1
14967 • Very strong and fast amplification:
14969 speechnorm=e=25:r=0.0001:l=1
14971 • Extreme and fast amplification:
14973 speechnorm=e=50:r=0.0001:l=1
14975 stereotools
14977 This filter has some handy utilities to manage stereo signals, for
14978 converting M/S stereo recordings to L/R signal while having control
14979 over the parameters or spreading the stereo image of master track.
14980 The filter accepts the following options:
14982 level_in
14984 Set input level before filtering for both channels. Defaults is 1.
14985 Allowed range is from 0.015625 to 64.
14986 level_out
14988 Set output level after filtering for both channels. Defaults is 1.
14989 Allowed range is from 0.015625 to 64.
14990 balance_in
14992 Set input balance between both channels. Default is 0. Allowed
14993 range is from -1 to 1.
14994 balance_out
14996 Set output balance between both channels. Default is 0. Allowed
14997 range is from -1 to 1.
14998 softclip
15000 Enable softclipping. Results in analog distortion instead of harsh
15001 digital 0dB clipping. Disabled by default.
15002 mutel
15004 Mute the left channel. Disabled by default.
15005 muter
15007 Mute the right channel. Disabled by default.
15008 phasel
15010 Change the phase of the left channel. Disabled by default.
15011 phaser
15013 Change the phase of the right channel. Disabled by default.
15014 mode
15016 Set stereo mode. Available values are:
15017 lr>lr
15019 Left/Right to Left/Right, this is default.
15020 lr>ms
15022 Left/Right to Mid/Side.
15023 ms>lr
15025 Mid/Side to Left/Right.
15026 lr>ll
15028 Left/Right to Left/Left.
15029 lr>rr
15031 Left/Right to Right/Right.
15032 lr>l+r
15034 Left/Right to Left + Right.
15035 lr>rl
15037 Left/Right to Right/Left.
15038 ms>ll
15040 Mid/Side to Left/Left.
15041 ms>rr
15043 Mid/Side to Right/Right.
15044 ms>rl
15046 Mid/Side to Right/Left.
15047 lr>l-r
15049 Left/Right to Left - Right.
15050 slev
15052 Set level of side signal. Default is 1. Allowed range is from
15053 0.015625 to 64.
15054 sbal
15056 Set balance of side signal. Default is 0. Allowed range is from -1
15057 to 1.
15058 mlev
15060 Set level of the middle signal. Default is 1. Allowed range is
15061 from 0.015625 to 64.
15062 mpan
15064 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
15065 base
15067 Set stereo base between mono and inversed channels. Default is 0.
15068 Allowed range is from -1 to 1.
15069 delay
15071 Set delay in milliseconds how much to delay left from right channel
15072 and vice versa. Default is 0. Allowed range is from -20 to 20.
15073 sclevel
15075 Set S/C level. Default is 1. Allowed range is from 1 to 100.
15076 phase
15078 Set the stereo phase in degrees. Default is 0. Allowed range is
15079 from 0 to 360.
15080 bmode_in, bmode_out
15082 Set balance mode for balance_in/balance_out option.
15083 Can be one of the following:
15085 balance
15087 Classic balance mode. Attenuate one channel at time. Gain is
15088 raised up to 1.
15089 amplitude
15091 Similar as classic mode above but gain is raised up to 2.
15092 power
15094 Equal power distribution, from -6dB to +6dB range.
15095 Commands
15097 This filter supports the all above options as commands.
15099 Examples
15101 • Apply karaoke like effect:
15103 stereotools=mlev=0.015625
15105 • Convert M/S signal to L/R:
15107 "stereotools=mode=ms>lr"
15109 stereowiden
15111 This filter enhance the stereo effect by suppressing signal common to
15112 both channels and by delaying the signal of left into right and vice
15113 versa, thereby widening the stereo effect.
15114 The filter accepts the following options:
15116 delay
15118 Time in milliseconds of the delay of left signal into right and
15119 vice versa. Default is 20 milliseconds.
15120 feedback
15122 Amount of gain in delayed signal into right and vice versa. Gives a
15123 delay effect of left signal in right output and vice versa which
15124 gives widening effect. Default is 0.3.
15125 crossfeed
15127 Cross feed of left into right with inverted phase. This helps in
15128 suppressing the mono. If the value is 1 it will cancel all the
15129 signal common to both channels. Default is 0.3.
15130 drymix
15132 Set level of input signal of original channel. Default is 0.8.
15133 Commands
15135 This filter supports the all above options except "delay" as commands.
15137 superequalizer
15139 Apply 18 band equalizer.
15140 The filter accepts the following options:
15142 1b Set 65Hz band gain.
15144 2b Set 92Hz band gain.
15146 3b Set 131Hz band gain.
15148 4b Set 185Hz band gain.
15150 5b Set 262Hz band gain.
15152 6b Set 370Hz band gain.
15154 7b Set 523Hz band gain.
15156 8b Set 740Hz band gain.
15158 9b Set 1047Hz band gain.
15160 10b Set 1480Hz band gain.
15162 11b Set 2093Hz band gain.
15164 12b Set 2960Hz band gain.
15166 13b Set 4186Hz band gain.
15168 14b Set 5920Hz band gain.
15170 15b Set 8372Hz band gain.
15172 16b Set 11840Hz band gain.
15174 17b Set 16744Hz band gain.
15176 18b Set 20000Hz band gain.
15178 surround
15180 Apply audio surround upmix filter.
15181 This filter allows to produce multichannel output from audio stream.
15183 The filter accepts the following options:
15185 chl_out
15187 Set output channel layout. By default, this is 5.1.
15188 See the Channel Layout section in the ffmpeg-utils(1) manual for
15190 the required syntax.
15191 chl_in
15193 Set input channel layout. By default, this is stereo.
15194 See the Channel Layout section in the ffmpeg-utils(1) manual for
15196 the required syntax.
15197 level_in
15199 Set input volume level. By default, this is 1.
15200 level_out
15202 Set output volume level. By default, this is 1.
15203 lfe Enable LFE channel output if output channel layout has it. By
15205 default, this is enabled.
15206 lfe_low
15208 Set LFE low cut off frequency. By default, this is 128 Hz.
15209 lfe_high
15211 Set LFE high cut off frequency. By default, this is 256 Hz.
15212 lfe_mode
15214 Set LFE mode, can be add or sub. Default is add. In add mode, LFE
15215 channel is created from input audio and added to output. In sub
15216 mode, LFE channel is created from input audio and added to output
15217 but also all non-LFE output channels are subtracted with output LFE
15218 channel.
15219 smooth
15221 Set temporal smoothness strength, used to gradually change factors
15222 when transforming stereo sound in time. Allowed range is from 0.0
15223 to 1.0. Useful to improve output quality with focus option values
15224 greater than 0.0. Default is 0.0. Only values inside this range
15225 and without edges are effective.
15226 angle
15228 Set angle of stereo surround transform, Allowed range is from 0 to
15229 360. Default is 90.
15230 focus
15232 Set focus of stereo surround transform, Allowed range is from -1 to
15233 1. Default is 0.
15234 fc_in
15236 Set front center input volume. By default, this is 1.
15237 fc_out
15239 Set front center output volume. By default, this is 1.
15240 fl_in
15242 Set front left input volume. By default, this is 1.
15243 fl_out
15245 Set front left output volume. By default, this is 1.
15246 fr_in
15248 Set front right input volume. By default, this is 1.
15249 fr_out
15251 Set front right output volume. By default, this is 1.
15252 sl_in
15254 Set side left input volume. By default, this is 1.
15255 sl_out
15257 Set side left output volume. By default, this is 1.
15258 sr_in
15260 Set side right input volume. By default, this is 1.
15261 sr_out
15263 Set side right output volume. By default, this is 1.
15264 bl_in
15266 Set back left input volume. By default, this is 1.
15267 bl_out
15269 Set back left output volume. By default, this is 1.
15270 br_in
15272 Set back right input volume. By default, this is 1.
15273 br_out
15275 Set back right output volume. By default, this is 1.
15276 bc_in
15278 Set back center input volume. By default, this is 1.
15279 bc_out
15281 Set back center output volume. By default, this is 1.
15282 lfe_in
15284 Set LFE input volume. By default, this is 1.
15285 lfe_out
15287 Set LFE output volume. By default, this is 1.
15288 allx
15290 Set spread usage of stereo image across X axis for all channels.
15291 Allowed range is from -1 to 15. By default this value is negative
15292 -1, and thus unused.
15293 ally
15295 Set spread usage of stereo image across Y axis for all channels.
15296 Allowed range is from -1 to 15. By default this value is negative
15297 -1, and thus unused.
15298 fcx, flx, frx, blx, brx, slx, srx, bcx
15300 Set spread usage of stereo image across X axis for each channel.
15301 Allowed range is from 0.06 to 15. By default this value is 0.5.
15302 fcy, fly, fry, bly, bry, sly, sry, bcy
15304 Set spread usage of stereo image across Y axis for each channel.
15305 Allowed range is from 0.06 to 15. By default this value is 0.5.
15306 win_size
15308 Set window size. Allowed range is from 1024 to 65536. Default size
15309 is 4096.
15310 win_func
15312 Set window function.
15313 It accepts the following values:
15315 rect
15317 bartlett
15318 hann, hanning
15319 hamming
15320 blackman
15321 welch
15322 flattop
15323 bharris
15324 bnuttall
15325 bhann
15326 sine
15327 nuttall
15328 lanczos
15329 gauss
15330 tukey
15331 dolph
15332 cauchy
15333 parzen
15334 poisson
15335 bohman
15336 kaiser
15337 Default is "hann".
15339 overlap
15341 Set window overlap. If set to 1, the recommended overlap for
15342 selected window function will be picked. Default is 0.5.
15343 tiltshelf
15345 Boost or cut the lower frequencies and cut or boost higher frequencies
15346 of the audio using a two-pole shelving filter with a response similar
15347 to that of a standard hi-fi's tone-controls. This is also known as
15348 shelving equalisation (EQ).
15349 The filter accepts the following options:
15351 gain, g
15353 Give the gain at 0 Hz. Its useful range is about -20 (for a large
15354 cut) to +20 (for a large boost). Beware of clipping when using a
15355 positive gain.
15356 frequency, f
15358 Set the filter's central frequency and so can be used to extend or
15359 reduce the frequency range to be boosted or cut. The default value
15360 is 3000 Hz.
15361 width_type, t
15363 Set method to specify band-width of filter.
15364 h Hz
15366 q Q-Factor
15368 o octave
15370 s slope
15372 k kHz
15374 width, w
15376 Determine how steep is the filter's shelf transition.
15377 poles, p
15379 Set number of poles. Default is 2.
15380 mix, m
15382 How much to use filtered signal in output. Default is 1. Range is
15383 between 0 and 1.
15384 channels, c
15386 Specify which channels to filter, by default all available are
15387 filtered.
15388 normalize, n
15390 Normalize biquad coefficients, by default is disabled. Enabling it
15391 will normalize magnitude response at DC to 0dB.
15392 transform, a
15394 Set transform type of IIR filter.
15395 di
15397 dii
15398 tdi
15399 tdii
15400 latt
15401 svf
15402 zdf
15403 precision, r
15404 Set precison of filtering.
15405 auto
15407 Pick automatic sample format depending on surround filters.
15408 s16 Always use signed 16-bit.
15410 s32 Always use signed 32-bit.
15412 f32 Always use float 32-bit.
15414 f64 Always use float 64-bit.
15416 block_size, b
15418 Set block size used for reverse IIR processing. If this value is
15419 set to high enough value (higher than impulse response length
15420 truncated when reaches near zero values) filtering will become
15421 linear phase otherwise if not big enough it will just produce nasty
15422 artifacts.
15423 Note that filter delay will be exactly this many samples when set
15425 to non-zero value.
15426 Commands
15428 This filter supports some options as commands.
15430 treble, highshelf
15432 Boost or cut treble (upper) frequencies of the audio using a two-pole
15433 shelving filter with a response similar to that of a standard hi-fi's
15434 tone-controls. This is also known as shelving equalisation (EQ).
15435 The filter accepts the following options:
15437 gain, g
15439 Give the gain at whichever is the lower of ~22 kHz and the Nyquist
15440 frequency. Its useful range is about -20 (for a large cut) to +20
15441 (for a large boost). Beware of clipping when using a positive gain.
15442 frequency, f
15444 Set the filter's central frequency and so can be used to extend or
15445 reduce the frequency range to be boosted or cut. The default value
15446 is 3000 Hz.
15447 width_type, t
15449 Set method to specify band-width of filter.
15450 h Hz
15452 q Q-Factor
15454 o octave
15456 s slope
15458 k kHz
15460 width, w
15462 Determine how steep is the filter's shelf transition.
15463 poles, p
15465 Set number of poles. Default is 2.
15466 mix, m
15468 How much to use filtered signal in output. Default is 1. Range is
15469 between 0 and 1.
15470 channels, c
15472 Specify which channels to filter, by default all available are
15473 filtered.
15474 normalize, n
15476 Normalize biquad coefficients, by default is disabled. Enabling it
15477 will normalize magnitude response at DC to 0dB.
15478 transform, a
15480 Set transform type of IIR filter.
15481 di
15483 dii
15484 tdi
15485 tdii
15486 latt
15487 svf
15488 zdf
15489 precision, r
15490 Set precison of filtering.
15491 auto
15493 Pick automatic sample format depending on surround filters.
15494 s16 Always use signed 16-bit.
15496 s32 Always use signed 32-bit.
15498 f32 Always use float 32-bit.
15500 f64 Always use float 64-bit.
15502 block_size, b
15504 Set block size used for reverse IIR processing. If this value is
15505 set to high enough value (higher than impulse response length
15506 truncated when reaches near zero values) filtering will become
15507 linear phase otherwise if not big enough it will just produce nasty
15508 artifacts.
15509 Note that filter delay will be exactly this many samples when set
15511 to non-zero value.
15512 Commands
15514 This filter supports the following commands:
15516 frequency, f
15518 Change treble frequency. Syntax for the command is : "frequency"
15519 width_type, t
15521 Change treble width_type. Syntax for the command is : "width_type"
15522 width, w
15524 Change treble width. Syntax for the command is : "width"
15525 gain, g
15527 Change treble gain. Syntax for the command is : "gain"
15528 mix, m
15530 Change treble mix. Syntax for the command is : "mix"
15531 tremolo
15533 Sinusoidal amplitude modulation.
15534 The filter accepts the following options:
15536 f Modulation frequency in Hertz. Modulation frequencies in the
15538 subharmonic range (20 Hz or lower) will result in a tremolo effect.
15539 This filter may also be used as a ring modulator by specifying a
15540 modulation frequency higher than 20 Hz. Range is 0.1 - 20000.0.
15541 Default value is 5.0 Hz.
15542 d Depth of modulation as a percentage. Range is 0.0 - 1.0. Default
15544 value is 0.5.
15545 vibrato
15547 Sinusoidal phase modulation.
15548 The filter accepts the following options:
15550 f Modulation frequency in Hertz. Range is 0.1 - 20000.0. Default
15552 value is 5.0 Hz.
15553 d Depth of modulation as a percentage. Range is 0.0 - 1.0. Default
15555 value is 0.5.
15556 virtualbass
15558 Apply audio Virtual Bass filter.
15559 This filter accepts stereo input and produce stereo with LFE (2.1)
15561 channels output. The newly produced LFE channel have enhanced virtual
15562 bass originally obtained from both stereo channels. This filter
15563 outputs front left and front right channels unchanged as available in
15564 stereo input.
15565 The filter accepts the following options:
15567 cutoff
15569 Set the virtual bass cutoff frequency. Default value is 250 Hz.
15570 Allowed range is from 100 to 500 Hz.
15571 strength
15573 Set the virtual bass strength. Allowed range is from 0.5 to 3.
15574 Default value is 3.
15575 volume
15577 Adjust the input audio volume.
15578 It accepts the following parameters:
15580 volume
15582 Set audio volume expression.
15583 Output values are clipped to the maximum value.
15585 The output audio volume is given by the relation:
15587 <output_volume> = <volume> * <input_volume>
15589 The default value for volume is "1.0".
15591 precision
15593 This parameter represents the mathematical precision.
15594 It determines which input sample formats will be allowed, which
15596 affects the precision of the volume scaling.
15597 fixed
15599 8-bit fixed-point; this limits input sample format to U8, S16,
15600 and S32.
15601 float
15603 32-bit floating-point; this limits input sample format to FLT.
15604 (default)
15605 double
15607 64-bit floating-point; this limits input sample format to DBL.
15608 replaygain
15610 Choose the behaviour on encountering ReplayGain side data in input
15611 frames.
15612 drop
15614 Remove ReplayGain side data, ignoring its contents (the
15615 default).
15616 ignore
15618 Ignore ReplayGain side data, but leave it in the frame.
15619 track
15621 Prefer the track gain, if present.
15622 album
15624 Prefer the album gain, if present.
15625 replaygain_preamp
15627 Pre-amplification gain in dB to apply to the selected replaygain
15628 gain.
15629 Default value for replaygain_preamp is 0.0.
15631 replaygain_noclip
15633 Prevent clipping by limiting the gain applied.
15634 Default value for replaygain_noclip is 1.
15636 eval
15638 Set when the volume expression is evaluated.
15639 It accepts the following values:
15641 once
15643 only evaluate expression once during the filter initialization,
15644 or when the volume command is sent
15645 frame
15647 evaluate expression for each incoming frame
15648 Default value is once.
15650 The volume expression can contain the following parameters.
15652 n frame number (starting at zero)
15654 nb_channels
15656 number of channels
15657 nb_consumed_samples
15659 number of samples consumed by the filter
15660 nb_samples
15662 number of samples in the current frame
15663 pos original frame position in the file
15665 pts frame PTS
15667 sample_rate
15669 sample rate
15670 startpts
15672 PTS at start of stream
15673 startt
15675 time at start of stream
15676 t frame time
15678 tb timestamp timebase
15680 volume
15682 last set volume value
15683 Note that when eval is set to once only the sample_rate and tb
15685 variables are available, all other variables will evaluate to NAN.
15686 Commands
15688 This filter supports the following commands:
15690 volume
15692 Modify the volume expression. The command accepts the same syntax
15693 of the corresponding option.
15694 If the specified expression is not valid, it is kept at its current
15696 value.
15697 Examples
15699 • Halve the input audio volume:
15701 volume=volume=0.5
15703 volume=volume=1/2
15704 volume=volume=-6.0206dB
15705 In all the above example the named key for volume can be omitted,
15707 for example like in:
15708 volume=0.5
15710 • Increase input audio power by 6 decibels using fixed-point
15712 precision:
15713 volume=volume=6dB:precision=fixed
15715 • Fade volume after time 10 with an annihilation period of 5 seconds:
15717 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
15719 volumedetect
15721 Detect the volume of the input video.
15722 The filter has no parameters. It supports only 16-bit signed integer
15724 samples, so the input will be converted when needed. Statistics about
15725 the volume will be printed in the log when the input stream end is
15726 reached.
15727 In particular it will show the mean volume (root mean square), maximum
15729 volume (on a per-sample basis), and the beginning of a histogram of the
15730 registered volume values (from the maximum value to a cumulated 1/1000
15731 of the samples).
15732 All volumes are in decibels relative to the maximum PCM value.
15734 Examples
15736 Here is an excerpt of the output:
15738 [Parsed_volumedetect_0 0xa23120] mean_volume: -27 dB
15740 [Parsed_volumedetect_0 0xa23120] max_volume: -4 dB
15741 [Parsed_volumedetect_0 0xa23120] histogram_4db: 6
15742 [Parsed_volumedetect_0 0xa23120] histogram_5db: 62
15743 [Parsed_volumedetect_0 0xa23120] histogram_6db: 286
15744 [Parsed_volumedetect_0 0xa23120] histogram_7db: 1042
15745 [Parsed_volumedetect_0 0xa23120] histogram_8db: 2551
15746 [Parsed_volumedetect_0 0xa23120] histogram_9db: 4609
15747 [Parsed_volumedetect_0 0xa23120] histogram_10db: 8409
15748 It means that:
15750 • The mean square energy is approximately -27 dB, or 10^-2.7.
15752 • The largest sample is at -4 dB, or more precisely between -4 dB and
15754 -5 dB.
15755 • There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
15757 In other words, raising the volume by +4 dB does not cause any
15759 clipping, raising it by +5 dB causes clipping for 6 samples, etc.
15760
15762 Below is a description of the currently available audio sources.
15763 abuffer
15765 Buffer audio frames, and make them available to the filter chain.
15766 This source is mainly intended for a programmatic use, in particular
15768 through the interface defined in libavfilter/buffersrc.h.
15769 It accepts the following parameters:
15771 time_base
15773 The timebase which will be used for timestamps of submitted frames.
15774 It must be either a floating-point number or in
15775 numerator/denominator form.
15776 sample_rate
15778 The sample rate of the incoming audio buffers.
15779 sample_fmt
15781 The sample format of the incoming audio buffers. Either a sample
15782 format name or its corresponding integer representation from the
15783 enum AVSampleFormat in libavutil/samplefmt.h
15784 channel_layout
15786 The channel layout of the incoming audio buffers. Either a channel
15787 layout name from channel_layout_map in libavutil/channel_layout.c
15788 or its corresponding integer representation from the AV_CH_LAYOUT_*
15789 macros in libavutil/channel_layout.h
15790 channels
15792 The number of channels of the incoming audio buffers. If both
15793 channels and channel_layout are specified, then they must be
15794 consistent.
15795 Examples
15797 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
15799 will instruct the source to accept planar 16bit signed stereo at
15801 44100Hz. Since the sample format with name "s16p" corresponds to the
15802 number 6 and the "stereo" channel layout corresponds to the value 0x3,
15803 this is equivalent to:
15804 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
15806 aevalsrc
15808 Generate an audio signal specified by an expression.
15809 This source accepts in input one or more expressions (one for each
15811 channel), which are evaluated and used to generate a corresponding
15812 audio signal.
15813 This source accepts the following options:
15815 exprs
15817 Set the '|'-separated expressions list for each separate channel.
15818 In case the channel_layout option is not specified, the selected
15819 channel layout depends on the number of provided expressions.
15820 Otherwise the last specified expression is applied to the remaining
15821 output channels.
15822 channel_layout, c
15824 Set the channel layout. The number of channels in the specified
15825 layout must be equal to the number of specified expressions.
15826 duration, d
15828 Set the minimum duration of the sourced audio. See the Time
15829 duration section in the ffmpeg-utils(1) manual for the accepted
15830 syntax. Note that the resulting duration may be greater than the
15831 specified duration, as the generated audio is always cut at the end
15832 of a complete frame.
15833 If not specified, or the expressed duration is negative, the audio
15835 is supposed to be generated forever.
15836 nb_samples, n
15838 Set the number of samples per channel per each output frame,
15839 default to 1024.
15840 sample_rate, s
15842 Specify the sample rate, default to 44100.
15843 Each expression in exprs can contain the following constants:
15845 n number of the evaluated sample, starting from 0
15847 t time of the evaluated sample expressed in seconds, starting from 0
15849 s sample rate
15851 Examples
15853 • Generate silence:
15855 aevalsrc=0
15857 • Generate a sin signal with frequency of 440 Hz, set sample rate to
15859 8000 Hz:
15860 aevalsrc="sin(440*2*PI*t):s=8000"
15862 • Generate a two channels signal, specify the channel layout (Front
15864 Center + Back Center) explicitly:
15865 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
15867 • Generate white noise:
15869 aevalsrc="-2+random(0)"
15871 • Generate an amplitude modulated signal:
15873 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
15875 • Generate 2.5 Hz binaural beats on a 360 Hz carrier:
15877 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
15879 afdelaysrc
15881 Generate a fractional delay FIR coefficients.
15882 The resulting stream can be used with afir filter for filtering the
15884 audio signal.
15885 The filter accepts the following options:
15887 delay, d
15889 Set the fractional delay. Default is 0.
15890 sample_rate, r
15892 Set the sample rate, default is 44100.
15893 nb_samples, n
15895 Set the number of samples per each frame. Default is 1024.
15896 taps, t
15898 Set the number of filter coefficents in output audio stream.
15899 Default value is 0.
15900 channel_layout, c
15902 Specifies the channel layout, and can be a string representing a
15903 channel layout. The default value of channel_layout is "stereo".
15904 afirsrc
15906 Generate a FIR coefficients using frequency sampling method.
15907 The resulting stream can be used with afir filter for filtering the
15909 audio signal.
15910 The filter accepts the following options:
15912 taps, t
15914 Set number of filter coefficents in output audio stream. Default
15915 value is 1025.
15916 frequency, f
15918 Set frequency points from where magnitude and phase are set. This
15919 must be in non decreasing order, and first element must be 0, while
15920 last element must be 1. Elements are separated by white spaces.
15921 magnitude, m
15923 Set magnitude value for every frequency point set by frequency.
15924 Number of values must be same as number of frequency points.
15925 Values are separated by white spaces.
15926 phase, p
15928 Set phase value for every frequency point set by frequency. Number
15929 of values must be same as number of frequency points. Values are
15930 separated by white spaces.
15931 sample_rate, r
15933 Set sample rate, default is 44100.
15934 nb_samples, n
15936 Set number of samples per each frame. Default is 1024.
15937 win_func, w
15939 Set window function. Default is blackman.
15940 anullsrc
15942 The null audio source, return unprocessed audio frames. It is mainly
15943 useful as a template and to be employed in analysis / debugging tools,
15944 or as the source for filters which ignore the input data (for example
15945 the sox synth filter).
15946 This source accepts the following options:
15948 channel_layout, cl
15950 Specifies the channel layout, and can be either an integer or a
15951 string representing a channel layout. The default value of
15952 channel_layout is "stereo".
15953 Check the channel_layout_map definition in
15955 libavutil/channel_layout.c for the mapping between strings and
15956 channel layout values.
15957 sample_rate, r
15959 Specifies the sample rate, and defaults to 44100.
15960 nb_samples, n
15962 Set the number of samples per requested frames.
15963 duration, d
15965 Set the duration of the sourced audio. See the Time duration
15966 section in the ffmpeg-utils(1) manual for the accepted syntax.
15967 If not specified, or the expressed duration is negative, the audio
15969 is supposed to be generated forever.
15970 Examples
15972 • Set the sample rate to 48000 Hz and the channel layout to
15974 AV_CH_LAYOUT_MONO.
15975 anullsrc=r=48000:cl=4
15977 • Do the same operation with a more obvious syntax:
15979 anullsrc=r=48000:cl=mono
15981 All the parameters need to be explicitly defined.
15983 flite
15985 Synthesize a voice utterance using the libflite library.
15986 To enable compilation of this filter you need to configure FFmpeg with
15988 "--enable-libflite".
15989 Note that versions of the flite library prior to 2.0 are not thread-
15991 safe.
15992 The filter accepts the following options:
15994 list_voices
15996 If set to 1, list the names of the available voices and exit
15997 immediately. Default value is 0.
15998 nb_samples, n
16000 Set the maximum number of samples per frame. Default value is 512.
16001 textfile
16003 Set the filename containing the text to speak.
16004 text
16006 Set the text to speak.
16007 voice, v
16009 Set the voice to use for the speech synthesis. Default value is
16010 "kal". See also the list_voices option.
16011 Examples
16013 • Read from file speech.txt, and synthesize the text using the
16015 standard flite voice:
16016 flite=textfile=speech.txt
16018 • Read the specified text selecting the "slt" voice:
16020 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
16022 • Input text to ffmpeg:
16024 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
16026 • Make ffplay speak the specified text, using "flite" and the "lavfi"
16028 device:
16029 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
16031 For more information about libflite, check:
16033 <http://www.festvox.org/flite/>
16034 anoisesrc
16036 Generate a noise audio signal.
16037 The filter accepts the following options:
16039 sample_rate, r
16041 Specify the sample rate. Default value is 48000 Hz.
16042 amplitude, a
16044 Specify the amplitude (0.0 - 1.0) of the generated audio stream.
16045 Default value is 1.0.
16046 duration, d
16048 Specify the duration of the generated audio stream. Not specifying
16049 this option results in noise with an infinite length.
16050 color, colour, c
16052 Specify the color of noise. Available noise colors are white, pink,
16053 brown, blue, violet and velvet. Default color is white.
16054 seed, s
16056 Specify a value used to seed the PRNG.
16057 nb_samples, n
16059 Set the number of samples per each output frame, default is 1024.
16060 Examples
16062 • Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate
16064 and an amplitude of 0.5:
16065 anoisesrc=d=60:c=pink:r=44100:a=0.5
16067 hilbert
16069 Generate odd-tap Hilbert transform FIR coefficients.
16070 The resulting stream can be used with afir filter for phase-shifting
16072 the signal by 90 degrees.
16073 This is used in many matrix coding schemes and for analytic signal
16075 generation. The process is often written as a multiplication by i (or
16076 j), the imaginary unit.
16077 The filter accepts the following options:
16079 sample_rate, s
16081 Set sample rate, default is 44100.
16082 taps, t
16084 Set length of FIR filter, default is 22051.
16085 nb_samples, n
16087 Set number of samples per each frame.
16088 win_func, w
16090 Set window function to be used when generating FIR coefficients.
16091 sinc
16093 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or
16094 band-reject FIR coefficients.
16095 The resulting stream can be used with afir filter for filtering the
16097 audio signal.
16098 The filter accepts the following options:
16100 sample_rate, r
16102 Set sample rate, default is 44100.
16103 nb_samples, n
16105 Set number of samples per each frame. Default is 1024.
16106 hp Set high-pass frequency. Default is 0.
16108 lp Set low-pass frequency. Default is 0. If high-pass frequency is
16110 lower than low-pass frequency and low-pass frequency is higher than
16111 0 then filter will create band-pass filter coefficients, otherwise
16112 band-reject filter coefficients.
16113 phase
16115 Set filter phase response. Default is 50. Allowed range is from 0
16116 to 100.
16117 beta
16119 Set Kaiser window beta.
16120 att Set stop-band attenuation. Default is 120dB, allowed range is from
16122 40 to 180 dB.
16123 round
16125 Enable rounding, by default is disabled.
16126 hptaps
16128 Set number of taps for high-pass filter.
16129 lptaps
16131 Set number of taps for low-pass filter.
16132 sine
16134 Generate an audio signal made of a sine wave with amplitude 1/8.
16135 The audio signal is bit-exact.
16137 The filter accepts the following options:
16139 frequency, f
16141 Set the carrier frequency. Default is 440 Hz.
16142 beep_factor, b
16144 Enable a periodic beep every second with frequency beep_factor
16145 times the carrier frequency. Default is 0, meaning the beep is
16146 disabled.
16147 sample_rate, r
16149 Specify the sample rate, default is 44100.
16150 duration, d
16152 Specify the duration of the generated audio stream.
16153 samples_per_frame
16155 Set the number of samples per output frame.
16156 The expression can contain the following constants:
16158 n The (sequential) number of the output audio frame, starting
16160 from 0.
16161 pts The PTS (Presentation TimeStamp) of the output audio frame,
16163 expressed in TB units.
16164 t The PTS of the output audio frame, expressed in seconds.
16166 TB The timebase of the output audio frames.
16168 Default is 1024.
16170 Examples
16172 • Generate a simple 440 Hz sine wave:
16174 sine
16176 • Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5
16178 seconds:
16179 sine=220:4:d=5
16181 sine=f=220:b=4:d=5
16182 sine=frequency=220:beep_factor=4:duration=5
16183 • Generate a 1 kHz sine wave following "1602,1601,1602,1601,1602"
16185 NTSC pattern:
16186 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
16188
16190 Below is a description of the currently available audio sinks.
16191 abuffersink
16193 Buffer audio frames, and make them available to the end of filter
16194 chain.
16195 This sink is mainly intended for programmatic use, in particular
16197 through the interface defined in libavfilter/buffersink.h or the
16198 options system.
16199 It accepts a pointer to an AVABufferSinkContext structure, which
16201 defines the incoming buffers' formats, to be passed as the opaque
16202 parameter to "avfilter_init_filter" for initialization.
16203 anullsink
16205 Null audio sink; do absolutely nothing with the input audio. It is
16206 mainly useful as a template and for use in analysis / debugging tools.
16207
16209 When you configure your FFmpeg build, you can disable any of the
16210 existing filters using "--disable-filters". The configure output will
16211 show the video filters included in your build.
16212 Below is a description of the currently available video filters.
16214 addroi
16216 Mark a region of interest in a video frame.
16217 The frame data is passed through unchanged, but metadata is attached to
16219 the frame indicating regions of interest which can affect the behaviour
16220 of later encoding. Multiple regions can be marked by applying the
16221 filter multiple times.
16222 x Region distance in pixels from the left edge of the frame.
16224 y Region distance in pixels from the top edge of the frame.
16226 w Region width in pixels.
16228 h Region height in pixels.
16230 The parameters x, y, w and h are expressions, and may contain the
16232 following variables:
16233 iw Width of the input frame.
16235 ih Height of the input frame.
16237 qoffset
16239 Quantisation offset to apply within the region.
16240 This must be a real value in the range -1 to +1. A value of zero
16242 indicates no quality change. A negative value asks for better
16243 quality (less quantisation), while a positive value asks for worse
16244 quality (greater quantisation).
16245 The range is calibrated so that the extreme values indicate the
16247 largest possible offset - if the rest of the frame is encoded with
16248 the worst possible quality, an offset of -1 indicates that this
16249 region should be encoded with the best possible quality anyway.
16250 Intermediate values are then interpolated in some codec-dependent
16251 way.
16252 For example, in 10-bit H.264 the quantisation parameter varies
16254 between -12 and 51. A typical qoffset value of -1/10 therefore
16255 indicates that this region should be encoded with a QP around one-
16256 tenth of the full range better than the rest of the frame. So, if
16257 most of the frame were to be encoded with a QP of around 30, this
16258 region would get a QP of around 24 (an offset of approximately
16259 -1/10 * (51 - -12) = -6.3). An extreme value of -1 would indicate
16260 that this region should be encoded with the best possible quality
16261 regardless of the treatment of the rest of the frame - that is,
16262 should be encoded at a QP of -12.
16263 clear
16265 If set to true, remove any existing regions of interest marked on
16266 the frame before adding the new one.
16267 Examples
16269 • Mark the centre quarter of the frame as interesting.
16271 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
16273 • Mark the 100-pixel-wide region on the left edge of the frame as
16275 very uninteresting (to be encoded at much lower quality than the
16276 rest of the frame).
16277 addroi=0:0:100:ih:+1/5
16279 alphaextract
16281 Extract the alpha component from the input as a grayscale video. This
16282 is especially useful with the alphamerge filter.
16283 alphamerge
16285 Add or replace the alpha component of the primary input with the
16286 grayscale value of a second input. This is intended for use with
16287 alphaextract to allow the transmission or storage of frame sequences
16288 that have alpha in a format that doesn't support an alpha channel.
16289 For example, to reconstruct full frames from a normal YUV-encoded video
16291 and a separate video created with alphaextract, you might use:
16292 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
16294 amplify
16296 Amplify differences between current pixel and pixels of adjacent frames
16297 in same pixel location.
16298 This filter accepts the following options:
16300 radius
16302 Set frame radius. Default is 2. Allowed range is from 1 to 63. For
16303 example radius of 3 will instruct filter to calculate average of 7
16304 frames.
16305 factor
16307 Set factor to amplify difference. Default is 2. Allowed range is
16308 from 0 to 65535.
16309 threshold
16311 Set threshold for difference amplification. Any difference greater
16312 or equal to this value will not alter source pixel. Default is 10.
16313 Allowed range is from 0 to 65535.
16314 tolerance
16316 Set tolerance for difference amplification. Any difference lower to
16317 this value will not alter source pixel. Default is 0. Allowed
16318 range is from 0 to 65535.
16319 low Set lower limit for changing source pixel. Default is 65535.
16321 Allowed range is from 0 to 65535. This option controls maximum
16322 possible value that will decrease source pixel value.
16323 high
16325 Set high limit for changing source pixel. Default is 65535. Allowed
16326 range is from 0 to 65535. This option controls maximum possible
16327 value that will increase source pixel value.
16328 planes
16330 Set which planes to filter. Default is all. Allowed range is from 0
16331 to 15.
16332 Commands
16334 This filter supports the following commands that corresponds to option
16336 of same name:
16337 factor
16339 threshold
16340 tolerance
16341 low
16342 high
16343 planes
16344 ass
16346 Same as the subtitles filter, except that it doesn't require libavcodec
16347 and libavformat to work. On the other hand, it is limited to ASS
16348 (Advanced Substation Alpha) subtitles files.
16349 This filter accepts the following option in addition to the common
16351 options from the subtitles filter:
16352 shaping
16354 Set the shaping engine
16355 Available values are:
16357 auto
16359 The default libass shaping engine, which is the best available.
16360 simple
16362 Fast, font-agnostic shaper that can do only substitutions
16363 complex
16365 Slower shaper using OpenType for substitutions and positioning
16366 The default is "auto".
16368 atadenoise
16370 Apply an Adaptive Temporal Averaging Denoiser to the video input.
16371 The filter accepts the following options:
16373 0a Set threshold A for 1st plane. Default is 0.02. Valid range is 0
16375 to 0.3.
16376 0b Set threshold B for 1st plane. Default is 0.04. Valid range is 0
16378 to 5.
16379 1a Set threshold A for 2nd plane. Default is 0.02. Valid range is 0
16381 to 0.3.
16382 1b Set threshold B for 2nd plane. Default is 0.04. Valid range is 0
16384 to 5.
16385 2a Set threshold A for 3rd plane. Default is 0.02. Valid range is 0
16387 to 0.3.
16388 2b Set threshold B for 3rd plane. Default is 0.04. Valid range is 0
16390 to 5.
16391 Threshold A is designed to react on abrupt changes in the input
16393 signal and threshold B is designed to react on continuous changes
16394 in the input signal.
16395 s Set number of frames filter will use for averaging. Default is 9.
16397 Must be odd number in range [5, 129].
16398 p Set what planes of frame filter will use for averaging. Default is
16400 all.
16401 a Set what variant of algorithm filter will use for averaging.
16403 Default is "p" parallel. Alternatively can be set to "s" serial.
16404 Parallel can be faster then serial, while other way around is never
16406 true. Parallel will abort early on first change being greater then
16407 thresholds, while serial will continue processing other side of
16408 frames if they are equal or below thresholds.
16409 0s
16411 1s
16412 2s Set sigma for 1st plane, 2nd plane or 3rd plane. Default is 32767.
16413 Valid range is from 0 to 32767. This options controls weight for
16414 each pixel in radius defined by size. Default value means every
16415 pixel have same weight. Setting this option to 0 effectively
16416 disables filtering.
16417 Commands
16419 This filter supports same commands as options except option "s". The
16421 command accepts the same syntax of the corresponding option.
16422 avgblur
16424 Apply average blur filter.
16425 The filter accepts the following options:
16427 sizeX
16429 Set horizontal radius size.
16430 planes
16432 Set which planes to filter. By default all planes are filtered.
16433 sizeY
16435 Set vertical radius size, if zero it will be same as "sizeX".
16436 Default is 0.
16437 Commands
16439 This filter supports same commands as options. The command accepts the
16441 same syntax of the corresponding option.
16442 If the specified expression is not valid, it is kept at its current
16444 value.
16445 backgroundkey
16447 Turns a static background into transparency.
16448 The filter accepts the following option:
16450 threshold
16452 Threshold for scene change detection.
16453 similarity
16455 Similarity percentage with the background.
16456 blend
16458 Set the blend amount for pixels that are not similar.
16459 Commands
16461 This filter supports the all above options as commands.
16463 bbox
16465 Compute the bounding box for the non-black pixels in the input frame
16466 luminance plane.
16467 This filter computes the bounding box containing all the pixels with a
16469 luminance value greater than the minimum allowed value. The parameters
16470 describing the bounding box are printed on the filter log.
16471 The filter accepts the following option:
16473 min_val
16475 Set the minimal luminance value. Default is 16.
16476 Commands
16478 This filter supports the all above options as commands.
16480 bilateral
16482 Apply bilateral filter, spatial smoothing while preserving edges.
16483 The filter accepts the following options:
16485 sigmaS
16487 Set sigma of gaussian function to calculate spatial weight.
16488 Allowed range is 0 to 512. Default is 0.1.
16489 sigmaR
16491 Set sigma of gaussian function to calculate range weight. Allowed
16492 range is 0 to 1. Default is 0.1.
16493 planes
16495 Set planes to filter. Default is first only.
16496 Commands
16498 This filter supports the all above options as commands.
16500 bilateral_cuda
16502 CUDA accelerated bilateral filter, an edge preserving filter. This
16503 filter is mathematically accurate thanks to the use of GPU
16504 acceleration. For best output quality, use one to one chroma
16505 subsampling, i.e. yuv444p format.
16506 The filter accepts the following options:
16508 sigmaS
16510 Set sigma of gaussian function to calculate spatial weight, also
16511 called sigma space. Allowed range is 0.1 to 512. Default is 0.1.
16512 sigmaR
16514 Set sigma of gaussian function to calculate color range weight,
16515 also called sigma color. Allowed range is 0.1 to 512. Default is
16516 0.1.
16517 window_size
16519 Set window size of the bilateral function to determine the number
16520 of neighbours to loop on. If the number entered is even, one will
16521 be added automatically. Allowed range is 1 to 255. Default is 1.
16522 Examples
16524 • Apply the bilateral filter on a video.
16526 ./ffmpeg -v verbose \
16528 -hwaccel cuda -hwaccel_output_format cuda -i input.mp4 \
16529 -init_hw_device cuda \
16530 -filter_complex \
16531 " \
16532 [0:v]scale_cuda=format=yuv444p[scaled_video];
16533 [scaled_video]bilateral_cuda=window_size=9:sigmaS=3.0:sigmaR=50.0" \
16534 -an -sn -c:v h264_nvenc -cq 20 out.mp4
16535 bitplanenoise
16537 Show and measure bit plane noise.
16538 The filter accepts the following options:
16540 bitplane
16542 Set which plane to analyze. Default is 1.
16543 filter
16545 Filter out noisy pixels from "bitplane" set above. Default is
16546 disabled.
16547 blackdetect
16549 Detect video intervals that are (almost) completely black. Can be
16550 useful to detect chapter transitions, commercials, or invalid
16551 recordings.
16552 The filter outputs its detection analysis to both the log as well as
16554 frame metadata. If a black segment of at least the specified minimum
16555 duration is found, a line with the start and end timestamps as well as
16556 duration is printed to the log with level "info". In addition, a log
16557 line with level "debug" is printed per frame showing the black amount
16558 detected for that frame.
16559 The filter also attaches metadata to the first frame of a black segment
16561 with key "lavfi.black_start" and to the first frame after the black
16562 segment ends with key "lavfi.black_end". The value is the frame's
16563 timestamp. This metadata is added regardless of the minimum duration
16564 specified.
16565 The filter accepts the following options:
16567 black_min_duration, d
16569 Set the minimum detected black duration expressed in seconds. It
16570 must be a non-negative floating point number.
16571 Default value is 2.0.
16573 picture_black_ratio_th, pic_th
16575 Set the threshold for considering a picture "black". Express the
16576 minimum value for the ratio:
16577 <nb_black_pixels> / <nb_pixels>
16579 for which a picture is considered black. Default value is 0.98.
16581 pixel_black_th, pix_th
16583 Set the threshold for considering a pixel "black".
16584 The threshold expresses the maximum pixel luminance value for which
16586 a pixel is considered "black". The provided value is scaled
16587 according to the following equation:
16588 <absolute_threshold> = <luminance_minimum_value> + <pixel_black_th> * <luminance_range_size>
16590 luminance_range_size and luminance_minimum_value depend on the
16592 input video format, the range is [0-255] for YUV full-range formats
16593 and [16-235] for YUV non full-range formats.
16594 Default value is 0.10.
16596 The following example sets the maximum pixel threshold to the minimum
16598 value, and detects only black intervals of 2 or more seconds:
16599 blackdetect=d=2:pix_th=0.00
16601 blackframe
16603 Detect frames that are (almost) completely black. Can be useful to
16604 detect chapter transitions or commercials. Output lines consist of the
16605 frame number of the detected frame, the percentage of blackness, the
16606 position in the file if known or -1 and the timestamp in seconds.
16607 In order to display the output lines, you need to set the loglevel at
16609 least to the AV_LOG_INFO value.
16610 This filter exports frame metadata "lavfi.blackframe.pblack". The
16612 value represents the percentage of pixels in the picture that are below
16613 the threshold value.
16614 It accepts the following parameters:
16616 amount
16618 The percentage of the pixels that have to be below the threshold;
16619 it defaults to 98.
16620 threshold, thresh
16622 The threshold below which a pixel value is considered black; it
16623 defaults to 32.
16624 blend
16626 Blend two video frames into each other.
16627 The "blend" filter takes two input streams and outputs one stream, the
16629 first input is the "top" layer and second input is "bottom" layer. By
16630 default, the output terminates when the longest input terminates.
16631 The "tblend" (time blend) filter takes two consecutive frames from one
16633 single stream, and outputs the result obtained by blending the new
16634 frame on top of the old frame.
16635 A description of the accepted options follows.
16637 c0_mode
16639 c1_mode
16640 c2_mode
16641 c3_mode
16642 all_mode
16643 Set blend mode for specific pixel component or all pixel components
16644 in case of all_mode. Default value is "normal".
16645 Available values for component modes are:
16647 addition
16649 and
16650 average
16651 bleach
16652 burn
16653 darken
16654 difference
16655 divide
16656 dodge
16657 exclusion
16658 extremity
16659 freeze
16660 geometric
16661 glow
16662 grainextract
16663 grainmerge
16664 hardlight
16665 hardmix
16666 hardoverlay
16667 harmonic
16668 heat
16669 interpolate
16670 lighten
16671 linearlight
16672 multiply
16673 multiply128
16674 negation
16675 normal
16676 or
16677 overlay
16678 phoenix
16679 pinlight
16680 reflect
16681 screen
16682 softdifference
16683 softlight
16684 stain
16685 subtract
16686 vividlight
16687 xor
16688 c0_opacity
16689 c1_opacity
16690 c2_opacity
16691 c3_opacity
16692 all_opacity
16693 Set blend opacity for specific pixel component or all pixel
16694 components in case of all_opacity. Only used in combination with
16695 pixel component blend modes.
16696 c0_expr
16698 c1_expr
16699 c2_expr
16700 c3_expr
16701 all_expr
16702 Set blend expression for specific pixel component or all pixel
16703 components in case of all_expr. Note that related mode options will
16704 be ignored if those are set.
16705 The expressions can use the following variables:
16707 N The sequential number of the filtered frame, starting from 0.
16709 X
16711 Y the coordinates of the current sample
16712 W
16714 H the width and height of currently filtered plane
16715 SW
16717 SH Width and height scale for the plane being filtered. It is the
16718 ratio between the dimensions of the current plane to the luma
16719 plane, e.g. for a "yuv420p" frame, the values are "1,1" for the
16720 luma plane and "0.5,0.5" for the chroma planes.
16721 T Time of the current frame, expressed in seconds.
16723 TOP, A
16725 Value of pixel component at current location for first video
16726 frame (top layer).
16727 BOTTOM, B
16729 Value of pixel component at current location for second video
16730 frame (bottom layer).
16731 The "blend" filter also supports the framesync options.
16733 Examples
16735 • Apply transition from bottom layer to top layer in first 10
16737 seconds:
16738 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
16740 • Apply linear horizontal transition from top layer to bottom layer:
16742 blend=all_expr='A*(X/W)+B*(1-X/W)'
16744 • Apply 1x1 checkerboard effect:
16746 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
16748 • Apply uncover left effect:
16750 blend=all_expr='if(gte(N*SW+X,W),A,B)'
16752 • Apply uncover down effect:
16754 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
16756 • Apply uncover up-left effect:
16758 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
16760 • Split diagonally video and shows top and bottom layer on each side:
16762 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
16764 • Display differences between the current and the previous frame:
16766 tblend=all_mode=grainextract
16768 Commands
16770 This filter supports same commands as options.
16772 blockdetect
16774 Determines blockiness of frames without altering the input frames.
16775 Based on Remco Muijs and Ihor Kirenko: "A no-reference blocking
16777 artifact measure for adaptive video processing." 2005 13th European
16778 signal processing conference.
16779 The filter accepts the following options:
16781 period_min
16783 period_max
16784 Set minimum and maximum values for determining pixel grids
16785 (periods). Default values are [3,24].
16786 planes
16788 Set planes to filter. Default is first only.
16789 Examples
16791 • Determine blockiness for the first plane and search for periods
16793 within [8,32]:
16794 blockdetect=period_min=8:period_max=32:planes=1
16796 blurdetect
16798 Determines blurriness of frames without altering the input frames.
16799 Based on Marziliano, Pina, et al. "A no-reference perceptual blur
16801 metric." Allows for a block-based abbreviation.
16802 The filter accepts the following options:
16804 low
16806 high
16807 Set low and high threshold values used by the Canny thresholding
16808 algorithm.
16809 The high threshold selects the "strong" edge pixels, which are then
16811 connected through 8-connectivity with the "weak" edge pixels
16812 selected by the low threshold.
16813 low and high threshold values must be chosen in the range [0,1],
16815 and low should be lesser or equal to high.
16816 Default value for low is "20/255", and default value for high is
16818 "50/255".
16819 radius
16821 Define the radius to search around an edge pixel for local maxima.
16822 block_pct
16824 Determine blurriness only for the most significant blocks, given in
16825 percentage.
16826 block_width
16828 Determine blurriness for blocks of width block_width. If set to any
16829 value smaller 1, no blocks are used and the whole image is
16830 processed as one no matter of block_height.
16831 block_height
16833 Determine blurriness for blocks of height block_height. If set to
16834 any value smaller 1, no blocks are used and the whole image is
16835 processed as one no matter of block_width.
16836 planes
16838 Set planes to filter. Default is first only.
16839 Examples
16841 • Determine blur for 80% of most significant 32x32 blocks:
16843 blurdetect=block_width=32:block_height=32:block_pct=80
16845 bm3d
16847 Denoise frames using Block-Matching 3D algorithm.
16848 The filter accepts the following options.
16850 sigma
16852 Set denoising strength. Default value is 1. Allowed range is from
16853 0 to 999.9. The denoising algorithm is very sensitive to sigma, so
16854 adjust it according to the source.
16855 block
16857 Set local patch size. This sets dimensions in 2D.
16858 bstep
16860 Set sliding step for processing blocks. Default value is 4.
16861 Allowed range is from 1 to 64. Smaller values allows processing
16862 more reference blocks and is slower.
16863 group
16865 Set maximal number of similar blocks for 3rd dimension. Default
16866 value is 1. When set to 1, no block matching is done. Larger
16867 values allows more blocks in single group. Allowed range is from 1
16868 to 256.
16869 range
16871 Set radius for search block matching. Default is 9. Allowed range
16872 is from 1 to INT32_MAX.
16873 mstep
16875 Set step between two search locations for block matching. Default
16876 is 1. Allowed range is from 1 to 64. Smaller is slower.
16877 thmse
16879 Set threshold of mean square error for block matching. Valid range
16880 is 0 to INT32_MAX.
16881 hdthr
16883 Set thresholding parameter for hard thresholding in 3D transformed
16884 domain. Larger values results in stronger hard-thresholding
16885 filtering in frequency domain.
16886 estim
16888 Set filtering estimation mode. Can be "basic" or "final". Default
16889 is "basic".
16890 ref If enabled, filter will use 2nd stream for block matching. Default
16892 is disabled for "basic" value of estim option, and always enabled
16893 if value of estim is "final".
16894 planes
16896 Set planes to filter. Default is all available except alpha.
16897 Examples
16899 • Basic filtering with bm3d:
16901 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
16903 • Same as above, but filtering only luma:
16905 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
16907 • Same as above, but with both estimation modes:
16909 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
16911 • Same as above, but prefilter with nlmeans filter instead:
16913 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
16915 boxblur
16917 Apply a boxblur algorithm to the input video.
16918 It accepts the following parameters:
16920 luma_radius, lr
16922 luma_power, lp
16923 chroma_radius, cr
16924 chroma_power, cp
16925 alpha_radius, ar
16926 alpha_power, ap
16927 A description of the accepted options follows.
16929 luma_radius, lr
16931 chroma_radius, cr
16932 alpha_radius, ar
16933 Set an expression for the box radius in pixels used for blurring
16934 the corresponding input plane.
16935 The radius value must be a non-negative number, and must not be
16937 greater than the value of the expression "min(w,h)/2" for the luma
16938 and alpha planes, and of "min(cw,ch)/2" for the chroma planes.
16939 Default value for luma_radius is "2". If not specified,
16941 chroma_radius and alpha_radius default to the corresponding value
16942 set for luma_radius.
16943 The expressions can contain the following constants:
16945 w
16947 h The input width and height in pixels.
16948 cw
16950 ch The input chroma image width and height in pixels.
16951 hsub
16953 vsub
16954 The horizontal and vertical chroma subsample values. For
16955 example, for the pixel format "yuv422p", hsub is 2 and vsub is
16956 1.
16957 luma_power, lp
16959 chroma_power, cp
16960 alpha_power, ap
16961 Specify how many times the boxblur filter is applied to the
16962 corresponding plane.
16963 Default value for luma_power is 2. If not specified, chroma_power
16965 and alpha_power default to the corresponding value set for
16966 luma_power.
16967 A value of 0 will disable the effect.
16969 Examples
16971 • Apply a boxblur filter with the luma, chroma, and alpha radii set
16973 to 2:
16974 boxblur=luma_radius=2:luma_power=1
16976 boxblur=2:1
16977 • Set the luma radius to 2, and alpha and chroma radius to 0:
16979 boxblur=2:1:cr=0:ar=0
16981 • Set the luma and chroma radii to a fraction of the video dimension:
16983 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
16985 bwdif
16987 Deinterlace the input video ("bwdif" stands for "Bob Weaver
16988 Deinterlacing Filter").
16989 Motion adaptive deinterlacing based on yadif with the use of w3fdif and
16991 cubic interpolation algorithms. It accepts the following parameters:
16992 mode
16994 The interlacing mode to adopt. It accepts one of the following
16995 values:
16996 0, send_frame
16998 Output one frame for each frame.
16999 1, send_field
17001 Output one frame for each field.
17002 The default value is "send_field".
17004 parity
17006 The picture field parity assumed for the input interlaced video. It
17007 accepts one of the following values:
17008 0, tff
17010 Assume the top field is first.
17011 1, bff
17013 Assume the bottom field is first.
17014 -1, auto
17016 Enable automatic detection of field parity.
17017 The default value is "auto". If the interlacing is unknown or the
17019 decoder does not export this information, top field first will be
17020 assumed.
17021 deint
17023 Specify which frames to deinterlace. Accepts one of the following
17024 values:
17025 0, all
17027 Deinterlace all frames.
17028 1, interlaced
17030 Only deinterlace frames marked as interlaced.
17031 The default value is "all".
17033 cas
17035 Apply Contrast Adaptive Sharpen filter to video stream.
17036 The filter accepts the following options:
17038 strength
17040 Set the sharpening strength. Default value is 0.
17041 planes
17043 Set planes to filter. Default value is to filter all planes except
17044 alpha plane.
17045 Commands
17047 This filter supports same commands as options.
17049 chromahold
17051 Remove all color information for all colors except for certain one.
17052 The filter accepts the following options:
17054 color
17056 The color which will not be replaced with neutral chroma.
17057 similarity
17059 Similarity percentage with the above color. 0.01 matches only the
17060 exact key color, while 1.0 matches everything.
17061 blend
17063 Blend percentage. 0.0 makes pixels either fully gray, or not gray
17064 at all. Higher values result in more preserved color.
17065 yuv Signals that the color passed is already in YUV instead of RGB.
17067 Literal colors like "green" or "red" don't make sense with this
17069 enabled anymore. This can be used to pass exact YUV values as
17070 hexadecimal numbers.
17071 Commands
17073 This filter supports same commands as options. The command accepts the
17075 same syntax of the corresponding option.
17076 If the specified expression is not valid, it is kept at its current
17078 value.
17079 chromakey
17081 YUV colorspace color/chroma keying.
17082 The filter accepts the following options:
17084 color
17086 The color which will be replaced with transparency.
17087 similarity
17089 Similarity percentage with the key color.
17090 0.01 matches only the exact key color, while 1.0 matches
17092 everything.
17093 blend
17095 Blend percentage.
17096 0.0 makes pixels either fully transparent, or not transparent at
17098 all.
17099 Higher values result in semi-transparent pixels, with a higher
17101 transparency the more similar the pixels color is to the key color.
17102 yuv Signals that the color passed is already in YUV instead of RGB.
17104 Literal colors like "green" or "red" don't make sense with this
17106 enabled anymore. This can be used to pass exact YUV values as
17107 hexadecimal numbers.
17108 Commands
17110 This filter supports same commands as options. The command accepts the
17112 same syntax of the corresponding option.
17113 If the specified expression is not valid, it is kept at its current
17115 value.
17116 Examples
17118 • Make every green pixel in the input image transparent:
17120 ffmpeg -i input.png -vf chromakey=green out.png
17122 • Overlay a greenscreen-video on top of a static black background.
17124 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
17126 chromakey_cuda
17128 CUDA accelerated YUV colorspace color/chroma keying.
17129 This filter works like normal chromakey filter but operates on CUDA
17131 frames. for more details and parameters see chromakey.
17132 Examples
17134 • Make all the green pixels in the input video transparent and use it
17136 as an overlay for another video:
17137 ./ffmpeg \
17139 -hwaccel cuda -hwaccel_output_format cuda -i input_green.mp4 \
17140 -hwaccel cuda -hwaccel_output_format cuda -i base_video.mp4 \
17141 -init_hw_device cuda \
17142 -filter_complex \
17143 " \
17144 [0:v]chromakey_cuda=0x25302D:0.1:0.12:1[overlay_video]; \
17145 [1:v]scale_cuda=format=yuv420p[base]; \
17146 [base][overlay_video]overlay_cuda" \
17147 -an -sn -c:v h264_nvenc -cq 20 output.mp4
17148 • Process two software sources, explicitly uploading the frames:
17150 ./ffmpeg -init_hw_device cuda=cuda -filter_hw_device cuda \
17152 -f lavfi -i color=size=800x600:color=white,format=yuv420p \
17153 -f lavfi -i yuvtestsrc=size=200x200,format=yuv420p \
17154 -filter_complex \
17155 " \
17156 [0]hwupload[under]; \
17157 [1]hwupload,chromakey_cuda=green:0.1:0.12[over]; \
17158 [under][over]overlay_cuda" \
17159 -c:v hevc_nvenc -cq 18 -preset slow output.mp4
17160 chromanr
17162 Reduce chrominance noise.
17163 The filter accepts the following options:
17165 thres
17167 Set threshold for averaging chrominance values. Sum of absolute
17168 difference of Y, U and V pixel components of current pixel and
17169 neighbour pixels lower than this threshold will be used in
17170 averaging. Luma component is left unchanged and is copied to
17171 output. Default value is 30. Allowed range is from 1 to 200.
17172 sizew
17174 Set horizontal radius of rectangle used for averaging. Allowed
17175 range is from 1 to 100. Default value is 5.
17176 sizeh
17178 Set vertical radius of rectangle used for averaging. Allowed range
17179 is from 1 to 100. Default value is 5.
17180 stepw
17182 Set horizontal step when averaging. Default value is 1. Allowed
17183 range is from 1 to 50. Mostly useful to speed-up filtering.
17184 steph
17186 Set vertical step when averaging. Default value is 1. Allowed
17187 range is from 1 to 50. Mostly useful to speed-up filtering.
17188 threy
17190 Set Y threshold for averaging chrominance values. Set finer
17191 control for max allowed difference between Y components of current
17192 pixel and neigbour pixels. Default value is 200. Allowed range is
17193 from 1 to 200.
17194 threu
17196 Set U threshold for averaging chrominance values. Set finer
17197 control for max allowed difference between U components of current
17198 pixel and neigbour pixels. Default value is 200. Allowed range is
17199 from 1 to 200.
17200 threv
17202 Set V threshold for averaging chrominance values. Set finer
17203 control for max allowed difference between V components of current
17204 pixel and neigbour pixels. Default value is 200. Allowed range is
17205 from 1 to 200.
17206 distance
17208 Set distance type used in calculations.
17209 manhattan
17211 Absolute difference.
17212 euclidean
17214 Difference squared.
17215 Default distance type is manhattan.
17217 Commands
17219 This filter supports same commands as options. The command accepts the
17221 same syntax of the corresponding option.
17222 chromashift
17224 Shift chroma pixels horizontally and/or vertically.
17225 The filter accepts the following options:
17227 cbh Set amount to shift chroma-blue horizontally.
17229 cbv Set amount to shift chroma-blue vertically.
17231 crh Set amount to shift chroma-red horizontally.
17233 crv Set amount to shift chroma-red vertically.
17235 edge
17237 Set edge mode, can be smear, default, or warp.
17238 Commands
17240 This filter supports the all above options as commands.
17242 ciescope
17244 Display CIE color diagram with pixels overlaid onto it.
17245 The filter accepts the following options:
17247 system
17249 Set color system.
17250 ntsc, 470m
17252 ebu, 470bg
17253 smpte
17254 240m
17255 apple
17256 widergb
17257 cie1931
17258 rec709, hdtv
17259 uhdtv, rec2020
17260 dcip3
17261 cie Set CIE system.
17262 xyy
17264 ucs
17265 luv
17266 gamuts
17267 Set what gamuts to draw.
17268 See "system" option for available values.
17270 size, s
17272 Set ciescope size, by default set to 512.
17273 intensity, i
17275 Set intensity used to map input pixel values to CIE diagram.
17276 contrast
17278 Set contrast used to draw tongue colors that are out of active
17279 color system gamut.
17280 corrgamma
17282 Correct gamma displayed on scope, by default enabled.
17283 showwhite
17285 Show white point on CIE diagram, by default disabled.
17286 gamma
17288 Set input gamma. Used only with XYZ input color space.
17289 fill
17291 Fill with CIE colors. By default is enabled.
17292 codecview
17294 Visualize information exported by some codecs.
17295 Some codecs can export information through frames using side-data or
17297 other means. For example, some MPEG based codecs export motion vectors
17298 through the export_mvs flag in the codec flags2 option.
17299 The filter accepts the following option:
17301 block
17303 Display block partition structure using the luma plane.
17304 mv Set motion vectors to visualize.
17306 Available flags for mv are:
17308 pf forward predicted MVs of P-frames
17310 bf forward predicted MVs of B-frames
17312 bb backward predicted MVs of B-frames
17314 qp Display quantization parameters using the chroma planes.
17316 mv_type, mvt
17318 Set motion vectors type to visualize. Includes MVs from all frames
17319 unless specified by frame_type option.
17320 Available flags for mv_type are:
17322 fp forward predicted MVs
17324 bp backward predicted MVs
17326 frame_type, ft
17328 Set frame type to visualize motion vectors of.
17329 Available flags for frame_type are:
17331 if intra-coded frames (I-frames)
17333 pf predicted frames (P-frames)
17335 bf bi-directionally predicted frames (B-frames)
17337 Examples
17339 • Visualize forward predicted MVs of all frames using ffplay:
17341 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
17343 • Visualize multi-directionals MVs of P and B-Frames using ffplay:
17345 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
17347 colorbalance
17349 Modify intensity of primary colors (red, green and blue) of input
17350 frames.
17351 The filter allows an input frame to be adjusted in the shadows,
17353 midtones or highlights regions for the red-cyan, green-magenta or blue-
17354 yellow balance.
17355 A positive adjustment value shifts the balance towards the primary
17357 color, a negative value towards the complementary color.
17358 The filter accepts the following options:
17360 rs
17362 gs
17363 bs Adjust red, green and blue shadows (darkest pixels).
17364 rm
17366 gm
17367 bm Adjust red, green and blue midtones (medium pixels).
17368 rh
17370 gh
17371 bh Adjust red, green and blue highlights (brightest pixels).
17372 Allowed ranges for options are "[-1.0, 1.0]". Defaults are 0.
17374 pl Preserve lightness when changing color balance. Default is
17376 disabled.
17377 Examples
17379 • Add red color cast to shadows:
17381 colorbalance=rs=.3
17383 Commands
17385 This filter supports the all above options as commands.
17387 colorcontrast
17389 Adjust color contrast between RGB components.
17390 The filter accepts the following options:
17392 rc Set the red-cyan contrast. Defaults is 0.0. Allowed range is from
17394 -1.0 to 1.0.
17395 gm Set the green-magenta contrast. Defaults is 0.0. Allowed range is
17397 from -1.0 to 1.0.
17398 by Set the blue-yellow contrast. Defaults is 0.0. Allowed range is
17400 from -1.0 to 1.0.
17401 rcw
17403 gmw
17404 byw Set the weight of each "rc", "gm", "by" option value. Default value
17405 is 0.0. Allowed range is from 0.0 to 1.0. If all weights are 0.0
17406 filtering is disabled.
17407 pl Set the amount of preserving lightness. Default value is 0.0.
17409 Allowed range is from 0.0 to 1.0.
17410 Commands
17412 This filter supports the all above options as commands.
17414 colorcorrect
17416 Adjust color white balance selectively for blacks and whites. This
17417 filter operates in YUV colorspace.
17418 The filter accepts the following options:
17420 rl Set the red shadow spot. Allowed range is from -1.0 to 1.0.
17422 Default value is 0.
17423 bl Set the blue shadow spot. Allowed range is from -1.0 to 1.0.
17425 Default value is 0.
17426 rh Set the red highlight spot. Allowed range is from -1.0 to 1.0.
17428 Default value is 0.
17429 bh Set the red highlight spot. Allowed range is from -1.0 to 1.0.
17431 Default value is 0.
17432 saturation
17434 Set the amount of saturation. Allowed range is from -3.0 to 3.0.
17435 Default value is 1.
17436 analyze
17438 If set to anything other than "manual" it will analyze every frame
17439 and use derived parameters for filtering output frame.
17440 Possible values are:
17442 manual
17444 average
17445 minmax
17446 median
17447 Default value is "manual".
17449 Commands
17451 This filter supports the all above options as commands.
17453 colorchannelmixer
17455 Adjust video input frames by re-mixing color channels.
17456 This filter modifies a color channel by adding the values associated to
17458 the other channels of the same pixels. For example if the value to
17459 modify is red, the output value will be:
17460 <red>=<red>*<rr> + <blue>*<rb> + <green>*<rg> + <alpha>*<ra>
17462 The filter accepts the following options:
17464 rr
17466 rg
17467 rb
17468 ra Adjust contribution of input red, green, blue and alpha channels
17469 for output red channel. Default is 1 for rr, and 0 for rg, rb and
17470 ra.
17471 gr
17473 gg
17474 gb
17475 ga Adjust contribution of input red, green, blue and alpha channels
17476 for output green channel. Default is 1 for gg, and 0 for gr, gb
17477 and ga.
17478 br
17480 bg
17481 bb
17482 ba Adjust contribution of input red, green, blue and alpha channels
17483 for output blue channel. Default is 1 for bb, and 0 for br, bg and
17484 ba.
17485 ar
17487 ag
17488 ab
17489 aa Adjust contribution of input red, green, blue and alpha channels
17490 for output alpha channel. Default is 1 for aa, and 0 for ar, ag
17491 and ab.
17492 Allowed ranges for options are "[-2.0, 2.0]".
17494 pc Set preserve color mode. The accepted values are:
17496 none
17498 Disable color preserving, this is default.
17499 lum Preserve luminance.
17501 max Preserve max value of RGB triplet.
17503 avg Preserve average value of RGB triplet.
17505 sum Preserve sum value of RGB triplet.
17507 nrm Preserve normalized value of RGB triplet.
17509 pwr Preserve power value of RGB triplet.
17511 pa Set the preserve color amount when changing colors. Allowed range
17513 is from "[0.0, 1.0]". Default is 0.0, thus disabled.
17514 Examples
17516 • Convert source to grayscale:
17518 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
17520 • Simulate sepia tones:
17522 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
17524 Commands
17526 This filter supports the all above options as commands.
17528 colorize
17530 Overlay a solid color on the video stream.
17531 The filter accepts the following options:
17533 hue Set the color hue. Allowed range is from 0 to 360. Default value
17535 is 0.
17536 saturation
17538 Set the color saturation. Allowed range is from 0 to 1. Default
17539 value is 0.5.
17540 lightness
17542 Set the color lightness. Allowed range is from 0 to 1. Default
17543 value is 0.5.
17544 mix Set the mix of source lightness. By default is set to 1.0. Allowed
17546 range is from 0.0 to 1.0.
17547 Commands
17549 This filter supports the all above options as commands.
17551 colorkey
17553 RGB colorspace color keying. This filter operates on 8-bit RGB format
17554 frames by setting the alpha component of each pixel which falls within
17555 the similarity radius of the key color to 0. The alpha value for pixels
17556 outside the similarity radius depends on the value of the blend option.
17557 The filter accepts the following options:
17559 color
17561 Set the color for which alpha will be set to 0 (full transparency).
17562 See "Color" section in the ffmpeg-utils manual. Default is
17563 "black".
17564 similarity
17566 Set the radius from the key color within which other colors also
17567 have full transparency. The computed distance is related to the
17568 unit fractional distance in 3D space between the RGB values of the
17569 key color and the pixel's color. Range is 0.01 to 1.0. 0.01 matches
17570 within a very small radius around the exact key color, while 1.0
17571 matches everything. Default is 0.01.
17572 blend
17574 Set how the alpha value for pixels that fall outside the similarity
17575 radius is computed. 0.0 makes pixels either fully transparent or
17576 fully opaque. Higher values result in semi-transparent pixels,
17577 with greater transparency the more similar the pixel color is to
17578 the key color. Range is 0.0 to 1.0. Default is 0.0.
17579 Examples
17581 • Make every green pixel in the input image transparent:
17583 ffmpeg -i input.png -vf colorkey=green out.png
17585 • Overlay a greenscreen-video on top of a static background image.
17587 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
17589 Commands
17591 This filter supports same commands as options. The command accepts the
17593 same syntax of the corresponding option.
17594 If the specified expression is not valid, it is kept at its current
17596 value.
17597 colorhold
17599 Remove all color information for all RGB colors except for certain one.
17600 The filter accepts the following options:
17602 color
17604 The color which will not be replaced with neutral gray.
17605 similarity
17607 Similarity percentage with the above color. 0.01 matches only the
17608 exact key color, while 1.0 matches everything.
17609 blend
17611 Blend percentage. 0.0 makes pixels fully gray. Higher values
17612 result in more preserved color.
17613 Commands
17615 This filter supports same commands as options. The command accepts the
17617 same syntax of the corresponding option.
17618 If the specified expression is not valid, it is kept at its current
17620 value.
17621 colorlevels
17623 Adjust video input frames using levels.
17624 The filter accepts the following options:
17626 rimin
17628 gimin
17629 bimin
17630 aimin
17631 Adjust red, green, blue and alpha input black point. Allowed
17632 ranges for options are "[-1.0, 1.0]". Defaults are 0.
17633 rimax
17635 gimax
17636 bimax
17637 aimax
17638 Adjust red, green, blue and alpha input white point. Allowed
17639 ranges for options are "[-1.0, 1.0]". Defaults are 1.
17640 Input levels are used to lighten highlights (bright tones), darken
17642 shadows (dark tones), change the balance of bright and dark tones.
17643 romin
17645 gomin
17646 bomin
17647 aomin
17648 Adjust red, green, blue and alpha output black point. Allowed
17649 ranges for options are "[0, 1.0]". Defaults are 0.
17650 romax
17652 gomax
17653 bomax
17654 aomax
17655 Adjust red, green, blue and alpha output white point. Allowed
17656 ranges for options are "[0, 1.0]". Defaults are 1.
17657 Output levels allows manual selection of a constrained output level
17659 range.
17660 preserve
17662 Set preserve color mode. The accepted values are:
17663 none
17665 Disable color preserving, this is default.
17666 lum Preserve luminance.
17668 max Preserve max value of RGB triplet.
17670 avg Preserve average value of RGB triplet.
17672 sum Preserve sum value of RGB triplet.
17674 nrm Preserve normalized value of RGB triplet.
17676 pwr Preserve power value of RGB triplet.
17678 Examples
17680 • Make video output darker:
17682 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
17684 • Increase contrast:
17686 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
17688 • Make video output lighter:
17690 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
17692 • Increase brightness:
17694 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
17696 Commands
17698 This filter supports the all above options as commands.
17700 colormap
17702 Apply custom color maps to video stream.
17703 This filter needs three input video streams. First stream is video
17705 stream that is going to be filtered out. Second and third video stream
17706 specify color patches for source color to target color mapping.
17707 The filter accepts the following options:
17709 patch_size
17711 Set the source and target video stream patch size in pixels.
17712 nb_patches
17714 Set the max number of used patches from source and target video
17715 stream. Default value is number of patches available in additional
17716 video streams. Max allowed number of patches is 64.
17717 type
17719 Set the adjustments used for target colors. Can be "relative" or
17720 "absolute". Defaults is "absolute".
17721 kernel
17723 Set the kernel used to measure color differences between mapped
17724 colors.
17725 The accepted values are:
17727 euclidean
17729 weuclidean
17730 Default is "euclidean".
17732 colormatrix
17734 Convert color matrix.
17735 The filter accepts the following options:
17737 src
17739 dst Specify the source and destination color matrix. Both values must
17740 be specified.
17741 The accepted values are:
17743 bt709
17745 BT.709
17746 fcc FCC
17748 bt601
17750 BT.601
17751 bt470
17753 BT.470
17754 bt470bg
17756 BT.470BG
17757 smpte170m
17759 SMPTE-170M
17760 smpte240m
17762 SMPTE-240M
17763 bt2020
17765 BT.2020
17766 For example to convert from BT.601 to SMPTE-240M, use the command:
17768 colormatrix=bt601:smpte240m
17770 colorspace
17772 Convert colorspace, transfer characteristics or color primaries. Input
17773 video needs to have an even size.
17774 The filter accepts the following options:
17776 all Specify all color properties at once.
17778 The accepted values are:
17780 bt470m
17782 BT.470M
17783 bt470bg
17785 BT.470BG
17786 bt601-6-525
17788 BT.601-6 525
17789 bt601-6-625
17791 BT.601-6 625
17792 bt709
17794 BT.709
17795 smpte170m
17797 SMPTE-170M
17798 smpte240m
17800 SMPTE-240M
17801 bt2020
17803 BT.2020
17804 space
17806 Specify output colorspace.
17807 The accepted values are:
17809 bt709
17811 BT.709
17812 fcc FCC
17814 bt470bg
17816 BT.470BG or BT.601-6 625
17817 smpte170m
17819 SMPTE-170M or BT.601-6 525
17820 smpte240m
17822 SMPTE-240M
17823 ycgco
17825 YCgCo
17826 bt2020ncl
17828 BT.2020 with non-constant luminance
17829 trc Specify output transfer characteristics.
17831 The accepted values are:
17833 bt709
17835 BT.709
17836 bt470m
17838 BT.470M
17839 bt470bg
17841 BT.470BG
17842 gamma22
17844 Constant gamma of 2.2
17845 gamma28
17847 Constant gamma of 2.8
17848 smpte170m
17850 SMPTE-170M, BT.601-6 625 or BT.601-6 525
17851 smpte240m
17853 SMPTE-240M
17854 srgb
17856 SRGB
17857 iec61966-2-1
17859 iec61966-2-1
17860 iec61966-2-4
17862 iec61966-2-4
17863 xvycc
17865 xvycc
17866 bt2020-10
17868 BT.2020 for 10-bits content
17869 bt2020-12
17871 BT.2020 for 12-bits content
17872 primaries
17874 Specify output color primaries.
17875 The accepted values are:
17877 bt709
17879 BT.709
17880 bt470m
17882 BT.470M
17883 bt470bg
17885 BT.470BG or BT.601-6 625
17886 smpte170m
17888 SMPTE-170M or BT.601-6 525
17889 smpte240m
17891 SMPTE-240M
17892 film
17894 film
17895 smpte431
17897 SMPTE-431
17898 smpte432
17900 SMPTE-432
17901 bt2020
17903 BT.2020
17904 jedec-p22
17906 JEDEC P22 phosphors
17907 range
17909 Specify output color range.
17910 The accepted values are:
17912 tv TV (restricted) range
17914 mpeg
17916 MPEG (restricted) range
17917 pc PC (full) range
17919 jpeg
17921 JPEG (full) range
17922 format
17924 Specify output color format.
17925 The accepted values are:
17927 yuv420p
17929 YUV 4:2:0 planar 8-bits
17930 yuv420p10
17932 YUV 4:2:0 planar 10-bits
17933 yuv420p12
17935 YUV 4:2:0 planar 12-bits
17936 yuv422p
17938 YUV 4:2:2 planar 8-bits
17939 yuv422p10
17941 YUV 4:2:2 planar 10-bits
17942 yuv422p12
17944 YUV 4:2:2 planar 12-bits
17945 yuv444p
17947 YUV 4:4:4 planar 8-bits
17948 yuv444p10
17950 YUV 4:4:4 planar 10-bits
17951 yuv444p12
17953 YUV 4:4:4 planar 12-bits
17954 fast
17956 Do a fast conversion, which skips gamma/primary correction. This
17957 will take significantly less CPU, but will be mathematically
17958 incorrect. To get output compatible with that produced by the
17959 colormatrix filter, use fast=1.
17960 dither
17962 Specify dithering mode.
17963 The accepted values are:
17965 none
17967 No dithering
17968 fsb Floyd-Steinberg dithering
17970 wpadapt
17972 Whitepoint adaptation mode.
17973 The accepted values are:
17975 bradford
17977 Bradford whitepoint adaptation
17978 vonkries
17980 von Kries whitepoint adaptation
17981 identity
17983 identity whitepoint adaptation (i.e. no whitepoint adaptation)
17984 iall
17986 Override all input properties at once. Same accepted values as all.
17987 ispace
17989 Override input colorspace. Same accepted values as space.
17990 iprimaries
17992 Override input color primaries. Same accepted values as primaries.
17993 itrc
17995 Override input transfer characteristics. Same accepted values as
17996 trc.
17997 irange
17999 Override input color range. Same accepted values as range.
18000 The filter converts the transfer characteristics, color space and color
18002 primaries to the specified user values. The output value, if not
18003 specified, is set to a default value based on the "all" property. If
18004 that property is also not specified, the filter will log an error. The
18005 output color range and format default to the same value as the input
18006 color range and format. The input transfer characteristics, color
18007 space, color primaries and color range should be set on the input data.
18008 If any of these are missing, the filter will log an error and no
18009 conversion will take place.
18010 For example to convert the input to SMPTE-240M, use the command:
18012 colorspace=smpte240m
18014 colorspace_cuda
18016 CUDA accelerated implementation of the colorspace filter.
18017 It is by no means feature complete compared to the software colorspace
18019 filter, and at the current time only supports color range conversion
18020 between jpeg/full and mpeg/limited range.
18021 The filter accepts the following options:
18023 range
18025 Specify output color range.
18026 The accepted values are:
18028 tv TV (restricted) range
18030 mpeg
18032 MPEG (restricted) range
18033 pc PC (full) range
18035 jpeg
18037 JPEG (full) range
18038 colortemperature
18040 Adjust color temperature in video to simulate variations in ambient
18041 color temperature.
18042 The filter accepts the following options:
18044 temperature
18046 Set the temperature in Kelvin. Allowed range is from 1000 to 40000.
18047 Default value is 6500 K.
18048 mix Set mixing with filtered output. Allowed range is from 0 to 1.
18050 Default value is 1.
18051 pl Set the amount of preserving lightness. Allowed range is from 0 to
18053 1. Default value is 0.
18054 Commands
18056 This filter supports same commands as options.
18058 convolution
18060 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49
18061 elements.
18062 The filter accepts the following options:
18064 0m
18066 1m
18067 2m
18068 3m Set matrix for each plane. Matrix is sequence of 9, 25 or 49
18069 signed integers in square mode, and from 1 to 49 odd number of
18070 signed integers in row mode.
