1FFPROBE-ALL(1) FFPROBE-ALL(1)
2
6 ffprobe - ffprobe media prober
7
9 ffprobe [options] input_url
10
12 ffprobe gathers information from multimedia streams and prints it in
13 human- and machine-readable fashion.
14 For example it can be used to check the format of the container used by
16 a multimedia stream and the format and type of each media stream
17 contained in it.
18 If a url is specified in input, ffprobe will try to open and probe the
20 url content. If the url cannot be opened or recognized as a multimedia
21 file, a positive exit code is returned.
22 If no output is specified as output with o ffprobe will write to
24 stdout.
25 ffprobe may be employed both as a standalone application or in
27 combination with a textual filter, which may perform more sophisticated
28 processing, e.g. statistical processing or plotting.
29 Options are used to list some of the formats supported by ffprobe or
31 for specifying which information to display, and for setting how
32 ffprobe will show it.
33 ffprobe output is designed to be easily parsable by a textual filter,
35 and consists of one or more sections of a form defined by the selected
36 writer, which is specified by the print_format option.
37 Sections may contain other nested sections, and are identified by a
39 name (which may be shared by other sections), and an unique name. See
40 the output of sections.
41 Metadata tags stored in the container or in the streams are recognized
43 and printed in the corresponding "FORMAT", "STREAM" or "PROGRAM_STREAM"
44 section.
45
47 All the numerical options, if not specified otherwise, accept a string
48 representing a number as input, which may be followed by one of the SI
49 unit prefixes, for example: 'K', 'M', or 'G'.
50 If 'i' is appended to the SI unit prefix, the complete prefix will be
52 interpreted as a unit prefix for binary multiples, which are based on
53 powers of 1024 instead of powers of 1000. Appending 'B' to the SI unit
54 prefix multiplies the value by 8. This allows using, for example: 'KB',
55 'MiB', 'G' and 'B' as number suffixes.
56 Options which do not take arguments are boolean options, and set the
58 corresponding value to true. They can be set to false by prefixing the
59 option name with "no". For example using "-nofoo" will set the boolean
60 option with name "foo" to false.
61 Stream specifiers
63 Some options are applied per-stream, e.g. bitrate or codec. Stream
64 specifiers are used to precisely specify which stream(s) a given option
65 belongs to.
66 A stream specifier is a string generally appended to the option name
68 and separated from it by a colon. E.g. "-codec:a:1 ac3" contains the
69 "a:1" stream specifier, which matches the second audio stream.
70 Therefore, it would select the ac3 codec for the second audio stream.
71 A stream specifier can match several streams, so that the option is
73 applied to all of them. E.g. the stream specifier in "-b:a 128k"
74 matches all audio streams.
75 An empty stream specifier matches all streams. For example, "-codec
77 copy" or "-codec: copy" would copy all the streams without reencoding.
78 Possible forms of stream specifiers are:
80 stream_index
82 Matches the stream with this index. E.g. "-threads:1 4" would set
83 the thread count for the second stream to 4. If stream_index is
84 used as an additional stream specifier (see below), then it selects
85 stream number stream_index from the matching streams. Stream
86 numbering is based on the order of the streams as detected by
87 libavformat except when a program ID is also specified. In this
88 case it is based on the ordering of the streams in the program.
89 stream_type[:additional_stream_specifier]
91 stream_type is one of following: 'v' or 'V' for video, 'a' for
92 audio, 's' for subtitle, 'd' for data, and 't' for attachments. 'v'
93 matches all video streams, 'V' only matches video streams which are
94 not attached pictures, video thumbnails or cover arts. If
95 additional_stream_specifier is used, then it matches streams which
96 both have this type and match the additional_stream_specifier.
97 Otherwise, it matches all streams of the specified type.
98 p:program_id[:additional_stream_specifier]
100 Matches streams which are in the program with the id program_id. If
101 additional_stream_specifier is used, then it matches streams which
102 both are part of the program and match the
103 additional_stream_specifier.
104 #stream_id or i:stream_id
106 Match the stream by stream id (e.g. PID in MPEG-TS container).
107 m:key[:value]
109 Matches streams with the metadata tag key having the specified
110 value. If value is not given, matches streams that contain the
111 given tag with any value.
112 u Matches streams with usable configuration, the codec must be
114 defined and the essential information such as video dimension or
115 audio sample rate must be present.
116 Note that in ffmpeg, matching by metadata will only work properly
118 for input files.
119 Generic options
121 These options are shared amongst the ff* tools.
122 -L Show license.
124 -h, -?, -help, --help [arg]
126 Show help. An optional parameter may be specified to print help
127 about a specific item. If no argument is specified, only basic (non
128 advanced) tool options are shown.
129 Possible values of arg are:
131 long
133 Print advanced tool options in addition to the basic tool
134 options.
135 full
137 Print complete list of options, including shared and private
138 options for encoders, decoders, demuxers, muxers, filters, etc.
139 decoder=decoder_name
141 Print detailed information about the decoder named
142 decoder_name. Use the -decoders option to get a list of all
143 decoders.
144 encoder=encoder_name
146 Print detailed information about the encoder named
147 encoder_name. Use the -encoders option to get a list of all
148 encoders.
149 demuxer=demuxer_name
151 Print detailed information about the demuxer named
152 demuxer_name. Use the -formats option to get a list of all
153 demuxers and muxers.
154 muxer=muxer_name
156 Print detailed information about the muxer named muxer_name.
157 Use the -formats option to get a list of all muxers and
158 demuxers.
159 filter=filter_name
161 Print detailed information about the filter named filter_name.
162 Use the -filters option to get a list of all filters.
163 bsf=bitstream_filter_name
165 Print detailed information about the bitstream filter named
166 bitstream_filter_name. Use the -bsfs option to get a list of
167 all bitstream filters.
168 protocol=protocol_name
170 Print detailed information about the protocol named
171 protocol_name. Use the -protocols option to get a list of all
172 protocols.
173 -version
175 Show version.
176 -buildconf
178 Show the build configuration, one option per line.
179 -formats
181 Show available formats (including devices).
182 -demuxers
184 Show available demuxers.
185 -muxers
187 Show available muxers.
188 -devices
190 Show available devices.
191 -codecs
193 Show all codecs known to libavcodec.
194 Note that the term 'codec' is used throughout this documentation as
196 a shortcut for what is more correctly called a media bitstream
197 format.
198 -decoders
200 Show available decoders.
201 -encoders
203 Show all available encoders.
204 -bsfs
206 Show available bitstream filters.
207 -protocols
209 Show available protocols.
210 -filters
212 Show available libavfilter filters.
213 -pix_fmts
215 Show available pixel formats.
216 -sample_fmts
218 Show available sample formats.
219 -layouts
221 Show channel names and standard channel layouts.
222 -dispositions
224 Show stream dispositions.
225 -colors
227 Show recognized color names.
228 -sources device[,opt1=val1[,opt2=val2]...]
230 Show autodetected sources of the input device. Some devices may
231 provide system-dependent source names that cannot be autodetected.
232 The returned list cannot be assumed to be always complete.
233 ffmpeg -sources pulse,server=192.168.0.4
235 -sinks device[,opt1=val1[,opt2=val2]...]
237 Show autodetected sinks of the output device. Some devices may
238 provide system-dependent sink names that cannot be autodetected.
239 The returned list cannot be assumed to be always complete.
240 ffmpeg -sinks pulse,server=192.168.0.4
242 -loglevel [flags+]loglevel | -v [flags+]loglevel
244 Set logging level and flags used by the library.
245 The optional flags prefix can consist of the following values:
247 repeat
249 Indicates that repeated log output should not be compressed to
250 the first line and the "Last message repeated n times" line
251 will be omitted.
252 level
254 Indicates that log output should add a "[level]" prefix to each
255 message line. This can be used as an alternative to log
256 coloring, e.g. when dumping the log to file.
257 Flags can also be used alone by adding a '+'/'-' prefix to
259 set/reset a single flag without affecting other flags or changing
260 loglevel. When setting both flags and loglevel, a '+' separator is
261 expected between the last flags value and before loglevel.
262 loglevel is a string or a number containing one of the following
264 values:
265 quiet, -8
267 Show nothing at all; be silent.
268 panic, 0
270 Only show fatal errors which could lead the process to crash,
271 such as an assertion failure. This is not currently used for
272 anything.
273 fatal, 8
275 Only show fatal errors. These are errors after which the
276 process absolutely cannot continue.
277 error, 16
279 Show all errors, including ones which can be recovered from.
280 warning, 24
282 Show all warnings and errors. Any message related to possibly
283 incorrect or unexpected events will be shown.
284 info, 32
286 Show informative messages during processing. This is in
287 addition to warnings and errors. This is the default value.
288 verbose, 40
290 Same as "info", except more verbose.
291 debug, 48
293 Show everything, including debugging information.
294 trace, 56
296 For example to enable repeated log output, add the "level" prefix,
298 and set loglevel to "verbose":
299 ffmpeg -loglevel repeat+level+verbose -i input output
301 Another example that enables repeated log output without affecting
303 current state of "level" prefix flag or loglevel:
304 ffmpeg [...] -loglevel +repeat
306 By default the program logs to stderr. If coloring is supported by
308 the terminal, colors are used to mark errors and warnings. Log
309 coloring can be disabled setting the environment variable
310 AV_LOG_FORCE_NOCOLOR, or can be forced setting the environment
311 variable AV_LOG_FORCE_COLOR.
312 -report
314 Dump full command line and log output to a file named
315 "program-YYYYMMDD-HHMMSS.log" in the current directory. This file
316 can be useful for bug reports. It also implies "-loglevel debug".
317 Setting the environment variable FFREPORT to any value has the same
319 effect. If the value is a ':'-separated key=value sequence, these
320 options will affect the report; option values must be escaped if
321 they contain special characters or the options delimiter ':' (see
322 the ``Quoting and escaping'' section in the ffmpeg-utils manual).
323 The following options are recognized:
325 file
327 set the file name to use for the report; %p is expanded to the
328 name of the program, %t is expanded to a timestamp, "%%" is
329 expanded to a plain "%"
330 level
332 set the log verbosity level using a numerical value (see
333 "-loglevel").
334 For example, to output a report to a file named ffreport.log using
336 a log level of 32 (alias for log level "info"):
337 FFREPORT=file=ffreport.log:level=32 ffmpeg -i input output
339 Errors in parsing the environment variable are not fatal, and will
341 not appear in the report.
342 -hide_banner
344 Suppress printing banner.
345 All FFmpeg tools will normally show a copyright notice, build
347 options and library versions. This option can be used to suppress
348 printing this information.
349 -cpuflags flags (global)
351 Allows setting and clearing cpu flags. This option is intended for
352 testing. Do not use it unless you know what you're doing.
353 ffmpeg -cpuflags -sse+mmx ...
355 ffmpeg -cpuflags mmx ...
356 ffmpeg -cpuflags 0 ...
357 Possible flags for this option are:
359 x86
361 mmx
362 mmxext
363 sse
364 sse2
365 sse2slow
366 sse3
367 sse3slow
368 ssse3
369 atom
370 sse4.1
371 sse4.2
372 avx
373 avx2
374 xop
375 fma3
376 fma4
377 3dnow
378 3dnowext
379 bmi1
380 bmi2
381 cmov
382 ARM
383 armv5te
384 armv6
385 armv6t2
386 vfp
387 vfpv3
388 neon
389 setend
390 AArch64
391 armv8
392 vfp
393 neon
394 PowerPC
395 altivec
396 Specific Processors
397 pentium2
398 pentium3
399 pentium4
400 k6
401 k62
402 athlon
403 athlonxp
404 k8
405 -cpucount count (global)
406 Override detection of CPU count. This option is intended for
407 testing. Do not use it unless you know what you're doing.
408 ffmpeg -cpucount 2
410 -max_alloc bytes
412 Set the maximum size limit for allocating a block on the heap by
413 ffmpeg's family of malloc functions. Exercise extreme caution when
414 using this option. Don't use if you do not understand the full
415 consequence of doing so. Default is INT_MAX.
416 AVOptions
418 These options are provided directly by the libavformat, libavdevice and
419 libavcodec libraries. To see the list of available AVOptions, use the
420 -help option. They are separated into two categories:
421 generic
423 These options can be set for any container, codec or device.
424 Generic options are listed under AVFormatContext options for
425 containers/devices and under AVCodecContext options for codecs.
426 private
428 These options are specific to the given container, device or codec.
429 Private options are listed under their corresponding
430 containers/devices/codecs.
431 For example to write an ID3v2.3 header instead of a default ID3v2.4 to
433 an MP3 file, use the id3v2_version private option of the MP3 muxer:
434 ffmpeg -i input.flac -id3v2_version 3 out.mp3
436 All codec AVOptions are per-stream, and thus a stream specifier should
438 be attached to them:
439 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
441 In the above example, a multichannel audio stream is mapped twice for
443 output. The first instance is encoded with codec ac3 and bitrate 640k.
444 The second instance is downmixed to 2 channels and encoded with codec
445 aac. A bitrate of 128k is specified for it using absolute index of the
446 output stream.
447 Note: the -nooption syntax cannot be used for boolean AVOptions, use
449 -option 0/-option 1.
450 Note: the old undocumented way of specifying per-stream AVOptions by
452 prepending v/a/s to the options name is now obsolete and will be
453 removed soon.
454 Main options
456 -f format
457 Force format to use.
458 -unit
460 Show the unit of the displayed values.
461 -prefix
463 Use SI prefixes for the displayed values. Unless the
464 "-byte_binary_prefix" option is used all the prefixes are decimal.
465 -byte_binary_prefix
467 Force the use of binary prefixes for byte values.
468 -sexagesimal
470 Use sexagesimal format HH:MM:SS.MICROSECONDS for time values.
471 -pretty
473 Prettify the format of the displayed values, it corresponds to the
474 options "-unit -prefix -byte_binary_prefix -sexagesimal".
475 -of, -print_format writer_name[=writer_options]
477 Set the output printing format.
478 writer_name specifies the name of the writer, and writer_options
480 specifies the options to be passed to the writer.
481 For example for printing the output in JSON format, specify:
483 -print_format json
485 For more details on the available output printing formats, see the
487 Writers section below.
488 -sections
490 Print sections structure and section information, and exit. The
491 output is not meant to be parsed by a machine.
492 -select_streams stream_specifier
494 Select only the streams specified by stream_specifier. This option
495 affects only the options related to streams (e.g. "show_streams",
496 "show_packets", etc.).
497 For example to show only audio streams, you can use the command:
499 ffprobe -show_streams -select_streams a INPUT
501 To show only video packets belonging to the video stream with index
503 1:
504 ffprobe -show_packets -select_streams v:1 INPUT
506 -show_data
508 Show payload data, as a hexadecimal and ASCII dump. Coupled with
509 -show_packets, it will dump the packets' data. Coupled with
510 -show_streams, it will dump the codec extradata.
511 The dump is printed as the "data" field. It may contain newlines.
513 -show_data_hash algorithm
515 Show a hash of payload data, for packets with -show_packets and for
516 codec extradata with -show_streams.
517 -show_error
519 Show information about the error found when trying to probe the
520 input.
521 The error information is printed within a section with name
523 "ERROR".
524 -show_format
526 Show information about the container format of the input multimedia
527 stream.
528 All the container format information is printed within a section
530 with name "FORMAT".
531 -show_format_entry name
533 Like -show_format, but only prints the specified entry of the
534 container format information, rather than all. This option may be
535 given more than once, then all specified entries will be shown.
536 This option is deprecated, use "show_entries" instead.
538 -show_entries section_entries
540 Set list of entries to show.
541 Entries are specified according to the following syntax.
543 section_entries contains a list of section entries separated by
544 ":". Each section entry is composed by a section name (or unique
545 name), optionally followed by a list of entries local to that
546 section, separated by ",".
547 If section name is specified but is followed by no "=", all entries
549 are printed to output, together with all the contained sections.
550 Otherwise only the entries specified in the local section entries
551 list are printed. In particular, if "=" is specified but the list
552 of local entries is empty, then no entries will be shown for that
553 section.
554 Note that the order of specification of the local section entries
556 is not honored in the output, and the usual display order will be
557 retained.
558 The formal syntax is given by:
560 <LOCAL_SECTION_ENTRIES> ::= <SECTION_ENTRY_NAME>[,<LOCAL_SECTION_ENTRIES>]
562 <SECTION_ENTRY> ::= <SECTION_NAME>[=[<LOCAL_SECTION_ENTRIES>]]
563 <SECTION_ENTRIES> ::= <SECTION_ENTRY>[:<SECTION_ENTRIES>]
564 For example, to show only the index and type of each stream, and
566 the PTS time, duration time, and stream index of the packets, you
567 can specify the argument:
568 packet=pts_time,duration_time,stream_index : stream=index,codec_type
570 To show all the entries in the section "format", but only the codec
572 type in the section "stream", specify the argument:
573 format : stream=codec_type
575 To show all the tags in the stream and format sections:
577 stream_tags : format_tags
579 To show only the "title" tag (if available) in the stream sections:
581 stream_tags=title
583 -show_packets
585 Show information about each packet contained in the input
586 multimedia stream.
587 The information for each single packet is printed within a
589 dedicated section with name "PACKET".
590 -show_frames
592 Show information about each frame and subtitle contained in the
593 input multimedia stream.
594 The information for each single frame is printed within a dedicated
596 section with name "FRAME" or "SUBTITLE".
597 -show_log loglevel
599 Show logging information from the decoder about each frame
600 according to the value set in loglevel, (see "-loglevel"). This
601 option requires "-show_frames".
602 The information for each log message is printed within a dedicated
604 section with name "LOG".
605 -show_streams
607 Show information about each media stream contained in the input
608 multimedia stream.
609 Each media stream information is printed within a dedicated section
611 with name "STREAM".
612 -show_programs
614 Show information about programs and their streams contained in the
615 input multimedia stream.
616 Each media stream information is printed within a dedicated section
618 with name "PROGRAM_STREAM".
619 -show_chapters
621 Show information about chapters stored in the format.
622 Each chapter is printed within a dedicated section with name
624 "CHAPTER".
625 -count_frames
627 Count the number of frames per stream and report it in the
628 corresponding stream section.
629 -count_packets
631 Count the number of packets per stream and report it in the
632 corresponding stream section.
633 -read_intervals read_intervals
635 Read only the specified intervals. read_intervals must be a
636 sequence of interval specifications separated by ",". ffprobe will
637 seek to the interval starting point, and will continue reading from
638 that.
639 Each interval is specified by two optional parts, separated by "%".
641 The first part specifies the interval start position. It is
643 interpreted as an absolute position, or as a relative offset from
644 the current position if it is preceded by the "+" character. If
645 this first part is not specified, no seeking will be performed when
646 reading this interval.
647 The second part specifies the interval end position. It is
649 interpreted as an absolute position, or as a relative offset from
650 the current position if it is preceded by the "+" character. If the
651 offset specification starts with "#", it is interpreted as the
652 number of packets to read (not including the flushing packets) from
653 the interval start. If no second part is specified, the program
654 will read until the end of the input.
655 Note that seeking is not accurate, thus the actual interval start
657 point may be different from the specified position. Also, when an
658 interval duration is specified, the absolute end time will be
659 computed by adding the duration to the interval start point found
660 by seeking the file, rather than to the specified start value.
661 The formal syntax is given by:
663 <INTERVAL> ::= [<START>|+<START_OFFSET>][%[<END>|+<END_OFFSET>]]
665 <INTERVALS> ::= <INTERVAL>[,<INTERVALS>]
666 A few examples follow.
668 • Seek to time 10, read packets until 20 seconds after the found
670 seek point, then seek to position "01:30" (1 minute and thirty
671 seconds) and read packets until position "01:45".
672 10%+20,01:30%01:45
674 • Read only 42 packets after seeking to position "01:23":
676 01:23%+#42
678 • Read only the first 20 seconds from the start:
680 %+20
682 • Read from the start until position "02:30":
684 %02:30
686 -show_private_data, -private
688 Show private data, that is data depending on the format of the
689 particular shown element. This option is enabled by default, but
690 you may need to disable it for specific uses, for example when
691 creating XSD-compliant XML output.
692 -show_program_version
694 Show information related to program version.
695 Version information is printed within a section with name
697 "PROGRAM_VERSION".
698 -show_library_versions
700 Show information related to library versions.
701 Version information for each library is printed within a section
703 with name "LIBRARY_VERSION".
704 -show_versions
706 Show information related to program and library versions. This is
707 the equivalent of setting both -show_program_version and
708 -show_library_versions options.
709 -show_pixel_formats
711 Show information about all pixel formats supported by FFmpeg.
712 Pixel format information for each format is printed within a
714 section with name "PIXEL_FORMAT".
715 -show_optional_fields value
717 Some writers viz. JSON and XML, omit the printing of fields with
718 invalid or non-applicable values, while other writers always print
719 them. This option enables one to control this behaviour. Valid
720 values are "always"/1, "never"/0 and "auto"/"-1". Default is auto.
721 -bitexact
723 Force bitexact output, useful to produce output which is not
724 dependent on the specific build.
725 -i input_url
727 Read input_url.
728 -o output_url
730 Write output to output_url. If not specified, the output is sent to
731 stdout.
732
734 A writer defines the output format adopted by ffprobe, and will be used
735 for printing all the parts of the output.
736 A writer may accept one or more arguments, which specify the options to
738 adopt. The options are specified as a list of key=value pairs,
739 separated by ":".
740 All writers support the following options:
742 string_validation, sv
744 Set string validation mode.
745 The following values are accepted.
747 fail
749 The writer will fail immediately in case an invalid string
750 (UTF-8) sequence or code point is found in the input. This is
751 especially useful to validate input metadata.
752 ignore
754 Any validation error will be ignored. This will result in
755 possibly broken output, especially with the json or xml writer.
756 replace
758 The writer will substitute invalid UTF-8 sequences or code
759 points with the string specified with the
760 string_validation_replacement.
761 Default value is replace.
763 string_validation_replacement, svr
765 Set replacement string to use in case string_validation is set to
766 replace.
767 In case the option is not specified, the writer will assume the
769 empty string, that is it will remove the invalid sequences from the
770 input strings.
771 A description of the currently available writers follows.
773 default
775 Default format.
776 Print each section in the form:
778 [SECTION]
780 key1=val1
781 ...
782 keyN=valN
783 [/SECTION]
784 Metadata tags are printed as a line in the corresponding FORMAT, STREAM
786 or PROGRAM_STREAM section, and are prefixed by the string "TAG:".
787 A description of the accepted options follows.
789 nokey, nk
791 If set to 1 specify not to print the key of each field. Default
792 value is 0.
793 noprint_wrappers, nw
795 If set to 1 specify not to print the section header and footer.
796 Default value is 0.
797 compact, csv
799 Compact and CSV format.
800 The "csv" writer is equivalent to "compact", but supports different
802 defaults.
803 Each section is printed on a single line. If no option is specified,
805 the output has the form:
806 section|key1=val1| ... |keyN=valN
808 Metadata tags are printed in the corresponding "format" or "stream"
810 section. A metadata tag key, if printed, is prefixed by the string
811 "tag:".
812 The description of the accepted options follows.
814 item_sep, s
816 Specify the character to use for separating fields in the output
817 line. It must be a single printable character, it is "|" by
818 default ("," for the "csv" writer).
819 nokey, nk
821 If set to 1 specify not to print the key of each field. Its default
822 value is 0 (1 for the "csv" writer).
823 escape, e
825 Set the escape mode to use, default to "c" ("csv" for the "csv"
826 writer).
827 It can assume one of the following values:
829 c Perform C-like escaping. Strings containing a newline (\n),
831 carriage return (\r), a tab (\t), a form feed (\f), the
832 escaping character (\) or the item separator character SEP are
833 escaped using C-like fashioned escaping, so that a newline is
834 converted to the sequence \n, a carriage return to \r, \ to \\
835 and the separator SEP is converted to \SEP.
836 csv Perform CSV-like escaping, as described in RFC4180. Strings
838 containing a newline (\n), a carriage return (\r), a double
839 quote ("), or SEP are enclosed in double-quotes.
840 none
842 Perform no escaping.
843 print_section, p
845 Print the section name at the beginning of each line if the value
846 is 1, disable it with value set to 0. Default value is 1.
847 flat
849 Flat format.
850 A free-form output where each line contains an explicit key=value, such
852 as "streams.stream.3.tags.foo=bar". The output is shell escaped, so it
853 can be directly embedded in sh scripts as long as the separator
854 character is an alphanumeric character or an underscore (see sep_char
855 option).
856 The description of the accepted options follows.
858 sep_char, s
860 Separator character used to separate the chapter, the section name,
861 IDs and potential tags in the printed field key.
862 Default value is ..
864 hierarchical, h
866 Specify if the section name specification should be hierarchical.
867 If set to 1, and if there is more than one section in the current
868 chapter, the section name will be prefixed by the name of the
869 chapter. A value of 0 will disable this behavior.
870 Default value is 1.
872 ini
874 INI format output.
875 Print output in an INI based format.
877 The following conventions are adopted:
879 • all key and values are UTF-8
881 • . is the subgroup separator
883 • newline, \t, \f, \b and the following characters are escaped
885 • \ is the escape character
887 • # is the comment indicator
889 • = is the key/value separator
891 • : is not used but usually parsed as key/value separator
893 This writer accepts options as a list of key=value pairs, separated by
895 :.
896 The description of the accepted options follows.
898 hierarchical, h
900 Specify if the section name specification should be hierarchical.
901 If set to 1, and if there is more than one section in the current
902 chapter, the section name will be prefixed by the name of the
903 chapter. A value of 0 will disable this behavior.
904 Default value is 1.
906 json
908 JSON based format.
909 Each section is printed using JSON notation.
911 The description of the accepted options follows.
913 compact, c
915 If set to 1 enable compact output, that is each section will be
916 printed on a single line. Default value is 0.
917 For more information about JSON, see <http://www.json.org/>.
919 xml
921 XML based format.
922 The XML output is described in the XML schema description file
924 ffprobe.xsd installed in the FFmpeg datadir.
925 An updated version of the schema can be retrieved at the url
927 <http://www.ffmpeg.org/schema/ffprobe.xsd>, which redirects to the
928 latest schema committed into the FFmpeg development source code tree.
929 Note that the output issued will be compliant to the ffprobe.xsd schema
931 only when no special global output options (unit, prefix,
932 byte_binary_prefix, sexagesimal etc.) are specified.
933 The description of the accepted options follows.
935 fully_qualified, q
937 If set to 1 specify if the output should be fully qualified.
938 Default value is 0. This is required for generating an XML file
939 which can be validated through an XSD file.
940 xsd_strict, x
942 If set to 1 perform more checks for ensuring that the output is XSD
943 compliant. Default value is 0. This option automatically sets
944 fully_qualified to 1.
945 For more information about the XML format, see
947 <https://www.w3.org/XML/>.
948
950 ffprobe supports Timecode extraction:
951 • MPEG1/2 timecode is extracted from the GOP, and is available in the
953 video stream details (-show_streams, see timecode).
954 • MOV timecode is extracted from tmcd track, so is available in the
956 tmcd stream metadata (-show_streams, see TAG:timecode).
957 • DV, GXF and AVI timecodes are available in format metadata
959 (-show_format, see TAG:timecode).
960
962 This section documents the syntax and formats employed by the FFmpeg
963 libraries and tools.
964 Quoting and escaping
966 FFmpeg adopts the following quoting and escaping mechanism, unless
967 explicitly specified. The following rules are applied:
968 • ' and \ are special characters (respectively used for quoting and
970 escaping). In addition to them, there might be other special
971 characters depending on the specific syntax where the escaping and
972 quoting are employed.
973 • A special character is escaped by prefixing it with a \.
975 • All characters enclosed between '' are included literally in the
977 parsed string. The quote character ' itself cannot be quoted, so
978 you may need to close the quote and escape it.
979 • Leading and trailing whitespaces, unless escaped or quoted, are
981 removed from the parsed string.
982 Note that you may need to add a second level of escaping when using the
984 command line or a script, which depends on the syntax of the adopted
985 shell language.
986 The function "av_get_token" defined in libavutil/avstring.h can be used
988 to parse a token quoted or escaped according to the rules defined
989 above.
990 The tool tools/ffescape in the FFmpeg source tree can be used to
992 automatically quote or escape a string in a script.
993 Examples
995 • Escape the string "Crime d'Amour" containing the "'" special
997 character:
998 Crime d\'Amour
1000 • The string above contains a quote, so the "'" needs to be escaped
1002 when quoting it:
1003 'Crime d'\''Amour'
1005 • Include leading or trailing whitespaces using quoting:
1007 ' this string starts and ends with whitespaces '
1009 • Escaping and quoting can be mixed together:
1011 ' The string '\'string\'' is a string '
1013 • To include a literal \ you can use either escaping or quoting:
1015 'c:\foo' can be written as c:\\foo
1017 Date
1019 The accepted syntax is:
1020 [(YYYY-MM-DD|YYYYMMDD)[T|t| ]]((HH:MM:SS[.m...]]])|(HHMMSS[.m...]]]))[Z]
1022 now
1023 If the value is "now" it takes the current time.
1025 Time is local time unless Z is appended, in which case it is
1027 interpreted as UTC. If the year-month-day part is not specified it
1028 takes the current year-month-day.
1029 Time duration
1031 There are two accepted syntaxes for expressing time duration.
1032 [-][<HH>:]<MM>:<SS>[.<m>...]
1034 HH expresses the number of hours, MM the number of minutes for a
1036 maximum of 2 digits, and SS the number of seconds for a maximum of 2
1037 digits. The m at the end expresses decimal value for SS.
1038 or
1040 [-]<S>+[.<m>...][s|ms|us]
1042 S expresses the number of seconds, with the optional decimal part m.
1044 The optional literal suffixes s, ms or us indicate to interpret the
1045 value as seconds, milliseconds or microseconds, respectively.
1046 In both expressions, the optional - indicates negative duration.
1048 Examples
1050 The following examples are all valid time duration:
1052 55 55 seconds
1054 0.2 0.2 seconds
1056 200ms
1058 200 milliseconds, that's 0.2s
1059 200000us
1061 200000 microseconds, that's 0.2s
1062 12:03:45
1064 12 hours, 03 minutes and 45 seconds
1065 23.189
1067 23.189 seconds
1068 Video size
1070 Specify the size of the sourced video, it may be a string of the form
1071 widthxheight, or the name of a size abbreviation.
1072 The following abbreviations are recognized:
1074 ntsc
1076 720x480
1077 pal 720x576
1079 qntsc
1081 352x240
1082 qpal
1084 352x288
1085 sntsc
1087 640x480
1088 spal
1090 768x576
1091 film
1093 352x240
1094 ntsc-film
1096 352x240
1097 sqcif
1099 128x96
1100 qcif
1102 176x144
1103 cif 352x288
1105 4cif
1107 704x576
1108 16cif
1110 1408x1152
1111 qqvga
1113 160x120
1114 qvga
1116 320x240
1117 vga 640x480
1119 svga
1121 800x600
1122 xga 1024x768
1124 uxga
1126 1600x1200
1127 qxga
1129 2048x1536
1130 sxga
1132 1280x1024
1133 qsxga
1135 2560x2048
1136 hsxga
1138 5120x4096
1139 wvga
1141 852x480
1142 wxga
1144 1366x768
1145 wsxga
1147 1600x1024
1148 wuxga
1150 1920x1200
1151 woxga
1153 2560x1600
1154 wqsxga
1156 3200x2048
1157 wquxga
1159 3840x2400
1160 whsxga
1162 6400x4096
1163 whuxga
1165 7680x4800
1166 cga 320x200
1168 ega 640x350
1170 hd480
1172 852x480
1173 hd720
1175 1280x720
1176 hd1080
1178 1920x1080
1179 2k 2048x1080
1181 2kflat
1183 1998x1080
1184 2kscope
1186 2048x858
1187 4k 4096x2160
1189 4kflat
1191 3996x2160
1192 4kscope
1194 4096x1716
1195 nhd 640x360
1197 hqvga
1199 240x160
1200 wqvga
1202 400x240
1203 fwqvga
1205 432x240
1206 hvga
1208 480x320
1209 qhd 960x540
1211 2kdci
1213 2048x1080
1214 4kdci
1216 4096x2160
1217 uhd2160
1219 3840x2160
1220 uhd4320
1222 7680x4320
1223 Video rate
1225 Specify the frame rate of a video, expressed as the number of frames
1226 generated per second. It has to be a string in the format
1227 frame_rate_num/frame_rate_den, an integer number, a float number or a
1228 valid video frame rate abbreviation.
1229 The following abbreviations are recognized:
1231 ntsc
1233 30000/1001
1234 pal 25/1
1236 qntsc
1238 30000/1001
1239 qpal
1241 25/1
1242 sntsc
1244 30000/1001
1245 spal
1247 25/1
1248 film
1250 24/1
1251 ntsc-film
1253 24000/1001
1254 Ratio
1256 A ratio can be expressed as an expression, or in the form
1257 numerator:denominator.
1258 Note that a ratio with infinite (1/0) or negative value is considered
1260 valid, so you should check on the returned value if you want to exclude
1261 those values.
1262 The undefined value can be expressed using the "0:0" string.
1264 Color
1266 It can be the name of a color as defined below (case insensitive match)
1267 or a "[0x|#]RRGGBB[AA]" sequence, possibly followed by @ and a string
1268 representing the alpha component.
1269 The alpha component may be a string composed by "0x" followed by an
1271 hexadecimal number or a decimal number between 0.0 and 1.0, which
1272 represents the opacity value (0x00 or 0.0 means completely transparent,
1273 0xff or 1.0 completely opaque). If the alpha component is not specified
1274 then 0xff is assumed.
1275 The string random will result in a random color.
1277 The following names of colors are recognized:
1279 AliceBlue
1281 0xF0F8FF
1282 AntiqueWhite
1284 0xFAEBD7
1285 Aqua
1287 0x00FFFF
1288 Aquamarine
1290 0x7FFFD4
1291 Azure
1293 0xF0FFFF
1294 Beige
1296 0xF5F5DC
1297 Bisque
1299 0xFFE4C4
1300 Black
1302 0x000000
1303 BlanchedAlmond
1305 0xFFEBCD
1306 Blue
1308 0x0000FF
1309 BlueViolet
1311 0x8A2BE2
1312 Brown
1314 0xA52A2A
1315 BurlyWood
1317 0xDEB887
1318 CadetBlue
1320 0x5F9EA0
1321 Chartreuse
1323 0x7FFF00
1324 Chocolate
1326 0xD2691E
1327 Coral
1329 0xFF7F50
1330 CornflowerBlue
1332 0x6495ED
1333 Cornsilk
1335 0xFFF8DC
1336 Crimson
1338 0xDC143C
1339 Cyan
1341 0x00FFFF
1342 DarkBlue
1344 0x00008B
1345 DarkCyan
1347 0x008B8B
1348 DarkGoldenRod
1350 0xB8860B
1351 DarkGray
1353 0xA9A9A9
1354 DarkGreen
1356 0x006400
1357 DarkKhaki
1359 0xBDB76B
1360 DarkMagenta
1362 0x8B008B
1363 DarkOliveGreen
1365 0x556B2F
1366 Darkorange
1368 0xFF8C00
1369 DarkOrchid
1371 0x9932CC
1372 DarkRed
1374 0x8B0000
1375 DarkSalmon
1377 0xE9967A
1378 DarkSeaGreen
1380 0x8FBC8F
1381 DarkSlateBlue
1383 0x483D8B
1384 DarkSlateGray
1386 0x2F4F4F
1387 DarkTurquoise
1389 0x00CED1
1390 DarkViolet
1392 0x9400D3
1393 DeepPink
1395 0xFF1493
1396 DeepSkyBlue
1398 0x00BFFF
1399 DimGray
1401 0x696969
1402 DodgerBlue
1404 0x1E90FF
1405 FireBrick
1407 0xB22222
1408 FloralWhite
1410 0xFFFAF0
1411 ForestGreen
1413 0x228B22
1414 Fuchsia
1416 0xFF00FF
1417 Gainsboro
1419 0xDCDCDC
1420 GhostWhite
1422 0xF8F8FF
1423 Gold
1425 0xFFD700
1426 GoldenRod
1428 0xDAA520
1429 Gray
1431 0x808080
1432 Green
1434 0x008000
1435 GreenYellow
1437 0xADFF2F
1438 HoneyDew
1440 0xF0FFF0
1441 HotPink
1443 0xFF69B4
1444 IndianRed
1446 0xCD5C5C
1447 Indigo
1449 0x4B0082
1450 Ivory
1452 0xFFFFF0
1453 Khaki
1455 0xF0E68C
1456 Lavender
1458 0xE6E6FA
1459 LavenderBlush
1461 0xFFF0F5
1462 LawnGreen
1464 0x7CFC00
1465 LemonChiffon
1467 0xFFFACD
1468 LightBlue
1470 0xADD8E6
1471 LightCoral
1473 0xF08080
1474 LightCyan
1476 0xE0FFFF
1477 LightGoldenRodYellow
1479 0xFAFAD2
1480 LightGreen
1482 0x90EE90
1483 LightGrey
1485 0xD3D3D3
1486 LightPink
1488 0xFFB6C1
1489 LightSalmon
1491 0xFFA07A
1492 LightSeaGreen
1494 0x20B2AA
1495 LightSkyBlue
1497 0x87CEFA
1498 LightSlateGray
1500 0x778899
1501 LightSteelBlue
1503 0xB0C4DE
1504 LightYellow
1506 0xFFFFE0
1507 Lime
1509 0x00FF00
1510 LimeGreen
1512 0x32CD32
1513 Linen
1515 0xFAF0E6
1516 Magenta
1518 0xFF00FF
1519 Maroon
1521 0x800000
1522 MediumAquaMarine
1524 0x66CDAA
1525 MediumBlue
1527 0x0000CD
1528 MediumOrchid
1530 0xBA55D3
1531 MediumPurple
1533 0x9370D8
1534 MediumSeaGreen
1536 0x3CB371
1537 MediumSlateBlue
1539 0x7B68EE
1540 MediumSpringGreen
1542 0x00FA9A
1543 MediumTurquoise
1545 0x48D1CC
1546 MediumVioletRed
1548 0xC71585
1549 MidnightBlue
1551 0x191970
1552 MintCream
1554 0xF5FFFA
1555 MistyRose
1557 0xFFE4E1
1558 Moccasin
1560 0xFFE4B5
1561 NavajoWhite
1563 0xFFDEAD
1564 Navy
1566 0x000080
1567 OldLace
1569 0xFDF5E6
1570 Olive
1572 0x808000
1573 OliveDrab
1575 0x6B8E23
1576 Orange
1578 0xFFA500
1579 OrangeRed
1581 0xFF4500
1582 Orchid
1584 0xDA70D6
1585 PaleGoldenRod
1587 0xEEE8AA
1588 PaleGreen
1590 0x98FB98
1591 PaleTurquoise
1593 0xAFEEEE
1594 PaleVioletRed
1596 0xD87093
1597 PapayaWhip
1599 0xFFEFD5
1600 PeachPuff
1602 0xFFDAB9
1603 Peru
1605 0xCD853F
1606 Pink
1608 0xFFC0CB
1609 Plum
1611 0xDDA0DD
1612 PowderBlue
1614 0xB0E0E6
1615 Purple
1617 0x800080
1618 Red 0xFF0000
1620 RosyBrown
1622 0xBC8F8F
1623 RoyalBlue
1625 0x4169E1
1626 SaddleBrown
1628 0x8B4513
1629 Salmon
1631 0xFA8072
1632 SandyBrown
1634 0xF4A460
1635 SeaGreen
1637 0x2E8B57
1638 SeaShell
1640 0xFFF5EE
1641 Sienna
1643 0xA0522D
1644 Silver
1646 0xC0C0C0
1647 SkyBlue
1649 0x87CEEB
1650 SlateBlue
1652 0x6A5ACD
1653 SlateGray
1655 0x708090
1656 Snow
1658 0xFFFAFA
1659 SpringGreen
1661 0x00FF7F
1662 SteelBlue
1664 0x4682B4
1665 Tan 0xD2B48C
1667 Teal
1669 0x008080
1670 Thistle
1672 0xD8BFD8
1673 Tomato
1675 0xFF6347
1676 Turquoise
1678 0x40E0D0
1679 Violet
1681 0xEE82EE
1682 Wheat
1684 0xF5DEB3
1685 White
1687 0xFFFFFF
1688 WhiteSmoke
1690 0xF5F5F5
1691 Yellow
1693 0xFFFF00
1694 YellowGreen
1696 0x9ACD32
1697 Channel Layout
1699 A channel layout specifies the spatial disposition of the channels in a
1700 multi-channel audio stream. To specify a channel layout, FFmpeg makes
1701 use of a special syntax.
1702 Individual channels are identified by an id, as given by the table
1704 below:
1705 FL front left
1707 FR front right
1709 FC front center
1711 LFE low frequency
1713 BL back left
1715 BR back right
1717 FLC front left-of-center
1719 FRC front right-of-center
1721 BC back center
1723 SL side left
1725 SR side right
1727 TC top center
1729 TFL top front left
1731 TFC top front center
1733 TFR top front right
1735 TBL top back left
1737 TBC top back center
1739 TBR top back right
1741 DL downmix left
1743 DR downmix right
1745 WL wide left
1747 WR wide right
1749 SDL surround direct left
1751 SDR surround direct right
1753 LFE2
1755 low frequency 2
1756 Standard channel layout compositions can be specified by using the
1758 following identifiers:
1759 mono
1761 FC
1762 stereo
1764 FL+FR
1765 2.1 FL+FR+LFE
1767 3.0 FL+FR+FC
1769 3.0(back)
1771 FL+FR+BC
1772 4.0 FL+FR+FC+BC
1774 quad
1776 FL+FR+BL+BR
1777 quad(side)
1779 FL+FR+SL+SR
1780 3.1 FL+FR+FC+LFE
1782 5.0 FL+FR+FC+BL+BR
1784 5.0(side)
1786 FL+FR+FC+SL+SR
1787 4.1 FL+FR+FC+LFE+BC
1789 5.1 FL+FR+FC+LFE+BL+BR
1791 5.1(side)
1793 FL+FR+FC+LFE+SL+SR
1794 6.0 FL+FR+FC+BC+SL+SR
1796 6.0(front)
1798 FL+FR+FLC+FRC+SL+SR
1799 hexagonal
1801 FL+FR+FC+BL+BR+BC
1802 6.1 FL+FR+FC+LFE+BC+SL+SR
1804 6.1 FL+FR+FC+LFE+BL+BR+BC
1806 6.1(front)
1808 FL+FR+LFE+FLC+FRC+SL+SR
1809 7.0 FL+FR+FC+BL+BR+SL+SR
1811 7.0(front)
1813 FL+FR+FC+FLC+FRC+SL+SR
1814 7.1 FL+FR+FC+LFE+BL+BR+SL+SR
1816 7.1(wide)
1818 FL+FR+FC+LFE+BL+BR+FLC+FRC
1819 7.1(wide-side)
1821 FL+FR+FC+LFE+FLC+FRC+SL+SR
1822 octagonal
1824 FL+FR+FC+BL+BR+BC+SL+SR
1825 hexadecagonal
1827 FL+FR+FC+BL+BR+BC+SL+SR+WL+WR+TBL+TBR+TBC+TFC+TFL+TFR
1828 downmix
1830 DL+DR
1831 22.2
1833 FL+FR+FC+LFE+BL+BR+FLC+FRC+BC+SL+SR+TC+TFL+TFC+TFR+TBL+TBC+TBR+LFE2+TSL+TSR+BFC+BFL+BFR
1834 A custom channel layout can be specified as a sequence of terms,
1836 separated by '+'. Each term can be:
1837 • the name of a single channel (e.g. FL, FR, FC, LFE, etc.), each
1839 optionally containing a custom name after a '@', (e.g. FL@Left,
1840 FR@Right, FC@Center, LFE@Low_Frequency, etc.)
1841 A standard channel layout can be specified by the following:
1843 • the name of a single channel (e.g. FL, FR, FC, LFE, etc.)
1845 • the name of a standard channel layout (e.g. mono, stereo, 4.0,
1847 quad, 5.0, etc.)
1848 • a number of channels, in decimal, followed by 'c', yielding the
1850 default channel layout for that number of channels (see the
1851 function "av_channel_layout_default"). Note that not all channel
1852 counts have a default layout.
1853 • a number of channels, in decimal, followed by 'C', yielding an
1855 unknown channel layout with the specified number of channels. Note
1856 that not all channel layout specification strings support unknown
1857 channel layouts.
1858 • a channel layout mask, in hexadecimal starting with "0x" (see the
1860 "AV_CH_*" macros in libavutil/channel_layout.h.
1861 Before libavutil version 53 the trailing character "c" to specify a
1863 number of channels was optional, but now it is required, while a
1864 channel layout mask can also be specified as a decimal number (if and
1865 only if not followed by "c" or "C").
1866 See also the function "av_channel_layout_from_string" defined in
1868 libavutil/channel_layout.h.
1869
1871 When evaluating an arithmetic expression, FFmpeg uses an internal
1872 formula evaluator, implemented through the libavutil/eval.h interface.
1873 An expression may contain unary, binary operators, constants, and
1875 functions.
1876 Two expressions expr1 and expr2 can be combined to form another
1878 expression "expr1;expr2". expr1 and expr2 are evaluated in turn, and
1879 the new expression evaluates to the value of expr2.
1880 The following binary operators are available: "+", "-", "*", "/", "^".
1882 The following unary operators are available: "+", "-".
1884 The following functions are available:
1886 abs(x)
1888 Compute absolute value of x.
1889 acos(x)
1891 Compute arccosine of x.
1892 asin(x)
1894 Compute arcsine of x.
1895 atan(x)
1897 Compute arctangent of x.
1898 atan2(x, y)
1900 Compute principal value of the arc tangent of y/x.
1901 between(x, min, max)
1903 Return 1 if x is greater than or equal to min and lesser than or
1904 equal to max, 0 otherwise.
1905 bitand(x, y)
1907 bitor(x, y)
1908 Compute bitwise and/or operation on x and y.
1909 The results of the evaluation of x and y are converted to integers
1911 before executing the bitwise operation.
1912 Note that both the conversion to integer and the conversion back to
1914 floating point can lose precision. Beware of unexpected results for
1915 large numbers (usually 2^53 and larger).
1916 ceil(expr)
1918 Round the value of expression expr upwards to the nearest integer.
1919 For example, "ceil(1.5)" is "2.0".
1920 clip(x, min, max)
1922 Return the value of x clipped between min and max.
1923 cos(x)
1925 Compute cosine of x.
1926 cosh(x)
1928 Compute hyperbolic cosine of x.
1929 eq(x, y)
1931 Return 1 if x and y are equivalent, 0 otherwise.
1932 exp(x)
1934 Compute exponential of x (with base "e", the Euler's number).
1935 floor(expr)
1937 Round the value of expression expr downwards to the nearest
1938 integer. For example, "floor(-1.5)" is "-2.0".
1939 gauss(x)
1941 Compute Gauss function of x, corresponding to "exp(-x*x/2) /
1942 sqrt(2*PI)".
1943 gcd(x, y)
1945 Return the greatest common divisor of x and y. If both x and y are
1946 0 or either or both are less than zero then behavior is undefined.
1947 gt(x, y)
1949 Return 1 if x is greater than y, 0 otherwise.
1950 gte(x, y)
1952 Return 1 if x is greater than or equal to y, 0 otherwise.
1953 hypot(x, y)
1955 This function is similar to the C function with the same name; it
1956 returns "sqrt(x*x + y*y)", the length of the hypotenuse of a right
1957 triangle with sides of length x and y, or the distance of the point
1958 (x, y) from the origin.
1959 if(x, y)
1961 Evaluate x, and if the result is non-zero return the result of the
1962 evaluation of y, return 0 otherwise.
1963 if(x, y, z)
1965 Evaluate x, and if the result is non-zero return the evaluation
1966 result of y, otherwise the evaluation result of z.
1967 ifnot(x, y)
1969 Evaluate x, and if the result is zero return the result of the
1970 evaluation of y, return 0 otherwise.
1971 ifnot(x, y, z)
1973 Evaluate x, and if the result is zero return the evaluation result
1974 of y, otherwise the evaluation result of z.
1975 isinf(x)
1977 Return 1.0 if x is +/-INFINITY, 0.0 otherwise.
1978 isnan(x)
1980 Return 1.0 if x is NAN, 0.0 otherwise.
1981 ld(var)
1983 Load the value of the internal variable with number var, which was
1984 previously stored with st(var, expr). The function returns the
1985 loaded value.
1986 lerp(x, y, z)
1988 Return linear interpolation between x and y by amount of z.
1989 log(x)
1991 Compute natural logarithm of x.
1992 lt(x, y)
1994 Return 1 if x is lesser than y, 0 otherwise.
1995 lte(x, y)
1997 Return 1 if x is lesser than or equal to y, 0 otherwise.
1998 max(x, y)
2000 Return the maximum between x and y.
2001 min(x, y)
2003 Return the minimum between x and y.
2004 mod(x, y)
2006 Compute the remainder of division of x by y.
2007 not(expr)
2009 Return 1.0 if expr is zero, 0.0 otherwise.
2010 pow(x, y)
2012 Compute the power of x elevated y, it is equivalent to "(x)^(y)".
2013 print(t)
2015 print(t, l)
2016 Print the value of expression t with loglevel l. If l is not
2017 specified then a default log level is used. Returns the value of
2018 the expression printed.
2019 Prints t with loglevel l
2021 random(x)
2023 Return a pseudo random value between 0.0 and 1.0. x is the index of
2024 the internal variable which will be used to save the seed/state.
2025 root(expr, max)
2027 Find an input value for which the function represented by expr with
2028 argument ld(0) is 0 in the interval 0..max.
2029 The expression in expr must denote a continuous function or the
2031 result is undefined.
2032 ld(0) is used to represent the function input value, which means
2034 that the given expression will be evaluated multiple times with
2035 various input values that the expression can access through ld(0).
2036 When the expression evaluates to 0 then the corresponding input
2037 value will be returned.
2038 round(expr)
2040 Round the value of expression expr to the nearest integer. For
2041 example, "round(1.5)" is "2.0".
2042 sgn(x)
2044 Compute sign of x.
2045 sin(x)
2047 Compute sine of x.
2048 sinh(x)
2050 Compute hyperbolic sine of x.
2051 sqrt(expr)
2053 Compute the square root of expr. This is equivalent to "(expr)^.5".
2054 squish(x)
2056 Compute expression "1/(1 + exp(4*x))".
2057 st(var, expr)
2059 Store the value of the expression expr in an internal variable. var
2060 specifies the number of the variable where to store the value, and
2061 it is a value ranging from 0 to 9. The function returns the value
2062 stored in the internal variable. Note, Variables are currently not
2063 shared between expressions.
2064 tan(x)
2066 Compute tangent of x.
2067 tanh(x)
2069 Compute hyperbolic tangent of x.
2070 taylor(expr, x)
2072 taylor(expr, x, id)
2073 Evaluate a Taylor series at x, given an expression representing the
2074 "ld(id)"-th derivative of a function at 0.
2075 When the series does not converge the result is undefined.
2077 ld(id) is used to represent the derivative order in expr, which
2079 means that the given expression will be evaluated multiple times
2080 with various input values that the expression can access through
2081 "ld(id)". If id is not specified then 0 is assumed.
2082 Note, when you have the derivatives at y instead of 0,
2084 "taylor(expr, x-y)" can be used.
2085 time(0)
2087 Return the current (wallclock) time in seconds.
2088 trunc(expr)
2090 Round the value of expression expr towards zero to the nearest
2091 integer. For example, "trunc(-1.5)" is "-1.0".
2092 while(cond, expr)
2094 Evaluate expression expr while the expression cond is non-zero, and
2095 returns the value of the last expr evaluation, or NAN if cond was
2096 always false.
2097 The following constants are available:
2099 PI area of the unit disc, approximately 3.14
2101 E exp(1) (Euler's number), approximately 2.718
2103 PHI golden ratio (1+sqrt(5))/2, approximately 1.618
2105 Assuming that an expression is considered "true" if it has a non-zero
2107 value, note that:
2108 "*" works like AND
2110 "+" works like OR
2112 For example the construct:
2114 if (A AND B) then C
2116 is equivalent to:
2118 if(A*B, C)
2120 In your C code, you can extend the list of unary and binary functions,
2122 and define recognized constants, so that they are available for your
2123 expressions.
2124 The evaluator also recognizes the International System unit prefixes.
2126 If 'i' is appended after the prefix, binary prefixes are used, which
2127 are based on powers of 1024 instead of powers of 1000. The 'B' postfix
2128 multiplies the value by 8, and can be appended after a unit prefix or
2129 used alone. This allows using for example 'KB', 'MiB', 'G' and 'B' as
2130 number postfix.
2131 The list of available International System prefixes follows, with
2133 indication of the corresponding powers of 10 and of 2.
2134 y 10^-24 / 2^-80
2136 z 10^-21 / 2^-70
2138 a 10^-18 / 2^-60
2140 f 10^-15 / 2^-50
2142 p 10^-12 / 2^-40
2144 n 10^-9 / 2^-30
2146 u 10^-6 / 2^-20
2148 m 10^-3 / 2^-10
2150 c 10^-2
2152 d 10^-1
2154 h 10^2
2156 k 10^3 / 2^10
2158 K 10^3 / 2^10
2160 M 10^6 / 2^20
2162 G 10^9 / 2^30
2164 T 10^12 / 2^40
2166 P 10^15 / 2^40
2168 E 10^18 / 2^50
2170 Z 10^21 / 2^60
2172 Y 10^24 / 2^70
2174
2176 libavcodec provides some generic global options, which can be set on
2177 all the encoders and decoders. In addition each codec may support so-
2178 called private options, which are specific for a given codec.
2179 Sometimes, a global option may only affect a specific kind of codec,
2181 and may be nonsensical or ignored by another, so you need to be aware
2182 of the meaning of the specified options. Also some options are meant
2183 only for decoding or encoding.
2184 Options may be set by specifying -option value in the FFmpeg tools, or
2186 by setting the value explicitly in the "AVCodecContext" options or
2187 using the libavutil/opt.h API for programmatic use.
2188 The list of supported options follow:
2190 b integer (encoding,audio,video)
2192 Set bitrate in bits/s. Default value is 200K.
2193 ab integer (encoding,audio)
2195 Set audio bitrate (in bits/s). Default value is 128K.
2196 bt integer (encoding,video)
2198 Set video bitrate tolerance (in bits/s). In 1-pass mode, bitrate
2199 tolerance specifies how far ratecontrol is willing to deviate from
2200 the target average bitrate value. This is not related to min/max
2201 bitrate. Lowering tolerance too much has an adverse effect on
2202 quality.
2203 flags flags (decoding/encoding,audio,video,subtitles)
2205 Set generic flags.
2206 Possible values:
2208 mv4 Use four motion vector by macroblock (mpeg4).
2210 qpel
2212 Use 1/4 pel motion compensation.
2213 loop
2215 Use loop filter.
2216 qscale
2218 Use fixed qscale.
2219 pass1
2221 Use internal 2pass ratecontrol in first pass mode.
2222 pass2
2224 Use internal 2pass ratecontrol in second pass mode.
2225 gray
2227 Only decode/encode grayscale.
2228 psnr
2230 Set error[?] variables during encoding.
2231 truncated
2233 Input bitstream might be randomly truncated.
2234 drop_changed
2236 Don't output frames whose parameters differ from first decoded
2237 frame in stream. Error AVERROR_INPUT_CHANGED is returned when
2238 a frame is dropped.
2239 ildct
2241 Use interlaced DCT.
2242 low_delay
2244 Force low delay.
2245 global_header
2247 Place global headers in extradata instead of every keyframe.
2248 bitexact
2250 Only write platform-, build- and time-independent data. (except
2251 (I)DCT). This ensures that file and data checksums are
2252 reproducible and match between platforms. Its primary use is
2253 for regression testing.
2254 aic Apply H263 advanced intra coding / mpeg4 ac prediction.
2256 ilme
2258 Apply interlaced motion estimation.
2259 cgop
2261 Use closed gop.
2262 output_corrupt
2264 Output even potentially corrupted frames.
2265 time_base rational number
2267 Set codec time base.
2268 It is the fundamental unit of time (in seconds) in terms of which
2270 frame timestamps are represented. For fixed-fps content, timebase
2271 should be "1 / frame_rate" and timestamp increments should be
2272 identically 1.
2273 g integer (encoding,video)
2275 Set the group of picture (GOP) size. Default value is 12.
2276 ar integer (decoding/encoding,audio)
2278 Set audio sampling rate (in Hz).
2279 ac integer (decoding/encoding,audio)
2281 Set number of audio channels.
2282 cutoff integer (encoding,audio)
2284 Set cutoff bandwidth. (Supported only by selected encoders, see
2285 their respective documentation sections.)
2286 frame_size integer (encoding,audio)
2288 Set audio frame size.
2289 Each submitted frame except the last must contain exactly
2291 frame_size samples per channel. May be 0 when the codec has
2292 CODEC_CAP_VARIABLE_FRAME_SIZE set, in that case the frame size is
2293 not restricted. It is set by some decoders to indicate constant
2294 frame size.
2295 frame_number integer
2297 Set the frame number.
2298 delay integer
2300 qcomp float (encoding,video)
2301 Set video quantizer scale compression (VBR). It is used as a
2302 constant in the ratecontrol equation. Recommended range for default
2303 rc_eq: 0.0-1.0.
2304 qblur float (encoding,video)
2306 Set video quantizer scale blur (VBR).
2307 qmin integer (encoding,video)
2309 Set min video quantizer scale (VBR). Must be included between -1
2310 and 69, default value is 2.
2311 qmax integer (encoding,video)
2313 Set max video quantizer scale (VBR). Must be included between -1
2314 and 1024, default value is 31.
2315 qdiff integer (encoding,video)
2317 Set max difference between the quantizer scale (VBR).
2318 bf integer (encoding,video)
2320 Set max number of B frames between non-B-frames.
2321 Must be an integer between -1 and 16. 0 means that B-frames are
2323 disabled. If a value of -1 is used, it will choose an automatic
2324 value depending on the encoder.
2325 Default value is 0.
2327 b_qfactor float (encoding,video)
2329 Set qp factor between P and B frames.
2330 codec_tag integer
2332 bug flags (decoding,video)
2333 Workaround not auto detected encoder bugs.
2334 Possible values:
2336 autodetect
2338 xvid_ilace
2339 Xvid interlacing bug (autodetected if fourcc==XVIX)
2340 ump4
2342 (autodetected if fourcc==UMP4)
2343 no_padding
2345 padding bug (autodetected)
2346 amv
2348 qpel_chroma
2349 std_qpel
2350 old standard qpel (autodetected per fourcc/version)
2351 qpel_chroma2
2353 direct_blocksize
2354 direct-qpel-blocksize bug (autodetected per fourcc/version)
2355 edge
2357 edge padding bug (autodetected per fourcc/version)
2358 hpel_chroma
2360 dc_clip
2361 ms Workaround various bugs in microsoft broken decoders.
2362 trunc
2364 trancated frames
2365 strict integer (decoding/encoding,audio,video)
2367 Specify how strictly to follow the standards.
2368 Possible values:
2370 very
2372 strictly conform to an older more strict version of the spec or
2373 reference software
2374 strict
2376 strictly conform to all the things in the spec no matter what
2377 consequences
2378 normal
2380 unofficial
2381 allow unofficial extensions
2382 experimental
2384 allow non standardized experimental things, experimental
2385 (unfinished/work in progress/not well tested) decoders and
2386 encoders. Note: experimental decoders can pose a security
2387 risk, do not use this for decoding untrusted input.
2388 b_qoffset float (encoding,video)
2390 Set QP offset between P and B frames.
2391 err_detect flags (decoding,audio,video)
2393 Set error detection flags.
2394 Possible values:
2396 crccheck
2398 verify embedded CRCs
2399 bitstream
2401 detect bitstream specification deviations
2402 buffer
2404 detect improper bitstream length
2405 explode
2407 abort decoding on minor error detection
2408 ignore_err
2410 ignore decoding errors, and continue decoding. This is useful
2411 if you want to analyze the content of a video and thus want
2412 everything to be decoded no matter what. This option will not
2413 result in a video that is pleasing to watch in case of errors.
2414 careful
2416 consider things that violate the spec and have not been seen in
2417 the wild as errors
2418 compliant
2420 consider all spec non compliancies as errors
2421 aggressive
2423 consider things that a sane encoder should not do as an error
2424 has_b_frames integer
2426 block_align integer
2427 rc_override_count integer
2428 maxrate integer (encoding,audio,video)
2429 Set max bitrate tolerance (in bits/s). Requires bufsize to be set.
2430 minrate integer (encoding,audio,video)
2432 Set min bitrate tolerance (in bits/s). Most useful in setting up a
2433 CBR encode. It is of little use elsewise.
2434 bufsize integer (encoding,audio,video)
2436 Set ratecontrol buffer size (in bits).
2437 i_qfactor float (encoding,video)
2439 Set QP factor between P and I frames.
2440 i_qoffset float (encoding,video)
2442 Set QP offset between P and I frames.
2443 dct integer (encoding,video)
2445 Set DCT algorithm.
2446 Possible values:
2448 auto
2450 autoselect a good one (default)
2451 fastint
2453 fast integer
2454 int accurate integer
2456 mmx
2458 altivec
2459 faan
2460 floating point AAN DCT
2461 lumi_mask float (encoding,video)
2463 Compress bright areas stronger than medium ones.
2464 tcplx_mask float (encoding,video)
2466 Set temporal complexity masking.
2467 scplx_mask float (encoding,video)
2469 Set spatial complexity masking.
2470 p_mask float (encoding,video)
2472 Set inter masking.
2473 dark_mask float (encoding,video)
2475 Compress dark areas stronger than medium ones.
2476 idct integer (decoding/encoding,video)
2478 Select IDCT implementation.
2479 Possible values:
2481 auto
2483 int
2484 simple
2485 simplemmx
2486 simpleauto
2487 Automatically pick a IDCT compatible with the simple one
2488 arm
2490 altivec
2491 sh4
2492 simplearm
2493 simplearmv5te
2494 simplearmv6
2495 simpleneon
2496 xvid
2497 faani
2498 floating point AAN IDCT
2499 slice_count integer
2501 ec flags (decoding,video)
2502 Set error concealment strategy.
2503 Possible values:
2505 guess_mvs
2507 iterative motion vector (MV) search (slow)
2508 deblock
2510 use strong deblock filter for damaged MBs
2511 favor_inter
2513 favor predicting from the previous frame instead of the current
2514 bits_per_coded_sample integer
2516 aspect rational number (encoding,video)
2517 Set sample aspect ratio.
2518 sar rational number (encoding,video)
2520 Set sample aspect ratio. Alias to aspect.
2521 debug flags (decoding/encoding,audio,video,subtitles)
2523 Print specific debug info.
2524 Possible values:
2526 pict
2528 picture info
2529 rc rate control
2531 bitstream
2533 mb_type
2534 macroblock (MB) type
2535 qp per-block quantization parameter (QP)
2537 dct_coeff
2539 green_metadata
2540 display complexity metadata for the upcoming frame, GoP or for
2541 a given duration.
2542 skip
2544 startcode
2545 er error recognition
2546 mmco
2548 memory management control operations (H.264)
2549 bugs
2551 buffers
2552 picture buffer allocations
2553 thread_ops
2555 threading operations
2556 nomc
2558 skip motion compensation
2559 cmp integer (encoding,video)
2561 Set full pel me compare function.
2562 Possible values:
2564 sad sum of absolute differences, fast (default)
2566 sse sum of squared errors
2568 satd
2570 sum of absolute Hadamard transformed differences
2571 dct sum of absolute DCT transformed differences
2573 psnr
2575 sum of squared quantization errors (avoid, low quality)
2576 bit number of bits needed for the block
2578 rd rate distortion optimal, slow
2580 zero
2582 0
2583 vsad
2585 sum of absolute vertical differences
2586 vsse
2588 sum of squared vertical differences
2589 nsse
2591 noise preserving sum of squared differences
2592 w53 5/3 wavelet, only used in snow
2594 w97 9/7 wavelet, only used in snow
2596 dctmax
2598 chroma
2599 subcmp integer (encoding,video)
2600 Set sub pel me compare function.
2601 Possible values:
2603 sad sum of absolute differences, fast (default)
2605 sse sum of squared errors
2607 satd
2609 sum of absolute Hadamard transformed differences
2610 dct sum of absolute DCT transformed differences
2612 psnr
2614 sum of squared quantization errors (avoid, low quality)
2615 bit number of bits needed for the block
2617 rd rate distortion optimal, slow
2619 zero
2621 0
2622 vsad
2624 sum of absolute vertical differences
2625 vsse
2627 sum of squared vertical differences
2628 nsse
2630 noise preserving sum of squared differences
2631 w53 5/3 wavelet, only used in snow
2633 w97 9/7 wavelet, only used in snow
2635 dctmax
2637 chroma
2638 mbcmp integer (encoding,video)
2639 Set macroblock compare function.
2640 Possible values:
2642 sad sum of absolute differences, fast (default)
2644 sse sum of squared errors
2646 satd
2648 sum of absolute Hadamard transformed differences
2649 dct sum of absolute DCT transformed differences
2651 psnr
2653 sum of squared quantization errors (avoid, low quality)
2654 bit number of bits needed for the block
2656 rd rate distortion optimal, slow
2658 zero
2660 0
2661 vsad
2663 sum of absolute vertical differences
2664 vsse
2666 sum of squared vertical differences
2667 nsse
2669 noise preserving sum of squared differences
2670 w53 5/3 wavelet, only used in snow
2672 w97 9/7 wavelet, only used in snow
2674 dctmax
2676 chroma
2677 ildctcmp integer (encoding,video)
2678 Set interlaced dct compare function.
2679 Possible values:
2681 sad sum of absolute differences, fast (default)
2683 sse sum of squared errors
2685 satd
2687 sum of absolute Hadamard transformed differences
2688 dct sum of absolute DCT transformed differences
2690 psnr
2692 sum of squared quantization errors (avoid, low quality)
2693 bit number of bits needed for the block
2695 rd rate distortion optimal, slow
2697 zero
2699 0
2700 vsad
2702 sum of absolute vertical differences
2703 vsse
2705 sum of squared vertical differences
2706 nsse
2708 noise preserving sum of squared differences
2709 w53 5/3 wavelet, only used in snow
2711 w97 9/7 wavelet, only used in snow
2713 dctmax
2715 chroma
2716 dia_size integer (encoding,video)
2717 Set diamond type & size for motion estimation.
2718 (1024, INT_MAX)
2720 full motion estimation(slowest)
2721 (768, 1024]
2723 umh motion estimation
2724 (512, 768]
2726 hex motion estimation
2727 (256, 512]
2729 l2s diamond motion estimation
2730 [2,256]
2732 var diamond motion estimation
2733 (-1, 2)
2735 small diamond motion estimation
2736 -1 funny diamond motion estimation
2738 (INT_MIN, -1)
2740 sab diamond motion estimation
2741 last_pred integer (encoding,video)
2743 Set amount of motion predictors from the previous frame.
2744 precmp integer (encoding,video)
2746 Set pre motion estimation compare function.
2747 Possible values:
2749 sad sum of absolute differences, fast (default)
2751 sse sum of squared errors
2753 satd
2755 sum of absolute Hadamard transformed differences
2756 dct sum of absolute DCT transformed differences
2758 psnr
2760 sum of squared quantization errors (avoid, low quality)
2761 bit number of bits needed for the block
2763 rd rate distortion optimal, slow
2765 zero
2767 0
2768 vsad
2770 sum of absolute vertical differences
2771 vsse
2773 sum of squared vertical differences
2774 nsse
2776 noise preserving sum of squared differences
2777 w53 5/3 wavelet, only used in snow
2779 w97 9/7 wavelet, only used in snow
2781 dctmax
2783 chroma
2784 pre_dia_size integer (encoding,video)
2785 Set diamond type & size for motion estimation pre-pass.
2786 subq integer (encoding,video)
2788 Set sub pel motion estimation quality.
2789 me_range integer (encoding,video)
2791 Set limit motion vectors range (1023 for DivX player).
2792 global_quality integer (encoding,audio,video)
2794 slice_flags integer
2795 mbd integer (encoding,video)
2796 Set macroblock decision algorithm (high quality mode).
2797 Possible values:
2799 simple
2801 use mbcmp (default)
2802 bits
2804 use fewest bits
2805 rd use best rate distortion
2807 rc_init_occupancy integer (encoding,video)
2809 Set number of bits which should be loaded into the rc buffer before
2810 decoding starts.
2811 flags2 flags (decoding/encoding,audio,video,subtitles)
2813 Possible values:
2814 fast
2816 Allow non spec compliant speedup tricks.
2817 noout
2819 Skip bitstream encoding.
2820 ignorecrop
2822 Ignore cropping information from sps.
2823 local_header
2825 Place global headers at every keyframe instead of in extradata.
2826 chunks
2828 Frame data might be split into multiple chunks.
2829 showall
2831 Show all frames before the first keyframe.
2832 export_mvs
2834 Export motion vectors into frame side-data (see
2835 "AV_FRAME_DATA_MOTION_VECTORS") for codecs that support it. See
2836 also doc/examples/export_mvs.c.
2837 skip_manual
2839 Do not skip samples and export skip information as frame side
2840 data.
2841 ass_ro_flush_noop
2843 Do not reset ASS ReadOrder field on flush.
2844 export_side_data flags (decoding/encoding,audio,video,subtitles)
2846 Possible values:
2847 mvs Export motion vectors into frame side-data (see
2849 "AV_FRAME_DATA_MOTION_VECTORS") for codecs that support it. See
2850 also doc/examples/export_mvs.c.
2851 prft
2853 Export encoder Producer Reference Time into packet side-data
2854 (see "AV_PKT_DATA_PRFT") for codecs that support it.
2855 venc_params
2857 Export video encoding parameters through frame side data (see
2858 "AV_FRAME_DATA_VIDEO_ENC_PARAMS") for codecs that support it.
2859 At present, those are H.264 and VP9.
2860 film_grain
2862 Export film grain parameters through frame side data (see
2863 "AV_FRAME_DATA_FILM_GRAIN_PARAMS"). Supported at present by
2864 AV1 decoders.
2865 threads integer (decoding/encoding,video)
2867 Set the number of threads to be used, in case the selected codec
2868 implementation supports multi-threading.
2869 Possible values:
2871 auto, 0
2873 automatically select the number of threads to set
2874 Default value is auto.
2876 dc integer (encoding,video)
2878 Set intra_dc_precision.
2879 nssew integer (encoding,video)
2881 Set nsse weight.
2882 skip_top integer (decoding,video)
2884 Set number of macroblock rows at the top which are skipped.
2885 skip_bottom integer (decoding,video)
2887 Set number of macroblock rows at the bottom which are skipped.
2888 profile integer (encoding,audio,video)
2890 Set encoder codec profile. Default value is unknown. Encoder
2891 specific profiles are documented in the relevant encoder
2892 documentation.
2893 level integer (encoding,audio,video)
2895 Possible values:
2896 unknown
2898 lowres integer (decoding,audio,video)
2899 Decode at 1= 1/2, 2=1/4, 3=1/8 resolutions.
2900 mblmin integer (encoding,video)
2902 Set min macroblock lagrange factor (VBR).
2903 mblmax integer (encoding,video)
2905 Set max macroblock lagrange factor (VBR).
2906 skip_loop_filter integer (decoding,video)
2908 skip_idct integer (decoding,video)
2909 skip_frame integer (decoding,video)
2910 Make decoder discard processing depending on the frame type
2911 selected by the option value.
2912 skip_loop_filter skips frame loop filtering, skip_idct skips frame
2914 IDCT/dequantization, skip_frame skips decoding.
2915 Possible values:
2917 none
2919 Discard no frame.
2920 default
2922 Discard useless frames like 0-sized frames.
2923 noref
2925 Discard all non-reference frames.
2926 bidir
2928 Discard all bidirectional frames.
2929 nokey
2931 Discard all frames excepts keyframes.
2932 nointra
2934 Discard all frames except I frames.
2935 all Discard all frames.
2937 Default value is default.
2939 bidir_refine integer (encoding,video)
2941 Refine the two motion vectors used in bidirectional macroblocks.
2942 keyint_min integer (encoding,video)
2944 Set minimum interval between IDR-frames.
2945 refs integer (encoding,video)
2947 Set reference frames to consider for motion compensation.
2948 trellis integer (encoding,audio,video)
2950 Set rate-distortion optimal quantization.
2951 mv0_threshold integer (encoding,video)
2953 compression_level integer (encoding,audio,video)
2954 bits_per_raw_sample integer
2955 channel_layout integer (decoding/encoding,audio)
2956 Possible values:
2957 request_channel_layout integer (decoding,audio)
2959 Possible values:
2960 rc_max_vbv_use float (encoding,video)
2962 rc_min_vbv_use float (encoding,video)
2963 ticks_per_frame integer (decoding/encoding,audio,video)
2964 color_primaries integer (decoding/encoding,video)
2965 Possible values:
2966 bt709
2968 BT.709
2969 bt470m
2971 BT.470 M
2972 bt470bg
2974 BT.470 BG
2975 smpte170m
2977 SMPTE 170 M
2978 smpte240m
2980 SMPTE 240 M
2981 film
2983 Film
2984 bt2020
2986 BT.2020
2987 smpte428
2989 smpte428_1
2990 SMPTE ST 428-1
2991 smpte431
2993 SMPTE 431-2
2994 smpte432
2996 SMPTE 432-1
2997 jedec-p22
2999 JEDEC P22
3000 color_trc integer (decoding/encoding,video)
3002 Possible values:
3003 bt709
3005 BT.709
3006 gamma22
3008 BT.470 M
3009 gamma28
3011 BT.470 BG
3012 smpte170m
3014 SMPTE 170 M
3015 smpte240m
3017 SMPTE 240 M
3018 linear
3020 Linear
3021 log
3023 log100
3024 Log
3025 log_sqrt
3027 log316
3028 Log square root
3029 iec61966_2_4
3031 iec61966-2-4
3032 IEC 61966-2-4
3033 bt1361
3035 bt1361e
3036 BT.1361
3037 iec61966_2_1
3039 iec61966-2-1
3040 IEC 61966-2-1
3041 bt2020_10
3043 bt2020_10bit
3044 BT.2020 - 10 bit
3045 bt2020_12
3047 bt2020_12bit
3048 BT.2020 - 12 bit
3049 smpte2084
3051 SMPTE ST 2084
3052 smpte428
3054 smpte428_1
3055 SMPTE ST 428-1
3056 arib-std-b67
3058 ARIB STD-B67
3059 colorspace integer (decoding/encoding,video)
3061 Possible values:
3062 rgb RGB
3064 bt709
3066 BT.709
3067 fcc FCC
3069 bt470bg
3071 BT.470 BG
3072 smpte170m
3074 SMPTE 170 M
3075 smpte240m
3077 SMPTE 240 M
3078 ycocg
3080 YCOCG
3081 bt2020nc
3083 bt2020_ncl
3084 BT.2020 NCL
3085 bt2020c
3087 bt2020_cl
3088 BT.2020 CL
3089 smpte2085
3091 SMPTE 2085
3092 chroma-derived-nc
3094 Chroma-derived NCL
3095 chroma-derived-c
3097 Chroma-derived CL
3098 ictcp
3100 ICtCp
3101 color_range integer (decoding/encoding,video)
3103 If used as input parameter, it serves as a hint to the decoder,
3104 which color_range the input has. Possible values:
3105 tv
3107 mpeg
3108 MPEG (219*2^(n-8))
3109 pc
3111 jpeg
3112 JPEG (2^n-1)
3113 chroma_sample_location integer (decoding/encoding,video)
3115 Possible values:
3116 left
3118 center
3119 topleft
3120 top
3121 bottomleft
3122 bottom
3123 log_level_offset integer
3124 Set the log level offset.
3125 slices integer (encoding,video)
3127 Number of slices, used in parallelized encoding.
3128 thread_type flags (decoding/encoding,video)
3130 Select which multithreading methods to use.
3131 Use of frame will increase decoding delay by one frame per thread,
3133 so clients which cannot provide future frames should not use it.
3134 Possible values:
3136 slice
3138 Decode more than one part of a single frame at once.
3139 Multithreading using slices works only when the video was
3141 encoded with slices.
3142 frame
3144 Decode more than one frame at once.
3145 Default value is slice+frame.
3147 audio_service_type integer (encoding,audio)
3149 Set audio service type.
3150 Possible values:
3152 ma Main Audio Service
3154 ef Effects
3156 vi Visually Impaired
3158 hi Hearing Impaired
3160 di Dialogue
3162 co Commentary
3164 em Emergency
3166 vo Voice Over
3168 ka Karaoke
3170 request_sample_fmt sample_fmt (decoding,audio)
3172 Set sample format audio decoders should prefer. Default value is
3173 "none".
3174 pkt_timebase rational number
3176 sub_charenc encoding (decoding,subtitles)
3177 Set the input subtitles character encoding.
3178 field_order field_order (video)
3180 Set/override the field order of the video. Possible values:
3181 progressive
3183 Progressive video
3184 tt Interlaced video, top field coded and displayed first
3186 bb Interlaced video, bottom field coded and displayed first
3188 tb Interlaced video, top coded first, bottom displayed first
3190 bt Interlaced video, bottom coded first, top displayed first
3192 skip_alpha bool (decoding,video)
3194 Set to 1 to disable processing alpha (transparency). This works
3195 like the gray flag in the flags option which skips chroma
3196 information instead of alpha. Default is 0.
3197 codec_whitelist list (input)
3199 "," separated list of allowed decoders. By default all are allowed.
3200 dump_separator string (input)
3202 Separator used to separate the fields printed on the command line
3203 about the Stream parameters. For example, to separate the fields
3204 with newlines and indentation:
3205 ffprobe -dump_separator "
3207 " -i ~/videos/matrixbench_mpeg2.mpg
3208 max_pixels integer (decoding/encoding,video)
3210 Maximum number of pixels per image. This value can be used to avoid
3211 out of memory failures due to large images.
3212 apply_cropping bool (decoding,video)
3214 Enable cropping if cropping parameters are multiples of the
3215 required alignment for the left and top parameters. If the
3216 alignment is not met the cropping will be partially applied to
3217 maintain alignment. Default is 1 (enabled). Note: The required
3218 alignment depends on if "AV_CODEC_FLAG_UNALIGNED" is set and the
3219 CPU. "AV_CODEC_FLAG_UNALIGNED" cannot be changed from the command
3220 line. Also hardware decoders will not apply left/top Cropping.
3221
3223 Decoders are configured elements in FFmpeg which allow the decoding of
3224 multimedia streams.
3225 When you configure your FFmpeg build, all the supported native decoders
3227 are enabled by default. Decoders requiring an external library must be
3228 enabled manually via the corresponding "--enable-lib" option. You can
3229 list all available decoders using the configure option
3230 "--list-decoders".
3231 You can disable all the decoders with the configure option
3233 "--disable-decoders" and selectively enable / disable single decoders
3234 with the options "--enable-decoder=DECODER" /
3235 "--disable-decoder=DECODER".
3236 The option "-decoders" of the ff* tools will display the list of
3238 enabled decoders.
3239
3241 A description of some of the currently available video decoders
3242 follows.
3243 av1
3245 AOMedia Video 1 (AV1) decoder.
3246 Options
3248 operating_point
3250 Select an operating point of a scalable AV1 bitstream (0 - 31).
3251 Default is 0.
3252 rawvideo
3254 Raw video decoder.
3255 This decoder decodes rawvideo streams.
3257 Options
3259 top top_field_first
3261 Specify the assumed field type of the input video.
3262 -1 the video is assumed to be progressive (default)
3264 0 bottom-field-first is assumed
3266 1 top-field-first is assumed
3268 libdav1d
3270 dav1d AV1 decoder.
3271 libdav1d allows libavcodec to decode the AOMedia Video 1 (AV1) codec.
3273 Requires the presence of the libdav1d headers and library during
3274 configuration. You need to explicitly configure the build with
3275 "--enable-libdav1d".
3276 Options
3278 The following options are supported by the libdav1d wrapper.
3280 framethreads
3282 Set amount of frame threads to use during decoding. The default
3283 value is 0 (autodetect). This option is deprecated for libdav1d >=
3284 1.0 and will be removed in the future. Use the global option
3285 "threads" instead.
3286 tilethreads
3288 Set amount of tile threads to use during decoding. The default
3289 value is 0 (autodetect). This option is deprecated for libdav1d >=
3290 1.0 and will be removed in the future. Use the global option
3291 "threads" instead.
3292 filmgrain
3294 Apply film grain to the decoded video if present in the bitstream.
3295 Defaults to the internal default of the library. This option is
3296 deprecated and will be removed in the future. See the global option
3297 "export_side_data" to export Film Grain parameters instead of
3298 applying it.
3299 oppoint
3301 Select an operating point of a scalable AV1 bitstream (0 - 31).
3302 Defaults to the internal default of the library.
3303 alllayers
3305 Output all spatial layers of a scalable AV1 bitstream. The default
3306 value is false.
3307 libdavs2
3309 AVS2-P2/IEEE1857.4 video decoder wrapper.
3310 This decoder allows libavcodec to decode AVS2 streams with davs2
3312 library.
3313 libuavs3d
3315 AVS3-P2/IEEE1857.10 video decoder.
3316 libuavs3d allows libavcodec to decode AVS3 streams. Requires the
3318 presence of the libuavs3d headers and library during configuration.
3319 You need to explicitly configure the build with "--enable-libuavs3d".
3320 Options
3322 The following option is supported by the libuavs3d wrapper.
3324 frame_threads
3326 Set amount of frame threads to use during decoding. The default
3327 value is 0 (autodetect).
3328 QSV Decoders
3330 The family of Intel QuickSync Video decoders (VC1, MPEG-2, H.264, HEVC,
3331 JPEG/MJPEG, VP8, VP9, AV1).
3332 Common Options
3334 The following options are supported by all qsv decoders.
3336 async_depth
3338 Internal parallelization depth, the higher the value the higher the
3339 latency.
3340 gpu_copy
3342 A GPU-accelerated copy between video and system memory
3343 default
3345 on
3346 off
3347 HEVC Options
3349 Extra options for hevc_qsv.
3351 load_plugin
3353 A user plugin to load in an internal session
3354 none
3356 hevc_sw
3357 hevc_hw
3358 load_plugins
3359 A :-separate list of hexadecimal plugin UIDs to load in an internal
3360 session
3361 v210
3363 Uncompressed 4:2:2 10-bit decoder.
3364 Options
3366 custom_stride
3368 Set the line size of the v210 data in bytes. The default value is 0
3369 (autodetect). You can use the special -1 value for a strideless
3370 v210 as seen in BOXX files.
3371
3373 A description of some of the currently available audio decoders
3374 follows.
3375 ac3
3377 AC-3 audio decoder.
3378 This decoder implements part of ATSC A/52:2010 and ETSI TS 102 366, as
3380 well as the undocumented RealAudio 3 (a.k.a. dnet).
3381 AC-3 Decoder Options
3383 -drc_scale value
3385 Dynamic Range Scale Factor. The factor to apply to dynamic range
3386 values from the AC-3 stream. This factor is applied exponentially.
3387 The default value is 1. There are 3 notable scale factor ranges:
3388 drc_scale == 0
3390 DRC disabled. Produces full range audio.
3391 0 < drc_scale <= 1
3393 DRC enabled. Applies a fraction of the stream DRC value.
3394 Audio reproduction is between full range and full compression.
3395 drc_scale > 1
3397 DRC enabled. Applies drc_scale asymmetrically. Loud sounds are
3398 fully compressed. Soft sounds are enhanced.
3399 flac
3401 FLAC audio decoder.
3402 This decoder aims to implement the complete FLAC specification from
3404 Xiph.
3405 FLAC Decoder options
3407 -use_buggy_lpc
3409 The lavc FLAC encoder used to produce buggy streams with high lpc
3410 values (like the default value). This option makes it possible to
3411 decode such streams correctly by using lavc's old buggy lpc logic
3412 for decoding.
3413 ffwavesynth
3415 Internal wave synthesizer.
3416 This decoder generates wave patterns according to predefined sequences.
3418 Its use is purely internal and the format of the data it accepts is not
3419 publicly documented.
3420 libcelt
3422 libcelt decoder wrapper.
3423 libcelt allows libavcodec to decode the Xiph CELT ultra-low delay audio
3425 codec. Requires the presence of the libcelt headers and library during
3426 configuration. You need to explicitly configure the build with
3427 "--enable-libcelt".
3428 libgsm
3430 libgsm decoder wrapper.
3431 libgsm allows libavcodec to decode the GSM full rate audio codec.
3433 Requires the presence of the libgsm headers and library during
3434 configuration. You need to explicitly configure the build with
3435 "--enable-libgsm".
3436 This decoder supports both the ordinary GSM and the Microsoft variant.
3438 libilbc
3440 libilbc decoder wrapper.
3441 libilbc allows libavcodec to decode the Internet Low Bitrate Codec
3443 (iLBC) audio codec. Requires the presence of the libilbc headers and
3444 library during configuration. You need to explicitly configure the
3445 build with "--enable-libilbc".
3446 Options
3448 The following option is supported by the libilbc wrapper.
3450 enhance
3452 Enable the enhancement of the decoded audio when set to 1. The
3453 default value is 0 (disabled).
3454 libopencore-amrnb
3456 libopencore-amrnb decoder wrapper.
3457 libopencore-amrnb allows libavcodec to decode the Adaptive Multi-Rate
3459 Narrowband audio codec. Using it requires the presence of the
3460 libopencore-amrnb headers and library during configuration. You need to
3461 explicitly configure the build with "--enable-libopencore-amrnb".
3462 An FFmpeg native decoder for AMR-NB exists, so users can decode AMR-NB
3464 without this library.
3465 libopencore-amrwb
3467 libopencore-amrwb decoder wrapper.
3468 libopencore-amrwb allows libavcodec to decode the Adaptive Multi-Rate
3470 Wideband audio codec. Using it requires the presence of the
3471 libopencore-amrwb headers and library during configuration. You need to
3472 explicitly configure the build with "--enable-libopencore-amrwb".
3473 An FFmpeg native decoder for AMR-WB exists, so users can decode AMR-WB
3475 without this library.
3476 libopus
3478 libopus decoder wrapper.
3479 libopus allows libavcodec to decode the Opus Interactive Audio Codec.
3481 Requires the presence of the libopus headers and library during
3482 configuration. You need to explicitly configure the build with
3483 "--enable-libopus".
3484 An FFmpeg native decoder for Opus exists, so users can decode Opus
3486 without this library.
3487
3489 libaribb24
3490 ARIB STD-B24 caption decoder.
3491 Implements profiles A and C of the ARIB STD-B24 standard.
3493 libaribb24 Decoder Options
3495 -aribb24-base-path path
3497 Sets the base path for the libaribb24 library. This is utilized for
3498 reading of configuration files (for custom unicode conversions),
3499 and for dumping of non-text symbols as images under that location.
3500 Unset by default.
3502 -aribb24-skip-ruby-text boolean
3504 Tells the decoder wrapper to skip text blocks that contain half-
3505 height ruby text.
3506 Enabled by default.
3508 dvbsub
3510 Options
3511 compute_clut
3513 -2 Compute clut once if no matching CLUT is in the stream.
3514 -1 Compute clut if no matching CLUT is in the stream.
3516 0 Never compute CLUT
3518 1 Always compute CLUT and override the one provided in the
3520 stream.
3521 dvb_substream
3523 Selects the dvb substream, or all substreams if -1 which is
3524 default.
3525 dvdsub
3527 This codec decodes the bitmap subtitles used in DVDs; the same
3528 subtitles can also be found in VobSub file pairs and in some Matroska
3529 files.
3530 Options
3532 palette
3534 Specify the global palette used by the bitmaps. When stored in
3535 VobSub, the palette is normally specified in the index file; in
3536 Matroska, the palette is stored in the codec extra-data in the same
3537 format as in VobSub. In DVDs, the palette is stored in the IFO
3538 file, and therefore not available when reading from dumped VOB
3539 files.
3540 The format for this option is a string containing 16 24-bits
3542 hexadecimal numbers (without 0x prefix) separated by commas, for
3543 example "0d00ee, ee450d, 101010, eaeaea, 0ce60b, ec14ed, ebff0b,
3544 0d617a, 7b7b7b, d1d1d1, 7b2a0e, 0d950c, 0f007b, cf0dec, cfa80c,
3545 7c127b".
3546 ifo_palette
3548 Specify the IFO file from which the global palette is obtained.
3549 (experimental)
3550 forced_subs_only
3552 Only decode subtitle entries marked as forced. Some titles have
3553 forced and non-forced subtitles in the same track. Setting this
3554 flag to 1 will only keep the forced subtitles. Default value is 0.
3555 libzvbi-teletext
3557 Libzvbi allows libavcodec to decode DVB teletext pages and DVB teletext
3558 subtitles. Requires the presence of the libzvbi headers and library
3559 during configuration. You need to explicitly configure the build with
3560 "--enable-libzvbi".
3561 Options
3563 txt_page
3565 List of teletext page numbers to decode. Pages that do not match
3566 the specified list are dropped. You may use the special "*" string
3567 to match all pages, or "subtitle" to match all subtitle pages.
3568 Default value is *.
3569 txt_default_region
3571 Set default character set used for decoding, a value between 0 and
3572 87 (see ETS 300 706, Section 15, Table 32). Default value is -1,
3573 which does not override the libzvbi default. This option is needed
3574 for some legacy level 1.0 transmissions which cannot signal the
3575 proper charset.
3576 txt_chop_top
3578 Discards the top teletext line. Default value is 1.
3579 txt_format
3581 Specifies the format of the decoded subtitles.
3582 bitmap
3584 The default format, you should use this for teletext pages,
3585 because certain graphics and colors cannot be expressed in
3586 simple text or even ASS.
3587 text
3589 Simple text based output without formatting.
3590 ass Formatted ASS output, subtitle pages and teletext pages are
3592 returned in different styles, subtitle pages are stripped down
3593 to text, but an effort is made to keep the text alignment and
3594 the formatting.
3595 txt_left
3597 X offset of generated bitmaps, default is 0.
3598 txt_top
3600 Y offset of generated bitmaps, default is 0.
3601 txt_chop_spaces
3603 Chops leading and trailing spaces and removes empty lines from the
3604 generated text. This option is useful for teletext based subtitles
3605 where empty spaces may be present at the start or at the end of the
3606 lines or empty lines may be present between the subtitle lines
3607 because of double-sized teletext characters. Default value is 1.
3608 txt_duration
3610 Sets the display duration of the decoded teletext pages or
3611 subtitles in milliseconds. Default value is -1 which means infinity
3612 or until the next subtitle event comes.
3613 txt_transparent
3615 Force transparent background of the generated teletext bitmaps.
3616 Default value is 0 which means an opaque background.
3617 txt_opacity
3619 Sets the opacity (0-255) of the teletext background. If
3620 txt_transparent is not set, it only affects characters between a
3621 start box and an end box, typically subtitles. Default value is 0
3622 if txt_transparent is set, 255 otherwise.
3623
3625 When you configure your FFmpeg build, all the supported bitstream
3626 filters are enabled by default. You can list all available ones using
3627 the configure option "--list-bsfs".
3628 You can disable all the bitstream filters using the configure option
3630 "--disable-bsfs", and selectively enable any bitstream filter using the
3631 option "--enable-bsf=BSF", or you can disable a particular bitstream
3632 filter using the option "--disable-bsf=BSF".
3633 The option "-bsfs" of the ff* tools will display the list of all the
3635 supported bitstream filters included in your build.
3636 The ff* tools have a -bsf option applied per stream, taking a comma-
3638 separated list of filters, whose parameters follow the filter name
3639 after a '='.
3640 ffmpeg -i INPUT -c:v copy -bsf:v filter1[=opt1=str1:opt2=str2][,filter2] OUTPUT
3642 Below is a description of the currently available bitstream filters,
3644 with their parameters, if any.
3645 aac_adtstoasc
3647 Convert MPEG-2/4 AAC ADTS to an MPEG-4 Audio Specific Configuration
3648 bitstream.
3649 This filter creates an MPEG-4 AudioSpecificConfig from an MPEG-2/4 ADTS
3651 header and removes the ADTS header.
3652 This filter is required for example when copying an AAC stream from a
3654 raw ADTS AAC or an MPEG-TS container to MP4A-LATM, to an FLV file, or
3655 to MOV/MP4 files and related formats such as 3GP or M4A. Please note
3656 that it is auto-inserted for MP4A-LATM and MOV/MP4 and related formats.
3657 av1_metadata
3659 Modify metadata embedded in an AV1 stream.
3660 td Insert or remove temporal delimiter OBUs in all temporal units of
3662 the stream.
3663 insert
3665 Insert a TD at the beginning of every TU which does not already
3666 have one.
3667 remove
3669 Remove the TD from the beginning of every TU which has one.
3670 color_primaries
3672 transfer_characteristics
3673 matrix_coefficients
3674 Set the color description fields in the stream (see AV1 section
3675 6.4.2).
3676 color_range
3678 Set the color range in the stream (see AV1 section 6.4.2; note that
3679 this cannot be set for streams using BT.709 primaries, sRGB
3680 transfer characteristic and identity (RGB) matrix coefficients).
3681 tv Limited range.
3683 pc Full range.
3685 chroma_sample_position
3687 Set the chroma sample location in the stream (see AV1 section
3688 6.4.2). This can only be set for 4:2:0 streams.
3689 vertical
3691 Left position (matching the default in MPEG-2 and H.264).
3692 colocated
3694 Top-left position.
3695 tick_rate
3697 Set the tick rate (time_scale / num_units_in_display_tick) in the
3698 timing info in the sequence header.
3699 num_ticks_per_picture
3701 Set the number of ticks in each picture, to indicate that the
3702 stream has a fixed framerate. Ignored if tick_rate is not also
3703 set.
3704 delete_padding
3706 Deletes Padding OBUs.
3707 chomp
3709 Remove zero padding at the end of a packet.
3710 dca_core
3712 Extract the core from a DCA/DTS stream, dropping extensions such as
3713 DTS-HD.
3714 dump_extra
3716 Add extradata to the beginning of the filtered packets except when said
3717 packets already exactly begin with the extradata that is intended to be
3718 added.
3719 freq
3721 The additional argument specifies which packets should be filtered.
3722 It accepts the values:
3723 k
3725 keyframe
3726 add extradata to all key packets
3727 e
3729 all add extradata to all packets
3730 If not specified it is assumed k.
3732 For example the following ffmpeg command forces a global header (thus
3734 disabling individual packet headers) in the H.264 packets generated by
3735 the "libx264" encoder, but corrects them by adding the header stored in
3736 extradata to the key packets:
3737 ffmpeg -i INPUT -map 0 -flags:v +global_header -c:v libx264 -bsf:v dump_extra out.ts
3739 dv_error_marker
3741 Blocks in DV which are marked as damaged are replaced by blocks of the
3742 specified color.
3743 color
3745 The color to replace damaged blocks by
3746 sta A 16 bit mask which specifies which of the 16 possible error status
3748 values are to be replaced by colored blocks. 0xFFFE is the default
3749 which replaces all non 0 error status values.
3750 ok No error, no concealment
3752 err Error, No concealment
3754 res Reserved
3756 notok
3758 Error or concealment
3759 notres
3761 Not reserved
3762 Aa, Ba, Ca, Ab, Bb, Cb, A, B, C, a, b, erri, erru
3764 The specific error status code
3765 see page 44-46 or section 5.5 of
3767 <http://web.archive.org/web/20060927044735/http://www.smpte.org/smpte_store/standards/pdf/s314m.pdf>
3768 eac3_core
3770 Extract the core from a E-AC-3 stream, dropping extra channels.
3771 extract_extradata
3773 Extract the in-band extradata.
3774 Certain codecs allow the long-term headers (e.g. MPEG-2 sequence
3776 headers, or H.264/HEVC (VPS/)SPS/PPS) to be transmitted either "in-
3777 band" (i.e. as a part of the bitstream containing the coded frames) or
3778 "out of band" (e.g. on the container level). This latter form is called
3779 "extradata" in FFmpeg terminology.
3780 This bitstream filter detects the in-band headers and makes them
3782 available as extradata.
3783 remove
3785 When this option is enabled, the long-term headers are removed from
3786 the bitstream after extraction.
3787 filter_units
3789 Remove units with types in or not in a given set from the stream.
3790 pass_types
3792 List of unit types or ranges of unit types to pass through while
3793 removing all others. This is specified as a '|'-separated list of
3794 unit type values or ranges of values with '-'.
3795 remove_types
3797 Identical to pass_types, except the units in the given set removed
3798 and all others passed through.
3799 Extradata is unchanged by this transformation, but note that if the
3801 stream contains inline parameter sets then the output may be unusable
3802 if they are removed.
3803 For example, to remove all non-VCL NAL units from an H.264 stream:
3805 ffmpeg -i INPUT -c:v copy -bsf:v 'filter_units=pass_types=1-5' OUTPUT
3807 To remove all AUDs, SEI and filler from an H.265 stream:
3809 ffmpeg -i INPUT -c:v copy -bsf:v 'filter_units=remove_types=35|38-40' OUTPUT
3811 hapqa_extract
3813 Extract Rgb or Alpha part of an HAPQA file, without recompression, in
3814 order to create an HAPQ or an HAPAlphaOnly file.
3815 texture
3817 Specifies the texture to keep.
3818 color
3820 alpha
3821 Convert HAPQA to HAPQ
3823 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
3825 Convert HAPQA to HAPAlphaOnly
3827 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
3829 h264_metadata
3831 Modify metadata embedded in an H.264 stream.
3832 aud Insert or remove AUD NAL units in all access units of the stream.
3834 pass
3836 insert
3837 remove
3838 Default is pass.
3840 sample_aspect_ratio
3842 Set the sample aspect ratio of the stream in the VUI parameters.
3843 See H.264 table E-1.
3844 overscan_appropriate_flag
3846 Set whether the stream is suitable for display using overscan or
3847 not (see H.264 section E.2.1).
3848 video_format
3850 video_full_range_flag
3851 Set the video format in the stream (see H.264 section E.2.1 and
3852 table E-2).
3853 colour_primaries
3855 transfer_characteristics
3856 matrix_coefficients
3857 Set the colour description in the stream (see H.264 section E.2.1
3858 and tables E-3, E-4 and E-5).
3859 chroma_sample_loc_type
3861 Set the chroma sample location in the stream (see H.264 section
3862 E.2.1 and figure E-1).
3863 tick_rate
3865 Set the tick rate (time_scale / num_units_in_tick) in the VUI
3866 parameters. This is the smallest time unit representable in the
3867 stream, and in many cases represents the field rate of the stream
3868 (double the frame rate).
3869 fixed_frame_rate_flag
3871 Set whether the stream has fixed framerate - typically this
3872 indicates that the framerate is exactly half the tick rate, but the
3873 exact meaning is dependent on interlacing and the picture structure
3874 (see H.264 section E.2.1 and table E-6).
3875 zero_new_constraint_set_flags
3877 Zero constraint_set4_flag and constraint_set5_flag in the SPS.
3878 These bits were reserved in a previous version of the H.264 spec,
3879 and thus some hardware decoders require these to be zero. The
3880 result of zeroing this is still a valid bitstream.
3881 crop_left
3883 crop_right
3884 crop_top
3885 crop_bottom
3886 Set the frame cropping offsets in the SPS. These values will
3887 replace the current ones if the stream is already cropped.
3888 These fields are set in pixels. Note that some sizes may not be
3890 representable if the chroma is subsampled or the stream is
3891 interlaced (see H.264 section 7.4.2.1.1).
3892 sei_user_data
3894 Insert a string as SEI unregistered user data. The argument must
3895 be of the form UUID+string, where the UUID is as hex digits
3896 possibly separated by hyphens, and the string can be anything.
3897 For example, 086f3693-b7b3-4f2c-9653-21492feee5b8+hello will insert
3899 the string ``hello'' associated with the given UUID.
3900 delete_filler
3902 Deletes both filler NAL units and filler SEI messages.
3903 display_orientation
3905 Insert, extract or remove Display orientation SEI messages. See
3906 H.264 section D.1.27 and D.2.27 for syntax and semantics.
3907 pass
3909 insert
3910 remove
3911 extract
3912 Default is pass.
3914 Insert mode works in conjunction with "rotate" and "flip" options.
3916 Any pre-existing Display orientation messages will be removed in
3917 insert or remove mode. Extract mode attaches the display matrix to
3918 the packet as side data.
3919 rotate
3921 Set rotation in display orientation SEI (anticlockwise angle in
3922 degrees). Range is -360 to +360. Default is NaN.
3923 flip
3925 Set flip in display orientation SEI.
3926 horizontal
3928 vertical
3929 Default is unset.
3931 level
3933 Set the level in the SPS. Refer to H.264 section A.3 and tables
3934 A-1 to A-5.
3935 The argument must be the name of a level (for example, 4.2), a
3937 level_idc value (for example, 42), or the special name auto
3938 indicating that the filter should attempt to guess the level from
3939 the input stream properties.
3940 h264_mp4toannexb
3942 Convert an H.264 bitstream from length prefixed mode to start code
3943 prefixed mode (as defined in the Annex B of the ITU-T H.264
3944 specification).
3945 This is required by some streaming formats, typically the MPEG-2
3947 transport stream format (muxer "mpegts").
3948 For example to remux an MP4 file containing an H.264 stream to mpegts
3950 format with ffmpeg, you can use the command:
3951 ffmpeg -i INPUT.mp4 -codec copy -bsf:v h264_mp4toannexb OUTPUT.ts
3953 Please note that this filter is auto-inserted for MPEG-TS (muxer
3955 "mpegts") and raw H.264 (muxer "h264") output formats.
3956 h264_redundant_pps
3958 This applies a specific fixup to some Blu-ray streams which contain
3959 redundant PPSs modifying irrelevant parameters of the stream which
3960 confuse other transformations which require correct extradata.
3961 A new single global PPS is created, and all of the redundant PPSs
3963 within the stream are removed.
3964 hevc_metadata
3966 Modify metadata embedded in an HEVC stream.
3967 aud Insert or remove AUD NAL units in all access units of the stream.
3969 insert
3971 remove
3972 sample_aspect_ratio
3973 Set the sample aspect ratio in the stream in the VUI parameters.
3974 video_format
3976 video_full_range_flag
3977 Set the video format in the stream (see H.265 section E.3.1 and
3978 table E.2).
3979 colour_primaries
3981 transfer_characteristics
3982 matrix_coefficients
3983 Set the colour description in the stream (see H.265 section E.3.1
3984 and tables E.3, E.4 and E.5).
3985 chroma_sample_loc_type
3987 Set the chroma sample location in the stream (see H.265 section
3988 E.3.1 and figure E.1).
3989 tick_rate
3991 Set the tick rate in the VPS and VUI parameters (time_scale /
3992 num_units_in_tick). Combined with num_ticks_poc_diff_one, this can
3993 set a constant framerate in the stream. Note that it is likely to
3994 be overridden by container parameters when the stream is in a
3995 container.
3996 num_ticks_poc_diff_one
3998 Set poc_proportional_to_timing_flag in VPS and VUI and use this
3999 value to set num_ticks_poc_diff_one_minus1 (see H.265 sections
4000 7.4.3.1 and E.3.1). Ignored if tick_rate is not also set.
4001 crop_left
4003 crop_right
4004 crop_top
4005 crop_bottom
4006 Set the conformance window cropping offsets in the SPS. These
4007 values will replace the current ones if the stream is already
4008 cropped.
4009 These fields are set in pixels. Note that some sizes may not be
4011 representable if the chroma is subsampled (H.265 section
4012 7.4.3.2.1).
4013 level
4015 Set the level in the VPS and SPS. See H.265 section A.4 and tables
4016 A.6 and A.7.
4017 The argument must be the name of a level (for example, 5.1), a
4019 general_level_idc value (for example, 153 for level 5.1), or the
4020 special name auto indicating that the filter should attempt to
4021 guess the level from the input stream properties.
4022 hevc_mp4toannexb
4024 Convert an HEVC/H.265 bitstream from length prefixed mode to start code
4025 prefixed mode (as defined in the Annex B of the ITU-T H.265
4026 specification).
4027 This is required by some streaming formats, typically the MPEG-2
4029 transport stream format (muxer "mpegts").
4030 For example to remux an MP4 file containing an HEVC stream to mpegts
4032 format with ffmpeg, you can use the command:
4033 ffmpeg -i INPUT.mp4 -codec copy -bsf:v hevc_mp4toannexb OUTPUT.ts
4035 Please note that this filter is auto-inserted for MPEG-TS (muxer
4037 "mpegts") and raw HEVC/H.265 (muxer "h265" or "hevc") output formats.
4038 imxdump
4040 Modifies the bitstream to fit in MOV and to be usable by the Final Cut
4041 Pro decoder. This filter only applies to the mpeg2video codec, and is
4042 likely not needed for Final Cut Pro 7 and newer with the appropriate
4043 -tag:v.