18071 0rdiv
18073 1rdiv
18074 2rdiv
18075 3rdiv
18076 Set multiplier for calculated value for each plane. If unset or 0,
18077 it will be sum of all matrix elements.
18078 0bias
18080 1bias
18081 2bias
18082 3bias
18083 Set bias for each plane. This value is added to the result of the
18084 multiplication. Useful for making the overall image brighter or
18085 darker. Default is 0.0.
18086 0mode
18088 1mode
18089 2mode
18090 3mode
18091 Set matrix mode for each plane. Can be square, row or column.
18092 Default is square.
18093 Commands
18095 This filter supports the all above options as commands.
18097 Examples
18099 • Apply sharpen:
18101 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"
18103 • Apply blur:
18105 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"
18107 • Apply edge enhance:
18109 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"
18111 • Apply edge detect:
18113 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"
18115 • Apply laplacian edge detector which includes diagonals:
18117 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"
18119 • Apply emboss:
18121 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"
18123 convolve
18125 Apply 2D convolution of video stream in frequency domain using second
18126 stream as impulse.
18127 The filter accepts the following options:
18129 planes
18131 Set which planes to process.
18132 impulse
18134 Set which impulse video frames will be processed, can be first or
18135 all. Default is all.
18136 The "convolve" filter also supports the framesync options.
18138 copy
18140 Copy the input video source unchanged to the output. This is mainly
18141 useful for testing purposes.
18142 coreimage
18144 Video filtering on GPU using Apple's CoreImage API on OSX.
18145 Hardware acceleration is based on an OpenGL context. Usually, this
18147 means it is processed by video hardware. However, software-based OpenGL
18148 implementations exist which means there is no guarantee for hardware
18149 processing. It depends on the respective OSX.
18150 There are many filters and image generators provided by Apple that come
18152 with a large variety of options. The filter has to be referenced by its
18153 name along with its options.
18154 The coreimage filter accepts the following options:
18156 list_filters
18158 List all available filters and generators along with all their
18159 respective options as well as possible minimum and maximum values
18160 along with the default values.
18161 list_filters=true
18163 filter
18165 Specify all filters by their respective name and options. Use
18166 list_filters to determine all valid filter names and options.
18167 Numerical options are specified by a float value and are
18168 automatically clamped to their respective value range. Vector and
18169 color options have to be specified by a list of space separated
18170 float values. Character escaping has to be done. A special option
18171 name "default" is available to use default options for a filter.
18172 It is required to specify either "default" or at least one of the
18174 filter options. All omitted options are used with their default
18175 values. The syntax of the filter string is as follows:
18176 filter=<NAME>@<OPTION>=<VALUE>[@<OPTION>=<VALUE>][@...][#<NAME>@<OPTION>=<VALUE>[@<OPTION>=<VALUE>][@...]][#...]
18178 output_rect
18180 Specify a rectangle where the output of the filter chain is copied
18181 into the input image. It is given by a list of space separated
18182 float values:
18183 output_rect=x\ y\ width\ height
18185 If not given, the output rectangle equals the dimensions of the
18187 input image. The output rectangle is automatically cropped at the
18188 borders of the input image. Negative values are valid for each
18189 component.
18190 output_rect=25\ 25\ 100\ 100
18192 Several filters can be chained for successive processing without GPU-
18194 HOST transfers allowing for fast processing of complex filter chains.
18195 Currently, only filters with zero (generators) or exactly one (filters)
18196 input image and one output image are supported. Also, transition
18197 filters are not yet usable as intended.
18198 Some filters generate output images with additional padding depending
18200 on the respective filter kernel. The padding is automatically removed
18201 to ensure the filter output has the same size as the input image.
18202 For image generators, the size of the output image is determined by the
18204 previous output image of the filter chain or the input image of the
18205 whole filterchain, respectively. The generators do not use the pixel
18206 information of this image to generate their output. However, the
18207 generated output is blended onto this image, resulting in partial or
18208 complete coverage of the output image.
18209 The coreimagesrc video source can be used for generating input images
18211 which are directly fed into the filter chain. By using it, providing
18212 input images by another video source or an input video is not required.
18213 Examples
18215 • List all filters available:
18217 coreimage=list_filters=true
18219 • Use the CIBoxBlur filter with default options to blur an image:
18221 coreimage=filter=CIBoxBlur@default
18223 • Use a filter chain with CISepiaTone at default values and
18225 CIVignetteEffect with its center at 100x100 and a radius of 50
18226 pixels:
18227 coreimage=filter=CIBoxBlur@default#CIVignetteEffect@inputCenter=100\ 100@inputRadius=50
18229 • Use nullsrc and CIQRCodeGenerator to create a QR code for the
18231 FFmpeg homepage, given as complete and escaped command-line for
18232 Apple's standard bash shell:
18233 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@inputMessage=https\\\\\://FFmpeg.org/@inputCorrectionLevel=H -frames:v 1 QRCode.png
18235 corr
18237 Obtain the correlation between two input videos.
18238 This filter takes two input videos.
18240 Both input videos must have the same resolution and pixel format for
18242 this filter to work correctly. Also it assumes that both inputs have
18243 the same number of frames, which are compared one by one.
18244 The obtained per component, average, min and max correlation is printed
18246 through the logging system.
18247 The filter stores the calculated correlation of each frame in frame
18249 metadata.
18250 This filter also supports the framesync options.
18252 In the below example the input file main.mpg being processed is
18254 compared with the reference file ref.mpg.
18255 ffmpeg -i main.mpg -i ref.mpg -lavfi corr -f null -
18257 cover_rect
18259 Cover a rectangular object
18260 It accepts the following options:
18262 cover
18264 Filepath of the optional cover image, needs to be in yuv420.
18265 mode
18267 Set covering mode.
18268 It accepts the following values:
18270 cover
18272 cover it by the supplied image
18273 blur
18275 cover it by interpolating the surrounding pixels
18276 Default value is blur.
18278 Examples
18280 • Cover a rectangular object by the supplied image of a given video
18282 using ffmpeg:
18283 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
18285 crop
18287 Crop the input video to given dimensions.
18288 It accepts the following parameters:
18290 w, out_w
18292 The width of the output video. It defaults to "iw". This
18293 expression is evaluated only once during the filter configuration,
18294 or when the w or out_w command is sent.
18295 h, out_h
18297 The height of the output video. It defaults to "ih". This
18298 expression is evaluated only once during the filter configuration,
18299 or when the h or out_h command is sent.
18300 x The horizontal position, in the input video, of the left edge of
18302 the output video. It defaults to "(in_w-out_w)/2". This expression
18303 is evaluated per-frame.
18304 y The vertical position, in the input video, of the top edge of the
18306 output video. It defaults to "(in_h-out_h)/2". This expression is
18307 evaluated per-frame.
18308 keep_aspect
18310 If set to 1 will force the output display aspect ratio to be the
18311 same of the input, by changing the output sample aspect ratio. It
18312 defaults to 0.
18313 exact
18315 Enable exact cropping. If enabled, subsampled videos will be
18316 cropped at exact width/height/x/y as specified and will not be
18317 rounded to nearest smaller value. It defaults to 0.
18318 The out_w, out_h, x, y parameters are expressions containing the
18320 following constants:
18321 x
18323 y The computed values for x and y. They are evaluated for each new
18324 frame.
18325 in_w
18327 in_h
18328 The input width and height.
18329 iw
18331 ih These are the same as in_w and in_h.
18332 out_w
18334 out_h
18335 The output (cropped) width and height.
18336 ow
18338 oh These are the same as out_w and out_h.
18339 a same as iw / ih
18341 sar input sample aspect ratio
18343 dar input display aspect ratio, it is the same as (iw / ih) * sar
18345 hsub
18347 vsub
18348 horizontal and vertical chroma subsample values. For example for
18349 the pixel format "yuv422p" hsub is 2 and vsub is 1.
18350 n The number of the input frame, starting from 0.
18352 pos the position in the file of the input frame, NAN if unknown
18354 t The timestamp expressed in seconds. It's NAN if the input timestamp
18356 is unknown.
18357 The expression for out_w may depend on the value of out_h, and the
18359 expression for out_h may depend on out_w, but they cannot depend on x
18360 and y, as x and y are evaluated after out_w and out_h.
18361 The x and y parameters specify the expressions for the position of the
18363 top-left corner of the output (non-cropped) area. They are evaluated
18364 for each frame. If the evaluated value is not valid, it is approximated
18365 to the nearest valid value.
18366 The expression for x may depend on y, and the expression for y may
18368 depend on x.
18369 Examples
18371 • Crop area with size 100x100 at position (12,34).
18373 crop=100:100:12:34
18375 Using named options, the example above becomes:
18377 crop=w=100:h=100:x=12:y=34
18379 • Crop the central input area with size 100x100:
18381 crop=100:100
18383 • Crop the central input area with size 2/3 of the input video:
18385 crop=2/3*in_w:2/3*in_h
18387 • Crop the input video central square:
18389 crop=out_w=in_h
18391 crop=in_h
18392 • Delimit the rectangle with the top-left corner placed at position
18394 100:100 and the right-bottom corner corresponding to the right-
18395 bottom corner of the input image.
18396 crop=in_w-100:in_h-100:100:100
18398 • Crop 10 pixels from the left and right borders, and 20 pixels from
18400 the top and bottom borders
18401 crop=in_w-2*10:in_h-2*20
18403 • Keep only the bottom right quarter of the input image:
18405 crop=in_w/2:in_h/2:in_w/2:in_h/2
18407 • Crop height for getting Greek harmony:
18409 crop=in_w:1/PHI*in_w
18411 • Apply trembling effect:
18413 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)
18415 • Apply erratic camera effect depending on timestamp:
18417 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)"
18419 • Set x depending on the value of y:
18421 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
18423 Commands
18425 This filter supports the following commands:
18427 w, out_w
18429 h, out_h
18430 x
18431 y Set width/height of the output video and the horizontal/vertical
18432 position in the input video. The command accepts the same syntax
18433 of the corresponding option.
18434 If the specified expression is not valid, it is kept at its current
18436 value.
18437 cropdetect
18439 Auto-detect the crop size.
18440 It calculates the necessary cropping parameters and prints the
18442 recommended parameters via the logging system. The detected dimensions
18443 correspond to the non-black or video area of the input video according
18444 to mode.
18445 It accepts the following parameters:
18447 mode
18449 Depending on mode crop detection is based on either the mere black
18450 value of surrounding pixels or a combination of motion vectors and
18451 edge pixels.
18452 black
18454 Detect black pixels surrounding the playing video. For fine
18455 control use option limit.
18456 mvedges
18458 Detect the playing video by the motion vectors inside the video
18459 and scanning for edge pixels typically forming the border of a
18460 playing video.
18461 limit
18463 Set higher black value threshold, which can be optionally specified
18464 from nothing (0) to everything (255 for 8-bit based formats). An
18465 intensity value greater to the set value is considered non-black.
18466 It defaults to 24. You can also specify a value between 0.0 and
18467 1.0 which will be scaled depending on the bitdepth of the pixel
18468 format.
18469 round
18471 The value which the width/height should be divisible by. It
18472 defaults to 16. The offset is automatically adjusted to center the
18473 video. Use 2 to get only even dimensions (needed for 4:2:2 video).
18474 16 is best when encoding to most video codecs.
18475 skip
18477 Set the number of initial frames for which evaluation is skipped.
18478 Default is 2. Range is 0 to INT_MAX.
18479 reset_count, reset
18481 Set the counter that determines after how many frames cropdetect
18482 will reset the previously detected largest video area and start
18483 over to detect the current optimal crop area. Default value is 0.
18484 This can be useful when channel logos distort the video area. 0
18486 indicates 'never reset', and returns the largest area encountered
18487 during playback.
18488 mv_threshold
18490 Set motion in pixel units as threshold for motion detection. It
18491 defaults to 8.
18492 low
18494 high
18495 Set low and high threshold values used by the Canny thresholding
18496 algorithm.
18497 The high threshold selects the "strong" edge pixels, which are then
18499 connected through 8-connectivity with the "weak" edge pixels
18500 selected by the low threshold.
18501 low and high threshold values must be chosen in the range [0,1],
18503 and low should be lesser or equal to high.
18504 Default value for low is "5/255", and default value for high is
18506 "15/255".
18507 Examples
18509 • Find video area surrounded by black borders:
18511 ffmpeg -i file.mp4 -vf cropdetect,metadata=mode=print -f null -
18513 • Find an embedded video area, generate motion vectors beforehand:
18515 ffmpeg -i file.mp4 -vf mestimate,cropdetect=mode=mvedges,metadata=mode=print -f null -
18517 • Find an embedded video area, use motion vectors from decoder:
18519 ffmpeg -flags2 +export_mvs -i file.mp4 -vf cropdetect=mode=mvedges,metadata=mode=print -f null -
18521 Commands
18523 This filter supports the following commands:
18525 limit
18527 The command accepts the same syntax of the corresponding option.
18528 If the specified expression is not valid, it is kept at its current
18529 value.
18530 cue
18532 Delay video filtering until a given wallclock timestamp. The filter
18533 first passes on preroll amount of frames, then it buffers at most
18534 buffer amount of frames and waits for the cue. After reaching the cue
18535 it forwards the buffered frames and also any subsequent frames coming
18536 in its input.
18537 The filter can be used synchronize the output of multiple ffmpeg
18539 processes for realtime output devices like decklink. By putting the
18540 delay in the filtering chain and pre-buffering frames the process can
18541 pass on data to output almost immediately after the target wallclock
18542 timestamp is reached.
18543 Perfect frame accuracy cannot be guaranteed, but the result is good
18545 enough for some use cases.
18546 cue The cue timestamp expressed in a UNIX timestamp in microseconds.
18548 Default is 0.
18549 preroll
18551 The duration of content to pass on as preroll expressed in seconds.
18552 Default is 0.
18553 buffer
18555 The maximum duration of content to buffer before waiting for the
18556 cue expressed in seconds. Default is 0.
18557 curves
18559 Apply color adjustments using curves.
18560 This filter is similar to the Adobe Photoshop and GIMP curves tools.
18562 Each component (red, green and blue) has its values defined by N key
18563 points tied from each other using a smooth curve. The x-axis represents
18564 the pixel values from the input frame, and the y-axis the new pixel
18565 values to be set for the output frame.
18566 By default, a component curve is defined by the two points (0;0) and
18568 (1;1). This creates a straight line where each original pixel value is
18569 "adjusted" to its own value, which means no change to the image.
18570 The filter allows you to redefine these two points and add some more. A
18572 new curve will be define to pass smoothly through all these new
18573 coordinates. The new defined points needs to be strictly increasing
18574 over the x-axis, and their x and y values must be in the [0;1]
18575 interval. The curve is formed by using a natural or monotonic cubic
18576 spline interpolation, depending on the interp option (default:
18577 "natural"). The "natural" spline produces a smoother curve in general
18578 while the monotonic ("pchip") spline guarantees the transitions between
18579 the specified points to be monotonic. If the computed curves happened
18580 to go outside the vector spaces, the values will be clipped
18581 accordingly.
18582 The filter accepts the following options:
18584 preset
18586 Select one of the available color presets. This option can be used
18587 in addition to the r, g, b parameters; in this case, the later
18588 options takes priority on the preset values. Available presets
18589 are:
18590 none
18592 color_negative
18593 cross_process
18594 darker
18595 increase_contrast
18596 lighter
18597 linear_contrast
18598 medium_contrast
18599 negative
18600 strong_contrast
18601 vintage
18602 Default is "none".
18604 master, m
18606 Set the master key points. These points will define a second pass
18607 mapping. It is sometimes called a "luminance" or "value" mapping.
18608 It can be used with r, g, b or all since it acts like a post-
18609 processing LUT.
18610 red, r
18612 Set the key points for the red component.
18613 green, g
18615 Set the key points for the green component.
18616 blue, b
18618 Set the key points for the blue component.
18619 all Set the key points for all components (not including master). Can
18621 be used in addition to the other key points component options. In
18622 this case, the unset component(s) will fallback on this all
18623 setting.
18624 psfile
18626 Specify a Photoshop curves file (".acv") to import the settings
18627 from.
18628 plot
18630 Save Gnuplot script of the curves in specified file.
18631 interp
18633 Specify the kind of interpolation. Available algorithms are:
18634 natural
18636 Natural cubic spline using a piece-wise cubic polynomial that
18637 is twice continuously differentiable.
18638 pchip
18640 Monotonic cubic spline using a piecewise cubic Hermite
18641 interpolating polynomial (PCHIP).
18642 To avoid some filtergraph syntax conflicts, each key points list need
18644 to be defined using the following syntax: "x0/y0 x1/y1 x2/y2 ...".
18645 Commands
18647 This filter supports same commands as options.
18649 Examples
18651 • Increase slightly the middle level of blue:
18653 curves=blue='0/0 0.5/0.58 1/1'
18655 • Vintage effect:
18657 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'
18659 Here we obtain the following coordinates for each components:
18661 red "(0;0.11) (0.42;0.51) (1;0.95)"
18663 green
18665 "(0;0) (0.50;0.48) (1;1)"
18666 blue
18668 "(0;0.22) (0.49;0.44) (1;0.80)"
18669 • The previous example can also be achieved with the associated
18671 built-in preset:
18672 curves=preset=vintage
18674 • Or simply:
18676 curves=vintage
18678 • Use a Photoshop preset and redefine the points of the green
18680 component:
18681 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
18683 • Check out the curves of the "cross_process" profile using ffmpeg
18685 and gnuplot:
18686 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
18688 gnuplot -p /tmp/curves.plt
18689 datascope
18691 Video data analysis filter.
18692 This filter shows hexadecimal pixel values of part of video.
18694 The filter accepts the following options:
18696 size, s
18698 Set output video size.
18699 x Set x offset from where to pick pixels.
18701 y Set y offset from where to pick pixels.
18703 mode
18705 Set scope mode, can be one of the following:
18706 mono
18708 Draw hexadecimal pixel values with white color on black
18709 background.
18710 color
18712 Draw hexadecimal pixel values with input video pixel color on
18713 black background.
18714 color2
18716 Draw hexadecimal pixel values on color background picked from
18717 input video, the text color is picked in such way so its always
18718 visible.
18719 axis
18721 Draw rows and columns numbers on left and top of video.
18722 opacity
18724 Set background opacity.
18725 format
18727 Set display number format. Can be "hex", or "dec". Default is
18728 "hex".
18729 components
18731 Set pixel components to display. By default all pixel components
18732 are displayed.
18733 Commands
18735 This filter supports same commands as options excluding "size" option.
18737 dblur
18739 Apply Directional blur filter.
18740 The filter accepts the following options:
18742 angle
18744 Set angle of directional blur. Default is 45.
18745 radius
18747 Set radius of directional blur. Default is 5.
18748 planes
18750 Set which planes to filter. By default all planes are filtered.
18751 Commands
18753 This filter supports same commands as options. The command accepts the
18755 same syntax of the corresponding option.
18756 If the specified expression is not valid, it is kept at its current
18758 value.
18759 dctdnoiz
18761 Denoise frames using 2D DCT (frequency domain filtering).
18762 This filter is not designed for real time.
18764 The filter accepts the following options:
18766 sigma, s
18768 Set the noise sigma constant.
18769 This sigma defines a hard threshold of "3 * sigma"; every DCT
18771 coefficient (absolute value) below this threshold with be dropped.
18772 If you need a more advanced filtering, see expr.
18774 Default is 0.
18776 overlap
18778 Set number overlapping pixels for each block. Since the filter can
18779 be slow, you may want to reduce this value, at the cost of a less
18780 effective filter and the risk of various artefacts.
18781 If the overlapping value doesn't permit processing the whole input
18783 width or height, a warning will be displayed and according borders
18784 won't be denoised.
18785 Default value is blocksize-1, which is the best possible setting.
18787 expr, e
18789 Set the coefficient factor expression.
18790 For each coefficient of a DCT block, this expression will be
18792 evaluated as a multiplier value for the coefficient.
18793 If this is option is set, the sigma option will be ignored.
18795 The absolute value of the coefficient can be accessed through the c
18797 variable.
18798 n Set the blocksize using the number of bits. "1<<n" defines the
18800 blocksize, which is the width and height of the processed blocks.
18801 The default value is 3 (8x8) and can be raised to 4 for a blocksize
18803 of 16x16. Note that changing this setting has huge consequences on
18804 the speed processing. Also, a larger block size does not
18805 necessarily means a better de-noising.
18806 Examples
18808 Apply a denoise with a sigma of 4.5:
18810 dctdnoiz=4.5
18812 The same operation can be achieved using the expression system:
18814 dctdnoiz=e='gte(c, 4.5*3)'
18816 Violent denoise using a block size of "16x16":
18818 dctdnoiz=15:n=4
18820 deband
18822 Remove banding artifacts from input video. It works by replacing
18823 banded pixels with average value of referenced pixels.
18824 The filter accepts the following options:
18826 1thr
18828 2thr
18829 3thr
18830 4thr
18831 Set banding detection threshold for each plane. Default is 0.02.
18832 Valid range is 0.00003 to 0.5. If difference between current pixel
18833 and reference pixel is less than threshold, it will be considered
18834 as banded.
18835 range, r
18837 Banding detection range in pixels. Default is 16. If positive,
18838 random number in range 0 to set value will be used. If negative,
18839 exact absolute value will be used. The range defines square of
18840 four pixels around current pixel.
18841 direction, d
18843 Set direction in radians from which four pixel will be compared. If
18844 positive, random direction from 0 to set direction will be picked.
18845 If negative, exact of absolute value will be picked. For example
18846 direction 0, -PI or -2*PI radians will pick only pixels on same row
18847 and -PI/2 will pick only pixels on same column.
18848 blur, b
18850 If enabled, current pixel is compared with average value of all
18851 four surrounding pixels. The default is enabled. If disabled
18852 current pixel is compared with all four surrounding pixels. The
18853 pixel is considered banded if only all four differences with
18854 surrounding pixels are less than threshold.
18855 coupling, c
18857 If enabled, current pixel is changed if and only if all pixel
18858 components are banded, e.g. banding detection threshold is
18859 triggered for all color components. The default is disabled.
18860 Commands
18862 This filter supports the all above options as commands.
18864 deblock
18866 Remove blocking artifacts from input video.
18867 The filter accepts the following options:
18869 filter
18871 Set filter type, can be weak or strong. Default is strong. This
18872 controls what kind of deblocking is applied.
18873 block
18875 Set size of block, allowed range is from 4 to 512. Default is 8.
18876 alpha
18878 beta
18879 gamma
18880 delta
18881 Set blocking detection thresholds. Allowed range is 0 to 1.
18882 Defaults are: 0.098 for alpha and 0.05 for the rest. Using higher
18883 threshold gives more deblocking strength. Setting alpha controls
18884 threshold detection at exact edge of block. Remaining options
18885 controls threshold detection near the edge. Each one for
18886 below/above or left/right. Setting any of those to 0 disables
18887 deblocking.
18888 planes
18890 Set planes to filter. Default is to filter all available planes.
18891 Examples
18893 • Deblock using weak filter and block size of 4 pixels.
18895 deblock=filter=weak:block=4
18897 • Deblock using strong filter, block size of 4 pixels and custom
18899 thresholds for deblocking more edges.
18900 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
18902 • Similar as above, but filter only first plane.
18904 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
18906 • Similar as above, but filter only second and third plane.
18908 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
18910 Commands
18912 This filter supports the all above options as commands.
18914 decimate
18916 Drop duplicated frames at regular intervals.
18917 The filter accepts the following options:
18919 cycle
18921 Set the number of frames from which one will be dropped. Setting
18922 this to N means one frame in every batch of N frames will be
18923 dropped. Default is 5.
18924 dupthresh
18926 Set the threshold for duplicate detection. If the difference metric
18927 for a frame is less than or equal to this value, then it is
18928 declared as duplicate. Default is 1.1
18929 scthresh
18931 Set scene change threshold. Default is 15.
18932 blockx
18934 blocky
18935 Set the size of the x and y-axis blocks used during metric
18936 calculations. Larger blocks give better noise suppression, but
18937 also give worse detection of small movements. Must be a power of
18938 two. Default is 32.
18939 ppsrc
18941 Mark main input as a pre-processed input and activate clean source
18942 input stream. This allows the input to be pre-processed with
18943 various filters to help the metrics calculation while keeping the
18944 frame selection lossless. When set to 1, the first stream is for
18945 the pre-processed input, and the second stream is the clean source
18946 from where the kept frames are chosen. Default is 0.
18947 chroma
18949 Set whether or not chroma is considered in the metric calculations.
18950 Default is 1.
18951 mixed
18953 Set whether or not the input only partially contains content to be
18954 decimated. Default is "false". If enabled video output stream
18955 will be in variable frame rate.
18956 deconvolve
18958 Apply 2D deconvolution of video stream in frequency domain using second
18959 stream as impulse.
18960 The filter accepts the following options:
18962 planes
18964 Set which planes to process.
18965 impulse
18967 Set which impulse video frames will be processed, can be first or
18968 all. Default is all.
18969 noise
18971 Set noise when doing divisions. Default is 0.0000001. Useful when
18972 width and height are not same and not power of 2 or if stream prior
18973 to convolving had noise.
18974 The "deconvolve" filter also supports the framesync options.
18976 dedot
18978 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from
18979 video.
18980 It accepts the following options:
18982 m Set mode of operation. Can be combination of dotcrawl for cross-
18984 luminance reduction and/or rainbows for cross-color reduction.
18985 lt Set spatial luma threshold. Lower values increases reduction of
18987 cross-luminance.
18988 tl Set tolerance for temporal luma. Higher values increases reduction
18990 of cross-luminance.
18991 tc Set tolerance for chroma temporal variation. Higher values
18993 increases reduction of cross-color.
18994 ct Set temporal chroma threshold. Lower values increases reduction of
18996 cross-color.
18997 deflate
18999 Apply deflate effect to the video.
19000 This filter replaces the pixel by the local(3x3) average by taking into
19002 account only values lower than the pixel.
19003 It accepts the following options:
19005 threshold0
19007 threshold1
19008 threshold2
19009 threshold3
19010 Limit the maximum change for each plane, default is 65535. If 0,
19011 plane will remain unchanged.
19012 Commands
19014 This filter supports the all above options as commands.
19016 deflicker
19018 Remove temporal frame luminance variations.
19019 It accepts the following options:
19021 size, s
19023 Set moving-average filter size in frames. Default is 5. Allowed
19024 range is 2 - 129.
19025 mode, m
19027 Set averaging mode to smooth temporal luminance variations.
19028 Available values are:
19030 am Arithmetic mean
19032 gm Geometric mean
19034 hm Harmonic mean
19036 qm Quadratic mean
19038 cm Cubic mean
19040 pm Power mean
19042 median
19044 Median
19045 bypass
19047 Do not actually modify frame. Useful when one only wants metadata.
19048 dejudder
19050 Remove judder produced by partially interlaced telecined content.
19051 Judder can be introduced, for instance, by pullup filter. If the
19053 original source was partially telecined content then the output of
19054 "pullup,dejudder" will have a variable frame rate. May change the
19055 recorded frame rate of the container. Aside from that change, this
19056 filter will not affect constant frame rate video.
19057 The option available in this filter is:
19059 cycle
19061 Specify the length of the window over which the judder repeats.
19062 Accepts any integer greater than 1. Useful values are:
19064 4 If the original was telecined from 24 to 30 fps (Film to NTSC).
19066 5 If the original was telecined from 25 to 30 fps (PAL to NTSC).
19068 20 If a mixture of the two.
19070 The default is 4.
19072 delogo
19074 Suppress a TV station logo by a simple interpolation of the surrounding
19075 pixels. Just set a rectangle covering the logo and watch it disappear
19076 (and sometimes something even uglier appear - your mileage may vary).
19077 It accepts the following parameters:
19079 x
19081 y Specify the top left corner coordinates of the logo. They must be
19082 specified.
19083 w
19085 h Specify the width and height of the logo to clear. They must be
19086 specified.
19087 show
19089 When set to 1, a green rectangle is drawn on the screen to simplify
19090 finding the right x, y, w, and h parameters. The default value is
19091 0.
19092 The rectangle is drawn on the outermost pixels which will be
19094 (partly) replaced with interpolated values. The values of the next
19095 pixels immediately outside this rectangle in each direction will be
19096 used to compute the interpolated pixel values inside the rectangle.
19097 Examples
19099 • Set a rectangle covering the area with top left corner coordinates
19101 0,0 and size 100x77:
19102 delogo=x=0:y=0:w=100:h=77
19104 derain
19106 Remove the rain in the input image/video by applying the derain methods
19107 based on convolutional neural networks. Supported models:
19108 • Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
19110 See
19111 <http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf>.
19112 Training as well as model generation scripts are provided in the
19114 repository at <https://github.com/XueweiMeng/derain_filter.git>.
19115 Native model files (.model) can be generated from TensorFlow model
19117 files (.pb) by using tools/python/convert.py
19118 The filter accepts the following options:
19120 filter_type
19122 Specify which filter to use. This option accepts the following
19123 values:
19124 derain
19126 Derain filter. To conduct derain filter, you need to use a
19127 derain model.
19128 dehaze
19130 Dehaze filter. To conduct dehaze filter, you need to use a
19131 dehaze model.
19132 Default value is derain.
19134 dnn_backend
19136 Specify which DNN backend to use for model loading and execution.
19137 This option accepts the following values:
19138 native
19140 Native implementation of DNN loading and execution.
19141 tensorflow
19143 TensorFlow backend. To enable this backend you need to install
19144 the TensorFlow for C library (see
19145 <https://www.tensorflow.org/install/lang_c>) and configure
19146 FFmpeg with "--enable-libtensorflow"
19147 Default value is native.
19149 model
19151 Set path to model file specifying network architecture and its
19152 parameters. Note that different backends use different file
19153 formats. TensorFlow and native backend can load files for only its
19154 format.
19155 To get full functionality (such as async execution), please use the
19157 dnn_processing filter.
19158 deshake
19160 Attempt to fix small changes in horizontal and/or vertical shift. This
19161 filter helps remove camera shake from hand-holding a camera, bumping a
19162 tripod, moving on a vehicle, etc.
19163 The filter accepts the following options:
19165 x
19167 y
19168 w
19169 h Specify a rectangular area where to limit the search for motion
19170 vectors. If desired the search for motion vectors can be limited
19171 to a rectangular area of the frame defined by its top left corner,
19172 width and height. These parameters have the same meaning as the
19173 drawbox filter which can be used to visualise the position of the
19174 bounding box.
19175 This is useful when simultaneous movement of subjects within the
19177 frame might be confused for camera motion by the motion vector
19178 search.
19179 If any or all of x, y, w and h are set to -1 then the full frame is
19181 used. This allows later options to be set without specifying the
19182 bounding box for the motion vector search.
19183 Default - search the whole frame.
19185 rx
19187 ry Specify the maximum extent of movement in x and y directions in the
19188 range 0-64 pixels. Default 16.
19189 edge
19191 Specify how to generate pixels to fill blanks at the edge of the
19192 frame. Available values are:
19193 blank, 0
19195 Fill zeroes at blank locations
19196 original, 1
19198 Original image at blank locations
19199 clamp, 2
19201 Extruded edge value at blank locations
19202 mirror, 3
19204 Mirrored edge at blank locations
19205 Default value is mirror.
19207 blocksize
19209 Specify the blocksize to use for motion search. Range 4-128 pixels,
19210 default 8.
19211 contrast
19213 Specify the contrast threshold for blocks. Only blocks with more
19214 than the specified contrast (difference between darkest and
19215 lightest pixels) will be considered. Range 1-255, default 125.
19216 search
19218 Specify the search strategy. Available values are:
19219 exhaustive, 0
19221 Set exhaustive search
19222 less, 1
19224 Set less exhaustive search.
19225 Default value is exhaustive.
19227 filename
19229 If set then a detailed log of the motion search is written to the
19230 specified file.
19231 despill
19233 Remove unwanted contamination of foreground colors, caused by reflected
19234 color of greenscreen or bluescreen.
19235 This filter accepts the following options:
19237 type
19239 Set what type of despill to use.
19240 mix Set how spillmap will be generated.
19242 expand
19244 Set how much to get rid of still remaining spill.
19245 red Controls amount of red in spill area.
19247 green
19249 Controls amount of green in spill area. Should be -1 for
19250 greenscreen.
19251 blue
19253 Controls amount of blue in spill area. Should be -1 for
19254 bluescreen.
19255 brightness
19257 Controls brightness of spill area, preserving colors.
19258 alpha
19260 Modify alpha from generated spillmap.
19261 Commands
19263 This filter supports the all above options as commands.
19265 detelecine
19267 Apply an exact inverse of the telecine operation. It requires a
19268 predefined pattern specified using the pattern option which must be the
19269 same as that passed to the telecine filter.
19270 This filter accepts the following options:
19272 first_field
19274 top, t
19275 top field first
19276 bottom, b
19278 bottom field first The default value is "top".
19279 pattern
19281 A string of numbers representing the pulldown pattern you wish to
19282 apply. The default value is 23.
19283 start_frame
19285 A number representing position of the first frame with respect to
19286 the telecine pattern. This is to be used if the stream is cut. The
19287 default value is 0.
19288 dilation
19290 Apply dilation effect to the video.
19291 This filter replaces the pixel by the local(3x3) maximum.
19293 It accepts the following options:
19295 threshold0
19297 threshold1
19298 threshold2
19299 threshold3
19300 Limit the maximum change for each plane, default is 65535. If 0,
19301 plane will remain unchanged.
19302 coordinates
19304 Flag which specifies the pixel to refer to. Default is 255 i.e. all
19305 eight pixels are used.
19306 Flags to local 3x3 coordinates maps like this:
19308 1 2 3
19310 4 5
19311 6 7 8
19312 Commands
19314 This filter supports the all above options as commands.
19316 displace
19318 Displace pixels as indicated by second and third input stream.
19319 It takes three input streams and outputs one stream, the first input is
19321 the source, and second and third input are displacement maps.
19322 The second input specifies how much to displace pixels along the
19324 x-axis, while the third input specifies how much to displace pixels
19325 along the y-axis. If one of displacement map streams terminates, last
19326 frame from that displacement map will be used.
19327 Note that once generated, displacements maps can be reused over and
19329 over again.
19330 A description of the accepted options follows.
19332 edge
19334 Set displace behavior for pixels that are out of range.
19335 Available values are:
19337 blank
19339 Missing pixels are replaced by black pixels.
19340 smear
19342 Adjacent pixels will spread out to replace missing pixels.
19343 wrap
19345 Out of range pixels are wrapped so they point to pixels of
19346 other side.
19347 mirror
19349 Out of range pixels will be replaced with mirrored pixels.
19350 Default is smear.
19352 Examples
19354 • Add ripple effect to rgb input of video size hd720:
19356 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
19358 • Add wave effect to rgb input of video size hd720:
19360 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
19362 dnn_classify
19364 Do classification with deep neural networks based on bounding boxes.
19365 The filter accepts the following options:
19367 dnn_backend
19369 Specify which DNN backend to use for model loading and execution.
19370 This option accepts only openvino now, tensorflow backends will be
19371 added.
19372 model
19374 Set path to model file specifying network architecture and its
19375 parameters. Note that different backends use different file
19376 formats.
19377 input
19379 Set the input name of the dnn network.
19380 output
19382 Set the output name of the dnn network.
19383 confidence
19385 Set the confidence threshold (default: 0.5).
19386 labels
19388 Set path to label file specifying the mapping between label id and
19389 name. Each label name is written in one line, tailing spaces and
19390 empty lines are skipped. The first line is the name of label id 0,
19391 and the second line is the name of label id 1, etc. The label id
19392 is considered as name if the label file is not provided.
19393 backend_configs
19395 Set the configs to be passed into backend
19396 For tensorflow backend, you can set its configs with sess_config
19398 options, please use tools/python/tf_sess_config.py to get the
19399 configs for your system.
19400 dnn_detect
19402 Do object detection with deep neural networks.
19403 The filter accepts the following options:
19405 dnn_backend
19407 Specify which DNN backend to use for model loading and execution.
19408 This option accepts only openvino now, tensorflow backends will be
19409 added.
19410 model
19412 Set path to model file specifying network architecture and its
19413 parameters. Note that different backends use different file
19414 formats.
19415 input
19417 Set the input name of the dnn network.
19418 output
19420 Set the output name of the dnn network.
19421 confidence
19423 Set the confidence threshold (default: 0.5).
19424 labels
19426 Set path to label file specifying the mapping between label id and
19427 name. Each label name is written in one line, tailing spaces and
19428 empty lines are skipped. The first line is the name of label id 0
19429 (usually it is 'background'), and the second line is the name of
19430 label id 1, etc. The label id is considered as name if the label
19431 file is not provided.
19432 backend_configs
19434 Set the configs to be passed into backend. To use async execution,
19435 set async (default: set). Roll back to sync execution if the
19436 backend does not support async.
19437 dnn_processing
19439 Do image processing with deep neural networks. It works together with
19440 another filter which converts the pixel format of the Frame to what the
19441 dnn network requires.
19442 The filter accepts the following options:
19444 dnn_backend
19446 Specify which DNN backend to use for model loading and execution.
19447 This option accepts the following values:
19448 native
19450 Native implementation of DNN loading and execution.
19451 tensorflow
19453 TensorFlow backend. To enable this backend you need to install
19454 the TensorFlow for C library (see
19455 <https://www.tensorflow.org/install/lang_c>) and configure
19456 FFmpeg with "--enable-libtensorflow"
19457 openvino
19459 OpenVINO backend. To enable this backend you need to build and
19460 install the OpenVINO for C library (see
19461 <https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md>)
19462 and configure FFmpeg with "--enable-libopenvino"
19463 (--extra-cflags=-I... --extra-ldflags=-L... might be needed if
19464 the header files and libraries are not installed into system
19465 path)
19466 Default value is native.
19468 model
19470 Set path to model file specifying network architecture and its
19471 parameters. Note that different backends use different file
19472 formats. TensorFlow, OpenVINO and native backend can load files for
19473 only its format.
19474 Native model file (.model) can be generated from TensorFlow model
19476 file (.pb) by using tools/python/convert.py
19477 input
19479 Set the input name of the dnn network.
19480 output
19482 Set the output name of the dnn network.
19483 backend_configs
19485 Set the configs to be passed into backend. To use async execution,
19486 set async (default: set). Roll back to sync execution if the
19487 backend does not support async.
19488 For tensorflow backend, you can set its configs with sess_config
19490 options, please use tools/python/tf_sess_config.py to get the
19491 configs of TensorFlow backend for your system.
19492 Examples
19494 • Remove rain in rgb24 frame with can.pb (see derain filter):
19496 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
19498 • Halve the pixel value of the frame with format gray32f:
19500 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
19502 • Handle the Y channel with srcnn.pb (see sr filter) for frame with
19504 yuv420p (planar YUV formats supported):
19505 ./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
19507 • Handle the Y channel with espcn.pb (see sr filter), which changes
19509 frame size, for format yuv420p (planar YUV formats supported),
19510 please use tools/python/tf_sess_config.py to get the configs of
19511 TensorFlow backend for your system.
19512 ./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
19514 drawbox
19516 Draw a colored box on the input image.
19517 It accepts the following parameters:
19519 x
19521 y The expressions which specify the top left corner coordinates of
19522 the box. It defaults to 0.
19523 width, w
19525 height, h
19526 The expressions which specify the width and height of the box; if 0
19527 they are interpreted as the input width and height. It defaults to
19528 0.
19529 color, c
19531 Specify the color of the box to write. For the general syntax of
19532 this option, check the "Color" section in the ffmpeg-utils manual.
19533 If the special value "invert" is used, the box edge color is the
19534 same as the video with inverted luma.
19535 thickness, t
19537 The expression which sets the thickness of the box edge. A value
19538 of "fill" will create a filled box. Default value is 3.
19539 See below for the list of accepted constants.
19541 replace
19543 Applicable if the input has alpha. With value 1, the pixels of the
19544 painted box will overwrite the video's color and alpha pixels.
19545 Default is 0, which composites the box onto the input, leaving the
19546 video's alpha intact.
19547 The parameters for x, y, w and h and t are expressions containing the
19549 following constants:
19550 dar The input display aspect ratio, it is the same as (w / h) * sar.
19552 hsub
19554 vsub
19555 horizontal and vertical chroma subsample values. For example for
19556 the pixel format "yuv422p" hsub is 2 and vsub is 1.
19557 in_h, ih
19559 in_w, iw
19560 The input width and height.
19561 sar The input sample aspect ratio.
19563 x
19565 y The x and y offset coordinates where the box is drawn.
19566 w
19568 h The width and height of the drawn box.
19569 box_source
19571 Box source can be set as side_data_detection_bboxes if you want to
19572 use box data in detection bboxes of side data.
19573 If box_source is set, the x, y, width and height will be ignored
19575 and still use box data in detection bboxes of side data. So please
19576 do not use this parameter if you were not sure about the box
19577 source.
19578 t The thickness of the drawn box.
19580 These constants allow the x, y, w, h and t expressions to refer to
19582 each other, so you may for example specify "y=x/dar" or "h=w/dar".
19583 Examples
19585 • Draw a black box around the edge of the input image:
19587 drawbox
19589 • Draw a box with color red and an opacity of 50%:
19591 drawbox=10:20:200:60:red@0.5
19593 The previous example can be specified as:
19595 drawbox=x=10:y=20:w=200:h=60:color=red@0.5
19597 • Fill the box with pink color:
19599 drawbox=x=10:y=10:w=100:h=100:color=pink@0.5:t=fill
19601 • Draw a 2-pixel red 2.40:1 mask:
19603 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
19605 Commands
19607 This filter supports same commands as options. The command accepts the
19609 same syntax of the corresponding option.
19610 If the specified expression is not valid, it is kept at its current
19612 value.
19613 drawgraph
19615 Draw a graph using input video metadata.
19616 It accepts the following parameters:
19618 m1 Set 1st frame metadata key from which metadata values will be used
19620 to draw a graph.
19621 fg1 Set 1st foreground color expression.
19623 m2 Set 2nd frame metadata key from which metadata values will be used
19625 to draw a graph.
19626 fg2 Set 2nd foreground color expression.
19628 m3 Set 3rd frame metadata key from which metadata values will be used
19630 to draw a graph.
19631 fg3 Set 3rd foreground color expression.
19633 m4 Set 4th frame metadata key from which metadata values will be used
19635 to draw a graph.
19636 fg4 Set 4th foreground color expression.
19638 min Set minimal value of metadata value.
19640 max Set maximal value of metadata value.
19642 bg Set graph background color. Default is white.
19644 mode
19646 Set graph mode.
19647 Available values for mode is:
19649 bar
19651 dot
19652 line
19653 Default is "line".
19655 slide
19657 Set slide mode.
19658 Available values for slide is:
19660 frame
19662 Draw new frame when right border is reached.
19663 replace
19665 Replace old columns with new ones.
19666 scroll
19668 Scroll from right to left.
19669 rscroll
19671 Scroll from left to right.
19672 picture
19674 Draw single picture.
19675 Default is "frame".
19677 size
19679 Set size of graph video. For the syntax of this option, check the
19680 "Video size" section in the ffmpeg-utils manual. The default value
19681 is "900x256".
19682 rate, r
19684 Set the output frame rate. Default value is 25.
19685 The foreground color expressions can use the following variables:
19687 MIN Minimal value of metadata value.
19689 MAX Maximal value of metadata value.
19691 VAL Current metadata key value.
19693 The color is defined as 0xAABBGGRR.
19695 Example using metadata from signalstats filter:
19697 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
19699 Example using metadata from ebur128 filter:
19701 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
19703 drawgrid
19705 Draw a grid on the input image.
19706 It accepts the following parameters:
19708 x
19710 y The expressions which specify the coordinates of some point of grid
19711 intersection (meant to configure offset). Both default to 0.
19712 width, w
19714 height, h
19715 The expressions which specify the width and height of the grid
19716 cell, if 0 they are interpreted as the input width and height,
19717 respectively, minus "thickness", so image gets framed. Default to
19718 0.
19719 color, c
19721 Specify the color of the grid. For the general syntax of this
19722 option, check the "Color" section in the ffmpeg-utils manual. If
19723 the special value "invert" is used, the grid color is the same as
19724 the video with inverted luma.
19725 thickness, t
19727 The expression which sets the thickness of the grid line. Default
19728 value is 1.
19729 See below for the list of accepted constants.
19731 replace
19733 Applicable if the input has alpha. With 1 the pixels of the painted
19734 grid will overwrite the video's color and alpha pixels. Default is
19735 0, which composites the grid onto the input, leaving the video's
19736 alpha intact.
19737 The parameters for x, y, w and h and t are expressions containing the
19739 following constants:
19740 dar The input display aspect ratio, it is the same as (w / h) * sar.
19742 hsub
19744 vsub
19745 horizontal and vertical chroma subsample values. For example for
19746 the pixel format "yuv422p" hsub is 2 and vsub is 1.
19747 in_h, ih
19749 in_w, iw
19750 The input grid cell width and height.
19751 sar The input sample aspect ratio.
19753 x
19755 y The x and y coordinates of some point of grid intersection (meant
19756 to configure offset).
19757 w
19759 h The width and height of the drawn cell.
19760 t The thickness of the drawn cell.
19762 These constants allow the x, y, w, h and t expressions to refer to
19764 each other, so you may for example specify "y=x/dar" or "h=w/dar".
19765 Examples
19767 • Draw a grid with cell 100x100 pixels, thickness 2 pixels, with
19769 color red and an opacity of 50%:
19770 drawgrid=width=100:height=100:thickness=2:color=red@0.5
19772 • Draw a white 3x3 grid with an opacity of 50%:
19774 drawgrid=w=iw/3:h=ih/3:t=2:c=white@0.5
19776 Commands
19778 This filter supports same commands as options. The command accepts the
19780 same syntax of the corresponding option.
19781 If the specified expression is not valid, it is kept at its current
19783 value.
19784 drawtext
19786 Draw a text string or text from a specified file on top of a video,
19787 using the libfreetype library.
19788 To enable compilation of this filter, you need to configure FFmpeg with
19790 "--enable-libfreetype". To enable default font fallback and the font
19791 option you need to configure FFmpeg with "--enable-libfontconfig". To
19792 enable the text_shaping option, you need to configure FFmpeg with
19793 "--enable-libfribidi".
19794 Syntax
19796 It accepts the following parameters:
19798 box Used to draw a box around text using the background color. The
19800 value must be either 1 (enable) or 0 (disable). The default value
19801 of box is 0.
19802 boxborderw
19804 Set the width of the border to be drawn around the box using
19805 boxcolor. The default value of boxborderw is 0.
19806 boxcolor
19808 The color to be used for drawing box around text. For the syntax of
19809 this option, check the "Color" section in the ffmpeg-utils manual.
19810 The default value of boxcolor is "white".
19812 line_spacing
19814 Set the line spacing in pixels of the border to be drawn around the
19815 box using box. The default value of line_spacing is 0.
19816 borderw
19818 Set the width of the border to be drawn around the text using
19819 bordercolor. The default value of borderw is 0.
19820 bordercolor
19822 Set the color to be used for drawing border around text. For the
19823 syntax of this option, check the "Color" section in the ffmpeg-
19824 utils manual.
19825 The default value of bordercolor is "black".
19827 expansion
19829 Select how the text is expanded. Can be either "none", "strftime"
19830 (deprecated) or "normal" (default). See the drawtext_expansion,
19831 Text expansion section below for details.
19832 basetime
19834 Set a start time for the count. Value is in microseconds. Only
19835 applied in the deprecated strftime expansion mode. To emulate in
19836 normal expansion mode use the "pts" function, supplying the start
19837 time (in seconds) as the second argument.
19838 fix_bounds
19840 If true, check and fix text coords to avoid clipping.
19841 fontcolor
19843 The color to be used for drawing fonts. For the syntax of this
19844 option, check the "Color" section in the ffmpeg-utils manual.
19845 The default value of fontcolor is "black".
19847 fontcolor_expr
19849 String which is expanded the same way as text to obtain dynamic
19850 fontcolor value. By default this option has empty value and is not
19851 processed. When this option is set, it overrides fontcolor option.
19852 font
19854 The font family to be used for drawing text. By default Sans.
19855 fontfile
19857 The font file to be used for drawing text. The path must be
19858 included. This parameter is mandatory if the fontconfig support is
19859 disabled.
19860 alpha
19862 Draw the text applying alpha blending. The value can be a number
19863 between 0.0 and 1.0. The expression accepts the same variables x,
19864 y as well. The default value is 1. Please see fontcolor_expr.
19865 fontsize
19867 The font size to be used for drawing text. The default value of
19868 fontsize is 16.
19869 text_shaping
19871 If set to 1, attempt to shape the text (for example, reverse the
19872 order of right-to-left text and join Arabic characters) before
19873 drawing it. Otherwise, just draw the text exactly as given. By
19874 default 1 (if supported).
19875 ft_load_flags
19877 The flags to be used for loading the fonts.
19878 The flags map the corresponding flags supported by libfreetype, and
19880 are a combination of the following values:
19881 default
19883 no_scale
19884 no_hinting
19885 render
19886 no_bitmap
19887 vertical_layout
19888 force_autohint
19889 crop_bitmap
19890 pedantic
19891 ignore_global_advance_width
19892 no_recurse
19893 ignore_transform
19894 monochrome
19895 linear_design
19896 no_autohint
19897 Default value is "default".
19899 For more information consult the documentation for the FT_LOAD_*
19901 libfreetype flags.
19902 shadowcolor
19904 The color to be used for drawing a shadow behind the drawn text.
19905 For the syntax of this option, check the "Color" section in the
19906 ffmpeg-utils manual.
19907 The default value of shadowcolor is "black".
19909 shadowx
19911 shadowy
19912 The x and y offsets for the text shadow position with respect to
19913 the position of the text. They can be either positive or negative
19914 values. The default value for both is "0".
19915 start_number
19917 The starting frame number for the n/frame_num variable. The default
19918 value is "0".
19919 tabsize
19921 The size in number of spaces to use for rendering the tab. Default
19922 value is 4.
19923 timecode
19925 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
19926 format. It can be used with or without text parameter.
19927 timecode_rate option must be specified.
19928 timecode_rate, rate, r
19930 Set the timecode frame rate (timecode only). Value will be rounded
19931 to nearest integer. Minimum value is "1". Drop-frame timecode is
19932 supported for frame rates 30 & 60.
19933 tc24hmax
19935 If set to 1, the output of the timecode option will wrap around at
19936 24 hours. Default is 0 (disabled).
19937 text
19939 The text string to be drawn. The text must be a sequence of UTF-8
19940 encoded characters. This parameter is mandatory if no file is
19941 specified with the parameter textfile.
19942 textfile
19944 A text file containing text to be drawn. The text must be a
19945 sequence of UTF-8 encoded characters.
19946 This parameter is mandatory if no text string is specified with the
19948 parameter text.
19949 If both text and textfile are specified, an error is thrown.
19951 text_source
19953 Text source should be set as side_data_detection_bboxes if you want
19954 to use text data in detection bboxes of side data.
19955 If text source is set, text and textfile will be ignored and still
19957 use text data in detection bboxes of side data. So please do not
19958 use this parameter if you are not sure about the text source.
19959 reload
19961 The textfile will be reloaded at specified frame interval. Be sure
19962 to update textfile atomically, or it may be read partially, or even
19963 fail. Range is 0 to INT_MAX. Default is 0.
19964 x
19966 y The expressions which specify the offsets where text will be drawn
19967 within the video frame. They are relative to the top/left border of
19968 the output image.
19969 The default value of x and y is "0".
19971 See below for the list of accepted constants and functions.
19973 The parameters for x and y are expressions containing the following
19975 constants and functions:
19976 dar input display aspect ratio, it is the same as (w / h) * sar
19978 hsub
19980 vsub
19981 horizontal and vertical chroma subsample values. For example for
19982 the pixel format "yuv422p" hsub is 2 and vsub is 1.
19983 line_h, lh
19985 the height of each text line
19986 main_h, h, H
19988 the input height
19989 main_w, w, W
19991 the input width
19992 max_glyph_a, ascent
19994 the maximum distance from the baseline to the highest/upper grid
19995 coordinate used to place a glyph outline point, for all the
19996 rendered glyphs. It is a positive value, due to the grid's
19997 orientation with the Y axis upwards.
19998 max_glyph_d, descent
20000 the maximum distance from the baseline to the lowest grid
20001 coordinate used to place a glyph outline point, for all the
20002 rendered glyphs. This is a negative value, due to the grid's
20003 orientation, with the Y axis upwards.
20004 max_glyph_h
20006 maximum glyph height, that is the maximum height for all the glyphs
20007 contained in the rendered text, it is equivalent to ascent -
20008 descent.
20009 max_glyph_w
20011 maximum glyph width, that is the maximum width for all the glyphs
20012 contained in the rendered text
20013 n the number of input frame, starting from 0
20015 rand(min, max)
20017 return a random number included between min and max
20018 sar The input sample aspect ratio.
20020 t timestamp expressed in seconds, NAN if the input timestamp is
20022 unknown
20023 text_h, th
20025 the height of the rendered text
20026 text_w, tw
20028 the width of the rendered text
20029 x
20031 y the x and y offset coordinates where the text is drawn.
20032 These parameters allow the x and y expressions to refer to each
20034 other, so you can for example specify "y=x/dar".
20035 pict_type
20037 A one character description of the current frame's picture type.
20038 pkt_pos
20040 The current packet's position in the input file or stream (in
20041 bytes, from the start of the input). A value of -1 indicates this
20042 info is not available.
20043 duration
20045 The current packet's duration, in seconds.
20046 pkt_size
20048 The current packet's size (in bytes).
20049 Text expansion
20051 If expansion is set to "strftime", the filter recognizes strftime()
20053 sequences in the provided text and expands them accordingly. Check the
20054 documentation of strftime(). This feature is deprecated.
20055 If expansion is set to "none", the text is printed verbatim.
20057 If expansion is set to "normal" (which is the default), the following
20059 expansion mechanism is used.
20060 The backslash character \, followed by any character, always expands to
20062 the second character.
20063 Sequences of the form "%{...}" are expanded. The text between the
20065 braces is a function name, possibly followed by arguments separated by
20066 ':'. If the arguments contain special characters or delimiters (':' or
20067 '}'), they should be escaped.
20068 Note that they probably must also be escaped as the value for the text
20070 option in the filter argument string and as the filter argument in the
20071 filtergraph description, and possibly also for the shell, that makes up
20072 to four levels of escaping; using a text file avoids these problems.
20073 The following functions are available:
20075 expr, e
20077 The expression evaluation result.
20078 It must take one argument specifying the expression to be
20080 evaluated, which accepts the same constants and functions as the x
20081 and y values. Note that not all constants should be used, for
20082 example the text size is not known when evaluating the expression,
20083 so the constants text_w and text_h will have an undefined value.
20084 expr_int_format, eif
20086 Evaluate the expression's value and output as formatted integer.
20087 The first argument is the expression to be evaluated, just as for
20089 the expr function. The second argument specifies the output
20090 format. Allowed values are x, X, d and u. They are treated exactly
20091 as in the "printf" function. The third parameter is optional and
20092 sets the number of positions taken by the output. It can be used
20093 to add padding with zeros from the left.
20094 gmtime
20096 The time at which the filter is running, expressed in UTC. It can
20097 accept an argument: a strftime() format string. The format string
20098 is extended to support the variable %[1-6]N which prints fractions
20099 of the second with optionally specified number of digits.
20100 localtime
20102 The time at which the filter is running, expressed in the local
20103 time zone. It can accept an argument: a strftime() format string.
20104 The format string is extended to support the variable %[1-6]N which
20105 prints fractions of the second with optionally specified number of
20106 digits.
20107 metadata
20109 Frame metadata. Takes one or two arguments.
20110 The first argument is mandatory and specifies the metadata key.
20112 The second argument is optional and specifies a default value, used
20114 when the metadata key is not found or empty.
20115 Available metadata can be identified by inspecting entries starting
20117 with TAG included within each frame section printed by running
20118 "ffprobe -show_frames".
20119 String metadata generated in filters leading to the drawtext filter
20121 are also available.
20122 n, frame_num
20124 The frame number, starting from 0.
20125 pict_type
20127 A one character description of the current picture type.
20128 pts The timestamp of the current frame. It can take up to three
20130 arguments.
20131 The first argument is the format of the timestamp; it defaults to
20133 "flt" for seconds as a decimal number with microsecond accuracy;
20134 "hms" stands for a formatted [-]HH:MM:SS.mmm timestamp with
20135 millisecond accuracy. "gmtime" stands for the timestamp of the
20136 frame formatted as UTC time; "localtime" stands for the timestamp
20137 of the frame formatted as local time zone time.
20138 The second argument is an offset added to the timestamp.
20140 If the format is set to "hms", a third argument "24HH" may be
20142 supplied to present the hour part of the formatted timestamp in 24h
20143 format (00-23).
20144 If the format is set to "localtime" or "gmtime", a third argument
20146 may be supplied: a strftime() format string. By default, YYYY-MM-
20147 DD HH:MM:SS format will be used.
20148 Commands
20150 This filter supports altering parameters via commands:
20152 reinit
20154 Alter existing filter parameters.
20155 Syntax for the argument is the same as for filter invocation, e.g.
20157 fontsize=56:fontcolor=green:text='Hello World'
20159 Full filter invocation with sendcmd would look like this:
20161 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
20163 If the entire argument can't be parsed or applied as valid values then
20165 the filter will continue with its existing parameters.
20166 Examples
20168 • Draw "Test Text" with font FreeSerif, using the default values for
20170 the optional parameters.
20171 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
20173 • Draw 'Test Text' with font FreeSerif of size 24 at position x=100
20175 and y=50 (counting from the top-left corner of the screen), text is
20176 yellow with a red box around it. Both the text and the box have an
20177 opacity of 20%.
20178 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
20180 x=100: y=50: fontsize=24: fontcolor=yellow@0.2: box=1: boxcolor=red@0.2"
20181 Note that the double quotes are not necessary if spaces are not
20183 used within the parameter list.
20184 • Show the text at the center of the video frame:
20186 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
20188 • Show the text at a random position, switching to a new position
20190 every 30 seconds:
20191 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)"
20193 • Show a text line sliding from right to left in the last row of the
20195 video frame. The file LONG_LINE is assumed to contain a single line
20196 with no newlines.
20197 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
20199 • Show the content of file CREDITS off the bottom of the frame and
20201 scroll up.
20202 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
20204 • Draw a single green letter "g", at the center of the input video.
20206 The glyph baseline is placed at half screen height.
20207 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
20209 • Show text for 1 second every 3 seconds:
20211 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
20213 • Use fontconfig to set the font. Note that the colons need to be
20215 escaped.
20216 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
20218 • Draw "Test Text" with font size dependent on height of the video.
20220 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
20222 • Print the date of a real-time encoding (see strftime(3)):
20224 drawtext='fontfile=FreeSans.ttf:text=%{localtime\:%a %b %d %Y}'
20226 • Show text fading in and out (appearing/disappearing):
20228 #!/bin/sh
20230 DS=1.0 # display start
20231 DE=10.0 # display end
20232 FID=1.5 # fade in duration
20233 FOD=5 # fade out duration
20234 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 }"
20235 • Horizontally align multiple separate texts. Note that max_glyph_a
20237 and the fontsize value are included in the y offset.
20238 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
20240 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
20241 • Plot special lavf.image2dec.source_basename metadata onto each
20243 frame if such metadata exists. Otherwise, plot the string "NA".
20244 Note that image2 demuxer must have option -export_path_metadata 1
20245 for the special metadata fields to be available for filters.
20246 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%{metadata\:lavf.image2dec.source_basename\:NA}':x=10:y=10"
20248 For more information about libfreetype, check:
20250 <http://www.freetype.org/>.
20251 For more information about fontconfig, check:
20253 <http://freedesktop.org/software/fontconfig/fontconfig-user.html>.
20254 For more information about libfribidi, check: <http://fribidi.org/>.
20256 edgedetect
20258 Detect and draw edges. The filter uses the Canny Edge Detection
20259 algorithm.
20260 The filter accepts the following options:
20262 low
20264 high
20265 Set low and high threshold values used by the Canny thresholding
20266 algorithm.
20267 The high threshold selects the "strong" edge pixels, which are then
20269 connected through 8-connectivity with the "weak" edge pixels
20270 selected by the low threshold.
20271 low and high threshold values must be chosen in the range [0,1],
20273 and low should be lesser or equal to high.
20274 Default value for low is "20/255", and default value for high is
20276 "50/255".
20277 mode
20279 Define the drawing mode.
20280 wires
20282 Draw white/gray wires on black background.
20283 colormix
20285 Mix the colors to create a paint/cartoon effect.
20286 canny
20288 Apply Canny edge detector on all selected planes.
20289 Default value is wires.
20291 planes
20293 Select planes for filtering. By default all available planes are
20294 filtered.
20295 Examples
20297 • Standard edge detection with custom values for the hysteresis
20299 thresholding:
20300 edgedetect=low=0.1:high=0.4
20302 • Painting effect without thresholding:
20304 edgedetect=mode=colormix:high=0
20306 elbg
20308 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
20309 For each input image, the filter will compute the optimal mapping from
20311 the input to the output given the codebook length, that is the number
20312 of distinct output colors.
20313 This filter accepts the following options.
20315 codebook_length, l
20317 Set codebook length. The value must be a positive integer, and
20318 represents the number of distinct output colors. Default value is
20319 256.
20320 nb_steps, n
20322 Set the maximum number of iterations to apply for computing the
20323 optimal mapping. The higher the value the better the result and the
20324 higher the computation time. Default value is 1.
20325 seed, s
20327 Set a random seed, must be an integer included between 0 and
20328 UINT32_MAX. If not specified, or if explicitly set to -1, the
20329 filter will try to use a good random seed on a best effort basis.
20330 pal8
20332 Set pal8 output pixel format. This option does not work with
20333 codebook length greater than 256. Default is disabled.
20334 use_alpha
20336 Include alpha values in the quantization calculation. Allows
20337 creating palettized output images (e.g. PNG8) with multiple alpha
20338 smooth blending.
20339 entropy
20341 Measure graylevel entropy in histogram of color channels of video
20342 frames.
20343 It accepts the following parameters:
20345 mode
20347 Can be either normal or diff. Default is normal.
20348 diff mode measures entropy of histogram delta values, absolute
20350 differences between neighbour histogram values.
20351 epx
20353 Apply the EPX magnification filter which is designed for pixel art.
20354 It accepts the following option:
20356 n Set the scaling dimension: 2 for "2xEPX", 3 for "3xEPX". Default
20358 is 3.
20359 eq
20361 Set brightness, contrast, saturation and approximate gamma adjustment.
20362 The filter accepts the following options:
20364 contrast
20366 Set the contrast expression. The value must be a float value in
20367 range -1000.0 to 1000.0. The default value is "1".
20368 brightness
20370 Set the brightness expression. The value must be a float value in
20371 range -1.0 to 1.0. The default value is "0".
20372 saturation
20374 Set the saturation expression. The value must be a float in range
20375 0.0 to 3.0. The default value is "1".
20376 gamma
20378 Set the gamma expression. The value must be a float in range 0.1 to
20379 10.0. The default value is "1".
20380 gamma_r
20382 Set the gamma expression for red. The value must be a float in
20383 range 0.1 to 10.0. The default value is "1".
20384 gamma_g
20386 Set the gamma expression for green. The value must be a float in
20387 range 0.1 to 10.0. The default value is "1".
20388 gamma_b
20390 Set the gamma expression for blue. The value must be a float in
20391 range 0.1 to 10.0. The default value is "1".
20392 gamma_weight
20394 Set the gamma weight expression. It can be used to reduce the
20395 effect of a high gamma value on bright image areas, e.g. keep them
20396 from getting overamplified and just plain white. The value must be
20397 a float in range 0.0 to 1.0. A value of 0.0 turns the gamma
20398 correction all the way down while 1.0 leaves it at its full
20399 strength. Default is "1".
20400 eval
20402 Set when the expressions for brightness, contrast, saturation and
20403 gamma expressions are evaluated.
20404 It accepts the following values:
20406 init
20408 only evaluate expressions once during the filter initialization
20409 or when a command is processed
20410 frame
20412 evaluate expressions for each incoming frame
20413 Default value is init.
20415 The expressions accept the following parameters:
20417 n frame count of the input frame starting from 0
20419 pos byte position of the corresponding packet in the input file, NAN if
20421 unspecified
20422 r frame rate of the input video, NAN if the input frame rate is
20424 unknown
20425 t timestamp expressed in seconds, NAN if the input timestamp is
20427 unknown
20428 Commands
20430 The filter supports the following commands:
20432 contrast
20434 Set the contrast expression.
20435 brightness
20437 Set the brightness expression.
20438 saturation
20440 Set the saturation expression.
20441 gamma
20443 Set the gamma expression.
20444 gamma_r
20446 Set the gamma_r expression.
20447 gamma_g
20449 Set gamma_g expression.
20450 gamma_b
20452 Set gamma_b expression.
20453 gamma_weight
20455 Set gamma_weight expression.
20456 The command accepts the same syntax of the corresponding option.
20458 If the specified expression is not valid, it is kept at its current
20460 value.
20461 erosion
20463 Apply erosion effect to the video.
20464 This filter replaces the pixel by the local(3x3) minimum.
20466 It accepts the following options:
20468 threshold0
20470 threshold1
20471 threshold2
20472 threshold3
20473 Limit the maximum change for each plane, default is 65535. If 0,
20474 plane will remain unchanged.
20475 coordinates
20477 Flag which specifies the pixel to refer to. Default is 255 i.e. all
20478 eight pixels are used.
20479 Flags to local 3x3 coordinates maps like this:
20481 1 2 3
20483 4 5
20484 6 7 8
20485 Commands
20487 This filter supports the all above options as commands.
20489 estdif
20491 Deinterlace the input video ("estdif" stands for "Edge Slope Tracing
20492 Deinterlacing Filter").
20493 Spatial only filter that uses edge slope tracing algorithm to
20495 interpolate missing lines. It accepts the following parameters:
20496 mode
20498 The interlacing mode to adopt. It accepts one of the following
20499 values:
20500 frame
20502 Output one frame for each frame.
20503 field
20505 Output one frame for each field.
20506 The default value is "field".
20508 parity
20510 The picture field parity assumed for the input interlaced video. It
20511 accepts one of the following values:
20512 tff Assume the top field is first.
20514 bff Assume the bottom field is first.
20516 auto
20518 Enable automatic detection of field parity.
20519 The default value is "auto". If the interlacing is unknown or the
20521 decoder does not export this information, top field first will be
20522 assumed.
20523 deint
20525 Specify which frames to deinterlace. Accepts one of the following
20526 values:
20527 all Deinterlace all frames.
20529 interlaced
20531 Only deinterlace frames marked as interlaced.
20532 The default value is "all".
20534 rslope
20536 Specify the search radius for edge slope tracing. Default value is
20537 1. Allowed range is from 1 to 15.
20538 redge
20540 Specify the search radius for best edge matching. Default value is
20541 2. Allowed range is from 0 to 15.
20542 ecost
20544 Specify the edge cost for edge matching. Default value is 1.0.
20545 Allowed range is from 0 to 9.
20546 mcost
20548 Specify the middle cost for edge matching. Default value is 0.5.
20549 Allowed range is from 0 to 1.
20550 dcost
20552 Specify the distance cost for edge matching. Default value is 0.5.
20553 Allowed range is from 0 to 1.
20554 interp
20556 Specify the interpolation used. Default is 4-point interpolation.
20557 It accepts one of the following values:
20558 2p Two-point interpolation.
20560 4p Four-point interpolation.
20562 6p Six-point interpolation.
20564 Commands
20566 This filter supports same commands as options.
20568 exposure
20570 Adjust exposure of the video stream.
20571 The filter accepts the following options:
20573 exposure
20575 Set the exposure correction in EV. Allowed range is from -3.0 to
20576 3.0 EV Default value is 0 EV.
20577 black
20579 Set the black level correction. Allowed range is from -1.0 to 1.0.
20580 Default value is 0.
20581 Commands
20583 This filter supports same commands as options.
20585 extractplanes
20587 Extract color channel components from input video stream into separate
20588 grayscale video streams.
20589 The filter accepts the following option:
20591 planes
20593 Set plane(s) to extract.
20594 Available values for planes are:
20596 y
20598 u
20599 v
20600 a
20601 r
20602 g
20603 b
20604 Choosing planes not available in the input will result in an error.
20606 That means you cannot select "r", "g", "b" planes with "y", "u",
20607 "v" planes at same time.
20608 Examples
20610 • Extract luma, u and v color channel component from input video
20612 frame into 3 grayscale outputs:
20613 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
20615 fade
20617 Apply a fade-in/out effect to the input video.
20618 It accepts the following parameters:
20620 type, t
20622 The effect type can be either "in" for a fade-in, or "out" for a
20623 fade-out effect. Default is "in".
20624 start_frame, s
20626 Specify the number of the frame to start applying the fade effect
20627 at. Default is 0.
20628 nb_frames, n
20630 The number of frames that the fade effect lasts. At the end of the
20631 fade-in effect, the output video will have the same intensity as
20632 the input video. At the end of the fade-out transition, the output
20633 video will be filled with the selected color. Default is 25.
20634 alpha
20636 If set to 1, fade only alpha channel, if one exists on the input.
20637 Default value is 0.
20638 start_time, st
20640 Specify the timestamp (in seconds) of the frame to start to apply
20641 the fade effect. If both start_frame and start_time are specified,
20642 the fade will start at whichever comes last. Default is 0.
20643 duration, d
20645 The number of seconds for which the fade effect has to last. At the
20646 end of the fade-in effect the output video will have the same
20647 intensity as the input video, at the end of the fade-out transition
20648 the output video will be filled with the selected color. If both
20649 duration and nb_frames are specified, duration is used. Default is
20650 0 (nb_frames is used by default).
20651 color, c
20653 Specify the color of the fade. Default is "black".
20654 Examples
20656 • Fade in the first 30 frames of video:
20658 fade=in:0:30
20660 The command above is equivalent to:
20662 fade=t=in:s=0:n=30
20664 • Fade out the last 45 frames of a 200-frame video:
20666 fade=out:155:45
20668 fade=type=out:start_frame=155:nb_frames=45
20669 • Fade in the first 25 frames and fade out the last 25 frames of a
20671 1000-frame video:
20672 fade=in:0:25, fade=out:975:25
20674 • Make the first 5 frames yellow, then fade in from frame 5-24:
20676 fade=in:5:20:color=yellow
20678 • Fade in alpha over first 25 frames of video:
20680 fade=in:0:25:alpha=1
20682 • Make the first 5.5 seconds black, then fade in for 0.5 seconds:
20684 fade=t=in:st=5.5:d=0.5
20686 feedback
20688 Apply feedback video filter.
20689 This filter pass cropped input frames to 2nd output. From there it can
20691 be filtered with other video filters. After filter receives frame from
20692 2nd input, that frame is combined on top of original frame from 1st
20693 input and passed to 1st output.
20694 The typical usage is filter only part of frame.
20696 The filter accepts the following options:
20698 x
20700 y Set the top left crop position.
20701 w
20703 h Set the crop size.
20704 Examples
20706 • Blur only top left rectangular part of video frame size 100x100
20708 with gblur filter.
20709 [in][blurin]feedback=x=0:y=0:w=100:h=100[out][blurout];[blurout]gblur=8[blurin]
20711 • Draw black box on top left part of video frame of size 100x100 with
20713 drawbox filter.
20714 [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]
20716 fftdnoiz
20718 Denoise frames using 3D FFT (frequency domain filtering).
20719 The filter accepts the following options:
20721 sigma
20723 Set the noise sigma constant. This sets denoising strength.
20724 Default value is 1. Allowed range is from 0 to 30. Using very high
20725 sigma with low overlap may give blocking artifacts.
20726 amount
20728 Set amount of denoising. By default all detected noise is reduced.
20729 Default value is 1. Allowed range is from 0 to 1.
20730 block
20732 Set size of block in pixels, Default is 32, can be 8 to 256.
20733 overlap
20735 Set block overlap. Default is 0.5. Allowed range is from 0.2 to
20736 0.8.
20737 method
20739 Set denoising method. Default is "wiener", can also be "hard".
20740 prev
20742 Set number of previous frames to use for denoising. By default is
20743 set to 0.
20744 next
20746 Set number of next frames to to use for denoising. By default is
20747 set to 0.
20748 planes
20750 Set planes which will be filtered, by default are all available
20751 filtered except alpha.
20752 fftfilt
20754 Apply arbitrary expressions to samples in frequency domain
20755 dc_Y
20757 Adjust the dc value (gain) of the luma plane of the image. The
20758 filter accepts an integer value in range 0 to 1000. The default
20759 value is set to 0.
20760 dc_U
20762 Adjust the dc value (gain) of the 1st chroma plane of the image.
20763 The filter accepts an integer value in range 0 to 1000. The default
20764 value is set to 0.
20765 dc_V
20767 Adjust the dc value (gain) of the 2nd chroma plane of the image.
20768 The filter accepts an integer value in range 0 to 1000. The default
20769 value is set to 0.
20770 weight_Y
20772 Set the frequency domain weight expression for the luma plane.
20773 weight_U
20775 Set the frequency domain weight expression for the 1st chroma
20776 plane.
20777 weight_V
20779 Set the frequency domain weight expression for the 2nd chroma
20780 plane.
20781 eval
20783 Set when the expressions are evaluated.
20784 It accepts the following values:
20786 init
20788 Only evaluate expressions once during the filter
20789 initialization.
20790 frame
20792 Evaluate expressions for each incoming frame.
20793 Default value is init.
20795 The filter accepts the following variables:
20797 X
20799 Y The coordinates of the current sample.
20800 W
20802 H The width and height of the image.
20803 N The number of input frame, starting from 0.
20805 WS
20807 HS The size of FFT array for horizontal and vertical processing.
20808 Examples
20810 • High-pass:
20812 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
20814 • Low-pass:
20816 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
20818 • Sharpen:
20820 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
20822 • Blur:
20824 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
20826 field
20828 Extract a single field from an interlaced image using stride arithmetic
20829 to avoid wasting CPU time. The output frames are marked as non-
20830 interlaced.
20831 The filter accepts the following options:
20833 type
20835 Specify whether to extract the top (if the value is 0 or "top") or
20836 the bottom field (if the value is 1 or "bottom").
20837 fieldhint
20839 Create new frames by copying the top and bottom fields from surrounding
20840 frames supplied as numbers by the hint file.
20841 hint
20843 Set file containing hints: absolute/relative frame numbers.
20844 There must be one line for each frame in a clip. Each line must
20846 contain two numbers separated by the comma, optionally followed by
20847 "-" or "+". Numbers supplied on each line of file can not be out
20848 of [N-1,N+1] where N is current frame number for "absolute" mode or
20849 out of [-1, 1] range for "relative" mode. First number tells from
20850 which frame to pick up top field and second number tells from which
20851 frame to pick up bottom field.
20852 If optionally followed by "+" output frame will be marked as
20854 interlaced, else if followed by "-" output frame will be marked as
20855 progressive, else it will be marked same as input frame. If
20856 optionally followed by "t" output frame will use only top field, or
20857 in case of "b" it will use only bottom field. If line starts with
20858 "#" or ";" that line is skipped.
20859 mode
20861 Can be item "absolute" or "relative" or "pattern". Default is
20862 "absolute". The "pattern" mode is same as "relative" mode, except
20863 at last entry of file if there are more frames to process than
20864 "hint" file is seek back to start.
20865 Example of first several lines of "hint" file for "relative" mode:
20867 0,0 - # first frame
20869 1,0 - # second frame, use third's frame top field and second's frame bottom field
20870 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
20871 1,0 -
20872 0,0 -
20873 0,0 -
20874 1,0 -
20875 1,0 -
20876 1,0 -
20877 0,0 -
20878 0,0 -
20879 1,0 -
20880 1,0 -
20881 1,0 -
20882 0,0 -
20883 fieldmatch
20885 Field matching filter for inverse telecine. It is meant to reconstruct
20886 the progressive frames from a telecined stream. The filter does not
20887 drop duplicated frames, so to achieve a complete inverse telecine
20888 "fieldmatch" needs to be followed by a decimation filter such as
20889 decimate in the filtergraph.
20890 The separation of the field matching and the decimation is notably
20892 motivated by the possibility of inserting a de-interlacing filter
20893 fallback between the two. If the source has mixed telecined and real
20894 interlaced content, "fieldmatch" will not be able to match fields for
20895 the interlaced parts. But these remaining combed frames will be marked
20896 as interlaced, and thus can be de-interlaced by a later filter such as
20897 yadif before decimation.
20898 In addition to the various configuration options, "fieldmatch" can take
20900 an optional second stream, activated through the ppsrc option. If
20901 enabled, the frames reconstruction will be based on the fields and
20902 frames from this second stream. This allows the first input to be pre-
20903 processed in order to help the various algorithms of the filter, while
20904 keeping the output lossless (assuming the fields are matched properly).
20905 Typically, a field-aware denoiser, or brightness/contrast adjustments
20906 can help.
20907 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth
20909 project) and VIVTC/VFM (VapourSynth project). The later is a light
20910 clone of TFM from which "fieldmatch" is based on. While the semantic
20911 and usage are very close, some behaviour and options names can differ.
20912 The decimate filter currently only works for constant frame rate input.
20914 If your input has mixed telecined (30fps) and progressive content with
20915 a lower framerate like 24fps use the following filterchain to produce
20916 the necessary cfr stream:
20917 "dejudder,fps=30000/1001,fieldmatch,decimate".
20918 The filter accepts the following options:
20920 order
20922 Specify the assumed field order of the input stream. Available
20923 values are:
20924 auto
20926 Auto detect parity (use FFmpeg's internal parity value).
20927 bff Assume bottom field first.
20929 tff Assume top field first.
20931 Note that it is sometimes recommended not to trust the parity
20933 announced by the stream.
20934 Default value is auto.
20936 mode
20938 Set the matching mode or strategy to use. pc mode is the safest in
20939 the sense that it won't risk creating jerkiness due to duplicate
20940 frames when possible, but if there are bad edits or blended fields
20941 it will end up outputting combed frames when a good match might
20942 actually exist. On the other hand, pcn_ub mode is the most risky in
20943 terms of creating jerkiness, but will almost always find a good
20944 frame if there is one. The other values are all somewhere in
20945 between pc and pcn_ub in terms of risking jerkiness and creating
20946 duplicate frames versus finding good matches in sections with bad
20947 edits, orphaned fields, blended fields, etc.
20948 More details about p/c/n/u/b are available in p/c/n/u/b meaning
20950 section.
20951 Available values are:
20953 pc 2-way matching (p/c)
20955 pc_n
20957 2-way matching, and trying 3rd match if still combed (p/c + n)
20958 pc_u
20960 2-way matching, and trying 3rd match (same order) if still
20961 combed (p/c + u)
20962 pc_n_ub
20964 2-way matching, trying 3rd match if still combed, and trying
20965 4th/5th matches if still combed (p/c + n + u/b)
20966 pcn 3-way matching (p/c/n)
20968 pcn_ub
20970 3-way matching, and trying 4th/5th matches if all 3 of the
20971 original matches are detected as combed (p/c/n + u/b)
20972 The parenthesis at the end indicate the matches that would be used
20974 for that mode assuming order=tff (and field on auto or top).
20975 In terms of speed pc mode is by far the fastest and pcn_ub is the
20977 slowest.
20978 Default value is pc_n.
20980 ppsrc
20982 Mark the main input stream as a pre-processed input, and enable the
20983 secondary input stream as the clean source to pick the fields from.
20984 See the filter introduction for more details. It is similar to the
20985 clip2 feature from VFM/TFM.
20986 Default value is 0 (disabled).
20988 field
20990 Set the field to match from. It is recommended to set this to the
20991 same value as order unless you experience matching failures with
20992 that setting. In certain circumstances changing the field that is
20993 used to match from can have a large impact on matching performance.
20994 Available values are:
20995 auto
20997 Automatic (same value as order).
20998 bottom
21000 Match from the bottom field.
21001 top Match from the top field.
21003 Default value is auto.
21005 mchroma
21007 Set whether or not chroma is included during the match comparisons.
21008 In most cases it is recommended to leave this enabled. You should
21009 set this to 0 only if your clip has bad chroma problems such as
21010 heavy rainbowing or other artifacts. Setting this to 0 could also
21011 be used to speed things up at the cost of some accuracy.
21012 Default value is 1.
21014 y0
21016 y1 These define an exclusion band which excludes the lines between y0
21017 and y1 from being included in the field matching decision. An
21018 exclusion band can be used to ignore subtitles, a logo, or other
21019 things that may interfere with the matching. y0 sets the starting
21020 scan line and y1 sets the ending line; all lines in between y0 and
21021 y1 (including y0 and y1) will be ignored. Setting y0 and y1 to the
21022 same value will disable the feature. y0 and y1 defaults to 0.
21023 scthresh
21025 Set the scene change detection threshold as a percentage of maximum
21026 change on the luma plane. Good values are in the "[8.0, 14.0]"
21027 range. Scene change detection is only relevant in case
21028 combmatch=sc. The range for scthresh is "[0.0, 100.0]".
21029 Default value is 12.0.
21031 combmatch
21033 When combatch is not none, "fieldmatch" will take into account the
21034 combed scores of matches when deciding what match to use as the
21035 final match. Available values are:
21036 none
21038 No final matching based on combed scores.
21039 sc Combed scores are only used when a scene change is detected.
21041 full
21043 Use combed scores all the time.
21044 Default is sc.
21046 combdbg
21048 Force "fieldmatch" to calculate the combed metrics for certain
21049 matches and print them. This setting is known as micout in TFM/VFM
21050 vocabulary. Available values are:
21051 none
21053 No forced calculation.
21054 pcn Force p/c/n calculations.
21056 pcnub
21058 Force p/c/n/u/b calculations.
21059 Default value is none.
21061 cthresh
21063 This is the area combing threshold used for combed frame detection.
21064 This essentially controls how "strong" or "visible" combing must be
21065 to be detected. Larger values mean combing must be more visible
21066 and smaller values mean combing can be less visible or strong and
21067 still be detected. Valid settings are from -1 (every pixel will be
21068 detected as combed) to 255 (no pixel will be detected as combed).
21069 This is basically a pixel difference value. A good range is "[8,
21070 12]".
21071 Default value is 9.
21073 chroma
21075 Sets whether or not chroma is considered in the combed frame
21076 decision. Only disable this if your source has chroma problems
21077 (rainbowing, etc.) that are causing problems for the combed frame
21078 detection with chroma enabled. Actually, using chroma=0 is usually
21079 more reliable, except for the case where there is chroma only
21080 combing in the source.
21081 Default value is 0.
21083 blockx
21085 blocky
21086 Respectively set the x-axis and y-axis size of the window used
21087 during combed frame detection. This has to do with the size of the
21088 area in which combpel pixels are required to be detected as combed
21089 for a frame to be declared combed. See the combpel parameter
21090 description for more info. Possible values are any number that is
21091 a power of 2 starting at 4 and going up to 512.
21092 Default value is 16.
21094 combpel
21096 The number of combed pixels inside any of the blocky by blockx size
21097 blocks on the frame for the frame to be detected as combed. While
21098 cthresh controls how "visible" the combing must be, this setting
21099 controls "how much" combing there must be in any localized area (a
21100 window defined by the blockx and blocky settings) on the frame.
21101 Minimum value is 0 and maximum is "blocky x blockx" (at which point
21102 no frames will ever be detected as combed). This setting is known
21103 as MI in TFM/VFM vocabulary.
21104 Default value is 80.
21106 p/c/n/u/b meaning
21108 p/c/n
21110 We assume the following telecined stream:
21112 Top fields: 1 2 2 3 4
21114 Bottom fields: 1 2 3 4 4
21115 The numbers correspond to the progressive frame the fields relate to.
21117 Here, the first two frames are progressive, the 3rd and 4th are combed,
21118 and so on.
21119 When "fieldmatch" is configured to run a matching from bottom
21121 (field=bottom) this is how this input stream get transformed:
21122 Input stream:
21124 T 1 2 2 3 4
21125 B 1 2 3 4 4 <-- matching reference
21126 Matches: c c n n c
21128 Output stream:
21130 T 1 2 3 4 4
21131 B 1 2 3 4 4
21132 As a result of the field matching, we can see that some frames get
21134 duplicated. To perform a complete inverse telecine, you need to rely
21135 on a decimation filter after this operation. See for instance the
21136 decimate filter.
21137 The same operation now matching from top fields (field=top) looks like
21139 this:
21140 Input stream:
21142 T 1 2 2 3 4 <-- matching reference
21143 B 1 2 3 4 4
21144 Matches: c c p p c
21146 Output stream:
21148 T 1 2 2 3 4
21149 B 1 2 2 3 4
21150 In these examples, we can see what p, c and n mean; basically, they
21152 refer to the frame and field of the opposite parity:
21153 *<p matches the field of the opposite parity in the previous frame>
21155 *<c matches the field of the opposite parity in the current frame>
21156 *<n matches the field of the opposite parity in the next frame>
21157 u/b
21159 The u and b matching are a bit special in the sense that they match
21161 from the opposite parity flag. In the following examples, we assume
21162 that we are currently matching the 2nd frame (Top:2, bottom:2).
21163 According to the match, a 'x' is placed above and below each matched
21164 fields.
21165 With bottom matching (field=bottom):
21167 Match: c p n b u
21169 x x x x x
21171 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
21172 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
21173 x x x x x
21174 Output frames:
21176 2 1 2 2 2
21177 2 2 2 1 3
21178 With top matching (field=top):
21180 Match: c p n b u
21182 x x x x x
21184 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
21185 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
21186 x x x x x
21187 Output frames:
21189 2 2 2 1 2
21190 2 1 3 2 2
21191 Examples
21193 Simple IVTC of a top field first telecined stream:
21195 fieldmatch=order=tff:combmatch=none, decimate
21197 Advanced IVTC, with fallback on yadif for still combed frames:
21199 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
21201 fieldorder
21203 Transform the field order of the input video.
21204 It accepts the following parameters:
21206 order
21208 The output field order. Valid values are tff for top field first or
21209 bff for bottom field first.
21210 The default value is tff.
21212 The transformation is done by shifting the picture content up or down
21214 by one line, and filling the remaining line with appropriate picture
21215 content. This method is consistent with most broadcast field order
21216 converters.
21217 If the input video is not flagged as being interlaced, or it is already
21219 flagged as being of the required output field order, then this filter
21220 does not alter the incoming video.
21221 It is very useful when converting to or from PAL DV material, which is
21223 bottom field first.
21224 For example:
21226 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
21228 fifo, afifo
21230 Buffer input images and send them when they are requested.
21231 It is mainly useful when auto-inserted by the libavfilter framework.
21233 It does not take parameters.
21235 fillborders
21237 Fill borders of the input video, without changing video stream
21238 dimensions. Sometimes video can have garbage at the four edges and you
21239 may not want to crop video input to keep size multiple of some number.
21240 This filter accepts the following options:
21242 left
21244 Number of pixels to fill from left border.
21245 right
21247 Number of pixels to fill from right border.
21248 top Number of pixels to fill from top border.
21250 bottom
21252 Number of pixels to fill from bottom border.
21253 mode
21255 Set fill mode.
21256 It accepts the following values:
21258 smear
21260 fill pixels using outermost pixels
21261 mirror
21263 fill pixels using mirroring (half sample symmetric)
21264 fixed
21266 fill pixels with constant value
21267 reflect
21269 fill pixels using reflecting (whole sample symmetric)
21270 wrap
21272 fill pixels using wrapping
21273 fade
21275 fade pixels to constant value
21276 margins
21278 fill pixels at top and bottom with weighted averages pixels
21279 near borders
21280 Default is smear.
21282 color
21284 Set color for pixels in fixed or fade mode. Default is black.
21285 Commands
21287 This filter supports same commands as options. The command accepts the
21289 same syntax of the corresponding option.
21290 If the specified expression is not valid, it is kept at its current
21292 value.
21293 find_rect
21295 Find a rectangular object
21296 It accepts the following options:
21298 object
21300 Filepath of the object image, needs to be in gray8.
21301 threshold
21303 Detection threshold, default is 0.5.
21304 mipmaps
21306 Number of mipmaps, default is 3.
21307 xmin, ymin, xmax, ymax
21309 Specifies the rectangle in which to search.
21310 discard
21312 Discard frames where object is not detected. Default is disabled.
21313 Examples
21315 • Cover a rectangular object by the supplied image of a given video
21317 using ffmpeg:
21318 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
21320 floodfill
21322 Flood area with values of same pixel components with another values.
21323 It accepts the following options:
21325 x Set pixel x coordinate.
21327 y Set pixel y coordinate.
21329 s0 Set source #0 component value.
21331 s1 Set source #1 component value.
21333 s2 Set source #2 component value.
21335 s3 Set source #3 component value.
21337 d0 Set destination #0 component value.
21339 d1 Set destination #1 component value.
21341 d2 Set destination #2 component value.
21343 d3 Set destination #3 component value.
21345 format
21347 Convert the input video to one of the specified pixel formats.
21348 Libavfilter will try to pick one that is suitable as input to the next
21349 filter.
21350 It accepts the following parameters:
21352 pix_fmts
21354 A '|'-separated list of pixel format names, such as
21355 "pix_fmts=yuv420p|monow|rgb24".
21356 Examples
21358 • Convert the input video to the yuv420p format
21360 format=pix_fmts=yuv420p
21362 Convert the input video to any of the formats in the list
21364 format=pix_fmts=yuv420p|yuv444p|yuv410p
21366 fps
21368 Convert the video to specified constant frame rate by duplicating or
21369 dropping frames as necessary.
21370 It accepts the following parameters:
21372 fps The desired output frame rate. It accepts expressions containing
21374 the following constants:
21375 source_fps
21377 The input's frame rate
21378 ntsc
21380 NTSC frame rate of "30000/1001"
21381 pal PAL frame rate of 25.0
21383 film
21385 Film frame rate of 24.0
21386 ntsc_film
21388 NTSC-film frame rate of "24000/1001"
21389 The default is 25.
21391 start_time
21393 Assume the first PTS should be the given value, in seconds. This
21394 allows for padding/trimming at the start of stream. By default, no
21395 assumption is made about the first frame's expected PTS, so no
21396 padding or trimming is done. For example, this could be set to 0
21397 to pad the beginning with duplicates of the first frame if a video
21398 stream starts after the audio stream or to trim any frames with a
21399 negative PTS.
21400 round
21402 Timestamp (PTS) rounding method.
21403 Possible values are:
21405 zero
21407 round towards 0
21408 inf round away from 0
21410 down
21412 round towards -infinity
21413 up round towards +infinity
21415 near
21417 round to nearest
21418 The default is "near".
21420 eof_action
21422 Action performed when reading the last frame.
21423 Possible values are:
21425 round
21427 Use same timestamp rounding method as used for other frames.
21428 pass
21430 Pass through last frame if input duration has not been reached
21431 yet.
21432 The default is "round".
21434 Alternatively, the options can be specified as a flat string:
21436 fps[:start_time[:round]].
21437 See also the setpts filter.
21439 Examples
21441 • A typical usage in order to set the fps to 25:
21443 fps=fps=25
21445 • Sets the fps to 24, using abbreviation and rounding method to round
21447 to nearest:
21448 fps=fps=film:round=near
21450 framepack
21452 Pack two different video streams into a stereoscopic video, setting
21453 proper metadata on supported codecs. The two views should have the same
21454 size and framerate and processing will stop when the shorter video
21455 ends. Please note that you may conveniently adjust view properties with
21456 the scale and fps filters.
21457 It accepts the following parameters:
21459 format
21461 The desired packing format. Supported values are:
21462 sbs The views are next to each other (default).
21464 tab The views are on top of each other.
21466 lines
21468 The views are packed by line.
21469 columns
21471 The views are packed by column.
21472 frameseq
21474 The views are temporally interleaved.
21475 Some examples:
21477 # Convert left and right views into a frame-sequential video
21479 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
21480 # Convert views into a side-by-side video with the same output resolution as the input
21482 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
21483 framerate
21485 Change the frame rate by interpolating new video output frames from the
21486 source frames.
21487 This filter is not designed to function correctly with interlaced
21489 media. If you wish to change the frame rate of interlaced media then
21490 you are required to deinterlace before this filter and re-interlace
21491 after this filter.
21492 A description of the accepted options follows.
21494 fps Specify the output frames per second. This option can also be
21496 specified as a value alone. The default is 50.
21497 interp_start
21499 Specify the start of a range where the output frame will be created
21500 as a linear interpolation of two frames. The range is [0-255], the
21501 default is 15.
21502 interp_end
21504 Specify the end of a range where the output frame will be created
21505 as a linear interpolation of two frames. The range is [0-255], the
21506 default is 240.
21507 scene
21509 Specify the level at which a scene change is detected as a value
21510 between 0 and 100 to indicate a new scene; a low value reflects a
21511 low probability for the current frame to introduce a new scene,
21512 while a higher value means the current frame is more likely to be
21513 one. The default is 8.2.
21514 flags
21516 Specify flags influencing the filter process.
21517 Available value for flags is:
21519 scene_change_detect, scd
21521 Enable scene change detection using the value of the option
21522 scene. This flag is enabled by default.
21523 framestep
21525 Select one frame every N-th frame.
21526 This filter accepts the following option:
21528 step
21530 Select frame after every "step" frames. Allowed values are
21531 positive integers higher than 0. Default value is 1.
21532 freezedetect
21534 Detect frozen video.
21535 This filter logs a message and sets frame metadata when it detects that
21537 the input video has no significant change in content during a specified
21538 duration. Video freeze detection calculates the mean average absolute
21539 difference of all the components of video frames and compares it to a
21540 noise floor.
21541 The printed times and duration are expressed in seconds. The
21543 "lavfi.freezedetect.freeze_start" metadata key is set on the first
21544 frame whose timestamp equals or exceeds the detection duration and it
21545 contains the timestamp of the first frame of the freeze. The
21546 "lavfi.freezedetect.freeze_duration" and
21547 "lavfi.freezedetect.freeze_end" metadata keys are set on the first
21548 frame after the freeze.
21549 The filter accepts the following options:
21551 noise, n
21553 Set noise tolerance. Can be specified in dB (in case "dB" is
21554 appended to the specified value) or as a difference ratio between 0
21555 and 1. Default is -60dB, or 0.001.
21556 duration, d
21558 Set freeze duration until notification (default is 2 seconds).
21559 freezeframes
21561 Freeze video frames.
21562 This filter freezes video frames using frame from 2nd input.
21564 The filter accepts the following options:
21566 first
21568 Set number of first frame from which to start freeze.
21569 last
21571 Set number of last frame from which to end freeze.
21572 replace
21574 Set number of frame from 2nd input which will be used instead of
21575 replaced frames.
21576 frei0r
21578 Apply a frei0r effect to the input video.
21579 To enable the compilation of this filter, you need to install the
21581 frei0r header and configure FFmpeg with "--enable-frei0r".
21582 It accepts the following parameters:
21584 filter_name
21586 The name of the frei0r effect to load. If the environment variable
21587 FREI0R_PATH is defined, the frei0r effect is searched for in each
21588 of the directories specified by the colon-separated list in
21589 FREI0R_PATH. Otherwise, the standard frei0r paths are searched, in
21590 this order: HOME/.frei0r-1/lib/, /usr/local/lib/frei0r-1/,
21591 /usr/lib/frei0r-1/.
21592 filter_params
21594 A '|'-separated list of parameters to pass to the frei0r effect.
21595 A frei0r effect parameter can be a boolean (its value is either "y" or
21597 "n"), a double, a color (specified as R/G/B, where R, G, and B are
21598 floating point numbers between 0.0 and 1.0, inclusive) or a color
21599 description as specified in the "Color" section in the ffmpeg-utils
21600 manual, a position (specified as X/Y, where X and Y are floating point
21601 numbers) and/or a string.
21602 The number and types of parameters depend on the loaded effect. If an
21604 effect parameter is not specified, the default value is set.
21605 Examples
21607 • Apply the distort0r effect, setting the first two double
21609 parameters:
21610 frei0r=filter_name=distort0r:filter_params=0.5|0.01
21612 • Apply the colordistance effect, taking a color as the first
21614 parameter:
21615 frei0r=colordistance:0.2/0.3/0.4
21617 frei0r=colordistance:violet
21618 frei0r=colordistance:0x112233
21619 • Apply the perspective effect, specifying the top left and top right
21621 image positions:
21622 frei0r=perspective:0.2/0.2|0.8/0.2
21624 For more information, see <http://frei0r.dyne.org>
21626 Commands
21628 This filter supports the filter_params option as commands.
21630 fspp
21632 Apply fast and simple postprocessing. It is a faster version of spp.
21633 It splits (I)DCT into horizontal/vertical passes. Unlike the simple
21635 post- processing filter, one of them is performed once per block, not
21636 per pixel. This allows for much higher speed.
21637 The filter accepts the following options:
21639 quality
21641 Set quality. This option defines the number of levels for
21642 averaging. It accepts an integer in the range 4-5. Default value is
21643 4.
21644 qp Force a constant quantization parameter. It accepts an integer in
21646 range 0-63. If not set, the filter will use the QP from the video
21647 stream (if available).
21648 strength
21650 Set filter strength. It accepts an integer in range -15 to 32.
21651 Lower values mean more details but also more artifacts, while
21652 higher values make the image smoother but also blurrier. Default
21653 value is 0 − PSNR optimal.
21654 use_bframe_qp
21656 Enable the use of the QP from the B-Frames if set to 1. Using this
21657 option may cause flicker since the B-Frames have often larger QP.
21658 Default is 0 (not enabled).
21659 gblur
21661 Apply Gaussian blur filter.
21662 The filter accepts the following options:
21664 sigma
21666 Set horizontal sigma, standard deviation of Gaussian blur. Default
21667 is 0.5.
21668 steps
21670 Set number of steps for Gaussian approximation. Default is 1.
21671 planes
21673 Set which planes to filter. By default all planes are filtered.
21674 sigmaV
21676 Set vertical sigma, if negative it will be same as "sigma".
21677 Default is -1.
21678 Commands
21680 This filter supports same commands as options. The command accepts the
21682 same syntax of the corresponding option.
21683 If the specified expression is not valid, it is kept at its current
21685 value.
21686 geq
21688 Apply generic equation to each pixel.
21689 The filter accepts the following options:
21691 lum_expr, lum
21693 Set the luminance expression.
21694 cb_expr, cb
21696 Set the chrominance blue expression.
21697 cr_expr, cr
21699 Set the chrominance red expression.
21700 alpha_expr, a
21702 Set the alpha expression.
21703 red_expr, r
21705 Set the red expression.
21706 green_expr, g
21708 Set the green expression.
21709 blue_expr, b
21711 Set the blue expression.
21712 The colorspace is selected according to the specified options. If one
21714 of the lum_expr, cb_expr, or cr_expr options is specified, the filter
21715 will automatically select a YCbCr colorspace. If one of the red_expr,
21716 green_expr, or blue_expr options is specified, it will select an RGB
21717 colorspace.
21718 If one of the chrominance expression is not defined, it falls back on
21720 the other one. If no alpha expression is specified it will evaluate to
21721 opaque value. If none of chrominance expressions are specified, they
21722 will evaluate to the luminance expression.
21723 The expressions can use the following variables and functions:
21725 N The sequential number of the filtered frame, starting from 0.
21727 X
21729 Y The coordinates of the current sample.
21730 W
21732 H The width and height of the image.
21733 SW
21735 SH Width and height scale depending on the currently filtered plane.
21736 It is the ratio between the corresponding luma plane number of
21737 pixels and the current plane ones. E.g. for YUV4:2:0 the values are
21738 "1,1" for the luma plane, and "0.5,0.5" for chroma planes.
21739 T Time of the current frame, expressed in seconds.
21741 p(x, y)
21743 Return the value of the pixel at location (x,y) of the current
21744 plane.
21745 lum(x, y)
21747 Return the value of the pixel at location (x,y) of the luminance
21748 plane.
21749 cb(x, y)
21751 Return the value of the pixel at location (x,y) of the blue-
21752 difference chroma plane. Return 0 if there is no such plane.
21753 cr(x, y)
21755 Return the value of the pixel at location (x,y) of the red-
21756 difference chroma plane. Return 0 if there is no such plane.
21757 r(x, y)
21759 g(x, y)
21760 b(x, y)
21761 Return the value of the pixel at location (x,y) of the
21762 red/green/blue component. Return 0 if there is no such component.
21763 alpha(x, y)
21765 Return the value of the pixel at location (x,y) of the alpha plane.
21766 Return 0 if there is no such plane.
21767 psum(x,y), lumsum(x, y), cbsum(x,y), crsum(x,y), rsum(x,y), gsum(x,y),
21769 bsum(x,y), alphasum(x,y)
21770 Sum of sample values in the rectangle from (0,0) to (x,y), this
21771 allows obtaining sums of samples within a rectangle. See the
21772 functions without the sum postfix.
21773 interpolation
21775 Set one of interpolation methods:
21776 nearest, n
21778 bilinear, b
21779 Default is bilinear.
21781 For functions, if x and y are outside the area, the value will be
21783 automatically clipped to the closer edge.
21784 Please note that this filter can use multiple threads in which case
21786 each slice will have its own expression state. If you want to use only
21787 a single expression state because your expressions depend on previous
21788 state then you should limit the number of filter threads to 1.
21789 Examples
21791 • Flip the image horizontally:
21793 geq=p(W-X\,Y)
21795 • Generate a bidimensional sine wave, with angle "PI/3" and a
21797 wavelength of 100 pixels:
21798 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
21800 • Generate a fancy enigmatic moving light:
21802 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
21804 • Generate a quick emboss effect:
21806 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
21808 • Modify RGB components depending on pixel position:
21810 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
21812 • Create a radial gradient that is the same size as the input (also
21814 see the vignette filter):
21815 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
21817 gradfun
21819 Fix the banding artifacts that are sometimes introduced into nearly
21820 flat regions by truncation to 8-bit color depth. Interpolate the
21821 gradients that should go where the bands are, and dither them.
21822 It is designed for playback only. Do not use it prior to lossy
21824 compression, because compression tends to lose the dither and bring
21825 back the bands.
21826 It accepts the following parameters:
21828 strength
21830 The maximum amount by which the filter will change any one pixel.
21831 This is also the threshold for detecting nearly flat regions.
21832 Acceptable values range from .51 to 64; the default value is 1.2.
21833 Out-of-range values will be clipped to the valid range.
21834 radius
21836 The neighborhood to fit the gradient to. A larger radius makes for
21837 smoother gradients, but also prevents the filter from modifying the
21838 pixels near detailed regions. Acceptable values are 8-32; the
21839 default value is 16. Out-of-range values will be clipped to the
21840 valid range.
21841 Alternatively, the options can be specified as a flat string:
21843 strength[:radius]
21844 Examples
21846 • Apply the filter with a 3.5 strength and radius of 8:
21848 gradfun=3.5:8
21850 • Specify radius, omitting the strength (which will fall-back to the
21852 default value):
21853 gradfun=radius=8
21855 graphmonitor
21857 Show various filtergraph stats.
21858 With this filter one can debug complete filtergraph. Especially issues
21860 with links filling with queued frames.
21861 The filter accepts the following options:
21863 size, s
21865 Set video output size. Default is hd720.
21866 opacity, o
21868 Set video opacity. Default is 0.9. Allowed range is from 0 to 1.
21869 mode, m
21871 Set output mode, can be fulll or compact. In compact mode only
21872 filters with some queued frames have displayed stats.
21873 flags, f
21875 Set flags which enable which stats are shown in video.
21876 Available values for flags are:
21878 queue
21880 Display number of queued frames in each link.
21881 frame_count_in
21883 Display number of frames taken from filter.
21884 frame_count_out
21886 Display number of frames given out from filter.
21887 frame_count_delta
21889 Display delta number of frames between above two values.
21890 pts Display current filtered frame pts.
21892 pts_delta
21894 Display pts delta between current and previous frame.
21895 time
21897 Display current filtered frame time.
21898 time_delta
21900 Display time delta between current and previous frame.
21901 timebase
21903 Display time base for filter link.
21904 format
21906 Display used format for filter link.
21907 size
21909 Display video size or number of audio channels in case of audio
21910 used by filter link.
21911 rate
21913 Display video frame rate or sample rate in case of audio used
21914 by filter link.
21915 eof Display link output status.
21917 sample_count_in
21919 Display number of samples taken from filter.
21920 sample_count_out
21922 Display number of samples given out from filter.
21923 sample_count_delta
21925 Display delta number of samples between above two values.
21926 rate, r
21928 Set upper limit for video rate of output stream, Default value is
21929 25. This guarantee that output video frame rate will not be higher
21930 than this value.
21931 grayworld
21933 A color constancy filter that applies color correction based on the
21934 grayworld assumption
21935 See:
21937 <https://www.researchgate.net/publication/275213614_A_New_Color_Correction_Method_for_Underwater_Imaging>
21938 The algorithm uses linear light, so input data should be linearized
21940 beforehand (and possibly correctly tagged).
21941 ffmpeg -i INPUT -vf zscale=transfer=linear,grayworld,zscale=transfer=bt709,format=yuv420p OUTPUT
21943 greyedge
21945 A color constancy variation filter which estimates scene illumination
21946 via grey edge algorithm and corrects the scene colors accordingly.
21947 See: <https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf>
21949 The filter accepts the following options:
21951 difford
21953 The order of differentiation to be applied on the scene. Must be
21954 chosen in the range [0,2] and default value is 1.
21955 minknorm
21957 The Minkowski parameter to be used for calculating the Minkowski
21958 distance. Must be chosen in the range [0,20] and default value is
21959 1. Set to 0 for getting max value instead of calculating Minkowski
21960 distance.
21961 sigma
21963 The standard deviation of Gaussian blur to be applied on the scene.
21964 Must be chosen in the range [0,1024.0] and default value = 1.
21965 floor( sigma * break_off_sigma(3) ) can't be equal to 0 if difford
21966 is greater than 0.
21967 Examples
21969 • Grey Edge:
21971 greyedge=difford=1:minknorm=5:sigma=2
21973 • Max Edge:
21975 greyedge=difford=1:minknorm=0:sigma=2
21977 guided
21979 Apply guided filter for edge-preserving smoothing, dehazing and so on.
21980 The filter accepts the following options:
21982 radius
21984 Set the box radius in pixels. Allowed range is 1 to 20. Default is
21985 3.
21986 eps Set regularization parameter (with square). Allowed range is 0 to
21988 1. Default is 0.01.
21989 mode
21991 Set filter mode. Can be "basic" or "fast". Default is "basic".
21992 sub Set subsampling ratio for "fast" mode. Range is 2 to 64. Default
21994 is 4. No subsampling occurs in "basic" mode.
21995 guidance
21997 Set guidance mode. Can be "off" or "on". Default is "off". If
21998 "off", single input is required. If "on", two inputs of the same
21999 resolution and pixel format are required. The second input serves
22000 as the guidance.
22001 planes
22003 Set planes to filter. Default is first only.
22004 Commands
22006 This filter supports the all above options as commands.
22008 Examples
22010 • Edge-preserving smoothing with guided filter:
22012 ffmpeg -i in.png -vf guided out.png
22014 • Dehazing, structure-transferring filtering, detail enhancement with
22016 guided filter. For the generation of guidance image, refer to
22017 paper "Guided Image Filtering". See:
22018 <http://kaiminghe.com/publications/pami12guidedfilter.pdf>.
22019 ffmpeg -i in.png -i guidance.png -filter_complex guided=guidance=on out.png
22021 haldclut
22023 Apply a Hald CLUT to a video stream.
22024 First input is the video stream to process, and second one is the Hald
22026 CLUT. The Hald CLUT input can be a simple picture or a complete video
22027 stream.
22028 The filter accepts the following options:
22030 clut
22032 Set which CLUT video frames will be processed from second input
22033 stream, can be first or all. Default is all.
22034 shortest
22036 Force termination when the shortest input terminates. Default is 0.
22037 repeatlast
22039 Continue applying the last CLUT after the end of the stream. A
22040 value of 0 disable the filter after the last frame of the CLUT is
22041 reached. Default is 1.
22042 "haldclut" also has the same interpolation options as lut3d (both
22044 filters share the same internals).
22045 This filter also supports the framesync options.
22047 More information about the Hald CLUT can be found on Eskil Steenberg's
22049 website (Hald CLUT author) at
22050 <http://www.quelsolaar.com/technology/clut.html>.
22051 Commands
22053 This filter supports the "interp" option as commands.
22055 Workflow examples
22057 Hald CLUT video stream
22059 Generate an identity Hald CLUT stream altered with various effects:
22061 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
22063 Note: make sure you use a lossless codec.
22065 Then use it with "haldclut" to apply it on some random stream:
22067 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
22069 The Hald CLUT will be applied to the 10 first seconds (duration of
22071 clut.nut), then the latest picture of that CLUT stream will be applied
22072 to the remaining frames of the "mandelbrot" stream.
22073 Hald CLUT with preview
22075 A Hald CLUT is supposed to be a squared image of "Level*Level*Level" by
22077 "Level*Level*Level" pixels. For a given Hald CLUT, FFmpeg will select
22078 the biggest possible square starting at the top left of the picture.
22079 The remaining padding pixels (bottom or right) will be ignored. This
22080 area can be used to add a preview of the Hald CLUT.
22081 Typically, the following generated Hald CLUT will be supported by the
22083 "haldclut" filter:
22084 ffmpeg -f lavfi -i B<haldclutsrc>=8 -vf "
22086 pad=iw+320 [padded_clut];
22087 smptebars=s=320x256, split [a][b];
22088 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
22089 [main][b] overlay=W-320" -frames:v 1 clut.png
22090 It contains the original and a preview of the effect of the CLUT: SMPTE
22092 color bars are displayed on the right-top, and below the same color
22093 bars processed by the color changes.
22094 Then, the effect of this Hald CLUT can be visualized with:
22096 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
22098 hflip
22100 Flip the input video horizontally.
22101 For example, to horizontally flip the input video with ffmpeg:
22103 ffmpeg -i in.avi -vf "hflip" out.avi
22105 histeq
22107 This filter applies a global color histogram equalization on a per-
22108 frame basis.
22109 It can be used to correct video that has a compressed range of pixel
22111 intensities. The filter redistributes the pixel intensities to
22112 equalize their distribution across the intensity range. It may be
22113 viewed as an "automatically adjusting contrast filter". This filter is
22114 useful only for correcting degraded or poorly captured source video.
22115 The filter accepts the following options:
22117 strength
22119 Determine the amount of equalization to be applied. As the
22120 strength is reduced, the distribution of pixel intensities more-
22121 and-more approaches that of the input frame. The value must be a
22122 float number in the range [0,1] and defaults to 0.200.
22123 intensity
22125 Set the maximum intensity that can generated and scale the output
22126 values appropriately. The strength should be set as desired and
22127 then the intensity can be limited if needed to avoid washing-out.
22128 The value must be a float number in the range [0,1] and defaults to
22129 0.210.
22130 antibanding
22132 Set the antibanding level. If enabled the filter will randomly vary
22133 the luminance of output pixels by a small amount to avoid banding
22134 of the histogram. Possible values are "none", "weak" or "strong".
22135 It defaults to "none".
22136 histogram
22138 Compute and draw a color distribution histogram for the input video.
22139 The computed histogram is a representation of the color component
22141 distribution in an image.
22142 Standard histogram displays the color components distribution in an
22144 image. Displays color graph for each color component. Shows
22145 distribution of the Y, U, V, A or R, G, B components, depending on
22146 input format, in the current frame. Below each graph a color component
22147 scale meter is shown.
22148 The filter accepts the following options:
22150 level_height
22152 Set height of level. Default value is 200. Allowed range is [50,
22153 2048].
22154 scale_height
22156 Set height of color scale. Default value is 12. Allowed range is
22157 [0, 40].
22158 display_mode
22160 Set display mode. It accepts the following values:
22161 stack
22163 Per color component graphs are placed below each other.
22164 parade
22166 Per color component graphs are placed side by side.
22167 overlay
22169 Presents information identical to that in the "parade", except
22170 that the graphs representing color components are superimposed
22171 directly over one another.
22172 Default is "stack".
22174 levels_mode
22176 Set mode. Can be either "linear", or "logarithmic". Default is
22177 "linear".
22178 components
22180 Set what color components to display. Default is 7.
22181 fgopacity
22183 Set foreground opacity. Default is 0.7.
22184 bgopacity
22186 Set background opacity. Default is 0.5.
22187 colors_mode
22189 Set colors mode. It accepts the following values:
22190 whiteonblack
22192 blackonwhite
22193 whiteongray
22194 blackongray
22195 coloronblack
22196 coloronwhite
22197 colorongray
22198 blackoncolor
22199 whiteoncolor
22200 grayoncolor
22201 Default is "whiteonblack".
22203 Examples
22205 • Calculate and draw histogram:
22207 ffplay -i input -vf histogram
22209 hqdn3d
22211 This is a high precision/quality 3d denoise filter. It aims to reduce
22212 image noise, producing smooth images and making still images really
22213 still. It should enhance compressibility.
22214 It accepts the following optional parameters:
22216 luma_spatial
22218 A non-negative floating point number which specifies spatial luma
22219 strength. It defaults to 4.0.
22220 chroma_spatial
22222 A non-negative floating point number which specifies spatial chroma
22223 strength. It defaults to 3.0*luma_spatial/4.0.
22224 luma_tmp
22226 A floating point number which specifies luma temporal strength. It
22227 defaults to 6.0*luma_spatial/4.0.
22228 chroma_tmp
22230 A floating point number which specifies chroma temporal strength.
22231 It defaults to luma_tmp*chroma_spatial/luma_spatial.
22232 Commands
22234 This filter supports same commands as options. The command accepts the
22236 same syntax of the corresponding option.
22237 If the specified expression is not valid, it is kept at its current
22239 value.
22240 hwdownload
22242 Download hardware frames to system memory.
22243 The input must be in hardware frames, and the output a non-hardware
22245 format. Not all formats will be supported on the output - it may be
22246 necessary to insert an additional format filter immediately following
22247 in the graph to get the output in a supported format.
22248 hwmap
22250 Map hardware frames to system memory or to another device.
22251 This filter has several different modes of operation; which one is used
22253 depends on the input and output formats:
22254 • Hardware frame input, normal frame output
22256 Map the input frames to system memory and pass them to the output.
22258 If the original hardware frame is later required (for example,
22259 after overlaying something else on part of it), the hwmap filter
22260 can be used again in the next mode to retrieve it.
22261 • Normal frame input, hardware frame output
22263 If the input is actually a software-mapped hardware frame, then
22265 unmap it - that is, return the original hardware frame.
22266 Otherwise, a device must be provided. Create new hardware surfaces
22268 on that device for the output, then map them back to the software
22269 format at the input and give those frames to the preceding filter.
22270 This will then act like the hwupload filter, but may be able to
22271 avoid an additional copy when the input is already in a compatible
22272 format.
22273 • Hardware frame input and output
22275 A device must be supplied for the output, either directly or with
22277 the derive_device option. The input and output devices must be of
22278 different types and compatible - the exact meaning of this is
22279 system-dependent, but typically it means that they must refer to
22280 the same underlying hardware context (for example, refer to the
22281 same graphics card).
22282 If the input frames were originally created on the output device,
22284 then unmap to retrieve the original frames.
22285 Otherwise, map the frames to the output device - create new
22287 hardware frames on the output corresponding to the frames on the
22288 input.
22289 The following additional parameters are accepted:
22291 mode
22293 Set the frame mapping mode. Some combination of:
22294 read
22296 The mapped frame should be readable.
22297 write
22299 The mapped frame should be writeable.
22300 overwrite
22302 The mapping will always overwrite the entire frame.
22303 This may improve performance in some cases, as the original
22305 contents of the frame need not be loaded.
22306 direct
22308 The mapping must not involve any copying.
22309 Indirect mappings to copies of frames are created in some cases
22311 where either direct mapping is not possible or it would have
22312 unexpected properties. Setting this flag ensures that the
22313 mapping is direct and will fail if that is not possible.
22314 Defaults to read+write if not specified.
22316 derive_device type
22318 Rather than using the device supplied at initialisation, instead
22319 derive a new device of type type from the device the input frames
22320 exist on.
22321 reverse
22323 In a hardware to hardware mapping, map in reverse - create frames
22324 in the sink and map them back to the source. This may be necessary
22325 in some cases where a mapping in one direction is required but only
22326 the opposite direction is supported by the devices being used.
22327 This option is dangerous - it may break the preceding filter in
22329 undefined ways if there are any additional constraints on that
22330 filter's output. Do not use it without fully understanding the
22331 implications of its use.
22332 hwupload
22334 Upload system memory frames to hardware surfaces.
22335 The device to upload to must be supplied when the filter is
22337 initialised. If using ffmpeg, select the appropriate device with the
22338 -filter_hw_device option or with the derive_device option. The input
22339 and output devices must be of different types and compatible - the
22340 exact meaning of this is system-dependent, but typically it means that
22341 they must refer to the same underlying hardware context (for example,
22342 refer to the same graphics card).
22343 The following additional parameters are accepted:
22345 derive_device type
22347 Rather than using the device supplied at initialisation, instead
22348 derive a new device of type type from the device the input frames
22349 exist on.
22350 hwupload_cuda
22352 Upload system memory frames to a CUDA device.
22353 It accepts the following optional parameters:
22355 device
22357 The number of the CUDA device to use
22358 hqx
22360 Apply a high-quality magnification filter designed for pixel art. This
22361 filter was originally created by Maxim Stepin.
22362 It accepts the following option:
22364 n Set the scaling dimension: 2 for "hq2x", 3 for "hq3x" and 4 for
22366 "hq4x". Default is 3.
22367 hstack
22369 Stack input videos horizontally.
22370 All streams must be of same pixel format and of same height.
22372 Note that this filter is faster than using overlay and pad filter to
22374 create same output.
22375 The filter accepts the following option:
22377 inputs
22379 Set number of input streams. Default is 2.
22380 shortest
22382 If set to 1, force the output to terminate when the shortest input
22383 terminates. Default value is 0.
22384 hsvhold
22386 Turns a certain HSV range into gray values.
22387 This filter measures color difference between set HSV color in options
22389 and ones measured in video stream. Depending on options, output colors
22390 can be changed to be gray or not.
22391 The filter accepts the following options:
22393 hue Set the hue value which will be used in color difference
22395 calculation. Allowed range is from -360 to 360. Default value is
22396 0.
22397 sat Set the saturation value which will be used in color difference
22399 calculation. Allowed range is from -1 to 1. Default value is 0.
22400 val Set the value which will be used in color difference calculation.
22402 Allowed range is from -1 to 1. Default value is 0.
22403 similarity
22405 Set similarity percentage with the key color. Allowed range is
22406 from 0 to 1. Default value is 0.01.
22407 0.00001 matches only the exact key color, while 1.0 matches
22409 everything.
22410 blend
22412 Blend percentage. Allowed range is from 0 to 1. Default value is
22413 0.
22414 0.0 makes pixels either fully gray, or not gray at all.
22416 Higher values result in more gray pixels, with a higher gray pixel
22418 the more similar the pixels color is to the key color.
22419 hsvkey
22421 Turns a certain HSV range into transparency.
22422 This filter measures color difference between set HSV color in options
22424 and ones measured in video stream. Depending on options, output colors
22425 can be changed to transparent by adding alpha channel.
22426 The filter accepts the following options:
22428 hue Set the hue value which will be used in color difference
22430 calculation. Allowed range is from -360 to 360. Default value is
22431 0.
22432 sat Set the saturation value which will be used in color difference
22434 calculation. Allowed range is from -1 to 1. Default value is 0.
22435 val Set the value which will be used in color difference calculation.
22437 Allowed range is from -1 to 1. Default value is 0.
22438 similarity
22440 Set similarity percentage with the key color. Allowed range is
22441 from 0 to 1. Default value is 0.01.
22442 0.00001 matches only the exact key color, while 1.0 matches
22444 everything.
22445 blend
22447 Blend percentage. Allowed range is from 0 to 1. Default value is
22448 0.
22449 0.0 makes pixels either fully transparent, or not transparent at
22451 all.
22452 Higher values result in semi-transparent pixels, with a higher
22454 transparency the more similar the pixels color is to the key color.
22455 hue
22457 Modify the hue and/or the saturation of the input.
22458 It accepts the following parameters:
22460 h Specify the hue angle as a number of degrees. It accepts an
22462 expression, and defaults to "0".
22463 s Specify the saturation in the [-10,10] range. It accepts an
22465 expression and defaults to "1".
22466 H Specify the hue angle as a number of radians. It accepts an
22468 expression, and defaults to "0".
22469 b Specify the brightness in the [-10,10] range. It accepts an
22471 expression and defaults to "0".
22472 h and H are mutually exclusive, and can't be specified at the same
22474 time.
22475 The b, h, H and s option values are expressions containing the
22477 following constants:
22478 n frame count of the input frame starting from 0
22480 pts presentation timestamp of the input frame expressed in time base
22482 units
22483 r frame rate of the input video, NAN if the input frame rate is
22485 unknown
22486 t timestamp expressed in seconds, NAN if the input timestamp is
22488 unknown
22489 tb time base of the input video
22491 Examples
22493 • Set the hue to 90 degrees and the saturation to 1.0:
22495 hue=h=90:s=1
22497 • Same command but expressing the hue in radians:
22499 hue=H=PI/2:s=1
22501 • Rotate hue and make the saturation swing between 0 and 2 over a
22503 period of 1 second:
22504 hue="H=2*PI*t: s=sin(2*PI*t)+1"
22506 • Apply a 3 seconds saturation fade-in effect starting at 0:
22508 hue="s=min(t/3\,1)"
22510 The general fade-in expression can be written as:
22512 hue="s=min(0\, max((t-START)/DURATION\, 1))"
22514 • Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
22516 hue="s=max(0\, min(1\, (8-t)/3))"
22518 The general fade-out expression can be written as:
22520 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
22522 Commands
22524 This filter supports the following commands:
22526 b
22528 s
22529 h
22530 H Modify the hue and/or the saturation and/or brightness of the input
22531 video. The command accepts the same syntax of the corresponding
22532 option.
22533 If the specified expression is not valid, it is kept at its current
22535 value.
22536 huesaturation
22538 Apply hue-saturation-intensity adjustments to input video stream.
22539 This filter operates in RGB colorspace.
22541 This filter accepts the following options:
22543 hue Set the hue shift in degrees to apply. Default is 0. Allowed range
22545 is from -180 to 180.
22546 saturation
22548 Set the saturation shift. Default is 0. Allowed range is from -1
22549 to 1.
22550 intensity
22552 Set the intensity shift. Default is 0. Allowed range is from -1 to
22553 1.
22554 colors
22556 Set which primary and complementary colors are going to be
22557 adjusted. This options is set by providing one or multiple values.
22558 This can select multiple colors at once. By default all colors are
22559 selected.
22560 r Adjust reds.
22562 y Adjust yellows.
22564 g Adjust greens.
22566 c Adjust cyans.
22568 b Adjust blues.
22570 m Adjust magentas.
22572 a Adjust all colors.
22574 strength
22576 Set strength of filtering. Allowed range is from 0 to 100. Default
22577 value is 1.
22578 rw, gw, bw
22580 Set weight for each RGB component. Allowed range is from 0 to 1.
22581 By default is set to 0.333, 0.334, 0.333. Those options are used
22582 in saturation and lightess processing.
22583 lightness
22585 Set preserving lightness, by default is disabled. Adjusting hues
22586 can change lightness from original RGB triplet, with this option
22587 enabled lightness is kept at same value.
22588 hysteresis
22590 Grow first stream into second stream by connecting components. This
22591 makes it possible to build more robust edge masks.
22592 This filter accepts the following options:
22594 planes
22596 Set which planes will be processed as bitmap, unprocessed planes
22597 will be copied from first stream. By default value 0xf, all planes
22598 will be processed.
22599 threshold
22601 Set threshold which is used in filtering. If pixel component value
22602 is higher than this value filter algorithm for connecting
22603 components is activated. By default value is 0.
22604 The "hysteresis" filter also supports the framesync options.
22606 iccdetect
22608 Detect the colorspace from an embedded ICC profile (if present), and
22609 update the frame's tags accordingly.
22610 This filter accepts the following options:
22612 force
22614 If true, the frame's existing colorspace tags will always be
22615 overridden by values detected from an ICC profile. Otherwise, they
22616 will only be assigned if they contain "unknown". Enabled by
22617 default.
22618 iccgen
22620 Generate ICC profiles and attach them to frames.
22621 This filter accepts the following options:
22623 color_primaries
22625 color_trc
22626 Configure the colorspace that the ICC profile will be generated
22627 for. The default value of "auto" infers the value from the input
22628 frame's metadata, defaulting to BT.709/sRGB as appropriate.
22629 See the setparams filter for a list of possible values, but note
22631 that "unknown" are not valid values for this filter.
22632 force
22634 If true, an ICC profile will be generated even if it would
22635 overwrite an already existing ICC profile. Disabled by default.
22636 identity
22638 Obtain the identity score between two input videos.
22639 This filter takes two input videos.
22641 Both input videos must have the same resolution and pixel format for
22643 this filter to work correctly. Also it assumes that both inputs have
22644 the same number of frames, which are compared one by one.
22645 The obtained per component, average, min and max identity score is
22647 printed through the logging system.
22648 The filter stores the calculated identity scores of each frame in frame
22650 metadata.
22651 This filter also supports the framesync options.
22653 In the below example the input file main.mpg being processed is
22655 compared with the reference file ref.mpg.
22656 ffmpeg -i main.mpg -i ref.mpg -lavfi identity -f null -
22658 idet
22660 Detect video interlacing type.
22661 This filter tries to detect if the input frames are interlaced,
22663 progressive, top or bottom field first. It will also try to detect
22664 fields that are repeated between adjacent frames (a sign of telecine).
22665 Single frame detection considers only immediately adjacent frames when
22667 classifying each frame. Multiple frame detection incorporates the
22668 classification history of previous frames.
22669 The filter will log these metadata values:
22671 single.current_frame
22673 Detected type of current frame using single-frame detection. One
22674 of: ``tff'' (top field first), ``bff'' (bottom field first),
22675 ``progressive'', or ``undetermined''
22676 single.tff
22678 Cumulative number of frames detected as top field first using
22679 single-frame detection.
22680 multiple.tff
22682 Cumulative number of frames detected as top field first using
22683 multiple-frame detection.
22684 single.bff
22686 Cumulative number of frames detected as bottom field first using
22687 single-frame detection.
22688 multiple.current_frame
22690 Detected type of current frame using multiple-frame detection. One
22691 of: ``tff'' (top field first), ``bff'' (bottom field first),
22692 ``progressive'', or ``undetermined''
22693 multiple.bff
22695 Cumulative number of frames detected as bottom field first using
22696 multiple-frame detection.
22697 single.progressive
22699 Cumulative number of frames detected as progressive using single-
22700 frame detection.
22701 multiple.progressive
22703 Cumulative number of frames detected as progressive using multiple-
22704 frame detection.
22705 single.undetermined
22707 Cumulative number of frames that could not be classified using
22708 single-frame detection.
22709 multiple.undetermined
22711 Cumulative number of frames that could not be classified using
22712 multiple-frame detection.
22713 repeated.current_frame
22715 Which field in the current frame is repeated from the last. One of
22716 ``neither'', ``top'', or ``bottom''.
22717 repeated.neither
22719 Cumulative number of frames with no repeated field.
22720 repeated.top
22722 Cumulative number of frames with the top field repeated from the
22723 previous frame's top field.
22724 repeated.bottom
22726 Cumulative number of frames with the bottom field repeated from the
22727 previous frame's bottom field.
22728 The filter accepts the following options:
22730 intl_thres
22732 Set interlacing threshold.
22733 prog_thres
22735 Set progressive threshold.
22736 rep_thres
22738 Threshold for repeated field detection.
22739 half_life
22741 Number of frames after which a given frame's contribution to the
22742 statistics is halved (i.e., it contributes only 0.5 to its
22743 classification). The default of 0 means that all frames seen are
22744 given full weight of 1.0 forever.
22745 analyze_interlaced_flag
22747 When this is not 0 then idet will use the specified number of
22748 frames to determine if the interlaced flag is accurate, it will not
22749 count undetermined frames. If the flag is found to be accurate it
22750 will be used without any further computations, if it is found to be
22751 inaccurate it will be cleared without any further computations.
22752 This allows inserting the idet filter as a low computational method
22753 to clean up the interlaced flag
22754 il
22756 Deinterleave or interleave fields.
22757 This filter allows one to process interlaced images fields without
22759 deinterlacing them. Deinterleaving splits the input frame into 2 fields
22760 (so called half pictures). Odd lines are moved to the top half of the
22761 output image, even lines to the bottom half. You can process (filter)
22762 them independently and then re-interleave them.
22763 The filter accepts the following options:
22765 luma_mode, l
22767 chroma_mode, c
22768 alpha_mode, a
22769 Available values for luma_mode, chroma_mode and alpha_mode are:
22770 none
22772 Do nothing.
22773 deinterleave, d
22775 Deinterleave fields, placing one above the other.
22776 interleave, i
22778 Interleave fields. Reverse the effect of deinterleaving.
22779 Default value is "none".
22781 luma_swap, ls
22783 chroma_swap, cs
22784 alpha_swap, as
22785 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default
22786 value is 0.
22787 Commands
22789 This filter supports the all above options as commands.
22791 inflate
22793 Apply inflate effect to the video.
22794 This filter replaces the pixel by the local(3x3) average by taking into
22796 account only values higher than the pixel.
22797 It accepts the following options:
22799 threshold0
22801 threshold1
22802 threshold2
22803 threshold3
22804 Limit the maximum change for each plane, default is 65535. If 0,
22805 plane will remain unchanged.
22806 Commands
22808 This filter supports the all above options as commands.
22810 interlace
22812 Simple interlacing filter from progressive contents. This interleaves
22813 upper (or lower) lines from odd frames with lower (or upper) lines from
22814 even frames, halving the frame rate and preserving image height.
22815 Original Original New Frame
22817 Frame 'j' Frame 'j+1' (tff)
22818 ========== =========== ==================
22819 Line 0 --------------------> Frame 'j' Line 0
22820 Line 1 Line 1 ----> Frame 'j+1' Line 1
22821 Line 2 ---------------------> Frame 'j' Line 2
22822 Line 3 Line 3 ----> Frame 'j+1' Line 3
22823 ... ... ...
22824 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
22825 It accepts the following optional parameters:
22827 scan
22829 This determines whether the interlaced frame is taken from the even
22830 (tff - default) or odd (bff) lines of the progressive frame.
22831 lowpass
22833 Vertical lowpass filter to avoid twitter interlacing and reduce
22834 moire patterns.
22835 0, off
22837 Disable vertical lowpass filter
22838 1, linear
22840 Enable linear filter (default)
22841 2, complex
22843 Enable complex filter. This will slightly less reduce twitter
22844 and moire but better retain detail and subjective sharpness
22845 impression.
22846 kerndeint
22848 Deinterlace input video by applying Donald Graft's adaptive kernel
22849 deinterling. Work on interlaced parts of a video to produce progressive
22850 frames.
22851 The description of the accepted parameters follows.
22853 thresh
22855 Set the threshold which affects the filter's tolerance when
22856 determining if a pixel line must be processed. It must be an
22857 integer in the range [0,255] and defaults to 10. A value of 0 will
22858 result in applying the process on every pixels.
22859 map Paint pixels exceeding the threshold value to white if set to 1.
22861 Default is 0.
22862 order
22864 Set the fields order. Swap fields if set to 1, leave fields alone
22865 if 0. Default is 0.
22866 sharp
22868 Enable additional sharpening if set to 1. Default is 0.
22869 twoway
22871 Enable twoway sharpening if set to 1. Default is 0.
22872 Examples
22874 • Apply default values:
22876 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
22878 • Enable additional sharpening:
22880 kerndeint=sharp=1
22882 • Paint processed pixels in white:
22884 kerndeint=map=1
22886 kirsch
22888 Apply kirsch operator to input video stream.
22889 The filter accepts the following option:
22891 planes
22893 Set which planes will be processed, unprocessed planes will be
22894 copied. By default value 0xf, all planes will be processed.
22895 scale
22897 Set value which will be multiplied with filtered result.
22898 delta
22900 Set value which will be added to filtered result.
22901 Commands
22903 This filter supports the all above options as commands.
22905 lagfun
22907 Slowly update darker pixels.
22908 This filter makes short flashes of light appear longer. This filter
22910 accepts the following options:
22911 decay
22913 Set factor for decaying. Default is .95. Allowed range is from 0 to
22914 1.
22915 planes
22917 Set which planes to filter. Default is all. Allowed range is from 0
22918 to 15.
22919 Commands
22921 This filter supports the all above options as commands.
22923 lenscorrection
22925 Correct radial lens distortion
22926 This filter can be used to correct for radial distortion as can result
22928 from the use of wide angle lenses, and thereby re-rectify the image. To
22929 find the right parameters one can use tools available for example as
22930 part of opencv or simply trial-and-error. To use opencv use the
22931 calibration sample (under samples/cpp) from the opencv sources and
22932 extract the k1 and k2 coefficients from the resulting matrix.
22933 Note that effectively the same filter is available in the open-source
22935 tools Krita and Digikam from the KDE project.
22936 In contrast to the vignette filter, which can also be used to
22938 compensate lens errors, this filter corrects the distortion of the
22939 image, whereas vignette corrects the brightness distribution, so you
22940 may want to use both filters together in certain cases, though you will
22941 have to take care of ordering, i.e. whether vignetting should be
22942 applied before or after lens correction.
22943 Options
22945 The filter accepts the following options:
22947 cx Relative x-coordinate of the focal point of the image, and thereby
22949 the center of the distortion. This value has a range [0,1] and is
22950 expressed as fractions of the image width. Default is 0.5.
22951 cy Relative y-coordinate of the focal point of the image, and thereby
22953 the center of the distortion. This value has a range [0,1] and is
22954 expressed as fractions of the image height. Default is 0.5.
22955 k1 Coefficient of the quadratic correction term. This value has a
22957 range [-1,1]. 0 means no correction. Default is 0.
22958 k2 Coefficient of the double quadratic correction term. This value has
22960 a range [-1,1]. 0 means no correction. Default is 0.
22961 i Set interpolation type. Can be "nearest" or "bilinear". Default is
22963 "nearest".
22964 fc Specify the color of the unmapped pixels. For the syntax of this
22966 option, check the "Color" section in the ffmpeg-utils manual.
22967 Default color is "black@0".
22968 The formula that generates the correction is:
22970 r_src = r_tgt * (1 + k1 * (r_tgt / r_0)^2 + k2 * (r_tgt / r_0)^4)
22972 where r_0 is halve of the image diagonal and r_src and r_tgt are the
22974 distances from the focal point in the source and target images,
22975 respectively.
22976 Commands
22978 This filter supports the all above options as commands.
22980 lensfun
22982 Apply lens correction via the lensfun library
22983 (<http://lensfun.sourceforge.net/>).
22984 The "lensfun" filter requires the camera make, camera model, and lens
22986 model to apply the lens correction. The filter will load the lensfun
22987 database and query it to find the corresponding camera and lens entries
22988 in the database. As long as these entries can be found with the given
22989 options, the filter can perform corrections on frames. Note that
22990 incomplete strings will result in the filter choosing the best match
22991 with the given options, and the filter will output the chosen camera
22992 and lens models (logged with level "info"). You must provide the make,
22993 camera model, and lens model as they are required.
22994 To obtain a list of available makes and models, leave out one or both
22996 of "make" and "model" options. The filter will send the full list to
22997 the log with level "INFO". The first column is the make and the second
22998 column is the model. To obtain a list of available lenses, set any
22999 values for make and model and leave out the "lens_model" option. The
23000 filter will send the full list of lenses in the log with level "INFO".
23001 The ffmpeg tool will exit after the list is printed.
23002 The filter accepts the following options:
23004 make
23006 The make of the camera (for example, "Canon"). This option is
23007 required.
23008 model
23010 The model of the camera (for example, "Canon EOS 100D"). This
23011 option is required.
23012 lens_model
23014 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6
23015 IS STM"). This option is required.
23016 db_path
23018 The full path to the lens database folder. If not set, the filter
23019 will attempt to load the database from the install path when the
23020 library was built. Default is unset.
23021 mode
23023 The type of correction to apply. The following values are valid
23024 options:
23025 vignetting
23027 Enables fixing lens vignetting.
23028 geometry
23030 Enables fixing lens geometry. This is the default.
23031 subpixel
23033 Enables fixing chromatic aberrations.
23034 vig_geo
23036 Enables fixing lens vignetting and lens geometry.
23037 vig_subpixel
23039 Enables fixing lens vignetting and chromatic aberrations.
23040 distortion
23042 Enables fixing both lens geometry and chromatic aberrations.
23043 all Enables all possible corrections.
23045 focal_length
23047 The focal length of the image/video (zoom; expected constant for
23048 video). For example, a 18--55mm lens has focal length range of
23049 [18--55], so a value in that range should be chosen when using that
23050 lens. Default 18.
23051 aperture
23053 The aperture of the image/video (expected constant for video). Note
23054 that aperture is only used for vignetting correction. Default 3.5.
23055 focus_distance
23057 The focus distance of the image/video (expected constant for
23058 video). Note that focus distance is only used for vignetting and
23059 only slightly affects the vignetting correction process. If
23060 unknown, leave it at the default value (which is 1000).
23061 scale
23063 The scale factor which is applied after transformation. After
23064 correction the video is no longer necessarily rectangular. This
23065 parameter controls how much of the resulting image is visible. The
23066 value 0 means that a value will be chosen automatically such that
23067 there is little or no unmapped area in the output image. 1.0 means
23068 that no additional scaling is done. Lower values may result in more
23069 of the corrected image being visible, while higher values may avoid
23070 unmapped areas in the output.
23071 target_geometry
23073 The target geometry of the output image/video. The following values
23074 are valid options:
23075 rectilinear (default)
23077 fisheye
23078 panoramic
23079 equirectangular
23080 fisheye_orthographic
23081 fisheye_stereographic
23082 fisheye_equisolid
23083 fisheye_thoby
23084 reverse
23085 Apply the reverse of image correction (instead of correcting
23086 distortion, apply it).
23087 interpolation
23089 The type of interpolation used when correcting distortion. The
23090 following values are valid options:
23091 nearest
23093 linear (default)
23094 lanczos
23095 Examples
23097 • Apply lens correction with make "Canon", camera model "Canon EOS
23099 100D", and lens model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with
23100 focal length of "18" and aperture of "8.0".
23101 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
23103 • Apply the same as before, but only for the first 5 seconds of
23105 video.
23106 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
23108 libplacebo
23110 Flexible GPU-accelerated processing filter based on libplacebo
23111 (<https://code.videolan.org/videolan/libplacebo>). Note that this
23112 filter currently only accepts Vulkan input frames.
23113 Options
23115 The options for this filter are divided into the following sections:
23117 Output mode
23119 These options control the overall output mode. By default, libplacebo
23121 will try to preserve the source colorimetry and size as best as it can,
23122 but it will apply any embedded film grain, dolby vision metadata or
23123 anamorphic SAR present in source frames.
23124 w
23126 h Set the output video dimension expression. Default value is the
23127 input dimension.
23128 Allows for the same expressions as the scale filter.
23130 format
23132 Set the output format override. If unset (the default), frames will
23133 be output in the same format as the respective input frames.
23134 Otherwise, format conversion will be performed.
23135 force_original_aspect_ratio
23137 force_divisible_by
23138 Work the same as the identical scale filter options.
23139 normalize_sar
23141 If enabled, output frames will always have a pixel aspect ratio of
23142 1:1. This will introduce padding/cropping as necessary. If disabled
23143 (the default), any aspect ratio mismatches, including those from
23144 e.g. anamorphic video sources, are forwarded to the output pixel
23145 aspect ratio.
23146 pad_crop_ratio
23148 Specifies a ratio (between 0.0 and 1.0) between padding and
23149 cropping when the input aspect ratio does not match the output
23150 aspect ratio and normalize_sar is in effect. The default of 0.0
23151 always pads the content with black borders, while a value of 1.0
23152 always crops off parts of the content. Intermediate values are
23153 possible, leading to a mix of the two approaches.
23154 colorspace
23156 color_primaries
23157 color_trc
23158 range
23159 Configure the colorspace that output frames will be delivered in.
23160 The default value of "auto" outputs frames in the same format as
23161 the input frames, leading to no change. For any other value,
23162 conversion will be performed.
23163 See the setparams filter for a list of possible values.
23165 apply_filmgrain
23167 Apply film grain (e.g. AV1 or H.274) if present in source frames,
23168 and strip it from the output. Enabled by default.
23169 apply_dolbyvision
23171 Apply Dolby Vision RPU metadata if present in source frames, and
23172 strip it from the output. Enabled by default. Note that Dolby
23173 Vision will always output BT.2020+PQ, overriding the usual input
23174 frame metadata. These will also be picked as the values of "auto"
23175 for the respective frame output options.
23176 Scaling
23178 The options in this section control how libplacebo performs upscaling
23180 and (if necessary) downscaling. Note that libplacebo will always
23181 internally operate on 4:4:4 content, so any sub-sampled chroma formats
23182 such as "yuv420p" will necessarily be upsampled and downsampled as part
23183 of the rendering process. That means scaling might be in effect even if
23184 the source and destination resolution are the same.
23185 upscaler
23187 downscaler
23188 Configure the filter kernel used for upscaling and downscaling. The
23189 respective defaults are "spline36" and "mitchell". For a full list
23190 of possible values, pass "help" to these options. The most
23191 important values are:
23192 none
23194 Forces the use of built-in GPU texture sampling (typically
23195 bilinear). Extremely fast but poor quality, especially when
23196 downscaling.
23197 bilinear
23199 Bilinear interpolation. Can generally be done for free on GPUs,
23200 except when doing so would lead to aliasing. Fast and low
23201 quality.
23202 nearest
23204 Nearest-neighbour interpolation. Sharp but highly aliasing.
23205 oversample
23207 Algorithm that looks visually similar to nearest-neighbour
23208 interpolation but tries to preserve pixel aspect ratio. Good
23209 for pixel art, since it results in minimal distortion of the
23210 artistic appearance.
23211 lanczos
23213 Standard sinc-sinc interpolation kernel.
23214 spline36
23216 Cubic spline approximation of lanczos. No difference in
23217 performance, but has very slightly less ringing.
23218 ewa_lanczos
23220 Elliptically weighted average version of lanczos, based on a
23221 jinc-sinc kernel. This is also popularly referred to as just
23222 "Jinc scaling". Slow but very high quality.
23223 gaussian
23225 Gaussian kernel. Has certain ideal mathematical properties, but
23226 subjectively very blurry.
23227 mitchell
23229 Cubic BC spline with parameters recommended by Mitchell and
23230 Netravali. Very little ringing.
23231 lut_entries
23233 Configures the size of scaler LUTs, ranging from 1 to 256. The
23234 default of 0 will pick libplacebo's internal default, typically 64.
23235 antiringing
23237 Enables anti-ringing (for non-EWA filters). The value (between 0.0
23238 and 1.0) configures the strength of the anti-ringing algorithm. May
23239 increase aliasing if set too high. Disabled by default.
23240 sigmoid
23242 Enable sigmoidal compression during upscaling. Reduces ringing
23243 slightly. Enabled by default.
23244 Debanding
23246 Libplacebo comes with a built-in debanding filter that is good at
23248 counteracting many common sources of banding and blocking. Turning this
23249 on is highly recommended whenever quality is desired.
23250 deband
23252 Enable (fast) debanding algorithm. Disabled by default.
23253 deband_iterations
23255 Number of deband iterations of the debanding algorithm. Each
23256 iteration is performed with progressively increased radius (and
23257 diminished threshold). Recommended values are in the range 1 to 4.
23258 Defaults to 1.
23259 deband_threshold
23261 Debanding filter strength. Higher numbers lead to more aggressive
23262 debanding. Defaults to 4.0.
23263 deband_radius
23265 Debanding filter radius. A higher radius is better for slow
23266 gradients, while a lower radius is better for steep gradients.
23267 Defaults to 16.0.
23268 deband_grain
23270 Amount of extra output grain to add. Helps hide imperfections.
23271 Defaults to 6.0.
23272 Color adjustment
23274 A collection of subjective color controls. Not very rigorous, so the
23276 exact effect will vary somewhat depending on the input primaries and
23277 colorspace.
23278 brightness
23280 Brightness boost, between -1.0 and 1.0. Defaults to 0.0.
23281 contrast
23283 Contrast gain, between 0.0 and 16.0. Defaults to 1.0.
23284 saturation
23286 Saturation gain, between 0.0 and 16.0. Defaults to 1.0.
23287 hue Hue shift in radians, between -3.14 and 3.14. Defaults to 0.0. This
23289 will rotate the UV subvector, defaulting to BT.709 coefficients for
23290 RGB inputs.
23291 gamma
23293 Gamma adjustment, between 0.0 and 16.0. Defaults to 1.0.
23294 cones
23296 Cone model to use for color blindness simulation. Accepts any
23297 combination of "l", "m" and "s". Here are some examples:
23298 m Deuteranomaly / deuteranopia (affecting 3%-4% of the
23300 population)
23301 l Protanomaly / protanopia (affecting 1%-2% of the population)
23303 l+m Monochromacy (very rare)
23305 l+m+s
23307 Achromatopsy (complete loss of daytime vision, extremely rare)
23308 cone-strength
23310 Gain factor for the cones specified by "cones", between 0.0 and
23311 10.0. A value of 1.0 results in no change to color vision. A value
23312 of 0.0 (the default) simulates complete loss of those cones. Values
23313 above 1.0 result in exaggerating the differences between cones,
23314 which may help compensate for reduced color vision.
23315 Peak detection
23317 To help deal with sources that only have static HDR10 metadata (or no
23319 tagging whatsoever), libplacebo uses its own internal frame analysis
23320 compute shader to analyze source frames and adapt the tone mapping
23321 function in realtime. If this is too slow, or if exactly reproducible
23322 frame-perfect results are needed, it's recommended to turn this feature
23323 off.
23324 peak_detect
23326 Enable HDR peak detection. Ignores static MaxCLL/MaxFALL values in
23327 favor of dynamic detection from the input. Note that the detected
23328 values do not get written back to the output frames, they merely
23329 guide the internal tone mapping process. Enabled by default.
23330 smoothing_period
23332 Peak detection smoothing period, between 0.0 and 1000.0. Higher
23333 values result in peak detection becoming less responsive to changes
23334 in the input. Defaults to 100.0.
23335 minimum_peak
23337 Lower bound on the detected peak (relative to SDR white), between
23338 0.0 and 100.0. Defaults to 1.0.
23339 scene_threshold_low
23341 scene_threshold_high
23342 Lower and upper thresholds for scene change detection. Expressed in
23343 a logarithmic scale between 0.0 and 100.0. Default to 5.5 and 10.0,
23344 respectively. Setting either to a negative value disables this
23345 functionality.
23346 overshoot
23348 Peak smoothing overshoot margin, between 0.0 and 1.0. Provides a
23349 safety margin to prevent clipping as a result of peak smoothing.
23350 Defaults to 0.05, corresponding to a margin of 5%.
23351 Tone mapping
23353 The options in this section control how libplacebo performs tone-
23355 mapping and gamut-mapping when dealing with mismatches between wide-
23356 gamut or HDR content. In general, libplacebo relies on accurate source
23357 tagging and mastering display gamut information to produce the best
23358 results.
23359 intent
23361 Rendering intent to use when adapting between different primary
23362 color gamuts (after tone-mapping).
23363 perceptual
23365 Perceptual gamut mapping. Currently equivalent to relative
23366 colorimetric.
23367 relative
23369 Relative colorimetric. This is the default.
23370 absolute
23372 Absolute colorimetric.
23373 saturation
23375 Saturation mapping. Forcibly stretches the source gamut to the
23376 target gamut.
23377 gamut_mode
23379 How to handle out-of-gamut colors that can occur as a result of
23380 colorimetric gamut mapping.
23381 clip
23383 Do nothing, simply clip out-of-range colors to the RGB volume.
23384 This is the default.
23385 warn
23387 Highlight out-of-gamut pixels (by coloring them pink).
23388 darken
23390 Linearly reduces content brightness to preserves saturated
23391 details, followed by clipping the remaining out-of-gamut
23392 colors. As the name implies, this makes everything darker, but
23393 provides a good balance between preserving details and colors.
23394 desaturate
23396 Hard-desaturates out-of-gamut colors towards white, while
23397 preserving the luminance. Has a tendency to shift colors.
23398 tonemapping
23400 Tone-mapping algorithm to use. Available values are:
23401 auto
23403 Automatic selection based on internal heuristics. This is the
23404 default.
23405 clip
23407 Performs no tone-mapping, just clips out-of-range colors.
23408 Retains perfect color accuracy for in-range colors but
23409 completely destroys out-of-range information. Does not perform
23410 any black point adaptation. Not configurable.
23411 st2094-40
23413 EETF from SMPTE ST 2094-40 Annex B, which applies the Bezier
23414 curves from HDR10+ dynamic metadata based on Bezier curves to
23415 perform tone-mapping. The OOTF used is adjusted based on the
23416 ratio between the targeted and actual display peak luminances.
23417 st2094-10
23419 EETF from SMPTE ST 2094-10 Annex B.2, which takes into account
23420 the input signal average luminance in addition to the
23421 maximum/minimum. The configurable contrast parameter influences
23422 the slope of the linear output segment, defaulting to 1.0 for
23423 no increase/decrease in contrast. Note that this does not
23424 currently include the subjective gain/offset/gamma controls
23425 defined in Annex B.3.
23426 bt.2390
23428 EETF from the ITU-R Report BT.2390, a hermite spline roll-off
23429 with linear segment. The knee point offset is configurable.
23430 Note that this parameter defaults to 1.0, rather than the value
23431 of 0.5 from the ITU-R spec.
23432 bt.2446a
23434 EETF from ITU-R Report BT.2446, method A. Designed for well-
23435 mastered HDR sources. Can be used for both forward and inverse
23436 tone mapping. Not configurable.
23437 spline
23439 Simple spline consisting of two polynomials, joined by a single
23440 pivot point. The parameter gives the pivot point (in PQ
23441 space), defaulting to 0.30. Can be used for both forward and
23442 inverse tone mapping.
23443 reinhard
23445 Simple non-linear, global tone mapping algorithm. The parameter
23446 specifies the local contrast coefficient at the display peak.
23447 Essentially, a parameter of 0.5 implies that the reference
23448 white will be about half as bright as when clipping. Defaults
23449 to 0.5, which results in the simplest formulation of this
23450 function.
23451 mobius
23453 Generalization of the reinhard tone mapping algorithm to
23454 support an additional linear slope near black. The tone mapping
23455 parameter indicates the trade-off between the linear section
23456 and the non-linear section. Essentially, for a given parameter
23457 x, every color value below x will be mapped linearly, while
23458 higher values get non-linearly tone-mapped. Values near 1.0
23459 make this curve behave like "clip", while values near 0.0 make
23460 this curve behave like "reinhard". The default value is 0.3,
23461 which provides a good balance between colorimetric accuracy and
23462 preserving out-of-gamut details.
23463 hable
23465 Piece-wise, filmic tone-mapping algorithm developed by John
23466 Hable for use in Uncharted 2, inspired by a similar tone-
23467 mapping algorithm used by Kodak. Popularized by its use in
23468 video games with HDR rendering. Preserves both dark and bright
23469 details very well, but comes with the drawback of changing the
23470 average brightness quite significantly. This is sort of similar
23471 to "reinhard" with parameter 0.24.
23472 gamma
23474 Fits a gamma (power) function to transfer between the source
23475 and target color spaces, effectively resulting in a perceptual
23476 hard-knee joining two roughly linear sections. This preserves
23477 details at all scales fairly accurately, but can result in an
23478 image with a muted or dull appearance. The parameter is used as
23479 the cutoff point, defaulting to 0.5.
23480 linear
23482 Linearly stretches the input range to the output range, in PQ
23483 space. This will preserve all details accurately, but results
23484 in a significantly different average brightness. Can be used
23485 for inverse tone-mapping in addition to regular tone-mapping.
23486 The parameter can be used as an additional linear gain
23487 coefficient (defaulting to 1.0).
23488 tonemapping_param
23490 For tunable tone mapping functions, this parameter can be used to
23491 fine-tune the curve behavior. Refer to the documentation of
23492 "tonemapping". The default value of 0.0 is replaced by the curve's
23493 preferred default setting.
23494 tonemapping_mode
23496 This option determines how the tone mapping function specified by
23497 "tonemapping" is applied to the colors in a scene. Possible values
23498 are:
23499 auto
23501 Automatic selection based on internal heuristics. This is the
23502 default.
23503 rgb Apply the function per-channel in the RGB colorspace. Per-
23505 channel tone-mapping in RGB. Guarantees no clipping and heavily
23506 desaturates the output, but distorts the colors quite
23507 significantly. Very similar to the "Hollywood" look and feel.
23508 max Tone-mapping is performed on the brightest component found in
23510 the signal. Good at preserving details in highlights, but has a
23511 tendency to crush blacks.
23512 hybrid
23514 Tone-map per-channel for highlights and linearly (luma-based)
23515 for midtones/shadows, based on a fixed gamma 2.4 coefficient
23516 curve.
23517 luma
23519 Tone-map linearly on the luma component (CIE Y), and adjust
23520 (desaturate) the chromaticities to compensate using a simple
23521 constant factor. This is essentially the mode used in ITU-R
23522 BT.2446 method A.
23523 inverse_tonemapping
23525 If enabled, this filter will also attempt stretching SDR signals to
23526 fill HDR output color volumes. Disabled by default.
23527 tonemapping_crosstalk
23529 Extra tone-mapping crosstalk factor, between 0.0 and 0.3. This can
23530 help reduce issues tone-mapping certain bright spectral colors.
23531 Defaults to 0.04.
23532 tonemapping_lut_size
23534 Size of the tone-mapping LUT, between 2 and 1024. Defaults to 256.
23535 Note that this figure is squared when combined with "peak_detect".
23536 Dithering
23538 By default, libplacebo will dither whenever necessary, which includes
23540 rendering to any integer format below 16-bit precision. It's
23541 recommended to always leave this on, since not doing so may result in
23542 visible banding in the output, even if the "debanding" filter is
23543 enabled. If maximum performance is needed, use "ordered_fixed" instead
23544 of disabling dithering.
23545 dithering
23547 Dithering method to use. Accepts the following values:
23548 none
23550 Disables dithering completely. May result in visible banding.
23551 blue
23553 Dither with pseudo-blue noise. This is the default.
23554 ordered
23556 Tunable ordered dither pattern.
23557 ordered_fixed
23559 Faster ordered dither with a fixed size of 6. Texture-less.
23560 white
23562 Dither with white noise. Texture-less.
23563 dither_lut_size
23565 Dither LUT size, as log base2 between 1 and 8. Defaults to 6,
23566 corresponding to a LUT size of "64x64".
23567 dither_temporal
23569 Enables temporal dithering. Disabled by default.
23570 Custom shaders
23572 libplacebo supports a number of custom shaders based on the mpv .hook
23574 GLSL syntax. A collection of such shaders can be found here:
23575 <https://github.com/mpv-player/mpv/wiki/User-Scripts#user-shaders>
23576 A full description of the mpv shader format is beyond the scope of this
23578 section, but a summary can be found here:
23579 <https://mpv.io/manual/master/#options-glsl-shader>
23580 custom_shader_path
23582 Specifies a path to a custom shader file to load at runtime.
23583 custom_shader_bin
23585 Specifies a complete custom shader as a raw string.
23586 Debugging / performance
23588 All of the options in this section default off. They may be of
23590 assistance when attempting to squeeze the maximum performance at the
23591 cost of quality.
23592 skip_aa
23594 Disable anti-aliasing when downscaling.
23595 polar_cutoff
23597 Truncate polar (EWA) scaler kernels below this absolute magnitude,
23598 between 0.0 and 1.0.
23599 disable_linear
23601 Disable linear light scaling.
23602 disable_builtin
23604 Disable built-in GPU sampling (forces LUT).
23605 disable_fbos
23607 Forcibly disable FBOs, resulting in loss of almost all
23608 functionality, but offering the maximum possible speed.
23609 Commands
23611 This filter supports almost all of the above options as commands.
23613 Examples
23615 • Complete example for how to initialize the Vulkan device, upload
23617 frames to the GPU, perform filter conversion to yuv420p, and
23618 download frames back to the CPU for output. Note that in specific
23619 cases you can get around the need to perform format conversion by
23620 specifying the correct "format" filter option corresponding to the
23621 input frames.
23622 ffmpeg -i $INPUT -init_hw_device vulkan -vf hwupload,libplacebo=format=yuv420p,hwdownload,format=yuv420p $OUTPUT
23624 • Tone-map input to standard gamut BT.709 output:
23626 libplacebo=colorspace=bt709:color_primaries=bt709:color_trc=bt709:range=tv
23628 • Rescale input to fit into standard 1080p, with high quality
23630 scaling:
23631 libplacebo=w=1920:h=1080:force_original_aspect_ratio=decrease:normalize_sar=true:upscaler=ewa_lanczos:downscaler=ewa_lanczos
23633 • Convert input to standard sRGB JPEG:
23635 libplacebo=format=yuv420p:colorspace=bt470bg:color_primaries=bt709:color_trc=iec61966-2-1:range=pc
23637 • Use higher quality debanding settings:
23639 libplacebo=deband=true:deband_iterations=3:deband_radius=8:deband_threshold=6
23641 • Run this filter on the CPU, on systems with Mesa installed (and
23643 with the most expensive options disabled):
23644 ffmpeg ... -init_hw_device vulkan:llvmpipe ... -vf libplacebo=upscaler=none:downscaler=none:peak_detect=false
23646 • Suppress CPU-based AV1/H.274 film grain application in the decoder,
23648 in favor of doing it with this filter. Note that this is only a
23649 gain if the frames are either already on the GPU, or if you're
23650 using libplacebo for other purposes, since otherwise the VRAM
23651 roundtrip will more than offset any expected speedup.
23652 ffmpeg -export_side_data +film_grain ... -vf libplacebo=apply_filmgrain=true
23654 libvmaf
23656 Calulate the VMAF (Video Multi-Method Assessment Fusion) score for a
23657 reference/distorted pair of input videos.
23658 The first input is the distorted video, and the second input is the
23660 reference video.
23661 The obtained VMAF score is printed through the logging system.
23663 It requires Netflix's vmaf library (libvmaf) as a pre-requisite. After
23665 installing the library it can be enabled using: "./configure
23666 --enable-libvmaf".
23667 The filter has following options:
23669 model
23671 A `|` delimited list of vmaf models. Each model can be configured
23672 with a number of parameters. Default value: "version=vmaf_v0.6.1"
23673 model_path
23675 Deprecated, use model='path=...'.
23676 enable_transform
23678 Deprecated, use model='enable_transform=true'.
23679 phone_model
23681 Deprecated, use model='enable_transform=true'.
23682 enable_conf_interval
23684 Deprecated, use model='enable_conf_interval=true'.
23685 feature
23687 A `|` delimited list of features. Each feature can be configured
23688 with a number of parameters.
23689 psnr
23691 Deprecated, use feature='name=psnr'.
23692 ssim
23694 Deprecated, use feature='name=ssim'.
23695 ms_ssim
23697 Deprecated, use feature='name=ms_ssim'.
23698 log_path
23700 Set the file path to be used to store log files.
23701 log_fmt
23703 Set the format of the log file (xml, json, csv, or sub).
23704 n_threads
23706 Set number of threads to be used when initializing libvmaf.
23707 Default value: 0, no threads.
23708 n_subsample
23710 Set frame subsampling interval to be used.
23711 This filter also supports the framesync options.
23713 Examples
23715 • In the examples below, a distorted video distorted.mpg is compared
23717 with a reference file reference.mpg.