4044 For example, to remux 30 MB/sec NTSC IMX to MOV:
4046 ffmpeg -i input.mxf -c copy -bsf:v imxdump -tag:v mx3n output.mov
4048 mjpeg2jpeg
4050 Convert MJPEG/AVI1 packets to full JPEG/JFIF packets.
4051 MJPEG is a video codec wherein each video frame is essentially a JPEG
4053 image. The individual frames can be extracted without loss, e.g. by
4054 ffmpeg -i ../some_mjpeg.avi -c:v copy frames_%d.jpg
4056 Unfortunately, these chunks are incomplete JPEG images, because they
4058 lack the DHT segment required for decoding. Quoting from
4059 <http://www.digitalpreservation.gov/formats/fdd/fdd000063.shtml>:
4060 Avery Lee, writing in the rec.video.desktop newsgroup in 2001,
4062 commented that "MJPEG, or at least the MJPEG in AVIs having the MJPG
4063 fourcc, is restricted JPEG with a fixed -- and *omitted* -- Huffman
4064 table. The JPEG must be YCbCr colorspace, it must be 4:2:2, and it must
4065 use basic Huffman encoding, not arithmetic or progressive. . . . You
4066 can indeed extract the MJPEG frames and decode them with a regular JPEG
4067 decoder, but you have to prepend the DHT segment to them, or else the
4068 decoder won't have any idea how to decompress the data. The exact table
4069 necessary is given in the OpenDML spec."
4070 This bitstream filter patches the header of frames extracted from an
4072 MJPEG stream (carrying the AVI1 header ID and lacking a DHT segment) to
4073 produce fully qualified JPEG images.
4074 ffmpeg -i mjpeg-movie.avi -c:v copy -bsf:v mjpeg2jpeg frame_%d.jpg
4076 exiftran -i -9 frame*.jpg
4077 ffmpeg -i frame_%d.jpg -c:v copy rotated.avi
4078 mjpegadump
4080 Add an MJPEG A header to the bitstream, to enable decoding by
4081 Quicktime.
4082 mov2textsub
4084 Extract a representable text file from MOV subtitles, stripping the
4085 metadata header from each subtitle packet.
4086 See also the text2movsub filter.
4088 mp3decomp
4090 Decompress non-standard compressed MP3 audio headers.
4091 mpeg2_metadata
4093 Modify metadata embedded in an MPEG-2 stream.
4094 display_aspect_ratio
4096 Set the display aspect ratio in the stream.
4097 The following fixed values are supported:
4099 4/3
4101 16/9
4102 221/100
4103 Any other value will result in square pixels being signalled
4105 instead (see H.262 section 6.3.3 and table 6-3).
4106 frame_rate
4108 Set the frame rate in the stream. This is constructed from a table
4109 of known values combined with a small multiplier and divisor - if
4110 the supplied value is not exactly representable, the nearest
4111 representable value will be used instead (see H.262 section 6.3.3
4112 and table 6-4).
4113 video_format
4115 Set the video format in the stream (see H.262 section 6.3.6 and
4116 table 6-6).
4117 colour_primaries
4119 transfer_characteristics
4120 matrix_coefficients
4121 Set the colour description in the stream (see H.262 section 6.3.6
4122 and tables 6-7, 6-8 and 6-9).
4123 mpeg4_unpack_bframes
4125 Unpack DivX-style packed B-frames.
4126 DivX-style packed B-frames are not valid MPEG-4 and were only a
4128 workaround for the broken Video for Windows subsystem. They use more
4129 space, can cause minor AV sync issues, require more CPU power to decode
4130 (unless the player has some decoded picture queue to compensate the
4131 2,0,2,0 frame per packet style) and cause trouble if copied into a
4132 standard container like mp4 or mpeg-ps/ts, because MPEG-4 decoders may
4133 not be able to decode them, since they are not valid MPEG-4.
4134 For example to fix an AVI file containing an MPEG-4 stream with DivX-
4136 style packed B-frames using ffmpeg, you can use the command:
4137 ffmpeg -i INPUT.avi -codec copy -bsf:v mpeg4_unpack_bframes OUTPUT.avi
4139 noise
4141 Damages the contents of packets or simply drops them without damaging
4142 the container. Can be used for fuzzing or testing error
4143 resilience/concealment.
4144 Parameters:
4146 amount
4148 Accepts an expression whose evaluation per-packet determines how
4149 often bytes in that packet will be modified. A value below 0 will
4150 result in a variable frequency. Default is 0 which results in no
4151 modification. However, if neither amount nor drop is specified,
4152 amount will be set to -1. See below for accepted variables.
4153 drop
4155 Accepts an expression evaluated per-packet whose value determines
4156 whether that packet is dropped. Evaluation to a positive value
4157 results in the packet being dropped. Evaluation to a negative value
4158 results in a variable chance of it being dropped, roughly inverse
4159 in proportion to the magnitude of the value. Default is 0 which
4160 results in no drops. See below for accepted variables.
4161 dropamount
4163 Accepts a non-negative integer, which assigns a variable chance of
4164 it being dropped, roughly inverse in proportion to the value.
4165 Default is 0 which results in no drops. This option is kept for
4166 backwards compatibility and is equivalent to setting drop to a
4167 negative value with the same magnitude i.e. "dropamount=4" is the
4168 same as "drop=-4". Ignored if drop is also specified.
4169 Both "amount" and "drop" accept expressions containing the following
4171 variables:
4172 n The index of the packet, starting from zero.
4174 tb The timebase for packet timestamps.
4176 pts Packet presentation timestamp.
4178 dts Packet decoding timestamp.
4180 nopts
4182 Constant representing AV_NOPTS_VALUE.
4183 startpts
4185 First non-AV_NOPTS_VALUE PTS seen in the stream.
4186 startdts
4188 First non-AV_NOPTS_VALUE DTS seen in the stream.
4189 duration
4191 d Packet duration, in timebase units.
4192 pos Packet position in input; may be -1 when unknown or not set.
4194 size
4196 Packet size, in bytes.
4197 key Whether packet is marked as a keyframe.
4199 state
4201 A pseudo random integer, primarily derived from the content of
4202 packet payload.
4203 Examples
4205 Apply modification to every byte but don't drop any packets.
4207 ffmpeg -i INPUT -c copy -bsf noise=1 output.mkv
4209 Drop every video packet not marked as a keyframe after timestamp 30s
4211 but do not modify any of the remaining packets.
4212 ffmpeg -i INPUT -c copy -bsf:v noise=drop='gt(t\,30)*not(key)' output.mkv
4214 Drop one second of audio every 10 seconds and add some random noise to
4216 the rest.
4217 ffmpeg -i INPUT -c copy -bsf:a noise=amount=-1:drop='between(mod(t\,10)\,9\,10)' output.mkv
4219 null
4221 This bitstream filter passes the packets through unchanged.
4222 pcm_rechunk
4224 Repacketize PCM audio to a fixed number of samples per packet or a
4225 fixed packet rate per second. This is similar to the asetnsamples audio
4226 filter but works on audio packets instead of audio frames.
4227 nb_out_samples, n
4229 Set the number of samples per each output audio packet. The number
4230 is intended as the number of samples per each channel. Default
4231 value is 1024.
4232 pad, p
4234 If set to 1, the filter will pad the last audio packet with
4235 silence, so that it will contain the same number of samples (or
4236 roughly the same number of samples, see frame_rate) as the previous
4237 ones. Default value is 1.
4238 frame_rate, r
4240 This option makes the filter output a fixed number of packets per
4241 second instead of a fixed number of samples per packet. If the
4242 audio sample rate is not divisible by the frame rate then the
4243 number of samples will not be constant but will vary slightly so
4244 that each packet will start as close to the frame boundary as
4245 possible. Using this option has precedence over nb_out_samples.
4246 You can generate the well known 1602-1601-1602-1601-1602 pattern of
4248 48kHz audio for NTSC frame rate using the frame_rate option.
4249 ffmpeg -f lavfi -i sine=r=48000:d=1 -c pcm_s16le -bsf pcm_rechunk=r=30000/1001 -f framecrc -
4251 pgs_frame_merge
4253 Merge a sequence of PGS Subtitle segments ending with an "end of
4254 display set" segment into a single packet.
4255 This is required by some containers that support PGS subtitles (muxer
4257 "matroska").
4258 prores_metadata
4260 Modify color property metadata embedded in prores stream.
4261 color_primaries
4263 Set the color primaries. Available values are:
4264 auto
4266 Keep the same color primaries property (default).
4267 unknown
4269 bt709
4270 bt470bg
4271 BT601 625
4272 smpte170m
4274 BT601 525
4275 bt2020
4277 smpte431
4278 DCI P3
4279 smpte432
4281 P3 D65
4282 transfer_characteristics
4284 Set the color transfer. Available values are:
4285 auto
4287 Keep the same transfer characteristics property (default).
4288 unknown
4290 bt709
4291 BT 601, BT 709, BT 2020
4292 smpte2084
4294 SMPTE ST 2084
4295 arib-std-b67
4297 ARIB STD-B67
4298 matrix_coefficients
4300 Set the matrix coefficient. Available values are:
4301 auto
4303 Keep the same colorspace property (default).
4304 unknown
4306 bt709
4307 smpte170m
4308 BT 601
4309 bt2020nc
4311 Set Rec709 colorspace for each frame of the file
4313 ffmpeg -i INPUT -c copy -bsf:v prores_metadata=color_primaries=bt709:color_trc=bt709:colorspace=bt709 output.mov
4315 Set Hybrid Log-Gamma parameters for each frame of the file
4317 ffmpeg -i INPUT -c copy -bsf:v prores_metadata=color_primaries=bt2020:color_trc=arib-std-b67:colorspace=bt2020nc output.mov
4319 remove_extra
4321 Remove extradata from packets.
4322 It accepts the following parameter:
4324 freq
4326 Set which frame types to remove extradata from.
4327 k Remove extradata from non-keyframes only.
4329 keyframe
4331 Remove extradata from keyframes only.
4332 e, all
4334 Remove extradata from all frames.
4335 setts
4337 Set PTS and DTS in packets.
4338 It accepts the following parameters:
4340 ts
4342 pts
4343 dts Set expressions for PTS, DTS or both.
4344 duration
4346 Set expression for duration.
4347 time_base
4349 Set output time base.
4350 The expressions are evaluated through the eval API and can contain the
4352 following constants:
4353 N The count of the input packet. Starting from 0.
4355 TS The demux timestamp in input in case of "ts" or "dts" option or
4357 presentation timestamp in case of "pts" option.
4358 POS The original position in the file of the packet, or undefined if
4360 undefined for the current packet
4361 DTS The demux timestamp in input.
4363 PTS The presentation timestamp in input.
4365 DURATION
4367 The duration in input.
4368 STARTDTS
4370 The DTS of the first packet.
4371 STARTPTS
4373 The PTS of the first packet.
4374 PREV_INDTS
4376 The previous input DTS.
4377 PREV_INPTS
4379 The previous input PTS.
4380 PREV_INDURATION
4382 The previous input duration.
4383 PREV_OUTDTS
4385 The previous output DTS.
4386 PREV_OUTPTS
4388 The previous output PTS.
4389 PREV_OUTDURATION
4391 The previous output duration.
4392 NEXT_DTS
4394 The next input DTS.
4395 NEXT_PTS
4397 The next input PTS.
4398 NEXT_DURATION
4400 The next input duration.
4401 TB The timebase of stream packet belongs.
4403 TB_OUT
4405 The output timebase.
4406 SR The sample rate of stream packet belongs.
4408 NOPTS
4410 The AV_NOPTS_VALUE constant.
4411 text2movsub
4413 Convert text subtitles to MOV subtitles (as used by the "mov_text"
4414 codec) with metadata headers.
4415 See also the mov2textsub filter.
4417 trace_headers
4419 Log trace output containing all syntax elements in the coded stream
4420 headers (everything above the level of individual coded blocks). This
4421 can be useful for debugging low-level stream issues.
4422 Supports AV1, H.264, H.265, (M)JPEG, MPEG-2 and VP9, but depending on
4424 the build only a subset of these may be available.
4425 truehd_core
4427 Extract the core from a TrueHD stream, dropping ATMOS data.
4428 vp9_metadata
4430 Modify metadata embedded in a VP9 stream.
4431 color_space
4433 Set the color space value in the frame header. Note that any frame
4434 set to RGB will be implicitly set to PC range and that RGB is
4435 incompatible with profiles 0 and 2.
4436 unknown
4438 bt601
4439 bt709
4440 smpte170
4441 smpte240
4442 bt2020
4443 rgb
4444 color_range
4445 Set the color range value in the frame header. Note that any value
4446 imposed by the color space will take precedence over this value.
4447 tv
4449 pc
4450 vp9_superframe
4452 Merge VP9 invisible (alt-ref) frames back into VP9 superframes. This
4453 fixes merging of split/segmented VP9 streams where the alt-ref frame
4454 was split from its visible counterpart.
4455 vp9_superframe_split
4457 Split VP9 superframes into single frames.
4458 vp9_raw_reorder
4460 Given a VP9 stream with correct timestamps but possibly out of order,
4461 insert additional show-existing-frame packets to correct the ordering.
4462
4464 The libavformat library provides some generic global options, which can
4465 be set on all the muxers and demuxers. In addition each muxer or
4466 demuxer may support so-called private options, which are specific for
4467 that component.
4468 Options may be set by specifying -option value in the FFmpeg tools, or
4470 by setting the value explicitly in the "AVFormatContext" options or
4471 using the libavutil/opt.h API for programmatic use.
4472 The list of supported options follows:
4474 avioflags flags (input/output)
4476 Possible values:
4477 direct
4479 Reduce buffering.
4480 probesize integer (input)
4482 Set probing size in bytes, i.e. the size of the data to analyze to
4483 get stream information. A higher value will enable detecting more
4484 information in case it is dispersed into the stream, but will
4485 increase latency. Must be an integer not lesser than 32. It is
4486 5000000 by default.
4487 max_probe_packets integer (input)
4489 Set the maximum number of buffered packets when probing a codec.
4490 Default is 2500 packets.
4491 packetsize integer (output)
4493 Set packet size.
4494 fflags flags
4496 Set format flags. Some are implemented for a limited number of
4497 formats.
4498 Possible values for input files:
4500 discardcorrupt
4502 Discard corrupted packets.
4503 fastseek
4505 Enable fast, but inaccurate seeks for some formats.
4506 genpts
4508 Generate missing PTS if DTS is present.
4509 igndts
4511 Ignore DTS if PTS is set. Inert when nofillin is set.
4512 ignidx
4514 Ignore index.
4515 nobuffer
4517 Reduce the latency introduced by buffering during initial input
4518 streams analysis.
4519 nofillin
4521 Do not fill in missing values in packet fields that can be
4522 exactly calculated.
4523 noparse
4525 Disable AVParsers, this needs "+nofillin" too.
4526 sortdts
4528 Try to interleave output packets by DTS. At present, available
4529 only for AVIs with an index.
4530 Possible values for output files:
4532 autobsf
4534 Automatically apply bitstream filters as required by the output
4535 format. Enabled by default.
4536 bitexact
4538 Only write platform-, build- and time-independent data. This
4539 ensures that file and data checksums are reproducible and match
4540 between platforms. Its primary use is for regression testing.
4541 flush_packets
4543 Write out packets immediately.
4544 shortest
4546 Stop muxing at the end of the shortest stream. It may be
4547 needed to increase max_interleave_delta to avoid flushing the
4548 longer streams before EOF.
4549 seek2any integer (input)
4551 Allow seeking to non-keyframes on demuxer level when supported if
4552 set to 1. Default is 0.
4553 analyzeduration integer (input)
4555 Specify how many microseconds are analyzed to probe the input. A
4556 higher value will enable detecting more accurate information, but
4557 will increase latency. It defaults to 5,000,000 microseconds = 5
4558 seconds.
4559 cryptokey hexadecimal string (input)
4561 Set decryption key.
4562 indexmem integer (input)
4564 Set max memory used for timestamp index (per stream).
4565 rtbufsize integer (input)
4567 Set max memory used for buffering real-time frames.
4568 fdebug flags (input/output)
4570 Print specific debug info.
4571 Possible values:
4573 ts
4575 max_delay integer (input/output)
4576 Set maximum muxing or demuxing delay in microseconds.
4577 fpsprobesize integer (input)
4579 Set number of frames used to probe fps.
4580 audio_preload integer (output)
4582 Set microseconds by which audio packets should be interleaved
4583 earlier.
4584 chunk_duration integer (output)
4586 Set microseconds for each chunk.
4587 chunk_size integer (output)
4589 Set size in bytes for each chunk.
4590 err_detect, f_err_detect flags (input)
4592 Set error detection flags. "f_err_detect" is deprecated and should
4593 be used only via the ffmpeg tool.
4594 Possible values:
4596 crccheck
4598 Verify embedded CRCs.
4599 bitstream
4601 Detect bitstream specification deviations.
4602 buffer
4604 Detect improper bitstream length.
4605 explode
4607 Abort decoding on minor error detection.
4608 careful
4610 Consider things that violate the spec and have not been seen in
4611 the wild as errors.
4612 compliant
4614 Consider all spec non compliancies as errors.
4615 aggressive
4617 Consider things that a sane encoder should not do as an error.
4618 max_interleave_delta integer (output)
4620 Set maximum buffering duration for interleaving. The duration is
4621 expressed in microseconds, and defaults to 10000000 (10 seconds).
4622 To ensure all the streams are interleaved correctly, libavformat
4624 will wait until it has at least one packet for each stream before
4625 actually writing any packets to the output file. When some streams
4626 are "sparse" (i.e. there are large gaps between successive
4627 packets), this can result in excessive buffering.
4628 This field specifies the maximum difference between the timestamps
4630 of the first and the last packet in the muxing queue, above which
4631 libavformat will output a packet regardless of whether it has
4632 queued a packet for all the streams.
4633 If set to 0, libavformat will continue buffering packets until it
4635 has a packet for each stream, regardless of the maximum timestamp
4636 difference between the buffered packets.
4637 use_wallclock_as_timestamps integer (input)
4639 Use wallclock as timestamps if set to 1. Default is 0.
4640 avoid_negative_ts integer (output)
4642 Possible values:
4643 make_non_negative
4645 Shift timestamps to make them non-negative. Also note that
4646 this affects only leading negative timestamps, and not non-
4647 monotonic negative timestamps.
4648 make_zero
4650 Shift timestamps so that the first timestamp is 0.
4651 auto (default)
4653 Enables shifting when required by the target format.
4654 disabled
4656 Disables shifting of timestamp.
4657 When shifting is enabled, all output timestamps are shifted by the
4659 same amount. Audio, video, and subtitles desynching and relative
4660 timestamp differences are preserved compared to how they would have
4661 been without shifting.
4662 skip_initial_bytes integer (input)
4664 Set number of bytes to skip before reading header and frames if set
4665 to 1. Default is 0.
4666 correct_ts_overflow integer (input)
4668 Correct single timestamp overflows if set to 1. Default is 1.
4669 flush_packets integer (output)
4671 Flush the underlying I/O stream after each packet. Default is -1
4672 (auto), which means that the underlying protocol will decide, 1
4673 enables it, and has the effect of reducing the latency, 0 disables
4674 it and may increase IO throughput in some cases.
4675 output_ts_offset offset (output)
4677 Set the output time offset.
4678 offset must be a time duration specification, see the Time duration
4680 section in the ffmpeg-utils(1) manual.
4681 The offset is added by the muxer to the output timestamps.
4683 Specifying a positive offset means that the corresponding streams
4685 are delayed bt the time duration specified in offset. Default value
4686 is 0 (meaning that no offset is applied).
4687 format_whitelist list (input)
4689 "," separated list of allowed demuxers. By default all are allowed.
4690 dump_separator string (input)
4692 Separator used to separate the fields printed on the command line
4693 about the Stream parameters. For example, to separate the fields
4694 with newlines and indentation:
4695 ffprobe -dump_separator "
4697 " -i ~/videos/matrixbench_mpeg2.mpg
4698 max_streams integer (input)
4700 Specifies the maximum number of streams. This can be used to reject
4701 files that would require too many resources due to a large number
4702 of streams.
4703 skip_estimate_duration_from_pts bool (input)
4705 Skip estimation of input duration when calculated using PTS. At
4706 present, applicable for MPEG-PS and MPEG-TS.
4707 strict, f_strict integer (input/output)
4709 Specify how strictly to follow the standards. "f_strict" is
4710 deprecated and should be used only via the ffmpeg tool.
4711 Possible values:
4713 very
4715 strictly conform to an older more strict version of the spec or
4716 reference software
4717 strict
4719 strictly conform to all the things in the spec no matter what
4720 consequences
4721 normal
4723 unofficial
4724 allow unofficial extensions
4725 experimental
4727 allow non standardized experimental things, experimental
4728 (unfinished/work in progress/not well tested) decoders and
4729 encoders. Note: experimental decoders can pose a security
4730 risk, do not use this for decoding untrusted input.
4731 Format stream specifiers
4733 Format stream specifiers allow selection of one or more streams that
4734 match specific properties.
4735 The exact semantics of stream specifiers is defined by the
4737 "avformat_match_stream_specifier()" function declared in the
4738 libavformat/avformat.h header and documented in the Stream specifiers
4739 section in the ffmpeg(1) manual.
4740
4742 Demuxers are configured elements in FFmpeg that can read the multimedia
4743 streams from a particular type of file.
4744 When you configure your FFmpeg build, all the supported demuxers are
4746 enabled by default. You can list all available ones using the configure
4747 option "--list-demuxers".
4748 You can disable all the demuxers using the configure option
4750 "--disable-demuxers", and selectively enable a single demuxer with the
4751 option "--enable-demuxer=DEMUXER", or disable it with the option
4752 "--disable-demuxer=DEMUXER".
4753 The option "-demuxers" of the ff* tools will display the list of
4755 enabled demuxers. Use "-formats" to view a combined list of enabled
4756 demuxers and muxers.
4757 The description of some of the currently available demuxers follows.
4759 aa
4761 Audible Format 2, 3, and 4 demuxer.
4762 This demuxer is used to demux Audible Format 2, 3, and 4 (.aa) files.
4764 aac
4766 Raw Audio Data Transport Stream AAC demuxer.
4767 This demuxer is used to demux an ADTS input containing a single AAC
4769 stream alongwith any ID3v1/2 or APE tags in it.
4770 apng
4772 Animated Portable Network Graphics demuxer.
4773 This demuxer is used to demux APNG files. All headers, but the PNG
4775 signature, up to (but not including) the first fcTL chunk are
4776 transmitted as extradata. Frames are then split as being all the
4777 chunks between two fcTL ones, or between the last fcTL and IEND chunks.
4778 -ignore_loop bool
4780 Ignore the loop variable in the file if set. Default is enabled.
4781 -max_fps int
4783 Maximum framerate in frames per second. Default of 0 imposes no
4784 limit.
4785 -default_fps int
4787 Default framerate in frames per second when none is specified in
4788 the file (0 meaning as fast as possible). Default is 15.
4789 asf
4791 Advanced Systems Format demuxer.
4792 This demuxer is used to demux ASF files and MMS network streams.
4794 -no_resync_search bool
4796 Do not try to resynchronize by looking for a certain optional start
4797 code.
4798 concat
4800 Virtual concatenation script demuxer.
4801 This demuxer reads a list of files and other directives from a text
4803 file and demuxes them one after the other, as if all their packets had
4804 been muxed together.
4805 The timestamps in the files are adjusted so that the first file starts
4807 at 0 and each next file starts where the previous one finishes. Note
4808 that it is done globally and may cause gaps if all streams do not have
4809 exactly the same length.
4810 All files must have the same streams (same codecs, same time base,
4812 etc.).
4813 The duration of each file is used to adjust the timestamps of the next
4815 file: if the duration is incorrect (because it was computed using the
4816 bit-rate or because the file is truncated, for example), it can cause
4817 artifacts. The "duration" directive can be used to override the
4818 duration stored in each file.
4819 Syntax
4821 The script is a text file in extended-ASCII, with one directive per
4823 line. Empty lines, leading spaces and lines starting with '#' are
4824 ignored. The following directive is recognized:
4825 "file path"
4827 Path to a file to read; special characters and spaces must be
4828 escaped with backslash or single quotes.
4829 All subsequent file-related directives apply to that file.
4831 "ffconcat version 1.0"
4833 Identify the script type and version.
4834 To make FFmpeg recognize the format automatically, this directive
4836 must appear exactly as is (no extra space or byte-order-mark) on
4837 the very first line of the script.
4838 "duration dur"
4840 Duration of the file. This information can be specified from the
4841 file; specifying it here may be more efficient or help if the
4842 information from the file is not available or accurate.
4843 If the duration is set for all files, then it is possible to seek
4845 in the whole concatenated video.
4846 "inpoint timestamp"
4848 In point of the file. When the demuxer opens the file it instantly
4849 seeks to the specified timestamp. Seeking is done so that all
4850 streams can be presented successfully at In point.
4851 This directive works best with intra frame codecs, because for non-
4853 intra frame ones you will usually get extra packets before the
4854 actual In point and the decoded content will most likely contain
4855 frames before In point too.
4856 For each file, packets before the file In point will have
4858 timestamps less than the calculated start timestamp of the file
4859 (negative in case of the first file), and the duration of the files
4860 (if not specified by the "duration" directive) will be reduced
4861 based on their specified In point.
4862 Because of potential packets before the specified In point, packet
4864 timestamps may overlap between two concatenated files.
4865 "outpoint timestamp"
4867 Out point of the file. When the demuxer reaches the specified
4868 decoding timestamp in any of the streams, it handles it as an end
4869 of file condition and skips the current and all the remaining
4870 packets from all streams.
4871 Out point is exclusive, which means that the demuxer will not
4873 output packets with a decoding timestamp greater or equal to Out
4874 point.
4875 This directive works best with intra frame codecs and formats where
4877 all streams are tightly interleaved. For non-intra frame codecs you
4878 will usually get additional packets with presentation timestamp
4879 after Out point therefore the decoded content will most likely
4880 contain frames after Out point too. If your streams are not tightly
4881 interleaved you may not get all the packets from all streams before
4882 Out point and you may only will be able to decode the earliest
4883 stream until Out point.
4884 The duration of the files (if not specified by the "duration"
4886 directive) will be reduced based on their specified Out point.
4887 "file_packet_metadata key=value"
4889 Metadata of the packets of the file. The specified metadata will be
4890 set for each file packet. You can specify this directive multiple
4891 times to add multiple metadata entries. This directive is
4892 deprecated, use "file_packet_meta" instead.
4893 "file_packet_meta key value"
4895 Metadata of the packets of the file. The specified metadata will be
4896 set for each file packet. You can specify this directive multiple
4897 times to add multiple metadata entries.
4898 "option key value"
4900 Option to access, open and probe the file. Can be present multiple
4901 times.
4902 "stream"
4904 Introduce a stream in the virtual file. All subsequent stream-
4905 related directives apply to the last introduced stream. Some
4906 streams properties must be set in order to allow identifying the
4907 matching streams in the subfiles. If no streams are defined in the
4908 script, the streams from the first file are copied.
4909 "exact_stream_id id"
4911 Set the id of the stream. If this directive is given, the string
4912 with the corresponding id in the subfiles will be used. This is
4913 especially useful for MPEG-PS (VOB) files, where the order of the
4914 streams is not reliable.
4915 "stream_meta key value"
4917 Metadata for the stream. Can be present multiple times.
4918 "stream_codec value"
4920 Codec for the stream.
4921 "stream_extradata hex_string"
4923 Extradata for the string, encoded in hexadecimal.
4924 "chapter id start end"
4926 Add a chapter. id is an unique identifier, possibly small and
4927 consecutive.
4928 Options
4930 This demuxer accepts the following option:
4932 safe
4934 If set to 1, reject unsafe file paths and directives. A file path
4935 is considered safe if it does not contain a protocol specification
4936 and is relative and all components only contain characters from the
4937 portable character set (letters, digits, period, underscore and
4938 hyphen) and have no period at the beginning of a component.
4939 If set to 0, any file name is accepted.
4941 The default is 1.
4943 auto_convert
4945 If set to 1, try to perform automatic conversions on packet data to
4946 make the streams concatenable. The default is 1.
4947 Currently, the only conversion is adding the h264_mp4toannexb
4949 bitstream filter to H.264 streams in MP4 format. This is necessary
4950 in particular if there are resolution changes.
4951 segment_time_metadata
4953 If set to 1, every packet will contain the lavf.concat.start_time
4954 and the lavf.concat.duration packet metadata values which are the
4955 start_time and the duration of the respective file segments in the
4956 concatenated output expressed in microseconds. The duration
4957 metadata is only set if it is known based on the concat file. The
4958 default is 0.
4959 Examples
4961 • Use absolute filenames and include some comments:
4963 # my first filename
4965 file /mnt/share/file-1.wav
4966 # my second filename including whitespace
4967 file '/mnt/share/file 2.wav'
4968 # my third filename including whitespace plus single quote
4969 file '/mnt/share/file 3'\''.wav'
4970 • Allow for input format auto-probing, use safe filenames and set the
4972 duration of the first file:
4973 ffconcat version 1.0
4975 file file-1.wav
4977 duration 20.0
4978 file subdir/file-2.wav
4980 dash
4982 Dynamic Adaptive Streaming over HTTP demuxer.
4983 This demuxer presents all AVStreams found in the manifest. By setting
4985 the discard flags on AVStreams the caller can decide which streams to
4986 actually receive. Each stream mirrors the "id" and "bandwidth"
4987 properties from the "<Representation>" as metadata keys named "id" and
4988 "variant_bitrate" respectively.
4989 Options
4991 This demuxer accepts the following option:
4993 cenc_decryption_key
4995 16-byte key, in hex, to decrypt files encrypted using ISO Common
4996 Encryption (CENC/AES-128 CTR; ISO/IEC 23001-7).
4997 imf
4999 Interoperable Master Format demuxer.
5000 This demuxer presents audio and video streams found in an IMF
5002 Composition.
5003 flv, live_flv, kux
5005 Adobe Flash Video Format demuxer.
5006 This demuxer is used to demux FLV files and RTMP network streams. In
5008 case of live network streams, if you force format, you may use live_flv
5009 option instead of flv to survive timestamp discontinuities. KUX is a
5010 flv variant used on the Youku platform.
5011 ffmpeg -f flv -i myfile.flv ...
5013 ffmpeg -f live_flv -i rtmp://<any.server>/anything/key ....
5014 -flv_metadata bool
5016 Allocate the streams according to the onMetaData array content.
5017 -flv_ignore_prevtag bool
5019 Ignore the size of previous tag value.
5020 -flv_full_metadata bool
5022 Output all context of the onMetadata.
5023 gif
5025 Animated GIF demuxer.
5026 It accepts the following options:
5028 min_delay
5030 Set the minimum valid delay between frames in hundredths of
5031 seconds. Range is 0 to 6000. Default value is 2.
5032 max_gif_delay
5034 Set the maximum valid delay between frames in hundredth of seconds.
5035 Range is 0 to 65535. Default value is 65535 (nearly eleven
5036 minutes), the maximum value allowed by the specification.
5037 default_delay
5039 Set the default delay between frames in hundredths of seconds.
5040 Range is 0 to 6000. Default value is 10.
5041 ignore_loop
5043 GIF files can contain information to loop a certain number of times
5044 (or infinitely). If ignore_loop is set to 1, then the loop setting
5045 from the input will be ignored and looping will not occur. If set
5046 to 0, then looping will occur and will cycle the number of times
5047 according to the GIF. Default value is 1.
5048 For example, with the overlay filter, place an infinitely looping GIF
5050 over another video:
5051 ffmpeg -i input.mp4 -ignore_loop 0 -i input.gif -filter_complex overlay=shortest=1 out.mkv
5053 Note that in the above example the shortest option for overlay filter
5055 is used to end the output video at the length of the shortest input
5056 file, which in this case is input.mp4 as the GIF in this example loops
5057 infinitely.
5058 hls
5060 HLS demuxer
5061 Apple HTTP Live Streaming demuxer.
5063 This demuxer presents all AVStreams from all variant streams. The id
5065 field is set to the bitrate variant index number. By setting the
5066 discard flags on AVStreams (by pressing 'a' or 'v' in ffplay), the
5067 caller can decide which variant streams to actually receive. The total
5068 bitrate of the variant that the stream belongs to is available in a
5069 metadata key named "variant_bitrate".
5070 It accepts the following options:
5072 live_start_index
5074 segment index to start live streams at (negative values are from
5075 the end).
5076 prefer_x_start
5078 prefer to use #EXT-X-START if it's in playlist instead of
5079 live_start_index.
5080 allowed_extensions
5082 ',' separated list of file extensions that hls is allowed to
5083 access.
5084 max_reload
5086 Maximum number of times a insufficient list is attempted to be
5087 reloaded. Default value is 1000.
5088 m3u8_hold_counters
5090 The maximum number of times to load m3u8 when it refreshes without
5091 new segments. Default value is 1000.
5092 http_persistent
5094 Use persistent HTTP connections. Applicable only for HTTP streams.
5095 Enabled by default.
5096 http_multiple
5098 Use multiple HTTP connections for downloading HTTP segments.
5099 Enabled by default for HTTP/1.1 servers.
5100 http_seekable
5102 Use HTTP partial requests for downloading HTTP segments. 0 =
5103 disable, 1 = enable, -1 = auto, Default is auto.
5104 seg_format_options
5106 Set options for the demuxer of media segments using a list of
5107 key=value pairs separated by ":".
5108 image2
5110 Image file demuxer.
5111 This demuxer reads from a list of image files specified by a pattern.
5113 The syntax and meaning of the pattern is specified by the option
5114 pattern_type.
5115 The pattern may contain a suffix which is used to automatically
5117 determine the format of the images contained in the files.
5118 The size, the pixel format, and the format of each image must be the
5120 same for all the files in the sequence.
5121 This demuxer accepts the following options:
5123 framerate
5125 Set the frame rate for the video stream. It defaults to 25.
5126 loop
5128 If set to 1, loop over the input. Default value is 0.
5129 pattern_type
5131 Select the pattern type used to interpret the provided filename.
5132 pattern_type accepts one of the following values.
5134 none
5136 Disable pattern matching, therefore the video will only contain
5137 the specified image. You should use this option if you do not
5138 want to create sequences from multiple images and your
5139 filenames may contain special pattern characters.
5140 sequence
5142 Select a sequence pattern type, used to specify a sequence of
5143 files indexed by sequential numbers.
5144 A sequence pattern may contain the string "%d" or "%0Nd", which
5146 specifies the position of the characters representing a
5147 sequential number in each filename matched by the pattern. If
5148 the form "%d0Nd" is used, the string representing the number in
5149 each filename is 0-padded and N is the total number of 0-padded
5150 digits representing the number. The literal character '%' can
5151 be specified in the pattern with the string "%%".
5152 If the sequence pattern contains "%d" or "%0Nd", the first
5154 filename of the file list specified by the pattern must contain
5155 a number inclusively contained between start_number and
5156 start_number+start_number_range-1, and all the following
5157 numbers must be sequential.
5158 For example the pattern "img-%03d.bmp" will match a sequence of
5160 filenames of the form img-001.bmp, img-002.bmp, ...,
5161 img-010.bmp, etc.; the pattern "i%%m%%g-%d.jpg" will match a
5162 sequence of filenames of the form i%m%g-1.jpg, i%m%g-2.jpg,
5163 ..., i%m%g-10.jpg, etc.
5164 Note that the pattern must not necessarily contain "%d" or
5166 "%0Nd", for example to convert a single image file img.jpeg you
5167 can employ the command:
5168 ffmpeg -i img.jpeg img.png
5170 glob
5172 Select a glob wildcard pattern type.
5173 The pattern is interpreted like a "glob()" pattern. This is
5175 only selectable if libavformat was compiled with globbing
5176 support.
5177 glob_sequence (deprecated, will be removed)
5179 Select a mixed glob wildcard/sequence pattern.
5180 If your version of libavformat was compiled with globbing
5182 support, and the provided pattern contains at least one glob
5183 meta character among "%*?[]{}" that is preceded by an unescaped
5184 "%", the pattern is interpreted like a "glob()" pattern,
5185 otherwise it is interpreted like a sequence pattern.
5186 All glob special characters "%*?[]{}" must be prefixed with
5188 "%". To escape a literal "%" you shall use "%%".
5189 For example the pattern "foo-%*.jpeg" will match all the
5191 filenames prefixed by "foo-" and terminating with ".jpeg", and
5192 "foo-%?%?%?.jpeg" will match all the filenames prefixed with
5193 "foo-", followed by a sequence of three characters, and
5194 terminating with ".jpeg".
5195 This pattern type is deprecated in favor of glob and sequence.
5197 Default value is glob_sequence.
5199 pixel_format
5201 Set the pixel format of the images to read. If not specified the
5202 pixel format is guessed from the first image file in the sequence.
5203 start_number
5205 Set the index of the file matched by the image file pattern to
5206 start to read from. Default value is 0.
5207 start_number_range
5209 Set the index interval range to check when looking for the first
5210 image file in the sequence, starting from start_number. Default
5211 value is 5.
5212 ts_from_file
5214 If set to 1, will set frame timestamp to modification time of image
5215 file. Note that monotonity of timestamps is not provided: images go
5216 in the same order as without this option. Default value is 0. If
5217 set to 2, will set frame timestamp to the modification time of the
5218 image file in nanosecond precision.
5219 video_size
5221 Set the video size of the images to read. If not specified the
5222 video size is guessed from the first image file in the sequence.
5223 export_path_metadata
5225 If set to 1, will add two extra fields to the metadata found in
5226 input, making them also available for other filters (see drawtext
5227 filter for examples). Default value is 0. The extra fields are
5228 described below:
5229 lavf.image2dec.source_path
5231 Corresponds to the full path to the input file being read.
5232 lavf.image2dec.source_basename
5234 Corresponds to the name of the file being read.
5235 Examples
5237 • Use ffmpeg for creating a video from the images in the file
5239 sequence img-001.jpeg, img-002.jpeg, ..., assuming an input frame
5240 rate of 10 frames per second:
5241 ffmpeg -framerate 10 -i 'img-%03d.jpeg' out.mkv
5243 • As above, but start by reading from a file with index 100 in the
5245 sequence:
5246 ffmpeg -framerate 10 -start_number 100 -i 'img-%03d.jpeg' out.mkv
5248 • Read images matching the "*.png" glob pattern , that is all the
5250 files terminating with the ".png" suffix:
5251 ffmpeg -framerate 10 -pattern_type glob -i "*.png" out.mkv
5253 libgme
5255 The Game Music Emu library is a collection of video game music file
5256 emulators.
5257 See <https://bitbucket.org/mpyne/game-music-emu/overview> for more
5259 information.
5260 It accepts the following options:
5262 track_index
5264 Set the index of which track to demux. The demuxer can only export
5265 one track. Track indexes start at 0. Default is to pick the first
5266 track. Number of tracks is exported as tracks metadata entry.
5267 sample_rate
5269 Set the sampling rate of the exported track. Range is 1000 to
5270 999999. Default is 44100.
5271 max_size (bytes)
5273 The demuxer buffers the entire file into memory. Adjust this value
5274 to set the maximum buffer size, which in turn, acts as a ceiling
5275 for the size of files that can be read. Default is 50 MiB.
5276 libmodplug
5278 ModPlug based module demuxer
5279 See <https://github.com/Konstanty/libmodplug>
5281 It will export one 2-channel 16-bit 44.1 kHz audio stream. Optionally,
5283 a "pal8" 16-color video stream can be exported with or without printed
5284 metadata.
5285 It accepts the following options:
5287 noise_reduction
5289 Apply a simple low-pass filter. Can be 1 (on) or 0 (off). Default
5290 is 0.
5291 reverb_depth
5293 Set amount of reverb. Range 0-100. Default is 0.
5294 reverb_delay
5296 Set delay in ms, clamped to 40-250 ms. Default is 0.
5297 bass_amount
5299 Apply bass expansion a.k.a. XBass or megabass. Range is 0 (quiet)
5300 to 100 (loud). Default is 0.
5301 bass_range
5303 Set cutoff i.e. upper-bound for bass frequencies. Range is 10-100
5304 Hz. Default is 0.
5305 surround_depth
5307 Apply a Dolby Pro-Logic surround effect. Range is 0 (quiet) to 100
5308 (heavy). Default is 0.
5309 surround_delay
5311 Set surround delay in ms, clamped to 5-40 ms. Default is 0.
5312 max_size
5314 The demuxer buffers the entire file into memory. Adjust this value
5315 to set the maximum buffer size, which in turn, acts as a ceiling
5316 for the size of files that can be read. Range is 0 to 100 MiB. 0
5317 removes buffer size limit (not recommended). Default is 5 MiB.
5318 video_stream_expr
5320 String which is evaluated using the eval API to assign colors to
5321 the generated video stream. Variables which can be used are "x",
5322 "y", "w", "h", "t", "speed", "tempo", "order", "pattern" and "row".
5323 video_stream
5325 Generate video stream. Can be 1 (on) or 0 (off). Default is 0.
5326 video_stream_w
5328 Set video frame width in 'chars' where one char indicates 8 pixels.
5329 Range is 20-512. Default is 30.
5330 video_stream_h
5332 Set video frame height in 'chars' where one char indicates 8
5333 pixels. Range is 20-512. Default is 30.
5334 video_stream_ptxt
5336 Print metadata on video stream. Includes "speed", "tempo", "order",
5337 "pattern", "row" and "ts" (time in ms). Can be 1 (on) or 0 (off).
5338 Default is 1.
5339 libopenmpt
5341 libopenmpt based module demuxer
5342 See <https://lib.openmpt.org/libopenmpt/> for more information.
5344 Some files have multiple subsongs (tracks) this can be set with the
5346 subsong option.
5347 It accepts the following options:
5349 subsong
5351 Set the subsong index. This can be either 'all', 'auto', or the
5352 index of the subsong. Subsong indexes start at 0. The default is
5353 'auto'.
5354 The default value is to let libopenmpt choose.
5356 layout
5358 Set the channel layout. Valid values are 1, 2, and 4 channel
5359 layouts. The default value is STEREO.
5360 sample_rate
5362 Set the sample rate for libopenmpt to output. Range is from 1000
5363 to INT_MAX. The value default is 48000.
5364 mov/mp4/3gp
5366 Demuxer for Quicktime File Format & ISO/IEC Base Media File Format
5367 (ISO/IEC 14496-12 or MPEG-4 Part 12, ISO/IEC 15444-12 or JPEG 2000 Part
5368 12).
5369 Registered extensions: mov, mp4, m4a, 3gp, 3g2, mj2, psp, m4b, ism,
5371 ismv, isma, f4v
5372 Options
5374 This demuxer accepts the following options:
5376 enable_drefs
5378 Enable loading of external tracks, disabled by default. Enabling
5379 this can theoretically leak information in some use cases.
5380 use_absolute_path
5382 Allows loading of external tracks via absolute paths, disabled by
5383 default. Enabling this poses a security risk. It should only be
5384 enabled if the source is known to be non-malicious.
5385 seek_streams_individually
5387 When seeking, identify the closest point in each stream
5388 individually and demux packets in that stream from identified
5389 point. This can lead to a different sequence of packets compared to
5390 demuxing linearly from the beginning. Default is true.
5391 ignore_editlist
5393 Ignore any edit list atoms. The demuxer, by default, modifies the
5394 stream index to reflect the timeline described by the edit list.
5395 Default is false.
5396 advanced_editlist
5398 Modify the stream index to reflect the timeline described by the
5399 edit list. "ignore_editlist" must be set to false for this option
5400 to be effective. If both "ignore_editlist" and this option are set
5401 to false, then only the start of the stream index is modified to
5402 reflect initial dwell time or starting timestamp described by the
5403 edit list. Default is true.
5404 ignore_chapters
5406 Don't parse chapters. This includes GoPro 'HiLight' tags/moments.
5407 Note that chapters are only parsed when input is seekable. Default
5408 is false.
5409 use_mfra_for
5411 For seekable fragmented input, set fragment's starting timestamp
5412 from media fragment random access box, if present.
5413 Following options are available:
5415 auto
5417 Auto-detect whether to set mfra timestamps as PTS or DTS
5418 (default)
5419 dts Set mfra timestamps as DTS
5421 pts Set mfra timestamps as PTS
5423 0 Don't use mfra box to set timestamps
5425 use_tfdt
5427 For fragmented input, set fragment's starting timestamp to
5428 "baseMediaDecodeTime" from the "tfdt" box. Default is enabled,
5429 which will prefer to use the "tfdt" box to set DTS. Disable to use
5430 the "earliest_presentation_time" from the "sidx" box. In either
5431 case, the timestamp from the "mfra" box will be used if it's
5432 available and "use_mfra_for" is set to pts or dts.
5433 export_all
5435 Export unrecognized boxes within the udta box as metadata entries.
5436 The first four characters of the box type are set as the key.
5437 Default is false.
5438 export_xmp
5440 Export entire contents of XMP_ box and uuid box as a string with
5441 key "xmp". Note that if "export_all" is set and this option isn't,
5442 the contents of XMP_ box are still exported but with key "XMP_".
5443 Default is false.
5444 activation_bytes
5446 4-byte key required to decrypt Audible AAX and AAX+ files. See
5447 Audible AAX subsection below.
5448 audible_fixed_key
5450 Fixed key used for handling Audible AAX/AAX+ files. It has been
5451 pre-set so should not be necessary to specify.
5452 decryption_key
5454 16-byte key, in hex, to decrypt files encrypted using ISO Common
5455 Encryption (CENC/AES-128 CTR; ISO/IEC 23001-7).
5456 max_stts_delta
5458 Very high sample deltas written in a trak's stts box may
5459 occasionally be intended but usually they are written in error or
5460 used to store a negative value for dts correction when treated as
5461 signed 32-bit integers. This option lets the user set an upper
5462 limit, beyond which the delta is clamped to 1. Values greater than
5463 the limit if negative when cast to int32 are used to adjust onward
5464 dts.
5465 Unit is the track time scale. Range is 0 to UINT_MAX. Default is
5467 "UINT_MAX - 48000*10" which allows upto a 10 second dts correction
5468 for 48 kHz audio streams while accommodating 99.9% of "uint32"
5469 range.
5470 Audible AAX
5472 Audible AAX files are encrypted M4B files, and they can be decrypted by
5474 specifying a 4 byte activation secret.
5475 ffmpeg -activation_bytes 1CEB00DA -i test.aax -vn -c:a copy output.mp4
5477 mpegts
5479 MPEG-2 transport stream demuxer.
5480 This demuxer accepts the following options:
5482 resync_size
5484 Set size limit for looking up a new synchronization. Default value
5485 is 65536.
5486 skip_unknown_pmt
5488 Skip PMTs for programs not defined in the PAT. Default value is 0.
5489 fix_teletext_pts
5491 Override teletext packet PTS and DTS values with the timestamps
5492 calculated from the PCR of the first program which the teletext
5493 stream is part of and is not discarded. Default value is 1, set
5494 this option to 0 if you want your teletext packet PTS and DTS
5495 values untouched.
5496 ts_packetsize
5498 Output option carrying the raw packet size in bytes. Show the
5499 detected raw packet size, cannot be set by the user.
5500 scan_all_pmts
5502 Scan and combine all PMTs. The value is an integer with value from
5503 -1 to 1 (-1 means automatic setting, 1 means enabled, 0 means
5504 disabled). Default value is -1.
5505 merge_pmt_versions
5507 Re-use existing streams when a PMT's version is updated and
5508 elementary streams move to different PIDs. Default value is 0.
5509 max_packet_size
5511 Set maximum size, in bytes, of packet emitted by the demuxer.
5512 Payloads above this size are split across multiple packets. Range
5513 is 1 to INT_MAX/2. Default is 204800 bytes.
5514 mpjpeg
5516 MJPEG encapsulated in multi-part MIME demuxer.
5517 This demuxer allows reading of MJPEG, where each frame is represented
5519 as a part of multipart/x-mixed-replace stream.
5520 strict_mime_boundary
5522 Default implementation applies a relaxed standard to multi-part
5523 MIME boundary detection, to prevent regression with numerous
5524 existing endpoints not generating a proper MIME MJPEG stream.
5525 Turning this option on by setting it to 1 will result in a stricter
5526 check of the boundary value.
5527 rawvideo
5529 Raw video demuxer.
5530 This demuxer allows one to read raw video data. Since there is no
5532 header specifying the assumed video parameters, the user must specify
5533 them in order to be able to decode the data correctly.
5534 This demuxer accepts the following options:
5536 framerate
5538 Set input video frame rate. Default value is 25.
5539 pixel_format
5541 Set the input video pixel format. Default value is "yuv420p".
5542 video_size
5544 Set the input video size. This value must be specified explicitly.
5545 For example to read a rawvideo file input.raw with ffplay, assuming a
5547 pixel format of "rgb24", a video size of "320x240", and a frame rate of
5548 10 images per second, use the command:
5549 ffplay -f rawvideo -pixel_format rgb24 -video_size 320x240 -framerate 10 input.raw
5551 sbg
5553 SBaGen script demuxer.
5554 This demuxer reads the script language used by SBaGen
5556 <http://uazu.net/sbagen/> to generate binaural beats sessions. A SBG
5557 script looks like that:
5558 -SE
5560 a: 300-2.5/3 440+4.5/0
5561 b: 300-2.5/0 440+4.5/3
5562 off: -
5563 NOW == a
5564 +0:07:00 == b
5565 +0:14:00 == a
5566 +0:21:00 == b
5567 +0:30:00 off
5568 A SBG script can mix absolute and relative timestamps. If the script
5570 uses either only absolute timestamps (including the script start time)
5571 or only relative ones, then its layout is fixed, and the conversion is
5572 straightforward. On the other hand, if the script mixes both kind of
5573 timestamps, then the NOW reference for relative timestamps will be
5574 taken from the current time of day at the time the script is read, and
5575 the script layout will be frozen according to that reference. That
5576 means that if the script is directly played, the actual times will
5577 match the absolute timestamps up to the sound controller's clock
5578 accuracy, but if the user somehow pauses the playback or seeks, all
5579 times will be shifted accordingly.
5580 tedcaptions
5582 JSON captions used for <http://www.ted.com/>.
5583 TED does not provide links to the captions, but they can be guessed
5585 from the page. The file tools/bookmarklets.html from the FFmpeg source
5586 tree contains a bookmarklet to expose them.
5587 This demuxer accepts the following option:
5589 start_time
5591 Set the start time of the TED talk, in milliseconds. The default is
5592 15000 (15s). It is used to sync the captions with the downloadable
5593 videos, because they include a 15s intro.
5594 Example: convert the captions to a format most players understand:
5596 ffmpeg -i http://www.ted.com/talks/subtitles/id/1/lang/en talk1-en.srt
5598 vapoursynth
5600 Vapoursynth wrapper.
5601 Due to security concerns, Vapoursynth scripts will not be autodetected
5603 so the input format has to be forced. For ff* CLI tools, add "-f
5604 vapoursynth" before the input "-i yourscript.vpy".
5605 This demuxer accepts the following option:
5607 max_script_size
5609 The demuxer buffers the entire script into memory. Adjust this
5610 value to set the maximum buffer size, which in turn, acts as a
5611 ceiling for the size of scripts that can be read. Default is 1
5612 MiB.
5613
5615 FFmpeg is able to dump metadata from media files into a simple
5616 UTF-8-encoded INI-like text file and then load it back using the
5617 metadata muxer/demuxer.
5618 The file format is as follows:
5620 1. A file consists of a header and a number of metadata tags divided
5622 into sections, each on its own line.
5623 2. The header is a ;FFMETADATA string, followed by a version number
5625 (now 1).
5626 3. Metadata tags are of the form key=value
5628 4. Immediately after header follows global metadata
5630 5. After global metadata there may be sections with
5632 per-stream/per-chapter metadata.
5633 6. A section starts with the section name in uppercase (i.e. STREAM or
5635 CHAPTER) in brackets ([, ]) and ends with next section or end of
5636 file.
5637 7. At the beginning of a chapter section there may be an optional
5639 timebase to be used for start/end values. It must be in form
5640 TIMEBASE=num/den, where num and den are integers. If the timebase
5641 is missing then start/end times are assumed to be in nanoseconds.
5642 Next a chapter section must contain chapter start and end times in
5644 form START=num, END=num, where num is a positive integer.
5645 8. Empty lines and lines starting with ; or # are ignored.
5647 9. Metadata keys or values containing special characters (=, ;, #, \
5649 and a newline) must be escaped with a backslash \.
5650 10. Note that whitespace in metadata (e.g. foo = bar) is considered to
5652 be a part of the tag (in the example above key is foo , value is
5653 bar).
5654 A ffmetadata file might look like this:
5656 ;FFMETADATA1
5658 title=bike\\shed
5659 ;this is a comment
5660 artist=FFmpeg troll team
5661 [CHAPTER]
5663 TIMEBASE=1/1000
5664 START=0
5665 #chapter ends at 0:01:00
5666 END=60000
5667 title=chapter \#1
5668 [STREAM]
5669 title=multi\
5670 line
5671 By using the ffmetadata muxer and demuxer it is possible to extract
5673 metadata from an input file to an ffmetadata file, and then transcode
5674 the file into an output file with the edited ffmetadata file.
5675 Extracting an ffmetadata file with ffmpeg goes as follows:
5677 ffmpeg -i INPUT -f ffmetadata FFMETADATAFILE
5679 Reinserting edited metadata information from the FFMETADATAFILE file
5681 can be done as:
5682 ffmpeg -i INPUT -i FFMETADATAFILE -map_metadata 1 -codec copy OUTPUT
5684
5686 The libavformat library provides some generic global options, which can
5687 be set on all the protocols. In addition each protocol may support so-
5688 called private options, which are specific for that component.
5689 Options may be set by specifying -option value in the FFmpeg tools, or
5691 by setting the value explicitly in the "AVFormatContext" options or
5692 using the libavutil/opt.h API for programmatic use.
5693 The list of supported options follows:
5695 protocol_whitelist list (input)
5697 Set a ","-separated list of allowed protocols. "ALL" matches all
5698 protocols. Protocols prefixed by "-" are disabled. All protocols
5699 are allowed by default but protocols used by an another protocol
5700 (nested protocols) are restricted to a per protocol subset.
5701
5703 Protocols are configured elements in FFmpeg that enable access to
5704 resources that require specific protocols.
5705 When you configure your FFmpeg build, all the supported protocols are
5707 enabled by default. You can list all available ones using the configure
5708 option "--list-protocols".
5709 You can disable all the protocols using the configure option
5711 "--disable-protocols", and selectively enable a protocol using the
5712 option "--enable-protocol=PROTOCOL", or you can disable a particular
5713 protocol using the option "--disable-protocol=PROTOCOL".
5714 The option "-protocols" of the ff* tools will display the list of
5716 supported protocols.
5717 All protocols accept the following options:
5719 rw_timeout
5721 Maximum time to wait for (network) read/write operations to
5722 complete, in microseconds.
5723 A description of the currently available protocols follows.
5725 amqp
5727 Advanced Message Queueing Protocol (AMQP) version 0-9-1 is a broker
5728 based publish-subscribe communication protocol.
5729 FFmpeg must be compiled with --enable-librabbitmq to support AMQP. A
5731 separate AMQP broker must also be run. An example open-source AMQP
5732 broker is RabbitMQ.
5733 After starting the broker, an FFmpeg client may stream data to the
5735 broker using the command:
5736 ffmpeg -re -i input -f mpegts amqp://[[user]:[password]@]hostname[:port][/vhost]
5738 Where hostname and port (default is 5672) is the address of the broker.
5740 The client may also set a user/password for authentication. The default
5741 for both fields is "guest". Name of virtual host on broker can be set
5742 with vhost. The default value is "/".
5743 Muliple subscribers may stream from the broker using the command:
5745 ffplay amqp://[[user]:[password]@]hostname[:port][/vhost]
5747 In RabbitMQ all data published to the broker flows through a specific
5749 exchange, and each subscribing client has an assigned queue/buffer.
5750 When a packet arrives at an exchange, it may be copied to a client's
5751 queue depending on the exchange and routing_key fields.
5752 The following options are supported:
5754 exchange
5756 Sets the exchange to use on the broker. RabbitMQ has several
5757 predefined exchanges: "amq.direct" is the default exchange, where
5758 the publisher and subscriber must have a matching routing_key;
5759 "amq.fanout" is the same as a broadcast operation (i.e. the data is
5760 forwarded to all queues on the fanout exchange independent of the
5761 routing_key); and "amq.topic" is similar to "amq.direct", but
5762 allows for more complex pattern matching (refer to the RabbitMQ
5763 documentation).
5764 routing_key
5766 Sets the routing key. The default value is "amqp". The routing key
5767 is used on the "amq.direct" and "amq.topic" exchanges to decide
5768 whether packets are written to the queue of a subscriber.
5769 pkt_size
5771 Maximum size of each packet sent/received to the broker. Default is
5772 131072. Minimum is 4096 and max is any large value (representable
5773 by an int). When receiving packets, this sets an internal buffer
5774 size in FFmpeg. It should be equal to or greater than the size of
5775 the published packets to the broker. Otherwise the received message
5776 may be truncated causing decoding errors.
5777 connection_timeout
5779 The timeout in seconds during the initial connection to the broker.
5780 The default value is rw_timeout, or 5 seconds if rw_timeout is not
5781 set.
5782 delivery_mode mode
5784 Sets the delivery mode of each message sent to broker. The
5785 following values are accepted:
5786 persistent
5788 Delivery mode set to "persistent" (2). This is the default
5789 value. Messages may be written to the broker's disk depending
5790 on its setup.
5791 non-persistent
5793 Delivery mode set to "non-persistent" (1). Messages will stay
5794 in broker's memory unless the broker is under memory pressure.
5795 async
5797 Asynchronous data filling wrapper for input stream.
5798 Fill data in a background thread, to decouple I/O operation from demux
5800 thread.
5801 async:<URL>
5803 async:http://host/resource
5804 async:cache:http://host/resource
5805 bluray
5807 Read BluRay playlist.
5808 The accepted options are:
5810 angle
5812 BluRay angle
5813 chapter
5815 Start chapter (1...N)
5816 playlist
5818 Playlist to read (BDMV/PLAYLIST/?????.mpls)
5819 Examples:
5821 Read longest playlist from BluRay mounted to /mnt/bluray:
5823 bluray:/mnt/bluray
5825 Read angle 2 of playlist 4 from BluRay mounted to /mnt/bluray, start
5827 from chapter 2:
5828 -playlist 4 -angle 2 -chapter 2 bluray:/mnt/bluray
5830 cache
5832 Caching wrapper for input stream.
5833 Cache the input stream to temporary file. It brings seeking capability
5835 to live streams.
5836 The accepted options are:
5838 read_ahead_limit
5840 Amount in bytes that may be read ahead when seeking isn't
5841 supported. Range is -1 to INT_MAX. -1 for unlimited. Default is
5842 65536.
5843 URL Syntax is
5845 cache:<URL>
5847 concat
5849 Physical concatenation protocol.
5850 Read and seek from many resources in sequence as if they were a unique
5852 resource.
5853 A URL accepted by this protocol has the syntax:
5855 concat:<URL1>|<URL2>|...|<URLN>
5857 where URL1, URL2, ..., URLN are the urls of the resource to be
5859 concatenated, each one possibly specifying a distinct protocol.
5860 For example to read a sequence of files split1.mpeg, split2.mpeg,
5862 split3.mpeg with ffplay use the command:
5863 ffplay concat:split1.mpeg\|split2.mpeg\|split3.mpeg
5865 Note that you may need to escape the character "|" which is special for
5867 many shells.
5868 concatf
5870 Physical concatenation protocol using a line break delimited list of
5871 resources.
5872 Read and seek from many resources in sequence as if they were a unique
5874 resource.
5875 A URL accepted by this protocol has the syntax:
5877 concatf:<URL>
5879 where URL is the url containing a line break delimited list of
5881 resources to be concatenated, each one possibly specifying a distinct
5882 protocol. Special characters must be escaped with backslash or single
5883 quotes. See the "Quoting and escaping" section in the ffmpeg-utils(1)
5884 manual.
5885 For example to read a sequence of files split1.mpeg, split2.mpeg,
5887 split3.mpeg listed in separate lines within a file split.txt with
5888 ffplay use the command:
5889 ffplay concatf:split.txt
5891 Where split.txt contains the lines:
5893 split1.mpeg
5895 split2.mpeg
5896 split3.mpeg
5897 crypto
5899 AES-encrypted stream reading protocol.
5900 The accepted options are:
5902 key Set the AES decryption key binary block from given hexadecimal
5904 representation.
5905 iv Set the AES decryption initialization vector binary block from
5907 given hexadecimal representation.
5908 Accepted URL formats:
5910 crypto:<URL>
5912 crypto+<URL>
5913 data
5915 Data in-line in the URI. See
5916 <http://en.wikipedia.org/wiki/Data_URI_scheme>.
5917 For example, to convert a GIF file given inline with ffmpeg:
5919 ffmpeg -i "data:image/gif;base64,R0lGODdhCAAIAMIEAAAAAAAA//8AAP//AP///////////////ywAAAAACAAIAAADF0gEDLojDgdGiJdJqUX02iB4E8Q9jUMkADs=" smiley.png
5921 file
5923 File access protocol.
5924 Read from or write to a file.
5926 A file URL can have the form:
5928 file:<filename>
5930 where filename is the path of the file to read.
5932 An URL that does not have a protocol prefix will be assumed to be a
5934 file URL. Depending on the build, an URL that looks like a Windows path
5935 with the drive letter at the beginning will also be assumed to be a
5936 file URL (usually not the case in builds for unix-like systems).
5937 For example to read from a file input.mpeg with ffmpeg use the command:
5939 ffmpeg -i file:input.mpeg output.mpeg
5941 This protocol accepts the following options:
5943 truncate
5945 Truncate existing files on write, if set to 1. A value of 0
5946 prevents truncating. Default value is 1.
5947 blocksize
5949 Set I/O operation maximum block size, in bytes. Default value is
5950 "INT_MAX", which results in not limiting the requested block size.
5951 Setting this value reasonably low improves user termination request
5952 reaction time, which is valuable for files on slow medium.
5953 follow
5955 If set to 1, the protocol will retry reading at the end of the
5956 file, allowing reading files that still are being written. In order
5957 for this to terminate, you either need to use the rw_timeout
5958 option, or use the interrupt callback (for API users).
5959 seekable
5961 Controls if seekability is advertised on the file. 0 means non-
5962 seekable, -1 means auto (seekable for normal files, non-seekable
5963 for named pipes).
5964 Many demuxers handle seekable and non-seekable resources
5966 differently, overriding this might speed up opening certain files
5967 at the cost of losing some features (e.g. accurate seeking).
5968 ftp
5970 FTP (File Transfer Protocol).
5971 Read from or write to remote resources using FTP protocol.
5973 Following syntax is required.
5975 ftp://[user[:password]@]server[:port]/path/to/remote/resource.mpeg
5977 This protocol accepts the following options.
5979 timeout
5981 Set timeout in microseconds of socket I/O operations used by the
5982 underlying low level operation. By default it is set to -1, which
5983 means that the timeout is not specified.
5984 ftp-user
5986 Set a user to be used for authenticating to the FTP server. This is
5987 overridden by the user in the FTP URL.
5988 ftp-password
5990 Set a password to be used for authenticating to the FTP server.
5991 This is overridden by the password in the FTP URL, or by ftp-
5992 anonymous-password if no user is set.
5993 ftp-anonymous-password
5995 Password used when login as anonymous user. Typically an e-mail
5996 address should be used.
5997 ftp-write-seekable
5999 Control seekability of connection during encoding. If set to 1 the
6000 resource is supposed to be seekable, if set to 0 it is assumed not
6001 to be seekable. Default value is 0.
6002 NOTE: Protocol can be used as output, but it is recommended to not do
6004 it, unless special care is taken (tests, customized server
6005 configuration etc.). Different FTP servers behave in different way
6006 during seek operation. ff* tools may produce incomplete content due to
6007 server limitations.
6008 gopher
6010 Gopher protocol.
6011 gophers
6013 Gophers protocol.
6014 The Gopher protocol with TLS encapsulation.
6016 hls
6018 Read Apple HTTP Live Streaming compliant segmented stream as a uniform
6019 one. The M3U8 playlists describing the segments can be remote HTTP
6020 resources or local files, accessed using the standard file protocol.
6021 The nested protocol is declared by specifying "+proto" after the hls
6022 URI scheme name, where proto is either "file" or "http".
6023 hls+http://host/path/to/remote/resource.m3u8
6025 hls+file://path/to/local/resource.m3u8
6026 Using this protocol is discouraged - the hls demuxer should work just
6028 as well (if not, please report the issues) and is more complete. To
6029 use the hls demuxer instead, simply use the direct URLs to the m3u8
6030 files.
6031 http
6033 HTTP (Hyper Text Transfer Protocol).
6034 This protocol accepts the following options:
6036 seekable
6038 Control seekability of connection. If set to 1 the resource is
6039 supposed to be seekable, if set to 0 it is assumed not to be
6040 seekable, if set to -1 it will try to autodetect if it is seekable.
6041 Default value is -1.
6042 chunked_post
6044 If set to 1 use chunked Transfer-Encoding for posts, default is 1.
6045 content_type
6047 Set a specific content type for the POST messages or for listen
6048 mode.
6049 http_proxy
6051 set HTTP proxy to tunnel through e.g. http://example.com:1234
6052 headers
6054 Set custom HTTP headers, can override built in default headers. The
6055 value must be a string encoding the headers.
6056 multiple_requests
6058 Use persistent connections if set to 1, default is 0.
6059 post_data
6061 Set custom HTTP post data.
6062 referer
6064 Set the Referer header. Include 'Referer: URL' header in HTTP
6065 request.
6066 user_agent
6068 Override the User-Agent header. If not specified the protocol will
6069 use a string describing the libavformat build. ("Lavf/<version>")
6070 reconnect_at_eof
6072 If set then eof is treated like an error and causes reconnection,
6073 this is useful for live / endless streams.
6074 reconnect_streamed
6076 If set then even streamed/non seekable streams will be reconnected
6077 on errors.
6078 reconnect_on_network_error
6080 Reconnect automatically in case of TCP/TLS errors during connect.
6081 reconnect_on_http_error
6083 A comma separated list of HTTP status codes to reconnect on. The
6084 list can include specific status codes (e.g. '503') or the strings
6085 '4xx' / '5xx'.
6086 reconnect_delay_max
6088 Sets the maximum delay in seconds after which to give up
6089 reconnecting
6090 mime_type
6092 Export the MIME type.
6093 http_version
6095 Exports the HTTP response version number. Usually "1.0" or "1.1".
6096 icy If set to 1 request ICY (SHOUTcast) metadata from the server. If
6098 the server supports this, the metadata has to be retrieved by the
6099 application by reading the icy_metadata_headers and
6100 icy_metadata_packet options. The default is 1.
6101 icy_metadata_headers
6103 If the server supports ICY metadata, this contains the ICY-specific
6104 HTTP reply headers, separated by newline characters.
6105 icy_metadata_packet
6107 If the server supports ICY metadata, and icy was set to 1, this
6108 contains the last non-empty metadata packet sent by the server. It
6109 should be polled in regular intervals by applications interested in
6110 mid-stream metadata updates.
6111 cookies
6113 Set the cookies to be sent in future requests. The format of each
6114 cookie is the same as the value of a Set-Cookie HTTP response
6115 field. Multiple cookies can be delimited by a newline character.
6116 offset
6118 Set initial byte offset.
6119 end_offset
6121 Try to limit the request to bytes preceding this offset.
6122 method
6124 When used as a client option it sets the HTTP method for the
6125 request.
6126 When used as a server option it sets the HTTP method that is going
6128 to be expected from the client(s). If the expected and the
6129 received HTTP method do not match the client will be given a Bad
6130 Request response. When unset the HTTP method is not checked for
6131 now. This will be replaced by autodetection in the future.
6132 listen
6134 If set to 1 enables experimental HTTP server. This can be used to
6135 send data when used as an output option, or read data from a client
6136 with HTTP POST when used as an input option. If set to 2 enables
6137 experimental multi-client HTTP server. This is not yet implemented
6138 in ffmpeg.c and thus must not be used as a command line option.
6139 # Server side (sending):
6141 ffmpeg -i somefile.ogg -c copy -listen 1 -f ogg http://<server>:<port>
6142 # Client side (receiving):
6144 ffmpeg -i http://<server>:<port> -c copy somefile.ogg
6145 # Client can also be done with wget:
6147 wget http://<server>:<port> -O somefile.ogg
6148 # Server side (receiving):
6150 ffmpeg -listen 1 -i http://<server>:<port> -c copy somefile.ogg
6151 # Client side (sending):
6153 ffmpeg -i somefile.ogg -chunked_post 0 -c copy -f ogg http://<server>:<port>
6154 # Client can also be done with wget:
6156 wget --post-file=somefile.ogg http://<server>:<port>
6157 send_expect_100
6159 Send an Expect: 100-continue header for POST. If set to 1 it will
6160 send, if set to 0 it won't, if set to -1 it will try to send if it
6161 is applicable. Default value is -1.
6162 auth_type
6164 Set HTTP authentication type. No option for Digest, since this
6165 method requires getting nonce parameters from the server first and
6166 can't be used straight away like Basic.
6167 none
6169 Choose the HTTP authentication type automatically. This is the
6170 default.
6171 basic
6173 Choose the HTTP basic authentication.
6174 Basic authentication sends a Base64-encoded string that
6176 contains a user name and password for the client. Base64 is not
6177 a form of encryption and should be considered the same as
6178 sending the user name and password in clear text (Base64 is a
6179 reversible encoding). If a resource needs to be protected,
6180 strongly consider using an authentication scheme other than
6181 basic authentication. HTTPS/TLS should be used with basic
6182 authentication. Without these additional security
6183 enhancements, basic authentication should not be used to
6184 protect sensitive or valuable information.
6185 HTTP Cookies
6187 Some HTTP requests will be denied unless cookie values are passed in
6189 with the request. The cookies option allows these cookies to be
6190 specified. At the very least, each cookie must specify a value along
6191 with a path and domain. HTTP requests that match both the domain and
6192 path will automatically include the cookie value in the HTTP Cookie
6193 header field. Multiple cookies can be delimited by a newline.
6194 The required syntax to play a stream specifying a cookie is:
6196 ffplay -cookies "nlqptid=nltid=tsn; path=/; domain=somedomain.com;" http://somedomain.com/somestream.m3u8
6198 Icecast
6200 Icecast protocol (stream to Icecast servers)
6201 This protocol accepts the following options:
6203 ice_genre
6205 Set the stream genre.
6206 ice_name
6208 Set the stream name.
6209 ice_description
6211 Set the stream description.
6212 ice_url
6214 Set the stream website URL.
6215 ice_public
6217 Set if the stream should be public. The default is 0 (not public).
6218 user_agent
6220 Override the User-Agent header. If not specified a string of the
6221 form "Lavf/<version>" will be used.
6222 password
6224 Set the Icecast mountpoint password.
6225 content_type
6227 Set the stream content type. This must be set if it is different
6228 from audio/mpeg.
6229 legacy_icecast
6231 This enables support for Icecast versions < 2.4.0, that do not
6232 support the HTTP PUT method but the SOURCE method.
6233 tls Establish a TLS (HTTPS) connection to Icecast.
6235 icecast://[<username>[:<password>]@]<server>:<port>/<mountpoint>
6237 ipfs
6239 InterPlanetary File System (IPFS) protocol support. One can access
6240 files stored on the IPFS network through so-called gateways. These are
6241 http(s) endpoints. This protocol wraps the IPFS native protocols
6242 (ipfs:// and ipns://) to be sent to such a gateway. Users can (and
6243 should) host their own node which means this protocol will use one's
6244 local gateway to access files on the IPFS network.
6245 If a user doesn't have a node of their own then the public gateway
6247 "https://dweb.link" is used by default.
6248 This protocol accepts the following options:
6250 gateway
6252 Defines the gateway to use. When not set, the protocol will first
6253 try locating the local gateway by looking at $IPFS_GATEWAY,
6254 $IPFS_PATH and "$HOME/.ipfs/", in that order. If that fails
6255 "https://dweb.link" will be used.
6256 One can use this protocol in 2 ways. Using IPFS:
6258 ffplay ipfs://QmbGtJg23skhvFmu9mJiePVByhfzu5rwo74MEkVDYAmF5T
6260 Or the IPNS protocol (IPNS is mutable IPFS):
6262 ffplay ipns://QmbGtJg23skhvFmu9mJiePVByhfzu5rwo74MEkVDYAmF5T
6264 mmst
6266 MMS (Microsoft Media Server) protocol over TCP.
6267 mmsh
6269 MMS (Microsoft Media Server) protocol over HTTP.
6270 The required syntax is:
6272 mmsh://<server>[:<port>][/<app>][/<playpath>]
6274 md5
6276 MD5 output protocol.
6277 Computes the MD5 hash of the data to be written, and on close writes
6279 this to the designated output or stdout if none is specified. It can be
6280 used to test muxers without writing an actual file.
6281 Some examples follow.
6283 # Write the MD5 hash of the encoded AVI file to the file output.avi.md5.
6285 ffmpeg -i input.flv -f avi -y md5:output.avi.md5
6286 # Write the MD5 hash of the encoded AVI file to stdout.
6288 ffmpeg -i input.flv -f avi -y md5:
6289 Note that some formats (typically MOV) require the output protocol to
6291 be seekable, so they will fail with the MD5 output protocol.
6292 pipe
6294 UNIX pipe access protocol.
6295 Read and write from UNIX pipes.
6297 The accepted syntax is:
6299 pipe:[<number>]
6301 number is the number corresponding to the file descriptor of the pipe
6303 (e.g. 0 for stdin, 1 for stdout, 2 for stderr). If number is not
6304 specified, by default the stdout file descriptor will be used for
6305 writing, stdin for reading.
6306 For example to read from stdin with ffmpeg:
6308 cat test.wav | ffmpeg -i pipe:0
6310 # ...this is the same as...
6311 cat test.wav | ffmpeg -i pipe:
6312 For writing to stdout with ffmpeg:
6314 ffmpeg -i test.wav -f avi pipe:1 | cat > test.avi
6316 # ...this is the same as...
6317 ffmpeg -i test.wav -f avi pipe: | cat > test.avi
6318 This protocol accepts the following options:
6320 blocksize
6322 Set I/O operation maximum block size, in bytes. Default value is
6323 "INT_MAX", which results in not limiting the requested block size.
6324 Setting this value reasonably low improves user termination request
6325 reaction time, which is valuable if data transmission is slow.
6326 Note that some formats (typically MOV), require the output protocol to
6328 be seekable, so they will fail with the pipe output protocol.
6329 prompeg
6331 Pro-MPEG Code of Practice #3 Release 2 FEC protocol.
6332 The Pro-MPEG CoP#3 FEC is a 2D parity-check forward error correction
6334 mechanism for MPEG-2 Transport Streams sent over RTP.
6335 This protocol must be used in conjunction with the "rtp_mpegts" muxer
6337 and the "rtp" protocol.
6338 The required syntax is:
6340 -f rtp_mpegts -fec prompeg=<option>=<val>... rtp://<hostname>:<port>
6342 The destination UDP ports are "port + 2" for the column FEC stream and
6344 "port + 4" for the row FEC stream.
6345 This protocol accepts the following options:
6347 l=n The number of columns (4-20, LxD <= 100)
6349 d=n The number of rows (4-20, LxD <= 100)
6351 Example usage:
6353 -f rtp_mpegts -fec prompeg=l=8:d=4 rtp://<hostname>:<port>
6355 rist
6357 Reliable Internet Streaming Transport protocol
6358 The accepted options are:
6360 rist_profile
6362 Supported values:
6363 simple
6365 main
6366 This one is default.
6367 advanced
6369 buffer_size
6370 Set internal RIST buffer size in milliseconds for retransmission of
6371 data. Default value is 0 which means the librist default (1 sec).
6372 Maximum value is 30 seconds.
6373 fifo_size
6375 Size of the librist receiver output fifo in number of packets. This
6376 must be a power of 2. Defaults to 8192 (vs the librist default of
6377 1024).
6378 overrun_nonfatal=1|0
6380 Survive in case of librist fifo buffer overrun. Default value is 0.
6381 pkt_size
6383 Set maximum packet size for sending data. 1316 by default.
6384 log_level
6386 Set loglevel for RIST logging messages. You only need to set this
6387 if you explicitly want to enable debug level messages or packet
6388 loss simulation, otherwise the regular loglevel is respected.
6389 secret
6391 Set override of encryption secret, by default is unset.
6392 encryption
6394 Set encryption type, by default is disabled. Acceptable values are
6395 128 and 256.
6396 rtmp
6398 Real-Time Messaging Protocol.
6399 The Real-Time Messaging Protocol (RTMP) is used for streaming
6401 multimedia content across a TCP/IP network.
6402 The required syntax is:
6404 rtmp://[<username>:<password>@]<server>[:<port>][/<app>][/<instance>][/<playpath>]
6406 The accepted parameters are:
6408 username
6410 An optional username (mostly for publishing).
6411 password
6413 An optional password (mostly for publishing).
6414 server
6416 The address of the RTMP server.
6417 port
6419 The number of the TCP port to use (by default is 1935).
6420 app It is the name of the application to access. It usually corresponds
6422 to the path where the application is installed on the RTMP server
6423 (e.g. /ondemand/, /flash/live/, etc.). You can override the value
6424 parsed from the URI through the "rtmp_app" option, too.
6425 playpath
6427 It is the path or name of the resource to play with reference to
6428 the application specified in app, may be prefixed by "mp4:". You
6429 can override the value parsed from the URI through the
6430 "rtmp_playpath" option, too.
6431 listen
6433 Act as a server, listening for an incoming connection.
6434 timeout
6436 Maximum time to wait for the incoming connection. Implies listen.
6437 Additionally, the following parameters can be set via command line
6439 options (or in code via "AVOption"s):
6440 rtmp_app
6442 Name of application to connect on the RTMP server. This option
6443 overrides the parameter specified in the URI.
6444 rtmp_buffer
6446 Set the client buffer time in milliseconds. The default is 3000.
6447 rtmp_conn
6449 Extra arbitrary AMF connection parameters, parsed from a string,
6450 e.g. like "B:1 S:authMe O:1 NN:code:1.23 NS:flag:ok O:0". Each
6451 value is prefixed by a single character denoting the type, B for
6452 Boolean, N for number, S for string, O for object, or Z for null,
6453 followed by a colon. For Booleans the data must be either 0 or 1
6454 for FALSE or TRUE, respectively. Likewise for Objects the data
6455 must be 0 or 1 to end or begin an object, respectively. Data items
6456 in subobjects may be named, by prefixing the type with 'N' and
6457 specifying the name before the value (i.e. "NB:myFlag:1"). This
6458 option may be used multiple times to construct arbitrary AMF
6459 sequences.
6460 rtmp_flashver
6462 Version of the Flash plugin used to run the SWF player. The default
6463 is LNX 9,0,124,2. (When publishing, the default is FMLE/3.0
6464 (compatible; <libavformat version>).)
6465 rtmp_flush_interval
6467 Number of packets flushed in the same request (RTMPT only). The
6468 default is 10.
6469 rtmp_live
6471 Specify that the media is a live stream. No resuming or seeking in
6472 live streams is possible. The default value is "any", which means
6473 the subscriber first tries to play the live stream specified in the
6474 playpath. If a live stream of that name is not found, it plays the
6475 recorded stream. The other possible values are "live" and
6476 "recorded".
6477 rtmp_pageurl
6479 URL of the web page in which the media was embedded. By default no
6480 value will be sent.
6481 rtmp_playpath
6483 Stream identifier to play or to publish. This option overrides the
6484 parameter specified in the URI.
6485 rtmp_subscribe
6487 Name of live stream to subscribe to. By default no value will be
6488 sent. It is only sent if the option is specified or if rtmp_live
6489 is set to live.
6490 rtmp_swfhash
6492 SHA256 hash of the decompressed SWF file (32 bytes).
6493 rtmp_swfsize
6495 Size of the decompressed SWF file, required for SWFVerification.
6496 rtmp_swfurl
6498 URL of the SWF player for the media. By default no value will be
6499 sent.
6500 rtmp_swfverify
6502 URL to player swf file, compute hash/size automatically.
6503 rtmp_tcurl
6505 URL of the target stream. Defaults to proto://host[:port]/app.
6506 tcp_nodelay=1|0
6508 Set TCP_NODELAY to disable Nagle's algorithm. Default value is 0.
6509 Remark: Writing to the socket is currently not optimized to
6511 minimize system calls and reduces the efficiency / effect of
6512 TCP_NODELAY.
6513 For example to read with ffplay a multimedia resource named "sample"
6515 from the application "vod" from an RTMP server "myserver":
6516 ffplay rtmp://myserver/vod/sample
6518 To publish to a password protected server, passing the playpath and app
6520 names separately:
6521 ffmpeg -re -i <input> -f flv -rtmp_playpath some/long/path -rtmp_app long/app/name rtmp://username:password@myserver/
6523 rtmpe
6525 Encrypted Real-Time Messaging Protocol.
6526 The Encrypted Real-Time Messaging Protocol (RTMPE) is used for
6528 streaming multimedia content within standard cryptographic primitives,
6529 consisting of Diffie-Hellman key exchange and HMACSHA256, generating a
6530 pair of RC4 keys.
6531 rtmps
6533 Real-Time Messaging Protocol over a secure SSL connection.
6534 The Real-Time Messaging Protocol (RTMPS) is used for streaming
6536 multimedia content across an encrypted connection.
6537 rtmpt
6539 Real-Time Messaging Protocol tunneled through HTTP.
6540 The Real-Time Messaging Protocol tunneled through HTTP (RTMPT) is used
6542 for streaming multimedia content within HTTP requests to traverse
6543 firewalls.
6544 rtmpte
6546 Encrypted Real-Time Messaging Protocol tunneled through HTTP.
6547 The Encrypted Real-Time Messaging Protocol tunneled through HTTP
6549 (RTMPTE) is used for streaming multimedia content within HTTP requests
6550 to traverse firewalls.
6551 rtmpts
6553 Real-Time Messaging Protocol tunneled through HTTPS.
6554 The Real-Time Messaging Protocol tunneled through HTTPS (RTMPTS) is
6556 used for streaming multimedia content within HTTPS requests to traverse
6557 firewalls.
6558 libsmbclient
6560 libsmbclient permits one to manipulate CIFS/SMB network resources.
6561 Following syntax is required.
6563 smb://[[domain:]user[:password@]]server[/share[/path[/file]]]
6565 This protocol accepts the following options.
6567 timeout
6569 Set timeout in milliseconds of socket I/O operations used by the
6570 underlying low level operation. By default it is set to -1, which
6571 means that the timeout is not specified.
6572 truncate
6574 Truncate existing files on write, if set to 1. A value of 0
6575 prevents truncating. Default value is 1.
6576 workgroup
6578 Set the workgroup used for making connections. By default workgroup
6579 is not specified.
6580 For more information see: <http://www.samba.org/>.
6582 libssh
6584 Secure File Transfer Protocol via libssh
6585 Read from or write to remote resources using SFTP protocol.
6587 Following syntax is required.
6589 sftp://[user[:password]@]server[:port]/path/to/remote/resource.mpeg
6591 This protocol accepts the following options.
6593 timeout
6595 Set timeout of socket I/O operations used by the underlying low
6596 level operation. By default it is set to -1, which means that the
6597 timeout is not specified.
6598 truncate
6600 Truncate existing files on write, if set to 1. A value of 0
6601 prevents truncating. Default value is 1.
6602 private_key
6604 Specify the path of the file containing private key to use during
6605 authorization. By default libssh searches for keys in the ~/.ssh/
6606 directory.
6607 Example: Play a file stored on remote server.
6609 ffplay sftp://user:password@server_address:22/home/user/resource.mpeg
6611 librtmp rtmp, rtmpe, rtmps, rtmpt, rtmpte
6613 Real-Time Messaging Protocol and its variants supported through
6614 librtmp.
6615 Requires the presence of the librtmp headers and library during
6617 configuration. You need to explicitly configure the build with
6618 "--enable-librtmp". If enabled this will replace the native RTMP
6619 protocol.
6620 This protocol provides most client functions and a few server functions
6622 needed to support RTMP, RTMP tunneled in HTTP (RTMPT), encrypted RTMP
6623 (RTMPE), RTMP over SSL/TLS (RTMPS) and tunneled variants of these
6624 encrypted types (RTMPTE, RTMPTS).
6625 The required syntax is:
6627 <rtmp_proto>://<server>[:<port>][/<app>][/<playpath>] <options>
6629 where rtmp_proto is one of the strings "rtmp", "rtmpt", "rtmpe",
6631 "rtmps", "rtmpte", "rtmpts" corresponding to each RTMP variant, and
6632 server, port, app and playpath have the same meaning as specified for
6633 the RTMP native protocol. options contains a list of space-separated
6634 options of the form key=val.
6635 See the librtmp manual page (man 3 librtmp) for more information.
6637 For example, to stream a file in real-time to an RTMP server using
6639 ffmpeg:
6640 ffmpeg -re -i myfile -f flv rtmp://myserver/live/mystream
6642 To play the same stream using ffplay:
6644 ffplay "rtmp://myserver/live/mystream live=1"
6646 rtp
6648 Real-time Transport Protocol.
6649 The required syntax for an RTP URL is:
6651 rtp://hostname[:port][?option=val...]
6652 port specifies the RTP port to use.
6654 The following URL options are supported:
6656 ttl=n
6658 Set the TTL (Time-To-Live) value (for multicast only).
6659 rtcpport=n
6661 Set the remote RTCP port to n.
6662 localrtpport=n
6664 Set the local RTP port to n.
6665 localrtcpport=n'
6667 Set the local RTCP port to n.
6668 pkt_size=n
6670 Set max packet size (in bytes) to n.
6671 buffer_size=size
6673 Set the maximum UDP socket buffer size in bytes.
6674 connect=0|1
6676 Do a "connect()" on the UDP socket (if set to 1) or not (if set to
6677 0).
6678 sources=ip[,ip]
6680 List allowed source IP addresses.
6681 block=ip[,ip]
6683 List disallowed (blocked) source IP addresses.
6684 write_to_source=0|1
6686 Send packets to the source address of the latest received packet
6687 (if set to 1) or to a default remote address (if set to 0).
6688 localport=n
6690 Set the local RTP port to n.
6691 localaddr=addr
6693 Local IP address of a network interface used for sending packets or
6694 joining multicast groups.
6695 timeout=n
6697 Set timeout (in microseconds) of socket I/O operations to n.
6698 This is a deprecated option. Instead, localrtpport should be used.
6700 Important notes:
6702 1. If rtcpport is not set the RTCP port will be set to the RTP port
6704 value plus 1.
6705 2. If localrtpport (the local RTP port) is not set any available port
6707 will be used for the local RTP and RTCP ports.
6708 3. If localrtcpport (the local RTCP port) is not set it will be set to
6710 the local RTP port value plus 1.
6711 rtsp
6713 Real-Time Streaming Protocol.
6714 RTSP is not technically a protocol handler in libavformat, it is a
6716 demuxer and muxer. The demuxer supports both normal RTSP (with data
6717 transferred over RTP; this is used by e.g. Apple and Microsoft) and
6718 Real-RTSP (with data transferred over RDT).
6719 The muxer can be used to send a stream using RTSP ANNOUNCE to a server
6721 supporting it (currently Darwin Streaming Server and Mischa
6722 Spiegelmock's <https://github.com/revmischa/rtsp-server>).
6723 The required syntax for a RTSP url is:
6725 rtsp://<hostname>[:<port>]/<path>
6727 Options can be set on the ffmpeg/ffplay command line, or set in code
6729 via "AVOption"s or in "avformat_open_input".
6730 The following options are supported.
6732 initial_pause
6734 Do not start playing the stream immediately if set to 1. Default
6735 value is 0.
6736 rtsp_transport
6738 Set RTSP transport protocols.
6739 It accepts the following values:
6741 udp Use UDP as lower transport protocol.
6743 tcp Use TCP (interleaving within the RTSP control channel) as lower
6745 transport protocol.
6746 udp_multicast
6748 Use UDP multicast as lower transport protocol.
6749 http
6751 Use HTTP tunneling as lower transport protocol, which is useful
6752 for passing proxies.
6753 Multiple lower transport protocols may be specified, in that case
6755 they are tried one at a time (if the setup of one fails, the next
6756 one is tried). For the muxer, only the tcp and udp options are
6757 supported.
6758 rtsp_flags
6760 Set RTSP flags.
6761 The following values are accepted:
6763 filter_src
6765 Accept packets only from negotiated peer address and port.
6766 listen
6768 Act as a server, listening for an incoming connection.
6769 prefer_tcp
6771 Try TCP for RTP transport first, if TCP is available as RTSP
6772 RTP transport.
6773 Default value is none.
6775 allowed_media_types
6777 Set media types to accept from the server.
6778 The following flags are accepted:
6780 video
6782 audio
6783 data
6784 By default it accepts all media types.
6786 min_port
6788 Set minimum local UDP port. Default value is 5000.
6789 max_port
6791 Set maximum local UDP port. Default value is 65000.
6792 listen_timeout
6794 Set maximum timeout (in seconds) to establish an initial
6795 connection. Setting listen_timeout > 0 sets rtsp_flags to listen.
6796 Default is -1 which means an infinite timeout when listen mode is
6797 set.
6798 reorder_queue_size
6800 Set number of packets to buffer for handling of reordered packets.
6801 timeout
6803 Set socket TCP I/O timeout in microseconds.
6804 user_agent
6806 Override User-Agent header. If not specified, it defaults to the
6807 libavformat identifier string.
6808 When receiving data over UDP, the demuxer tries to reorder received
6810 packets (since they may arrive out of order, or packets may get lost
6811 totally). This can be disabled by setting the maximum demuxing delay to
6812 zero (via the "max_delay" field of AVFormatContext).
6813 When watching multi-bitrate Real-RTSP streams with ffplay, the streams
6815 to display can be chosen with "-vst" n and "-ast" n for video and audio
6816 respectively, and can be switched on the fly by pressing "v" and "a".
6817 Examples
6819 The following examples all make use of the ffplay and ffmpeg tools.
6821 • Watch a stream over UDP, with a max reordering delay of 0.5
6823 seconds:
6824 ffplay -max_delay 500000 -rtsp_transport udp rtsp://server/video.mp4
6826 • Watch a stream tunneled over HTTP:
6828 ffplay -rtsp_transport http rtsp://server/video.mp4
6830 • Send a stream in realtime to a RTSP server, for others to watch:
6832 ffmpeg -re -i <input> -f rtsp -muxdelay 0.1 rtsp://server/live.sdp
6834 • Receive a stream in realtime:
6836 ffmpeg -rtsp_flags listen -i rtsp://ownaddress/live.sdp <output>
6838 sap
6840 Session Announcement Protocol (RFC 2974). This is not technically a
6841 protocol handler in libavformat, it is a muxer and demuxer. It is used
6842 for signalling of RTP streams, by announcing the SDP for the streams
6843 regularly on a separate port.
6844 Muxer
6846 The syntax for a SAP url given to the muxer is:
6848 sap://<destination>[:<port>][?<options>]
6850 The RTP packets are sent to destination on port port, or to port 5004
6852 if no port is specified. options is a "&"-separated list. The
6853 following options are supported:
6854 announce_addr=address
6856 Specify the destination IP address for sending the announcements
6857 to. If omitted, the announcements are sent to the commonly used
6858 SAP announcement multicast address 224.2.127.254 (sap.mcast.net),
6859 or ff0e::2:7ffe if destination is an IPv6 address.
6860 announce_port=port
6862 Specify the port to send the announcements on, defaults to 9875 if
6863 not specified.
6864 ttl=ttl
6866 Specify the time to live value for the announcements and RTP
6867 packets, defaults to 255.
6868 same_port=0|1
6870 If set to 1, send all RTP streams on the same port pair. If zero
6871 (the default), all streams are sent on unique ports, with each
6872 stream on a port 2 numbers higher than the previous. VLC/Live555
6873 requires this to be set to 1, to be able to receive the stream.
6874 The RTP stack in libavformat for receiving requires all streams to
6875 be sent on unique ports.
6876 Example command lines follow.
6878 To broadcast a stream on the local subnet, for watching in VLC:
6880 ffmpeg -re -i <input> -f sap sap://224.0.0.255?same_port=1
6882 Similarly, for watching in ffplay:
6884 ffmpeg -re -i <input> -f sap sap://224.0.0.255
6886 And for watching in ffplay, over IPv6:
6888 ffmpeg -re -i <input> -f sap sap://[ff0e::1:2:3:4]
6890 Demuxer
6892 The syntax for a SAP url given to the demuxer is:
6894 sap://[<address>][:<port>]
6896 address is the multicast address to listen for announcements on, if
6898 omitted, the default 224.2.127.254 (sap.mcast.net) is used. port is the
6899 port that is listened on, 9875 if omitted.
6900 The demuxers listens for announcements on the given address and port.
6902 Once an announcement is received, it tries to receive that particular
6903 stream.
6904 Example command lines follow.
6906 To play back the first stream announced on the normal SAP multicast
6908 address:
6909 ffplay sap://
6911 To play back the first stream announced on one the default IPv6 SAP
6913 multicast address:
6914 ffplay sap://[ff0e::2:7ffe]
6916 sctp
6918 Stream Control Transmission Protocol.
6919 The accepted URL syntax is:
6921 sctp://<host>:<port>[?<options>]
6923 The protocol accepts the following options:
6925 listen
6927 If set to any value, listen for an incoming connection. Outgoing
6928 connection is done by default.
6929 max_streams
6931 Set the maximum number of streams. By default no limit is set.
6932 srt
6934 Haivision Secure Reliable Transport Protocol via libsrt.
6935 The supported syntax for a SRT URL is:
6937 srt://<hostname>:<port>[?<options>]
6939 options contains a list of &-separated options of the form key=val.
6941 or
6943 <options> srt://<hostname>:<port>
6945 options contains a list of '-key val' options.
6947 This protocol accepts the following options.
6949 connect_timeout=milliseconds
6951 Connection timeout; SRT cannot connect for RTT > 1500 msec (2
6952 handshake exchanges) with the default connect timeout of 3 seconds.
6953 This option applies to the caller and rendezvous connection modes.
6954 The connect timeout is 10 times the value set for the rendezvous
6955 mode (which can be used as a workaround for this connection problem
6956 with earlier versions).
6957 ffs=bytes
6959 Flight Flag Size (Window Size), in bytes. FFS is actually an
6960 internal parameter and you should set it to not less than
6961 recv_buffer_size and mss. The default value is relatively large,
6962 therefore unless you set a very large receiver buffer, you do not
6963 need to change this option. Default value is 25600.
6964 inputbw=bytes/seconds
6966 Sender nominal input rate, in bytes per seconds. Used along with
6967 oheadbw, when maxbw is set to relative (0), to calculate maximum
6968 sending rate when recovery packets are sent along with the main
6969 media stream: inputbw * (100 + oheadbw) / 100 if inputbw is not set
6970 while maxbw is set to relative (0), the actual input rate is
6971 evaluated inside the library. Default value is 0.
6972 iptos=tos
6974 IP Type of Service. Applies to sender only. Default value is 0xB8.
6975 ipttl=ttl
6977 IP Time To Live. Applies to sender only. Default value is 64.
6978 latency=microseconds
6980 Timestamp-based Packet Delivery Delay. Used to absorb bursts of
6981 missed packet retransmissions. This flag sets both rcvlatency and
6982 peerlatency to the same value. Note that prior to version 1.3.0
6983 this is the only flag to set the latency, however this is
6984 effectively equivalent to setting peerlatency, when side is sender
6985 and rcvlatency when side is receiver, and the bidirectional stream
6986 sending is not supported.
6987 listen_timeout=microseconds
6989 Set socket listen timeout.
6990 maxbw=bytes/seconds
6992 Maximum sending bandwidth, in bytes per seconds. -1 infinite
6993 (CSRTCC limit is 30mbps) 0 relative to input rate (see inputbw) >0
6994 absolute limit value Default value is 0 (relative)
6995 mode=caller|listener|rendezvous
6997 Connection mode. caller opens client connection. listener starts
6998 server to listen for incoming connections. rendezvous use Rendez-
6999 Vous connection mode. Default value is caller.
7000 mss=bytes
7002 Maximum Segment Size, in bytes. Used for buffer allocation and rate
7003 calculation using a packet counter assuming fully filled packets.
7004 The smallest MSS between the peers is used. This is 1500 by default
7005 in the overall internet. This is the maximum size of the UDP
7006 packet and can be only decreased, unless you have some unusual
7007 dedicated network settings. Default value is 1500.
7008 nakreport=1|0
7010 If set to 1, Receiver will send `UMSG_LOSSREPORT` messages
7011 periodically until a lost packet is retransmitted or intentionally
7012 dropped. Default value is 1.
7013 oheadbw=percents
7015 Recovery bandwidth overhead above input rate, in percents. See
7016 inputbw. Default value is 25%.
7017 passphrase=string
7019 HaiCrypt Encryption/Decryption Passphrase string, length from 10 to
7020 79 characters. The passphrase is the shared secret between the
7021 sender and the receiver. It is used to generate the Key Encrypting
7022 Key using PBKDF2 (Password-Based Key Derivation Function). It is
7023 used only if pbkeylen is non-zero. It is used on the receiver only
7024 if the received data is encrypted. The configured passphrase
7025 cannot be recovered (write-only).
7026 enforced_encryption=1|0
7028 If true, both connection parties must have the same password set
7029 (including empty, that is, with no encryption). If the password
7030 doesn't match or only one side is unencrypted, the connection is
7031 rejected. Default is true.
7032 kmrefreshrate=packets
7034 The number of packets to be transmitted after which the encryption
7035 key is switched to a new key. Default is -1. -1 means auto
7036 (0x1000000 in srt library). The range for this option is integers
7037 in the 0 - "INT_MAX".
7038 kmpreannounce=packets
7040 The interval between when a new encryption key is sent and when
7041 switchover occurs. This value also applies to the subsequent
7042 interval between when switchover occurs and when the old encryption
7043 key is decommissioned. Default is -1. -1 means auto (0x1000 in srt
7044 library). The range for this option is integers in the 0 -
7045 "INT_MAX".
7046 snddropdelay=microseconds
7048 The sender's extra delay before dropping packets. This delay is
7049 added to the default drop delay time interval value.
7050 Special value -1: Do not drop packets on the sender at all.
7052 payload_size=bytes
7054 Sets the maximum declared size of a packet transferred during the
7055 single call to the sending function in Live mode. Use 0 if this
7056 value isn't used (which is default in file mode). Default is -1
7057 (automatic), which typically means MPEG-TS; if you are going to use
7058 SRT to send any different kind of payload, such as, for example,
7059 wrapping a live stream in very small frames, then you can use a
7060 bigger maximum frame size, though not greater than 1456 bytes.
7061 pkt_size=bytes
7063 Alias for payload_size.
7064 peerlatency=microseconds
7066 The latency value (as described in rcvlatency) that is set by the
7067 sender side as a minimum value for the receiver.
7068 pbkeylen=bytes
7070 Sender encryption key length, in bytes. Only can be set to 0, 16,
7071 24 and 32. Enable sender encryption if not 0. Not required on
7072 receiver (set to 0), key size obtained from sender in HaiCrypt
7073 handshake. Default value is 0.
7074 rcvlatency=microseconds
7076 The time that should elapse since the moment when the packet was
7077 sent and the moment when it's delivered to the receiver application
7078 in the receiving function. This time should be a buffer time large
7079 enough to cover the time spent for sending, unexpectedly extended
7080 RTT time, and the time needed to retransmit the lost UDP packet.
7081 The effective latency value will be the maximum of this options'
7082 value and the value of peerlatency set by the peer side. Before
7083 version 1.3.0 this option is only available as latency.
7084 recv_buffer_size=bytes
7086 Set UDP receive buffer size, expressed in bytes.
7087 send_buffer_size=bytes
7089 Set UDP send buffer size, expressed in bytes.
7090 timeout=microseconds
7092 Set raise error timeouts for read, write and connect operations.
7093 Note that the SRT library has internal timeouts which can be
7094 controlled separately, the value set here is only a cap on those.
7095 tlpktdrop=1|0
7097 Too-late Packet Drop. When enabled on receiver, it skips missing
7098 packets that have not been delivered in time and delivers the
7099 following packets to the application when their time-to-play has
7100 come. It also sends a fake ACK to the sender. When enabled on
7101 sender and enabled on the receiving peer, the sender drops the
7102 older packets that have no chance of being delivered in time. It
7103 was automatically enabled in the sender if the receiver supports
7104 it.
7105 sndbuf=bytes
7107 Set send buffer size, expressed in bytes.
7108 rcvbuf=bytes
7110 Set receive buffer size, expressed in bytes.
7111 Receive buffer must not be greater than ffs.
7113 lossmaxttl=packets
7115 The value up to which the Reorder Tolerance may grow. When Reorder
7116 Tolerance is > 0, then packet loss report is delayed until that
7117 number of packets come in. Reorder Tolerance increases every time a
7118 "belated" packet has come, but it wasn't due to retransmission
7119 (that is, when UDP packets tend to come out of order), with the
7120 difference between the latest sequence and this packet's sequence,
7121 and not more than the value of this option. By default it's 0,
7122 which means that this mechanism is turned off, and the loss report
7123 is always sent immediately upon experiencing a "gap" in sequences.
7124 minversion
7126 The minimum SRT version that is required from the peer. A
7127 connection to a peer that does not satisfy the minimum version
7128 requirement will be rejected.
7129 The version format in hex is 0xXXYYZZ for x.y.z in human readable
7131 form.
7132 streamid=string
7134 A string limited to 512 characters that can be set on the socket
7135 prior to connecting. This stream ID will be able to be retrieved by
7136 the listener side from the socket that is returned from srt_accept
7137 and was connected by a socket with that set stream ID. SRT does not
7138 enforce any special interpretation of the contents of this string.
7139 This option doesnXt make sense in Rendezvous connection; the result
7140 might be that simply one side will override the value from the
7141 other side and itXs the matter of luck which one would win
7142 srt_streamid=string
7144 Alias for streamid to avoid conflict with ffmpeg command line
7145 option.
7146 smoother=live|file
7148 The type of Smoother used for the transmission for that socket,
7149 which is responsible for the transmission and congestion control.
7150 The Smoother type must be exactly the same on both connecting
7151 parties, otherwise the connection is rejected.
7152 messageapi=1|0
7154 When set, this socket uses the Message API, otherwise it uses
7155 Buffer API. Note that in live mode (see transtype) thereXs only
7156 message API available. In File mode you can chose to use one of two
7157 modes:
7158 Stream API (default, when this option is false). In this mode you
7160 may send as many data as you wish with one sending instruction, or
7161 even use dedicated functions that read directly from a file. The
7162 internal facility will take care of any speed and congestion
7163 control. When receiving, you can also receive as many data as
7164 desired, the data not extracted will be waiting for the next call.
7165 There is no boundary between data portions in the Stream mode.
7166 Message API. In this mode your single sending instruction passes
7168 exactly one piece of data that has boundaries (a message). Contrary
7169 to Live mode, this message may span across multiple UDP packets and
7170 the only size limitation is that it shall fit as a whole in the
7171 sending buffer. The receiver shall use as large buffer as necessary
7172 to receive the message, otherwise the message will not be given up.
7173 When the message is not complete (not all packets received or there
7174 was a packet loss) it will not be given up.
7175 transtype=live|file
7177 Sets the transmission type for the socket, in particular, setting
7178 this option sets multiple other parameters to their default values
7179 as required for a particular transmission type.
7180 live: Set options as for live transmission. In this mode, you
7182 should send by one sending instruction only so many data that fit
7183 in one UDP packet, and limited to the value defined first in
7184 payload_size (1316 is default in this mode). There is no speed
7185 control in this mode, only the bandwidth control, if configured, in
7186 order to not exceed the bandwidth with the overhead transmission
7187 (retransmitted and control packets).
7188 file: Set options as for non-live transmission. See messageapi for
7190 further explanations
7191 linger=seconds
7193 The number of seconds that the socket waits for unsent data when
7194 closing. Default is -1. -1 means auto (off with 0 seconds in live
7195 mode, on with 180 seconds in file mode). The range for this option
7196 is integers in the 0 - "INT_MAX".
7197 tsbpd=1|0
7199 When true, use Timestamp-based Packet Delivery mode. The default
7200 behavior depends on the transmission type: enabled in live mode,
7201 disabled in file mode.
7202 For more information see: <https://github.com/Haivision/srt>.
7204 srtp
7206 Secure Real-time Transport Protocol.
7207 The accepted options are:
7209 srtp_in_suite
7211 srtp_out_suite
7212 Select input and output encoding suites.
7213 Supported values:
7215 AES_CM_128_HMAC_SHA1_80
7217 SRTP_AES128_CM_HMAC_SHA1_80
7218 AES_CM_128_HMAC_SHA1_32
7219 SRTP_AES128_CM_HMAC_SHA1_32
7220 srtp_in_params
7221 srtp_out_params
7222 Set input and output encoding parameters, which are expressed by a
7223 base64-encoded representation of a binary block. The first 16 bytes
7224 of this binary block are used as master key, the following 14 bytes
7225 are used as master salt.
7226 subfile
7228 Virtually extract a segment of a file or another stream. The
7229 underlying stream must be seekable.
7230 Accepted options:
7232 start
7234 Start offset of the extracted segment, in bytes.
7235 end End offset of the extracted segment, in bytes. If set to 0,
7237 extract till end of file.
7238 Examples:
7240 Extract a chapter from a DVD VOB file (start and end sectors obtained
7242 externally and multiplied by 2048):
7243 subfile,,start,153391104,end,268142592,,:/media/dvd/VIDEO_TS/VTS_08_1.VOB
7245 Play an AVI file directly from a TAR archive:
7247 subfile,,start,183241728,end,366490624,,:archive.tar
7249 Play a MPEG-TS file from start offset till end:
7251 subfile,,start,32815239,end,0,,:video.ts
7253 tee
7255 Writes the output to multiple protocols. The individual outputs are
7256 separated by |
7257 tee:file://path/to/local/this.avi|file://path/to/local/that.avi
7259 tcp
7261 Transmission Control Protocol.
7262 The required syntax for a TCP url is:
7264 tcp://<hostname>:<port>[?<options>]
7266 options contains a list of &-separated options of the form key=val.
7268 The list of supported options follows.
7270 listen=2|1|0
7272 Listen for an incoming connection. 0 disables listen, 1 enables
7273 listen in single client mode, 2 enables listen in multi-client
7274 mode. Default value is 0.
7275 timeout=microseconds
7277 Set raise error timeout, expressed in microseconds.
7278 This option is only relevant in read mode: if no data arrived in
7280 more than this time interval, raise error.
7281 listen_timeout=milliseconds
7283 Set listen timeout, expressed in milliseconds.
7284 recv_buffer_size=bytes
7286 Set receive buffer size, expressed bytes.
7287 send_buffer_size=bytes
7289 Set send buffer size, expressed bytes.
7290 tcp_nodelay=1|0
7292 Set TCP_NODELAY to disable Nagle's algorithm. Default value is 0.
7293 Remark: Writing to the socket is currently not optimized to
7295 minimize system calls and reduces the efficiency / effect of
7296 TCP_NODELAY.
7297 tcp_mss=bytes
7299 Set maximum segment size for outgoing TCP packets, expressed in
7300 bytes.
7301 The following example shows how to setup a listening TCP connection
7303 with ffmpeg, which is then accessed with ffplay:
7304 ffmpeg -i <input> -f <format> tcp://<hostname>:<port>?listen
7306 ffplay tcp://<hostname>:<port>
7307 tls
7309 Transport Layer Security (TLS) / Secure Sockets Layer (SSL)
7310 The required syntax for a TLS/SSL url is:
7312 tls://<hostname>:<port>[?<options>]
7314 The following parameters can be set via command line options (or in
7316 code via "AVOption"s):
7317 ca_file, cafile=filename
7319 A file containing certificate authority (CA) root certificates to
7320 treat as trusted. If the linked TLS library contains a default this
7321 might not need to be specified for verification to work, but not
7322 all libraries and setups have defaults built in. The file must be
7323 in OpenSSL PEM format.
7324 tls_verify=1|0
7326 If enabled, try to verify the peer that we are communicating with.
7327 Note, if using OpenSSL, this currently only makes sure that the
7328 peer certificate is signed by one of the root certificates in the
7329 CA database, but it does not validate that the certificate actually
7330 matches the host name we are trying to connect to. (With other
7331 backends, the host name is validated as well.)
7332 This is disabled by default since it requires a CA database to be
7334 provided by the caller in many cases.
7335 cert_file, cert=filename
7337 A file containing a certificate to use in the handshake with the
7338 peer. (When operating as server, in listen mode, this is more
7339 often required by the peer, while client certificates only are
7340 mandated in certain setups.)
7341 key_file, key=filename
7343 A file containing the private key for the certificate.
7344 listen=1|0
7346 If enabled, listen for connections on the provided port, and assume
7347 the server role in the handshake instead of the client role.
7348 http_proxy
7350 The HTTP proxy to tunnel through, e.g. "http://example.com:1234".
7351 The proxy must support the CONNECT method.
7352 Example command lines:
7354 To create a TLS/SSL server that serves an input stream.
7356 ffmpeg -i <input> -f <format> tls://<hostname>:<port>?listen&cert=<server.crt>&key=<server.key>
7358 To play back a stream from the TLS/SSL server using ffplay:
7360 ffplay tls://<hostname>:<port>
7362 udp
7364 User Datagram Protocol.
7365 The required syntax for an UDP URL is:
7367 udp://<hostname>:<port>[?<options>]
7369 options contains a list of &-separated options of the form key=val.
7371 In case threading is enabled on the system, a circular buffer is used
7373 to store the incoming data, which allows one to reduce loss of data due
7374 to UDP socket buffer overruns. The fifo_size and overrun_nonfatal
7375 options are related to this buffer.
7376 The list of supported options follows.
7378 buffer_size=size
7380 Set the UDP maximum socket buffer size in bytes. This is used to
7381 set either the receive or send buffer size, depending on what the
7382 socket is used for. Default is 32 KB for output, 384 KB for input.
7383 See also fifo_size.
7384 bitrate=bitrate
7386 If set to nonzero, the output will have the specified constant
7387 bitrate if the input has enough packets to sustain it.
7388 burst_bits=bits
7390 When using bitrate this specifies the maximum number of bits in
7391 packet bursts.
7392 localport=port
7394 Override the local UDP port to bind with.
7395 localaddr=addr
7397 Local IP address of a network interface used for sending packets or
7398 joining multicast groups.
7399 pkt_size=size
7401 Set the size in bytes of UDP packets.
7402 reuse=1|0
7404 Explicitly allow or disallow reusing UDP sockets.
7405 ttl=ttl
7407 Set the time to live value (for multicast only).
7408 connect=1|0
7410 Initialize the UDP socket with "connect()". In this case, the
7411 destination address can't be changed with ff_udp_set_remote_url
7412 later. If the destination address isn't known at the start, this
7413 option can be specified in ff_udp_set_remote_url, too. This allows
7414 finding out the source address for the packets with getsockname,
7415 and makes writes return with AVERROR(ECONNREFUSED) if "destination
7416 unreachable" is received. For receiving, this gives the benefit of
7417 only receiving packets from the specified peer address/port.
7418 sources=address[,address]
7420 Only receive packets sent from the specified addresses. In case of
7421 multicast, also subscribe to multicast traffic coming from these
7422 addresses only.
7423 block=address[,address]
7425 Ignore packets sent from the specified addresses. In case of
7426 multicast, also exclude the source addresses in the multicast
7427 subscription.
7428 fifo_size=units
7430 Set the UDP receiving circular buffer size, expressed as a number
7431 of packets with size of 188 bytes. If not specified defaults to
7432 7*4096.
7433 overrun_nonfatal=1|0
7435 Survive in case of UDP receiving circular buffer overrun. Default
7436 value is 0.
7437 timeout=microseconds
7439 Set raise error timeout, expressed in microseconds.
7440 This option is only relevant in read mode: if no data arrived in
7442 more than this time interval, raise error.
7443 broadcast=1|0
7445 Explicitly allow or disallow UDP broadcasting.
7446 Note that broadcasting may not work properly on networks having a
7448 broadcast storm protection.
7449 Examples
7451 • Use ffmpeg to stream over UDP to a remote endpoint:
7453 ffmpeg -i <input> -f <format> udp://<hostname>:<port>
7455 • Use ffmpeg to stream in mpegts format over UDP using 188 sized UDP
7457 packets, using a large input buffer:
7458 ffmpeg -i <input> -f mpegts udp://<hostname>:<port>?pkt_size=188&buffer_size=65535
7460 • Use ffmpeg to receive over UDP from a remote endpoint:
7462 ffmpeg -i udp://[<multicast-address>]:<port> ...
7464 unix
7466 Unix local socket
7467 The required syntax for a Unix socket URL is:
7469 unix://<filepath>
7471 The following parameters can be set via command line options (or in
7473 code via "AVOption"s):
7474 timeout
7476 Timeout in ms.
7477 listen
7479 Create the Unix socket in listening mode.
7480 zmq
7482 ZeroMQ asynchronous messaging using the libzmq library.
7483 This library supports unicast streaming to multiple clients without
7485 relying on an external server.
7486 The required syntax for streaming or connecting to a stream is:
7488 zmq:tcp://ip-address:port
7490 Example: Create a localhost stream on port 5555:
7492 ffmpeg -re -i input -f mpegts zmq:tcp://127.0.0.1:5555
7494 Multiple clients may connect to the stream using:
7496 ffplay zmq:tcp://127.0.0.1:5555
7498 Streaming to multiple clients is implemented using a ZeroMQ Pub-Sub
7500 pattern. The server side binds to a port and publishes data. Clients
7501 connect to the server (via IP address/port) and subscribe to the
7502 stream. The order in which the server and client start generally does
7503 not matter.
7504 ffmpeg must be compiled with the --enable-libzmq option to support this
7506 protocol.
7507 Options can be set on the ffmpeg/ffplay command line. The following
7509 options are supported:
7510 pkt_size
7512 Forces the maximum packet size for sending/receiving data. The
7513 default value is 131,072 bytes. On the server side, this sets the
7514 maximum size of sent packets via ZeroMQ. On the clients, it sets an
7515 internal buffer size for receiving packets. Note that pkt_size on
7516 the clients should be equal to or greater than pkt_size on the
7517 server. Otherwise the received message may be truncated causing
7518 decoding errors.
7519
7521 The libavdevice library provides the same interface as libavformat.
7522 Namely, an input device is considered like a demuxer, and an output
7523 device like a muxer, and the interface and generic device options are
7524 the same provided by libavformat (see the ffmpeg-formats manual).
7525 In addition each input or output device may support so-called private
7527 options, which are specific for that component.
7528 Options may be set by specifying -option value in the FFmpeg tools, or
7530 by setting the value explicitly in the device "AVFormatContext" options
7531 or using the libavutil/opt.h API for programmatic use.
7532
7534 Input devices are configured elements in FFmpeg which enable accessing
7535 the data coming from a multimedia device attached to your system.
7536 When you configure your FFmpeg build, all the supported input devices
7538 are enabled by default. You can list all available ones using the
7539 configure option "--list-indevs".
7540 You can disable all the input devices using the configure option
7542 "--disable-indevs", and selectively enable an input device using the
7543 option "--enable-indev=INDEV", or you can disable a particular input
7544 device using the option "--disable-indev=INDEV".
7545 The option "-devices" of the ff* tools will display the list of
7547 supported input devices.
7548 A description of the currently available input devices follows.
7550 alsa
7552 ALSA (Advanced Linux Sound Architecture) input device.
7553 To enable this input device during configuration you need libasound
7555 installed on your system.
7556 This device allows capturing from an ALSA device. The name of the
7558 device to capture has to be an ALSA card identifier.
7559 An ALSA identifier has the syntax:
7561 hw:<CARD>[,<DEV>[,<SUBDEV>]]
7563 where the DEV and SUBDEV components are optional.
7565 The three arguments (in order: CARD,DEV,SUBDEV) specify card number or
7567 identifier, device number and subdevice number (-1 means any).
7568 To see the list of cards currently recognized by your system check the
7570 files /proc/asound/cards and /proc/asound/devices.
7571 For example to capture with ffmpeg from an ALSA device with card id 0,
7573 you may run the command:
7574 ffmpeg -f alsa -i hw:0 alsaout.wav
7576 For more information see:
7578 <http://www.alsa-project.org/alsa-doc/alsa-lib/pcm.html>
7579 Options
7581 sample_rate
7583 Set the sample rate in Hz. Default is 48000.
7584 channels
7586 Set the number of channels. Default is 2.
7587 android_camera
7589 Android camera input device.
7590 This input devices uses the Android Camera2 NDK API which is available
7592 on devices with API level 24+. The availability of android_camera is
7593 autodetected during configuration.
7594 This device allows capturing from all cameras on an Android device,
7596 which are integrated into the Camera2 NDK API.
7597 The available cameras are enumerated internally and can be selected
7599 with the camera_index parameter. The input file string is discarded.
7600 Generally the back facing camera has index 0 while the front facing
7602 camera has index 1.
7603 Options
7605 video_size
7607 Set the video size given as a string such as 640x480 or hd720.
7608 Falls back to the first available configuration reported by Android
7609 if requested video size is not available or by default.
7610 framerate
7612 Set the video framerate. Falls back to the first available
7613 configuration reported by Android if requested framerate is not
7614 available or by default (-1).
7615 camera_index
7617 Set the index of the camera to use. Default is 0.
7618 input_queue_size
7620 Set the maximum number of frames to buffer. Default is 5.
7621 avfoundation
7623 AVFoundation input device.
7624 AVFoundation is the currently recommended framework by Apple for
7626 streamgrabbing on OSX >= 10.7 as well as on iOS.
7627 The input filename has to be given in the following syntax:
7629 -i "[[VIDEO]:[AUDIO]]"
7631 The first entry selects the video input while the latter selects the
7633 audio input. The stream has to be specified by the device name or the
7634 device index as shown by the device list. Alternatively, the video
7635 and/or audio input device can be chosen by index using the
7636 B<-video_device_index E<lt>INDEXE<gt>>
7638 and/or
7640 B<-audio_device_index E<lt>INDEXE<gt>>
7642 , overriding any device name or index given in the input filename.
7644 All available devices can be enumerated by using -list_devices true,
7646 listing all device names and corresponding indices.
7647 There are two device name aliases:
7649 "default"
7651 Select the AVFoundation default device of the corresponding type.
7652 "none"
7654 Do not record the corresponding media type. This is equivalent to
7655 specifying an empty device name or index.
7656 Options
7658 AVFoundation supports the following options:
7660 -list_devices <TRUE|FALSE>
7662 If set to true, a list of all available input devices is given
7663 showing all device names and indices.
7664 -video_device_index <INDEX>
7666 Specify the video device by its index. Overrides anything given in
7667 the input filename.
7668 -audio_device_index <INDEX>
7670 Specify the audio device by its index. Overrides anything given in
7671 the input filename.
7672 -pixel_format <FORMAT>
7674 Request the video device to use a specific pixel format. If the
7675 specified format is not supported, a list of available formats is
7676 given and the first one in this list is used instead. Available
7677 pixel formats are: "monob, rgb555be, rgb555le, rgb565be, rgb565le,
7678 rgb24, bgr24, 0rgb, bgr0, 0bgr, rgb0,
7679 bgr48be, uyvy422, yuva444p, yuva444p16le, yuv444p, yuv422p16,
7680 yuv422p10, yuv444p10,
7681 yuv420p, nv12, yuyv422, gray"
7682 -framerate
7684 Set the grabbing frame rate. Default is "ntsc", corresponding to a
7685 frame rate of "30000/1001".
7686 -video_size
7688 Set the video frame size.
7689 -capture_cursor
7691 Capture the mouse pointer. Default is 0.
7692 -capture_mouse_clicks
7694 Capture the screen mouse clicks. Default is 0.
7695 -capture_raw_data
7697 Capture the raw device data. Default is 0. Using this option may
7698 result in receiving the underlying data delivered to the
7699 AVFoundation framework. E.g. for muxed devices that sends raw DV
7700 data to the framework (like tape-based camcorders), setting this
7701 option to false results in extracted video frames captured in the
7702 designated pixel format only. Setting this option to true results
7703 in receiving the raw DV stream untouched.
7704 Examples
7706 • Print the list of AVFoundation supported devices and exit:
7708 $ ffmpeg -f avfoundation -list_devices true -i ""
7710 • Record video from video device 0 and audio from audio device 0 into
7712 out.avi:
7713 $ ffmpeg -f avfoundation -i "0:0" out.avi
7715 • Record video from video device 2 and audio from audio device 1 into
7717 out.avi:
7718 $ ffmpeg -f avfoundation -video_device_index 2 -i ":1" out.avi
7720 • Record video from the system default video device using the pixel
7722 format bgr0 and do not record any audio into out.avi:
7723 $ ffmpeg -f avfoundation -pixel_format bgr0 -i "default:none" out.avi
7725 • Record raw DV data from a suitable input device and write the
7727 output into out.dv:
7728 $ ffmpeg -f avfoundation -capture_raw_data true -i "zr100:none" out.dv
7730 bktr
7732 BSD video input device.
7733 Options
7735 framerate
7737 Set the frame rate.
7738 video_size
7740 Set the video frame size. Default is "vga".
7741 standard
7743 Available values are:
7744 pal
7746 ntsc
7747 secam
7748 paln
7749 palm
7750 ntscj
7751 decklink
7753 The decklink input device provides capture capabilities for Blackmagic
7754 DeckLink devices.
7755 To enable this input device, you need the Blackmagic DeckLink SDK and
7757 you need to configure with the appropriate "--extra-cflags" and
7758 "--extra-ldflags". On Windows, you need to run the IDL files through
7759 widl.
7760 DeckLink is very picky about the formats it supports. Pixel format of
7762 the input can be set with raw_format. Framerate and video size must be
7763 determined for your device with -list_formats 1. Audio sample rate is
7764 always 48 kHz and the number of channels can be 2, 8 or 16. Note that
7765 all audio channels are bundled in one single audio track.
7766 Options
7768 list_devices
7770 If set to true, print a list of devices and exit. Defaults to
7771 false. This option is deprecated, please use the "-sources" option
7772 of ffmpeg to list the available input devices.
7773 list_formats
7775 If set to true, print a list of supported formats and exit.
7776 Defaults to false.
7777 format_code <FourCC>
7779 This sets the input video format to the format given by the FourCC.
7780 To see the supported values of your device(s) use list_formats.
7781 Note that there is a FourCC 'pal ' that can also be used as pal (3
7782 letters). Default behavior is autodetection of the input video
7783 format, if the hardware supports it.
7784 raw_format
7786 Set the pixel format of the captured video. Available values are:
7787 auto
7789 This is the default which means 8-bit YUV 422 or 8-bit ARGB if
7790 format autodetection is used, 8-bit YUV 422 otherwise.
7791 uyvy422
7793 8-bit YUV 422.
7794 yuv422p10
7796 10-bit YUV 422.
7797 argb
7799 8-bit RGB.
7800 bgra
7802 8-bit RGB.
7803 rgb10
7805 10-bit RGB.
7806 teletext_lines
7808 If set to nonzero, an additional teletext stream will be captured
7809 from the vertical ancillary data. Both SD PAL (576i) and HD (1080i
7810 or 1080p) sources are supported. In case of HD sources, OP47
7811 packets are decoded.
7812 This option is a bitmask of the SD PAL VBI lines captured,
7814 specifically lines 6 to 22, and lines 318 to 335. Line 6 is the LSB
7815 in the mask. Selected lines which do not contain teletext
7816 information will be ignored. You can use the special all constant
7817 to select all possible lines, or standard to skip lines 6, 318 and
7818 319, which are not compatible with all receivers.
7819 For SD sources, ffmpeg needs to be compiled with
7821 "--enable-libzvbi". For HD sources, on older (pre-4K) DeckLink card
7822 models you have to capture in 10 bit mode.
7823 channels
7825 Defines number of audio channels to capture. Must be 2, 8 or 16.
7826 Defaults to 2.
7827 duplex_mode
7829 Sets the decklink device duplex/profile mode. Must be unset, half,
7830 full, one_sub_device_full, one_sub_device_half,
7831 two_sub_device_full, four_sub_device_half Defaults to unset.
7832 Note: DeckLink SDK 11.0 have replaced the duplex property by a
7834 profile property. For the DeckLink Duo 2 and DeckLink Quad 2, a
7835 profile is shared between any 2 sub-devices that utilize the same
7836 connectors. For the DeckLink 8K Pro, a profile is shared between
7837 all 4 sub-devices. So DeckLink 8K Pro support four profiles.
7838 Valid profile modes for DeckLink 8K Pro(with DeckLink SDK >= 11.0):
7840 one_sub_device_full, one_sub_device_half, two_sub_device_full,
7841 four_sub_device_half
7842 Valid profile modes for DeckLink Quad 2 and DeckLink Duo 2: half,
7844 full
7845 timecode_format
7847 Timecode type to include in the frame and video stream metadata.
7848 Must be none, rp188vitc, rp188vitc2, rp188ltc, rp188hfr, rp188any,
7849 vitc, vitc2, or serial. Defaults to none (not included).
7850 In order to properly support 50/60 fps timecodes, the ordering of
7852 the queried timecode types for rp188any is HFR, VITC1, VITC2 and
7853 LTC for >30 fps content. Note that this is slightly different to
7854 the ordering used by the DeckLink API, which is HFR, VITC1, LTC,
7855 VITC2.
7856 video_input
7858 Sets the video input source. Must be unset, sdi, hdmi, optical_sdi,
7859 component, composite or s_video. Defaults to unset.
7860 audio_input
7862 Sets the audio input source. Must be unset, embedded, aes_ebu,
7863 analog, analog_xlr, analog_rca or microphone. Defaults to unset.
7864 video_pts
7866 Sets the video packet timestamp source. Must be video, audio,
7867 reference, wallclock or abs_wallclock. Defaults to video.
7868 audio_pts
7870 Sets the audio packet timestamp source. Must be video, audio,
7871 reference, wallclock or abs_wallclock. Defaults to audio.
7872 draw_bars
7874 If set to true, color bars are drawn in the event of a signal loss.
7875 Defaults to true.
7876 queue_size
7878 Sets maximum input buffer size in bytes. If the buffering reaches
7879 this value, incoming frames will be dropped. Defaults to
7880 1073741824.
7881 audio_depth
7883 Sets the audio sample bit depth. Must be 16 or 32. Defaults to 16.
7884 decklink_copyts
7886 If set to true, timestamps are forwarded as they are without
7887 removing the initial offset. Defaults to false.
7888 timestamp_align
7890 Capture start time alignment in seconds. If set to nonzero, input
7891 frames are dropped till the system timestamp aligns with configured
7892 value. Alignment difference of up to one frame duration is
7893 tolerated. This is useful for maintaining input synchronization
7894 across N different hardware devices deployed for 'N-way'
7895 redundancy. The system time of different hardware devices should be
7896 synchronized with protocols such as NTP or PTP, before using this
7897 option. Note that this method is not foolproof. In some border
7898 cases input synchronization may not happen due to thread scheduling
7899 jitters in the OS. Either sync could go wrong by 1 frame or in a
7900 rarer case timestamp_align seconds. Defaults to 0.
7901 wait_for_tc (bool)
7903 Drop frames till a frame with timecode is received. Sometimes
7904 serial timecode isn't received with the first input frame. If that
7905 happens, the stored stream timecode will be inaccurate. If this
7906 option is set to true, input frames are dropped till a frame with
7907 timecode is received. Option timecode_format must be specified.
7908 Defaults to false.
7909 enable_klv(bool)
7911 If set to true, extracts KLV data from VANC and outputs KLV
7912 packets. KLV VANC packets are joined based on MID and PSC fields
7913 and aggregated into one KLV packet. Defaults to false.
7914 Examples
7916 • List input devices:
7918 ffmpeg -sources decklink
7920 • List supported formats:
7922 ffmpeg -f decklink -list_formats 1 -i 'Intensity Pro'
7924 • Capture video clip at 1080i50:
7926 ffmpeg -format_code Hi50 -f decklink -i 'Intensity Pro' -c:a copy -c:v copy output.avi
7928 • Capture video clip at 1080i50 10 bit:
7930 ffmpeg -raw_format yuv422p10 -format_code Hi50 -f decklink -i 'UltraStudio Mini Recorder' -c:a copy -c:v copy output.avi
7932 • Capture video clip at 1080i50 with 16 audio channels:
7934 ffmpeg -channels 16 -format_code Hi50 -f decklink -i 'UltraStudio Mini Recorder' -c:a copy -c:v copy output.avi
7936 dshow
7938 Windows DirectShow input device.
7939 DirectShow support is enabled when FFmpeg is built with the mingw-w64
7941 project. Currently only audio and video devices are supported.
7942 Multiple devices may be opened as separate inputs, but they may also be
7944 opened on the same input, which should improve synchronism between
7945 them.
7946 The input name should be in the format:
7948 <TYPE>=<NAME>[:<TYPE>=<NAME>]
7950 where TYPE can be either audio or video, and NAME is the device's name
7952 or alternative name..
7953 Options
7955 If no options are specified, the device's defaults are used. If the
7957 device does not support the requested options, it will fail to open.
7958 video_size
7960 Set the video size in the captured video.
7961 framerate
7963 Set the frame rate in the captured video.
7964 sample_rate
7966 Set the sample rate (in Hz) of the captured audio.
7967 sample_size
7969 Set the sample size (in bits) of the captured audio.
7970 channels
7972 Set the number of channels in the captured audio.
7973 list_devices
7975 If set to true, print a list of devices and exit.
7976 list_options
7978 If set to true, print a list of selected device's options and exit.
7979 video_device_number
7981 Set video device number for devices with the same name (starts at
7982 0, defaults to 0).
7983 audio_device_number
7985 Set audio device number for devices with the same name (starts at
7986 0, defaults to 0).
7987 pixel_format
7989 Select pixel format to be used by DirectShow. This may only be set
7990 when the video codec is not set or set to rawvideo.
7991 audio_buffer_size
7993 Set audio device buffer size in milliseconds (which can directly
7994 impact latency, depending on the device). Defaults to using the
7995 audio device's default buffer size (typically some multiple of
7996 500ms). Setting this value too low can degrade performance. See
7997 also
7998 <http://msdn.microsoft.com/en-us/library/windows/desktop/dd377582(v=vs.85).aspx>
7999 video_pin_name
8001 Select video capture pin to use by name or alternative name.
8002 audio_pin_name
8004 Select audio capture pin to use by name or alternative name.
8005 crossbar_video_input_pin_number
8007 Select video input pin number for crossbar device. This will be
8008 routed to the crossbar device's Video Decoder output pin. Note
8009 that changing this value can affect future invocations (sets a new
8010 default) until system reboot occurs.
8011 crossbar_audio_input_pin_number
8013 Select audio input pin number for crossbar device. This will be
8014 routed to the crossbar device's Audio Decoder output pin. Note
8015 that changing this value can affect future invocations (sets a new
8016 default) until system reboot occurs.
8017 show_video_device_dialog
8019 If set to true, before capture starts, popup a display dialog to
8020 the end user, allowing them to change video filter properties and
8021 configurations manually. Note that for crossbar devices, adjusting
8022 values in this dialog may be needed at times to toggle between PAL
8023 (25 fps) and NTSC (29.97) input frame rates, sizes, interlacing,
8024 etc. Changing these values can enable different scan rates/frame
8025 rates and avoiding green bars at the bottom, flickering scan lines,
8026 etc. Note that with some devices, changing these properties can
8027 also affect future invocations (sets new defaults) until system
8028 reboot occurs.
8029 show_audio_device_dialog
8031 If set to true, before capture starts, popup a display dialog to
8032 the end user, allowing them to change audio filter properties and
8033 configurations manually.
8034 show_video_crossbar_connection_dialog
8036 If set to true, before capture starts, popup a display dialog to
8037 the end user, allowing them to manually modify crossbar pin
8038 routings, when it opens a video device.
8039 show_audio_crossbar_connection_dialog
8041 If set to true, before capture starts, popup a display dialog to
8042 the end user, allowing them to manually modify crossbar pin
8043 routings, when it opens an audio device.
8044 show_analog_tv_tuner_dialog
8046 If set to true, before capture starts, popup a display dialog to
8047 the end user, allowing them to manually modify TV channels and
8048 frequencies.
8049 show_analog_tv_tuner_audio_dialog
8051 If set to true, before capture starts, popup a display dialog to
8052 the end user, allowing them to manually modify TV audio (like mono
8053 vs. stereo, Language A,B or C).
8054 audio_device_load
8056 Load an audio capture filter device from file instead of searching
8057 it by name. It may load additional parameters too, if the filter
8058 supports the serialization of its properties to. To use this an
8059 audio capture source has to be specified, but it can be anything
8060 even fake one.
8061 audio_device_save
8063 Save the currently used audio capture filter device and its
8064 parameters (if the filter supports it) to a file. If a file with
8065 the same name exists it will be overwritten.
8066 video_device_load
8068 Load a video capture filter device from file instead of searching
8069 it by name. It may load additional parameters too, if the filter
8070 supports the serialization of its properties to. To use this a
8071 video capture source has to be specified, but it can be anything
8072 even fake one.
8073 video_device_save
8075 Save the currently used video capture filter device and its
8076 parameters (if the filter supports it) to a file. If a file with
8077 the same name exists it will be overwritten.
8078 use_video_device_timestamps
8080 If set to false, the timestamp for video frames will be derived
8081 from the wallclock instead of the timestamp provided by the capture
8082 device. This allows working around devices that provide unreliable
8083 timestamps.
8084 Examples
8086 • Print the list of DirectShow supported devices and exit:
8088 $ ffmpeg -list_devices true -f dshow -i dummy
8090 • Open video device Camera:
8092 $ ffmpeg -f dshow -i video="Camera"
8094 • Open second video device with name Camera:
8096 $ ffmpeg -f dshow -video_device_number 1 -i video="Camera"
8098 • Open video device Camera and audio device Microphone:
8100 $ ffmpeg -f dshow -i video="Camera":audio="Microphone"
8102 • Print the list of supported options in selected device and exit:
8104 $ ffmpeg -list_options true -f dshow -i video="Camera"
8106 • Specify pin names to capture by name or alternative name, specify
8108 alternative device name:
8109 $ 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"
8111 • Configure a crossbar device, specifying crossbar pins, allow user
8113 to adjust video capture properties at startup:
8114 $ ffmpeg -f dshow -show_video_device_dialog true -crossbar_video_input_pin_number 0
8116 -crossbar_audio_input_pin_number 3 -i video="AVerMedia BDA Analog Capture":audio="AVerMedia BDA Analog Capture"
8117 fbdev
8119 Linux framebuffer input device.
8120 The Linux framebuffer is a graphic hardware-independent abstraction
8122 layer to show graphics on a computer monitor, typically on the console.
8123 It is accessed through a file device node, usually /dev/fb0.
8124 For more detailed information read the file
8126 Documentation/fb/framebuffer.txt included in the Linux source tree.
8127 See also <http://linux-fbdev.sourceforge.net/>, and fbset(1).
8129 To record from the framebuffer device /dev/fb0 with ffmpeg:
8131 ffmpeg -f fbdev -framerate 10 -i /dev/fb0 out.avi
8133 You can take a single screenshot image with the command:
8135 ffmpeg -f fbdev -framerate 1 -i /dev/fb0 -frames:v 1 screenshot.jpeg
8137 Options
8139 framerate
8141 Set the frame rate. Default is 25.
8142 gdigrab
8144 Win32 GDI-based screen capture device.
8145 This device allows you to capture a region of the display on Windows.
8147 There are two options for the input filename:
8149 desktop
8151 or
8153 title=<window_title>
8155 The first option will capture the entire desktop, or a fixed region of
8157 the desktop. The second option will instead capture the contents of a
8158 single window, regardless of its position on the screen.
8159 For example, to grab the entire desktop using ffmpeg:
8161 ffmpeg -f gdigrab -framerate 6 -i desktop out.mpg
8163 Grab a 640x480 region at position "10,20":
8165 ffmpeg -f gdigrab -framerate 6 -offset_x 10 -offset_y 20 -video_size vga -i desktop out.mpg
8167 Grab the contents of the window named "Calculator"
8169 ffmpeg -f gdigrab -framerate 6 -i title=Calculator out.mpg
8171 Options
8173 draw_mouse
8175 Specify whether to draw the mouse pointer. Use the value 0 to not
8176 draw the pointer. Default value is 1.
8177 framerate
8179 Set the grabbing frame rate. Default value is "ntsc", corresponding
8180 to a frame rate of "30000/1001".
8181 show_region
8183 Show grabbed region on screen.
8184 If show_region is specified with 1, then the grabbing region will
8186 be indicated on screen. With this option, it is easy to know what
8187 is being grabbed if only a portion of the screen is grabbed.
8188 Note that show_region is incompatible with grabbing the contents of
8190 a single window.
8191 For example:
8193 ffmpeg -f gdigrab -show_region 1 -framerate 6 -video_size cif -offset_x 10 -offset_y 20 -i desktop out.mpg
8195 video_size
8197 Set the video frame size. The default is to capture the full screen
8198 if desktop is selected, or the full window size if
8199 title=window_title is selected.
8200 offset_x
8202 When capturing a region with video_size, set the distance from the
8203 left edge of the screen or desktop.
8204 Note that the offset calculation is from the top left corner of the
8206 primary monitor on Windows. If you have a monitor positioned to the
8207 left of your primary monitor, you will need to use a negative
8208 offset_x value to move the region to that monitor.
8209 offset_y
8211 When capturing a region with video_size, set the distance from the
8212 top edge of the screen or desktop.
8213 Note that the offset calculation is from the top left corner of the
8215 primary monitor on Windows. If you have a monitor positioned above
8216 your primary monitor, you will need to use a negative offset_y
8217 value to move the region to that monitor.
8218 iec61883
8220 FireWire DV/HDV input device using libiec61883.
8221 To enable this input device, you need libiec61883, libraw1394 and
8223 libavc1394 installed on your system. Use the configure option
8224 "--enable-libiec61883" to compile with the device enabled.
8225 The iec61883 capture device supports capturing from a video device
8227 connected via IEEE1394 (FireWire), using libiec61883 and the new Linux
8228 FireWire stack (juju). This is the default DV/HDV input method in Linux
8229 Kernel 2.6.37 and later, since the old FireWire stack was removed.
8230 Specify the FireWire port to be used as input file, or "auto" to choose
8232 the first port connected.
8233 Options
8235 dvtype
8237 Override autodetection of DV/HDV. This should only be used if auto
8238 detection does not work, or if usage of a different device type
8239 should be prohibited. Treating a DV device as HDV (or vice versa)
8240 will not work and result in undefined behavior. The values auto,
8241 dv and hdv are supported.
8242 dvbuffer
8244 Set maximum size of buffer for incoming data, in frames. For DV,
8245 this is an exact value. For HDV, it is not frame exact, since HDV
8246 does not have a fixed frame size.
8247 dvguid
8249 Select the capture device by specifying its GUID. Capturing will
8250 only be performed from the specified device and fails if no device
8251 with the given GUID is found. This is useful to select the input if
8252 multiple devices are connected at the same time. Look at
8253 /sys/bus/firewire/devices to find out the GUIDs.
8254 Examples
8256 • Grab and show the input of a FireWire DV/HDV device.
8258 ffplay -f iec61883 -i auto
8260 • Grab and record the input of a FireWire DV/HDV device, using a
8262 packet buffer of 100000 packets if the source is HDV.
8263 ffmpeg -f iec61883 -i auto -dvbuffer 100000 out.mpg
8265 jack
8267 JACK input device.
8268 To enable this input device during configuration you need libjack
8270 installed on your system.
8271 A JACK input device creates one or more JACK writable clients, one for
8273 each audio channel, with name client_name:input_N, where client_name is
8274 the name provided by the application, and N is a number which
8275 identifies the channel. Each writable client will send the acquired
8276 data to the FFmpeg input device.
8277 Once you have created one or more JACK readable clients, you need to
8279 connect them to one or more JACK writable clients.
8280 To connect or disconnect JACK clients you can use the jack_connect and
8282 jack_disconnect programs, or do it through a graphical interface, for
8283 example with qjackctl.
8284 To list the JACK clients and their properties you can invoke the
8286 command jack_lsp.
8287 Follows an example which shows how to capture a JACK readable client
8289 with ffmpeg.
8290 # Create a JACK writable client with name "ffmpeg".
8292 $ ffmpeg -f jack -i ffmpeg -y out.wav
8293 # Start the sample jack_metro readable client.
8295 $ jack_metro -b 120 -d 0.2 -f 4000
8296 # List the current JACK clients.
8298 $ jack_lsp -c
8299 system:capture_1
8300 system:capture_2
8301 system:playback_1
8302 system:playback_2
8303 ffmpeg:input_1
8304 metro:120_bpm
8305 # Connect metro to the ffmpeg writable client.
8307 $ jack_connect metro:120_bpm ffmpeg:input_1
8308 For more information read: <http://jackaudio.org/>
8310 Options
8312 channels
8314 Set the number of channels. Default is 2.
8315 kmsgrab
8317 KMS video input device.
8318 Captures the KMS scanout framebuffer associated with a specified CRTC
8320 or plane as a DRM object that can be passed to other hardware
8321 functions.
8322 Requires either DRM master or CAP_SYS_ADMIN to run.
8324 If you don't understand what all of that means, you probably don't want
8326 this. Look at x11grab instead.
8327 Options
8329 device
8331 DRM device to capture on. Defaults to /dev/dri/card0.
8332 format
8334 Pixel format of the framebuffer. This can be autodetected if you
8335 are running Linux 5.7 or later, but needs to be provided for
8336 earlier versions. Defaults to bgr0, which is the most common
8337 format used by the Linux console and Xorg X server.
8338 format_modifier
8340 Format modifier to signal on output frames. This is necessary to
8341 import correctly into some APIs. It can be autodetected if you are
8342 running Linux 5.7 or later, but will need to be provided explicitly
8343 when needed in earlier versions. See the libdrm documentation for
8344 possible values.
8345 crtc_id
8347 KMS CRTC ID to define the capture source. The first active plane
8348 on the given CRTC will be used.
8349 plane_id
8351 KMS plane ID to define the capture source. Defaults to the first
8352 active plane found if neither crtc_id nor plane_id are specified.
8353 framerate
8355 Framerate to capture at. This is not synchronised to any page
8356 flipping or framebuffer changes - it just defines the interval at
8357 which the framebuffer is sampled. Sampling faster than the
8358 framebuffer update rate will generate independent frames with the
8359 same content. Defaults to 30.
8360 Examples
8362 • Capture from the first active plane, download the result to normal
8364 frames and encode. This will only work if the framebuffer is both
8365 linear and mappable - if not, the result may be scrambled or fail
8366 to download.
8367 ffmpeg -f kmsgrab -i - -vf 'hwdownload,format=bgr0' output.mp4
8369 • Capture from CRTC ID 42 at 60fps, map the result to VAAPI, convert
8371 to NV12 and encode as H.264.
8372 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
8374 • To capture only part of a plane the output can be cropped - this
8376 can be used to capture a single window, as long as it has a known
8377 absolute position and size. For example, to capture and encode the
8378 middle quarter of a 1920x1080 plane:
8379 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
8381 lavfi
8383 Libavfilter input virtual device.
8384 This input device reads data from the open output pads of a libavfilter
8386 filtergraph.
8387 For each filtergraph open output, the input device will create a
8389 corresponding stream which is mapped to the generated output. Currently
8390 only video data is supported. The filtergraph is specified through the
8391 option graph.
8392 Options
8394 graph
8396 Specify the filtergraph to use as input. Each video open output
8397 must be labelled by a unique string of the form "outN", where N is
8398 a number starting from 0 corresponding to the mapped input stream
8399 generated by the device. The first unlabelled output is
8400 automatically assigned to the "out0" label, but all the others need
8401 to be specified explicitly.
8402 The suffix "+subcc" can be appended to the output label to create
8404 an extra stream with the closed captions packets attached to that
8405 output (experimental; only for EIA-608 / CEA-708 for now). The
8406 subcc streams are created after all the normal streams, in the
8407 order of the corresponding stream. For example, if there is
8408 "out19+subcc", "out7+subcc" and up to "out42", the stream #43 is
8409 subcc for stream #7 and stream #44 is subcc for stream #19.
8410 If not specified defaults to the filename specified for the input
8412 device.
8413 graph_file
8415 Set the filename of the filtergraph to be read and sent to the
8416 other filters. Syntax of the filtergraph is the same as the one
8417 specified by the option graph.
8418 dumpgraph
8420 Dump graph to stderr.
8421 Examples
8423 • Create a color video stream and play it back with ffplay:
8425 ffplay -f lavfi -graph "color=c=pink [out0]" dummy
8427 • As the previous example, but use filename for specifying the graph
8429 description, and omit the "out0" label:
8430 ffplay -f lavfi color=c=pink
8432 • Create three different video test filtered sources and play them:
8434 ffplay -f lavfi -graph "testsrc [out0]; testsrc,hflip [out1]; testsrc,negate [out2]" test3
8436 • Read an audio stream from a file using the amovie source and play
8438 it back with ffplay:
8439 ffplay -f lavfi "amovie=test.wav"
8441 • Read an audio stream and a video stream and play it back with
8443 ffplay:
8444 ffplay -f lavfi "movie=test.avi[out0];amovie=test.wav[out1]"
8446 • Dump decoded frames to images and closed captions to a file
8448 (experimental):
8449 ffmpeg -f lavfi -i "movie=test.ts[out0+subcc]" -map v frame%08d.png -map s -c copy -f rawvideo subcc.bin
8451 libcdio
8453 Audio-CD input device based on libcdio.
8454 To enable this input device during configuration you need libcdio
8456 installed on your system. It requires the configure option
8457 "--enable-libcdio".
8458 This device allows playing and grabbing from an Audio-CD.
8460 For example to copy with ffmpeg the entire Audio-CD in /dev/sr0, you
8462 may run the command:
8463 ffmpeg -f libcdio -i /dev/sr0 cd.wav
8465 Options
8467 speed
8469 Set drive reading speed. Default value is 0.
8470 The speed is specified CD-ROM speed units. The speed is set through
8472 the libcdio "cdio_cddap_speed_set" function. On many CD-ROM drives,
8473 specifying a value too large will result in using the fastest
8474 speed.
8475 paranoia_mode
8477 Set paranoia recovery mode flags. It accepts one of the following
8478 values:
8479 disable
8481 verify
8482 overlap
8483 neverskip
8484 full
8485 Default value is disable.
8487 For more information about the available recovery modes, consult
8489 the paranoia project documentation.
8490 libdc1394
8492 IIDC1394 input device, based on libdc1394 and libraw1394.
8493 Requires the configure option "--enable-libdc1394".
8495 Options
8497 framerate
8499 Set the frame rate. Default is "ntsc", corresponding to a frame
8500 rate of "30000/1001".
8501 pixel_format
8503 Select the pixel format. Default is "uyvy422".
8504 video_size
8506 Set the video size given as a string such as "640x480" or "hd720".
8507 Default is "qvga".
8508 openal
8510 The OpenAL input device provides audio capture on all systems with a
8511 working OpenAL 1.1 implementation.
8512 To enable this input device during configuration, you need OpenAL
8514 headers and libraries installed on your system, and need to configure
8515 FFmpeg with "--enable-openal".
8516 OpenAL headers and libraries should be provided as part of your OpenAL
8518 implementation, or as an additional download (an SDK). Depending on
8519 your installation you may need to specify additional flags via the
8520 "--extra-cflags" and "--extra-ldflags" for allowing the build system to
8521 locate the OpenAL headers and libraries.
8522 An incomplete list of OpenAL implementations follows:
8524 Creative
8526 The official Windows implementation, providing hardware
8527 acceleration with supported devices and software fallback. See
8528 <http://openal.org/>.
8529 OpenAL Soft
8531 Portable, open source (LGPL) software implementation. Includes
8532 backends for the most common sound APIs on the Windows, Linux,
8533 Solaris, and BSD operating systems. See
8534 <http://kcat.strangesoft.net/openal.html>.
8535 Apple
8537 OpenAL is part of Core Audio, the official Mac OS X Audio
8538 interface. See
8539 <http://developer.apple.com/technologies/mac/audio-and-video.html>
8540 This device allows one to capture from an audio input device handled
8542 through OpenAL.
8543 You need to specify the name of the device to capture in the provided
8545 filename. If the empty string is provided, the device will
8546 automatically select the default device. You can get the list of the
8547 supported devices by using the option list_devices.
8548 Options
8550 channels
8552 Set the number of channels in the captured audio. Only the values 1
8553 (monaural) and 2 (stereo) are currently supported. Defaults to 2.
8554 sample_size
8556 Set the sample size (in bits) of the captured audio. Only the
8557 values 8 and 16 are currently supported. Defaults to 16.
8558 sample_rate
8560 Set the sample rate (in Hz) of the captured audio. Defaults to
8561 44.1k.
8562 list_devices
8564 If set to true, print a list of devices and exit. Defaults to
8565 false.
8566 Examples
8568 Print the list of OpenAL supported devices and exit:
8570 $ ffmpeg -list_devices true -f openal -i dummy out.ogg
8572 Capture from the OpenAL device DR-BT101 via PulseAudio:
8574 $ ffmpeg -f openal -i 'DR-BT101 via PulseAudio' out.ogg
8576 Capture from the default device (note the empty string '' as filename):
8578 $ ffmpeg -f openal -i '' out.ogg
8580 Capture from two devices simultaneously, writing to two different
8582 files, within the same ffmpeg command:
8583 $ ffmpeg -f openal -i 'DR-BT101 via PulseAudio' out1.ogg -f openal -i 'ALSA Default' out2.ogg
8585 Note: not all OpenAL implementations support multiple simultaneous
8587 capture - try the latest OpenAL Soft if the above does not work.
8588 oss
8590 Open Sound System input device.
8591 The filename to provide to the input device is the device node
8593 representing the OSS input device, and is usually set to /dev/dsp.
8594 For example to grab from /dev/dsp using ffmpeg use the command:
8596 ffmpeg -f oss -i /dev/dsp /tmp/oss.wav
8598 For more information about OSS see:
8600 <http://manuals.opensound.com/usersguide/dsp.html>
8601 Options
8603 sample_rate
8605 Set the sample rate in Hz. Default is 48000.
8606 channels
8608 Set the number of channels. Default is 2.
8609 pulse
8611 PulseAudio input device.
8612 To enable this output device you need to configure FFmpeg with
8614 "--enable-libpulse".
8615 The filename to provide to the input device is a source device or the
8617 string "default"
8618 To list the PulseAudio source devices and their properties you can
8620 invoke the command pactl list sources.
8621 More information about PulseAudio can be found on
8623 <http://www.pulseaudio.org>.
8624 Options
8626 server
8628 Connect to a specific PulseAudio server, specified by an IP
8629 address. Default server is used when not provided.
8630 name
8632 Specify the application name PulseAudio will use when showing
8633 active clients, by default it is the "LIBAVFORMAT_IDENT" string.
8634 stream_name
8636 Specify the stream name PulseAudio will use when showing active
8637 streams, by default it is "record".
8638 sample_rate
8640 Specify the samplerate in Hz, by default 48kHz is used.
8641 channels
8643 Specify the channels in use, by default 2 (stereo) is set.
8644 frame_size
8646 This option does nothing and is deprecated.
8647 fragment_size
8649 Specify the size in bytes of the minimal buffering fragment in
8650 PulseAudio, it will affect the audio latency. By default it is set
8651 to 50 ms amount of data.
8652 wallclock
8654 Set the initial PTS using the current time. Default is 1.
8655 Examples
8657 Record a stream from default device:
8659 ffmpeg -f pulse -i default /tmp/pulse.wav
8661 sndio
8663 sndio input device.
8664 To enable this input device during configuration you need libsndio
8666 installed on your system.
8667 The filename to provide to the input device is the device node
8669 representing the sndio input device, and is usually set to /dev/audio0.
8670 For example to grab from /dev/audio0 using ffmpeg use the command:
8672 ffmpeg -f sndio -i /dev/audio0 /tmp/oss.wav
8674 Options
8676 sample_rate
8678 Set the sample rate in Hz. Default is 48000.
8679 channels
8681 Set the number of channels. Default is 2.
8682 video4linux2, v4l2
8684 Video4Linux2 input video device.
8685 "v4l2" can be used as alias for "video4linux2".
8687 If FFmpeg is built with v4l-utils support (by using the
8689 "--enable-libv4l2" configure option), it is possible to use it with the
8690 "-use_libv4l2" input device option.
8691 The name of the device to grab is a file device node, usually Linux
8693 systems tend to automatically create such nodes when the device (e.g.
8694 an USB webcam) is plugged into the system, and has a name of the kind
8695 /dev/videoN, where N is a number associated to the device.
8696 Video4Linux2 devices usually support a limited set of widthxheight
8698 sizes and frame rates. You can check which are supported using
8699 -list_formats all for Video4Linux2 devices. Some devices, like TV
8700 cards, support one or more standards. It is possible to list all the
8701 supported standards using -list_standards all.
8702 The time base for the timestamps is 1 microsecond. Depending on the
8704 kernel version and configuration, the timestamps may be derived from
8705 the real time clock (origin at the Unix Epoch) or the monotonic clock
8706 (origin usually at boot time, unaffected by NTP or manual changes to
8707 the clock). The -timestamps abs or -ts abs option can be used to force
8708 conversion into the real time clock.
8709 Some usage examples of the video4linux2 device with ffmpeg and ffplay:
8711 • List supported formats for a video4linux2 device:
8713 ffplay -f video4linux2 -list_formats all /dev/video0
8715 • Grab and show the input of a video4linux2 device:
8717 ffplay -f video4linux2 -framerate 30 -video_size hd720 /dev/video0
8719 • Grab and record the input of a video4linux2 device, leave the frame
8721 rate and size as previously set:
8722 ffmpeg -f video4linux2 -input_format mjpeg -i /dev/video0 out.mpeg
8724 For more information about Video4Linux, check <http://linuxtv.org/>.
8726 Options
8728 standard
8730 Set the standard. Must be the name of a supported standard. To get
8731 a list of the supported standards, use the list_standards option.
8732 channel
8734 Set the input channel number. Default to -1, which means using the
8735 previously selected channel.
8736 video_size
8738 Set the video frame size. The argument must be a string in the form
8739 WIDTHxHEIGHT or a valid size abbreviation.
8740 pixel_format
8742 Select the pixel format (only valid for raw video input).
8743 input_format
8745 Set the preferred pixel format (for raw video) or a codec name.
8746 This option allows one to select the input format, when several are
8747 available.
8748 framerate
8750 Set the preferred video frame rate.
8751 list_formats
8753 List available formats (supported pixel formats, codecs, and frame
8754 sizes) and exit.
8755 Available values are:
8757 all Show all available (compressed and non-compressed) formats.
8759 raw Show only raw video (non-compressed) formats.
8761 compressed
8763 Show only compressed formats.
8764 list_standards
8766 List supported standards and exit.
8767 Available values are:
8769 all Show all supported standards.
8771 timestamps, ts
8773 Set type of timestamps for grabbed frames.
8774 Available values are:
8776 default
8778 Use timestamps from the kernel.
8779 abs Use absolute timestamps (wall clock).
8781 mono2abs
8783 Force conversion from monotonic to absolute timestamps.
8784 Default value is "default".
8786 use_libv4l2
8788 Use libv4l2 (v4l-utils) conversion functions. Default is 0.
8789 vfwcap
8791 VfW (Video for Windows) capture input device.
8792 The filename passed as input is the capture driver number, ranging from
8794 0 to 9. You may use "list" as filename to print a list of drivers. Any
8795 other filename will be interpreted as device number 0.
8796 Options
8798 video_size
8800 Set the video frame size.
8801 framerate
8803 Set the grabbing frame rate. Default value is "ntsc", corresponding
8804 to a frame rate of "30000/1001".
8805 x11grab
8807 X11 video input device.
8808 To enable this input device during configuration you need libxcb
8810 installed on your system. It will be automatically detected during
8811 configuration.
8812 This device allows one to capture a region of an X11 display.
8814 The filename passed as input has the syntax:
8816 [<hostname>]:<display_number>.<screen_number>[+<x_offset>,<y_offset>]
8818 hostname:display_number.screen_number specifies the X11 display name of
8820 the screen to grab from. hostname can be omitted, and defaults to
8821 "localhost". The environment variable DISPLAY contains the default
8822 display name.
8823 x_offset and y_offset specify the offsets of the grabbed area with
8825 respect to the top-left border of the X11 screen. They default to 0.
8826 Check the X11 documentation (e.g. man X) for more detailed information.
8828 Use the xdpyinfo program for getting basic information about the
8830 properties of your X11 display (e.g. grep for "name" or "dimensions").
8831 For example to grab from :0.0 using ffmpeg:
8833 ffmpeg -f x11grab -framerate 25 -video_size cif -i :0.0 out.mpg
8835 Grab at position "10,20":
8837 ffmpeg -f x11grab -framerate 25 -video_size cif -i :0.0+10,20 out.mpg
8839 Options
8841 select_region
8843 Specify whether to select the grabbing area graphically using the
8844 pointer. A value of 1 prompts the user to select the grabbing area
8845 graphically by clicking and dragging. A single click with no
8846 dragging will select the whole screen. A region with zero width or
8847 height will also select the whole screen. This option overwrites
8848 the video_size, grab_x, and grab_y options. Default value is 0.
8849 draw_mouse
8851 Specify whether to draw the mouse pointer. A value of 0 specifies
8852 not to draw the pointer. Default value is 1.
8853 follow_mouse
8855 Make the grabbed area follow the mouse. The argument can be
8856 "centered" or a number of pixels PIXELS.
8857 When it is specified with "centered", the grabbing region follows
8859 the mouse pointer and keeps the pointer at the center of region;
8860 otherwise, the region follows only when the mouse pointer reaches
8861 within PIXELS (greater than zero) to the edge of region.
8862 For example:
8864 ffmpeg -f x11grab -follow_mouse centered -framerate 25 -video_size cif -i :0.0 out.mpg
8866 To follow only when the mouse pointer reaches within 100 pixels to
8868 edge:
8869 ffmpeg -f x11grab -follow_mouse 100 -framerate 25 -video_size cif -i :0.0 out.mpg
8871 framerate
8873 Set the grabbing frame rate. Default value is "ntsc", corresponding
8874 to a frame rate of "30000/1001".
8875 show_region
8877 Show grabbed region on screen.
8878 If show_region is specified with 1, then the grabbing region will
8880 be indicated on screen. With this option, it is easy to know what
8881 is being grabbed if only a portion of the screen is grabbed.
8882 region_border
8884 Set the region border thickness if -show_region 1 is used. Range
8885 is 1 to 128 and default is 3 (XCB-based x11grab only).
8886 For example:
8888 ffmpeg -f x11grab -show_region 1 -framerate 25 -video_size cif -i :0.0+10,20 out.mpg
8890 With follow_mouse:
8892 ffmpeg -f x11grab -follow_mouse centered -show_region 1 -framerate 25 -video_size cif -i :0.0 out.mpg
8894 window_id
8896 Grab this window, instead of the whole screen. Default value is 0,
8897 which maps to the whole screen (root window).
8898 The id of a window can be found using the xwininfo program,
8900 possibly with options -tree and -root.
8901 If the window is later enlarged, the new area is not recorded.
8903 Video ends when the window is closed, unmapped (i.e., iconified) or
8904 shrunk beyond the video size (which defaults to the initial window
8905 size).
8906 This option disables options follow_mouse and select_region.
8908 video_size
8910 Set the video frame size. Default is the full desktop or window.
8911 grab_x
8913 grab_y
8914 Set the grabbing region coordinates. They are expressed as offset
8915 from the top left corner of the X11 window and correspond to the
8916 x_offset and y_offset parameters in the device name. The default
8917 value for both options is 0.
8918
8920 The audio resampler supports the following named options.
8921 Options may be set by specifying -option value in the FFmpeg tools,
8923 option=value for the aresample filter, by setting the value explicitly
8924 in the "SwrContext" options or using the libavutil/opt.h API for
8925 programmatic use.
8926 ich, in_channel_count
8928 Set the number of input channels. Default value is 0. Setting this
8929 value is not mandatory if the corresponding channel layout
8930 in_channel_layout is set.
8931 och, out_channel_count
8933 Set the number of output channels. Default value is 0. Setting this
8934 value is not mandatory if the corresponding channel layout
8935 out_channel_layout is set.
8936 uch, used_channel_count
8938 Set the number of used input channels. Default value is 0. This
8939 option is only used for special remapping.
8940 isr, in_sample_rate
8942 Set the input sample rate. Default value is 0.
8943 osr, out_sample_rate
8945 Set the output sample rate. Default value is 0.
8946 isf, in_sample_fmt
8948 Specify the input sample format. It is set by default to "none".
8949 osf, out_sample_fmt
8951 Specify the output sample format. It is set by default to "none".
8952 tsf, internal_sample_fmt
8954 Set the internal sample format. Default value is "none". This will
8955 automatically be chosen when it is not explicitly set.
8956 icl, in_channel_layout
8958 ocl, out_channel_layout
8959 Set the input/output channel layout.
8960 See the Channel Layout section in the ffmpeg-utils(1) manual for
8962 the required syntax.
8963 clev, center_mix_level
8965 Set the center mix level. It is a value expressed in deciBel, and
8966 must be in the interval [-32,32].
8967 slev, surround_mix_level
8969 Set the surround mix level. It is a value expressed in deciBel, and
8970 must be in the interval [-32,32].
8971 lfe_mix_level
8973 Set LFE mix into non LFE level. It is used when there is a LFE
8974 input but no LFE output. It is a value expressed in deciBel, and
8975 must be in the interval [-32,32].
8976 rmvol, rematrix_volume
8978 Set rematrix volume. Default value is 1.0.
8979 rematrix_maxval
8981 Set maximum output value for rematrixing. This can be used to
8982 prevent clipping vs. preventing volume reduction. A value of 1.0
8983 prevents clipping.
8984 flags, swr_flags
8986 Set flags used by the converter. Default value is 0.
8987 It supports the following individual flags:
8989 res force resampling, this flag forces resampling to be used even
8991 when the input and output sample rates match.
8992 dither_scale
8994 Set the dither scale. Default value is 1.
8995 dither_method
8997 Set dither method. Default value is 0.
8998 Supported values:
9000 rectangular
9002 select rectangular dither
9003 triangular
9005 select triangular dither
9006 triangular_hp
9008 select triangular dither with high pass
9009 lipshitz
9011 select Lipshitz noise shaping dither.
9012 shibata
9014 select Shibata noise shaping dither.
9015 low_shibata
9017 select low Shibata noise shaping dither.
9018 high_shibata
9020 select high Shibata noise shaping dither.
9021 f_weighted
9023 select f-weighted noise shaping dither
9024 modified_e_weighted
9026 select modified-e-weighted noise shaping dither
9027 improved_e_weighted
9029 select improved-e-weighted noise shaping dither
9030 resampler
9032 Set resampling engine. Default value is swr.
9033 Supported values:
9035 swr select the native SW Resampler; filter options precision and
9037 cheby are not applicable in this case.
9038 soxr
9040 select the SoX Resampler (where available); compensation, and
9041 filter options filter_size, phase_shift, exact_rational,
9042 filter_type & kaiser_beta, are not applicable in this case.
9043 filter_size
9045 For swr only, set resampling filter size, default value is 32.
9046 phase_shift
9048 For swr only, set resampling phase shift, default value is 10, and
9049 must be in the interval [0,30].
9050 linear_interp
9052 Use linear interpolation when enabled (the default). Disable it if
9053 you want to preserve speed instead of quality when exact_rational
9054 fails.
9055 exact_rational
9057 For swr only, when enabled, try to use exact phase_count based on
9058 input and output sample rate. However, if it is larger than "1 <<
9059 phase_shift", the phase_count will be "1 << phase_shift" as
9060 fallback. Default is enabled.
9061 cutoff
9063 Set cutoff frequency (swr: 6dB point; soxr: 0dB point) ratio; must
9064 be a float value between 0 and 1. Default value is 0.97 with swr,
9065 and 0.91 with soxr (which, with a sample-rate of 44100, preserves
9066 the entire audio band to 20kHz).
9067 precision
9069 For soxr only, the precision in bits to which the resampled signal
9070 will be calculated. The default value of 20 (which, with suitable
9071 dithering, is appropriate for a destination bit-depth of 16) gives
9072 SoX's 'High Quality'; a value of 28 gives SoX's 'Very High
9073 Quality'.
9074 cheby
9076 For soxr only, selects passband rolloff none (Chebyshev) & higher-
9077 precision approximation for 'irrational' ratios. Default value is
9078 0.
9079 async
9081 For swr only, simple 1 parameter audio sync to timestamps using
9082 stretching, squeezing, filling and trimming. Setting this to 1 will
9083 enable filling and trimming, larger values represent the maximum
9084 amount in samples that the data may be stretched or squeezed for
9085 each second. Default value is 0, thus no compensation is applied
9086 to make the samples match the audio timestamps.
9087 first_pts
9089 For swr only, assume the first pts should be this value. The time
9090 unit is 1 / sample rate. This allows for padding/trimming at the
9091 start of stream. By default, no assumption is made about the first
9092 frame's expected pts, so no padding or trimming is done. For
9093 example, this could be set to 0 to pad the beginning with silence
9094 if an audio stream starts after the video stream or to trim any
9095 samples with a negative pts due to encoder delay.
9096 min_comp
9098 For swr only, set the minimum difference between timestamps and
9099 audio data (in seconds) to trigger stretching/squeezing/filling or
9100 trimming of the data to make it match the timestamps. The default
9101 is that stretching/squeezing/filling and trimming is disabled
9102 (min_comp = "FLT_MAX").
9103 min_hard_comp
9105 For swr only, set the minimum difference between timestamps and
9106 audio data (in seconds) to trigger adding/dropping samples to make
9107 it match the timestamps. This option effectively is a threshold to
9108 select between hard (trim/fill) and soft (squeeze/stretch)
9109 compensation. Note that all compensation is by default disabled
9110 through min_comp. The default is 0.1.
9111 comp_duration
9113 For swr only, set duration (in seconds) over which data is
9114 stretched/squeezed to make it match the timestamps. Must be a non-
9115 negative double float value, default value is 1.0.
9116 max_soft_comp
9118 For swr only, set maximum factor by which data is
9119 stretched/squeezed to make it match the timestamps. Must be a non-
9120 negative double float value, default value is 0.
9121 matrix_encoding
9123 Select matrixed stereo encoding.
9124 It accepts the following values:
9126 none
9128 select none
9129 dolby
9131 select Dolby
9132 dplii
9134 select Dolby Pro Logic II
9135 Default value is "none".
9137 filter_type
9139 For swr only, select resampling filter type. This only affects
9140 resampling operations.
9141 It accepts the following values:
9143 cubic
9145 select cubic
9146 blackman_nuttall
9148 select Blackman Nuttall windowed sinc
9149 kaiser
9151 select Kaiser windowed sinc
9152 kaiser_beta
9154 For swr only, set Kaiser window beta value. Must be a double float
9155 value in the interval [2,16], default value is 9.
9156 output_sample_bits
9158 For swr only, set number of used output sample bits for dithering.
9159 Must be an integer in the interval [0,64], default value is 0,
9160 which means it's not used.
9161
9163 The video scaler supports the following named options.
9164 Options may be set by specifying -option value in the FFmpeg tools,
9166 with a few API-only exceptions noted below. For programmatic use, they
9167 can be set explicitly in the "SwsContext" options or through the
9168 libavutil/opt.h API.
9169 sws_flags
9171 Set the scaler flags. This is also used to set the scaling
9172 algorithm. Only a single algorithm should be selected. Default
9173 value is bicubic.
9174 It accepts the following values:
9176 fast_bilinear
9178 Select fast bilinear scaling algorithm.
9179 bilinear
9181 Select bilinear scaling algorithm.
9182 bicubic
9184 Select bicubic scaling algorithm.
9185 experimental
9187 Select experimental scaling algorithm.
9188 neighbor
9190 Select nearest neighbor rescaling algorithm.
9191 area
9193 Select averaging area rescaling algorithm.
9194 bicublin
9196 Select bicubic scaling algorithm for the luma component,
9197 bilinear for chroma components.
9198 gauss
9200 Select Gaussian rescaling algorithm.
9201 sinc
9203 Select sinc rescaling algorithm.
9204 lanczos
9206 Select Lanczos rescaling algorithm. The default width (alpha)
9207 is 3 and can be changed by setting "param0".
9208 spline
9210 Select natural bicubic spline rescaling algorithm.
9211 print_info
9213 Enable printing/debug logging.
9214 accurate_rnd
9216 Enable accurate rounding.
9217 full_chroma_int
9219 Enable full chroma interpolation.
9220 full_chroma_inp
9222 Select full chroma input.
9223 bitexact
9225 Enable bitexact output.
9226 srcw (API only)
9228 Set source width.
9229 srch (API only)
9231 Set source height.
9232 dstw (API only)
9234 Set destination width.
9235 dsth (API only)
9237 Set destination height.
9238 src_format (API only)
9240 Set source pixel format (must be expressed as an integer).
9241 dst_format (API only)
9243 Set destination pixel format (must be expressed as an integer).
9244 src_range (boolean)
9246 If value is set to 1, indicates source is full range. Default value
9247 is 0, which indicates source is limited range.
9248 dst_range (boolean)
9250 If value is set to 1, enable full range for destination. Default
9251 value is 0, which enables limited range.
9252 param0, param1
9254 Set scaling algorithm parameters. The specified values are specific
9255 of some scaling algorithms and ignored by others. The specified
9256 values are floating point number values.
9257 sws_dither
9259 Set the dithering algorithm. Accepts one of the following values.
9260 Default value is auto.
9261 auto
9263 automatic choice
9264 none
9266 no dithering
9267 bayer
9269 bayer dither
9270 ed error diffusion dither
9272 a_dither
9274 arithmetic dither, based using addition
9275 x_dither
9277 arithmetic dither, based using xor (more random/less apparent
9278 patterning that a_dither).
9279 alphablend
9281 Set the alpha blending to use when the input has alpha but the
9282 output does not. Default value is none.
9283 uniform_color
9285 Blend onto a uniform background color
9286 checkerboard
9288 Blend onto a checkerboard
9289 none
9291 No blending
9292
9294 Filtering in FFmpeg is enabled through the libavfilter library.
9295 In libavfilter, a filter can have multiple inputs and multiple outputs.
9297 To illustrate the sorts of things that are possible, we consider the
9298 following filtergraph.
9299 [main]
9301 input --> split ---------------------> overlay --> output
9302 | ^
9303 |[tmp] [flip]|
9304 +-----> crop --> vflip -------+
9305 This filtergraph splits the input stream in two streams, then sends one
9307 stream through the crop filter and the vflip filter, before merging it
9308 back with the other stream by overlaying it on top. You can use the
9309 following command to achieve this:
9310 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
9312 The result will be that the top half of the video is mirrored onto the
9314 bottom half of the output video.
9315 Filters in the same linear chain are separated by commas, and distinct
9317 linear chains of filters are separated by semicolons. In our example,
9318 crop,vflip are in one linear chain, split and overlay are separately in
9319 another. The points where the linear chains join are labelled by names
9320 enclosed in square brackets. In the example, the split filter generates
9321 two outputs that are associated to the labels [main] and [tmp].
9322 The stream sent to the second output of split, labelled as [tmp], is
9324 processed through the crop filter, which crops away the lower half part
9325 of the video, and then vertically flipped. The overlay filter takes in
9326 input the first unchanged output of the split filter (which was
9327 labelled as [main]), and overlay on its lower half the output generated
9328 by the crop,vflip filterchain.
9329 Some filters take in input a list of parameters: they are specified
9331 after the filter name and an equal sign, and are separated from each
9332 other by a colon.
9333 There exist so-called source filters that do not have an audio/video
9335 input, and sink filters that will not have audio/video output.
9336
9338 The graph2dot program included in the FFmpeg tools directory can be
9339 used to parse a filtergraph description and issue a corresponding
9340 textual representation in the dot language.
9341 Invoke the command:
9343 graph2dot -h
9345 to see how to use graph2dot.
9347 You can then pass the dot description to the dot program (from the
9349 graphviz suite of programs) and obtain a graphical representation of
9350 the filtergraph.
9351 For example the sequence of commands:
9353 echo <GRAPH_DESCRIPTION> | \
9355 tools/graph2dot -o graph.tmp && \
9356 dot -Tpng graph.tmp -o graph.png && \
9357 display graph.png
9358 can be used to create and display an image representing the graph
9360 described by the GRAPH_DESCRIPTION string. Note that this string must
9361 be a complete self-contained graph, with its inputs and outputs
9362 explicitly defined. For example if your command line is of the form:
9363 ffmpeg -i infile -vf scale=640:360 outfile
9365 your GRAPH_DESCRIPTION string will need to be of the form:
9367 nullsrc,scale=640:360,nullsink
9369 you may also need to set the nullsrc parameters and add a format filter
9371 in order to simulate a specific input file.
9372
9374 A filtergraph is a directed graph of connected filters. It can contain
9375 cycles, and there can be multiple links between a pair of filters. Each
9376 link has one input pad on one side connecting it to one filter from
9377 which it takes its input, and one output pad on the other side
9378 connecting it to one filter accepting its output.
9379 Each filter in a filtergraph is an instance of a filter class
9381 registered in the application, which defines the features and the
9382 number of input and output pads of the filter.
9383 A filter with no input pads is called a "source", and a filter with no
9385 output pads is called a "sink".
9386 Filtergraph syntax
9388 A filtergraph has a textual representation, which is recognized by the
9389 -filter/-vf/-af and -filter_complex options in ffmpeg and -vf/-af in
9390 ffplay, and by the "avfilter_graph_parse_ptr()" function defined in
9391 libavfilter/avfilter.h.
9392 A filterchain consists of a sequence of connected filters, each one
9394 connected to the previous one in the sequence. A filterchain is
9395 represented by a list of ","-separated filter descriptions.
9396 A filtergraph consists of a sequence of filterchains. A sequence of
9398 filterchains is represented by a list of ";"-separated filterchain
9399 descriptions.
9400 A filter is represented by a string of the form:
9402 [in_link_1]...[in_link_N]filter_name@id=arguments[out_link_1]...[out_link_M]
9403 filter_name is the name of the filter class of which the described
9405 filter is an instance of, and has to be the name of one of the filter
9406 classes registered in the program optionally followed by "@id". The
9407 name of the filter class is optionally followed by a string
9408 "=arguments".
9409 arguments is a string which contains the parameters used to initialize
9411 the filter instance. It may have one of two forms:
9412 • A ':'-separated list of key=value pairs.
9414 • A ':'-separated list of value. In this case, the keys are assumed
9416 to be the option names in the order they are declared. E.g. the
9417 "fade" filter declares three options in this order -- type,
9418 start_frame and nb_frames. Then the parameter list in:0:30 means
9419 that the value in is assigned to the option type, 0 to start_frame
9420 and 30 to nb_frames.
9421 • A ':'-separated list of mixed direct value and long key=value
9423 pairs. The direct value must precede the key=value pairs, and
9424 follow the same constraints order of the previous point. The
9425 following key=value pairs can be set in any preferred order.
9426 If the option value itself is a list of items (e.g. the "format" filter
9428 takes a list of pixel formats), the items in the list are usually
9429 separated by |.
9430 The list of arguments can be quoted using the character ' as initial
9432 and ending mark, and the character \ for escaping the characters within
9433 the quoted text; otherwise the argument string is considered terminated
9434 when the next special character (belonging to the set []=;,) is
9435 encountered.
9436 The name and arguments of the filter are optionally preceded and
9438 followed by a list of link labels. A link label allows one to name a
9439 link and associate it to a filter output or input pad. The preceding
9440 labels in_link_1 ... in_link_N, are associated to the filter input
9441 pads, the following labels out_link_1 ... out_link_M, are associated to
9442 the output pads.
9443 When two link labels with the same name are found in the filtergraph, a
9445 link between the corresponding input and output pad is created.
9446 If an output pad is not labelled, it is linked by default to the first
9448 unlabelled input pad of the next filter in the filterchain. For
9449 example in the filterchain
9450 nullsrc, split[L1], [L2]overlay, nullsink
9452 the split filter instance has two output pads, and the overlay filter
9454 instance two input pads. The first output pad of split is labelled
9455 "L1", the first input pad of overlay is labelled "L2", and the second
9456 output pad of split is linked to the second input pad of overlay, which
9457 are both unlabelled.
9458 In a filter description, if the input label of the first filter is not
9460 specified, "in" is assumed; if the output label of the last filter is
9461 not specified, "out" is assumed.
9462 In a complete filterchain all the unlabelled filter input and output
9464 pads must be connected. A filtergraph is considered valid if all the
9465 filter input and output pads of all the filterchains are connected.
9466 Libavfilter will automatically insert scale filters where format
9468 conversion is required. It is possible to specify swscale flags for
9469 those automatically inserted scalers by prepending "sws_flags=flags;"
9470 to the filtergraph description.
9471 Here is a BNF description of the filtergraph syntax:
9473 <NAME> ::= sequence of alphanumeric characters and '_'
9475 <FILTER_NAME> ::= <NAME>["@"<NAME>]
9476 <LINKLABEL> ::= "[" <NAME> "]"
9477 <LINKLABELS> ::= <LINKLABEL> [<LINKLABELS>]
9478 <FILTER_ARGUMENTS> ::= sequence of chars (possibly quoted)
9479 <FILTER> ::= [<LINKLABELS>] <FILTER_NAME> ["=" <FILTER_ARGUMENTS>] [<LINKLABELS>]
9480 <FILTERCHAIN> ::= <FILTER> [,<FILTERCHAIN>]
9481 <FILTERGRAPH> ::= [sws_flags=<flags>;] <FILTERCHAIN> [;<FILTERGRAPH>]
9482 Notes on filtergraph escaping
9484 Filtergraph description composition entails several levels of escaping.
9485 See the "Quoting and escaping" section in the ffmpeg-utils(1) manual
9486 for more information about the employed escaping procedure.
9487 A first level escaping affects the content of each filter option value,
9489 which may contain the special character ":" used to separate values, or
9490 one of the escaping characters "\'".
9491 A second level escaping affects the whole filter description, which may
9493 contain the escaping characters "\'" or the special characters "[],;"
9494 used by the filtergraph description.
9495 Finally, when you specify a filtergraph on a shell commandline, you
9497 need to perform a third level escaping for the shell special characters
9498 contained within it.
9499 For example, consider the following string to be embedded in the
9501 drawtext filter description text value:
9502 this is a 'string': may contain one, or more, special characters
9504 This string contains the "'" special escaping character, and the ":"
9506 special character, so it needs to be escaped in this way:
9507 text=this is a \'string\'\: may contain one, or more, special characters
9509 A second level of escaping is required when embedding the filter
9511 description in a filtergraph description, in order to escape all the
9512 filtergraph special characters. Thus the example above becomes:
9513 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
9515 (note that in addition to the "\'" escaping special characters, also
9517 "," needs to be escaped).
9518 Finally an additional level of escaping is needed when writing the
9520 filtergraph description in a shell command, which depends on the
9521 escaping rules of the adopted shell. For example, assuming that "\" is
9522 special and needs to be escaped with another "\", the previous string
9523 will finally result in:
9524 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
9526
9528 Some filters support a generic enable option. For the filters
9529 supporting timeline editing, this option can be set to an expression
9530 which is evaluated before sending a frame to the filter. If the
9531 evaluation is non-zero, the filter will be enabled, otherwise the frame
9532 will be sent unchanged to the next filter in the filtergraph.
9533 The expression accepts the following values:
9535 t timestamp expressed in seconds, NAN if the input timestamp is
9537 unknown
9538 n sequential number of the input frame, starting from 0
9540 pos the position in the file of the input frame, NAN if unknown
9542 w
9544 h width and height of the input frame if video
9545 Additionally, these filters support an enable command that can be used
9547 to re-define the expression.
9548 Like any other filtering option, the enable option follows the same
9550 rules.
9551 For example, to enable a blur filter (smartblur) from 10 seconds to 3
9553 minutes, and a curves filter starting at 3 seconds:
9554 smartblur = enable='between(t,10,3*60)',
9556 curves = enable='gte(t,3)' : preset=cross_process
9557 See "ffmpeg -filters" to view which filters have timeline support.
9559
9561 Some options can be changed during the operation of the filter using a
9562 command. These options are marked 'T' on the output of ffmpeg -h
9563 filter=<name of filter>. The name of the command is the name of the
9564 option and the argument is the new value.
9565
9567 Some filters with several inputs support a common set of options.
9568 These options can only be set by name, not with the short notation.
9569 eof_action
9571 The action to take when EOF is encountered on the secondary input;
9572 it accepts one of the following values:
9573 repeat
9575 Repeat the last frame (the default).
9576 endall
9578 End both streams.
9579 pass
9581 Pass the main input through.
9582 shortest
9584 If set to 1, force the output to terminate when the shortest input
9585 terminates. Default value is 0.
9586 repeatlast
9588 If set to 1, force the filter to extend the last frame of secondary
9589 streams until the end of the primary stream. A value of 0 disables
9590 this behavior. Default value is 1.
9591
9593 When you configure your FFmpeg build, you can disable any of the
9594 existing filters using "--disable-filters". The configure output will
9595 show the audio filters included in your build.
9596 Below is a description of the currently available audio filters.
9598 acompressor
9600 A compressor is mainly used to reduce the dynamic range of a signal.
9601 Especially modern music is mostly compressed at a high ratio to improve
9602 the overall loudness. It's done to get the highest attention of a
9603 listener, "fatten" the sound and bring more "power" to the track. If a
9604 signal is compressed too much it may sound dull or "dead" afterwards or
9605 it may start to "pump" (which could be a powerful effect but can also
9606 destroy a track completely). The right compression is the key to reach
9607 a professional sound and is the high art of mixing and mastering.
9608 Because of its complex settings it may take a long time to get the
9609 right feeling for this kind of effect.
9610 Compression is done by detecting the volume above a chosen level
9612 "threshold" and dividing it by the factor set with "ratio". So if you
9613 set the threshold to -12dB and your signal reaches -6dB a ratio of 2:1
9614 will result in a signal at -9dB. Because an exact manipulation of the
9615 signal would cause distortion of the waveform the reduction can be
9616 levelled over the time. This is done by setting "Attack" and "Release".
9617 "attack" determines how long the signal has to rise above the threshold
9618 before any reduction will occur and "release" sets the time the signal
9619 has to fall below the threshold to reduce the reduction again. Shorter
9620 signals than the chosen attack time will be left untouched. The
9621 overall reduction of the signal can be made up afterwards with the
9622 "makeup" setting. So compressing the peaks of a signal about 6dB and
9623 raising the makeup to this level results in a signal twice as loud than
9624 the source. To gain a softer entry in the compression the "knee"
9625 flattens the hard edge at the threshold in the range of the chosen
9626 decibels.
9627 The filter accepts the following options:
9629 level_in
9631 Set input gain. Default is 1. Range is between 0.015625 and 64.
9632 mode
9634 Set mode of compressor operation. Can be "upward" or "downward".
9635 Default is "downward".
9636 threshold
9638 If a signal of stream rises above this level it will affect the
9639 gain reduction. By default it is 0.125. Range is between
9640 0.00097563 and 1.
9641 ratio
9643 Set a ratio by which the signal is reduced. 1:2 means that if the
9644 level rose 4dB above the threshold, it will be only 2dB above after
9645 the reduction. Default is 2. Range is between 1 and 20.
9646 attack
9648 Amount of milliseconds the signal has to rise above the threshold
9649 before gain reduction starts. Default is 20. Range is between 0.01
9650 and 2000.
9651 release
9653 Amount of milliseconds the signal has to fall below the threshold
9654 before reduction is decreased again. Default is 250. Range is
9655 between 0.01 and 9000.
9656 makeup
9658 Set the amount by how much signal will be amplified after
9659 processing. Default is 1. Range is from 1 to 64.
9660 knee
9662 Curve the sharp knee around the threshold to enter gain reduction
9663 more softly. Default is 2.82843. Range is between 1 and 8.
9664 link
9666 Choose if the "average" level between all channels of input stream
9667 or the louder("maximum") channel of input stream affects the
9668 reduction. Default is "average".
9669 detection
9671 Should the exact signal be taken in case of "peak" or an RMS one in
9672 case of "rms". Default is "rms" which is mostly smoother.
9673 mix How much to use compressed signal in output. Default is 1. Range
9675 is between 0 and 1.
9676 Commands
9678 This filter supports the all above options as commands.
9680 acontrast
9682 Simple audio dynamic range compression/expansion filter.
9683 The filter accepts the following options:
9685 contrast
9687 Set contrast. Default is 33. Allowed range is between 0 and 100.
9688 acopy
9690 Copy the input audio source unchanged to the output. This is mainly
9691 useful for testing purposes.
9692 acrossfade
9694 Apply cross fade from one input audio stream to another input audio
9695 stream. The cross fade is applied for specified duration near the end
9696 of first stream.
9697 The filter accepts the following options:
9699 nb_samples, ns
9701 Specify the number of samples for which the cross fade effect has
9702 to last. At the end of the cross fade effect the first input audio
9703 will be completely silent. Default is 44100.
9704 duration, d
9706 Specify the duration of the cross fade effect. See the Time
9707 duration section in the ffmpeg-utils(1) manual for the accepted
9708 syntax. By default the duration is determined by nb_samples. If
9709 set this option is used instead of nb_samples.
9710 overlap, o
9712 Should first stream end overlap with second stream start. Default
9713 is enabled.
9714 curve1
9716 Set curve for cross fade transition for first stream.
9717 curve2
9719 Set curve for cross fade transition for second stream.
9720 For description of available curve types see afade filter
9722 description.
9723 Examples
9725 • Cross fade from one input to another:
9727 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
9729 • Cross fade from one input to another but without overlapping:
9731 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
9733 acrossover
9735 Split audio stream into several bands.
9736 This filter splits audio stream into two or more frequency ranges.
9738 Summing all streams back will give flat output.
9739 The filter accepts the following options:
9741 split
9743 Set split frequencies. Those must be positive and increasing.
9744 order
9746 Set filter order for each band split. This controls filter roll-off
9747 or steepness of filter transfer function. Available values are:
9748 2nd 12 dB per octave.
9750 4th 24 dB per octave.
9752 6th 36 dB per octave.
9754 8th 48 dB per octave.
9756 10th
9758 60 dB per octave.
9759 12th
9761 72 dB per octave.
9762 14th
9764 84 dB per octave.
9765 16th
9767 96 dB per octave.
9768 18th
9770 108 dB per octave.
9771 20th
9773 120 dB per octave.
9774 Default is 4th.
9776 level
9778 Set input gain level. Allowed range is from 0 to 1. Default value
9779 is 1.
9780 gains
9782 Set output gain for each band. Default value is 1 for all bands.
9783 precision
9785 Set which precision to use when processing samples.
9786 auto
9788 Auto pick internal sample format depending on other filters.
9789 float
9791 Always use single-floating point precision sample format.
9792 double
9794 Always use double-floating point precision sample format.
9795 Default value is "auto".
9797 Examples
9799 • Split input audio stream into two bands (low and high) with split
9801 frequency of 1500 Hz, each band will be in separate stream:
9802 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
9804 • Same as above, but with higher filter order:
9806 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
9808 • Same as above, but also with additional middle band (frequencies
9810 between 1500 and 8000):
9811 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
9813 acrusher
9815 Reduce audio bit resolution.
9816 This filter is bit crusher with enhanced functionality. A bit crusher
9818 is used to audibly reduce number of bits an audio signal is sampled
9819 with. This doesn't change the bit depth at all, it just produces the
9820 effect. Material reduced in bit depth sounds more harsh and "digital".
9821 This filter is able to even round to continuous values instead of
9822 discrete bit depths. Additionally it has a D/C offset which results in
9823 different crushing of the lower and the upper half of the signal. An
9824 Anti-Aliasing setting is able to produce "softer" crushing sounds.
9825 Another feature of this filter is the logarithmic mode. This setting
9827 switches from linear distances between bits to logarithmic ones. The
9828 result is a much more "natural" sounding crusher which doesn't gate low
9829 signals for example. The human ear has a logarithmic perception, so
9830 this kind of crushing is much more pleasant. Logarithmic crushing is
9831 also able to get anti-aliased.
9832 The filter accepts the following options:
9834 level_in
9836 Set level in.
9837 level_out
9839 Set level out.
9840 bits
9842 Set bit reduction.
9843 mix Set mixing amount.
9845 mode
9847 Can be linear: "lin" or logarithmic: "log".
9848 dc Set DC.
9850 aa Set anti-aliasing.
9852 samples
9854 Set sample reduction.
9855 lfo Enable LFO. By default disabled.
9857 lforange
9859 Set LFO range.
9860 lforate
9862 Set LFO rate.
9863 Commands
9865 This filter supports the all above options as commands.
9867 acue
9869 Delay audio filtering until a given wallclock timestamp. See the cue
9870 filter.
9871 adeclick
9873 Remove impulsive noise from input audio.
9874 Samples detected as impulsive noise are replaced by interpolated
9876 samples using autoregressive modelling.
9877 window, w
9879 Set window size, in milliseconds. Allowed range is from 10 to 100.
9880 Default value is 55 milliseconds. This sets size of window which
9881 will be processed at once.
9882 overlap, o
9884 Set window overlap, in percentage of window size. Allowed range is
9885 from 50 to 95. Default value is 75 percent. Setting this to a very
9886 high value increases impulsive noise removal but makes whole
9887 process much slower.
9888 arorder, a
9890 Set autoregression order, in percentage of window size. Allowed
9891 range is from 0 to 25. Default value is 2 percent. This option also
9892 controls quality of interpolated samples using neighbour good
9893 samples.
9894 threshold, t
9896 Set threshold value. Allowed range is from 1 to 100. Default value
9897 is 2. This controls the strength of impulsive noise which is going
9898 to be removed. The lower value, the more samples will be detected
9899 as impulsive noise.
9900 burst, b
9902 Set burst fusion, in percentage of window size. Allowed range is 0
9903 to 10. Default value is 2. If any two samples detected as noise
9904 are spaced less than this value then any sample between those two
9905 samples will be also detected as noise.
9906 method, m
9908 Set overlap method.
9909 It accepts the following values:
9911 add, a
9913 Select overlap-add method. Even not interpolated samples are
9914 slightly changed with this method.
9915 save, s
9917 Select overlap-save method. Not interpolated samples remain
9918 unchanged.
9919 Default value is "a".
9921 adeclip
9923 Remove clipped samples from input audio.
9924 Samples detected as clipped are replaced by interpolated samples using
9926 autoregressive modelling.
9927 window, w
9929 Set window size, in milliseconds. Allowed range is from 10 to 100.
9930 Default value is 55 milliseconds. This sets size of window which
9931 will be processed at once.
9932 overlap, o
9934 Set window overlap, in percentage of window size. Allowed range is
9935 from 50 to 95. Default value is 75 percent.
9936 arorder, a
9938 Set autoregression order, in percentage of window size. Allowed
9939 range is from 0 to 25. Default value is 8 percent. This option also
9940 controls quality of interpolated samples using neighbour good
9941 samples.
9942 threshold, t
9944 Set threshold value. Allowed range is from 1 to 100. Default value
9945 is 10. Higher values make clip detection less aggressive.
9946 hsize, n
9948 Set size of histogram used to detect clips. Allowed range is from
9949 100 to 9999. Default value is 1000. Higher values make clip
9950 detection less aggressive.
9951 method, m
9953 Set overlap method.
9954 It accepts the following values:
9956 add, a
9958 Select overlap-add method. Even not interpolated samples are
9959 slightly changed with this method.
9960 save, s
9962 Select overlap-save method. Not interpolated samples remain
9963 unchanged.
9964 Default value is "a".
9966 adecorrelate
9968 Apply decorrelation to input audio stream.
9969 The filter accepts the following options:
9971 stages
9973 Set decorrelation stages of filtering. Allowed range is from 1 to
9974 16. Default value is 6.
9975 seed
9977 Set random seed used for setting delay in samples across channels.
9978 adelay
9980 Delay one or more audio channels.
9981 Samples in delayed channel are filled with silence.
9983 The filter accepts the following option:
9985 delays
9987 Set list of delays in milliseconds for each channel separated by
9988 '|'. Unused delays will be silently ignored. If number of given
9989 delays is smaller than number of channels all remaining channels
9990 will not be delayed. If you want to delay exact number of samples,
9991 append 'S' to number. If you want instead to delay in seconds,
9992 append 's' to number.
9993 all Use last set delay for all remaining channels. By default is
9995 disabled. This option if enabled changes how option "delays" is
9996 interpreted.
9997 Examples
9999 • Delay first channel by 1.5 seconds, the third channel by 0.5
10001 seconds and leave the second channel (and any other channels that
10002 may be present) unchanged.
10003 adelay=1500|0|500
10005 • Delay second channel by 500 samples, the third channel by 700
10007 samples and leave the first channel (and any other channels that
10008 may be present) unchanged.
10009 adelay=0|500S|700S
10011 • Delay all channels by same number of samples:
10013 adelay=delays=64S:all=1
10015 adenorm
10017 Remedy denormals in audio by adding extremely low-level noise.
10018 This filter shall be placed before any filter that can produce
10020 denormals.
10021 A description of the accepted parameters follows.
10023 level
10025 Set level of added noise in dB. Default is "-351". Allowed range
10026 is from -451 to -90.
10027 type
10029 Set type of added noise.
10030 dc Add DC signal.
10032 ac Add AC signal.
10034 square
10036 Add square signal.
10037 pulse
10039 Add pulse signal.
10040 Default is "dc".
10042 Commands
10044 This filter supports the all above options as commands.
10046 aderivative, aintegral
10048 Compute derivative/integral of audio stream.
10049 Applying both filters one after another produces original audio.
10051 adynamicequalizer
10053 Apply dynamic equalization to input audio stream.
10054 A description of the accepted options follows.
10056 threshold
10058 Set the detection threshold used to trigger equalization.
10059 Threshold detection is using bandpass filter. Default value is 0.
10060 Allowed range is from 0 to 100.
10061 dfrequency
10063 Set the detection frequency in Hz used for bandpass filter used to
10064 trigger equalization. Default value is 1000 Hz. Allowed range is
10065 between 2 and 1000000 Hz.
10066 dqfactor
10068 Set the detection resonance factor for bandpass filter used to
10069 trigger equalization. Default value is 1. Allowed range is from
10070 0.001 to 1000.
10071 tfrequency
10073 Set the target frequency of equalization filter. Default value is
10074 1000 Hz. Allowed range is between 2 and 1000000 Hz.
10075 tqfactor
10077 Set the target resonance factor for target equalization filter.
10078 Default value is 1. Allowed range is from 0.001 to 1000.
10079 attack
10081 Set the amount of milliseconds the signal from detection has to
10082 rise above the detection threshold before equalization starts.
10083 Default is 20. Allowed range is between 1 and 2000.
10084 release
10086 Set the amount of milliseconds the signal from detection has to
10087 fall below the detection threshold before equalization ends.
10088 Default is 200. Allowed range is between 1 and 2000.
10089 knee
10091 Curve the sharp knee around the detection threshold to calculate
10092 equalization gain more softly. Default is 1. Allowed range is
10093 between 0 and 8.
10094 ratio
10096 Set the ratio by which the equalization gain is raised. Default is
10097 1. Allowed range is between 1 and 20.
10098 makeup
10100 Set the makeup offset in dB by which the equalization gain is
10101 raised. Default is 0. Allowed range is between 0 and 30.
10102 range
10104 Set the max allowed cut/boost amount in dB. Default is 0. Allowed
10105 range is from 0 to 200.
10106 slew
10108 Set the slew factor. Default is 1. Allowed range is from 1 to 200.
10109 mode
10111 Set the mode of filter operation, can be one of the following:
10112 listen
10114 Output only isolated bandpass signal.
10115 cut Cut frequencies above detection threshold.
10117 boost
10119 Boost frequencies bellow detection threshold.
10120 Default mode is cut.
10122 tftype
10124 Set the type of target filter, can be one of the following:
10125 bell
10127 lowshelf
10128 highshelf
10129 Default type is bell.
10131 Commands
10133 This filter supports the all above options as commands.
10135 adynamicsmooth
10137 Apply dynamic smoothing to input audio stream.
10138 A description of the accepted options follows.
10140 sensitivity
10142 Set an amount of sensitivity to frequency fluctations. Default is
10143 2. Allowed range is from 0 to 1e+06.
10144 basefreq
10146 Set a base frequency for smoothing. Default value is 22050.
10147 Allowed range is from 2 to 1e+06.
10148 Commands
10150 This filter supports the all above options as commands.
10152 aecho
10154 Apply echoing to the input audio.
10155 Echoes are reflected sound and can occur naturally amongst mountains
10157 (and sometimes large buildings) when talking or shouting; digital echo
10158 effects emulate this behaviour and are often used to help fill out the
10159 sound of a single instrument or vocal. The time difference between the
10160 original signal and the reflection is the "delay", and the loudness of
10161 the reflected signal is the "decay". Multiple echoes can have
10162 different delays and decays.
10163 A description of the accepted parameters follows.
10165 in_gain
10167 Set input gain of reflected signal. Default is 0.6.
10168 out_gain
10170 Set output gain of reflected signal. Default is 0.3.
10171 delays
10173 Set list of time intervals in milliseconds between original signal
10174 and reflections separated by '|'. Allowed range for each "delay" is
10175 "(0 - 90000.0]". Default is 1000.
10176 decays
10178 Set list of loudness of reflected signals separated by '|'.
10179 Allowed range for each "decay" is "(0 - 1.0]". Default is 0.5.
10180 Examples
10182 • Make it sound as if there are twice as many instruments as are
10184 actually playing:
10185 aecho=0.8:0.88:60:0.4
10187 • If delay is very short, then it sounds like a (metallic) robot
10189 playing music:
10190 aecho=0.8:0.88:6:0.4
10192 • A longer delay will sound like an open air concert in the
10194 mountains:
10195 aecho=0.8:0.9:1000:0.3
10197 • Same as above but with one more mountain:
10199 aecho=0.8:0.9:1000|1800:0.3|0.25
10201 aemphasis
10203 Audio emphasis filter creates or restores material directly taken from
10204 LPs or emphased CDs with different filter curves. E.g. to store music
10205 on vinyl the signal has to be altered by a filter first to even out the
10206 disadvantages of this recording medium. Once the material is played
10207 back the inverse filter has to be applied to restore the distortion of
10208 the frequency response.
10209 The filter accepts the following options:
10211 level_in
10213 Set input gain.
10214 level_out
10216 Set output gain.
10217 mode
10219 Set filter mode. For restoring material use "reproduction" mode,
10220 otherwise use "production" mode. Default is "reproduction" mode.
10221 type
10223 Set filter type. Selects medium. Can be one of the following:
10224 col select Columbia.
10226 emi select EMI.
10228 bsi select BSI (78RPM).
10230 riaa
10232 select RIAA.
10233 cd select Compact Disc (CD).
10235 50fm
10237 select 50Xs (FM).
10238 75fm
10240 select 75Xs (FM).
10241 50kf
10243 select 50Xs (FM-KF).
10244 75kf
10246 select 75Xs (FM-KF).
10247 Commands
10249 This filter supports the all above options as commands.
10251 aeval
10253 Modify an audio signal according to the specified expressions.
10254 This filter accepts one or more expressions (one for each channel),
10256 which are evaluated and used to modify a corresponding audio signal.
10257 It accepts the following parameters:
10259 exprs
10261 Set the '|'-separated expressions list for each separate channel.
10262 If the number of input channels is greater than the number of
10263 expressions, the last specified expression is used for the
10264 remaining output channels.
10265 channel_layout, c
10267 Set output channel layout. If not specified, the channel layout is
10268 specified by the number of expressions. If set to same, it will use
10269 by default the same input channel layout.
10270 Each expression in exprs can contain the following constants and
10272 functions:
10273 ch channel number of the current expression
10275 n number of the evaluated sample, starting from 0
10277 s sample rate
10279 t time of the evaluated sample expressed in seconds
10281 nb_in_channels
10283 nb_out_channels
10284 input and output number of channels
10285 val(CH)
10287 the value of input channel with number CH
10288 Note: this filter is slow. For faster processing you should use a
10290 dedicated filter.
10291 Examples
10293 • Half volume:
10295 aeval=val(ch)/2:c=same
10297 • Invert phase of the second channel:
10299 aeval=val(0)|-val(1)
10301 aexciter
10303 An exciter is used to produce high sound that is not present in the
10304 original signal. This is done by creating harmonic distortions of the
10305 signal which are restricted in range and added to the original signal.
10306 An Exciter raises the upper end of an audio signal without simply
10307 raising the higher frequencies like an equalizer would do to create a
10308 more "crisp" or "brilliant" sound.
10309 The filter accepts the following options:
10311 level_in
10313 Set input level prior processing of signal. Allowed range is from
10314 0 to 64. Default value is 1.
10315 level_out
10317 Set output level after processing of signal. Allowed range is from
10318 0 to 64. Default value is 1.
10319 amount
10321 Set the amount of harmonics added to original signal. Allowed
10322 range is from 0 to 64. Default value is 1.
10323 drive
10325 Set the amount of newly created harmonics. Allowed range is from
10326 0.1 to 10. Default value is 8.5.
10327 blend
10329 Set the octave of newly created harmonics. Allowed range is from
10330 -10 to 10. Default value is 0.
10331 freq
10333 Set the lower frequency limit of producing harmonics in Hz.
10334 Allowed range is from 2000 to 12000 Hz. Default is 7500 Hz.
10335 ceil
10337 Set the upper frequency limit of producing harmonics. Allowed
10338 range is from 9999 to 20000 Hz. If value is lower than 10000 Hz no
10339 limit is applied.
10340 listen
10342 Mute the original signal and output only added harmonics. By
10343 default is disabled.
10344 Commands
10346 This filter supports the all above options as commands.
10348 afade
10350 Apply fade-in/out effect to input audio.
10351 A description of the accepted parameters follows.
10353 type, t
10355 Specify the effect type, can be either "in" for fade-in, or "out"
10356 for a fade-out effect. Default is "in".
10357 start_sample, ss
10359 Specify the number of the start sample for starting to apply the
10360 fade effect. Default is 0.
10361 nb_samples, ns
10363 Specify the number of samples for which the fade effect has to
10364 last. At the end of the fade-in effect the output audio will have
10365 the same volume as the input audio, at the end of the fade-out
10366 transition the output audio will be silence. Default is 44100.
10367 start_time, st
10369 Specify the start time of the fade effect. Default is 0. The value
10370 must be specified as a time duration; see the Time duration section
10371 in the ffmpeg-utils(1) manual for the accepted syntax. If set this
10372 option is used instead of start_sample.
10373 duration, d
10375 Specify the duration of the fade effect. See the Time duration
10376 section in the ffmpeg-utils(1) manual for the accepted syntax. At
10377 the end of the fade-in effect the output audio will have the same
10378 volume as the input audio, at the end of the fade-out transition
10379 the output audio will be silence. By default the duration is
10380 determined by nb_samples. If set this option is used instead of
10381 nb_samples.
10382 curve
10384 Set curve for fade transition.
10385 It accepts the following values:
10387 tri select triangular, linear slope (default)
10389 qsin
10391 select quarter of sine wave
10392 hsin
10394 select half of sine wave
10395 esin
10397 select exponential sine wave
10398 log select logarithmic
10400 ipar
10402 select inverted parabola
10403 qua select quadratic
10405 cub select cubic
10407 squ select square root
10409 cbr select cubic root
10411 par select parabola
10413 exp select exponential
10415 iqsin
10417 select inverted quarter of sine wave
10418 ihsin
10420 select inverted half of sine wave
10421 dese
10423 select double-exponential seat
10424 desi
10426 select double-exponential sigmoid
10427 losi
10429 select logistic sigmoid
10430 sinc
10432 select sine cardinal function
10433 isinc
10435 select inverted sine cardinal function
10436 nofade
10438 no fade applied
10439 Commands
10441 This filter supports the all above options as commands.
10443 Examples
10445 • Fade in first 15 seconds of audio:
10447 afade=t=in:ss=0:d=15
10449 • Fade out last 25 seconds of a 900 seconds audio:
10451 afade=t=out:st=875:d=25
10453 afftdn
10455 Denoise audio samples with FFT.
10456 A description of the accepted parameters follows.
10458 noise_reduction, nr
10460 Set the noise reduction in dB, allowed range is 0.01 to 97.
10461 Default value is 12 dB.
10462 noise_floor, nf
10464 Set the noise floor in dB, allowed range is -80 to -20. Default
10465 value is -50 dB.
10466 noise_type, nt
10468 Set the noise type.
10469 It accepts the following values:
10471 white, w
10473 Select white noise.
10474 vinyl, v
10476 Select vinyl noise.
10477 shellac, s
10479 Select shellac noise.
10480 custom, c
10482 Select custom noise, defined in "bn" option.
10483 Default value is white noise.
10485 band_noise, bn
10487 Set custom band noise profile for every one of 15 bands. Bands are
10488 separated by ' ' or '|'.
10489 residual_floor, rf
10491 Set the residual floor in dB, allowed range is -80 to -20. Default
10492 value is -38 dB.
10493 track_noise, tn
10495 Enable noise floor tracking. By default is disabled. With this
10496 enabled, noise floor is automatically adjusted.
10497 track_residual, tr
10499 Enable residual tracking. By default is disabled.
10500 output_mode, om
10502 Set the output mode.
10503 It accepts the following values:
10505 input, i
10507 Pass input unchanged.
10508 output, o
10510 Pass noise filtered out.
10511 noise, n
10513 Pass only noise.
10514 Default value is output.
10516 adaptivity, ad
10518 Set the adaptivity factor, used how fast to adapt gains adjustments
10519 per each frequency bin. Value 0 enables instant adaptation, while
10520 higher values react much slower. Allowed range is from 0 to 1.
10521 Default value is 0.5.
10522 floor_offset, fo
10524 Set the noise floor offset factor. This option is used to adjust
10525 offset applied to measured noise floor. It is only effective when
10526 noise floor tracking is enabled. Allowed range is from -2.0 to
10527 2.0. Default value is 1.0.
10528 noise_link, nl
10530 Set the noise link used for multichannel audio.
10531 It accepts the following values:
10533 none
10535 Use unchanged channel's noise floor.
10536 min Use measured min noise floor of all channels.
10538 max Use measured max noise floor of all channels.
10540 average
10542 Use measured average noise floor of all channels.
10543 Default value is min.
10545 band_multiplier, bm
10547 Set the band multiplier factor, used how much to spread bands
10548 across frequency bins. Allowed range is from 0.2 to 5. Default
10549 value is 1.25.
10550 sample_noise, sn
10552 Toggle capturing and measurement of noise profile from input audio.
10553 It accepts the following values:
10555 start, begin
10557 Start sample noise capture.
10558 stop, end
10560 Stop sample noise capture and measure new noise band profile.
10561 Default value is "none".
10563 gain_smooth, gs
10565 Set gain smooth spatial radius, used to smooth gains applied to
10566 each frequency bin. Useful to reduce random music noise artefacts.
10567 Higher values increases smoothing of gains. Allowed range is from
10568 0 to 50. Default value is 0.
10569 Commands
10571 This filter supports the some above mentioned options as commands.
10573 Examples
10575 • Reduce white noise by 10dB, and use previously measured noise floor
10577 of -40dB:
10578 afftdn=nr=10:nf=-40
10580 • Reduce white noise by 10dB, also set initial noise floor to -80dB
10582 and enable automatic tracking of noise floor so noise floor will
10583 gradually change during processing:
10584 afftdn=nr=10:nf=-80:tn=1
10586 • Reduce noise by 20dB, using noise floor of -40dB and using commands
10588 to take noise profile of first 0.4 seconds of input audio:
10589 asendcmd=0.0 afftdn sn start,asendcmd=0.4 afftdn sn stop,afftdn=nr=20:nf=-40
10591 afftfilt
10593 Apply arbitrary expressions to samples in frequency domain.
10594 real
10596 Set frequency domain real expression for each separate channel
10597 separated by '|'. Default is "re". If the number of input channels
10598 is greater than the number of expressions, the last specified
10599 expression is used for the remaining output channels.
10600 imag
10602 Set frequency domain imaginary expression for each separate channel
10603 separated by '|'. Default is "im".
10604 Each expression in real and imag can contain the following
10606 constants and functions:
10607 sr sample rate
10609 b current frequency bin number
10611 nb number of available bins
10613 ch channel number of the current expression
10615 chs number of channels
10617 pts current frame pts
10619 re current real part of frequency bin of current channel
10621 im current imaginary part of frequency bin of current channel
10623 real(b, ch)
10625 Return the value of real part of frequency bin at location
10626 (bin,channel)
10627 imag(b, ch)
10629 Return the value of imaginary part of frequency bin at location
10630 (bin,channel)
10631 win_size
10633 Set window size. Allowed range is from 16 to 131072. Default is
10634 4096
10635 win_func
10637 Set window function.
10638 It accepts the following values:
10640 rect
10642 bartlett
10643 hann, hanning
10644 hamming
10645 blackman
10646 welch
10647 flattop
10648 bharris
10649 bnuttall
10650 bhann
10651 sine
10652 nuttall
10653 lanczos
10654 gauss
10655 tukey
10656 dolph
10657 cauchy
10658 parzen
10659 poisson
10660 bohman
10661 Default is "hann".
10663 overlap
10665 Set window overlap. If set to 1, the recommended overlap for
10666 selected window function will be picked. Default is 0.75.
10667 Examples
10669 • Leave almost only low frequencies in audio:
10671 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
10673 • Apply robotize effect:
10675 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
10677 • Apply whisper effect:
10679 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"
10681 afir
10683 Apply an arbitrary Finite Impulse Response filter.
10684 This filter is designed for applying long FIR filters, up to 60 seconds
10686 long.
10687 It can be used as component for digital crossover filters, room
10689 equalization, cross talk cancellation, wavefield synthesis,
10690 auralization, ambiophonics, ambisonics and spatialization.
10691 This filter uses the streams higher than first one as FIR coefficients.
10693 If the non-first stream holds a single channel, it will be used for all
10694 input channels in the first stream, otherwise the number of channels in
10695 the non-first stream must be same as the number of channels in the
10696 first stream.
10697 It accepts the following parameters:
10699 dry Set dry gain. This sets input gain.
10701 wet Set wet gain. This sets final output gain.