23718 • Basic usage:
23720 ffmpeg -i distorted.mpg -i reference.mpg -lavfi libvmaf=log_path=output.xml -f null -
23722 • Example with multiple models:
23724 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 -
23726 • Example with multiple addtional features:
23728 ffmpeg -i distorted.mpg -i reference.mpg -lavfi libvmaf='feature=name=psnr|name=ciede' -f null -
23730 • Example with options and different containers:
23732 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 -
23734 limitdiff
23736 Apply limited difference filter using second and optionally third video
23737 stream.
23738 The filter accepts the following options:
23740 threshold
23742 Set the threshold to use when allowing certain differences between
23743 video streams. Any absolute difference value lower or exact than
23744 this threshold will pick pixel components from first video stream.
23745 elasticity
23747 Set the elasticity of soft thresholding when processing video
23748 streams. This value multiplied with first one sets second
23749 threshold. Any absolute difference value greater or exact than
23750 second threshold will pick pixel components from second video
23751 stream. For values between those two threshold linear interpolation
23752 between first and second video stream will be used.
23753 reference
23755 Enable the reference (third) video stream processing. By default is
23756 disabled. If set, this video stream will be used for calculating
23757 absolute difference with first video stream.
23758 planes
23760 Specify which planes will be processed. Defaults to all available.
23761 Commands
23763 This filter supports the all above options as commands except option
23765 reference.
23766 limiter
23768 Limits the pixel components values to the specified range [min, max].
23769 The filter accepts the following options:
23771 min Lower bound. Defaults to the lowest allowed value for the input.
23773 max Upper bound. Defaults to the highest allowed value for the input.
23775 planes
23777 Specify which planes will be processed. Defaults to all available.
23778 Commands
23780 This filter supports the all above options as commands.
23782 loop
23784 Loop video frames.
23785 The filter accepts the following options:
23787 loop
23789 Set the number of loops. Setting this value to -1 will result in
23790 infinite loops. Default is 0.
23791 size
23793 Set maximal size in number of frames. Default is 0.
23794 start
23796 Set first frame of loop. Default is 0.
23797 Examples
23799 • Loop single first frame infinitely:
23801 loop=loop=-1:size=1:start=0
23803 • Loop single first frame 10 times:
23805 loop=loop=10:size=1:start=0
23807 • Loop 10 first frames 5 times:
23809 loop=loop=5:size=10:start=0
23811 lut1d
23813 Apply a 1D LUT to an input video.
23814 The filter accepts the following options:
23816 file
23818 Set the 1D LUT file name.
23819 Currently supported formats:
23821 cube
23823 Iridas
23824 csp cineSpace
23826 interp
23828 Select interpolation mode.
23829 Available values are:
23831 nearest
23833 Use values from the nearest defined point.
23834 linear
23836 Interpolate values using the linear interpolation.
23837 cosine
23839 Interpolate values using the cosine interpolation.
23840 cubic
23842 Interpolate values using the cubic interpolation.
23843 spline
23845 Interpolate values using the spline interpolation.
23846 Commands
23848 This filter supports the all above options as commands.
23850 lut3d
23852 Apply a 3D LUT to an input video.
23853 The filter accepts the following options:
23855 file
23857 Set the 3D LUT file name.
23858 Currently supported formats:
23860 3dl AfterEffects
23862 cube
23864 Iridas
23865 dat DaVinci
23867 m3d Pandora
23869 csp cineSpace
23871 interp
23873 Select interpolation mode.
23874 Available values are:
23876 nearest
23878 Use values from the nearest defined point.
23879 trilinear
23881 Interpolate values using the 8 points defining a cube.
23882 tetrahedral
23884 Interpolate values using a tetrahedron.
23885 pyramid
23887 Interpolate values using a pyramid.
23888 prism
23890 Interpolate values using a prism.
23891 Commands
23893 This filter supports the "interp" option as commands.
23895 lumakey
23897 Turn certain luma values into transparency.
23898 The filter accepts the following options:
23900 threshold
23902 Set the luma which will be used as base for transparency. Default
23903 value is 0.
23904 tolerance
23906 Set the range of luma values to be keyed out. Default value is
23907 0.01.
23908 softness
23910 Set the range of softness. Default value is 0. Use this to control
23911 gradual transition from zero to full transparency.
23912 Commands
23914 This filter supports same commands as options. The command accepts the
23916 same syntax of the corresponding option.
23917 If the specified expression is not valid, it is kept at its current
23919 value.
23920 lut, lutrgb, lutyuv
23922 Compute a look-up table for binding each pixel component input value to
23923 an output value, and apply it to the input video.
23924 lutyuv applies a lookup table to a YUV input video, lutrgb to an RGB
23926 input video.
23927 These filters accept the following parameters:
23929 c0 set first pixel component expression
23931 c1 set second pixel component expression
23933 c2 set third pixel component expression
23935 c3 set fourth pixel component expression, corresponds to the alpha
23937 component
23938 r set red component expression
23940 g set green component expression
23942 b set blue component expression
23944 a alpha component expression
23946 y set Y/luminance component expression
23948 u set U/Cb component expression
23950 v set V/Cr component expression
23952 Each of them specifies the expression to use for computing the lookup
23954 table for the corresponding pixel component values.
23955 The exact component associated to each of the c* options depends on the
23957 format in input.
23958 The lut filter requires either YUV or RGB pixel formats in input,
23960 lutrgb requires RGB pixel formats in input, and lutyuv requires YUV.
23961 The expressions can contain the following constants and functions:
23963 w
23965 h The input width and height.
23966 val The input value for the pixel component.
23968 clipval
23970 The input value, clipped to the minval-maxval range.
23971 maxval
23973 The maximum value for the pixel component.
23974 minval
23976 The minimum value for the pixel component.
23977 negval
23979 The negated value for the pixel component value, clipped to the
23980 minval-maxval range; it corresponds to the expression
23981 "maxval-clipval+minval".
23982 clip(val)
23984 The computed value in val, clipped to the minval-maxval range.
23985 gammaval(gamma)
23987 The computed gamma correction value of the pixel component value,
23988 clipped to the minval-maxval range. It corresponds to the
23989 expression
23990 "pow((clipval-minval)/(maxval-minval)\,gamma)*(maxval-minval)+minval"
23991 All expressions default to "clipval".
23993 Commands
23995 This filter supports same commands as options.
23997 Examples
23999 • Negate input video:
24001 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
24003 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
24004 The above is the same as:
24006 lutrgb="r=negval:g=negval:b=negval"
24008 lutyuv="y=negval:u=negval:v=negval"
24009 • Negate luminance:
24011 lutyuv=y=negval
24013 • Remove chroma components, turning the video into a graytone image:
24015 lutyuv="u=128:v=128"
24017 • Apply a luma burning effect:
24019 lutyuv="y=2*val"
24021 • Remove green and blue components:
24023 lutrgb="g=0:b=0"
24025 • Set a constant alpha channel value on input:
24027 format=rgba,lutrgb=a="maxval-minval/2"
24029 • Correct luminance gamma by a factor of 0.5:
24031 lutyuv=y=gammaval(0.5)
24033 • Discard least significant bits of luma:
24035 lutyuv=y='bitand(val, 128+64+32)'
24037 • Technicolor like effect:
24039 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
24041 lut2, tlut2
24043 The "lut2" filter takes two input streams and outputs one stream.
24044 The "tlut2" (time lut2) filter takes two consecutive frames from one
24046 single stream.
24047 This filter accepts the following parameters:
24049 c0 set first pixel component expression
24051 c1 set second pixel component expression
24053 c2 set third pixel component expression
24055 c3 set fourth pixel component expression, corresponds to the alpha
24057 component
24058 d set output bit depth, only available for "lut2" filter. By default
24060 is 0, which means bit depth is automatically picked from first
24061 input format.
24062 The "lut2" filter also supports the framesync options.
24064 Each of them specifies the expression to use for computing the lookup
24066 table for the corresponding pixel component values.
24067 The exact component associated to each of the c* options depends on the
24069 format in inputs.
24070 The expressions can contain the following constants:
24072 w
24074 h The input width and height.
24075 x The first input value for the pixel component.
24077 y The second input value for the pixel component.
24079 bdx The first input video bit depth.
24081 bdy The second input video bit depth.
24083 All expressions default to "x".
24085 Commands
24087 This filter supports the all above options as commands except option
24089 "d".
24090 Examples
24092 • Highlight differences between two RGB video streams:
24094 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)'
24096 • Highlight differences between two YUV video streams:
24098 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)'
24100 • Show max difference between two video streams:
24102 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)))'
24104 maskedclamp
24106 Clamp the first input stream with the second input and third input
24107 stream.
24108 Returns the value of first stream to be between second input stream -
24110 "undershoot" and third input stream + "overshoot".
24111 This filter accepts the following options:
24113 undershoot
24115 Default value is 0.
24116 overshoot
24118 Default value is 0.
24119 planes
24121 Set which planes will be processed as bitmap, unprocessed planes
24122 will be copied from first stream. By default value 0xf, all planes
24123 will be processed.
24124 Commands
24126 This filter supports the all above options as commands.
24128 maskedmax
24130 Merge the second and third input stream into output stream using
24131 absolute differences between second input stream and first input stream
24132 and absolute difference between third input stream and first input
24133 stream. The picked value will be from second input stream if second
24134 absolute difference is greater than first one or from third input
24135 stream otherwise.
24136 This filter accepts the following options:
24138 planes
24140 Set which planes will be processed as bitmap, unprocessed planes
24141 will be copied from first stream. By default value 0xf, all planes
24142 will be processed.
24143 Commands
24145 This filter supports the all above options as commands.
24147 maskedmerge
24149 Merge the first input stream with the second input stream using per
24150 pixel weights in the third input stream.
24151 A value of 0 in the third stream pixel component means that pixel
24153 component from first stream is returned unchanged, while maximum value
24154 (eg. 255 for 8-bit videos) means that pixel component from second
24155 stream is returned unchanged. Intermediate values define the amount of
24156 merging between both input stream's pixel components.
24157 This filter accepts the following options:
24159 planes
24161 Set which planes will be processed as bitmap, unprocessed planes
24162 will be copied from first stream. By default value 0xf, all planes
24163 will be processed.
24164 Commands
24166 This filter supports the all above options as commands.
24168 maskedmin
24170 Merge the second and third input stream into output stream using
24171 absolute differences between second input stream and first input stream
24172 and absolute difference between third input stream and first input
24173 stream. The picked value will be from second input stream if second
24174 absolute difference is less than first one or from third input stream
24175 otherwise.
24176 This filter accepts the following options:
24178 planes
24180 Set which planes will be processed as bitmap, unprocessed planes
24181 will be copied from first stream. By default value 0xf, all planes
24182 will be processed.
24183 Commands
24185 This filter supports the all above options as commands.
24187 maskedthreshold
24189 Pick pixels comparing absolute difference of two video streams with
24190 fixed threshold.
24191 If absolute difference between pixel component of first and second
24193 video stream is equal or lower than user supplied threshold than pixel
24194 component from first video stream is picked, otherwise pixel component
24195 from second video stream is picked.
24196 This filter accepts the following options:
24198 threshold
24200 Set threshold used when picking pixels from absolute difference
24201 from two input video streams.
24202 planes
24204 Set which planes will be processed as bitmap, unprocessed planes
24205 will be copied from second stream. By default value 0xf, all
24206 planes will be processed.
24207 mode
24209 Set mode of filter operation. Can be "abs" or "diff". Default is
24210 "abs".
24211 Commands
24213 This filter supports the all above options as commands.
24215 maskfun
24217 Create mask from input video.
24218 For example it is useful to create motion masks after "tblend" filter.
24220 This filter accepts the following options:
24222 low Set low threshold. Any pixel component lower or exact than this
24224 value will be set to 0.
24225 high
24227 Set high threshold. Any pixel component higher than this value will
24228 be set to max value allowed for current pixel format.
24229 planes
24231 Set planes to filter, by default all available planes are filtered.
24232 fill
24234 Fill all frame pixels with this value.
24235 sum Set max average pixel value for frame. If sum of all pixel
24237 components is higher that this average, output frame will be
24238 completely filled with value set by fill option. Typically useful
24239 for scene changes when used in combination with "tblend" filter.
24240 Commands
24242 This filter supports the all above options as commands.
24244 mcdeint
24246 Apply motion-compensation deinterlacing.
24247 It needs one field per frame as input and must thus be used together
24249 with yadif=1/3 or equivalent.
24250 This filter is only available in ffmpeg version 4.4 or earlier.
24252 This filter accepts the following options:
24254 mode
24256 Set the deinterlacing mode.
24257 It accepts one of the following values:
24259 fast
24261 medium
24262 slow
24263 use iterative motion estimation
24264 extra_slow
24266 like slow, but use multiple reference frames.
24267 Default value is fast.
24269 parity
24271 Set the picture field parity assumed for the input video. It must
24272 be one of the following values:
24273 0, tff
24275 assume top field first
24276 1, bff
24278 assume bottom field first
24279 Default value is bff.
24281 qp Set per-block quantization parameter (QP) used by the internal
24283 encoder.
24284 Higher values should result in a smoother motion vector field but
24286 less optimal individual vectors. Default value is 1.
24287 median
24289 Pick median pixel from certain rectangle defined by radius.
24290 This filter accepts the following options:
24292 radius
24294 Set horizontal radius size. Default value is 1. Allowed range is
24295 integer from 1 to 127.
24296 planes
24298 Set which planes to process. Default is 15, which is all available
24299 planes.
24300 radiusV
24302 Set vertical radius size. Default value is 0. Allowed range is
24303 integer from 0 to 127. If it is 0, value will be picked from
24304 horizontal "radius" option.
24305 percentile
24307 Set median percentile. Default value is 0.5. Default value of 0.5
24308 will pick always median values, while 0 will pick minimum values,
24309 and 1 maximum values.
24310 Commands
24312 This filter supports same commands as options. The command accepts the
24314 same syntax of the corresponding option.
24315 If the specified expression is not valid, it is kept at its current
24317 value.
24318 mergeplanes
24320 Merge color channel components from several video streams.
24321 The filter accepts up to 4 input streams, and merge selected input
24323 planes to the output video.
24324 This filter accepts the following options:
24326 mapping
24328 Set input to output plane mapping. Default is 0.
24329 The mappings is specified as a bitmap. It should be specified as a
24331 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
24332 mapping for the first plane of the output stream. 'A' sets the
24333 number of the input stream to use (from 0 to 3), and 'a' the plane
24334 number of the corresponding input to use (from 0 to 3). The rest of
24335 the mappings is similar, 'Bb' describes the mapping for the output
24336 stream second plane, 'Cc' describes the mapping for the output
24337 stream third plane and 'Dd' describes the mapping for the output
24338 stream fourth plane.
24339 format
24341 Set output pixel format. Default is "yuva444p".
24342 map0s
24344 map1s
24345 map2s
24346 map3s
24347 Set input to output stream mapping for output Nth plane. Default is
24348 0.
24349 map0p
24351 map1p
24352 map2p
24353 map3p
24354 Set input to output plane mapping for output Nth plane. Default is
24355 0.
24356 Examples
24358 • Merge three gray video streams of same width and height into single
24360 video stream:
24361 [a0][a1][a2]mergeplanes=0x001020:yuv444p
24363 • Merge 1st yuv444p stream and 2nd gray video stream into yuva444p
24365 video stream:
24366 [a0][a1]mergeplanes=0x00010210:yuva444p
24368 • Swap Y and A plane in yuva444p stream:
24370 format=yuva444p,mergeplanes=0x03010200:yuva444p
24372 • Swap U and V plane in yuv420p stream:
24374 format=yuv420p,mergeplanes=0x000201:yuv420p
24376 • Cast a rgb24 clip to yuv444p:
24378 format=rgb24,mergeplanes=0x000102:yuv444p
24380 mestimate
24382 Estimate and export motion vectors using block matching algorithms.
24383 Motion vectors are stored in frame side data to be used by other
24384 filters.
24385 This filter accepts the following options:
24387 method
24389 Specify the motion estimation method. Accepts one of the following
24390 values:
24391 esa Exhaustive search algorithm.
24393 tss Three step search algorithm.
24395 tdls
24397 Two dimensional logarithmic search algorithm.
24398 ntss
24400 New three step search algorithm.
24401 fss Four step search algorithm.
24403 ds Diamond search algorithm.
24405 hexbs
24407 Hexagon-based search algorithm.
24408 epzs
24410 Enhanced predictive zonal search algorithm.
24411 umh Uneven multi-hexagon search algorithm.
24413 Default value is esa.
24415 mb_size
24417 Macroblock size. Default 16.
24418 search_param
24420 Search parameter. Default 7.
24421 midequalizer
24423 Apply Midway Image Equalization effect using two video streams.
24424 Midway Image Equalization adjusts a pair of images to have the same
24426 histogram, while maintaining their dynamics as much as possible. It's
24427 useful for e.g. matching exposures from a pair of stereo cameras.
24428 This filter has two inputs and one output, which must be of same pixel
24430 format, but may be of different sizes. The output of filter is first
24431 input adjusted with midway histogram of both inputs.
24432 This filter accepts the following option:
24434 planes
24436 Set which planes to process. Default is 15, which is all available
24437 planes.
24438 minterpolate
24440 Convert the video to specified frame rate using motion interpolation.
24441 This filter accepts the following options:
24443 fps Specify the output frame rate. This can be rational e.g.
24445 "60000/1001". Frames are dropped if fps is lower than source fps.
24446 Default 60.
24447 mi_mode
24449 Motion interpolation mode. Following values are accepted:
24450 dup Duplicate previous or next frame for interpolating new ones.
24452 blend
24454 Blend source frames. Interpolated frame is mean of previous and
24455 next frames.
24456 mci Motion compensated interpolation. Following options are
24458 effective when this mode is selected:
24459 mc_mode
24461 Motion compensation mode. Following values are accepted:
24462 obmc
24464 Overlapped block motion compensation.
24465 aobmc
24467 Adaptive overlapped block motion compensation. Window
24468 weighting coefficients are controlled adaptively
24469 according to the reliabilities of the neighboring
24470 motion vectors to reduce oversmoothing.
24471 Default mode is obmc.
24473 me_mode
24475 Motion estimation mode. Following values are accepted:
24476 bidir
24478 Bidirectional motion estimation. Motion vectors are
24479 estimated for each source frame in both forward and
24480 backward directions.
24481 bilat
24483 Bilateral motion estimation. Motion vectors are
24484 estimated directly for interpolated frame.
24485 Default mode is bilat.
24487 me The algorithm to be used for motion estimation. Following
24489 values are accepted:
24490 esa Exhaustive search algorithm.
24492 tss Three step search algorithm.
24494 tdls
24496 Two dimensional logarithmic search algorithm.
24497 ntss
24499 New three step search algorithm.
24500 fss Four step search algorithm.
24502 ds Diamond search algorithm.
24504 hexbs
24506 Hexagon-based search algorithm.
24507 epzs
24509 Enhanced predictive zonal search algorithm.
24510 umh Uneven multi-hexagon search algorithm.
24512 Default algorithm is epzs.
24514 mb_size
24516 Macroblock size. Default 16.
24517 search_param
24519 Motion estimation search parameter. Default 32.
24520 vsbmc
24522 Enable variable-size block motion compensation. Motion
24523 estimation is applied with smaller block sizes at object
24524 boundaries in order to make the them less blur. Default is
24525 0 (disabled).
24526 scd Scene change detection method. Scene change leads motion vectors to
24528 be in random direction. Scene change detection replace interpolated
24529 frames by duplicate ones. May not be needed for other modes.
24530 Following values are accepted:
24531 none
24533 Disable scene change detection.
24534 fdiff
24536 Frame difference. Corresponding pixel values are compared and
24537 if it satisfies scd_threshold scene change is detected.
24538 Default method is fdiff.
24540 scd_threshold
24542 Scene change detection threshold. Default is 10..
24543 mix
24545 Mix several video input streams into one video stream.
24546 A description of the accepted options follows.
24548 inputs
24550 The number of inputs. If unspecified, it defaults to 2.
24551 weights
24553 Specify weight of each input video stream as sequence. Each weight
24554 is separated by space. If number of weights is smaller than number
24555 of frames last specified weight will be used for all remaining
24556 unset weights.
24557 scale
24559 Specify scale, if it is set it will be multiplied with sum of each
24560 weight multiplied with pixel values to give final destination pixel
24561 value. By default scale is auto scaled to sum of weights.
24562 planes
24564 Set which planes to filter. Default is all. Allowed range is from 0
24565 to 15.
24566 duration
24568 Specify how end of stream is determined.
24569 longest
24571 The duration of the longest input. (default)
24572 shortest
24574 The duration of the shortest input.
24575 first
24577 The duration of the first input.
24578 Commands
24580 This filter supports the following commands:
24582 weights
24584 scale
24585 planes
24586 Syntax is same as option with same name.
24587 monochrome
24589 Convert video to gray using custom color filter.
24590 A description of the accepted options follows.
24592 cb Set the chroma blue spot. Allowed range is from -1 to 1. Default
24594 value is 0.
24595 cr Set the chroma red spot. Allowed range is from -1 to 1. Default
24597 value is 0.
24598 size
24600 Set the color filter size. Allowed range is from .1 to 10. Default
24601 value is 1.
24602 high
24604 Set the highlights strength. Allowed range is from 0 to 1. Default
24605 value is 0.
24606 Commands
24608 This filter supports the all above options as commands.
24610 morpho
24612 This filter allows to apply main morphological grayscale transforms,
24613 erode and dilate with arbitrary structures set in second input stream.
24614 Unlike naive implementation and much slower performance in erosion and
24616 dilation filters, when speed is critical "morpho" filter should be used
24617 instead.
24618 A description of accepted options follows,
24620 mode
24622 Set morphological transform to apply, can be:
24623 erode
24625 dilate
24626 open
24627 close
24628 gradient
24629 tophat
24630 blackhat
24631 Default is "erode".
24633 planes
24635 Set planes to filter, by default all planes except alpha are
24636 filtered.
24637 structure
24639 Set which structure video frames will be processed from second
24640 input stream, can be first or all. Default is all.
24641 The "morpho" filter also supports the framesync options.
24643 Commands
24645 This filter supports same commands as options.
24647 mpdecimate
24649 Drop frames that do not differ greatly from the previous frame in order
24650 to reduce frame rate.
24651 The main use of this filter is for very-low-bitrate encoding (e.g.
24653 streaming over dialup modem), but it could in theory be used for fixing
24654 movies that were inverse-telecined incorrectly.
24655 A description of the accepted options follows.
24657 max Set the maximum number of consecutive frames which can be dropped
24659 (if positive), or the minimum interval between dropped frames (if
24660 negative). If the value is 0, the frame is dropped disregarding the
24661 number of previous sequentially dropped frames.
24662 Default value is 0.
24664 hi
24666 lo
24667 frac
24668 Set the dropping threshold values.
24669 Values for hi and lo are for 8x8 pixel blocks and represent actual
24671 pixel value differences, so a threshold of 64 corresponds to 1 unit
24672 of difference for each pixel, or the same spread out differently
24673 over the block.
24674 A frame is a candidate for dropping if no 8x8 blocks differ by more
24676 than a threshold of hi, and if no more than frac blocks (1 meaning
24677 the whole image) differ by more than a threshold of lo.
24678 Default value for hi is 64*12, default value for lo is 64*5, and
24680 default value for frac is 0.33.
24681 msad
24683 Obtain the MSAD (Mean Sum of Absolute Differences) between two input
24684 videos.
24685 This filter takes two input videos.
24687 Both input videos must have the same resolution and pixel format for
24689 this filter to work correctly. Also it assumes that both inputs have
24690 the same number of frames, which are compared one by one.
24691 The obtained per component, average, min and max MSAD is printed
24693 through the logging system.
24694 The filter stores the calculated MSAD of each frame in frame metadata.
24696 This filter also supports the framesync options.
24698 In the below example the input file main.mpg being processed is
24700 compared with the reference file ref.mpg.
24701 ffmpeg -i main.mpg -i ref.mpg -lavfi msad -f null -
24703 multiply
24705 Multiply first video stream pixels values with second video stream
24706 pixels values.
24707 The filter accepts the following options:
24709 scale
24711 Set the scale applied to second video stream. By default is 1.
24712 Allowed range is from 0 to 9.
24713 offset
24715 Set the offset applied to second video stream. By default is 0.5.
24716 Allowed range is from -1 to 1.
24717 planes
24719 Specify planes from input video stream that will be processed. By
24720 default all planes are processed.
24721 Commands
24723 This filter supports same commands as options.
24725 negate
24727 Negate (invert) the input video.
24728 It accepts the following option:
24730 components
24732 Set components to negate.
24733 Available values for components are:
24735 y
24737 u
24738 v
24739 a
24740 r
24741 g
24742 b
24743 negate_alpha
24744 With value 1, it negates the alpha component, if present. Default
24745 value is 0.
24746 Commands
24748 This filter supports same commands as options.
24750 nlmeans
24752 Denoise frames using Non-Local Means algorithm.
24753 Each pixel is adjusted by looking for other pixels with similar
24755 contexts. This context similarity is defined by comparing their
24756 surrounding patches of size pxp. Patches are searched in an area of rxr
24757 around the pixel.
24758 Note that the research area defines centers for patches, which means
24760 some patches will be made of pixels outside that research area.
24761 The filter accepts the following options.
24763 s Set denoising strength. Default is 1.0. Must be in range [1.0,
24765 30.0].
24766 p Set patch size. Default is 7. Must be odd number in range [0, 99].
24768 pc Same as p but for chroma planes.
24770 The default value is 0 and means automatic.
24772 r Set research size. Default is 15. Must be odd number in range [0,
24774 99].
24775 rc Same as r but for chroma planes.
24777 The default value is 0 and means automatic.
24779 nnedi
24781 Deinterlace video using neural network edge directed interpolation.
24782 This filter accepts the following options:
24784 weights
24786 Mandatory option, without binary file filter can not work.
24787 Currently file can be found here:
24788 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
24789 deint
24791 Set which frames to deinterlace, by default it is "all". Can be
24792 "all" or "interlaced".
24793 field
24795 Set mode of operation.
24796 Can be one of the following:
24798 af Use frame flags, both fields.
24800 a Use frame flags, single field.
24802 t Use top field only.
24804 b Use bottom field only.
24806 tf Use both fields, top first.
24808 bf Use both fields, bottom first.
24810 planes
24812 Set which planes to process, by default filter process all frames.
24813 nsize
24815 Set size of local neighborhood around each pixel, used by the
24816 predictor neural network.
24817 Can be one of the following:
24819 s8x6
24821 s16x6
24822 s32x6
24823 s48x6
24824 s8x4
24825 s16x4
24826 s32x4
24827 nns Set the number of neurons in predictor neural network. Can be one
24828 of the following:
24829 n16
24831 n32
24832 n64
24833 n128
24834 n256
24835 qual
24836 Controls the number of different neural network predictions that
24837 are blended together to compute the final output value. Can be
24838 "fast", default or "slow".
24839 etype
24841 Set which set of weights to use in the predictor. Can be one of
24842 the following:
24843 a, abs
24845 weights trained to minimize absolute error
24846 s, mse
24848 weights trained to minimize squared error
24849 pscrn
24851 Controls whether or not the prescreener neural network is used to
24852 decide which pixels should be processed by the predictor neural
24853 network and which can be handled by simple cubic interpolation.
24854 The prescreener is trained to know whether cubic interpolation will
24855 be sufficient for a pixel or whether it should be predicted by the
24856 predictor nn. The computational complexity of the prescreener nn
24857 is much less than that of the predictor nn. Since most pixels can
24858 be handled by cubic interpolation, using the prescreener generally
24859 results in much faster processing. The prescreener is pretty
24860 accurate, so the difference between using it and not using it is
24861 almost always unnoticeable.
24862 Can be one of the following:
24864 none
24866 original
24867 new
24868 new2
24869 new3
24870 Default is "new".
24872 Commands
24874 This filter supports same commands as options, excluding weights
24876 option.
24877 noformat
24879 Force libavfilter not to use any of the specified pixel formats for the
24880 input to the next filter.
24881 It accepts the following parameters:
24883 pix_fmts
24885 A '|'-separated list of pixel format names, such as
24886 pix_fmts=yuv420p|monow|rgb24".
24887 Examples
24889 • Force libavfilter to use a format different from yuv420p for the
24891 input to the vflip filter:
24892 noformat=pix_fmts=yuv420p,vflip
24894 • Convert the input video to any of the formats not contained in the
24896 list:
24897 noformat=yuv420p|yuv444p|yuv410p
24899 noise
24901 Add noise on video input frame.
24902 The filter accepts the following options:
24904 all_seed
24906 c0_seed
24907 c1_seed
24908 c2_seed
24909 c3_seed
24910 Set noise seed for specific pixel component or all pixel components
24911 in case of all_seed. Default value is 123457.
24912 all_strength, alls
24914 c0_strength, c0s
24915 c1_strength, c1s
24916 c2_strength, c2s
24917 c3_strength, c3s
24918 Set noise strength for specific pixel component or all pixel
24919 components in case all_strength. Default value is 0. Allowed range
24920 is [0, 100].
24921 all_flags, allf
24923 c0_flags, c0f
24924 c1_flags, c1f
24925 c2_flags, c2f
24926 c3_flags, c3f
24927 Set pixel component flags or set flags for all components if
24928 all_flags. Available values for component flags are:
24929 a averaged temporal noise (smoother)
24931 p mix random noise with a (semi)regular pattern
24933 t temporal noise (noise pattern changes between frames)
24935 u uniform noise (gaussian otherwise)
24937 Examples
24939 Add temporal and uniform noise to input video:
24941 noise=alls=20:allf=t+u
24943 normalize
24945 Normalize RGB video (aka histogram stretching, contrast stretching).
24946 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
24947 For each channel of each frame, the filter computes the input range and
24949 maps it linearly to the user-specified output range. The output range
24950 defaults to the full dynamic range from pure black to pure white.
24951 Temporal smoothing can be used on the input range to reduce flickering
24953 (rapid changes in brightness) caused when small dark or bright objects
24954 enter or leave the scene. This is similar to the auto-exposure
24955 (automatic gain control) on a video camera, and, like a video camera,
24956 it may cause a period of over- or under-exposure of the video.
24957 The R,G,B channels can be normalized independently, which may cause
24959 some color shifting, or linked together as a single channel, which
24960 prevents color shifting. Linked normalization preserves hue.
24961 Independent normalization does not, so it can be used to remove some
24962 color casts. Independent and linked normalization can be combined in
24963 any ratio.
24964 The normalize filter accepts the following options:
24966 blackpt
24968 whitept
24969 Colors which define the output range. The minimum input value is
24970 mapped to the blackpt. The maximum input value is mapped to the
24971 whitept. The defaults are black and white respectively. Specifying
24972 white for blackpt and black for whitept will give color-inverted,
24973 normalized video. Shades of grey can be used to reduce the dynamic
24974 range (contrast). Specifying saturated colors here can create some
24975 interesting effects.
24976 smoothing
24978 The number of previous frames to use for temporal smoothing. The
24979 input range of each channel is smoothed using a rolling average
24980 over the current frame and the smoothing previous frames. The
24981 default is 0 (no temporal smoothing).
24982 independence
24984 Controls the ratio of independent (color shifting) channel
24985 normalization to linked (color preserving) normalization. 0.0 is
24986 fully linked, 1.0 is fully independent. Defaults to 1.0 (fully
24987 independent).
24988 strength
24990 Overall strength of the filter. 1.0 is full strength. 0.0 is a
24991 rather expensive no-op. Defaults to 1.0 (full strength).
24992 Commands
24994 This filter supports same commands as options, excluding smoothing
24996 option. The command accepts the same syntax of the corresponding
24997 option.
24998 If the specified expression is not valid, it is kept at its current
25000 value.
25001 Examples
25003 Stretch video contrast to use the full dynamic range, with no temporal
25005 smoothing; may flicker depending on the source content:
25006 normalize=blackpt=black:whitept=white:smoothing=0
25008 As above, but with 50 frames of temporal smoothing; flicker should be
25010 reduced, depending on the source content:
25011 normalize=blackpt=black:whitept=white:smoothing=50
25013 As above, but with hue-preserving linked channel normalization:
25015 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
25017 As above, but with half strength:
25019 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
25021 Map the darkest input color to red, the brightest input color to cyan:
25023 normalize=blackpt=red:whitept=cyan
25025 null
25027 Pass the video source unchanged to the output.
25028 ocr
25030 Optical Character Recognition
25031 This filter uses Tesseract for optical character recognition. To enable
25033 compilation of this filter, you need to configure FFmpeg with
25034 "--enable-libtesseract".
25035 It accepts the following options:
25037 datapath
25039 Set datapath to tesseract data. Default is to use whatever was set
25040 at installation.
25041 language
25043 Set language, default is "eng".
25044 whitelist
25046 Set character whitelist.
25047 blacklist
25049 Set character blacklist.
25050 The filter exports recognized text as the frame metadata
25052 "lavfi.ocr.text". The filter exports confidence of recognized words as
25053 the frame metadata "lavfi.ocr.confidence".
25054 ocv
25056 Apply a video transform using libopencv.
25057 To enable this filter, install the libopencv library and headers and
25059 configure FFmpeg with "--enable-libopencv".
25060 It accepts the following parameters:
25062 filter_name
25064 The name of the libopencv filter to apply.
25065 filter_params
25067 The parameters to pass to the libopencv filter. If not specified,
25068 the default values are assumed.
25069 Refer to the official libopencv documentation for more precise
25071 information:
25072 <http://docs.opencv.org/master/modules/imgproc/doc/filtering.html>
25073 Several libopencv filters are supported; see the following subsections.
25075 dilate
25077 Dilate an image by using a specific structuring element. It
25079 corresponds to the libopencv function "cvDilate".
25080 It accepts the parameters: struct_el|nb_iterations.
25082 struct_el represents a structuring element, and has the syntax:
25084 colsxrows+anchor_xxanchor_y/shape
25085 cols and rows represent the number of columns and rows of the
25087 structuring element, anchor_x and anchor_y the anchor point, and shape
25088 the shape for the structuring element. shape must be "rect", "cross",
25089 "ellipse", or "custom".
25090 If the value for shape is "custom", it must be followed by a string of
25092 the form "=filename". The file with name filename is assumed to
25093 represent a binary image, with each printable character corresponding
25094 to a bright pixel. When a custom shape is used, cols and rows are
25095 ignored, the number or columns and rows of the read file are assumed
25096 instead.
25097 The default value for struct_el is "3x3+0x0/rect".
25099 nb_iterations specifies the number of times the transform is applied to
25101 the image, and defaults to 1.
25102 Some examples:
25104 # Use the default values
25106 ocv=dilate
25107 # Dilate using a structuring element with a 5x5 cross, iterating two times
25109 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
25110 # Read the shape from the file diamond.shape, iterating two times.
25112 # The file diamond.shape may contain a pattern of characters like this
25113 # *
25114 # ***
25115 # *****
25116 # ***
25117 # *
25118 # The specified columns and rows are ignored
25119 # but the anchor point coordinates are not
25120 ocv=dilate:0x0+2x2/custom=diamond.shape|2
25121 erode
25123 Erode an image by using a specific structuring element. It corresponds
25125 to the libopencv function "cvErode".
25126 It accepts the parameters: struct_el:nb_iterations, with the same
25128 syntax and semantics as the dilate filter.
25129 smooth
25131 Smooth the input video.
25133 The filter takes the following parameters:
25135 type|param1|param2|param3|param4.
25136 type is the type of smooth filter to apply, and must be one of the
25138 following values: "blur", "blur_no_scale", "median", "gaussian", or
25139 "bilateral". The default value is "gaussian".
25140 The meaning of param1, param2, param3, and param4 depends on the smooth
25142 type. param1 and param2 accept integer positive values or 0. param3 and
25143 param4 accept floating point values.
25144 The default value for param1 is 3. The default value for the other
25146 parameters is 0.
25147 These parameters correspond to the parameters assigned to the libopencv
25149 function "cvSmooth".
25150 oscilloscope
25152 2D Video Oscilloscope.
25153 Useful to measure spatial impulse, step responses, chroma delays, etc.
25155 It accepts the following parameters:
25157 x Set scope center x position.
25159 y Set scope center y position.
25161 s Set scope size, relative to frame diagonal.
25163 t Set scope tilt/rotation.
25165 o Set trace opacity.
25167 tx Set trace center x position.
25169 ty Set trace center y position.
25171 tw Set trace width, relative to width of frame.
25173 th Set trace height, relative to height of frame.
25175 c Set which components to trace. By default it traces first three
25177 components.
25178 g Draw trace grid. By default is enabled.
25180 st Draw some statistics. By default is enabled.
25182 sc Draw scope. By default is enabled.
25184 Commands
25186 This filter supports same commands as options. The command accepts the
25188 same syntax of the corresponding option.
25189 If the specified expression is not valid, it is kept at its current
25191 value.
25192 Examples
25194 • Inspect full first row of video frame.
25196 oscilloscope=x=0.5:y=0:s=1
25198 • Inspect full last row of video frame.
25200 oscilloscope=x=0.5:y=1:s=1
25202 • Inspect full 5th line of video frame of height 1080.
25204 oscilloscope=x=0.5:y=5/1080:s=1
25206 • Inspect full last column of video frame.
25208 oscilloscope=x=1:y=0.5:s=1:t=1
25210 overlay
25212 Overlay one video on top of another.
25213 It takes two inputs and has one output. The first input is the "main"
25215 video on which the second input is overlaid.
25216 It accepts the following parameters:
25218 A description of the accepted options follows.
25220 x
25222 y Set the expression for the x and y coordinates of the overlaid
25223 video on the main video. Default value is "0" for both expressions.
25224 In case the expression is invalid, it is set to a huge value
25225 (meaning that the overlay will not be displayed within the output
25226 visible area).
25227 eof_action
25229 See framesync.
25230 eval
25232 Set when the expressions for x, and y are evaluated.
25233 It accepts the following values:
25235 init
25237 only evaluate expressions once during the filter initialization
25238 or when a command is processed
25239 frame
25241 evaluate expressions for each incoming frame
25242 Default value is frame.
25244 shortest
25246 See framesync.
25247 format
25249 Set the format for the output video.
25250 It accepts the following values:
25252 yuv420
25254 force YUV420 output
25255 yuv420p10
25257 force YUV420p10 output
25258 yuv422
25260 force YUV422 output
25261 yuv422p10
25263 force YUV422p10 output
25264 yuv444
25266 force YUV444 output
25267 rgb force packed RGB output
25269 gbrp
25271 force planar RGB output
25272 auto
25274 automatically pick format
25275 Default value is yuv420.
25277 repeatlast
25279 See framesync.
25280 alpha
25282 Set format of alpha of the overlaid video, it can be straight or
25283 premultiplied. Default is straight.
25284 The x, and y expressions can contain the following parameters.
25286 main_w, W
25288 main_h, H
25289 The main input width and height.
25290 overlay_w, w
25292 overlay_h, h
25293 The overlay input width and height.
25294 x
25296 y The computed values for x and y. They are evaluated for each new
25297 frame.
25298 hsub
25300 vsub
25301 horizontal and vertical chroma subsample values of the output
25302 format. For example for the pixel format "yuv422p" hsub is 2 and
25303 vsub is 1.
25304 n the number of input frame, starting from 0
25306 pos the position in the file of the input frame, NAN if unknown
25308 t The timestamp, expressed in seconds. It's NAN if the input
25310 timestamp is unknown.
25311 This filter also supports the framesync options.
25313 Note that the n, pos, t variables are available only when evaluation is
25315 done per frame, and will evaluate to NAN when eval is set to init.
25316 Be aware that frames are taken from each input video in timestamp
25318 order, hence, if their initial timestamps differ, it is a good idea to
25319 pass the two inputs through a setpts=PTS-STARTPTS filter to have them
25320 begin in the same zero timestamp, as the example for the movie filter
25321 does.
25322 You can chain together more overlays but you should test the efficiency
25324 of such approach.
25325 Commands
25327 This filter supports the following commands:
25329 x
25331 y Modify the x and y of the overlay input. The command accepts the
25332 same syntax of the corresponding option.
25333 If the specified expression is not valid, it is kept at its current
25335 value.
25336 Examples
25338 • Draw the overlay at 10 pixels from the bottom right corner of the
25340 main video:
25341 overlay=main_w-overlay_w-10:main_h-overlay_h-10
25343 Using named options the example above becomes:
25345 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
25347 • Insert a transparent PNG logo in the bottom left corner of the
25349 input, using the ffmpeg tool with the "-filter_complex" option:
25350 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
25352 • Insert 2 different transparent PNG logos (second logo on bottom
25354 right corner) using the ffmpeg tool:
25355 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
25357 • Add a transparent color layer on top of the main video; "WxH" must
25359 specify the size of the main input to the overlay filter:
25360 color=color=red@.3:size=WxH [over]; [in][over] overlay [out]
25362 • Play an original video and a filtered version (here with the
25364 deshake filter) side by side using the ffplay tool:
25365 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
25367 The above command is the same as:
25369 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
25371 • Make a sliding overlay appearing from the left to the right top
25373 part of the screen starting since time 2:
25374 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
25376 • Compose output by putting two input videos side to side:
25378 ffmpeg -i left.avi -i right.avi -filter_complex "
25380 nullsrc=size=200x100 [background];
25381 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
25382 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
25383 [background][left] overlay=shortest=1 [background+left];
25384 [background+left][right] overlay=shortest=1:x=100 [left+right]
25385 "
25386 • Mask 10-20 seconds of a video by applying the delogo filter to a
25388 section
25389 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
25391 -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]'
25392 masked.avi
25393 • Chain several overlays in cascade:
25395 nullsrc=s=200x200 [bg];
25397 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
25398 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
25399 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
25400 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
25401 [in3] null, [mid2] overlay=100:100 [out0]
25402 overlay_cuda
25404 Overlay one video on top of another.
25405 This is the CUDA variant of the overlay filter. It only accepts CUDA
25407 frames. The underlying input pixel formats have to match.
25408 It takes two inputs and has one output. The first input is the "main"
25410 video on which the second input is overlaid.
25411 It accepts the following parameters:
25413 x
25415 y Set expressions for the x and y coordinates of the overlaid video
25416 on the main video.
25417 They can contain the following parameters:
25419 main_w, W
25421 main_h, H
25422 The main input width and height.
25423 overlay_w, w
25425 overlay_h, h
25426 The overlay input width and height.
25427 x
25429 y The computed values for x and y. They are evaluated for each
25430 new frame.
25431 n The ordinal index of the main input frame, starting from 0.
25433 pos The byte offset position in the file of the main input frame,
25435 NAN if unknown.
25436 t The timestamp of the main input frame, expressed in seconds,
25438 NAN if unknown.
25439 Default value is "0" for both expressions.
25441 eval
25443 Set when the expressions for x and y are evaluated.
25444 It accepts the following values:
25446 init
25448 Evaluate expressions once during filter initialization or when
25449 a command is processed.
25450 frame
25452 Evaluate expressions for each incoming frame
25453 Default value is frame.
25455 eof_action
25457 See framesync.
25458 shortest
25460 See framesync.
25461 repeatlast
25463 See framesync.
25464 This filter also supports the framesync options.
25466 owdenoise
25468 Apply Overcomplete Wavelet denoiser.
25469 The filter accepts the following options:
25471 depth
25473 Set depth.
25474 Larger depth values will denoise lower frequency components more,
25476 but slow down filtering.
25477 Must be an int in the range 8-16, default is 8.
25479 luma_strength, ls
25481 Set luma strength.
25482 Must be a double value in the range 0-1000, default is 1.0.
25484 chroma_strength, cs
25486 Set chroma strength.
25487 Must be a double value in the range 0-1000, default is 1.0.
25489 pad
25491 Add paddings to the input image, and place the original input at the
25492 provided x, y coordinates.
25493 It accepts the following parameters:
25495 width, w
25497 height, h
25498 Specify an expression for the size of the output image with the
25499 paddings added. If the value for width or height is 0, the
25500 corresponding input size is used for the output.
25501 The width expression can reference the value set by the height
25503 expression, and vice versa.
25504 The default value of width and height is 0.
25506 x
25508 y Specify the offsets to place the input image at within the padded
25509 area, with respect to the top/left border of the output image.
25510 The x expression can reference the value set by the y expression,
25512 and vice versa.
25513 The default value of x and y is 0.
25515 If x or y evaluate to a negative number, they'll be changed so the
25517 input image is centered on the padded area.
25518 color
25520 Specify the color of the padded area. For the syntax of this
25521 option, check the "Color" section in the ffmpeg-utils manual.
25522 The default value of color is "black".
25524 eval
25526 Specify when to evaluate width, height, x and y expression.
25527 It accepts the following values:
25529 init
25531 Only evaluate expressions once during the filter initialization
25532 or when a command is processed.
25533 frame
25535 Evaluate expressions for each incoming frame.
25536 Default value is init.
25538 aspect
25540 Pad to aspect instead to a resolution.
25541 The value for the width, height, x, and y options are expressions
25543 containing the following constants:
25544 in_w
25546 in_h
25547 The input video width and height.
25548 iw
25550 ih These are the same as in_w and in_h.
25551 out_w
25553 out_h
25554 The output width and height (the size of the padded area), as
25555 specified by the width and height expressions.
25556 ow
25558 oh These are the same as out_w and out_h.
25559 x
25561 y The x and y offsets as specified by the x and y expressions, or NAN
25562 if not yet specified.
25563 a same as iw / ih
25565 sar input sample aspect ratio
25567 dar input display aspect ratio, it is the same as (iw / ih) * sar
25569 hsub
25571 vsub
25572 The horizontal and vertical chroma subsample values. For example
25573 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
25574 Examples
25576 • Add paddings with the color "violet" to the input video. The output
25578 video size is 640x480, and the top-left corner of the input video
25579 is placed at column 0, row 40
25580 pad=640:480:0:40:violet
25582 The example above is equivalent to the following command:
25584 pad=width=640:height=480:x=0:y=40:color=violet
25586 • Pad the input to get an output with dimensions increased by 3/2,
25588 and put the input video at the center of the padded area:
25589 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
25591 • Pad the input to get a squared output with size equal to the
25593 maximum value between the input width and height, and put the input
25594 video at the center of the padded area:
25595 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
25597 • Pad the input to get a final w/h ratio of 16:9:
25599 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
25601 • In case of anamorphic video, in order to set the output display
25603 aspect correctly, it is necessary to use sar in the expression,
25604 according to the relation:
25605 (ih * X / ih) * sar = output_dar
25607 X = output_dar / sar
25608 Thus the previous example needs to be modified to:
25610 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
25612 • Double the output size and put the input video in the bottom-right
25614 corner of the output padded area:
25615 pad="2*iw:2*ih:ow-iw:oh-ih"
25617 palettegen
25619 Generate one palette for a whole video stream.
25620 It accepts the following options:
25622 max_colors
25624 Set the maximum number of colors to quantize in the palette. Note:
25625 the palette will still contain 256 colors; the unused palette
25626 entries will be black.
25627 reserve_transparent
25629 Create a palette of 255 colors maximum and reserve the last one for
25630 transparency. Reserving the transparency color is useful for GIF
25631 optimization. If not set, the maximum of colors in the palette
25632 will be 256. You probably want to disable this option for a
25633 standalone image. Set by default.
25634 transparency_color
25636 Set the color that will be used as background for transparency.
25637 stats_mode
25639 Set statistics mode.
25640 It accepts the following values:
25642 full
25644 Compute full frame histograms.
25645 diff
25647 Compute histograms only for the part that differs from previous
25648 frame. This might be relevant to give more importance to the
25649 moving part of your input if the background is static.
25650 single
25652 Compute new histogram for each frame.
25653 Default value is full.
25655 The filter also exports the frame metadata "lavfi.color_quant_ratio"
25657 ("nb_color_in / nb_color_out") which you can use to evaluate the degree
25658 of color quantization of the palette. This information is also visible
25659 at info logging level.
25660 Examples
25662 • Generate a representative palette of a given video using ffmpeg:
25664 ffmpeg -i input.mkv -vf palettegen palette.png
25666 paletteuse
25668 Use a palette to downsample an input video stream.
25669 The filter takes two inputs: one video stream and a palette. The
25671 palette must be a 256 pixels image.
25672 It accepts the following options:
25674 dither
25676 Select dithering mode. Available algorithms are:
25677 bayer
25679 Ordered 8x8 bayer dithering (deterministic)
25680 heckbert
25682 Dithering as defined by Paul Heckbert in 1982 (simple error
25683 diffusion). Note: this dithering is sometimes considered
25684 "wrong" and is included as a reference.
25685 floyd_steinberg
25687 Floyd and Steingberg dithering (error diffusion)
25688 sierra2
25690 Frankie Sierra dithering v2 (error diffusion)
25691 sierra2_4a
25693 Frankie Sierra dithering v2 "Lite" (error diffusion)
25694 sierra3
25696 Frankie Sierra dithering v3 (error diffusion)
25697 burkes
25699 Burkes dithering (error diffusion)
25700 atkinson
25702 Atkinson dithering by Bill Atkinson at Apple Computer (error
25703 diffusion)
25704 Default is sierra2_4a.
25706 bayer_scale
25708 When bayer dithering is selected, this option defines the scale of
25709 the pattern (how much the crosshatch pattern is visible). A low
25710 value means more visible pattern for less banding, and higher value
25711 means less visible pattern at the cost of more banding.
25712 The option must be an integer value in the range [0,5]. Default is
25714 2.
25715 diff_mode
25717 If set, define the zone to process
25718 rectangle
25720 Only the changing rectangle will be reprocessed. This is
25721 similar to GIF cropping/offsetting compression mechanism. This
25722 option can be useful for speed if only a part of the image is
25723 changing, and has use cases such as limiting the scope of the
25724 error diffusal dither to the rectangle that bounds the moving
25725 scene (it leads to more deterministic output if the scene
25726 doesn't change much, and as a result less moving noise and
25727 better GIF compression).
25728 Default is none.
25730 new Take new palette for each output frame.
25732 alpha_threshold
25734 Sets the alpha threshold for transparency. Alpha values above this
25735 threshold will be treated as completely opaque, and values below
25736 this threshold will be treated as completely transparent.
25737 The option must be an integer value in the range [0,255]. Default
25739 is 128.
25740 Examples
25742 • Use a palette (generated for example with palettegen) to encode a
25744 GIF using ffmpeg:
25745 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
25747 perspective
25749 Correct perspective of video not recorded perpendicular to the screen.
25750 A description of the accepted parameters follows.
25752 x0
25754 y0
25755 x1
25756 y1
25757 x2
25758 y2
25759 x3
25760 y3 Set coordinates expression for top left, top right, bottom left and
25761 bottom right corners. Default values are "0:0:W:0:0:H:W:H" with
25762 which perspective will remain unchanged. If the "sense" option is
25763 set to "source", then the specified points will be sent to the
25764 corners of the destination. If the "sense" option is set to
25765 "destination", then the corners of the source will be sent to the
25766 specified coordinates.
25767 The expressions can use the following variables:
25769 W
25771 H the width and height of video frame.
25772 in Input frame count.
25774 on Output frame count.
25776 interpolation
25778 Set interpolation for perspective correction.
25779 It accepts the following values:
25781 linear
25783 cubic
25784 Default value is linear.
25786 sense
25788 Set interpretation of coordinate options.
25789 It accepts the following values:
25791 0, source
25793 Send point in the source specified by the given coordinates to
25794 the corners of the destination.
25795 1, destination
25797 Send the corners of the source to the point in the destination
25798 specified by the given coordinates.
25799 Default value is source.
25801 eval
25803 Set when the expressions for coordinates x0,y0,...x3,y3 are
25804 evaluated.
25805 It accepts the following values:
25807 init
25809 only evaluate expressions once during the filter initialization
25810 or when a command is processed
25811 frame
25813 evaluate expressions for each incoming frame
25814 Default value is init.
25816 phase
25818 Delay interlaced video by one field time so that the field order
25819 changes.
25820 The intended use is to fix PAL movies that have been captured with the
25822 opposite field order to the film-to-video transfer.
25823 A description of the accepted parameters follows.
25825 mode
25827 Set phase mode.
25828 It accepts the following values:
25830 t Capture field order top-first, transfer bottom-first. Filter
25832 will delay the bottom field.
25833 b Capture field order bottom-first, transfer top-first. Filter
25835 will delay the top field.
25836 p Capture and transfer with the same field order. This mode only
25838 exists for the documentation of the other options to refer to,
25839 but if you actually select it, the filter will faithfully do
25840 nothing.
25841 a Capture field order determined automatically by field flags,
25843 transfer opposite. Filter selects among t and b modes on a
25844 frame by frame basis using field flags. If no field information
25845 is available, then this works just like u.
25846 u Capture unknown or varying, transfer opposite. Filter selects
25848 among t and b on a frame by frame basis by analyzing the images
25849 and selecting the alternative that produces best match between
25850 the fields.
25851 T Capture top-first, transfer unknown or varying. Filter selects
25853 among t and p using image analysis.
25854 B Capture bottom-first, transfer unknown or varying. Filter
25856 selects among b and p using image analysis.
25857 A Capture determined by field flags, transfer unknown or varying.
25859 Filter selects among t, b and p using field flags and image
25860 analysis. If no field information is available, then this works
25861 just like U. This is the default mode.
25862 U Both capture and transfer unknown or varying. Filter selects
25864 among t, b and p using image analysis only.
25865 Commands
25867 This filter supports the all above options as commands.
25869 photosensitivity
25871 Reduce various flashes in video, so to help users with epilepsy.
25872 It accepts the following options:
25874 frames, f
25876 Set how many frames to use when filtering. Default is 30.
25877 threshold, t
25879 Set detection threshold factor. Default is 1. Lower is stricter.
25880 skip
25882 Set how many pixels to skip when sampling frames. Default is 1.
25883 Allowed range is from 1 to 1024.
25884 bypass
25886 Leave frames unchanged. Default is disabled.
25887 pixdesctest
25889 Pixel format descriptor test filter, mainly useful for internal
25890 testing. The output video should be equal to the input video.
25891 For example:
25893 format=monow, pixdesctest
25895 can be used to test the monowhite pixel format descriptor definition.
25897 pixelize
25899 Apply pixelization to video stream.
25900 The filter accepts the following options:
25902 width, w
25904 height, h
25905 Set block dimensions that will be used for pixelization. Default
25906 value is 16.
25907 mode, m
25909 Set the mode of pixelization used.
25910 Possible values are:
25912 avg
25914 min
25915 max
25916 Default value is "avg".
25918 planes, p
25920 Set what planes to filter. Default is to filter all planes.
25921 Commands
25923 This filter supports all options as commands.
25925 pixscope
25927 Display sample values of color channels. Mainly useful for checking
25928 color and levels. Minimum supported resolution is 640x480.
25929 The filters accept the following options:
25931 x Set scope X position, relative offset on X axis.
25933 y Set scope Y position, relative offset on Y axis.
25935 w Set scope width.
25937 h Set scope height.
25939 o Set window opacity. This window also holds statistics about pixel
25941 area.
25942 wx Set window X position, relative offset on X axis.
25944 wy Set window Y position, relative offset on Y axis.
25946 Commands
25948 This filter supports same commands as options.
25950 pp
25952 Enable the specified chain of postprocessing subfilters using
25953 libpostproc. This library should be automatically selected with a GPL
25954 build ("--enable-gpl"). Subfilters must be separated by '/' and can be
25955 disabled by prepending a '-'. Each subfilter and some options have a
25956 short and a long name that can be used interchangeably, i.e. dr/dering
25957 are the same.
25958 The filters accept the following options:
25960 subfilters
25962 Set postprocessing subfilters string.
25963 All subfilters share common options to determine their scope:
25965 a/autoq
25967 Honor the quality commands for this subfilter.
25968 c/chrom
25970 Do chrominance filtering, too (default).
25971 y/nochrom
25973 Do luminance filtering only (no chrominance).
25974 n/noluma
25976 Do chrominance filtering only (no luminance).
25977 These options can be appended after the subfilter name, separated by a
25979 '|'.
25980 Available subfilters are:
25982 hb/hdeblock[|difference[|flatness]]
25984 Horizontal deblocking filter
25985 difference
25987 Difference factor where higher values mean more deblocking
25988 (default: 32).
25989 flatness
25991 Flatness threshold where lower values mean more deblocking
25992 (default: 39).
25993 vb/vdeblock[|difference[|flatness]]
25995 Vertical deblocking filter
25996 difference
25998 Difference factor where higher values mean more deblocking
25999 (default: 32).
26000 flatness
26002 Flatness threshold where lower values mean more deblocking
26003 (default: 39).
26004 ha/hadeblock[|difference[|flatness]]
26006 Accurate horizontal deblocking filter
26007 difference
26009 Difference factor where higher values mean more deblocking
26010 (default: 32).
26011 flatness
26013 Flatness threshold where lower values mean more deblocking
26014 (default: 39).
26015 va/vadeblock[|difference[|flatness]]
26017 Accurate vertical deblocking filter
26018 difference
26020 Difference factor where higher values mean more deblocking
26021 (default: 32).
26022 flatness
26024 Flatness threshold where lower values mean more deblocking
26025 (default: 39).
26026 The horizontal and vertical deblocking filters share the difference and
26028 flatness values so you cannot set different horizontal and vertical
26029 thresholds.
26030 h1/x1hdeblock
26032 Experimental horizontal deblocking filter
26033 v1/x1vdeblock
26035 Experimental vertical deblocking filter
26036 dr/dering
26038 Deringing filter
26039 tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise
26041 reducer
26042 threshold1
26043 larger -> stronger filtering
26044 threshold2
26046 larger -> stronger filtering
26047 threshold3
26049 larger -> stronger filtering
26050 al/autolevels[:f/fullyrange], automatic brightness / contrast
26052 correction
26053 f/fullyrange
26054 Stretch luminance to "0-255".
26055 lb/linblenddeint
26057 Linear blend deinterlacing filter that deinterlaces the given block
26058 by filtering all lines with a "(1 2 1)" filter.
26059 li/linipoldeint
26061 Linear interpolating deinterlacing filter that deinterlaces the
26062 given block by linearly interpolating every second line.
26063 ci/cubicipoldeint
26065 Cubic interpolating deinterlacing filter deinterlaces the given
26066 block by cubically interpolating every second line.
26067 md/mediandeint
26069 Median deinterlacing filter that deinterlaces the given block by
26070 applying a median filter to every second line.
26071 fd/ffmpegdeint
26073 FFmpeg deinterlacing filter that deinterlaces the given block by
26074 filtering every second line with a "(-1 4 2 4 -1)" filter.
26075 l5/lowpass5
26077 Vertically applied FIR lowpass deinterlacing filter that
26078 deinterlaces the given block by filtering all lines with a "(-1 2 6
26079 2 -1)" filter.
26080 fq/forceQuant[|quantizer]
26082 Overrides the quantizer table from the input with the constant
26083 quantizer you specify.
26084 quantizer
26086 Quantizer to use
26087 de/default
26089 Default pp filter combination ("hb|a,vb|a,dr|a")
26090 fa/fast
26092 Fast pp filter combination ("h1|a,v1|a,dr|a")
26093 ac High quality pp filter combination ("ha|a|128|7,va|a,dr|a")
26095 Examples
26097 • Apply horizontal and vertical deblocking, deringing and automatic
26099 brightness/contrast:
26100 pp=hb/vb/dr/al
26102 • Apply default filters without brightness/contrast correction:
26104 pp=de/-al
26106 • Apply default filters and temporal denoiser:
26108 pp=default/tmpnoise|1|2|3
26110 • Apply deblocking on luminance only, and switch vertical deblocking
26112 on or off automatically depending on available CPU time:
26113 pp=hb|y/vb|a
26115 pp7
26117 Apply Postprocessing filter 7. It is variant of the spp filter, similar
26118 to spp = 6 with 7 point DCT, where only the center sample is used after
26119 IDCT.
26120 The filter accepts the following options:
26122 qp Force a constant quantization parameter. It accepts an integer in
26124 range 0 to 63. If not set, the filter will use the QP from the
26125 video stream (if available).
26126 mode
26128 Set thresholding mode. Available modes are:
26129 hard
26131 Set hard thresholding.
26132 soft
26134 Set soft thresholding (better de-ringing effect, but likely
26135 blurrier).
26136 medium
26138 Set medium thresholding (good results, default).
26139 premultiply
26141 Apply alpha premultiply effect to input video stream using first plane
26142 of second stream as alpha.
26143 Both streams must have same dimensions and same pixel format.
26145 The filter accepts the following option:
26147 planes
26149 Set which planes will be processed, unprocessed planes will be
26150 copied. By default value 0xf, all planes will be processed.
26151 inplace
26153 Do not require 2nd input for processing, instead use alpha plane
26154 from input stream.
26155 prewitt
26157 Apply prewitt operator to input video stream.
26158 The filter accepts the following option:
26160 planes
26162 Set which planes will be processed, unprocessed planes will be
26163 copied. By default value 0xf, all planes will be processed.
26164 scale
26166 Set value which will be multiplied with filtered result.
26167 delta
26169 Set value which will be added to filtered result.
26170 Commands
26172 This filter supports the all above options as commands.
26174 pseudocolor
26176 Alter frame colors in video with pseudocolors.
26177 This filter accepts the following options:
26179 c0 set pixel first component expression
26181 c1 set pixel second component expression
26183 c2 set pixel third component expression
26185 c3 set pixel fourth component expression, corresponds to the alpha
26187 component
26188 index, i
26190 set component to use as base for altering colors
26191 preset, p
26193 Pick one of built-in LUTs. By default is set to none.
26194 Available LUTs:
26196 magma
26198 inferno
26199 plasma
26200 viridis
26201 turbo
26202 cividis
26203 range1
26204 range2
26205 shadows
26206 highlights
26207 solar
26208 nominal
26209 preferred
26210 total
26211 spectral
26212 opacity
26213 Set opacity of output colors. Allowed range is from 0 to 1.
26214 Default value is set to 1.
26215 Each of the expression options specifies the expression to use for
26217 computing the lookup table for the corresponding pixel component
26218 values.
26219 The expressions can contain the following constants and functions:
26221 w
26223 h The input width and height.
26224 val The input value for the pixel component.
26226 ymin, umin, vmin, amin
26228 The minimum allowed component value.
26229 ymax, umax, vmax, amax
26231 The maximum allowed component value.
26232 All expressions default to "val".
26234 Commands
26236 This filter supports the all above options as commands.
26238 Examples
26240 • Change too high luma values to gradient:
26242 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'"
26244 psnr
26246 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
26247 Ratio) between two input videos.
26248 This filter takes in input two input videos, the first input is
26250 considered the "main" source and is passed unchanged to the output. The
26251 second input is used as a "reference" video for computing the PSNR.
26252 Both video inputs must have the same resolution and pixel format for
26254 this filter to work correctly. Also it assumes that both inputs have
26255 the same number of frames, which are compared one by one.
26256 The obtained average PSNR is printed through the logging system.
26258 The filter stores the accumulated MSE (mean squared error) of each
26260 frame, and at the end of the processing it is averaged across all
26261 frames equally, and the following formula is applied to obtain the
26262 PSNR:
26263 PSNR = 10*log10(MAX^2/MSE)
26265 Where MAX is the average of the maximum values of each component of the
26267 image.
26268 The description of the accepted parameters follows.
26270 stats_file, f
26272 If specified the filter will use the named file to save the PSNR of
26273 each individual frame. When filename equals "-" the data is sent to
26274 standard output.
26275 stats_version
26277 Specifies which version of the stats file format to use. Details of
26278 each format are written below. Default value is 1.
26279 stats_add_max
26281 Determines whether the max value is output to the stats log.
26282 Default value is 0. Requires stats_version >= 2. If this is set
26283 and stats_version < 2, the filter will return an error.
26284 This filter also supports the framesync options.
26286 The file printed if stats_file is selected, contains a sequence of
26288 key/value pairs of the form key:value for each compared couple of
26289 frames.
26290 If a stats_version greater than 1 is specified, a header line precedes
26292 the list of per-frame-pair stats, with key value pairs following the
26293 frame format with the following parameters:
26294 psnr_log_version
26296 The version of the log file format. Will match stats_version.
26297 fields
26299 A comma separated list of the per-frame-pair parameters included in
26300 the log.
26301 A description of each shown per-frame-pair parameter follows:
26303 n sequential number of the input frame, starting from 1
26305 mse_avg
26307 Mean Square Error pixel-by-pixel average difference of the compared
26308 frames, averaged over all the image components.
26309 mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
26311 Mean Square Error pixel-by-pixel average difference of the compared
26312 frames for the component specified by the suffix.
26313 psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
26315 Peak Signal to Noise ratio of the compared frames for the component
26316 specified by the suffix.
26317 max_avg, max_y, max_u, max_v
26319 Maximum allowed value for each channel, and average over all
26320 channels.
26321 Examples
26323 • For example:
26325 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
26327 [main][ref] psnr="stats_file=stats.log" [out]
26328 On this example the input file being processed is compared with the
26330 reference file ref_movie.mpg. The PSNR of each individual frame is
26331 stored in stats.log.
26332 • Another example with different containers:
26334 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 -
26336 pullup
26338 Pulldown reversal (inverse telecine) filter, capable of handling mixed
26339 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps
26340 progressive content.
26341 The pullup filter is designed to take advantage of future context in
26343 making its decisions. This filter is stateless in the sense that it
26344 does not lock onto a pattern to follow, but it instead looks forward to
26345 the following fields in order to identify matches and rebuild
26346 progressive frames.
26347 To produce content with an even framerate, insert the fps filter after
26349 pullup, use "fps=24000/1001" if the input frame rate is 29.97fps,
26350 "fps=24" for 30fps and the (rare) telecined 25fps input.
26351 The filter accepts the following options:
26353 jl
26355 jr
26356 jt
26357 jb These options set the amount of "junk" to ignore at the left,
26358 right, top, and bottom of the image, respectively. Left and right
26359 are in units of 8 pixels, while top and bottom are in units of 2
26360 lines. The default is 8 pixels on each side.
26361 sb Set the strict breaks. Setting this option to 1 will reduce the
26363 chances of filter generating an occasional mismatched frame, but it
26364 may also cause an excessive number of frames to be dropped during
26365 high motion sequences. Conversely, setting it to -1 will make
26366 filter match fields more easily. This may help processing of video
26367 where there is slight blurring between the fields, but may also
26368 cause there to be interlaced frames in the output. Default value
26369 is 0.
26370 mp Set the metric plane to use. It accepts the following values:
26372 l Use luma plane.
26374 u Use chroma blue plane.
26376 v Use chroma red plane.
26378 This option may be set to use chroma plane instead of the default
26380 luma plane for doing filter's computations. This may improve
26381 accuracy on very clean source material, but more likely will
26382 decrease accuracy, especially if there is chroma noise (rainbow
26383 effect) or any grayscale video. The main purpose of setting mp to
26384 a chroma plane is to reduce CPU load and make pullup usable in
26385 realtime on slow machines.
26386 For best results (without duplicated frames in the output file) it is
26388 necessary to change the output frame rate. For example, to inverse
26389 telecine NTSC input:
26390 ffmpeg -i input -vf pullup -r 24000/1001 ...
26392 qp
26394 Change video quantization parameters (QP).
26395 The filter accepts the following option:
26397 qp Set expression for quantization parameter.
26399 The expression is evaluated through the eval API and can contain, among
26401 others, the following constants:
26402 known
26404 1 if index is not 129, 0 otherwise.
26405 qp Sequential index starting from -129 to 128.
26407 Examples
26409 • Some equation like:
26411 qp=2+2*sin(PI*qp)
26413 random
26415 Flush video frames from internal cache of frames into a random order.
26416 No frame is discarded. Inspired by frei0r nervous filter.
26417 frames
26419 Set size in number of frames of internal cache, in range from 2 to
26420 512. Default is 30.
26421 seed
26423 Set seed for random number generator, must be an integer included
26424 between 0 and "UINT32_MAX". If not specified, or if explicitly set
26425 to less than 0, the filter will try to use a good random seed on a
26426 best effort basis.
26427 readeia608
26429 Read closed captioning (EIA-608) information from the top lines of a
26430 video frame.
26431 This filter adds frame metadata for "lavfi.readeia608.X.cc" and
26433 "lavfi.readeia608.X.line", where "X" is the number of the identified
26434 line with EIA-608 data (starting from 0). A description of each
26435 metadata value follows:
26436 lavfi.readeia608.X.cc
26438 The two bytes stored as EIA-608 data (printed in hexadecimal).
26439 lavfi.readeia608.X.line
26441 The number of the line on which the EIA-608 data was identified and
26442 read.
26443 This filter accepts the following options:
26445 scan_min
26447 Set the line to start scanning for EIA-608 data. Default is 0.
26448 scan_max
26450 Set the line to end scanning for EIA-608 data. Default is 29.
26451 spw Set the ratio of width reserved for sync code detection. Default
26453 is 0.27. Allowed range is "[0.1 - 0.7]".
26454 chp Enable checking the parity bit. In the event of a parity error, the
26456 filter will output 0x00 for that character. Default is false.
26457 lp Lowpass lines prior to further processing. Default is enabled.
26459 Commands
26461 This filter supports the all above options as commands.
26463 Examples
26465 • Output a csv with presentation time and the first two lines of
26467 identified EIA-608 captioning data.
26468 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
26470 readvitc
26472 Read vertical interval timecode (VITC) information from the top lines
26473 of a video frame.
26474 The filter adds frame metadata key "lavfi.readvitc.tc_str" with the
26476 timecode value, if a valid timecode has been detected. Further metadata
26477 key "lavfi.readvitc.found" is set to 0/1 depending on whether timecode
26478 data has been found or not.
26479 This filter accepts the following options:
26481 scan_max
26483 Set the maximum number of lines to scan for VITC data. If the value
26484 is set to -1 the full video frame is scanned. Default is 45.
26485 thr_b
26487 Set the luma threshold for black. Accepts float numbers in the
26488 range [0.0,1.0], default value is 0.2. The value must be equal or
26489 less than "thr_w".
26490 thr_w
26492 Set the luma threshold for white. Accepts float numbers in the
26493 range [0.0,1.0], default value is 0.6. The value must be equal or
26494 greater than "thr_b".
26495 Examples
26497 • Detect and draw VITC data onto the video frame; if no valid VITC is
26499 detected, draw "--:--:--:--" as a placeholder:
26500 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--}:x=(w-tw)/2:y=400-ascent'
26502 remap
26504 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
26505 Destination pixel at position (X, Y) will be picked from source (x, y)
26507 position where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are
26508 out of range, zero value for pixel will be used for destination pixel.
26509 Xmap and Ymap input video streams must be of same dimensions. Output
26511 video stream will have Xmap/Ymap video stream dimensions. Xmap and
26512 Ymap input video streams are 16bit depth, single channel.
26513 format
26515 Specify pixel format of output from this filter. Can be "color" or
26516 "gray". Default is "color".
26517 fill
26519 Specify the color of the unmapped pixels. For the syntax of this
26520 option, check the "Color" section in the ffmpeg-utils manual.
26521 Default color is "black".
26522 removegrain
26524 The removegrain filter is a spatial denoiser for progressive video.
26525 m0 Set mode for the first plane.
26527 m1 Set mode for the second plane.
26529 m2 Set mode for the third plane.
26531 m3 Set mode for the fourth plane.
26533 Range of mode is from 0 to 24. Description of each mode follows:
26535 0 Leave input plane unchanged. Default.
26537 1 Clips the pixel with the minimum and maximum of the 8 neighbour
26539 pixels.
26540 2 Clips the pixel with the second minimum and maximum of the 8
26542 neighbour pixels.
26543 3 Clips the pixel with the third minimum and maximum of the 8
26545 neighbour pixels.
26546 4 Clips the pixel with the fourth minimum and maximum of the 8
26548 neighbour pixels. This is equivalent to a median filter.
26549 5 Line-sensitive clipping giving the minimal change.
26551 6 Line-sensitive clipping, intermediate.
26553 7 Line-sensitive clipping, intermediate.
26555 8 Line-sensitive clipping, intermediate.
26557 9 Line-sensitive clipping on a line where the neighbours pixels are
26559 the closest.
26560 10 Replaces the target pixel with the closest neighbour.
26562 11 [1 2 1] horizontal and vertical kernel blur.
26564 12 Same as mode 11.
26566 13 Bob mode, interpolates top field from the line where the neighbours
26568 pixels are the closest.
26569 14 Bob mode, interpolates bottom field from the line where the
26571 neighbours pixels are the closest.
26572 15 Bob mode, interpolates top field. Same as 13 but with a more
26574 complicated interpolation formula.
26575 16 Bob mode, interpolates bottom field. Same as 14 but with a more
26577 complicated interpolation formula.
26578 17 Clips the pixel with the minimum and maximum of respectively the
26580 maximum and minimum of each pair of opposite neighbour pixels.
26581 18 Line-sensitive clipping using opposite neighbours whose greatest
26583 distance from the current pixel is minimal.
26584 19 Replaces the pixel with the average of its 8 neighbours.
26586 20 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
26588 21 Clips pixels using the averages of opposite neighbour.
26590 22 Same as mode 21 but simpler and faster.
26592 23 Small edge and halo removal, but reputed useless.
26594 24 Similar as 23.
26596 removelogo
26598 Suppress a TV station logo, using an image file to determine which
26599 pixels comprise the logo. It works by filling in the pixels that
26600 comprise the logo with neighboring pixels.
26601 The filter accepts the following options:
26603 filename, f
26605 Set the filter bitmap file, which can be any image format supported
26606 by libavformat. The width and height of the image file must match
26607 those of the video stream being processed.
26608 Pixels in the provided bitmap image with a value of zero are not
26610 considered part of the logo, non-zero pixels are considered part of the
26611 logo. If you use white (255) for the logo and black (0) for the rest,
26612 you will be safe. For making the filter bitmap, it is recommended to
26613 take a screen capture of a black frame with the logo visible, and then
26614 using a threshold filter followed by the erode filter once or twice.
26615 If needed, little splotches can be fixed manually. Remember that if
26617 logo pixels are not covered, the filter quality will be much reduced.
26618 Marking too many pixels as part of the logo does not hurt as much, but
26619 it will increase the amount of blurring needed to cover over the image
26620 and will destroy more information than necessary, and extra pixels will
26621 slow things down on a large logo.
26622 repeatfields
26624 This filter uses the repeat_field flag from the Video ES headers and
26625 hard repeats fields based on its value.
26626 reverse
26628 Reverse a video clip.
26629 Warning: This filter requires memory to buffer the entire clip, so
26631 trimming is suggested.
26632 Examples
26634 • Take the first 5 seconds of a clip, and reverse it.
26636 trim=end=5,reverse
26638 rgbashift
26640 Shift R/G/B/A pixels horizontally and/or vertically.
26641 The filter accepts the following options:
26643 rh Set amount to shift red horizontally.
26645 rv Set amount to shift red vertically.
26647 gh Set amount to shift green horizontally.
26649 gv Set amount to shift green vertically.
26651 bh Set amount to shift blue horizontally.
26653 bv Set amount to shift blue vertically.
26655 ah Set amount to shift alpha horizontally.
26657 av Set amount to shift alpha vertically.
26659 edge
26661 Set edge mode, can be smear, default, or warp.
26662 Commands
26664 This filter supports the all above options as commands.
26666 roberts
26668 Apply roberts cross operator to input video stream.
26669 The filter accepts the following option:
26671 planes
26673 Set which planes will be processed, unprocessed planes will be
26674 copied. By default value 0xf, all planes will be processed.
26675 scale
26677 Set value which will be multiplied with filtered result.
26678 delta
26680 Set value which will be added to filtered result.
26681 Commands
26683 This filter supports the all above options as commands.
26685 rotate
26687 Rotate video by an arbitrary angle expressed in radians.
26688 The filter accepts the following options:
26690 A description of the optional parameters follows.
26692 angle, a
26694 Set an expression for the angle by which to rotate the input video
26695 clockwise, expressed as a number of radians. A negative value will
26696 result in a counter-clockwise rotation. By default it is set to
26697 "0".
26698 This expression is evaluated for each frame.
26700 out_w, ow
26702 Set the output width expression, default value is "iw". This
26703 expression is evaluated just once during configuration.
26704 out_h, oh
26706 Set the output height expression, default value is "ih". This
26707 expression is evaluated just once during configuration.
26708 bilinear
26710 Enable bilinear interpolation if set to 1, a value of 0 disables
26711 it. Default value is 1.
26712 fillcolor, c
26714 Set the color used to fill the output area not covered by the
26715 rotated image. For the general syntax of this option, check the
26716 "Color" section in the ffmpeg-utils manual. If the special value
26717 "none" is selected then no background is printed (useful for
26718 example if the background is never shown).
26719 Default value is "black".
26721 The expressions for the angle and the output size can contain the
26723 following constants and functions:
26724 n sequential number of the input frame, starting from 0. It is always
26726 NAN before the first frame is filtered.
26727 t time in seconds of the input frame, it is set to 0 when the filter
26729 is configured. It is always NAN before the first frame is filtered.
26730 hsub
26732 vsub
26733 horizontal and vertical chroma subsample values. For example for
26734 the pixel format "yuv422p" hsub is 2 and vsub is 1.
26735 in_w, iw
26737 in_h, ih
26738 the input video width and height
26739 out_w, ow
26741 out_h, oh
26742 the output width and height, that is the size of the padded area as
26743 specified by the width and height expressions
26744 rotw(a)
26746 roth(a)
26747 the minimal width/height required for completely containing the
26748 input video rotated by a radians.
26749 These are only available when computing the out_w and out_h
26751 expressions.
26752 Examples
26754 • Rotate the input by PI/6 radians clockwise:
26756 rotate=PI/6
26758 • Rotate the input by PI/6 radians counter-clockwise:
26760 rotate=-PI/6
26762 • Rotate the input by 45 degrees clockwise:
26764 rotate=45*PI/180
26766 • Apply a constant rotation with period T, starting from an angle of
26768 PI/3:
26769 rotate=PI/3+2*PI*t/T
26771 • Make the input video rotation oscillating with a period of T
26773 seconds and an amplitude of A radians:
26774 rotate=A*sin(2*PI/T*t)
26776 • Rotate the video, output size is chosen so that the whole rotating
26778 input video is always completely contained in the output:
26779 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
26781 • Rotate the video, reduce the output size so that no background is
26783 ever shown:
26784 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
26786 Commands
26788 The filter supports the following commands:
26790 a, angle
26792 Set the angle expression. The command accepts the same syntax of
26793 the corresponding option.
26794 If the specified expression is not valid, it is kept at its current
26796 value.
26797 sab
26799 Apply Shape Adaptive Blur.
26800 The filter accepts the following options:
26802 luma_radius, lr
26804 Set luma blur filter strength, must be a value in range 0.1-4.0,
26805 default value is 1.0. A greater value will result in a more blurred
26806 image, and in slower processing.
26807 luma_pre_filter_radius, lpfr
26809 Set luma pre-filter radius, must be a value in the 0.1-2.0 range,
26810 default value is 1.0.
26811 luma_strength, ls
26813 Set luma maximum difference between pixels to still be considered,
26814 must be a value in the 0.1-100.0 range, default value is 1.0.
26815 chroma_radius, cr
26817 Set chroma blur filter strength, must be a value in range -0.9-4.0.
26818 A greater value will result in a more blurred image, and in slower
26819 processing.
26820 chroma_pre_filter_radius, cpfr
26822 Set chroma pre-filter radius, must be a value in the -0.9-2.0
26823 range.
26824 chroma_strength, cs
26826 Set chroma maximum difference between pixels to still be
26827 considered, must be a value in the -0.9-100.0 range.
26828 Each chroma option value, if not explicitly specified, is set to the
26830 corresponding luma option value.