10703 length
10705 Set Impulse Response filter length. Default is 1, which means whole
10706 IR is processed.
10707 gtype
10709 Enable applying gain measured from power of IR.
10710 Set which approach to use for auto gain measurement.
10712 none
10714 Do not apply any gain.
10715 peak
10717 select peak gain, very conservative approach. This is default
10718 value.
10719 dc select DC gain, limited application.
10721 gn select gain to noise approach, this is most popular one.
10723 irgain
10725 Set gain to be applied to IR coefficients before filtering.
10726 Allowed range is 0 to 1. This gain is applied after any gain
10727 applied with gtype option.
10728 irfmt
10730 Set format of IR stream. Can be "mono" or "input". Default is
10731 "input".
10732 maxir
10734 Set max allowed Impulse Response filter duration in seconds.
10735 Default is 30 seconds. Allowed range is 0.1 to 60 seconds.
10736 response
10738 Show IR frequency response, magnitude(magenta), phase(green) and
10739 group delay(yellow) in additional video stream. By default it is
10740 disabled.
10741 channel
10743 Set for which IR channel to display frequency response. By default
10744 is first channel displayed. This option is used only when response
10745 is enabled.
10746 size
10748 Set video stream size. This option is used only when response is
10749 enabled.
10750 rate
10752 Set video stream frame rate. This option is used only when response
10753 is enabled.
10754 minp
10756 Set minimal partition size used for convolution. Default is 8192.
10757 Allowed range is from 1 to 32768. Lower values decreases latency
10758 at cost of higher CPU usage.
10759 maxp
10761 Set maximal partition size used for convolution. Default is 8192.
10762 Allowed range is from 8 to 32768. Lower values may increase CPU
10763 usage.
10764 nbirs
10766 Set number of input impulse responses streams which will be
10767 switchable at runtime. Allowed range is from 1 to 32. Default is
10768 1.
10769 ir Set IR stream which will be used for convolution, starting from 0,
10771 should always be lower than supplied value by "nbirs" option.
10772 Default is 0. This option can be changed at runtime via commands.
10773 precision
10775 Set which precision to use when processing samples.
10776 auto
10778 Auto pick internal sample format depending on other filters.
10779 float
10781 Always use single-floating point precision sample format.
10782 double
10784 Always use double-floating point precision sample format.
10785 Default value is auto.
10787 Examples
10789 • Apply reverb to stream using mono IR file as second input, complete
10791 command using ffmpeg:
10792 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
10794 aformat
10796 Set output format constraints for the input audio. The framework will
10797 negotiate the most appropriate format to minimize conversions.
10798 It accepts the following parameters:
10800 sample_fmts, f
10802 A '|'-separated list of requested sample formats.
10803 sample_rates, r
10805 A '|'-separated list of requested sample rates.
10806 channel_layouts, cl
10808 A '|'-separated list of requested channel layouts.
10809 See the Channel Layout section in the ffmpeg-utils(1) manual for
10811 the required syntax.
10812 If a parameter is omitted, all values are allowed.
10814 Force the output to either unsigned 8-bit or signed 16-bit stereo
10816 aformat=sample_fmts=u8|s16:channel_layouts=stereo
10818 afreqshift
10820 Apply frequency shift to input audio samples.
10821 The filter accepts the following options:
10823 shift
10825 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
10826 Default value is 0.0.
10827 level
10829 Set output gain applied to final output. Allowed range is from 0.0
10830 to 1.0. Default value is 1.0.
10831 order
10833 Set filter order used for filtering. Allowed range is from 1 to 16.
10834 Default value is 8.
10835 Commands
10837 This filter supports the all above options as commands.
10839 afwtdn
10841 Reduce broadband noise from input samples using Wavelets.
10842 A description of the accepted options follows.
10844 sigma
10846 Set the noise sigma, allowed range is from 0 to 1. Default value
10847 is 0. This option controls strength of denoising applied to input
10848 samples. Most useful way to set this option is via decibels, eg.
10849 -45dB.
10850 levels
10852 Set the number of wavelet levels of decomposition. Allowed range
10853 is from 1 to 12. Default value is 10. Setting this too low make
10854 denoising performance very poor.
10855 wavet
10857 Set wavelet type for decomposition of input frame. They are sorted
10858 by number of coefficients, from lowest to highest. More
10859 coefficients means worse filtering speed, but overall better
10860 quality. Available wavelets are:
10861 sym2
10863 sym4
10864 rbior68
10865 deb10
10866 sym10
10867 coif5
10868 bl3
10869 percent
10870 Set percent of full denoising. Allowed range is from 0 to 100
10871 percent. Default value is 85 percent or partial denoising.
10872 profile
10874 If enabled, first input frame will be used as noise profile. If
10875 first frame samples contain non-noise performance will be very
10876 poor.
10877 adaptive
10879 If enabled, input frames are analyzed for presence of noise. If
10880 noise is detected with high possibility then input frame profile
10881 will be used for processing following frames, until new noise frame
10882 is detected.
10883 samples
10885 Set size of single frame in number of samples. Allowed range is
10886 from 512 to 65536. Default frame size is 8192 samples.
10887 softness
10889 Set softness applied inside thresholding function. Allowed range is
10890 from 0 to 10. Default softness is 1.
10891 Commands
10893 This filter supports the all above options as commands.
10895 agate
10897 A gate is mainly used to reduce lower parts of a signal. This kind of
10898 signal processing reduces disturbing noise between useful signals.
10899 Gating is done by detecting the volume below a chosen level threshold
10901 and dividing it by the factor set with ratio. The bottom of the noise
10902 floor is set via range. Because an exact manipulation of the signal
10903 would cause distortion of the waveform the reduction can be levelled
10904 over time. This is done by setting attack and release.
10905 attack determines how long the signal has to fall below the threshold
10907 before any reduction will occur and release sets the time the signal
10908 has to rise above the threshold to reduce the reduction again. Shorter
10909 signals than the chosen attack time will be left untouched.
10910 level_in
10912 Set input level before filtering. Default is 1. Allowed range is
10913 from 0.015625 to 64.
10914 mode
10916 Set the mode of operation. Can be "upward" or "downward". Default
10917 is "downward". If set to "upward" mode, higher parts of signal will
10918 be amplified, expanding dynamic range in upward direction.
10919 Otherwise, in case of "downward" lower parts of signal will be
10920 reduced.
10921 range
10923 Set the level of gain reduction when the signal is below the
10924 threshold. Default is 0.06125. Allowed range is from 0 to 1.
10925 Setting this to 0 disables reduction and then filter behaves like
10926 expander.
10927 threshold
10929 If a signal rises above this level the gain reduction is released.
10930 Default is 0.125. Allowed range is from 0 to 1.
10931 ratio
10933 Set a ratio by which the signal is reduced. Default is 2. Allowed
10934 range is from 1 to 9000.
10935 attack
10937 Amount of milliseconds the signal has to rise above the threshold
10938 before gain reduction stops. Default is 20 milliseconds. Allowed
10939 range is from 0.01 to 9000.
10940 release
10942 Amount of milliseconds the signal has to fall below the threshold
10943 before the reduction is increased again. Default is 250
10944 milliseconds. Allowed range is from 0.01 to 9000.
10945 makeup
10947 Set amount of amplification of signal after processing. Default is
10948 1. Allowed range is from 1 to 64.
10949 knee
10951 Curve the sharp knee around the threshold to enter gain reduction
10952 more softly. Default is 2.828427125. Allowed range is from 1 to 8.
10953 detection
10955 Choose if exact signal should be taken for detection or an RMS like
10956 one. Default is "rms". Can be "peak" or "rms".
10957 link
10959 Choose if the average level between all channels or the louder
10960 channel affects the reduction. Default is "average". Can be
10961 "average" or "maximum".
10962 Commands
10964 This filter supports the all above options as commands.
10966 aiir
10968 Apply an arbitrary Infinite Impulse Response filter.
10969 It accepts the following parameters:
10971 zeros, z
10973 Set B/numerator/zeros/reflection coefficients.
10974 poles, p
10976 Set A/denominator/poles/ladder coefficients.
10977 gains, k
10979 Set channels gains.
10980 dry_gain
10982 Set input gain.
10983 wet_gain
10985 Set output gain.
10986 format, f
10988 Set coefficients format.
10989 ll lattice-ladder function
10991 sf analog transfer function
10993 tf digital transfer function
10995 zp Z-plane zeros/poles, cartesian (default)
10997 pr Z-plane zeros/poles, polar radians
10999 pd Z-plane zeros/poles, polar degrees
11001 sp S-plane zeros/poles
11003 process, r
11005 Set type of processing.
11006 d direct processing
11008 s serial processing
11010 p parallel processing
11012 precision, e
11014 Set filtering precision.
11015 dbl double-precision floating-point (default)
11017 flt single-precision floating-point
11019 i32 32-bit integers
11021 i16 16-bit integers
11023 normalize, n
11025 Normalize filter coefficients, by default is enabled. Enabling it
11026 will normalize magnitude response at DC to 0dB.
11027 mix How much to use filtered signal in output. Default is 1. Range is
11029 between 0 and 1.
11030 response
11032 Show IR frequency response, magnitude(magenta), phase(green) and
11033 group delay(yellow) in additional video stream. By default it is
11034 disabled.
11035 channel
11037 Set for which IR channel to display frequency response. By default
11038 is first channel displayed. This option is used only when response
11039 is enabled.
11040 size
11042 Set video stream size. This option is used only when response is
11043 enabled.
11044 Coefficients in "tf" and "sf" format are separated by spaces and are in
11046 ascending order.
11047 Coefficients in "zp" format are separated by spaces and order of
11049 coefficients doesn't matter. Coefficients in "zp" format are complex
11050 numbers with i imaginary unit.
11051 Different coefficients and gains can be provided for every channel, in
11053 such case use '|' to separate coefficients or gains. Last provided
11054 coefficients will be used for all remaining channels.
11055 Examples
11057 • Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample
11059 rate:
11060 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
11062 • Same as above but in "zp" format:
11064 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
11066 • Apply 3-rd order analog normalized Butterworth low-pass filter,
11068 using analog transfer function format:
11069 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
11071 alimiter
11073 The limiter prevents an input signal from rising over a desired
11074 threshold. This limiter uses lookahead technology to prevent your
11075 signal from distorting. It means that there is a small delay after the
11076 signal is processed. Keep in mind that the delay it produces is the
11077 attack time you set.
11078 The filter accepts the following options:
11080 level_in
11082 Set input gain. Default is 1.
11083 level_out
11085 Set output gain. Default is 1.
11086 limit
11088 Don't let signals above this level pass the limiter. Default is 1.
11089 attack
11091 The limiter will reach its attenuation level in this amount of time
11092 in milliseconds. Default is 5 milliseconds.
11093 release
11095 Come back from limiting to attenuation 1.0 in this amount of
11096 milliseconds. Default is 50 milliseconds.
11097 asc When gain reduction is always needed ASC takes care of releasing to
11099 an average reduction level rather than reaching a reduction of 0 in
11100 the release time.
11101 asc_level
11103 Select how much the release time is affected by ASC, 0 means nearly
11104 no changes in release time while 1 produces higher release times.
11105 level
11107 Auto level output signal. Default is enabled. This normalizes
11108 audio back to 0dB if enabled.
11109 latency
11111 Compensate the delay introduced by using the lookahead buffer set
11112 with attack parameter. Also flush the valid audio data in the
11113 lookahead buffer when the stream hits EOF.
11114 Depending on picked setting it is recommended to upsample input 2x or
11116 4x times with aresample before applying this filter.
11117 allpass
11119 Apply a two-pole all-pass filter with central frequency (in Hz)
11120 frequency, and filter-width width. An all-pass filter changes the
11121 audio's frequency to phase relationship without changing its frequency
11122 to amplitude relationship.
11123 The filter accepts the following options:
11125 frequency, f
11127 Set frequency in Hz.
11128 width_type, t
11130 Set method to specify band-width of filter.
11131 h Hz
11133 q Q-Factor
11135 o octave
11137 s slope
11139 k kHz
11141 width, w
11143 Specify the band-width of a filter in width_type units.
11144 mix, m
11146 How much to use filtered signal in output. Default is 1. Range is
11147 between 0 and 1.
11148 channels, c
11150 Specify which channels to filter, by default all available are
11151 filtered.
11152 normalize, n
11154 Normalize biquad coefficients, by default is disabled. Enabling it
11155 will normalize magnitude response at DC to 0dB.
11156 order, o
11158 Set the filter order, can be 1 or 2. Default is 2.
11159 transform, a
11161 Set transform type of IIR filter.
11162 di
11164 dii
11165 tdi
11166 tdii
11167 latt
11168 svf
11169 zdf
11170 precision, r
11171 Set precison of filtering.
11172 auto
11174 Pick automatic sample format depending on surround filters.
11175 s16 Always use signed 16-bit.
11177 s32 Always use signed 32-bit.
11179 f32 Always use float 32-bit.
11181 f64 Always use float 64-bit.
11183 Commands
11185 This filter supports the following commands:
11187 frequency, f
11189 Change allpass frequency. Syntax for the command is : "frequency"
11190 width_type, t
11192 Change allpass width_type. Syntax for the command is :
11193 "width_type"
11194 width, w
11196 Change allpass width. Syntax for the command is : "width"
11197 mix, m
11199 Change allpass mix. Syntax for the command is : "mix"
11200 aloop
11202 Loop audio samples.
11203 The filter accepts the following options:
11205 loop
11207 Set the number of loops. Setting this value to -1 will result in
11208 infinite loops. Default is 0.
11209 size
11211 Set maximal number of samples. Default is 0.
11212 start
11214 Set first sample of loop. Default is 0.
11215 amerge
11217 Merge two or more audio streams into a single multi-channel stream.
11218 The filter accepts the following options:
11220 inputs
11222 Set the number of inputs. Default is 2.
11223 If the channel layouts of the inputs are disjoint, and therefore
11225 compatible, the channel layout of the output will be set accordingly
11226 and the channels will be reordered as necessary. If the channel layouts
11227 of the inputs are not disjoint, the output will have all the channels
11228 of the first input then all the channels of the second input, in that
11229 order, and the channel layout of the output will be the default value
11230 corresponding to the total number of channels.
11231 For example, if the first input is in 2.1 (FL+FR+LF) and the second
11233 input is FC+BL+BR, then the output will be in 5.1, with the channels in
11234 the following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of
11235 the first input, b1 is the first channel of the second input).
11236 On the other hand, if both input are in stereo, the output channels
11238 will be in the default order: a1, a2, b1, b2, and the channel layout
11239 will be arbitrarily set to 4.0, which may or may not be the expected
11240 value.
11241 All inputs must have the same sample rate, and format.
11243 If inputs do not have the same duration, the output will stop with the
11245 shortest.
11246 Examples
11248 • Merge two mono files into a stereo stream:
11250 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
11252 • Multiple merges assuming 1 video stream and 6 audio streams in
11254 input.mkv:
11255 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
11257 amix
11259 Mixes multiple audio inputs into a single output.
11260 Note that this filter only supports float samples (the amerge and pan
11262 audio filters support many formats). If the amix input has integer
11263 samples then aresample will be automatically inserted to perform the
11264 conversion to float samples.
11265 For example
11267 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
11269 will mix 3 input audio streams to a single output with the same
11271 duration as the first input and a dropout transition time of 3 seconds.
11272 It accepts the following parameters:
11274 inputs
11276 The number of inputs. If unspecified, it defaults to 2.
11277 duration
11279 How to determine the end-of-stream.
11280 longest
11282 The duration of the longest input. (default)
11283 shortest
11285 The duration of the shortest input.
11286 first
11288 The duration of the first input.
11289 dropout_transition
11291 The transition time, in seconds, for volume renormalization when an
11292 input stream ends. The default value is 2 seconds.
11293 weights
11295 Specify weight of each input audio stream as sequence. Each weight
11296 is separated by space. By default all inputs have same weight.
11297 normalize
11299 Always scale inputs instead of only doing summation of samples.
11300 Beware of heavy clipping if inputs are not normalized prior or
11301 after filtering by this filter if this option is disabled. By
11302 default is enabled.
11303 Commands
11305 This filter supports the following commands:
11307 weights
11309 normalize
11310 Syntax is same as option with same name.
11311 amultiply
11313 Multiply first audio stream with second audio stream and store result
11314 in output audio stream. Multiplication is done by multiplying each
11315 sample from first stream with sample at same position from second
11316 stream.
11317 With this element-wise multiplication one can create amplitude fades
11319 and amplitude modulations.
11320 anequalizer
11322 High-order parametric multiband equalizer for each channel.
11323 It accepts the following parameters:
11325 params
11327 This option string is in format: "cchn f=cf w=w g=g t=f | ..."
11328 Each equalizer band is separated by '|'.
11329 chn Set channel number to which equalization will be applied. If
11331 input doesn't have that channel the entry is ignored.
11332 f Set central frequency for band. If input doesn't have that
11334 frequency the entry is ignored.
11335 w Set band width in Hertz.
11337 g Set band gain in dB.
11339 t Set filter type for band, optional, can be:
11341 0 Butterworth, this is default.
11343 1 Chebyshev type 1.
11345 2 Chebyshev type 2.
11347 curves
11349 With this option activated frequency response of anequalizer is
11350 displayed in video stream.
11351 size
11353 Set video stream size. Only useful if curves option is activated.
11354 mgain
11356 Set max gain that will be displayed. Only useful if curves option
11357 is activated. Setting this to a reasonable value makes it possible
11358 to display gain which is derived from neighbour bands which are too
11359 close to each other and thus produce higher gain when both are
11360 activated.
11361 fscale
11363 Set frequency scale used to draw frequency response in video
11364 output. Can be linear or logarithmic. Default is logarithmic.
11365 colors
11367 Set color for each channel curve which is going to be displayed in
11368 video stream. This is list of color names separated by space or by
11369 '|'. Unrecognised or missing colors will be replaced by white
11370 color.
11371 Examples
11373 • Lower gain by 10 of central frequency 200Hz and width 100 Hz for
11375 first 2 channels using Chebyshev type 1 filter:
11376 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
11378 Commands
11380 This filter supports the following commands:
11382 change
11384 Alter existing filter parameters. Syntax for the commands is :
11385 "fN|f=freq|w=width|g=gain"
11386 fN is existing filter number, starting from 0, if no such filter is
11388 available error is returned. freq set new frequency parameter.
11389 width set new width parameter in Hertz. gain set new gain
11390 parameter in dB.
11391 Full filter invocation with asendcmd may look like this:
11393 asendcmd=c='4.0 anequalizer change
11394 0|f=200|w=50|g=1',anequalizer=...
11395 anlmdn
11397 Reduce broadband noise in audio samples using Non-Local Means
11398 algorithm.
11399 Each sample is adjusted by looking for other samples with similar
11401 contexts. This context similarity is defined by comparing their
11402 surrounding patches of size p. Patches are searched in an area of r
11403 around the sample.
11404 The filter accepts the following options:
11406 strength, s
11408 Set denoising strength. Allowed range is from 0.00001 to 10000.
11409 Default value is 0.00001.
11410 patch, p
11412 Set patch radius duration. Allowed range is from 1 to 100
11413 milliseconds. Default value is 2 milliseconds.
11414 research, r
11416 Set research radius duration. Allowed range is from 2 to 300
11417 milliseconds. Default value is 6 milliseconds.
11418 output, o
11420 Set the output mode.
11421 It accepts the following values:
11423 i Pass input unchanged.
11425 o Pass noise filtered out.
11427 n Pass only noise.
11429 Default value is o.
11431 smooth, m
11433 Set smooth factor. Default value is 11. Allowed range is from 1 to
11434 1000.
11435 Commands
11437 This filter supports the all above options as commands.
11439 anlmf, anlms
11441 Apply Normalized Least-Mean-(Squares|Fourth) algorithm to the first
11442 audio stream using the second audio stream.
11443 This adaptive filter is used to mimic a desired filter by finding the
11445 filter coefficients that relate to producing the least mean square of
11446 the error signal (difference between the desired, 2nd input audio
11447 stream and the actual signal, the 1st input audio stream).
11448 A description of the accepted options follows.
11450 order
11452 Set filter order.
11453 mu Set filter mu.
11455 eps Set the filter eps.
11457 leakage
11459 Set the filter leakage.
11460 out_mode
11462 It accepts the following values:
11463 i Pass the 1st input.
11465 d Pass the 2nd input.
11467 o Pass filtered samples.
11469 n Pass difference between desired and filtered samples.
11471 Default value is o.
11473 Examples
11475 • One of many usages of this filter is noise reduction, input audio
11477 is filtered with same samples that are delayed by fixed amount, one
11478 such example for stereo audio is:
11479 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
11481 Commands
11483 This filter supports the same commands as options, excluding option
11485 "order".
11486 anull
11488 Pass the audio source unchanged to the output.
11489 apad
11491 Pad the end of an audio stream with silence.
11492 This can be used together with ffmpeg -shortest to extend audio streams
11494 to the same length as the video stream.
11495 A description of the accepted options follows.
11497 packet_size
11499 Set silence packet size. Default value is 4096.
11500 pad_len
11502 Set the number of samples of silence to add to the end. After the
11503 value is reached, the stream is terminated. This option is mutually
11504 exclusive with whole_len.
11505 whole_len
11507 Set the minimum total number of samples in the output audio stream.
11508 If the value is longer than the input audio length, silence is
11509 added to the end, until the value is reached. This option is
11510 mutually exclusive with pad_len.
11511 pad_dur
11513 Specify the duration of samples of silence to add. See the Time
11514 duration section in the ffmpeg-utils(1) manual for the accepted
11515 syntax. Used only if set to non-negative value.
11516 whole_dur
11518 Specify the minimum total duration in the output audio stream. See
11519 the Time duration section in the ffmpeg-utils(1) manual for the
11520 accepted syntax. Used only if set to non-negative value. If the
11521 value is longer than the input audio length, silence is added to
11522 the end, until the value is reached. This option is mutually
11523 exclusive with pad_dur
11524 If neither the pad_len nor the whole_len nor pad_dur nor whole_dur
11526 option is set, the filter will add silence to the end of the input
11527 stream indefinitely.
11528 Note that for ffmpeg 4.4 and earlier a zero pad_dur or whole_dur also
11530 caused the filter to add silence indefinitely.
11531 Examples
11533 • Add 1024 samples of silence to the end of the input:
11535 apad=pad_len=1024
11537 • Make sure the audio output will contain at least 10000 samples, pad
11539 the input with silence if required:
11540 apad=whole_len=10000
11542 • Use ffmpeg to pad the audio input with silence, so that the video
11544 stream will always result the shortest and will be converted until
11545 the end in the output file when using the shortest option:
11546 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
11548 aphaser
11550 Add a phasing effect to the input audio.
11551 A phaser filter creates series of peaks and troughs in the frequency
11553 spectrum. The position of the peaks and troughs are modulated so that
11554 they vary over time, creating a sweeping effect.
11555 A description of the accepted parameters follows.
11557 in_gain
11559 Set input gain. Default is 0.4.
11560 out_gain
11562 Set output gain. Default is 0.74
11563 delay
11565 Set delay in milliseconds. Default is 3.0.
11566 decay
11568 Set decay. Default is 0.4.
11569 speed
11571 Set modulation speed in Hz. Default is 0.5.
11572 type
11574 Set modulation type. Default is triangular.
11575 It accepts the following values:
11577 triangular, t
11579 sinusoidal, s
11580 aphaseshift
11582 Apply phase shift to input audio samples.
11583 The filter accepts the following options:
11585 shift
11587 Specify phase shift. Allowed range is from -1.0 to 1.0. Default
11588 value is 0.0.
11589 level
11591 Set output gain applied to final output. Allowed range is from 0.0
11592 to 1.0. Default value is 1.0.
11593 order
11595 Set filter order used for filtering. Allowed range is from 1 to 16.
11596 Default value is 8.
11597 Commands
11599 This filter supports the all above options as commands.
11601 apsyclip
11603 Apply Psychoacoustic clipper to input audio stream.
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 clip
11614 Set the clipping start value. Default value is 0dBFS or 1.
11615 diff
11617 Output only difference samples, useful to hear introduced
11618 distortions. By default is disabled.
11619 adaptive
11621 Set strength of adaptive distortion applied. Default value is 0.5.
11622 Allowed range is from 0 to 1.
11623 iterations
11625 Set number of iterations of psychoacoustic clipper. Allowed range
11626 is from 1 to 20. Default value is 10.
11627 level
11629 Auto level output signal. Default is disabled. This normalizes
11630 audio back to 0dBFS if enabled.
11631 Commands
11633 This filter supports the all above options as commands.
11635 apulsator
11637 Audio pulsator is something between an autopanner and a tremolo. But
11638 it can produce funny stereo effects as well. Pulsator changes the
11639 volume of the left and right channel based on a LFO (low frequency
11640 oscillator) with different waveforms and shifted phases. This filter
11641 have the ability to define an offset between left and right channel. An
11642 offset of 0 means that both LFO shapes match each other. The left and
11643 right channel are altered equally - a conventional tremolo. An offset
11644 of 50% means that the shape of the right channel is exactly shifted in
11645 phase (or moved backwards about half of the frequency) - pulsator acts
11646 as an autopanner. At 1 both curves match again. Every setting in
11647 between moves the phase shift gapless between all stages and produces
11648 some "bypassing" sounds with sine and triangle waveforms. The more you
11649 set the offset near 1 (starting from the 0.5) the faster the signal
11650 passes from the left to the right speaker.
11651 The filter accepts the following options:
11653 level_in
11655 Set input gain. By default it is 1. Range is [0.015625 - 64].
11656 level_out
11658 Set output gain. By default it is 1. Range is [0.015625 - 64].
11659 mode
11661 Set waveform shape the LFO will use. Can be one of: sine, triangle,
11662 square, sawup or sawdown. Default is sine.
11663 amount
11665 Set modulation. Define how much of original signal is affected by
11666 the LFO.
11667 offset_l
11669 Set left channel offset. Default is 0. Allowed range is [0 - 1].
11670 offset_r
11672 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
11673 width
11675 Set pulse width. Default is 1. Allowed range is [0 - 2].
11676 timing
11678 Set possible timing mode. Can be one of: bpm, ms or hz. Default is
11679 hz.
11680 bpm Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if
11682 timing is set to bpm.
11683 ms Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if
11685 timing is set to ms.
11686 hz Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100].
11688 Only used if timing is set to hz.
11689 aresample
11691 Resample the input audio to the specified parameters, using the
11692 libswresample library. If none are specified then the filter will
11693 automatically convert between its input and output.
11694 This filter is also able to stretch/squeeze the audio data to make it
11696 match the timestamps or to inject silence / cut out audio to make it
11697 match the timestamps, do a combination of both or do neither.
11698 The filter accepts the syntax [sample_rate:]resampler_options, where
11700 sample_rate expresses a sample rate and resampler_options is a list of
11701 key=value pairs, separated by ":". See the "Resampler Options" section
11702 in the ffmpeg-resampler(1) manual for the complete list of supported
11703 options.
11704 Examples
11706 • Resample the input audio to 44100Hz:
11708 aresample=44100
11710 • Stretch/squeeze samples to the given timestamps, with a maximum of
11712 1000 samples per second compensation:
11713 aresample=async=1000
11715 areverse
11717 Reverse an audio clip.
11718 Warning: This filter requires memory to buffer the entire clip, so
11720 trimming is suggested.
11721 Examples
11723 • Take the first 5 seconds of a clip, and reverse it.
11725 atrim=end=5,areverse
11727 arnndn
11729 Reduce noise from speech using Recurrent Neural Networks.
11730 This filter accepts the following options:
11732 model, m
11734 Set train model file to load. This option is always required.
11735 mix Set how much to mix filtered samples into final output. Allowed
11737 range is from -1 to 1. Default value is 1. Negative values are
11738 special, they set how much to keep filtered noise in the final
11739 filter output. Set this option to -1 to hear actual noise removed
11740 from input signal.
11741 Commands
11743 This filter supports the all above options as commands.
11745 asdr
11747 Measure Audio Signal-to-Distortion Ratio.
11748 This filter takes two audio streams for input, and outputs first audio
11750 stream. Results are in dB per channel at end of either input.
11751 asetnsamples
11753 Set the number of samples per each output audio frame.
11754 The last output packet may contain a different number of samples, as
11756 the filter will flush all the remaining samples when the input audio
11757 signals its end.
11758 The filter accepts the following options:
11760 nb_out_samples, n
11762 Set the number of frames per each output audio frame. The number is
11763 intended as the number of samples per each channel. Default value
11764 is 1024.
11765 pad, p
11767 If set to 1, the filter will pad the last audio frame with zeroes,
11768 so that the last frame will contain the same number of samples as
11769 the previous ones. Default value is 1.
11770 For example, to set the number of per-frame samples to 1234 and disable
11772 padding for the last frame, use:
11773 asetnsamples=n=1234:p=0
11775 asetrate
11777 Set the sample rate without altering the PCM data. This will result in
11778 a change of speed and pitch.
11779 The filter accepts the following options:
11781 sample_rate, r
11783 Set the output sample rate. Default is 44100 Hz.
11784 ashowinfo
11786 Show a line containing various information for each input audio frame.
11787 The input audio is not modified.
11788 The shown line contains a sequence of key/value pairs of the form
11790 key:value.
11791 The following values are shown in the output:
11793 n The (sequential) number of the input frame, starting from 0.
11795 pts The presentation timestamp of the input frame, in time base units;
11797 the time base depends on the filter input pad, and is usually
11798 1/sample_rate.
11799 pts_time
11801 The presentation timestamp of the input frame in seconds.
11802 pos position of the frame in the input stream, -1 if this information
11804 in unavailable and/or meaningless (for example in case of synthetic
11805 audio)
11806 fmt The sample format.
11808 chlayout
11810 The channel layout.
11811 rate
11813 The sample rate for the audio frame.
11814 nb_samples
11816 The number of samples (per channel) in the frame.
11817 checksum
11819 The Adler-32 checksum (printed in hexadecimal) of the audio data.
11820 For planar audio, the data is treated as if all the planes were
11821 concatenated.
11822 plane_checksums
11824 A list of Adler-32 checksums for each data plane.
11825 asoftclip
11827 Apply audio soft clipping.
11828 Soft clipping is a type of distortion effect where the amplitude of a
11830 signal is saturated along a smooth curve, rather than the abrupt shape
11831 of hard-clipping.
11832 This filter accepts the following options:
11834 type
11836 Set type of soft-clipping.
11837 It accepts the following values:
11839 hard
11841 tanh
11842 atan
11843 cubic
11844 exp
11845 alg
11846 quintic
11847 sin
11848 erf
11849 threshold
11850 Set threshold from where to start clipping. Default value is 0dB or
11851 1.
11852 output
11854 Set gain applied to output. Default value is 0dB or 1.
11855 param
11857 Set additional parameter which controls sigmoid function.
11858 oversample
11860 Set oversampling factor.
11861 Commands
11863 This filter supports the all above options as commands.
11865 aspectralstats
11867 Display frequency domain statistical information about the audio
11868 channels. Statistics are calculated and stored as metadata for each
11869 audio channel and for each audio frame.
11870 It accepts the following option:
11872 win_size
11874 Set the window length in samples. Default value is 2048. Allowed
11875 range is from 32 to 65536.
11876 win_func
11878 Set window function.
11879 It accepts the following values:
11881 rect
11883 bartlett
11884 hann, hanning
11885 hamming
11886 blackman
11887 welch
11888 flattop
11889 bharris
11890 bnuttall
11891 bhann
11892 sine
11893 nuttall
11894 lanczos
11895 gauss
11896 tukey
11897 dolph
11898 cauchy
11899 parzen
11900 poisson
11901 bohman
11902 Default is "hann".
11904 overlap
11906 Set window overlap. Allowed range is from 0 to 1. Default value is
11907 0.5.
11908 A list of each metadata key follows:
11910 mean
11912 variance
11913 centroid
11914 spread
11915 skewness
11916 kurtosis
11917 entropy
11918 flatness
11919 crest
11920 flux
11921 slope
11922 decrease
11923 rolloff
11924 asr
11926 Automatic Speech Recognition
11927 This filter uses PocketSphinx for speech recognition. To enable
11929 compilation of this filter, you need to configure FFmpeg with
11930 "--enable-pocketsphinx".
11931 It accepts the following options:
11933 rate
11935 Set sampling rate of input audio. Defaults is 16000. This need to
11936 match speech models, otherwise one will get poor results.
11937 hmm Set dictionary containing acoustic model files.
11939 dict
11941 Set pronunciation dictionary.
11942 lm Set language model file.
11944 lmctl
11946 Set language model set.
11947 lmname
11949 Set which language model to use.
11950 logfn
11952 Set output for log messages.
11953 The filter exports recognized speech as the frame metadata
11955 "lavfi.asr.text".
11956 astats
11958 Display time domain statistical information about the audio channels.
11959 Statistics are calculated and displayed for each audio channel and,
11960 where applicable, an overall figure is also given.
11961 It accepts the following option:
11963 length
11965 Short window length in seconds, used for peak and trough RMS
11966 measurement. Default is 0.05 (50 milliseconds). Allowed range is
11967 "[0 - 10]".
11968 metadata
11970 Set metadata injection. All the metadata keys are prefixed with
11971 "lavfi.astats.X", where "X" is channel number starting from 1 or
11972 string "Overall". Default is disabled.
11973 Available keys for each channel are: DC_offset Min_level Max_level
11975 Min_difference Max_difference Mean_difference RMS_difference
11976 Peak_level RMS_peak RMS_trough Crest_factor Flat_factor Peak_count
11977 Noise_floor Noise_floor_count Entropy Bit_depth Dynamic_range
11978 Zero_crossings Zero_crossings_rate Number_of_NaNs Number_of_Infs
11979 Number_of_denormals
11980 and for Overall: DC_offset Min_level Max_level Min_difference
11982 Max_difference Mean_difference RMS_difference Peak_level RMS_level
11983 RMS_peak RMS_trough Flat_factor Peak_count Noise_floor
11984 Noise_floor_count Entropy Bit_depth Number_of_samples
11985 Number_of_NaNs Number_of_Infs Number_of_denormals
11986 For example full key look like this "lavfi.astats.1.DC_offset" or
11988 this "lavfi.astats.Overall.Peak_count".
11989 For description what each key means read below.
11991 reset
11993 Set the number of frames over which cumulative stats are calculated
11994 before being reset Default is disabled.
11995 measure_perchannel
11997 Select the parameters which are measured per channel. The metadata
11998 keys can be used as flags, default is all which measures
11999 everything. none disables all per channel measurement.
12000 measure_overall
12002 Select the parameters which are measured overall. The metadata keys
12003 can be used as flags, default is all which measures everything.
12004 none disables all overall measurement.
12005 A description of each shown parameter follows:
12007 DC offset
12009 Mean amplitude displacement from zero.
12010 Min level
12012 Minimal sample level.
12013 Max level
12015 Maximal sample level.
12016 Min difference
12018 Minimal difference between two consecutive samples.
12019 Max difference
12021 Maximal difference between two consecutive samples.
12022 Mean difference
12024 Mean difference between two consecutive samples. The average of
12025 each difference between two consecutive samples.
12026 RMS difference
12028 Root Mean Square difference between two consecutive samples.
12029 Peak level dB
12031 RMS level dB
12032 Standard peak and RMS level measured in dBFS.
12033 RMS peak dB
12035 RMS trough dB
12036 Peak and trough values for RMS level measured over a short window.
12037 Crest factor
12039 Standard ratio of peak to RMS level (note: not in dB).
12040 Flat factor
12042 Flatness (i.e. consecutive samples with the same value) of the
12043 signal at its peak levels (i.e. either Min level or Max level).
12044 Peak count
12046 Number of occasions (not the number of samples) that the signal
12047 attained either Min level or Max level.
12048 Noise floor dB
12050 Minimum local peak measured in dBFS over a short window.
12051 Noise floor count
12053 Number of occasions (not the number of samples) that the signal
12054 attained Noise floor.
12055 Entropy
12057 Entropy measured across whole audio. Entropy of value near 1.0 is
12058 typically measured for white noise.
12059 Bit depth
12061 Overall bit depth of audio. Number of bits used for each sample.
12062 Dynamic range
12064 Measured dynamic range of audio in dB.
12065 Zero crossings
12067 Number of points where the waveform crosses the zero level axis.
12068 Zero crossings rate
12070 Rate of Zero crossings and number of audio samples.
12071 asubboost
12073 Boost subwoofer frequencies.
12074 The filter accepts the following options:
12076 dry Set dry gain, how much of original signal is kept. Allowed range is
12078 from 0 to 1. Default value is 1.0.
12079 wet Set wet gain, how much of filtered signal is kept. Allowed range is
12081 from 0 to 1. Default value is 1.0.
12082 boost
12084 Set max boost factor. Allowed range is from 1 to 12. Default value
12085 is 2.
12086 decay
12088 Set delay line decay gain value. Allowed range is from 0 to 1.
12089 Default value is 0.0.
12090 feedback
12092 Set delay line feedback gain value. Allowed range is from 0 to 1.
12093 Default value is 0.9.
12094 cutoff
12096 Set cutoff frequency in Hertz. Allowed range is 50 to 900. Default
12097 value is 100.
12098 slope
12100 Set slope amount for cutoff frequency. Allowed range is 0.0001 to
12101 1. Default value is 0.5.
12102 delay
12104 Set delay. Allowed range is from 1 to 100. Default value is 20.
12105 channels
12107 Set the channels to process. Default value is all available.
12108 Commands
12110 This filter supports the all above options as commands.
12112 asubcut
12114 Cut subwoofer frequencies.
12115 This filter allows to set custom, steeper roll off than highpass
12117 filter, and thus is able to more attenuate frequency content in stop-
12118 band.
12119 The filter accepts the following options:
12121 cutoff
12123 Set cutoff frequency in Hertz. Allowed range is 2 to 200. Default
12124 value is 20.
12125 order
12127 Set filter order. Available values are from 3 to 20. Default value
12128 is 10.
12129 level
12131 Set input gain level. Allowed range is from 0 to 1. Default value
12132 is 1.
12133 Commands
12135 This filter supports the all above options as commands.
12137 asupercut
12139 Cut super frequencies.
12140 The filter accepts the following options:
12142 cutoff
12144 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
12145 Default value is 20000.
12146 order
12148 Set filter order. Available values are from 3 to 20. Default value
12149 is 10.
12150 level
12152 Set input gain level. Allowed range is from 0 to 1. Default value
12153 is 1.
12154 Commands
12156 This filter supports the all above options as commands.
12158 asuperpass
12160 Apply high order Butterworth band-pass filter.
12161 The filter accepts the following options:
12163 centerf
12165 Set center frequency in Hertz. Allowed range is 2 to 999999.
12166 Default value is 1000.
12167 order
12169 Set filter order. Available values are from 4 to 20. Default value
12170 is 4.
12171 qfactor
12173 Set Q-factor. Allowed range is from 0.01 to 100. Default value is
12174 1.
12175 level
12177 Set input gain level. Allowed range is from 0 to 2. Default value
12178 is 1.
12179 Commands
12181 This filter supports the all above options as commands.
12183 asuperstop
12185 Apply high order Butterworth band-stop filter.
12186 The filter accepts the following options:
12188 centerf
12190 Set center frequency in Hertz. Allowed range is 2 to 999999.
12191 Default value is 1000.
12192 order
12194 Set filter order. Available values are from 4 to 20. Default value
12195 is 4.
12196 qfactor
12198 Set Q-factor. Allowed range is from 0.01 to 100. Default value is
12199 1.
12200 level
12202 Set input gain level. Allowed range is from 0 to 2. Default value
12203 is 1.
12204 Commands
12206 This filter supports the all above options as commands.
12208 atempo
12210 Adjust audio tempo.
12211 The filter accepts exactly one parameter, the audio tempo. If not
12213 specified then the filter will assume nominal 1.0 tempo. Tempo must be
12214 in the [0.5, 100.0] range.
12215 Note that tempo greater than 2 will skip some samples rather than blend
12217 them in. If for any reason this is a concern it is always possible to
12218 daisy-chain several instances of atempo to achieve the desired product
12219 tempo.
12220 Examples
12222 • Slow down audio to 80% tempo:
12224 atempo=0.8
12226 • To speed up audio to 300% tempo:
12228 atempo=3
12230 • To speed up audio to 300% tempo by daisy-chaining two atempo
12232 instances:
12233 atempo=sqrt(3),atempo=sqrt(3)
12235 Commands
12237 This filter supports the following commands:
12239 tempo
12241 Change filter tempo scale factor. Syntax for the command is :
12242 "tempo"
12243 atilt
12245 Apply spectral tilt filter to audio stream.
12246 This filter apply any spectral roll-off slope over any specified
12248 frequency band.
12249 The filter accepts the following options:
12251 freq
12253 Set central frequency of tilt in Hz. Default is 10000 Hz.
12254 slope
12256 Set slope direction of tilt. Default is 0. Allowed range is from -1
12257 to 1.
12258 width
12260 Set width of tilt. Default is 1000. Allowed range is from 100 to
12261 10000.
12262 order
12264 Set order of tilt filter.
12265 level
12267 Set input volume level. Allowed range is from 0 to 4. Defalt is 1.
12268 Commands
12270 This filter supports the all above options as commands.
12272 atrim
12274 Trim the input so that the output contains one continuous subpart of
12275 the input.
12276 It accepts the following parameters:
12278 start
12280 Timestamp (in seconds) of the start of the section to keep. I.e.
12281 the audio sample with the timestamp start will be the first sample
12282 in the output.
12283 end Specify time of the first audio sample that will be dropped, i.e.
12285 the audio sample immediately preceding the one with the timestamp
12286 end will be the last sample in the output.
12287 start_pts
12289 Same as start, except this option sets the start timestamp in
12290 samples instead of seconds.
12291 end_pts
12293 Same as end, except this option sets the end timestamp in samples
12294 instead of seconds.
12295 duration
12297 The maximum duration of the output in seconds.
12298 start_sample
12300 The number of the first sample that should be output.
12301 end_sample
12303 The number of the first sample that should be dropped.
12304 start, end, and duration are expressed as time duration specifications;
12306 see the Time duration section in the ffmpeg-utils(1) manual.
12307 Note that the first two sets of the start/end options and the duration
12309 option look at the frame timestamp, while the _sample options simply
12310 count the samples that pass through the filter. So start/end_pts and
12311 start/end_sample will give different results when the timestamps are
12312 wrong, inexact or do not start at zero. Also note that this filter does
12313 not modify the timestamps. If you wish to have the output timestamps
12314 start at zero, insert the asetpts filter after the atrim filter.
12315 If multiple start or end options are set, this filter tries to be
12317 greedy and keep all samples that match at least one of the specified
12318 constraints. To keep only the part that matches all the constraints at
12319 once, chain multiple atrim filters.
12320 The defaults are such that all the input is kept. So it is possible to
12322 set e.g. just the end values to keep everything before the specified
12323 time.
12324 Examples:
12326 • Drop everything except the second minute of input:
12328 ffmpeg -i INPUT -af atrim=60:120
12330 • Keep only the first 1000 samples:
12332 ffmpeg -i INPUT -af atrim=end_sample=1000
12334 axcorrelate
12336 Calculate normalized windowed cross-correlation between two input audio
12337 streams.
12338 Resulted samples are always between -1 and 1 inclusive. If result is 1
12340 it means two input samples are highly correlated in that selected
12341 segment. Result 0 means they are not correlated at all. If result is
12342 -1 it means two input samples are out of phase, which means they cancel
12343 each other.
12344 The filter accepts the following options:
12346 size
12348 Set size of segment over which cross-correlation is calculated.
12349 Default is 256. Allowed range is from 2 to 131072.
12350 algo
12352 Set algorithm for cross-correlation. Can be "slow" or "fast".
12353 Default is "slow". Fast algorithm assumes mean values over any
12354 given segment are always zero and thus need much less calculations
12355 to make. This is generally not true, but is valid for typical
12356 audio streams.
12357 Examples
12359 • Calculate correlation between channels in stereo audio stream:
12361 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
12363 bandpass
12365 Apply a two-pole Butterworth band-pass filter with central frequency
12366 frequency, and (3dB-point) band-width width. The csg option selects a
12367 constant skirt gain (peak gain = Q) instead of the default: constant
12368 0dB peak gain. The filter roll off at 6dB per octave (20dB per
12369 decade).
12370 The filter accepts the following options:
12372 frequency, f
12374 Set the filter's central frequency. Default is 3000.
12375 csg Constant skirt gain if set to 1. Defaults to 0.
12377 width_type, t
12379 Set method to specify band-width of filter.
12380 h Hz
12382 q Q-Factor
12384 o octave
12386 s slope
12388 k kHz
12390 width, w
12392 Specify the band-width of a filter in width_type units.
12393 mix, m
12395 How much to use filtered signal in output. Default is 1. Range is
12396 between 0 and 1.
12397 channels, c
12399 Specify which channels to filter, by default all available are
12400 filtered.
12401 normalize, n
12403 Normalize biquad coefficients, by default is disabled. Enabling it
12404 will normalize magnitude response at DC to 0dB.
12405 transform, a
12407 Set transform type of IIR filter.
12408 di
12410 dii
12411 tdi
12412 tdii
12413 latt
12414 svf
12415 zdf
12416 precision, r
12417 Set precison of filtering.
12418 auto
12420 Pick automatic sample format depending on surround filters.
12421 s16 Always use signed 16-bit.
12423 s32 Always use signed 32-bit.
12425 f32 Always use float 32-bit.
12427 f64 Always use float 64-bit.
12429 block_size, b
12431 Set block size used for reverse IIR processing. If this value is
12432 set to high enough value (higher than impulse response length
12433 truncated when reaches near zero values) filtering will become
12434 linear phase otherwise if not big enough it will just produce nasty
12435 artifacts.
12436 Note that filter delay will be exactly this many samples when set
12438 to non-zero value.
12439 Commands
12441 This filter supports the following commands:
12443 frequency, f
12445 Change bandpass frequency. Syntax for the command is : "frequency"
12446 width_type, t
12448 Change bandpass width_type. Syntax for the command is :
12449 "width_type"
12450 width, w
12452 Change bandpass width. Syntax for the command is : "width"
12453 mix, m
12455 Change bandpass mix. Syntax for the command is : "mix"
12456 bandreject
12458 Apply a two-pole Butterworth band-reject filter with central frequency
12459 frequency, and (3dB-point) band-width width. The filter roll off at
12460 6dB per octave (20dB per decade).
12461 The filter accepts the following options:
12463 frequency, f
12465 Set the filter's central frequency. Default is 3000.
12466 width_type, t
12468 Set method to specify band-width of filter.
12469 h Hz
12471 q Q-Factor
12473 o octave
12475 s slope
12477 k kHz
12479 width, w
12481 Specify the band-width of a filter in width_type units.
12482 mix, m
12484 How much to use filtered signal in output. Default is 1. Range is
12485 between 0 and 1.
12486 channels, c
12488 Specify which channels to filter, by default all available are
12489 filtered.
12490 normalize, n
12492 Normalize biquad coefficients, by default is disabled. Enabling it
12493 will normalize magnitude response at DC to 0dB.
12494 transform, a
12496 Set transform type of IIR filter.
12497 di
12499 dii
12500 tdi
12501 tdii
12502 latt
12503 svf
12504 zdf
12505 precision, r
12506 Set precison of filtering.
12507 auto
12509 Pick automatic sample format depending on surround filters.
12510 s16 Always use signed 16-bit.
12512 s32 Always use signed 32-bit.
12514 f32 Always use float 32-bit.
12516 f64 Always use float 64-bit.
12518 block_size, b
12520 Set block size used for reverse IIR processing. If this value is
12521 set to high enough value (higher than impulse response length
12522 truncated when reaches near zero values) filtering will become
12523 linear phase otherwise if not big enough it will just produce nasty
12524 artifacts.
12525 Note that filter delay will be exactly this many samples when set
12527 to non-zero value.
12528 Commands
12530 This filter supports the following commands:
12532 frequency, f
12534 Change bandreject frequency. Syntax for the command is :
12535 "frequency"
12536 width_type, t
12538 Change bandreject width_type. Syntax for the command is :
12539 "width_type"
12540 width, w
12542 Change bandreject width. Syntax for the command is : "width"
12543 mix, m
12545 Change bandreject mix. Syntax for the command is : "mix"
12546 bass, lowshelf
12548 Boost or cut the bass (lower) frequencies of the audio using a two-pole
12549 shelving filter with a response similar to that of a standard hi-fi's
12550 tone-controls. This is also known as shelving equalisation (EQ).
12551 The filter accepts the following options:
12553 gain, g
12555 Give the gain at 0 Hz. Its useful range is about -20 (for a large
12556 cut) to +20 (for a large boost). Beware of clipping when using a
12557 positive gain.
12558 frequency, f
12560 Set the filter's central frequency and so can be used to extend or
12561 reduce the frequency range to be boosted or cut. The default value
12562 is 100 Hz.
12563 width_type, t
12565 Set method to specify band-width of filter.
12566 h Hz
12568 q Q-Factor
12570 o octave
12572 s slope
12574 k kHz
12576 width, w
12578 Determine how steep is the filter's shelf transition.
12579 poles, p
12581 Set number of poles. Default is 2.
12582 mix, m
12584 How much to use filtered signal in output. Default is 1. Range is
12585 between 0 and 1.
12586 channels, c
12588 Specify which channels to filter, by default all available are
12589 filtered.
12590 normalize, n
12592 Normalize biquad coefficients, by default is disabled. Enabling it
12593 will normalize magnitude response at DC to 0dB.
12594 transform, a
12596 Set transform type of IIR filter.
12597 di
12599 dii
12600 tdi
12601 tdii
12602 latt
12603 svf
12604 zdf
12605 precision, r
12606 Set precison of filtering.
12607 auto
12609 Pick automatic sample format depending on surround filters.
12610 s16 Always use signed 16-bit.
12612 s32 Always use signed 32-bit.
12614 f32 Always use float 32-bit.
12616 f64 Always use float 64-bit.
12618 block_size, b
12620 Set block size used for reverse IIR processing. If this value is
12621 set to high enough value (higher than impulse response length
12622 truncated when reaches near zero values) filtering will become
12623 linear phase otherwise if not big enough it will just produce nasty
12624 artifacts.
12625 Note that filter delay will be exactly this many samples when set
12627 to non-zero value.
12628 Commands
12630 This filter supports the following commands:
12632 frequency, f
12634 Change bass frequency. Syntax for the command is : "frequency"
12635 width_type, t
12637 Change bass width_type. Syntax for the command is : "width_type"
12638 width, w
12640 Change bass width. Syntax for the command is : "width"
12641 gain, g
12643 Change bass gain. Syntax for the command is : "gain"
12644 mix, m
12646 Change bass mix. Syntax for the command is : "mix"
12647 biquad
12649 Apply a biquad IIR filter with the given coefficients. Where b0, b1,
12650 b2 and a0, a1, a2 are the numerator and denominator coefficients
12651 respectively. and channels, c specify which channels to filter, by
12652 default all available are filtered.
12653 Commands
12655 This filter supports the following commands:
12657 a0
12659 a1
12660 a2
12661 b0
12662 b1
12663 b2 Change biquad parameter. Syntax for the command is : "value"
12664 mix, m
12666 How much to use filtered signal in output. Default is 1. Range is
12667 between 0 and 1.
12668 channels, c
12670 Specify which channels to filter, by default all available are
12671 filtered.
12672 normalize, n
12674 Normalize biquad coefficients, by default is disabled. Enabling it
12675 will normalize magnitude response at DC to 0dB.
12676 transform, a
12678 Set transform type of IIR filter.
12679 di
12681 dii
12682 tdi
12683 tdii
12684 latt
12685 svf
12686 zdf
12687 precision, r
12688 Set precison of filtering.
12689 auto
12691 Pick automatic sample format depending on surround filters.
12692 s16 Always use signed 16-bit.
12694 s32 Always use signed 32-bit.
12696 f32 Always use float 32-bit.
12698 f64 Always use float 64-bit.
12700 block_size, b
12702 Set block size used for reverse IIR processing. If this value is
12703 set to high enough value (higher than impulse response length
12704 truncated when reaches near zero values) filtering will become
12705 linear phase otherwise if not big enough it will just produce nasty
12706 artifacts.
12707 Note that filter delay will be exactly this many samples when set
12709 to non-zero value.
12710 bs2b
12712 Bauer stereo to binaural transformation, which improves headphone
12713 listening of stereo audio records.
12714 To enable compilation of this filter you need to configure FFmpeg with
12716 "--enable-libbs2b".
12717 It accepts the following parameters:
12719 profile
12721 Pre-defined crossfeed level.
12722 default
12724 Default level (fcut=700, feed=50).
12725 cmoy
12727 Chu Moy circuit (fcut=700, feed=60).
12728 jmeier
12730 Jan Meier circuit (fcut=650, feed=95).
12731 fcut
12733 Cut frequency (in Hz).
12734 feed
12736 Feed level (in Hz).
12737 channelmap
12739 Remap input channels to new locations.
12740 It accepts the following parameters:
12742 map Map channels from input to output. The argument is a '|'-separated
12744 list of mappings, each in the "in_channel-out_channel" or
12745 in_channel form. in_channel can be either the name of the input
12746 channel (e.g. FL for front left) or its index in the input channel
12747 layout. out_channel is the name of the output channel or its index
12748 in the output channel layout. If out_channel is not given then it
12749 is implicitly an index, starting with zero and increasing by one
12750 for each mapping.
12751 channel_layout
12753 The channel layout of the output stream.
12754 If no mapping is present, the filter will implicitly map input channels
12756 to output channels, preserving indices.
12757 Examples
12759 • For example, assuming a 5.1+downmix input MOV file,
12761 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
12763 will create an output WAV file tagged as stereo from the downmix
12765 channels of the input.
12766 • To fix a 5.1 WAV improperly encoded in AAC's native channel order
12768 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
12770 channelsplit
12772 Split each channel from an input audio stream into a separate output
12773 stream.
12774 It accepts the following parameters:
12776 channel_layout
12778 The channel layout of the input stream. The default is "stereo".
12779 channels
12781 A channel layout describing the channels to be extracted as
12782 separate output streams or "all" to extract each input channel as a
12783 separate stream. The default is "all".
12784 Choosing channels not present in channel layout in the input will
12786 result in an error.
12787 Examples
12789 • For example, assuming a stereo input MP3 file,
12791 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
12793 will create an output Matroska file with two audio streams, one
12795 containing only the left channel and the other the right channel.
12796 • Split a 5.1 WAV file into per-channel files:
12798 ffmpeg -i in.wav -filter_complex
12800 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
12801 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
12802 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
12803 side_right.wav
12804 • Extract only LFE from a 5.1 WAV file:
12806 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
12808 -map '[LFE]' lfe.wav
12809 chorus
12811 Add a chorus effect to the audio.
12812 Can make a single vocal sound like a chorus, but can also be applied to
12814 instrumentation.
12815 Chorus resembles an echo effect with a short delay, but whereas with
12817 echo the delay is constant, with chorus, it is varied using using
12818 sinusoidal or triangular modulation. The modulation depth defines the
12819 range the modulated delay is played before or after the delay. Hence
12820 the delayed sound will sound slower or faster, that is the delayed
12821 sound tuned around the original one, like in a chorus where some vocals
12822 are slightly off key.
12823 It accepts the following parameters:
12825 in_gain
12827 Set input gain. Default is 0.4.
12828 out_gain
12830 Set output gain. Default is 0.4.
12831 delays
12833 Set delays. A typical delay is around 40ms to 60ms.
12834 decays
12836 Set decays.
12837 speeds
12839 Set speeds.
12840 depths
12842 Set depths.
12843 Examples
12845 • A single delay:
12847 chorus=0.7:0.9:55:0.4:0.25:2
12849 • Two delays:
12851 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
12853 • Fuller sounding chorus with three delays:
12855 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
12857 compand
12859 Compress or expand the audio's dynamic range.
12860 It accepts the following parameters:
12862 attacks
12864 decays
12865 A list of times in seconds for each channel over which the
12866 instantaneous level of the input signal is averaged to determine
12867 its volume. attacks refers to increase of volume and decays refers
12868 to decrease of volume. For most situations, the attack time
12869 (response to the audio getting louder) should be shorter than the
12870 decay time, because the human ear is more sensitive to sudden loud
12871 audio than sudden soft audio. A typical value for attack is 0.3
12872 seconds and a typical value for decay is 0.8 seconds. If specified
12873 number of attacks & decays is lower than number of channels, the
12874 last set attack/decay will be used for all remaining channels.
12875 points
12877 A list of points for the transfer function, specified in dB
12878 relative to the maximum possible signal amplitude. Each key points
12879 list must be defined using the following syntax:
12880 "x0/y0|x1/y1|x2/y2|...." or "x0/y0 x1/y1 x2/y2 ...."
12881 The input values must be in strictly increasing order but the
12883 transfer function does not have to be monotonically rising. The
12884 point "0/0" is assumed but may be overridden (by "0/out-dBn").
12885 Typical values for the transfer function are "-70/-70|-60/-20|1/0".
12886 soft-knee
12888 Set the curve radius in dB for all joints. It defaults to 0.01.
12889 gain
12891 Set the additional gain in dB to be applied at all points on the
12892 transfer function. This allows for easy adjustment of the overall
12893 gain. It defaults to 0.
12894 volume
12896 Set an initial volume, in dB, to be assumed for each channel when
12897 filtering starts. This permits the user to supply a nominal level
12898 initially, so that, for example, a very large gain is not applied
12899 to initial signal levels before the companding has begun to
12900 operate. A typical value for audio which is initially quiet is -90
12901 dB. It defaults to 0.
12902 delay
12904 Set a delay, in seconds. The input audio is analyzed immediately,
12905 but audio is delayed before being fed to the volume adjuster.
12906 Specifying a delay approximately equal to the attack/decay times
12907 allows the filter to effectively operate in predictive rather than
12908 reactive mode. It defaults to 0.
12909 Examples
12911 • Make music with both quiet and loud passages suitable for listening
12913 to in a noisy environment:
12914 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
12916 Another example for audio with whisper and explosion parts:
12918 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
12920 • A noise gate for when the noise is at a lower level than the
12922 signal:
12923 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
12925 • Here is another noise gate, this time for when the noise is at a
12927 higher level than the signal (making it, in some ways, similar to
12928 squelch):
12929 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
12931 • 2:1 compression starting at -6dB:
12933 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
12935 • 2:1 compression starting at -9dB:
12937 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
12939 • 2:1 compression starting at -12dB:
12941 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
12943 • 2:1 compression starting at -18dB:
12945 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
12947 • 3:1 compression starting at -15dB:
12949 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
12951 • Compressor/Gate:
12953 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
12955 • Expander:
12957 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
12959 • Hard limiter at -6dB:
12961 compand=attacks=0:points=-80/-80|-6/-6|20/-6
12963 • Hard limiter at -12dB:
12965 compand=attacks=0:points=-80/-80|-12/-12|20/-12
12967 • Hard noise gate at -35 dB:
12969 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
12971 • Soft limiter:
12973 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
12975 compensationdelay
12977 Compensation Delay Line is a metric based delay to compensate differing
12978 positions of microphones or speakers.
12979 For example, you have recorded guitar with two microphones placed in
12981 different locations. Because the front of sound wave has fixed speed in
12982 normal conditions, the phasing of microphones can vary and depends on
12983 their location and interposition. The best sound mix can be achieved
12984 when these microphones are in phase (synchronized). Note that a
12985 distance of ~30 cm between microphones makes one microphone capture the
12986 signal in antiphase to the other microphone. That makes the final mix
12987 sound moody. This filter helps to solve phasing problems by adding
12988 different delays to each microphone track and make them synchronized.
12989 The best result can be reached when you take one track as base and
12991 synchronize other tracks one by one with it. Remember that
12992 synchronization/delay tolerance depends on sample rate, too. Higher
12993 sample rates will give more tolerance.
12994 The filter accepts the following parameters:
12996 mm Set millimeters distance. This is compensation distance for fine
12998 tuning. Default is 0.
12999 cm Set cm distance. This is compensation distance for tightening
13001 distance setup. Default is 0.
13002 m Set meters distance. This is compensation distance for hard
13004 distance setup. Default is 0.
13005 dry Set dry amount. Amount of unprocessed (dry) signal. Default is 0.
13007 wet Set wet amount. Amount of processed (wet) signal. Default is 1.
13009 temp
13011 Set temperature in degrees Celsius. This is the temperature of the
13012 environment. Default is 20.
13013 Commands
13015 This filter supports the all above options as commands.
13017 crossfeed
13019 Apply headphone crossfeed filter.
13020 Crossfeed is the process of blending the left and right channels of
13022 stereo audio recording. It is mainly used to reduce extreme stereo
13023 separation of low frequencies.
13024 The intent is to produce more speaker like sound to the listener.
13026 The filter accepts the following options:
13028 strength
13030 Set strength of crossfeed. Default is 0.2. Allowed range is from 0
13031 to 1. This sets gain of low shelf filter for side part of stereo
13032 image. Default is -6dB. Max allowed is -30db when strength is set
13033 to 1.
13034 range
13036 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to
13037 1. This sets cut off frequency of low shelf filter. Default is cut
13038 off near 1550 Hz. With range set to 1 cut off frequency is set to
13039 2100 Hz.
13040 slope
13042 Set curve slope of low shelf filter. Default is 0.5. Allowed range
13043 is from 0.01 to 1.
13044 level_in
13046 Set input gain. Default is 0.9.
13047 level_out
13049 Set output gain. Default is 1.
13050 block_size
13052 Set block size used for reverse IIR processing. If this value is
13053 set to high enough value (higher than impulse response length
13054 truncated when reaches near zero values) filtering will become
13055 linear phase otherwise if not big enough it will just produce nasty
13056 artifacts.
13057 Note that filter delay will be exactly this many samples when set
13059 to non-zero value.
13060 Commands
13062 This filter supports the all above options as commands.
13064 crystalizer
13066 Simple algorithm for audio noise sharpening.
13067 This filter linearly increases differences betweeen each audio sample.
13069 The filter accepts the following options:
13071 i Sets the intensity of effect (default: 2.0). Must be in range
13073 between -10.0 to 0 (unchanged sound) to 10.0 (maximum effect). To
13074 inverse filtering use negative value.
13075 c Enable clipping. By default is enabled.
13077 Commands
13079 This filter supports the all above options as commands.
13081 dcshift
13083 Apply a DC shift to the audio.
13084 This can be useful to remove a DC offset (caused perhaps by a hardware
13086 problem in the recording chain) from the audio. The effect of a DC
13087 offset is reduced headroom and hence volume. The astats filter can be
13088 used to determine if a signal has a DC offset.
13089 shift
13091 Set the DC shift, allowed range is [-1, 1]. It indicates the amount
13092 to shift the audio.
13093 limitergain
13095 Optional. It should have a value much less than 1 (e.g. 0.05 or
13096 0.02) and is used to prevent clipping.
13097 deesser
13099 Apply de-essing to the audio samples.
13100 i Set intensity for triggering de-essing. Allowed range is from 0 to
13102 1. Default is 0.
13103 m Set amount of ducking on treble part of sound. Allowed range is
13105 from 0 to 1. Default is 0.5.
13106 f How much of original frequency content to keep when de-essing.
13108 Allowed range is from 0 to 1. Default is 0.5.
13109 s Set the output mode.
13111 It accepts the following values:
13113 i Pass input unchanged.
13115 o Pass ess filtered out.
13117 e Pass only ess.
13119 Default value is o.
13121 dialoguenhance
13123 Enhance dialogue in stereo audio.
13124 This filter accepts stereo input and produce surround (3.0) channels
13126 output. The newly produced front center channel have enhanced speech
13127 dialogue originally available in both stereo channels. This filter
13128 outputs front left and front right channels same as available in stereo
13129 input.
13130 The filter accepts the following options:
13132 original
13134 Set the original center factor to keep in front center channel
13135 output. Allowed range is from 0 to 1. Default value is 1.
13136 enhance
13138 Set the dialogue enhance factor to put in front center channel
13139 output. Allowed range is from 0 to 3. Default value is 1.
13140 voice
13142 Set the voice detection factor. Allowed range is from 2 to 32.
13143 Default value is 2.
13144 Commands
13146 This filter supports the all above options as commands.
13148 drmeter
13150 Measure audio dynamic range.
13151 DR values of 14 and higher is found in very dynamic material. DR of 8
13153 to 13 is found in transition material. And anything less that 8 have
13154 very poor dynamics and is very compressed.
13155 The filter accepts the following options:
13157 length
13159 Set window length in seconds used to split audio into segments of
13160 equal length. Default is 3 seconds.
13161 dynaudnorm
13163 Dynamic Audio Normalizer.
13164 This filter applies a certain amount of gain to the input audio in
13166 order to bring its peak magnitude to a target level (e.g. 0 dBFS).
13167 However, in contrast to more "simple" normalization algorithms, the
13168 Dynamic Audio Normalizer *dynamically* re-adjusts the gain factor to
13169 the input audio. This allows for applying extra gain to the "quiet"
13170 sections of the audio while avoiding distortions or clipping the "loud"
13171 sections. In other words: The Dynamic Audio Normalizer will "even out"
13172 the volume of quiet and loud sections, in the sense that the volume of
13173 each section is brought to the same target level. Note, however, that
13174 the Dynamic Audio Normalizer achieves this goal *without* applying
13175 "dynamic range compressing". It will retain 100% of the dynamic range
13176 *within* each section of the audio file.
13177 framelen, f
13179 Set the frame length in milliseconds. In range from 10 to 8000
13180 milliseconds. Default is 500 milliseconds. The Dynamic Audio
13181 Normalizer processes the input audio in small chunks, referred to
13182 as frames. This is required, because a peak magnitude has no
13183 meaning for just a single sample value. Instead, we need to
13184 determine the peak magnitude for a contiguous sequence of sample
13185 values. While a "standard" normalizer would simply use the peak
13186 magnitude of the complete file, the Dynamic Audio Normalizer
13187 determines the peak magnitude individually for each frame. The
13188 length of a frame is specified in milliseconds. By default, the
13189 Dynamic Audio Normalizer uses a frame length of 500 milliseconds,
13190 which has been found to give good results with most files. Note
13191 that the exact frame length, in number of samples, will be
13192 determined automatically, based on the sampling rate of the
13193 individual input audio file.
13194 gausssize, g
13196 Set the Gaussian filter window size. In range from 3 to 301, must
13197 be odd number. Default is 31. Probably the most important
13198 parameter of the Dynamic Audio Normalizer is the "window size" of
13199 the Gaussian smoothing filter. The filter's window size is
13200 specified in frames, centered around the current frame. For the
13201 sake of simplicity, this must be an odd number. Consequently, the
13202 default value of 31 takes into account the current frame, as well
13203 as the 15 preceding frames and the 15 subsequent frames. Using a
13204 larger window results in a stronger smoothing effect and thus in
13205 less gain variation, i.e. slower gain adaptation. Conversely, using
13206 a smaller window results in a weaker smoothing effect and thus in
13207 more gain variation, i.e. faster gain adaptation. In other words,
13208 the more you increase this value, the more the Dynamic Audio
13209 Normalizer will behave like a "traditional" normalization filter.
13210 On the contrary, the more you decrease this value, the more the
13211 Dynamic Audio Normalizer will behave like a dynamic range
13212 compressor.
13213 peak, p
13215 Set the target peak value. This specifies the highest permissible
13216 magnitude level for the normalized audio input. This filter will
13217 try to approach the target peak magnitude as closely as possible,
13218 but at the same time it also makes sure that the normalized signal
13219 will never exceed the peak magnitude. A frame's maximum local gain
13220 factor is imposed directly by the target peak magnitude. The
13221 default value is 0.95 and thus leaves a headroom of 5%*. It is not
13222 recommended to go above this value.
13223 maxgain, m
13225 Set the maximum gain factor. In range from 1.0 to 100.0. Default is
13226 10.0. The Dynamic Audio Normalizer determines the maximum possible
13227 (local) gain factor for each input frame, i.e. the maximum gain
13228 factor that does not result in clipping or distortion. The maximum
13229 gain factor is determined by the frame's highest magnitude sample.
13230 However, the Dynamic Audio Normalizer additionally bounds the
13231 frame's maximum gain factor by a predetermined (global) maximum
13232 gain factor. This is done in order to avoid excessive gain factors
13233 in "silent" or almost silent frames. By default, the maximum gain
13234 factor is 10.0, For most inputs the default value should be
13235 sufficient and it usually is not recommended to increase this
13236 value. Though, for input with an extremely low overall volume
13237 level, it may be necessary to allow even higher gain factors. Note,
13238 however, that the Dynamic Audio Normalizer does not simply apply a
13239 "hard" threshold (i.e. cut off values above the threshold).
13240 Instead, a "sigmoid" threshold function will be applied. This way,
13241 the gain factors will smoothly approach the threshold value, but
13242 never exceed that value.
13243 targetrms, r
13245 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 -
13246 disabled. By default, the Dynamic Audio Normalizer performs "peak"
13247 normalization. This means that the maximum local gain factor for
13248 each frame is defined (only) by the frame's highest magnitude
13249 sample. This way, the samples can be amplified as much as possible
13250 without exceeding the maximum signal level, i.e. without clipping.
13251 Optionally, however, the Dynamic Audio Normalizer can also take
13252 into account the frame's root mean square, abbreviated RMS. In
13253 electrical engineering, the RMS is commonly used to determine the
13254 power of a time-varying signal. It is therefore considered that the
13255 RMS is a better approximation of the "perceived loudness" than just
13256 looking at the signal's peak magnitude. Consequently, by adjusting
13257 all frames to a constant RMS value, a uniform "perceived loudness"
13258 can be established. If a target RMS value has been specified, a
13259 frame's local gain factor is defined as the factor that would
13260 result in exactly that RMS value. Note, however, that the maximum
13261 local gain factor is still restricted by the frame's highest
13262 magnitude sample, in order to prevent clipping.
13263 coupling, n
13265 Enable channels coupling. By default is enabled. By default, the
13266 Dynamic Audio Normalizer will amplify all channels by the same
13267 amount. This means the same gain factor will be applied to all
13268 channels, i.e. the maximum possible gain factor is determined by
13269 the "loudest" channel. However, in some recordings, it may happen
13270 that the volume of the different channels is uneven, e.g. one
13271 channel may be "quieter" than the other one(s). In this case, this
13272 option can be used to disable the channel coupling. This way, the
13273 gain factor will be determined independently for each channel,
13274 depending only on the individual channel's highest magnitude
13275 sample. This allows for harmonizing the volume of the different
13276 channels.
13277 correctdc, c
13279 Enable DC bias correction. By default is disabled. An audio signal
13280 (in the time domain) is a sequence of sample values. In the
13281 Dynamic Audio Normalizer these sample values are represented in the
13282 -1.0 to 1.0 range, regardless of the original input format.
13283 Normally, the audio signal, or "waveform", should be centered
13284 around the zero point. That means if we calculate the mean value
13285 of all samples in a file, or in a single frame, then the result
13286 should be 0.0 or at least very close to that value. If, however,
13287 there is a significant deviation of the mean value from 0.0, in
13288 either positive or negative direction, this is referred to as a DC
13289 bias or DC offset. Since a DC bias is clearly undesirable, the
13290 Dynamic Audio Normalizer provides optional DC bias correction.
13291 With DC bias correction enabled, the Dynamic Audio Normalizer will
13292 determine the mean value, or "DC correction" offset, of each input
13293 frame and subtract that value from all of the frame's sample values
13294 which ensures those samples are centered around 0.0 again. Also, in
13295 order to avoid "gaps" at the frame boundaries, the DC correction
13296 offset values will be interpolated smoothly between neighbouring
13297 frames.
13298 altboundary, b
13300 Enable alternative boundary mode. By default is disabled. The
13301 Dynamic Audio Normalizer takes into account a certain neighbourhood
13302 around each frame. This includes the preceding frames as well as
13303 the subsequent frames. However, for the "boundary" frames, located
13304 at the very beginning and at the very end of the audio file, not
13305 all neighbouring frames are available. In particular, for the first
13306 few frames in the audio file, the preceding frames are not known.
13307 And, similarly, for the last few frames in the audio file, the
13308 subsequent frames are not known. Thus, the question arises which
13309 gain factors should be assumed for the missing frames in the
13310 "boundary" region. The Dynamic Audio Normalizer implements two
13311 modes to deal with this situation. The default boundary mode
13312 assumes a gain factor of exactly 1.0 for the missing frames,
13313 resulting in a smooth "fade in" and "fade out" at the beginning and
13314 at the end of the input, respectively.
13315 compress, s
13317 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
13318 By default, the Dynamic Audio Normalizer does not apply
13319 "traditional" compression. This means that signal peaks will not be
13320 pruned and thus the full dynamic range will be retained within each
13321 local neighbourhood. However, in some cases it may be desirable to
13322 combine the Dynamic Audio Normalizer's normalization algorithm with
13323 a more "traditional" compression. For this purpose, the Dynamic
13324 Audio Normalizer provides an optional compression (thresholding)
13325 function. If (and only if) the compression feature is enabled, all
13326 input frames will be processed by a soft knee thresholding function
13327 prior to the actual normalization process. Put simply, the
13328 thresholding function is going to prune all samples whose magnitude
13329 exceeds a certain threshold value. However, the Dynamic Audio
13330 Normalizer does not simply apply a fixed threshold value. Instead,
13331 the threshold value will be adjusted for each individual frame. In
13332 general, smaller parameters result in stronger compression, and
13333 vice versa. Values below 3.0 are not recommended, because audible
13334 distortion may appear.
13335 threshold, t
13337 Set the target threshold value. This specifies the lowest
13338 permissible magnitude level for the audio input which will be
13339 normalized. If input frame volume is above this value frame will
13340 be normalized. Otherwise frame may not be normalized at all. The
13341 default value is set to 0, which means all input frames will be
13342 normalized. This option is mostly useful if digital noise is not
13343 wanted to be amplified.
13344 channels, h
13346 Specify which channels to filter, by default all available channels
13347 are filtered.
13348 overlap, o
13350 Specify overlap for frames. If set to 0 (default) no frame
13351 overlapping is done. Using >0 and <1 values will make less
13352 conservative gain adjustments, like when framelen option is set to
13353 smaller value, if framelen option value is compensated for non-zero
13354 overlap then gain adjustments will be smoother across time compared
13355 to zero overlap case.
13356 Commands
13358 This filter supports the all above options as commands.
13360 earwax
13362 Make audio easier to listen to on headphones.
13363 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
13365 so that when listened to on headphones the stereo image is moved from
13366 inside your head (standard for headphones) to outside and in front of
13367 the listener (standard for speakers).
13368 Ported from SoX.
13370 equalizer
13372 Apply a two-pole peaking equalisation (EQ) filter. With this filter,
13373 the signal-level at and around a selected frequency can be increased or
13374 decreased, whilst (unlike bandpass and bandreject filters) that at all
13375 other frequencies is unchanged.
13376 In order to produce complex equalisation curves, this filter can be
13378 given several times, each with a different central frequency.
13379 The filter accepts the following options:
13381 frequency, f
13383 Set the filter's central frequency in Hz.
13384 width_type, t
13386 Set method to specify band-width of filter.
13387 h Hz
13389 q Q-Factor
13391 o octave
13393 s slope
13395 k kHz
13397 width, w
13399 Specify the band-width of a filter in width_type units.
13400 gain, g
13402 Set the required gain or attenuation in dB. Beware of clipping
13403 when using a positive gain.
13404 mix, m
13406 How much to use filtered signal in output. Default is 1. Range is
13407 between 0 and 1.
13408 channels, c
13410 Specify which channels to filter, by default all available are
13411 filtered.
13412 normalize, n
13414 Normalize biquad coefficients, by default is disabled. Enabling it
13415 will normalize magnitude response at DC to 0dB.
13416 transform, a
13418 Set transform type of IIR filter.
13419 di
13421 dii
13422 tdi
13423 tdii
13424 latt
13425 svf
13426 zdf
13427 precision, r
13428 Set precison of filtering.
13429 auto
13431 Pick automatic sample format depending on surround filters.
13432 s16 Always use signed 16-bit.
13434 s32 Always use signed 32-bit.
13436 f32 Always use float 32-bit.
13438 f64 Always use float 64-bit.
13440 block_size, b
13442 Set block size used for reverse IIR processing. If this value is
13443 set to high enough value (higher than impulse response length
13444 truncated when reaches near zero values) filtering will become
13445 linear phase otherwise if not big enough it will just produce nasty
13446 artifacts.
13447 Note that filter delay will be exactly this many samples when set
13449 to non-zero value.
13450 Examples
13452 • Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
13454 equalizer=f=1000:t=h:width=200:g=-10
13456 • Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz
13458 with Q 2:
13459 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
13461 Commands
13463 This filter supports the following commands:
13465 frequency, f
13467 Change equalizer frequency. Syntax for the command is :
13468 "frequency"
13469 width_type, t
13471 Change equalizer width_type. Syntax for the command is :
13472 "width_type"
13473 width, w
13475 Change equalizer width. Syntax for the command is : "width"
13476 gain, g
13478 Change equalizer gain. Syntax for the command is : "gain"
13479 mix, m
13481 Change equalizer mix. Syntax for the command is : "mix"
13482 extrastereo
13484 Linearly increases the difference between left and right channels which
13485 adds some sort of "live" effect to playback.
13486 The filter accepts the following options:
13488 m Sets the difference coefficient (default: 2.5). 0.0 means mono
13490 sound (average of both channels), with 1.0 sound will be unchanged,
13491 with -1.0 left and right channels will be swapped.
13492 c Enable clipping. By default is enabled.
13494 Commands
13496 This filter supports the all above options as commands.
13498 firequalizer
13500 Apply FIR Equalization using arbitrary frequency response.
13501 The filter accepts the following option:
13503 gain
13505 Set gain curve equation (in dB). The expression can contain
13506 variables:
13507 f the evaluated frequency
13509 sr sample rate
13511 ch channel number, set to 0 when multichannels evaluation is
13513 disabled
13514 chid
13516 channel id, see libavutil/channel_layout.h, set to the first
13517 channel id when multichannels evaluation is disabled
13518 chs number of channels
13520 chlayout
13522 channel_layout, see libavutil/channel_layout.h
13523 and functions:
13525 gain_interpolate(f)
13527 interpolate gain on frequency f based on gain_entry
13528 cubic_interpolate(f)
13530 same as gain_interpolate, but smoother
13531 This option is also available as command. Default is
13533 gain_interpolate(f).
13534 gain_entry
13536 Set gain entry for gain_interpolate function. The expression can
13537 contain functions:
13538 entry(f, g)
13540 store gain entry at frequency f with value g
13541 This option is also available as command.
13543 delay
13545 Set filter delay in seconds. Higher value means more accurate.
13546 Default is 0.01.
13547 accuracy
13549 Set filter accuracy in Hz. Lower value means more accurate.
13550 Default is 5.
13551 wfunc
13553 Set window function. Acceptable values are:
13554 rectangular
13556 rectangular window, useful when gain curve is already smooth
13557 hann
13559 hann window (default)
13560 hamming
13562 hamming window
13563 blackman
13565 blackman window
13566 nuttall3
13568 3-terms continuous 1st derivative nuttall window
13569 mnuttall3
13571 minimum 3-terms discontinuous nuttall window
13572 nuttall
13574 4-terms continuous 1st derivative nuttall window
13575 bnuttall
13577 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
13578 bharris
13580 blackman-harris window
13581 tukey
13583 tukey window
13584 fixed
13586 If enabled, use fixed number of audio samples. This improves speed
13587 when filtering with large delay. Default is disabled.
13588 multi
13590 Enable multichannels evaluation on gain. Default is disabled.
13591 zero_phase
13593 Enable zero phase mode by subtracting timestamp to compensate
13594 delay. Default is disabled.
13595 scale
13597 Set scale used by gain. Acceptable values are:
13598 linlin
13600 linear frequency, linear gain
13601 linlog
13603 linear frequency, logarithmic (in dB) gain (default)
13604 loglin
13606 logarithmic (in octave scale where 20 Hz is 0) frequency,
13607 linear gain
13608 loglog
13610 logarithmic frequency, logarithmic gain
13611 dumpfile
13613 Set file for dumping, suitable for gnuplot.
13614 dumpscale
13616 Set scale for dumpfile. Acceptable values are same with scale
13617 option. Default is linlog.
13618 fft2
13620 Enable 2-channel convolution using complex FFT. This improves speed
13621 significantly. Default is disabled.
13622 min_phase
13624 Enable minimum phase impulse response. Default is disabled.
13625 Examples
13627 • lowpass at 1000 Hz:
13629 firequalizer=gain='if(lt(f,1000), 0, -INF)'
13631 • lowpass at 1000 Hz with gain_entry:
13633 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
13635 • custom equalization:
13637 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
13639 • higher delay with zero phase to compensate delay:
13641 firequalizer=delay=0.1:fixed=on:zero_phase=on
13643 • lowpass on left channel, highpass on right channel:
13645 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
13647 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
13648 flanger
13650 Apply a flanging effect to the audio.
13651 The filter accepts the following options:
13653 delay
13655 Set base delay in milliseconds. Range from 0 to 30. Default value
13656 is 0.
13657 depth
13659 Set added sweep delay in milliseconds. Range from 0 to 10. Default
13660 value is 2.
13661 regen
13663 Set percentage regeneration (delayed signal feedback). Range from
13664 -95 to 95. Default value is 0.
13665 width
13667 Set percentage of delayed signal mixed with original. Range from 0
13668 to 100. Default value is 71.
13669 speed
13671 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is
13672 0.5.
13673 shape
13675 Set swept wave shape, can be triangular or sinusoidal. Default
13676 value is sinusoidal.
13677 phase
13679 Set swept wave percentage-shift for multi channel. Range from 0 to
13680 100. Default value is 25.
13681 interp
13683 Set delay-line interpolation, linear or quadratic. Default is
13684 linear.
13685 haas
13687 Apply Haas effect to audio.
13688 Note that this makes most sense to apply on mono signals. With this
13690 filter applied to mono signals it give some directionality and
13691 stretches its stereo image.
13692 The filter accepts the following options:
13694 level_in
13696 Set input level. By default is 1, or 0dB
13697 level_out
13699 Set output level. By default is 1, or 0dB.
13700 side_gain
13702 Set gain applied to side part of signal. By default is 1.
13703 middle_source
13705 Set kind of middle source. Can be one of the following:
13706 left
13708 Pick left channel.
13709 right
13711 Pick right channel.
13712 mid Pick middle part signal of stereo image.
13714 side
13716 Pick side part signal of stereo image.
13717 middle_phase
13719 Change middle phase. By default is disabled.
13720 left_delay
13722 Set left channel delay. By default is 2.05 milliseconds.
13723 left_balance
13725 Set left channel balance. By default is -1.
13726 left_gain
13728 Set left channel gain. By default is 1.
13729 left_phase
13731 Change left phase. By default is disabled.
13732 right_delay
13734 Set right channel delay. By defaults is 2.12 milliseconds.
13735 right_balance
13737 Set right channel balance. By default is 1.
13738 right_gain
13740 Set right channel gain. By default is 1.
13741 right_phase
13743 Change right phase. By default is enabled.
13744 hdcd
13746 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM
13747 stream with embedded HDCD codes is expanded into a 20-bit PCM stream.
13748 The filter supports the Peak Extend and Low-level Gain Adjustment
13750 features of HDCD, and detects the Transient Filter flag.
13751 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
13753 When using the filter with wav, note the default encoding for wav is
13755 16-bit, so the resulting 20-bit stream will be truncated back to
13756 16-bit. Use something like -acodec pcm_s24le after the filter to get
13757 24-bit PCM output.
13758 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
13760 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
13761 The filter accepts the following options:
13763 disable_autoconvert
13765 Disable any automatic format conversion or resampling in the filter
13766 graph.
13767 process_stereo
13769 Process the stereo channels together. If target_gain does not match
13770 between channels, consider it invalid and use the last valid
13771 target_gain.
13772 cdt_ms
13774 Set the code detect timer period in ms.
13775 force_pe
13777 Always extend peaks above -3dBFS even if PE isn't signaled.
13778 analyze_mode
13780 Replace audio with a solid tone and adjust the amplitude to signal
13781 some specific aspect of the decoding process. The output file can
13782 be loaded in an audio editor alongside the original to aid
13783 analysis.
13784 "analyze_mode=pe:force_pe=true" can be used to see all samples
13786 above the PE level.
13787 Modes are:
13789 0, off
13791 Disabled
13792 1, lle
13794 Gain adjustment level at each sample
13795 2, pe
13797 Samples where peak extend occurs
13798 3, cdt
13800 Samples where the code detect timer is active
13801 4, tgm
13803 Samples where the target gain does not match between channels
13804 headphone
13806 Apply head-related transfer functions (HRTFs) to create virtual
13807 loudspeakers around the user for binaural listening via headphones.
13808 The HRIRs are provided via additional streams, for each channel one
13809 stereo input stream is needed.
13810 The filter accepts the following options:
13812 map Set mapping of input streams for convolution. The argument is a
13814 '|'-separated list of channel names in order as they are given as
13815 additional stream inputs for filter. This also specify number of
13816 input streams. Number of input streams must be not less than number
13817 of channels in first stream plus one.
13818 gain
13820 Set gain applied to audio. Value is in dB. Default is 0.
13821 type
13823 Set processing type. Can be time or freq. time is processing audio
13824 in time domain which is slow. freq is processing audio in
13825 frequency domain which is fast. Default is freq.
13826 lfe Set custom gain for LFE channels. Value is in dB. Default is 0.
13828 size
13830 Set size of frame in number of samples which will be processed at
13831 once. Default value is 1024. Allowed range is from 1024 to 96000.
13832 hrir
13834 Set format of hrir stream. Default value is stereo. Alternative
13835 value is multich. If value is set to stereo, number of additional
13836 streams should be greater or equal to number of input channels in
13837 first input stream. Also each additional stream should have stereo
13838 number of channels. If value is set to multich, number of
13839 additional streams should be exactly one. Also number of input
13840 channels of additional stream should be equal or greater than twice
13841 number of channels of first input stream.
13842 Examples
13844 • Full example using wav files as coefficients with amovie filters
13846 for 7.1 downmix, each amovie filter use stereo file with IR
13847 coefficients as input. The files give coefficients for each
13848 position of virtual loudspeaker:
13849 ffmpeg -i input.wav
13851 -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"
13852 output.wav
13853 • Full example using wav files as coefficients with amovie filters
13855 for 7.1 downmix, but now in multich hrir format.
13856 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"
13858 output.wav
13859 highpass
13861 Apply a high-pass filter with 3dB point frequency. The filter can be
13862 either single-pole, or double-pole (the default). The filter roll off
13863 at 6dB per pole per octave (20dB per pole per decade).
13864 The filter accepts the following options:
13866 frequency, f
13868 Set frequency in Hz. Default is 3000.
13869 poles, p
13871 Set number of poles. Default is 2.
13872 width_type, t
13874 Set method to specify band-width of filter.
13875 h Hz
13877 q Q-Factor
13879 o octave
13881 s slope
13883 k kHz
13885 width, w
13887 Specify the band-width of a filter in width_type units. Applies
13888 only to double-pole filter. The default is 0.707q and gives a
13889 Butterworth response.
13890 mix, m
13892 How much to use filtered signal in output. Default is 1. Range is
13893 between 0 and 1.
13894 channels, c
13896 Specify which channels to filter, by default all available are
13897 filtered.
13898 normalize, n
13900 Normalize biquad coefficients, by default is disabled. Enabling it
13901 will normalize magnitude response at DC to 0dB.
13902 transform, a
13904 Set transform type of IIR filter.
13905 di
13907 dii
13908 tdi
13909 tdii
13910 latt
13911 svf
13912 zdf
13913 precision, r
13914 Set precison of filtering.
13915 auto
13917 Pick automatic sample format depending on surround filters.
13918 s16 Always use signed 16-bit.
13920 s32 Always use signed 32-bit.
13922 f32 Always use float 32-bit.
13924 f64 Always use float 64-bit.
13926 block_size, b
13928 Set block size used for reverse IIR processing. If this value is
13929 set to high enough value (higher than impulse response length
13930 truncated when reaches near zero values) filtering will become
13931 linear phase otherwise if not big enough it will just produce nasty
13932 artifacts.
13933 Note that filter delay will be exactly this many samples when set
13935 to non-zero value.
13936 Commands
13938 This filter supports the following commands:
13940 frequency, f
13942 Change highpass frequency. Syntax for the command is : "frequency"
13943 width_type, t
13945 Change highpass width_type. Syntax for the command is :
13946 "width_type"
13947 width, w
13949 Change highpass width. Syntax for the command is : "width"
13950 mix, m
13952 Change highpass mix. Syntax for the command is : "mix"
13953 join
13955 Join multiple input streams into one multi-channel stream.
13956 It accepts the following parameters:
13958 inputs
13960 The number of input streams. It defaults to 2.
13961 channel_layout
13963 The desired output channel layout. It defaults to stereo.
13964 map Map channels from inputs to output. The argument is a '|'-separated
13966 list of mappings, each in the "input_idx.in_channel-out_channel"
13967 form. input_idx is the 0-based index of the input stream.
13968 in_channel can be either the name of the input channel (e.g. FL for
13969 front left) or its index in the specified input stream. out_channel
13970 is the name of the output channel.
13971 The filter will attempt to guess the mappings when they are not
13973 specified explicitly. It does so by first trying to find an unused
13974 matching input channel and if that fails it picks the first unused
13975 input channel.
13976 Join 3 inputs (with properly set channel layouts):
13978 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
13980 Build a 5.1 output from 6 single-channel streams:
13982 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
13984 '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'
13985 out
13986 ladspa
13988 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
13989 To enable compilation of this filter you need to configure FFmpeg with
13991 "--enable-ladspa".
13992 file, f
13994 Specifies the name of LADSPA plugin library to load. If the
13995 environment variable LADSPA_PATH is defined, the LADSPA plugin is
13996 searched in each one of the directories specified by the colon
13997 separated list in LADSPA_PATH, otherwise in the standard LADSPA
13998 paths, which are in this order: HOME/.ladspa/lib/,
13999 /usr/local/lib/ladspa/, /usr/lib/ladspa/.
14000 plugin, p
14002 Specifies the plugin within the library. Some libraries contain
14003 only one plugin, but others contain many of them. If this is not
14004 set filter will list all available plugins within the specified
14005 library.
14006 controls, c
14008 Set the '|' separated list of controls which are zero or more
14009 floating point values that determine the behavior of the loaded
14010 plugin (for example delay, threshold or gain). Controls need to be
14011 defined using the following syntax:
14012 c0=value0|c1=value1|c2=value2|..., where valuei is the value set on
14013 the i-th control. Alternatively they can be also defined using the
14014 following syntax: value0|value1|value2|..., where valuei is the
14015 value set on the i-th control. If controls is set to "help", all
14016 available controls and their valid ranges are printed.
14017 sample_rate, s
14019 Specify the sample rate, default to 44100. Only used if plugin have
14020 zero inputs.
14021 nb_samples, n
14023 Set the number of samples per channel per each output frame,
14024 default is 1024. Only used if plugin have zero inputs.
14025 duration, d
14027 Set the minimum duration of the sourced audio. See the Time
14028 duration section in the ffmpeg-utils(1) manual for the accepted
14029 syntax. Note that the resulting duration may be greater than the
14030 specified duration, as the generated audio is always cut at the end
14031 of a complete frame. If not specified, or the expressed duration
14032 is negative, the audio is supposed to be generated forever. Only
14033 used if plugin have zero inputs.
14034 latency, l
14036 Enable latency compensation, by default is disabled. Only used if
14037 plugin have inputs.
14038 Examples
14040 • List all available plugins within amp (LADSPA example plugin)
14042 library:
14043 ladspa=file=amp
14045 • List all available controls and their valid ranges for "vcf_notch"
14047 plugin from "VCF" library:
14048 ladspa=f=vcf:p=vcf_notch:c=help
14050 • Simulate low quality audio equipment using "Computer Music Toolkit"
14052 (CMT) plugin library:
14053 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
14055 • Add reverberation to the audio using TAP-plugins (Tom's Audio
14057 Processing plugins):
14058 ladspa=file=tap_reverb:tap_reverb
14060 • Generate white noise, with 0.2 amplitude:
14062 ladspa=file=cmt:noise_source_white:c=c0=.2
14064 • Generate 20 bpm clicks using plugin "C* Click - Metronome" from the
14066 "C* Audio Plugin Suite" (CAPS) library:
14067 ladspa=file=caps:Click:c=c1=20'
14069 • Apply "C* Eq10X2 - Stereo 10-band equaliser" effect:
14071 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
14073 • Increase volume by 20dB using fast lookahead limiter from Steve
14075 Harris "SWH Plugins" collection:
14076 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
14078 • Attenuate low frequencies using Multiband EQ from Steve Harris "SWH
14080 Plugins" collection:
14081 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
14083 • Reduce stereo image using "Narrower" from the "C* Audio Plugin
14085 Suite" (CAPS) library:
14086 ladspa=caps:Narrower
14088 • Another white noise, now using "C* Audio Plugin Suite" (CAPS)
14090 library:
14091 ladspa=caps:White:.2
14093 • Some fractal noise, using "C* Audio Plugin Suite" (CAPS) library:
14095 ladspa=caps:Fractal:c=c1=1
14097 • Dynamic volume normalization using "VLevel" plugin:
14099 ladspa=vlevel-ladspa:vlevel_mono
14101 Commands
14103 This filter supports the following commands:
14105 cN Modify the N-th control value.
14107 If the specified value is not valid, it is ignored and prior one is
14109 kept.
14110 loudnorm
14112 EBU R128 loudness normalization. Includes both dynamic and linear
14113 normalization modes. Support for both single pass (livestreams, files)
14114 and double pass (files) modes. This algorithm can target IL, LRA, and
14115 maximum true peak. In dynamic mode, to accurately detect true peaks,
14116 the audio stream will be upsampled to 192 kHz. Use the "-ar" option or
14117 "aresample" filter to explicitly set an output sample rate.
14118 The filter accepts the following options:
14120 I, i
14122 Set integrated loudness target. Range is -70.0 - -5.0. Default
14123 value is -24.0.
14124 LRA, lra
14126 Set loudness range target. Range is 1.0 - 50.0. Default value is
14127 7.0.
14128 TP, tp
14130 Set maximum true peak. Range is -9.0 - +0.0. Default value is
14131 -2.0.
14132 measured_I, measured_i
14134 Measured IL of input file. Range is -99.0 - +0.0.
14135 measured_LRA, measured_lra
14137 Measured LRA of input file. Range is 0.0 - 99.0.
14138 measured_TP, measured_tp
14140 Measured true peak of input file. Range is -99.0 - +99.0.
14141 measured_thresh
14143 Measured threshold of input file. Range is -99.0 - +0.0.
14144 offset
14146 Set offset gain. Gain is applied before the true-peak limiter.
14147 Range is -99.0 - +99.0. Default is +0.0.
14148 linear
14150 Normalize by linearly scaling the source audio. "measured_I",
14151 "measured_LRA", "measured_TP", and "measured_thresh" must all be
14152 specified. Target LRA shouldn't be lower than source LRA and the
14153 change in integrated loudness shouldn't result in a true peak which
14154 exceeds the target TP. If any of these conditions aren't met,
14155 normalization mode will revert to dynamic. Options are "true" or
14156 "false". Default is "true".
14157 dual_mono
14159 Treat mono input files as "dual-mono". If a mono file is intended
14160 for playback on a stereo system, its EBU R128 measurement will be
14161 perceptually incorrect. If set to "true", this option will
14162 compensate for this effect. Multi-channel input files are not
14163 affected by this option. Options are true or false. Default is
14164 false.
14165 print_format
14167 Set print format for stats. Options are summary, json, or none.
14168 Default value is none.
14169 lowpass
14171 Apply a low-pass filter with 3dB point frequency. The filter can be
14172 either single-pole or double-pole (the default). The filter roll off
14173 at 6dB per pole per octave (20dB per pole per decade).
14174 The filter accepts the following options:
14176 frequency, f
14178 Set frequency in Hz. Default is 500.
14179 poles, p
14181 Set number of poles. Default is 2.
14182 width_type, t
14184 Set method to specify band-width of filter.
14185 h Hz
14187 q Q-Factor
14189 o octave
14191 s slope
14193 k kHz
14195 width, w
14197 Specify the band-width of a filter in width_type units. Applies
14198 only to double-pole filter. The default is 0.707q and gives a
14199 Butterworth response.
14200 mix, m
14202 How much to use filtered signal in output. Default is 1. Range is
14203 between 0 and 1.
14204 channels, c
14206 Specify which channels to filter, by default all available are
14207 filtered.
14208 normalize, n
14210 Normalize biquad coefficients, by default is disabled. Enabling it
14211 will normalize magnitude response at DC to 0dB.
14212 transform, a
14214 Set transform type of IIR filter.
14215 di
14217 dii
14218 tdi
14219 tdii
14220 latt
14221 svf
14222 zdf
14223 precision, r
14224 Set precison of filtering.
14225 auto
14227 Pick automatic sample format depending on surround filters.
14228 s16 Always use signed 16-bit.
14230 s32 Always use signed 32-bit.
14232 f32 Always use float 32-bit.
14234 f64 Always use float 64-bit.
14236 block_size, b
14238 Set block size used for reverse IIR processing. If this value is
14239 set to high enough value (higher than impulse response length
14240 truncated when reaches near zero values) filtering will become
14241 linear phase otherwise if not big enough it will just produce nasty
14242 artifacts.
14243 Note that filter delay will be exactly this many samples when set
14245 to non-zero value.
14246 Examples
14248 • Lowpass only LFE channel, it LFE is not present it does nothing:
14250 lowpass=c=LFE
14252 Commands
14254 This filter supports the following commands:
14256 frequency, f
14258 Change lowpass frequency. Syntax for the command is : "frequency"
14259 width_type, t
14261 Change lowpass width_type. Syntax for the command is :
14262 "width_type"
14263 width, w
14265 Change lowpass width. Syntax for the command is : "width"
14266 mix, m
14268 Change lowpass mix. Syntax for the command is : "mix"
14269 lv2
14271 Load a LV2 (LADSPA Version 2) plugin.
14272 To enable compilation of this filter you need to configure FFmpeg with
14274 "--enable-lv2".
14275 plugin, p
14277 Specifies the plugin URI. You may need to escape ':'.
14278 controls, c
14280 Set the '|' separated list of controls which are zero or more
14281 floating point values that determine the behavior of the loaded
14282 plugin (for example delay, threshold or gain). If controls is set
14283 to "help", all available controls and their valid ranges are
14284 printed.
14285 sample_rate, s
14287 Specify the sample rate, default to 44100. Only used if plugin have
14288 zero inputs.
14289 nb_samples, n
14291 Set the number of samples per channel per each output frame,
14292 default is 1024. Only used if plugin have zero inputs.
14293 duration, d
14295 Set the minimum duration of the sourced audio. See the Time
14296 duration section in the ffmpeg-utils(1) manual for the accepted
14297 syntax. Note that the resulting duration may be greater than the
14298 specified duration, as the generated audio is always cut at the end
14299 of a complete frame. If not specified, or the expressed duration
14300 is negative, the audio is supposed to be generated forever. Only
14301 used if plugin have zero inputs.
14302 Examples
14304 • Apply bass enhancer plugin from Calf:
14306 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
14308 • Apply vinyl plugin from Calf:
14310 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
14312 • Apply bit crusher plugin from ArtyFX:
14314 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
14316 Commands
14318 This filter supports all options that are exported by plugin as
14320 commands.
14321 mcompand
14323 Multiband Compress or expand the audio's dynamic range.
14324 The input audio is divided into bands using 4th order Linkwitz-Riley
14326 IIRs. This is akin to the crossover of a loudspeaker, and results in
14327 flat frequency response when absent compander action.
14328 It accepts the following parameters:
14330 args
14332 This option syntax is: attack,decay,[attack,decay..] soft-knee
14333 points crossover_frequency [delay [initial_volume [gain]]] |
14334 attack,decay ... For explanation of each item refer to compand
14335 filter documentation.
14336 pan
14338 Mix channels with specific gain levels. The filter accepts the output
14339 channel layout followed by a set of channels definitions.
14340 This filter is also designed to efficiently remap the channels of an
14342 audio stream.
14343 The filter accepts parameters of the form: "l|outdef|outdef|..."
14345 l output channel layout or number of channels
14347 outdef
14349 output channel specification, of the form:
14350 "out_name=[gain*]in_name[(+-)[gain*]in_name...]"
14351 out_name
14353 output channel to define, either a channel name (FL, FR, etc.) or a
14354 channel number (c0, c1, etc.)
14355 gain
14357 multiplicative coefficient for the channel, 1 leaving the volume
14358 unchanged
14359 in_name
14361 input channel to use, see out_name for details; it is not possible
14362 to mix named and numbered input channels
14363 If the `=' in a channel specification is replaced by `<', then the
14365 gains for that specification will be renormalized so that the total is
14366 1, thus avoiding clipping noise.
14367 Mixing examples
14369 For example, if you want to down-mix from stereo to mono, but with a
14371 bigger factor for the left channel:
14372 pan=1c|c0=0.9*c0+0.1*c1
14374 A customized down-mix to stereo that works automatically for 3-, 4-, 5-
14376 and 7-channels surround:
14377 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
14379 Note that ffmpeg integrates a default down-mix (and up-mix) system that
14381 should be preferred (see "-ac" option) unless you have very specific
14382 needs.
14383 Remapping examples
14385 The channel remapping will be effective if, and only if:
14387 *<gain coefficients are zeroes or ones,>
14389 *<only one input per channel output,>
14390 If all these conditions are satisfied, the filter will notify the user
14392 ("Pure channel mapping detected"), and use an optimized and lossless
14393 method to do the remapping.
14394 For example, if you have a 5.1 source and want a stereo audio stream by
14396 dropping the extra channels:
14397 pan="stereo| c0=FL | c1=FR"
14399 Given the same source, you can also switch front left and front right
14401 channels and keep the input channel layout:
14402 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
14404 If the input is a stereo audio stream, you can mute the front left
14406 channel (and still keep the stereo channel layout) with:
14407 pan="stereo|c1=c1"
14409 Still with a stereo audio stream input, you can copy the right channel
14411 in both front left and right:
14412 pan="stereo| c0=FR | c1=FR"
14414 replaygain
14416 ReplayGain scanner filter. This filter takes an audio stream as an
14417 input and outputs it unchanged. At end of filtering it displays
14418 "track_gain" and "track_peak".
14419 resample
14421 Convert the audio sample format, sample rate and channel layout. It is
14422 not meant to be used directly.
14423 rubberband
14425 Apply time-stretching and pitch-shifting with librubberband.
14426 To enable compilation of this filter, you need to configure FFmpeg with
14428 "--enable-librubberband".
14429 The filter accepts the following options:
14431 tempo
14433 Set tempo scale factor.
14434 pitch
14436 Set pitch scale factor.
14437 transients
14439 Set transients detector. Possible values are:
14440 crisp
14442 mixed
14443 smooth
14444 detector
14445 Set detector. Possible values are:
14446 compound
14448 percussive
14449 soft
14450 phase
14451 Set phase. Possible values are:
14452 laminar
14454 independent
14455 window
14456 Set processing window size. Possible values are:
14457 standard
14459 short
14460 long
14461 smoothing
14462 Set smoothing. Possible values are:
14463 off
14465 on
14466 formant
14467 Enable formant preservation when shift pitching. Possible values
14468 are:
14469 shifted
14471 preserved
14472 pitchq
14473 Set pitch quality. Possible values are:
14474 quality
14476 speed
14477 consistency
14478 channels
14479 Set channels. Possible values are:
14480 apart
14482 together
14483 Commands
14485 This filter supports the following commands:
14487 tempo
14489 Change filter tempo scale factor. Syntax for the command is :
14490 "tempo"
14491 pitch
14493 Change filter pitch scale factor. Syntax for the command is :
14494 "pitch"
14495 sidechaincompress
14497 This filter acts like normal compressor but has the ability to compress
14498 detected signal using second input signal. It needs two input streams
14499 and returns one output stream. First input stream will be processed
14500 depending on second stream signal. The filtered signal then can be
14501 filtered with other filters in later stages of processing. See pan and
14502 amerge filter.
14503 The filter accepts the following options:
14505 level_in
14507 Set input gain. Default is 1. Range is between 0.015625 and 64.
14508 mode
14510 Set mode of compressor operation. Can be "upward" or "downward".
14511 Default is "downward".
14512 threshold
14514 If a signal of second stream raises above this level it will affect
14515 the gain reduction of first stream. By default is 0.125. Range is
14516 between 0.00097563 and 1.
14517 ratio
14519 Set a ratio about which the signal is reduced. 1:2 means that if
14520 the level raised 4dB above the threshold, it will be only 2dB above
14521 after the reduction. Default is 2. Range is between 1 and 20.
14522 attack
14524 Amount of milliseconds the signal has to rise above the threshold
14525 before gain reduction starts. Default is 20. Range is between 0.01
14526 and 2000.
14527 release
14529 Amount of milliseconds the signal has to fall below the threshold
14530 before reduction is decreased again. Default is 250. Range is
14531 between 0.01 and 9000.
14532 makeup
14534 Set the amount by how much signal will be amplified after
14535 processing. Default is 1. Range is from 1 to 64.
14536 knee
14538 Curve the sharp knee around the threshold to enter gain reduction
14539 more softly. Default is 2.82843. Range is between 1 and 8.
14540 link
14542 Choose if the "average" level between all channels of side-chain
14543 stream or the louder("maximum") channel of side-chain stream
14544 affects the reduction. Default is "average".
14545 detection
14547 Should the exact signal be taken in case of "peak" or an RMS one in
14548 case of "rms". Default is "rms" which is mainly smoother.
14549 level_sc
14551 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
14552 mix How much to use compressed signal in output. Default is 1. Range
14554 is between 0 and 1.
14555 Commands
14557 This filter supports the all above options as commands.
14559 Examples
14561 • Full ffmpeg example taking 2 audio inputs, 1st input to be
14563 compressed depending on the signal of 2nd input and later
14564 compressed signal to be merged with 2nd input:
14565 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
14567 sidechaingate
14569 A sidechain gate acts like a normal (wideband) gate but has the ability
14570 to filter the detected signal before sending it to the gain reduction
14571 stage. Normally a gate uses the full range signal to detect a level
14572 above the threshold. For example: If you cut all lower frequencies
14573 from your sidechain signal the gate will decrease the volume of your
14574 track only if not enough highs appear. With this technique you are able
14575 to reduce the resonation of a natural drum or remove "rumbling" of
14576 muted strokes from a heavily distorted guitar. It needs two input
14577 streams and returns one output stream. First input stream will be
14578 processed depending on second stream signal.
14579 The filter accepts the following options:
14581 level_in
14583 Set input level before filtering. Default is 1. Allowed range is
14584 from 0.015625 to 64.
14585 mode
14587 Set the mode of operation. Can be "upward" or "downward". Default
14588 is "downward". If set to "upward" mode, higher parts of signal will
14589 be amplified, expanding dynamic range in upward direction.
14590 Otherwise, in case of "downward" lower parts of signal will be
14591 reduced.
14592 range
14594 Set the level of gain reduction when the signal is below the
14595 threshold. Default is 0.06125. Allowed range is from 0 to 1.
14596 Setting this to 0 disables reduction and then filter behaves like
14597 expander.
14598 threshold
14600 If a signal rises above this level the gain reduction is released.
14601 Default is 0.125. Allowed range is from 0 to 1.
14602 ratio
14604 Set a ratio about which the signal is reduced. Default is 2.
14605 Allowed range is from 1 to 9000.
14606 attack
14608 Amount of milliseconds the signal has to rise above the threshold
14609 before gain reduction stops. Default is 20 milliseconds. Allowed
14610 range is from 0.01 to 9000.
14611 release
14613 Amount of milliseconds the signal has to fall below the threshold
14614 before the reduction is increased again. Default is 250
14615 milliseconds. Allowed range is from 0.01 to 9000.
14616 makeup
14618 Set amount of amplification of signal after processing. Default is
14619 1. Allowed range is from 1 to 64.
14620 knee
14622 Curve the sharp knee around the threshold to enter gain reduction
14623 more softly. Default is 2.828427125. Allowed range is from 1 to 8.
14624 detection
14626 Choose if exact signal should be taken for detection or an RMS like
14627 one. Default is rms. Can be peak or rms.
14628 link
14630 Choose if the average level between all channels or the louder
14631 channel affects the reduction. Default is average. Can be average
14632 or maximum.
14633 level_sc
14635 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
14636 Commands
14638 This filter supports the all above options as commands.
14640 silencedetect
14642 Detect silence in an audio stream.
14643 This filter logs a message when it detects that the input audio volume
14645 is less or equal to a noise tolerance value for a duration greater or
14646 equal to the minimum detected noise duration.
14647 The printed times and duration are expressed in seconds. The
14649 "lavfi.silence_start" or "lavfi.silence_start.X" metadata key is set on
14650 the first frame whose timestamp equals or exceeds the detection
14651 duration and it contains the timestamp of the first frame of the
14652 silence.
14653 The "lavfi.silence_duration" or "lavfi.silence_duration.X" and
14655 "lavfi.silence_end" or "lavfi.silence_end.X" metadata keys are set on
14656 the first frame after the silence. If mono is enabled, and each channel
14657 is evaluated separately, the ".X" suffixed keys are used, and "X"
14658 corresponds to the channel number.
14659 The filter accepts the following options:
14661 noise, n
14663 Set noise tolerance. Can be specified in dB (in case "dB" is
14664 appended to the specified value) or amplitude ratio. Default is
14665 -60dB, or 0.001.
14666 duration, d
14668 Set silence duration until notification (default is 2 seconds). See
14669 the Time duration section in the ffmpeg-utils(1) manual for the
14670 accepted syntax.
14671 mono, m
14673 Process each channel separately, instead of combined. By default is
14674 disabled.
14675 Examples
14677 • Detect 5 seconds of silence with -50dB noise tolerance:
14679 silencedetect=n=-50dB:d=5
14681 • Complete example with ffmpeg to detect silence with 0.0001 noise
14683 tolerance in silence.mp3:
14684 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
14686 silenceremove
14688 Remove silence from the beginning, middle or end of the audio.
14689 The filter accepts the following options:
14691 start_periods
14693 This value is used to indicate if audio should be trimmed at
14694 beginning of the audio. A value of zero indicates no silence should
14695 be trimmed from the beginning. When specifying a non-zero value, it
14696 trims audio up until it finds non-silence. Normally, when trimming
14697 silence from beginning of audio the start_periods will be 1 but it
14698 can be increased to higher values to trim all audio up to specific
14699 count of non-silence periods. Default value is 0.
14700 start_duration
14702 Specify the amount of time that non-silence must be detected before
14703 it stops trimming audio. By increasing the duration, bursts of
14704 noises can be treated as silence and trimmed off. Default value is
14705 0.
14706 start_threshold
14708 This indicates what sample value should be treated as silence. For
14709 digital audio, a value of 0 may be fine but for audio recorded from
14710 analog, you may wish to increase the value to account for
14711 background noise. Can be specified in dB (in case "dB" is appended
14712 to the specified value) or amplitude ratio. Default value is 0.
14713 start_silence
14715 Specify max duration of silence at beginning that will be kept
14716 after trimming. Default is 0, which is equal to trimming all
14717 samples detected as silence.
14718 start_mode
14720 Specify mode of detection of silence end in start of multi-channel
14721 audio. Can be any or all. Default is any. With any, any sample
14722 that is detected as non-silence will cause stopped trimming of
14723 silence. With all, only if all channels are detected as non-
14724 silence will cause stopped trimming of silence.
14725 stop_periods
14727 Set the count for trimming silence from the end of audio. To
14728 remove silence from the middle of a file, specify a stop_periods
14729 that is negative. This value is then treated as a positive value
14730 and is used to indicate the effect should restart processing as
14731 specified by start_periods, making it suitable for removing periods
14732 of silence in the middle of the audio. Default value is 0.
14733 stop_duration
14735 Specify a duration of silence that must exist before audio is not
14736 copied any more. By specifying a higher duration, silence that is
14737 wanted can be left in the audio. Default value is 0.
14738 stop_threshold
14740 This is the same as start_threshold but for trimming silence from
14741 the end of audio. Can be specified in dB (in case "dB" is appended
14742 to the specified value) or amplitude ratio. Default value is 0.
14743 stop_silence
14745 Specify max duration of silence at end that will be kept after
14746 trimming. Default is 0, which is equal to trimming all samples
14747 detected as silence.
14748 stop_mode
14750 Specify mode of detection of silence start in end of multi-channel
14751 audio. Can be any or all. Default is any. With any, any sample
14752 that is detected as non-silence will cause stopped trimming of
14753 silence. With all, only if all channels are detected as non-
14754 silence will cause stopped trimming of silence.
14755 detection
14757 Set how is silence detected. Can be "rms" or "peak". Second is
14758 faster and works better with digital silence which is exactly 0.
14759 Default value is "rms".
14760 window
14762 Set duration in number of seconds used to calculate size of window
14763 in number of samples for detecting silence. Default value is 0.02.
14764 Allowed range is from 0 to 10.
14765 Examples
14767 • The following example shows how this filter can be used to start a
14769 recording that does not contain the delay at the start which
14770 usually occurs between pressing the record button and the start of
14771 the performance:
14772 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
14774 • Trim all silence encountered from beginning to end where there is
14776 more than 1 second of silence in audio:
14777 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
14779 • Trim all digital silence samples, using peak detection, from
14781 beginning to end where there is more than 0 samples of digital
14782 silence in audio and digital silence is detected in all channels at
14783 same positions in stream:
14784 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
14786 sofalizer
14788 SOFAlizer uses head-related transfer functions (HRTFs) to create
14789 virtual loudspeakers around the user for binaural listening via
14790 headphones (audio formats up to 9 channels supported). The HRTFs are
14791 stored in SOFA files (see <http://www.sofacoustics.org/> for a
14792 database). SOFAlizer is developed at the Acoustics Research Institute
14793 (ARI) of the Austrian Academy of Sciences.
14794 To enable compilation of this filter you need to configure FFmpeg with
14796 "--enable-libmysofa".
14797 The filter accepts the following options:
14799 sofa
14801 Set the SOFA file used for rendering.
14802 gain
14804 Set gain applied to audio. Value is in dB. Default is 0.
14805 rotation
14807 Set rotation of virtual loudspeakers in deg. Default is 0.
14808 elevation
14810 Set elevation of virtual speakers in deg. Default is 0.
14811 radius
14813 Set distance in meters between loudspeakers and the listener with
14814 near-field HRTFs. Default is 1.
14815 type
14817 Set processing type. Can be time or freq. time is processing audio
14818 in time domain which is slow. freq is processing audio in
14819 frequency domain which is fast. Default is freq.
14820 speakers
14822 Set custom positions of virtual loudspeakers. Syntax for this
14823 option is: <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...]. Each
14824 virtual loudspeaker is described with short channel name following
14825 with azimuth and elevation in degrees. Each virtual loudspeaker
14826 description is separated by '|'. For example to override front
14827 left and front right channel positions use: 'speakers=FL 45 15|FR
14828 345 15'. Descriptions with unrecognised channel names are ignored.
14829 lfegain
14831 Set custom gain for LFE channels. Value is in dB. Default is 0.
14832 framesize
14834 Set custom frame size in number of samples. Default is 1024.
14835 Allowed range is from 1024 to 96000. Only used if option type is
14836 set to freq.
14837 normalize
14839 Should all IRs be normalized upon importing SOFA file. By default
14840 is enabled.
14841 interpolate
14843 Should nearest IRs be interpolated with neighbor IRs if exact
14844 position does not match. By default is disabled.
14845 minphase
14847 Minphase all IRs upon loading of SOFA file. By default is disabled.
14848 anglestep
14850 Set neighbor search angle step. Only used if option interpolate is
14851 enabled.
14852 radstep
14854 Set neighbor search radius step. Only used if option interpolate is
14855 enabled.
14856 Examples
14858 • Using ClubFritz6 sofa file:
14860 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
14862 • Using ClubFritz12 sofa file and bigger radius with small rotation:
14864 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
14866 • Similar as above but with custom speaker positions for front left,
14868 front right, back left and back right and also with custom gain:
14869 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
14871 speechnorm
14873 Speech Normalizer.
14874 This filter expands or compresses each half-cycle of audio samples
14876 (local set of samples all above or all below zero and between two
14877 nearest zero crossings) depending on threshold value, so audio reaches
14878 target peak value under conditions controlled by below options.
14879 The filter accepts the following options:
14881 peak, p
14883 Set the expansion target peak value. This specifies the highest
14884 allowed absolute amplitude level for the normalized audio input.
14885 Default value is 0.95. Allowed range is from 0.0 to 1.0.
14886 expansion, e
14888 Set the maximum expansion factor. Allowed range is from 1.0 to
14889 50.0. Default value is 2.0. This option controls maximum local
14890 half-cycle of samples expansion. The maximum expansion would be
14891 such that local peak value reaches target peak value but never to
14892 surpass it and that ratio between new and previous peak value does
14893 not surpass this option value.
14894 compression, c
14896 Set the maximum compression factor. Allowed range is from 1.0 to
14897 50.0. Default value is 2.0. This option controls maximum local
14898 half-cycle of samples compression. This option is used only if
14899 threshold option is set to value greater than 0.0, then in such
14900 cases when local peak is lower or same as value set by threshold
14901 all samples belonging to that peak's half-cycle will be compressed
14902 by current compression factor.
14903 threshold, t
14905 Set the threshold value. Default value is 0.0. Allowed range is
14906 from 0.0 to 1.0. This option specifies which half-cycles of
14907 samples will be compressed and which will be expanded. Any half-
14908 cycle samples with their local peak value below or same as this
14909 option value will be compressed by current compression factor,
14910 otherwise, if greater than threshold value they will be expanded
14911 with expansion factor so that it could reach peak target value but
14912 never surpass it.
14913 raise, r
14915 Set the expansion raising amount per each half-cycle of samples.
14916 Default value is 0.001. Allowed range is from 0.0 to 1.0. This
14917 controls how fast expansion factor is raised per each new half-
14918 cycle until it reaches expansion value. Setting this options too
14919 high may lead to distortions.
14920 fall, f
14922 Set the compression raising amount per each half-cycle of samples.
14923 Default value is 0.001. Allowed range is from 0.0 to 1.0. This
14924 controls how fast compression factor is raised per each new half-
14925 cycle until it reaches compression value.
14926 channels, h
14928 Specify which channels to filter, by default all available channels
14929 are filtered.
14930 invert, i
14932 Enable inverted filtering, by default is disabled. This inverts
14933 interpretation of threshold option. When enabled any half-cycle of
14934 samples with their local peak value below or same as threshold
14935 option will be expanded otherwise it will be compressed.
14936 link, l
14938 Link channels when calculating gain applied to each filtered
14939 channel sample, by default is disabled. When disabled each
14940 filtered channel gain calculation is independent, otherwise when
14941 this option is enabled the minimum of all possible gains for each
14942 filtered channel is used.
14943 Commands
14945 This filter supports the all above options as commands.
14947 stereotools
14949 This filter has some handy utilities to manage stereo signals, for
14950 converting M/S stereo recordings to L/R signal while having control
14951 over the parameters or spreading the stereo image of master track.
14952 The filter accepts the following options:
14954 level_in
14956 Set input level before filtering for both channels. Defaults is 1.
14957 Allowed range is from 0.015625 to 64.
14958 level_out
14960 Set output level after filtering for both channels. Defaults is 1.
14961 Allowed range is from 0.015625 to 64.
14962 balance_in
14964 Set input balance between both channels. Default is 0. Allowed
14965 range is from -1 to 1.
14966 balance_out
14968 Set output balance between both channels. Default is 0. Allowed
14969 range is from -1 to 1.
14970 softclip
14972 Enable softclipping. Results in analog distortion instead of harsh
14973 digital 0dB clipping. Disabled by default.
14974 mutel
14976 Mute the left channel. Disabled by default.
14977 muter
14979 Mute the right channel. Disabled by default.
14980 phasel
14982 Change the phase of the left channel. Disabled by default.
14983 phaser
14985 Change the phase of the right channel. Disabled by default.
14986 mode
14988 Set stereo mode. Available values are:
14989 lr>lr
14991 Left/Right to Left/Right, this is default.
14992 lr>ms
14994 Left/Right to Mid/Side.
14995 ms>lr
14997 Mid/Side to Left/Right.
14998 lr>ll
15000 Left/Right to Left/Left.
15001 lr>rr
15003 Left/Right to Right/Right.
15004 lr>l+r
15006 Left/Right to Left + Right.
15007 lr>rl
15009 Left/Right to Right/Left.
15010 ms>ll
15012 Mid/Side to Left/Left.
15013 ms>rr
15015 Mid/Side to Right/Right.
15016 ms>rl
15018 Mid/Side to Right/Left.
15019 lr>l-r
15021 Left/Right to Left - Right.
15022 slev
15024 Set level of side signal. Default is 1. Allowed range is from
15025 0.015625 to 64.
15026 sbal
15028 Set balance of side signal. Default is 0. Allowed range is from -1
15029 to 1.
15030 mlev
15032 Set level of the middle signal. Default is 1. Allowed range is
15033 from 0.015625 to 64.
15034 mpan
15036 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
15037 base
15039 Set stereo base between mono and inversed channels. Default is 0.
15040 Allowed range is from -1 to 1.
15041 delay
15043 Set delay in milliseconds how much to delay left from right channel
15044 and vice versa. Default is 0. Allowed range is from -20 to 20.
15045 sclevel
15047 Set S/C level. Default is 1. Allowed range is from 1 to 100.
15048 phase
15050 Set the stereo phase in degrees. Default is 0. Allowed range is
15051 from 0 to 360.
15052 bmode_in, bmode_out
15054 Set balance mode for balance_in/balance_out option.
15055 Can be one of the following:
15057 balance
15059 Classic balance mode. Attenuate one channel at time. Gain is
15060 raised up to 1.
15061 amplitude
15063 Similar as classic mode above but gain is raised up to 2.
15064 power
15066 Equal power distribution, from -6dB to +6dB range.
15067 Commands
15069 This filter supports the all above options as commands.
15071 Examples
15073 • Apply karaoke like effect:
15075 stereotools=mlev=0.015625
15077 • Convert M/S signal to L/R:
15079 "stereotools=mode=ms>lr"
15081 stereowiden
15083 This filter enhance the stereo effect by suppressing signal common to
15084 both channels and by delaying the signal of left into right and vice
15085 versa, thereby widening the stereo effect.
15086 The filter accepts the following options:
15088 delay
15090 Time in milliseconds of the delay of left signal into right and
15091 vice versa. Default is 20 milliseconds.
15092 feedback
15094 Amount of gain in delayed signal into right and vice versa. Gives a
15095 delay effect of left signal in right output and vice versa which
15096 gives widening effect. Default is 0.3.
15097 crossfeed
15099 Cross feed of left into right with inverted phase. This helps in
15100 suppressing the mono. If the value is 1 it will cancel all the
15101 signal common to both channels. Default is 0.3.
15102 drymix
15104 Set level of input signal of original channel. Default is 0.8.
15105 Commands
15107 This filter supports the all above options except "delay" as commands.
15109 superequalizer
15111 Apply 18 band equalizer.
15112 The filter accepts the following options:
15114 1b Set 65Hz band gain.
15116 2b Set 92Hz band gain.
15118 3b Set 131Hz band gain.
15120 4b Set 185Hz band gain.
15122 5b Set 262Hz band gain.
15124 6b Set 370Hz band gain.
15126 7b Set 523Hz band gain.
15128 8b Set 740Hz band gain.
15130 9b Set 1047Hz band gain.
15132 10b Set 1480Hz band gain.
15134 11b Set 2093Hz band gain.
15136 12b Set 2960Hz band gain.
15138 13b Set 4186Hz band gain.
15140 14b Set 5920Hz band gain.
15142 15b Set 8372Hz band gain.
15144 16b Set 11840Hz band gain.
15146 17b Set 16744Hz band gain.
15148 18b Set 20000Hz band gain.
15150 surround
15152 Apply audio surround upmix filter.
15153 This filter allows to produce multichannel output from audio stream.
15155 The filter accepts the following options:
15157 chl_out
15159 Set output channel layout. By default, this is 5.1.
15160 See the Channel Layout section in the ffmpeg-utils(1) manual for
15162 the required syntax.
15163 chl_in
15165 Set input channel layout. By default, this is stereo.
15166 See the Channel Layout section in the ffmpeg-utils(1) manual for
15168 the required syntax.
15169 level_in
15171 Set input volume level. By default, this is 1.
15172 level_out
15174 Set output volume level. By default, this is 1.
15175 lfe Enable LFE channel output if output channel layout has it. By
15177 default, this is enabled.
15178 lfe_low
15180 Set LFE low cut off frequency. By default, this is 128 Hz.
15181 lfe_high
15183 Set LFE high cut off frequency. By default, this is 256 Hz.
15184 lfe_mode
15186 Set LFE mode, can be add or sub. Default is add. In add mode, LFE
15187 channel is created from input audio and added to output. In sub
15188 mode, LFE channel is created from input audio and added to output
15189 but also all non-LFE output channels are subtracted with output LFE
15190 channel.
15191 angle
15193 Set angle of stereo surround transform, Allowed range is from 0 to
15194 360. Default is 90.
15195 fc_in
15197 Set front center input volume. By default, this is 1.
15198 fc_out
15200 Set front center output volume. By default, this is 1.
15201 fl_in
15203 Set front left input volume. By default, this is 1.
15204 fl_out
15206 Set front left output volume. By default, this is 1.
15207 fr_in
15209 Set front right input volume. By default, this is 1.
15210 fr_out
15212 Set front right output volume. By default, this is 1.
15213 sl_in
15215 Set side left input volume. By default, this is 1.
15216 sl_out
15218 Set side left output volume. By default, this is 1.
15219 sr_in
15221 Set side right input volume. By default, this is 1.
15222 sr_out
15224 Set side right output volume. By default, this is 1.
15225 bl_in
15227 Set back left input volume. By default, this is 1.
15228 bl_out
15230 Set back left output volume. By default, this is 1.
15231 br_in
15233 Set back right input volume. By default, this is 1.
15234 br_out
15236 Set back right output volume. By default, this is 1.
15237 bc_in
15239 Set back center input volume. By default, this is 1.
15240 bc_out
15242 Set back center output volume. By default, this is 1.
15243 lfe_in
15245 Set LFE input volume. By default, this is 1.
15246 lfe_out
15248 Set LFE output volume. By default, this is 1.
15249 allx
15251 Set spread usage of stereo image across X axis for all channels.
15252 Allowed range is from -1 to 15. By default this value is negative
15253 -1, and thus unused.
15254 ally
15256 Set spread usage of stereo image across Y axis for all channels.
15257 Allowed range is from -1 to 15. By default this value is negative
15258 -1, and thus unused.
15259 fcx, flx, frx, blx, brx, slx, srx, bcx
15261 Set spread usage of stereo image across X axis for each channel.
15262 Allowed range is from 0.06 to 15. By default this value is 0.5.
15263 fcy, fly, fry, bly, bry, sly, sry, bcy
15265 Set spread usage of stereo image across Y axis for each channel.
15266 Allowed range is from 0.06 to 15. By default this value is 0.5.
15267 win_size
15269 Set window size. Allowed range is from 1024 to 65536. Default size
15270 is 4096.
15271 win_func
15273 Set window function.
15274 It accepts the following values:
15276 rect
15278 bartlett
15279 hann, hanning
15280 hamming
15281 blackman
15282 welch
15283 flattop
15284 bharris
15285 bnuttall
15286 bhann
15287 sine
15288 nuttall
15289 lanczos
15290 gauss
15291 tukey
15292 dolph
15293 cauchy
15294 parzen
15295 poisson
15296 bohman
15297 Default is "hann".
15299 overlap
15301 Set window overlap. If set to 1, the recommended overlap for
15302 selected window function will be picked. Default is 0.5.
15303 tiltshelf
15305 Boost or cut the lower frequencies and cut or boost higher frequencies
15306 of the audio using a two-pole shelving filter with a response similar
15307 to that of a standard hi-fi's tone-controls. This is also known as
15308 shelving equalisation (EQ).
15309 The filter accepts the following options:
15311 gain, g
15313 Give the gain at 0 Hz. Its useful range is about -20 (for a large
15314 cut) to +20 (for a large boost). Beware of clipping when using a
15315 positive gain.
15316 frequency, f
15318 Set the filter's central frequency and so can be used to extend or
15319 reduce the frequency range to be boosted or cut. The default value
15320 is 3000 Hz.
15321 width_type, t
15323 Set method to specify band-width of filter.
15324 h Hz
15326 q Q-Factor
15328 o octave
15330 s slope
15332 k kHz
15334 width, w
15336 Determine how steep is the filter's shelf transition.
15337 poles, p
15339 Set number of poles. Default is 2.
15340 mix, m
15342 How much to use filtered signal in output. Default is 1. Range is
15343 between 0 and 1.
15344 channels, c
15346 Specify which channels to filter, by default all available are
15347 filtered.
15348 normalize, n
15350 Normalize biquad coefficients, by default is disabled. Enabling it
15351 will normalize magnitude response at DC to 0dB.
15352 transform, a
15354 Set transform type of IIR filter.
15355 di
15357 dii
15358 tdi
15359 tdii
15360 latt
15361 svf
15362 zdf
15363 precision, r
15364 Set precison of filtering.
15365 auto
15367 Pick automatic sample format depending on surround filters.
15368 s16 Always use signed 16-bit.
15370 s32 Always use signed 32-bit.
15372 f32 Always use float 32-bit.
15374 f64 Always use float 64-bit.
15376 block_size, b
15378 Set block size used for reverse IIR processing. If this value is
15379 set to high enough value (higher than impulse response length
15380 truncated when reaches near zero values) filtering will become
15381 linear phase otherwise if not big enough it will just produce nasty
15382 artifacts.
15383 Note that filter delay will be exactly this many samples when set
15385 to non-zero value.
15386 Commands
15388 This filter supports some options as commands.
15390 treble, highshelf
15392 Boost or cut treble (upper) frequencies of the audio using a two-pole
15393 shelving filter with a response similar to that of a standard hi-fi's
15394 tone-controls. This is also known as shelving equalisation (EQ).
15395 The filter accepts the following options:
15397 gain, g
15399 Give the gain at whichever is the lower of ~22 kHz and the Nyquist
15400 frequency. Its useful range is about -20 (for a large cut) to +20
15401 (for a large boost). Beware of clipping when using a positive gain.
15402 frequency, f
15404 Set the filter's central frequency and so can be used to extend or
15405 reduce the frequency range to be boosted or cut. The default value
15406 is 3000 Hz.
15407 width_type, t
15409 Set method to specify band-width of filter.
15410 h Hz
15412 q Q-Factor
15414 o octave
15416 s slope
15418 k kHz
15420 width, w
15422 Determine how steep is the filter's shelf transition.
15423 poles, p
15425 Set number of poles. Default is 2.
15426 mix, m
15428 How much to use filtered signal in output. Default is 1. Range is
15429 between 0 and 1.
15430 channels, c
15432 Specify which channels to filter, by default all available are
15433 filtered.
15434 normalize, n
15436 Normalize biquad coefficients, by default is disabled. Enabling it
15437 will normalize magnitude response at DC to 0dB.
15438 transform, a
15440 Set transform type of IIR filter.
15441 di
15443 dii
15444 tdi
15445 tdii
15446 latt
15447 svf
15448 zdf
15449 precision, r
15450 Set precison of filtering.
15451 auto
15453 Pick automatic sample format depending on surround filters.
15454 s16 Always use signed 16-bit.
15456 s32 Always use signed 32-bit.
15458 f32 Always use float 32-bit.
15460 f64 Always use float 64-bit.
15462 block_size, b
15464 Set block size used for reverse IIR processing. If this value is
15465 set to high enough value (higher than impulse response length
15466 truncated when reaches near zero values) filtering will become
15467 linear phase otherwise if not big enough it will just produce nasty
15468 artifacts.
15469 Note that filter delay will be exactly this many samples when set
15471 to non-zero value.
15472 Commands
15474 This filter supports the following commands:
15476 frequency, f
15478 Change treble frequency. Syntax for the command is : "frequency"
15479 width_type, t
15481 Change treble width_type. Syntax for the command is : "width_type"
15482 width, w
15484 Change treble width. Syntax for the command is : "width"
15485 gain, g
15487 Change treble gain. Syntax for the command is : "gain"
15488 mix, m
15490 Change treble mix. Syntax for the command is : "mix"
15491 tremolo
15493 Sinusoidal amplitude modulation.
15494 The filter accepts the following options:
15496 f Modulation frequency in Hertz. Modulation frequencies in the
15498 subharmonic range (20 Hz or lower) will result in a tremolo effect.
15499 This filter may also be used as a ring modulator by specifying a
15500 modulation frequency higher than 20 Hz. Range is 0.1 - 20000.0.
15501 Default value is 5.0 Hz.
15502 d Depth of modulation as a percentage. Range is 0.0 - 1.0. Default
15504 value is 0.5.
15505 vibrato
15507 Sinusoidal phase modulation.
15508 The filter accepts the following options:
15510 f Modulation frequency in Hertz. Range is 0.1 - 20000.0. Default
15512 value is 5.0 Hz.
15513 d Depth of modulation as a percentage. Range is 0.0 - 1.0. Default
15515 value is 0.5.
15516 virtualbass
15518 Apply audio Virtual Bass filter.
15519 This filter accepts stereo input and produce stereo with LFE (2.1)
15521 channels output. The newly produced LFE channel have enhanced virtual
15522 bass originally obtained from both stereo channels. This filter
15523 outputs front left and front right channels unchanged as available in
15524 stereo input.
15525 The filter accepts the following options:
15527 cutoff
15529 Set the virtual bass cutoff frequency. Default value is 250 Hz.
15530 Allowed range is from 100 to 500 Hz.
15531 strength
15533 Set the virtual bass strength. Allowed range is from 0.5 to 3.
15534 Default value is 3.
15535 volume
15537 Adjust the input audio volume.
15538 It accepts the following parameters:
15540 volume
15542 Set audio volume expression.
15543 Output values are clipped to the maximum value.
15545 The output audio volume is given by the relation:
15547 <output_volume> = <volume> * <input_volume>
15549 The default value for volume is "1.0".
15551 precision
15553 This parameter represents the mathematical precision.
15554 It determines which input sample formats will be allowed, which
15556 affects the precision of the volume scaling.
15557 fixed
15559 8-bit fixed-point; this limits input sample format to U8, S16,
15560 and S32.
15561 float
15563 32-bit floating-point; this limits input sample format to FLT.
15564 (default)
15565 double
15567 64-bit floating-point; this limits input sample format to DBL.
15568 replaygain
15570 Choose the behaviour on encountering ReplayGain side data in input
15571 frames.
15572 drop
15574 Remove ReplayGain side data, ignoring its contents (the
15575 default).
15576 ignore
15578 Ignore ReplayGain side data, but leave it in the frame.
15579 track
15581 Prefer the track gain, if present.
15582 album
15584 Prefer the album gain, if present.
15585 replaygain_preamp
15587 Pre-amplification gain in dB to apply to the selected replaygain
15588 gain.
15589 Default value for replaygain_preamp is 0.0.
15591 replaygain_noclip
15593 Prevent clipping by limiting the gain applied.
15594 Default value for replaygain_noclip is 1.
15596 eval
15598 Set when the volume expression is evaluated.
15599 It accepts the following values:
15601 once
15603 only evaluate expression once during the filter initialization,
15604 or when the volume command is sent
15605 frame
15607 evaluate expression for each incoming frame
15608 Default value is once.
15610 The volume expression can contain the following parameters.
15612 n frame number (starting at zero)
15614 nb_channels
15616 number of channels
15617 nb_consumed_samples
15619 number of samples consumed by the filter
15620 nb_samples
15622 number of samples in the current frame
15623 pos original frame position in the file
15625 pts frame PTS
15627 sample_rate
15629 sample rate
15630 startpts
15632 PTS at start of stream
15633 startt
15635 time at start of stream
15636 t frame time
15638 tb timestamp timebase
15640 volume
15642 last set volume value
15643 Note that when eval is set to once only the sample_rate and tb
15645 variables are available, all other variables will evaluate to NAN.
15646 Commands
15648 This filter supports the following commands:
15650 volume
15652 Modify the volume expression. The command accepts the same syntax
15653 of the corresponding option.
15654 If the specified expression is not valid, it is kept at its current
15656 value.
15657 Examples
15659 • Halve the input audio volume:
15661 volume=volume=0.5
15663 volume=volume=1/2
15664 volume=volume=-6.0206dB
15665 In all the above example the named key for volume can be omitted,
15667 for example like in:
15668 volume=0.5
15670 • Increase input audio power by 6 decibels using fixed-point
15672 precision:
15673 volume=volume=6dB:precision=fixed
15675 • Fade volume after time 10 with an annihilation period of 5 seconds:
15677 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
15679 volumedetect
15681 Detect the volume of the input video.
15682 The filter has no parameters. It supports only 16-bit signed integer
15684 samples, so the input will be converted when needed. Statistics about
15685 the volume will be printed in the log when the input stream end is
15686 reached.
15687 In particular it will show the mean volume (root mean square), maximum
15689 volume (on a per-sample basis), and the beginning of a histogram of the
15690 registered volume values (from the maximum value to a cumulated 1/1000
15691 of the samples).
15692 All volumes are in decibels relative to the maximum PCM value.
15694 Examples
15696 Here is an excerpt of the output:
15698 [Parsed_volumedetect_0 0xa23120] mean_volume: -27 dB
15700 [Parsed_volumedetect_0 0xa23120] max_volume: -4 dB
15701 [Parsed_volumedetect_0 0xa23120] histogram_4db: 6
15702 [Parsed_volumedetect_0 0xa23120] histogram_5db: 62
15703 [Parsed_volumedetect_0 0xa23120] histogram_6db: 286
15704 [Parsed_volumedetect_0 0xa23120] histogram_7db: 1042
15705 [Parsed_volumedetect_0 0xa23120] histogram_8db: 2551
15706 [Parsed_volumedetect_0 0xa23120] histogram_9db: 4609
15707 [Parsed_volumedetect_0 0xa23120] histogram_10db: 8409
15708 It means that:
15710 • The mean square energy is approximately -27 dB, or 10^-2.7.
15712 • The largest sample is at -4 dB, or more precisely between -4 dB and
15714 -5 dB.
15715 • There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
15717 In other words, raising the volume by +4 dB does not cause any
15719 clipping, raising it by +5 dB causes clipping for 6 samples, etc.
15720
15722 Below is a description of the currently available audio sources.
15723 abuffer
15725 Buffer audio frames, and make them available to the filter chain.
15726 This source is mainly intended for a programmatic use, in particular
15728 through the interface defined in libavfilter/buffersrc.h.
15729 It accepts the following parameters:
15731 time_base
15733 The timebase which will be used for timestamps of submitted frames.
15734 It must be either a floating-point number or in
15735 numerator/denominator form.
15736 sample_rate
15738 The sample rate of the incoming audio buffers.
15739 sample_fmt
15741 The sample format of the incoming audio buffers. Either a sample
15742 format name or its corresponding integer representation from the
15743 enum AVSampleFormat in libavutil/samplefmt.h
15744 channel_layout
15746 The channel layout of the incoming audio buffers. Either a channel
15747 layout name from channel_layout_map in libavutil/channel_layout.c
15748 or its corresponding integer representation from the AV_CH_LAYOUT_*
15749 macros in libavutil/channel_layout.h
15750 channels
15752 The number of channels of the incoming audio buffers. If both
15753 channels and channel_layout are specified, then they must be
15754 consistent.
15755 Examples
15757 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
15759 will instruct the source to accept planar 16bit signed stereo at
15761 44100Hz. Since the sample format with name "s16p" corresponds to the
15762 number 6 and the "stereo" channel layout corresponds to the value 0x3,
15763 this is equivalent to:
15764 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
15766 aevalsrc
15768 Generate an audio signal specified by an expression.
15769 This source accepts in input one or more expressions (one for each
15771 channel), which are evaluated and used to generate a corresponding
15772 audio signal.
15773 This source accepts the following options:
15775 exprs
15777 Set the '|'-separated expressions list for each separate channel.
15778 In case the channel_layout option is not specified, the selected
15779 channel layout depends on the number of provided expressions.
15780 Otherwise the last specified expression is applied to the remaining
15781 output channels.
15782 channel_layout, c
15784 Set the channel layout. The number of channels in the specified
15785 layout must be equal to the number of specified expressions.
15786 duration, d
15788 Set the minimum duration of the sourced audio. See the Time
15789 duration section in the ffmpeg-utils(1) manual for the accepted
15790 syntax. Note that the resulting duration may be greater than the
15791 specified duration, as the generated audio is always cut at the end
15792 of a complete frame.
15793 If not specified, or the expressed duration is negative, the audio
15795 is supposed to be generated forever.
15796 nb_samples, n
15798 Set the number of samples per channel per each output frame,
15799 default to 1024.
15800 sample_rate, s
15802 Specify the sample rate, default to 44100.
15803 Each expression in exprs can contain the following constants:
15805 n number of the evaluated sample, starting from 0
15807 t time of the evaluated sample expressed in seconds, starting from 0
15809 s sample rate
15811 Examples
15813 • Generate silence:
15815 aevalsrc=0
15817 • Generate a sin signal with frequency of 440 Hz, set sample rate to
15819 8000 Hz:
15820 aevalsrc="sin(440*2*PI*t):s=8000"
15822 • Generate a two channels signal, specify the channel layout (Front
15824 Center + Back Center) explicitly:
15825 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
15827 • Generate white noise:
15829 aevalsrc="-2+random(0)"
15831 • Generate an amplitude modulated signal:
15833 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
15835 • Generate 2.5 Hz binaural beats on a 360 Hz carrier:
15837 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
15839 afirsrc
15841 Generate a FIR coefficients using frequency sampling method.
15842 The resulting stream can be used with afir filter for filtering the
15844 audio signal.
15845 The filter accepts the following options:
15847 taps, t
15849 Set number of filter coefficents in output audio stream. Default
15850 value is 1025.
15851 frequency, f
15853 Set frequency points from where magnitude and phase are set. This
15854 must be in non decreasing order, and first element must be 0, while
15855 last element must be 1. Elements are separated by white spaces.
15856 magnitude, m
15858 Set magnitude value for every frequency point set by frequency.
15859 Number of values must be same as number of frequency points.
15860 Values are separated by white spaces.
15861 phase, p
15863 Set phase value for every frequency point set by frequency. Number
15864 of values must be same as number of frequency points. Values are
15865 separated by white spaces.
15866 sample_rate, r
15868 Set sample rate, default is 44100.
15869 nb_samples, n
15871 Set number of samples per each frame. Default is 1024.
15872 win_func, w
15874 Set window function. Default is blackman.
15875 anullsrc
15877 The null audio source, return unprocessed audio frames. It is mainly
15878 useful as a template and to be employed in analysis / debugging tools,
15879 or as the source for filters which ignore the input data (for example
15880 the sox synth filter).
15881 This source accepts the following options:
15883 channel_layout, cl
15885 Specifies the channel layout, and can be either an integer or a
15886 string representing a channel layout. The default value of
15887 channel_layout is "stereo".
15888 Check the channel_layout_map definition in
15890 libavutil/channel_layout.c for the mapping between strings and
15891 channel layout values.
15892 sample_rate, r
15894 Specifies the sample rate, and defaults to 44100.
15895 nb_samples, n
15897 Set the number of samples per requested frames.
15898 duration, d
15900 Set the duration of the sourced audio. See the Time duration
15901 section in the ffmpeg-utils(1) manual for the accepted syntax.
15902 If not specified, or the expressed duration is negative, the audio
15904 is supposed to be generated forever.
15905 Examples
15907 • Set the sample rate to 48000 Hz and the channel layout to
15909 AV_CH_LAYOUT_MONO.
15910 anullsrc=r=48000:cl=4
15912 • Do the same operation with a more obvious syntax:
15914 anullsrc=r=48000:cl=mono
15916 All the parameters need to be explicitly defined.
15918 flite
15920 Synthesize a voice utterance using the libflite library.
15921 To enable compilation of this filter you need to configure FFmpeg with
15923 "--enable-libflite".
15924 Note that versions of the flite library prior to 2.0 are not thread-
15926 safe.
15927 The filter accepts the following options:
15929 list_voices
15931 If set to 1, list the names of the available voices and exit
15932 immediately. Default value is 0.
15933 nb_samples, n
15935 Set the maximum number of samples per frame. Default value is 512.
15936 textfile
15938 Set the filename containing the text to speak.
15939 text
15941 Set the text to speak.
15942 voice, v
15944 Set the voice to use for the speech synthesis. Default value is
15945 "kal". See also the list_voices option.
15946 Examples
15948 • Read from file speech.txt, and synthesize the text using the
15950 standard flite voice:
15951 flite=textfile=speech.txt
15953 • Read the specified text selecting the "slt" voice:
15955 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
15957 • Input text to ffmpeg:
15959 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
15961 • Make ffplay speak the specified text, using "flite" and the "lavfi"
15963 device:
15964 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
15966 For more information about libflite, check:
15968 <http://www.festvox.org/flite/>
15969 anoisesrc
15971 Generate a noise audio signal.
15972 The filter accepts the following options:
15974 sample_rate, r
15976 Specify the sample rate. Default value is 48000 Hz.
15977 amplitude, a
15979 Specify the amplitude (0.0 - 1.0) of the generated audio stream.
15980 Default value is 1.0.
15981 duration, d
15983 Specify the duration of the generated audio stream. Not specifying
15984 this option results in noise with an infinite length.
15985 color, colour, c
15987 Specify the color of noise. Available noise colors are white, pink,
15988 brown, blue, violet and velvet. Default color is white.
15989 seed, s
15991 Specify a value used to seed the PRNG.
15992 nb_samples, n
15994 Set the number of samples per each output frame, default is 1024.
15995 Examples
15997 • Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate
15999 and an amplitude of 0.5:
16000 anoisesrc=d=60:c=pink:r=44100:a=0.5
16002 hilbert
16004 Generate odd-tap Hilbert transform FIR coefficients.
16005 The resulting stream can be used with afir filter for phase-shifting
16007 the signal by 90 degrees.
16008 This is used in many matrix coding schemes and for analytic signal
16010 generation. The process is often written as a multiplication by i (or
16011 j), the imaginary unit.
16012 The filter accepts the following options:
16014 sample_rate, s
16016 Set sample rate, default is 44100.
16017 taps, t
16019 Set length of FIR filter, default is 22051.
16020 nb_samples, n
16022 Set number of samples per each frame.
16023 win_func, w
16025 Set window function to be used when generating FIR coefficients.
16026 sinc
16028 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or
16029 band-reject FIR coefficients.
16030 The resulting stream can be used with afir filter for filtering the
16032 audio signal.
16033 The filter accepts the following options:
16035 sample_rate, r
16037 Set sample rate, default is 44100.
16038 nb_samples, n
16040 Set number of samples per each frame. Default is 1024.
16041 hp Set high-pass frequency. Default is 0.
16043 lp Set low-pass frequency. Default is 0. If high-pass frequency is
16045 lower than low-pass frequency and low-pass frequency is higher than
16046 0 then filter will create band-pass filter coefficients, otherwise
16047 band-reject filter coefficients.
16048 phase
16050 Set filter phase response. Default is 50. Allowed range is from 0
16051 to 100.
16052 beta
16054 Set Kaiser window beta.
16055 att Set stop-band attenuation. Default is 120dB, allowed range is from
16057 40 to 180 dB.
16058 round
16060 Enable rounding, by default is disabled.
16061 hptaps
16063 Set number of taps for high-pass filter.
16064 lptaps
16066 Set number of taps for low-pass filter.
16067 sine
16069 Generate an audio signal made of a sine wave with amplitude 1/8.
16070 The audio signal is bit-exact.
16072 The filter accepts the following options:
16074 frequency, f
16076 Set the carrier frequency. Default is 440 Hz.
16077 beep_factor, b
16079 Enable a periodic beep every second with frequency beep_factor
16080 times the carrier frequency. Default is 0, meaning the beep is
16081 disabled.
16082 sample_rate, r
16084 Specify the sample rate, default is 44100.
16085 duration, d
16087 Specify the duration of the generated audio stream.
16088 samples_per_frame
16090 Set the number of samples per output frame.
16091 The expression can contain the following constants:
16093 n The (sequential) number of the output audio frame, starting
16095 from 0.
16096 pts The PTS (Presentation TimeStamp) of the output audio frame,
16098 expressed in TB units.
16099 t The PTS of the output audio frame, expressed in seconds.
16101 TB The timebase of the output audio frames.
16103 Default is 1024.
16105 Examples
16107 • Generate a simple 440 Hz sine wave:
16109 sine
16111 • Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5
16113 seconds:
16114 sine=220:4:d=5
16116 sine=f=220:b=4:d=5
16117 sine=frequency=220:beep_factor=4:duration=5
16118 • Generate a 1 kHz sine wave following "1602,1601,1602,1601,1602"
16120 NTSC pattern:
16121 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
16123
16125 Below is a description of the currently available audio sinks.
16126 abuffersink
16128 Buffer audio frames, and make them available to the end of filter
16129 chain.
16130 This sink is mainly intended for programmatic use, in particular
16132 through the interface defined in libavfilter/buffersink.h or the
16133 options system.
16134 It accepts a pointer to an AVABufferSinkContext structure, which
16136 defines the incoming buffers' formats, to be passed as the opaque
16137 parameter to "avfilter_init_filter" for initialization.
16138 anullsink
16140 Null audio sink; do absolutely nothing with the input audio. It is
16141 mainly useful as a template and for use in analysis / debugging tools.
16142
16144 When you configure your FFmpeg build, you can disable any of the
16145 existing filters using "--disable-filters". The configure output will
16146 show the video filters included in your build.
16147 Below is a description of the currently available video filters.
16149 addroi
16151 Mark a region of interest in a video frame.
16152 The frame data is passed through unchanged, but metadata is attached to
16154 the frame indicating regions of interest which can affect the behaviour
16155 of later encoding. Multiple regions can be marked by applying the
16156 filter multiple times.
16157 x Region distance in pixels from the left edge of the frame.
16159 y Region distance in pixels from the top edge of the frame.
16161 w Region width in pixels.
16163 h Region height in pixels.
16165 The parameters x, y, w and h are expressions, and may contain the
16167 following variables:
16168 iw Width of the input frame.
16170 ih Height of the input frame.
16172 qoffset
16174 Quantisation offset to apply within the region.
16175 This must be a real value in the range -1 to +1. A value of zero
16177 indicates no quality change. A negative value asks for better
16178 quality (less quantisation), while a positive value asks for worse
16179 quality (greater quantisation).
16180 The range is calibrated so that the extreme values indicate the
16182 largest possible offset - if the rest of the frame is encoded with
16183 the worst possible quality, an offset of -1 indicates that this
16184 region should be encoded with the best possible quality anyway.
16185 Intermediate values are then interpolated in some codec-dependent
16186 way.
16187 For example, in 10-bit H.264 the quantisation parameter varies
16189 between -12 and 51. A typical qoffset value of -1/10 therefore
16190 indicates that this region should be encoded with a QP around one-
16191 tenth of the full range better than the rest of the frame. So, if
16192 most of the frame were to be encoded with a QP of around 30, this
16193 region would get a QP of around 24 (an offset of approximately
16194 -1/10 * (51 - -12) = -6.3). An extreme value of -1 would indicate
16195 that this region should be encoded with the best possible quality
16196 regardless of the treatment of the rest of the frame - that is,
16197 should be encoded at a QP of -12.
16198 clear
16200 If set to true, remove any existing regions of interest marked on
16201 the frame before adding the new one.
16202 Examples
16204 • Mark the centre quarter of the frame as interesting.
16206 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
16208 • Mark the 100-pixel-wide region on the left edge of the frame as
16210 very uninteresting (to be encoded at much lower quality than the
16211 rest of the frame).
16212 addroi=0:0:100:ih:+1/5
16214 alphaextract
16216 Extract the alpha component from the input as a grayscale video. This
16217 is especially useful with the alphamerge filter.
16218 alphamerge
16220 Add or replace the alpha component of the primary input with the
16221 grayscale value of a second input. This is intended for use with
16222 alphaextract to allow the transmission or storage of frame sequences
16223 that have alpha in a format that doesn't support an alpha channel.
16224 For example, to reconstruct full frames from a normal YUV-encoded video
16226 and a separate video created with alphaextract, you might use:
16227 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
16229 amplify
16231 Amplify differences between current pixel and pixels of adjacent frames
16232 in same pixel location.
16233 This filter accepts the following options:
16235 radius
16237 Set frame radius. Default is 2. Allowed range is from 1 to 63. For
16238 example radius of 3 will instruct filter to calculate average of 7
16239 frames.
16240 factor
16242 Set factor to amplify difference. Default is 2. Allowed range is
16243 from 0 to 65535.
16244 threshold
16246 Set threshold for difference amplification. Any difference greater
16247 or equal to this value will not alter source pixel. Default is 10.
16248 Allowed range is from 0 to 65535.
16249 tolerance
16251 Set tolerance for difference amplification. Any difference lower to
16252 this value will not alter source pixel. Default is 0. Allowed
16253 range is from 0 to 65535.
16254 low Set lower limit for changing source pixel. Default is 65535.
16256 Allowed range is from 0 to 65535. This option controls maximum
16257 possible value that will decrease source pixel value.
16258 high
16260 Set high limit for changing source pixel. Default is 65535. Allowed
16261 range is from 0 to 65535. This option controls maximum possible
16262 value that will increase source pixel value.
16263 planes
16265 Set which planes to filter. Default is all. Allowed range is from 0
16266 to 15.
16267 Commands
16269 This filter supports the following commands that corresponds to option
16271 of same name:
16272 factor
16274 threshold
16275 tolerance
16276 low
16277 high
16278 planes
16279 ass
16281 Same as the subtitles filter, except that it doesn't require libavcodec
16282 and libavformat to work. On the other hand, it is limited to ASS
16283 (Advanced Substation Alpha) subtitles files.
16284 This filter accepts the following option in addition to the common
16286 options from the subtitles filter:
16287 shaping
16289 Set the shaping engine
16290 Available values are:
16292 auto
16294 The default libass shaping engine, which is the best available.
16295 simple
16297 Fast, font-agnostic shaper that can do only substitutions
16298 complex
16300 Slower shaper using OpenType for substitutions and positioning
16301 The default is "auto".
16303 atadenoise
16305 Apply an Adaptive Temporal Averaging Denoiser to the video input.
16306 The filter accepts the following options:
16308 0a Set threshold A for 1st plane. Default is 0.02. Valid range is 0
16310 to 0.3.
16311 0b Set threshold B for 1st plane. Default is 0.04. Valid range is 0
16313 to 5.
16314 1a Set threshold A for 2nd plane. Default is 0.02. Valid range is 0
16316 to 0.3.
16317 1b Set threshold B for 2nd plane. Default is 0.04. Valid range is 0
16319 to 5.
16320 2a Set threshold A for 3rd plane. Default is 0.02. Valid range is 0
16322 to 0.3.
16323 2b Set threshold B for 3rd plane. Default is 0.04. Valid range is 0
16325 to 5.
16326 Threshold A is designed to react on abrupt changes in the input
16328 signal and threshold B is designed to react on continuous changes
16329 in the input signal.
16330 s Set number of frames filter will use for averaging. Default is 9.
16332 Must be odd number in range [5, 129].
16333 p Set what planes of frame filter will use for averaging. Default is
16335 all.
16336 a Set what variant of algorithm filter will use for averaging.
16338 Default is "p" parallel. Alternatively can be set to "s" serial.
16339 Parallel can be faster then serial, while other way around is never
16341 true. Parallel will abort early on first change being greater then
16342 thresholds, while serial will continue processing other side of
16343 frames if they are equal or below thresholds.
16344 0s
16346 1s
16347 2s Set sigma for 1st plane, 2nd plane or 3rd plane. Default is 32767.
16348 Valid range is from 0 to 32767. This options controls weight for
16349 each pixel in radius defined by size. Default value means every
16350 pixel have same weight. Setting this option to 0 effectively
16351 disables filtering.
16352 Commands
16354 This filter supports same commands as options except option "s". The
16356 command accepts the same syntax of the corresponding option.
16357 avgblur
16359 Apply average blur filter.
16360 The filter accepts the following options:
16362 sizeX
16364 Set horizontal radius size.
16365 planes
16367 Set which planes to filter. By default all planes are filtered.
16368 sizeY
16370 Set vertical radius size, if zero it will be same as "sizeX".
16371 Default is 0.
16372 Commands
16374 This filter supports same commands as options. The command accepts the
16376 same syntax of the corresponding option.
16377 If the specified expression is not valid, it is kept at its current
16379 value.
16380 bbox
16382 Compute the bounding box for the non-black pixels in the input frame
16383 luminance plane.
16384 This filter computes the bounding box containing all the pixels with a
16386 luminance value greater than the minimum allowed value. The parameters
16387 describing the bounding box are printed on the filter log.
16388 The filter accepts the following option:
16390 min_val
16392 Set the minimal luminance value. Default is 16.
16393 Commands
16395 This filter supports the all above options as commands.
16397 bilateral
16399 Apply bilateral filter, spatial smoothing while preserving edges.
16400 The filter accepts the following options:
16402 sigmaS
16404 Set sigma of gaussian function to calculate spatial weight.
16405 Allowed range is 0 to 512. Default is 0.1.
16406 sigmaR
16408 Set sigma of gaussian function to calculate range weight. Allowed
16409 range is 0 to 1. Default is 0.1.
16410 planes
16412 Set planes to filter. Default is first only.
16413 Commands
16415 This filter supports the all above options as commands.
16417 bitplanenoise
16419 Show and measure bit plane noise.
16420 The filter accepts the following options:
16422 bitplane
16424 Set which plane to analyze. Default is 1.
16425 filter
16427 Filter out noisy pixels from "bitplane" set above. Default is
16428 disabled.
16429 blackdetect
16431 Detect video intervals that are (almost) completely black. Can be
16432 useful to detect chapter transitions, commercials, or invalid
16433 recordings.
16434 The filter outputs its detection analysis to both the log as well as
16436 frame metadata. If a black segment of at least the specified minimum
16437 duration is found, a line with the start and end timestamps as well as
16438 duration is printed to the log with level "info". In addition, a log
16439 line with level "debug" is printed per frame showing the black amount
16440 detected for that frame.
16441 The filter also attaches metadata to the first frame of a black segment
16443 with key "lavfi.black_start" and to the first frame after the black
16444 segment ends with key "lavfi.black_end". The value is the frame's
16445 timestamp. This metadata is added regardless of the minimum duration
16446 specified.
16447 The filter accepts the following options:
16449 black_min_duration, d
16451 Set the minimum detected black duration expressed in seconds. It
16452 must be a non-negative floating point number.
16453 Default value is 2.0.
16455 picture_black_ratio_th, pic_th
16457 Set the threshold for considering a picture "black". Express the
16458 minimum value for the ratio:
16459 <nb_black_pixels> / <nb_pixels>
16461 for which a picture is considered black. Default value is 0.98.
16463 pixel_black_th, pix_th
16465 Set the threshold for considering a pixel "black".
16466 The threshold expresses the maximum pixel luminance value for which
16468 a pixel is considered "black". The provided value is scaled
16469 according to the following equation:
16470 <absolute_threshold> = <luminance_minimum_value> + <pixel_black_th> * <luminance_range_size>
16472 luminance_range_size and luminance_minimum_value depend on the
16474 input video format, the range is [0-255] for YUV full-range formats
16475 and [16-235] for YUV non full-range formats.
16476 Default value is 0.10.
16478 The following example sets the maximum pixel threshold to the minimum
16480 value, and detects only black intervals of 2 or more seconds:
16481 blackdetect=d=2:pix_th=0.00
16483 blackframe
16485 Detect frames that are (almost) completely black. Can be useful to
16486 detect chapter transitions or commercials. Output lines consist of the
16487 frame number of the detected frame, the percentage of blackness, the
16488 position in the file if known or -1 and the timestamp in seconds.
16489 In order to display the output lines, you need to set the loglevel at
16491 least to the AV_LOG_INFO value.
16492 This filter exports frame metadata "lavfi.blackframe.pblack". The
16494 value represents the percentage of pixels in the picture that are below
16495 the threshold value.
16496 It accepts the following parameters:
16498 amount
16500 The percentage of the pixels that have to be below the threshold;
16501 it defaults to 98.
16502 threshold, thresh
16504 The threshold below which a pixel value is considered black; it
16505 defaults to 32.
16506 blend
16508 Blend two video frames into each other.
16509 The "blend" filter takes two input streams and outputs one stream, the
16511 first input is the "top" layer and second input is "bottom" layer. By
16512 default, the output terminates when the longest input terminates.
16513 The "tblend" (time blend) filter takes two consecutive frames from one
16515 single stream, and outputs the result obtained by blending the new
16516 frame on top of the old frame.
16517 A description of the accepted options follows.
16519 c0_mode
16521 c1_mode
16522 c2_mode
16523 c3_mode
16524 all_mode
16525 Set blend mode for specific pixel component or all pixel components
16526 in case of all_mode. Default value is "normal".
16527 Available values for component modes are:
16529 addition
16531 and
16532 average
16533 bleach
16534 burn
16535 darken
16536 difference
16537 divide
16538 dodge
16539 exclusion
16540 extremity
16541 freeze
16542 geometric
16543 glow
16544 grainextract
16545 grainmerge
16546 hardlight
16547 hardmix
16548 hardoverlay
16549 harmonic
16550 heat
16551 interpolate
16552 lighten
16553 linearlight
16554 multiply
16555 multiply128
16556 negation
16557 normal
16558 or
16559 overlay
16560 phoenix
16561 pinlight
16562 reflect
16563 screen
16564 softdifference
16565 softlight
16566 stain
16567 subtract
16568 vividlight
16569 xor
16570 c0_opacity
16571 c1_opacity
16572 c2_opacity
16573 c3_opacity
16574 all_opacity
16575 Set blend opacity for specific pixel component or all pixel
16576 components in case of all_opacity. Only used in combination with
16577 pixel component blend modes.
16578 c0_expr
16580 c1_expr
16581 c2_expr
16582 c3_expr
16583 all_expr
16584 Set blend expression for specific pixel component or all pixel
16585 components in case of all_expr. Note that related mode options will
16586 be ignored if those are set.
16587 The expressions can use the following variables:
16589 N The sequential number of the filtered frame, starting from 0.
16591 X
16593 Y the coordinates of the current sample
16594 W
16596 H the width and height of currently filtered plane
16597 SW
16599 SH Width and height scale for the plane being filtered. It is the
16600 ratio between the dimensions of the current plane to the luma
16601 plane, e.g. for a "yuv420p" frame, the values are "1,1" for the
16602 luma plane and "0.5,0.5" for the chroma planes.
16603 T Time of the current frame, expressed in seconds.
16605 TOP, A
16607 Value of pixel component at current location for first video
16608 frame (top layer).
16609 BOTTOM, B
16611 Value of pixel component at current location for second video
16612 frame (bottom layer).
16613 The "blend" filter also supports the framesync options.
16615 Examples
16617 • Apply transition from bottom layer to top layer in first 10
16619 seconds:
16620 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
16622 • Apply linear horizontal transition from top layer to bottom layer:
16624 blend=all_expr='A*(X/W)+B*(1-X/W)'
16626 • Apply 1x1 checkerboard effect:
16628 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
16630 • Apply uncover left effect:
16632 blend=all_expr='if(gte(N*SW+X,W),A,B)'
16634 • Apply uncover down effect:
16636 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
16638 • Apply uncover up-left effect:
16640 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
16642 • Split diagonally video and shows top and bottom layer on each side:
16644 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
16646 • Display differences between the current and the previous frame:
16648 tblend=all_mode=grainextract
16650 Commands
16652 This filter supports same commands as options.
16654 blockdetect
16656 Determines blockiness of frames without altering the input frames.
16657 Based on Remco Muijs and Ihor Kirenko: "A no-reference blocking
16659 artifact measure for adaptive video processing." 2005 13th European
16660 signal processing conference.
16661 The filter accepts the following options:
16663 period_min
16665 period_max
16666 Set minimum and maximum values for determining pixel grids
16667 (periods). Default values are [3,24].
16668 planes
16670 Set planes to filter. Default is first only.
16671 Examples
16673 • Determine blockiness for the first plane and search for periods
16675 within [8,32]:
16676 blockdetect=period_min=8:period_max=32:planes=1
16678 blurdetect
16680 Determines blurriness of frames without altering the input frames.
16681 Based on Marziliano, Pina, et al. "A no-reference perceptual blur
16683 metric." Allows for a block-based abbreviation.
16684 The filter accepts the following options:
16686 low
16688 high
16689 Set low and high threshold values used by the Canny thresholding
16690 algorithm.
16691 The high threshold selects the "strong" edge pixels, which are then
16693 connected through 8-connectivity with the "weak" edge pixels
16694 selected by the low threshold.
16695 low and high threshold values must be chosen in the range [0,1],
16697 and low should be lesser or equal to high.
16698 Default value for low is "20/255", and default value for high is
16700 "50/255".
16701 radius
16703 Define the radius to search around an edge pixel for local maxima.
16704 block_pct
16706 Determine blurriness only for the most significant blocks, given in
16707 percentage.
16708 block_width
16710 Determine blurriness for blocks of width block_width. If set to any
16711 value smaller 1, no blocks are used and the whole image is
16712 processed as one no matter of block_height.
16713 block_height
16715 Determine blurriness for blocks of height block_height. If set to
16716 any value smaller 1, no blocks are used and the whole image is
16717 processed as one no matter of block_width.
16718 planes
16720 Set planes to filter. Default is first only.
16721 Examples
16723 • Determine blur for 80% of most significant 32x32 blocks:
16725 blurdetect=block_width=32:block_height=32:block_pct=80
16727 bm3d
16729 Denoise frames using Block-Matching 3D algorithm.
16730 The filter accepts the following options.
16732 sigma
16734 Set denoising strength. Default value is 1. Allowed range is from
16735 0 to 999.9. The denoising algorithm is very sensitive to sigma, so
16736 adjust it according to the source.
16737 block
16739 Set local patch size. This sets dimensions in 2D.
16740 bstep
16742 Set sliding step for processing blocks. Default value is 4.
16743 Allowed range is from 1 to 64. Smaller values allows processing
16744 more reference blocks and is slower.
16745 group
16747 Set maximal number of similar blocks for 3rd dimension. Default
16748 value is 1. When set to 1, no block matching is done. Larger
16749 values allows more blocks in single group. Allowed range is from 1
16750 to 256.
16751 range
16753 Set radius for search block matching. Default is 9. Allowed range
16754 is from 1 to INT32_MAX.
16755 mstep
16757 Set step between two search locations for block matching. Default
16758 is 1. Allowed range is from 1 to 64. Smaller is slower.
16759 thmse
16761 Set threshold of mean square error for block matching. Valid range
16762 is 0 to INT32_MAX.
16763 hdthr
16765 Set thresholding parameter for hard thresholding in 3D transformed
16766 domain. Larger values results in stronger hard-thresholding
16767 filtering in frequency domain.
16768 estim
16770 Set filtering estimation mode. Can be "basic" or "final". Default
16771 is "basic".
16772 ref If enabled, filter will use 2nd stream for block matching. Default
16774 is disabled for "basic" value of estim option, and always enabled
16775 if value of estim is "final".
16776 planes
16778 Set planes to filter. Default is all available except alpha.
16779 Examples
16781 • Basic filtering with bm3d:
16783 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
16785 • Same as above, but filtering only luma:
16787 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
16789 • Same as above, but with both estimation modes:
16791 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
16793 • Same as above, but prefilter with nlmeans filter instead:
16795 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
16797 boxblur
16799 Apply a boxblur algorithm to the input video.
16800 It accepts the following parameters:
16802 luma_radius, lr
16804 luma_power, lp
16805 chroma_radius, cr
16806 chroma_power, cp
16807 alpha_radius, ar
16808 alpha_power, ap
16809 A description of the accepted options follows.
16811 luma_radius, lr
16813 chroma_radius, cr
16814 alpha_radius, ar
16815 Set an expression for the box radius in pixels used for blurring
16816 the corresponding input plane.
16817 The radius value must be a non-negative number, and must not be
16819 greater than the value of the expression "min(w,h)/2" for the luma
16820 and alpha planes, and of "min(cw,ch)/2" for the chroma planes.
16821 Default value for luma_radius is "2". If not specified,
16823 chroma_radius and alpha_radius default to the corresponding value
16824 set for luma_radius.
16825 The expressions can contain the following constants:
16827 w
16829 h The input width and height in pixels.
16830 cw
16832 ch The input chroma image width and height in pixels.
16833 hsub
16835 vsub
16836 The horizontal and vertical chroma subsample values. For
16837 example, for the pixel format "yuv422p", hsub is 2 and vsub is
16838 1.
16839 luma_power, lp
16841 chroma_power, cp
16842 alpha_power, ap
16843 Specify how many times the boxblur filter is applied to the
16844 corresponding plane.
16845 Default value for luma_power is 2. If not specified, chroma_power
16847 and alpha_power default to the corresponding value set for
16848 luma_power.
16849 A value of 0 will disable the effect.
16851 Examples
16853 • Apply a boxblur filter with the luma, chroma, and alpha radii set
16855 to 2:
16856 boxblur=luma_radius=2:luma_power=1
16858 boxblur=2:1
16859 • Set the luma radius to 2, and alpha and chroma radius to 0:
16861 boxblur=2:1:cr=0:ar=0
16863 • Set the luma and chroma radii to a fraction of the video dimension:
16865 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
16867 bwdif
16869 Deinterlace the input video ("bwdif" stands for "Bob Weaver
16870 Deinterlacing Filter").
16871 Motion adaptive deinterlacing based on yadif with the use of w3fdif and
16873 cubic interpolation algorithms. It accepts the following parameters:
16874 mode
16876 The interlacing mode to adopt. It accepts one of the following
16877 values:
16878 0, send_frame
16880 Output one frame for each frame.
16881 1, send_field
16883 Output one frame for each field.
16884 The default value is "send_field".
16886 parity
16888 The picture field parity assumed for the input interlaced video. It
16889 accepts one of the following values:
16890 0, tff
16892 Assume the top field is first.
16893 1, bff
16895 Assume the bottom field is first.
16896 -1, auto
16898 Enable automatic detection of field parity.
16899 The default value is "auto". If the interlacing is unknown or the
16901 decoder does not export this information, top field first will be
16902 assumed.
16903 deint
16905 Specify which frames to deinterlace. Accepts one of the following
16906 values:
16907 0, all
16909 Deinterlace all frames.
16910 1, interlaced
16912 Only deinterlace frames marked as interlaced.
16913 The default value is "all".
16915 cas
16917 Apply Contrast Adaptive Sharpen filter to video stream.
16918 The filter accepts the following options:
16920 strength
16922 Set the sharpening strength. Default value is 0.
16923 planes
16925 Set planes to filter. Default value is to filter all planes except
16926 alpha plane.
16927 Commands
16929 This filter supports same commands as options.
16931 chromahold
16933 Remove all color information for all colors except for certain one.
16934 The filter accepts the following options:
16936 color
16938 The color which will not be replaced with neutral chroma.
16939 similarity
16941 Similarity percentage with the above color. 0.01 matches only the
16942 exact key color, while 1.0 matches everything.
16943 blend
16945 Blend percentage. 0.0 makes pixels either fully gray, or not gray
16946 at all. Higher values result in more preserved color.
16947 yuv Signals that the color passed is already in YUV instead of RGB.
16949 Literal colors like "green" or "red" don't make sense with this
16951 enabled anymore. This can be used to pass exact YUV values as
16952 hexadecimal numbers.
16953 Commands
16955 This filter supports same commands as options. The command accepts the
16957 same syntax of the corresponding option.
16958 If the specified expression is not valid, it is kept at its current
16960 value.
16961 chromakey
16963 YUV colorspace color/chroma keying.
16964 The filter accepts the following options:
16966 color
16968 The color which will be replaced with transparency.
16969 similarity
16971 Similarity percentage with the key color.
16972 0.01 matches only the exact key color, while 1.0 matches
16974 everything.
16975 blend
16977 Blend percentage.
16978 0.0 makes pixels either fully transparent, or not transparent at
16980 all.
16981 Higher values result in semi-transparent pixels, with a higher
16983 transparency the more similar the pixels color is to the key color.
16984 yuv Signals that the color passed is already in YUV instead of RGB.
16986 Literal colors like "green" or "red" don't make sense with this
16988 enabled anymore. This can be used to pass exact YUV values as
16989 hexadecimal numbers.
16990 Commands
16992 This filter supports same commands as options. The command accepts the
16994 same syntax of the corresponding option.
16995 If the specified expression is not valid, it is kept at its current
16997 value.
16998 Examples
17000 • Make every green pixel in the input image transparent:
17002 ffmpeg -i input.png -vf chromakey=green out.png
17004 • Overlay a greenscreen-video on top of a static black background.
17006 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
17008 chromakey_cuda
17010 CUDA accelerated YUV colorspace color/chroma keying.
17011 This filter works like normal chromakey filter but operates on CUDA
17013 frames. for more details and parameters see chromakey.
17014 Examples
17016 • Make all the green pixels in the input video transparent and use it
17018 as an overlay for another video:
17019 ./ffmpeg \
17021 -hwaccel cuda -hwaccel_output_format cuda -i input_green.mp4 \
17022 -hwaccel cuda -hwaccel_output_format cuda -i base_video.mp4 \
17023 -init_hw_device cuda \
17024 -filter_complex \
17025 " \
17026 [0:v]chromakey_cuda=0x25302D:0.1:0.12:1[overlay_video]; \
17027 [1:v]scale_cuda=format=yuv420p[base]; \
17028 [base][overlay_video]overlay_cuda" \
17029 -an -sn -c:v h264_nvenc -cq 20 output.mp4
17030 • Process two software sources, explicitly uploading the frames:
17032 ./ffmpeg -init_hw_device cuda=cuda -filter_hw_device cuda \
17034 -f lavfi -i color=size=800x600:color=white,format=yuv420p \
17035 -f lavfi -i yuvtestsrc=size=200x200,format=yuv420p \
17036 -filter_complex \
17037 " \
17038 [0]hwupload[under]; \
17039 [1]hwupload,chromakey_cuda=green:0.1:0.12[over]; \
17040 [under][over]overlay_cuda" \
17041 -c:v hevc_nvenc -cq 18 -preset slow output.mp4
17042 chromanr
17044 Reduce chrominance noise.
17045 The filter accepts the following options:
17047 thres
17049 Set threshold for averaging chrominance values. Sum of absolute
17050 difference of Y, U and V pixel components of current pixel and
17051 neighbour pixels lower than this threshold will be used in
17052 averaging. Luma component is left unchanged and is copied to
17053 output. Default value is 30. Allowed range is from 1 to 200.
17054 sizew
17056 Set horizontal radius of rectangle used for averaging. Allowed
17057 range is from 1 to 100. Default value is 5.
17058 sizeh
17060 Set vertical radius of rectangle used for averaging. Allowed range
17061 is from 1 to 100. Default value is 5.
17062 stepw
17064 Set horizontal step when averaging. Default value is 1. Allowed
17065 range is from 1 to 50. Mostly useful to speed-up filtering.
17066 steph
17068 Set vertical step when averaging. Default value is 1. Allowed
17069 range is from 1 to 50. Mostly useful to speed-up filtering.
17070 threy
17072 Set Y threshold for averaging chrominance values. Set finer
17073 control for max allowed difference between Y components of current
17074 pixel and neigbour pixels. Default value is 200. Allowed range is
17075 from 1 to 200.
17076 threu
17078 Set U threshold for averaging chrominance values. Set finer
17079 control for max allowed difference between U components of current
17080 pixel and neigbour pixels. Default value is 200. Allowed range is
17081 from 1 to 200.
17082 threv
17084 Set V threshold for averaging chrominance values. Set finer
17085 control for max allowed difference between V components of current
17086 pixel and neigbour pixels. Default value is 200. Allowed range is
17087 from 1 to 200.
17088 distance
17090 Set distance type used in calculations.
17091 manhattan
17093 Absolute difference.
17094 euclidean
17096 Difference squared.
17097 Default distance type is manhattan.
17099 Commands
17101 This filter supports same commands as options. The command accepts the
17103 same syntax of the corresponding option.
17104 chromashift
17106 Shift chroma pixels horizontally and/or vertically.
17107 The filter accepts the following options:
17109 cbh Set amount to shift chroma-blue horizontally.
17111 cbv Set amount to shift chroma-blue vertically.
17113 crh Set amount to shift chroma-red horizontally.
17115 crv Set amount to shift chroma-red vertically.
17117 edge
17119 Set edge mode, can be smear, default, or warp.
17120 Commands
17122 This filter supports the all above options as commands.
17124 ciescope
17126 Display CIE color diagram with pixels overlaid onto it.
17127 The filter accepts the following options:
17129 system
17131 Set color system.
17132 ntsc, 470m
17134 ebu, 470bg
17135 smpte
17136 240m
17137 apple
17138 widergb
17139 cie1931
17140 rec709, hdtv
17141 uhdtv, rec2020
17142 dcip3
17143 cie Set CIE system.
17144 xyy
17146 ucs
17147 luv
17148 gamuts
17149 Set what gamuts to draw.
17150 See "system" option for available values.
17152 size, s
17154 Set ciescope size, by default set to 512.
17155 intensity, i
17157 Set intensity used to map input pixel values to CIE diagram.
17158 contrast
17160 Set contrast used to draw tongue colors that are out of active
17161 color system gamut.
17162 corrgamma
17164 Correct gamma displayed on scope, by default enabled.
17165 showwhite
17167 Show white point on CIE diagram, by default disabled.
17168 gamma
17170 Set input gamma. Used only with XYZ input color space.
17171 fill
17173 Fill with CIE colors. By default is enabled.
17174 codecview
17176 Visualize information exported by some codecs.
17177 Some codecs can export information through frames using side-data or
17179 other means. For example, some MPEG based codecs export motion vectors
17180 through the export_mvs flag in the codec flags2 option.
17181 The filter accepts the following option:
17183 block
17185 Display block partition structure using the luma plane.
17186 mv Set motion vectors to visualize.
17188 Available flags for mv are:
17190 pf forward predicted MVs of P-frames
17192 bf forward predicted MVs of B-frames
17194 bb backward predicted MVs of B-frames
17196 qp Display quantization parameters using the chroma planes.
17198 mv_type, mvt
17200 Set motion vectors type to visualize. Includes MVs from all frames
17201 unless specified by frame_type option.
17202 Available flags for mv_type are:
17204 fp forward predicted MVs
17206 bp backward predicted MVs
17208 frame_type, ft
17210 Set frame type to visualize motion vectors of.
17211 Available flags for frame_type are:
17213 if intra-coded frames (I-frames)
17215 pf predicted frames (P-frames)
17217 bf bi-directionally predicted frames (B-frames)
17219 Examples
17221 • Visualize forward predicted MVs of all frames using ffplay:
17223 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
17225 • Visualize multi-directionals MVs of P and B-Frames using ffplay:
17227 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
17229 colorbalance
17231 Modify intensity of primary colors (red, green and blue) of input
17232 frames.
17233 The filter allows an input frame to be adjusted in the shadows,
17235 midtones or highlights regions for the red-cyan, green-magenta or blue-
17236 yellow balance.
17237 A positive adjustment value shifts the balance towards the primary
17239 color, a negative value towards the complementary color.
17240 The filter accepts the following options:
17242 rs
17244 gs
17245 bs Adjust red, green and blue shadows (darkest pixels).
17246 rm
17248 gm
17249 bm Adjust red, green and blue midtones (medium pixels).
17250 rh
17252 gh
17253 bh Adjust red, green and blue highlights (brightest pixels).
17254 Allowed ranges for options are "[-1.0, 1.0]". Defaults are 0.
17256 pl Preserve lightness when changing color balance. Default is
17258 disabled.
17259 Examples
17261 • Add red color cast to shadows:
17263 colorbalance=rs=.3
17265 Commands
17267 This filter supports the all above options as commands.
17269 colorcontrast
17271 Adjust color contrast between RGB components.
17272 The filter accepts the following options:
17274 rc Set the red-cyan contrast. Defaults is 0.0. Allowed range is from
17276 -1.0 to 1.0.
17277 gm Set the green-magenta contrast. Defaults is 0.0. Allowed range is
17279 from -1.0 to 1.0.
17280 by Set the blue-yellow contrast. Defaults is 0.0. Allowed range is
17282 from -1.0 to 1.0.
17283 rcw
17285 gmw
17286 byw Set the weight of each "rc", "gm", "by" option value. Default value
17287 is 0.0. Allowed range is from 0.0 to 1.0. If all weights are 0.0
17288 filtering is disabled.
17289 pl Set the amount of preserving lightness. Default value is 0.0.
17291 Allowed range is from 0.0 to 1.0.
17292 Commands
17294 This filter supports the all above options as commands.
17296 colorcorrect
17298 Adjust color white balance selectively for blacks and whites. This
17299 filter operates in YUV colorspace.
17300 The filter accepts the following options:
17302 rl Set the red shadow spot. Allowed range is from -1.0 to 1.0.
17304 Default value is 0.
17305 bl Set the blue shadow spot. Allowed range is from -1.0 to 1.0.
17307 Default value is 0.
17308 rh Set the red highlight spot. Allowed range is from -1.0 to 1.0.
17310 Default value is 0.
17311 bh Set the red highlight spot. Allowed range is from -1.0 to 1.0.
17313 Default value is 0.
17314 saturation
17316 Set the amount of saturation. Allowed range is from -3.0 to 3.0.
17317 Default value is 1.
17318 analyze
17320 If set to anything other than "manual" it will analyze every frame
17321 and use derived parameters for filtering output frame.
17322 Possible values are:
17324 manual
17326 average
17327 minmax
17328 median
17329 Default value is "manual".
17331 Commands
17333 This filter supports the all above options as commands.
17335 colorchannelmixer
17337 Adjust video input frames by re-mixing color channels.
17338 This filter modifies a color channel by adding the values associated to
17340 the other channels of the same pixels. For example if the value to
17341 modify is red, the output value will be:
17342 <red>=<red>*<rr> + <blue>*<rb> + <green>*<rg> + <alpha>*<ra>
17344 The filter accepts the following options:
17346 rr
17348 rg
17349 rb
17350 ra Adjust contribution of input red, green, blue and alpha channels
17351 for output red channel. Default is 1 for rr, and 0 for rg, rb and
17352 ra.
17353 gr
17355 gg
17356 gb
17357 ga Adjust contribution of input red, green, blue and alpha channels
17358 for output green channel. Default is 1 for gg, and 0 for gr, gb
17359 and ga.
17360 br
17362 bg
17363 bb
17364 ba Adjust contribution of input red, green, blue and alpha channels
17365 for output blue channel. Default is 1 for bb, and 0 for br, bg and
17366 ba.
17367 ar
17369 ag
17370 ab
17371 aa Adjust contribution of input red, green, blue and alpha channels
17372 for output alpha channel. Default is 1 for aa, and 0 for ar, ag
17373 and ab.
17374 Allowed ranges for options are "[-2.0, 2.0]".
17376 pc Set preserve color mode. The accepted values are:
17378 none
17380 Disable color preserving, this is default.
17381 lum Preserve luminance.
17383 max Preserve max value of RGB triplet.
17385 avg Preserve average value of RGB triplet.
17387 sum Preserve sum value of RGB triplet.
17389 nrm Preserve normalized value of RGB triplet.
17391 pwr Preserve power value of RGB triplet.
17393 pa Set the preserve color amount when changing colors. Allowed range
17395 is from "[0.0, 1.0]". Default is 0.0, thus disabled.
17396 Examples
17398 • Convert source to grayscale:
17400 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
17402 • Simulate sepia tones:
17404 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
17406 Commands
17408 This filter supports the all above options as commands.
17410 colorize
17412 Overlay a solid color on the video stream.
17413 The filter accepts the following options:
17415 hue Set the color hue. Allowed range is from 0 to 360. Default value
17417 is 0.
17418 saturation
17420 Set the color saturation. Allowed range is from 0 to 1. Default
17421 value is 0.5.
17422 lightness
17424 Set the color lightness. Allowed range is from 0 to 1. Default
17425 value is 0.5.
17426 mix Set the mix of source lightness. By default is set to 1.0. Allowed
17428 range is from 0.0 to 1.0.
17429 Commands
17431 This filter supports the all above options as commands.
17433 colorkey
17435 RGB colorspace color keying. This filter operates on 8-bit RGB format
17436 frames by setting the alpha component of each pixel which falls within
17437 the similarity radius of the key color to 0. The alpha value for pixels
17438 outside the similarity radius depends on the value of the blend option.
17439 The filter accepts the following options:
17441 color
17443 Set the color for which alpha will be set to 0 (full transparency).
17444 See "Color" section in the ffmpeg-utils manual. Default is
17445 "black".
17446 similarity
17448 Set the radius from the key color within which other colors also
17449 have full transparency. The computed distance is related to the
17450 unit fractional distance in 3D space between the RGB values of the
17451 key color and the pixel's color. Range is 0.01 to 1.0. 0.01 matches
17452 within a very small radius around the exact key color, while 1.0
17453 matches everything. Default is 0.01.
17454 blend
17456 Set how the alpha value for pixels that fall outside the similarity
17457 radius is computed. 0.0 makes pixels either fully transparent or
17458 fully opaque. Higher values result in semi-transparent pixels,
17459 with greater transparency the more similar the pixel color is to
17460 the key color. Range is 0.0 to 1.0. Default is 0.0.
17461 Examples
17463 • Make every green pixel in the input image transparent:
17465 ffmpeg -i input.png -vf colorkey=green out.png
17467 • Overlay a greenscreen-video on top of a static background image.
17469 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
17471 Commands
17473 This filter supports same commands as options. The command accepts the
17475 same syntax of the corresponding option.
17476 If the specified expression is not valid, it is kept at its current
17478 value.
17479 colorhold
17481 Remove all color information for all RGB colors except for certain one.
17482 The filter accepts the following options:
17484 color
17486 The color which will not be replaced with neutral gray.
17487 similarity
17489 Similarity percentage with the above color. 0.01 matches only the
17490 exact key color, while 1.0 matches everything.
17491 blend
17493 Blend percentage. 0.0 makes pixels fully gray. Higher values
17494 result in more preserved color.
17495 Commands
17497 This filter supports same commands as options. The command accepts the
17499 same syntax of the corresponding option.
17500 If the specified expression is not valid, it is kept at its current
17502 value.
17503 colorlevels
17505 Adjust video input frames using levels.
17506 The filter accepts the following options:
17508 rimin
17510 gimin
17511 bimin
17512 aimin
17513 Adjust red, green, blue and alpha input black point. Allowed
17514 ranges for options are "[-1.0, 1.0]". Defaults are 0.
17515 rimax
17517 gimax
17518 bimax
17519 aimax
17520 Adjust red, green, blue and alpha input white point. Allowed
17521 ranges for options are "[-1.0, 1.0]". Defaults are 1.
17522 Input levels are used to lighten highlights (bright tones), darken
17524 shadows (dark tones), change the balance of bright and dark tones.
17525 romin
17527 gomin
17528 bomin
17529 aomin
17530 Adjust red, green, blue and alpha output black point. Allowed
17531 ranges for options are "[0, 1.0]". Defaults are 0.
17532 romax
17534 gomax
17535 bomax
17536 aomax
17537 Adjust red, green, blue and alpha output white point. Allowed
17538 ranges for options are "[0, 1.0]". Defaults are 1.
17539 Output levels allows manual selection of a constrained output level
17541 range.
17542 preserve
17544 Set preserve color mode. The accepted values are:
17545 none
17547 Disable color preserving, this is default.
17548 lum Preserve luminance.
17550 max Preserve max value of RGB triplet.
17552 avg Preserve average value of RGB triplet.
17554 sum Preserve sum value of RGB triplet.
17556 nrm Preserve normalized value of RGB triplet.
17558 pwr Preserve power value of RGB triplet.
17560 Examples
17562 • Make video output darker:
17564 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
17566 • Increase contrast:
17568 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
17570 • Make video output lighter:
17572 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
17574 • Increase brightness:
17576 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
17578 Commands
17580 This filter supports the all above options as commands.
17582 colormap
17584 Apply custom color maps to video stream.
17585 This filter needs three input video streams. First stream is video
17587 stream that is going to be filtered out. Second and third video stream
17588 specify color patches for source color to target color mapping.
17589 The filter accepts the following options:
17591 patch_size
17593 Set the source and target video stream patch size in pixels.
17594 nb_patches
17596 Set the max number of used patches from source and target video
17597 stream. Default value is number of patches available in additional
17598 video streams. Max allowed number of patches is 64.
17599 type
17601 Set the adjustments used for target colors. Can be "relative" or
17602 "absolute". Defaults is "absolute".
17603 kernel
17605 Set the kernel used to measure color differences between mapped
17606 colors.
17607 The accepted values are:
17609 euclidean
17611 weuclidean
17612 Default is "euclidean".
17614 colormatrix
17616 Convert color matrix.
17617 The filter accepts the following options:
17619 src
17621 dst Specify the source and destination color matrix. Both values must
17622 be specified.
17623 The accepted values are:
17625 bt709
17627 BT.709
17628 fcc FCC
17630 bt601
17632 BT.601
17633 bt470
17635 BT.470
17636 bt470bg
17638 BT.470BG
17639 smpte170m
17641 SMPTE-170M
17642 smpte240m
17644 SMPTE-240M
17645 bt2020
17647 BT.2020
17648 For example to convert from BT.601 to SMPTE-240M, use the command:
17650 colormatrix=bt601:smpte240m
17652 colorspace
17654 Convert colorspace, transfer characteristics or color primaries. Input
17655 video needs to have an even size.
17656 The filter accepts the following options:
17658 all Specify all color properties at once.
17660 The accepted values are:
17662 bt470m
17664 BT.470M
17665 bt470bg
17667 BT.470BG
17668 bt601-6-525
17670 BT.601-6 525
17671 bt601-6-625
17673 BT.601-6 625
17674 bt709
17676 BT.709
17677 smpte170m
17679 SMPTE-170M
17680 smpte240m
17682 SMPTE-240M
17683 bt2020
17685 BT.2020
17686 space
17688 Specify output colorspace.
17689 The accepted values are:
17691 bt709
17693 BT.709
17694 fcc FCC
17696 bt470bg
17698 BT.470BG or BT.601-6 625
17699 smpte170m
17701 SMPTE-170M or BT.601-6 525
17702 smpte240m
17704 SMPTE-240M
17705 ycgco
17707 YCgCo
17708 bt2020ncl
17710 BT.2020 with non-constant luminance
17711 trc Specify output transfer characteristics.
17713 The accepted values are:
17715 bt709
17717 BT.709
17718 bt470m
17720 BT.470M
17721 bt470bg
17723 BT.470BG
17724 gamma22
17726 Constant gamma of 2.2
17727 gamma28
17729 Constant gamma of 2.8
17730 smpte170m
17732 SMPTE-170M, BT.601-6 625 or BT.601-6 525
17733 smpte240m
17735 SMPTE-240M
17736 srgb
17738 SRGB
17739 iec61966-2-1
17741 iec61966-2-1
17742 iec61966-2-4
17744 iec61966-2-4
17745 xvycc
17747 xvycc
17748 bt2020-10
17750 BT.2020 for 10-bits content
17751 bt2020-12
17753 BT.2020 for 12-bits content
17754 primaries
17756 Specify output color primaries.
17757 The accepted values are:
17759 bt709
17761 BT.709
17762 bt470m
17764 BT.470M
17765 bt470bg
17767 BT.470BG or BT.601-6 625
17768 smpte170m
17770 SMPTE-170M or BT.601-6 525
17771 smpte240m
17773 SMPTE-240M
17774 film
17776 film
17777 smpte431
17779 SMPTE-431
17780 smpte432
17782 SMPTE-432
17783 bt2020
17785 BT.2020
17786 jedec-p22
17788 JEDEC P22 phosphors
17789 range
17791 Specify output color range.
17792 The accepted values are:
17794 tv TV (restricted) range
17796 mpeg
17798 MPEG (restricted) range
17799 pc PC (full) range
17801 jpeg
17803 JPEG (full) range
17804 format
17806 Specify output color format.
17807 The accepted values are:
17809 yuv420p
17811 YUV 4:2:0 planar 8-bits
17812 yuv420p10
17814 YUV 4:2:0 planar 10-bits
17815 yuv420p12
17817 YUV 4:2:0 planar 12-bits
17818 yuv422p
17820 YUV 4:2:2 planar 8-bits
17821 yuv422p10
17823 YUV 4:2:2 planar 10-bits
17824 yuv422p12
17826 YUV 4:2:2 planar 12-bits
17827 yuv444p
17829 YUV 4:4:4 planar 8-bits
17830 yuv444p10
17832 YUV 4:4:4 planar 10-bits
17833 yuv444p12
17835 YUV 4:4:4 planar 12-bits
17836 fast
17838 Do a fast conversion, which skips gamma/primary correction. This
17839 will take significantly less CPU, but will be mathematically
17840 incorrect. To get output compatible with that produced by the
17841 colormatrix filter, use fast=1.
17842 dither
17844 Specify dithering mode.
17845 The accepted values are:
17847 none
17849 No dithering
17850 fsb Floyd-Steinberg dithering
17852 wpadapt
17854 Whitepoint adaptation mode.
17855 The accepted values are:
17857 bradford
17859 Bradford whitepoint adaptation
17860 vonkries
17862 von Kries whitepoint adaptation
17863 identity
17865 identity whitepoint adaptation (i.e. no whitepoint adaptation)
17866 iall
17868 Override all input properties at once. Same accepted values as all.
17869 ispace
17871 Override input colorspace. Same accepted values as space.
17872 iprimaries
17874 Override input color primaries. Same accepted values as primaries.
17875 itrc
17877 Override input transfer characteristics. Same accepted values as
17878 trc.
17879 irange
17881 Override input color range. Same accepted values as range.
17882 The filter converts the transfer characteristics, color space and color
17884 primaries to the specified user values. The output value, if not
17885 specified, is set to a default value based on the "all" property. If
17886 that property is also not specified, the filter will log an error. The
17887 output color range and format default to the same value as the input
17888 color range and format. The input transfer characteristics, color
17889 space, color primaries and color range should be set on the input data.
17890 If any of these are missing, the filter will log an error and no
17891 conversion will take place.
17892 For example to convert the input to SMPTE-240M, use the command:
17894 colorspace=smpte240m
17896 colortemperature
17898 Adjust color temperature in video to simulate variations in ambient
17899 color temperature.
17900 The filter accepts the following options:
17902 temperature
17904 Set the temperature in Kelvin. Allowed range is from 1000 to 40000.
17905 Default value is 6500 K.
17906 mix Set mixing with filtered output. Allowed range is from 0 to 1.
17908 Default value is 1.
17909 pl Set the amount of preserving lightness. Allowed range is from 0 to
17911 1. Default value is 0.
17912 Commands
17914 This filter supports same commands as options.
17916 convolution
17918 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49
17919 elements.
17920 The filter accepts the following options:
17922 0m
17924 1m
17925 2m
17926 3m Set matrix for each plane. Matrix is sequence of 9, 25 or 49
17927 signed integers in square mode, and from 1 to 49 odd number of
17928 signed integers in row mode.
17929 0rdiv
17931 1rdiv
17932 2rdiv
17933 3rdiv
17934 Set multiplier for calculated value for each plane. If unset or 0,
17935 it will be sum of all matrix elements.
17936 0bias
17938 1bias
17939 2bias
17940 3bias
17941 Set bias for each plane. This value is added to the result of the
17942 multiplication. Useful for making the overall image brighter or
17943 darker. Default is 0.0.
17944 0mode
17946 1mode
17947 2mode
17948 3mode
17949 Set matrix mode for each plane. Can be square, row or column.
17950 Default is square.
17951 Commands
17953 This filter supports the all above options as commands.
17955 Examples
17957 • Apply sharpen:
17959 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"
17961 • Apply blur:
17963 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"
17965 • Apply edge enhance:
17967 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"
17969 • Apply edge detect:
17971 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"
17973 • Apply laplacian edge detector which includes diagonals:
17975 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"
17977 • Apply emboss:
17979 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"
17981 convolve
17983 Apply 2D convolution of video stream in frequency domain using second
17984 stream as impulse.
17985 The filter accepts the following options:
17987 planes
17989 Set which planes to process.
17990 impulse
17992 Set which impulse video frames will be processed, can be first or
17993 all. Default is all.
17994 The "convolve" filter also supports the framesync options.
17996 copy
17998 Copy the input video source unchanged to the output. This is mainly
17999 useful for testing purposes.
18000 coreimage
18002 Video filtering on GPU using Apple's CoreImage API on OSX.
18003 Hardware acceleration is based on an OpenGL context. Usually, this
18005 means it is processed by video hardware. However, software-based OpenGL
18006 implementations exist which means there is no guarantee for hardware
18007 processing. It depends on the respective OSX.
18008 There are many filters and image generators provided by Apple that come
18010 with a large variety of options. The filter has to be referenced by its
18011 name along with its options.
18012 The coreimage filter accepts the following options:
18014 list_filters
18016 List all available filters and generators along with all their
18017 respective options as well as possible minimum and maximum values
18018 along with the default values.
18019 list_filters=true
18021 filter
18023 Specify all filters by their respective name and options. Use
18024 list_filters to determine all valid filter names and options.
18025 Numerical options are specified by a float value and are
18026 automatically clamped to their respective value range. Vector and
18027 color options have to be specified by a list of space separated
18028 float values. Character escaping has to be done. A special option
18029 name "default" is available to use default options for a filter.
18030 It is required to specify either "default" or at least one of the
18032 filter options. All omitted options are used with their default
18033 values. The syntax of the filter string is as follows:
18034 filter=<NAME>@<OPTION>=<VALUE>[@<OPTION>=<VALUE>][@...][#<NAME>@<OPTION>=<VALUE>[@<OPTION>=<VALUE>][@...]][#...]
18036 output_rect
18038 Specify a rectangle where the output of the filter chain is copied
18039 into the input image. It is given by a list of space separated
18040 float values:
18041 output_rect=x\ y\ width\ height
18043 If not given, the output rectangle equals the dimensions of the
18045 input image. The output rectangle is automatically cropped at the
18046 borders of the input image. Negative values are valid for each
18047 component.
18048 output_rect=25\ 25\ 100\ 100
18050 Several filters can be chained for successive processing without GPU-
18052 HOST transfers allowing for fast processing of complex filter chains.
18053 Currently, only filters with zero (generators) or exactly one (filters)
18054 input image and one output image are supported. Also, transition
18055 filters are not yet usable as intended.
18056 Some filters generate output images with additional padding depending
18058 on the respective filter kernel. The padding is automatically removed
18059 to ensure the filter output has the same size as the input image.
18060 For image generators, the size of the output image is determined by the
18062 previous output image of the filter chain or the input image of the
18063 whole filterchain, respectively. The generators do not use the pixel
18064 information of this image to generate their output. However, the
18065 generated output is blended onto this image, resulting in partial or
18066 complete coverage of the output image.
18067 The coreimagesrc video source can be used for generating input images
18069 which are directly fed into the filter chain. By using it, providing
18070 input images by another video source or an input video is not required.
18071 Examples
18073 • List all filters available:
18075 coreimage=list_filters=true
18077 • Use the CIBoxBlur filter with default options to blur an image:
18079 coreimage=filter=CIBoxBlur@default
18081 • Use a filter chain with CISepiaTone at default values and
18083 CIVignetteEffect with its center at 100x100 and a radius of 50
18084 pixels:
18085 coreimage=filter=CIBoxBlur@default#CIVignetteEffect@inputCenter=100\ 100@inputRadius=50
18087 • Use nullsrc and CIQRCodeGenerator to create a QR code for the
18089 FFmpeg homepage, given as complete and escaped command-line for
18090 Apple's standard bash shell:
18091 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@inputMessage=https\\\\\://FFmpeg.org/@inputCorrectionLevel=H -frames:v 1 QRCode.png
18093 cover_rect
18095 Cover a rectangular object
18096 It accepts the following options:
18098 cover
18100 Filepath of the optional cover image, needs to be in yuv420.
18101 mode
18103 Set covering mode.
18104 It accepts the following values:
18106 cover
18108 cover it by the supplied image
18109 blur
18111 cover it by interpolating the surrounding pixels
18112 Default value is blur.
18114 Examples
18116 • Cover a rectangular object by the supplied image of a given video
18118 using ffmpeg:
18119 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
18121 crop
18123 Crop the input video to given dimensions.
18124 It accepts the following parameters:
18126 w, out_w
18128 The width of the output video. It defaults to "iw". This
18129 expression is evaluated only once during the filter configuration,
18130 or when the w or out_w command is sent.
18131 h, out_h
18133 The height of the output video. It defaults to "ih". This
18134 expression is evaluated only once during the filter configuration,
18135 or when the h or out_h command is sent.
18136 x The horizontal position, in the input video, of the left edge of
18138 the output video. It defaults to "(in_w-out_w)/2". This expression
18139 is evaluated per-frame.
18140 y The vertical position, in the input video, of the top edge of the
18142 output video. It defaults to "(in_h-out_h)/2". This expression is
18143 evaluated per-frame.
18144 keep_aspect
18146 If set to 1 will force the output display aspect ratio to be the
18147 same of the input, by changing the output sample aspect ratio. It
18148 defaults to 0.
18149 exact
18151 Enable exact cropping. If enabled, subsampled videos will be
18152 cropped at exact width/height/x/y as specified and will not be
18153 rounded to nearest smaller value. It defaults to 0.
18154 The out_w, out_h, x, y parameters are expressions containing the
18156 following constants:
18157 x
18159 y The computed values for x and y. They are evaluated for each new
18160 frame.
18161 in_w
18163 in_h
18164 The input width and height.
18165 iw
18167 ih These are the same as in_w and in_h.
18168 out_w
18170 out_h
18171 The output (cropped) width and height.
18172 ow
18174 oh These are the same as out_w and out_h.
18175 a same as iw / ih
18177 sar input sample aspect ratio
18179 dar input display aspect ratio, it is the same as (iw / ih) * sar
18181 hsub
18183 vsub
18184 horizontal and vertical chroma subsample values. For example for
18185 the pixel format "yuv422p" hsub is 2 and vsub is 1.
18186 n The number of the input frame, starting from 0.
18188 pos the position in the file of the input frame, NAN if unknown
18190 t The timestamp expressed in seconds. It's NAN if the input timestamp
18192 is unknown.
18193 The expression for out_w may depend on the value of out_h, and the
18195 expression for out_h may depend on out_w, but they cannot depend on x
18196 and y, as x and y are evaluated after out_w and out_h.
18197 The x and y parameters specify the expressions for the position of the
18199 top-left corner of the output (non-cropped) area. They are evaluated
18200 for each frame. If the evaluated value is not valid, it is approximated
18201 to the nearest valid value.
18202 The expression for x may depend on y, and the expression for y may
18204 depend on x.
18205 Examples
18207 • Crop area with size 100x100 at position (12,34).
18209 crop=100:100:12:34
18211 Using named options, the example above becomes:
18213 crop=w=100:h=100:x=12:y=34
18215 • Crop the central input area with size 100x100:
18217 crop=100:100
18219 • Crop the central input area with size 2/3 of the input video:
18221 crop=2/3*in_w:2/3*in_h
18223 • Crop the input video central square:
18225 crop=out_w=in_h
18227 crop=in_h
18228 • Delimit the rectangle with the top-left corner placed at position
18230 100:100 and the right-bottom corner corresponding to the right-
18231 bottom corner of the input image.
18232 crop=in_w-100:in_h-100:100:100
18234 • Crop 10 pixels from the left and right borders, and 20 pixels from
18236 the top and bottom borders
18237 crop=in_w-2*10:in_h-2*20
18239 • Keep only the bottom right quarter of the input image:
18241 crop=in_w/2:in_h/2:in_w/2:in_h/2
18243 • Crop height for getting Greek harmony:
18245 crop=in_w:1/PHI*in_w
18247 • Apply trembling effect:
18249 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)
18251 • Apply erratic camera effect depending on timestamp:
18253 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)"
18255 • Set x depending on the value of y:
18257 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
18259 Commands
18261 This filter supports the following commands:
18263 w, out_w
18265 h, out_h
18266 x
18267 y Set width/height of the output video and the horizontal/vertical
18268 position in the input video. The command accepts the same syntax
18269 of the corresponding option.
18270 If the specified expression is not valid, it is kept at its current
18272 value.
18273 cropdetect
18275 Auto-detect the crop size.
18276 It calculates the necessary cropping parameters and prints the
18278 recommended parameters via the logging system. The detected dimensions
18279 correspond to the non-black area of the input video.
18280 It accepts the following parameters:
18282 limit
18284 Set higher black value threshold, which can be optionally specified
18285 from nothing (0) to everything (255 for 8-bit based formats). An
18286 intensity value greater to the set value is considered non-black.
18287 It defaults to 24. You can also specify a value between 0.0 and
18288 1.0 which will be scaled depending on the bitdepth of the pixel
18289 format.
18290 round
18292 The value which the width/height should be divisible by. It
18293 defaults to 16. The offset is automatically adjusted to center the
18294 video. Use 2 to get only even dimensions (needed for 4:2:2 video).
18295 16 is best when encoding to most video codecs.
18296 skip
18298 Set the number of initial frames for which evaluation is skipped.
18299 Default is 2. Range is 0 to INT_MAX.
18300 reset_count, reset
18302 Set the counter that determines after how many frames cropdetect
18303 will reset the previously detected largest video area and start
18304 over to detect the current optimal crop area. Default value is 0.
18305 This can be useful when channel logos distort the video area. 0
18307 indicates 'never reset', and returns the largest area encountered
18308 during playback.
18309 cue
18311 Delay video filtering until a given wallclock timestamp. The filter
18312 first passes on preroll amount of frames, then it buffers at most
18313 buffer amount of frames and waits for the cue. After reaching the cue
18314 it forwards the buffered frames and also any subsequent frames coming
18315 in its input.
18316 The filter can be used synchronize the output of multiple ffmpeg
18318 processes for realtime output devices like decklink. By putting the
18319 delay in the filtering chain and pre-buffering frames the process can
18320 pass on data to output almost immediately after the target wallclock
18321 timestamp is reached.
18322 Perfect frame accuracy cannot be guaranteed, but the result is good
18324 enough for some use cases.
18325 cue The cue timestamp expressed in a UNIX timestamp in microseconds.
18327 Default is 0.
18328 preroll
18330 The duration of content to pass on as preroll expressed in seconds.
18331 Default is 0.
18332 buffer
18334 The maximum duration of content to buffer before waiting for the
18335 cue expressed in seconds. Default is 0.
18336 curves
18338 Apply color adjustments using curves.
18339 This filter is similar to the Adobe Photoshop and GIMP curves tools.
18341 Each component (red, green and blue) has its values defined by N key
18342 points tied from each other using a smooth curve. The x-axis represents
18343 the pixel values from the input frame, and the y-axis the new pixel
18344 values to be set for the output frame.
18345 By default, a component curve is defined by the two points (0;0) and
18347 (1;1). This creates a straight line where each original pixel value is
18348 "adjusted" to its own value, which means no change to the image.
18349 The filter allows you to redefine these two points and add some more. A
18351 new curve (using a natural cubic spline interpolation) will be define
18352 to pass smoothly through all these new coordinates. The new defined
18353 points needs to be strictly increasing over the x-axis, and their x and
18354 y values must be in the [0;1] interval. If the computed curves
18355 happened to go outside the vector spaces, the values will be clipped
18356 accordingly.
18357 The filter accepts the following options:
18359 preset
18361 Select one of the available color presets. This option can be used
18362 in addition to the r, g, b parameters; in this case, the later
18363 options takes priority on the preset values. Available presets
18364 are:
18365 none
18367 color_negative
18368 cross_process
18369 darker
18370 increase_contrast
18371 lighter
18372 linear_contrast
18373 medium_contrast
18374 negative
18375 strong_contrast
18376 vintage
18377 Default is "none".
18379 master, m
18381 Set the master key points. These points will define a second pass
18382 mapping. It is sometimes called a "luminance" or "value" mapping.
18383 It can be used with r, g, b or all since it acts like a post-
18384 processing LUT.
18385 red, r
18387 Set the key points for the red component.
18388 green, g
18390 Set the key points for the green component.
18391 blue, b
18393 Set the key points for the blue component.
18394 all Set the key points for all components (not including master). Can
18396 be used in addition to the other key points component options. In
18397 this case, the unset component(s) will fallback on this all
18398 setting.
18399 psfile
18401 Specify a Photoshop curves file (".acv") to import the settings
18402 from.
18403 plot
18405 Save Gnuplot script of the curves in specified file.
18406 To avoid some filtergraph syntax conflicts, each key points list need
18408 to be defined using the following syntax: "x0/y0 x1/y1 x2/y2 ...".
18409 Commands
18411 This filter supports same commands as options.
18413 Examples
18415 • Increase slightly the middle level of blue:
18417 curves=blue='0/0 0.5/0.58 1/1'
18419 • Vintage effect:
18421 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'
18423 Here we obtain the following coordinates for each components:
18425 red "(0;0.11) (0.42;0.51) (1;0.95)"
18427 green
18429 "(0;0) (0.50;0.48) (1;1)"
18430 blue
18432 "(0;0.22) (0.49;0.44) (1;0.80)"
18433 • The previous example can also be achieved with the associated
18435 built-in preset:
18436 curves=preset=vintage
18438 • Or simply:
18440 curves=vintage
18442 • Use a Photoshop preset and redefine the points of the green
18444 component:
18445 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
18447 • Check out the curves of the "cross_process" profile using ffmpeg
18449 and gnuplot:
18450 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
18452 gnuplot -p /tmp/curves.plt
18453 datascope
18455 Video data analysis filter.
18456 This filter shows hexadecimal pixel values of part of video.
18458 The filter accepts the following options:
18460 size, s
18462 Set output video size.
18463 x Set x offset from where to pick pixels.
18465 y Set y offset from where to pick pixels.
18467 mode
18469 Set scope mode, can be one of the following:
18470 mono
18472 Draw hexadecimal pixel values with white color on black
18473 background.
18474 color
18476 Draw hexadecimal pixel values with input video pixel color on
18477 black background.
18478 color2
18480 Draw hexadecimal pixel values on color background picked from
18481 input video, the text color is picked in such way so its always
18482 visible.