26831 scale
26833 Scale (resize) the input video, using the libswscale library.
26834 The scale filter forces the output display aspect ratio to be the same
26836 of the input, by changing the output sample aspect ratio.
26837 If the input image format is different from the format requested by the
26839 next filter, the scale filter will convert the input to the requested
26840 format.
26841 Options
26843 The filter accepts the following options, or any of the options
26845 supported by the libswscale scaler.
26846 See the ffmpeg-scaler manual for the complete list of scaler options.
26848 width, w
26850 height, h
26851 Set the output video dimension expression. Default value is the
26852 input dimension.
26853 If the width or w value is 0, the input width is used for the
26855 output. If the height or h value is 0, the input height is used for
26856 the output.
26857 If one and only one of the values is -n with n >= 1, the scale
26859 filter will use a value that maintains the aspect ratio of the
26860 input image, calculated from the other specified dimension. After
26861 that it will, however, make sure that the calculated dimension is
26862 divisible by n and adjust the value if necessary.
26863 If both values are -n with n >= 1, the behavior will be identical
26865 to both values being set to 0 as previously detailed.
26866 See below for the list of accepted constants for use in the
26868 dimension expression.
26869 eval
26871 Specify when to evaluate width and height expression. It accepts
26872 the following values:
26873 init
26875 Only evaluate expressions once during the filter initialization
26876 or when a command is processed.
26877 frame
26879 Evaluate expressions for each incoming frame.
26880 Default value is init.
26882 interl
26884 Set the interlacing mode. It accepts the following values:
26885 1 Force interlaced aware scaling.
26887 0 Do not apply interlaced scaling.
26889 -1 Select interlaced aware scaling depending on whether the source
26891 frames are flagged as interlaced or not.
26892 Default value is 0.
26894 flags
26896 Set libswscale scaling flags. See the ffmpeg-scaler manual for the
26897 complete list of values. If not explicitly specified the filter
26898 applies the default flags.
26899 param0, param1
26901 Set libswscale input parameters for scaling algorithms that need
26902 them. See the ffmpeg-scaler manual for the complete documentation.
26903 If not explicitly specified the filter applies empty parameters.
26904 size, s
26906 Set the video size. For the syntax of this option, check the "Video
26907 size" section in the ffmpeg-utils manual.
26908 in_color_matrix
26910 out_color_matrix
26911 Set in/output YCbCr color space type.
26912 This allows the autodetected value to be overridden as well as
26914 allows forcing a specific value used for the output and encoder.
26915 If not specified, the color space type depends on the pixel format.
26917 Possible values:
26919 auto
26921 Choose automatically.
26922 bt709
26924 Format conforming to International Telecommunication Union
26925 (ITU) Recommendation BT.709.
26926 fcc Set color space conforming to the United States Federal
26928 Communications Commission (FCC) Code of Federal Regulations
26929 (CFR) Title 47 (2003) 73.682 (a).
26930 bt601
26932 bt470
26933 smpte170m
26934 Set color space conforming to:
26935 • ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
26937 • ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
26939 • Society of Motion Picture and Television Engineers (SMPTE)
26941 ST 170:2004
26942 smpte240m
26944 Set color space conforming to SMPTE ST 240:1999.
26945 bt2020
26947 Set color space conforming to ITU-R BT.2020 non-constant
26948 luminance system.
26949 in_range
26951 out_range
26952 Set in/output YCbCr sample range.
26953 This allows the autodetected value to be overridden as well as
26955 allows forcing a specific value used for the output and encoder. If
26956 not specified, the range depends on the pixel format. Possible
26957 values:
26958 auto/unknown
26960 Choose automatically.
26961 jpeg/full/pc
26963 Set full range (0-255 in case of 8-bit luma).
26964 mpeg/limited/tv
26966 Set "MPEG" range (16-235 in case of 8-bit luma).
26967 force_original_aspect_ratio
26969 Enable decreasing or increasing output video width or height if
26970 necessary to keep the original aspect ratio. Possible values:
26971 disable
26973 Scale the video as specified and disable this feature.
26974 decrease
26976 The output video dimensions will automatically be decreased if
26977 needed.
26978 increase
26980 The output video dimensions will automatically be increased if
26981 needed.
26982 One useful instance of this option is that when you know a specific
26984 device's maximum allowed resolution, you can use this to limit the
26985 output video to that, while retaining the aspect ratio. For
26986 example, device A allows 1280x720 playback, and your video is
26987 1920x800. Using this option (set it to decrease) and specifying
26988 1280x720 to the command line makes the output 1280x533.
26989 Please note that this is a different thing than specifying -1 for w
26991 or h, you still need to specify the output resolution for this
26992 option to work.
26993 force_divisible_by
26995 Ensures that both the output dimensions, width and height, are
26996 divisible by the given integer when used together with
26997 force_original_aspect_ratio. This works similar to using "-n" in
26998 the w and h options.
26999 This option respects the value set for force_original_aspect_ratio,
27001 increasing or decreasing the resolution accordingly. The video's
27002 aspect ratio may be slightly modified.
27003 This option can be handy if you need to have a video fit within or
27005 exceed a defined resolution using force_original_aspect_ratio but
27006 also have encoder restrictions on width or height divisibility.
27007 The values of the w and h options are expressions containing the
27009 following constants:
27010 in_w
27012 in_h
27013 The input width and height
27014 iw
27016 ih These are the same as in_w and in_h.
27017 out_w
27019 out_h
27020 The output (scaled) width and height
27021 ow
27023 oh These are the same as out_w and out_h
27024 a The same as iw / ih
27026 sar input sample aspect ratio
27028 dar The input display aspect ratio. Calculated from "(iw / ih) * sar".
27030 hsub
27032 vsub
27033 horizontal and vertical input chroma subsample values. For example
27034 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
27035 ohsub
27037 ovsub
27038 horizontal and vertical output chroma subsample values. For example
27039 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
27040 n The (sequential) number of the input frame, starting from 0. Only
27042 available with "eval=frame".
27043 t The presentation timestamp of the input frame, expressed as a
27045 number of seconds. Only available with "eval=frame".
27046 pos The position (byte offset) of the frame in the input stream, or NaN
27048 if this information is unavailable and/or meaningless (for example
27049 in case of synthetic video). Only available with "eval=frame".
27050 Examples
27052 • Scale the input video to a size of 200x100
27054 scale=w=200:h=100
27056 This is equivalent to:
27058 scale=200:100
27060 or:
27062 scale=200x100
27064 • Specify a size abbreviation for the output size:
27066 scale=qcif
27068 which can also be written as:
27070 scale=size=qcif
27072 • Scale the input to 2x:
27074 scale=w=2*iw:h=2*ih
27076 • The above is the same as:
27078 scale=2*in_w:2*in_h
27080 • Scale the input to 2x with forced interlaced scaling:
27082 scale=2*iw:2*ih:interl=1
27084 • Scale the input to half size:
27086 scale=w=iw/2:h=ih/2
27088 • Increase the width, and set the height to the same size:
27090 scale=3/2*iw:ow
27092 • Seek Greek harmony:
27094 scale=iw:1/PHI*iw
27096 scale=ih*PHI:ih
27097 • Increase the height, and set the width to 3/2 of the height:
27099 scale=w=3/2*oh:h=3/5*ih
27101 • Increase the size, making the size a multiple of the chroma
27103 subsample values:
27104 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
27106 • Increase the width to a maximum of 500 pixels, keeping the same
27108 aspect ratio as the input:
27109 scale=w='min(500\, iw*3/2):h=-1'
27111 • Make pixels square by combining scale and setsar:
27113 scale='trunc(ih*dar):ih',setsar=1/1
27115 • Make pixels square by combining scale and setsar, making sure the
27117 resulting resolution is even (required by some codecs):
27118 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
27120 Commands
27122 This filter supports the following commands:
27124 width, w
27126 height, h
27127 Set the output video dimension expression. The command accepts the
27128 same syntax of the corresponding option.
27129 If the specified expression is not valid, it is kept at its current
27131 value.
27132 scale_cuda
27134 Scale (resize) and convert (pixel format) the input video, using
27135 accelerated CUDA kernels. Setting the output width and height works in
27136 the same way as for the scale filter.
27137 The filter accepts the following options:
27139 w
27141 h Set the output video dimension expression. Default value is the
27142 input dimension.
27143 Allows for the same expressions as the scale filter.
27145 interp_algo
27147 Sets the algorithm used for scaling:
27148 nearest
27150 Nearest neighbour
27151 Used by default if input parameters match the desired output.
27153 bilinear
27155 Bilinear
27156 bicubic
27158 Bicubic
27159 This is the default.
27161 lanczos
27163 Lanczos
27164 format
27166 Controls the output pixel format. By default, or if none is
27167 specified, the input pixel format is used.
27168 The filter does not support converting between YUV and RGB pixel
27170 formats.
27171 passthrough
27173 If set to 0, every frame is processed, even if no conversion is
27174 neccesary. This mode can be useful to use the filter as a buffer
27175 for a downstream frame-consumer that exhausts the limited decoder
27176 frame pool.
27177 If set to 1, frames are passed through as-is if they match the
27179 desired output parameters. This is the default behaviour.
27180 param
27182 Algorithm-Specific parameter.
27183 Affects the curves of the bicubic algorithm.
27185 force_original_aspect_ratio
27187 force_divisible_by
27188 Work the same as the identical scale filter options.
27189 Examples
27191 • Scale input to 720p, keeping aspect ratio and ensuring the output
27193 is yuv420p.
27194 scale_cuda=-2:720:format=yuv420p
27196 • Upscale to 4K using nearest neighbour algorithm.
27198 scale_cuda=4096:2160:interp_algo=nearest
27200 • Don't do any conversion or scaling, but copy all input frames into
27202 newly allocated ones. This can be useful to deal with a filter and
27203 encode chain that otherwise exhausts the decoders frame pool.
27204 scale_cuda=passthrough=0
27206 scale_npp
27208 Use the NVIDIA Performance Primitives (libnpp) to perform scaling
27209 and/or pixel format conversion on CUDA video frames. Setting the output
27210 width and height works in the same way as for the scale filter.
27211 The following additional options are accepted:
27213 format
27215 The pixel format of the output CUDA frames. If set to the string
27216 "same" (the default), the input format will be kept. Note that
27217 automatic format negotiation and conversion is not yet supported
27218 for hardware frames
27219 interp_algo
27221 The interpolation algorithm used for resizing. One of the
27222 following:
27223 nn Nearest neighbour.
27225 linear
27227 cubic
27228 cubic2p_bspline
27229 2-parameter cubic (B=1, C=0)
27230 cubic2p_catmullrom
27232 2-parameter cubic (B=0, C=1/2)
27233 cubic2p_b05c03
27235 2-parameter cubic (B=1/2, C=3/10)
27236 super
27238 Supersampling
27239 lanczos
27241 force_original_aspect_ratio
27242 Enable decreasing or increasing output video width or height if
27243 necessary to keep the original aspect ratio. Possible values:
27244 disable
27246 Scale the video as specified and disable this feature.
27247 decrease
27249 The output video dimensions will automatically be decreased if
27250 needed.
27251 increase
27253 The output video dimensions will automatically be increased if
27254 needed.
27255 One useful instance of this option is that when you know a specific
27257 device's maximum allowed resolution, you can use this to limit the
27258 output video to that, while retaining the aspect ratio. For
27259 example, device A allows 1280x720 playback, and your video is
27260 1920x800. Using this option (set it to decrease) and specifying
27261 1280x720 to the command line makes the output 1280x533.
27262 Please note that this is a different thing than specifying -1 for w
27264 or h, you still need to specify the output resolution for this
27265 option to work.
27266 force_divisible_by
27268 Ensures that both the output dimensions, width and height, are
27269 divisible by the given integer when used together with
27270 force_original_aspect_ratio. This works similar to using "-n" in
27271 the w and h options.
27272 This option respects the value set for force_original_aspect_ratio,
27274 increasing or decreasing the resolution accordingly. The video's
27275 aspect ratio may be slightly modified.
27276 This option can be handy if you need to have a video fit within or
27278 exceed a defined resolution using force_original_aspect_ratio but
27279 also have encoder restrictions on width or height divisibility.
27280 eval
27282 Specify when to evaluate width and height expression. It accepts
27283 the following values:
27284 init
27286 Only evaluate expressions once during the filter initialization
27287 or when a command is processed.
27288 frame
27290 Evaluate expressions for each incoming frame.
27291 The values of the w and h options are expressions containing the
27293 following constants:
27294 in_w
27296 in_h
27297 The input width and height
27298 iw
27300 ih These are the same as in_w and in_h.
27301 out_w
27303 out_h
27304 The output (scaled) width and height
27305 ow
27307 oh These are the same as out_w and out_h
27308 a The same as iw / ih
27310 sar input sample aspect ratio
27312 dar The input display aspect ratio. Calculated from "(iw / ih) * sar".
27314 n The (sequential) number of the input frame, starting from 0. Only
27316 available with "eval=frame".
27317 t The presentation timestamp of the input frame, expressed as a
27319 number of seconds. Only available with "eval=frame".
27320 pos The position (byte offset) of the frame in the input stream, or NaN
27322 if this information is unavailable and/or meaningless (for example
27323 in case of synthetic video). Only available with "eval=frame".
27324 scale2ref
27326 Scale (resize) the input video, based on a reference video.
27327 See the scale filter for available options, scale2ref supports the same
27329 but uses the reference video instead of the main input as basis.
27330 scale2ref also supports the following additional constants for the w
27331 and h options:
27332 main_w
27334 main_h
27335 The main input video's width and height
27336 main_a
27338 The same as main_w / main_h
27339 main_sar
27341 The main input video's sample aspect ratio
27342 main_dar, mdar
27344 The main input video's display aspect ratio. Calculated from
27345 "(main_w / main_h) * main_sar".
27346 main_hsub
27348 main_vsub
27349 The main input video's horizontal and vertical chroma subsample
27350 values. For example for the pixel format "yuv422p" hsub is 2 and
27351 vsub is 1.
27352 main_n
27354 The (sequential) number of the main input frame, starting from 0.
27355 Only available with "eval=frame".
27356 main_t
27358 The presentation timestamp of the main input frame, expressed as a
27359 number of seconds. Only available with "eval=frame".
27360 main_pos
27362 The position (byte offset) of the frame in the main input stream,
27363 or NaN if this information is unavailable and/or meaningless (for
27364 example in case of synthetic video). Only available with
27365 "eval=frame".
27366 Examples
27368 • Scale a subtitle stream (b) to match the main video (a) in size
27370 before overlaying
27371 'scale2ref[b][a];[a][b]overlay'
27373 • Scale a logo to 1/10th the height of a video, while preserving its
27375 display aspect ratio.
27376 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
27378 Commands
27380 This filter supports the following commands:
27382 width, w
27384 height, h
27385 Set the output video dimension expression. The command accepts the
27386 same syntax of the corresponding option.
27387 If the specified expression is not valid, it is kept at its current
27389 value.
27390 scale2ref_npp
27392 Use the NVIDIA Performance Primitives (libnpp) to scale (resize) the
27393 input video, based on a reference video.
27394 See the scale_npp filter for available options, scale2ref_npp supports
27396 the same but uses the reference video instead of the main input as
27397 basis. scale2ref_npp also supports the following additional constants
27398 for the w and h options:
27399 main_w
27401 main_h
27402 The main input video's width and height
27403 main_a
27405 The same as main_w / main_h
27406 main_sar
27408 The main input video's sample aspect ratio
27409 main_dar, mdar
27411 The main input video's display aspect ratio. Calculated from
27412 "(main_w / main_h) * main_sar".
27413 main_n
27415 The (sequential) number of the main input frame, starting from 0.
27416 Only available with "eval=frame".
27417 main_t
27419 The presentation timestamp of the main input frame, expressed as a
27420 number of seconds. Only available with "eval=frame".
27421 main_pos
27423 The position (byte offset) of the frame in the main input stream,
27424 or NaN if this information is unavailable and/or meaningless (for
27425 example in case of synthetic video). Only available with
27426 "eval=frame".
27427 Examples
27429 • Scale a subtitle stream (b) to match the main video (a) in size
27431 before overlaying
27432 'scale2ref_npp[b][a];[a][b]overlay_cuda'
27434 • Scale a logo to 1/10th the height of a video, while preserving its
27436 display aspect ratio.
27437 [logo-in][video-in]scale2ref_npp=w=oh*mdar:h=ih/10[logo-out][video-out]
27439 scharr
27441 Apply scharr operator to input video stream.
27442 The filter accepts the following option:
27444 planes
27446 Set which planes will be processed, unprocessed planes will be
27447 copied. By default value 0xf, all planes will be processed.
27448 scale
27450 Set value which will be multiplied with filtered result.
27451 delta
27453 Set value which will be added to filtered result.
27454 Commands
27456 This filter supports the all above options as commands.
27458 scroll
27460 Scroll input video horizontally and/or vertically by constant speed.
27461 The filter accepts the following options:
27463 horizontal, h
27465 Set the horizontal scrolling speed. Default is 0. Allowed range is
27466 from -1 to 1. Negative values changes scrolling direction.
27467 vertical, v
27469 Set the vertical scrolling speed. Default is 0. Allowed range is
27470 from -1 to 1. Negative values changes scrolling direction.
27471 hpos
27473 Set the initial horizontal scrolling position. Default is 0.
27474 Allowed range is from 0 to 1.
27475 vpos
27477 Set the initial vertical scrolling position. Default is 0. Allowed
27478 range is from 0 to 1.
27479 Commands
27481 This filter supports the following commands:
27483 horizontal, h
27485 Set the horizontal scrolling speed.
27486 vertical, v
27488 Set the vertical scrolling speed.
27489 scdet
27491 Detect video scene change.
27492 This filter sets frame metadata with mafd between frame, the scene
27494 score, and forward the frame to the next filter, so they can use these
27495 metadata to detect scene change or others.
27496 In addition, this filter logs a message and sets frame metadata when it
27498 detects a scene change by threshold.
27499 "lavfi.scd.mafd" metadata keys are set with mafd for every frame.
27501 "lavfi.scd.score" metadata keys are set with scene change score for
27503 every frame to detect scene change.
27504 "lavfi.scd.time" metadata keys are set with current filtered frame time
27506 which detect scene change with threshold.
27507 The filter accepts the following options:
27509 threshold, t
27511 Set the scene change detection threshold as a percentage of maximum
27512 change. Good values are in the "[8.0, 14.0]" range. The range for
27513 threshold is "[0., 100.]".
27514 Default value is 10..
27516 sc_pass, s
27518 Set the flag to pass scene change frames to the next filter.
27519 Default value is 0 You can enable it if you want to get snapshot of
27520 scene change frames only.
27521 selectivecolor
27523 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of
27524 colors (such as "reds", "yellows", "greens", "cyans", ...). The
27525 adjustment range is defined by the "purity" of the color (that is, how
27526 saturated it already is).
27527 This filter is similar to the Adobe Photoshop Selective Color tool.
27529 The filter accepts the following options:
27531 correction_method
27533 Select color correction method.
27534 Available values are:
27536 absolute
27538 Specified adjustments are applied "as-is" (added/subtracted to
27539 original pixel component value).
27540 relative
27542 Specified adjustments are relative to the original component
27543 value.
27544 Default is "absolute".
27546 reds
27548 Adjustments for red pixels (pixels where the red component is the
27549 maximum)
27550 yellows
27552 Adjustments for yellow pixels (pixels where the blue component is
27553 the minimum)
27554 greens
27556 Adjustments for green pixels (pixels where the green component is
27557 the maximum)
27558 cyans
27560 Adjustments for cyan pixels (pixels where the red component is the
27561 minimum)
27562 blues
27564 Adjustments for blue pixels (pixels where the blue component is the
27565 maximum)
27566 magentas
27568 Adjustments for magenta pixels (pixels where the green component is
27569 the minimum)
27570 whites
27572 Adjustments for white pixels (pixels where all components are
27573 greater than 128)
27574 neutrals
27576 Adjustments for all pixels except pure black and pure white
27577 blacks
27579 Adjustments for black pixels (pixels where all components are
27580 lesser than 128)
27581 psfile
27583 Specify a Photoshop selective color file (".asv") to import the
27584 settings from.
27585 All the adjustment settings (reds, yellows, ...) accept up to 4 space
27587 separated floating point adjustment values in the [-1,1] range,
27588 respectively to adjust the amount of cyan, magenta, yellow and black
27589 for the pixels of its range.
27590 Examples
27592 • Increase cyan by 50% and reduce yellow by 33% in every green areas,
27594 and increase magenta by 27% in blue areas:
27595 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
27597 • Use a Photoshop selective color preset:
27599 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
27601 separatefields
27603 The "separatefields" takes a frame-based video input and splits each
27604 frame into its components fields, producing a new half height clip with
27605 twice the frame rate and twice the frame count.
27606 This filter use field-dominance information in frame to decide which of
27608 each pair of fields to place first in the output. If it gets it wrong
27609 use setfield filter before "separatefields" filter.
27610 setdar, setsar
27612 The "setdar" filter sets the Display Aspect Ratio for the filter output
27613 video.
27614 This is done by changing the specified Sample (aka Pixel) Aspect Ratio,
27616 according to the following equation:
27617 <DAR> = <HORIZONTAL_RESOLUTION> / <VERTICAL_RESOLUTION> * <SAR>
27619 Keep in mind that the "setdar" filter does not modify the pixel
27621 dimensions of the video frame. Also, the display aspect ratio set by
27622 this filter may be changed by later filters in the filterchain, e.g. in
27623 case of scaling or if another "setdar" or a "setsar" filter is applied.
27624 The "setsar" filter sets the Sample (aka Pixel) Aspect Ratio for the
27626 filter output video.
27627 Note that as a consequence of the application of this filter, the
27629 output display aspect ratio will change according to the equation
27630 above.
27631 Keep in mind that the sample aspect ratio set by the "setsar" filter
27633 may be changed by later filters in the filterchain, e.g. if another
27634 "setsar" or a "setdar" filter is applied.
27635 It accepts the following parameters:
27637 r, ratio, dar ("setdar" only), sar ("setsar" only)
27639 Set the aspect ratio used by the filter.
27640 The parameter can be a floating point number string, an expression,
27642 or a string of the form num:den, where num and den are the
27643 numerator and denominator of the aspect ratio. If the parameter is
27644 not specified, it is assumed the value "0". In case the form
27645 "num:den" is used, the ":" character should be escaped.
27646 max Set the maximum integer value to use for expressing numerator and
27648 denominator when reducing the expressed aspect ratio to a rational.
27649 Default value is 100.
27650 The parameter sar is an expression containing the following constants:
27652 E, PI, PHI
27654 These are approximated values for the mathematical constants e
27655 (Euler's number), pi (Greek pi), and phi (the golden ratio).
27656 w, h
27658 The input width and height.
27659 a These are the same as w / h.
27661 sar The input sample aspect ratio.
27663 dar The input display aspect ratio. It is the same as (w / h) * sar.
27665 hsub, vsub
27667 Horizontal and vertical chroma subsample values. For example, for
27668 the pixel format "yuv422p" hsub is 2 and vsub is 1.
27669 Examples
27671 • To change the display aspect ratio to 16:9, specify one of the
27673 following:
27674 setdar=dar=1.77777
27676 setdar=dar=16/9
27677 • To change the sample aspect ratio to 10:11, specify:
27679 setsar=sar=10/11
27681 • To set a display aspect ratio of 16:9, and specify a maximum
27683 integer value of 1000 in the aspect ratio reduction, use the
27684 command:
27685 setdar=ratio=16/9:max=1000
27687 setfield
27689 Force field for the output video frame.
27690 The "setfield" filter marks the interlace type field for the output
27692 frames. It does not change the input frame, but only sets the
27693 corresponding property, which affects how the frame is treated by
27694 following filters (e.g. "fieldorder" or "yadif").
27695 The filter accepts the following options:
27697 mode
27699 Available values are:
27700 auto
27702 Keep the same field property.
27703 bff Mark the frame as bottom-field-first.
27705 tff Mark the frame as top-field-first.
27707 prog
27709 Mark the frame as progressive.
27710 setparams
27712 Force frame parameter for the output video frame.
27713 The "setparams" filter marks interlace and color range for the output
27715 frames. It does not change the input frame, but only sets the
27716 corresponding property, which affects how the frame is treated by
27717 filters/encoders.
27718 field_mode
27720 Available values are:
27721 auto
27723 Keep the same field property (default).
27724 bff Mark the frame as bottom-field-first.
27726 tff Mark the frame as top-field-first.
27728 prog
27730 Mark the frame as progressive.
27731 range
27733 Available values are:
27734 auto
27736 Keep the same color range property (default).
27737 unspecified, unknown
27739 Mark the frame as unspecified color range.
27740 limited, tv, mpeg
27742 Mark the frame as limited range.
27743 full, pc, jpeg
27745 Mark the frame as full range.
27746 color_primaries
27748 Set the color primaries. Available values are:
27749 auto
27751 Keep the same color primaries property (default).
27752 bt709
27754 unknown
27755 bt470m
27756 bt470bg
27757 smpte170m
27758 smpte240m
27759 film
27760 bt2020
27761 smpte428
27762 smpte431
27763 smpte432
27764 jedec-p22
27765 color_trc
27766 Set the color transfer. Available values are:
27767 auto
27769 Keep the same color trc property (default).
27770 bt709
27772 unknown
27773 bt470m
27774 bt470bg
27775 smpte170m
27776 smpte240m
27777 linear
27778 log100
27779 log316
27780 iec61966-2-4
27781 bt1361e
27782 iec61966-2-1
27783 bt2020-10
27784 bt2020-12
27785 smpte2084
27786 smpte428
27787 arib-std-b67
27788 colorspace
27789 Set the colorspace. Available values are:
27790 auto
27792 Keep the same colorspace property (default).
27793 gbr
27795 bt709
27796 unknown
27797 fcc
27798 bt470bg
27799 smpte170m
27800 smpte240m
27801 ycgco
27802 bt2020nc
27803 bt2020c
27804 smpte2085
27805 chroma-derived-nc
27806 chroma-derived-c
27807 ictcp
27808 sharpen_npp
27810 Use the NVIDIA Performance Primitives (libnpp) to perform image
27811 sharpening with border control.
27812 The following additional options are accepted:
27814 border_type
27816 Type of sampling to be used ad frame borders. One of the following:
27817 replicate
27819 Replicate pixel values.
27820 shear
27822 Apply shear transform to input video.
27823 This filter supports the following options:
27825 shx Shear factor in X-direction. Default value is 0. Allowed range is
27827 from -2 to 2.
27828 shy Shear factor in Y-direction. Default value is 0. Allowed range is
27830 from -2 to 2.
27831 fillcolor, c
27833 Set the color used to fill the output area not covered by the
27834 transformed video. For the general syntax of this option, check the
27835 "Color" section in the ffmpeg-utils manual. If the special value
27836 "none" is selected then no background is printed (useful for
27837 example if the background is never shown).
27838 Default value is "black".
27840 interp
27842 Set interpolation type. Can be "bilinear" or "nearest". Default is
27843 "bilinear".
27844 Commands
27846 This filter supports the all above options as commands.
27848 showinfo
27850 Show a line containing various information for each input video frame.
27851 The input video is not modified.
27852 This filter supports the following options:
27854 checksum
27856 Calculate checksums of each plane. By default enabled.
27857 The shown line contains a sequence of key/value pairs of the form
27859 key:value.
27860 The following values are shown in the output:
27862 n The (sequential) number of the input frame, starting from 0.
27864 pts The Presentation TimeStamp of the input frame, expressed as a
27866 number of time base units. The time base unit depends on the filter
27867 input pad.
27868 pts_time
27870 The Presentation TimeStamp of the input frame, expressed as a
27871 number of seconds.
27872 pos The position of the frame in the input stream, or -1 if this
27874 information is unavailable and/or meaningless (for example in case
27875 of synthetic video).
27876 fmt The pixel format name.
27878 sar The sample aspect ratio of the input frame, expressed in the form
27880 num/den.
27881 s The size of the input frame. For the syntax of this option, check
27883 the "Video size" section in the ffmpeg-utils manual.
27884 i The type of interlaced mode ("P" for "progressive", "T" for top
27886 field first, "B" for bottom field first).
27887 iskey
27889 This is 1 if the frame is a key frame, 0 otherwise.
27890 type
27892 The picture type of the input frame ("I" for an I-frame, "P" for a
27893 P-frame, "B" for a B-frame, or "?" for an unknown type). Also
27894 refer to the documentation of the "AVPictureType" enum and of the
27895 "av_get_picture_type_char" function defined in libavutil/avutil.h.
27896 checksum
27898 The Adler-32 checksum (printed in hexadecimal) of all the planes of
27899 the input frame.
27900 plane_checksum
27902 The Adler-32 checksum (printed in hexadecimal) of each plane of the
27903 input frame, expressed in the form "[c0 c1 c2 c3]".
27904 mean
27906 The mean value of pixels in each plane of the input frame,
27907 expressed in the form "[mean0 mean1 mean2 mean3]".
27908 stdev
27910 The standard deviation of pixel values in each plane of the input
27911 frame, expressed in the form "[stdev0 stdev1 stdev2 stdev3]".
27912 showpalette
27914 Displays the 256 colors palette of each frame. This filter is only
27915 relevant for pal8 pixel format frames.
27916 It accepts the following option:
27918 s Set the size of the box used to represent one palette color entry.
27920 Default is 30 (for a "30x30" pixel box).
27921 shuffleframes
27923 Reorder and/or duplicate and/or drop video frames.
27924 It accepts the following parameters:
27926 mapping
27928 Set the destination indexes of input frames. This is space or '|'
27929 separated list of indexes that maps input frames to output frames.
27930 Number of indexes also sets maximal value that each index may have.
27931 '-1' index have special meaning and that is to drop frame.
27932 The first frame has the index 0. The default is to keep the input
27934 unchanged.
27935 Examples
27937 • Swap second and third frame of every three frames of the input:
27939 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
27941 • Swap 10th and 1st frame of every ten frames of the input:
27943 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
27945 shufflepixels
27947 Reorder pixels in video frames.
27948 This filter accepts the following options:
27950 direction, d
27952 Set shuffle direction. Can be forward or inverse direction.
27953 Default direction is forward.
27954 mode, m
27956 Set shuffle mode. Can be horizontal, vertical or block mode.
27957 width, w
27959 height, h
27960 Set shuffle block_size. In case of horizontal shuffle mode only
27961 width part of size is used, and in case of vertical shuffle mode
27962 only height part of size is used.
27963 seed, s
27965 Set random seed used with shuffling pixels. Mainly useful to set to
27966 be able to reverse filtering process to get original input. For
27967 example, to reverse forward shuffle you need to use same parameters
27968 and exact same seed and to set direction to inverse.
27969 shuffleplanes
27971 Reorder and/or duplicate video planes.
27972 It accepts the following parameters:
27974 map0
27976 The index of the input plane to be used as the first output plane.
27977 map1
27979 The index of the input plane to be used as the second output plane.
27980 map2
27982 The index of the input plane to be used as the third output plane.
27983 map3
27985 The index of the input plane to be used as the fourth output plane.
27986 The first plane has the index 0. The default is to keep the input
27988 unchanged.
27989 Examples
27991 • Swap the second and third planes of the input:
27993 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
27995 signalstats
27997 Evaluate various visual metrics that assist in determining issues
27998 associated with the digitization of analog video media.
27999 By default the filter will log these metadata values:
28001 YMIN
28003 Display the minimal Y value contained within the input frame.
28004 Expressed in range of [0-255].
28005 YLOW
28007 Display the Y value at the 10% percentile within the input frame.
28008 Expressed in range of [0-255].
28009 YAVG
28011 Display the average Y value within the input frame. Expressed in
28012 range of [0-255].
28013 YHIGH
28015 Display the Y value at the 90% percentile within the input frame.
28016 Expressed in range of [0-255].
28017 YMAX
28019 Display the maximum Y value contained within the input frame.
28020 Expressed in range of [0-255].
28021 UMIN
28023 Display the minimal U value contained within the input frame.
28024 Expressed in range of [0-255].
28025 ULOW
28027 Display the U value at the 10% percentile within the input frame.
28028 Expressed in range of [0-255].
28029 UAVG
28031 Display the average U value within the input frame. Expressed in
28032 range of [0-255].
28033 UHIGH
28035 Display the U value at the 90% percentile within the input frame.
28036 Expressed in range of [0-255].
28037 UMAX
28039 Display the maximum U value contained within the input frame.
28040 Expressed in range of [0-255].
28041 VMIN
28043 Display the minimal V value contained within the input frame.
28044 Expressed in range of [0-255].
28045 VLOW
28047 Display the V value at the 10% percentile within the input frame.
28048 Expressed in range of [0-255].
28049 VAVG
28051 Display the average V value within the input frame. Expressed in
28052 range of [0-255].
28053 VHIGH
28055 Display the V value at the 90% percentile within the input frame.
28056 Expressed in range of [0-255].
28057 VMAX
28059 Display the maximum V value contained within the input frame.
28060 Expressed in range of [0-255].
28061 SATMIN
28063 Display the minimal saturation value contained within the input
28064 frame. Expressed in range of [0-~181.02].
28065 SATLOW
28067 Display the saturation value at the 10% percentile within the input
28068 frame. Expressed in range of [0-~181.02].
28069 SATAVG
28071 Display the average saturation value within the input frame.
28072 Expressed in range of [0-~181.02].
28073 SATHIGH
28075 Display the saturation value at the 90% percentile within the input
28076 frame. Expressed in range of [0-~181.02].
28077 SATMAX
28079 Display the maximum saturation value contained within the input
28080 frame. Expressed in range of [0-~181.02].
28081 HUEMED
28083 Display the median value for hue within the input frame. Expressed
28084 in range of [0-360].
28085 HUEAVG
28087 Display the average value for hue within the input frame. Expressed
28088 in range of [0-360].
28089 YDIF
28091 Display the average of sample value difference between all values
28092 of the Y plane in the current frame and corresponding values of the
28093 previous input frame. Expressed in range of [0-255].
28094 UDIF
28096 Display the average of sample value difference between all values
28097 of the U plane in the current frame and corresponding values of the
28098 previous input frame. Expressed in range of [0-255].
28099 VDIF
28101 Display the average of sample value difference between all values
28102 of the V plane in the current frame and corresponding values of the
28103 previous input frame. Expressed in range of [0-255].
28104 YBITDEPTH
28106 Display bit depth of Y plane in current frame. Expressed in range
28107 of [0-16].
28108 UBITDEPTH
28110 Display bit depth of U plane in current frame. Expressed in range
28111 of [0-16].
28112 VBITDEPTH
28114 Display bit depth of V plane in current frame. Expressed in range
28115 of [0-16].
28116 The filter accepts the following options:
28118 stat
28120 out stat specify an additional form of image analysis. out output
28121 video with the specified type of pixel highlighted.
28122 Both options accept the following values:
28124 tout
28126 Identify temporal outliers pixels. A temporal outlier is a
28127 pixel unlike the neighboring pixels of the same field. Examples
28128 of temporal outliers include the results of video dropouts,
28129 head clogs, or tape tracking issues.
28130 vrep
28132 Identify vertical line repetition. Vertical line repetition
28133 includes similar rows of pixels within a frame. In born-digital
28134 video vertical line repetition is common, but this pattern is
28135 uncommon in video digitized from an analog source. When it
28136 occurs in video that results from the digitization of an analog
28137 source it can indicate concealment from a dropout compensator.
28138 brng
28140 Identify pixels that fall outside of legal broadcast range.
28141 color, c
28143 Set the highlight color for the out option. The default color is
28144 yellow.
28145 Examples
28147 • Output data of various video metrics:
28149 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
28151 • Output specific data about the minimum and maximum values of the Y
28153 plane per frame:
28154 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
28156 • Playback video while highlighting pixels that are outside of
28158 broadcast range in red.
28159 ffplay example.mov -vf signalstats="out=brng:color=red"
28161 • Playback video with signalstats metadata drawn over the frame.
28163 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
28165 The contents of signalstat_drawtext.txt used in the command are:
28167 time %{pts:hms}
28169 Y (%{metadata:lavfi.signalstats.YMIN}-%{metadata:lavfi.signalstats.YMAX})
28170 U (%{metadata:lavfi.signalstats.UMIN}-%{metadata:lavfi.signalstats.UMAX})
28171 V (%{metadata:lavfi.signalstats.VMIN}-%{metadata:lavfi.signalstats.VMAX})
28172 saturation maximum: %{metadata:lavfi.signalstats.SATMAX}
28173 signature
28175 Calculates the MPEG-7 Video Signature. The filter can handle more than
28176 one input. In this case the matching between the inputs can be
28177 calculated additionally. The filter always passes through the first
28178 input. The signature of each stream can be written into a file.
28179 It accepts the following options:
28181 detectmode
28183 Enable or disable the matching process.
28184 Available values are:
28186 off Disable the calculation of a matching (default).
28188 full
28190 Calculate the matching for the whole video and output whether
28191 the whole video matches or only parts.
28192 fast
28194 Calculate only until a matching is found or the video ends.
28195 Should be faster in some cases.
28196 nb_inputs
28198 Set the number of inputs. The option value must be a non negative
28199 integer. Default value is 1.
28200 filename
28202 Set the path to which the output is written. If there is more than
28203 one input, the path must be a prototype, i.e. must contain %d or
28204 %0nd (where n is a positive integer), that will be replaced with
28205 the input number. If no filename is specified, no output will be
28206 written. This is the default.
28207 format
28209 Choose the output format.
28210 Available values are:
28212 binary
28214 Use the specified binary representation (default).
28215 xml Use the specified xml representation.
28217 th_d
28219 Set threshold to detect one word as similar. The option value must
28220 be an integer greater than zero. The default value is 9000.
28221 th_dc
28223 Set threshold to detect all words as similar. The option value must
28224 be an integer greater than zero. The default value is 60000.
28225 th_xh
28227 Set threshold to detect frames as similar. The option value must be
28228 an integer greater than zero. The default value is 116.
28229 th_di
28231 Set the minimum length of a sequence in frames to recognize it as
28232 matching sequence. The option value must be a non negative integer
28233 value. The default value is 0.
28234 th_it
28236 Set the minimum relation, that matching frames to all frames must
28237 have. The option value must be a double value between 0 and 1. The
28238 default value is 0.5.
28239 Examples
28241 • To calculate the signature of an input video and store it in
28243 signature.bin:
28244 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
28246 • To detect whether two videos match and store the signatures in XML
28248 format in signature0.xml and signature1.xml:
28249 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 -
28251 siti
28253 Calculate Spatial Info (SI) and Temporal Info (TI) scores for a video,
28254 as defined in ITU-T P.910: Subjective video quality assessment methods
28255 for multimedia applications. Available PDF at
28256 <https://www.itu.int/rec/T-REC-P.910-199909-S/en >.
28257 It accepts the following option:
28259 print_summary
28261 If set to 1, Summary statistics will be printed to the console.
28262 Default 0.
28263 Examples
28265 • To calculate SI/TI metrics and print summary:
28267 ffmpeg -i input.mp4 -vf siti=print_summary=1 -f null -
28269 smartblur
28271 Blur the input video without impacting the outlines.
28272 It accepts the following options:
28274 luma_radius, lr
28276 Set the luma radius. The option value must be a float number in the
28277 range [0.1,5.0] that specifies the variance of the gaussian filter
28278 used to blur the image (slower if larger). Default value is 1.0.
28279 luma_strength, ls
28281 Set the luma strength. The option value must be a float number in
28282 the range [-1.0,1.0] that configures the blurring. A value included
28283 in [0.0,1.0] will blur the image whereas a value included in
28284 [-1.0,0.0] will sharpen the image. Default value is 1.0.
28285 luma_threshold, lt
28287 Set the luma threshold used as a coefficient to determine whether a
28288 pixel should be blurred or not. The option value must be an integer
28289 in the range [-30,30]. A value of 0 will filter all the image, a
28290 value included in [0,30] will filter flat areas and a value
28291 included in [-30,0] will filter edges. Default value is 0.
28292 chroma_radius, cr
28294 Set the chroma radius. The option value must be a float number in
28295 the range [0.1,5.0] that specifies the variance of the gaussian
28296 filter used to blur the image (slower if larger). Default value is
28297 luma_radius.
28298 chroma_strength, cs
28300 Set the chroma strength. The option value must be a float number in
28301 the range [-1.0,1.0] that configures the blurring. A value included
28302 in [0.0,1.0] will blur the image whereas a value included in
28303 [-1.0,0.0] will sharpen the image. Default value is luma_strength.
28304 chroma_threshold, ct
28306 Set the chroma threshold used as a coefficient to determine whether
28307 a pixel should be blurred or not. The option value must be an
28308 integer in the range [-30,30]. A value of 0 will filter all the
28309 image, a value included in [0,30] will filter flat areas and a
28310 value included in [-30,0] will filter edges. Default value is
28311 luma_threshold.
28312 If a chroma option is not explicitly set, the corresponding luma value
28314 is set.
28315 sobel
28317 Apply sobel operator to input video stream.
28318 The filter accepts the following option:
28320 planes
28322 Set which planes will be processed, unprocessed planes will be
28323 copied. By default value 0xf, all planes will be processed.
28324 scale
28326 Set value which will be multiplied with filtered result.
28327 delta
28329 Set value which will be added to filtered result.
28330 Commands
28332 This filter supports the all above options as commands.
28334 spp
28336 Apply a simple postprocessing filter that compresses and decompresses
28337 the image at several (or - in the case of quality level 6 - all) shifts
28338 and average the results.
28339 The filter accepts the following options:
28341 quality
28343 Set quality. This option defines the number of levels for
28344 averaging. It accepts an integer in the range 0-6. If set to 0, the
28345 filter will have no effect. A value of 6 means the higher quality.
28346 For each increment of that value the speed drops by a factor of
28347 approximately 2. Default value is 3.
28348 qp Force a constant quantization parameter. If not set, the filter
28350 will use the QP from the video stream (if available).
28351 mode
28353 Set thresholding mode. Available modes are:
28354 hard
28356 Set hard thresholding (default).
28357 soft
28359 Set soft thresholding (better de-ringing effect, but likely
28360 blurrier).
28361 use_bframe_qp
28363 Enable the use of the QP from the B-Frames if set to 1. Using this
28364 option may cause flicker since the B-Frames have often larger QP.
28365 Default is 0 (not enabled).
28366 Commands
28368 This filter supports the following commands:
28370 quality, level
28372 Set quality level. The value "max" can be used to set the maximum
28373 level, currently 6.
28374 sr
28376 Scale the input by applying one of the super-resolution methods based
28377 on convolutional neural networks. Supported models:
28378 • Super-Resolution Convolutional Neural Network model (SRCNN). See
28380 <https://arxiv.org/abs/1501.00092>.
28381 • Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
28383 See <https://arxiv.org/abs/1609.05158>.
28384 Training scripts as well as scripts for model file (.pb) saving can be
28386 found at <https://github.com/XueweiMeng/sr/tree/sr_dnn_native>.
28387 Original repository is at
28388 <https://github.com/HighVoltageRocknRoll/sr.git>.
28389 Native model files (.model) can be generated from TensorFlow model
28391 files (.pb) by using tools/python/convert.py
28392 The filter accepts the following options:
28394 dnn_backend
28396 Specify which DNN backend to use for model loading and execution.
28397 This option accepts the following values:
28398 native
28400 Native implementation of DNN loading and execution.
28401 tensorflow
28403 TensorFlow backend. To enable this backend you need to install
28404 the TensorFlow for C library (see
28405 <https://www.tensorflow.org/install/lang_c>) and configure
28406 FFmpeg with "--enable-libtensorflow"
28407 Default value is native.
28409 model
28411 Set path to model file specifying network architecture and its
28412 parameters. Note that different backends use different file
28413 formats. TensorFlow backend can load files for both formats, while
28414 native backend can load files for only its format.
28415 scale_factor
28417 Set scale factor for SRCNN model. Allowed values are 2, 3 and 4.
28418 Default value is 2. Scale factor is necessary for SRCNN model,
28419 because it accepts input upscaled using bicubic upscaling with
28420 proper scale factor.
28421 To get full functionality (such as async execution), please use the
28423 dnn_processing filter.
28424 ssim
28426 Obtain the SSIM (Structural SImilarity Metric) between two input
28427 videos.
28428 This filter takes in input two input videos, the first input is
28430 considered the "main" source and is passed unchanged to the output. The
28431 second input is used as a "reference" video for computing the SSIM.
28432 Both video inputs must have the same resolution and pixel format for
28434 this filter to work correctly. Also it assumes that both inputs have
28435 the same number of frames, which are compared one by one.
28436 The filter stores the calculated SSIM of each frame.
28438 The description of the accepted parameters follows.
28440 stats_file, f
28442 If specified the filter will use the named file to save the SSIM of
28443 each individual frame. When filename equals "-" the data is sent to
28444 standard output.
28445 The file printed if stats_file is selected, contains a sequence of
28447 key/value pairs of the form key:value for each compared couple of
28448 frames.
28449 A description of each shown parameter follows:
28451 n sequential number of the input frame, starting from 1
28453 Y, U, V, R, G, B
28455 SSIM of the compared frames for the component specified by the
28456 suffix.
28457 All SSIM of the compared frames for the whole frame.
28459 dB Same as above but in dB representation.
28461 This filter also supports the framesync options.
28463 Examples
28465 • For example:
28467 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
28469 [main][ref] ssim="stats_file=stats.log" [out]
28470 On this example the input file being processed is compared with the
28472 reference file ref_movie.mpg. The SSIM of each individual frame is
28473 stored in stats.log.
28474 • Another example with both psnr and ssim at same time:
28476 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
28478 • Another example with different containers:
28480 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 -
28482 stereo3d
28484 Convert between different stereoscopic image formats.
28485 The filters accept the following options:
28487 in Set stereoscopic image format of input.
28489 Available values for input image formats are:
28491 sbsl
28493 side by side parallel (left eye left, right eye right)
28494 sbsr
28496 side by side crosseye (right eye left, left eye right)
28497 sbs2l
28499 side by side parallel with half width resolution (left eye
28500 left, right eye right)
28501 sbs2r
28503 side by side crosseye with half width resolution (right eye
28504 left, left eye right)
28505 abl
28507 tbl above-below (left eye above, right eye below)
28508 abr
28510 tbr above-below (right eye above, left eye below)
28511 ab2l
28513 tb2l
28514 above-below with half height resolution (left eye above, right
28515 eye below)
28516 ab2r
28518 tb2r
28519 above-below with half height resolution (right eye above, left
28520 eye below)
28521 al alternating frames (left eye first, right eye second)
28523 ar alternating frames (right eye first, left eye second)
28525 irl interleaved rows (left eye has top row, right eye starts on
28527 next row)
28528 irr interleaved rows (right eye has top row, left eye starts on
28530 next row)
28531 icl interleaved columns, left eye first
28533 icr interleaved columns, right eye first
28535 Default value is sbsl.
28537 out Set stereoscopic image format of output.
28539 sbsl
28541 side by side parallel (left eye left, right eye right)
28542 sbsr
28544 side by side crosseye (right eye left, left eye right)
28545 sbs2l
28547 side by side parallel with half width resolution (left eye
28548 left, right eye right)
28549 sbs2r
28551 side by side crosseye with half width resolution (right eye
28552 left, left eye right)
28553 abl
28555 tbl above-below (left eye above, right eye below)
28556 abr
28558 tbr above-below (right eye above, left eye below)
28559 ab2l
28561 tb2l
28562 above-below with half height resolution (left eye above, right
28563 eye below)
28564 ab2r
28566 tb2r
28567 above-below with half height resolution (right eye above, left
28568 eye below)
28569 al alternating frames (left eye first, right eye second)
28571 ar alternating frames (right eye first, left eye second)
28573 irl interleaved rows (left eye has top row, right eye starts on
28575 next row)
28576 irr interleaved rows (right eye has top row, left eye starts on
28578 next row)
28579 arbg
28581 anaglyph red/blue gray (red filter on left eye, blue filter on
28582 right eye)
28583 argg
28585 anaglyph red/green gray (red filter on left eye, green filter
28586 on right eye)
28587 arcg
28589 anaglyph red/cyan gray (red filter on left eye, cyan filter on
28590 right eye)
28591 arch
28593 anaglyph red/cyan half colored (red filter on left eye, cyan
28594 filter on right eye)
28595 arcc
28597 anaglyph red/cyan color (red filter on left eye, cyan filter on
28598 right eye)
28599 arcd
28601 anaglyph red/cyan color optimized with the least squares
28602 projection of dubois (red filter on left eye, cyan filter on
28603 right eye)
28604 agmg
28606 anaglyph green/magenta gray (green filter on left eye, magenta
28607 filter on right eye)
28608 agmh
28610 anaglyph green/magenta half colored (green filter on left eye,
28611 magenta filter on right eye)
28612 agmc
28614 anaglyph green/magenta colored (green filter on left eye,
28615 magenta filter on right eye)
28616 agmd
28618 anaglyph green/magenta color optimized with the least squares
28619 projection of dubois (green filter on left eye, magenta filter
28620 on right eye)
28621 aybg
28623 anaglyph yellow/blue gray (yellow filter on left eye, blue
28624 filter on right eye)
28625 aybh
28627 anaglyph yellow/blue half colored (yellow filter on left eye,
28628 blue filter on right eye)
28629 aybc
28631 anaglyph yellow/blue colored (yellow filter on left eye, blue
28632 filter on right eye)
28633 aybd
28635 anaglyph yellow/blue color optimized with the least squares
28636 projection of dubois (yellow filter on left eye, blue filter on
28637 right eye)
28638 ml mono output (left eye only)
28640 mr mono output (right eye only)
28642 chl checkerboard, left eye first
28644 chr checkerboard, right eye first
28646 icl interleaved columns, left eye first
28648 icr interleaved columns, right eye first
28650 hdmi
28652 HDMI frame pack
28653 Default value is arcd.
28655 Examples
28657 • Convert input video from side by side parallel to anaglyph
28659 yellow/blue dubois:
28660 stereo3d=sbsl:aybd
28662 • Convert input video from above below (left eye above, right eye
28664 below) to side by side crosseye.
28665 stereo3d=abl:sbsr
28667 streamselect, astreamselect
28669 Select video or audio streams.
28670 The filter accepts the following options:
28672 inputs
28674 Set number of inputs. Default is 2.
28675 map Set input indexes to remap to outputs.
28677 Commands
28679 The "streamselect" and "astreamselect" filter supports the following
28681 commands:
28682 map Set input indexes to remap to outputs.
28684 Examples
28686 • Select first 5 seconds 1st stream and rest of time 2nd stream:
28688 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
28690 • Same as above, but for audio:
28692 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
28694 subtitles
28696 Draw subtitles on top of input video using the libass library.
28697 To enable compilation of this filter you need to configure FFmpeg with
28699 "--enable-libass". This filter also requires a build with libavcodec
28700 and libavformat to convert the passed subtitles file to ASS (Advanced
28701 Substation Alpha) subtitles format.
28702 The filter accepts the following options:
28704 filename, f
28706 Set the filename of the subtitle file to read. It must be
28707 specified.
28708 original_size
28710 Specify the size of the original video, the video for which the ASS
28711 file was composed. For the syntax of this option, check the "Video
28712 size" section in the ffmpeg-utils manual. Due to a misdesign in
28713 ASS aspect ratio arithmetic, this is necessary to correctly scale
28714 the fonts if the aspect ratio has been changed.
28715 fontsdir
28717 Set a directory path containing fonts that can be used by the
28718 filter. These fonts will be used in addition to whatever the font
28719 provider uses.
28720 alpha
28722 Process alpha channel, by default alpha channel is untouched.
28723 charenc
28725 Set subtitles input character encoding. "subtitles" filter only.
28726 Only useful if not UTF-8.
28727 stream_index, si
28729 Set subtitles stream index. "subtitles" filter only.
28730 force_style
28732 Override default style or script info parameters of the subtitles.
28733 It accepts a string containing ASS style format "KEY=VALUE" couples
28734 separated by ",".
28735 If the first key is not specified, it is assumed that the first value
28737 specifies the filename.
28738 For example, to render the file sub.srt on top of the input video, use
28740 the command:
28741 subtitles=sub.srt
28743 which is equivalent to:
28745 subtitles=filename=sub.srt
28747 To render the default subtitles stream from file video.mkv, use:
28749 subtitles=video.mkv
28751 To render the second subtitles stream from that file, use:
28753 subtitles=video.mkv:si=1
28755 To make the subtitles stream from sub.srt appear in 80% transparent
28757 blue "DejaVu Serif", use:
28758 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
28760 super2xsai
28762 Scale the input by 2x and smooth using the Super2xSaI (Scale and
28763 Interpolate) pixel art scaling algorithm.
28764 Useful for enlarging pixel art images without reducing sharpness.
28766 swaprect
28768 Swap two rectangular objects in video.
28769 This filter accepts the following options:
28771 w Set object width.
28773 h Set object height.
28775 x1 Set 1st rect x coordinate.
28777 y1 Set 1st rect y coordinate.
28779 x2 Set 2nd rect x coordinate.
28781 y2 Set 2nd rect y coordinate.
28783 All expressions are evaluated once for each frame.
28785 The all options are expressions containing the following constants:
28787 w
28789 h The input width and height.
28790 a same as w / h
28792 sar input sample aspect ratio
28794 dar input display aspect ratio, it is the same as (w / h) * sar
28796 n The number of the input frame, starting from 0.
28798 t The timestamp expressed in seconds. It's NAN if the input timestamp
28800 is unknown.
28801 pos the position in the file of the input frame, NAN if unknown
28803 Commands
28805 This filter supports the all above options as commands.
28807 swapuv
28809 Swap U & V plane.
28810 tblend
28812 Blend successive video frames.
28813 See blend
28815 telecine
28817 Apply telecine process to the video.
28818 This filter accepts the following options:
28820 first_field
28822 top, t
28823 top field first
28824 bottom, b
28826 bottom field first The default value is "top".
28827 pattern
28829 A string of numbers representing the pulldown pattern you wish to
28830 apply. The default value is 23.
28831 Some typical patterns:
28833 NTSC output (30i):
28835 27.5p: 32222
28836 24p: 23 (classic)
28837 24p: 2332 (preferred)
28838 20p: 33
28839 18p: 334
28840 16p: 3444
28841 PAL output (25i):
28843 27.5p: 12222
28844 24p: 222222222223 ("Euro pulldown")
28845 16.67p: 33
28846 16p: 33333334
28847 thistogram
28849 Compute and draw a color distribution histogram for the input video
28850 across time.
28851 Unlike histogram video filter which only shows histogram of single
28853 input frame at certain time, this filter shows also past histograms of
28854 number of frames defined by "width" option.
28855 The computed histogram is a representation of the color component
28857 distribution in an image.
28858 The filter accepts the following options:
28860 width, w
28862 Set width of single color component output. Default value is 0.
28863 Value of 0 means width will be picked from input video. This also
28864 set number of passed histograms to keep. Allowed range is [0,
28865 8192].
28866 display_mode, d
28868 Set display mode. It accepts the following values:
28869 stack
28871 Per color component graphs are placed below each other.
28872 parade
28874 Per color component graphs are placed side by side.
28875 overlay
28877 Presents information identical to that in the "parade", except
28878 that the graphs representing color components are superimposed
28879 directly over one another.
28880 Default is "stack".
28882 levels_mode, m
28884 Set mode. Can be either "linear", or "logarithmic". Default is
28885 "linear".
28886 components, c
28888 Set what color components to display. Default is 7.
28889 bgopacity, b
28891 Set background opacity. Default is 0.9.
28892 envelope, e
28894 Show envelope. Default is disabled.
28895 ecolor, ec
28897 Set envelope color. Default is "gold".
28898 slide
28900 Set slide mode.
28901 Available values for slide is:
28903 frame
28905 Draw new frame when right border is reached.
28906 replace
28908 Replace old columns with new ones.
28909 scroll
28911 Scroll from right to left.
28912 rscroll
28914 Scroll from left to right.
28915 picture
28917 Draw single picture.
28918 Default is "replace".
28920 threshold
28922 Apply threshold effect to video stream.
28923 This filter needs four video streams to perform thresholding. First
28925 stream is stream we are filtering. Second stream is holding threshold
28926 values, third stream is holding min values, and last, fourth stream is
28927 holding max values.
28928 The filter accepts the following option:
28930 planes
28932 Set which planes will be processed, unprocessed planes will be
28933 copied. By default value 0xf, all planes will be processed.
28934 For example if first stream pixel's component value is less then
28936 threshold value of pixel component from 2nd threshold stream, third
28937 stream value will picked, otherwise fourth stream pixel component value
28938 will be picked.
28939 Using color source filter one can perform various types of
28941 thresholding:
28942 Commands
28944 This filter supports the all options as commands.
28946 Examples
28948 • Binary threshold, using gray color as threshold:
28950 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
28952 • Inverted binary threshold, using gray color as threshold:
28954 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
28956 • Truncate binary threshold, using gray color as threshold:
28958 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
28960 • Threshold to zero, using gray color as threshold:
28962 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
28964 • Inverted threshold to zero, using gray color as threshold:
28966 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
28968 thumbnail
28970 Select the most representative frame in a given sequence of consecutive
28971 frames.
28972 The filter accepts the following options:
28974 n Set the frames batch size to analyze; in a set of n frames, the
28976 filter will pick one of them, and then handle the next batch of n
28977 frames until the end. Default is 100.
28978 log Set the log level to display picked frame stats. Default is
28980 "info".
28981 Since the filter keeps track of the whole frames sequence, a bigger n
28983 value will result in a higher memory usage, so a high value is not
28984 recommended.
28985 Examples
28987 • Extract one picture each 50 frames:
28989 thumbnail=50
28991 • Complete example of a thumbnail creation with ffmpeg:
28993 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
28995 tile
28997 Tile several successive frames together.
28998 The untile filter can do the reverse.
29000 The filter accepts the following options:
29002 layout
29004 Set the grid size in the form "COLUMNSxROWS". Range is upto
29005 UINT_MAX cells. Default is "6x5".
29006 nb_frames
29008 Set the maximum number of frames to render in the given area. It
29009 must be less than or equal to wxh. The default value is 0, meaning
29010 all the area will be used.
29011 margin
29013 Set the outer border margin in pixels. Range is 0 to 1024. Default
29014 is 0.
29015 padding
29017 Set the inner border thickness (i.e. the number of pixels between
29018 frames). For more advanced padding options (such as having
29019 different values for the edges), refer to the pad video filter.
29020 Range is 0 to 1024. Default is 0.
29021 color
29023 Specify the color of the unused area. For the syntax of this
29024 option, check the "Color" section in the ffmpeg-utils manual. The
29025 default value of color is "black".
29026 overlap
29028 Set the number of frames to overlap when tiling several successive
29029 frames together. The value must be between 0 and nb_frames - 1.
29030 Default is 0.
29031 init_padding
29033 Set the number of frames to initially be empty before displaying
29034 first output frame. This controls how soon will one get first
29035 output frame. The value must be between 0 and nb_frames - 1.
29036 Default is 0.
29037 Examples
29039 • Produce 8x8 PNG tiles of all keyframes (-skip_frame nokey) in a
29041 movie:
29042 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
29044 The -vsync 0 is necessary to prevent ffmpeg from duplicating each
29046 output frame to accommodate the originally detected frame rate.
29047 • Display 5 pictures in an area of "3x2" frames, with 7 pixels
29049 between them, and 2 pixels of initial margin, using mixed flat and
29050 named options:
29051 tile=3x2:nb_frames=5:padding=7:margin=2
29053 tinterlace
29055 Perform various types of temporal field interlacing.
29056 Frames are counted starting from 1, so the first input frame is
29058 considered odd.
29059 The filter accepts the following options:
29061 mode
29063 Specify the mode of the interlacing. This option can also be
29064 specified as a value alone. See below for a list of values for this
29065 option.
29066 Available values are:
29068 merge, 0
29070 Move odd frames into the upper field, even into the lower
29071 field, generating a double height frame at half frame rate.
29072 ------> time
29074 Input:
29075 Frame 1 Frame 2 Frame 3 Frame 4
29076 11111 22222 33333 44444
29078 11111 22222 33333 44444
29079 11111 22222 33333 44444
29080 11111 22222 33333 44444
29081 Output:
29083 11111 33333
29084 22222 44444
29085 11111 33333
29086 22222 44444
29087 11111 33333
29088 22222 44444
29089 11111 33333
29090 22222 44444
29091 drop_even, 1
29093 Only output odd frames, even frames are dropped, generating a
29094 frame with unchanged height at half frame rate.
29095 ------> time
29097 Input:
29098 Frame 1 Frame 2 Frame 3 Frame 4
29099 11111 22222 33333 44444
29101 11111 22222 33333 44444
29102 11111 22222 33333 44444
29103 11111 22222 33333 44444
29104 Output:
29106 11111 33333
29107 11111 33333
29108 11111 33333
29109 11111 33333
29110 drop_odd, 2
29112 Only output even frames, odd frames are dropped, generating a
29113 frame with unchanged height at half frame rate.
29114 ------> time
29116 Input:
29117 Frame 1 Frame 2 Frame 3 Frame 4
29118 11111 22222 33333 44444
29120 11111 22222 33333 44444
29121 11111 22222 33333 44444
29122 11111 22222 33333 44444
29123 Output:
29125 22222 44444
29126 22222 44444
29127 22222 44444
29128 22222 44444
29129 pad, 3
29131 Expand each frame to full height, but pad alternate lines with
29132 black, generating a frame with double height at the same input
29133 frame rate.
29134 ------> time
29136 Input:
29137 Frame 1 Frame 2 Frame 3 Frame 4
29138 11111 22222 33333 44444
29140 11111 22222 33333 44444
29141 11111 22222 33333 44444
29142 11111 22222 33333 44444
29143 Output:
29145 11111 ..... 33333 .....
29146 ..... 22222 ..... 44444
29147 11111 ..... 33333 .....
29148 ..... 22222 ..... 44444
29149 11111 ..... 33333 .....
29150 ..... 22222 ..... 44444
29151 11111 ..... 33333 .....
29152 ..... 22222 ..... 44444
29153 interleave_top, 4
29155 Interleave the upper field from odd frames with the lower field
29156 from even frames, generating a frame with unchanged height at
29157 half frame rate.
29158 ------> time
29160 Input:
29161 Frame 1 Frame 2 Frame 3 Frame 4
29162 11111<- 22222 33333<- 44444
29164 11111 22222<- 33333 44444<-
29165 11111<- 22222 33333<- 44444
29166 11111 22222<- 33333 44444<-
29167 Output:
29169 11111 33333
29170 22222 44444
29171 11111 33333
29172 22222 44444
29173 interleave_bottom, 5
29175 Interleave the lower field from odd frames with the upper field
29176 from even frames, generating a frame with unchanged height at
29177 half frame rate.
29178 ------> time
29180 Input:
29181 Frame 1 Frame 2 Frame 3 Frame 4
29182 11111 22222<- 33333 44444<-
29184 11111<- 22222 33333<- 44444
29185 11111 22222<- 33333 44444<-
29186 11111<- 22222 33333<- 44444
29187 Output:
29189 22222 44444
29190 11111 33333
29191 22222 44444
29192 11111 33333
29193 interlacex2, 6
29195 Double frame rate with unchanged height. Frames are inserted
29196 each containing the second temporal field from the previous
29197 input frame and the first temporal field from the next input
29198 frame. This mode relies on the top_field_first flag. Useful for
29199 interlaced video displays with no field synchronisation.
29200 ------> time
29202 Input:
29203 Frame 1 Frame 2 Frame 3 Frame 4
29204 11111 22222 33333 44444
29206 11111 22222 33333 44444
29207 11111 22222 33333 44444
29208 11111 22222 33333 44444
29209 Output:
29211 11111 22222 22222 33333 33333 44444 44444
29212 11111 11111 22222 22222 33333 33333 44444
29213 11111 22222 22222 33333 33333 44444 44444
29214 11111 11111 22222 22222 33333 33333 44444
29215 mergex2, 7
29217 Move odd frames into the upper field, even into the lower
29218 field, generating a double height frame at same frame rate.
29219 ------> time
29221 Input:
29222 Frame 1 Frame 2 Frame 3 Frame 4
29223 11111 22222 33333 44444
29225 11111 22222 33333 44444
29226 11111 22222 33333 44444
29227 11111 22222 33333 44444
29228 Output:
29230 11111 33333 33333 55555
29231 22222 22222 44444 44444
29232 11111 33333 33333 55555
29233 22222 22222 44444 44444
29234 11111 33333 33333 55555
29235 22222 22222 44444 44444
29236 11111 33333 33333 55555
29237 22222 22222 44444 44444
29238 Numeric values are deprecated but are accepted for backward
29240 compatibility reasons.
29241 Default mode is "merge".
29243 flags
29245 Specify flags influencing the filter process.
29246 Available value for flags is:
29248 low_pass_filter, vlpf
29250 Enable linear vertical low-pass filtering in the filter.
29251 Vertical low-pass filtering is required when creating an
29252 interlaced destination from a progressive source which contains
29253 high-frequency vertical detail. Filtering will reduce interlace
29254 'twitter' and Moire patterning.
29255 complex_filter, cvlpf
29257 Enable complex vertical low-pass filtering. This will slightly
29258 less reduce interlace 'twitter' and Moire patterning but better
29259 retain detail and subjective sharpness impression.
29260 bypass_il
29262 Bypass already interlaced frames, only adjust the frame rate.
29263 Vertical low-pass filtering and bypassing already interlaced frames
29265 can only be enabled for mode interleave_top and interleave_bottom.
29266 tmedian
29268 Pick median pixels from several successive input video frames.
29269 The filter accepts the following options:
29271 radius
29273 Set radius of median filter. Default is 1. Allowed range is from 1
29274 to 127.
29275 planes
29277 Set which planes to filter. Default value is 15, by which all
29278 planes are processed.
29279 percentile
29281 Set median percentile. Default value is 0.5. Default value of 0.5
29282 will pick always median values, while 0 will pick minimum values,
29283 and 1 maximum values.
29284 Commands
29286 This filter supports all above options as commands, excluding option
29288 "radius".
29289 tmidequalizer
29291 Apply Temporal Midway Video Equalization effect.
29292 Midway Video Equalization adjusts a sequence of video frames to have
29294 the same histograms, while maintaining their dynamics as much as
29295 possible. It's useful for e.g. matching exposures from a video frames
29296 sequence.
29297 This filter accepts the following option:
29299 radius
29301 Set filtering radius. Default is 5. Allowed range is from 1 to 127.
29302 sigma
29304 Set filtering sigma. Default is 0.5. This controls strength of
29305 filtering. Setting this option to 0 effectively does nothing.
29306 planes
29308 Set which planes to process. Default is 15, which is all available
29309 planes.
29310 tmix
29312 Mix successive video frames.
29313 A description of the accepted options follows.
29315 frames
29317 The number of successive frames to mix. If unspecified, it defaults
29318 to 3.
29319 weights
29321 Specify weight of each input video frame. Each weight is separated
29322 by space. If number of weights is smaller than number of frames
29323 last specified weight will be used for all remaining unset weights.
29324 scale
29326 Specify scale, if it is set it will be multiplied with sum of each
29327 weight multiplied with pixel values to give final destination pixel
29328 value. By default scale is auto scaled to sum of weights.
29329 planes
29331 Set which planes to filter. Default is all. Allowed range is from 0
29332 to 15.
29333 Examples
29335 • Average 7 successive frames:
29337 tmix=frames=7:weights="1 1 1 1 1 1 1"
29339 • Apply simple temporal convolution:
29341 tmix=frames=3:weights="-1 3 -1"
29343 • Similar as above but only showing temporal differences:
29345 tmix=frames=3:weights="-1 2 -1":scale=1
29347 Commands
29349 This filter supports the following commands:
29351 weights
29353 scale
29354 planes
29355 Syntax is same as option with same name.
29356 tonemap
29358 Tone map colors from different dynamic ranges.
29359 This filter expects data in single precision floating point, as it
29361 needs to operate on (and can output) out-of-range values. Another
29362 filter, such as zscale, is needed to convert the resulting frame to a
29363 usable format.
29364 The tonemapping algorithms implemented only work on linear light, so
29366 input data should be linearized beforehand (and possibly correctly
29367 tagged).
29368 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
29370 Options
29372 The filter accepts the following options.
29374 tonemap
29376 Set the tone map algorithm to use.
29377 Possible values are:
29379 none
29381 Do not apply any tone map, only desaturate overbright pixels.
29382 clip
29384 Hard-clip any out-of-range values. Use it for perfect color
29385 accuracy for in-range values, while distorting out-of-range
29386 values.
29387 linear
29389 Stretch the entire reference gamut to a linear multiple of the
29390 display.
29391 gamma
29393 Fit a logarithmic transfer between the tone curves.
29394 reinhard
29396 Preserve overall image brightness with a simple curve, using
29397 nonlinear contrast, which results in flattening details and
29398 degrading color accuracy.
29399 hable
29401 Preserve both dark and bright details better than reinhard, at
29402 the cost of slightly darkening everything. Use it when detail
29403 preservation is more important than color and brightness
29404 accuracy.
29405 mobius
29407 Smoothly map out-of-range values, while retaining contrast and
29408 colors for in-range material as much as possible. Use it when
29409 color accuracy is more important than detail preservation.
29410 Default is none.
29412 param
29414 Tune the tone mapping algorithm.
29415 This affects the following algorithms:
29417 none
29419 Ignored.
29420 linear
29422 Specifies the scale factor to use while stretching. Default to
29423 1.0.
29424 gamma
29426 Specifies the exponent of the function. Default to 1.8.
29427 clip
29429 Specify an extra linear coefficient to multiply into the signal
29430 before clipping. Default to 1.0.
29431 reinhard
29433 Specify the local contrast coefficient at the display peak.
29434 Default to 0.5, which means that in-gamut values will be about
29435 half as bright as when clipping.
29436 hable
29438 Ignored.
29439 mobius
29441 Specify the transition point from linear to mobius transform.
29442 Every value below this point is guaranteed to be mapped 1:1.
29443 The higher the value, the more accurate the result will be, at
29444 the cost of losing bright details. Default to 0.3, which due
29445 to the steep initial slope still preserves in-range colors
29446 fairly accurately.
29447 desat
29449 Apply desaturation for highlights that exceed this level of
29450 brightness. The higher the parameter, the more color information
29451 will be preserved. This setting helps prevent unnaturally blown-out
29452 colors for super-highlights, by (smoothly) turning into white
29453 instead. This makes images feel more natural, at the cost of
29454 reducing information about out-of-range colors.
29455 The default of 2.0 is somewhat conservative and will mostly just
29457 apply to skies or directly sunlit surfaces. A setting of 0.0
29458 disables this option.
29459 This option works only if the input frame has a supported color
29461 tag.
29462 peak
29464 Override signal/nominal/reference peak with this value. Useful when
29465 the embedded peak information in display metadata is not reliable
29466 or when tone mapping from a lower range to a higher range.
29467 tpad
29469 Temporarily pad video frames.
29470 The filter accepts the following options:
29472 start
29474 Specify number of delay frames before input video stream. Default
29475 is 0.
29476 stop
29478 Specify number of padding frames after input video stream. Set to
29479 -1 to pad indefinitely. Default is 0.
29480 start_mode
29482 Set kind of frames added to beginning of stream. Can be either add
29483 or clone. With add frames of solid-color are added. With clone
29484 frames are clones of first frame. Default is add.
29485 stop_mode
29487 Set kind of frames added to end of stream. Can be either add or
29488 clone. With add frames of solid-color are added. With clone
29489 frames are clones of last frame. Default is add.
29490 start_duration, stop_duration
29492 Specify the duration of the start/stop delay. See the Time duration
29493 section in the ffmpeg-utils(1) manual for the accepted syntax.
29494 These options override start and stop. Default is 0.
29495 color
29497 Specify the color of the padded area. For the syntax of this
29498 option, check the "Color" section in the ffmpeg-utils manual.
29499 The default value of color is "black".
29501 transpose
29503 Transpose rows with columns in the input video and optionally flip it.
29504 It accepts the following parameters:
29506 dir Specify the transposition direction.
29508 Can assume the following values:
29510 0, 4, cclock_flip
29512 Rotate by 90 degrees counterclockwise and vertically flip
29513 (default), that is:
29514 L.R L.l
29516 . . -> . .
29517 l.r R.r
29518 1, 5, clock
29520 Rotate by 90 degrees clockwise, that is:
29521 L.R l.L
29523 . . -> . .
29524 l.r r.R
29525 2, 6, cclock
29527 Rotate by 90 degrees counterclockwise, that is:
29528 L.R R.r
29530 . . -> . .
29531 l.r L.l
29532 3, 7, clock_flip
29534 Rotate by 90 degrees clockwise and vertically flip, that is:
29535 L.R r.R
29537 . . -> . .
29538 l.r l.L
29539 For values between 4-7, the transposition is only done if the input
29541 video geometry is portrait and not landscape. These values are
29542 deprecated, the "passthrough" option should be used instead.
29543 Numerical values are deprecated, and should be dropped in favor of
29545 symbolic constants.
29546 passthrough
29548 Do not apply the transposition if the input geometry matches the
29549 one specified by the specified value. It accepts the following
29550 values:
29551 none
29553 Always apply transposition.
29554 portrait
29556 Preserve portrait geometry (when height >= width).
29557 landscape
29559 Preserve landscape geometry (when width >= height).
29560 Default value is "none".
29562 For example to rotate by 90 degrees clockwise and preserve portrait
29564 layout:
29565 transpose=dir=1:passthrough=portrait
29567 The command above can also be specified as:
29569 transpose=1:portrait
29571 transpose_npp
29573 Transpose rows with columns in the input video and optionally flip it.
29574 For more in depth examples see the transpose video filter, which shares
29575 mostly the same options.
29576 It accepts the following parameters:
29578 dir Specify the transposition direction.
29580 Can assume the following values:
29582 cclock_flip
29584 Rotate by 90 degrees counterclockwise and vertically flip.
29585 (default)
29586 clock
29588 Rotate by 90 degrees clockwise.
29589 cclock
29591 Rotate by 90 degrees counterclockwise.
29592 clock_flip
29594 Rotate by 90 degrees clockwise and vertically flip.
29595 passthrough
29597 Do not apply the transposition if the input geometry matches the
29598 one specified by the specified value. It accepts the following
29599 values:
29600 none
29602 Always apply transposition. (default)
29603 portrait
29605 Preserve portrait geometry (when height >= width).
29606 landscape
29608 Preserve landscape geometry (when width >= height).
29609 trim
29611 Trim the input so that the output contains one continuous subpart of
29612 the input.
29613 It accepts the following parameters:
29615 start
29617 Specify the time of the start of the kept section, i.e. the frame
29618 with the timestamp start will be the first frame in the output.
29619 end Specify the time of the first frame that will be dropped, i.e. the
29621 frame immediately preceding the one with the timestamp end will be
29622 the last frame in the output.
29623 start_pts
29625 This is the same as start, except this option sets the start
29626 timestamp in timebase units instead of seconds.
29627 end_pts
29629 This is the same as end, except this option sets the end timestamp
29630 in timebase units instead of seconds.
29631 duration
29633 The maximum duration of the output in seconds.
29634 start_frame
29636 The number of the first frame that should be passed to the output.
29637 end_frame
29639 The number of the first frame that should be dropped.
29640 start, end, and duration are expressed as time duration specifications;
29642 see the Time duration section in the ffmpeg-utils(1) manual for the
29643 accepted syntax.
29644 Note that the first two sets of the start/end options and the duration
29646 option look at the frame timestamp, while the _frame variants simply
29647 count the frames that pass through the filter. Also note that this
29648 filter does not modify the timestamps. If you wish for the output
29649 timestamps to start at zero, insert a setpts filter after the trim
29650 filter.
29651 If multiple start or end options are set, this filter tries to be
29653 greedy and keep all the frames that match at least one of the specified
29654 constraints. To keep only the part that matches all the constraints at
29655 once, chain multiple trim filters.
29656 The defaults are such that all the input is kept. So it is possible to
29658 set e.g. just the end values to keep everything before the specified
29659 time.
29660 Examples:
29662 • Drop everything except the second minute of input:
29664 ffmpeg -i INPUT -vf trim=60:120
29666 • Keep only the first second:
29668 ffmpeg -i INPUT -vf trim=duration=1
29670 unpremultiply
29672 Apply alpha unpremultiply effect to input video stream using first
29673 plane of second stream as alpha.
29674 Both streams must have same dimensions and same pixel format.
29676 The filter accepts the following option:
29678 planes
29680 Set which planes will be processed, unprocessed planes will be
29681 copied. By default value 0xf, all planes will be processed.
29682 If the format has 1 or 2 components, then luma is bit 0. If the
29684 format has 3 or 4 components: for RGB formats bit 0 is green, bit 1
29685 is blue and bit 2 is red; for YUV formats bit 0 is luma, bit 1 is
29686 chroma-U and bit 2 is chroma-V. If present, the alpha channel is
29687 always the last bit.
29688 inplace
29690 Do not require 2nd input for processing, instead use alpha plane
29691 from input stream.
29692 unsharp
29694 Sharpen or blur the input video.
29695 It accepts the following parameters:
29697 luma_msize_x, lx
29699 Set the luma matrix horizontal size. It must be an odd integer
29700 between 3 and 23. The default value is 5.
29701 luma_msize_y, ly
29703 Set the luma matrix vertical size. It must be an odd integer
29704 between 3 and 23. The default value is 5.
29705 luma_amount, la
29707 Set the luma effect strength. It must be a floating point number,
29708 reasonable values lay between -1.5 and 1.5.
29709 Negative values will blur the input video, while positive values
29711 will sharpen it, a value of zero will disable the effect.
29712 Default value is 1.0.
29714 chroma_msize_x, cx
29716 Set the chroma matrix horizontal size. It must be an odd integer
29717 between 3 and 23. The default value is 5.
29718 chroma_msize_y, cy
29720 Set the chroma matrix vertical size. It must be an odd integer
29721 between 3 and 23. The default value is 5.
29722 chroma_amount, ca
29724 Set the chroma effect strength. It must be a floating point number,
29725 reasonable values lay between -1.5 and 1.5.
29726 Negative values will blur the input video, while positive values
29728 will sharpen it, a value of zero will disable the effect.
29729 Default value is 0.0.
29731 alpha_msize_x, ax
29733 Set the alpha matrix horizontal size. It must be an odd integer
29734 between 3 and 23. The default value is 5.
29735 alpha_msize_y, ay
29737 Set the alpha matrix vertical size. It must be an odd integer
29738 between 3 and 23. The default value is 5.
29739 alpha_amount, aa
29741 Set the alpha effect strength. It must be a floating point number,
29742 reasonable values lay between -1.5 and 1.5.
29743 Negative values will blur the input video, while positive values
29745 will sharpen it, a value of zero will disable the effect.
29746 Default value is 0.0.
29748 All parameters are optional and default to the equivalent of the string
29750 '5:5:1.0:5:5:0.0'.
29751 Examples
29753 • Apply strong luma sharpen effect:
29755 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
29757 • Apply a strong blur of both luma and chroma parameters:
29759 unsharp=7:7:-2:7:7:-2
29761 untile
29763 Decompose a video made of tiled images into the individual images.
29764 The frame rate of the output video is the frame rate of the input video
29766 multiplied by the number of tiles.
29767 This filter does the reverse of tile.
29769 The filter accepts the following options:
29771 layout
29773 Set the grid size (i.e. the number of lines and columns). For the
29774 syntax of this option, check the "Video size" section in the
29775 ffmpeg-utils manual.
29776 Examples
29778 • Produce a 1-second video from a still image file made of 25 frames
29780 stacked vertically, like an analogic film reel:
29781 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
29783 uspp
29785 Apply ultra slow/simple postprocessing filter that compresses and
29786 decompresses the image at several (or - in the case of quality level 8
29787 - all) shifts and average the results.
29788 The way this differs from the behavior of spp is that uspp actually
29790 encodes & decodes each case with libavcodec Snow, whereas spp uses a
29791 simplified intra only 8x8 DCT similar to MJPEG.