18483 axis
18485 Draw rows and columns numbers on left and top of video.
18486 opacity
18488 Set background opacity.
18489 format
18491 Set display number format. Can be "hex", or "dec". Default is
18492 "hex".
18493 components
18495 Set pixel components to display. By default all pixel components
18496 are displayed.
18497 Commands
18499 This filter supports same commands as options excluding "size" option.
18501 dblur
18503 Apply Directional blur filter.
18504 The filter accepts the following options:
18506 angle
18508 Set angle of directional blur. Default is 45.
18509 radius
18511 Set radius of directional blur. Default is 5.
18512 planes
18514 Set which planes to filter. By default all planes are filtered.
18515 Commands
18517 This filter supports same commands as options. The command accepts the
18519 same syntax of the corresponding option.
18520 If the specified expression is not valid, it is kept at its current
18522 value.
18523 dctdnoiz
18525 Denoise frames using 2D DCT (frequency domain filtering).
18526 This filter is not designed for real time.
18528 The filter accepts the following options:
18530 sigma, s
18532 Set the noise sigma constant.
18533 This sigma defines a hard threshold of "3 * sigma"; every DCT
18535 coefficient (absolute value) below this threshold with be dropped.
18536 If you need a more advanced filtering, see expr.
18538 Default is 0.
18540 overlap
18542 Set number overlapping pixels for each block. Since the filter can
18543 be slow, you may want to reduce this value, at the cost of a less
18544 effective filter and the risk of various artefacts.
18545 If the overlapping value doesn't permit processing the whole input
18547 width or height, a warning will be displayed and according borders
18548 won't be denoised.
18549 Default value is blocksize-1, which is the best possible setting.
18551 expr, e
18553 Set the coefficient factor expression.
18554 For each coefficient of a DCT block, this expression will be
18556 evaluated as a multiplier value for the coefficient.
18557 If this is option is set, the sigma option will be ignored.
18559 The absolute value of the coefficient can be accessed through the c
18561 variable.
18562 n Set the blocksize using the number of bits. "1<<n" defines the
18564 blocksize, which is the width and height of the processed blocks.
18565 The default value is 3 (8x8) and can be raised to 4 for a blocksize
18567 of 16x16. Note that changing this setting has huge consequences on
18568 the speed processing. Also, a larger block size does not
18569 necessarily means a better de-noising.
18570 Examples
18572 Apply a denoise with a sigma of 4.5:
18574 dctdnoiz=4.5
18576 The same operation can be achieved using the expression system:
18578 dctdnoiz=e='gte(c, 4.5*3)'
18580 Violent denoise using a block size of "16x16":
18582 dctdnoiz=15:n=4
18584 deband
18586 Remove banding artifacts from input video. It works by replacing
18587 banded pixels with average value of referenced pixels.
18588 The filter accepts the following options:
18590 1thr
18592 2thr
18593 3thr
18594 4thr
18595 Set banding detection threshold for each plane. Default is 0.02.
18596 Valid range is 0.00003 to 0.5. If difference between current pixel
18597 and reference pixel is less than threshold, it will be considered
18598 as banded.
18599 range, r
18601 Banding detection range in pixels. Default is 16. If positive,
18602 random number in range 0 to set value will be used. If negative,
18603 exact absolute value will be used. The range defines square of
18604 four pixels around current pixel.
18605 direction, d
18607 Set direction in radians from which four pixel will be compared. If
18608 positive, random direction from 0 to set direction will be picked.
18609 If negative, exact of absolute value will be picked. For example
18610 direction 0, -PI or -2*PI radians will pick only pixels on same row
18611 and -PI/2 will pick only pixels on same column.
18612 blur, b
18614 If enabled, current pixel is compared with average value of all
18615 four surrounding pixels. The default is enabled. If disabled
18616 current pixel is compared with all four surrounding pixels. The
18617 pixel is considered banded if only all four differences with
18618 surrounding pixels are less than threshold.
18619 coupling, c
18621 If enabled, current pixel is changed if and only if all pixel
18622 components are banded, e.g. banding detection threshold is
18623 triggered for all color components. The default is disabled.
18624 Commands
18626 This filter supports the all above options as commands.
18628 deblock
18630 Remove blocking artifacts from input video.
18631 The filter accepts the following options:
18633 filter
18635 Set filter type, can be weak or strong. Default is strong. This
18636 controls what kind of deblocking is applied.
18637 block
18639 Set size of block, allowed range is from 4 to 512. Default is 8.
18640 alpha
18642 beta
18643 gamma
18644 delta
18645 Set blocking detection thresholds. Allowed range is 0 to 1.
18646 Defaults are: 0.098 for alpha and 0.05 for the rest. Using higher
18647 threshold gives more deblocking strength. Setting alpha controls
18648 threshold detection at exact edge of block. Remaining options
18649 controls threshold detection near the edge. Each one for
18650 below/above or left/right. Setting any of those to 0 disables
18651 deblocking.
18652 planes
18654 Set planes to filter. Default is to filter all available planes.
18655 Examples
18657 • Deblock using weak filter and block size of 4 pixels.
18659 deblock=filter=weak:block=4
18661 • Deblock using strong filter, block size of 4 pixels and custom
18663 thresholds for deblocking more edges.
18664 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
18666 • Similar as above, but filter only first plane.
18668 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
18670 • Similar as above, but filter only second and third plane.
18672 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
18674 Commands
18676 This filter supports the all above options as commands.
18678 decimate
18680 Drop duplicated frames at regular intervals.
18681 The filter accepts the following options:
18683 cycle
18685 Set the number of frames from which one will be dropped. Setting
18686 this to N means one frame in every batch of N frames will be
18687 dropped. Default is 5.
18688 dupthresh
18690 Set the threshold for duplicate detection. If the difference metric
18691 for a frame is less than or equal to this value, then it is
18692 declared as duplicate. Default is 1.1
18693 scthresh
18695 Set scene change threshold. Default is 15.
18696 blockx
18698 blocky
18699 Set the size of the x and y-axis blocks used during metric
18700 calculations. Larger blocks give better noise suppression, but
18701 also give worse detection of small movements. Must be a power of
18702 two. Default is 32.
18703 ppsrc
18705 Mark main input as a pre-processed input and activate clean source
18706 input stream. This allows the input to be pre-processed with
18707 various filters to help the metrics calculation while keeping the
18708 frame selection lossless. When set to 1, the first stream is for
18709 the pre-processed input, and the second stream is the clean source
18710 from where the kept frames are chosen. Default is 0.
18711 chroma
18713 Set whether or not chroma is considered in the metric calculations.
18714 Default is 1.
18715 deconvolve
18717 Apply 2D deconvolution of video stream in frequency domain using second
18718 stream as impulse.
18719 The filter accepts the following options:
18721 planes
18723 Set which planes to process.
18724 impulse
18726 Set which impulse video frames will be processed, can be first or
18727 all. Default is all.
18728 noise
18730 Set noise when doing divisions. Default is 0.0000001. Useful when
18731 width and height are not same and not power of 2 or if stream prior
18732 to convolving had noise.
18733 The "deconvolve" filter also supports the framesync options.
18735 dedot
18737 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from
18738 video.
18739 It accepts the following options:
18741 m Set mode of operation. Can be combination of dotcrawl for cross-
18743 luminance reduction and/or rainbows for cross-color reduction.
18744 lt Set spatial luma threshold. Lower values increases reduction of
18746 cross-luminance.
18747 tl Set tolerance for temporal luma. Higher values increases reduction
18749 of cross-luminance.
18750 tc Set tolerance for chroma temporal variation. Higher values
18752 increases reduction of cross-color.
18753 ct Set temporal chroma threshold. Lower values increases reduction of
18755 cross-color.
18756 deflate
18758 Apply deflate effect to the video.
18759 This filter replaces the pixel by the local(3x3) average by taking into
18761 account only values lower than the pixel.
18762 It accepts the following options:
18764 threshold0
18766 threshold1
18767 threshold2
18768 threshold3
18769 Limit the maximum change for each plane, default is 65535. If 0,
18770 plane will remain unchanged.
18771 Commands
18773 This filter supports the all above options as commands.
18775 deflicker
18777 Remove temporal frame luminance variations.
18778 It accepts the following options:
18780 size, s
18782 Set moving-average filter size in frames. Default is 5. Allowed
18783 range is 2 - 129.
18784 mode, m
18786 Set averaging mode to smooth temporal luminance variations.
18787 Available values are:
18789 am Arithmetic mean
18791 gm Geometric mean
18793 hm Harmonic mean
18795 qm Quadratic mean
18797 cm Cubic mean
18799 pm Power mean
18801 median
18803 Median
18804 bypass
18806 Do not actually modify frame. Useful when one only wants metadata.
18807 dejudder
18809 Remove judder produced by partially interlaced telecined content.
18810 Judder can be introduced, for instance, by pullup filter. If the
18812 original source was partially telecined content then the output of
18813 "pullup,dejudder" will have a variable frame rate. May change the
18814 recorded frame rate of the container. Aside from that change, this
18815 filter will not affect constant frame rate video.
18816 The option available in this filter is:
18818 cycle
18820 Specify the length of the window over which the judder repeats.
18821 Accepts any integer greater than 1. Useful values are:
18823 4 If the original was telecined from 24 to 30 fps (Film to NTSC).
18825 5 If the original was telecined from 25 to 30 fps (PAL to NTSC).
18827 20 If a mixture of the two.
18829 The default is 4.
18831 delogo
18833 Suppress a TV station logo by a simple interpolation of the surrounding
18834 pixels. Just set a rectangle covering the logo and watch it disappear
18835 (and sometimes something even uglier appear - your mileage may vary).
18836 It accepts the following parameters:
18838 x
18840 y Specify the top left corner coordinates of the logo. They must be
18841 specified.
18842 w
18844 h Specify the width and height of the logo to clear. They must be
18845 specified.
18846 show
18848 When set to 1, a green rectangle is drawn on the screen to simplify
18849 finding the right x, y, w, and h parameters. The default value is
18850 0.
18851 The rectangle is drawn on the outermost pixels which will be
18853 (partly) replaced with interpolated values. The values of the next
18854 pixels immediately outside this rectangle in each direction will be
18855 used to compute the interpolated pixel values inside the rectangle.
18856 Examples
18858 • Set a rectangle covering the area with top left corner coordinates
18860 0,0 and size 100x77:
18861 delogo=x=0:y=0:w=100:h=77
18863 derain
18865 Remove the rain in the input image/video by applying the derain methods
18866 based on convolutional neural networks. Supported models:
18867 • Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
18869 See
18870 <http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf>.
18871 Training as well as model generation scripts are provided in the
18873 repository at <https://github.com/XueweiMeng/derain_filter.git>.
18874 Native model files (.model) can be generated from TensorFlow model
18876 files (.pb) by using tools/python/convert.py
18877 The filter accepts the following options:
18879 filter_type
18881 Specify which filter to use. This option accepts the following
18882 values:
18883 derain
18885 Derain filter. To conduct derain filter, you need to use a
18886 derain model.
18887 dehaze
18889 Dehaze filter. To conduct dehaze filter, you need to use a
18890 dehaze model.
18891 Default value is derain.
18893 dnn_backend
18895 Specify which DNN backend to use for model loading and execution.
18896 This option accepts the following values:
18897 native
18899 Native implementation of DNN loading and execution.
18900 tensorflow
18902 TensorFlow backend. To enable this backend you need to install
18903 the TensorFlow for C library (see
18904 <https://www.tensorflow.org/install/lang_c>) and configure
18905 FFmpeg with "--enable-libtensorflow"
18906 Default value is native.
18908 model
18910 Set path to model file specifying network architecture and its
18911 parameters. Note that different backends use different file
18912 formats. TensorFlow and native backend can load files for only its
18913 format.
18914 To get full functionality (such as async execution), please use the
18916 dnn_processing filter.
18917 deshake
18919 Attempt to fix small changes in horizontal and/or vertical shift. This
18920 filter helps remove camera shake from hand-holding a camera, bumping a
18921 tripod, moving on a vehicle, etc.
18922 The filter accepts the following options:
18924 x
18926 y
18927 w
18928 h Specify a rectangular area where to limit the search for motion
18929 vectors. If desired the search for motion vectors can be limited
18930 to a rectangular area of the frame defined by its top left corner,
18931 width and height. These parameters have the same meaning as the
18932 drawbox filter which can be used to visualise the position of the
18933 bounding box.
18934 This is useful when simultaneous movement of subjects within the
18936 frame might be confused for camera motion by the motion vector
18937 search.
18938 If any or all of x, y, w and h are set to -1 then the full frame is
18940 used. This allows later options to be set without specifying the
18941 bounding box for the motion vector search.
18942 Default - search the whole frame.
18944 rx
18946 ry Specify the maximum extent of movement in x and y directions in the
18947 range 0-64 pixels. Default 16.
18948 edge
18950 Specify how to generate pixels to fill blanks at the edge of the
18951 frame. Available values are:
18952 blank, 0
18954 Fill zeroes at blank locations
18955 original, 1
18957 Original image at blank locations
18958 clamp, 2
18960 Extruded edge value at blank locations
18961 mirror, 3
18963 Mirrored edge at blank locations
18964 Default value is mirror.
18966 blocksize
18968 Specify the blocksize to use for motion search. Range 4-128 pixels,
18969 default 8.
18970 contrast
18972 Specify the contrast threshold for blocks. Only blocks with more
18973 than the specified contrast (difference between darkest and
18974 lightest pixels) will be considered. Range 1-255, default 125.
18975 search
18977 Specify the search strategy. Available values are:
18978 exhaustive, 0
18980 Set exhaustive search
18981 less, 1
18983 Set less exhaustive search.
18984 Default value is exhaustive.
18986 filename
18988 If set then a detailed log of the motion search is written to the
18989 specified file.
18990 despill
18992 Remove unwanted contamination of foreground colors, caused by reflected
18993 color of greenscreen or bluescreen.
18994 This filter accepts the following options:
18996 type
18998 Set what type of despill to use.
18999 mix Set how spillmap will be generated.
19001 expand
19003 Set how much to get rid of still remaining spill.
19004 red Controls amount of red in spill area.
19006 green
19008 Controls amount of green in spill area. Should be -1 for
19009 greenscreen.
19010 blue
19012 Controls amount of blue in spill area. Should be -1 for
19013 bluescreen.
19014 brightness
19016 Controls brightness of spill area, preserving colors.
19017 alpha
19019 Modify alpha from generated spillmap.
19020 Commands
19022 This filter supports the all above options as commands.
19024 detelecine
19026 Apply an exact inverse of the telecine operation. It requires a
19027 predefined pattern specified using the pattern option which must be the
19028 same as that passed to the telecine filter.
19029 This filter accepts the following options:
19031 first_field
19033 top, t
19034 top field first
19035 bottom, b
19037 bottom field first The default value is "top".
19038 pattern
19040 A string of numbers representing the pulldown pattern you wish to
19041 apply. The default value is 23.
19042 start_frame
19044 A number representing position of the first frame with respect to
19045 the telecine pattern. This is to be used if the stream is cut. The
19046 default value is 0.
19047 dilation
19049 Apply dilation effect to the video.
19050 This filter replaces the pixel by the local(3x3) maximum.
19052 It accepts the following options:
19054 threshold0
19056 threshold1
19057 threshold2
19058 threshold3
19059 Limit the maximum change for each plane, default is 65535. If 0,
19060 plane will remain unchanged.
19061 coordinates
19063 Flag which specifies the pixel to refer to. Default is 255 i.e. all
19064 eight pixels are used.
19065 Flags to local 3x3 coordinates maps like this:
19067 1 2 3
19069 4 5
19070 6 7 8
19071 Commands
19073 This filter supports the all above options as commands.
19075 displace
19077 Displace pixels as indicated by second and third input stream.
19078 It takes three input streams and outputs one stream, the first input is
19080 the source, and second and third input are displacement maps.
19081 The second input specifies how much to displace pixels along the
19083 x-axis, while the third input specifies how much to displace pixels
19084 along the y-axis. If one of displacement map streams terminates, last
19085 frame from that displacement map will be used.
19086 Note that once generated, displacements maps can be reused over and
19088 over again.
19089 A description of the accepted options follows.
19091 edge
19093 Set displace behavior for pixels that are out of range.
19094 Available values are:
19096 blank
19098 Missing pixels are replaced by black pixels.
19099 smear
19101 Adjacent pixels will spread out to replace missing pixels.
19102 wrap
19104 Out of range pixels are wrapped so they point to pixels of
19105 other side.
19106 mirror
19108 Out of range pixels will be replaced with mirrored pixels.
19109 Default is smear.
19111 Examples
19113 • Add ripple effect to rgb input of video size hd720:
19115 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
19117 • Add wave effect to rgb input of video size hd720:
19119 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
19121 dnn_classify
19123 Do classification with deep neural networks based on bounding boxes.
19124 The filter accepts the following options:
19126 dnn_backend
19128 Specify which DNN backend to use for model loading and execution.
19129 This option accepts only openvino now, tensorflow backends will be
19130 added.
19131 model
19133 Set path to model file specifying network architecture and its
19134 parameters. Note that different backends use different file
19135 formats.
19136 input
19138 Set the input name of the dnn network.
19139 output
19141 Set the output name of the dnn network.
19142 confidence
19144 Set the confidence threshold (default: 0.5).
19145 labels
19147 Set path to label file specifying the mapping between label id and
19148 name. Each label name is written in one line, tailing spaces and
19149 empty lines are skipped. The first line is the name of label id 0,
19150 and the second line is the name of label id 1, etc. The label id
19151 is considered as name if the label file is not provided.
19152 backend_configs
19154 Set the configs to be passed into backend
19155 For tensorflow backend, you can set its configs with sess_config
19157 options, please use tools/python/tf_sess_config.py to get the
19158 configs for your system.
19159 dnn_detect
19161 Do object detection with deep neural networks.
19162 The filter accepts the following options:
19164 dnn_backend
19166 Specify which DNN backend to use for model loading and execution.
19167 This option accepts only openvino now, tensorflow backends will be
19168 added.
19169 model
19171 Set path to model file specifying network architecture and its
19172 parameters. Note that different backends use different file
19173 formats.
19174 input
19176 Set the input name of the dnn network.
19177 output
19179 Set the output name of the dnn network.
19180 confidence
19182 Set the confidence threshold (default: 0.5).
19183 labels
19185 Set path to label file specifying the mapping between label id and
19186 name. Each label name is written in one line, tailing spaces and
19187 empty lines are skipped. The first line is the name of label id 0
19188 (usually it is 'background'), and the second line is the name of
19189 label id 1, etc. The label id is considered as name if the label
19190 file is not provided.
19191 backend_configs
19193 Set the configs to be passed into backend. To use async execution,
19194 set async (default: set). Roll back to sync execution if the
19195 backend does not support async.
19196 dnn_processing
19198 Do image processing with deep neural networks. It works together with
19199 another filter which converts the pixel format of the Frame to what the
19200 dnn network requires.
19201 The filter accepts the following options:
19203 dnn_backend
19205 Specify which DNN backend to use for model loading and execution.
19206 This option accepts the following values:
19207 native
19209 Native implementation of DNN loading and execution.
19210 tensorflow
19212 TensorFlow backend. To enable this backend you need to install
19213 the TensorFlow for C library (see
19214 <https://www.tensorflow.org/install/lang_c>) and configure
19215 FFmpeg with "--enable-libtensorflow"
19216 openvino
19218 OpenVINO backend. To enable this backend you need to build and
19219 install the OpenVINO for C library (see
19220 <https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md>)
19221 and configure FFmpeg with "--enable-libopenvino"
19222 (--extra-cflags=-I... --extra-ldflags=-L... might be needed if
19223 the header files and libraries are not installed into system
19224 path)
19225 Default value is native.
19227 model
19229 Set path to model file specifying network architecture and its
19230 parameters. Note that different backends use different file
19231 formats. TensorFlow, OpenVINO and native backend can load files for
19232 only its format.
19233 Native model file (.model) can be generated from TensorFlow model
19235 file (.pb) by using tools/python/convert.py
19236 input
19238 Set the input name of the dnn network.
19239 output
19241 Set the output name of the dnn network.
19242 backend_configs
19244 Set the configs to be passed into backend. To use async execution,
19245 set async (default: set). Roll back to sync execution if the
19246 backend does not support async.
19247 For tensorflow backend, you can set its configs with sess_config
19249 options, please use tools/python/tf_sess_config.py to get the
19250 configs of TensorFlow backend for your system.
19251 Examples
19253 • Remove rain in rgb24 frame with can.pb (see derain filter):
19255 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
19257 • Halve the pixel value of the frame with format gray32f:
19259 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
19261 • Handle the Y channel with srcnn.pb (see sr filter) for frame with
19263 yuv420p (planar YUV formats supported):
19264 ./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
19266 • Handle the Y channel with espcn.pb (see sr filter), which changes
19268 frame size, for format yuv420p (planar YUV formats supported),
19269 please use tools/python/tf_sess_config.py to get the configs of
19270 TensorFlow backend for your system.
19271 ./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
19273 drawbox
19275 Draw a colored box on the input image.
19276 It accepts the following parameters:
19278 x
19280 y The expressions which specify the top left corner coordinates of
19281 the box. It defaults to 0.
19282 width, w
19284 height, h
19285 The expressions which specify the width and height of the box; if 0
19286 they are interpreted as the input width and height. It defaults to
19287 0.
19288 color, c
19290 Specify the color of the box to write. For the general syntax of
19291 this option, check the "Color" section in the ffmpeg-utils manual.
19292 If the special value "invert" is used, the box edge color is the
19293 same as the video with inverted luma.
19294 thickness, t
19296 The expression which sets the thickness of the box edge. A value
19297 of "fill" will create a filled box. Default value is 3.
19298 See below for the list of accepted constants.
19300 replace
19302 Applicable if the input has alpha. With value 1, the pixels of the
19303 painted box will overwrite the video's color and alpha pixels.
19304 Default is 0, which composites the box onto the input, leaving the
19305 video's alpha intact.
19306 The parameters for x, y, w and h and t are expressions containing the
19308 following constants:
19309 dar The input display aspect ratio, it is the same as (w / h) * sar.
19311 hsub
19313 vsub
19314 horizontal and vertical chroma subsample values. For example for
19315 the pixel format "yuv422p" hsub is 2 and vsub is 1.
19316 in_h, ih
19318 in_w, iw
19319 The input width and height.
19320 sar The input sample aspect ratio.
19322 x
19324 y The x and y offset coordinates where the box is drawn.
19325 w
19327 h The width and height of the drawn box.
19328 box_source
19330 Box source can be set as side_data_detection_bboxes if you want to
19331 use box data in detection bboxes of side data.
19332 If box_source is set, the x, y, width and height will be ignored
19334 and still use box data in detection bboxes of side data. So please
19335 do not use this parameter if you were not sure about the box
19336 source.
19337 t The thickness of the drawn box.
19339 These constants allow the x, y, w, h and t expressions to refer to
19341 each other, so you may for example specify "y=x/dar" or "h=w/dar".
19342 Examples
19344 • Draw a black box around the edge of the input image:
19346 drawbox
19348 • Draw a box with color red and an opacity of 50%:
19350 drawbox=10:20:200:60:red@0.5
19352 The previous example can be specified as:
19354 drawbox=x=10:y=20:w=200:h=60:color=red@0.5
19356 • Fill the box with pink color:
19358 drawbox=x=10:y=10:w=100:h=100:color=pink@0.5:t=fill
19360 • Draw a 2-pixel red 2.40:1 mask:
19362 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
19364 Commands
19366 This filter supports same commands as options. The command accepts the
19368 same syntax of the corresponding option.
19369 If the specified expression is not valid, it is kept at its current
19371 value.
19372 drawgraph
19374 Draw a graph using input video metadata.
19375 It accepts the following parameters:
19377 m1 Set 1st frame metadata key from which metadata values will be used
19379 to draw a graph.
19380 fg1 Set 1st foreground color expression.
19382 m2 Set 2nd frame metadata key from which metadata values will be used
19384 to draw a graph.
19385 fg2 Set 2nd foreground color expression.
19387 m3 Set 3rd frame metadata key from which metadata values will be used
19389 to draw a graph.
19390 fg3 Set 3rd foreground color expression.
19392 m4 Set 4th frame metadata key from which metadata values will be used
19394 to draw a graph.
19395 fg4 Set 4th foreground color expression.
19397 min Set minimal value of metadata value.
19399 max Set maximal value of metadata value.
19401 bg Set graph background color. Default is white.
19403 mode
19405 Set graph mode.
19406 Available values for mode is:
19408 bar
19410 dot
19411 line
19412 Default is "line".
19414 slide
19416 Set slide mode.
19417 Available values for slide is:
19419 frame
19421 Draw new frame when right border is reached.
19422 replace
19424 Replace old columns with new ones.
19425 scroll
19427 Scroll from right to left.
19428 rscroll
19430 Scroll from left to right.
19431 picture
19433 Draw single picture.
19434 Default is "frame".
19436 size
19438 Set size of graph video. For the syntax of this option, check the
19439 "Video size" section in the ffmpeg-utils manual. The default value
19440 is "900x256".
19441 rate, r
19443 Set the output frame rate. Default value is 25.
19444 The foreground color expressions can use the following variables:
19446 MIN Minimal value of metadata value.
19448 MAX Maximal value of metadata value.
19450 VAL Current metadata key value.
19452 The color is defined as 0xAABBGGRR.
19454 Example using metadata from signalstats filter:
19456 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
19458 Example using metadata from ebur128 filter:
19460 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
19462 drawgrid
19464 Draw a grid on the input image.
19465 It accepts the following parameters:
19467 x
19469 y The expressions which specify the coordinates of some point of grid
19470 intersection (meant to configure offset). Both default to 0.
19471 width, w
19473 height, h
19474 The expressions which specify the width and height of the grid
19475 cell, if 0 they are interpreted as the input width and height,
19476 respectively, minus "thickness", so image gets framed. Default to
19477 0.
19478 color, c
19480 Specify the color of the grid. For the general syntax of this
19481 option, check the "Color" section in the ffmpeg-utils manual. If
19482 the special value "invert" is used, the grid color is the same as
19483 the video with inverted luma.
19484 thickness, t
19486 The expression which sets the thickness of the grid line. Default
19487 value is 1.
19488 See below for the list of accepted constants.
19490 replace
19492 Applicable if the input has alpha. With 1 the pixels of the painted
19493 grid will overwrite the video's color and alpha pixels. Default is
19494 0, which composites the grid onto the input, leaving the video's
19495 alpha intact.
19496 The parameters for x, y, w and h and t are expressions containing the
19498 following constants:
19499 dar The input display aspect ratio, it is the same as (w / h) * sar.
19501 hsub
19503 vsub
19504 horizontal and vertical chroma subsample values. For example for
19505 the pixel format "yuv422p" hsub is 2 and vsub is 1.
19506 in_h, ih
19508 in_w, iw
19509 The input grid cell width and height.
19510 sar The input sample aspect ratio.
19512 x
19514 y The x and y coordinates of some point of grid intersection (meant
19515 to configure offset).
19516 w
19518 h The width and height of the drawn cell.
19519 t The thickness of the drawn cell.
19521 These constants allow the x, y, w, h and t expressions to refer to
19523 each other, so you may for example specify "y=x/dar" or "h=w/dar".
19524 Examples
19526 • Draw a grid with cell 100x100 pixels, thickness 2 pixels, with
19528 color red and an opacity of 50%:
19529 drawgrid=width=100:height=100:thickness=2:color=red@0.5
19531 • Draw a white 3x3 grid with an opacity of 50%:
19533 drawgrid=w=iw/3:h=ih/3:t=2:c=white@0.5
19535 Commands
19537 This filter supports same commands as options. The command accepts the
19539 same syntax of the corresponding option.
19540 If the specified expression is not valid, it is kept at its current
19542 value.
19543 drawtext
19545 Draw a text string or text from a specified file on top of a video,
19546 using the libfreetype library.
19547 To enable compilation of this filter, you need to configure FFmpeg with
19549 "--enable-libfreetype". To enable default font fallback and the font
19550 option you need to configure FFmpeg with "--enable-libfontconfig". To
19551 enable the text_shaping option, you need to configure FFmpeg with
19552 "--enable-libfribidi".
19553 Syntax
19555 It accepts the following parameters:
19557 box Used to draw a box around text using the background color. The
19559 value must be either 1 (enable) or 0 (disable). The default value
19560 of box is 0.
19561 boxborderw
19563 Set the width of the border to be drawn around the box using
19564 boxcolor. The default value of boxborderw is 0.
19565 boxcolor
19567 The color to be used for drawing box around text. For the syntax of
19568 this option, check the "Color" section in the ffmpeg-utils manual.
19569 The default value of boxcolor is "white".
19571 line_spacing
19573 Set the line spacing in pixels of the border to be drawn around the
19574 box using box. The default value of line_spacing is 0.
19575 borderw
19577 Set the width of the border to be drawn around the text using
19578 bordercolor. The default value of borderw is 0.
19579 bordercolor
19581 Set the color to be used for drawing border around text. For the
19582 syntax of this option, check the "Color" section in the ffmpeg-
19583 utils manual.
19584 The default value of bordercolor is "black".
19586 expansion
19588 Select how the text is expanded. Can be either "none", "strftime"
19589 (deprecated) or "normal" (default). See the drawtext_expansion,
19590 Text expansion section below for details.
19591 basetime
19593 Set a start time for the count. Value is in microseconds. Only
19594 applied in the deprecated strftime expansion mode. To emulate in
19595 normal expansion mode use the "pts" function, supplying the start
19596 time (in seconds) as the second argument.
19597 fix_bounds
19599 If true, check and fix text coords to avoid clipping.
19600 fontcolor
19602 The color to be used for drawing fonts. For the syntax of this
19603 option, check the "Color" section in the ffmpeg-utils manual.
19604 The default value of fontcolor is "black".
19606 fontcolor_expr
19608 String which is expanded the same way as text to obtain dynamic
19609 fontcolor value. By default this option has empty value and is not
19610 processed. When this option is set, it overrides fontcolor option.
19611 font
19613 The font family to be used for drawing text. By default Sans.
19614 fontfile
19616 The font file to be used for drawing text. The path must be
19617 included. This parameter is mandatory if the fontconfig support is
19618 disabled.
19619 alpha
19621 Draw the text applying alpha blending. The value can be a number
19622 between 0.0 and 1.0. The expression accepts the same variables x,
19623 y as well. The default value is 1. Please see fontcolor_expr.
19624 fontsize
19626 The font size to be used for drawing text. The default value of
19627 fontsize is 16.
19628 text_shaping
19630 If set to 1, attempt to shape the text (for example, reverse the
19631 order of right-to-left text and join Arabic characters) before
19632 drawing it. Otherwise, just draw the text exactly as given. By
19633 default 1 (if supported).
19634 ft_load_flags
19636 The flags to be used for loading the fonts.
19637 The flags map the corresponding flags supported by libfreetype, and
19639 are a combination of the following values:
19640 default
19642 no_scale
19643 no_hinting
19644 render
19645 no_bitmap
19646 vertical_layout
19647 force_autohint
19648 crop_bitmap
19649 pedantic
19650 ignore_global_advance_width
19651 no_recurse
19652 ignore_transform
19653 monochrome
19654 linear_design
19655 no_autohint
19656 Default value is "default".
19658 For more information consult the documentation for the FT_LOAD_*
19660 libfreetype flags.
19661 shadowcolor
19663 The color to be used for drawing a shadow behind the drawn text.
19664 For the syntax of this option, check the "Color" section in the
19665 ffmpeg-utils manual.
19666 The default value of shadowcolor is "black".
19668 shadowx
19670 shadowy
19671 The x and y offsets for the text shadow position with respect to
19672 the position of the text. They can be either positive or negative
19673 values. The default value for both is "0".
19674 start_number
19676 The starting frame number for the n/frame_num variable. The default
19677 value is "0".
19678 tabsize
19680 The size in number of spaces to use for rendering the tab. Default
19681 value is 4.
19682 timecode
19684 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
19685 format. It can be used with or without text parameter.
19686 timecode_rate option must be specified.
19687 timecode_rate, rate, r
19689 Set the timecode frame rate (timecode only). Value will be rounded
19690 to nearest integer. Minimum value is "1". Drop-frame timecode is
19691 supported for frame rates 30 & 60.
19692 tc24hmax
19694 If set to 1, the output of the timecode option will wrap around at
19695 24 hours. Default is 0 (disabled).
19696 text
19698 The text string to be drawn. The text must be a sequence of UTF-8
19699 encoded characters. This parameter is mandatory if no file is
19700 specified with the parameter textfile.
19701 textfile
19703 A text file containing text to be drawn. The text must be a
19704 sequence of UTF-8 encoded characters.
19705 This parameter is mandatory if no text string is specified with the
19707 parameter text.
19708 If both text and textfile are specified, an error is thrown.
19710 text_source
19712 Text source should be set as side_data_detection_bboxes if you want
19713 to use text data in detection bboxes of side data.
19714 If text source is set, text and textfile will be ignored and still
19716 use text data in detection bboxes of side data. So please do not
19717 use this parameter if you are not sure about the text source.
19718 reload
19720 The textfile will be reloaded at specified frame interval. Be sure
19721 to update textfile atomically, or it may be read partially, or even
19722 fail. Range is 0 to INT_MAX. Default is 0.
19723 x
19725 y The expressions which specify the offsets where text will be drawn
19726 within the video frame. They are relative to the top/left border of
19727 the output image.
19728 The default value of x and y is "0".
19730 See below for the list of accepted constants and functions.
19732 The parameters for x and y are expressions containing the following
19734 constants and functions:
19735 dar input display aspect ratio, it is the same as (w / h) * sar
19737 hsub
19739 vsub
19740 horizontal and vertical chroma subsample values. For example for
19741 the pixel format "yuv422p" hsub is 2 and vsub is 1.
19742 line_h, lh
19744 the height of each text line
19745 main_h, h, H
19747 the input height
19748 main_w, w, W
19750 the input width
19751 max_glyph_a, ascent
19753 the maximum distance from the baseline to the highest/upper grid
19754 coordinate used to place a glyph outline point, for all the
19755 rendered glyphs. It is a positive value, due to the grid's
19756 orientation with the Y axis upwards.
19757 max_glyph_d, descent
19759 the maximum distance from the baseline to the lowest grid
19760 coordinate used to place a glyph outline point, for all the
19761 rendered glyphs. This is a negative value, due to the grid's
19762 orientation, with the Y axis upwards.
19763 max_glyph_h
19765 maximum glyph height, that is the maximum height for all the glyphs
19766 contained in the rendered text, it is equivalent to ascent -
19767 descent.
19768 max_glyph_w
19770 maximum glyph width, that is the maximum width for all the glyphs
19771 contained in the rendered text
19772 n the number of input frame, starting from 0
19774 rand(min, max)
19776 return a random number included between min and max
19777 sar The input sample aspect ratio.
19779 t timestamp expressed in seconds, NAN if the input timestamp is
19781 unknown
19782 text_h, th
19784 the height of the rendered text
19785 text_w, tw
19787 the width of the rendered text
19788 x
19790 y the x and y offset coordinates where the text is drawn.
19791 These parameters allow the x and y expressions to refer to each
19793 other, so you can for example specify "y=x/dar".
19794 pict_type
19796 A one character description of the current frame's picture type.
19797 pkt_pos
19799 The current packet's position in the input file or stream (in
19800 bytes, from the start of the input). A value of -1 indicates this
19801 info is not available.
19802 pkt_duration
19804 The current packet's duration, in seconds.
19805 pkt_size
19807 The current packet's size (in bytes).
19808 Text expansion
19810 If expansion is set to "strftime", the filter recognizes strftime()
19812 sequences in the provided text and expands them accordingly. Check the
19813 documentation of strftime(). This feature is deprecated.
19814 If expansion is set to "none", the text is printed verbatim.
19816 If expansion is set to "normal" (which is the default), the following
19818 expansion mechanism is used.
19819 The backslash character \, followed by any character, always expands to
19821 the second character.
19822 Sequences of the form "%{...}" are expanded. The text between the
19824 braces is a function name, possibly followed by arguments separated by
19825 ':'. If the arguments contain special characters or delimiters (':' or
19826 '}'), they should be escaped.
19827 Note that they probably must also be escaped as the value for the text
19829 option in the filter argument string and as the filter argument in the
19830 filtergraph description, and possibly also for the shell, that makes up
19831 to four levels of escaping; using a text file avoids these problems.
19832 The following functions are available:
19834 expr, e
19836 The expression evaluation result.
19837 It must take one argument specifying the expression to be
19839 evaluated, which accepts the same constants and functions as the x
19840 and y values. Note that not all constants should be used, for
19841 example the text size is not known when evaluating the expression,
19842 so the constants text_w and text_h will have an undefined value.
19843 expr_int_format, eif
19845 Evaluate the expression's value and output as formatted integer.
19846 The first argument is the expression to be evaluated, just as for
19848 the expr function. The second argument specifies the output
19849 format. Allowed values are x, X, d and u. They are treated exactly
19850 as in the "printf" function. The third parameter is optional and
19851 sets the number of positions taken by the output. It can be used
19852 to add padding with zeros from the left.
19853 gmtime
19855 The time at which the filter is running, expressed in UTC. It can
19856 accept an argument: a strftime() format string. The format string
19857 is extended to support the variable %[1-6]N which prints fractions
19858 of the second with optionally specified number of digits.
19859 localtime
19861 The time at which the filter is running, expressed in the local
19862 time zone. It can accept an argument: a strftime() format string.
19863 The format string is extended to support the variable %[1-6]N which
19864 prints fractions of the second with optionally specified number of
19865 digits.
19866 metadata
19868 Frame metadata. Takes one or two arguments.
19869 The first argument is mandatory and specifies the metadata key.
19871 The second argument is optional and specifies a default value, used
19873 when the metadata key is not found or empty.
19874 Available metadata can be identified by inspecting entries starting
19876 with TAG included within each frame section printed by running
19877 "ffprobe -show_frames".
19878 String metadata generated in filters leading to the drawtext filter
19880 are also available.
19881 n, frame_num
19883 The frame number, starting from 0.
19884 pict_type
19886 A one character description of the current picture type.
19887 pts The timestamp of the current frame. It can take up to three
19889 arguments.
19890 The first argument is the format of the timestamp; it defaults to
19892 "flt" for seconds as a decimal number with microsecond accuracy;
19893 "hms" stands for a formatted [-]HH:MM:SS.mmm timestamp with
19894 millisecond accuracy. "gmtime" stands for the timestamp of the
19895 frame formatted as UTC time; "localtime" stands for the timestamp
19896 of the frame formatted as local time zone time.
19897 The second argument is an offset added to the timestamp.
19899 If the format is set to "hms", a third argument "24HH" may be
19901 supplied to present the hour part of the formatted timestamp in 24h
19902 format (00-23).
19903 If the format is set to "localtime" or "gmtime", a third argument
19905 may be supplied: a strftime() format string. By default, YYYY-MM-
19906 DD HH:MM:SS format will be used.
19907 Commands
19909 This filter supports altering parameters via commands:
19911 reinit
19913 Alter existing filter parameters.
19914 Syntax for the argument is the same as for filter invocation, e.g.
19916 fontsize=56:fontcolor=green:text='Hello World'
19918 Full filter invocation with sendcmd would look like this:
19920 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
19922 If the entire argument can't be parsed or applied as valid values then
19924 the filter will continue with its existing parameters.
19925 Examples
19927 • Draw "Test Text" with font FreeSerif, using the default values for
19929 the optional parameters.
19930 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
19932 • Draw 'Test Text' with font FreeSerif of size 24 at position x=100
19934 and y=50 (counting from the top-left corner of the screen), text is
19935 yellow with a red box around it. Both the text and the box have an
19936 opacity of 20%.
19937 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
19939 x=100: y=50: fontsize=24: fontcolor=yellow@0.2: box=1: boxcolor=red@0.2"
19940 Note that the double quotes are not necessary if spaces are not
19942 used within the parameter list.
19943 • Show the text at the center of the video frame:
19945 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
19947 • Show the text at a random position, switching to a new position
19949 every 30 seconds:
19950 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)"
19952 • Show a text line sliding from right to left in the last row of the
19954 video frame. The file LONG_LINE is assumed to contain a single line
19955 with no newlines.
19956 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
19958 • Show the content of file CREDITS off the bottom of the frame and
19960 scroll up.
19961 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
19963 • Draw a single green letter "g", at the center of the input video.
19965 The glyph baseline is placed at half screen height.
19966 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
19968 • Show text for 1 second every 3 seconds:
19970 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
19972 • Use fontconfig to set the font. Note that the colons need to be
19974 escaped.
19975 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
19977 • Draw "Test Text" with font size dependent on height of the video.
19979 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
19981 • Print the date of a real-time encoding (see strftime(3)):
19983 drawtext='fontfile=FreeSans.ttf:text=%{localtime\:%a %b %d %Y}'
19985 • Show text fading in and out (appearing/disappearing):
19987 #!/bin/sh
19989 DS=1.0 # display start
19990 DE=10.0 # display end
19991 FID=1.5 # fade in duration
19992 FOD=5 # fade out duration
19993 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 }"
19994 • Horizontally align multiple separate texts. Note that max_glyph_a
19996 and the fontsize value are included in the y offset.
19997 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
19999 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
20000 • Plot special lavf.image2dec.source_basename metadata onto each
20002 frame if such metadata exists. Otherwise, plot the string "NA".
20003 Note that image2 demuxer must have option -export_path_metadata 1
20004 for the special metadata fields to be available for filters.
20005 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%{metadata\:lavf.image2dec.source_basename\:NA}':x=10:y=10"
20007 For more information about libfreetype, check:
20009 <http://www.freetype.org/>.
20010 For more information about fontconfig, check:
20012 <http://freedesktop.org/software/fontconfig/fontconfig-user.html>.
20013 For more information about libfribidi, check: <http://fribidi.org/>.
20015 edgedetect
20017 Detect and draw edges. The filter uses the Canny Edge Detection
20018 algorithm.
20019 The filter accepts the following options:
20021 low
20023 high
20024 Set low and high threshold values used by the Canny thresholding
20025 algorithm.
20026 The high threshold selects the "strong" edge pixels, which are then
20028 connected through 8-connectivity with the "weak" edge pixels
20029 selected by the low threshold.
20030 low and high threshold values must be chosen in the range [0,1],
20032 and low should be lesser or equal to high.
20033 Default value for low is "20/255", and default value for high is
20035 "50/255".
20036 mode
20038 Define the drawing mode.
20039 wires
20041 Draw white/gray wires on black background.
20042 colormix
20044 Mix the colors to create a paint/cartoon effect.
20045 canny
20047 Apply Canny edge detector on all selected planes.
20048 Default value is wires.
20050 planes
20052 Select planes for filtering. By default all available planes are
20053 filtered.
20054 Examples
20056 • Standard edge detection with custom values for the hysteresis
20058 thresholding:
20059 edgedetect=low=0.1:high=0.4
20061 • Painting effect without thresholding:
20063 edgedetect=mode=colormix:high=0
20065 elbg
20067 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
20068 For each input image, the filter will compute the optimal mapping from
20070 the input to the output given the codebook length, that is the number
20071 of distinct output colors.
20072 This filter accepts the following options.
20074 codebook_length, l
20076 Set codebook length. The value must be a positive integer, and
20077 represents the number of distinct output colors. Default value is
20078 256.
20079 nb_steps, n
20081 Set the maximum number of iterations to apply for computing the
20082 optimal mapping. The higher the value the better the result and the
20083 higher the computation time. Default value is 1.
20084 seed, s
20086 Set a random seed, must be an integer included between 0 and
20087 UINT32_MAX. If not specified, or if explicitly set to -1, the
20088 filter will try to use a good random seed on a best effort basis.
20089 pal8
20091 Set pal8 output pixel format. This option does not work with
20092 codebook length greater than 256. Default is disabled.
20093 use_alpha
20095 Include alpha values in the quantization calculation. Allows
20096 creating palettized output images (e.g. PNG8) with multiple alpha
20097 smooth blending.
20098 entropy
20100 Measure graylevel entropy in histogram of color channels of video
20101 frames.
20102 It accepts the following parameters:
20104 mode
20106 Can be either normal or diff. Default is normal.
20107 diff mode measures entropy of histogram delta values, absolute
20109 differences between neighbour histogram values.
20110 epx
20112 Apply the EPX magnification filter which is designed for pixel art.
20113 It accepts the following option:
20115 n Set the scaling dimension: 2 for "2xEPX", 3 for "3xEPX". Default
20117 is 3.
20118 eq
20120 Set brightness, contrast, saturation and approximate gamma adjustment.
20121 The filter accepts the following options:
20123 contrast
20125 Set the contrast expression. The value must be a float value in
20126 range "-1000.0" to 1000.0. The default value is "1".
20127 brightness
20129 Set the brightness expression. The value must be a float value in
20130 range "-1.0" to 1.0. The default value is "0".
20131 saturation
20133 Set the saturation expression. The value must be a float in range
20134 0.0 to 3.0. The default value is "1".
20135 gamma
20137 Set the gamma expression. The value must be a float in range 0.1 to
20138 10.0. The default value is "1".
20139 gamma_r
20141 Set the gamma expression for red. The value must be a float in
20142 range 0.1 to 10.0. The default value is "1".
20143 gamma_g
20145 Set the gamma expression for green. The value must be a float in
20146 range 0.1 to 10.0. The default value is "1".
20147 gamma_b
20149 Set the gamma expression for blue. The value must be a float in
20150 range 0.1 to 10.0. The default value is "1".
20151 gamma_weight
20153 Set the gamma weight expression. It can be used to reduce the
20154 effect of a high gamma value on bright image areas, e.g. keep them
20155 from getting overamplified and just plain white. The value must be
20156 a float in range 0.0 to 1.0. A value of 0.0 turns the gamma
20157 correction all the way down while 1.0 leaves it at its full
20158 strength. Default is "1".
20159 eval
20161 Set when the expressions for brightness, contrast, saturation and
20162 gamma expressions are evaluated.
20163 It accepts the following values:
20165 init
20167 only evaluate expressions once during the filter initialization
20168 or when a command is processed
20169 frame
20171 evaluate expressions for each incoming frame
20172 Default value is init.
20174 The expressions accept the following parameters:
20176 n frame count of the input frame starting from 0
20178 pos byte position of the corresponding packet in the input file, NAN if
20180 unspecified
20181 r frame rate of the input video, NAN if the input frame rate is
20183 unknown
20184 t timestamp expressed in seconds, NAN if the input timestamp is
20186 unknown
20187 Commands
20189 The filter supports the following commands:
20191 contrast
20193 Set the contrast expression.
20194 brightness
20196 Set the brightness expression.
20197 saturation
20199 Set the saturation expression.
20200 gamma
20202 Set the gamma expression.
20203 gamma_r
20205 Set the gamma_r expression.
20206 gamma_g
20208 Set gamma_g expression.
20209 gamma_b
20211 Set gamma_b expression.
20212 gamma_weight
20214 Set gamma_weight expression.
20215 The command accepts the same syntax of the corresponding option.
20217 If the specified expression is not valid, it is kept at its current
20219 value.
20220 erosion
20222 Apply erosion effect to the video.
20223 This filter replaces the pixel by the local(3x3) minimum.
20225 It accepts the following options:
20227 threshold0
20229 threshold1
20230 threshold2
20231 threshold3
20232 Limit the maximum change for each plane, default is 65535. If 0,
20233 plane will remain unchanged.
20234 coordinates
20236 Flag which specifies the pixel to refer to. Default is 255 i.e. all
20237 eight pixels are used.
20238 Flags to local 3x3 coordinates maps like this:
20240 1 2 3
20242 4 5
20243 6 7 8
20244 Commands
20246 This filter supports the all above options as commands.
20248 estdif
20250 Deinterlace the input video ("estdif" stands for "Edge Slope Tracing
20251 Deinterlacing Filter").
20252 Spatial only filter that uses edge slope tracing algorithm to
20254 interpolate missing lines. It accepts the following parameters:
20255 mode
20257 The interlacing mode to adopt. It accepts one of the following
20258 values:
20259 frame
20261 Output one frame for each frame.
20262 field
20264 Output one frame for each field.
20265 The default value is "field".
20267 parity
20269 The picture field parity assumed for the input interlaced video. It
20270 accepts one of the following values:
20271 tff Assume the top field is first.
20273 bff Assume the bottom field is first.
20275 auto
20277 Enable automatic detection of field parity.
20278 The default value is "auto". If the interlacing is unknown or the
20280 decoder does not export this information, top field first will be
20281 assumed.
20282 deint
20284 Specify which frames to deinterlace. Accepts one of the following
20285 values:
20286 all Deinterlace all frames.
20288 interlaced
20290 Only deinterlace frames marked as interlaced.
20291 The default value is "all".
20293 rslope
20295 Specify the search radius for edge slope tracing. Default value is
20296 1. Allowed range is from 1 to 15.
20297 redge
20299 Specify the search radius for best edge matching. Default value is
20300 2. Allowed range is from 0 to 15.
20301 ecost
20303 Specify the edge cost for edge matching. Default value is 1.0.
20304 Allowed range is from 0 to 9.
20305 mcost
20307 Specify the middle cost for edge matching. Default value is 0.5.
20308 Allowed range is from 0 to 1.
20309 dcost
20311 Specify the distance cost for edge matching. Default value is 0.5.
20312 Allowed range is from 0 to 1.
20313 interp
20315 Specify the interpolation used. Default is 4-point interpolation.
20316 It accepts one of the following values:
20317 2p Two-point interpolation.
20319 4p Four-point interpolation.
20321 6p Six-point interpolation.
20323 Commands
20325 This filter supports same commands as options.
20327 exposure
20329 Adjust exposure of the video stream.
20330 The filter accepts the following options:
20332 exposure
20334 Set the exposure correction in EV. Allowed range is from -3.0 to
20335 3.0 EV Default value is 0 EV.
20336 black
20338 Set the black level correction. Allowed range is from -1.0 to 1.0.
20339 Default value is 0.
20340 Commands
20342 This filter supports same commands as options.
20344 extractplanes
20346 Extract color channel components from input video stream into separate
20347 grayscale video streams.
20348 The filter accepts the following option:
20350 planes
20352 Set plane(s) to extract.
20353 Available values for planes are:
20355 y
20357 u
20358 v
20359 a
20360 r
20361 g
20362 b
20363 Choosing planes not available in the input will result in an error.
20365 That means you cannot select "r", "g", "b" planes with "y", "u",
20366 "v" planes at same time.
20367 Examples
20369 • Extract luma, u and v color channel component from input video
20371 frame into 3 grayscale outputs:
20372 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
20374 fade
20376 Apply a fade-in/out effect to the input video.
20377 It accepts the following parameters:
20379 type, t
20381 The effect type can be either "in" for a fade-in, or "out" for a
20382 fade-out effect. Default is "in".
20383 start_frame, s
20385 Specify the number of the frame to start applying the fade effect
20386 at. Default is 0.
20387 nb_frames, n
20389 The number of frames that the fade effect lasts. At the end of the
20390 fade-in effect, the output video will have the same intensity as
20391 the input video. At the end of the fade-out transition, the output
20392 video will be filled with the selected color. Default is 25.
20393 alpha
20395 If set to 1, fade only alpha channel, if one exists on the input.
20396 Default value is 0.
20397 start_time, st
20399 Specify the timestamp (in seconds) of the frame to start to apply
20400 the fade effect. If both start_frame and start_time are specified,
20401 the fade will start at whichever comes last. Default is 0.
20402 duration, d
20404 The number of seconds for which the fade effect has to last. At the
20405 end of the fade-in effect the output video will have the same
20406 intensity as the input video, at the end of the fade-out transition
20407 the output video will be filled with the selected color. If both
20408 duration and nb_frames are specified, duration is used. Default is
20409 0 (nb_frames is used by default).
20410 color, c
20412 Specify the color of the fade. Default is "black".
20413 Examples
20415 • Fade in the first 30 frames of video:
20417 fade=in:0:30
20419 The command above is equivalent to:
20421 fade=t=in:s=0:n=30
20423 • Fade out the last 45 frames of a 200-frame video:
20425 fade=out:155:45
20427 fade=type=out:start_frame=155:nb_frames=45
20428 • Fade in the first 25 frames and fade out the last 25 frames of a
20430 1000-frame video:
20431 fade=in:0:25, fade=out:975:25
20433 • Make the first 5 frames yellow, then fade in from frame 5-24:
20435 fade=in:5:20:color=yellow
20437 • Fade in alpha over first 25 frames of video:
20439 fade=in:0:25:alpha=1
20441 • Make the first 5.5 seconds black, then fade in for 0.5 seconds:
20443 fade=t=in:st=5.5:d=0.5
20445 feedback
20447 Apply feedback video filter.
20448 This filter pass cropped input frames to 2nd output. From there it can
20450 be filtered with other video filters. After filter receives frame from
20451 2nd input, that frame is combined on top of original frame from 1st
20452 input and passed to 1st output.
20453 The typical usage is filter only part of frame.
20455 The filter accepts the following options:
20457 x
20459 y Set the top left crop position.
20460 w
20462 h Set the crop size.
20463 Examples
20465 • Blur only top left rectangular part of video frame size 100x100
20467 with gblur filter.
20468 [in][blurin]feedback=x=0:y=0:w=100:h=100[out][blurout];[blurout]gblur=8[blurin]
20470 • Draw black box on top left part of video frame of size 100x100 with
20472 drawbox filter.
20473 [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]
20475 fftdnoiz
20477 Denoise frames using 3D FFT (frequency domain filtering).
20478 The filter accepts the following options:
20480 sigma
20482 Set the noise sigma constant. This sets denoising strength.
20483 Default value is 1. Allowed range is from 0 to 30. Using very high
20484 sigma with low overlap may give blocking artifacts.
20485 amount
20487 Set amount of denoising. By default all detected noise is reduced.
20488 Default value is 1. Allowed range is from 0 to 1.
20489 block
20491 Set size of block in pixels, Default is 32, can be 8 to 256.
20492 overlap
20494 Set block overlap. Default is 0.5. Allowed range is from 0.2 to
20495 0.8.
20496 method
20498 Set denoising method. Default is "wiener", can also be "hard".
20499 prev
20501 Set number of previous frames to use for denoising. By default is
20502 set to 0.
20503 next
20505 Set number of next frames to to use for denoising. By default is
20506 set to 0.
20507 planes
20509 Set planes which will be filtered, by default are all available
20510 filtered except alpha.
20511 fftfilt
20513 Apply arbitrary expressions to samples in frequency domain
20514 dc_Y
20516 Adjust the dc value (gain) of the luma plane of the image. The
20517 filter accepts an integer value in range 0 to 1000. The default
20518 value is set to 0.
20519 dc_U
20521 Adjust the dc value (gain) of the 1st chroma plane of the image.
20522 The filter accepts an integer value in range 0 to 1000. The default
20523 value is set to 0.
20524 dc_V
20526 Adjust the dc value (gain) of the 2nd chroma plane of the image.
20527 The filter accepts an integer value in range 0 to 1000. The default
20528 value is set to 0.
20529 weight_Y
20531 Set the frequency domain weight expression for the luma plane.
20532 weight_U
20534 Set the frequency domain weight expression for the 1st chroma
20535 plane.
20536 weight_V
20538 Set the frequency domain weight expression for the 2nd chroma
20539 plane.
20540 eval
20542 Set when the expressions are evaluated.
20543 It accepts the following values:
20545 init
20547 Only evaluate expressions once during the filter
20548 initialization.
20549 frame
20551 Evaluate expressions for each incoming frame.
20552 Default value is init.
20554 The filter accepts the following variables:
20556 X
20558 Y The coordinates of the current sample.
20559 W
20561 H The width and height of the image.
20562 N The number of input frame, starting from 0.
20564 WS
20566 HS The size of FFT array for horizontal and vertical processing.
20567 Examples
20569 • High-pass:
20571 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
20573 • Low-pass:
20575 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
20577 • Sharpen:
20579 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
20581 • Blur:
20583 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
20585 field
20587 Extract a single field from an interlaced image using stride arithmetic
20588 to avoid wasting CPU time. The output frames are marked as non-
20589 interlaced.
20590 The filter accepts the following options:
20592 type
20594 Specify whether to extract the top (if the value is 0 or "top") or
20595 the bottom field (if the value is 1 or "bottom").
20596 fieldhint
20598 Create new frames by copying the top and bottom fields from surrounding
20599 frames supplied as numbers by the hint file.
20600 hint
20602 Set file containing hints: absolute/relative frame numbers.
20603 There must be one line for each frame in a clip. Each line must
20605 contain two numbers separated by the comma, optionally followed by
20606 "-" or "+". Numbers supplied on each line of file can not be out
20607 of [N-1,N+1] where N is current frame number for "absolute" mode or
20608 out of [-1, 1] range for "relative" mode. First number tells from
20609 which frame to pick up top field and second number tells from which
20610 frame to pick up bottom field.
20611 If optionally followed by "+" output frame will be marked as
20613 interlaced, else if followed by "-" output frame will be marked as
20614 progressive, else it will be marked same as input frame. If
20615 optionally followed by "t" output frame will use only top field, or
20616 in case of "b" it will use only bottom field. If line starts with
20617 "#" or ";" that line is skipped.
20618 mode
20620 Can be item "absolute" or "relative" or "pattern". Default is
20621 "absolute". The "pattern" mode is same as "relative" mode, except
20622 at last entry of file if there are more frames to process than
20623 "hint" file is seek back to start.
20624 Example of first several lines of "hint" file for "relative" mode:
20626 0,0 - # first frame
20628 1,0 - # second frame, use third's frame top field and second's frame bottom field
20629 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
20630 1,0 -
20631 0,0 -
20632 0,0 -
20633 1,0 -
20634 1,0 -
20635 1,0 -
20636 0,0 -
20637 0,0 -
20638 1,0 -
20639 1,0 -
20640 1,0 -
20641 0,0 -
20642 fieldmatch
20644 Field matching filter for inverse telecine. It is meant to reconstruct
20645 the progressive frames from a telecined stream. The filter does not
20646 drop duplicated frames, so to achieve a complete inverse telecine
20647 "fieldmatch" needs to be followed by a decimation filter such as
20648 decimate in the filtergraph.
20649 The separation of the field matching and the decimation is notably
20651 motivated by the possibility of inserting a de-interlacing filter
20652 fallback between the two. If the source has mixed telecined and real
20653 interlaced content, "fieldmatch" will not be able to match fields for
20654 the interlaced parts. But these remaining combed frames will be marked
20655 as interlaced, and thus can be de-interlaced by a later filter such as
20656 yadif before decimation.
20657 In addition to the various configuration options, "fieldmatch" can take
20659 an optional second stream, activated through the ppsrc option. If
20660 enabled, the frames reconstruction will be based on the fields and
20661 frames from this second stream. This allows the first input to be pre-
20662 processed in order to help the various algorithms of the filter, while
20663 keeping the output lossless (assuming the fields are matched properly).
20664 Typically, a field-aware denoiser, or brightness/contrast adjustments
20665 can help.
20666 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth
20668 project) and VIVTC/VFM (VapourSynth project). The later is a light
20669 clone of TFM from which "fieldmatch" is based on. While the semantic
20670 and usage are very close, some behaviour and options names can differ.
20671 The decimate filter currently only works for constant frame rate input.
20673 If your input has mixed telecined (30fps) and progressive content with
20674 a lower framerate like 24fps use the following filterchain to produce
20675 the necessary cfr stream:
20676 "dejudder,fps=30000/1001,fieldmatch,decimate".
20677 The filter accepts the following options:
20679 order
20681 Specify the assumed field order of the input stream. Available
20682 values are:
20683 auto
20685 Auto detect parity (use FFmpeg's internal parity value).
20686 bff Assume bottom field first.
20688 tff Assume top field first.
20690 Note that it is sometimes recommended not to trust the parity
20692 announced by the stream.
20693 Default value is auto.
20695 mode
20697 Set the matching mode or strategy to use. pc mode is the safest in
20698 the sense that it won't risk creating jerkiness due to duplicate
20699 frames when possible, but if there are bad edits or blended fields
20700 it will end up outputting combed frames when a good match might
20701 actually exist. On the other hand, pcn_ub mode is the most risky in
20702 terms of creating jerkiness, but will almost always find a good
20703 frame if there is one. The other values are all somewhere in
20704 between pc and pcn_ub in terms of risking jerkiness and creating
20705 duplicate frames versus finding good matches in sections with bad
20706 edits, orphaned fields, blended fields, etc.
20707 More details about p/c/n/u/b are available in p/c/n/u/b meaning
20709 section.
20710 Available values are:
20712 pc 2-way matching (p/c)
20714 pc_n
20716 2-way matching, and trying 3rd match if still combed (p/c + n)
20717 pc_u
20719 2-way matching, and trying 3rd match (same order) if still
20720 combed (p/c + u)
20721 pc_n_ub
20723 2-way matching, trying 3rd match if still combed, and trying
20724 4th/5th matches if still combed (p/c + n + u/b)
20725 pcn 3-way matching (p/c/n)
20727 pcn_ub
20729 3-way matching, and trying 4th/5th matches if all 3 of the
20730 original matches are detected as combed (p/c/n + u/b)
20731 The parenthesis at the end indicate the matches that would be used
20733 for that mode assuming order=tff (and field on auto or top).
20734 In terms of speed pc mode is by far the fastest and pcn_ub is the
20736 slowest.
20737 Default value is pc_n.
20739 ppsrc
20741 Mark the main input stream as a pre-processed input, and enable the
20742 secondary input stream as the clean source to pick the fields from.
20743 See the filter introduction for more details. It is similar to the
20744 clip2 feature from VFM/TFM.
20745 Default value is 0 (disabled).
20747 field
20749 Set the field to match from. It is recommended to set this to the
20750 same value as order unless you experience matching failures with
20751 that setting. In certain circumstances changing the field that is
20752 used to match from can have a large impact on matching performance.
20753 Available values are:
20754 auto
20756 Automatic (same value as order).
20757 bottom
20759 Match from the bottom field.
20760 top Match from the top field.
20762 Default value is auto.
20764 mchroma
20766 Set whether or not chroma is included during the match comparisons.
20767 In most cases it is recommended to leave this enabled. You should
20768 set this to 0 only if your clip has bad chroma problems such as
20769 heavy rainbowing or other artifacts. Setting this to 0 could also
20770 be used to speed things up at the cost of some accuracy.
20771 Default value is 1.
20773 y0
20775 y1 These define an exclusion band which excludes the lines between y0
20776 and y1 from being included in the field matching decision. An
20777 exclusion band can be used to ignore subtitles, a logo, or other
20778 things that may interfere with the matching. y0 sets the starting
20779 scan line and y1 sets the ending line; all lines in between y0 and
20780 y1 (including y0 and y1) will be ignored. Setting y0 and y1 to the
20781 same value will disable the feature. y0 and y1 defaults to 0.
20782 scthresh
20784 Set the scene change detection threshold as a percentage of maximum
20785 change on the luma plane. Good values are in the "[8.0, 14.0]"
20786 range. Scene change detection is only relevant in case
20787 combmatch=sc. The range for scthresh is "[0.0, 100.0]".
20788 Default value is 12.0.
20790 combmatch
20792 When combatch is not none, "fieldmatch" will take into account the
20793 combed scores of matches when deciding what match to use as the
20794 final match. Available values are:
20795 none
20797 No final matching based on combed scores.
20798 sc Combed scores are only used when a scene change is detected.
20800 full
20802 Use combed scores all the time.
20803 Default is sc.
20805 combdbg
20807 Force "fieldmatch" to calculate the combed metrics for certain
20808 matches and print them. This setting is known as micout in TFM/VFM
20809 vocabulary. Available values are:
20810 none
20812 No forced calculation.
20813 pcn Force p/c/n calculations.
20815 pcnub
20817 Force p/c/n/u/b calculations.
20818 Default value is none.
20820 cthresh
20822 This is the area combing threshold used for combed frame detection.
20823 This essentially controls how "strong" or "visible" combing must be
20824 to be detected. Larger values mean combing must be more visible
20825 and smaller values mean combing can be less visible or strong and
20826 still be detected. Valid settings are from "-1" (every pixel will
20827 be detected as combed) to 255 (no pixel will be detected as
20828 combed). This is basically a pixel difference value. A good range
20829 is "[8, 12]".
20830 Default value is 9.
20832 chroma
20834 Sets whether or not chroma is considered in the combed frame
20835 decision. Only disable this if your source has chroma problems
20836 (rainbowing, etc.) that are causing problems for the combed frame
20837 detection with chroma enabled. Actually, using chroma=0 is usually
20838 more reliable, except for the case where there is chroma only
20839 combing in the source.
20840 Default value is 0.
20842 blockx
20844 blocky
20845 Respectively set the x-axis and y-axis size of the window used
20846 during combed frame detection. This has to do with the size of the
20847 area in which combpel pixels are required to be detected as combed
20848 for a frame to be declared combed. See the combpel parameter
20849 description for more info. Possible values are any number that is
20850 a power of 2 starting at 4 and going up to 512.
20851 Default value is 16.
20853 combpel
20855 The number of combed pixels inside any of the blocky by blockx size
20856 blocks on the frame for the frame to be detected as combed. While
20857 cthresh controls how "visible" the combing must be, this setting
20858 controls "how much" combing there must be in any localized area (a
20859 window defined by the blockx and blocky settings) on the frame.
20860 Minimum value is 0 and maximum is "blocky x blockx" (at which point
20861 no frames will ever be detected as combed). This setting is known
20862 as MI in TFM/VFM vocabulary.
20863 Default value is 80.
20865 p/c/n/u/b meaning
20867 p/c/n
20869 We assume the following telecined stream:
20871 Top fields: 1 2 2 3 4
20873 Bottom fields: 1 2 3 4 4
20874 The numbers correspond to the progressive frame the fields relate to.
20876 Here, the first two frames are progressive, the 3rd and 4th are combed,
20877 and so on.
20878 When "fieldmatch" is configured to run a matching from bottom
20880 (field=bottom) this is how this input stream get transformed:
20881 Input stream:
20883 T 1 2 2 3 4
20884 B 1 2 3 4 4 <-- matching reference
20885 Matches: c c n n c
20887 Output stream:
20889 T 1 2 3 4 4
20890 B 1 2 3 4 4
20891 As a result of the field matching, we can see that some frames get
20893 duplicated. To perform a complete inverse telecine, you need to rely
20894 on a decimation filter after this operation. See for instance the
20895 decimate filter.
20896 The same operation now matching from top fields (field=top) looks like
20898 this:
20899 Input stream:
20901 T 1 2 2 3 4 <-- matching reference
20902 B 1 2 3 4 4
20903 Matches: c c p p c
20905 Output stream:
20907 T 1 2 2 3 4
20908 B 1 2 2 3 4
20909 In these examples, we can see what p, c and n mean; basically, they
20911 refer to the frame and field of the opposite parity:
20912 *<p matches the field of the opposite parity in the previous frame>
20914 *<c matches the field of the opposite parity in the current frame>
20915 *<n matches the field of the opposite parity in the next frame>
20916 u/b
20918 The u and b matching are a bit special in the sense that they match
20920 from the opposite parity flag. In the following examples, we assume
20921 that we are currently matching the 2nd frame (Top:2, bottom:2).
20922 According to the match, a 'x' is placed above and below each matched
20923 fields.
20924 With bottom matching (field=bottom):
20926 Match: c p n b u
20928 x x x x x
20930 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
20931 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
20932 x x x x x
20933 Output frames:
20935 2 1 2 2 2
20936 2 2 2 1 3
20937 With top matching (field=top):
20939 Match: c p n b u
20941 x x x x x
20943 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
20944 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
20945 x x x x x
20946 Output frames:
20948 2 2 2 1 2
20949 2 1 3 2 2
20950 Examples
20952 Simple IVTC of a top field first telecined stream:
20954 fieldmatch=order=tff:combmatch=none, decimate
20956 Advanced IVTC, with fallback on yadif for still combed frames:
20958 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
20960 fieldorder
20962 Transform the field order of the input video.
20963 It accepts the following parameters:
20965 order
20967 The output field order. Valid values are tff for top field first or
20968 bff for bottom field first.
20969 The default value is tff.
20971 The transformation is done by shifting the picture content up or down
20973 by one line, and filling the remaining line with appropriate picture
20974 content. This method is consistent with most broadcast field order
20975 converters.
20976 If the input video is not flagged as being interlaced, or it is already
20978 flagged as being of the required output field order, then this filter
20979 does not alter the incoming video.
20980 It is very useful when converting to or from PAL DV material, which is
20982 bottom field first.
20983 For example:
20985 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
20987 fifo, afifo
20989 Buffer input images and send them when they are requested.
20990 It is mainly useful when auto-inserted by the libavfilter framework.
20992 It does not take parameters.
20994 fillborders
20996 Fill borders of the input video, without changing video stream
20997 dimensions. Sometimes video can have garbage at the four edges and you
20998 may not want to crop video input to keep size multiple of some number.
20999 This filter accepts the following options:
21001 left
21003 Number of pixels to fill from left border.
21004 right
21006 Number of pixels to fill from right border.
21007 top Number of pixels to fill from top border.
21009 bottom
21011 Number of pixels to fill from bottom border.
21012 mode
21014 Set fill mode.
21015 It accepts the following values:
21017 smear
21019 fill pixels using outermost pixels
21020 mirror
21022 fill pixels using mirroring (half sample symmetric)
21023 fixed
21025 fill pixels with constant value
21026 reflect
21028 fill pixels using reflecting (whole sample symmetric)
21029 wrap
21031 fill pixels using wrapping
21032 fade
21034 fade pixels to constant value
21035 margins
21037 fill pixels at top and bottom with weighted averages pixels
21038 near borders
21039 Default is smear.
21041 color
21043 Set color for pixels in fixed or fade mode. Default is black.
21044 Commands
21046 This filter supports same commands as options. The command accepts the
21048 same syntax of the corresponding option.
21049 If the specified expression is not valid, it is kept at its current
21051 value.
21052 find_rect
21054 Find a rectangular object
21055 It accepts the following options:
21057 object
21059 Filepath of the object image, needs to be in gray8.
21060 threshold
21062 Detection threshold, default is 0.5.
21063 mipmaps
21065 Number of mipmaps, default is 3.
21066 xmin, ymin, xmax, ymax
21068 Specifies the rectangle in which to search.
21069 discard
21071 Discard frames where object is not detected. Default is disabled.
21072 Examples
21074 • Cover a rectangular object by the supplied image of a given video
21076 using ffmpeg:
21077 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
21079 floodfill
21081 Flood area with values of same pixel components with another values.
21082 It accepts the following options:
21084 x Set pixel x coordinate.
21086 y Set pixel y coordinate.
21088 s0 Set source #0 component value.
21090 s1 Set source #1 component value.
21092 s2 Set source #2 component value.
21094 s3 Set source #3 component value.
21096 d0 Set destination #0 component value.
21098 d1 Set destination #1 component value.
21100 d2 Set destination #2 component value.
21102 d3 Set destination #3 component value.
21104 format
21106 Convert the input video to one of the specified pixel formats.
21107 Libavfilter will try to pick one that is suitable as input to the next
21108 filter.
21109 It accepts the following parameters:
21111 pix_fmts
21113 A '|'-separated list of pixel format names, such as
21114 "pix_fmts=yuv420p|monow|rgb24".
21115 Examples
21117 • Convert the input video to the yuv420p format
21119 format=pix_fmts=yuv420p
21121 Convert the input video to any of the formats in the list
21123 format=pix_fmts=yuv420p|yuv444p|yuv410p
21125 fps
21127 Convert the video to specified constant frame rate by duplicating or
21128 dropping frames as necessary.
21129 It accepts the following parameters:
21131 fps The desired output frame rate. It accepts expressions containing
21133 the following constants:
21134 source_fps
21136 The input's frame rate
21137 ntsc
21139 NTSC frame rate of "30000/1001"
21140 pal PAL frame rate of 25.0
21142 film
21144 Film frame rate of 24.0
21145 ntsc_film
21147 NTSC-film frame rate of "24000/1001"
21148 The default is 25.
21150 start_time
21152 Assume the first PTS should be the given value, in seconds. This
21153 allows for padding/trimming at the start of stream. By default, no
21154 assumption is made about the first frame's expected PTS, so no
21155 padding or trimming is done. For example, this could be set to 0
21156 to pad the beginning with duplicates of the first frame if a video
21157 stream starts after the audio stream or to trim any frames with a
21158 negative PTS.
21159 round
21161 Timestamp (PTS) rounding method.
21162 Possible values are:
21164 zero
21166 round towards 0
21167 inf round away from 0
21169 down
21171 round towards -infinity
21172 up round towards +infinity
21174 near
21176 round to nearest
21177 The default is "near".
21179 eof_action
21181 Action performed when reading the last frame.
21182 Possible values are:
21184 round
21186 Use same timestamp rounding method as used for other frames.
21187 pass
21189 Pass through last frame if input duration has not been reached
21190 yet.
21191 The default is "round".
21193 Alternatively, the options can be specified as a flat string:
21195 fps[:start_time[:round]].
21196 See also the setpts filter.
21198 Examples
21200 • A typical usage in order to set the fps to 25:
21202 fps=fps=25
21204 • Sets the fps to 24, using abbreviation and rounding method to round
21206 to nearest:
21207 fps=fps=film:round=near
21209 framepack
21211 Pack two different video streams into a stereoscopic video, setting
21212 proper metadata on supported codecs. The two views should have the same
21213 size and framerate and processing will stop when the shorter video
21214 ends. Please note that you may conveniently adjust view properties with
21215 the scale and fps filters.
21216 It accepts the following parameters:
21218 format
21220 The desired packing format. Supported values are:
21221 sbs The views are next to each other (default).
21223 tab The views are on top of each other.
21225 lines
21227 The views are packed by line.
21228 columns
21230 The views are packed by column.
21231 frameseq
21233 The views are temporally interleaved.
21234 Some examples:
21236 # Convert left and right views into a frame-sequential video
21238 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
21239 # Convert views into a side-by-side video with the same output resolution as the input
21241 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
21242 framerate
21244 Change the frame rate by interpolating new video output frames from the
21245 source frames.
21246 This filter is not designed to function correctly with interlaced
21248 media. If you wish to change the frame rate of interlaced media then
21249 you are required to deinterlace before this filter and re-interlace
21250 after this filter.
21251 A description of the accepted options follows.
21253 fps Specify the output frames per second. This option can also be
21255 specified as a value alone. The default is 50.
21256 interp_start
21258 Specify the start of a range where the output frame will be created
21259 as a linear interpolation of two frames. The range is [0-255], the
21260 default is 15.
21261 interp_end
21263 Specify the end of a range where the output frame will be created
21264 as a linear interpolation of two frames. The range is [0-255], the
21265 default is 240.
21266 scene
21268 Specify the level at which a scene change is detected as a value
21269 between 0 and 100 to indicate a new scene; a low value reflects a
21270 low probability for the current frame to introduce a new scene,
21271 while a higher value means the current frame is more likely to be
21272 one. The default is 8.2.
21273 flags
21275 Specify flags influencing the filter process.
21276 Available value for flags is:
21278 scene_change_detect, scd
21280 Enable scene change detection using the value of the option
21281 scene. This flag is enabled by default.
21282 framestep
21284 Select one frame every N-th frame.
21285 This filter accepts the following option:
21287 step
21289 Select frame after every "step" frames. Allowed values are
21290 positive integers higher than 0. Default value is 1.
21291 freezedetect
21293 Detect frozen video.
21294 This filter logs a message and sets frame metadata when it detects that
21296 the input video has no significant change in content during a specified
21297 duration. Video freeze detection calculates the mean average absolute
21298 difference of all the components of video frames and compares it to a
21299 noise floor.
21300 The printed times and duration are expressed in seconds. The
21302 "lavfi.freezedetect.freeze_start" metadata key is set on the first
21303 frame whose timestamp equals or exceeds the detection duration and it
21304 contains the timestamp of the first frame of the freeze. The
21305 "lavfi.freezedetect.freeze_duration" and
21306 "lavfi.freezedetect.freeze_end" metadata keys are set on the first
21307 frame after the freeze.
21308 The filter accepts the following options:
21310 noise, n
21312 Set noise tolerance. Can be specified in dB (in case "dB" is
21313 appended to the specified value) or as a difference ratio between 0
21314 and 1. Default is -60dB, or 0.001.
21315 duration, d
21317 Set freeze duration until notification (default is 2 seconds).
21318 freezeframes
21320 Freeze video frames.
21321 This filter freezes video frames using frame from 2nd input.
21323 The filter accepts the following options:
21325 first
21327 Set number of first frame from which to start freeze.
21328 last
21330 Set number of last frame from which to end freeze.
21331 replace
21333 Set number of frame from 2nd input which will be used instead of
21334 replaced frames.
21335 frei0r
21337 Apply a frei0r effect to the input video.
21338 To enable the compilation of this filter, you need to install the
21340 frei0r header and configure FFmpeg with "--enable-frei0r".
21341 It accepts the following parameters:
21343 filter_name
21345 The name of the frei0r effect to load. If the environment variable
21346 FREI0R_PATH is defined, the frei0r effect is searched for in each
21347 of the directories specified by the colon-separated list in
21348 FREI0R_PATH. Otherwise, the standard frei0r paths are searched, in
21349 this order: HOME/.frei0r-1/lib/, /usr/local/lib/frei0r-1/,
21350 /usr/lib/frei0r-1/.
21351 filter_params
21353 A '|'-separated list of parameters to pass to the frei0r effect.
21354 A frei0r effect parameter can be a boolean (its value is either "y" or
21356 "n"), a double, a color (specified as R/G/B, where R, G, and B are
21357 floating point numbers between 0.0 and 1.0, inclusive) or a color
21358 description as specified in the "Color" section in the ffmpeg-utils
21359 manual, a position (specified as X/Y, where X and Y are floating point
21360 numbers) and/or a string.
21361 The number and types of parameters depend on the loaded effect. If an
21363 effect parameter is not specified, the default value is set.
21364 Examples
21366 • Apply the distort0r effect, setting the first two double
21368 parameters:
21369 frei0r=filter_name=distort0r:filter_params=0.5|0.01
21371 • Apply the colordistance effect, taking a color as the first
21373 parameter:
21374 frei0r=colordistance:0.2/0.3/0.4
21376 frei0r=colordistance:violet
21377 frei0r=colordistance:0x112233
21378 • Apply the perspective effect, specifying the top left and top right
21380 image positions:
21381 frei0r=perspective:0.2/0.2|0.8/0.2
21383 For more information, see <http://frei0r.dyne.org>
21385 Commands
21387 This filter supports the filter_params option as commands.
21389 fspp
21391 Apply fast and simple postprocessing. It is a faster version of spp.
21392 It splits (I)DCT into horizontal/vertical passes. Unlike the simple
21394 post- processing filter, one of them is performed once per block, not
21395 per pixel. This allows for much higher speed.
21396 The filter accepts the following options:
21398 quality
21400 Set quality. This option defines the number of levels for
21401 averaging. It accepts an integer in the range 4-5. Default value is
21402 4.
21403 qp Force a constant quantization parameter. It accepts an integer in
21405 range 0-63. If not set, the filter will use the QP from the video
21406 stream (if available).
21407 strength
21409 Set filter strength. It accepts an integer in range -15 to 32.
21410 Lower values mean more details but also more artifacts, while
21411 higher values make the image smoother but also blurrier. Default
21412 value is 0 X PSNR optimal.
21413 use_bframe_qp
21415 Enable the use of the QP from the B-Frames if set to 1. Using this
21416 option may cause flicker since the B-Frames have often larger QP.
21417 Default is 0 (not enabled).
21418 gblur
21420 Apply Gaussian blur filter.
21421 The filter accepts the following options:
21423 sigma
21425 Set horizontal sigma, standard deviation of Gaussian blur. Default
21426 is 0.5.
21427 steps
21429 Set number of steps for Gaussian approximation. Default is 1.
21430 planes
21432 Set which planes to filter. By default all planes are filtered.
21433 sigmaV
21435 Set vertical sigma, if negative it will be same as "sigma".
21436 Default is "-1".
21437 Commands
21439 This filter supports same commands as options. The command accepts the
21441 same syntax of the corresponding option.
21442 If the specified expression is not valid, it is kept at its current
21444 value.
21445 geq
21447 Apply generic equation to each pixel.
21448 The filter accepts the following options:
21450 lum_expr, lum
21452 Set the luminance expression.
21453 cb_expr, cb
21455 Set the chrominance blue expression.
21456 cr_expr, cr
21458 Set the chrominance red expression.
21459 alpha_expr, a
21461 Set the alpha expression.
21462 red_expr, r
21464 Set the red expression.
21465 green_expr, g
21467 Set the green expression.
21468 blue_expr, b
21470 Set the blue expression.
21471 The colorspace is selected according to the specified options. If one
21473 of the lum_expr, cb_expr, or cr_expr options is specified, the filter
21474 will automatically select a YCbCr colorspace. If one of the red_expr,
21475 green_expr, or blue_expr options is specified, it will select an RGB
21476 colorspace.
21477 If one of the chrominance expression is not defined, it falls back on
21479 the other one. If no alpha expression is specified it will evaluate to
21480 opaque value. If none of chrominance expressions are specified, they
21481 will evaluate to the luminance expression.
21482 The expressions can use the following variables and functions:
21484 N The sequential number of the filtered frame, starting from 0.
21486 X
21488 Y The coordinates of the current sample.
21489 W
21491 H The width and height of the image.
21492 SW
21494 SH Width and height scale depending on the currently filtered plane.
21495 It is the ratio between the corresponding luma plane number of
21496 pixels and the current plane ones. E.g. for YUV4:2:0 the values are
21497 "1,1" for the luma plane, and "0.5,0.5" for chroma planes.
21498 T Time of the current frame, expressed in seconds.
21500 p(x, y)
21502 Return the value of the pixel at location (x,y) of the current
21503 plane.
21504 lum(x, y)
21506 Return the value of the pixel at location (x,y) of the luminance
21507 plane.
21508 cb(x, y)
21510 Return the value of the pixel at location (x,y) of the blue-
21511 difference chroma plane. Return 0 if there is no such plane.
21512 cr(x, y)
21514 Return the value of the pixel at location (x,y) of the red-
21515 difference chroma plane. Return 0 if there is no such plane.
21516 r(x, y)
21518 g(x, y)
21519 b(x, y)
21520 Return the value of the pixel at location (x,y) of the
21521 red/green/blue component. Return 0 if there is no such component.
21522 alpha(x, y)
21524 Return the value of the pixel at location (x,y) of the alpha plane.
21525 Return 0 if there is no such plane.
21526 psum(x,y), lumsum(x, y), cbsum(x,y), crsum(x,y), rsum(x,y), gsum(x,y),
21528 bsum(x,y), alphasum(x,y)
21529 Sum of sample values in the rectangle from (0,0) to (x,y), this
21530 allows obtaining sums of samples within a rectangle. See the
21531 functions without the sum postfix.
21532 interpolation
21534 Set one of interpolation methods:
21535 nearest, n
21537 bilinear, b
21538 Default is bilinear.
21540 For functions, if x and y are outside the area, the value will be
21542 automatically clipped to the closer edge.
21543 Please note that this filter can use multiple threads in which case
21545 each slice will have its own expression state. If you want to use only
21546 a single expression state because your expressions depend on previous
21547 state then you should limit the number of filter threads to 1.
21548 Examples
21550 • Flip the image horizontally:
21552 geq=p(W-X\,Y)
21554 • Generate a bidimensional sine wave, with angle "PI/3" and a
21556 wavelength of 100 pixels:
21557 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
21559 • Generate a fancy enigmatic moving light:
21561 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
21563 • Generate a quick emboss effect:
21565 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
21567 • Modify RGB components depending on pixel position:
21569 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
21571 • Create a radial gradient that is the same size as the input (also
21573 see the vignette filter):
21574 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
21576 gradfun
21578 Fix the banding artifacts that are sometimes introduced into nearly
21579 flat regions by truncation to 8-bit color depth. Interpolate the
21580 gradients that should go where the bands are, and dither them.
21581 It is designed for playback only. Do not use it prior to lossy
21583 compression, because compression tends to lose the dither and bring
21584 back the bands.
21585 It accepts the following parameters:
21587 strength
21589 The maximum amount by which the filter will change any one pixel.
21590 This is also the threshold for detecting nearly flat regions.
21591 Acceptable values range from .51 to 64; the default value is 1.2.
21592 Out-of-range values will be clipped to the valid range.
21593 radius
21595 The neighborhood to fit the gradient to. A larger radius makes for
21596 smoother gradients, but also prevents the filter from modifying the
21597 pixels near detailed regions. Acceptable values are 8-32; the
21598 default value is 16. Out-of-range values will be clipped to the
21599 valid range.
21600 Alternatively, the options can be specified as a flat string:
21602 strength[:radius]
21603 Examples
21605 • Apply the filter with a 3.5 strength and radius of 8:
21607 gradfun=3.5:8
21609 • Specify radius, omitting the strength (which will fall-back to the
21611 default value):
21612 gradfun=radius=8
21614 graphmonitor
21616 Show various filtergraph stats.
21617 With this filter one can debug complete filtergraph. Especially issues
21619 with links filling with queued frames.
21620 The filter accepts the following options:
21622 size, s
21624 Set video output size. Default is hd720.
21625 opacity, o
21627 Set video opacity. Default is 0.9. Allowed range is from 0 to 1.
21628 mode, m
21630 Set output mode, can be fulll or compact. In compact mode only
21631 filters with some queued frames have displayed stats.
21632 flags, f
21634 Set flags which enable which stats are shown in video.
21635 Available values for flags are:
21637 queue
21639 Display number of queued frames in each link.
21640 frame_count_in
21642 Display number of frames taken from filter.
21643 frame_count_out
21645 Display number of frames given out from filter.
21646 frame_count_delta
21648 Display delta number of frames between above two values.
21649 pts Display current filtered frame pts.
21651 pts_delta
21653 Display pts delta between current and previous frame.
21654 time
21656 Display current filtered frame time.
21657 time_delta
21659 Display time delta between current and previous frame.
21660 timebase
21662 Display time base for filter link.
21663 format
21665 Display used format for filter link.
21666 size
21668 Display video size or number of audio channels in case of audio
21669 used by filter link.
21670 rate
21672 Display video frame rate or sample rate in case of audio used
21673 by filter link.
21674 eof Display link output status.
21676 sample_count_in
21678 Display number of samples taken from filter.
21679 sample_count_out
21681 Display number of samples given out from filter.
21682 sample_count_delta
21684 Display delta number of samples between above two values.
21685 rate, r
21687 Set upper limit for video rate of output stream, Default value is
21688 25. This guarantee that output video frame rate will not be higher
21689 than this value.
21690 grayworld
21692 A color constancy filter that applies color correction based on the
21693 grayworld assumption
21694 See:
21696 <https://www.researchgate.net/publication/275213614_A_New_Color_Correction_Method_for_Underwater_Imaging>
21697 The algorithm uses linear light, so input data should be linearized
21699 beforehand (and possibly correctly tagged).
21700 ffmpeg -i INPUT -vf zscale=transfer=linear,grayworld,zscale=transfer=bt709,format=yuv420p OUTPUT
21702 greyedge
21704 A color constancy variation filter which estimates scene illumination
21705 via grey edge algorithm and corrects the scene colors accordingly.
21706 See: <https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf>
21708 The filter accepts the following options:
21710 difford
21712 The order of differentiation to be applied on the scene. Must be
21713 chosen in the range [0,2] and default value is 1.
21714 minknorm
21716 The Minkowski parameter to be used for calculating the Minkowski
21717 distance. Must be chosen in the range [0,20] and default value is
21718 1. Set to 0 for getting max value instead of calculating Minkowski
21719 distance.
21720 sigma
21722 The standard deviation of Gaussian blur to be applied on the scene.
21723 Must be chosen in the range [0,1024.0] and default value = 1.
21724 floor( sigma * break_off_sigma(3) ) can't be equal to 0 if difford
21725 is greater than 0.
21726 Examples
21728 • Grey Edge:
21730 greyedge=difford=1:minknorm=5:sigma=2
21732 • Max Edge:
21734 greyedge=difford=1:minknorm=0:sigma=2
21736 guided
21738 Apply guided filter for edge-preserving smoothing, dehazing and so on.
21739 The filter accepts the following options:
21741 radius
21743 Set the box radius in pixels. Allowed range is 1 to 20. Default is
21744 3.
21745 eps Set regularization parameter (with square). Allowed range is 0 to
21747 1. Default is 0.01.
21748 mode
21750 Set filter mode. Can be "basic" or "fast". Default is "basic".
21751 sub Set subsampling ratio for "fast" mode. Range is 2 to 64. Default
21753 is 4. No subsampling occurs in "basic" mode.
21754 guidance
21756 Set guidance mode. Can be "off" or "on". Default is "off". If
21757 "off", single input is required. If "on", two inputs of the same
21758 resolution and pixel format are required. The second input serves
21759 as the guidance.
21760 planes
21762 Set planes to filter. Default is first only.
21763 Commands
21765 This filter supports the all above options as commands.
21767 Examples
21769 • Edge-preserving smoothing with guided filter:
21771 ffmpeg -i in.png -vf guided out.png
21773 • Dehazing, structure-transferring filtering, detail enhancement with
21775 guided filter. For the generation of guidance image, refer to
21776 paper "Guided Image Filtering". See:
21777 <http://kaiminghe.com/publications/pami12guidedfilter.pdf>.
21778 ffmpeg -i in.png -i guidance.png -filter_complex guided=guidance=on out.png
21780 haldclut
21782 Apply a Hald CLUT to a video stream.
21783 First input is the video stream to process, and second one is the Hald
21785 CLUT. The Hald CLUT input can be a simple picture or a complete video
21786 stream.
21787 The filter accepts the following options:
21789 clut
21791 Set which CLUT video frames will be processed from second input
21792 stream, can be first or all. Default is all.
21793 shortest
21795 Force termination when the shortest input terminates. Default is 0.
21796 repeatlast
21798 Continue applying the last CLUT after the end of the stream. A
21799 value of 0 disable the filter after the last frame of the CLUT is
21800 reached. Default is 1.
21801 "haldclut" also has the same interpolation options as lut3d (both
21803 filters share the same internals).
21804 This filter also supports the framesync options.
21806 More information about the Hald CLUT can be found on Eskil Steenberg's
21808 website (Hald CLUT author) at
21809 <http://www.quelsolaar.com/technology/clut.html>.
21810 Commands
21812 This filter supports the "interp" option as commands.
21814 Workflow examples
21816 Hald CLUT video stream
21818 Generate an identity Hald CLUT stream altered with various effects:
21820 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
21822 Note: make sure you use a lossless codec.
21824 Then use it with "haldclut" to apply it on some random stream:
21826 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
21828 The Hald CLUT will be applied to the 10 first seconds (duration of
21830 clut.nut), then the latest picture of that CLUT stream will be applied
21831 to the remaining frames of the "mandelbrot" stream.
21832 Hald CLUT with preview
21834 A Hald CLUT is supposed to be a squared image of "Level*Level*Level" by
21836 "Level*Level*Level" pixels. For a given Hald CLUT, FFmpeg will select
21837 the biggest possible square starting at the top left of the picture.
21838 The remaining padding pixels (bottom or right) will be ignored. This
21839 area can be used to add a preview of the Hald CLUT.
21840 Typically, the following generated Hald CLUT will be supported by the
21842 "haldclut" filter:
21843 ffmpeg -f lavfi -i B<haldclutsrc>=8 -vf "
21845 pad=iw+320 [padded_clut];
21846 smptebars=s=320x256, split [a][b];
21847 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
21848 [main][b] overlay=W-320" -frames:v 1 clut.png
21849 It contains the original and a preview of the effect of the CLUT: SMPTE
21851 color bars are displayed on the right-top, and below the same color
21852 bars processed by the color changes.
21853 Then, the effect of this Hald CLUT can be visualized with:
21855 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
21857 hflip
21859 Flip the input video horizontally.
21860 For example, to horizontally flip the input video with ffmpeg:
21862 ffmpeg -i in.avi -vf "hflip" out.avi
21864 histeq
21866 This filter applies a global color histogram equalization on a per-
21867 frame basis.
21868 It can be used to correct video that has a compressed range of pixel
21870 intensities. The filter redistributes the pixel intensities to
21871 equalize their distribution across the intensity range. It may be
21872 viewed as an "automatically adjusting contrast filter". This filter is
21873 useful only for correcting degraded or poorly captured source video.
21874 The filter accepts the following options:
21876 strength
21878 Determine the amount of equalization to be applied. As the
21879 strength is reduced, the distribution of pixel intensities more-
21880 and-more approaches that of the input frame. The value must be a
21881 float number in the range [0,1] and defaults to 0.200.
21882 intensity
21884 Set the maximum intensity that can generated and scale the output
21885 values appropriately. The strength should be set as desired and
21886 then the intensity can be limited if needed to avoid washing-out.
21887 The value must be a float number in the range [0,1] and defaults to
21888 0.210.
21889 antibanding
21891 Set the antibanding level. If enabled the filter will randomly vary
21892 the luminance of output pixels by a small amount to avoid banding
21893 of the histogram. Possible values are "none", "weak" or "strong".
21894 It defaults to "none".
21895 histogram
21897 Compute and draw a color distribution histogram for the input video.
21898 The computed histogram is a representation of the color component
21900 distribution in an image.
21901 Standard histogram displays the color components distribution in an
21903 image. Displays color graph for each color component. Shows
21904 distribution of the Y, U, V, A or R, G, B components, depending on
21905 input format, in the current frame. Below each graph a color component
21906 scale meter is shown.
21907 The filter accepts the following options:
21909 level_height
21911 Set height of level. Default value is 200. Allowed range is [50,
21912 2048].
21913 scale_height
21915 Set height of color scale. Default value is 12. Allowed range is
21916 [0, 40].
21917 display_mode
21919 Set display mode. It accepts the following values:
21920 stack
21922 Per color component graphs are placed below each other.
21923 parade
21925 Per color component graphs are placed side by side.
21926 overlay
21928 Presents information identical to that in the "parade", except
21929 that the graphs representing color components are superimposed
21930 directly over one another.
21931 Default is "stack".
21933 levels_mode
21935 Set mode. Can be either "linear", or "logarithmic". Default is
21936 "linear".
21937 components
21939 Set what color components to display. Default is 7.
21940 fgopacity
21942 Set foreground opacity. Default is 0.7.
21943 bgopacity
21945 Set background opacity. Default is 0.5.
21946 colors_mode
21948 Set colors mode. It accepts the following values:
21949 whiteonblack
21951 blackonwhite
21952 whiteongray
21953 blackongray
21954 coloronblack
21955 coloronwhite
21956 colorongray
21957 blackoncolor
21958 whiteoncolor
21959 grayoncolor
21960 Default is "whiteonblack".
21962 Examples
21964 • Calculate and draw histogram:
21966 ffplay -i input -vf histogram
21968 hqdn3d
21970 This is a high precision/quality 3d denoise filter. It aims to reduce
21971 image noise, producing smooth images and making still images really
21972 still. It should enhance compressibility.
21973 It accepts the following optional parameters:
21975 luma_spatial
21977 A non-negative floating point number which specifies spatial luma
21978 strength. It defaults to 4.0.
21979 chroma_spatial
21981 A non-negative floating point number which specifies spatial chroma
21982 strength. It defaults to 3.0*luma_spatial/4.0.
21983 luma_tmp
21985 A floating point number which specifies luma temporal strength. It
21986 defaults to 6.0*luma_spatial/4.0.
21987 chroma_tmp
21989 A floating point number which specifies chroma temporal strength.
21990 It defaults to luma_tmp*chroma_spatial/luma_spatial.
21991 Commands
21993 This filter supports same commands as options. The command accepts the
21995 same syntax of the corresponding option.
21996 If the specified expression is not valid, it is kept at its current
21998 value.
21999 hwdownload
22001 Download hardware frames to system memory.
22002 The input must be in hardware frames, and the output a non-hardware
22004 format. Not all formats will be supported on the output - it may be
22005 necessary to insert an additional format filter immediately following
22006 in the graph to get the output in a supported format.
22007 hwmap
22009 Map hardware frames to system memory or to another device.
22010 This filter has several different modes of operation; which one is used
22012 depends on the input and output formats:
22013 • Hardware frame input, normal frame output
22015 Map the input frames to system memory and pass them to the output.
22017 If the original hardware frame is later required (for example,
22018 after overlaying something else on part of it), the hwmap filter
22019 can be used again in the next mode to retrieve it.
22020 • Normal frame input, hardware frame output
22022 If the input is actually a software-mapped hardware frame, then
22024 unmap it - that is, return the original hardware frame.
22025 Otherwise, a device must be provided. Create new hardware surfaces
22027 on that device for the output, then map them back to the software
22028 format at the input and give those frames to the preceding filter.
22029 This will then act like the hwupload filter, but may be able to
22030 avoid an additional copy when the input is already in a compatible
22031 format.
22032 • Hardware frame input and output
22034 A device must be supplied for the output, either directly or with
22036 the derive_device option. The input and output devices must be of
22037 different types and compatible - the exact meaning of this is
22038 system-dependent, but typically it means that they must refer to
22039 the same underlying hardware context (for example, refer to the
22040 same graphics card).
22041 If the input frames were originally created on the output device,
22043 then unmap to retrieve the original frames.
22044 Otherwise, map the frames to the output device - create new
22046 hardware frames on the output corresponding to the frames on the
22047 input.
22048 The following additional parameters are accepted:
22050 mode
22052 Set the frame mapping mode. Some combination of:
22053 read
22055 The mapped frame should be readable.
22056 write
22058 The mapped frame should be writeable.
22059 overwrite
22061 The mapping will always overwrite the entire frame.
22062 This may improve performance in some cases, as the original
22064 contents of the frame need not be loaded.
22065 direct
22067 The mapping must not involve any copying.
22068 Indirect mappings to copies of frames are created in some cases
22070 where either direct mapping is not possible or it would have
22071 unexpected properties. Setting this flag ensures that the
22072 mapping is direct and will fail if that is not possible.
22073 Defaults to read+write if not specified.
22075 derive_device type
22077 Rather than using the device supplied at initialisation, instead
22078 derive a new device of type type from the device the input frames
22079 exist on.
22080 reverse
22082 In a hardware to hardware mapping, map in reverse - create frames
22083 in the sink and map them back to the source. This may be necessary
22084 in some cases where a mapping in one direction is required but only
22085 the opposite direction is supported by the devices being used.
22086 This option is dangerous - it may break the preceding filter in
22088 undefined ways if there are any additional constraints on that
22089 filter's output. Do not use it without fully understanding the
22090 implications of its use.
22091 hwupload
22093 Upload system memory frames to hardware surfaces.
22094 The device to upload to must be supplied when the filter is
22096 initialised. If using ffmpeg, select the appropriate device with the
22097 -filter_hw_device option or with the derive_device option. The input
22098 and output devices must be of different types and compatible - the
22099 exact meaning of this is system-dependent, but typically it means that
22100 they must refer to the same underlying hardware context (for example,
22101 refer to the same graphics card).
22102 The following additional parameters are accepted:
22104 derive_device type
22106 Rather than using the device supplied at initialisation, instead
22107 derive a new device of type type from the device the input frames
22108 exist on.
22109 hwupload_cuda
22111 Upload system memory frames to a CUDA device.
22112 It accepts the following optional parameters:
22114 device
22116 The number of the CUDA device to use
22117 hqx
22119 Apply a high-quality magnification filter designed for pixel art. This
22120 filter was originally created by Maxim Stepin.
22121 It accepts the following option:
22123 n Set the scaling dimension: 2 for "hq2x", 3 for "hq3x" and 4 for
22125 "hq4x". Default is 3.
22126 hstack
22128 Stack input videos horizontally.
22129 All streams must be of same pixel format and of same height.
22131 Note that this filter is faster than using overlay and pad filter to
22133 create same output.
22134 The filter accepts the following option:
22136 inputs
22138 Set number of input streams. Default is 2.
22139 shortest
22141 If set to 1, force the output to terminate when the shortest input
22142 terminates. Default value is 0.
22143 hsvhold
22145 Turns a certain HSV range into gray values.
22146 This filter measures color difference between set HSV color in options
22148 and ones measured in video stream. Depending on options, output colors
22149 can be changed to be gray or not.
22150 The filter accepts the following options:
22152 hue Set the hue value which will be used in color difference
22154 calculation. Allowed range is from -360 to 360. Default value is
22155 0.
22156 sat Set the saturation value which will be used in color difference
22158 calculation. Allowed range is from -1 to 1. Default value is 0.
22159 val Set the value which will be used in color difference calculation.
22161 Allowed range is from -1 to 1. Default value is 0.
22162 similarity
22164 Set similarity percentage with the key color. Allowed range is
22165 from 0 to 1. Default value is 0.01.
22166 0.00001 matches only the exact key color, while 1.0 matches
22168 everything.
22169 blend
22171 Blend percentage. Allowed range is from 0 to 1. Default value is
22172 0.
22173 0.0 makes pixels either fully gray, or not gray at all.
22175 Higher values result in more gray pixels, with a higher gray pixel
22177 the more similar the pixels color is to the key color.
22178 hsvkey
22180 Turns a certain HSV range into transparency.
22181 This filter measures color difference between set HSV color in options
22183 and ones measured in video stream. Depending on options, output colors
22184 can be changed to transparent by adding alpha channel.
22185 The filter accepts the following options:
22187 hue Set the hue value which will be used in color difference
22189 calculation. Allowed range is from -360 to 360. Default value is
22190 0.
22191 sat Set the saturation value which will be used in color difference
22193 calculation. Allowed range is from -1 to 1. Default value is 0.
22194 val Set the value which will be used in color difference calculation.
22196 Allowed range is from -1 to 1. Default value is 0.
22197 similarity
22199 Set similarity percentage with the key color. Allowed range is
22200 from 0 to 1. Default value is 0.01.
22201 0.00001 matches only the exact key color, while 1.0 matches
22203 everything.
22204 blend
22206 Blend percentage. Allowed range is from 0 to 1. Default value is
22207 0.
22208 0.0 makes pixels either fully transparent, or not transparent at
22210 all.
22211 Higher values result in semi-transparent pixels, with a higher
22213 transparency the more similar the pixels color is to the key color.
22214 hue
22216 Modify the hue and/or the saturation of the input.
22217 It accepts the following parameters:
22219 h Specify the hue angle as a number of degrees. It accepts an
22221 expression, and defaults to "0".
22222 s Specify the saturation in the [-10,10] range. It accepts an
22224 expression and defaults to "1".
22225 H Specify the hue angle as a number of radians. It accepts an
22227 expression, and defaults to "0".
22228 b Specify the brightness in the [-10,10] range. It accepts an
22230 expression and defaults to "0".
22231 h and H are mutually exclusive, and can't be specified at the same
22233 time.
22234 The b, h, H and s option values are expressions containing the
22236 following constants:
22237 n frame count of the input frame starting from 0
22239 pts presentation timestamp of the input frame expressed in time base
22241 units
22242 r frame rate of the input video, NAN if the input frame rate is
22244 unknown
22245 t timestamp expressed in seconds, NAN if the input timestamp is
22247 unknown
22248 tb time base of the input video
22250 Examples
22252 • Set the hue to 90 degrees and the saturation to 1.0:
22254 hue=h=90:s=1
22256 • Same command but expressing the hue in radians:
22258 hue=H=PI/2:s=1
22260 • Rotate hue and make the saturation swing between 0 and 2 over a
22262 period of 1 second:
22263 hue="H=2*PI*t: s=sin(2*PI*t)+1"
22265 • Apply a 3 seconds saturation fade-in effect starting at 0:
22267 hue="s=min(t/3\,1)"
22269 The general fade-in expression can be written as:
22271 hue="s=min(0\, max((t-START)/DURATION\, 1))"
22273 • Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
22275 hue="s=max(0\, min(1\, (8-t)/3))"
22277 The general fade-out expression can be written as:
22279 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
22281 Commands
22283 This filter supports the following commands:
22285 b
22287 s
22288 h
22289 H Modify the hue and/or the saturation and/or brightness of the input
22290 video. The command accepts the same syntax of the corresponding
22291 option.
22292 If the specified expression is not valid, it is kept at its current
22294 value.
22295 huesaturation
22297 Apply hue-saturation-intensity adjustments to input video stream.
22298 This filter operates in RGB colorspace.
22300 This filter accepts the following options:
22302 hue Set the hue shift in degrees to apply. Default is 0. Allowed range
22304 is from -180 to 180.
22305 saturation
22307 Set the saturation shift. Default is 0. Allowed range is from -1
22308 to 1.
22309 intensity
22311 Set the intensity shift. Default is 0. Allowed range is from -1 to
22312 1.
22313 colors
22315 Set which primary and complementary colors are going to be
22316 adjusted. This options is set by providing one or multiple values.
22317 This can select multiple colors at once. By default all colors are
22318 selected.
22319 r Adjust reds.
22321 y Adjust yellows.
22323 g Adjust greens.
22325 c Adjust cyans.
22327 b Adjust blues.
22329 m Adjust magentas.
22331 a Adjust all colors.
22333 strength
22335 Set strength of filtering. Allowed range is from 0 to 100. Default
22336 value is 1.
22337 rw, gw, bw
22339 Set weight for each RGB component. Allowed range is from 0 to 1.
22340 By default is set to 0.333, 0.334, 0.333. Those options are used
22341 in saturation and lightess processing.