29792 This filter is only available in ffmpeg version 4.4 or earlier.
29794 The filter accepts the following options:
29796 quality
29798 Set quality. This option defines the number of levels for
29799 averaging. It accepts an integer in the range 0-8. If set to 0, the
29800 filter will have no effect. A value of 8 means the higher quality.
29801 For each increment of that value the speed drops by a factor of
29802 approximately 2. Default value is 3.
29803 qp Force a constant quantization parameter. If not set, the filter
29805 will use the QP from the video stream (if available).
29806 v360
29808 Convert 360 videos between various formats.
29809 The filter accepts the following options:
29811 input
29813 output
29814 Set format of the input/output video.
29815 Available formats:
29817 e
29819 equirect
29820 Equirectangular projection.
29821 c3x2
29823 c6x1
29824 c1x6
29825 Cubemap with 3x2/6x1/1x6 layout.
29826 Format specific options:
29828 in_pad
29830 out_pad
29831 Set padding proportion for the input/output cubemap. Values
29832 in decimals.
29833 Example values:
29835 0 No padding.
29837 0.01
29839 1% of face is padding. For example, with 1920x1280
29840 resolution face size would be 640x640 and padding would
29841 be 3 pixels from each side. (640 * 0.01 = 6 pixels)
29842 Default value is @samp{0}. Maximum value is @samp{0.1}.
29844 fin_pad
29846 fout_pad
29847 Set fixed padding for the input/output cubemap. Values in
29848 pixels.
29849 Default value is @samp{0}. If greater than zero it
29851 overrides other padding options.
29852 in_forder
29854 out_forder
29855 Set order of faces for the input/output cubemap. Choose one
29856 direction for each position.
29857 Designation of directions:
29859 r right
29861 l left
29863 u up
29865 d down
29867 f forward
29869 b back
29871 Default value is @samp{rludfb}.
29873 in_frot
29875 out_frot
29876 Set rotation of faces for the input/output cubemap. Choose
29877 one angle for each position.
29878 Designation of angles:
29880 0 0 degrees clockwise
29882 1 90 degrees clockwise
29884 2 180 degrees clockwise
29886 3 270 degrees clockwise
29888 Default value is @samp{000000}.
29890 eac Equi-Angular Cubemap.
29892 flat
29894 gnomonic
29895 rectilinear
29896 Regular video.
29897 Format specific options:
29899 h_fov
29901 v_fov
29902 d_fov
29903 Set output horizontal/vertical/diagonal field of view.
29904 Values in degrees.
29905 If diagonal field of view is set it overrides horizontal
29907 and vertical field of view.
29908 ih_fov
29910 iv_fov
29911 id_fov
29912 Set input horizontal/vertical/diagonal field of view.
29913 Values in degrees.
29914 If diagonal field of view is set it overrides horizontal
29916 and vertical field of view.
29917 dfisheye
29919 Dual fisheye.
29920 Format specific options:
29922 h_fov
29924 v_fov
29925 d_fov
29926 Set output horizontal/vertical/diagonal field of view.
29927 Values in degrees.
29928 If diagonal field of view is set it overrides horizontal
29930 and vertical field of view.
29931 ih_fov
29933 iv_fov
29934 id_fov
29935 Set input horizontal/vertical/diagonal field of view.
29936 Values in degrees.
29937 If diagonal field of view is set it overrides horizontal
29939 and vertical field of view.
29940 barrel
29942 fb
29943 barrelsplit
29944 Facebook's 360 formats.
29945 sg Stereographic format.
29947 Format specific options:
29949 h_fov
29951 v_fov
29952 d_fov
29953 Set output horizontal/vertical/diagonal field of view.
29954 Values in degrees.
29955 If diagonal field of view is set it overrides horizontal
29957 and vertical field of view.
29958 ih_fov
29960 iv_fov
29961 id_fov
29962 Set input horizontal/vertical/diagonal field of view.
29963 Values in degrees.
29964 If diagonal field of view is set it overrides horizontal
29966 and vertical field of view.
29967 mercator
29969 Mercator format.
29970 ball
29972 Ball format, gives significant distortion toward the back.
29973 hammer
29975 Hammer-Aitoff map projection format.
29976 sinusoidal
29978 Sinusoidal map projection format.
29979 fisheye
29981 Fisheye projection.
29982 Format specific options:
29984 h_fov
29986 v_fov
29987 d_fov
29988 Set output horizontal/vertical/diagonal field of view.
29989 Values in degrees.
29990 If diagonal field of view is set it overrides horizontal
29992 and vertical field of view.
29993 ih_fov
29995 iv_fov
29996 id_fov
29997 Set input horizontal/vertical/diagonal field of view.
29998 Values in degrees.
29999 If diagonal field of view is set it overrides horizontal
30001 and vertical field of view.
30002 pannini
30004 Pannini projection.
30005 Format specific options:
30007 h_fov
30009 Set output pannini parameter.
30010 ih_fov
30012 Set input pannini parameter.
30013 cylindrical
30015 Cylindrical projection.
30016 Format specific options:
30018 h_fov
30020 v_fov
30021 d_fov
30022 Set output horizontal/vertical/diagonal field of view.
30023 Values in degrees.
30024 If diagonal field of view is set it overrides horizontal
30026 and vertical field of view.
30027 ih_fov
30029 iv_fov
30030 id_fov
30031 Set input horizontal/vertical/diagonal field of view.
30032 Values in degrees.
30033 If diagonal field of view is set it overrides horizontal
30035 and vertical field of view.
30036 perspective
30038 Perspective projection. (output only)
30039 Format specific options:
30041 v_fov
30043 Set perspective parameter.
30044 tetrahedron
30046 Tetrahedron projection.
30047 tsp Truncated square pyramid projection.
30049 he
30051 hequirect
30052 Half equirectangular projection.
30053 equisolid
30055 Equisolid format.
30056 Format specific options:
30058 h_fov
30060 v_fov
30061 d_fov
30062 Set output horizontal/vertical/diagonal field of view.
30063 Values in degrees.
30064 If diagonal field of view is set it overrides horizontal
30066 and vertical field of view.
30067 ih_fov
30069 iv_fov
30070 id_fov
30071 Set input horizontal/vertical/diagonal field of view.
30072 Values in degrees.
30073 If diagonal field of view is set it overrides horizontal
30075 and vertical field of view.
30076 og Orthographic format.
30078 Format specific options:
30080 h_fov
30082 v_fov
30083 d_fov
30084 Set output horizontal/vertical/diagonal field of view.
30085 Values in degrees.
30086 If diagonal field of view is set it overrides horizontal
30088 and vertical field of view.
30089 ih_fov
30091 iv_fov
30092 id_fov
30093 Set input horizontal/vertical/diagonal field of view.
30094 Values in degrees.
30095 If diagonal field of view is set it overrides horizontal
30097 and vertical field of view.
30098 octahedron
30100 Octahedron projection.
30101 cylindricalea
30103 Cylindrical Equal Area projection.
30104 interp
30106 Set interpolation method.Note: more complex interpolation methods
30107 require much more memory to run.
30108 Available methods:
30110 near
30112 nearest
30113 Nearest neighbour.
30114 line
30116 linear
30117 Bilinear interpolation.
30118 lagrange9
30120 Lagrange9 interpolation.
30121 cube
30123 cubic
30124 Bicubic interpolation.
30125 lanc
30127 lanczos
30128 Lanczos interpolation.
30129 sp16
30131 spline16
30132 Spline16 interpolation.
30133 gauss
30135 gaussian
30136 Gaussian interpolation.
30137 mitchell
30139 Mitchell interpolation.
30140 Default value is @samp{line}.
30142 w
30144 h Set the output video resolution.
30145 Default resolution depends on formats.
30147 in_stereo
30149 out_stereo
30150 Set the input/output stereo format.
30151 2d 2D mono
30153 sbs Side by side
30155 tb Top bottom
30157 Default value is @samp{2d} for input and output format.
30159 yaw
30161 pitch
30162 roll
30163 Set rotation for the output video. Values in degrees.
30164 rorder
30166 Set rotation order for the output video. Choose one item for each
30167 position.
30168 y, Y
30170 yaw
30171 p, P
30173 pitch
30174 r, R
30176 roll
30177 Default value is @samp{ypr}.
30179 h_flip
30181 v_flip
30182 d_flip
30183 Flip the output video horizontally(swaps
30184 left-right)/vertically(swaps up-down)/in-depth(swaps back-forward).
30185 Boolean values.
30186 ih_flip
30188 iv_flip
30189 Set if input video is flipped horizontally/vertically. Boolean
30190 values.
30191 in_trans
30193 Set if input video is transposed. Boolean value, by default
30194 disabled.
30195 out_trans
30197 Set if output video needs to be transposed. Boolean value, by
30198 default disabled.
30199 h_offset
30201 v_offset
30202 Set output horizontal/vertical off-axis offset. Default is set to
30203 0. Allowed range is from -1 to 1.
30204 alpha_mask
30206 Build mask in alpha plane for all unmapped pixels by marking them
30207 fully transparent. Boolean value, by default disabled.
30208 reset_rot
30210 Reset rotation of output video. Boolean value, by default disabled.
30211 Examples
30213 • Convert equirectangular video to cubemap with 3x2 layout and 1%
30215 padding using bicubic interpolation:
30216 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
30218 • Extract back view of Equi-Angular Cubemap:
30220 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
30222 • Convert transposed and horizontally flipped Equi-Angular Cubemap in
30224 side-by-side stereo format to equirectangular top-bottom stereo
30225 format:
30226 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
30228 Commands
30230 This filter supports subset of above options as commands.
30232 vaguedenoiser
30234 Apply a wavelet based denoiser.
30235 It transforms each frame from the video input into the wavelet domain,
30237 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
30238 the obtained coefficients. It does an inverse wavelet transform after.
30239 Due to wavelet properties, it should give a nice smoothed result, and
30240 reduced noise, without blurring picture features.
30241 This filter accepts the following options:
30243 threshold
30245 The filtering strength. The higher, the more filtered the video
30246 will be. Hard thresholding can use a higher threshold than soft
30247 thresholding before the video looks overfiltered. Default value is
30248 2.
30249 method
30251 The filtering method the filter will use.
30252 It accepts the following values:
30254 hard
30256 All values under the threshold will be zeroed.
30257 soft
30259 All values under the threshold will be zeroed. All values above
30260 will be reduced by the threshold.
30261 garrote
30263 Scales or nullifies coefficients - intermediary between (more)
30264 soft and (less) hard thresholding.
30265 Default is garrote.
30267 nsteps
30269 Number of times, the wavelet will decompose the picture. Picture
30270 can't be decomposed beyond a particular point (typically, 8 for a
30271 640x480 frame - as 2^9 = 512 > 480). Valid values are integers
30272 between 1 and 32. Default value is 6.
30273 percent
30275 Partial of full denoising (limited coefficients shrinking), from 0
30276 to 100. Default value is 85.
30277 planes
30279 A list of the planes to process. By default all planes are
30280 processed.
30281 type
30283 The threshold type the filter will use.
30284 It accepts the following values:
30286 universal
30288 Threshold used is same for all decompositions.
30289 bayes
30291 Threshold used depends also on each decomposition coefficients.
30292 Default is universal.
30294 varblur
30296 Apply variable blur filter by using 2nd video stream to set blur
30297 radius. The 2nd stream must have the same dimensions.
30298 This filter accepts the following options:
30300 min_r
30302 Set min allowed radius. Allowed range is from 0 to 254. Default is
30303 0.
30304 max_r
30306 Set max allowed radius. Allowed range is from 1 to 255. Default is
30307 8.
30308 planes
30310 Set which planes to process. By default, all are used.
30311 The "varblur" filter also supports the framesync options.
30313 Commands
30315 This filter supports all the above options as commands.
30317 vectorscope
30319 Display 2 color component values in the two dimensional graph (which is
30320 called a vectorscope).
30321 This filter accepts the following options:
30323 mode, m
30325 Set vectorscope mode.
30326 It accepts the following values:
30328 gray
30330 tint
30331 Gray values are displayed on graph, higher brightness means
30332 more pixels have same component color value on location in
30333 graph. This is the default mode.
30334 color
30336 Gray values are displayed on graph. Surrounding pixels values
30337 which are not present in video frame are drawn in gradient of 2
30338 color components which are set by option "x" and "y". The 3rd
30339 color component is static.
30340 color2
30342 Actual color components values present in video frame are
30343 displayed on graph.
30344 color3
30346 Similar as color2 but higher frequency of same values "x" and
30347 "y" on graph increases value of another color component, which
30348 is luminance by default values of "x" and "y".
30349 color4
30351 Actual colors present in video frame are displayed on graph. If
30352 two different colors map to same position on graph then color
30353 with higher value of component not present in graph is picked.
30354 color5
30356 Gray values are displayed on graph. Similar to "color" but with
30357 3rd color component picked from radial gradient.
30358 x Set which color component will be represented on X-axis. Default is
30360 1.
30361 y Set which color component will be represented on Y-axis. Default is
30363 2.
30364 intensity, i
30366 Set intensity, used by modes: gray, color, color3 and color5 for
30367 increasing brightness of color component which represents frequency
30368 of (X, Y) location in graph.
30369 envelope, e
30371 none
30372 No envelope, this is default.
30373 instant
30375 Instant envelope, even darkest single pixel will be clearly
30376 highlighted.
30377 peak
30379 Hold maximum and minimum values presented in graph over time.
30380 This way you can still spot out of range values without
30381 constantly looking at vectorscope.
30382 peak+instant
30384 Peak and instant envelope combined together.
30385 graticule, g
30387 Set what kind of graticule to draw.
30388 none
30390 green
30391 color
30392 invert
30393 opacity, o
30394 Set graticule opacity.
30395 flags, f
30397 Set graticule flags.
30398 white
30400 Draw graticule for white point.
30401 black
30403 Draw graticule for black point.
30404 name
30406 Draw color points short names.
30407 bgopacity, b
30409 Set background opacity.
30410 lthreshold, l
30412 Set low threshold for color component not represented on X or Y
30413 axis. Values lower than this value will be ignored. Default is 0.
30414 Note this value is multiplied with actual max possible value one
30415 pixel component can have. So for 8-bit input and low threshold
30416 value of 0.1 actual threshold is 0.1 * 255 = 25.
30417 hthreshold, h
30419 Set high threshold for color component not represented on X or Y
30420 axis. Values higher than this value will be ignored. Default is 1.
30421 Note this value is multiplied with actual max possible value one
30422 pixel component can have. So for 8-bit input and high threshold
30423 value of 0.9 actual threshold is 0.9 * 255 = 230.
30424 colorspace, c
30426 Set what kind of colorspace to use when drawing graticule.
30427 auto
30429 601
30430 709
30431 Default is auto.
30433 tint0, t0
30435 tint1, t1
30436 Set color tint for gray/tint vectorscope mode. By default both
30437 options are zero. This means no tint, and output will remain gray.
30438 vidstabdetect
30440 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
30441 vidstabtransform for pass 2.
30442 This filter generates a file with relative translation and rotation
30444 transform information about subsequent frames, which is then used by
30445 the vidstabtransform filter.
30446 To enable compilation of this filter you need to configure FFmpeg with
30448 "--enable-libvidstab".
30449 This filter accepts the following options:
30451 result
30453 Set the path to the file used to write the transforms information.
30454 Default value is transforms.trf.
30455 shakiness
30457 Set how shaky the video is and how quick the camera is. It accepts
30458 an integer in the range 1-10, a value of 1 means little shakiness,
30459 a value of 10 means strong shakiness. Default value is 5.
30460 accuracy
30462 Set the accuracy of the detection process. It must be a value in
30463 the range 1-15. A value of 1 means low accuracy, a value of 15
30464 means high accuracy. Default value is 15.
30465 stepsize
30467 Set stepsize of the search process. The region around minimum is
30468 scanned with 1 pixel resolution. Default value is 6.
30469 mincontrast
30471 Set minimum contrast. Below this value a local measurement field is
30472 discarded. Must be a floating point value in the range 0-1. Default
30473 value is 0.3.
30474 tripod
30476 Set reference frame number for tripod mode.
30477 If enabled, the motion of the frames is compared to a reference
30479 frame in the filtered stream, identified by the specified number.
30480 The idea is to compensate all movements in a more-or-less static
30481 scene and keep the camera view absolutely still.
30482 If set to 0, it is disabled. The frames are counted starting from
30484 1.
30485 show
30487 Show fields and transforms in the resulting frames. It accepts an
30488 integer in the range 0-2. Default value is 0, which disables any
30489 visualization.
30490 Examples
30492 • Use default values:
30494 vidstabdetect
30496 • Analyze strongly shaky movie and put the results in file
30498 mytransforms.trf:
30499 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
30501 • Visualize the result of internal transformations in the resulting
30503 video:
30504 vidstabdetect=show=1
30506 • Analyze a video with medium shakiness using ffmpeg:
30508 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
30510 vidstabtransform
30512 Video stabilization/deshaking: pass 2 of 2, see vidstabdetect for pass
30513 1.
30514 Read a file with transform information for each frame and
30516 apply/compensate them. Together with the vidstabdetect filter this can
30517 be used to deshake videos. See also
30518 <http://public.hronopik.de/vid.stab>. It is important to also use the
30519 unsharp filter, see below.
30520 To enable compilation of this filter you need to configure FFmpeg with
30522 "--enable-libvidstab".
30523 Options
30525 input
30527 Set path to the file used to read the transforms. Default value is
30528 transforms.trf.
30529 smoothing
30531 Set the number of frames (value*2 + 1) used for lowpass filtering
30532 the camera movements. Default value is 10.
30533 For example a number of 10 means that 21 frames are used (10 in the
30535 past and 10 in the future) to smoothen the motion in the video. A
30536 larger value leads to a smoother video, but limits the acceleration
30537 of the camera (pan/tilt movements). 0 is a special case where a
30538 static camera is simulated.
30539 optalgo
30541 Set the camera path optimization algorithm.
30542 Accepted values are:
30544 gauss
30546 gaussian kernel low-pass filter on camera motion (default)
30547 avg averaging on transformations
30549 maxshift
30551 Set maximal number of pixels to translate frames. Default value is
30552 -1, meaning no limit.
30553 maxangle
30555 Set maximal angle in radians (degree*PI/180) to rotate frames.
30556 Default value is -1, meaning no limit.
30557 crop
30559 Specify how to deal with borders that may be visible due to
30560 movement compensation.
30561 Available values are:
30563 keep
30565 keep image information from previous frame (default)
30566 black
30568 fill the border black
30569 invert
30571 Invert transforms if set to 1. Default value is 0.
30572 relative
30574 Consider transforms as relative to previous frame if set to 1,
30575 absolute if set to 0. Default value is 0.
30576 zoom
30578 Set percentage to zoom. A positive value will result in a zoom-in
30579 effect, a negative value in a zoom-out effect. Default value is 0
30580 (no zoom).
30581 optzoom
30583 Set optimal zooming to avoid borders.
30584 Accepted values are:
30586 0 disabled
30588 1 optimal static zoom value is determined (only very strong
30590 movements will lead to visible borders) (default)
30591 2 optimal adaptive zoom value is determined (no borders will be
30593 visible), see zoomspeed
30594 Note that the value given at zoom is added to the one calculated
30596 here.
30597 zoomspeed
30599 Set percent to zoom maximally each frame (enabled when optzoom is
30600 set to 2). Range is from 0 to 5, default value is 0.25.
30601 interpol
30603 Specify type of interpolation.
30604 Available values are:
30606 no no interpolation
30608 linear
30610 linear only horizontal
30611 bilinear
30613 linear in both directions (default)
30614 bicubic
30616 cubic in both directions (slow)
30617 tripod
30619 Enable virtual tripod mode if set to 1, which is equivalent to
30620 "relative=0:smoothing=0". Default value is 0.
30621 Use also "tripod" option of vidstabdetect.
30623 debug
30625 Increase log verbosity if set to 1. Also the detected global
30626 motions are written to the temporary file global_motions.trf.
30627 Default value is 0.
30628 Examples
30630 • Use ffmpeg for a typical stabilization with default values:
30632 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
30634 Note the use of the unsharp filter which is always recommended.
30636 • Zoom in a bit more and load transform data from a given file:
30638 vidstabtransform=zoom=5:input="mytransforms.trf"
30640 • Smoothen the video even more:
30642 vidstabtransform=smoothing=30
30644 vflip
30646 Flip the input video vertically.
30647 For example, to vertically flip a video with ffmpeg:
30649 ffmpeg -i in.avi -vf "vflip" out.avi
30651 vfrdet
30653 Detect variable frame rate video.
30654 This filter tries to detect if the input is variable or constant frame
30656 rate.
30657 At end it will output number of frames detected as having variable
30659 delta pts, and ones with constant delta pts. If there was frames with
30660 variable delta, than it will also show min, max and average delta
30661 encountered.
30662 vibrance
30664 Boost or alter saturation.
30665 The filter accepts the following options:
30667 intensity
30669 Set strength of boost if positive value or strength of alter if
30670 negative value. Default is 0. Allowed range is from -2 to 2.
30671 rbal
30673 Set the red balance. Default is 1. Allowed range is from -10 to 10.
30674 gbal
30676 Set the green balance. Default is 1. Allowed range is from -10 to
30677 10.
30678 bbal
30680 Set the blue balance. Default is 1. Allowed range is from -10 to
30681 10.
30682 rlum
30684 Set the red luma coefficient.
30685 glum
30687 Set the green luma coefficient.
30688 blum
30690 Set the blue luma coefficient.
30691 alternate
30693 If "intensity" is negative and this is set to 1, colors will
30694 change, otherwise colors will be less saturated, more towards gray.
30695 Commands
30697 This filter supports the all above options as commands.
30699 vif
30701 Obtain the average VIF (Visual Information Fidelity) between two input
30702 videos.
30703 This filter takes two input videos.
30705 Both input videos must have the same resolution and pixel format for
30707 this filter to work correctly. Also it assumes that both inputs have
30708 the same number of frames, which are compared one by one.
30709 The obtained average VIF score is printed through the logging system.
30711 The filter stores the calculated VIF score of each frame.
30713 This filter also supports the framesync options.
30715 In the below example the input file main.mpg being processed is
30717 compared with the reference file ref.mpg.
30718 ffmpeg -i main.mpg -i ref.mpg -lavfi vif -f null -
30720 vignette
30722 Make or reverse a natural vignetting effect.
30723 The filter accepts the following options:
30725 angle, a
30727 Set lens angle expression as a number of radians.
30728 The value is clipped in the "[0,PI/2]" range.
30730 Default value: "PI/5"
30732 x0
30734 y0 Set center coordinates expressions. Respectively "w/2" and "h/2" by
30735 default.
30736 mode
30738 Set forward/backward mode.
30739 Available modes are:
30741 forward
30743 The larger the distance from the central point, the darker the
30744 image becomes.
30745 backward
30747 The larger the distance from the central point, the brighter
30748 the image becomes. This can be used to reverse a vignette
30749 effect, though there is no automatic detection to extract the
30750 lens angle and other settings (yet). It can also be used to
30751 create a burning effect.
30752 Default value is forward.
30754 eval
30756 Set evaluation mode for the expressions (angle, x0, y0).
30757 It accepts the following values:
30759 init
30761 Evaluate expressions only once during the filter
30762 initialization.
30763 frame
30765 Evaluate expressions for each incoming frame. This is way
30766 slower than the init mode since it requires all the scalers to
30767 be re-computed, but it allows advanced dynamic expressions.
30768 Default value is init.
30770 dither
30772 Set dithering to reduce the circular banding effects. Default is 1
30773 (enabled).
30774 aspect
30776 Set vignette aspect. This setting allows one to adjust the shape of
30777 the vignette. Setting this value to the SAR of the input will make
30778 a rectangular vignetting following the dimensions of the video.
30779 Default is "1/1".
30781 Expressions
30783 The alpha, x0 and y0 expressions can contain the following parameters.
30785 w
30787 h input width and height
30788 n the number of input frame, starting from 0
30790 pts the PTS (Presentation TimeStamp) time of the filtered video frame,
30792 expressed in TB units, NAN if undefined
30793 r frame rate of the input video, NAN if the input frame rate is
30795 unknown
30796 t the PTS (Presentation TimeStamp) of the filtered video frame,
30798 expressed in seconds, NAN if undefined
30799 tb time base of the input video
30801 Examples
30803 • Apply simple strong vignetting effect:
30805 vignette=PI/4
30807 • Make a flickering vignetting:
30809 vignette='PI/4+random(1)*PI/50':eval=frame
30811 vmafmotion
30813 Obtain the average VMAF motion score of a video. It is one of the
30814 component metrics of VMAF.
30815 The obtained average motion score is printed through the logging
30817 system.
30818 The filter accepts the following options:
30820 stats_file
30822 If specified, the filter will use the named file to save the motion
30823 score of each frame with respect to the previous frame. When
30824 filename equals "-" the data is sent to standard output.
30825 Example:
30827 ffmpeg -i ref.mpg -vf vmafmotion -f null -
30829 vstack
30831 Stack input videos vertically.
30832 All streams must be of same pixel format and of same width.
30834 Note that this filter is faster than using overlay and pad filter to
30836 create same output.
30837 The filter accepts the following options:
30839 inputs
30841 Set number of input streams. Default is 2.
30842 shortest
30844 If set to 1, force the output to terminate when the shortest input
30845 terminates. Default value is 0.
30846 w3fdif
30848 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
30849 Deinterlacing Filter").
30850 Based on the process described by Martin Weston for BBC R&D, and
30852 implemented based on the de-interlace algorithm written by Jim
30853 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter uses
30854 filter coefficients calculated by BBC R&D.
30855 This filter uses field-dominance information in frame to decide which
30857 of each pair of fields to place first in the output. If it gets it
30858 wrong use setfield filter before "w3fdif" filter.
30859 There are two sets of filter coefficients, so called "simple" and
30861 "complex". Which set of filter coefficients is used can be set by
30862 passing an optional parameter:
30863 filter
30865 Set the interlacing filter coefficients. Accepts one of the
30866 following values:
30867 simple
30869 Simple filter coefficient set.
30870 complex
30872 More-complex filter coefficient set.
30873 Default value is complex.
30875 mode
30877 The interlacing mode to adopt. It accepts one of the following
30878 values:
30879 frame
30881 Output one frame for each frame.
30882 field
30884 Output one frame for each field.
30885 The default value is "field".
30887 parity
30889 The picture field parity assumed for the input interlaced video. It
30890 accepts one of the following values:
30891 tff Assume the top field is first.
30893 bff Assume the bottom field is first.
30895 auto
30897 Enable automatic detection of field parity.
30898 The default value is "auto". If the interlacing is unknown or the
30900 decoder does not export this information, top field first will be
30901 assumed.
30902 deint
30904 Specify which frames to deinterlace. Accepts one of the following
30905 values:
30906 all Deinterlace all frames,
30908 interlaced
30910 Only deinterlace frames marked as interlaced.
30911 Default value is all.
30913 Commands
30915 This filter supports same commands as options.
30917 waveform
30919 Video waveform monitor.
30920 The waveform monitor plots color component intensity. By default
30922 luminance only. Each column of the waveform corresponds to a column of
30923 pixels in the source video.
30924 It accepts the following options:
30926 mode, m
30928 Can be either "row", or "column". Default is "column". In row
30929 mode, the graph on the left side represents color component value 0
30930 and the right side represents value = 255. In column mode, the top
30931 side represents color component value = 0 and bottom side
30932 represents value = 255.
30933 intensity, i
30935 Set intensity. Smaller values are useful to find out how many
30936 values of the same luminance are distributed across input
30937 rows/columns. Default value is 0.04. Allowed range is [0, 1].
30938 mirror, r
30940 Set mirroring mode. 0 means unmirrored, 1 means mirrored. In
30941 mirrored mode, higher values will be represented on the left side
30942 for "row" mode and at the top for "column" mode. Default is 1
30943 (mirrored).
30944 display, d
30946 Set display mode. It accepts the following values:
30947 overlay
30949 Presents information identical to that in the "parade", except
30950 that the graphs representing color components are superimposed
30951 directly over one another.
30952 This display mode makes it easier to spot relative differences
30954 or similarities in overlapping areas of the color components
30955 that are supposed to be identical, such as neutral whites,
30956 grays, or blacks.
30957 stack
30959 Display separate graph for the color components side by side in
30960 "row" mode or one below the other in "column" mode.
30961 parade
30963 Display separate graph for the color components side by side in
30964 "column" mode or one below the other in "row" mode.
30965 Using this display mode makes it easy to spot color casts in
30967 the highlights and shadows of an image, by comparing the
30968 contours of the top and the bottom graphs of each waveform.
30969 Since whites, grays, and blacks are characterized by exactly
30970 equal amounts of red, green, and blue, neutral areas of the
30971 picture should display three waveforms of roughly equal
30972 width/height. If not, the correction is easy to perform by
30973 making level adjustments the three waveforms.
30974 Default is "stack".
30976 components, c
30978 Set which color components to display. Default is 1, which means
30979 only luminance or red color component if input is in RGB
30980 colorspace. If is set for example to 7 it will display all 3 (if)
30981 available color components.
30982 envelope, e
30984 none
30985 No envelope, this is default.
30986 instant
30988 Instant envelope, minimum and maximum values presented in graph
30989 will be easily visible even with small "step" value.
30990 peak
30992 Hold minimum and maximum values presented in graph across time.
30993 This way you can still spot out of range values without
30994 constantly looking at waveforms.
30995 peak+instant
30997 Peak and instant envelope combined together.
30998 filter, f
31000 lowpass
31001 No filtering, this is default.
31002 flat
31004 Luma and chroma combined together.
31005 aflat
31007 Similar as above, but shows difference between blue and red
31008 chroma.
31009 xflat
31011 Similar as above, but use different colors.
31012 yflat
31014 Similar as above, but again with different colors.
31015 chroma
31017 Displays only chroma.
31018 color
31020 Displays actual color value on waveform.
31021 acolor
31023 Similar as above, but with luma showing frequency of chroma
31024 values.
31025 graticule, g
31027 Set which graticule to display.
31028 none
31030 Do not display graticule.
31031 green
31033 Display green graticule showing legal broadcast ranges.
31034 orange
31036 Display orange graticule showing legal broadcast ranges.
31037 invert
31039 Display invert graticule showing legal broadcast ranges.
31040 opacity, o
31042 Set graticule opacity.
31043 flags, fl
31045 Set graticule flags.
31046 numbers
31048 Draw numbers above lines. By default enabled.
31049 dots
31051 Draw dots instead of lines.
31052 scale, s
31054 Set scale used for displaying graticule.
31055 digital
31057 millivolts
31058 ire
31059 Default is digital.
31061 bgopacity, b
31063 Set background opacity.
31064 tint0, t0
31066 tint1, t1
31067 Set tint for output. Only used with lowpass filter and when
31068 display is not overlay and input pixel formats are not RGB.
31069 fitmode, fm
31071 Set sample aspect ratio of video output frames. Can be used to
31072 configure waveform so it is not streched too much in one of
31073 directions.
31074 none
31076 Set sample aspect ration to 1/1.
31077 size
31079 Set sample aspect ratio to match input size of video
31080 Default is none.
31082 weave, doubleweave
31084 The "weave" takes a field-based video input and join each two
31085 sequential fields into single frame, producing a new double height clip
31086 with half the frame rate and half the frame count.
31087 The "doubleweave" works same as "weave" but without halving frame rate
31089 and frame count.
31090 It accepts the following option:
31092 first_field
31094 Set first field. Available values are:
31095 top, t
31097 Set the frame as top-field-first.
31098 bottom, b
31100 Set the frame as bottom-field-first.
31101 Examples
31103 • Interlace video using select and separatefields filter:
31105 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
31107 xbr
31109 Apply the xBR high-quality magnification filter which is designed for
31110 pixel art. It follows a set of edge-detection rules, see
31111 <https://forums.libretro.com/t/xbr-algorithm-tutorial/123>.
31112 It accepts the following option:
31114 n Set the scaling dimension: 2 for "2xBR", 3 for "3xBR" and 4 for
31116 "4xBR". Default is 3.
31117 xcorrelate
31119 Apply normalized cross-correlation between first and second input video
31120 stream.
31121 Second input video stream dimensions must be lower than first input
31123 video stream.
31124 The filter accepts the following options:
31126 planes
31128 Set which planes to process.
31129 secondary
31131 Set which secondary video frames will be processed from second
31132 input video stream, can be first or all. Default is all.
31133 The "xcorrelate" filter also supports the framesync options.
31135 xfade
31137 Apply cross fade from one input video stream to another input video
31138 stream. The cross fade is applied for specified duration.
31139 Both inputs must be constant frame-rate and have the same resolution,
31141 pixel format, frame rate and timebase.
31142 The filter accepts the following options:
31144 transition
31146 Set one of available transition effects:
31147 custom
31149 fade
31150 wipeleft
31151 wiperight
31152 wipeup
31153 wipedown
31154 slideleft
31155 slideright
31156 slideup
31157 slidedown
31158 circlecrop
31159 rectcrop
31160 distance
31161 fadeblack
31162 fadewhite
31163 radial
31164 smoothleft
31165 smoothright
31166 smoothup
31167 smoothdown
31168 circleopen
31169 circleclose
31170 vertopen
31171 vertclose
31172 horzopen
31173 horzclose
31174 dissolve
31175 pixelize
31176 diagtl
31177 diagtr
31178 diagbl
31179 diagbr
31180 hlslice
31181 hrslice
31182 vuslice
31183 vdslice
31184 hblur
31185 fadegrays
31186 wipetl
31187 wipetr
31188 wipebl
31189 wipebr
31190 squeezeh
31191 squeezev
31192 zoomin
31193 fadefast
31194 fadeslow
31195 Default transition effect is fade.
31197 duration
31199 Set cross fade duration in seconds. Range is 0 to 60 seconds.
31200 Default duration is 1 second.
31201 offset
31203 Set cross fade start relative to first input stream in seconds.
31204 Default offset is 0.
31205 expr
31207 Set expression for custom transition effect.
31208 The expressions can use the following variables and functions:
31210 X
31212 Y The coordinates of the current sample.
31213 W
31215 H The width and height of the image.
31216 P Progress of transition effect.
31218 PLANE
31220 Currently processed plane.
31221 A Return value of first input at current location and plane.
31223 B Return value of second input at current location and plane.
31225 a0(x, y)
31227 a1(x, y)
31228 a2(x, y)
31229 a3(x, y)
31230 Return the value of the pixel at location (x,y) of the
31231 first/second/third/fourth component of first input.
31232 b0(x, y)
31234 b1(x, y)
31235 b2(x, y)
31236 b3(x, y)
31237 Return the value of the pixel at location (x,y) of the
31238 first/second/third/fourth component of second input.
31239 Examples
31241 • Cross fade from one input video to another input video, with fade
31243 transition and duration of transition of 2 seconds starting at
31244 offset of 5 seconds:
31245 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
31247 xmedian
31249 Pick median pixels from several input videos.
31250 The filter accepts the following options:
31252 inputs
31254 Set number of inputs. Default is 3. Allowed range is from 3 to
31255 255. If number of inputs is even number, than result will be mean
31256 value between two median values.
31257 planes
31259 Set which planes to filter. Default value is 15, by which all
31260 planes are processed.
31261 percentile
31263 Set median percentile. Default value is 0.5. Default value of 0.5
31264 will pick always median values, while 0 will pick minimum values,
31265 and 1 maximum values.
31266 Commands
31268 This filter supports all above options as commands, excluding option
31270 "inputs".
31271 xstack
31273 Stack video inputs into custom layout.
31274 All streams must be of same pixel format.
31276 The filter accepts the following options:
31278 inputs
31280 Set number of input streams. Default is 2.
31281 layout
31283 Specify layout of inputs. This option requires the desired layout
31284 configuration to be explicitly set by the user. This sets position
31285 of each video input in output. Each input is separated by '|'. The
31286 first number represents the column, and the second number
31287 represents the row. Numbers start at 0 and are separated by '_'.
31288 Optionally one can use wX and hX, where X is video input from which
31289 to take width or height. Multiple values can be used when
31290 separated by '+'. In such case values are summed together.
31291 Note that if inputs are of different sizes gaps may appear, as not
31293 all of the output video frame will be filled. Similarly, videos can
31294 overlap each other if their position doesn't leave enough space for
31295 the full frame of adjoining videos.
31296 For 2 inputs, a default layout of "0_0|w0_0" (equivalent to
31298 "grid=2x1") is set. In all other cases, a layout or a grid must be
31299 set by the user. Either "grid" or "layout" can be specified at a
31300 time. Specifying both will result in an error.
31301 grid
31303 Specify a fixed size grid of inputs. This option is used to create
31304 a fixed size grid of the input streams. Set the grid size in the
31305 form "COLUMNSxROWS". There must be "ROWS * COLUMNS" input streams
31306 and they will be arranged as a grid with "ROWS" rows and "COLUMNS"
31307 columns. When using this option, each input stream within a row
31308 must have the same height and all the rows must have the same
31309 width.
31310 If "grid" is set, then "inputs" option is ignored and is implicitly
31312 set to "ROWS * COLUMNS".
31313 For 2 inputs, a default grid of "2x1" (equivalent to
31315 "layout=0_0|w0_0") is set. In all other cases, a layout or a grid
31316 must be set by the user. Either "grid" or "layout" can be specified
31317 at a time. Specifying both will result in an error.
31318 shortest
31320 If set to 1, force the output to terminate when the shortest input
31321 terminates. Default value is 0.
31322 fill
31324 If set to valid color, all unused pixels will be filled with that
31325 color. By default fill is set to none, so it is disabled.
31326 Examples
31328 • Display 4 inputs into 2x2 grid.
31330 Layout:
31332 input1(0, 0) | input3(w0, 0)
31334 input2(0, h0) | input4(w0, h0)
31335 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
31339 Note that if inputs are of different sizes, gaps or overlaps may
31341 occur.
31342 • Display 4 inputs into 1x4 grid.
31344 Layout:
31346 input1(0, 0)
31348 input2(0, h0)
31349 input3(0, h0+h1)
31350 input4(0, h0+h1+h2)
31351 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
31355 Note that if inputs are of different widths, unused space will
31357 appear.
31358 • Display 9 inputs into 3x3 grid.
31360 Layout:
31362 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
31364 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
31365 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
31366 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
31370 Note that if inputs are of different sizes, gaps or overlaps may
31372 occur.
31373 • Display 16 inputs into 4x4 grid.
31375 Layout:
31377 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
31379 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
31380 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
31381 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
31382 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|
31386 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
31387 Note that if inputs are of different sizes, gaps or overlaps may
31389 occur.
31390 yadif
31392 Deinterlace the input video ("yadif" means "yet another deinterlacing
31393 filter").
31394 It accepts the following parameters:
31396 mode
31398 The interlacing mode to adopt. It accepts one of the following
31399 values:
31400 0, send_frame
31402 Output one frame for each frame.
31403 1, send_field
31405 Output one frame for each field.
31406 2, send_frame_nospatial
31408 Like "send_frame", but it skips the spatial interlacing check.
31409 3, send_field_nospatial
31411 Like "send_field", but it skips the spatial interlacing check.
31412 The default value is "send_frame".
31414 parity
31416 The picture field parity assumed for the input interlaced video. It
31417 accepts one of the following values:
31418 0, tff
31420 Assume the top field is first.
31421 1, bff
31423 Assume the bottom field is first.
31424 -1, auto
31426 Enable automatic detection of field parity.
31427 The default value is "auto". If the interlacing is unknown or the
31429 decoder does not export this information, top field first will be
31430 assumed.
31431 deint
31433 Specify which frames to deinterlace. Accepts one of the following
31434 values:
31435 0, all
31437 Deinterlace all frames.
31438 1, interlaced
31440 Only deinterlace frames marked as interlaced.
31441 The default value is "all".
31443 yadif_cuda
31445 Deinterlace the input video using the yadif algorithm, but implemented
31446 in CUDA so that it can work as part of a GPU accelerated pipeline with
31447 nvdec and/or nvenc.
31448 It accepts the following parameters:
31450 mode
31452 The interlacing mode to adopt. It accepts one of the following
31453 values:
31454 0, send_frame
31456 Output one frame for each frame.
31457 1, send_field
31459 Output one frame for each field.
31460 2, send_frame_nospatial
31462 Like "send_frame", but it skips the spatial interlacing check.
31463 3, send_field_nospatial
31465 Like "send_field", but it skips the spatial interlacing check.
31466 The default value is "send_frame".
31468 parity
31470 The picture field parity assumed for the input interlaced video. It
31471 accepts one of the following values:
31472 0, tff
31474 Assume the top field is first.
31475 1, bff
31477 Assume the bottom field is first.
31478 -1, auto
31480 Enable automatic detection of field parity.
31481 The default value is "auto". If the interlacing is unknown or the
31483 decoder does not export this information, top field first will be
31484 assumed.
31485 deint
31487 Specify which frames to deinterlace. Accepts one of the following
31488 values:
31489 0, all
31491 Deinterlace all frames.
31492 1, interlaced
31494 Only deinterlace frames marked as interlaced.
31495 The default value is "all".
31497 yaepblur
31499 Apply blur filter while preserving edges ("yaepblur" means "yet another
31500 edge preserving blur filter"). The algorithm is described in "J. S.
31501 Lee, Digital image enhancement and noise filtering by use of local
31502 statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
31503 It accepts the following parameters:
31505 radius, r
31507 Set the window radius. Default value is 3.
31508 planes, p
31510 Set which planes to filter. Default is only the first plane.
31511 sigma, s
31513 Set blur strength. Default value is 128.
31514 Commands
31516 This filter supports same commands as options.
31518 zoompan
31520 Apply Zoom & Pan effect.
31521 This filter accepts the following options:
31523 zoom, z
31525 Set the zoom expression. Range is 1-10. Default is 1.
31526 x
31528 y Set the x and y expression. Default is 0.
31529 d Set the duration expression in number of frames. This sets for how
31531 many number of frames effect will last for single input image.
31532 Default is 90.
31533 s Set the output image size, default is 'hd720'.
31535 fps Set the output frame rate, default is '25'.
31537 Each expression can contain the following constants:
31539 in_w, iw
31541 Input width.
31542 in_h, ih
31544 Input height.
31545 out_w, ow
31547 Output width.
31548 out_h, oh
31550 Output height.
31551 in Input frame count.
31553 on Output frame count.
31555 in_time, it
31557 The input timestamp expressed in seconds. It's NAN if the input
31558 timestamp is unknown.
31559 out_time, time, ot
31561 The output timestamp expressed in seconds.
31562 x
31564 y Last calculated 'x' and 'y' position from 'x' and 'y' expression
31565 for current input frame.
31566 px
31568 py 'x' and 'y' of last output frame of previous input frame or 0 when
31569 there was not yet such frame (first input frame).
31570 zoom
31572 Last calculated zoom from 'z' expression for current input frame.
31573 pzoom
31575 Last calculated zoom of last output frame of previous input frame.
31576 duration
31578 Number of output frames for current input frame. Calculated from
31579 'd' expression for each input frame.
31580 pduration
31582 number of output frames created for previous input frame
31583 a Rational number: input width / input height
31585 sar sample aspect ratio
31587 dar display aspect ratio
31589 Examples
31591 • Zoom in up to 1.5x and pan at same time to some spot near center of
31593 picture:
31594 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
31596 • Zoom in up to 1.5x and pan always at center of picture:
31598 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
31600 • Same as above but without pausing:
31602 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
31604 • Zoom in 2x into center of picture only for the first second of the
31606 input video:
31607 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
31609 zscale
31611 Scale (resize) the input video, using the z.lib library:
31612 <https://github.com/sekrit-twc/zimg>. To enable compilation of this
31613 filter, you need to configure FFmpeg with "--enable-libzimg".
31614 The zscale filter forces the output display aspect ratio to be the same
31616 as the input, by changing the output sample aspect ratio.
31617 If the input image format is different from the format requested by the
31619 next filter, the zscale filter will convert the input to the requested
31620 format.
31621 Options
31623 The filter accepts the following options.
31625 width, w
31627 height, h
31628 Set the output video dimension expression. Default value is the
31629 input dimension.
31630 If the width or w value is 0, the input width is used for the
31632 output. If the height or h value is 0, the input height is used for
31633 the output.
31634 If one and only one of the values is -n with n >= 1, the zscale
31636 filter will use a value that maintains the aspect ratio of the
31637 input image, calculated from the other specified dimension. After
31638 that it will, however, make sure that the calculated dimension is
31639 divisible by n and adjust the value if necessary.
31640 If both values are -n with n >= 1, the behavior will be identical
31642 to both values being set to 0 as previously detailed.
31643 See below for the list of accepted constants for use in the
31645 dimension expression.
31646 size, s
31648 Set the video size. For the syntax of this option, check the "Video
31649 size" section in the ffmpeg-utils manual.
31650 dither, d
31652 Set the dither type.
31653 Possible values are:
31655 none
31657 ordered
31658 random
31659 error_diffusion
31660 Default is none.
31662 filter, f
31664 Set the resize filter type.
31665 Possible values are:
31667 point
31669 bilinear
31670 bicubic
31671 spline16
31672 spline36
31673 lanczos
31674 Default is bilinear.
31676 range, r
31678 Set the color range.
31679 Possible values are:
31681 input
31683 limited
31684 full
31685 Default is same as input.
31687 primaries, p
31689 Set the color primaries.
31690 Possible values are:
31692 input
31694 709
31695 unspecified
31696 170m
31697 240m
31698 2020
31699 Default is same as input.
31701 transfer, t
31703 Set the transfer characteristics.
31704 Possible values are:
31706 input
31708 709
31709 unspecified
31710 601
31711 linear
31712 2020_10
31713 2020_12
31714 smpte2084
31715 iec61966-2-1
31716 arib-std-b67
31717 Default is same as input.
31719 matrix, m
31721 Set the colorspace matrix.
31722 Possible value are:
31724 input
31726 709
31727 unspecified
31728 470bg
31729 170m
31730 2020_ncl
31731 2020_cl
31732 Default is same as input.
31734 rangein, rin
31736 Set the input color range.
31737 Possible values are:
31739 input
31741 limited
31742 full
31743 Default is same as input.
31745 primariesin, pin
31747 Set the input color primaries.
31748 Possible values are:
31750 input
31752 709
31753 unspecified
31754 170m
31755 240m
31756 2020
31757 Default is same as input.
31759 transferin, tin
31761 Set the input transfer characteristics.
31762 Possible values are:
31764 input
31766 709
31767 unspecified
31768 601
31769 linear
31770 2020_10
31771 2020_12
31772 Default is same as input.
31774 matrixin, min
31776 Set the input colorspace matrix.
31777 Possible value are:
31779 input
31781 709
31782 unspecified
31783 470bg
31784 170m
31785 2020_ncl
31786 2020_cl
31787 chromal, c
31788 Set the output chroma location.
31789 Possible values are:
31791 input
31793 left
31794 center
31795 topleft
31796 top
31797 bottomleft
31798 bottom
31799 chromalin, cin
31800 Set the input chroma location.
31801 Possible values are:
31803 input
31805 left
31806 center
31807 topleft
31808 top
31809 bottomleft
31810 bottom
31811 npl Set the nominal peak luminance.
31812 param_a
31814 Parameter A for scaling filters. Parameter "b" for bicubic, and the
31815 number of filter taps for lanczos.
31816 param_b
31818 Parameter B for scaling filters. Parameter "c" for bicubic.
31819 The values of the w and h options are expressions containing the
31821 following constants:
31822 in_w
31824 in_h
31825 The input width and height
31826 iw
31828 ih These are the same as in_w and in_h.
31829 out_w
31831 out_h
31832 The output (scaled) width and height
31833 ow
31835 oh These are the same as out_w and out_h
31836 a The same as iw / ih
31838 sar input sample aspect ratio
31840 dar The input display aspect ratio. Calculated from "(iw / ih) * sar".
31842 hsub
31844 vsub
31845 horizontal and vertical input chroma subsample values. For example
31846 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
31847 ohsub
31849 ovsub
31850 horizontal and vertical output chroma subsample values. For example
31851 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
31852 Commands
31854 This filter supports the following commands:
31856 width, w
31858 height, h
31859 Set the output video dimension expression. The command accepts the
31860 same syntax of the corresponding option.
31861 If the specified expression is not valid, it is kept at its current
31863 value.
31864
31866 Below is a description of the currently available OpenCL video filters.
31867 To enable compilation of these filters you need to configure FFmpeg
31869 with "--enable-opencl".
31870 Running OpenCL filters requires you to initialize a hardware device and
31872 to pass that device to all filters in any filter graph.
31873 -init_hw_device opencl[=name][:device[,key=value...]]
31875 Initialise a new hardware device of type opencl called name, using
31876 the given device parameters.
31877 -filter_hw_device name
31879 Pass the hardware device called name to all filters in any filter
31880 graph.
31881 For more detailed information see
31883 <https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options>
31884 • Example of choosing the first device on the second platform and
31886 running avgblur_opencl filter with default parameters on it.
31887 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
31889 Since OpenCL filters are not able to access frame data in normal
31891 memory, all frame data needs to be uploaded(hwupload) to hardware
31892 surfaces connected to the appropriate device before being used and then
31893 downloaded(hwdownload) back to normal memory. Note that hwupload will
31894 upload to a surface with the same layout as the software frame, so it
31895 may be necessary to add a format filter immediately before to get the
31896 input into the right format and hwdownload does not support all formats
31897 on the output - it may be necessary to insert an additional format
31898 filter immediately following in the graph to get the output in a
31899 supported format.
31900 avgblur_opencl
31902 Apply average blur filter.
31903 The filter accepts the following options:
31905 sizeX
31907 Set horizontal radius size. Range is "[1, 1024]" and default value
31908 is 1.
31909 planes
31911 Set which planes to filter. Default value is 0xf, by which all
31912 planes are processed.
31913 sizeY
31915 Set vertical radius size. Range is "[1, 1024]" and default value is
31916 0. If zero, "sizeX" value will be used.
31917 Example
31919 • Apply average blur filter with horizontal and vertical size of 3,
31921 setting each pixel of the output to the average value of the 7x7
31922 region centered on it in the input. For pixels on the edges of the
31923 image, the region does not extend beyond the image boundaries, and
31924 so out-of-range coordinates are not used in the calculations.
31925 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
31927 boxblur_opencl
31929 Apply a boxblur algorithm to the input video.
31930 It accepts the following parameters:
31932 luma_radius, lr
31934 luma_power, lp
31935 chroma_radius, cr
31936 chroma_power, cp
31937 alpha_radius, ar
31938 alpha_power, ap
31939 A description of the accepted options follows.
31941 luma_radius, lr
31943 chroma_radius, cr
31944 alpha_radius, ar
31945 Set an expression for the box radius in pixels used for blurring
31946 the corresponding input plane.
31947 The radius value must be a non-negative number, and must not be
31949 greater than the value of the expression "min(w,h)/2" for the luma
31950 and alpha planes, and of "min(cw,ch)/2" for the chroma planes.
31951 Default value for luma_radius is "2". If not specified,
31953 chroma_radius and alpha_radius default to the corresponding value
31954 set for luma_radius.
31955 The expressions can contain the following constants:
31957 w
31959 h The input width and height in pixels.
31960 cw
31962 ch The input chroma image width and height in pixels.
31963 hsub
31965 vsub
31966 The horizontal and vertical chroma subsample values. For
31967 example, for the pixel format "yuv422p", hsub is 2 and vsub is
31968 1.
31969 luma_power, lp
31971 chroma_power, cp
31972 alpha_power, ap
31973 Specify how many times the boxblur filter is applied to the
31974 corresponding plane.
31975 Default value for luma_power is 2. If not specified, chroma_power
31977 and alpha_power default to the corresponding value set for
31978 luma_power.
31979 A value of 0 will disable the effect.
31981 Examples
31983 Apply boxblur filter, setting each pixel of the output to the average
31985 value of box-radiuses luma_radius, chroma_radius, alpha_radius for each
31986 plane respectively. The filter will apply luma_power, chroma_power,
31987 alpha_power times onto the corresponding plane. For pixels on the edges
31988 of the image, the radius does not extend beyond the image boundaries,
31989 and so out-of-range coordinates are not used in the calculations.
31990 • Apply a boxblur filter with the luma, chroma, and alpha radius set
31992 to 2 and luma, chroma, and alpha power set to 3. The filter will
31993 run 3 times with box-radius set to 2 for every plane of the image.
31994 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
31996 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
31997 • Apply a boxblur filter with luma radius set to 2, luma_power to 1,
31999 chroma_radius to 4, chroma_power to 5, alpha_radius to 3 and
32000 alpha_power to 7.
32001 For the luma plane, a 2x2 box radius will be run once.
32003 For the chroma plane, a 4x4 box radius will be run 5 times.
32005 For the alpha plane, a 3x3 box radius will be run 7 times.
32007 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
32009 colorkey_opencl
32011 RGB colorspace color keying.
32012 The filter accepts the following options:
32014 color
32016 The color which will be replaced with transparency.
32017 similarity
32019 Similarity percentage with the key color.
32020 0.01 matches only the exact key color, while 1.0 matches
32022 everything.
32023 blend
32025 Blend percentage.
32026 0.0 makes pixels either fully transparent, or not transparent at
32028 all.
32029 Higher values result in semi-transparent pixels, with a higher
32031 transparency the more similar the pixels color is to the key color.
32032 Examples
32034 • Make every semi-green pixel in the input transparent with some
32036 slight blending:
32037 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
32039 convolution_opencl
32041 Apply convolution of 3x3, 5x5, 7x7 matrix.
32042 The filter accepts the following options:
32044 0m
32046 1m
32047 2m
32048 3m Set matrix for each plane. Matrix is sequence of 9, 25 or 49
32049 signed numbers. Default value for each plane is "0 0 0 0 1 0 0 0
32050 0".
32051 0rdiv
32053 1rdiv
32054 2rdiv
32055 3rdiv
32056 Set multiplier for calculated value for each plane. If unset or 0,
32057 it will be sum of all matrix elements. The option value must be a
32058 float number greater or equal to 0.0. Default value is 1.0.
32059 0bias
32061 1bias
32062 2bias
32063 3bias
32064 Set bias for each plane. This value is added to the result of the
32065 multiplication. Useful for making the overall image brighter or
32066 darker. The option value must be a float number greater or equal
32067 to 0.0. Default value is 0.0.
32068 Examples
32070 • Apply sharpen:
32072 -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
32074 • Apply blur:
32076 -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
32078 • Apply edge enhance:
32080 -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
32082 • Apply edge detect:
32084 -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
32086 • Apply laplacian edge detector which includes diagonals:
32088 -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
32090 • Apply emboss:
32092 -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
32094 erosion_opencl
32096 Apply erosion effect to the video.
32097 This filter replaces the pixel by the local(3x3) minimum.
32099 It accepts the following options:
32101 threshold0
32103 threshold1
32104 threshold2
32105 threshold3
32106 Limit the maximum change for each plane. Range is "[0, 65535]" and
32107 default value is 65535. If 0, plane will remain unchanged.
32108 coordinates
32110 Flag which specifies the pixel to refer to. Range is "[0, 255]"
32111 and default value is 255, i.e. all eight pixels are used.
32112 Flags to local 3x3 coordinates region centered on "x":
32114 1 2 3
32116 4 x 5
32118 6 7 8
32120 Example
32122 • Apply erosion filter with threshold0 set to 30, threshold1 set 40,
32124 threshold2 set to 50 and coordinates set to 231, setting each pixel
32125 of the output to the local minimum between pixels: 1, 2, 3, 6, 7, 8
32126 of the 3x3 region centered on it in the input. If the difference
32127 between input pixel and local minimum is more then threshold of the
32128 corresponding plane, output pixel will be set to input pixel -
32129 threshold of corresponding plane.
32130 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
32132 deshake_opencl
32134 Feature-point based video stabilization filter.
32135 The filter accepts the following options:
32137 tripod
32139 Simulates a tripod by preventing any camera movement whatsoever
32140 from the original frame. Defaults to 0.
32141 debug
32143 Whether or not additional debug info should be displayed, both in
32144 the processed output and in the console.
32145 Note that in order to see console debug output you will also need
32147 to pass "-v verbose" to ffmpeg.
32148 Viewing point matches in the output video is only supported for RGB
32150 input.
32151 Defaults to 0.
32153 adaptive_crop
32155 Whether or not to do a tiny bit of cropping at the borders to cut
32156 down on the amount of mirrored pixels.
32157 Defaults to 1.
32159 refine_features
32161 Whether or not feature points should be refined at a sub-pixel
32162 level.
32163 This can be turned off for a slight performance gain at the cost of
32165 precision.
32166 Defaults to 1.
32168 smooth_strength
32170 The strength of the smoothing applied to the camera path from 0.0
32171 to 1.0.
32172 1.0 is the maximum smoothing strength while values less than that
32174 result in less smoothing.
32175 0.0 causes the filter to adaptively choose a smoothing strength on
32177 a per-frame basis.
32178 Defaults to 0.0.
32180 smooth_window_multiplier
32182 Controls the size of the smoothing window (the number of frames
32183 buffered to determine motion information from).
32184 The size of the smoothing window is determined by multiplying the
32186 framerate of the video by this number.
32187 Acceptable values range from 0.1 to 10.0.
32189 Larger values increase the amount of motion data available for
32191 determining how to smooth the camera path, potentially improving
32192 smoothness, but also increase latency and memory usage.
32193 Defaults to 2.0.
32195 Examples
32197 • Stabilize a video with a fixed, medium smoothing strength:
32199 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
32201 • Stabilize a video with debugging (both in console and in rendered
32203 video):
32204 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
32206 dilation_opencl
32208 Apply dilation effect to the video.
32209 This filter replaces the pixel by the local(3x3) maximum.
32211 It accepts the following options:
32213 threshold0
32215 threshold1
32216 threshold2
32217 threshold3
32218 Limit the maximum change for each plane. Range is "[0, 65535]" and
32219 default value is 65535. If 0, plane will remain unchanged.
32220 coordinates
32222 Flag which specifies the pixel to refer to. Range is "[0, 255]"
32223 and default value is 255, i.e. all eight pixels are used.
32224 Flags to local 3x3 coordinates region centered on "x":
32226 1 2 3
32228 4 x 5
32230 6 7 8
32232 Example
32234 • Apply dilation filter with threshold0 set to 30, threshold1 set 40,
32236 threshold2 set to 50 and coordinates set to 231, setting each pixel
32237 of the output to the local maximum between pixels: 1, 2, 3, 6, 7, 8
32238 of the 3x3 region centered on it in the input. If the difference
32239 between input pixel and local maximum is more then threshold of the
32240 corresponding plane, output pixel will be set to input pixel +
32241 threshold of corresponding plane.
32242 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
32244 nlmeans_opencl
32246 Non-local Means denoise filter through OpenCL, this filter accepts same
32247 options as nlmeans.
32248 overlay_opencl
32250 Overlay one video on top of another.
32251 It takes two inputs and has one output. The first input is the "main"
32253 video on which the second input is overlaid. This filter requires same
32254 memory layout for all the inputs. So, format conversion may be needed.
32255 The filter accepts the following options:
32257 x Set the x coordinate of the overlaid video on the main video.
32259 Default value is 0.
32260 y Set the y coordinate of the overlaid video on the main video.
32262 Default value is 0.
32263 Examples
32265 • Overlay an image LOGO at the top-left corner of the INPUT video.
32267 Both inputs are yuv420p format.
32268 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
32270 • The inputs have same memory layout for color channels , the overlay
32272 has additional alpha plane, like INPUT is yuv420p, and the LOGO is
32273 yuva420p.
32274 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
32276 pad_opencl
32278 Add paddings to the input image, and place the original input at the
32279 provided x, y coordinates.
32280 It accepts the following options:
32282 width, w
32284 height, h
32285 Specify an expression for the size of the output image with the
32286 paddings added. If the value for width or height is 0, the
32287 corresponding input size is used for the output.
32288 The width expression can reference the value set by the height
32290 expression, and vice versa.
32291 The default value of width and height is 0.
32293 x
32295 y Specify the offsets to place the input image at within the padded
32296 area, with respect to the top/left border of the output image.
32297 The x expression can reference the value set by the y expression,
32299 and vice versa.
32300 The default value of x and y is 0.
32302 If x or y evaluate to a negative number, they'll be changed so the
32304 input image is centered on the padded area.
32305 color
32307 Specify the color of the padded area. For the syntax of this
32308 option, check the "Color" section in the ffmpeg-utils manual.
32309 aspect
32311 Pad to an aspect instead to a resolution.
32312 The value for the width, height, x, and y options are expressions
32314 containing the following constants:
32315 in_w
32317 in_h
32318 The input video width and height.
32319 iw
32321 ih These are the same as in_w and in_h.
32322 out_w
32324 out_h
32325 The output width and height (the size of the padded area), as
32326 specified by the width and height expressions.
32327 ow
32329 oh These are the same as out_w and out_h.
32330 x
32332 y The x and y offsets as specified by the x and y expressions, or NAN
32333 if not yet specified.
32334 a same as iw / ih
32336 sar input sample aspect ratio
32338 dar input display aspect ratio, it is the same as (iw / ih) * sar
32340 prewitt_opencl
32342 Apply the Prewitt operator
32343 (<https://en.wikipedia.org/wiki/Prewitt_operator>) to input video
32344 stream.
32345 The filter accepts the following option:
32347 planes
32349 Set which planes to filter. Default value is 0xf, by which all
32350 planes are processed.
32351 scale
32353 Set value which will be multiplied with filtered result. Range is
32354 "[0.0, 65535]" and default value is 1.0.
32355 delta
32357 Set value which will be added to filtered result. Range is
32358 "[-65535, 65535]" and default value is 0.0.
32359 Example
32361 • Apply the Prewitt operator with scale set to 2 and delta set to 10.
32363 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
32365 program_opencl
32367 Filter video using an OpenCL program.
32368 source
32370 OpenCL program source file.
32371 kernel
32373 Kernel name in program.
32374 inputs
32376 Number of inputs to the filter. Defaults to 1.
32377 size, s
32379 Size of output frames. Defaults to the same as the first input.
32380 The "program_opencl" filter also supports the framesync options.
32382 The program source file must contain a kernel function with the given
32384 name, which will be run once for each plane of the output. Each run on
32385 a plane gets enqueued as a separate 2D global NDRange with one work-
32386 item for each pixel to be generated. The global ID offset for each
32387 work-item is therefore the coordinates of a pixel in the destination
32388 image.
32389 The kernel function needs to take the following arguments:
32391 • Destination image, __write_only image2d_t.
32393 This image will become the output; the kernel should write all of
32395 it.
32396 • Frame index, unsigned int.
32398 This is a counter starting from zero and increasing by one for each
32400 frame.
32401 • Source images, __read_only image2d_t.
32403 These are the most recent images on each input. The kernel may
32405 read from them to generate the output, but they can't be written
32406 to.
32407 Example programs:
32409 • Copy the input to the output (output must be the same size as the
32411 input).
32412 __kernel void copy(__write_only image2d_t destination,
32414 unsigned int index,
32415 __read_only image2d_t source)
32416 {
32417 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
32418 int2 location = (int2)(get_global_id(0), get_global_id(1));
32420 float4 value = read_imagef(source, sampler, location);
32422 write_imagef(destination, location, value);
32424 }
32425 • Apply a simple transformation, rotating the input by an amount
32427 increasing with the index counter. Pixel values are linearly
32428 interpolated by the sampler, and the output need not have the same
32429 dimensions as the input.
32430 __kernel void rotate_image(__write_only image2d_t dst,
32432 unsigned int index,
32433 __read_only image2d_t src)
32434 {
32435 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
32436 CLK_FILTER_LINEAR);
32437 float angle = (float)index / 100.0f;
32439 float2 dst_dim = convert_float2(get_image_dim(dst));
32441 float2 src_dim = convert_float2(get_image_dim(src));
32442 float2 dst_cen = dst_dim / 2.0f;
32444 float2 src_cen = src_dim / 2.0f;
32445 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
32447 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
32449 float2 src_pos = {
32450 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
32451 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
32452 };
32453 src_pos = src_pos * src_dim / dst_dim;
32454 float2 src_loc = src_pos + src_cen;
32456 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
32458 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
32459 write_imagef(dst, dst_loc, 0.5f);
32460 else
32461 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
32462 }
32463 • Blend two inputs together, with the amount of each input used
32465 varying with the index counter.
32466 __kernel void blend_images(__write_only image2d_t dst,
32468 unsigned int index,
32469 __read_only image2d_t src1,
32470 __read_only image2d_t src2)
32471 {
32472 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
32473 CLK_FILTER_LINEAR);
32474 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
32476 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
32478 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
32479 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
32480 float4 val1 = read_imagef(src1, sampler, src1_loc);
32482 float4 val2 = read_imagef(src2, sampler, src2_loc);
32483 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
32485 }
32486 remap_opencl
32488 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
32489 Destination pixel at position (X, Y) will be picked from source (x, y)
32491 position where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are
32492 out of range, zero value for pixel will be used for destination pixel.
32493 Xmap and Ymap input video streams must be of same dimensions. Output
32495 video stream will have Xmap/Ymap video stream dimensions. Xmap and
32496 Ymap input video streams are 32bit float pixel format, single channel.
32497 interp
32499 Specify interpolation used for remapping of pixels. Allowed values
32500 are "near" and "linear". Default value is "linear".
32501 fill
32503 Specify the color of the unmapped pixels. For the syntax of this
32504 option, check the "Color" section in the ffmpeg-utils manual.
32505 Default color is "black".
32506 roberts_opencl
32508 Apply the Roberts cross operator
32509 (<https://en.wikipedia.org/wiki/Roberts_cross>) to input video stream.
32510 The filter accepts the following option:
32512 planes
32514 Set which planes to filter. Default value is 0xf, by which all
32515 planes are processed.
32516 scale
32518 Set value which will be multiplied with filtered result. Range is
32519 "[0.0, 65535]" and default value is 1.0.
32520 delta
32522 Set value which will be added to filtered result. Range is
32523 "[-65535, 65535]" and default value is 0.0.
32524 Example
32526 • Apply the Roberts cross operator with scale set to 2 and delta set
32528 to 10
32529 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
32531 sobel_opencl
32533 Apply the Sobel operator
32534 (<https://en.wikipedia.org/wiki/Sobel_operator>) to input video stream.
32535 The filter accepts the following option:
32537 planes
32539 Set which planes to filter. Default value is 0xf, by which all
32540 planes are processed.
32541 scale
32543 Set value which will be multiplied with filtered result. Range is
32544 "[0.0, 65535]" and default value is 1.0.
32545 delta
32547 Set value which will be added to filtered result. Range is
32548 "[-65535, 65535]" and default value is 0.0.
32549 Example
32551 • Apply sobel operator with scale set to 2 and delta set to 10
32553 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
32555 tonemap_opencl
32557 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
32558 It accepts the following parameters:
32560 tonemap
32562 Specify the tone-mapping operator to be used. Same as tonemap
32563 option in tonemap.
32564 param
32566 Tune the tone mapping algorithm. same as param option in tonemap.
32567 desat
32569 Apply desaturation for highlights that exceed this level of
32570 brightness. The higher the parameter, the more color information
32571 will be preserved. This setting helps prevent unnaturally blown-out
32572 colors for super-highlights, by (smoothly) turning into white
32573 instead. This makes images feel more natural, at the cost of
32574 reducing information about out-of-range colors.
32575 The default value is 0.5, and the algorithm here is a little
32577 different from the cpu version tonemap currently. A setting of 0.0
32578 disables this option.
32579 threshold
32581 The tonemapping algorithm parameters is fine-tuned per each scene.
32582 And a threshold is used to detect whether the scene has changed or
32583 not. If the distance between the current frame average brightness
32584 and the current running average exceeds a threshold value, we would
32585 re-calculate scene average and peak brightness. The default value
32586 is 0.2.
32587 format
32589 Specify the output pixel format.
32590 Currently supported formats are:
32592 p010
32594 nv12
32595 range, r
32596 Set the output color range.
32597 Possible values are:
32599 tv/mpeg
32601 pc/jpeg
32602 Default is same as input.
32604 primaries, p
32606 Set the output color primaries.
32607 Possible values are:
32609 bt709
32611 bt2020
32612 Default is same as input.
32614 transfer, t
32616 Set the output transfer characteristics.
32617 Possible values are:
32619 bt709
32621 bt2020
32622 Default is bt709.
32624 matrix, m
32626 Set the output colorspace matrix.
32627 Possible value are:
32629 bt709
32631 bt2020
32632 Default is same as input.
32634 Example
32636 • Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010
32638 format using linear operator.
32639 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
32641 unsharp_opencl
32643 Sharpen or blur the input video.
32644 It accepts the following parameters:
32646 luma_msize_x, lx
32648 Set the luma matrix horizontal size. Range is "[1, 23]" and
32649 default value is 5.
32650 luma_msize_y, ly
32652 Set the luma matrix vertical size. Range is "[1, 23]" and default
32653 value is 5.
32654 luma_amount, la
32656 Set the luma effect strength. Range is "[-10, 10]" and default
32657 value is 1.0.
32658 Negative values will blur the input video, while positive values
32660 will sharpen it, a value of zero will disable the effect.
32661 chroma_msize_x, cx
32663 Set the chroma matrix horizontal size. Range is "[1, 23]" and
32664 default value is 5.
32665 chroma_msize_y, cy
32667 Set the chroma matrix vertical size. Range is "[1, 23]" and
32668 default value is 5.
32669 chroma_amount, ca
32671 Set the chroma effect strength. Range is "[-10, 10]" and default
32672 value is 0.0.
32673 Negative values will blur the input video, while positive values
32675 will sharpen it, a value of zero will disable the effect.
32676 All parameters are optional and default to the equivalent of the string
32678 '5:5:1.0:5:5:0.0'.
32679 Examples
32681 • Apply strong luma sharpen effect:
32683 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
32685 • Apply a strong blur of both luma and chroma parameters:
32687 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
32689 xfade_opencl
32691 Cross fade two videos with custom transition effect by using OpenCL.
32692 It accepts the following options:
32694 transition
32696 Set one of possible transition effects.
32697 custom
32699 Select custom transition effect, the actual transition
32700 description will be picked from source and kernel options.
32701 fade
32703 wipeleft
32704 wiperight
32705 wipeup
32706 wipedown
32707 slideleft
32708 slideright
32709 slideup
32710 slidedown
32711 Default transition is fade.
32712 source
32714 OpenCL program source file for custom transition.
32715 kernel
32717 Set name of kernel to use for custom transition from program source
32718 file.
32719 duration
32721 Set duration of video transition.
32722 offset
32724 Set time of start of transition relative to first video.
32725 The program source file must contain a kernel function with the given
32727 name, which will be run once for each plane of the output. Each run on
32728 a plane gets enqueued as a separate 2D global NDRange with one work-
32729 item for each pixel to be generated. The global ID offset for each
32730 work-item is therefore the coordinates of a pixel in the destination
32731 image.
32732 The kernel function needs to take the following arguments:
32734 • Destination image, __write_only image2d_t.
32736 This image will become the output; the kernel should write all of
32738 it.
32739 • First Source image, __read_only image2d_t. Second Source image,
32741 __read_only image2d_t.
32742 These are the most recent images on each input. The kernel may
32744 read from them to generate the output, but they can't be written
32745 to.
32746 • Transition progress, float. This value is always between 0 and 1
32748 inclusive.
32749 Example programs:
32751 • Apply dots curtain transition effect:
32753 __kernel void blend_images(__write_only image2d_t dst,
32755 __read_only image2d_t src1,
32756 __read_only image2d_t src2,
32757 float progress)
32758 {
32759 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
32760 CLK_FILTER_LINEAR);
32761 int2 p = (int2)(get_global_id(0), get_global_id(1));
32762 float2 rp = (float2)(get_global_id(0), get_global_id(1));
32763 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
32764 rp = rp / dim;
32765 float2 dots = (float2)(20.0, 20.0);
32767 float2 center = (float2)(0,0);
32768 float2 unused;
32769 float4 val1 = read_imagef(src1, sampler, p);
32771 float4 val2 = read_imagef(src2, sampler, p);
32772 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
32773 write_imagef(dst, p, next ? val1 : val2);
32775 }
32776
32778 VAAPI Video filters are usually used with VAAPI decoder and VAAPI
32779 encoder. Below is a description of VAAPI video filters.
32780 To enable compilation of these filters you need to configure FFmpeg
32782 with "--enable-vaapi".
32783 To use vaapi filters, you need to setup the vaapi device correctly. For
32785 more information, please read
32786 <https://trac.ffmpeg.org/wiki/Hardware/VAAPI>
32787 overlay_vaapi
32789 Overlay one video on the top of another.
32790 It takes two inputs and has one output. The first input is the "main"
32792 video on which the second input is overlaid.
32793 The filter accepts the following options:
32795 x
32797 y Set expressions for the x and y coordinates of the overlaid video
32798 on the main video.
32799 Default value is "0" for both expressions.
32801 w
32803 h Set expressions for the width and height the overlaid video on the
32804 main video.
32805 Default values are 'overlay_iw' for 'w' and
32807 'overlay_ih*w/overlay_iw' for 'h'.
32808 The expressions can contain the following parameters:
32810 main_w, W
32812 main_h, H
32813 The main input width and height.
32814 overlay_iw
32816 overlay_ih
32817 The overlay input width and height.
32818 overlay_w, w
32820 overlay_h, h
32821 The overlay output width and height.
32822 overlay_x, x
32824 overlay_y, y
32825 Position of the overlay layer inside of main
32826 alpha
32828 Set transparency of overlaid video. Allowed range is 0.0 to 1.0.
32829 Higher value means lower transparency. Default value is 1.0.
32830 eof_action
32832 See framesync.
32833 shortest
32835 See framesync.
32836 repeatlast
32838 See framesync.
32839 This filter also supports the framesync options.
32841 Examples
32843 • Overlay an image LOGO at the top-left corner of the INPUT video.
32845 Both inputs for this filter are yuv420p format.
32846 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_vaapi" OUTPUT
32848 • Overlay an image LOGO at the offset (200, 100) from the top-left
32850 corner of the INPUT video. The inputs have same memory layout for
32851 color channels, the overlay has additional alpha plane, like INPUT
32852 is yuv420p, and the LOGO is yuva420p.
32853 -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
32855 tonemap_vaapi
32857 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range)
32858 conversion with tone-mapping. It maps the dynamic range of HDR10
32859 content to the SDR content. It currently only accepts HDR10 as input.
32860 It accepts the following parameters:
32862 format
32864 Specify the output pixel format.
32865 Currently supported formats are:
32867 p010
32869 nv12
32870 Default is nv12.
32872 primaries, p
32874 Set the output color primaries.
32875 Default is same as input.
32877 transfer, t
32879 Set the output transfer characteristics.
32880 Default is bt709.
32882 matrix, m
32884 Set the output colorspace matrix.
32885 Default is same as input.
32887 Example
32889 • Convert HDR(HDR10) video to bt2020-transfer-characteristic p010
32891 format
32892 tonemap_vaapi=format=p010:t=bt2020-10
32894 hstack_vaapi
32896 Stack input videos horizontally.
32897 This is the VA-API variant of the hstack filter, each input stream may
32899 have different height, this filter will scale down/up each input stream
32900 while keeping the orignal aspect.
32901 It accepts the following options:
32903 inputs
32905 See hstack.
32906 shortest
32908 See hstack.
32909 height
32911 Set height of output. If set to 0, this filter will set height of
32912 output to height of the first input stream. Default value is 0.
32913 vstack_vaapi
32915 Stack input videos vertically.
32916 This is the VA-API variant of the vstack filter, each input stream may
32918 have different width, this filter will scale down/up each input stream
32919 while keeping the orignal aspect.
32920 It accepts the following options:
32922 inputs
32924 See vstack.
32925 shortest
32927 See vstack.
32928 width
32930 Set width of output. If set to 0, this filter will set width of
32931 output to width of the first input stream. Default value is 0.
32932 xstack_vaapi
32934 Stack video inputs into custom layout.
32935 This is the VA-API variant of the xstack filter, each input stream may
32937 have different size, this filter will scale down/up each input stream
32938 to the given output size, or the size of the first input stream.
32939 It accepts the following options:
32941 inputs
32943 See xstack.
32944 shortest
32946 See xstack.
32947 layout
32949 See xstack. Moreover, this permits the user to supply output size
32950 for each input stream.
32951 xstack_vaapi=inputs=4:layout=0_0_1920x1080|0_h0_1920x1080|w0_0_1920x1080|w0_h0_1920x1080
32953 grid
32955 See xstack.
32956 grid_tile_size
32958 Set output size for each input stream when grid is set. If this
32959 option is not set, this filter will set output size by default to
32960 the size of the first input stream. For the syntax of this option,
32961 check the "Video size" section in the ffmpeg-utils manual.
32962 fill
32964 See xstack.
32965
32967 Below is a description of the currently available QSV video filters.
32968 To enable compilation of these filters you need to configure FFmpeg
32970 with "--enable-libmfx" or "--enable-libvpl".
32971 To use QSV filters, you need to setup the QSV device correctly. For
32973 more information, please read
32974 <https://trac.ffmpeg.org/wiki/Hardware/QuickSync>
32975 hstack_qsv
32977 Stack input videos horizontally.
32978 This is the QSV variant of the hstack filter, each input stream may
32980 have different height, this filter will scale down/up each input stream
32981 while keeping the orignal aspect.
32982 It accepts the following options:
32984 inputs
32986 See hstack.
32987 shortest
32989 See hstack.
32990 height
32992 Set height of output. If set to 0, this filter will set height of
32993 output to height of the first input stream. Default value is 0.
32994 vstack_qsv
32996 Stack input videos vertically.
32997 This is the QSV variant of the vstack filter, each input stream may
32999 have different width, this filter will scale down/up each input stream
33000 while keeping the orignal aspect.
33001 It accepts the following options:
33003 inputs
33005 See vstack.
33006 shortest
33008 See vstack.
33009 width
33011 Set width of output. If set to 0, this filter will set width of
33012 output to width of the first input stream. Default value is 0.
33013 xstack_qsv
33015 Stack video inputs into custom layout.
33016 This is the QSV variant of the xstack filter.
33018 It accepts the following options:
33020 inputs
33022 See xstack.
33023 shortest
33025 See xstack.
33026 layout
33028 See xstack. Moreover, this permits the user to supply output size
33029 for each input stream.
33030 xstack_qsv=inputs=4:layout=0_0_1920x1080|0_h0_1920x1080|w0_0_1920x1080|w0_h0_1920x1080
33032 grid
33034 See xstack.
33035 grid_tile_size
33037 Set output size for each input stream when grid is set. If this
33038 option is not set, this filter will set output size by default to
33039 the size of the first input stream. For the syntax of this option,
33040 check the "Video size" section in the ffmpeg-utils manual.
33041 fill
33043 See xstack.
33044
33046 Below is a description of the currently available video sources.
33047 buffer
33049 Buffer video frames, and make them available to the filter chain.
33050 This source is mainly intended for a programmatic use, in particular
33052 through the interface defined in libavfilter/buffersrc.h.
33053 It accepts the following parameters:
33055 video_size
33057 Specify the size (width and height) of the buffered video frames.
33058 For the syntax of this option, check the "Video size" section in
33059 the ffmpeg-utils manual.
33060 width
33062 The input video width.
33063 height
33065 The input video height.
33066 pix_fmt
33068 A string representing the pixel format of the buffered video
33069 frames. It may be a number corresponding to a pixel format, or a
33070 pixel format name.
33071 time_base
33073 Specify the timebase assumed by the timestamps of the buffered
33074 frames.
33075 frame_rate
33077 Specify the frame rate expected for the video stream.
33078 pixel_aspect, sar
33080 The sample (pixel) aspect ratio of the input video.
33081 hw_frames_ctx
33083 When using a hardware pixel format, this should be a reference to
33084 an AVHWFramesContext describing input frames.