22342 lightness
22344 Set preserving lightness, by default is disabled. Adjusting hues
22345 can change lightness from original RGB triplet, with this option
22346 enabled lightness is kept at same value.
22347 hysteresis
22349 Grow first stream into second stream by connecting components. This
22350 makes it possible to build more robust edge masks.
22351 This filter accepts the following options:
22353 planes
22355 Set which planes will be processed as bitmap, unprocessed planes
22356 will be copied from first stream. By default value 0xf, all planes
22357 will be processed.
22358 threshold
22360 Set threshold which is used in filtering. If pixel component value
22361 is higher than this value filter algorithm for connecting
22362 components is activated. By default value is 0.
22363 The "hysteresis" filter also supports the framesync options.
22365 iccdetect
22367 Detect the colorspace from an embedded ICC profile (if present), and
22368 update the frame's tags accordingly.
22369 This filter accepts the following options:
22371 force
22373 If true, the frame's existing colorspace tags will always be
22374 overridden by values detected from an ICC profile. Otherwise, they
22375 will only be assigned if they contain "unknown". Enabled by
22376 default.
22377 iccgen
22379 Generate ICC profiles and attach them to frames.
22380 This filter accepts the following options:
22382 color_primaries
22384 color_trc
22385 Configure the colorspace that the ICC profile will be generated
22386 for. The default value of "auto" infers the value from the input
22387 frame's metadata, defaulting to BT.709/sRGB as appropriate.
22388 See the setparams filter for a list of possible values, but note
22390 that "unknown" are not valid values for this filter.
22391 force
22393 If true, an ICC profile will be generated even if it would
22394 overwrite an already existing ICC profile. Disabled by default.
22395 identity
22397 Obtain the identity score between two input videos.
22398 This filter takes two input videos.
22400 Both input videos must have the same resolution and pixel format for
22402 this filter to work correctly. Also it assumes that both inputs have
22403 the same number of frames, which are compared one by one.
22404 The obtained per component, average, min and max identity score is
22406 printed through the logging system.
22407 The filter stores the calculated identity scores of each frame in frame
22409 metadata.
22410 In the below example the input file main.mpg being processed is
22412 compared with the reference file ref.mpg.
22413 ffmpeg -i main.mpg -i ref.mpg -lavfi identity -f null -
22415 idet
22417 Detect video interlacing type.
22418 This filter tries to detect if the input frames are interlaced,
22420 progressive, top or bottom field first. It will also try to detect
22421 fields that are repeated between adjacent frames (a sign of telecine).
22422 Single frame detection considers only immediately adjacent frames when
22424 classifying each frame. Multiple frame detection incorporates the
22425 classification history of previous frames.
22426 The filter will log these metadata values:
22428 single.current_frame
22430 Detected type of current frame using single-frame detection. One
22431 of: ``tff'' (top field first), ``bff'' (bottom field first),
22432 ``progressive'', or ``undetermined''
22433 single.tff
22435 Cumulative number of frames detected as top field first using
22436 single-frame detection.
22437 multiple.tff
22439 Cumulative number of frames detected as top field first using
22440 multiple-frame detection.
22441 single.bff
22443 Cumulative number of frames detected as bottom field first using
22444 single-frame detection.
22445 multiple.current_frame
22447 Detected type of current frame using multiple-frame detection. One
22448 of: ``tff'' (top field first), ``bff'' (bottom field first),
22449 ``progressive'', or ``undetermined''
22450 multiple.bff
22452 Cumulative number of frames detected as bottom field first using
22453 multiple-frame detection.
22454 single.progressive
22456 Cumulative number of frames detected as progressive using single-
22457 frame detection.
22458 multiple.progressive
22460 Cumulative number of frames detected as progressive using multiple-
22461 frame detection.
22462 single.undetermined
22464 Cumulative number of frames that could not be classified using
22465 single-frame detection.
22466 multiple.undetermined
22468 Cumulative number of frames that could not be classified using
22469 multiple-frame detection.
22470 repeated.current_frame
22472 Which field in the current frame is repeated from the last. One of
22473 ``neither'', ``top'', or ``bottom''.
22474 repeated.neither
22476 Cumulative number of frames with no repeated field.
22477 repeated.top
22479 Cumulative number of frames with the top field repeated from the
22480 previous frame's top field.
22481 repeated.bottom
22483 Cumulative number of frames with the bottom field repeated from the
22484 previous frame's bottom field.
22485 The filter accepts the following options:
22487 intl_thres
22489 Set interlacing threshold.
22490 prog_thres
22492 Set progressive threshold.
22493 rep_thres
22495 Threshold for repeated field detection.
22496 half_life
22498 Number of frames after which a given frame's contribution to the
22499 statistics is halved (i.e., it contributes only 0.5 to its
22500 classification). The default of 0 means that all frames seen are
22501 given full weight of 1.0 forever.
22502 analyze_interlaced_flag
22504 When this is not 0 then idet will use the specified number of
22505 frames to determine if the interlaced flag is accurate, it will not
22506 count undetermined frames. If the flag is found to be accurate it
22507 will be used without any further computations, if it is found to be
22508 inaccurate it will be cleared without any further computations.
22509 This allows inserting the idet filter as a low computational method
22510 to clean up the interlaced flag
22511 il
22513 Deinterleave or interleave fields.
22514 This filter allows one to process interlaced images fields without
22516 deinterlacing them. Deinterleaving splits the input frame into 2 fields
22517 (so called half pictures). Odd lines are moved to the top half of the
22518 output image, even lines to the bottom half. You can process (filter)
22519 them independently and then re-interleave them.
22520 The filter accepts the following options:
22522 luma_mode, l
22524 chroma_mode, c
22525 alpha_mode, a
22526 Available values for luma_mode, chroma_mode and alpha_mode are:
22527 none
22529 Do nothing.
22530 deinterleave, d
22532 Deinterleave fields, placing one above the other.
22533 interleave, i
22535 Interleave fields. Reverse the effect of deinterleaving.
22536 Default value is "none".
22538 luma_swap, ls
22540 chroma_swap, cs
22541 alpha_swap, as
22542 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default
22543 value is 0.
22544 Commands
22546 This filter supports the all above options as commands.
22548 inflate
22550 Apply inflate effect to the video.
22551 This filter replaces the pixel by the local(3x3) average by taking into
22553 account only values higher than the pixel.
22554 It accepts the following options:
22556 threshold0
22558 threshold1
22559 threshold2
22560 threshold3
22561 Limit the maximum change for each plane, default is 65535. If 0,
22562 plane will remain unchanged.
22563 Commands
22565 This filter supports the all above options as commands.
22567 interlace
22569 Simple interlacing filter from progressive contents. This interleaves
22570 upper (or lower) lines from odd frames with lower (or upper) lines from
22571 even frames, halving the frame rate and preserving image height.
22572 Original Original New Frame
22574 Frame 'j' Frame 'j+1' (tff)
22575 ========== =========== ==================
22576 Line 0 --------------------> Frame 'j' Line 0
22577 Line 1 Line 1 ----> Frame 'j+1' Line 1
22578 Line 2 ---------------------> Frame 'j' Line 2
22579 Line 3 Line 3 ----> Frame 'j+1' Line 3
22580 ... ... ...
22581 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
22582 It accepts the following optional parameters:
22584 scan
22586 This determines whether the interlaced frame is taken from the even
22587 (tff - default) or odd (bff) lines of the progressive frame.
22588 lowpass
22590 Vertical lowpass filter to avoid twitter interlacing and reduce
22591 moire patterns.
22592 0, off
22594 Disable vertical lowpass filter
22595 1, linear
22597 Enable linear filter (default)
22598 2, complex
22600 Enable complex filter. This will slightly less reduce twitter
22601 and moire but better retain detail and subjective sharpness
22602 impression.
22603 kerndeint
22605 Deinterlace input video by applying Donald Graft's adaptive kernel
22606 deinterling. Work on interlaced parts of a video to produce progressive
22607 frames.
22608 The description of the accepted parameters follows.
22610 thresh
22612 Set the threshold which affects the filter's tolerance when
22613 determining if a pixel line must be processed. It must be an
22614 integer in the range [0,255] and defaults to 10. A value of 0 will
22615 result in applying the process on every pixels.
22616 map Paint pixels exceeding the threshold value to white if set to 1.
22618 Default is 0.
22619 order
22621 Set the fields order. Swap fields if set to 1, leave fields alone
22622 if 0. Default is 0.
22623 sharp
22625 Enable additional sharpening if set to 1. Default is 0.
22626 twoway
22628 Enable twoway sharpening if set to 1. Default is 0.
22629 Examples
22631 • Apply default values:
22633 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
22635 • Enable additional sharpening:
22637 kerndeint=sharp=1
22639 • Paint processed pixels in white:
22641 kerndeint=map=1
22643 kirsch
22645 Apply kirsch operator to input video stream.
22646 The filter accepts the following option:
22648 planes
22650 Set which planes will be processed, unprocessed planes will be
22651 copied. By default value 0xf, all planes will be processed.
22652 scale
22654 Set value which will be multiplied with filtered result.
22655 delta
22657 Set value which will be added to filtered result.
22658 Commands
22660 This filter supports the all above options as commands.
22662 lagfun
22664 Slowly update darker pixels.
22665 This filter makes short flashes of light appear longer. This filter
22667 accepts the following options:
22668 decay
22670 Set factor for decaying. Default is .95. Allowed range is from 0 to
22671 1.
22672 planes
22674 Set which planes to filter. Default is all. Allowed range is from 0
22675 to 15.
22676 Commands
22678 This filter supports the all above options as commands.
22680 lenscorrection
22682 Correct radial lens distortion
22683 This filter can be used to correct for radial distortion as can result
22685 from the use of wide angle lenses, and thereby re-rectify the image. To
22686 find the right parameters one can use tools available for example as
22687 part of opencv or simply trial-and-error. To use opencv use the
22688 calibration sample (under samples/cpp) from the opencv sources and
22689 extract the k1 and k2 coefficients from the resulting matrix.
22690 Note that effectively the same filter is available in the open-source
22692 tools Krita and Digikam from the KDE project.
22693 In contrast to the vignette filter, which can also be used to
22695 compensate lens errors, this filter corrects the distortion of the
22696 image, whereas vignette corrects the brightness distribution, so you
22697 may want to use both filters together in certain cases, though you will
22698 have to take care of ordering, i.e. whether vignetting should be
22699 applied before or after lens correction.
22700 Options
22702 The filter accepts the following options:
22704 cx Relative x-coordinate of the focal point of the image, and thereby
22706 the center of the distortion. This value has a range [0,1] and is
22707 expressed as fractions of the image width. Default is 0.5.
22708 cy Relative y-coordinate of the focal point of the image, and thereby
22710 the center of the distortion. This value has a range [0,1] and is
22711 expressed as fractions of the image height. Default is 0.5.
22712 k1 Coefficient of the quadratic correction term. This value has a
22714 range [-1,1]. 0 means no correction. Default is 0.
22715 k2 Coefficient of the double quadratic correction term. This value has
22717 a range [-1,1]. 0 means no correction. Default is 0.
22718 i Set interpolation type. Can be "nearest" or "bilinear". Default is
22720 "nearest".
22721 fc Specify the color of the unmapped pixels. For the syntax of this
22723 option, check the "Color" section in the ffmpeg-utils manual.
22724 Default color is "black@0".
22725 The formula that generates the correction is:
22727 r_src = r_tgt * (1 + k1 * (r_tgt / r_0)^2 + k2 * (r_tgt / r_0)^4)
22729 where r_0 is halve of the image diagonal and r_src and r_tgt are the
22731 distances from the focal point in the source and target images,
22732 respectively.
22733 Commands
22735 This filter supports the all above options as commands.
22737 lensfun
22739 Apply lens correction via the lensfun library
22740 (<http://lensfun.sourceforge.net/>).
22741 The "lensfun" filter requires the camera make, camera model, and lens
22743 model to apply the lens correction. The filter will load the lensfun
22744 database and query it to find the corresponding camera and lens entries
22745 in the database. As long as these entries can be found with the given
22746 options, the filter can perform corrections on frames. Note that
22747 incomplete strings will result in the filter choosing the best match
22748 with the given options, and the filter will output the chosen camera
22749 and lens models (logged with level "info"). You must provide the make,
22750 camera model, and lens model as they are required.
22751 To obtain a list of available makes and models, leave out one or both
22753 of "make" and "model" options. The filter will send the full list to
22754 the log with level "INFO". The first column is the make and the second
22755 column is the model. To obtain a list of available lenses, set any
22756 values for make and model and leave out the "lens_model" option. The
22757 filter will send the full list of lenses in the log with level "INFO".
22758 The ffmpeg tool will exit after the list is printed.
22759 The filter accepts the following options:
22761 make
22763 The make of the camera (for example, "Canon"). This option is
22764 required.
22765 model
22767 The model of the camera (for example, "Canon EOS 100D"). This
22768 option is required.
22769 lens_model
22771 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6
22772 IS STM"). This option is required.
22773 db_path
22775 The full path to the lens database folder. If not set, the filter
22776 will attempt to load the database from the install path when the
22777 library was built. Default is unset.
22778 mode
22780 The type of correction to apply. The following values are valid
22781 options:
22782 vignetting
22784 Enables fixing lens vignetting.
22785 geometry
22787 Enables fixing lens geometry. This is the default.
22788 subpixel
22790 Enables fixing chromatic aberrations.
22791 vig_geo
22793 Enables fixing lens vignetting and lens geometry.
22794 vig_subpixel
22796 Enables fixing lens vignetting and chromatic aberrations.
22797 distortion
22799 Enables fixing both lens geometry and chromatic aberrations.
22800 all Enables all possible corrections.
22802 focal_length
22804 The focal length of the image/video (zoom; expected constant for
22805 video). For example, a 18--55mm lens has focal length range of
22806 [18--55], so a value in that range should be chosen when using that
22807 lens. Default 18.
22808 aperture
22810 The aperture of the image/video (expected constant for video). Note
22811 that aperture is only used for vignetting correction. Default 3.5.
22812 focus_distance
22814 The focus distance of the image/video (expected constant for
22815 video). Note that focus distance is only used for vignetting and
22816 only slightly affects the vignetting correction process. If
22817 unknown, leave it at the default value (which is 1000).
22818 scale
22820 The scale factor which is applied after transformation. After
22821 correction the video is no longer necessarily rectangular. This
22822 parameter controls how much of the resulting image is visible. The
22823 value 0 means that a value will be chosen automatically such that
22824 there is little or no unmapped area in the output image. 1.0 means
22825 that no additional scaling is done. Lower values may result in more
22826 of the corrected image being visible, while higher values may avoid
22827 unmapped areas in the output.
22828 target_geometry
22830 The target geometry of the output image/video. The following values
22831 are valid options:
22832 rectilinear (default)
22834 fisheye
22835 panoramic
22836 equirectangular
22837 fisheye_orthographic
22838 fisheye_stereographic
22839 fisheye_equisolid
22840 fisheye_thoby
22841 reverse
22842 Apply the reverse of image correction (instead of correcting
22843 distortion, apply it).
22844 interpolation
22846 The type of interpolation used when correcting distortion. The
22847 following values are valid options:
22848 nearest
22850 linear (default)
22851 lanczos
22852 Examples
22854 • Apply lens correction with make "Canon", camera model "Canon EOS
22856 100D", and lens model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with
22857 focal length of "18" and aperture of "8.0".
22858 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
22860 • Apply the same as before, but only for the first 5 seconds of
22862 video.
22863 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
22865 libplacebo
22867 Flexible GPU-accelerated processing filter based on libplacebo
22868 (<https://code.videolan.org/videolan/libplacebo>). Note that this
22869 filter currently only accepts Vulkan input frames.
22870 Options
22872 The options for this filter are divided into the following sections:
22874 Output mode
22876 These options control the overall output mode. By default, libplacebo
22878 will try to preserve the source colorimetry and size as best as it can,
22879 but it will apply any embedded film grain, dolby vision metadata or
22880 anamorphic SAR present in source frames.
22881 w
22883 h Set the output video dimension expression. Default value is the
22884 input dimension.
22885 Allows for the same expressions as the scale filter.
22887 format
22889 Set the output format override. If unset (the default), frames will
22890 be output in the same format as the respective input frames.
22891 Otherwise, format conversion will be performed.
22892 force_original_aspect_ratio
22894 force_divisible_by
22895 Work the same as the identical scale filter options.
22896 normalize_sar
22898 If enabled (the default), output frames will always have a pixel
22899 aspect ratio of 1:1. If disabled, any aspect ratio mismatches,
22900 including those from e.g. anamorphic video sources, are forwarded
22901 to the output pixel aspect ratio.
22902 pad_crop_ratio
22904 Specifies a ratio (between 0.0 and 1.0) between padding and
22905 cropping when the input aspect ratio does not match the output
22906 aspect ratio and normalize_sar is in effect. The default of 0.0
22907 always pads the content with black borders, while a value of 1.0
22908 always crops off parts of the content. Intermediate values are
22909 possible, leading to a mix of the two approaches.
22910 colorspace
22912 color_primaries
22913 color_trc
22914 range
22915 Configure the colorspace that output frames will be delivered in.
22916 The default value of "auto" outputs frames in the same format as
22917 the input frames, leading to no change. For any other value,
22918 conversion will be performed.
22919 See the setparams filter for a list of possible values.
22921 apply_filmgrain
22923 Apply film grain (e.g. AV1 or H.274) if present in source frames,
22924 and strip it from the output. Enabled by default.
22925 apply_dolbyvision
22927 Apply Dolby Vision RPU metadata if present in source frames, and
22928 strip it from the output. Enabled by default. Note that Dolby
22929 Vision will always output BT.2020+PQ, overriding the usual input
22930 frame metadata. These will also be picked as the values of "auto"
22931 for the respective frame output options.
22932 Scaling
22934 The options in this section control how libplacebo performs upscaling
22936 and (if necessary) downscaling. Note that libplacebo will always
22937 internally operate on 4:4:4 content, so any sub-sampled chroma formats
22938 such as "yuv420p" will necessarily be upsampled and downsampled as part
22939 of the rendering process. That means scaling might be in effect even if
22940 the source and destination resolution are the same.
22941 upscaler
22943 downscaler
22944 Configure the filter kernel used for upscaling and downscaling. The
22945 respective defaults are "spline36" and "mitchell". For a full list
22946 of possible values, pass "help" to these options. The most
22947 important values are:
22948 none
22950 Forces the use of built-in GPU texture sampling (typically
22951 bilinear). Extremely fast but poor quality, especially when
22952 downscaling.
22953 bilinear
22955 Bilinear interpolation. Can generally be done for free on GPUs,
22956 except when doing so would lead to aliasing. Fast and low
22957 quality.
22958 nearest
22960 Nearest-neighbour interpolation. Sharp but highly aliasing.
22961 oversample
22963 Algorithm that looks visually similar to nearest-neighbour
22964 interpolation but tries to preserve pixel aspect ratio. Good
22965 for pixel art, since it results in minimal distortion of the
22966 artistic appearance.
22967 lanczos
22969 Standard sinc-sinc interpolation kernel.
22970 spline36
22972 Cubic spline approximation of lanczos. No difference in
22973 performance, but has very slightly less ringing.
22974 ewa_lanczos
22976 Elliptically weighted average version of lanczos, based on a
22977 jinc-sinc kernel. This is also popularly referred to as just
22978 "Jinc scaling". Slow but very high quality.
22979 gaussian
22981 Gaussian kernel. Has certain ideal mathematical properties, but
22982 subjectively very blurry.
22983 mitchell
22985 Cubic BC spline with parameters recommended by Mitchell and
22986 Netravali. Very little ringing.
22987 lut_entries
22989 Configures the size of scaler LUTs, ranging from 1 to 256. The
22990 default of 0 will pick libplacebo's internal default, typically 64.
22991 antiringing
22993 Enables anti-ringing (for non-EWA filters). The value (between 0.0
22994 and 1.0) configures the strength of the anti-ringing algorithm. May
22995 increase aliasing if set too high. Disabled by default.
22996 sigmoid
22998 Enable sigmoidal compression during upscaling. Reduces ringing
22999 slightly. Enabled by default.
23000 Debanding
23002 Libplacebo comes with a built-in debanding filter that is good at
23004 counteracting many common sources of banding and blocking. Turning this
23005 on is highly recommended whenever quality is desired.
23006 deband
23008 Enable (fast) debanding algorithm. Disabled by default.
23009 deband_iterations
23011 Number of deband iterations of the debanding algorithm. Each
23012 iteration is performed with progressively increased radius (and
23013 diminished threshold). Recommended values are in the range 1 to 4.
23014 Defaults to 1.
23015 deband_threshold
23017 Debanding filter strength. Higher numbers lead to more aggressive
23018 debanding. Defaults to 4.0.
23019 deband_radius
23021 Debanding filter radius. A higher radius is better for slow
23022 gradients, while a lower radius is better for steep gradients.
23023 Defaults to 16.0.
23024 deband_grain
23026 Amount of extra output grain to add. Helps hide imperfections.
23027 Defaults to 6.0.
23028 Color adjustment
23030 A collection of subjective color controls. Not very rigorous, so the
23032 exact effect will vary somewhat depending on the input primaries and
23033 colorspace.
23034 brightness
23036 Brightness boost, between "-1.0" and 1.0. Defaults to 0.0.
23037 contrast
23039 Contrast gain, between 0.0 and 16.0. Defaults to 1.0.
23040 saturation
23042 Saturation gain, between 0.0 and 16.0. Defaults to 1.0.
23043 hue Hue shift in radians, between "-3.14" and 3.14. Defaults to 0.0.
23045 This will rotate the UV subvector, defaulting to BT.709
23046 coefficients for RGB inputs.
23047 gamma
23049 Gamma adjustment, between 0.0 and 16.0. Defaults to 1.0.
23050 cones
23052 Cone model to use for color blindness simulation. Accepts any
23053 combination of "l", "m" and "s". Here are some examples:
23054 m Deuteranomaly / deuteranopia (affecting 3%-4% of the
23056 population)
23057 l Protanomaly / protanopia (affecting 1%-2% of the population)
23059 l+m Monochromacy (very rare)
23061 l+m+s
23063 Achromatopsy (complete loss of daytime vision, extremely rare)
23064 cone-strength
23066 Gain factor for the cones specified by "cones", between 0.0 and
23067 10.0. A value of 1.0 results in no change to color vision. A value
23068 of 0.0 (the default) simulates complete loss of those cones. Values
23069 above 1.0 result in exaggerating the differences between cones,
23070 which may help compensate for reduced color vision.
23071 Peak detection
23073 To help deal with sources that only have static HDR10 metadata (or no
23075 tagging whatsoever), libplacebo uses its own internal frame analysis
23076 compute shader to analyze source frames and adapt the tone mapping
23077 function in realtime. If this is too slow, or if exactly reproducible
23078 frame-perfect results are needed, it's recommended to turn this feature
23079 off.
23080 peak_detect
23082 Enable HDR peak detection. Ignores static MaxCLL/MaxFALL values in
23083 favor of dynamic detection from the input. Note that the detected
23084 values do not get written back to the output frames, they merely
23085 guide the internal tone mapping process. Enabled by default.
23086 smoothing_period
23088 Peak detection smoothing period, between 0.0 and 1000.0. Higher
23089 values result in peak detection becoming less responsive to changes
23090 in the input. Defaults to 100.0.
23091 minimum_peak
23093 Lower bound on the detected peak (relative to SDR white), between
23094 0.0 and 100.0. Defaults to 1.0.
23095 scene_threshold_low
23097 scene_threshold_high
23098 Lower and upper thresholds for scene change detection. Expressed in
23099 a logarithmic scale between 0.0 and 100.0. Default to 5.5 and 10.0,
23100 respectively. Setting either to a negative value disables this
23101 functionality.
23102 overshoot
23104 Peak smoothing overshoot margin, between 0.0 and 1.0. Provides a
23105 safety margin to prevent clipping as a result of peak smoothing.
23106 Defaults to 0.05, corresponding to a margin of 5%.
23107 Tone mapping
23109 The options in this section control how libplacebo performs tone-
23111 mapping and gamut-mapping when dealing with mismatches between wide-
23112 gamut or HDR content. In general, libplacebo relies on accurate source
23113 tagging and mastering display gamut information to produce the best
23114 results.
23115 intent
23117 Rendering intent to use when adapting between different primary
23118 color gamuts (after tone-mapping).
23119 perceptual
23121 Perceptual gamut mapping. Currently equivalent to relative
23122 colorimetric.
23123 relative
23125 Relative colorimetric. This is the default.
23126 absolute
23128 Absolute colorimetric.
23129 saturation
23131 Saturation mapping. Forcibly stretches the source gamut to the
23132 target gamut.
23133 gamut_mode
23135 How to handle out-of-gamut colors that can occur as a result of
23136 colorimetric gamut mapping.
23137 clip
23139 Do nothing, simply clip out-of-range colors to the RGB volume.
23140 This is the default.
23141 warn
23143 Highlight out-of-gamut pixels (by coloring them pink).
23144 darken
23146 Linearly reduces content brightness to preserves saturated
23147 details, followed by clipping the remaining out-of-gamut
23148 colors. As the name implies, this makes everything darker, but
23149 provides a good balance between preserving details and colors.
23150 desaturate
23152 Hard-desaturates out-of-gamut colors towards white, while
23153 preserving the luminance. Has a tendency to shift colors.
23154 tonemapping
23156 Tone-mapping algorithm to use. Available values are:
23157 auto
23159 Automatic selection based on internal heuristics. This is the
23160 default.
23161 clip
23163 Performs no tone-mapping, just clips out-of-range colors.
23164 Retains perfect color accuracy for in-range colors but
23165 completely destroys out-of-range information. Does not perform
23166 any black point adaptation. Not configurable.
23167 bt.2390
23169 EETF from the ITU-R Report BT.2390, a hermite spline roll-off
23170 with linear segment. The knee point offset is configurable.
23171 Note that this parameter defaults to 1.0, rather than the value
23172 of 0.5 from the ITU-R spec.
23173 bt.2446a
23175 EETF from ITU-R Report BT.2446, method A. Designed for well-
23176 mastered HDR sources. Can be used for both forward and inverse
23177 tone mapping. Not configurable.
23178 spline
23180 Simple spline consisting of two polynomials, joined by a single
23181 pivot point. The parameter gives the pivot point (in PQ
23182 space), defaulting to 0.30. Can be used for both forward and
23183 inverse tone mapping.
23184 reinhard
23186 Simple non-linear, global tone mapping algorithm. The parameter
23187 specifies the local contrast coefficient at the display peak.
23188 Essentially, a parameter of 0.5 implies that the reference
23189 white will be about half as bright as when clipping. Defaults
23190 to 0.5, which results in the simplest formulation of this
23191 function.
23192 mobius
23194 Generalization of the reinhard tone mapping algorithm to
23195 support an additional linear slope near black. The tone mapping
23196 parameter indicates the trade-off between the linear section
23197 and the non-linear section. Essentially, for a given parameter
23198 x, every color value below x will be mapped linearly, while
23199 higher values get non-linearly tone-mapped. Values near 1.0
23200 make this curve behave like "clip", while values near 0.0 make
23201 this curve behave like "reinhard". The default value is 0.3,
23202 which provides a good balance between colorimetric accuracy and
23203 preserving out-of-gamut details.
23204 hable
23206 Piece-wise, filmic tone-mapping algorithm developed by John
23207 Hable for use in Uncharted 2, inspired by a similar tone-
23208 mapping algorithm used by Kodak. Popularized by its use in
23209 video games with HDR rendering. Preserves both dark and bright
23210 details very well, but comes with the drawback of changing the
23211 average brightness quite significantly. This is sort of similar
23212 to "reinhard" with parameter 0.24.
23213 gamma
23215 Fits a gamma (power) function to transfer between the source
23216 and target color spaces, effectively resulting in a perceptual
23217 hard-knee joining two roughly linear sections. This preserves
23218 details at all scales fairly accurately, but can result in an
23219 image with a muted or dull appearance. The parameter is used as
23220 the cutoff point, defaulting to 0.5.
23221 linear
23223 Linearly stretches the input range to the output range, in PQ
23224 space. This will preserve all details accurately, but results
23225 in a significantly different average brightness. Can be used
23226 for inverse tone-mapping in addition to regular tone-mapping.
23227 The parameter can be used as an additional linear gain
23228 coefficient (defaulting to 1.0).
23229 tonemapping_param
23231 For tunable tone mapping functions, this parameter can be used to
23232 fine-tune the curve behavior. Refer to the documentation of
23233 "tonemapping". The default value of 0.0 is replaced by the curve's
23234 preferred default setting.
23235 tonemapping_mode
23237 This option determines how the tone mapping function specified by
23238 "tonemapping" is applied to the colors in a scene. Possible values
23239 are:
23240 auto
23242 Automatic selection based on internal heuristics. This is the
23243 default.
23244 rgb Apply the function per-channel in the RGB colorspace. Per-
23246 channel tone-mapping in RGB. Guarantees no clipping and heavily
23247 desaturates the output, but distorts the colors quite
23248 significantly. Very similar to the "Hollywood" look and feel.
23249 max Tone-mapping is performed on the brightest component found in
23251 the signal. Good at preserving details in highlights, but has a
23252 tendency to crush blacks.
23253 hybrid
23255 Tone-map per-channel for highlights and linearly (luma-based)
23256 for midtones/shadows, based on a fixed gamma 2.4 coefficient
23257 curve.
23258 luma
23260 Tone-map linearly on the luma component (CIE Y), and adjust
23261 (desaturate) the chromaticities to compensate using a simple
23262 constant factor. This is essentially the mode used in ITU-R
23263 BT.2446 method A.
23264 inverse_tonemapping
23266 If enabled, this filter will also attempt stretching SDR signals to
23267 fill HDR output color volumes. Disabled by default.
23268 tonemapping_crosstalk
23270 Extra tone-mapping crosstalk factor, between 0.0 and 0.3. This can
23271 help reduce issues tone-mapping certain bright spectral colors.
23272 Defaults to 0.04.
23273 tonemapping_lut_size
23275 Size of the tone-mapping LUT, between 2 and 1024. Defaults to 256.
23276 Note that this figure is squared when combined with "peak_detect".
23277 Dithering
23279 By default, libplacebo will dither whenever necessary, which includes
23281 rendering to any integer format below 16-bit precision. It's
23282 recommended to always leave this on, since not doing so may result in
23283 visible banding in the output, even if the "debanding" filter is
23284 enabled. If maximum performance is needed, use "ordered_fixed" instead
23285 of disabling dithering.
23286 dithering
23288 Dithering method to use. Accepts the following values:
23289 none
23291 Disables dithering completely. May result in visible banding.
23292 blue
23294 Dither with pseudo-blue noise. This is the default.
23295 ordered
23297 Tunable ordered dither pattern.
23298 ordered_fixed
23300 Faster ordered dither with a fixed size of 6. Texture-less.
23301 white
23303 Dither with white noise. Texture-less.
23304 dither_lut_size
23306 Dither LUT size, as log base2 between 1 and 8. Defaults to 6,
23307 corresponding to a LUT size of "64x64".
23308 dither_temporal
23310 Enables temporal dithering. Disabled by default.
23311 Custom shaders
23313 libplacebo supports a number of custom shaders based on the mpv .hook
23315 GLSL syntax. A collection of such shaders can be found here:
23316 <https://github.com/mpv-player/mpv/wiki/User-Scripts#user-shaders>
23317 A full description of the mpv shader format is beyond the scope of this
23319 section, but a summary can be found here:
23320 <https://mpv.io/manual/master/#options-glsl-shader>
23321 custom_shader_path
23323 Specifies a path to a custom shader file to load at runtime.
23324 custom_shader_bin
23326 Specifies a complete custom shader as a raw string.
23327 Debugging / performance
23329 All of the options in this section default off. They may be of
23331 assistance when attempting to squeeze the maximum performance at the
23332 cost of quality.
23333 skip_aa
23335 Disable anti-aliasing when downscaling.
23336 polar_cutoff
23338 Truncate polar (EWA) scaler kernels below this absolute magnitude,
23339 between 0.0 and 1.0.
23340 disable_linear
23342 Disable linear light scaling.
23343 disable_builtin
23345 Disable built-in GPU sampling (forces LUT).
23346 force_icc_lut
23348 Force the use of a full ICC 3DLUT for gamut mapping.
23349 disable_fbos
23351 Forcibly disable FBOs, resulting in loss of almost all
23352 functionality, but offering the maximum possible speed.
23353 Commands
23355 This filter supports almost all of the above options as commands.
23357 Examples
23359 • Complete example for how to initialize the Vulkan device, upload
23361 frames to the GPU, perform filter conversion to yuv420p, and
23362 download frames back to the CPU for output. Note that in specific
23363 cases you can get around the need to perform format conversion by
23364 specifying the correct "format" filter option corresponding to the
23365 input frames.
23366 ffmpeg -i $INPUT -init_hw_device vulkan -vf hwupload,libplacebo=format=yuv420p,hwdownload,format=yuv420p $OUTPUT
23368 • Tone-map input to standard gamut BT.709 output:
23370 libplacebo=colorspace=bt709:color_primaries=bt709:color_trc=bt709:range=tv
23372 • Rescale input to fit into standard 1080p, with high quality
23374 scaling:
23375 libplacebo=w=1920:h=1080:force_original_aspect_ratio=decrease:normalize_sar=true:upscaler=ewa_lanczos:downscaler=ewa_lanczos
23377 • Convert input to standard sRGB JPEG:
23379 libplacebo=format=yuv420p:colorspace=bt470bg:color_primaries=bt709:color_trc=iec61966-2-1:range=pc
23381 • Use higher quality debanding settings:
23383 libplacebo=deband=true:deband_iterations=3:deband_radius=8:deband_threshold=6
23385 • Run this filter on the CPU, on systems with Mesa installed (and
23387 with the most expensive options disabled):
23388 ffmpeg ... -init_hw_device vulkan:llvmpipe ... -vf libplacebo=upscaler=none:downscaler=none:peak_detect=false
23390 • Suppress CPU-based AV1/H.274 film grain application in the decoder,
23392 in favor of doing it with this filter. Note that this is only a
23393 gain if the frames are either already on the GPU, or if you're
23394 using libplacebo for other purposes, since otherwise the VRAM
23395 roundtrip will more than offset any expected speedup.
23396 ffmpeg -export_side_data +film_grain ... -vf libplacebo=apply_filmgrain=true
23398 libvmaf
23400 Calulate the VMAF (Video Multi-Method Assessment Fusion) score for a
23401 reference/distorted pair of input videos.
23402 The first input is the distorted video, and the second input is the
23404 reference video.
23405 The obtained VMAF score is printed through the logging system.
23407 It requires Netflix's vmaf library (libvmaf) as a pre-requisite. After
23409 installing the library it can be enabled using: "./configure
23410 --enable-libvmaf".
23411 The filter has following options:
23413 model
23415 A `|` delimited list of vmaf models. Each model can be configured
23416 with a number of parameters. Default value: "version=vmaf_v0.6.1"
23417 model_path
23419 Deprecated, use model='path=...'.
23420 enable_transform
23422 Deprecated, use model='enable_transform=true'.
23423 phone_model
23425 Deprecated, use model='enable_transform=true'.
23426 enable_conf_interval
23428 Deprecated, use model='enable_conf_interval=true'.
23429 feature
23431 A `|` delimited list of features. Each feature can be configured
23432 with a number of parameters.
23433 psnr
23435 Deprecated, use feature='name=psnr'.
23436 ssim
23438 Deprecated, use feature='name=ssim'.
23439 ms_ssim
23441 Deprecated, use feature='name=ms_ssim'.
23442 log_path
23444 Set the file path to be used to store log files.
23445 log_fmt
23447 Set the format of the log file (xml, json, csv, or sub).
23448 n_threads
23450 Set number of threads to be used when initializing libvmaf.
23451 Default value: 0, no threads.
23452 n_subsample
23454 Set frame subsampling interval to be used.
23455 This filter also supports the framesync options.
23457 Examples
23459 • In the examples below, a distorted video distorted.mpg is compared
23461 with a reference file reference.mpg.
23462 • Basic usage:
23464 ffmpeg -i distorted.mpg -i reference.mpg -lavfi libvmaf=log_path=output.xml -f null -
23466 • Example with multiple models:
23468 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 -
23470 • Example with multiple addtional features:
23472 ffmpeg -i distorted.mpg -i reference.mpg -lavfi libvmaf='feature=name=psnr|name=ciede' -f null -
23474 • Example with options and different containers:
23476 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 -
23478 limitdiff
23480 Apply limited difference filter using second and optionally third video
23481 stream.
23482 The filter accepts the following options:
23484 threshold
23486 Set the threshold to use when allowing certain differences between
23487 video streams. Any absolute difference value lower or exact than
23488 this threshold will pick pixel components from first video stream.
23489 elasticity
23491 Set the elasticity of soft thresholding when processing video
23492 streams. This value multiplied with first one sets second
23493 threshold. Any absolute difference value greater or exact than
23494 second threshold will pick pixel components from second video
23495 stream. For values between those two threshold linear interpolation
23496 between first and second video stream will be used.
23497 reference
23499 Enable the reference (third) video stream processing. By default is
23500 disabled. If set, this video stream will be used for calculating
23501 absolute difference with first video stream.
23502 planes
23504 Specify which planes will be processed. Defaults to all available.
23505 Commands
23507 This filter supports the all above options as commands except option
23509 reference.
23510 limiter
23512 Limits the pixel components values to the specified range [min, max].
23513 The filter accepts the following options:
23515 min Lower bound. Defaults to the lowest allowed value for the input.
23517 max Upper bound. Defaults to the highest allowed value for the input.
23519 planes
23521 Specify which planes will be processed. Defaults to all available.
23522 Commands
23524 This filter supports the all above options as commands.
23526 loop
23528 Loop video frames.
23529 The filter accepts the following options:
23531 loop
23533 Set the number of loops. Setting this value to -1 will result in
23534 infinite loops. Default is 0.
23535 size
23537 Set maximal size in number of frames. Default is 0.
23538 start
23540 Set first frame of loop. Default is 0.
23541 Examples
23543 • Loop single first frame infinitely:
23545 loop=loop=-1:size=1:start=0
23547 • Loop single first frame 10 times:
23549 loop=loop=10:size=1:start=0
23551 • Loop 10 first frames 5 times:
23553 loop=loop=5:size=10:start=0
23555 lut1d
23557 Apply a 1D LUT to an input video.
23558 The filter accepts the following options:
23560 file
23562 Set the 1D LUT file name.
23563 Currently supported formats:
23565 cube
23567 Iridas
23568 csp cineSpace
23570 interp
23572 Select interpolation mode.
23573 Available values are:
23575 nearest
23577 Use values from the nearest defined point.
23578 linear
23580 Interpolate values using the linear interpolation.
23581 cosine
23583 Interpolate values using the cosine interpolation.
23584 cubic
23586 Interpolate values using the cubic interpolation.
23587 spline
23589 Interpolate values using the spline interpolation.
23590 Commands
23592 This filter supports the all above options as commands.
23594 lut3d
23596 Apply a 3D LUT to an input video.
23597 The filter accepts the following options:
23599 file
23601 Set the 3D LUT file name.
23602 Currently supported formats:
23604 3dl AfterEffects
23606 cube
23608 Iridas
23609 dat DaVinci
23611 m3d Pandora
23613 csp cineSpace
23615 interp
23617 Select interpolation mode.
23618 Available values are:
23620 nearest
23622 Use values from the nearest defined point.
23623 trilinear
23625 Interpolate values using the 8 points defining a cube.
23626 tetrahedral
23628 Interpolate values using a tetrahedron.
23629 pyramid
23631 Interpolate values using a pyramid.
23632 prism
23634 Interpolate values using a prism.
23635 Commands
23637 This filter supports the "interp" option as commands.
23639 lumakey
23641 Turn certain luma values into transparency.
23642 The filter accepts the following options:
23644 threshold
23646 Set the luma which will be used as base for transparency. Default
23647 value is 0.
23648 tolerance
23650 Set the range of luma values to be keyed out. Default value is
23651 0.01.
23652 softness
23654 Set the range of softness. Default value is 0. Use this to control
23655 gradual transition from zero to full transparency.
23656 Commands
23658 This filter supports same commands as options. The command accepts the
23660 same syntax of the corresponding option.
23661 If the specified expression is not valid, it is kept at its current
23663 value.
23664 lut, lutrgb, lutyuv
23666 Compute a look-up table for binding each pixel component input value to
23667 an output value, and apply it to the input video.
23668 lutyuv applies a lookup table to a YUV input video, lutrgb to an RGB
23670 input video.
23671 These filters accept the following parameters:
23673 c0 set first pixel component expression
23675 c1 set second pixel component expression
23677 c2 set third pixel component expression
23679 c3 set fourth pixel component expression, corresponds to the alpha
23681 component
23682 r set red component expression
23684 g set green component expression
23686 b set blue component expression
23688 a alpha component expression
23690 y set Y/luminance component expression
23692 u set U/Cb component expression
23694 v set V/Cr component expression
23696 Each of them specifies the expression to use for computing the lookup
23698 table for the corresponding pixel component values.
23699 The exact component associated to each of the c* options depends on the
23701 format in input.
23702 The lut filter requires either YUV or RGB pixel formats in input,
23704 lutrgb requires RGB pixel formats in input, and lutyuv requires YUV.
23705 The expressions can contain the following constants and functions:
23707 w
23709 h The input width and height.
23710 val The input value for the pixel component.
23712 clipval
23714 The input value, clipped to the minval-maxval range.
23715 maxval
23717 The maximum value for the pixel component.
23718 minval
23720 The minimum value for the pixel component.
23721 negval
23723 The negated value for the pixel component value, clipped to the
23724 minval-maxval range; it corresponds to the expression
23725 "maxval-clipval+minval".
23726 clip(val)
23728 The computed value in val, clipped to the minval-maxval range.
23729 gammaval(gamma)
23731 The computed gamma correction value of the pixel component value,
23732 clipped to the minval-maxval range. It corresponds to the
23733 expression
23734 "pow((clipval-minval)/(maxval-minval)\,gamma)*(maxval-minval)+minval"
23735 All expressions default to "clipval".
23737 Commands
23739 This filter supports same commands as options.
23741 Examples
23743 • Negate input video:
23745 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
23747 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
23748 The above is the same as:
23750 lutrgb="r=negval:g=negval:b=negval"
23752 lutyuv="y=negval:u=negval:v=negval"
23753 • Negate luminance:
23755 lutyuv=y=negval
23757 • Remove chroma components, turning the video into a graytone image:
23759 lutyuv="u=128:v=128"
23761 • Apply a luma burning effect:
23763 lutyuv="y=2*val"
23765 • Remove green and blue components:
23767 lutrgb="g=0:b=0"
23769 • Set a constant alpha channel value on input:
23771 format=rgba,lutrgb=a="maxval-minval/2"
23773 • Correct luminance gamma by a factor of 0.5:
23775 lutyuv=y=gammaval(0.5)
23777 • Discard least significant bits of luma:
23779 lutyuv=y='bitand(val, 128+64+32)'
23781 • Technicolor like effect:
23783 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
23785 lut2, tlut2
23787 The "lut2" filter takes two input streams and outputs one stream.
23788 The "tlut2" (time lut2) filter takes two consecutive frames from one
23790 single stream.
23791 This filter accepts the following parameters:
23793 c0 set first pixel component expression
23795 c1 set second pixel component expression
23797 c2 set third pixel component expression
23799 c3 set fourth pixel component expression, corresponds to the alpha
23801 component
23802 d set output bit depth, only available for "lut2" filter. By default
23804 is 0, which means bit depth is automatically picked from first
23805 input format.
23806 The "lut2" filter also supports the framesync options.
23808 Each of them specifies the expression to use for computing the lookup
23810 table for the corresponding pixel component values.
23811 The exact component associated to each of the c* options depends on the
23813 format in inputs.
23814 The expressions can contain the following constants:
23816 w
23818 h The input width and height.
23819 x The first input value for the pixel component.
23821 y The second input value for the pixel component.
23823 bdx The first input video bit depth.
23825 bdy The second input video bit depth.
23827 All expressions default to "x".
23829 Commands
23831 This filter supports the all above options as commands except option
23833 "d".
23834 Examples
23836 • Highlight differences between two RGB video streams:
23838 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)'
23840 • Highlight differences between two YUV video streams:
23842 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)'
23844 • Show max difference between two video streams:
23846 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)))'
23848 maskedclamp
23850 Clamp the first input stream with the second input and third input
23851 stream.
23852 Returns the value of first stream to be between second input stream -
23854 "undershoot" and third input stream + "overshoot".
23855 This filter accepts the following options:
23857 undershoot
23859 Default value is 0.
23860 overshoot
23862 Default value is 0.
23863 planes
23865 Set which planes will be processed as bitmap, unprocessed planes
23866 will be copied from first stream. By default value 0xf, all planes
23867 will be processed.
23868 Commands
23870 This filter supports the all above options as commands.
23872 maskedmax
23874 Merge the second and third input stream into output stream using
23875 absolute differences between second input stream and first input stream
23876 and absolute difference between third input stream and first input
23877 stream. The picked value will be from second input stream if second
23878 absolute difference is greater than first one or from third input
23879 stream otherwise.
23880 This filter accepts the following options:
23882 planes
23884 Set which planes will be processed as bitmap, unprocessed planes
23885 will be copied from first stream. By default value 0xf, all planes
23886 will be processed.
23887 Commands
23889 This filter supports the all above options as commands.
23891 maskedmerge
23893 Merge the first input stream with the second input stream using per
23894 pixel weights in the third input stream.
23895 A value of 0 in the third stream pixel component means that pixel
23897 component from first stream is returned unchanged, while maximum value
23898 (eg. 255 for 8-bit videos) means that pixel component from second
23899 stream is returned unchanged. Intermediate values define the amount of
23900 merging between both input stream's pixel components.
23901 This filter accepts the following options:
23903 planes
23905 Set which planes will be processed as bitmap, unprocessed planes
23906 will be copied from first stream. By default value 0xf, all planes
23907 will be processed.
23908 Commands
23910 This filter supports the all above options as commands.
23912 maskedmin
23914 Merge the second and third input stream into output stream using
23915 absolute differences between second input stream and first input stream
23916 and absolute difference between third input stream and first input
23917 stream. The picked value will be from second input stream if second
23918 absolute difference is less than first one or from third input stream
23919 otherwise.
23920 This filter accepts the following options:
23922 planes
23924 Set which planes will be processed as bitmap, unprocessed planes
23925 will be copied from first stream. By default value 0xf, all planes
23926 will be processed.
23927 Commands
23929 This filter supports the all above options as commands.
23931 maskedthreshold
23933 Pick pixels comparing absolute difference of two video streams with
23934 fixed threshold.
23935 If absolute difference between pixel component of first and second
23937 video stream is equal or lower than user supplied threshold than pixel
23938 component from first video stream is picked, otherwise pixel component
23939 from second video stream is picked.
23940 This filter accepts the following options:
23942 threshold
23944 Set threshold used when picking pixels from absolute difference
23945 from two input video streams.
23946 planes
23948 Set which planes will be processed as bitmap, unprocessed planes
23949 will be copied from second stream. By default value 0xf, all
23950 planes will be processed.
23951 Commands
23953 This filter supports the all above options as commands.
23955 maskfun
23957 Create mask from input video.
23958 For example it is useful to create motion masks after "tblend" filter.
23960 This filter accepts the following options:
23962 low Set low threshold. Any pixel component lower or exact than this
23964 value will be set to 0.
23965 high
23967 Set high threshold. Any pixel component higher than this value will
23968 be set to max value allowed for current pixel format.
23969 planes
23971 Set planes to filter, by default all available planes are filtered.
23972 fill
23974 Fill all frame pixels with this value.
23975 sum Set max average pixel value for frame. If sum of all pixel
23977 components is higher that this average, output frame will be
23978 completely filled with value set by fill option. Typically useful
23979 for scene changes when used in combination with "tblend" filter.
23980 Commands
23982 This filter supports the all above options as commands.
23984 mcdeint
23986 Apply motion-compensation deinterlacing.
23987 It needs one field per frame as input and must thus be used together
23989 with yadif=1/3 or equivalent.
23990 This filter is only available in ffmpeg version 4.4 or earlier.
23992 This filter accepts the following options:
23994 mode
23996 Set the deinterlacing mode.
23997 It accepts one of the following values:
23999 fast
24001 medium
24002 slow
24003 use iterative motion estimation
24004 extra_slow
24006 like slow, but use multiple reference frames.
24007 Default value is fast.
24009 parity
24011 Set the picture field parity assumed for the input video. It must
24012 be one of the following values:
24013 0, tff
24015 assume top field first
24016 1, bff
24018 assume bottom field first
24019 Default value is bff.
24021 qp Set per-block quantization parameter (QP) used by the internal
24023 encoder.
24024 Higher values should result in a smoother motion vector field but
24026 less optimal individual vectors. Default value is 1.
24027 median
24029 Pick median pixel from certain rectangle defined by radius.
24030 This filter accepts the following options:
24032 radius
24034 Set horizontal radius size. Default value is 1. Allowed range is
24035 integer from 1 to 127.
24036 planes
24038 Set which planes to process. Default is 15, which is all available
24039 planes.
24040 radiusV
24042 Set vertical radius size. Default value is 0. Allowed range is
24043 integer from 0 to 127. If it is 0, value will be picked from
24044 horizontal "radius" option.
24045 percentile
24047 Set median percentile. Default value is 0.5. Default value of 0.5
24048 will pick always median values, while 0 will pick minimum values,
24049 and 1 maximum values.
24050 Commands
24052 This filter supports same commands as options. The command accepts the
24054 same syntax of the corresponding option.
24055 If the specified expression is not valid, it is kept at its current
24057 value.
24058 mergeplanes
24060 Merge color channel components from several video streams.
24061 The filter accepts up to 4 input streams, and merge selected input
24063 planes to the output video.
24064 This filter accepts the following options:
24066 mapping
24068 Set input to output plane mapping. Default is 0.
24069 The mappings is specified as a bitmap. It should be specified as a
24071 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
24072 mapping for the first plane of the output stream. 'A' sets the
24073 number of the input stream to use (from 0 to 3), and 'a' the plane
24074 number of the corresponding input to use (from 0 to 3). The rest of
24075 the mappings is similar, 'Bb' describes the mapping for the output
24076 stream second plane, 'Cc' describes the mapping for the output
24077 stream third plane and 'Dd' describes the mapping for the output
24078 stream fourth plane.
24079 format
24081 Set output pixel format. Default is "yuva444p".
24082 map0s
24084 map1s
24085 map2s
24086 map3s
24087 Set input to output stream mapping for output Nth plane. Default is
24088 0.
24089 map0p
24091 map1p
24092 map2p
24093 map3p
24094 Set input to output plane mapping for output Nth plane. Default is
24095 0.
24096 Examples
24098 • Merge three gray video streams of same width and height into single
24100 video stream:
24101 [a0][a1][a2]mergeplanes=0x001020:yuv444p
24103 • Merge 1st yuv444p stream and 2nd gray video stream into yuva444p
24105 video stream:
24106 [a0][a1]mergeplanes=0x00010210:yuva444p
24108 • Swap Y and A plane in yuva444p stream:
24110 format=yuva444p,mergeplanes=0x03010200:yuva444p
24112 • Swap U and V plane in yuv420p stream:
24114 format=yuv420p,mergeplanes=0x000201:yuv420p
24116 • Cast a rgb24 clip to yuv444p:
24118 format=rgb24,mergeplanes=0x000102:yuv444p
24120 mestimate
24122 Estimate and export motion vectors using block matching algorithms.
24123 Motion vectors are stored in frame side data to be used by other
24124 filters.
24125 This filter accepts the following options:
24127 method
24129 Specify the motion estimation method. Accepts one of the following
24130 values:
24131 esa Exhaustive search algorithm.
24133 tss Three step search algorithm.
24135 tdls
24137 Two dimensional logarithmic search algorithm.
24138 ntss
24140 New three step search algorithm.
24141 fss Four step search algorithm.
24143 ds Diamond search algorithm.
24145 hexbs
24147 Hexagon-based search algorithm.
24148 epzs
24150 Enhanced predictive zonal search algorithm.
24151 umh Uneven multi-hexagon search algorithm.
24153 Default value is esa.
24155 mb_size
24157 Macroblock size. Default 16.
24158 search_param
24160 Search parameter. Default 7.
24161 midequalizer
24163 Apply Midway Image Equalization effect using two video streams.
24164 Midway Image Equalization adjusts a pair of images to have the same
24166 histogram, while maintaining their dynamics as much as possible. It's
24167 useful for e.g. matching exposures from a pair of stereo cameras.
24168 This filter has two inputs and one output, which must be of same pixel
24170 format, but may be of different sizes. The output of filter is first
24171 input adjusted with midway histogram of both inputs.
24172 This filter accepts the following option:
24174 planes
24176 Set which planes to process. Default is 15, which is all available
24177 planes.
24178 minterpolate
24180 Convert the video to specified frame rate using motion interpolation.
24181 This filter accepts the following options:
24183 fps Specify the output frame rate. This can be rational e.g.
24185 "60000/1001". Frames are dropped if fps is lower than source fps.
24186 Default 60.
24187 mi_mode
24189 Motion interpolation mode. Following values are accepted:
24190 dup Duplicate previous or next frame for interpolating new ones.
24192 blend
24194 Blend source frames. Interpolated frame is mean of previous and
24195 next frames.
24196 mci Motion compensated interpolation. Following options are
24198 effective when this mode is selected:
24199 mc_mode
24201 Motion compensation mode. Following values are accepted:
24202 obmc
24204 Overlapped block motion compensation.
24205 aobmc
24207 Adaptive overlapped block motion compensation. Window
24208 weighting coefficients are controlled adaptively
24209 according to the reliabilities of the neighboring
24210 motion vectors to reduce oversmoothing.
24211 Default mode is obmc.
24213 me_mode
24215 Motion estimation mode. Following values are accepted:
24216 bidir
24218 Bidirectional motion estimation. Motion vectors are
24219 estimated for each source frame in both forward and
24220 backward directions.
24221 bilat
24223 Bilateral motion estimation. Motion vectors are
24224 estimated directly for interpolated frame.
24225 Default mode is bilat.
24227 me The algorithm to be used for motion estimation. Following
24229 values are accepted:
24230 esa Exhaustive search algorithm.
24232 tss Three step search algorithm.
24234 tdls
24236 Two dimensional logarithmic search algorithm.
24237 ntss
24239 New three step search algorithm.
24240 fss Four step search algorithm.
24242 ds Diamond search algorithm.
24244 hexbs
24246 Hexagon-based search algorithm.
24247 epzs
24249 Enhanced predictive zonal search algorithm.
24250 umh Uneven multi-hexagon search algorithm.
24252 Default algorithm is epzs.
24254 mb_size
24256 Macroblock size. Default 16.
24257 search_param
24259 Motion estimation search parameter. Default 32.
24260 vsbmc
24262 Enable variable-size block motion compensation. Motion
24263 estimation is applied with smaller block sizes at object
24264 boundaries in order to make the them less blur. Default is
24265 0 (disabled).
24266 scd Scene change detection method. Scene change leads motion vectors to
24268 be in random direction. Scene change detection replace interpolated
24269 frames by duplicate ones. May not be needed for other modes.
24270 Following values are accepted:
24271 none
24273 Disable scene change detection.
24274 fdiff
24276 Frame difference. Corresponding pixel values are compared and
24277 if it satisfies scd_threshold scene change is detected.
24278 Default method is fdiff.
24280 scd_threshold
24282 Scene change detection threshold. Default is 10..
24283 mix
24285 Mix several video input streams into one video stream.
24286 A description of the accepted options follows.
24288 inputs
24290 The number of inputs. If unspecified, it defaults to 2.
24291 weights
24293 Specify weight of each input video stream as sequence. Each weight
24294 is separated by space. If number of weights is smaller than number
24295 of frames last specified weight will be used for all remaining
24296 unset weights.
24297 scale
24299 Specify scale, if it is set it will be multiplied with sum of each
24300 weight multiplied with pixel values to give final destination pixel
24301 value. By default scale is auto scaled to sum of weights.
24302 planes
24304 Set which planes to filter. Default is all. Allowed range is from 0
24305 to 15.
24306 duration
24308 Specify how end of stream is determined.
24309 longest
24311 The duration of the longest input. (default)
24312 shortest
24314 The duration of the shortest input.
24315 first
24317 The duration of the first input.
24318 Commands
24320 This filter supports the following commands:
24322 weights
24324 scale
24325 planes
24326 Syntax is same as option with same name.
24327 monochrome
24329 Convert video to gray using custom color filter.
24330 A description of the accepted options follows.
24332 cb Set the chroma blue spot. Allowed range is from -1 to 1. Default
24334 value is 0.
24335 cr Set the chroma red spot. Allowed range is from -1 to 1. Default
24337 value is 0.
24338 size
24340 Set the color filter size. Allowed range is from .1 to 10. Default
24341 value is 1.
24342 high
24344 Set the highlights strength. Allowed range is from 0 to 1. Default
24345 value is 0.
24346 Commands
24348 This filter supports the all above options as commands.
24350 morpho
24352 This filter allows to apply main morphological grayscale transforms,
24353 erode and dilate with arbitrary structures set in second input stream.
24354 Unlike naive implementation and much slower performance in erosion and
24356 dilation filters, when speed is critical "morpho" filter should be used
24357 instead.
24358 A description of accepted options follows,
24360 mode
24362 Set morphological transform to apply, can be:
24363 erode
24365 dilate
24366 open
24367 close
24368 gradient
24369 tophat
24370 blackhat
24371 Default is "erode".
24373 planes
24375 Set planes to filter, by default all planes except alpha are
24376 filtered.
24377 structure
24379 Set which structure video frames will be processed from second
24380 input stream, can be first or all. Default is all.
24381 The "morpho" filter also supports the framesync options.
24383 Commands
24385 This filter supports same commands as options.
24387 mpdecimate
24389 Drop frames that do not differ greatly from the previous frame in order
24390 to reduce frame rate.
24391 The main use of this filter is for very-low-bitrate encoding (e.g.
24393 streaming over dialup modem), but it could in theory be used for fixing
24394 movies that were inverse-telecined incorrectly.
24395 A description of the accepted options follows.
24397 max Set the maximum number of consecutive frames which can be dropped
24399 (if positive), or the minimum interval between dropped frames (if
24400 negative). If the value is 0, the frame is dropped disregarding the
24401 number of previous sequentially dropped frames.
24402 Default value is 0.
24404 hi
24406 lo
24407 frac
24408 Set the dropping threshold values.
24409 Values for hi and lo are for 8x8 pixel blocks and represent actual
24411 pixel value differences, so a threshold of 64 corresponds to 1 unit
24412 of difference for each pixel, or the same spread out differently
24413 over the block.
24414 A frame is a candidate for dropping if no 8x8 blocks differ by more
24416 than a threshold of hi, and if no more than frac blocks (1 meaning
24417 the whole image) differ by more than a threshold of lo.
24418 Default value for hi is 64*12, default value for lo is 64*5, and
24420 default value for frac is 0.33.
24421 msad
24423 Obtain the MSAD (Mean Sum of Absolute Differences) between two input
24424 videos.
24425 This filter takes two input videos.
24427 Both input videos must have the same resolution and pixel format for
24429 this filter to work correctly. Also it assumes that both inputs have
24430 the same number of frames, which are compared one by one.
24431 The obtained per component, average, min and max MSAD is printed
24433 through the logging system.
24434 The filter stores the calculated MSAD of each frame in frame metadata.
24436 In the below example the input file main.mpg being processed is
24438 compared with the reference file ref.mpg.
24439 ffmpeg -i main.mpg -i ref.mpg -lavfi msad -f null -
24441 multiply
24443 Multiply first video stream pixels values with second video stream
24444 pixels values.
24445 The filter accepts the following options:
24447 scale
24449 Set the scale applied to second video stream. By default is 1.
24450 Allowed range is from 0 to 9.
24451 offset
24453 Set the offset applied to second video stream. By default is 0.5.
24454 Allowed range is from "-1" to 1.
24455 planes
24457 Specify planes from input video stream that will be processed. By
24458 default all planes are processed.
24459 Commands
24461 This filter supports same commands as options.
24463 negate
24465 Negate (invert) the input video.
24466 It accepts the following option:
24468 components
24470 Set components to negate.
24471 Available values for components are:
24473 y
24475 u
24476 v
24477 a
24478 r
24479 g
24480 b
24481 negate_alpha
24482 With value 1, it negates the alpha component, if present. Default
24483 value is 0.
24484 Commands
24486 This filter supports same commands as options.
24488 nlmeans
24490 Denoise frames using Non-Local Means algorithm.
24491 Each pixel is adjusted by looking for other pixels with similar
24493 contexts. This context similarity is defined by comparing their
24494 surrounding patches of size pxp. Patches are searched in an area of rxr
24495 around the pixel.
24496 Note that the research area defines centers for patches, which means
24498 some patches will be made of pixels outside that research area.
24499 The filter accepts the following options.
24501 s Set denoising strength. Default is 1.0. Must be in range [1.0,
24503 30.0].
24504 p Set patch size. Default is 7. Must be odd number in range [0, 99].
24506 pc Same as p but for chroma planes.
24508 The default value is 0 and means automatic.
24510 r Set research size. Default is 15. Must be odd number in range [0,
24512 99].
24513 rc Same as r but for chroma planes.
24515 The default value is 0 and means automatic.
24517 nnedi
24519 Deinterlace video using neural network edge directed interpolation.
24520 This filter accepts the following options:
24522 weights
24524 Mandatory option, without binary file filter can not work.
24525 Currently file can be found here:
24526 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
24527 deint
24529 Set which frames to deinterlace, by default it is "all". Can be
24530 "all" or "interlaced".
24531 field
24533 Set mode of operation.
24534 Can be one of the following:
24536 af Use frame flags, both fields.
24538 a Use frame flags, single field.
24540 t Use top field only.
24542 b Use bottom field only.
24544 tf Use both fields, top first.
24546 bf Use both fields, bottom first.
24548 planes
24550 Set which planes to process, by default filter process all frames.
24551 nsize
24553 Set size of local neighborhood around each pixel, used by the
24554 predictor neural network.
24555 Can be one of the following:
24557 s8x6
24559 s16x6
24560 s32x6
24561 s48x6
24562 s8x4
24563 s16x4
24564 s32x4
24565 nns Set the number of neurons in predictor neural network. Can be one
24566 of the following:
24567 n16
24569 n32
24570 n64
24571 n128
24572 n256
24573 qual
24574 Controls the number of different neural network predictions that
24575 are blended together to compute the final output value. Can be
24576 "fast", default or "slow".
24577 etype
24579 Set which set of weights to use in the predictor. Can be one of
24580 the following:
24581 a, abs
24583 weights trained to minimize absolute error
24584 s, mse
24586 weights trained to minimize squared error
24587 pscrn
24589 Controls whether or not the prescreener neural network is used to
24590 decide which pixels should be processed by the predictor neural
24591 network and which can be handled by simple cubic interpolation.
24592 The prescreener is trained to know whether cubic interpolation will
24593 be sufficient for a pixel or whether it should be predicted by the
24594 predictor nn. The computational complexity of the prescreener nn
24595 is much less than that of the predictor nn. Since most pixels can
24596 be handled by cubic interpolation, using the prescreener generally
24597 results in much faster processing. The prescreener is pretty
24598 accurate, so the difference between using it and not using it is
24599 almost always unnoticeable.
24600 Can be one of the following:
24602 none
24604 original
24605 new
24606 new2
24607 new3
24608 Default is "new".
24610 Commands
24612 This filter supports same commands as options, excluding weights
24614 option.
24615 noformat
24617 Force libavfilter not to use any of the specified pixel formats for the
24618 input to the next filter.
24619 It accepts the following parameters:
24621 pix_fmts
24623 A '|'-separated list of pixel format names, such as
24624 pix_fmts=yuv420p|monow|rgb24".
24625 Examples
24627 • Force libavfilter to use a format different from yuv420p for the
24629 input to the vflip filter:
24630 noformat=pix_fmts=yuv420p,vflip
24632 • Convert the input video to any of the formats not contained in the
24634 list:
24635 noformat=yuv420p|yuv444p|yuv410p
24637 noise
24639 Add noise on video input frame.
24640 The filter accepts the following options:
24642 all_seed
24644 c0_seed
24645 c1_seed
24646 c2_seed
24647 c3_seed
24648 Set noise seed for specific pixel component or all pixel components
24649 in case of all_seed. Default value is 123457.
24650 all_strength, alls
24652 c0_strength, c0s
24653 c1_strength, c1s
24654 c2_strength, c2s
24655 c3_strength, c3s
24656 Set noise strength for specific pixel component or all pixel
24657 components in case all_strength. Default value is 0. Allowed range
24658 is [0, 100].
24659 all_flags, allf
24661 c0_flags, c0f
24662 c1_flags, c1f
24663 c2_flags, c2f
24664 c3_flags, c3f
24665 Set pixel component flags or set flags for all components if
24666 all_flags. Available values for component flags are:
24667 a averaged temporal noise (smoother)
24669 p mix random noise with a (semi)regular pattern
24671 t temporal noise (noise pattern changes between frames)
24673 u uniform noise (gaussian otherwise)
24675 Examples
24677 Add temporal and uniform noise to input video:
24679 noise=alls=20:allf=t+u
24681 normalize
24683 Normalize RGB video (aka histogram stretching, contrast stretching).
24684 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
24685 For each channel of each frame, the filter computes the input range and
24687 maps it linearly to the user-specified output range. The output range
24688 defaults to the full dynamic range from pure black to pure white.
24689 Temporal smoothing can be used on the input range to reduce flickering
24691 (rapid changes in brightness) caused when small dark or bright objects
24692 enter or leave the scene. This is similar to the auto-exposure
24693 (automatic gain control) on a video camera, and, like a video camera,
24694 it may cause a period of over- or under-exposure of the video.
24695 The R,G,B channels can be normalized independently, which may cause
24697 some color shifting, or linked together as a single channel, which
24698 prevents color shifting. Linked normalization preserves hue.
24699 Independent normalization does not, so it can be used to remove some
24700 color casts. Independent and linked normalization can be combined in
24701 any ratio.
24702 The normalize filter accepts the following options:
24704 blackpt
24706 whitept
24707 Colors which define the output range. The minimum input value is
24708 mapped to the blackpt. The maximum input value is mapped to the
24709 whitept. The defaults are black and white respectively. Specifying
24710 white for blackpt and black for whitept will give color-inverted,
24711 normalized video. Shades of grey can be used to reduce the dynamic
24712 range (contrast). Specifying saturated colors here can create some
24713 interesting effects.
24714 smoothing
24716 The number of previous frames to use for temporal smoothing. The
24717 input range of each channel is smoothed using a rolling average
24718 over the current frame and the smoothing previous frames. The
24719 default is 0 (no temporal smoothing).
24720 independence
24722 Controls the ratio of independent (color shifting) channel
24723 normalization to linked (color preserving) normalization. 0.0 is
24724 fully linked, 1.0 is fully independent. Defaults to 1.0 (fully
24725 independent).
24726 strength
24728 Overall strength of the filter. 1.0 is full strength. 0.0 is a
24729 rather expensive no-op. Defaults to 1.0 (full strength).
24730 Commands
24732 This filter supports same commands as options, excluding smoothing
24734 option. The command accepts the same syntax of the corresponding
24735 option.
24736 If the specified expression is not valid, it is kept at its current
24738 value.
24739 Examples
24741 Stretch video contrast to use the full dynamic range, with no temporal
24743 smoothing; may flicker depending on the source content:
24744 normalize=blackpt=black:whitept=white:smoothing=0
24746 As above, but with 50 frames of temporal smoothing; flicker should be
24748 reduced, depending on the source content:
24749 normalize=blackpt=black:whitept=white:smoothing=50
24751 As above, but with hue-preserving linked channel normalization:
24753 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
24755 As above, but with half strength:
24757 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
24759 Map the darkest input color to red, the brightest input color to cyan:
24761 normalize=blackpt=red:whitept=cyan
24763 null
24765 Pass the video source unchanged to the output.
24766 ocr
24768 Optical Character Recognition
24769 This filter uses Tesseract for optical character recognition. To enable
24771 compilation of this filter, you need to configure FFmpeg with
24772 "--enable-libtesseract".
24773 It accepts the following options:
24775 datapath
24777 Set datapath to tesseract data. Default is to use whatever was set
24778 at installation.
24779 language
24781 Set language, default is "eng".
24782 whitelist
24784 Set character whitelist.
24785 blacklist
24787 Set character blacklist.
24788 The filter exports recognized text as the frame metadata
24790 "lavfi.ocr.text". The filter exports confidence of recognized words as
24791 the frame metadata "lavfi.ocr.confidence".
24792 ocv
24794 Apply a video transform using libopencv.
24795 To enable this filter, install the libopencv library and headers and
24797 configure FFmpeg with "--enable-libopencv".
24798 It accepts the following parameters:
24800 filter_name
24802 The name of the libopencv filter to apply.
24803 filter_params
24805 The parameters to pass to the libopencv filter. If not specified,
24806 the default values are assumed.
24807 Refer to the official libopencv documentation for more precise
24809 information:
24810 <http://docs.opencv.org/master/modules/imgproc/doc/filtering.html>
24811 Several libopencv filters are supported; see the following subsections.
24813 dilate
24815 Dilate an image by using a specific structuring element. It
24817 corresponds to the libopencv function "cvDilate".
24818 It accepts the parameters: struct_el|nb_iterations.
24820 struct_el represents a structuring element, and has the syntax:
24822 colsxrows+anchor_xxanchor_y/shape
24823 cols and rows represent the number of columns and rows of the
24825 structuring element, anchor_x and anchor_y the anchor point, and shape
24826 the shape for the structuring element. shape must be "rect", "cross",
24827 "ellipse", or "custom".
24828 If the value for shape is "custom", it must be followed by a string of
24830 the form "=filename". The file with name filename is assumed to
24831 represent a binary image, with each printable character corresponding
24832 to a bright pixel. When a custom shape is used, cols and rows are
24833 ignored, the number or columns and rows of the read file are assumed
24834 instead.
24835 The default value for struct_el is "3x3+0x0/rect".
24837 nb_iterations specifies the number of times the transform is applied to
24839 the image, and defaults to 1.
24840 Some examples:
24842 # Use the default values
24844 ocv=dilate
24845 # Dilate using a structuring element with a 5x5 cross, iterating two times
24847 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
24848 # Read the shape from the file diamond.shape, iterating two times.
24850 # The file diamond.shape may contain a pattern of characters like this
24851 # *
24852 # ***
24853 # *****
24854 # ***
24855 # *
24856 # The specified columns and rows are ignored
24857 # but the anchor point coordinates are not
24858 ocv=dilate:0x0+2x2/custom=diamond.shape|2
24859 erode
24861 Erode an image by using a specific structuring element. It corresponds
24863 to the libopencv function "cvErode".
24864 It accepts the parameters: struct_el:nb_iterations, with the same
24866 syntax and semantics as the dilate filter.
24867 smooth
24869 Smooth the input video.
24871 The filter takes the following parameters:
24873 type|param1|param2|param3|param4.
24874 type is the type of smooth filter to apply, and must be one of the
24876 following values: "blur", "blur_no_scale", "median", "gaussian", or
24877 "bilateral". The default value is "gaussian".
24878 The meaning of param1, param2, param3, and param4 depends on the smooth
24880 type. param1 and param2 accept integer positive values or 0. param3 and
24881 param4 accept floating point values.
24882 The default value for param1 is 3. The default value for the other
24884 parameters is 0.
24885 These parameters correspond to the parameters assigned to the libopencv
24887 function "cvSmooth".
24888 oscilloscope
24890 2D Video Oscilloscope.
24891 Useful to measure spatial impulse, step responses, chroma delays, etc.
24893 It accepts the following parameters:
24895 x Set scope center x position.
24897 y Set scope center y position.
24899 s Set scope size, relative to frame diagonal.
24901 t Set scope tilt/rotation.
24903 o Set trace opacity.
24905 tx Set trace center x position.
24907 ty Set trace center y position.
24909 tw Set trace width, relative to width of frame.
24911 th Set trace height, relative to height of frame.
24913 c Set which components to trace. By default it traces first three
24915 components.
24916 g Draw trace grid. By default is enabled.
24918 st Draw some statistics. By default is enabled.
24920 sc Draw scope. By default is enabled.
24922 Commands
24924 This filter supports same commands as options. The command accepts the
24926 same syntax of the corresponding option.
24927 If the specified expression is not valid, it is kept at its current
24929 value.
24930 Examples
24932 • Inspect full first row of video frame.
24934 oscilloscope=x=0.5:y=0:s=1
24936 • Inspect full last row of video frame.
24938 oscilloscope=x=0.5:y=1:s=1
24940 • Inspect full 5th line of video frame of height 1080.
24942 oscilloscope=x=0.5:y=5/1080:s=1
24944 • Inspect full last column of video frame.
24946 oscilloscope=x=1:y=0.5:s=1:t=1
24948 overlay
24950 Overlay one video on top of another.
24951 It takes two inputs and has one output. The first input is the "main"
24953 video on which the second input is overlaid.
24954 It accepts the following parameters:
24956 A description of the accepted options follows.
24958 x
24960 y Set the expression for the x and y coordinates of the overlaid
24961 video on the main video. Default value is "0" for both expressions.
24962 In case the expression is invalid, it is set to a huge value
24963 (meaning that the overlay will not be displayed within the output
24964 visible area).
24965 eof_action
24967 See framesync.
24968 eval
24970 Set when the expressions for x, and y are evaluated.
24971 It accepts the following values:
24973 init
24975 only evaluate expressions once during the filter initialization
24976 or when a command is processed
24977 frame
24979 evaluate expressions for each incoming frame
24980 Default value is frame.
24982 shortest
24984 See framesync.
24985 format
24987 Set the format for the output video.
24988 It accepts the following values:
24990 yuv420
24992 force YUV420 output
24993 yuv420p10
24995 force YUV420p10 output
24996 yuv422
24998 force YUV422 output
24999 yuv422p10
25001 force YUV422p10 output
25002 yuv444
25004 force YUV444 output
25005 rgb force packed RGB output
25007 gbrp
25009 force planar RGB output
25010 auto
25012 automatically pick format
25013 Default value is yuv420.
25015 repeatlast
25017 See framesync.
25018 alpha
25020 Set format of alpha of the overlaid video, it can be straight or
25021 premultiplied. Default is straight.
25022 The x, and y expressions can contain the following parameters.
25024 main_w, W
25026 main_h, H
25027 The main input width and height.
25028 overlay_w, w
25030 overlay_h, h
25031 The overlay input width and height.
25032 x
25034 y The computed values for x and y. They are evaluated for each new
25035 frame.
25036 hsub
25038 vsub
25039 horizontal and vertical chroma subsample values of the output
25040 format. For example for the pixel format "yuv422p" hsub is 2 and
25041 vsub is 1.
25042 n the number of input frame, starting from 0
25044 pos the position in the file of the input frame, NAN if unknown
25046 t The timestamp, expressed in seconds. It's NAN if the input
25048 timestamp is unknown.
25049 This filter also supports the framesync options.
25051 Note that the n, pos, t variables are available only when evaluation is
25053 done per frame, and will evaluate to NAN when eval is set to init.
25054 Be aware that frames are taken from each input video in timestamp
25056 order, hence, if their initial timestamps differ, it is a good idea to
25057 pass the two inputs through a setpts=PTS-STARTPTS filter to have them
25058 begin in the same zero timestamp, as the example for the movie filter
25059 does.
25060 You can chain together more overlays but you should test the efficiency
25062 of such approach.
25063 Commands
25065 This filter supports the following commands:
25067 x
25069 y Modify the x and y of the overlay input. The command accepts the
25070 same syntax of the corresponding option.
25071 If the specified expression is not valid, it is kept at its current
25073 value.
25074 Examples
25076 • Draw the overlay at 10 pixels from the bottom right corner of the
25078 main video:
25079 overlay=main_w-overlay_w-10:main_h-overlay_h-10
25081 Using named options the example above becomes:
25083 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
25085 • Insert a transparent PNG logo in the bottom left corner of the
25087 input, using the ffmpeg tool with the "-filter_complex" option:
25088 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
25090 • Insert 2 different transparent PNG logos (second logo on bottom
25092 right corner) using the ffmpeg tool:
25093 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
25095 • Add a transparent color layer on top of the main video; "WxH" must
25097 specify the size of the main input to the overlay filter:
25098 color=color=red@.3:size=WxH [over]; [in][over] overlay [out]
25100 • Play an original video and a filtered version (here with the
25102 deshake filter) side by side using the ffplay tool:
25103 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
25105 The above command is the same as:
25107 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
25109 • Make a sliding overlay appearing from the left to the right top
25111 part of the screen starting since time 2:
25112 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
25114 • Compose output by putting two input videos side to side:
25116 ffmpeg -i left.avi -i right.avi -filter_complex "
25118 nullsrc=size=200x100 [background];
25119 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
25120 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
25121 [background][left] overlay=shortest=1 [background+left];
25122 [background+left][right] overlay=shortest=1:x=100 [left+right]
25123 "
25124 • Mask 10-20 seconds of a video by applying the delogo filter to a
25126 section
25127 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
25129 -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]'
25130 masked.avi
25131 • Chain several overlays in cascade:
25133 nullsrc=s=200x200 [bg];
25135 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
25136 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
25137 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
25138 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
25139 [in3] null, [mid2] overlay=100:100 [out0]
25140 overlay_cuda
25142 Overlay one video on top of another.
25143 This is the CUDA variant of the overlay filter. It only accepts CUDA
25145 frames. The underlying input pixel formats have to match.
25146 It takes two inputs and has one output. The first input is the "main"
25148 video on which the second input is overlaid.
25149 It accepts the following parameters:
25151 x
25153 y Set expressions for the x and y coordinates of the overlaid video
25154 on the main video.
25155 They can contain the following parameters:
25157 main_w, W
25159 main_h, H
25160 The main input width and height.
25161 overlay_w, w
25163 overlay_h, h
25164 The overlay input width and height.
25165 x
25167 y The computed values for x and y. They are evaluated for each
25168 new frame.
25169 n The ordinal index of the main input frame, starting from 0.
25171 pos The byte offset position in the file of the main input frame,
25173 NAN if unknown.
25174 t The timestamp of the main input frame, expressed in seconds,
25176 NAN if unknown.
25177 Default value is "0" for both expressions.
25179 eval
25181 Set when the expressions for x and y are evaluated.
25182 It accepts the following values:
25184 init
25186 Evaluate expressions once during filter initialization or when
25187 a command is processed.
25188 frame
25190 Evaluate expressions for each incoming frame
25191 Default value is frame.
25193 eof_action
25195 See framesync.
25196 shortest
25198 See framesync.
25199 repeatlast
25201 See framesync.
25202 This filter also supports the framesync options.
25204 owdenoise
25206 Apply Overcomplete Wavelet denoiser.
25207 The filter accepts the following options:
25209 depth
25211 Set depth.
25212 Larger depth values will denoise lower frequency components more,
25214 but slow down filtering.
25215 Must be an int in the range 8-16, default is 8.
25217 luma_strength, ls
25219 Set luma strength.
25220 Must be a double value in the range 0-1000, default is 1.0.
25222 chroma_strength, cs
25224 Set chroma strength.
25225 Must be a double value in the range 0-1000, default is 1.0.
25227 pad
25229 Add paddings to the input image, and place the original input at the
25230 provided x, y coordinates.
25231 It accepts the following parameters:
25233 width, w
25235 height, h
25236 Specify an expression for the size of the output image with the
25237 paddings added. If the value for width or height is 0, the
25238 corresponding input size is used for the output.
25239 The width expression can reference the value set by the height
25241 expression, and vice versa.
25242 The default value of width and height is 0.
25244 x
25246 y Specify the offsets to place the input image at within the padded
25247 area, with respect to the top/left border of the output image.
25248 The x expression can reference the value set by the y expression,
25250 and vice versa.
25251 The default value of x and y is 0.
25253 If x or y evaluate to a negative number, they'll be changed so the
25255 input image is centered on the padded area.
25256 color
25258 Specify the color of the padded area. For the syntax of this
25259 option, check the "Color" section in the ffmpeg-utils manual.
25260 The default value of color is "black".
25262 eval
25264 Specify when to evaluate width, height, x and y expression.
25265 It accepts the following values:
25267 init
25269 Only evaluate expressions once during the filter initialization
25270 or when a command is processed.
25271 frame
25273 Evaluate expressions for each incoming frame.
25274 Default value is init.
25276 aspect
25278 Pad to aspect instead to a resolution.
25279 The value for the width, height, x, and y options are expressions
25281 containing the following constants:
25282 in_w
25284 in_h
25285 The input video width and height.
25286 iw
25288 ih These are the same as in_w and in_h.
25289 out_w
25291 out_h
25292 The output width and height (the size of the padded area), as
25293 specified by the width and height expressions.
25294 ow
25296 oh These are the same as out_w and out_h.
25297 x
25299 y The x and y offsets as specified by the x and y expressions, or NAN
25300 if not yet specified.
25301 a same as iw / ih
25303 sar input sample aspect ratio
25305 dar input display aspect ratio, it is the same as (iw / ih) * sar
25307 hsub
25309 vsub
25310 The horizontal and vertical chroma subsample values. For example
25311 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
25312 Examples
25314 • Add paddings with the color "violet" to the input video. The output
25316 video size is 640x480, and the top-left corner of the input video
25317 is placed at column 0, row 40
25318 pad=640:480:0:40:violet
25320 The example above is equivalent to the following command:
25322 pad=width=640:height=480:x=0:y=40:color=violet
25324 • Pad the input to get an output with dimensions increased by 3/2,
25326 and put the input video at the center of the padded area:
25327 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
25329 • Pad the input to get a squared output with size equal to the
25331 maximum value between the input width and height, and put the input
25332 video at the center of the padded area:
25333 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
25335 • Pad the input to get a final w/h ratio of 16:9:
25337 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
25339 • In case of anamorphic video, in order to set the output display
25341 aspect correctly, it is necessary to use sar in the expression,
25342 according to the relation:
25343 (ih * X / ih) * sar = output_dar
25345 X = output_dar / sar
25346 Thus the previous example needs to be modified to:
25348 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
25350 • Double the output size and put the input video in the bottom-right
25352 corner of the output padded area:
25353 pad="2*iw:2*ih:ow-iw:oh-ih"
25355 palettegen
25357 Generate one palette for a whole video stream.
25358 It accepts the following options:
25360 max_colors
25362 Set the maximum number of colors to quantize in the palette. Note:
25363 the palette will still contain 256 colors; the unused palette
25364 entries will be black.
25365 reserve_transparent
25367 Create a palette of 255 colors maximum and reserve the last one for
25368 transparency. Reserving the transparency color is useful for GIF
25369 optimization. If not set, the maximum of colors in the palette
25370 will be 256. You probably want to disable this option for a
25371 standalone image. Set by default.
25372 transparency_color
25374 Set the color that will be used as background for transparency.
25375 stats_mode
25377 Set statistics mode.
25378 It accepts the following values:
25380 full
25382 Compute full frame histograms.
25383 diff
25385 Compute histograms only for the part that differs from previous
25386 frame. This might be relevant to give more importance to the
25387 moving part of your input if the background is static.
25388 single
25390 Compute new histogram for each frame.
25391 Default value is full.
25393 use_alpha
25395 Create a palette of colors with alpha components. Setting this,
25396 will automatically disable 'reserve_transparent'.
25397 The filter also exports the frame metadata "lavfi.color_quant_ratio"
25399 ("nb_color_in / nb_color_out") which you can use to evaluate the degree
25400 of color quantization of the palette. This information is also visible
25401 at info logging level.
25402 Examples
25404 • Generate a representative palette of a given video using ffmpeg:
25406 ffmpeg -i input.mkv -vf palettegen palette.png
25408 paletteuse
25410 Use a palette to downsample an input video stream.
25411 The filter takes two inputs: one video stream and a palette. The
25413 palette must be a 256 pixels image.
25414 It accepts the following options:
25416 dither
25418 Select dithering mode. Available algorithms are:
25419 bayer
25421 Ordered 8x8 bayer dithering (deterministic)
25422 heckbert
25424 Dithering as defined by Paul Heckbert in 1982 (simple error
25425 diffusion). Note: this dithering is sometimes considered
25426 "wrong" and is included as a reference.
25427 floyd_steinberg
25429 Floyd and Steingberg dithering (error diffusion)
25430 sierra2
25432 Frankie Sierra dithering v2 (error diffusion)
25433 sierra2_4a
25435 Frankie Sierra dithering v2 "Lite" (error diffusion)
25436 Default is sierra2_4a.
25438 bayer_scale
25440 When bayer dithering is selected, this option defines the scale of
25441 the pattern (how much the crosshatch pattern is visible). A low
25442 value means more visible pattern for less banding, and higher value
25443 means less visible pattern at the cost of more banding.
25444 The option must be an integer value in the range [0,5]. Default is
25446 2.
25447 diff_mode
25449 If set, define the zone to process
25450 rectangle
25452 Only the changing rectangle will be reprocessed. This is
25453 similar to GIF cropping/offsetting compression mechanism. This
25454 option can be useful for speed if only a part of the image is
25455 changing, and has use cases such as limiting the scope of the
25456 error diffusal dither to the rectangle that bounds the moving
25457 scene (it leads to more deterministic output if the scene
25458 doesn't change much, and as a result less moving noise and
25459 better GIF compression).
25460 Default is none.
25462 new Take new palette for each output frame.
25464 alpha_threshold
25466 Sets the alpha threshold for transparency. Alpha values above this
25467 threshold will be treated as completely opaque, and values below
25468 this threshold will be treated as completely transparent.
25469 The option must be an integer value in the range [0,255]. Default
25471 is 128.
25472 use_alpha
25474 Apply the palette by taking alpha values into account. Only useful
25475 with palettes that are containing multiple colors with alpha
25476 components. Setting this will automatically disable
25477 'alpha_treshold'.
25478 Examples
25480 • Use a palette (generated for example with palettegen) to encode a
25482 GIF using ffmpeg:
25483 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
25485 perspective
25487 Correct perspective of video not recorded perpendicular to the screen.
25488 A description of the accepted parameters follows.
25490 x0
25492 y0
25493 x1
25494 y1
25495 x2
25496 y2
25497 x3
25498 y3 Set coordinates expression for top left, top right, bottom left and
25499 bottom right corners. Default values are "0:0:W:0:0:H:W:H" with
25500 which perspective will remain unchanged. If the "sense" option is
25501 set to "source", then the specified points will be sent to the
25502 corners of the destination. If the "sense" option is set to
25503 "destination", then the corners of the source will be sent to the
25504 specified coordinates.
25505 The expressions can use the following variables:
25507 W
25509 H the width and height of video frame.
25510 in Input frame count.
25512 on Output frame count.
25514 interpolation
25516 Set interpolation for perspective correction.
25517 It accepts the following values:
25519 linear
25521 cubic
25522 Default value is linear.
25524 sense
25526 Set interpretation of coordinate options.
25527 It accepts the following values:
25529 0, source
25531 Send point in the source specified by the given coordinates to
25532 the corners of the destination.
25533 1, destination
25535 Send the corners of the source to the point in the destination
25536 specified by the given coordinates.
25537 Default value is source.
25539 eval
25541 Set when the expressions for coordinates x0,y0,...x3,y3 are
25542 evaluated.
25543 It accepts the following values:
25545 init
25547 only evaluate expressions once during the filter initialization
25548 or when a command is processed
25549 frame
25551 evaluate expressions for each incoming frame
25552 Default value is init.
25554 phase
25556 Delay interlaced video by one field time so that the field order
25557 changes.
25558 The intended use is to fix PAL movies that have been captured with the
25560 opposite field order to the film-to-video transfer.
25561 A description of the accepted parameters follows.
25563 mode
25565 Set phase mode.
25566 It accepts the following values:
25568 t Capture field order top-first, transfer bottom-first. Filter
25570 will delay the bottom field.
25571 b Capture field order bottom-first, transfer top-first. Filter
25573 will delay the top field.
25574 p Capture and transfer with the same field order. This mode only
25576 exists for the documentation of the other options to refer to,
25577 but if you actually select it, the filter will faithfully do
25578 nothing.
25579 a Capture field order determined automatically by field flags,
25581 transfer opposite. Filter selects among t and b modes on a
25582 frame by frame basis using field flags. If no field information
25583 is available, then this works just like u.
25584 u Capture unknown or varying, transfer opposite. Filter selects
25586 among t and b on a frame by frame basis by analyzing the images
25587 and selecting the alternative that produces best match between
25588 the fields.
25589 T Capture top-first, transfer unknown or varying. Filter selects
25591 among t and p using image analysis.
25592 B Capture bottom-first, transfer unknown or varying. Filter
25594 selects among b and p using image analysis.
25595 A Capture determined by field flags, transfer unknown or varying.
25597 Filter selects among t, b and p using field flags and image
25598 analysis. If no field information is available, then this works
25599 just like U. This is the default mode.
25600 U Both capture and transfer unknown or varying. Filter selects
25602 among t, b and p using image analysis only.
25603 Commands
25605 This filter supports the all above options as commands.
25607 photosensitivity
25609 Reduce various flashes in video, so to help users with epilepsy.
25610 It accepts the following options:
25612 frames, f
25614 Set how many frames to use when filtering. Default is 30.
25615 threshold, t
25617 Set detection threshold factor. Default is 1. Lower is stricter.
25618 skip
25620 Set how many pixels to skip when sampling frames. Default is 1.
25621 Allowed range is from 1 to 1024.
25622 bypass
25624 Leave frames unchanged. Default is disabled.
25625 pixdesctest
25627 Pixel format descriptor test filter, mainly useful for internal
25628 testing. The output video should be equal to the input video.
25629 For example:
25631 format=monow, pixdesctest
25633 can be used to test the monowhite pixel format descriptor definition.
25635 pixelize
25637 Apply pixelization to video stream.
25638 The filter accepts the following options:
25640 width, w
25642 height, h
25643 Set block dimensions that will be used for pixelization. Default
25644 value is 16.
25645 mode, m
25647 Set the mode of pixelization used.
25648 Possible values are:
25650 avg
25652 min
25653 max
25654 Default value is "avg".
25656 planes, p
25658 Set what planes to filter. Default is to filter all planes.
25659 Commands
25661 This filter supports all options as commands.
25663 pixscope
25665 Display sample values of color channels. Mainly useful for checking
25666 color and levels. Minimum supported resolution is 640x480.
25667 The filters accept the following options:
25669 x Set scope X position, relative offset on X axis.
25671 y Set scope Y position, relative offset on Y axis.
25673 w Set scope width.
25675 h Set scope height.
25677 o Set window opacity. This window also holds statistics about pixel
25679 area.
25680 wx Set window X position, relative offset on X axis.
25682 wy Set window Y position, relative offset on Y axis.
25684 Commands
25686 This filter supports same commands as options.
25688 pp
25690 Enable the specified chain of postprocessing subfilters using
25691 libpostproc. This library should be automatically selected with a GPL
25692 build ("--enable-gpl"). Subfilters must be separated by '/' and can be
25693 disabled by prepending a '-'. Each subfilter and some options have a
25694 short and a long name that can be used interchangeably, i.e. dr/dering
25695 are the same.
25696 The filters accept the following options:
25698 subfilters
25700 Set postprocessing subfilters string.
25701 All subfilters share common options to determine their scope:
25703 a/autoq
25705 Honor the quality commands for this subfilter.
25706 c/chrom
25708 Do chrominance filtering, too (default).
25709 y/nochrom
25711 Do luminance filtering only (no chrominance).
25712 n/noluma
25714 Do chrominance filtering only (no luminance).
25715 These options can be appended after the subfilter name, separated by a
25717 '|'.
25718 Available subfilters are:
25720 hb/hdeblock[|difference[|flatness]]
25722 Horizontal deblocking filter
25723 difference
25725 Difference factor where higher values mean more deblocking
25726 (default: 32).
25727 flatness
25729 Flatness threshold where lower values mean more deblocking
25730 (default: 39).
25731 vb/vdeblock[|difference[|flatness]]
25733 Vertical deblocking filter
25734 difference
25736 Difference factor where higher values mean more deblocking
25737 (default: 32).
25738 flatness
25740 Flatness threshold where lower values mean more deblocking
25741 (default: 39).
25742 ha/hadeblock[|difference[|flatness]]
25744 Accurate horizontal deblocking filter
25745 difference
25747 Difference factor where higher values mean more deblocking
25748 (default: 32).
25749 flatness
25751 Flatness threshold where lower values mean more deblocking
25752 (default: 39).
25753 va/vadeblock[|difference[|flatness]]
25755 Accurate vertical deblocking filter
25756 difference
25758 Difference factor where higher values mean more deblocking
25759 (default: 32).
25760 flatness
25762 Flatness threshold where lower values mean more deblocking
25763 (default: 39).
25764 The horizontal and vertical deblocking filters share the difference and
25766 flatness values so you cannot set different horizontal and vertical
25767 thresholds.
25768 h1/x1hdeblock
25770 Experimental horizontal deblocking filter
25771 v1/x1vdeblock
25773 Experimental vertical deblocking filter
25774 dr/dering
25776 Deringing filter
25777 tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise
25779 reducer
25780 threshold1
25781 larger -> stronger filtering
25782 threshold2
25784 larger -> stronger filtering
25785 threshold3
25787 larger -> stronger filtering
25788 al/autolevels[:f/fullyrange], automatic brightness / contrast
25790 correction
25791 f/fullyrange
25792 Stretch luminance to "0-255".
25793 lb/linblenddeint
25795 Linear blend deinterlacing filter that deinterlaces the given block
25796 by filtering all lines with a "(1 2 1)" filter.
25797 li/linipoldeint
25799 Linear interpolating deinterlacing filter that deinterlaces the
25800 given block by linearly interpolating every second line.
25801 ci/cubicipoldeint
25803 Cubic interpolating deinterlacing filter deinterlaces the given
25804 block by cubically interpolating every second line.
25805 md/mediandeint
25807 Median deinterlacing filter that deinterlaces the given block by
25808 applying a median filter to every second line.
25809 fd/ffmpegdeint
25811 FFmpeg deinterlacing filter that deinterlaces the given block by
25812 filtering every second line with a "(-1 4 2 4 -1)" filter.
25813 l5/lowpass5
25815 Vertically applied FIR lowpass deinterlacing filter that
25816 deinterlaces the given block by filtering all lines with a "(-1 2 6
25817 2 -1)" filter.
25818 fq/forceQuant[|quantizer]
25820 Overrides the quantizer table from the input with the constant
25821 quantizer you specify.
25822 quantizer
25824 Quantizer to use
25825 de/default
25827 Default pp filter combination ("hb|a,vb|a,dr|a")
25828 fa/fast
25830 Fast pp filter combination ("h1|a,v1|a,dr|a")
25831 ac High quality pp filter combination ("ha|a|128|7,va|a,dr|a")
25833 Examples
25835 • Apply horizontal and vertical deblocking, deringing and automatic
25837 brightness/contrast:
25838 pp=hb/vb/dr/al
25840 • Apply default filters without brightness/contrast correction:
25842 pp=de/-al
25844 • Apply default filters and temporal denoiser:
25846 pp=default/tmpnoise|1|2|3
25848 • Apply deblocking on luminance only, and switch vertical deblocking
25850 on or off automatically depending on available CPU time:
25851 pp=hb|y/vb|a
25853 pp7
25855 Apply Postprocessing filter 7. It is variant of the spp filter, similar
25856 to spp = 6 with 7 point DCT, where only the center sample is used after
25857 IDCT.
25858 The filter accepts the following options:
25860 qp Force a constant quantization parameter. It accepts an integer in
25862 range 0 to 63. If not set, the filter will use the QP from the
25863 video stream (if available).
25864 mode
25866 Set thresholding mode. Available modes are:
25867 hard
25869 Set hard thresholding.
25870 soft
25872 Set soft thresholding (better de-ringing effect, but likely
25873 blurrier).
25874 medium
25876 Set medium thresholding (good results, default).
25877 premultiply
25879 Apply alpha premultiply effect to input video stream using first plane
25880 of second stream as alpha.
25881 Both streams must have same dimensions and same pixel format.
25883 The filter accepts the following option:
25885 planes
25887 Set which planes will be processed, unprocessed planes will be
25888 copied. By default value 0xf, all planes will be processed.
25889 inplace
25891 Do not require 2nd input for processing, instead use alpha plane
25892 from input stream.
25893 prewitt
25895 Apply prewitt operator to input video stream.
25896 The filter accepts the following option:
25898 planes
25900 Set which planes will be processed, unprocessed planes will be
25901 copied. By default value 0xf, all planes will be processed.
25902 scale
25904 Set value which will be multiplied with filtered result.
25905 delta
25907 Set value which will be added to filtered result.
25908 Commands
25910 This filter supports the all above options as commands.
25912 pseudocolor
25914 Alter frame colors in video with pseudocolors.
25915 This filter accepts the following options:
25917 c0 set pixel first component expression
25919 c1 set pixel second component expression
25921 c2 set pixel third component expression
25923 c3 set pixel fourth component expression, corresponds to the alpha
25925 component
25926 index, i
25928 set component to use as base for altering colors
25929 preset, p
25931 Pick one of built-in LUTs. By default is set to none.
25932 Available LUTs:
25934 magma
25936 inferno
25937 plasma
25938 viridis
25939 turbo
25940 cividis
25941 range1
25942 range2
25943 shadows
25944 highlights
25945 solar
25946 nominal
25947 preferred
25948 total
25949 opacity
25950 Set opacity of output colors. Allowed range is from 0 to 1.
25951 Default value is set to 1.
25952 Each of the expression options specifies the expression to use for
25954 computing the lookup table for the corresponding pixel component
25955 values.
25956 The expressions can contain the following constants and functions:
25958 w
25960 h The input width and height.
25961 val The input value for the pixel component.
25963 ymin, umin, vmin, amin
25965 The minimum allowed component value.
25966 ymax, umax, vmax, amax
25968 The maximum allowed component value.
25969 All expressions default to "val".
25971 Commands
25973 This filter supports the all above options as commands.
25975 Examples
25977 • Change too high luma values to gradient:
25979 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'"
25981 psnr
25983 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
25984 Ratio) between two input videos.
25985 This filter takes in input two input videos, the first input is
25987 considered the "main" source and is passed unchanged to the output. The
25988 second input is used as a "reference" video for computing the PSNR.
25989 Both video inputs must have the same resolution and pixel format for
25991 this filter to work correctly. Also it assumes that both inputs have
25992 the same number of frames, which are compared one by one.
25993 The obtained average PSNR is printed through the logging system.
25995 The filter stores the accumulated MSE (mean squared error) of each
25997 frame, and at the end of the processing it is averaged across all
25998 frames equally, and the following formula is applied to obtain the
25999 PSNR:
26000 PSNR = 10*log10(MAX^2/MSE)
26002 Where MAX is the average of the maximum values of each component of the
26004 image.
26005 The description of the accepted parameters follows.
26007 stats_file, f
26009 If specified the filter will use the named file to save the PSNR of
26010 each individual frame. When filename equals "-" the data is sent to
26011 standard output.
26012 stats_version
26014 Specifies which version of the stats file format to use. Details of
26015 each format are written below. Default value is 1.
26016 stats_add_max
26018 Determines whether the max value is output to the stats log.
26019 Default value is 0. Requires stats_version >= 2. If this is set
26020 and stats_version < 2, the filter will return an error.
26021 This filter also supports the framesync options.
26023 The file printed if stats_file is selected, contains a sequence of
26025 key/value pairs of the form key:value for each compared couple of
26026 frames.
26027 If a stats_version greater than 1 is specified, a header line precedes
26029 the list of per-frame-pair stats, with key value pairs following the
26030 frame format with the following parameters:
26031 psnr_log_version
26033 The version of the log file format. Will match stats_version.
26034 fields
26036 A comma separated list of the per-frame-pair parameters included in
26037 the log.
26038 A description of each shown per-frame-pair parameter follows:
26040 n sequential number of the input frame, starting from 1
26042 mse_avg
26044 Mean Square Error pixel-by-pixel average difference of the compared
26045 frames, averaged over all the image components.
26046 mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
26048 Mean Square Error pixel-by-pixel average difference of the compared
26049 frames for the component specified by the suffix.
26050 psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
26052 Peak Signal to Noise ratio of the compared frames for the component
26053 specified by the suffix.
26054 max_avg, max_y, max_u, max_v
26056 Maximum allowed value for each channel, and average over all
26057 channels.
26058 Examples
26060 • For example:
26062 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
26064 [main][ref] psnr="stats_file=stats.log" [out]
26065 On this example the input file being processed is compared with the
26067 reference file ref_movie.mpg. The PSNR of each individual frame is
26068 stored in stats.log.
26069 • Another example with different containers:
26071 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 -
26073 pullup
26075 Pulldown reversal (inverse telecine) filter, capable of handling mixed
26076 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps
26077 progressive content.
26078 The pullup filter is designed to take advantage of future context in
26080 making its decisions. This filter is stateless in the sense that it
26081 does not lock onto a pattern to follow, but it instead looks forward to
26082 the following fields in order to identify matches and rebuild
26083 progressive frames.
26084 To produce content with an even framerate, insert the fps filter after
26086 pullup, use "fps=24000/1001" if the input frame rate is 29.97fps,
26087 "fps=24" for 30fps and the (rare) telecined 25fps input.
26088 The filter accepts the following options:
26090 jl
26092 jr
26093 jt
26094 jb These options set the amount of "junk" to ignore at the left,
26095 right, top, and bottom of the image, respectively. Left and right
26096 are in units of 8 pixels, while top and bottom are in units of 2
26097 lines. The default is 8 pixels on each side.
26098 sb Set the strict breaks. Setting this option to 1 will reduce the
26100 chances of filter generating an occasional mismatched frame, but it
26101 may also cause an excessive number of frames to be dropped during
26102 high motion sequences. Conversely, setting it to -1 will make
26103 filter match fields more easily. This may help processing of video
26104 where there is slight blurring between the fields, but may also
26105 cause there to be interlaced frames in the output. Default value
26106 is 0.
26107 mp Set the metric plane to use. It accepts the following values:
26109 l Use luma plane.
26111 u Use chroma blue plane.
26113 v Use chroma red plane.
26115 This option may be set to use chroma plane instead of the default
26117 luma plane for doing filter's computations. This may improve
26118 accuracy on very clean source material, but more likely will
26119 decrease accuracy, especially if there is chroma noise (rainbow
26120 effect) or any grayscale video. The main purpose of setting mp to
26121 a chroma plane is to reduce CPU load and make pullup usable in
26122 realtime on slow machines.
26123 For best results (without duplicated frames in the output file) it is
26125 necessary to change the output frame rate. For example, to inverse
26126 telecine NTSC input:
26127 ffmpeg -i input -vf pullup -r 24000/1001 ...
26129 qp
26131 Change video quantization parameters (QP).
26132 The filter accepts the following option:
26134 qp Set expression for quantization parameter.
26136 The expression is evaluated through the eval API and can contain, among
26138 others, the following constants:
26139 known
26141 1 if index is not 129, 0 otherwise.
26142 qp Sequential index starting from -129 to 128.
26144 Examples
26146 • Some equation like:
26148 qp=2+2*sin(PI*qp)
26150 random
26152 Flush video frames from internal cache of frames into a random order.
26153 No frame is discarded. Inspired by frei0r nervous filter.
26154 frames
26156 Set size in number of frames of internal cache, in range from 2 to
26157 512. Default is 30.
26158 seed
26160 Set seed for random number generator, must be an integer included
26161 between 0 and "UINT32_MAX". If not specified, or if explicitly set
26162 to less than 0, the filter will try to use a good random seed on a
26163 best effort basis.
26164 readeia608
26166 Read closed captioning (EIA-608) information from the top lines of a
26167 video frame.
26168 This filter adds frame metadata for "lavfi.readeia608.X.cc" and
26170 "lavfi.readeia608.X.line", where "X" is the number of the identified
26171 line with EIA-608 data (starting from 0). A description of each
26172 metadata value follows:
26173 lavfi.readeia608.X.cc
26175 The two bytes stored as EIA-608 data (printed in hexadecimal).
26176 lavfi.readeia608.X.line
26178 The number of the line on which the EIA-608 data was identified and
26179 read.
26180 This filter accepts the following options:
26182 scan_min
26184 Set the line to start scanning for EIA-608 data. Default is 0.
26185 scan_max
26187 Set the line to end scanning for EIA-608 data. Default is 29.
26188 spw Set the ratio of width reserved for sync code detection. Default
26190 is 0.27. Allowed range is "[0.1 - 0.7]".
26191 chp Enable checking the parity bit. In the event of a parity error, the
26193 filter will output 0x00 for that character. Default is false.
26194 lp Lowpass lines prior to further processing. Default is enabled.
26196 Commands
26198 This filter supports the all above options as commands.
26200 Examples
26202 • Output a csv with presentation time and the first two lines of
26204 identified EIA-608 captioning data.
26205 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
26207 readvitc
26209 Read vertical interval timecode (VITC) information from the top lines
26210 of a video frame.
26211 The filter adds frame metadata key "lavfi.readvitc.tc_str" with the
26213 timecode value, if a valid timecode has been detected. Further metadata
26214 key "lavfi.readvitc.found" is set to 0/1 depending on whether timecode
26215 data has been found or not.