33085 For example:
33087 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
33089 will instruct the source to accept video frames with size 320x240 and
33091 with format "yuv410p", assuming 1/24 as the timestamps timebase and
33092 square pixels (1:1 sample aspect ratio). Since the pixel format with
33093 name "yuv410p" corresponds to the number 6 (check the enum
33094 AVPixelFormat definition in libavutil/pixfmt.h), this example
33095 corresponds to:
33096 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
33098 Alternatively, the options can be specified as a flat string, but this
33100 syntax is deprecated:
33101 width:height:pix_fmt:time_base.num:time_base.den:pixel_aspect.num:pixel_aspect.den
33103 cellauto
33105 Create a pattern generated by an elementary cellular automaton.
33106 The initial state of the cellular automaton can be defined through the
33108 filename and pattern options. If such options are not specified an
33109 initial state is created randomly.
33110 At each new frame a new row in the video is filled with the result of
33112 the cellular automaton next generation. The behavior when the whole
33113 frame is filled is defined by the scroll option.
33114 This source accepts the following options:
33116 filename, f
33118 Read the initial cellular automaton state, i.e. the starting row,
33119 from the specified file. In the file, each non-whitespace
33120 character is considered an alive cell, a newline will terminate the
33121 row, and further characters in the file will be ignored.
33122 pattern, p
33124 Read the initial cellular automaton state, i.e. the starting row,
33125 from the specified string.
33126 Each non-whitespace character in the string is considered an alive
33128 cell, a newline will terminate the row, and further characters in
33129 the string will be ignored.
33130 rate, r
33132 Set the video rate, that is the number of frames generated per
33133 second. Default is 25.
33134 random_fill_ratio, ratio
33136 Set the random fill ratio for the initial cellular automaton row.
33137 It is a floating point number value ranging from 0 to 1, defaults
33138 to 1/PHI.
33139 This option is ignored when a file or a pattern is specified.
33141 random_seed, seed
33143 Set the seed for filling randomly the initial row, must be an
33144 integer included between 0 and UINT32_MAX. If not specified, or if
33145 explicitly set to -1, the filter will try to use a good random seed
33146 on a best effort basis.
33147 rule
33149 Set the cellular automaton rule, it is a number ranging from 0 to
33150 255. Default value is 110.
33151 size, s
33153 Set the size of the output video. For the syntax of this option,
33154 check the "Video size" section in the ffmpeg-utils manual.
33155 If filename or pattern is specified, the size is set by default to
33157 the width of the specified initial state row, and the height is set
33158 to width * PHI.
33159 If size is set, it must contain the width of the specified pattern
33161 string, and the specified pattern will be centered in the larger
33162 row.
33163 If a filename or a pattern string is not specified, the size value
33165 defaults to "320x518" (used for a randomly generated initial
33166 state).
33167 scroll
33169 If set to 1, scroll the output upward when all the rows in the
33170 output have been already filled. If set to 0, the new generated row
33171 will be written over the top row just after the bottom row is
33172 filled. Defaults to 1.
33173 start_full, full
33175 If set to 1, completely fill the output with generated rows before
33176 outputting the first frame. This is the default behavior, for
33177 disabling set the value to 0.
33178 stitch
33180 If set to 1, stitch the left and right row edges together. This is
33181 the default behavior, for disabling set the value to 0.
33182 Examples
33184 • Read the initial state from pattern, and specify an output of size
33186 200x400.
33187 cellauto=f=pattern:s=200x400
33189 • Generate a random initial row with a width of 200 cells, with a
33191 fill ratio of 2/3:
33192 cellauto=ratio=2/3:s=200x200
33194 • Create a pattern generated by rule 18 starting by a single alive
33196 cell centered on an initial row with width 100:
33197 cellauto=p=@s=100x400:full=0:rule=18
33199 • Specify a more elaborated initial pattern:
33201 cellauto=p='@@ @ @@':s=100x400:full=0:rule=18
33203 coreimagesrc
33205 Video source generated on GPU using Apple's CoreImage API on OSX.
33206 This video source is a specialized version of the coreimage video
33208 filter. Use a core image generator at the beginning of the applied
33209 filterchain to generate the content.
33210 The coreimagesrc video source accepts the following options:
33212 list_generators
33214 List all available generators along with all their respective
33215 options as well as possible minimum and maximum values along with
33216 the default values.
33217 list_generators=true
33219 size, s
33221 Specify the size of the sourced video. For the syntax of this
33222 option, check the "Video size" section in the ffmpeg-utils manual.
33223 The default value is "320x240".
33224 rate, r
33226 Specify the frame rate of the sourced video, as the number of
33227 frames generated per second. It has to be a string in the format
33228 frame_rate_num/frame_rate_den, an integer number, a floating point
33229 number or a valid video frame rate abbreviation. The default value
33230 is "25".
33231 sar Set the sample aspect ratio of the sourced video.
33233 duration, d
33235 Set the duration of the sourced video. See the Time duration
33236 section in the ffmpeg-utils(1) manual for the accepted syntax.
33237 If not specified, or the expressed duration is negative, the video
33239 is supposed to be generated forever.
33240 Additionally, all options of the coreimage video filter are accepted.
33242 A complete filterchain can be used for further processing of the
33243 generated input without CPU-HOST transfer. See coreimage documentation
33244 and examples for details.
33245 Examples
33247 • Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
33249 given as complete and escaped command-line for Apple's standard
33250 bash shell:
33251 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@inputMessage=https\\\\\://FFmpeg.org/@inputCorrectionLevel=H -frames:v 1 QRCode.png
33253 This example is equivalent to the QRCode example of coreimage
33255 without the need for a nullsrc video source.
33256 ddagrab
33258 Captures the Windows Desktop via Desktop Duplication API.
33259 The filter exclusively returns D3D11 Hardware Frames, for on-gpu
33261 encoding or processing. So an explicit hwdownload is needed for any
33262 kind of software processing.
33263 It accepts the following options:
33265 output_idx
33267 DXGI Output Index to capture.
33268 Usually corresponds to the index Windows has given the screen minus
33270 one, so it's starting at 0.
33271 Defaults to output 0.
33273 draw_mouse
33275 Whether to draw the mouse cursor.
33276 Defaults to true.
33278 Only affects hardware cursors. If a game or application renders its
33280 own cursor, it'll always be captured.
33281 framerate
33283 Framerate at which the desktop will be captured.
33284 Defaults to 30 FPS.
33286 video_size
33288 Specify the size of the captured video.
33289 Defaults to the full size of the screen.
33291 Cropped from the bottom/right if smaller than screen size.
33293 offset_x
33295 Horizontal offset of the captured video.
33296 offset_y
33298 Vertical offset of the captured video.
33299 output_fmt
33301 Desired filter output format. Defaults to 8 Bit BGRA.
33302 It accepts the following values:
33304 auto
33306 Passes all supported output formats to DDA and returns what DDA
33307 decides to use.
33308 8bit
33310 bgra
33311 8 Bit formats always work, and DDA will convert to them if
33312 neccesary.
33313 10bit
33315 x2bgr10
33316 Filter initialization will fail if 10 bit format is requested
33317 but unavailable.
33318 Examples
33320 Capture primary screen and encode using nvenc:
33322 ffmpeg -f lavfi -i ddagrab -c:v h264_nvenc -cq 18 output.mp4
33324 You can also skip the lavfi device and directly use the filter. Also
33326 demonstrates downloading the frame and encoding with libx264. Explicit
33327 output format specification is required in this case:
33328 ffmpeg -filter_complex ddagrab=output_idx=1:framerate=60,hwdownload,format=bgra -c:v libx264 -crf 18 output.mp4
33330 If you want to capture only a subsection of the desktop, this can be
33332 achieved by specifying a smaller size and its offsets into the screen:
33333 ddagrab=video_size=800x600:offset_x=100:offset_y=100
33335 gradients
33337 Generate several gradients.
33338 size, s
33340 Set frame size. For the syntax of this option, check the "Video
33341 size" section in the ffmpeg-utils manual. Default value is
33342 "640x480".
33343 rate, r
33345 Set frame rate, expressed as number of frames per second. Default
33346 value is "25".
33347 c0, c1, c2, c3, c4, c5, c6, c7
33349 Set 8 colors. Default values for colors is to pick random one.
33350 x0, y0, y0, y1
33352 Set gradient line source and destination points. If negative or out
33353 of range, random ones are picked.
33354 nb_colors, n
33356 Set number of colors to use at once. Allowed range is from 2 to 8.
33357 Default value is 2.
33358 seed
33360 Set seed for picking gradient line points.
33361 duration, d
33363 Set the duration of the sourced video. See the Time duration
33364 section in the ffmpeg-utils(1) manual for the accepted syntax.
33365 If not specified, or the expressed duration is negative, the video
33367 is supposed to be generated forever.
33368 speed
33370 Set speed of gradients rotation.
33371 type, t
33373 Set type of gradients, can be "linear" or "radial" or "circular" or
33374 "spiral".
33375 mandelbrot
33377 Generate a Mandelbrot set fractal, and progressively zoom towards the
33378 point specified with start_x and start_y.
33379 This source accepts the following options:
33381 end_pts
33383 Set the terminal pts value. Default value is 400.
33384 end_scale
33386 Set the terminal scale value. Must be a floating point value.
33387 Default value is 0.3.
33388 inner
33390 Set the inner coloring mode, that is the algorithm used to draw the
33391 Mandelbrot fractal internal region.
33392 It shall assume one of the following values:
33394 black
33396 Set black mode.
33397 convergence
33399 Show time until convergence.
33400 mincol
33402 Set color based on point closest to the origin of the
33403 iterations.
33404 period
33406 Set period mode.
33407 Default value is mincol.
33409 bailout
33411 Set the bailout value. Default value is 10.0.
33412 maxiter
33414 Set the maximum of iterations performed by the rendering algorithm.
33415 Default value is 7189.
33416 outer
33418 Set outer coloring mode. It shall assume one of following values:
33419 iteration_count
33421 Set iteration count mode.
33422 normalized_iteration_count
33424 set normalized iteration count mode.
33425 Default value is normalized_iteration_count.
33427 rate, r
33429 Set frame rate, expressed as number of frames per second. Default
33430 value is "25".
33431 size, s
33433 Set frame size. For the syntax of this option, check the "Video
33434 size" section in the ffmpeg-utils manual. Default value is
33435 "640x480".
33436 start_scale
33438 Set the initial scale value. Default value is 3.0.
33439 start_x
33441 Set the initial x position. Must be a floating point value between
33442 -100 and 100. Default value is
33443 -0.743643887037158704752191506114774.
33444 start_y
33446 Set the initial y position. Must be a floating point value between
33447 -100 and 100. Default value is
33448 -0.131825904205311970493132056385139.
33449 mptestsrc
33451 Generate various test patterns, as generated by the MPlayer test
33452 filter.
33453 The size of the generated video is fixed, and is 256x256. This source
33455 is useful in particular for testing encoding features.
33456 This source accepts the following options:
33458 rate, r
33460 Specify the frame rate of the sourced video, as the number of
33461 frames generated per second. It has to be a string in the format
33462 frame_rate_num/frame_rate_den, an integer number, a floating point
33463 number or a valid video frame rate abbreviation. The default value
33464 is "25".
33465 duration, d
33467 Set the duration of the sourced video. See the Time duration
33468 section in the ffmpeg-utils(1) manual for the accepted syntax.
33469 If not specified, or the expressed duration is negative, the video
33471 is supposed to be generated forever.
33472 test, t
33474 Set the number or the name of the test to perform. Supported tests
33475 are:
33476 dc_luma
33478 dc_chroma
33479 freq_luma
33480 freq_chroma
33481 amp_luma
33482 amp_chroma
33483 cbp
33484 mv
33485 ring1
33486 ring2
33487 all
33488 max_frames, m
33489 Set the maximum number of frames generated for each test,
33490 default value is 30.
33491 Default value is "all", which will cycle through the list of all
33493 tests.
33494 Some examples:
33496 mptestsrc=t=dc_luma
33498 will generate a "dc_luma" test pattern.
33500 frei0r_src
33502 Provide a frei0r source.
33503 To enable compilation of this filter you need to install the frei0r
33505 header and configure FFmpeg with "--enable-frei0r".
33506 This source accepts the following parameters:
33508 size
33510 The size of the video to generate. For the syntax of this option,
33511 check the "Video size" section in the ffmpeg-utils manual.
33512 framerate
33514 The framerate of the generated video. It may be a string of the
33515 form num/den or a frame rate abbreviation.
33516 filter_name
33518 The name to the frei0r source to load. For more information
33519 regarding frei0r and how to set the parameters, read the frei0r
33520 section in the video filters documentation.
33521 filter_params
33523 A '|'-separated list of parameters to pass to the frei0r source.
33524 For example, to generate a frei0r partik0l source with size 200x200 and
33526 frame rate 10 which is overlaid on the overlay filter main input:
33527 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
33529 life
33531 Generate a life pattern.
33532 This source is based on a generalization of John Conway's life game.
33534 The sourced input represents a life grid, each pixel represents a cell
33536 which can be in one of two possible states, alive or dead. Every cell
33537 interacts with its eight neighbours, which are the cells that are
33538 horizontally, vertically, or diagonally adjacent.
33539 At each interaction the grid evolves according to the adopted rule,
33541 which specifies the number of neighbor alive cells which will make a
33542 cell stay alive or born. The rule option allows one to specify the rule
33543 to adopt.
33544 This source accepts the following options:
33546 filename, f
33548 Set the file from which to read the initial grid state. In the
33549 file, each non-whitespace character is considered an alive cell,
33550 and newline is used to delimit the end of each row.
33551 If this option is not specified, the initial grid is generated
33553 randomly.
33554 rate, r
33556 Set the video rate, that is the number of frames generated per
33557 second. Default is 25.
33558 random_fill_ratio, ratio
33560 Set the random fill ratio for the initial random grid. It is a
33561 floating point number value ranging from 0 to 1, defaults to 1/PHI.
33562 It is ignored when a file is specified.
33563 random_seed, seed
33565 Set the seed for filling the initial random grid, must be an
33566 integer included between 0 and UINT32_MAX. If not specified, or if
33567 explicitly set to -1, the filter will try to use a good random seed
33568 on a best effort basis.
33569 rule
33571 Set the life rule.
33572 A rule can be specified with a code of the kind "SNS/BNB", where NS
33574 and NB are sequences of numbers in the range 0-8, NS specifies the
33575 number of alive neighbor cells which make a live cell stay alive,
33576 and NB the number of alive neighbor cells which make a dead cell to
33577 become alive (i.e. to "born"). "s" and "b" can be used in place of
33578 "S" and "B", respectively.
33579 Alternatively a rule can be specified by an 18-bits integer. The 9
33581 high order bits are used to encode the next cell state if it is
33582 alive for each number of neighbor alive cells, the low order bits
33583 specify the rule for "borning" new cells. Higher order bits encode
33584 for an higher number of neighbor cells. For example the number
33585 6153 = "(12<<9)+9" specifies a stay alive rule of 12 and a born
33586 rule of 9, which corresponds to "S23/B03".
33587 Default value is "S23/B3", which is the original Conway's game of
33589 life rule, and will keep a cell alive if it has 2 or 3 neighbor
33590 alive cells, and will born a new cell if there are three alive
33591 cells around a dead cell.
33592 size, s
33594 Set the size of the output video. For the syntax of this option,
33595 check the "Video size" section in the ffmpeg-utils manual.
33596 If filename is specified, the size is set by default to the same
33598 size of the input file. If size is set, it must contain the size
33599 specified in the input file, and the initial grid defined in that
33600 file is centered in the larger resulting area.
33601 If a filename is not specified, the size value defaults to
33603 "320x240" (used for a randomly generated initial grid).
33604 stitch
33606 If set to 1, stitch the left and right grid edges together, and the
33607 top and bottom edges also. Defaults to 1.
33608 mold
33610 Set cell mold speed. If set, a dead cell will go from death_color
33611 to mold_color with a step of mold. mold can have a value from 0 to
33612 255.
33613 life_color
33615 Set the color of living (or new born) cells.
33616 death_color
33618 Set the color of dead cells. If mold is set, this is the first
33619 color used to represent a dead cell.
33620 mold_color
33622 Set mold color, for definitely dead and moldy cells.
33623 For the syntax of these 3 color options, check the "Color" section
33625 in the ffmpeg-utils manual.
33626 Examples
33628 • Read a grid from pattern, and center it on a grid of size 300x300
33630 pixels:
33631 life=f=pattern:s=300x300
33633 • Generate a random grid of size 200x200, with a fill ratio of 2/3:
33635 life=ratio=2/3:s=200x200
33637 • Specify a custom rule for evolving a randomly generated grid:
33639 life=rule=S14/B34
33641 • Full example with slow death effect (mold) using ffplay:
33643 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
33645 allrgb, allyuv, color, colorchart, colorspectrum, haldclutsrc, nullsrc,
33647 pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc,
33648 testsrc2, yuvtestsrc
33649 The "allrgb" source returns frames of size 4096x4096 of all rgb colors.
33650 The "allyuv" source returns frames of size 4096x4096 of all yuv colors.
33652 The "color" source provides an uniformly colored input.
33654 The "colorchart" source provides a colors checker chart.
33656 The "colorspectrum" source provides a color spectrum input.
33658 The "haldclutsrc" source provides an identity Hald CLUT. See also
33660 haldclut filter.
33661 The "nullsrc" source returns unprocessed video frames. It is mainly
33663 useful to be employed in analysis / debugging tools, or as the source
33664 for filters which ignore the input data.
33665 The "pal75bars" source generates a color bars pattern, based on EBU PAL
33667 recommendations with 75% color levels.
33668 The "pal100bars" source generates a color bars pattern, based on EBU
33670 PAL recommendations with 100% color levels.
33671 The "rgbtestsrc" source generates an RGB test pattern useful for
33673 detecting RGB vs BGR issues. You should see a red, green and blue
33674 stripe from top to bottom.
33675 The "smptebars" source generates a color bars pattern, based on the
33677 SMPTE Engineering Guideline EG 1-1990.
33678 The "smptehdbars" source generates a color bars pattern, based on the
33680 SMPTE RP 219-2002.
33681 The "testsrc" source generates a test video pattern, showing a color
33683 pattern, a scrolling gradient and a timestamp. This is mainly intended
33684 for testing purposes.
33685 The "testsrc2" source is similar to testsrc, but supports more pixel
33687 formats instead of just "rgb24". This allows using it as an input for
33688 other tests without requiring a format conversion.
33689 The "yuvtestsrc" source generates an YUV test pattern. You should see a
33691 y, cb and cr stripe from top to bottom.
33692 The sources accept the following parameters:
33694 level
33696 Specify the level of the Hald CLUT, only available in the
33697 "haldclutsrc" source. A level of "N" generates a picture of "N*N*N"
33698 by "N*N*N" pixels to be used as identity matrix for 3D lookup
33699 tables. Each component is coded on a "1/(N*N)" scale.
33700 color, c
33702 Specify the color of the source, only available in the "color"
33703 source. For the syntax of this option, check the "Color" section in
33704 the ffmpeg-utils manual.
33705 size, s
33707 Specify the size of the sourced video. For the syntax of this
33708 option, check the "Video size" section in the ffmpeg-utils manual.
33709 The default value is "320x240".
33710 This option is not available with the "allrgb", "allyuv", and
33712 "haldclutsrc" filters.
33713 rate, r
33715 Specify the frame rate of the sourced video, as the number of
33716 frames generated per second. It has to be a string in the format
33717 frame_rate_num/frame_rate_den, an integer number, a floating point
33718 number or a valid video frame rate abbreviation. The default value
33719 is "25".
33720 duration, d
33722 Set the duration of the sourced video. See the Time duration
33723 section in the ffmpeg-utils(1) manual for the accepted syntax.
33724 If not specified, or the expressed duration is negative, the video
33726 is supposed to be generated forever.
33727 Since the frame rate is used as time base, all frames including the
33729 last one will have their full duration. If the specified duration
33730 is not a multiple of the frame duration, it will be rounded up.
33731 sar Set the sample aspect ratio of the sourced video.
33733 alpha
33735 Specify the alpha (opacity) of the background, only available in
33736 the "testsrc2" source. The value must be between 0 (fully
33737 transparent) and 255 (fully opaque, the default).
33738 decimals, n
33740 Set the number of decimals to show in the timestamp, only available
33741 in the "testsrc" source.
33742 The displayed timestamp value will correspond to the original
33744 timestamp value multiplied by the power of 10 of the specified
33745 value. Default value is 0.
33746 type
33748 Set the type of the color spectrum, only available in the
33749 "colorspectrum" source. Can be one of the following:
33750 black
33752 white
33753 all
33754 patch_size
33755 Set patch size of single color patch, only available in the
33756 "colorchart" source. Default is "64x64".
33757 preset
33759 Set colorchecker colors preset, only available in the "colorchart"
33760 source.
33761 Available values are:
33763 reference
33765 skintones
33766 Default value is "reference".
33768 Examples
33770 • Generate a video with a duration of 5.3 seconds, with size 176x144
33772 and a frame rate of 10 frames per second:
33773 testsrc=duration=5.3:size=qcif:rate=10
33775 • The following graph description will generate a red source with an
33777 opacity of 0.2, with size "qcif" and a frame rate of 10 frames per
33778 second:
33779 color=c=red@0.2:s=qcif:r=10
33781 • If the input content is to be ignored, "nullsrc" can be used. The
33783 following command generates noise in the luminance plane by
33784 employing the "geq" filter:
33785 nullsrc=s=256x256, geq=random(1)*255:128:128
33787 Commands
33789 The "color" source supports the following commands:
33791 c, color
33793 Set the color of the created image. Accepts the same syntax of the
33794 corresponding color option.
33795 openclsrc
33797 Generate video using an OpenCL program.
33798 source
33800 OpenCL program source file.
33801 kernel
33803 Kernel name in program.
33804 size, s
33806 Size of frames to generate. This must be set.
33807 format
33809 Pixel format to use for the generated frames. This must be set.
33810 rate, r
33812 Number of frames generated every second. Default value is '25'.
33813 For details of how the program loading works, see the program_opencl
33815 filter.
33816 Example programs:
33818 • Generate a colour ramp by setting pixel values from the position of
33820 the pixel in the output image. (Note that this will work with all
33821 pixel formats, but the generated output will not be the same.)
33822 __kernel void ramp(__write_only image2d_t dst,
33824 unsigned int index)
33825 {
33826 int2 loc = (int2)(get_global_id(0), get_global_id(1));
33827 float4 val;
33829 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
33830 write_imagef(dst, loc, val);
33832 }
33833 • Generate a Sierpinski carpet pattern, panning by a single pixel
33835 each frame.
33836 __kernel void sierpinski_carpet(__write_only image2d_t dst,
33838 unsigned int index)
33839 {
33840 int2 loc = (int2)(get_global_id(0), get_global_id(1));
33841 float4 value = 0.0f;
33843 int x = loc.x + index;
33844 int y = loc.y + index;
33845 while (x > 0 || y > 0) {
33846 if (x % 3 == 1 && y % 3 == 1) {
33847 value = 1.0f;
33848 break;
33849 }
33850 x /= 3;
33851 y /= 3;
33852 }
33853 write_imagef(dst, loc, value);
33855 }
33856 sierpinski
33858 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
33859 This source accepts the following options:
33861 size, s
33863 Set frame size. For the syntax of this option, check the "Video
33864 size" section in the ffmpeg-utils manual. Default value is
33865 "640x480".
33866 rate, r
33868 Set frame rate, expressed as number of frames per second. Default
33869 value is "25".
33870 seed
33872 Set seed which is used for random panning.
33873 jump
33875 Set max jump for single pan destination. Allowed range is from 1 to
33876 10000.
33877 type
33879 Set fractal type, can be default "carpet" or "triangle".
33880
33882 Below is a description of the currently available video sinks.
33883 buffersink
33885 Buffer video frames, and make them available to the end of the filter
33886 graph.
33887 This sink is mainly intended for programmatic use, in particular
33889 through the interface defined in libavfilter/buffersink.h or the
33890 options system.
33891 It accepts a pointer to an AVBufferSinkContext structure, which defines
33893 the incoming buffers' formats, to be passed as the opaque parameter to
33894 "avfilter_init_filter" for initialization.
33895 nullsink
33897 Null video sink: do absolutely nothing with the input video. It is
33898 mainly useful as a template and for use in analysis / debugging tools.
33899
33901 Below is a description of the currently available multimedia filters.
33902 a3dscope
33904 Convert input audio to 3d scope video output.
33905 The filter accepts the following options:
33907 rate, r
33909 Set frame rate, expressed as number of frames per second. Default
33910 value is "25".
33911 size, s
33913 Specify the video size for the output. For the syntax of this
33914 option, check the "Video size" section in the ffmpeg-utils manual.
33915 Default value is "hd720".
33916 fov Set the camera field of view. Default is 90 degrees. Allowed range
33918 is from 40 to 150.
33919 roll
33921 Set the camera roll.
33922 pitch
33924 Set the camera pitch.
33925 yaw Set the camera yaw.
33927 xzoom
33929 Set the camera zoom on X-axis.
33930 yzoom
33932 Set the camera zoom on Y-axis.
33933 zzoom
33935 Set the camera zoom on Z-axis.
33936 xpos
33938 Set the camera position on X-axis.
33939 ypos
33941 Set the camera position on Y-axis.
33942 zpos
33944 Set the camera position on Z-axis.
33945 length
33947 Set the length of displayed audio waves in number of frames.
33948 Commands
33950 Filter supports the some above options as commands.
33952 abitscope
33954 Convert input audio to a video output, displaying the audio bit scope.
33955 The filter accepts the following options:
33957 rate, r
33959 Set frame rate, expressed as number of frames per second. Default
33960 value is "25".
33961 size, s
33963 Specify the video size for the output. For the syntax of this
33964 option, check the "Video size" section in the ffmpeg-utils manual.
33965 Default value is "1024x256".
33966 colors
33968 Specify list of colors separated by space or by '|' which will be
33969 used to draw channels. Unrecognized or missing colors will be
33970 replaced by white color.
33971 mode, m
33973 Set output mode. Can be "bars" or "trace". Default is "bars".
33974 adrawgraph
33976 Draw a graph using input audio metadata.
33977 See drawgraph
33979 agraphmonitor
33981 See graphmonitor.
33982 ahistogram
33984 Convert input audio to a video output, displaying the volume histogram.
33985 The filter accepts the following options:
33987 dmode
33989 Specify how histogram is calculated.
33990 It accepts the following values:
33992 single
33994 Use single histogram for all channels.
33995 separate
33997 Use separate histogram for each channel.
33998 Default is "single".
34000 rate, r
34002 Set frame rate, expressed as number of frames per second. Default
34003 value is "25".
34004 size, s
34006 Specify the video size for the output. For the syntax of this
34007 option, check the "Video size" section in the ffmpeg-utils manual.
34008 Default value is "hd720".
34009 scale
34011 Set display scale.
34012 It accepts the following values:
34014 log logarithmic
34016 sqrt
34018 square root
34019 cbrt
34021 cubic root
34022 lin linear
34024 rlog
34026 reverse logarithmic
34027 Default is "log".
34029 ascale
34031 Set amplitude scale.
34032 It accepts the following values:
34034 log logarithmic
34036 lin linear
34038 Default is "log".
34040 acount
34042 Set how much frames to accumulate in histogram. Default is 1.
34043 Setting this to -1 accumulates all frames.
34044 rheight
34046 Set histogram ratio of window height.
34047 slide
34049 Set sonogram sliding.
34050 It accepts the following values:
34052 replace
34054 replace old rows with new ones.
34055 scroll
34057 scroll from top to bottom.
34058 Default is "replace".
34060 hmode
34062 Set histogram mode.
34063 It accepts the following values:
34065 abs Use absolute values of samples.
34067 sign
34069 Use untouched values of samples.
34070 Default is "abs".
34072 aphasemeter
34074 Measures phase of input audio, which is exported as metadata
34075 "lavfi.aphasemeter.phase", representing mean phase of current audio
34076 frame. A video output can also be produced and is enabled by default.
34077 The audio is passed through as first output.
34078 Audio will be rematrixed to stereo if it has a different channel
34080 layout. Phase value is in range "[-1, 1]" where -1 means left and right
34081 channels are completely out of phase and 1 means channels are in phase.
34082 The filter accepts the following options, all related to its video
34084 output:
34085 rate, r
34087 Set the output frame rate. Default value is 25.
34088 size, s
34090 Set the video size for the output. For the syntax of this option,
34091 check the "Video size" section in the ffmpeg-utils manual. Default
34092 value is "800x400".
34093 rc
34095 gc
34096 bc Specify the red, green, blue contrast. Default values are 2, 7 and
34097 1. Allowed range is "[0, 255]".
34098 mpc Set color which will be used for drawing median phase. If color is
34100 "none" which is default, no median phase value will be drawn.
34101 video
34103 Enable video output. Default is enabled.
34104 phasing detection
34106 The filter also detects out of phase and mono sequences in stereo
34108 streams. It logs the sequence start, end and duration when it lasts
34109 longer or as long as the minimum set.
34110 The filter accepts the following options for this detection:
34112 phasing
34114 Enable mono and out of phase detection. Default is disabled.
34115 tolerance, t
34117 Set phase tolerance for mono detection, in amplitude ratio. Default
34118 is 0. Allowed range is "[0, 1]".
34119 angle, a
34121 Set angle threshold for out of phase detection, in degree. Default
34122 is 170. Allowed range is "[90, 180]".
34123 duration, d
34125 Set mono or out of phase duration until notification, expressed in
34126 seconds. Default is 2.
34127 Examples
34129 • Complete example with ffmpeg to detect 1 second of mono with 0.001
34131 phase tolerance:
34132 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
34134 avectorscope
34136 Convert input audio to a video output, representing the audio vector
34137 scope.
34138 The filter is used to measure the difference between channels of stereo
34140 audio stream. A monaural signal, consisting of identical left and right
34141 signal, results in straight vertical line. Any stereo separation is
34142 visible as a deviation from this line, creating a Lissajous figure. If
34143 the straight (or deviation from it) but horizontal line appears this
34144 indicates that the left and right channels are out of phase.
34145 The filter accepts the following options:
34147 mode, m
34149 Set the vectorscope mode.
34150 Available values are:
34152 lissajous
34154 Lissajous rotated by 45 degrees.
34155 lissajous_xy
34157 Same as above but not rotated.
34158 polar
34160 Shape resembling half of circle.
34161 Default value is lissajous.
34163 size, s
34165 Set the video size for the output. For the syntax of this option,
34166 check the "Video size" section in the ffmpeg-utils manual. Default
34167 value is "400x400".
34168 rate, r
34170 Set the output frame rate. Default value is 25.
34171 rc
34173 gc
34174 bc
34175 ac Specify the red, green, blue and alpha contrast. Default values are
34176 40, 160, 80 and 255. Allowed range is "[0, 255]".
34177 rf
34179 gf
34180 bf
34181 af Specify the red, green, blue and alpha fade. Default values are 15,
34182 10, 5 and 5. Allowed range is "[0, 255]".
34183 zoom
34185 Set the zoom factor. Default value is 1. Allowed range is "[0,
34186 10]". Values lower than 1 will auto adjust zoom factor to maximal
34187 possible value.
34188 draw
34190 Set the vectorscope drawing mode.
34191 Available values are:
34193 dot Draw dot for each sample.
34195 line
34197 Draw line between previous and current sample.
34198 aaline
34200 Draw anti-aliased line between previous and current sample.
34201 Default value is dot.
34203 scale
34205 Specify amplitude scale of audio samples.
34206 Available values are:
34208 lin Linear.
34210 sqrt
34212 Square root.
34213 cbrt
34215 Cubic root.
34216 log Logarithmic.
34218 swap
34220 Swap left channel axis with right channel axis.
34221 mirror
34223 Mirror axis.
34224 none
34226 No mirror.
34227 x Mirror only x axis.
34229 y Mirror only y axis.
34231 xy Mirror both axis.
34233 Examples
34235 • Complete example using ffplay:
34237 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
34239 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
34240 Commands
34242 This filter supports the all above options as commands except options
34244 "size" and "rate".
34245 bench, abench
34247 Benchmark part of a filtergraph.
34248 The filter accepts the following options:
34250 action
34252 Start or stop a timer.
34253 Available values are:
34255 start
34257 Get the current time, set it as frame metadata (using the key
34258 "lavfi.bench.start_time"), and forward the frame to the next
34259 filter.
34260 stop
34262 Get the current time and fetch the "lavfi.bench.start_time"
34263 metadata from the input frame metadata to get the time
34264 difference. Time difference, average, maximum and minimum time
34265 (respectively "t", "avg", "max" and "min") are then printed.
34266 The timestamps are expressed in seconds.
34267 Examples
34269 • Benchmark selectivecolor filter:
34271 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
34273 concat
34275 Concatenate audio and video streams, joining them together one after
34276 the other.
34277 The filter works on segments of synchronized video and audio streams.
34279 All segments must have the same number of streams of each type, and
34280 that will also be the number of streams at output.
34281 The filter accepts the following options:
34283 n Set the number of segments. Default is 2.
34285 v Set the number of output video streams, that is also the number of
34287 video streams in each segment. Default is 1.
34288 a Set the number of output audio streams, that is also the number of
34290 audio streams in each segment. Default is 0.
34291 unsafe
34293 Activate unsafe mode: do not fail if segments have a different
34294 format.
34295 The filter has v+a outputs: first v video outputs, then a audio
34297 outputs.
34298 There are nx(v+a) inputs: first the inputs for the first segment, in
34300 the same order as the outputs, then the inputs for the second segment,
34301 etc.
34302 Related streams do not always have exactly the same duration, for
34304 various reasons including codec frame size or sloppy authoring. For
34305 that reason, related synchronized streams (e.g. a video and its audio
34306 track) should be concatenated at once. The concat filter will use the
34307 duration of the longest stream in each segment (except the last one),
34308 and if necessary pad shorter audio streams with silence.
34309 For this filter to work correctly, all segments must start at timestamp
34311 0.
34312 All corresponding streams must have the same parameters in all
34314 segments; the filtering system will automatically select a common pixel
34315 format for video streams, and a common sample format, sample rate and
34316 channel layout for audio streams, but other settings, such as
34317 resolution, must be converted explicitly by the user.
34318 Different frame rates are acceptable but will result in variable frame
34320 rate at output; be sure to configure the output file to handle it.
34321 Examples
34323 • Concatenate an opening, an episode and an ending, all in bilingual
34325 version (video in stream 0, audio in streams 1 and 2):
34326 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
34328 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
34329 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
34330 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
34331 • Concatenate two parts, handling audio and video separately, using
34333 the (a)movie sources, and adjusting the resolution:
34334 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
34336 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
34337 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
34338 Note that a desync will happen at the stitch if the audio and video
34340 streams do not have exactly the same duration in the first file.
34341 Commands
34343 This filter supports the following commands:
34345 next
34347 Close the current segment and step to the next one
34348 ebur128
34350 EBU R128 scanner filter. This filter takes an audio stream and analyzes
34351 its loudness level. By default, it logs a message at a frequency of
34352 10Hz with the Momentary loudness (identified by "M"), Short-term
34353 loudness ("S"), Integrated loudness ("I") and Loudness Range ("LRA").
34354 The filter can only analyze streams which have sample format is double-
34356 precision floating point. The input stream will be converted to this
34357 specification, if needed. Users may need to insert aformat and/or
34358 aresample filters after this filter to obtain the original parameters.
34359 The filter also has a video output (see the video option) with a real
34361 time graph to observe the loudness evolution. The graphic contains the
34362 logged message mentioned above, so it is not printed anymore when this
34363 option is set, unless the verbose logging is set. The main graphing
34364 area contains the short-term loudness (3 seconds of analysis), and the
34365 gauge on the right is for the momentary loudness (400 milliseconds),
34366 but can optionally be configured to instead display short-term loudness
34367 (see gauge).
34368 The green area marks a +/- 1LU target range around the target loudness
34370 (-23LUFS by default, unless modified through target).
34371 More information about the Loudness Recommendation EBU R128 on
34373 <http://tech.ebu.ch/loudness>.
34374 The filter accepts the following options:
34376 video
34378 Activate the video output. The audio stream is passed unchanged
34379 whether this option is set or no. The video stream will be the
34380 first output stream if activated. Default is 0.
34381 size
34383 Set the video size. This option is for video only. For the syntax
34384 of this option, check the "Video size" section in the ffmpeg-utils
34385 manual. Default and minimum resolution is "640x480".
34386 meter
34388 Set the EBU scale meter. Default is 9. Common values are 9 and 18,
34389 respectively for EBU scale meter +9 and EBU scale meter +18. Any
34390 other integer value between this range is allowed.
34391 metadata
34393 Set metadata injection. If set to 1, the audio input will be
34394 segmented into 100ms output frames, each of them containing various
34395 loudness information in metadata. All the metadata keys are
34396 prefixed with "lavfi.r128.".
34397 Default is 0.
34399 framelog
34401 Force the frame logging level.
34402 Available values are:
34404 quiet
34406 logging disabled
34407 info
34409 information logging level
34410 verbose
34412 verbose logging level
34413 By default, the logging level is set to info. If the video or the
34415 metadata options are set, it switches to verbose.
34416 peak
34418 Set peak mode(s).
34419 Available modes can be cumulated (the option is a "flag" type).
34421 Possible values are:
34422 none
34424 Disable any peak mode (default).
34425 sample
34427 Enable sample-peak mode.
34428 Simple peak mode looking for the higher sample value. It logs a
34430 message for sample-peak (identified by "SPK").
34431 true
34433 Enable true-peak mode.
34434 If enabled, the peak lookup is done on an over-sampled version
34436 of the input stream for better peak accuracy. It logs a message
34437 for true-peak. (identified by "TPK") and true-peak per frame
34438 (identified by "FTPK"). This mode requires a build with
34439 "libswresample".
34440 dualmono
34442 Treat mono input files as "dual mono". If a mono file is intended
34443 for playback on a stereo system, its EBU R128 measurement will be
34444 perceptually incorrect. If set to "true", this option will
34445 compensate for this effect. Multi-channel input files are not
34446 affected by this option.
34447 panlaw
34449 Set a specific pan law to be used for the measurement of dual mono
34450 files. This parameter is optional, and has a default value of
34451 -3.01dB.
34452 target
34454 Set a specific target level (in LUFS) used as relative zero in the
34455 visualization. This parameter is optional and has a default value
34456 of -23LUFS as specified by EBU R128. However, material published
34457 online may prefer a level of -16LUFS (e.g. for use with podcasts or
34458 video platforms).
34459 gauge
34461 Set the value displayed by the gauge. Valid values are "momentary"
34462 and s "shortterm". By default the momentary value will be used, but
34463 in certain scenarios it may be more useful to observe the short
34464 term value instead (e.g. live mixing).
34465 scale
34467 Sets the display scale for the loudness. Valid parameters are
34468 "absolute" (in LUFS) or "relative" (LU) relative to the target.
34469 This only affects the video output, not the summary or continuous
34470 log output.
34471 Examples
34473 • Real-time graph using ffplay, with a EBU scale meter +18:
34475 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
34477 • Run an analysis with ffmpeg:
34479 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
34481 interleave, ainterleave
34483 Temporally interleave frames from several inputs.
34484 "interleave" works with video inputs, "ainterleave" with audio.
34486 These filters read frames from several inputs and send the oldest
34488 queued frame to the output.
34489 Input streams must have well defined, monotonically increasing frame
34491 timestamp values.
34492 In order to submit one frame to output, these filters need to enqueue
34494 at least one frame for each input, so they cannot work in case one
34495 input is not yet terminated and will not receive incoming frames.
34496 For example consider the case when one input is a "select" filter which
34498 always drops input frames. The "interleave" filter will keep reading
34499 from that input, but it will never be able to send new frames to output
34500 until the input sends an end-of-stream signal.
34501 Also, depending on inputs synchronization, the filters will drop frames
34503 in case one input receives more frames than the other ones, and the
34504 queue is already filled.
34505 These filters accept the following options:
34507 nb_inputs, n
34509 Set the number of different inputs, it is 2 by default.
34510 duration
34512 How to determine the end-of-stream.
34513 longest
34515 The duration of the longest input. (default)
34516 shortest
34518 The duration of the shortest input.
34519 first
34521 The duration of the first input.
34522 Examples
34524 • Interleave frames belonging to different streams using ffmpeg:
34526 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
34528 • Add flickering blur effect:
34530 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
34532 latency, alatency
34534 Measure filtering latency.
34535 Report previous filter filtering latency, delay in number of audio
34537 samples for audio filters or number of video frames for video filters.
34538 On end of input stream, filter will report min and max measured latency
34540 for previous running filter in filtergraph.
34541 metadata, ametadata
34543 Manipulate frame metadata.
34544 This filter accepts the following options:
34546 mode
34548 Set mode of operation of the filter.
34549 Can be one of the following:
34551 select
34553 If both "value" and "key" is set, select frames which have such
34554 metadata. If only "key" is set, select every frame that has
34555 such key in metadata.
34556 add Add new metadata "key" and "value". If key is already available
34558 do nothing.
34559 modify
34561 Modify value of already present key.
34562 delete
34564 If "value" is set, delete only keys that have such value.
34565 Otherwise, delete key. If "key" is not set, delete all metadata
34566 values in the frame.
34567 print
34569 Print key and its value if metadata was found. If "key" is not
34570 set print all metadata values available in frame.
34571 key Set key used with all modes. Must be set for all modes except
34573 "print" and "delete".
34574 value
34576 Set metadata value which will be used. This option is mandatory for
34577 "modify" and "add" mode.
34578 function
34580 Which function to use when comparing metadata value and "value".
34581 Can be one of following:
34583 same_str
34585 Values are interpreted as strings, returns true if metadata
34586 value is same as "value".
34587 starts_with
34589 Values are interpreted as strings, returns true if metadata
34590 value starts with the "value" option string.
34591 less
34593 Values are interpreted as floats, returns true if metadata
34594 value is less than "value".
34595 equal
34597 Values are interpreted as floats, returns true if "value" is
34598 equal with metadata value.
34599 greater
34601 Values are interpreted as floats, returns true if metadata
34602 value is greater than "value".
34603 expr
34605 Values are interpreted as floats, returns true if expression
34606 from option "expr" evaluates to true.
34607 ends_with
34609 Values are interpreted as strings, returns true if metadata
34610 value ends with the "value" option string.
34611 expr
34613 Set expression which is used when "function" is set to "expr". The
34614 expression is evaluated through the eval API and can contain the
34615 following constants:
34616 VALUE1, FRAMEVAL
34618 Float representation of "value" from metadata key.
34619 VALUE2, USERVAL
34621 Float representation of "value" as supplied by user in "value"
34622 option.
34623 file
34625 If specified in "print" mode, output is written to the named file.
34626 Instead of plain filename any writable url can be specified.
34627 Filename ``-'' is a shorthand for standard output. If "file" option
34628 is not set, output is written to the log with AV_LOG_INFO loglevel.
34629 direct
34631 Reduces buffering in print mode when output is written to a URL set
34632 using file.
34633 Examples
34635 • Print all metadata values for frames with key
34637 "lavfi.signalstats.YDIF" with values between 0 and 1.
34638 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
34640 • Print silencedetect output to file metadata.txt.
34642 silencedetect,ametadata=mode=print:file=metadata.txt
34644 • Direct all metadata to a pipe with file descriptor 4.
34646 metadata=mode=print:file='pipe\:4'
34648 perms, aperms
34650 Set read/write permissions for the output frames.
34651 These filters are mainly aimed at developers to test direct path in the
34653 following filter in the filtergraph.
34654 The filters accept the following options:
34656 mode
34658 Select the permissions mode.
34659 It accepts the following values:
34661 none
34663 Do nothing. This is the default.
34664 ro Set all the output frames read-only.
34666 rw Set all the output frames directly writable.
34668 toggle
34670 Make the frame read-only if writable, and writable if read-
34671 only.
34672 random
34674 Set each output frame read-only or writable randomly.
34675 seed
34677 Set the seed for the random mode, must be an integer included
34678 between 0 and "UINT32_MAX". If not specified, or if explicitly set
34679 to -1, the filter will try to use a good random seed on a best
34680 effort basis.
34681 Note: in case of auto-inserted filter between the permission filter and
34683 the following one, the permission might not be received as expected in
34684 that following filter. Inserting a format or aformat filter before the
34685 perms/aperms filter can avoid this problem.
34686 realtime, arealtime
34688 Slow down filtering to match real time approximately.
34689 These filters will pause the filtering for a variable amount of time to
34691 match the output rate with the input timestamps. They are similar to
34692 the re option to "ffmpeg".
34693 They accept the following options:
34695 limit
34697 Time limit for the pauses. Any pause longer than that will be
34698 considered a timestamp discontinuity and reset the timer. Default
34699 is 2 seconds.
34700 speed
34702 Speed factor for processing. The value must be a float larger than
34703 zero. Values larger than 1.0 will result in faster than realtime
34704 processing, smaller will slow processing down. The limit is
34705 automatically adapted accordingly. Default is 1.0.
34706 A processing speed faster than what is possible without these
34708 filters cannot be achieved.
34709 Commands
34711 Both filters supports the all above options as commands.
34713 segment, asegment
34715 Split single input stream into multiple streams.
34716 This filter does opposite of concat filters.
34718 "segment" works on video frames, "asegment" on audio samples.
34720 This filter accepts the following options:
34722 timestamps
34724 Timestamps of output segments separated by '|'. The first segment
34725 will run from the beginning of the input stream. The last segment
34726 will run until the end of the input stream
34727 frames, samples
34729 Exact frame/sample count to split the segments.
34730 In all cases, prefixing an each segment with '+' will make it relative
34732 to the previous segment.
34733 Examples
34735 • Split input audio stream into three output audio streams, starting
34737 at start of input audio stream and storing that in 1st output audio
34738 stream, then following at 60th second and storing than in 2nd
34739 output audio stream, and last after 150th second of input audio
34740 stream store in 3rd output audio stream:
34741 asegment=timestamps="60|150"
34743 select, aselect
34745 Select frames to pass in output.
34746 This filter accepts the following options:
34748 expr, e
34750 Set expression, which is evaluated for each input frame.
34751 If the expression is evaluated to zero, the frame is discarded.
34753 If the evaluation result is negative or NaN, the frame is sent to
34755 the first output; otherwise it is sent to the output with index
34756 "ceil(val)-1", assuming that the input index starts from 0.
34757 For example a value of 1.2 corresponds to the output with index
34759 "ceil(1.2)-1 = 2-1 = 1", that is the second output.
34760 outputs, n
34762 Set the number of outputs. The output to which to send the selected
34763 frame is based on the result of the evaluation. Default value is 1.
34764 The expression can contain the following constants:
34766 n The (sequential) number of the filtered frame, starting from 0.
34768 selected_n
34770 The (sequential) number of the selected frame, starting from 0.
34771 prev_selected_n
34773 The sequential number of the last selected frame. It's NAN if
34774 undefined.
34775 TB The timebase of the input timestamps.
34777 pts The PTS (Presentation TimeStamp) of the filtered frame, expressed
34779 in TB units. It's NAN if undefined.
34780 t The PTS of the filtered frame, expressed in seconds. It's NAN if
34782 undefined.
34783 prev_pts
34785 The PTS of the previously filtered frame. It's NAN if undefined.
34786 prev_selected_pts
34788 The PTS of the last previously filtered frame. It's NAN if
34789 undefined.
34790 prev_selected_t
34792 The PTS of the last previously selected frame, expressed in
34793 seconds. It's NAN if undefined.
34794 start_pts
34796 The first PTS in the stream which is not NAN. It remains NAN if not
34797 found.
34798 start_t
34800 The first PTS, in seconds, in the stream which is not NAN. It
34801 remains NAN if not found.
34802 pict_type (video only)
34804 The type of the filtered frame. It can assume one of the following
34805 values:
34806 I
34808 P
34809 B
34810 S
34811 SI
34812 SP
34813 BI
34814 interlace_type (video only)
34815 The frame interlace type. It can assume one of the following
34816 values:
34817 PROGRESSIVE
34819 The frame is progressive (not interlaced).
34820 TOPFIRST
34822 The frame is top-field-first.
34823 BOTTOMFIRST
34825 The frame is bottom-field-first.
34826 consumed_sample_n (audio only)
34828 the number of selected samples before the current frame
34829 samples_n (audio only)
34831 the number of samples in the current frame
34832 sample_rate (audio only)
34834 the input sample rate
34835 key This is 1 if the filtered frame is a key-frame, 0 otherwise.
34837 pos the position in the file of the filtered frame, -1 if the
34839 information is not available (e.g. for synthetic video)
34840 scene (video only)
34842 value between 0 and 1 to indicate a new scene; a low value reflects
34843 a low probability for the current frame to introduce a new scene,
34844 while a higher value means the current frame is more likely to be
34845 one (see the example below)
34846 concatdec_select
34848 The concat demuxer can select only part of a concat input file by
34849 setting an inpoint and an outpoint, but the output packets may not
34850 be entirely contained in the selected interval. By using this
34851 variable, it is possible to skip frames generated by the concat
34852 demuxer which are not exactly contained in the selected interval.
34853 This works by comparing the frame pts against the
34855 lavf.concat.start_time and the lavf.concat.duration packet metadata
34856 values which are also present in the decoded frames.
34857 The concatdec_select variable is -1 if the frame pts is at least
34859 start_time and either the duration metadata is missing or the frame
34860 pts is less than start_time + duration, 0 otherwise, and NaN if the
34861 start_time metadata is missing.
34862 That basically means that an input frame is selected if its pts is
34864 within the interval set by the concat demuxer.
34865 The default value of the select expression is "1".
34867 Examples
34869 • Select all frames in input:
34871 select
34873 The example above is the same as:
34875 select=1
34877 • Skip all frames:
34879 select=0
34881 • Select only I-frames:
34883 select='eq(pict_type\,I)'
34885 • Select one frame every 100:
34887 select='not(mod(n\,100))'
34889 • Select only frames contained in the 10-20 time interval:
34891 select=between(t\,10\,20)
34893 • Select only I-frames contained in the 10-20 time interval:
34895 select=between(t\,10\,20)*eq(pict_type\,I)
34897 • Select frames with a minimum distance of 10 seconds:
34899 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
34901 • Use aselect to select only audio frames with samples number > 100:
34903 aselect='gt(samples_n\,100)'
34905 • Create a mosaic of the first scenes:
34907 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
34909 Comparing scene against a value between 0.3 and 0.5 is generally a
34911 sane choice.
34912 • Send even and odd frames to separate outputs, and compose them:
34914 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
34916 • Select useful frames from an ffconcat file which is using inpoints
34918 and outpoints but where the source files are not intra frame only.
34919 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
34921 sendcmd, asendcmd
34923 Send commands to filters in the filtergraph.
34924 These filters read commands to be sent to other filters in the
34926 filtergraph.
34927 "sendcmd" must be inserted between two video filters, "asendcmd" must
34929 be inserted between two audio filters, but apart from that they act the
34930 same way.
34931 The specification of commands can be provided in the filter arguments
34933 with the commands option, or in a file specified by the filename
34934 option.
34935 These filters accept the following options:
34937 commands, c
34939 Set the commands to be read and sent to the other filters.
34940 filename, f
34942 Set the filename of the commands to be read and sent to the other
34943 filters.
34944 Commands syntax
34946 A commands description consists of a sequence of interval
34948 specifications, comprising a list of commands to be executed when a
34949 particular event related to that interval occurs. The occurring event
34950 is typically the current frame time entering or leaving a given time
34951 interval.
34952 An interval is specified by the following syntax:
34954 <START>[-<END>] <COMMANDS>;
34956 The time interval is specified by the START and END times. END is
34958 optional and defaults to the maximum time.
34959 The current frame time is considered within the specified interval if
34961 it is included in the interval [START, END), that is when the time is
34962 greater or equal to START and is lesser than END.
34963 COMMANDS consists of a sequence of one or more command specifications,
34965 separated by ",", relating to that interval. The syntax of a command
34966 specification is given by:
34967 [<FLAGS>] <TARGET> <COMMAND> <ARG>
34969 FLAGS is optional and specifies the type of events relating to the time
34971 interval which enable sending the specified command, and must be a non-
34972 null sequence of identifier flags separated by "+" or "|" and enclosed
34973 between "[" and "]".
34974 The following flags are recognized:
34976 enter
34978 The command is sent when the current frame timestamp enters the
34979 specified interval. In other words, the command is sent when the
34980 previous frame timestamp was not in the given interval, and the
34981 current is.
34982 leave
34984 The command is sent when the current frame timestamp leaves the
34985 specified interval. In other words, the command is sent when the
34986 previous frame timestamp was in the given interval, and the current
34987 is not.
34988 expr
34990 The command ARG is interpreted as expression and result of
34991 expression is passed as ARG.
34992 The expression is evaluated through the eval API and can contain
34994 the following constants:
34995 POS Original position in the file of the frame, or undefined if
34997 undefined for the current frame.
34998 PTS The presentation timestamp in input.
35000 N The count of the input frame for video or audio, starting from
35002 0.
35003 T The time in seconds of the current frame.
35005 TS The start time in seconds of the current command interval.
35007 TE The end time in seconds of the current command interval.
35009 TI The interpolated time of the current command interval, TI = (T
35011 - TS) / (TE - TS).
35012 W The video frame width.
35014 H The video frame height.
35016 If FLAGS is not specified, a default value of "[enter]" is assumed.
35018 TARGET specifies the target of the command, usually the name of the
35020 filter class or a specific filter instance name.
35021 COMMAND specifies the name of the command for the target filter.
35023 ARG is optional and specifies the optional list of argument for the
35025 given COMMAND.
35026 Between one interval specification and another, whitespaces, or
35028 sequences of characters starting with "#" until the end of line, are
35029 ignored and can be used to annotate comments.
35030 A simplified BNF description of the commands specification syntax
35032 follows:
35033 <COMMAND_FLAG> ::= "enter" | "leave"
35035 <COMMAND_FLAGS> ::= <COMMAND_FLAG> [(+|"|")<COMMAND_FLAG>]
35036 <COMMAND> ::= ["[" <COMMAND_FLAGS> "]"] <TARGET> <COMMAND> [<ARG>]
35037 <COMMANDS> ::= <COMMAND> [,<COMMANDS>]
35038 <INTERVAL> ::= <START>[-<END>] <COMMANDS>
35039 <INTERVALS> ::= <INTERVAL>[;<INTERVALS>]
35040 Examples
35042 • Specify audio tempo change at second 4:
35044 asendcmd=c='4.0 atempo tempo 1.5',atempo
35046 • Target a specific filter instance:
35048 asendcmd=c='4.0 atempo@my tempo 1.5',atempo@my
35050 • Specify a list of drawtext and hue commands in a file.
35052 # show text in the interval 5-10
35054 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
35055 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
35056 # desaturate the image in the interval 15-20
35058 15.0-20.0 [enter] hue s 0,
35059 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
35060 [leave] hue s 1,
35061 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
35062 # apply an exponential saturation fade-out effect, starting from time 25
35064 25 [enter] hue s exp(25-t)
35065 A filtergraph allowing to read and process the above command list
35067 stored in a file test.cmd, can be specified with:
35068 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
35070 setpts, asetpts
35072 Change the PTS (presentation timestamp) of the input frames.
35073 "setpts" works on video frames, "asetpts" on audio frames.
35075 This filter accepts the following options:
35077 expr
35079 The expression which is evaluated for each frame to construct its
35080 timestamp.
35081 The expression is evaluated through the eval API and can contain the
35083 following constants:
35084 FRAME_RATE, FR
35086 frame rate, only defined for constant frame-rate video
35087 PTS The presentation timestamp in input
35089 N The count of the input frame for video or the number of consumed
35091 samples, not including the current frame for audio, starting from
35092 0.
35093 NB_CONSUMED_SAMPLES
35095 The number of consumed samples, not including the current frame
35096 (only audio)
35097 NB_SAMPLES, S
35099 The number of samples in the current frame (only audio)
35100 SAMPLE_RATE, SR
35102 The audio sample rate.
35103 STARTPTS
35105 The PTS of the first frame.
35106 STARTT
35108 the time in seconds of the first frame
35109 INTERLACED
35111 State whether the current frame is interlaced.
35112 T the time in seconds of the current frame
35114 POS original position in the file of the frame, or undefined if
35116 undefined for the current frame
35117 PREV_INPTS
35119 The previous input PTS.
35120 PREV_INT
35122 previous input time in seconds
35123 PREV_OUTPTS
35125 The previous output PTS.
35126 PREV_OUTT
35128 previous output time in seconds
35129 RTCTIME
35131 The wallclock (RTC) time in microseconds. This is deprecated, use
35132 time(0) instead.
35133 RTCSTART
35135 The wallclock (RTC) time at the start of the movie in microseconds.
35136 TB The timebase of the input timestamps.
35138 Examples
35140 • Start counting PTS from zero
35142 setpts=PTS-STARTPTS
35144 • Apply fast motion effect:
35146 setpts=0.5*PTS
35148 • Apply slow motion effect:
35150 setpts=2.0*PTS
35152 • Set fixed rate of 25 frames per second:
35154 setpts=N/(25*TB)
35156 • Set fixed rate 25 fps with some jitter:
35158 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
35160 • Apply an offset of 10 seconds to the input PTS:
35162 setpts=PTS+10/TB
35164 • Generate timestamps from a "live source" and rebase onto the
35166 current timebase:
35167 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
35169 • Generate timestamps by counting samples:
35171 asetpts=N/SR/TB
35173 setrange
35175 Force color range for the output video frame.
35176 The "setrange" filter marks the color range property for the output
35178 frames. It does not change the input frame, but only sets the
35179 corresponding property, which affects how the frame is treated by
35180 following filters.
35181 The filter accepts the following options:
35183 range
35185 Available values are:
35186 auto
35188 Keep the same color range property.
35189 unspecified, unknown
35191 Set the color range as unspecified.
35192 limited, tv, mpeg
35194 Set the color range as limited.
35195 full, pc, jpeg
35197 Set the color range as full.
35198 settb, asettb
35200 Set the timebase to use for the output frames timestamps. It is mainly
35201 useful for testing timebase configuration.
35202 It accepts the following parameters:
35204 expr, tb
35206 The expression which is evaluated into the output timebase.
35207 The value for tb is an arithmetic expression representing a rational.
35209 The expression can contain the constants "AVTB" (the default timebase),
35210 "intb" (the input timebase) and "sr" (the sample rate, audio only).
35211 Default value is "intb".
35212 Examples
35214 • Set the timebase to 1/25:
35216 settb=expr=1/25
35218 • Set the timebase to 1/10:
35220 settb=expr=0.1
35222 • Set the timebase to 1001/1000:
35224 settb=1+0.001
35226 • Set the timebase to 2*intb:
35228 settb=2*intb
35230 • Set the default timebase value:
35232 settb=AVTB
35234 showcqt
35236 Convert input audio to a video output representing frequency spectrum
35237 logarithmically using Brown-Puckette constant Q transform algorithm
35238 with direct frequency domain coefficient calculation (but the transform
35239 itself is not really constant Q, instead the Q factor is actually
35240 variable/clamped), with musical tone scale, from E0 to D#10.
35241 The filter accepts the following options:
35243 size, s
35245 Specify the video size for the output. It must be even. For the
35246 syntax of this option, check the "Video size" section in the
35247 ffmpeg-utils manual. Default value is "1920x1080".
35248 fps, rate, r
35250 Set the output frame rate. Default value is 25.
35251 bar_h
35253 Set the bargraph height. It must be even. Default value is -1 which
35254 computes the bargraph height automatically.
35255 axis_h
35257 Set the axis height. It must be even. Default value is -1 which
35258 computes the axis height automatically.
35259 sono_h
35261 Set the sonogram height. It must be even. Default value is -1 which
35262 computes the sonogram height automatically.
35263 fullhd
35265 Set the fullhd resolution. This option is deprecated, use size, s
35266 instead. Default value is 1.
35267 sono_v, volume
35269 Specify the sonogram volume expression. It can contain variables:
35270 bar_v
35272 the bar_v evaluated expression
35273 frequency, freq, f
35275 the frequency where it is evaluated
35276 timeclamp, tc
35278 the value of timeclamp option
35279 and functions:
35281 a_weighting(f)
35283 A-weighting of equal loudness
35284 b_weighting(f)
35286 B-weighting of equal loudness
35287 c_weighting(f)
35289 C-weighting of equal loudness.
35290 Default value is 16.
35292 bar_v, volume2
35294 Specify the bargraph volume expression. It can contain variables:
35295 sono_v
35297 the sono_v evaluated expression
35298 frequency, freq, f
35300 the frequency where it is evaluated
35301 timeclamp, tc
35303 the value of timeclamp option
35304 and functions:
35306 a_weighting(f)
35308 A-weighting of equal loudness
35309 b_weighting(f)
35311 B-weighting of equal loudness
35312 c_weighting(f)
35314 C-weighting of equal loudness.
35315 Default value is "sono_v".
35317 sono_g, gamma
35319 Specify the sonogram gamma. Lower gamma makes the spectrum more
35320 contrast, higher gamma makes the spectrum having more range.
35321 Default value is 3. Acceptable range is "[1, 7]".
35322 bar_g, gamma2
35324 Specify the bargraph gamma. Default value is 1. Acceptable range is
35325 "[1, 7]".
35326 bar_t
35328 Specify the bargraph transparency level. Lower value makes the
35329 bargraph sharper. Default value is 1. Acceptable range is "[0,
35330 1]".
35331 timeclamp, tc
35333 Specify the transform timeclamp. At low frequency, there is trade-
35334 off between accuracy in time domain and frequency domain. If
35335 timeclamp is lower, event in time domain is represented more
35336 accurately (such as fast bass drum), otherwise event in frequency
35337 domain is represented more accurately (such as bass guitar).
35338 Acceptable range is "[0.002, 1]". Default value is 0.17.
35339 attack
35341 Set attack time in seconds. The default is 0 (disabled). Otherwise,
35342 it limits future samples by applying asymmetric windowing in time
35343 domain, useful when low latency is required. Accepted range is "[0,
35344 1]".
35345 basefreq
35347 Specify the transform base frequency. Default value is
35348 20.01523126408007475, which is frequency 50 cents below E0.
35349 Acceptable range is "[10, 100000]".
35350 endfreq
35352 Specify the transform end frequency. Default value is
35353 20495.59681441799654, which is frequency 50 cents above D#10.
35354 Acceptable range is "[10, 100000]".
35355 coeffclamp
35357 This option is deprecated and ignored.
35358 tlength
35360 Specify the transform length in time domain. Use this option to
35361 control accuracy trade-off between time domain and frequency domain
35362 at every frequency sample. It can contain variables:
35363 frequency, freq, f
35365 the frequency where it is evaluated
35366 timeclamp, tc
35368 the value of timeclamp option.
35369 Default value is "384*tc/(384+tc*f)".
35371 count
35373 Specify the transform count for every video frame. Default value is
35374 6. Acceptable range is "[1, 30]".
35375 fcount
35377 Specify the transform count for every single pixel. Default value
35378 is 0, which makes it computed automatically. Acceptable range is
35379 "[0, 10]".
35380 fontfile
35382 Specify font file for use with freetype to draw the axis. If not
35383 specified, use embedded font. Note that drawing with font file or
35384 embedded font is not implemented with custom basefreq and endfreq,
35385 use axisfile option instead.
35386 font
35388 Specify fontconfig pattern. This has lower priority than fontfile.
35389 The ":" in the pattern may be replaced by "|" to avoid unnecessary
35390 escaping.
35391 fontcolor
35393 Specify font color expression. This is arithmetic expression that
35394 should return integer value 0xRRGGBB. It can contain variables:
35395 frequency, freq, f
35397 the frequency where it is evaluated
35398 timeclamp, tc
35400 the value of timeclamp option
35401 and functions:
35403 midi(f)
35405 midi number of frequency f, some midi numbers: E0(16), C1(24),
35406 C2(36), A4(69)
35407 r(x), g(x), b(x)
35409 red, green, and blue value of intensity x.
35410 Default value is "st(0, (midi(f)-59.5)/12); st(1,
35412 if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0)); r(1-ld(1)) +
35413 b(ld(1))".
35414 axisfile
35416 Specify image file to draw the axis. This option override fontfile
35417 and fontcolor option.
35418 axis, text
35420 Enable/disable drawing text to the axis. If it is set to 0, drawing
35421 to the axis is disabled, ignoring fontfile and axisfile option.
35422 Default value is 1.
35423 csp Set colorspace. The accepted values are:
35425 unspecified
35427 Unspecified (default)
35428 bt709
35430 BT.709
35431 fcc FCC
35433 bt470bg
35435 BT.470BG or BT.601-6 625
35436 smpte170m
35438 SMPTE-170M or BT.601-6 525
35439 smpte240m
35441 SMPTE-240M
35442 bt2020ncl
35444 BT.2020 with non-constant luminance
35445 cscheme
35447 Set spectrogram color scheme. This is list of floating point values
35448 with format "left_r|left_g|left_b|right_r|right_g|right_b". The
35449 default is "1|0.5|0|0|0.5|1".
35450 Examples
35452 • Playing audio while showing the spectrum:
35454 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
35456 • Same as above, but with frame rate 30 fps:
35458 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
35460 • Playing at 1280x720:
35462 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
35464 • Disable sonogram display:
35466 sono_h=0
35468 • A1 and its harmonics: A1, A2, (near)E3, A3:
35470 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),
35472 asplit[a][out1]; [a] showcqt [out0]'
35473 • Same as above, but with more accuracy in frequency domain:
35475 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),
35477 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
35478 • Custom volume:
35480 bar_v=10:sono_v=bar_v*a_weighting(f)
35482 • Custom gamma, now spectrum is linear to the amplitude.
35484 bar_g=2:sono_g=2
35486 • Custom tlength equation:
35488 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)))'
35490 • Custom fontcolor and fontfile, C-note is colored green, others are
35492 colored blue:
35493 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
35495 • Custom font using fontconfig:
35497 font='Courier New,Monospace,mono|bold'
35499 • Custom frequency range with custom axis using image file:
35501 axisfile=myaxis.png:basefreq=40:endfreq=10000
35503 showcwt
35505 Convert input audio to video output representing frequency spectrum
35506 using Continuous Wavelet Transform and Morlet wavelet.
35507 The filter accepts the following options:
35509 size, s
35511 Specify the video size for the output. For the syntax of this
35512 option, check the "Video size" section in the ffmpeg-utils manual.
35513 Default value is "640x512".
35514 rate, r
35516 Set the output frame rate. Default value is 25.
35517 scale
35519 Set the frequency scale used. Allowed values are:
35520 linear
35522 log2
35523 bark
35524 mel
35525 erbs
35526 Default value is "linear".
35528 min Set the minimum frequency that will be used in output. Default is
35530 20 Hz.
35531 max Set the maximum frequency that will be used in output. Default is
35533 20000 Hz. The real frequency upper limit depends on input audio's
35534 sample rate and such will be enforced on this value when it is set
35535 to value greater than Nyquist frequency.
35536 logb
35538 Set the logarithmic basis for brightness strength when mapping
35539 calculated magnitude values to pixel values. Allowed range is from
35540 0 to 1. Default value is 0.0001.
35541 deviation
35543 Set the frequency deviation. Lower values than 1 are more
35544 frequency oriented, while higher values than 1 are more time
35545 oriented. Allowed range is from 0 to 10. Default value is 1.
35546 pps Set the number of pixel output per each second in one row. Allowed
35548 range is from 1 to 1024. Default value is 64.
35549 mode
35551 Set the output visual mode. Allowed values are:
35552 magnitude
35554 Show magnitude.
35555 phase
35557 Show only phase.
35558 magphase
35560 Show combination of magnitude and phase. Magnitude is mapped
35561 to brightness and phase to color.
35562 channel
35564 Show unique color per channel magnitude.
35565 stereo
35567 Show unique color per stereo difference.
35568 Default value is "magnitude".
35570 slide
35572 Set the output slide method. Allowed values are:
35573 replace
35575 scroll
35576 frame
35577 direction
35578 Set the direction method for output slide method. Allowed values
35579 are:
35580 lr Direction from left to right.
35582 rl Direction from right to left.
35584 ud Direction from up to down.
35586 du Direction from down to up.
35588 showfreqs
35590 Convert input audio to video output representing the audio power
35591 spectrum. Audio amplitude is on Y-axis while frequency is on X-axis.
35592 The filter accepts the following options:
35594 size, s
35596 Specify size of video. For the syntax of this option, check the
35597 "Video size" section in the ffmpeg-utils manual. Default is
35598 "1024x512".
35599 rate, r
35601 Set video rate. Default is 25.
35602 mode
35604 Set display mode. This set how each frequency bin will be
35605 represented.
35606 It accepts the following values:
35608 line
35610 bar
35611 dot
35612 Default is "bar".
35614 ascale
35616 Set amplitude scale.
35617 It accepts the following values:
35619 lin Linear scale.
35621 sqrt
35623 Square root scale.
35624 cbrt
35626 Cubic root scale.
35627 log Logarithmic scale.
35629 Default is "log".
35631 fscale
35633 Set frequency scale.
35634 It accepts the following values:
35636 lin Linear scale.
35638 log Logarithmic scale.
35640 rlog
35642 Reverse logarithmic scale.
35643 Default is "lin".
35645 win_size
35647 Set window size. Allowed range is from 16 to 65536.
35648 Default is 2048
35650 win_func
35652 Set windowing function.
35653 It accepts the following values:
35655 rect
35657 bartlett
35658 hanning
35659 hamming
35660 blackman
35661 welch
35662 flattop
35663 bharris
35664 bnuttall
35665 bhann
35666 sine
35667 nuttall
35668 lanczos
35669 gauss
35670 tukey
35671 dolph
35672 cauchy
35673 parzen
35674 poisson
35675 bohman
35676 kaiser
35677 Default is "hanning".
35679 overlap
35681 Set window overlap. In range "[0, 1]". Default is 1, which means
35682 optimal overlap for selected window function will be picked.
35683 averaging
35685 Set time averaging. Setting this to 0 will display current maximal
35686 peaks. Default is 1, which means time averaging is disabled.
35687 colors
35689 Specify list of colors separated by space or by '|' which will be
35690 used to draw channel frequencies. Unrecognized or missing colors
35691 will be replaced by white color.
35692 cmode
35694 Set channel display mode.
35695 It accepts the following values:
35697 combined
35699 separate
35700 Default is "combined".
35702 minamp
35704 Set minimum amplitude used in "log" amplitude scaler.
35705 data
35707 Set data display mode.
35708 It accepts the following values:
35710 magnitude
35712 phase
35713 delay
35714 Default is "magnitude".
35716 channels
35718 Set channels to use when processing audio. By default all are
35719 processed.
35720 showspatial
35722 Convert stereo input audio to a video output, representing the spatial
35723 relationship between two channels.
35724 The filter accepts the following options:
35726 size, s
35728 Specify the video size for the output. For the syntax of this
35729 option, check the "Video size" section in the ffmpeg-utils manual.
35730 Default value is "512x512".
35731 win_size
35733 Set window size. Allowed range is from 1024 to 65536. Default size
35734 is 4096.
35735 win_func
35737 Set window function.
35738 It accepts the following values:
35740 rect
35742 bartlett
35743 hann
35744 hanning
35745 hamming
35746 blackman
35747 welch
35748 flattop
35749 bharris
35750 bnuttall
35751 bhann
35752 sine
35753 nuttall
35754 lanczos
35755 gauss
35756 tukey
35757 dolph
35758 cauchy
35759 parzen
35760 poisson
35761 bohman
35762 kaiser
35763 Default value is "hann".
35765 rate, r
35767 Set output framerate.
35768 showspectrum
35770 Convert input audio to a video output, representing the audio frequency
35771 spectrum.
35772 The filter accepts the following options:
35774 size, s
35776 Specify the video size for the output. For the syntax of this
35777 option, check the "Video size" section in the ffmpeg-utils manual.
35778 Default value is "640x512".
35779 slide
35781 Specify how the spectrum should slide along the window.
35782 It accepts the following values:
35784 replace
35786 the samples start again on the left when they reach the right
35787 scroll
35789 the samples scroll from right to left
35790 fullframe
35792 frames are only produced when the samples reach the right
35793 rscroll
35795 the samples scroll from left to right
35796 lreplace
35798 the samples start again on the right when they reach the left
35799 Default value is "replace".
35801 mode
35803 Specify display mode.
35804 It accepts the following values:
35806 combined
35808 all channels are displayed in the same row
35809 separate
35811 all channels are displayed in separate rows
35812 Default value is combined.
35814 color
35816 Specify display color mode.
35817 It accepts the following values:
35819 channel
35821 each channel is displayed in a separate color
35822 intensity
35824 each channel is displayed using the same color scheme
35825 rainbow
35827 each channel is displayed using the rainbow color scheme
35828 moreland
35830 each channel is displayed using the moreland color scheme
35831 nebulae
35833 each channel is displayed using the nebulae color scheme
35834 fire
35836 each channel is displayed using the fire color scheme
35837 fiery
35839 each channel is displayed using the fiery color scheme
35840 fruit
35842 each channel is displayed using the fruit color scheme
35843 cool
35845 each channel is displayed using the cool color scheme
35846 magma
35848 each channel is displayed using the magma color scheme
35849 green
35851 each channel is displayed using the green color scheme
35852 viridis
35854 each channel is displayed using the viridis color scheme
35855 plasma
35857 each channel is displayed using the plasma color scheme
35858 cividis
35860 each channel is displayed using the cividis color scheme
35861 terrain
35863 each channel is displayed using the terrain color scheme
35864 Default value is channel.
35866 scale
35868 Specify scale used for calculating intensity color values.
35869 It accepts the following values:
35871 lin linear
35873 sqrt
35875 square root, default
35876 cbrt
35878 cubic root
35879 log logarithmic
35881 4thrt
35883 4th root
35884 5thrt
35886 5th root
35887 Default value is sqrt.
35889 fscale
35891 Specify frequency scale.
35892 It accepts the following values:
35894 lin linear
35896 log logarithmic
35898 Default value is lin.
35900 saturation
35902 Set saturation modifier for displayed colors. Negative values
35903 provide alternative color scheme. 0 is no saturation at all.
35904 Saturation must be in [-10.0, 10.0] range. Default value is 1.
35905 win_func
35907 Set window function.
35908 It accepts the following values:
35910 rect
35912 bartlett
35913 hann
35914 hanning
35915 hamming
35916 blackman
35917 welch
35918 flattop
35919 bharris
35920 bnuttall
35921 bhann
35922 sine
35923 nuttall
35924 lanczos
35925 gauss
35926 tukey
35927 dolph
35928 cauchy
35929 parzen
35930 poisson
35931 bohman
35932 kaiser
35933 Default value is "hann".
35935 orientation
35937 Set orientation of time vs frequency axis. Can be "vertical" or
35938 "horizontal". Default is "vertical".
35939 overlap
35941 Set ratio of overlap window. Default value is 0. When value is 1
35942 overlap is set to recommended size for specific window function
35943 currently used.
35944 gain
35946 Set scale gain for calculating intensity color values. Default
35947 value is 1.
35948 data
35950 Set which data to display. Can be "magnitude", default or "phase",
35951 or unwrapped phase: "uphase".
35952 rotation
35954 Set color rotation, must be in [-1.0, 1.0] range. Default value is
35955 0.
35956 start
35958 Set start frequency from which to display spectrogram. Default is
35959 0.
35960 stop
35962 Set stop frequency to which to display spectrogram. Default is 0.
35963 fps Set upper frame rate limit. Default is "auto", unlimited.
35965 legend
35967 Draw time and frequency axes and legends. Default is disabled.
35968 drange
35970 Set dynamic range used to calculate intensity color values. Default
35971 is 120 dBFS. Allowed range is from 10 to 200.
35972 limit
35974 Set upper limit of input audio samples volume in dBFS. Default is 0
35975 dBFS. Allowed range is from -100 to 100.
35976 opacity
35978 Set opacity strength when using pixel format output with alpha
35979 component.
35980 The usage is very similar to the showwaves filter; see the examples in
35982 that section.
35983 Examples
35985 • Large window with logarithmic color scaling:
35987 showspectrum=s=1280x480:scale=log
35989 • Complete example for a colored and sliding spectrum per channel
35991 using ffplay:
35992 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
35994 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
35995 showspectrumpic
35997 Convert input audio to a single video frame, representing the audio
35998 frequency spectrum.
35999 The filter accepts the following options:
36001 size, s
36003 Specify the video size for the output. For the syntax of this
36004 option, check the "Video size" section in the ffmpeg-utils manual.
36005 Default value is "4096x2048".
36006 mode
36008 Specify display mode.
36009 It accepts the following values:
36011 combined
36013 all channels are displayed in the same row
36014 separate
36016 all channels are displayed in separate rows
36017 Default value is combined.
36019 color
36021 Specify display color mode.
36022 It accepts the following values:
36024 channel
36026 each channel is displayed in a separate color
36027 intensity
36029 each channel is displayed using the same color scheme
36030 rainbow
36032 each channel is displayed using the rainbow color scheme
36033 moreland
36035 each channel is displayed using the moreland color scheme
36036 nebulae
36038 each channel is displayed using the nebulae color scheme
36039 fire
36041 each channel is displayed using the fire color scheme
36042 fiery
36044 each channel is displayed using the fiery color scheme
36045 fruit
36047 each channel is displayed using the fruit color scheme
36048 cool
36050 each channel is displayed using the cool color scheme
36051 magma
36053 each channel is displayed using the magma color scheme
36054 green
36056 each channel is displayed using the green color scheme
36057 viridis
36059 each channel is displayed using the viridis color scheme
36060 plasma
36062 each channel is displayed using the plasma color scheme
36063 cividis
36065 each channel is displayed using the cividis color scheme
36066 terrain
36068 each channel is displayed using the terrain color scheme
36069 Default value is intensity.
36071 scale
36073 Specify scale used for calculating intensity color values.
36074 It accepts the following values:
36076 lin linear
36078 sqrt
36080 square root, default
36081 cbrt
36083 cubic root
36084 log logarithmic
36086 4thrt
36088 4th root
36089 5thrt
36091 5th root
36092 Default value is log.
36094 fscale
36096 Specify frequency scale.
36097 It accepts the following values:
36099 lin linear
36101 log logarithmic
36103 Default value is lin.
36105 saturation
36107 Set saturation modifier for displayed colors. Negative values
36108 provide alternative color scheme. 0 is no saturation at all.
36109 Saturation must be in [-10.0, 10.0] range. Default value is 1.
36110 win_func
36112 Set window function.
36113 It accepts the following values:
36115 rect
36117 bartlett
36118 hann
36119 hanning
36120 hamming
36121 blackman
36122 welch
36123 flattop
36124 bharris
36125 bnuttall
36126 bhann
36127 sine
36128 nuttall
36129 lanczos
36130 gauss
36131 tukey
36132 dolph
36133 cauchy
36134 parzen
36135 poisson
36136 bohman
36137 kaiser
36138 Default value is "hann".
36140 orientation
36142 Set orientation of time vs frequency axis. Can be "vertical" or
36143 "horizontal". Default is "vertical".
36144 gain
36146 Set scale gain for calculating intensity color values. Default
36147 value is 1.
36148 legend
36150 Draw time and frequency axes and legends. Default is enabled.
36151 rotation
36153 Set color rotation, must be in [-1.0, 1.0] range. Default value is
36154 0.
36155 start
36157 Set start frequency from which to display spectrogram. Default is
36158 0.
36159 stop
36161 Set stop frequency to which to display spectrogram. Default is 0.
36162 drange
36164 Set dynamic range used to calculate intensity color values. Default
36165 is 120 dBFS. Allowed range is from 10 to 200.
36166 limit
36168 Set upper limit of input audio samples volume in dBFS. Default is 0
36169 dBFS. Allowed range is from -100 to 100.
36170 opacity
36172 Set opacity strength when using pixel format output with alpha
36173 component.
36174 Examples
36176 • Extract an audio spectrogram of a whole audio track in a 1024x1024
36178 picture using ffmpeg:
36179 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
36181 showvolume
36183 Convert input audio volume to a video output.
36184 The filter accepts the following options:
36186 rate, r
36188 Set video rate.
36189 b Set border width, allowed range is [0, 5]. Default is 1.
36191 w Set channel width, allowed range is [80, 8192]. Default is 400.
36193 h Set channel height, allowed range is [1, 900]. Default is 20.
36195 f Set fade, allowed range is [0, 1]. Default is 0.95.
36197 c Set volume color expression.
36199 The expression can use the following variables:
36201 VOLUME
36203 Current max volume of channel in dB.
36204 PEAK
36206 Current peak.
36207 CHANNEL
36209 Current channel number, starting from 0.
36210 t If set, displays channel names. Default is enabled.
36212 v If set, displays volume values. Default is enabled.
36214 o Set orientation, can be horizontal: "h" or vertical: "v", default
36216 is "h".
36217 s Set step size, allowed range is [0, 5]. Default is 0, which means
36219 step is disabled.
36220 p Set background opacity, allowed range is [0, 1]. Default is 0.
36222 m Set metering mode, can be peak: "p" or rms: "r", default is "p".
36224 ds Set display scale, can be linear: "lin" or log: "log", default is
36226 "lin".
36227 dm In second. If set to > 0., display a line for the max level in the
36229 previous seconds. default is disabled: 0.
36230 dmc The color of the max line. Use when "dm" option is set to > 0.
36232 default is: "orange"
36233 showwaves
36235 Convert input audio to a video output, representing the samples waves.
36236 The filter accepts the following options:
36238 size, s
36240 Specify the video size for the output. For the syntax of this
36241 option, check the "Video size" section in the ffmpeg-utils manual.
36242 Default value is "600x240".
36243 mode
36245 Set display mode.
36246 Available values are:
36248 point
36250 Draw a point for each sample.
36251 line
36253 Draw a vertical line for each sample.
36254 p2p Draw a point for each sample and a line between them.
36256 cline
36258 Draw a centered vertical line for each sample.
36259 Default value is "point".
36261 n Set the number of samples which are printed on the same column. A
36263 larger value will decrease the frame rate. Must be a positive
36264 integer. This option can be set only if the value for rate is not
36265 explicitly specified.
36266 rate, r
36268 Set the (approximate) output frame rate. This is done by setting
36269 the option n. Default value is "25".
36270 split_channels
36272 Set if channels should be drawn separately or overlap. Default
36273 value is 0.
36274 colors
36276 Set colors separated by '|' which are going to be used for drawing
36277 of each channel.
36278 scale
36280 Set amplitude scale.
36281 Available values are:
36283 lin Linear.
36285 log Logarithmic.
36287 sqrt
36289 Square root.
36290 cbrt
36292 Cubic root.
36293 Default is linear.
36295 draw
36297 Set the draw mode. This is mostly useful to set for high n.
36298 Available values are:
36300 scale
36302 Scale pixel values for each drawn sample.
36303 full
36305 Draw every sample directly.
36306 Default value is "scale".
36308 Examples
36310 • Output the input file audio and the corresponding video
36312 representation at the same time:
36313 amovie=a.mp3,asplit[out0],showwaves[out1]
36315 • Create a synthetic signal and show it with showwaves, forcing a
36317 frame rate of 30 frames per second:
36318 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
36320 showwavespic
36322 Convert input audio to a single video frame, representing the samples
36323 waves.
36324 The filter accepts the following options:
36326 size, s
36328 Specify the video size for the output. For the syntax of this
36329 option, check the "Video size" section in the ffmpeg-utils manual.
36330 Default value is "600x240".
36331 split_channels
36333 Set if channels should be drawn separately or overlap. Default
36334 value is 0.
36335 colors
36337 Set colors separated by '|' which are going to be used for drawing
36338 of each channel.
36339 scale
36341 Set amplitude scale.
36342 Available values are:
36344 lin Linear.
36346 log Logarithmic.
36348 sqrt
36350 Square root.
36351 cbrt
36353 Cubic root.
36354 Default is linear.
36356 draw
36358 Set the draw mode.
36359 Available values are:
36361 scale
36363 Scale pixel values for each drawn sample.
36364 full
36366 Draw every sample directly.
36367 Default value is "scale".
36369 filter
36371 Set the filter mode.
36372 Available values are:
36374 average
36376 Use average samples values for each drawn sample.
36377 peak
36379 Use peak samples values for each drawn sample.
36380 Default value is "average".
36382 Examples
36384 • Extract a channel split representation of the wave form of a whole
36386 audio track in a 1024x800 picture using ffmpeg:
36387 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
36389 sidedata, asidedata
36391 Delete frame side data, or select frames based on it.
36392 This filter accepts the following options:
36394 mode
36396 Set mode of operation of the filter.
36397 Can be one of the following:
36399 select
36401 Select every frame with side data of "type".
36402 delete
36404 Delete side data of "type". If "type" is not set, delete all
36405 side data in the frame.
36406 type
36408 Set side data type used with all modes. Must be set for "select"
36409 mode. For the list of frame side data types, refer to the
36410 "AVFrameSideDataType" enum in libavutil/frame.h. For example, to
36411 choose "AV_FRAME_DATA_PANSCAN" side data, you must specify
36412 "PANSCAN".
36413 spectrumsynth
36415 Synthesize audio from 2 input video spectrums, first input stream
36416 represents magnitude across time and second represents phase across
36417 time. The filter will transform from frequency domain as displayed in
36418 videos back to time domain as presented in audio output.
36419 This filter is primarily created for reversing processed showspectrum
36421 filter outputs, but can synthesize sound from other spectrograms too.
36422 But in such case results are going to be poor if the phase data is not
36423 available, because in such cases phase data need to be recreated,
36424 usually it's just recreated from random noise. For best results use
36425 gray only output ("channel" color mode in showspectrum filter) and
36426 "log" scale for magnitude video and "lin" scale for phase video. To
36427 produce phase, for 2nd video, use "data" option. Inputs videos should
36428 generally use "fullframe" slide mode as that saves resources needed for
36429 decoding video.
36430 The filter accepts the following options:
36432 sample_rate
36434 Specify sample rate of output audio, the sample rate of audio from
36435 which spectrum was generated may differ.
36436 channels
36438 Set number of channels represented in input video spectrums.
36439 scale
36441 Set scale which was used when generating magnitude input spectrum.
36442 Can be "lin" or "log". Default is "log".
36443 slide
36445 Set slide which was used when generating inputs spectrums. Can be
36446 "replace", "scroll", "fullframe" or "rscroll". Default is
36447 "fullframe".
36448 win_func
36450 Set window function used for resynthesis.
36451 overlap
36453 Set window overlap. In range "[0, 1]". Default is 1, which means
36454 optimal overlap for selected window function will be picked.
36455 orientation
36457 Set orientation of input videos. Can be "vertical" or "horizontal".
36458 Default is "vertical".
36459 Examples
36461 • First create magnitude and phase videos from audio, assuming audio
36463 is stereo with 44100 sample rate, then resynthesize videos back to
36464 audio with spectrumsynth:
36465 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
36467 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
36468 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
36469 split, asplit
36471 Split input into several identical outputs.
36472 "asplit" works with audio input, "split" with video.
36474 The filter accepts a single parameter which specifies the number of
36476 outputs. If unspecified, it defaults to 2.
36477 Examples
36479 • Create two separate outputs from the same input:
36481 [in] split [out0][out1]
36483 • To create 3 or more outputs, you need to specify the number of
36485 outputs, like in:
36486 [in] asplit=3 [out0][out1][out2]
36488 • Create two separate outputs from the same input, one cropped and
36490 one padded:
36491 [in] split [splitout1][splitout2];
36493 [splitout1] crop=100:100:0:0 [cropout];
36494 [splitout2] pad=200:200:100:100 [padout];
36495 • Create 5 copies of the input audio with ffmpeg:
36497 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
36499 zmq, azmq
36501 Receive commands sent through a libzmq client, and forward them to
36502 filters in the filtergraph.
36503 "zmq" and "azmq" work as a pass-through filters. "zmq" must be inserted
36505 between two video filters, "azmq" between two audio filters. Both are
36506 capable to send messages to any filter type.
36507 To enable these filters you need to install the libzmq library and
36509 headers and configure FFmpeg with "--enable-libzmq".
36510 For more information about libzmq see: <http://www.zeromq.org/>
36512 The "zmq" and "azmq" filters work as a libzmq server, which receives
36514 messages sent through a network interface defined by the bind_address
36515 (or the abbreviation "b") option. Default value of this option is
36516 tcp://localhost:5555. You may want to alter this value to your needs,
36517 but do not forget to escape any ':' signs (see filtergraph escaping).
36518 The received message must be in the form:
36520 <TARGET> <COMMAND> [<ARG>]
36522 TARGET specifies the target of the command, usually the name of the
36524 filter class or a specific filter instance name. The default filter
36525 instance name uses the pattern Parsed_<filter_name>_<index>, but you
36526 can override this by using the filter_name@id syntax (see Filtergraph
36527 syntax).
36528 COMMAND specifies the name of the command for the target filter.
36530 ARG is optional and specifies the optional argument list for the given
36532 COMMAND.
36533 Upon reception, the message is processed and the corresponding command
36535 is injected into the filtergraph. Depending on the result, the filter
36536 will send a reply to the client, adopting the format:
36537 <ERROR_CODE> <ERROR_REASON>
36539 <MESSAGE>
36540 MESSAGE is optional.
36542 Examples
36544 Look at tools/zmqsend for an example of a zmq client which can be used
36546 to send commands processed by these filters.
36547 Consider the following filtergraph generated by ffplay. In this
36549 example the last overlay filter has an instance name. All other filters
36550 will have default instance names.
36551 ffplay -dumpgraph 1 -f lavfi "
36553 color=s=100x100:c=red [l];
36554 color=s=100x100:c=blue [r];
36555 nullsrc=s=200x100, zmq [bg];
36556 [bg][l] overlay [bg+l];
36557 [bg+l][r] overlay@my=x=100 "
36558 To change the color of the left side of the video, the following
36560 command can be used:
36561 echo Parsed_color_0 c yellow | tools/zmqsend
36563 To change the right side:
36565 echo Parsed_color_1 c pink | tools/zmqsend
36567 To change the position of the right side:
36569 echo overlay@my x 150 | tools/zmqsend
36571
36573 Below is a description of the currently available multimedia sources.
36574 amovie
36576 This is the same as movie source, except it selects an audio stream by
36577 default.
36578 avsynctest
36580 Generate an Audio/Video Sync Test.
36581 Generated stream periodically shows flash video frame and emits beep in
36583 audio. Useful to inspect A/V sync issues.
36584 It accepts the following options:
36586 size, s
36588 Set output video size. Default value is "hd720".
36589 framerate, fr
36591 Set output video frame rate. Default value is 30.
36592 samplerate, sr
36594 Set output audio sample rate. Default value is 44100.
36595 amplitude, a
36597 Set output audio beep amplitude. Default value is 0.7.
36598 period, p
36600 Set output audio beep period in seconds. Default value is 3.
36601 delay, dl
36603 Set output video flash delay in number of frames. Default value is
36604 0.
36605 cycle, c
36607 Enable cycling of video delays, by default is disabled.
36608 duration, d
36610 Set stream output duration. By default duration is unlimited.
36611 fg, bg, ag
36613 Set foreground/background/additional color.
36614 movie
36616 Read audio and/or video stream(s) from a movie container.
36617 It accepts the following parameters:
36619 filename
36621 The name of the resource to read (not necessarily a file; it can
36622 also be a device or a stream accessed through some protocol).
36623 format_name, f
36625 Specifies the format assumed for the movie to read, and can be
36626 either the name of a container or an input device. If not
36627 specified, the format is guessed from movie_name or by probing.
36628 seek_point, sp
36630 Specifies the seek point in seconds. The frames will be output
36631 starting from this seek point. The parameter is evaluated with
36632 "av_strtod", so the numerical value may be suffixed by an IS
36633 postfix. The default value is "0".
36634 streams, s
36636 Specifies the streams to read. Several streams can be specified,
36637 separated by "+". The source will then have as many outputs, in the
36638 same order. The syntax is explained in the "Stream specifiers"
36639 section in the ffmpeg manual. Two special names, "dv" and "da"
36640 specify respectively the default (best suited) video and audio
36641 stream. Default is "dv", or "da" if the filter is called as
36642 "amovie".
36643 stream_index, si
36645 Specifies the index of the video stream to read. If the value is
36646 -1, the most suitable video stream will be automatically selected.
36647 The default value is "-1". Deprecated. If the filter is called
36648 "amovie", it will select audio instead of video.
36649 loop
36651 Specifies how many times to read the stream in sequence. If the
36652 value is 0, the stream will be looped infinitely. Default value is
36653 "1".
36654 Note that when the movie is looped the source timestamps are not
36656 changed, so it will generate non monotonically increasing
36657 timestamps.
36658 discontinuity
36660 Specifies the time difference between frames above which the point
36661 is considered a timestamp discontinuity which is removed by
36662 adjusting the later timestamps.
36663 dec_threads
36665 Specifies the number of threads for decoding
36666 format_opts
36668 Specify format options for the opened file. Format options can be
36669 specified as a list of key=value pairs separated by ':'. The
36670 following example shows how to add protocol_whitelist and
36671 protocol_blacklist options:
36672 ffplay -f lavfi
36674 "movie=filename='1.sdp':format_opts='protocol_whitelist=file,rtp,udp\:protocol_blacklist=http'"
36675 It allows overlaying a second video on top of the main input of a
36677 filtergraph, as shown in this graph:
36678 input -----------> deltapts0 --> overlay --> output
36680 ^
36681 |
36682 movie --> scale--> deltapts1 -------+
36683 Examples
36685 • Skip 3.2 seconds from the start of the AVI file in.avi, and overlay
36687 it on top of the input labelled "in":
36688 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
36690 [in] setpts=PTS-STARTPTS [main];
36691 [main][over] overlay=16:16 [out]
36692 • Read from a video4linux2 device, and overlay it on top of the input
36694 labelled "in":
36695 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
36697 [in] setpts=PTS-STARTPTS [main];
36698 [main][over] overlay=16:16 [out]
36699 • Read the first video stream and the audio stream with id 0x81 from
36701 dvd.vob; the video is connected to the pad named "video" and the
36702 audio is connected to the pad named "audio":
36703 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
36705 Commands
36707 Both movie and amovie support the following commands:
36709 seek
36711 Perform seek using "av_seek_frame". The syntax is: seek
36712 stream_index|timestamp|flags
36713 • stream_index: If stream_index is -1, a default stream is
36715 selected, and timestamp is automatically converted from
36716 AV_TIME_BASE units to the stream specific time_base.
36717 • timestamp: Timestamp in AVStream.time_base units or, if no
36719 stream is specified, in AV_TIME_BASE units.
36720 • flags: Flags which select direction and seeking mode.
36722 get_duration
36724 Get movie duration in AV_TIME_BASE units.
36725
36727 FFmpeg can be hooked up with a number of external libraries to add
36728 support for more formats. None of them are used by default, their use
36729 has to be explicitly requested by passing the appropriate flags to
36730 ./configure.
36731 Alliance for Open Media (AOM)
36733 FFmpeg can make use of the AOM library for AV1 decoding and encoding.
36734 Go to <http://aomedia.org/> and follow the instructions for installing
36736 the library. Then pass "--enable-libaom" to configure to enable it.
36737 AMD AMF/VCE
36739 FFmpeg can use the AMD Advanced Media Framework library for accelerated
36740 H.264 and HEVC(only windows) encoding on hardware with Video Coding
36741 Engine (VCE).
36742 To enable support you must obtain the AMF framework header
36744 files(version 1.4.9+) from
36745 <https://github.com/GPUOpen-LibrariesAndSDKs/AMF.git>.
36746 Create an "AMF/" directory in the system include path. Copy the
36748 contents of "AMF/amf/public/include/" into that directory. Then
36749 configure FFmpeg with "--enable-amf".
36750 Initialization of amf encoder occurs in this order: 1) trying to
36752 initialize through dx11(only windows) 2) trying to initialize through
36753 dx9(only windows) 3) trying to initialize through vulkan
36754 To use h.264(AMD VCE) encoder on linux amdgru-pro version 19.20+ and
36756 amf-amdgpu-pro package(amdgru-pro contains, but does not install
36757 automatically) are required.
36758 This driver can be installed using amdgpu-pro-install script in
36760 official amd driver archive.
36761 AviSynth
36763 FFmpeg can read AviSynth scripts as input. To enable support, pass
36764 "--enable-avisynth" to configure after installing the headers provided
36765 by <https://github.com/AviSynth/AviSynthPlus>. AviSynth+ can be
36766 configured to install only the headers by either passing
36767 "-DHEADERS_ONLY:bool=on" to the normal CMake-based build system, or by
36768 using the supplied "GNUmakefile".
36769 For Windows, supported AviSynth variants are <http://avisynth.nl> for
36771 32-bit builds and <http://avisynth.nl/index.php/AviSynth+> for 32-bit
36772 and 64-bit builds.
36773 For Linux, macOS, and BSD, the only supported AviSynth variant is
36775 <https://github.com/AviSynth/AviSynthPlus>, starting with version 3.5.
36776 In 2016, AviSynth+ added support for building with GCC. However,
36778 due to the eccentricities of Windows' calling conventions, 32-bit
36779 GCC builds of AviSynth+ are not compatible with typical 32-bit
36780 builds of FFmpeg.
36781 By default, FFmpeg assumes compatibility with 32-bit MSVC builds of
36783 AviSynth+ since that is the most widely-used and entrenched build
36784 configuration. Users can override this and enable support for
36785 32-bit GCC builds of AviSynth+ by passing "-DAVSC_WIN32_GCC32" to
36786 "--extra-cflags" when configuring FFmpeg.
36787 64-bit builds of FFmpeg are not affected, and can use either MSVC
36789 or GCC builds of AviSynth+ without any special flags.
36790 AviSynth(+) is loaded dynamically. Distributors can build FFmpeg
36792 with "--enable-avisynth", and the binaries will work regardless of
36793 the end user having AviSynth installed. If/when an end user would
36794 like to use AviSynth scripts, then they can install AviSynth(+) and
36795 FFmpeg will be able to find and use it to open scripts.
36796 Chromaprint
36798 FFmpeg can make use of the Chromaprint library for generating audio
36799 fingerprints. Pass "--enable-chromaprint" to configure to enable it.
36800 See <https://acoustid.org/chromaprint>.
36801 codec2
36803 FFmpeg can make use of the codec2 library for codec2 decoding and
36804 encoding. There is currently no native decoder, so libcodec2 must be
36805 used for decoding.
36806 Go to <http://freedv.org/>, download "Codec 2 source archive". Build
36808 and install using CMake. Debian users can install the libcodec2-dev
36809 package instead. Once libcodec2 is installed you can pass
36810 "--enable-libcodec2" to configure to enable it.
36811 The easiest way to use codec2 is with .c2 files, since they contain the
36813 mode information required for decoding. To encode such a file, use a
36814 .c2 file extension and give the libcodec2 encoder the -mode option:
36815 "ffmpeg -i input.wav -mode 700C output.c2". Playback is as simple as
36816 "ffplay output.c2". For a list of supported modes, run "ffmpeg -h
36817 encoder=libcodec2". Raw codec2 files are also supported. To make
36818 sense of them the mode in use needs to be specified as a format option:
36819 "ffmpeg -f codec2raw -mode 1300 -i input.raw output.wav".
36820 dav1d
36822 FFmpeg can make use of the dav1d library for AV1 video decoding.
36823 Go to <https://code.videolan.org/videolan/dav1d> and follow the
36825 instructions for installing the library. Then pass "--enable-libdav1d"
36826 to configure to enable it.
36827 davs2
36829 FFmpeg can make use of the davs2 library for AVS2-P2/IEEE1857.4 video
36830 decoding.
36831 Go to <https://github.com/pkuvcl/davs2> and follow the instructions for
36833 installing the library. Then pass "--enable-libdavs2" to configure to
36834 enable it.
36835 libdavs2 is under the GNU Public License Version 2 or later (see
36837 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
36838 details), you must upgrade FFmpeg's license to GPL in order to use
36839 it.
36840 uavs3d
36842 FFmpeg can make use of the uavs3d library for AVS3-P2/IEEE1857.10 video
36843 decoding.
36844 Go to <https://github.com/uavs3/uavs3d> and follow the instructions for
36846 installing the library. Then pass "--enable-libuavs3d" to configure to
36847 enable it.
36848 Game Music Emu
36850 FFmpeg can make use of the Game Music Emu library to read audio from
36851 supported video game music file formats. Pass "--enable-libgme" to
36852 configure to enable it. See
36853 <https://bitbucket.org/mpyne/game-music-emu/overview>.
36854 Intel QuickSync Video
36856 FFmpeg can use Intel QuickSync Video (QSV) for accelerated decoding and
36857 encoding of multiple codecs. To use QSV, FFmpeg must be linked against
36858 the "libmfx" dispatcher, which loads the actual decoding libraries.
36859 The dispatcher is open source and can be downloaded from
36861 <https://github.com/lu-zero/mfx_dispatch.git>. FFmpeg needs to be
36862 configured with the "--enable-libmfx" option and "pkg-config" needs to
36863 be able to locate the dispatcher's ".pc" files.
36864 Kvazaar
36866 FFmpeg can make use of the Kvazaar library for HEVC encoding.
36867 Go to <https://github.com/ultravideo/kvazaar> and follow the
36869 instructions for installing the library. Then pass
36870 "--enable-libkvazaar" to configure to enable it.
36871 LAME
36873 FFmpeg can make use of the LAME library for MP3 encoding.
36874 Go to <http://lame.sourceforge.net/> and follow the instructions for
36876 installing the library. Then pass "--enable-libmp3lame" to configure
36877 to enable it.
36878 libilbc
36880 iLBC is a narrowband speech codec that has been made freely available
36881 by Google as part of the WebRTC project. libilbc is a packaging
36882 friendly copy of the iLBC codec. FFmpeg can make use of the libilbc
36883 library for iLBC decoding and encoding.
36884 Go to <https://github.com/TimothyGu/libilbc> and follow the
36886 instructions for installing the library. Then pass "--enable-libilbc"
36887 to configure to enable it.
36888 libjxl
36890 JPEG XL is an image format intended to fully replace legacy JPEG for an
36891 extended period of life. See <https://jpegxl.info/> for more
36892 information, and see <https://github.com/libjxl/libjxl> for the library
36893 source. You can pass "--enable-libjxl" to configure in order enable the
36894 libjxl wrapper.
36895 libvpx
36897 FFmpeg can make use of the libvpx library for VP8/VP9 decoding and
36898 encoding.
36899 Go to <http://www.webmproject.org/> and follow the instructions for
36901 installing the library. Then pass "--enable-libvpx" to configure to
36902 enable it.
36903 ModPlug
36905 FFmpeg can make use of this library, originating in Modplug-XMMS, to
36906 read from MOD-like music files. See
36907 <https://github.com/Konstanty/libmodplug>. Pass "--enable-libmodplug"
36908 to configure to enable it.
36909 OpenCORE, VisualOn, and Fraunhofer libraries
36911 Spun off Google Android sources, OpenCore, VisualOn and Fraunhofer
36912 libraries provide encoders for a number of audio codecs.
36913 OpenCORE and VisualOn libraries are under the Apache License 2.0
36915 (see <http://www.apache.org/licenses/LICENSE-2.0> for details),
36916 which is incompatible to the LGPL version 2.1 and GPL version 2.
36917 You have to upgrade FFmpeg's license to LGPL version 3 (or if you
36918 have enabled GPL components, GPL version 3) by passing
36919 "--enable-version3" to configure in order to use it.
36920 The license of the Fraunhofer AAC library is incompatible with the
36922 GPL. Therefore, for GPL builds, you have to pass
36923 "--enable-nonfree" to configure in order to use it. To the best of
36924 our knowledge, it is compatible with the LGPL.
36925 OpenCORE AMR
36927 FFmpeg can make use of the OpenCORE libraries for AMR-NB
36929 decoding/encoding and AMR-WB decoding.
36930 Go to <http://sourceforge.net/projects/opencore-amr/> and follow the
36932 instructions for installing the libraries. Then pass
36933 "--enable-libopencore-amrnb" and/or "--enable-libopencore-amrwb" to
36934 configure to enable them.
36935 VisualOn AMR-WB encoder library
36937 FFmpeg can make use of the VisualOn AMR-WBenc library for AMR-WB
36939 encoding.
36940 Go to <http://sourceforge.net/projects/opencore-amr/> and follow the
36942 instructions for installing the library. Then pass
36943 "--enable-libvo-amrwbenc" to configure to enable it.
36944 Fraunhofer AAC library
36946 FFmpeg can make use of the Fraunhofer AAC library for AAC decoding &
36948 encoding.
36949 Go to <http://sourceforge.net/projects/opencore-amr/> and follow the
36951 instructions for installing the library. Then pass
36952 "--enable-libfdk-aac" to configure to enable it.
36953 OpenH264
36955 FFmpeg can make use of the OpenH264 library for H.264 decoding and
36956 encoding.
36957 Go to <http://www.openh264.org/> and follow the instructions for
36959 installing the library. Then pass "--enable-libopenh264" to configure
36960 to enable it.
36961 For decoding, this library is much more limited than the built-in
36963 decoder in libavcodec; currently, this library lacks support for
36964 decoding B-frames and some other main/high profile features. (It
36965 currently only supports constrained baseline profile and CABAC.) Using
36966 it is mostly useful for testing and for taking advantage of Cisco's
36967 patent portfolio license
36968 (<http://www.openh264.org/BINARY_LICENSE.txt>).
36969 OpenJPEG
36971 FFmpeg can use the OpenJPEG libraries for decoding/encoding J2K videos.
36972 Go to <http://www.openjpeg.org/> to get the libraries and follow the
36973 installation instructions. To enable using OpenJPEG in FFmpeg, pass
36974 "--enable-libopenjpeg" to ./configure.
36975 rav1e
36977 FFmpeg can make use of rav1e (Rust AV1 Encoder) via its C bindings to
36978 encode videos. Go to <https://github.com/xiph/rav1e/> and follow the
36979 instructions to build the C library. To enable using rav1e in FFmpeg,
36980 pass "--enable-librav1e" to ./configure.
36981 SVT-AV1
36983 FFmpeg can make use of the Scalable Video Technology for AV1 library
36984 for AV1 encoding.
36985 Go to <https://gitlab.com/AOMediaCodec/SVT-AV1/> and follow the
36987 instructions for installing the library. Then pass "--enable-libsvtav1"
36988 to configure to enable it.
36989 TwoLAME
36991 FFmpeg can make use of the TwoLAME library for MP2 encoding.
36992 Go to <http://www.twolame.org/> and follow the instructions for
36994 installing the library. Then pass "--enable-libtwolame" to configure
36995 to enable it.
36996 VapourSynth
36998 FFmpeg can read VapourSynth scripts as input. To enable support, pass
36999 "--enable-vapoursynth" to configure. Vapoursynth is detected via
37000 "pkg-config". Versions 42 or greater supported. See
37001 <http://www.vapoursynth.com/>.
37002 Due to security concerns, Vapoursynth scripts will not be autodetected
37004 so the input format has to be forced. For ff* CLI tools, add "-f
37005 vapoursynth" before the input "-i yourscript.vpy".
37006 x264
37008 FFmpeg can make use of the x264 library for H.264 encoding.
37009 Go to <http://www.videolan.org/developers/x264.html> and follow the
37011 instructions for installing the library. Then pass "--enable-libx264"
37012 to configure to enable it.
37013 x264 is under the GNU Public License Version 2 or later (see
37015 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
37016 details), you must upgrade FFmpeg's license to GPL in order to use
37017 it.
37018 x265
37020 FFmpeg can make use of the x265 library for HEVC encoding.
37021 Go to <http://x265.org/developers.html> and follow the instructions for
37023 installing the library. Then pass "--enable-libx265" to configure to
37024 enable it.
37025 x265 is under the GNU Public License Version 2 or later (see
37027 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
37028 details), you must upgrade FFmpeg's license to GPL in order to use
37029 it.
37030 xavs
37032 FFmpeg can make use of the xavs library for AVS encoding.
37033 Go to <http://xavs.sf.net/> and follow the instructions for installing
37035 the library. Then pass "--enable-libxavs" to configure to enable it.
37036 xavs2
37038 FFmpeg can make use of the xavs2 library for AVS2-P2/IEEE1857.4 video
37039 encoding.
37040 Go to <https://github.com/pkuvcl/xavs2> and follow the instructions for
37042 installing the library. Then pass "--enable-libxavs2" to configure to
37043 enable it.
37044 libxavs2 is under the GNU Public License Version 2 or later (see
37046 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
37047 details), you must upgrade FFmpeg's license to GPL in order to use
37048 it.
37049 ZVBI
37051 ZVBI is a VBI decoding library which can be used by FFmpeg to decode
37052 DVB teletext pages and DVB teletext subtitles.
37053 Go to <http://sourceforge.net/projects/zapping/> and follow the
37055 instructions for installing the library. Then pass "--enable-libzvbi"
37056 to configure to enable it.
37057
37059 You can use the "-formats" and "-codecs" options to have an exhaustive
37060 list.
37061 File Formats
37063 FFmpeg supports the following file formats through the "libavformat"
37064 library:
37065 Name : Encoding @tab Decoding @tab Comments
37067 3dostr : @tab X
37068 4xm : @tab X
37069 @tab 4X Technologies format, used in some games.
37070 8088flex TMV : @tab X
37072 AAX : @tab X
37073 @tab Audible Enhanced Audio format, used in audiobooks.
37074 AA : @tab X
37076 @tab Audible Format 2, 3, and 4, used in audiobooks.
37077 ACT Voice : @tab X
37079 @tab contains G.729 audio
37080 Adobe Filmstrip : X @tab X
37082 Audio IFF (AIFF) : X @tab X
37083 American Laser Games MM : @tab X
37084 @tab Multimedia format used in games like Mad Dog McCree.
37085 3GPP AMR : X @tab X
37087 Amazing Studio Packed Animation File : @tab X
37088 @tab Multimedia format used in game Heart Of Darkness.
37089 Apple HTTP Live Streaming : @tab X
37091 Artworx Data Format : @tab X
37092 Interplay ACM : @tab X
37093 @tab Audio only format used in some Interplay games.
37094 ADP : @tab X
37096 @tab Audio format used on the Nintendo Gamecube.
37097 AFC : @tab X
37099 @tab Audio format used on the Nintendo Gamecube.
37100 ADS/SS2 : @tab X
37102 @tab Audio format used on the PS2.
37103 APNG : X @tab X
37105 ASF : X @tab X
37106 @tab Advanced / Active Streaming Format.
37107 AST : X @tab X
37109 @tab Audio format used on the Nintendo Wii.
37110 AVI : X @tab X
37112 AviSynth : @tab X
37113 AVR : @tab X
37114 @tab Audio format used on Mac.
37115 AVS : @tab X
37117 @tab Multimedia format used by the Creature Shock game.
37118 Beam Software SIFF : @tab X
37120 @tab Audio and video format used in some games by Beam Software.
37121 Bethesda Softworks VID : @tab X
37123 @tab Used in some games from Bethesda Softworks.
37124 Binary text : @tab X
37126 Bink : @tab X
37127 @tab Multimedia format used by many games.
37128 Bink Audio : @tab X
37130 @tab Audio only multimedia format used by some games.
37131 Bitmap Brothers JV : @tab X
37133 @tab Used in Z and Z95 games.
37134 BRP : @tab X
37136 @tab Argonaut Games format.
37137 Brute Force & Ignorance : @tab X
37139 @tab Used in the game Flash Traffic: City of Angels.
37140 BFSTM : @tab X
37142 @tab Audio format used on the Nintendo WiiU (based on BRSTM).
37143 BRSTM : @tab X
37145 @tab Audio format used on the Nintendo Wii.
37146 BW64 : @tab X
37148 @tab Broadcast Wave 64bit.
37149 BWF : X @tab X
37151 codec2 (raw) : X @tab X
37152 @tab Must be given -mode format option to decode correctly.
37153 codec2 (.c2 files) : X @tab X
37155 @tab Contains header with version and mode info, simplifying playback.
37156 CRI ADX : X @tab X
37158 @tab Audio-only format used in console video games.
37159 CRI AIX : @tab X
37161 CRI HCA : @tab X
37162 @tab Audio-only format used in console video games.
37163 Discworld II BMV : @tab X
37165 Interplay C93 : @tab X
37166 @tab Used in the game Cyberia from Interplay.
37167 Delphine Software International CIN : @tab X
37169 @tab Multimedia format used by Delphine Software games.
37170 Digital Speech Standard (DSS) : @tab X
37172 CD+G : @tab X
37173 @tab Video format used by CD+G karaoke disks
37174 Phantom Cine : @tab X
37176 Commodore CDXL : @tab X
37177 @tab Amiga CD video format
37178 Core Audio Format : X @tab X
37180 @tab Apple Core Audio Format
37181 CRC testing format : X @tab
37183 Creative Voice : X @tab X
37184 @tab Created for the Sound Blaster Pro.
37185 CRYO APC : @tab X
37187 @tab Audio format used in some games by CRYO Interactive Entertainment.
37188 D-Cinema audio : X @tab X
37190 Deluxe Paint Animation : @tab X
37191 DCSTR : @tab X
37192 DFA : @tab X
37193 @tab This format is used in Chronomaster game
37194 DirectDraw Surface : @tab X
37196 DSD Stream File (DSF) : @tab X
37197 DV video : X @tab X
37198 DXA : @tab X
37199 @tab This format is used in the non-Windows version of the Feeble Files
37200 game and different game cutscenes repacked for use with ScummVM.
37201 Electronic Arts cdata : @tab X
37203 Electronic Arts Multimedia : @tab X
37204 @tab Used in various EA games; files have extensions like WVE and UV2.
37205 Ensoniq Paris Audio File : @tab X
37207 FFM (FFserver live feed) : X @tab X
37208 Flash (SWF) : X @tab X
37209 Flash 9 (AVM2) : X @tab X
37210 @tab Only embedded audio is decoded.
37211 FLI/FLC/FLX animation : @tab X
37213 @tab .fli/.flc files
37214 Flash Video (FLV) : X @tab X
37216 @tab Macromedia Flash video files
37217 framecrc testing format : X @tab
37219 FunCom ISS : @tab X
37220 @tab Audio format used in various games from FunCom like The Longest Journey.
37221 G.723.1 : X @tab X
37223 G.726 : @tab X @tab Both left- and right-
37224 justified.
37225 G.729 BIT : X @tab X
37226 G.729 raw : @tab X
37227 GENH : @tab X
37228 @tab Audio format for various games.
37229 GIF Animation : X @tab X
37231 GXF : X @tab X
37232 @tab General eXchange Format SMPTE 360M, used by Thomson Grass Valley
37233 playout servers.
37234 HNM : @tab X
37236 @tab Only version 4 supported, used in some games from Cryo Interactive
37237 iCEDraw File : @tab X
37239 ICO : X @tab X
37240 @tab Microsoft Windows ICO
37241 id Quake II CIN video : @tab X
37243 id RoQ : X @tab X
37244 @tab Used in Quake III, Jedi Knight 2 and other computer games.
37245 IEC61937 encapsulation : X @tab X
37247 IFF : @tab X
37248 @tab Interchange File Format
37249 IFV : @tab X
37251 @tab A format used by some old CCTV DVRs.
37252 iLBC : X @tab X
37254 Interplay MVE : @tab X
37255 @tab Format used in various Interplay computer games.
37256 Iterated Systems ClearVideo : @tab X
37258 @tab I-frames only
37259 IV8 : @tab X
37261 @tab A format generated by IndigoVision 8000 video server.
37262 IVF (On2) : X @tab X
37264 @tab A format used by libvpx
37265 Internet Video Recording : @tab X
37267 IRCAM : X @tab X
37268 LAF : @tab X
37269 @tab Limitless Audio Format
37270 LATM : X @tab X
37272 LMLM4 : @tab X
37273 @tab Used by Linux Media Labs MPEG-4 PCI boards
37274 LOAS : @tab X
37276 @tab contains LATM multiplexed AAC audio
37277 LRC : X @tab X
37279 LVF : @tab X
37280 LXF : @tab X
37281 @tab VR native stream format, used by Leitch/Harris' video servers.
37282 Magic Lantern Video (MLV) : @tab X
37284 Matroska : X @tab X
37285 Matroska audio : X @tab
37286 FFmpeg metadata : X @tab X
37287 @tab Metadata in text format.
37288 MAXIS XA : @tab X
37290 @tab Used in Sim City 3000; file extension .xa.
37291 MCA : @tab X
37293 @tab Used in some games from Capcom; file extension .mca.
37294 MD Studio : @tab X
37296 Metal Gear Solid: The Twin Snakes : @tab X
37297 Megalux Frame : @tab X
37298 @tab Used by Megalux Ultimate Paint
37299 MobiClip MODS : @tab X
37301 MobiClip MOFLEX : @tab X
37302 Mobotix .mxg : @tab X
37303 Monkey's Audio : @tab X
37304 Motion Pixels MVI : @tab X
37305 MOV/QuickTime/MP4 : X @tab X
37306 @tab 3GP, 3GP2, PSP, iPod variants supported
37307 MP2 : X @tab X
37309 MP3 : X @tab X
37310 MPEG-1 System : X @tab X
37311 @tab muxed audio and video, VCD format supported
37312 MPEG-PS (program stream) : X @tab X
37314 @tab also known as C<VOB> file, SVCD and DVD format supported
37315 MPEG-TS (transport stream) : X @tab X
37317 @tab also known as DVB Transport Stream
37318 MPEG-4 : X @tab X
37320 @tab MPEG-4 is a variant of QuickTime.
37321 MSF : @tab X
37323 @tab Audio format used on the PS3.
37324 Mirillis FIC video : @tab X
37326 @tab No cursor rendering.
37327 MIDI Sample Dump Standard : @tab X
37329 MIME multipart JPEG : X @tab
37330 MSN TCP webcam : @tab X
37331 @tab Used by MSN Messenger webcam streams.
37332 MTV : @tab X
37334 Musepack : @tab X
37335 Musepack SV8 : @tab X
37336 Material eXchange Format (MXF) : X @tab X
37337 @tab SMPTE 377M, used by D-Cinema, broadcast industry.
37338 Material eXchange Format (MXF), D-10 Mapping : X @tab X
37340 @tab SMPTE 386M, D-10/IMX Mapping.
37341 NC camera feed : @tab X
37343 @tab NC (AVIP NC4600) camera streams
37344 NIST SPeech HEader REsources : @tab X
37346 Computerized Speech Lab NSP : @tab X
37347 NTT TwinVQ (VQF) : @tab X
37348 @tab Nippon Telegraph and Telephone Corporation TwinVQ.
37349 Nullsoft Streaming Video : @tab X
37351 NuppelVideo : @tab X
37352 NUT : X @tab X
37353 @tab NUT Open Container Format
37354 Ogg : X @tab X
37356 Playstation Portable PMP : @tab X
37357 Portable Voice Format : @tab X
37358 RK Audio (RKA) : @tab X
37359 TechnoTrend PVA : @tab X
37360 @tab Used by TechnoTrend DVB PCI boards.
37361 QCP : @tab X
37363 raw ADTS (AAC) : X @tab X
37364 raw AC-3 : X @tab X
37365 raw AMR-NB : @tab X
37366 raw AMR-WB : @tab X
37367 raw APAC : @tab X
37368 raw aptX : X @tab X
37369 raw aptX HD : X @tab X
37370 raw Bonk : @tab X
37371 raw Chinese AVS video : X @tab X
37372 raw DFPWM : X @tab X
37373 raw Dirac : X @tab X
37374 raw DNxHD : X @tab X
37375 raw DTS : X @tab X
37376 raw DTS-HD : @tab X
37377 raw E-AC-3 : X @tab X
37378 raw FLAC : X @tab X
37379 raw GSM : @tab X
37380 raw H.261 : X @tab X
37381 raw H.263 : X @tab X
37382 raw H.264 : X @tab X
37383 raw HEVC : X @tab X
37384 raw Ingenient MJPEG : @tab X
37385 raw MJPEG : X @tab X
37386 raw MLP : @tab X
37387 raw MPEG : @tab X
37388 raw MPEG-1 : @tab X
37389 raw MPEG-2 : @tab X
37390 raw MPEG-4 : X @tab X
37391 raw NULL : X @tab
37392 raw video : X @tab X
37393 raw id RoQ : X @tab
37394 raw OBU : X @tab X
37395 raw SBC : X @tab X
37396 raw Shorten : @tab X
37397 raw TAK : @tab X
37398 raw TrueHD : X @tab X
37399 raw VC-1 : X @tab X
37400 raw PCM A-law : X @tab X
37401 raw PCM mu-law : X @tab X
37402 raw PCM Archimedes VIDC : X @tab X
37403 raw PCM signed 8 bit : X @tab X
37404 raw PCM signed 16 bit big-endian : X @tab X
37405 raw PCM signed 16 bit little-endian : X @tab X
37406 raw PCM signed 24 bit big-endian : X @tab X
37407 raw PCM signed 24 bit little-endian : X @tab X
37408 raw PCM signed 32 bit big-endian : X @tab X
37409 raw PCM signed 32 bit little-endian : X @tab X
37410 raw PCM signed 64 bit big-endian : X @tab X
37411 raw PCM signed 64 bit little-endian : X @tab X
37412 raw PCM unsigned 8 bit : X @tab X
37413 raw PCM unsigned 16 bit big-endian : X @tab X
37414 raw PCM unsigned 16 bit little-endian : X @tab X
37415 raw PCM unsigned 24 bit big-endian : X @tab X
37416 raw PCM unsigned 24 bit little-endian : X @tab X
37417 raw PCM unsigned 32 bit big-endian : X @tab X
37418 raw PCM unsigned 32 bit little-endian : X @tab X
37419 raw PCM 16.8 floating point little-endian : @tab X
37420 raw PCM 24.0 floating point little-endian : @tab X
37421 raw PCM floating-point 32 bit big-endian : X @tab X
37422 raw PCM floating-point 32 bit little-endian : X @tab X
37423 raw PCM floating-point 64 bit big-endian : X @tab X
37424 raw PCM floating-point 64 bit little-endian : X @tab X
37425 RDT : @tab X
37426 REDCODE R3D : @tab X
37427 @tab File format used by RED Digital cameras, contains JPEG 2000 frames and PCM audio.
37428 RealMedia : X @tab X
37430 Redirector : @tab X
37431 RedSpark : @tab X
37432 Renderware TeXture Dictionary : @tab X
37433 Resolume DXV : @tab X
37434 RF64 : @tab X
37435 RL2 : @tab X
37436 @tab Audio and video format used in some games by Entertainment Software Partners.
37437 RPL/ARMovie : @tab X
37439 Lego Mindstorms RSO : X @tab X
37440 RSD : @tab X
37441 RTMP : X @tab X
37442 @tab Output is performed by publishing stream to RTMP server
37443 RTP : X @tab X
37445 RTSP : X @tab X
37446 Sample Dump eXchange : @tab X
37447 SAP : X @tab X
37448 SBG : @tab X
37449 SDNS : @tab X
37450 SDP : @tab X
37451 SER : @tab X
37452 Digital Pictures SGA : @tab X
37453 Sega FILM/CPK : X @tab X
37454 @tab Used in many Sega Saturn console games.
37455 Silicon Graphics Movie : @tab X
37457 Sierra SOL : @tab X
37458 @tab .sol files used in Sierra Online games.
37459 Sierra VMD : @tab X
37461 @tab Used in Sierra CD-ROM games.
37462 Smacker : @tab X
37464 @tab Multimedia format used by many games.
37465 SMJPEG : X @tab X
37467 @tab Used in certain Loki game ports.
37468 SMPTE 337M encapsulation : @tab X
37470 Smush : @tab X
37471 @tab Multimedia format used in some LucasArts games.
37472 Sony OpenMG (OMA) : X @tab X
37474 @tab Audio format used in Sony Sonic Stage and Sony Vegas.
37475 Sony PlayStation STR : @tab X
37477 Sony Wave64 (W64) : X @tab X
37478 SoX native format : X @tab X
37479 SUN AU format : X @tab X
37480 SUP raw PGS subtitles : X @tab X
37481 SVAG : @tab X
37482 @tab Audio format used in Konami PS2 games.
37483 TDSC : @tab X
37485 Text files : @tab X
37486 THP : @tab X
37487 @tab Used on the Nintendo GameCube.
37488 Tiertex Limited SEQ : @tab X
37490 @tab Tiertex .seq files used in the DOS CD-ROM version of the game Flashback.
37491 True Audio : X @tab X
37493 VAG : @tab X
37494 @tab Audio format used in many Sony PS2 games.
37495 VC-1 test bitstream : X @tab X
37497 Vidvox Hap : X @tab X
37498 Vivo : @tab X
37499 VPK : @tab X
37500 @tab Audio format used in Sony PS games.
37501 Marble WADY : @tab X
37503 WAV : X @tab X
37504 Waveform Archiver : @tab X
37505 WavPack : X @tab X
37506 WebM : X @tab X
37507 Windows Televison (WTV) : X @tab X
37508 Wing Commander III movie : @tab X
37509 @tab Multimedia format used in Origin's Wing Commander III computer game.
37510 Westwood Studios audio : X @tab X
37512 @tab Multimedia format used in Westwood Studios games.
37513 Westwood Studios VQA : @tab X
37515 @tab Multimedia format used in Westwood Studios games.
37516 Wideband Single-bit Data (WSD) : @tab X
37518 WVE : @tab X
37519 Konami XMD : @tab X
37520 XMV : @tab X
37521 @tab Microsoft video container used in Xbox games.
37522 XVAG : @tab X
37524 @tab Audio format used on the PS3.
37525 xWMA : @tab X
37527 @tab Microsoft audio container used by XAudio 2.
37528 eXtended BINary text (XBIN) : @tab X
37530 YUV4MPEG pipe : X @tab X
37531 Psygnosis YOP : @tab X
37532 "X" means that the feature in that column (encoding / decoding) is
37534 supported.
37535 Image Formats
37537 FFmpeg can read and write images for each frame of a video sequence.
37538 The following image formats are supported:
37539 Name : Encoding @tab Decoding @tab Comments
37541 .Y.U.V : X @tab X
37542 @tab one raw file per component
37543 Alias PIX : X @tab X
37545 @tab Alias/Wavefront PIX image format
37546 animated GIF : X @tab X
37548 APNG : X @tab X
37549 @tab Animated Portable Network Graphics
37550 BMP : X @tab X
37552 @tab Microsoft BMP image
37553 BRender PIX : @tab X
37555 @tab Argonaut BRender 3D engine image format.
37556 CRI : @tab X
37558 @tab Cintel RAW
37559 DPX : X @tab X
37561 @tab Digital Picture Exchange
37562 EXR : @tab X
37564 @tab OpenEXR
37565 FITS : X @tab X
37567 @tab Flexible Image Transport System
37568 HDR : X @tab X
37570 @tab Radiance HDR RGBE Image format
37571 IMG : @tab X
37573 @tab GEM Raster image
37574 JPEG : X @tab X
37576 @tab Progressive JPEG is not supported.
37577 JPEG 2000 : X @tab X
37579 JPEG-LS : X @tab X
37580 LJPEG : X @tab
37581 @tab Lossless JPEG
37582 Media 100 : @tab X
37584 MSP : @tab X
37585 @tab Microsoft Paint image
37586 PAM : X @tab X
37588 @tab PAM is a PNM extension with alpha support.
37589 PBM : X @tab X
37591 @tab Portable BitMap image
37592 PCD : @tab X
37594 @tab PhotoCD
37595 PCX : X @tab X
37597 @tab PC Paintbrush
37598 PFM : X @tab X
37600 @tab Portable FloatMap image
37601 PGM : X @tab X
37603 @tab Portable GrayMap image
37604 PGMYUV : X @tab X
37606 @tab PGM with U and V components in YUV 4:2:0
37607 PGX : @tab X
37609 @tab PGX file decoder
37610 PHM : X @tab X
37612 @tab Portable HalfFloatMap image
37613 PIC : @tab X
37615 @tab Pictor/PC Paint
37616 PNG : X @tab X
37618 @tab Portable Network Graphics image
37619 PPM : X @tab X
37621 @tab Portable PixelMap image
37622 PSD : @tab X
37624 @tab Photoshop
37625 PTX : @tab X
37627 @tab V.Flash PTX format
37628 QOI : X @tab X
37630 @tab Quite OK Image format
37631 SGI : X @tab X
37633 @tab SGI RGB image format
37634 Sun Rasterfile : X @tab X
37636 @tab Sun RAS image format
37637 TIFF : X @tab X
37639 @tab YUV, JPEG and some extension is not supported yet.
37640 Truevision Targa : X @tab X
37642 @tab Targa (.TGA) image format
37643 VBN : X @tab X
37645 @tab Vizrt Binary Image format
37646 WBMP : X @tab X
37648 @tab Wireless Application Protocol Bitmap image format
37649 WebP : E @tab X
37651 @tab WebP image format, encoding supported through external library libwebp
37652 XBM : X @tab X
37654 @tab X BitMap image format
37655 XFace : X @tab X
37657 @tab X-Face image format
37658 XPM : @tab X
37660 @tab X PixMap image format
37661 XWD : X @tab X
37663 @tab X Window Dump image format
37664 "X" means that the feature in that column (encoding / decoding) is
37666 supported.
37667 "E" means that support is provided through an external library.
37669 Video Codecs
37671 Name : Encoding @tab Decoding @tab Comments
37672 4X Movie : @tab X
37673 @tab Used in certain computer games.
37674 8088flex TMV : @tab X
37676 A64 multicolor : X @tab
37677 @tab Creates video suitable to be played on a commodore 64 (multicolor mode).
37678 Amazing Studio PAF Video : @tab X
37680 American Laser Games MM : @tab X
37681 @tab Used in games like Mad Dog McCree.
37682 Amuse Graphics Movie : @tab X
37684 AMV Video : X @tab X
37685 @tab Used in Chinese MP3 players.
37686 ANSI/ASCII art : @tab X
37688 Apple Intermediate Codec : @tab X
37689 Apple MJPEG-B : @tab X
37690 Apple Pixlet : @tab X
37691 Apple ProRes : X @tab X
37692 @tab fourcc: apch,apcn,apcs,apco,ap4h,ap4x
37693 Apple QuickDraw : @tab X
37695 @tab fourcc: qdrw
37696 Argonaut Video : @tab X
37698 @tab Used in some Argonaut games.
37699 Asus v1 : X @tab X
37701 @tab fourcc: ASV1
37702 Asus v2 : X @tab X
37704 @tab fourcc: ASV2
37705 ATI VCR1 : @tab X
37707 @tab fourcc: VCR1
37708 ATI VCR2 : @tab X
37710 @tab fourcc: VCR2
37711 Auravision Aura : @tab X
37713 Auravision Aura 2 : @tab X
37714 Autodesk Animator Flic video : @tab X
37715 Autodesk RLE : @tab X
37716 @tab fourcc: AASC
37717 AV1 : E @tab E
37719 @tab Supported through external libraries libaom, libdav1d, librav1e and libsvtav1
37720 Avid 1:1 10-bit RGB Packer : X @tab X
37722 @tab fourcc: AVrp
37723 AVS (Audio Video Standard) video : @tab X
37725 @tab Video encoding used by the Creature Shock game.
37726 AVS2-P2/IEEE1857.4 : E @tab E
37728 @tab Supported through external libraries libxavs2 and libdavs2
37729 AVS3-P2/IEEE1857.10 : @tab E
37731 @tab Supported through external library libuavs3d
37732 AYUV : X @tab X
37734 @tab Microsoft uncompressed packed 4:4:4:4
37735 Beam Software VB : @tab X
37737 Bethesda VID video : @tab X
37738 @tab Used in some games from Bethesda Softworks.
37739 Bink Video : @tab X
37741 BitJazz SheerVideo : @tab X
37742 Bitmap Brothers JV video : @tab X
37743 y41p Brooktree uncompressed 4:1:1 12-bit : X @tab X
37744 Brooktree ProSumer Video : @tab X
37745 @tab fourcc: BT20
37746 Brute Force & Ignorance : @tab X
37748 @tab Used in the game Flash Traffic: City of Angels.
37749 C93 video : @tab X
37751 @tab Codec used in Cyberia game.
37752 CamStudio : @tab X
37754 @tab fourcc: CSCD
37755 CD+G : @tab X
37757 @tab Video codec for CD+G karaoke disks
37758 CDXL : @tab X
37760 @tab Amiga CD video codec
37761 Chinese AVS video : E @tab X
37763 @tab AVS1-P2, JiZhun profile, encoding through external library libxavs
37764 Delphine Software International CIN video : @tab X
37766 @tab Codec used in Delphine Software International games.
37767 Discworld II BMV Video : @tab X
37769 CineForm HD : X @tab X
37770 Canopus HQ : @tab X
37771 Canopus HQA : @tab X
37772 Canopus HQX : @tab X
37773 Canopus Lossless Codec : @tab X
37774 CDToons : @tab X
37775 @tab Codec used in various Broderbund games.
37776 Cinepak : @tab X
37778 Cirrus Logic AccuPak : X @tab X
37779 @tab fourcc: CLJR
37780 CPiA Video Format : @tab X
37782 Creative YUV (CYUV) : @tab X
37783 DFA : @tab X
37784 @tab Codec used in Chronomaster game.
37785 Dirac : E @tab X
37787 @tab supported though the native vc2 (Dirac Pro) encoder
37788 Deluxe Paint Animation : @tab X
37790 DNxHD : X @tab X
37791 @tab aka SMPTE VC3
37792 Duck TrueMotion 1.0 : @tab X
37794 @tab fourcc: DUCK
37795 Duck TrueMotion 2.0 : @tab X
37797 @tab fourcc: TM20
37798 Duck TrueMotion 2.0 RT : @tab X
37800 @tab fourcc: TR20
37801 DV (Digital Video) : X @tab X
37803 Dxtory capture format : @tab X
37804 Feeble Files/ScummVM DXA : @tab X
37805 @tab Codec originally used in Feeble Files game.
37806 Electronic Arts CMV video : @tab X
37808 @tab Used in NHL 95 game.
37809 Electronic Arts Madcow video : @tab X
37811 Electronic Arts TGV video : @tab X
37812 Electronic Arts TGQ video : @tab X
37813 Electronic Arts TQI video : @tab X
37814 Escape 124 : @tab X
37815 Escape 130 : @tab X
37816 FFmpeg video codec #1 : X @tab X
37817 @tab lossless codec (fourcc: FFV1)
37818 Flash Screen Video v1 : X @tab X
37820 @tab fourcc: FSV1
37821 Flash Screen Video v2 : X @tab X
37823 Flash Video (FLV) : X @tab X
37824 @tab Sorenson H.263 used in Flash
37825 FM Screen Capture Codec : @tab X
37827 Forward Uncompressed : @tab X
37828 Fraps : @tab X
37829 Go2Meeting : @tab X
37830 @tab fourcc: G2M2, G2M3
37831 Go2Webinar : @tab X
37833 @tab fourcc: G2M4
37834 Gremlin Digital Video : @tab X
37836 H.261 : X @tab X
37837 H.263 / H.263-1996 : X @tab X
37838 H.263+ / H.263-1998 / H.263 version 2 : X @tab X
37839 H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10 : E @tab X
37840 @tab encoding supported through external library libx264 and OpenH264
37841 HEVC : X @tab X
37843 @tab encoding supported through external library libx265 and libkvazaar
37844 HNM version 4 : @tab X
37846 HuffYUV : X @tab X
37847 HuffYUV FFmpeg variant : X @tab X
37848 IBM Ultimotion : @tab X
37849 @tab fourcc: ULTI
37850 id Cinematic video : @tab X
37852 @tab Used in Quake II.
37853 id RoQ video : X @tab X
37855 @tab Used in Quake III, Jedi Knight 2, other computer games.
37856 IFF ILBM : @tab X
37858 @tab IFF interleaved bitmap
37859 IFF ByteRun1 : @tab X
37861 @tab IFF run length encoded bitmap
37862 Infinity IMM4 : @tab X
37864 Intel H.263 : @tab X
37865 Intel Indeo 2 : @tab X
37866 Intel Indeo 3 : @tab X
37867 Intel Indeo 4 : @tab X
37868 Intel Indeo 5 : @tab X
37869 Interplay C93 : @tab X
37870 @tab Used in the game Cyberia from Interplay.
37871 Interplay MVE video : @tab X
37873 @tab Used in Interplay .MVE files.
37874 J2K : X @tab X
37876 Karl Morton's video codec : @tab X
37877 @tab Codec used in Worms games.
37878 Kega Game Video (KGV1) : @tab X
37880 @tab Kega emulator screen capture codec.
37881 Lagarith : @tab X
37883 LCL (LossLess Codec Library) MSZH : @tab X
37884 LCL (LossLess Codec Library) ZLIB : E @tab E
37885 LOCO : @tab X
37886 LucasArts SANM/Smush : @tab X
37887 @tab Used in LucasArts games / SMUSH animations.
37888 lossless MJPEG : X @tab X
37890 MagicYUV Video : X @tab X
37891 Mandsoft Screen Capture Codec : @tab X
37892 Microsoft ATC Screen : @tab X
37893 @tab Also known as Microsoft Screen 3.
37894 Microsoft Expression Encoder Screen : @tab X
37896 @tab Also known as Microsoft Titanium Screen 2.
37897 Microsoft RLE : @tab X
37899 Microsoft Screen 1 : @tab X
37900 @tab Also known as Windows Media Video V7 Screen.
37901 Microsoft Screen 2 : @tab X
37903 @tab Also known as Windows Media Video V9 Screen.
37904 Microsoft Video 1 : @tab X
37906 Mimic : @tab X
37907 @tab Used in MSN Messenger Webcam streams.
37908 Miro VideoXL : @tab X
37910 @tab fourcc: VIXL
37911 MJPEG (Motion JPEG) : X @tab X
37913 Mobotix MxPEG video : @tab X
37914 Motion Pixels video : @tab X
37915 MPEG-1 video : X @tab X
37916 MPEG-2 video : X @tab X
37917 MPEG-4 part 2 : X @tab X
37918 @tab libxvidcore can be used alternatively for encoding.
37919 MPEG-4 part 2 Microsoft variant version 1 : @tab X
37921 MPEG-4 part 2 Microsoft variant version 2 : X @tab X
37922 MPEG-4 part 2 Microsoft variant version 3 : X @tab X
37923 Newtek SpeedHQ : X @tab X
37924 Nintendo Gamecube THP video : @tab X
37925 NotchLC : @tab X
37926 NuppelVideo/RTjpeg : @tab X
37927 @tab Video encoding used in NuppelVideo files.
37928 On2 VP3 : @tab X
37930 @tab still experimental
37931 On2 VP4 : @tab X
37933 @tab fourcc: VP40
37934 On2 VP5 : @tab X
37936 @tab fourcc: VP50
37937 On2 VP6 : @tab X
37939 @tab fourcc: VP60,VP61,VP62
37940 On2 VP7 : @tab X
37942 @tab fourcc: VP70,VP71
37943 VP8 : E @tab X
37945 @tab fourcc: VP80, encoding supported through external library libvpx
37946 VP9 : E @tab X
37948 @tab encoding supported through external library libvpx
37949 Pinnacle TARGA CineWave YUV16 : @tab X
37951 @tab fourcc: Y216
37952 Q-team QPEG : @tab X
37954 @tab fourccs: QPEG, Q1.0, Q1.1
37955 QuickTime 8BPS video : @tab X
37957 QuickTime Animation (RLE) video : X @tab X
37958 @tab fourcc: 'rle '
37959 QuickTime Graphics (SMC) : X @tab X
37961 @tab fourcc: 'smc '
37962 QuickTime video (RPZA) : X @tab X
37964 @tab fourcc: rpza
37965 R10K AJA Kona 10-bit RGB Codec : X @tab X
37967 R210 Quicktime Uncompressed RGB 10-bit : X @tab X
37968 Raw Video : X @tab X
37969 RealVideo 1.0 : X @tab X
37970 RealVideo 2.0 : X @tab X
37971 RealVideo 3.0 : @tab X
37972 @tab still far from ideal
37973 RealVideo 4.0 : @tab X
37975 Renderware TXD (TeXture Dictionary) : @tab X
37976 @tab Texture dictionaries used by the Renderware Engine.
37977 RL2 video : @tab X
37979 @tab used in some games by Entertainment Software Partners
37980 ScreenPressor : @tab X
37982 Screenpresso : @tab X
37983 Screen Recorder Gold Codec : @tab X
37984 Sierra VMD video : @tab X
37985 @tab Used in Sierra VMD files.
37986 Silicon Graphics Motion Video Compressor 1 (MVC1) : @tab X
37988 Silicon Graphics Motion Video Compressor 2 (MVC2) : @tab X
37989 Silicon Graphics RLE 8-bit video : @tab X
37990 Smacker video : @tab X
37991 @tab Video encoding used in Smacker.
37992 SMPTE VC-1 : @tab X
37994 Snow : X @tab X
37995 @tab experimental wavelet codec (fourcc: SNOW)
37996 Sony PlayStation MDEC (Motion DECoder) : @tab X
37998 Sorenson Vector Quantizer 1 : X @tab X
37999 @tab fourcc: SVQ1
38000 Sorenson Vector Quantizer 3 : @tab X
38002 @tab fourcc: SVQ3
38003 Sunplus JPEG (SP5X) : @tab X
38005 @tab fourcc: SP5X
38006 TechSmith Screen Capture Codec : @tab X
38008 @tab fourcc: TSCC
38009 TechSmith Screen Capture Codec 2 : @tab X
38011 @tab fourcc: TSC2
38012 Theora : E @tab X
38014 @tab encoding supported through external library libtheora
38015 Tiertex Limited SEQ video : @tab X
38017 @tab Codec used in DOS CD-ROM FlashBack game.
38018 Ut Video : X @tab X
38020 v210 QuickTime uncompressed 4:2:2 10-bit : X @tab X
38021 v308 QuickTime uncompressed 4:4:4 : X @tab X
38022 v408 QuickTime uncompressed 4:4:4:4 : X @tab X
38023 v410 QuickTime uncompressed 4:4:4 10-bit : X @tab X
38024 VBLE Lossless Codec : @tab X
38025 VMware Screen Codec / VMware Video : @tab X
38026 @tab Codec used in videos captured by VMware.
38027 Westwood Studios VQA (Vector Quantized Animation) video : @tab
38029 X
38030 Windows Media Image : @tab X
38031 Windows Media Video 7 : X @tab X
38032 Windows Media Video 8 : X @tab X
38033 Windows Media Video 9 : @tab X
38034 @tab not completely working
38035 Wing Commander III / Xan : @tab X
38037 @tab Used in Wing Commander III .MVE files.
38038 Wing Commander IV / Xan : @tab X
38040 @tab Used in Wing Commander IV.
38041 Winnov WNV1 : @tab X
38043 WMV7 : X @tab X
38044 YAMAHA SMAF : X @tab X
38045 Psygnosis YOP Video : @tab X
38046 yuv4 : X @tab X
38047 @tab libquicktime uncompressed packed 4:2:0
38048 ZeroCodec Lossless Video : @tab X
38050 ZLIB : X @tab X
38051 @tab part of LCL, encoder experimental
38052 Zip Motion Blocks Video : X @tab X
38054 @tab Encoder works only in PAL8.
38055 "X" means that the feature in that column (encoding / decoding) is
38057 supported.
38058 "E" means that support is provided through an external library.
38060 Audio Codecs
38062 Name : Encoding @tab Decoding @tab Comments
38063 8SVX exponential : @tab X
38064 8SVX fibonacci : @tab X
38065 AAC : EX @tab X
38066 @tab encoding supported through internal encoder and external library libfdk-aac
38067 AAC+ : E @tab IX
38069 @tab encoding supported through external library libfdk-aac
38070 AC-3 : IX @tab IX
38072 ACELP.KELVIN : @tab X
38073 ADPCM 4X Movie : @tab X
38074 ADPCM Yamaha AICA : @tab X
38075 ADPCM AmuseGraphics Movie : @tab X
38076 ADPCM Argonaut Games : X @tab X
38077 ADPCM CDROM XA : @tab X
38078 ADPCM Creative Technology : @tab X
38079 @tab 16 -E<gt> 4, 8 -E<gt> 4, 8 -E<gt> 3, 8 -E<gt> 2
38080 ADPCM Electronic Arts : @tab X
38082 @tab Used in various EA titles.
38083 ADPCM Electronic Arts Maxis CDROM XS : @tab X
38085 @tab Used in Sim City 3000.
38086 ADPCM Electronic Arts R1 : @tab X
38088 ADPCM Electronic Arts R2 : @tab X
38089 ADPCM Electronic Arts R3 : @tab X
38090 ADPCM Electronic Arts XAS : @tab X
38091 ADPCM G.722 : X @tab X
38092 ADPCM G.726 : X @tab X
38093 ADPCM IMA Acorn Replay : @tab X
38094 ADPCM IMA AMV : X @tab X
38095 @tab Used in AMV files
38096 ADPCM IMA Cunning Developments : @tab X
38098 ADPCM IMA Electronic Arts EACS : @tab X
38099 ADPCM IMA Electronic Arts SEAD : @tab X
38100 ADPCM IMA Funcom : @tab X
38101 ADPCM IMA High Voltage Software ALP : X @tab X
38102 ADPCM IMA Mobiclip MOFLEX : @tab X
38103 ADPCM IMA QuickTime : X @tab X
38104 ADPCM IMA Simon & Schuster Interactive : X @tab X
38105 ADPCM IMA Ubisoft APM : X @tab X
38106 ADPCM IMA Loki SDL MJPEG : @tab X
38107 ADPCM IMA WAV : X @tab X
38108 ADPCM IMA Westwood : @tab X
38109 ADPCM ISS IMA : @tab X
38110 @tab Used in FunCom games.
38111 ADPCM IMA Dialogic : @tab X
38113 ADPCM IMA Duck DK3 : @tab X
38114 @tab Used in some Sega Saturn console games.
38115 ADPCM IMA Duck DK4 : @tab X
38117 @tab Used in some Sega Saturn console games.
38118 ADPCM IMA Radical : @tab X
38120 ADPCM Microsoft : X @tab X
38121 ADPCM MS IMA : X @tab X
38122 ADPCM Nintendo Gamecube AFC : @tab X
38123 ADPCM Nintendo Gamecube DTK : @tab X
38124 ADPCM Nintendo THP : @tab X
38125 ADPCM Playstation : @tab X
38126 ADPCM QT IMA : X @tab X
38127 ADPCM SEGA CRI ADX : X @tab X
38128 @tab Used in Sega Dreamcast games.
38129 ADPCM Shockwave Flash : X @tab X
38131 ADPCM Sound Blaster Pro 2-bit : @tab X
38132 ADPCM Sound Blaster Pro 2.6-bit : @tab X
38133 ADPCM Sound Blaster Pro 4-bit : @tab X
38134 ADPCM VIMA : @tab X
38135 @tab Used in LucasArts SMUSH animations.
38136 ADPCM Konami XMD : @tab X
38138 ADPCM Westwood Studios IMA : X @tab X
38139 @tab Used in Westwood Studios games like Command and Conquer.
38140 ADPCM Yamaha : X @tab X
38142 ADPCM Zork : @tab X
38143 AMR-NB : E @tab X
38144 @tab encoding supported through external library libopencore-amrnb
38145 AMR-WB : E @tab X
38147 @tab encoding supported through external library libvo-amrwbenc
38148 Amazing Studio PAF Audio : @tab X
38150 Apple lossless audio : X @tab X
38151 @tab QuickTime fourcc 'alac'
38152 aptX : X @tab X
38154 @tab Used in Bluetooth A2DP
38155 aptX HD : X @tab X
38157 @tab Used in Bluetooth A2DP
38158 ATRAC1 : @tab X
38160 ATRAC3 : @tab X
38161 ATRAC3+ : @tab X
38162 ATRAC9 : @tab X
38163 Bink Audio : @tab X
38164 @tab Used in Bink and Smacker files in many games.
38165 Bonk audio : @tab X
38167 CELT : @tab E
38168 @tab decoding supported through external library libcelt
38169 codec2 : E @tab E
38171 @tab en/decoding supported through external library libcodec2
38172 CRI HCA : @tab X
38174 Delphine Software International CIN audio : @tab X
38175 @tab Codec used in Delphine Software International games.
38176 DFPWM : X @tab X
38178 Digital Speech Standard - Standard Play mode (DSS SP) : @tab X
38179 Discworld II BMV Audio : @tab X
38180 COOK : @tab X
38181 @tab All versions except 5.1 are supported.
38182 DCA (DTS Coherent Acoustics) : X @tab X
38184 @tab supported extensions: XCh, XXCH, X96, XBR, XLL, LBR (partially)
38185 Dolby E : @tab X
38187 DPCM Cuberoot-Delta-Exact : @tab X
38188 @tab Used in few games.
38189 DPCM Gremlin : @tab X
38191 DPCM id RoQ : X @tab X
38192 @tab Used in Quake III, Jedi Knight 2 and other computer games.
38193 DPCM Marble WADY : @tab X
38195 DPCM Interplay : @tab X
38196 @tab Used in various Interplay computer games.
38197 DPCM Squareroot-Delta-Exact : @tab X
38199 @tab Used in various games.
38200 DPCM Sierra Online : @tab X
38202 @tab Used in Sierra Online game audio files.
38203 DPCM Sol : @tab X
38205 DPCM Xan : @tab X
38206 @tab Used in Origin's Wing Commander IV AVI files.
38207 DPCM Xilam DERF : @tab X
38209 DSD (Direct Stream Digital), least significant bit first : @tab X
38210 DSD (Direct Stream Digital), most significant bit first : @tab X
38211 DSD (Direct Stream Digital), least significant bit first, planar :
38212 @tab X
38213 DSD (Direct Stream Digital), most significant bit first, planar :
38214 @tab X
38215 DSP Group TrueSpeech : @tab X
38216 DST (Direct Stream Transfer) : @tab X
38217 DV audio : @tab X
38218 Enhanced AC-3 : X @tab X
38219 EVRC (Enhanced Variable Rate Codec) : @tab X
38220 FLAC (Free Lossless Audio Codec) : X @tab IX
38221 FTR Voice : @tab X
38222 G.723.1 : X @tab X
38223 G.729 : @tab X
38224 GSM : E @tab X
38225 @tab encoding supported through external library libgsm
38226 GSM Microsoft variant : E @tab X
38228 @tab encoding supported through external library libgsm
38229 IAC (Indeo Audio Coder) : @tab X
38231 iLBC (Internet Low Bitrate Codec) : E @tab EX
38232 @tab encoding and decoding supported through external library libilbc
38233 IMC (Intel Music Coder) : @tab X
38235 Interplay ACM : @tab X
38236 MACE (Macintosh Audio Compression/Expansion) 3:1 : @tab X
38237 MACE (Macintosh Audio Compression/Expansion) 6:1 : @tab X
38238 Marian's A-pac audio : @tab X
38239 MI-SC4 (Micronas SC-4 Audio) : @tab X
38240 MLP (Meridian Lossless Packing) : X @tab X
38241 @tab Used in DVD-Audio discs.
38242 Monkey's Audio : @tab X
38244 MP1 (MPEG audio layer 1) : @tab IX
38245 MP2 (MPEG audio layer 2) : IX @tab IX
38246 @tab encoding supported also through external library TwoLAME
38247 MP3 (MPEG audio layer 3) : E @tab IX
38249 @tab encoding supported through external library LAME, ADU MP3 and MP3onMP4 also supported
38250 MPEG-4 Audio Lossless Coding (ALS) : @tab X
38252 MobiClip FastAudio : @tab X
38253 Musepack SV7 : @tab X
38254 Musepack SV8 : @tab X
38255 Nellymoser Asao : X @tab X
38256 On2 AVC (Audio for Video Codec) : @tab X
38257 Opus : E @tab X
38258 @tab encoding supported through external library libopus
38259 PCM A-law : X @tab X
38261 PCM mu-law : X @tab X
38262 PCM Archimedes VIDC : X @tab X
38263 PCM signed 8-bit planar : X @tab X
38264 PCM signed 16-bit big-endian planar : X @tab X
38265 PCM signed 16-bit little-endian planar : X @tab X
38266 PCM signed 24-bit little-endian planar : X @tab X
38267 PCM signed 32-bit little-endian planar : X @tab X
38268 PCM 32-bit floating point big-endian : X @tab X
38269 PCM 32-bit floating point little-endian : X @tab X
38270 PCM 64-bit floating point big-endian : X @tab X
38271 PCM 64-bit floating point little-endian : X @tab X
38272 PCM D-Cinema audio signed 24-bit : X @tab X
38273 PCM signed 8-bit : X @tab X
38274 PCM signed 16-bit big-endian : X @tab X
38275 PCM signed 16-bit little-endian : X @tab X
38276 PCM signed 24-bit big-endian : X @tab X
38277 PCM signed 24-bit little-endian : X @tab X
38278 PCM signed 32-bit big-endian : X @tab X
38279 PCM signed 32-bit little-endian : X @tab X
38280 PCM signed 16/20/24-bit big-endian in MPEG-TS : @tab X
38281 PCM unsigned 8-bit : X @tab X
38282 PCM unsigned 16-bit big-endian : X @tab X
38283 PCM unsigned 16-bit little-endian : X @tab X
38284 PCM unsigned 24-bit big-endian : X @tab X
38285 PCM unsigned 24-bit little-endian : X @tab X
38286 PCM unsigned 32-bit big-endian : X @tab X
38287 PCM unsigned 32-bit little-endian : X @tab X
38288 PCM SGA : @tab X
38289 QCELP / PureVoice : @tab X
38290 QDesign Music Codec 1 : @tab X
38291 QDesign Music Codec 2 : @tab X
38292 @tab There are still some distortions.
38293 RealAudio 1.0 (14.4K) : X @tab X
38295 @tab Real 14400 bit/s codec
38296 RealAudio 2.0 (28.8K) : @tab X
38298 @tab Real 28800 bit/s codec
38299 RealAudio 3.0 (dnet) : IX @tab X
38301 @tab Real low bitrate AC-3 codec
38302 RealAudio Lossless : @tab X
38304 RealAudio SIPR / ACELP.NET : @tab X
38305 RK Audio (RKA) : @tab X
38306 SBC (low-complexity subband codec) : X @tab X
38307 @tab Used in Bluetooth A2DP
38308 Shorten : @tab X
38310 Sierra VMD audio : @tab X
38311 @tab Used in Sierra VMD files.
38312 Smacker audio : @tab X
38314 SMPTE 302M AES3 audio : X @tab X
38315 Sonic : X @tab X
38316 @tab experimental codec
38317 Sonic lossless : X @tab X
38319 @tab experimental codec
38320 Speex : E @tab EX
38322 @tab supported through external library libspeex
38323 TAK (Tom's lossless Audio Kompressor) : @tab X
38325 True Audio (TTA) : X @tab X
38326 TrueHD : X @tab X
38327 @tab Used in HD-DVD and Blu-Ray discs.
38328 TwinVQ (VQF flavor) : @tab X
38330 VIMA : @tab X
38331 @tab Used in LucasArts SMUSH animations.
38332 ViewQuest VQC : @tab X
38334 Vorbis : E @tab X
38335 @tab A native but very primitive encoder exists.
38336 Voxware MetaSound : @tab X
38338 Waveform Archiver : @tab X
38339 WavPack : X @tab X
38340 Westwood Audio (SND1) : @tab X
38341 Windows Media Audio 1 : X @tab X
38342 Windows Media Audio 2 : X @tab X
38343 Windows Media Audio Lossless : @tab X
38344 Windows Media Audio Pro : @tab X
38345 Windows Media Audio Voice : @tab X
38346 Xbox Media Audio 1 : @tab X
38347 Xbox Media Audio 2 : @tab X
38348 "X" means that the feature in that column (encoding / decoding) is
38350 supported.
38351 "E" means that support is provided through an external library.
38353 "I" means that an integer-only version is available, too (ensures high
38355 performance on systems without hardware floating point support).
38356 Subtitle Formats
38358 Name : Muxing @tab Demuxing @tab Encoding @tab Decoding
38359 3GPP Timed Text : @tab @tab X @tab X
38360 AQTitle : @tab X @tab @tab X
38361 DVB : X @tab X @tab X @tab X
38362 DVB teletext : @tab X @tab @tab E
38363 DVD : X @tab X @tab X @tab X
38364 JACOsub : X @tab X @tab @tab X
38365 MicroDVD : X @tab X @tab @tab X
38366 MPL2 : @tab X @tab @tab X
38367 MPsub (MPlayer) : @tab X @tab @tab X
38368 PGS : @tab @tab @tab X
38369 PJS (Phoenix) : @tab X @tab @tab X
38370 RealText : @tab X @tab @tab X
38371 SAMI : @tab X @tab @tab X
38372 Spruce format (STL) : @tab X @tab @tab X
38373 SSA/ASS : X @tab X @tab X @tab X
38374 SubRip (SRT) : X @tab X @tab X @tab X
38375 SubViewer v1 : @tab X @tab @tab X
38376 SubViewer : @tab X @tab @tab X
38377 TED Talks captions : @tab X @tab @tab X
38378 TTML : X @tab @tab X @tab
38379 VobSub (IDX+SUB) : @tab X @tab @tab X
38380 VPlayer : @tab X @tab @tab X
38381 WebVTT : X @tab X @tab X @tab X
38382 XSUB : @tab @tab X @tab X
38383 "X" means that the feature is supported.
38385 "E" means that support is provided through an external library.
38387 Network Protocols
38389 Name : Support
38390 AMQP : E
38391 file : X
38392 FTP : X
38393 Gopher : X
38394 Gophers : X
38395 HLS : X
38396 HTTP : X
38397 HTTPS : X
38398 Icecast : X
38399 MMSH : X
38400 MMST : X
38401 pipe : X
38402 Pro-MPEG FEC : X
38403 RTMP : X
38404 RTMPE : X
38405 RTMPS : X
38406 RTMPT : X
38407 RTMPTE : X
38408 RTMPTS : X
38409 RTP : X
38410 SAMBA : E
38411 SCTP : X
38412 SFTP : E
38413 TCP : X
38414 TLS : X
38415 UDP : X
38416 ZMQ : E
38417 "X" means that the protocol is supported.
38419 "E" means that support is provided through an external library.
38421 Input/Output Devices
38423 Name : Input @tab Output
38424 ALSA : X @tab X
38425 BKTR : X @tab
38426 caca : @tab X
38427 DV1394 : X @tab
38428 Lavfi virtual device : X @tab
38429 Linux framebuffer : X @tab X
38430 JACK : X @tab
38431 LIBCDIO : X
38432 LIBDC1394 : X @tab
38433 OpenAL : X
38434 OpenGL : @tab X
38435 OSS : X @tab X
38436 PulseAudio : X @tab X
38437 SDL : @tab X
38438 Video4Linux2 : X @tab X
38439 VfW capture : X @tab
38440 X11 grabbing : X @tab
38441 Win32 grabbing : X @tab
38442 "X" means that input/output is supported.
38444 Timecode
38446 Codec/format : Read @tab Write
38447 AVI : X @tab X
38448 DV : X @tab X
38449 GXF : X @tab X
38450 MOV : X @tab X
38451 MPEG1/2 : X @tab X
38452 MXF : X @tab X
38453
38455 ffplay(1), ffmpeg(1), ffprobe(1), ffmpeg-utils(1), ffmpeg-scaler(1),
38456 ffmpeg-resampler(1), ffmpeg-codecs(1), ffmpeg-bitstream-filters(1),
38457 ffmpeg-formats(1), ffmpeg-devices(1), ffmpeg-protocols(1),
38458 ffmpeg-filters(1)
38459
38461 The FFmpeg developers.
38462 For details about the authorship, see the Git history of the project
38464 (https://git.ffmpeg.org/ffmpeg), e.g. by typing the command git log in
38465 the FFmpeg source directory, or browsing the online repository at
38466 <https://git.ffmpeg.org/ffmpeg>.
38467 Maintainers for the specific components are listed in the file
38469 MAINTAINERS in the source code tree.
38470 FFPLAY-ALL(1)