26216 This filter accepts the following options:
26218 scan_max
26220 Set the maximum number of lines to scan for VITC data. If the value
26221 is set to "-1" the full video frame is scanned. Default is 45.
26222 thr_b
26224 Set the luma threshold for black. Accepts float numbers in the
26225 range [0.0,1.0], default value is 0.2. The value must be equal or
26226 less than "thr_w".
26227 thr_w
26229 Set the luma threshold for white. Accepts float numbers in the
26230 range [0.0,1.0], default value is 0.6. The value must be equal or
26231 greater than "thr_b".
26232 Examples
26234 • Detect and draw VITC data onto the video frame; if no valid VITC is
26236 detected, draw "--:--:--:--" as a placeholder:
26237 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--}:x=(w-tw)/2:y=400-ascent'
26239 remap
26241 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
26242 Destination pixel at position (X, Y) will be picked from source (x, y)
26244 position where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are
26245 out of range, zero value for pixel will be used for destination pixel.
26246 Xmap and Ymap input video streams must be of same dimensions. Output
26248 video stream will have Xmap/Ymap video stream dimensions. Xmap and
26249 Ymap input video streams are 16bit depth, single channel.
26250 format
26252 Specify pixel format of output from this filter. Can be "color" or
26253 "gray". Default is "color".
26254 fill
26256 Specify the color of the unmapped pixels. For the syntax of this
26257 option, check the "Color" section in the ffmpeg-utils manual.
26258 Default color is "black".
26259 removegrain
26261 The removegrain filter is a spatial denoiser for progressive video.
26262 m0 Set mode for the first plane.
26264 m1 Set mode for the second plane.
26266 m2 Set mode for the third plane.
26268 m3 Set mode for the fourth plane.
26270 Range of mode is from 0 to 24. Description of each mode follows:
26272 0 Leave input plane unchanged. Default.
26274 1 Clips the pixel with the minimum and maximum of the 8 neighbour
26276 pixels.
26277 2 Clips the pixel with the second minimum and maximum of the 8
26279 neighbour pixels.
26280 3 Clips the pixel with the third minimum and maximum of the 8
26282 neighbour pixels.
26283 4 Clips the pixel with the fourth minimum and maximum of the 8
26285 neighbour pixels. This is equivalent to a median filter.
26286 5 Line-sensitive clipping giving the minimal change.
26288 6 Line-sensitive clipping, intermediate.
26290 7 Line-sensitive clipping, intermediate.
26292 8 Line-sensitive clipping, intermediate.
26294 9 Line-sensitive clipping on a line where the neighbours pixels are
26296 the closest.
26297 10 Replaces the target pixel with the closest neighbour.
26299 11 [1 2 1] horizontal and vertical kernel blur.
26301 12 Same as mode 11.
26303 13 Bob mode, interpolates top field from the line where the neighbours
26305 pixels are the closest.
26306 14 Bob mode, interpolates bottom field from the line where the
26308 neighbours pixels are the closest.
26309 15 Bob mode, interpolates top field. Same as 13 but with a more
26311 complicated interpolation formula.
26312 16 Bob mode, interpolates bottom field. Same as 14 but with a more
26314 complicated interpolation formula.
26315 17 Clips the pixel with the minimum and maximum of respectively the
26317 maximum and minimum of each pair of opposite neighbour pixels.
26318 18 Line-sensitive clipping using opposite neighbours whose greatest
26320 distance from the current pixel is minimal.
26321 19 Replaces the pixel with the average of its 8 neighbours.
26323 20 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
26325 21 Clips pixels using the averages of opposite neighbour.
26327 22 Same as mode 21 but simpler and faster.
26329 23 Small edge and halo removal, but reputed useless.
26331 24 Similar as 23.
26333 removelogo
26335 Suppress a TV station logo, using an image file to determine which
26336 pixels comprise the logo. It works by filling in the pixels that
26337 comprise the logo with neighboring pixels.
26338 The filter accepts the following options:
26340 filename, f
26342 Set the filter bitmap file, which can be any image format supported
26343 by libavformat. The width and height of the image file must match
26344 those of the video stream being processed.
26345 Pixels in the provided bitmap image with a value of zero are not
26347 considered part of the logo, non-zero pixels are considered part of the
26348 logo. If you use white (255) for the logo and black (0) for the rest,
26349 you will be safe. For making the filter bitmap, it is recommended to
26350 take a screen capture of a black frame with the logo visible, and then
26351 using a threshold filter followed by the erode filter once or twice.
26352 If needed, little splotches can be fixed manually. Remember that if
26354 logo pixels are not covered, the filter quality will be much reduced.
26355 Marking too many pixels as part of the logo does not hurt as much, but
26356 it will increase the amount of blurring needed to cover over the image
26357 and will destroy more information than necessary, and extra pixels will
26358 slow things down on a large logo.
26359 repeatfields
26361 This filter uses the repeat_field flag from the Video ES headers and
26362 hard repeats fields based on its value.
26363 reverse
26365 Reverse a video clip.
26366 Warning: This filter requires memory to buffer the entire clip, so
26368 trimming is suggested.
26369 Examples
26371 • Take the first 5 seconds of a clip, and reverse it.
26373 trim=end=5,reverse
26375 rgbashift
26377 Shift R/G/B/A pixels horizontally and/or vertically.
26378 The filter accepts the following options:
26380 rh Set amount to shift red horizontally.
26382 rv Set amount to shift red vertically.
26384 gh Set amount to shift green horizontally.
26386 gv Set amount to shift green vertically.
26388 bh Set amount to shift blue horizontally.
26390 bv Set amount to shift blue vertically.
26392 ah Set amount to shift alpha horizontally.
26394 av Set amount to shift alpha vertically.
26396 edge
26398 Set edge mode, can be smear, default, or warp.
26399 Commands
26401 This filter supports the all above options as commands.
26403 roberts
26405 Apply roberts cross operator to input video stream.
26406 The filter accepts the following option:
26408 planes
26410 Set which planes will be processed, unprocessed planes will be
26411 copied. By default value 0xf, all planes will be processed.
26412 scale
26414 Set value which will be multiplied with filtered result.
26415 delta
26417 Set value which will be added to filtered result.
26418 Commands
26420 This filter supports the all above options as commands.
26422 rotate
26424 Rotate video by an arbitrary angle expressed in radians.
26425 The filter accepts the following options:
26427 A description of the optional parameters follows.
26429 angle, a
26431 Set an expression for the angle by which to rotate the input video
26432 clockwise, expressed as a number of radians. A negative value will
26433 result in a counter-clockwise rotation. By default it is set to
26434 "0".
26435 This expression is evaluated for each frame.
26437 out_w, ow
26439 Set the output width expression, default value is "iw". This
26440 expression is evaluated just once during configuration.
26441 out_h, oh
26443 Set the output height expression, default value is "ih". This
26444 expression is evaluated just once during configuration.
26445 bilinear
26447 Enable bilinear interpolation if set to 1, a value of 0 disables
26448 it. Default value is 1.
26449 fillcolor, c
26451 Set the color used to fill the output area not covered by the
26452 rotated image. For the general syntax of this option, check the
26453 "Color" section in the ffmpeg-utils manual. If the special value
26454 "none" is selected then no background is printed (useful for
26455 example if the background is never shown).
26456 Default value is "black".
26458 The expressions for the angle and the output size can contain the
26460 following constants and functions:
26461 n sequential number of the input frame, starting from 0. It is always
26463 NAN before the first frame is filtered.
26464 t time in seconds of the input frame, it is set to 0 when the filter
26466 is configured. It is always NAN before the first frame is filtered.
26467 hsub
26469 vsub
26470 horizontal and vertical chroma subsample values. For example for
26471 the pixel format "yuv422p" hsub is 2 and vsub is 1.
26472 in_w, iw
26474 in_h, ih
26475 the input video width and height
26476 out_w, ow
26478 out_h, oh
26479 the output width and height, that is the size of the padded area as
26480 specified by the width and height expressions
26481 rotw(a)
26483 roth(a)
26484 the minimal width/height required for completely containing the
26485 input video rotated by a radians.
26486 These are only available when computing the out_w and out_h
26488 expressions.
26489 Examples
26491 • Rotate the input by PI/6 radians clockwise:
26493 rotate=PI/6
26495 • Rotate the input by PI/6 radians counter-clockwise:
26497 rotate=-PI/6
26499 • Rotate the input by 45 degrees clockwise:
26501 rotate=45*PI/180
26503 • Apply a constant rotation with period T, starting from an angle of
26505 PI/3:
26506 rotate=PI/3+2*PI*t/T
26508 • Make the input video rotation oscillating with a period of T
26510 seconds and an amplitude of A radians:
26511 rotate=A*sin(2*PI/T*t)
26513 • Rotate the video, output size is chosen so that the whole rotating
26515 input video is always completely contained in the output:
26516 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
26518 • Rotate the video, reduce the output size so that no background is
26520 ever shown:
26521 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
26523 Commands
26525 The filter supports the following commands:
26527 a, angle
26529 Set the angle expression. The command accepts the same syntax of
26530 the corresponding option.
26531 If the specified expression is not valid, it is kept at its current
26533 value.
26534 sab
26536 Apply Shape Adaptive Blur.
26537 The filter accepts the following options:
26539 luma_radius, lr
26541 Set luma blur filter strength, must be a value in range 0.1-4.0,
26542 default value is 1.0. A greater value will result in a more blurred
26543 image, and in slower processing.
26544 luma_pre_filter_radius, lpfr
26546 Set luma pre-filter radius, must be a value in the 0.1-2.0 range,
26547 default value is 1.0.
26548 luma_strength, ls
26550 Set luma maximum difference between pixels to still be considered,
26551 must be a value in the 0.1-100.0 range, default value is 1.0.
26552 chroma_radius, cr
26554 Set chroma blur filter strength, must be a value in range -0.9-4.0.
26555 A greater value will result in a more blurred image, and in slower
26556 processing.
26557 chroma_pre_filter_radius, cpfr
26559 Set chroma pre-filter radius, must be a value in the -0.9-2.0
26560 range.
26561 chroma_strength, cs
26563 Set chroma maximum difference between pixels to still be
26564 considered, must be a value in the -0.9-100.0 range.
26565 Each chroma option value, if not explicitly specified, is set to the
26567 corresponding luma option value.
26568 scale
26570 Scale (resize) the input video, using the libswscale library.
26571 The scale filter forces the output display aspect ratio to be the same
26573 of the input, by changing the output sample aspect ratio.
26574 If the input image format is different from the format requested by the
26576 next filter, the scale filter will convert the input to the requested
26577 format.
26578 Options
26580 The filter accepts the following options, or any of the options
26582 supported by the libswscale scaler.
26583 See the ffmpeg-scaler manual for the complete list of scaler options.
26585 width, w
26587 height, h
26588 Set the output video dimension expression. Default value is the
26589 input dimension.
26590 If the width or w value is 0, the input width is used for the
26592 output. If the height or h value is 0, the input height is used for
26593 the output.
26594 If one and only one of the values is -n with n >= 1, the scale
26596 filter will use a value that maintains the aspect ratio of the
26597 input image, calculated from the other specified dimension. After
26598 that it will, however, make sure that the calculated dimension is
26599 divisible by n and adjust the value if necessary.
26600 If both values are -n with n >= 1, the behavior will be identical
26602 to both values being set to 0 as previously detailed.
26603 See below for the list of accepted constants for use in the
26605 dimension expression.
26606 eval
26608 Specify when to evaluate width and height expression. It accepts
26609 the following values:
26610 init
26612 Only evaluate expressions once during the filter initialization
26613 or when a command is processed.
26614 frame
26616 Evaluate expressions for each incoming frame.
26617 Default value is init.
26619 interl
26621 Set the interlacing mode. It accepts the following values:
26622 1 Force interlaced aware scaling.
26624 0 Do not apply interlaced scaling.
26626 -1 Select interlaced aware scaling depending on whether the source
26628 frames are flagged as interlaced or not.
26629 Default value is 0.
26631 flags
26633 Set libswscale scaling flags. See the ffmpeg-scaler manual for the
26634 complete list of values. If not explicitly specified the filter
26635 applies the default flags.
26636 param0, param1
26638 Set libswscale input parameters for scaling algorithms that need
26639 them. See the ffmpeg-scaler manual for the complete documentation.
26640 If not explicitly specified the filter applies empty parameters.
26641 size, s
26643 Set the video size. For the syntax of this option, check the "Video
26644 size" section in the ffmpeg-utils manual.
26645 in_color_matrix
26647 out_color_matrix
26648 Set in/output YCbCr color space type.
26649 This allows the autodetected value to be overridden as well as
26651 allows forcing a specific value used for the output and encoder.
26652 If not specified, the color space type depends on the pixel format.
26654 Possible values:
26656 auto
26658 Choose automatically.
26659 bt709
26661 Format conforming to International Telecommunication Union
26662 (ITU) Recommendation BT.709.
26663 fcc Set color space conforming to the United States Federal
26665 Communications Commission (FCC) Code of Federal Regulations
26666 (CFR) Title 47 (2003) 73.682 (a).
26667 bt601
26669 bt470
26670 smpte170m
26671 Set color space conforming to:
26672 • ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
26674 • ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
26676 • Society of Motion Picture and Television Engineers (SMPTE)
26678 ST 170:2004
26679 smpte240m
26681 Set color space conforming to SMPTE ST 240:1999.
26682 bt2020
26684 Set color space conforming to ITU-R BT.2020 non-constant
26685 luminance system.
26686 in_range
26688 out_range
26689 Set in/output YCbCr sample range.
26690 This allows the autodetected value to be overridden as well as
26692 allows forcing a specific value used for the output and encoder. If
26693 not specified, the range depends on the pixel format. Possible
26694 values:
26695 auto/unknown
26697 Choose automatically.
26698 jpeg/full/pc
26700 Set full range (0-255 in case of 8-bit luma).
26701 mpeg/limited/tv
26703 Set "MPEG" range (16-235 in case of 8-bit luma).
26704 force_original_aspect_ratio
26706 Enable decreasing or increasing output video width or height if
26707 necessary to keep the original aspect ratio. Possible values:
26708 disable
26710 Scale the video as specified and disable this feature.
26711 decrease
26713 The output video dimensions will automatically be decreased if
26714 needed.
26715 increase
26717 The output video dimensions will automatically be increased if
26718 needed.
26719 One useful instance of this option is that when you know a specific
26721 device's maximum allowed resolution, you can use this to limit the
26722 output video to that, while retaining the aspect ratio. For
26723 example, device A allows 1280x720 playback, and your video is
26724 1920x800. Using this option (set it to decrease) and specifying
26725 1280x720 to the command line makes the output 1280x533.
26726 Please note that this is a different thing than specifying -1 for w
26728 or h, you still need to specify the output resolution for this
26729 option to work.
26730 force_divisible_by
26732 Ensures that both the output dimensions, width and height, are
26733 divisible by the given integer when used together with
26734 force_original_aspect_ratio. This works similar to using "-n" in
26735 the w and h options.
26736 This option respects the value set for force_original_aspect_ratio,
26738 increasing or decreasing the resolution accordingly. The video's
26739 aspect ratio may be slightly modified.
26740 This option can be handy if you need to have a video fit within or
26742 exceed a defined resolution using force_original_aspect_ratio but
26743 also have encoder restrictions on width or height divisibility.
26744 The values of the w and h options are expressions containing the
26746 following constants:
26747 in_w
26749 in_h
26750 The input width and height
26751 iw
26753 ih These are the same as in_w and in_h.
26754 out_w
26756 out_h
26757 The output (scaled) width and height
26758 ow
26760 oh These are the same as out_w and out_h
26761 a The same as iw / ih
26763 sar input sample aspect ratio
26765 dar The input display aspect ratio. Calculated from "(iw / ih) * sar".
26767 hsub
26769 vsub
26770 horizontal and vertical input chroma subsample values. For example
26771 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
26772 ohsub
26774 ovsub
26775 horizontal and vertical output chroma subsample values. For example
26776 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
26777 n The (sequential) number of the input frame, starting from 0. Only
26779 available with "eval=frame".
26780 t The presentation timestamp of the input frame, expressed as a
26782 number of seconds. Only available with "eval=frame".
26783 pos The position (byte offset) of the frame in the input stream, or NaN
26785 if this information is unavailable and/or meaningless (for example
26786 in case of synthetic video). Only available with "eval=frame".
26787 Examples
26789 • Scale the input video to a size of 200x100
26791 scale=w=200:h=100
26793 This is equivalent to:
26795 scale=200:100
26797 or:
26799 scale=200x100
26801 • Specify a size abbreviation for the output size:
26803 scale=qcif
26805 which can also be written as:
26807 scale=size=qcif
26809 • Scale the input to 2x:
26811 scale=w=2*iw:h=2*ih
26813 • The above is the same as:
26815 scale=2*in_w:2*in_h
26817 • Scale the input to 2x with forced interlaced scaling:
26819 scale=2*iw:2*ih:interl=1
26821 • Scale the input to half size:
26823 scale=w=iw/2:h=ih/2
26825 • Increase the width, and set the height to the same size:
26827 scale=3/2*iw:ow
26829 • Seek Greek harmony:
26831 scale=iw:1/PHI*iw
26833 scale=ih*PHI:ih
26834 • Increase the height, and set the width to 3/2 of the height:
26836 scale=w=3/2*oh:h=3/5*ih
26838 • Increase the size, making the size a multiple of the chroma
26840 subsample values:
26841 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
26843 • Increase the width to a maximum of 500 pixels, keeping the same
26845 aspect ratio as the input:
26846 scale=w='min(500\, iw*3/2):h=-1'
26848 • Make pixels square by combining scale and setsar:
26850 scale='trunc(ih*dar):ih',setsar=1/1
26852 • Make pixels square by combining scale and setsar, making sure the
26854 resulting resolution is even (required by some codecs):
26855 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
26857 Commands
26859 This filter supports the following commands:
26861 width, w
26863 height, h
26864 Set the output video dimension expression. The command accepts the
26865 same syntax of the corresponding option.
26866 If the specified expression is not valid, it is kept at its current
26868 value.
26869 scale_cuda
26871 Scale (resize) and convert (pixel format) the input video, using
26872 accelerated CUDA kernels. Setting the output width and height works in
26873 the same way as for the scale filter.
26874 The filter accepts the following options:
26876 w
26878 h Set the output video dimension expression. Default value is the
26879 input dimension.
26880 Allows for the same expressions as the scale filter.
26882 interp_algo
26884 Sets the algorithm used for scaling:
26885 nearest
26887 Nearest neighbour
26888 Used by default if input parameters match the desired output.
26890 bilinear
26892 Bilinear
26893 bicubic
26895 Bicubic
26896 This is the default.
26898 lanczos
26900 Lanczos
26901 format
26903 Controls the output pixel format. By default, or if none is
26904 specified, the input pixel format is used.
26905 The filter does not support converting between YUV and RGB pixel
26907 formats.
26908 passthrough
26910 If set to 0, every frame is processed, even if no conversion is
26911 neccesary. This mode can be useful to use the filter as a buffer
26912 for a downstream frame-consumer that exhausts the limited decoder
26913 frame pool.
26914 If set to 1, frames are passed through as-is if they match the
26916 desired output parameters. This is the default behaviour.
26917 param
26919 Algorithm-Specific parameter.
26920 Affects the curves of the bicubic algorithm.
26922 force_original_aspect_ratio
26924 force_divisible_by
26925 Work the same as the identical scale filter options.
26926 Examples
26928 • Scale input to 720p, keeping aspect ratio and ensuring the output
26930 is yuv420p.
26931 scale_cuda=-2:720:format=yuv420p
26933 • Upscale to 4K using nearest neighbour algorithm.
26935 scale_cuda=4096:2160:interp_algo=nearest
26937 • Don't do any conversion or scaling, but copy all input frames into
26939 newly allocated ones. This can be useful to deal with a filter and
26940 encode chain that otherwise exhausts the decoders frame pool.
26941 scale_cuda=passthrough=0
26943 scale_npp
26945 Use the NVIDIA Performance Primitives (libnpp) to perform scaling
26946 and/or pixel format conversion on CUDA video frames. Setting the output
26947 width and height works in the same way as for the scale filter.
26948 The following additional options are accepted:
26950 format
26952 The pixel format of the output CUDA frames. If set to the string
26953 "same" (the default), the input format will be kept. Note that
26954 automatic format negotiation and conversion is not yet supported
26955 for hardware frames
26956 interp_algo
26958 The interpolation algorithm used for resizing. One of the
26959 following:
26960 nn Nearest neighbour.
26962 linear
26964 cubic
26965 cubic2p_bspline
26966 2-parameter cubic (B=1, C=0)
26967 cubic2p_catmullrom
26969 2-parameter cubic (B=0, C=1/2)
26970 cubic2p_b05c03
26972 2-parameter cubic (B=1/2, C=3/10)
26973 super
26975 Supersampling
26976 lanczos
26978 force_original_aspect_ratio
26979 Enable decreasing or increasing output video width or height if
26980 necessary to keep the original aspect ratio. Possible values:
26981 disable
26983 Scale the video as specified and disable this feature.
26984 decrease
26986 The output video dimensions will automatically be decreased if
26987 needed.
26988 increase
26990 The output video dimensions will automatically be increased if
26991 needed.
26992 One useful instance of this option is that when you know a specific
26994 device's maximum allowed resolution, you can use this to limit the
26995 output video to that, while retaining the aspect ratio. For
26996 example, device A allows 1280x720 playback, and your video is
26997 1920x800. Using this option (set it to decrease) and specifying
26998 1280x720 to the command line makes the output 1280x533.
26999 Please note that this is a different thing than specifying -1 for w
27001 or h, you still need to specify the output resolution for this
27002 option to work.
27003 force_divisible_by
27005 Ensures that both the output dimensions, width and height, are
27006 divisible by the given integer when used together with
27007 force_original_aspect_ratio. This works similar to using "-n" in
27008 the w and h options.
27009 This option respects the value set for force_original_aspect_ratio,
27011 increasing or decreasing the resolution accordingly. The video's
27012 aspect ratio may be slightly modified.
27013 This option can be handy if you need to have a video fit within or
27015 exceed a defined resolution using force_original_aspect_ratio but
27016 also have encoder restrictions on width or height divisibility.
27017 eval
27019 Specify when to evaluate width and height expression. It accepts
27020 the following values:
27021 init
27023 Only evaluate expressions once during the filter initialization
27024 or when a command is processed.
27025 frame
27027 Evaluate expressions for each incoming frame.
27028 The values of the w and h options are expressions containing the
27030 following constants:
27031 in_w
27033 in_h
27034 The input width and height
27035 iw
27037 ih These are the same as in_w and in_h.
27038 out_w
27040 out_h
27041 The output (scaled) width and height
27042 ow
27044 oh These are the same as out_w and out_h
27045 a The same as iw / ih
27047 sar input sample aspect ratio
27049 dar The input display aspect ratio. Calculated from "(iw / ih) * sar".
27051 n The (sequential) number of the input frame, starting from 0. Only
27053 available with "eval=frame".
27054 t The presentation timestamp of the input frame, expressed as a
27056 number of seconds. Only available with "eval=frame".
27057 pos The position (byte offset) of the frame in the input stream, or NaN
27059 if this information is unavailable and/or meaningless (for example
27060 in case of synthetic video). Only available with "eval=frame".
27061 scale2ref
27063 Scale (resize) the input video, based on a reference video.
27064 See the scale filter for available options, scale2ref supports the same
27066 but uses the reference video instead of the main input as basis.
27067 scale2ref also supports the following additional constants for the w
27068 and h options:
27069 main_w
27071 main_h
27072 The main input video's width and height
27073 main_a
27075 The same as main_w / main_h
27076 main_sar
27078 The main input video's sample aspect ratio
27079 main_dar, mdar
27081 The main input video's display aspect ratio. Calculated from
27082 "(main_w / main_h) * main_sar".
27083 main_hsub
27085 main_vsub
27086 The main input video's horizontal and vertical chroma subsample
27087 values. For example for the pixel format "yuv422p" hsub is 2 and
27088 vsub is 1.
27089 main_n
27091 The (sequential) number of the main input frame, starting from 0.
27092 Only available with "eval=frame".
27093 main_t
27095 The presentation timestamp of the main input frame, expressed as a
27096 number of seconds. Only available with "eval=frame".
27097 main_pos
27099 The position (byte offset) of the frame in the main input stream,
27100 or NaN if this information is unavailable and/or meaningless (for
27101 example in case of synthetic video). Only available with
27102 "eval=frame".
27103 Examples
27105 • Scale a subtitle stream (b) to match the main video (a) in size
27107 before overlaying
27108 'scale2ref[b][a];[a][b]overlay'
27110 • Scale a logo to 1/10th the height of a video, while preserving its
27112 display aspect ratio.
27113 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
27115 Commands
27117 This filter supports the following commands:
27119 width, w
27121 height, h
27122 Set the output video dimension expression. The command accepts the
27123 same syntax of the corresponding option.
27124 If the specified expression is not valid, it is kept at its current
27126 value.
27127 scale2ref_npp
27129 Use the NVIDIA Performance Primitives (libnpp) to scale (resize) the
27130 input video, based on a reference video.
27131 See the scale_npp filter for available options, scale2ref_npp supports
27133 the same but uses the reference video instead of the main input as
27134 basis. scale2ref_npp also supports the following additional constants
27135 for the w and h options:
27136 main_w
27138 main_h
27139 The main input video's width and height
27140 main_a
27142 The same as main_w / main_h
27143 main_sar
27145 The main input video's sample aspect ratio
27146 main_dar, mdar
27148 The main input video's display aspect ratio. Calculated from
27149 "(main_w / main_h) * main_sar".
27150 main_n
27152 The (sequential) number of the main input frame, starting from 0.
27153 Only available with "eval=frame".
27154 main_t
27156 The presentation timestamp of the main input frame, expressed as a
27157 number of seconds. Only available with "eval=frame".
27158 main_pos
27160 The position (byte offset) of the frame in the main input stream,
27161 or NaN if this information is unavailable and/or meaningless (for
27162 example in case of synthetic video). Only available with
27163 "eval=frame".
27164 Examples
27166 • Scale a subtitle stream (b) to match the main video (a) in size
27168 before overlaying
27169 'scale2ref_npp[b][a];[a][b]overlay_cuda'
27171 • Scale a logo to 1/10th the height of a video, while preserving its
27173 display aspect ratio.
27174 [logo-in][video-in]scale2ref_npp=w=oh*mdar:h=ih/10[logo-out][video-out]
27176 scharr
27178 Apply scharr operator to input video stream.
27179 The filter accepts the following option:
27181 planes
27183 Set which planes will be processed, unprocessed planes will be
27184 copied. By default value 0xf, all planes will be processed.
27185 scale
27187 Set value which will be multiplied with filtered result.
27188 delta
27190 Set value which will be added to filtered result.
27191 Commands
27193 This filter supports the all above options as commands.
27195 scroll
27197 Scroll input video horizontally and/or vertically by constant speed.
27198 The filter accepts the following options:
27200 horizontal, h
27202 Set the horizontal scrolling speed. Default is 0. Allowed range is
27203 from -1 to 1. Negative values changes scrolling direction.
27204 vertical, v
27206 Set the vertical scrolling speed. Default is 0. Allowed range is
27207 from -1 to 1. Negative values changes scrolling direction.
27208 hpos
27210 Set the initial horizontal scrolling position. Default is 0.
27211 Allowed range is from 0 to 1.
27212 vpos
27214 Set the initial vertical scrolling position. Default is 0. Allowed
27215 range is from 0 to 1.
27216 Commands
27218 This filter supports the following commands:
27220 horizontal, h
27222 Set the horizontal scrolling speed.
27223 vertical, v
27225 Set the vertical scrolling speed.
27226 scdet
27228 Detect video scene change.
27229 This filter sets frame metadata with mafd between frame, the scene
27231 score, and forward the frame to the next filter, so they can use these
27232 metadata to detect scene change or others.
27233 In addition, this filter logs a message and sets frame metadata when it
27235 detects a scene change by threshold.
27236 "lavfi.scd.mafd" metadata keys are set with mafd for every frame.
27238 "lavfi.scd.score" metadata keys are set with scene change score for
27240 every frame to detect scene change.
27241 "lavfi.scd.time" metadata keys are set with current filtered frame time
27243 which detect scene change with threshold.
27244 The filter accepts the following options:
27246 threshold, t
27248 Set the scene change detection threshold as a percentage of maximum
27249 change. Good values are in the "[8.0, 14.0]" range. The range for
27250 threshold is "[0., 100.]".
27251 Default value is 10..
27253 sc_pass, s
27255 Set the flag to pass scene change frames to the next filter.
27256 Default value is 0 You can enable it if you want to get snapshot of
27257 scene change frames only.
27258 selectivecolor
27260 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of
27261 colors (such as "reds", "yellows", "greens", "cyans", ...). The
27262 adjustment range is defined by the "purity" of the color (that is, how
27263 saturated it already is).
27264 This filter is similar to the Adobe Photoshop Selective Color tool.
27266 The filter accepts the following options:
27268 correction_method
27270 Select color correction method.
27271 Available values are:
27273 absolute
27275 Specified adjustments are applied "as-is" (added/subtracted to
27276 original pixel component value).
27277 relative
27279 Specified adjustments are relative to the original component
27280 value.
27281 Default is "absolute".
27283 reds
27285 Adjustments for red pixels (pixels where the red component is the
27286 maximum)
27287 yellows
27289 Adjustments for yellow pixels (pixels where the blue component is
27290 the minimum)
27291 greens
27293 Adjustments for green pixels (pixels where the green component is
27294 the maximum)
27295 cyans
27297 Adjustments for cyan pixels (pixels where the red component is the
27298 minimum)
27299 blues
27301 Adjustments for blue pixels (pixels where the blue component is the
27302 maximum)
27303 magentas
27305 Adjustments for magenta pixels (pixels where the green component is
27306 the minimum)
27307 whites
27309 Adjustments for white pixels (pixels where all components are
27310 greater than 128)
27311 neutrals
27313 Adjustments for all pixels except pure black and pure white
27314 blacks
27316 Adjustments for black pixels (pixels where all components are
27317 lesser than 128)
27318 psfile
27320 Specify a Photoshop selective color file (".asv") to import the
27321 settings from.
27322 All the adjustment settings (reds, yellows, ...) accept up to 4 space
27324 separated floating point adjustment values in the [-1,1] range,
27325 respectively to adjust the amount of cyan, magenta, yellow and black
27326 for the pixels of its range.
27327 Examples
27329 • Increase cyan by 50% and reduce yellow by 33% in every green areas,
27331 and increase magenta by 27% in blue areas:
27332 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
27334 • Use a Photoshop selective color preset:
27336 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
27338 separatefields
27340 The "separatefields" takes a frame-based video input and splits each
27341 frame into its components fields, producing a new half height clip with
27342 twice the frame rate and twice the frame count.
27343 This filter use field-dominance information in frame to decide which of
27345 each pair of fields to place first in the output. If it gets it wrong
27346 use setfield filter before "separatefields" filter.
27347 setdar, setsar
27349 The "setdar" filter sets the Display Aspect Ratio for the filter output
27350 video.
27351 This is done by changing the specified Sample (aka Pixel) Aspect Ratio,
27353 according to the following equation:
27354 <DAR> = <HORIZONTAL_RESOLUTION> / <VERTICAL_RESOLUTION> * <SAR>
27356 Keep in mind that the "setdar" filter does not modify the pixel
27358 dimensions of the video frame. Also, the display aspect ratio set by
27359 this filter may be changed by later filters in the filterchain, e.g. in
27360 case of scaling or if another "setdar" or a "setsar" filter is applied.
27361 The "setsar" filter sets the Sample (aka Pixel) Aspect Ratio for the
27363 filter output video.
27364 Note that as a consequence of the application of this filter, the
27366 output display aspect ratio will change according to the equation
27367 above.
27368 Keep in mind that the sample aspect ratio set by the "setsar" filter
27370 may be changed by later filters in the filterchain, e.g. if another
27371 "setsar" or a "setdar" filter is applied.
27372 It accepts the following parameters:
27374 r, ratio, dar ("setdar" only), sar ("setsar" only)
27376 Set the aspect ratio used by the filter.
27377 The parameter can be a floating point number string, an expression,
27379 or a string of the form num:den, where num and den are the
27380 numerator and denominator of the aspect ratio. If the parameter is
27381 not specified, it is assumed the value "0". In case the form
27382 "num:den" is used, the ":" character should be escaped.
27383 max Set the maximum integer value to use for expressing numerator and
27385 denominator when reducing the expressed aspect ratio to a rational.
27386 Default value is 100.
27387 The parameter sar is an expression containing the following constants:
27389 E, PI, PHI
27391 These are approximated values for the mathematical constants e
27392 (Euler's number), pi (Greek pi), and phi (the golden ratio).
27393 w, h
27395 The input width and height.
27396 a These are the same as w / h.
27398 sar The input sample aspect ratio.
27400 dar The input display aspect ratio. It is the same as (w / h) * sar.
27402 hsub, vsub
27404 Horizontal and vertical chroma subsample values. For example, for
27405 the pixel format "yuv422p" hsub is 2 and vsub is 1.
27406 Examples
27408 • To change the display aspect ratio to 16:9, specify one of the
27410 following:
27411 setdar=dar=1.77777
27413 setdar=dar=16/9
27414 • To change the sample aspect ratio to 10:11, specify:
27416 setsar=sar=10/11
27418 • To set a display aspect ratio of 16:9, and specify a maximum
27420 integer value of 1000 in the aspect ratio reduction, use the
27421 command:
27422 setdar=ratio=16/9:max=1000
27424 setfield
27426 Force field for the output video frame.
27427 The "setfield" filter marks the interlace type field for the output
27429 frames. It does not change the input frame, but only sets the
27430 corresponding property, which affects how the frame is treated by
27431 following filters (e.g. "fieldorder" or "yadif").
27432 The filter accepts the following options:
27434 mode
27436 Available values are:
27437 auto
27439 Keep the same field property.
27440 bff Mark the frame as bottom-field-first.
27442 tff Mark the frame as top-field-first.
27444 prog
27446 Mark the frame as progressive.
27447 setparams
27449 Force frame parameter for the output video frame.
27450 The "setparams" filter marks interlace and color range for the output
27452 frames. It does not change the input frame, but only sets the
27453 corresponding property, which affects how the frame is treated by
27454 filters/encoders.
27455 field_mode
27457 Available values are:
27458 auto
27460 Keep the same field property (default).
27461 bff Mark the frame as bottom-field-first.
27463 tff Mark the frame as top-field-first.
27465 prog
27467 Mark the frame as progressive.
27468 range
27470 Available values are:
27471 auto
27473 Keep the same color range property (default).
27474 unspecified, unknown
27476 Mark the frame as unspecified color range.
27477 limited, tv, mpeg
27479 Mark the frame as limited range.
27480 full, pc, jpeg
27482 Mark the frame as full range.
27483 color_primaries
27485 Set the color primaries. Available values are:
27486 auto
27488 Keep the same color primaries property (default).
27489 bt709
27491 unknown
27492 bt470m
27493 bt470bg
27494 smpte170m
27495 smpte240m
27496 film
27497 bt2020
27498 smpte428
27499 smpte431
27500 smpte432
27501 jedec-p22
27502 color_trc
27503 Set the color transfer. Available values are:
27504 auto
27506 Keep the same color trc property (default).
27507 bt709
27509 unknown
27510 bt470m
27511 bt470bg
27512 smpte170m
27513 smpte240m
27514 linear
27515 log100
27516 log316
27517 iec61966-2-4
27518 bt1361e
27519 iec61966-2-1
27520 bt2020-10
27521 bt2020-12
27522 smpte2084
27523 smpte428
27524 arib-std-b67
27525 colorspace
27526 Set the colorspace. Available values are:
27527 auto
27529 Keep the same colorspace property (default).
27530 gbr
27532 bt709
27533 unknown
27534 fcc
27535 bt470bg
27536 smpte170m
27537 smpte240m
27538 ycgco
27539 bt2020nc
27540 bt2020c
27541 smpte2085
27542 chroma-derived-nc
27543 chroma-derived-c
27544 ictcp
27545 sharpen_npp
27547 Use the NVIDIA Performance Primitives (libnpp) to perform image
27548 sharpening with border control.
27549 The following additional options are accepted:
27551 border_type
27553 Type of sampling to be used ad frame borders. One of the following:
27554 replicate
27556 Replicate pixel values.
27557 shear
27559 Apply shear transform to input video.
27560 This filter supports the following options:
27562 shx Shear factor in X-direction. Default value is 0. Allowed range is
27564 from -2 to 2.
27565 shy Shear factor in Y-direction. Default value is 0. Allowed range is
27567 from -2 to 2.
27568 fillcolor, c
27570 Set the color used to fill the output area not covered by the
27571 transformed video. For the general syntax of this option, check the
27572 "Color" section in the ffmpeg-utils manual. If the special value
27573 "none" is selected then no background is printed (useful for
27574 example if the background is never shown).
27575 Default value is "black".
27577 interp
27579 Set interpolation type. Can be "bilinear" or "nearest". Default is
27580 "bilinear".
27581 Commands
27583 This filter supports the all above options as commands.
27585 showinfo
27587 Show a line containing various information for each input video frame.
27588 The input video is not modified.
27589 This filter supports the following options:
27591 checksum
27593 Calculate checksums of each plane. By default enabled.
27594 The shown line contains a sequence of key/value pairs of the form
27596 key:value.
27597 The following values are shown in the output:
27599 n The (sequential) number of the input frame, starting from 0.
27601 pts The Presentation TimeStamp of the input frame, expressed as a
27603 number of time base units. The time base unit depends on the filter
27604 input pad.
27605 pts_time
27607 The Presentation TimeStamp of the input frame, expressed as a
27608 number of seconds.
27609 pos The position of the frame in the input stream, or -1 if this
27611 information is unavailable and/or meaningless (for example in case
27612 of synthetic video).
27613 fmt The pixel format name.
27615 sar The sample aspect ratio of the input frame, expressed in the form
27617 num/den.
27618 s The size of the input frame. For the syntax of this option, check
27620 the "Video size" section in the ffmpeg-utils manual.
27621 i The type of interlaced mode ("P" for "progressive", "T" for top
27623 field first, "B" for bottom field first).
27624 iskey
27626 This is 1 if the frame is a key frame, 0 otherwise.
27627 type
27629 The picture type of the input frame ("I" for an I-frame, "P" for a
27630 P-frame, "B" for a B-frame, or "?" for an unknown type). Also
27631 refer to the documentation of the "AVPictureType" enum and of the
27632 "av_get_picture_type_char" function defined in libavutil/avutil.h.
27633 checksum
27635 The Adler-32 checksum (printed in hexadecimal) of all the planes of
27636 the input frame.
27637 plane_checksum
27639 The Adler-32 checksum (printed in hexadecimal) of each plane of the
27640 input frame, expressed in the form "[c0 c1 c2 c3]".
27641 mean
27643 The mean value of pixels in each plane of the input frame,
27644 expressed in the form "[mean0 mean1 mean2 mean3]".
27645 stdev
27647 The standard deviation of pixel values in each plane of the input
27648 frame, expressed in the form "[stdev0 stdev1 stdev2 stdev3]".
27649 showpalette
27651 Displays the 256 colors palette of each frame. This filter is only
27652 relevant for pal8 pixel format frames.
27653 It accepts the following option:
27655 s Set the size of the box used to represent one palette color entry.
27657 Default is 30 (for a "30x30" pixel box).
27658 shuffleframes
27660 Reorder and/or duplicate and/or drop video frames.
27661 It accepts the following parameters:
27663 mapping
27665 Set the destination indexes of input frames. This is space or '|'
27666 separated list of indexes that maps input frames to output frames.
27667 Number of indexes also sets maximal value that each index may have.
27668 '-1' index have special meaning and that is to drop frame.
27669 The first frame has the index 0. The default is to keep the input
27671 unchanged.
27672 Examples
27674 • Swap second and third frame of every three frames of the input:
27676 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
27678 • Swap 10th and 1st frame of every ten frames of the input:
27680 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
27682 shufflepixels
27684 Reorder pixels in video frames.
27685 This filter accepts the following options:
27687 direction, d
27689 Set shuffle direction. Can be forward or inverse direction.
27690 Default direction is forward.
27691 mode, m
27693 Set shuffle mode. Can be horizontal, vertical or block mode.
27694 width, w
27696 height, h
27697 Set shuffle block_size. In case of horizontal shuffle mode only
27698 width part of size is used, and in case of vertical shuffle mode
27699 only height part of size is used.
27700 seed, s
27702 Set random seed used with shuffling pixels. Mainly useful to set to
27703 be able to reverse filtering process to get original input. For
27704 example, to reverse forward shuffle you need to use same parameters
27705 and exact same seed and to set direction to inverse.
27706 shuffleplanes
27708 Reorder and/or duplicate video planes.
27709 It accepts the following parameters:
27711 map0
27713 The index of the input plane to be used as the first output plane.
27714 map1
27716 The index of the input plane to be used as the second output plane.
27717 map2
27719 The index of the input plane to be used as the third output plane.
27720 map3
27722 The index of the input plane to be used as the fourth output plane.
27723 The first plane has the index 0. The default is to keep the input
27725 unchanged.
27726 Examples
27728 • Swap the second and third planes of the input:
27730 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
27732 signalstats
27734 Evaluate various visual metrics that assist in determining issues
27735 associated with the digitization of analog video media.
27736 By default the filter will log these metadata values:
27738 YMIN
27740 Display the minimal Y value contained within the input frame.
27741 Expressed in range of [0-255].
27742 YLOW
27744 Display the Y value at the 10% percentile within the input frame.
27745 Expressed in range of [0-255].
27746 YAVG
27748 Display the average Y value within the input frame. Expressed in
27749 range of [0-255].
27750 YHIGH
27752 Display the Y value at the 90% percentile within the input frame.
27753 Expressed in range of [0-255].
27754 YMAX
27756 Display the maximum Y value contained within the input frame.
27757 Expressed in range of [0-255].
27758 UMIN
27760 Display the minimal U value contained within the input frame.
27761 Expressed in range of [0-255].
27762 ULOW
27764 Display the U value at the 10% percentile within the input frame.
27765 Expressed in range of [0-255].
27766 UAVG
27768 Display the average U value within the input frame. Expressed in
27769 range of [0-255].
27770 UHIGH
27772 Display the U value at the 90% percentile within the input frame.
27773 Expressed in range of [0-255].
27774 UMAX
27776 Display the maximum U value contained within the input frame.
27777 Expressed in range of [0-255].
27778 VMIN
27780 Display the minimal V value contained within the input frame.
27781 Expressed in range of [0-255].
27782 VLOW
27784 Display the V value at the 10% percentile within the input frame.
27785 Expressed in range of [0-255].
27786 VAVG
27788 Display the average V value within the input frame. Expressed in
27789 range of [0-255].
27790 VHIGH
27792 Display the V value at the 90% percentile within the input frame.
27793 Expressed in range of [0-255].
27794 VMAX
27796 Display the maximum V value contained within the input frame.
27797 Expressed in range of [0-255].
27798 SATMIN
27800 Display the minimal saturation value contained within the input
27801 frame. Expressed in range of [0-~181.02].
27802 SATLOW
27804 Display the saturation value at the 10% percentile within the input
27805 frame. Expressed in range of [0-~181.02].
27806 SATAVG
27808 Display the average saturation value within the input frame.
27809 Expressed in range of [0-~181.02].
27810 SATHIGH
27812 Display the saturation value at the 90% percentile within the input
27813 frame. Expressed in range of [0-~181.02].
27814 SATMAX
27816 Display the maximum saturation value contained within the input
27817 frame. Expressed in range of [0-~181.02].
27818 HUEMED
27820 Display the median value for hue within the input frame. Expressed
27821 in range of [0-360].
27822 HUEAVG
27824 Display the average value for hue within the input frame. Expressed
27825 in range of [0-360].
27826 YDIF
27828 Display the average of sample value difference between all values
27829 of the Y plane in the current frame and corresponding values of the
27830 previous input frame. Expressed in range of [0-255].
27831 UDIF
27833 Display the average of sample value difference between all values
27834 of the U plane in the current frame and corresponding values of the
27835 previous input frame. Expressed in range of [0-255].
27836 VDIF
27838 Display the average of sample value difference between all values
27839 of the V plane in the current frame and corresponding values of the
27840 previous input frame. Expressed in range of [0-255].
27841 YBITDEPTH
27843 Display bit depth of Y plane in current frame. Expressed in range
27844 of [0-16].
27845 UBITDEPTH
27847 Display bit depth of U plane in current frame. Expressed in range
27848 of [0-16].
27849 VBITDEPTH
27851 Display bit depth of V plane in current frame. Expressed in range
27852 of [0-16].
27853 The filter accepts the following options:
27855 stat
27857 out stat specify an additional form of image analysis. out output
27858 video with the specified type of pixel highlighted.
27859 Both options accept the following values:
27861 tout
27863 Identify temporal outliers pixels. A temporal outlier is a
27864 pixel unlike the neighboring pixels of the same field. Examples
27865 of temporal outliers include the results of video dropouts,
27866 head clogs, or tape tracking issues.
27867 vrep
27869 Identify vertical line repetition. Vertical line repetition
27870 includes similar rows of pixels within a frame. In born-digital
27871 video vertical line repetition is common, but this pattern is
27872 uncommon in video digitized from an analog source. When it
27873 occurs in video that results from the digitization of an analog
27874 source it can indicate concealment from a dropout compensator.
27875 brng
27877 Identify pixels that fall outside of legal broadcast range.
27878 color, c
27880 Set the highlight color for the out option. The default color is
27881 yellow.
27882 Examples
27884 • Output data of various video metrics:
27886 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
27888 • Output specific data about the minimum and maximum values of the Y
27890 plane per frame:
27891 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
27893 • Playback video while highlighting pixels that are outside of
27895 broadcast range in red.
27896 ffplay example.mov -vf signalstats="out=brng:color=red"
27898 • Playback video with signalstats metadata drawn over the frame.
27900 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
27902 The contents of signalstat_drawtext.txt used in the command are:
27904 time %{pts:hms}
27906 Y (%{metadata:lavfi.signalstats.YMIN}-%{metadata:lavfi.signalstats.YMAX})
27907 U (%{metadata:lavfi.signalstats.UMIN}-%{metadata:lavfi.signalstats.UMAX})
27908 V (%{metadata:lavfi.signalstats.VMIN}-%{metadata:lavfi.signalstats.VMAX})
27909 saturation maximum: %{metadata:lavfi.signalstats.SATMAX}
27910 signature
27912 Calculates the MPEG-7 Video Signature. The filter can handle more than
27913 one input. In this case the matching between the inputs can be
27914 calculated additionally. The filter always passes through the first
27915 input. The signature of each stream can be written into a file.
27916 It accepts the following options:
27918 detectmode
27920 Enable or disable the matching process.
27921 Available values are:
27923 off Disable the calculation of a matching (default).
27925 full
27927 Calculate the matching for the whole video and output whether
27928 the whole video matches or only parts.
27929 fast
27931 Calculate only until a matching is found or the video ends.
27932 Should be faster in some cases.
27933 nb_inputs
27935 Set the number of inputs. The option value must be a non negative
27936 integer. Default value is 1.
27937 filename
27939 Set the path to which the output is written. If there is more than
27940 one input, the path must be a prototype, i.e. must contain %d or
27941 %0nd (where n is a positive integer), that will be replaced with
27942 the input number. If no filename is specified, no output will be
27943 written. This is the default.
27944 format
27946 Choose the output format.
27947 Available values are:
27949 binary
27951 Use the specified binary representation (default).
27952 xml Use the specified xml representation.
27954 th_d
27956 Set threshold to detect one word as similar. The option value must
27957 be an integer greater than zero. The default value is 9000.
27958 th_dc
27960 Set threshold to detect all words as similar. The option value must
27961 be an integer greater than zero. The default value is 60000.
27962 th_xh
27964 Set threshold to detect frames as similar. The option value must be
27965 an integer greater than zero. The default value is 116.
27966 th_di
27968 Set the minimum length of a sequence in frames to recognize it as
27969 matching sequence. The option value must be a non negative integer
27970 value. The default value is 0.
27971 th_it
27973 Set the minimum relation, that matching frames to all frames must
27974 have. The option value must be a double value between 0 and 1. The
27975 default value is 0.5.
27976 Examples
27978 • To calculate the signature of an input video and store it in
27980 signature.bin:
27981 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
27983 • To detect whether two videos match and store the signatures in XML
27985 format in signature0.xml and signature1.xml:
27986 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 -
27988 siti
27990 Calculate Spatial Info (SI) and Temporal Info (TI) scores for a video,
27991 as defined in ITU-T P.910: Subjective video quality assessment methods
27992 for multimedia applications. Available PDF at
27993 <https://www.itu.int/rec/T-REC-P.910-199909-S/en >.
27994 It accepts the following option:
27996 print_summary
27998 If set to 1, Summary statistics will be printed to the console.
27999 Default 0.
28000 Examples
28002 • To calculate SI/TI metrics and print summary:
28004 ffmpeg -i input.mp4 -vf siti=print_summary=1 -f null -
28006 smartblur
28008 Blur the input video without impacting the outlines.
28009 It accepts the following options:
28011 luma_radius, lr
28013 Set the luma radius. The option value must be a float number in the
28014 range [0.1,5.0] that specifies the variance of the gaussian filter
28015 used to blur the image (slower if larger). Default value is 1.0.
28016 luma_strength, ls
28018 Set the luma strength. The option value must be a float number in
28019 the range [-1.0,1.0] that configures the blurring. A value included
28020 in [0.0,1.0] will blur the image whereas a value included in
28021 [-1.0,0.0] will sharpen the image. Default value is 1.0.
28022 luma_threshold, lt
28024 Set the luma threshold used as a coefficient to determine whether a
28025 pixel should be blurred or not. The option value must be an integer
28026 in the range [-30,30]. A value of 0 will filter all the image, a
28027 value included in [0,30] will filter flat areas and a value
28028 included in [-30,0] will filter edges. Default value is 0.
28029 chroma_radius, cr
28031 Set the chroma radius. The option value must be a float number in
28032 the range [0.1,5.0] that specifies the variance of the gaussian
28033 filter used to blur the image (slower if larger). Default value is
28034 luma_radius.
28035 chroma_strength, cs
28037 Set the chroma strength. The option value must be a float number in
28038 the range [-1.0,1.0] that configures the blurring. A value included
28039 in [0.0,1.0] will blur the image whereas a value included in
28040 [-1.0,0.0] will sharpen the image. Default value is luma_strength.
28041 chroma_threshold, ct
28043 Set the chroma threshold used as a coefficient to determine whether
28044 a pixel should be blurred or not. The option value must be an
28045 integer in the range [-30,30]. A value of 0 will filter all the
28046 image, a value included in [0,30] will filter flat areas and a
28047 value included in [-30,0] will filter edges. Default value is
28048 luma_threshold.
28049 If a chroma option is not explicitly set, the corresponding luma value
28051 is set.
28052 sobel
28054 Apply sobel operator to input video stream.
28055 The filter accepts the following option:
28057 planes
28059 Set which planes will be processed, unprocessed planes will be
28060 copied. By default value 0xf, all planes will be processed.
28061 scale
28063 Set value which will be multiplied with filtered result.
28064 delta
28066 Set value which will be added to filtered result.
28067 Commands
28069 This filter supports the all above options as commands.
28071 spp
28073 Apply a simple postprocessing filter that compresses and decompresses
28074 the image at several (or - in the case of quality level 6 - all) shifts
28075 and average the results.
28076 The filter accepts the following options:
28078 quality
28080 Set quality. This option defines the number of levels for
28081 averaging. It accepts an integer in the range 0-6. If set to 0, the
28082 filter will have no effect. A value of 6 means the higher quality.
28083 For each increment of that value the speed drops by a factor of
28084 approximately 2. Default value is 3.
28085 qp Force a constant quantization parameter. If not set, the filter
28087 will use the QP from the video stream (if available).
28088 mode
28090 Set thresholding mode. Available modes are:
28091 hard
28093 Set hard thresholding (default).
28094 soft
28096 Set soft thresholding (better de-ringing effect, but likely
28097 blurrier).
28098 use_bframe_qp
28100 Enable the use of the QP from the B-Frames if set to 1. Using this
28101 option may cause flicker since the B-Frames have often larger QP.
28102 Default is 0 (not enabled).
28103 Commands
28105 This filter supports the following commands:
28107 quality, level
28109 Set quality level. The value "max" can be used to set the maximum
28110 level, currently 6.
28111 sr
28113 Scale the input by applying one of the super-resolution methods based
28114 on convolutional neural networks. Supported models:
28115 • Super-Resolution Convolutional Neural Network model (SRCNN). See
28117 <https://arxiv.org/abs/1501.00092>.
28118 • Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
28120 See <https://arxiv.org/abs/1609.05158>.
28121 Training scripts as well as scripts for model file (.pb) saving can be
28123 found at <https://github.com/XueweiMeng/sr/tree/sr_dnn_native>.
28124 Original repository is at
28125 <https://github.com/HighVoltageRocknRoll/sr.git>.
28126 Native model files (.model) can be generated from TensorFlow model
28128 files (.pb) by using tools/python/convert.py
28129 The filter accepts the following options:
28131 dnn_backend
28133 Specify which DNN backend to use for model loading and execution.
28134 This option accepts the following values:
28135 native
28137 Native implementation of DNN loading and execution.
28138 tensorflow
28140 TensorFlow backend. To enable this backend you need to install
28141 the TensorFlow for C library (see
28142 <https://www.tensorflow.org/install/lang_c>) and configure
28143 FFmpeg with "--enable-libtensorflow"
28144 Default value is native.
28146 model
28148 Set path to model file specifying network architecture and its
28149 parameters. Note that different backends use different file
28150 formats. TensorFlow backend can load files for both formats, while
28151 native backend can load files for only its format.
28152 scale_factor
28154 Set scale factor for SRCNN model. Allowed values are 2, 3 and 4.
28155 Default value is 2. Scale factor is necessary for SRCNN model,
28156 because it accepts input upscaled using bicubic upscaling with
28157 proper scale factor.
28158 To get full functionality (such as async execution), please use the
28160 dnn_processing filter.
28161 ssim
28163 Obtain the SSIM (Structural SImilarity Metric) between two input
28164 videos.
28165 This filter takes in input two input videos, the first input is
28167 considered the "main" source and is passed unchanged to the output. The
28168 second input is used as a "reference" video for computing the SSIM.
28169 Both video inputs must have the same resolution and pixel format for
28171 this filter to work correctly. Also it assumes that both inputs have
28172 the same number of frames, which are compared one by one.
28173 The filter stores the calculated SSIM of each frame.
28175 The description of the accepted parameters follows.
28177 stats_file, f
28179 If specified the filter will use the named file to save the SSIM of
28180 each individual frame. When filename equals "-" the data is sent to
28181 standard output.
28182 The file printed if stats_file is selected, contains a sequence of
28184 key/value pairs of the form key:value for each compared couple of
28185 frames.
28186 A description of each shown parameter follows:
28188 n sequential number of the input frame, starting from 1
28190 Y, U, V, R, G, B
28192 SSIM of the compared frames for the component specified by the
28193 suffix.
28194 All SSIM of the compared frames for the whole frame.
28196 dB Same as above but in dB representation.
28198 This filter also supports the framesync options.
28200 Examples
28202 • For example:
28204 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
28206 [main][ref] ssim="stats_file=stats.log" [out]
28207 On this example the input file being processed is compared with the
28209 reference file ref_movie.mpg. The SSIM of each individual frame is
28210 stored in stats.log.
28211 • Another example with both psnr and ssim at same time:
28213 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
28215 • Another example with different containers:
28217 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 -
28219 stereo3d
28221 Convert between different stereoscopic image formats.
28222 The filters accept the following options:
28224 in Set stereoscopic image format of input.
28226 Available values for input image formats are:
28228 sbsl
28230 side by side parallel (left eye left, right eye right)
28231 sbsr
28233 side by side crosseye (right eye left, left eye right)
28234 sbs2l
28236 side by side parallel with half width resolution (left eye
28237 left, right eye right)
28238 sbs2r
28240 side by side crosseye with half width resolution (right eye
28241 left, left eye right)
28242 abl
28244 tbl above-below (left eye above, right eye below)
28245 abr
28247 tbr above-below (right eye above, left eye below)
28248 ab2l
28250 tb2l
28251 above-below with half height resolution (left eye above, right
28252 eye below)
28253 ab2r
28255 tb2r
28256 above-below with half height resolution (right eye above, left
28257 eye below)
28258 al alternating frames (left eye first, right eye second)
28260 ar alternating frames (right eye first, left eye second)
28262 irl interleaved rows (left eye has top row, right eye starts on
28264 next row)
28265 irr interleaved rows (right eye has top row, left eye starts on
28267 next row)
28268 icl interleaved columns, left eye first
28270 icr interleaved columns, right eye first
28272 Default value is sbsl.
28274 out Set stereoscopic image format of output.
28276 sbsl
28278 side by side parallel (left eye left, right eye right)
28279 sbsr
28281 side by side crosseye (right eye left, left eye right)
28282 sbs2l
28284 side by side parallel with half width resolution (left eye
28285 left, right eye right)
28286 sbs2r
28288 side by side crosseye with half width resolution (right eye
28289 left, left eye right)
28290 abl
28292 tbl above-below (left eye above, right eye below)
28293 abr
28295 tbr above-below (right eye above, left eye below)
28296 ab2l
28298 tb2l
28299 above-below with half height resolution (left eye above, right
28300 eye below)
28301 ab2r
28303 tb2r
28304 above-below with half height resolution (right eye above, left
28305 eye below)
28306 al alternating frames (left eye first, right eye second)
28308 ar alternating frames (right eye first, left eye second)
28310 irl interleaved rows (left eye has top row, right eye starts on
28312 next row)
28313 irr interleaved rows (right eye has top row, left eye starts on
28315 next row)
28316 arbg
28318 anaglyph red/blue gray (red filter on left eye, blue filter on
28319 right eye)
28320 argg
28322 anaglyph red/green gray (red filter on left eye, green filter
28323 on right eye)
28324 arcg
28326 anaglyph red/cyan gray (red filter on left eye, cyan filter on
28327 right eye)
28328 arch
28330 anaglyph red/cyan half colored (red filter on left eye, cyan
28331 filter on right eye)
28332 arcc
28334 anaglyph red/cyan color (red filter on left eye, cyan filter on
28335 right eye)
28336 arcd
28338 anaglyph red/cyan color optimized with the least squares
28339 projection of dubois (red filter on left eye, cyan filter on
28340 right eye)
28341 agmg
28343 anaglyph green/magenta gray (green filter on left eye, magenta
28344 filter on right eye)
28345 agmh
28347 anaglyph green/magenta half colored (green filter on left eye,
28348 magenta filter on right eye)
28349 agmc
28351 anaglyph green/magenta colored (green filter on left eye,
28352 magenta filter on right eye)
28353 agmd
28355 anaglyph green/magenta color optimized with the least squares
28356 projection of dubois (green filter on left eye, magenta filter
28357 on right eye)
28358 aybg
28360 anaglyph yellow/blue gray (yellow filter on left eye, blue
28361 filter on right eye)
28362 aybh
28364 anaglyph yellow/blue half colored (yellow filter on left eye,
28365 blue filter on right eye)
28366 aybc
28368 anaglyph yellow/blue colored (yellow filter on left eye, blue
28369 filter on right eye)
28370 aybd
28372 anaglyph yellow/blue color optimized with the least squares
28373 projection of dubois (yellow filter on left eye, blue filter on
28374 right eye)
28375 ml mono output (left eye only)
28377 mr mono output (right eye only)
28379 chl checkerboard, left eye first
28381 chr checkerboard, right eye first
28383 icl interleaved columns, left eye first
28385 icr interleaved columns, right eye first
28387 hdmi
28389 HDMI frame pack
28390 Default value is arcd.
28392 Examples
28394 • Convert input video from side by side parallel to anaglyph
28396 yellow/blue dubois:
28397 stereo3d=sbsl:aybd
28399 • Convert input video from above below (left eye above, right eye
28401 below) to side by side crosseye.
28402 stereo3d=abl:sbsr
28404 streamselect, astreamselect
28406 Select video or audio streams.
28407 The filter accepts the following options:
28409 inputs
28411 Set number of inputs. Default is 2.
28412 map Set input indexes to remap to outputs.
28414 Commands
28416 The "streamselect" and "astreamselect" filter supports the following
28418 commands:
28419 map Set input indexes to remap to outputs.
28421 Examples
28423 • Select first 5 seconds 1st stream and rest of time 2nd stream:
28425 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
28427 • Same as above, but for audio:
28429 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
28431 subtitles
28433 Draw subtitles on top of input video using the libass library.
28434 To enable compilation of this filter you need to configure FFmpeg with
28436 "--enable-libass". This filter also requires a build with libavcodec
28437 and libavformat to convert the passed subtitles file to ASS (Advanced
28438 Substation Alpha) subtitles format.
28439 The filter accepts the following options:
28441 filename, f
28443 Set the filename of the subtitle file to read. It must be
28444 specified.
28445 original_size
28447 Specify the size of the original video, the video for which the ASS
28448 file was composed. For the syntax of this option, check the "Video
28449 size" section in the ffmpeg-utils manual. Due to a misdesign in
28450 ASS aspect ratio arithmetic, this is necessary to correctly scale
28451 the fonts if the aspect ratio has been changed.
28452 fontsdir
28454 Set a directory path containing fonts that can be used by the
28455 filter. These fonts will be used in addition to whatever the font
28456 provider uses.
28457 alpha
28459 Process alpha channel, by default alpha channel is untouched.
28460 charenc
28462 Set subtitles input character encoding. "subtitles" filter only.
28463 Only useful if not UTF-8.
28464 stream_index, si
28466 Set subtitles stream index. "subtitles" filter only.
28467 force_style
28469 Override default style or script info parameters of the subtitles.
28470 It accepts a string containing ASS style format "KEY=VALUE" couples
28471 separated by ",".
28472 If the first key is not specified, it is assumed that the first value
28474 specifies the filename.
28475 For example, to render the file sub.srt on top of the input video, use
28477 the command:
28478 subtitles=sub.srt
28480 which is equivalent to:
28482 subtitles=filename=sub.srt
28484 To render the default subtitles stream from file video.mkv, use:
28486 subtitles=video.mkv
28488 To render the second subtitles stream from that file, use:
28490 subtitles=video.mkv:si=1
28492 To make the subtitles stream from sub.srt appear in 80% transparent
28494 blue "DejaVu Serif", use:
28495 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
28497 super2xsai
28499 Scale the input by 2x and smooth using the Super2xSaI (Scale and
28500 Interpolate) pixel art scaling algorithm.
28501 Useful for enlarging pixel art images without reducing sharpness.
28503 swaprect
28505 Swap two rectangular objects in video.
28506 This filter accepts the following options:
28508 w Set object width.
28510 h Set object height.
28512 x1 Set 1st rect x coordinate.
28514 y1 Set 1st rect y coordinate.
28516 x2 Set 2nd rect x coordinate.
28518 y2 Set 2nd rect y coordinate.
28520 All expressions are evaluated once for each frame.
28522 The all options are expressions containing the following constants:
28524 w
28526 h The input width and height.
28527 a same as w / h
28529 sar input sample aspect ratio
28531 dar input display aspect ratio, it is the same as (w / h) * sar
28533 n The number of the input frame, starting from 0.
28535 t The timestamp expressed in seconds. It's NAN if the input timestamp
28537 is unknown.
28538 pos the position in the file of the input frame, NAN if unknown
28540 Commands
28542 This filter supports the all above options as commands.
28544 swapuv
28546 Swap U & V plane.
28547 tblend
28549 Blend successive video frames.
28550 See blend
28552 telecine
28554 Apply telecine process to the video.
28555 This filter accepts the following options:
28557 first_field
28559 top, t
28560 top field first
28561 bottom, b
28563 bottom field first The default value is "top".
28564 pattern
28566 A string of numbers representing the pulldown pattern you wish to
28567 apply. The default value is 23.
28568 Some typical patterns:
28570 NTSC output (30i):
28572 27.5p: 32222
28573 24p: 23 (classic)
28574 24p: 2332 (preferred)
28575 20p: 33
28576 18p: 334
28577 16p: 3444
28578 PAL output (25i):
28580 27.5p: 12222
28581 24p: 222222222223 ("Euro pulldown")
28582 16.67p: 33
28583 16p: 33333334
28584 thistogram
28586 Compute and draw a color distribution histogram for the input video
28587 across time.
28588 Unlike histogram video filter which only shows histogram of single
28590 input frame at certain time, this filter shows also past histograms of
28591 number of frames defined by "width" option.
28592 The computed histogram is a representation of the color component
28594 distribution in an image.
28595 The filter accepts the following options:
28597 width, w
28599 Set width of single color component output. Default value is 0.
28600 Value of 0 means width will be picked from input video. This also
28601 set number of passed histograms to keep. Allowed range is [0,
28602 8192].
28603 display_mode, d
28605 Set display mode. It accepts the following values:
28606 stack
28608 Per color component graphs are placed below each other.
28609 parade
28611 Per color component graphs are placed side by side.
28612 overlay
28614 Presents information identical to that in the "parade", except
28615 that the graphs representing color components are superimposed
28616 directly over one another.
28617 Default is "stack".
28619 levels_mode, m
28621 Set mode. Can be either "linear", or "logarithmic". Default is
28622 "linear".
28623 components, c
28625 Set what color components to display. Default is 7.
28626 bgopacity, b
28628 Set background opacity. Default is 0.9.
28629 envelope, e
28631 Show envelope. Default is disabled.
28632 ecolor, ec
28634 Set envelope color. Default is "gold".
28635 slide
28637 Set slide mode.
28638 Available values for slide is:
28640 frame
28642 Draw new frame when right border is reached.
28643 replace
28645 Replace old columns with new ones.
28646 scroll
28648 Scroll from right to left.
28649 rscroll
28651 Scroll from left to right.
28652 picture
28654 Draw single picture.
28655 Default is "replace".
28657 threshold
28659 Apply threshold effect to video stream.
28660 This filter needs four video streams to perform thresholding. First
28662 stream is stream we are filtering. Second stream is holding threshold
28663 values, third stream is holding min values, and last, fourth stream is
28664 holding max values.
28665 The filter accepts the following option:
28667 planes
28669 Set which planes will be processed, unprocessed planes will be
28670 copied. By default value 0xf, all planes will be processed.
28671 For example if first stream pixel's component value is less then
28673 threshold value of pixel component from 2nd threshold stream, third
28674 stream value will picked, otherwise fourth stream pixel component value
28675 will be picked.
28676 Using color source filter one can perform various types of
28678 thresholding:
28679 Commands
28681 This filter supports the all options as commands.
28683 Examples
28685 • Binary threshold, using gray color as threshold:
28687 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
28689 • Inverted binary threshold, using gray color as threshold:
28691 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
28693 • Truncate binary threshold, using gray color as threshold:
28695 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
28697 • Threshold to zero, using gray color as threshold:
28699 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
28701 • Inverted threshold to zero, using gray color as threshold:
28703 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
28705 thumbnail
28707 Select the most representative frame in a given sequence of consecutive
28708 frames.
28709 The filter accepts the following options:
28711 n Set the frames batch size to analyze; in a set of n frames, the
28713 filter will pick one of them, and then handle the next batch of n
28714 frames until the end. Default is 100.
28715 Since the filter keeps track of the whole frames sequence, a bigger n
28717 value will result in a higher memory usage, so a high value is not
28718 recommended.
28719 Examples
28721 • Extract one picture each 50 frames:
28723 thumbnail=50
28725 • Complete example of a thumbnail creation with ffmpeg:
28727 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
28729 tile
28731 Tile several successive frames together.
28732 The untile filter can do the reverse.
28734 The filter accepts the following options:
28736 layout
28738 Set the grid size in the form "COLUMNSxROWS". Range is upto
28739 UINT_MAX cells. Default is "6x5".
28740 nb_frames
28742 Set the maximum number of frames to render in the given area. It
28743 must be less than or equal to wxh. The default value is 0, meaning
28744 all the area will be used.
28745 margin
28747 Set the outer border margin in pixels. Range is 0 to 1024. Default
28748 is 0.
28749 padding
28751 Set the inner border thickness (i.e. the number of pixels between
28752 frames). For more advanced padding options (such as having
28753 different values for the edges), refer to the pad video filter.
28754 Range is 0 to 1024. Default is 0.
28755 color
28757 Specify the color of the unused area. For the syntax of this
28758 option, check the "Color" section in the ffmpeg-utils manual. The
28759 default value of color is "black".
28760 overlap
28762 Set the number of frames to overlap when tiling several successive
28763 frames together. The value must be between 0 and nb_frames - 1.
28764 Default is 0.
28765 init_padding
28767 Set the number of frames to initially be empty before displaying
28768 first output frame. This controls how soon will one get first
28769 output frame. The value must be between 0 and nb_frames - 1.
28770 Default is 0.
28771 Examples
28773 • Produce 8x8 PNG tiles of all keyframes (-skip_frame nokey) in a
28775 movie:
28776 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
28778 The -vsync 0 is necessary to prevent ffmpeg from duplicating each
28780 output frame to accommodate the originally detected frame rate.
28781 • Display 5 pictures in an area of "3x2" frames, with 7 pixels
28783 between them, and 2 pixels of initial margin, using mixed flat and
28784 named options:
28785 tile=3x2:nb_frames=5:padding=7:margin=2
28787 tinterlace
28789 Perform various types of temporal field interlacing.
28790 Frames are counted starting from 1, so the first input frame is
28792 considered odd.
28793 The filter accepts the following options:
28795 mode
28797 Specify the mode of the interlacing. This option can also be
28798 specified as a value alone. See below for a list of values for this
28799 option.
28800 Available values are:
28802 merge, 0
28804 Move odd frames into the upper field, even into the lower
28805 field, generating a double height frame at half frame rate.
28806 ------> time
28808 Input:
28809 Frame 1 Frame 2 Frame 3 Frame 4
28810 11111 22222 33333 44444
28812 11111 22222 33333 44444
28813 11111 22222 33333 44444
28814 11111 22222 33333 44444
28815 Output:
28817 11111 33333
28818 22222 44444
28819 11111 33333
28820 22222 44444
28821 11111 33333
28822 22222 44444
28823 11111 33333
28824 22222 44444
28825 drop_even, 1
28827 Only output odd frames, even frames are dropped, generating a
28828 frame with unchanged height at half frame rate.
28829 ------> time
28831 Input:
28832 Frame 1 Frame 2 Frame 3 Frame 4
28833 11111 22222 33333 44444
28835 11111 22222 33333 44444
28836 11111 22222 33333 44444
28837 11111 22222 33333 44444
28838 Output:
28840 11111 33333
28841 11111 33333
28842 11111 33333
28843 11111 33333
28844 drop_odd, 2
28846 Only output even frames, odd frames are dropped, generating a
28847 frame with unchanged height at half frame rate.
28848 ------> time
28850 Input:
28851 Frame 1 Frame 2 Frame 3 Frame 4
28852 11111 22222 33333 44444
28854 11111 22222 33333 44444
28855 11111 22222 33333 44444
28856 11111 22222 33333 44444
28857 Output:
28859 22222 44444
28860 22222 44444
28861 22222 44444
28862 22222 44444
28863 pad, 3
28865 Expand each frame to full height, but pad alternate lines with
28866 black, generating a frame with double height at the same input
28867 frame rate.
28868 ------> time
28870 Input:
28871 Frame 1 Frame 2 Frame 3 Frame 4
28872 11111 22222 33333 44444
28874 11111 22222 33333 44444
28875 11111 22222 33333 44444
28876 11111 22222 33333 44444
28877 Output:
28879 11111 ..... 33333 .....
28880 ..... 22222 ..... 44444
28881 11111 ..... 33333 .....
28882 ..... 22222 ..... 44444
28883 11111 ..... 33333 .....
28884 ..... 22222 ..... 44444
28885 11111 ..... 33333 .....
28886 ..... 22222 ..... 44444
28887 interleave_top, 4
28889 Interleave the upper field from odd frames with the lower field
28890 from even frames, generating a frame with unchanged height at
28891 half frame rate.
28892 ------> time
28894 Input:
28895 Frame 1 Frame 2 Frame 3 Frame 4
28896 11111<- 22222 33333<- 44444
28898 11111 22222<- 33333 44444<-
28899 11111<- 22222 33333<- 44444
28900 11111 22222<- 33333 44444<-
28901 Output:
28903 11111 33333
28904 22222 44444
28905 11111 33333
28906 22222 44444
28907 interleave_bottom, 5
28909 Interleave the lower field from odd frames with the upper field
28910 from even frames, generating a frame with unchanged height at
28911 half frame rate.
28912 ------> time
28914 Input:
28915 Frame 1 Frame 2 Frame 3 Frame 4
28916 11111 22222<- 33333 44444<-
28918 11111<- 22222 33333<- 44444
28919 11111 22222<- 33333 44444<-
28920 11111<- 22222 33333<- 44444
28921 Output:
28923 22222 44444
28924 11111 33333
28925 22222 44444
28926 11111 33333
28927 interlacex2, 6
28929 Double frame rate with unchanged height. Frames are inserted
28930 each containing the second temporal field from the previous
28931 input frame and the first temporal field from the next input
28932 frame. This mode relies on the top_field_first flag. Useful for
28933 interlaced video displays with no field synchronisation.
28934 ------> time
28936 Input:
28937 Frame 1 Frame 2 Frame 3 Frame 4
28938 11111 22222 33333 44444
28940 11111 22222 33333 44444
28941 11111 22222 33333 44444
28942 11111 22222 33333 44444
28943 Output:
28945 11111 22222 22222 33333 33333 44444 44444
28946 11111 11111 22222 22222 33333 33333 44444
28947 11111 22222 22222 33333 33333 44444 44444
28948 11111 11111 22222 22222 33333 33333 44444
28949 mergex2, 7
28951 Move odd frames into the upper field, even into the lower
28952 field, generating a double height frame at same frame rate.
28953 ------> time
28955 Input:
28956 Frame 1 Frame 2 Frame 3 Frame 4
28957 11111 22222 33333 44444
28959 11111 22222 33333 44444
28960 11111 22222 33333 44444
28961 11111 22222 33333 44444
28962 Output:
28964 11111 33333 33333 55555
28965 22222 22222 44444 44444
28966 11111 33333 33333 55555
28967 22222 22222 44444 44444
28968 11111 33333 33333 55555
28969 22222 22222 44444 44444
28970 11111 33333 33333 55555
28971 22222 22222 44444 44444
28972 Numeric values are deprecated but are accepted for backward
28974 compatibility reasons.
28975 Default mode is "merge".
28977 flags
28979 Specify flags influencing the filter process.
28980 Available value for flags is:
28982 low_pass_filter, vlpf
28984 Enable linear vertical low-pass filtering in the filter.
28985 Vertical low-pass filtering is required when creating an
28986 interlaced destination from a progressive source which contains
28987 high-frequency vertical detail. Filtering will reduce interlace
28988 'twitter' and Moire patterning.
28989 complex_filter, cvlpf
28991 Enable complex vertical low-pass filtering. This will slightly
28992 less reduce interlace 'twitter' and Moire patterning but better
28993 retain detail and subjective sharpness impression.
28994 bypass_il
28996 Bypass already interlaced frames, only adjust the frame rate.
28997 Vertical low-pass filtering and bypassing already interlaced frames
28999 can only be enabled for mode interleave_top and interleave_bottom.
29000 tmedian
29002 Pick median pixels from several successive input video frames.
29003 The filter accepts the following options:
29005 radius
29007 Set radius of median filter. Default is 1. Allowed range is from 1
29008 to 127.
29009 planes
29011 Set which planes to filter. Default value is 15, by which all
29012 planes are processed.
29013 percentile
29015 Set median percentile. Default value is 0.5. Default value of 0.5
29016 will pick always median values, while 0 will pick minimum values,
29017 and 1 maximum values.
29018 Commands
29020 This filter supports all above options as commands, excluding option
29022 "radius".
29023 tmidequalizer
29025 Apply Temporal Midway Video Equalization effect.
29026 Midway Video Equalization adjusts a sequence of video frames to have
29028 the same histograms, while maintaining their dynamics as much as
29029 possible. It's useful for e.g. matching exposures from a video frames
29030 sequence.
29031 This filter accepts the following option:
29033 radius
29035 Set filtering radius. Default is 5. Allowed range is from 1 to 127.
29036 sigma
29038 Set filtering sigma. Default is 0.5. This controls strength of
29039 filtering. Setting this option to 0 effectively does nothing.
29040 planes
29042 Set which planes to process. Default is 15, which is all available
29043 planes.
29044 tmix
29046 Mix successive video frames.
29047 A description of the accepted options follows.
29049 frames
29051 The number of successive frames to mix. If unspecified, it defaults
29052 to 3.
29053 weights
29055 Specify weight of each input video frame. Each weight is separated
29056 by space. If number of weights is smaller than number of frames
29057 last specified weight will be used for all remaining unset weights.
29058 scale
29060 Specify scale, if it is set it will be multiplied with sum of each
29061 weight multiplied with pixel values to give final destination pixel
29062 value. By default scale is auto scaled to sum of weights.
29063 planes
29065 Set which planes to filter. Default is all. Allowed range is from 0
29066 to 15.
29067 Examples
29069 • Average 7 successive frames:
29071 tmix=frames=7:weights="1 1 1 1 1 1 1"
29073 • Apply simple temporal convolution:
29075 tmix=frames=3:weights="-1 3 -1"
29077 • Similar as above but only showing temporal differences:
29079 tmix=frames=3:weights="-1 2 -1":scale=1
29081 Commands
29083 This filter supports the following commands:
29085 weights
29087 scale
29088 planes
29089 Syntax is same as option with same name.
29090 tonemap
29092 Tone map colors from different dynamic ranges.
29093 This filter expects data in single precision floating point, as it
29095 needs to operate on (and can output) out-of-range values. Another
29096 filter, such as zscale, is needed to convert the resulting frame to a
29097 usable format.
29098 The tonemapping algorithms implemented only work on linear light, so
29100 input data should be linearized beforehand (and possibly correctly
29101 tagged).
29102 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
29104 Options
29106 The filter accepts the following options.
29108 tonemap
29110 Set the tone map algorithm to use.
29111 Possible values are:
29113 none
29115 Do not apply any tone map, only desaturate overbright pixels.
29116 clip
29118 Hard-clip any out-of-range values. Use it for perfect color
29119 accuracy for in-range values, while distorting out-of-range
29120 values.
29121 linear
29123 Stretch the entire reference gamut to a linear multiple of the
29124 display.
29125 gamma
29127 Fit a logarithmic transfer between the tone curves.
29128 reinhard
29130 Preserve overall image brightness with a simple curve, using
29131 nonlinear contrast, which results in flattening details and
29132 degrading color accuracy.
29133 hable
29135 Preserve both dark and bright details better than reinhard, at
29136 the cost of slightly darkening everything. Use it when detail
29137 preservation is more important than color and brightness
29138 accuracy.
29139 mobius
29141 Smoothly map out-of-range values, while retaining contrast and
29142 colors for in-range material as much as possible. Use it when
29143 color accuracy is more important than detail preservation.
29144 Default is none.
29146 param
29148 Tune the tone mapping algorithm.
29149 This affects the following algorithms:
29151 none
29153 Ignored.
29154 linear
29156 Specifies the scale factor to use while stretching. Default to
29157 1.0.
29158 gamma
29160 Specifies the exponent of the function. Default to 1.8.
29161 clip
29163 Specify an extra linear coefficient to multiply into the signal
29164 before clipping. Default to 1.0.
29165 reinhard
29167 Specify the local contrast coefficient at the display peak.
29168 Default to 0.5, which means that in-gamut values will be about
29169 half as bright as when clipping.
29170 hable
29172 Ignored.
29173 mobius
29175 Specify the transition point from linear to mobius transform.
29176 Every value below this point is guaranteed to be mapped 1:1.
29177 The higher the value, the more accurate the result will be, at
29178 the cost of losing bright details. Default to 0.3, which due
29179 to the steep initial slope still preserves in-range colors
29180 fairly accurately.
29181 desat
29183 Apply desaturation for highlights that exceed this level of
29184 brightness. The higher the parameter, the more color information
29185 will be preserved. This setting helps prevent unnaturally blown-out
29186 colors for super-highlights, by (smoothly) turning into white
29187 instead. This makes images feel more natural, at the cost of
29188 reducing information about out-of-range colors.
29189 The default of 2.0 is somewhat conservative and will mostly just
29191 apply to skies or directly sunlit surfaces. A setting of 0.0
29192 disables this option.
29193 This option works only if the input frame has a supported color
29195 tag.
29196 peak
29198 Override signal/nominal/reference peak with this value. Useful when
29199 the embedded peak information in display metadata is not reliable
29200 or when tone mapping from a lower range to a higher range.
29201 tpad
29203 Temporarily pad video frames.
29204 The filter accepts the following options:
29206 start
29208 Specify number of delay frames before input video stream. Default
29209 is 0.
29210 stop
29212 Specify number of padding frames after input video stream. Set to
29213 -1 to pad indefinitely. Default is 0.
29214 start_mode
29216 Set kind of frames added to beginning of stream. Can be either add
29217 or clone. With add frames of solid-color are added. With clone
29218 frames are clones of first frame. Default is add.
29219 stop_mode
29221 Set kind of frames added to end of stream. Can be either add or
29222 clone. With add frames of solid-color are added. With clone
29223 frames are clones of last frame. Default is add.
29224 start_duration, stop_duration
29226 Specify the duration of the start/stop delay. See the Time duration
29227 section in the ffmpeg-utils(1) manual for the accepted syntax.
29228 These options override start and stop. Default is 0.
29229 color
29231 Specify the color of the padded area. For the syntax of this
29232 option, check the "Color" section in the ffmpeg-utils manual.
29233 The default value of color is "black".
29235 transpose
29237 Transpose rows with columns in the input video and optionally flip it.
29238 It accepts the following parameters:
29240 dir Specify the transposition direction.
29242 Can assume the following values:
29244 0, 4, cclock_flip
29246 Rotate by 90 degrees counterclockwise and vertically flip
29247 (default), that is:
29248 L.R L.l
29250 . . -> . .
29251 l.r R.r
29252 1, 5, clock
29254 Rotate by 90 degrees clockwise, that is:
29255 L.R l.L
29257 . . -> . .
29258 l.r r.R
29259 2, 6, cclock
29261 Rotate by 90 degrees counterclockwise, that is:
29262 L.R R.r
29264 . . -> . .
29265 l.r L.l
29266 3, 7, clock_flip
29268 Rotate by 90 degrees clockwise and vertically flip, that is:
29269 L.R r.R
29271 . . -> . .
29272 l.r l.L
29273 For values between 4-7, the transposition is only done if the input
29275 video geometry is portrait and not landscape. These values are
29276 deprecated, the "passthrough" option should be used instead.
29277 Numerical values are deprecated, and should be dropped in favor of
29279 symbolic constants.
29280 passthrough
29282 Do not apply the transposition if the input geometry matches the
29283 one specified by the specified value. It accepts the following
29284 values:
29285 none
29287 Always apply transposition.
29288 portrait
29290 Preserve portrait geometry (when height >= width).
29291 landscape
29293 Preserve landscape geometry (when width >= height).
29294 Default value is "none".
29296 For example to rotate by 90 degrees clockwise and preserve portrait
29298 layout:
29299 transpose=dir=1:passthrough=portrait
29301 The command above can also be specified as:
29303 transpose=1:portrait
29305 transpose_npp
29307 Transpose rows with columns in the input video and optionally flip it.
29308 For more in depth examples see the transpose video filter, which shares
29309 mostly the same options.
29310 It accepts the following parameters:
29312 dir Specify the transposition direction.
29314 Can assume the following values:
29316 cclock_flip
29318 Rotate by 90 degrees counterclockwise and vertically flip.
29319 (default)
29320 clock
29322 Rotate by 90 degrees clockwise.
29323 cclock
29325 Rotate by 90 degrees counterclockwise.
29326 clock_flip
29328 Rotate by 90 degrees clockwise and vertically flip.
29329 passthrough
29331 Do not apply the transposition if the input geometry matches the
29332 one specified by the specified value. It accepts the following
29333 values:
29334 none
29336 Always apply transposition. (default)
29337 portrait
29339 Preserve portrait geometry (when height >= width).
29340 landscape
29342 Preserve landscape geometry (when width >= height).
29343 trim
29345 Trim the input so that the output contains one continuous subpart of
29346 the input.
29347 It accepts the following parameters:
29349 start
29351 Specify the time of the start of the kept section, i.e. the frame
29352 with the timestamp start will be the first frame in the output.
29353 end Specify the time of the first frame that will be dropped, i.e. the
29355 frame immediately preceding the one with the timestamp end will be
29356 the last frame in the output.
29357 start_pts
29359 This is the same as start, except this option sets the start
29360 timestamp in timebase units instead of seconds.
29361 end_pts
29363 This is the same as end, except this option sets the end timestamp
29364 in timebase units instead of seconds.
29365 duration
29367 The maximum duration of the output in seconds.
29368 start_frame
29370 The number of the first frame that should be passed to the output.
29371 end_frame
29373 The number of the first frame that should be dropped.
29374 start, end, and duration are expressed as time duration specifications;
29376 see the Time duration section in the ffmpeg-utils(1) manual for the
29377 accepted syntax.
29378 Note that the first two sets of the start/end options and the duration
29380 option look at the frame timestamp, while the _frame variants simply
29381 count the frames that pass through the filter. Also note that this
29382 filter does not modify the timestamps. If you wish for the output
29383 timestamps to start at zero, insert a setpts filter after the trim
29384 filter.
29385 If multiple start or end options are set, this filter tries to be
29387 greedy and keep all the frames that match at least one of the specified
29388 constraints. To keep only the part that matches all the constraints at
29389 once, chain multiple trim filters.
29390 The defaults are such that all the input is kept. So it is possible to
29392 set e.g. just the end values to keep everything before the specified
29393 time.
29394 Examples:
29396 • Drop everything except the second minute of input:
29398 ffmpeg -i INPUT -vf trim=60:120
29400 • Keep only the first second:
29402 ffmpeg -i INPUT -vf trim=duration=1
29404 unpremultiply
29406 Apply alpha unpremultiply effect to input video stream using first
29407 plane of second stream as alpha.
29408 Both streams must have same dimensions and same pixel format.
29410 The filter accepts the following option:
29412 planes
29414 Set which planes will be processed, unprocessed planes will be
29415 copied. By default value 0xf, all planes will be processed.
29416 If the format has 1 or 2 components, then luma is bit 0. If the
29418 format has 3 or 4 components: for RGB formats bit 0 is green, bit 1
29419 is blue and bit 2 is red; for YUV formats bit 0 is luma, bit 1 is
29420 chroma-U and bit 2 is chroma-V. If present, the alpha channel is
29421 always the last bit.
29422 inplace
29424 Do not require 2nd input for processing, instead use alpha plane
29425 from input stream.
29426 unsharp
29428 Sharpen or blur the input video.
29429 It accepts the following parameters:
29431 luma_msize_x, lx
29433 Set the luma matrix horizontal size. It must be an odd integer
29434 between 3 and 23. The default value is 5.
29435 luma_msize_y, ly
29437 Set the luma matrix vertical size. It must be an odd integer
29438 between 3 and 23. The default value is 5.
29439 luma_amount, la
29441 Set the luma effect strength. It must be a floating point number,
29442 reasonable values lay between -1.5 and 1.5.
29443 Negative values will blur the input video, while positive values
29445 will sharpen it, a value of zero will disable the effect.
29446 Default value is 1.0.
29448 chroma_msize_x, cx
29450 Set the chroma matrix horizontal size. It must be an odd integer
29451 between 3 and 23. The default value is 5.
29452 chroma_msize_y, cy
29454 Set the chroma matrix vertical size. It must be an odd integer
29455 between 3 and 23. The default value is 5.
29456 chroma_amount, ca
29458 Set the chroma effect strength. It must be a floating point number,
29459 reasonable values lay between -1.5 and 1.5.
29460 Negative values will blur the input video, while positive values
29462 will sharpen it, a value of zero will disable the effect.
29463 Default value is 0.0.
29465 alpha_msize_x, ax
29467 Set the alpha matrix horizontal size. It must be an odd integer
29468 between 3 and 23. The default value is 5.
29469 alpha_msize_y, ay
29471 Set the alpha matrix vertical size. It must be an odd integer
29472 between 3 and 23. The default value is 5.
29473 alpha_amount, aa
29475 Set the alpha effect strength. It must be a floating point number,
29476 reasonable values lay between -1.5 and 1.5.
29477 Negative values will blur the input video, while positive values
29479 will sharpen it, a value of zero will disable the effect.
29480 Default value is 0.0.
29482 All parameters are optional and default to the equivalent of the string
29484 '5:5:1.0:5:5:0.0'.
29485 Examples
29487 • Apply strong luma sharpen effect:
29489 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
29491 • Apply a strong blur of both luma and chroma parameters:
29493 unsharp=7:7:-2:7:7:-2
29495 untile
29497 Decompose a video made of tiled images into the individual images.
29498 The frame rate of the output video is the frame rate of the input video
29500 multiplied by the number of tiles.
29501 This filter does the reverse of tile.
29503 The filter accepts the following options:
29505 layout
29507 Set the grid size (i.e. the number of lines and columns). For the
29508 syntax of this option, check the "Video size" section in the
29509 ffmpeg-utils manual.
29510 Examples
29512 • Produce a 1-second video from a still image file made of 25 frames
29514 stacked vertically, like an analogic film reel:
29515 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
29517 uspp
29519 Apply ultra slow/simple postprocessing filter that compresses and
29520 decompresses the image at several (or - in the case of quality level 8
29521 - all) shifts and average the results.
29522 The way this differs from the behavior of spp is that uspp actually
29524 encodes & decodes each case with libavcodec Snow, whereas spp uses a
29525 simplified intra only 8x8 DCT similar to MJPEG.
29526 This filter is only available in ffmpeg version 4.4 or earlier.
29528 The filter accepts the following options:
29530 quality
29532 Set quality. This option defines the number of levels for
29533 averaging. It accepts an integer in the range 0-8. If set to 0, the
29534 filter will have no effect. A value of 8 means the higher quality.
29535 For each increment of that value the speed drops by a factor of
29536 approximately 2. Default value is 3.
29537 qp Force a constant quantization parameter. If not set, the filter
29539 will use the QP from the video stream (if available).
29540 v360
29542 Convert 360 videos between various formats.
29543 The filter accepts the following options:
29545 input
29547 output
29548 Set format of the input/output video.
29549 Available formats:
29551 e
29553 equirect
29554 Equirectangular projection.
29555 c3x2
29557 c6x1
29558 c1x6
29559 Cubemap with 3x2/6x1/1x6 layout.
29560 Format specific options:
29562 in_pad
29564 out_pad
29565 Set padding proportion for the input/output cubemap. Values
29566 in decimals.
29567 Example values:
29569 0 No padding.
29571 0.01
29573 1% of face is padding. For example, with 1920x1280
29574 resolution face size would be 640x640 and padding would
29575 be 3 pixels from each side. (640 * 0.01 = 6 pixels)
29576 Default value is @samp{0}. Maximum value is @samp{0.1}.
29578 fin_pad
29580 fout_pad
29581 Set fixed padding for the input/output cubemap. Values in
29582 pixels.
29583 Default value is @samp{0}. If greater than zero it
29585 overrides other padding options.
29586 in_forder
29588 out_forder
29589 Set order of faces for the input/output cubemap. Choose one
29590 direction for each position.
29591 Designation of directions:
29593 r right
29595 l left
29597 u up
29599 d down
29601 f forward
29603 b back
29605 Default value is @samp{rludfb}.
29607 in_frot
29609 out_frot
29610 Set rotation of faces for the input/output cubemap. Choose
29611 one angle for each position.
29612 Designation of angles:
29614 0 0 degrees clockwise
29616 1 90 degrees clockwise
29618 2 180 degrees clockwise
29620 3 270 degrees clockwise
29622 Default value is @samp{000000}.
29624 eac Equi-Angular Cubemap.
29626 flat
29628 gnomonic
29629 rectilinear
29630 Regular video.
29631 Format specific options:
29633 h_fov
29635 v_fov
29636 d_fov
29637 Set output horizontal/vertical/diagonal field of view.
29638 Values in degrees.
29639 If diagonal field of view is set it overrides horizontal
29641 and vertical field of view.
29642 ih_fov
29644 iv_fov
29645 id_fov
29646 Set input horizontal/vertical/diagonal field of view.
29647 Values in degrees.
29648 If diagonal field of view is set it overrides horizontal
29650 and vertical field of view.
29651 dfisheye
29653 Dual fisheye.
29654 Format specific options:
29656 h_fov
29658 v_fov
29659 d_fov
29660 Set output horizontal/vertical/diagonal field of view.
29661 Values in degrees.
29662 If diagonal field of view is set it overrides horizontal
29664 and vertical field of view.
29665 ih_fov
29667 iv_fov
29668 id_fov
29669 Set input horizontal/vertical/diagonal field of view.
29670 Values in degrees.
29671 If diagonal field of view is set it overrides horizontal
29673 and vertical field of view.
29674 barrel
29676 fb
29677 barrelsplit
29678 Facebook's 360 formats.
29679 sg Stereographic format.
29681 Format specific options:
29683 h_fov
29685 v_fov
29686 d_fov
29687 Set output horizontal/vertical/diagonal field of view.
29688 Values in degrees.
29689 If diagonal field of view is set it overrides horizontal
29691 and vertical field of view.
29692 ih_fov
29694 iv_fov
29695 id_fov
29696 Set input horizontal/vertical/diagonal field of view.
29697 Values in degrees.
29698 If diagonal field of view is set it overrides horizontal
29700 and vertical field of view.
29701 mercator
29703 Mercator format.
29704 ball
29706 Ball format, gives significant distortion toward the back.
29707 hammer
29709 Hammer-Aitoff map projection format.
29710 sinusoidal
29712 Sinusoidal map projection format.
29713 fisheye
29715 Fisheye projection.
29716 Format specific options:
29718 h_fov
29720 v_fov
29721 d_fov
29722 Set output horizontal/vertical/diagonal field of view.
29723 Values in degrees.
29724 If diagonal field of view is set it overrides horizontal
29726 and vertical field of view.
29727 ih_fov
29729 iv_fov
29730 id_fov
29731 Set input horizontal/vertical/diagonal field of view.
29732 Values in degrees.
29733 If diagonal field of view is set it overrides horizontal
29735 and vertical field of view.
29736 pannini
29738 Pannini projection.
29739 Format specific options:
29741 h_fov
29743 Set output pannini parameter.
29744 ih_fov
29746 Set input pannini parameter.
29747 cylindrical
29749 Cylindrical projection.
29750 Format specific options:
29752 h_fov
29754 v_fov
29755 d_fov
29756 Set output horizontal/vertical/diagonal field of view.
29757 Values in degrees.
29758 If diagonal field of view is set it overrides horizontal
29760 and vertical field of view.
29761 ih_fov
29763 iv_fov
29764 id_fov
29765 Set input horizontal/vertical/diagonal field of view.
29766 Values in degrees.
29767 If diagonal field of view is set it overrides horizontal
29769 and vertical field of view.
29770 perspective
29772 Perspective projection. (output only)
29773 Format specific options:
29775 v_fov
29777 Set perspective parameter.
29778 tetrahedron
29780 Tetrahedron projection.
29781 tsp Truncated square pyramid projection.
29783 he
29785 hequirect
29786 Half equirectangular projection.
29787 equisolid
29789 Equisolid format.
29790 Format specific options:
29792 h_fov
29794 v_fov
29795 d_fov
29796 Set output horizontal/vertical/diagonal field of view.
29797 Values in degrees.
29798 If diagonal field of view is set it overrides horizontal
29800 and vertical field of view.
29801 ih_fov
29803 iv_fov
29804 id_fov
29805 Set input horizontal/vertical/diagonal field of view.
29806 Values in degrees.
29807 If diagonal field of view is set it overrides horizontal
29809 and vertical field of view.
29810 og Orthographic format.
29812 Format specific options:
29814 h_fov
29816 v_fov
29817 d_fov
29818 Set output horizontal/vertical/diagonal field of view.
29819 Values in degrees.
29820 If diagonal field of view is set it overrides horizontal
29822 and vertical field of view.
29823 ih_fov
29825 iv_fov
29826 id_fov
29827 Set input horizontal/vertical/diagonal field of view.
29828 Values in degrees.
29829 If diagonal field of view is set it overrides horizontal
29831 and vertical field of view.
29832 octahedron
29834 Octahedron projection.
29835 cylindricalea
29837 Cylindrical Equal Area projection.
29838 interp
29840 Set interpolation method.Note: more complex interpolation methods
29841 require much more memory to run.
29842 Available methods:
29844 near
29846 nearest
29847 Nearest neighbour.
29848 line
29850 linear
29851 Bilinear interpolation.
29852 lagrange9
29854 Lagrange9 interpolation.
29855 cube
29857 cubic
29858 Bicubic interpolation.
29859 lanc
29861 lanczos
29862 Lanczos interpolation.
29863 sp16
29865 spline16
29866 Spline16 interpolation.
29867 gauss
29869 gaussian
29870 Gaussian interpolation.
29871 mitchell
29873 Mitchell interpolation.
29874 Default value is @samp{line}.
29876 w
29878 h Set the output video resolution.
29879 Default resolution depends on formats.
29881 in_stereo
29883 out_stereo
29884 Set the input/output stereo format.
29885 2d 2D mono
29887 sbs Side by side
29889 tb Top bottom
29891 Default value is @samp{2d} for input and output format.
29893 yaw
29895 pitch
29896 roll
29897 Set rotation for the output video. Values in degrees.
29898 rorder
29900 Set rotation order for the output video. Choose one item for each
29901 position.
29902 y, Y
29904 yaw
29905 p, P
29907 pitch
29908 r, R
29910 roll
29911 Default value is @samp{ypr}.
29913 h_flip
29915 v_flip
29916 d_flip
29917 Flip the output video horizontally(swaps
29918 left-right)/vertically(swaps up-down)/in-depth(swaps back-forward).
29919 Boolean values.
29920 ih_flip
29922 iv_flip
29923 Set if input video is flipped horizontally/vertically. Boolean
29924 values.
29925 in_trans
29927 Set if input video is transposed. Boolean value, by default
29928 disabled.
29929 out_trans
29931 Set if output video needs to be transposed. Boolean value, by
29932 default disabled.
29933 h_offset
29935 v_offset
29936 Set output horizontal/vertical off-axis offset. Default is set to
29937 0. Allowed range is from -1 to 1.
29938 alpha_mask
29940 Build mask in alpha plane for all unmapped pixels by marking them
29941 fully transparent. Boolean value, by default disabled.
29942 reset_rot
29944 Reset rotation of output video. Boolean value, by default disabled.
29945 Examples
29947 • Convert equirectangular video to cubemap with 3x2 layout and 1%
29949 padding using bicubic interpolation:
29950 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
29952 • Extract back view of Equi-Angular Cubemap:
29954 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
29956 • Convert transposed and horizontally flipped Equi-Angular Cubemap in
29958 side-by-side stereo format to equirectangular top-bottom stereo
29959 format:
29960 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
29962 Commands
29964 This filter supports subset of above options as commands.
29966 vaguedenoiser
29968 Apply a wavelet based denoiser.
29969 It transforms each frame from the video input into the wavelet domain,
29971 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
29972 the obtained coefficients. It does an inverse wavelet transform after.
29973 Due to wavelet properties, it should give a nice smoothed result, and
29974 reduced noise, without blurring picture features.
29975 This filter accepts the following options:
29977 threshold
29979 The filtering strength. The higher, the more filtered the video
29980 will be. Hard thresholding can use a higher threshold than soft
29981 thresholding before the video looks overfiltered. Default value is
29982 2.
29983 method
29985 The filtering method the filter will use.
29986 It accepts the following values:
29988 hard
29990 All values under the threshold will be zeroed.
29991 soft
29993 All values under the threshold will be zeroed. All values above
29994 will be reduced by the threshold.
29995 garrote
29997 Scales or nullifies coefficients - intermediary between (more)
29998 soft and (less) hard thresholding.
29999 Default is garrote.
30001 nsteps
30003 Number of times, the wavelet will decompose the picture. Picture
30004 can't be decomposed beyond a particular point (typically, 8 for a
30005 640x480 frame - as 2^9 = 512 > 480). Valid values are integers
30006 between 1 and 32. Default value is 6.
30007 percent
30009 Partial of full denoising (limited coefficients shrinking), from 0
30010 to 100. Default value is 85.
30011 planes
30013 A list of the planes to process. By default all planes are
30014 processed.
30015 type
30017 The threshold type the filter will use.
30018 It accepts the following values:
30020 universal
30022 Threshold used is same for all decompositions.
30023 bayes
30025 Threshold used depends also on each decomposition coefficients.
30026 Default is universal.
30028 varblur
30030 Apply variable blur filter by using 2nd video stream to set blur
30031 radius. The 2nd stream must have the same dimensions.
30032 This filter accepts the following options:
30034 min_r
30036 Set min allowed radius. Allowed range is from 0 to 254. Default is
30037 0.
30038 max_r
30040 Set max allowed radius. Allowed range is from 1 to 255. Default is
30041 8.
30042 planes
30044 Set which planes to process. By default, all are used.
30045 The "varblur" filter also supports the framesync options.
30047 Commands
30049 This filter supports all the above options as commands.
30051 vectorscope
30053 Display 2 color component values in the two dimensional graph (which is
30054 called a vectorscope).
30055 This filter accepts the following options:
30057 mode, m
30059 Set vectorscope mode.
30060 It accepts the following values:
30062 gray
30064 tint
30065 Gray values are displayed on graph, higher brightness means
30066 more pixels have same component color value on location in
30067 graph. This is the default mode.
30068 color
30070 Gray values are displayed on graph. Surrounding pixels values
30071 which are not present in video frame are drawn in gradient of 2
30072 color components which are set by option "x" and "y". The 3rd
30073 color component is static.
30074 color2
30076 Actual color components values present in video frame are
30077 displayed on graph.
30078 color3
30080 Similar as color2 but higher frequency of same values "x" and
30081 "y" on graph increases value of another color component, which
30082 is luminance by default values of "x" and "y".
30083 color4
30085 Actual colors present in video frame are displayed on graph. If
30086 two different colors map to same position on graph then color
30087 with higher value of component not present in graph is picked.
30088 color5
30090 Gray values are displayed on graph. Similar to "color" but with
30091 3rd color component picked from radial gradient.
30092 x Set which color component will be represented on X-axis. Default is
30094 1.
30095 y Set which color component will be represented on Y-axis. Default is
30097 2.
30098 intensity, i
30100 Set intensity, used by modes: gray, color, color3 and color5 for
30101 increasing brightness of color component which represents frequency
30102 of (X, Y) location in graph.
30103 envelope, e
30105 none
30106 No envelope, this is default.
30107 instant
30109 Instant envelope, even darkest single pixel will be clearly
30110 highlighted.
30111 peak
30113 Hold maximum and minimum values presented in graph over time.
30114 This way you can still spot out of range values without
30115 constantly looking at vectorscope.
30116 peak+instant
30118 Peak and instant envelope combined together.
30119 graticule, g
30121 Set what kind of graticule to draw.
30122 none
30124 green
30125 color
30126 invert
30127 opacity, o
30128 Set graticule opacity.
30129 flags, f
30131 Set graticule flags.
30132 white
30134 Draw graticule for white point.
30135 black
30137 Draw graticule for black point.
30138 name
30140 Draw color points short names.
30141 bgopacity, b
30143 Set background opacity.
30144 lthreshold, l
30146 Set low threshold for color component not represented on X or Y
30147 axis. Values lower than this value will be ignored. Default is 0.
30148 Note this value is multiplied with actual max possible value one
30149 pixel component can have. So for 8-bit input and low threshold
30150 value of 0.1 actual threshold is 0.1 * 255 = 25.
30151 hthreshold, h
30153 Set high threshold for color component not represented on X or Y
30154 axis. Values higher than this value will be ignored. Default is 1.
30155 Note this value is multiplied with actual max possible value one
30156 pixel component can have. So for 8-bit input and high threshold
30157 value of 0.9 actual threshold is 0.9 * 255 = 230.
30158 colorspace, c
30160 Set what kind of colorspace to use when drawing graticule.
30161 auto
30163 601
30164 709
30165 Default is auto.
30167 tint0, t0
30169 tint1, t1
30170 Set color tint for gray/tint vectorscope mode. By default both
30171 options are zero. This means no tint, and output will remain gray.
30172 vidstabdetect
30174 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
30175 vidstabtransform for pass 2.
30176 This filter generates a file with relative translation and rotation
30178 transform information about subsequent frames, which is then used by
30179 the vidstabtransform filter.
30180 To enable compilation of this filter you need to configure FFmpeg with
30182 "--enable-libvidstab".
30183 This filter accepts the following options:
30185 result
30187 Set the path to the file used to write the transforms information.
30188 Default value is transforms.trf.
30189 shakiness
30191 Set how shaky the video is and how quick the camera is. It accepts
30192 an integer in the range 1-10, a value of 1 means little shakiness,
30193 a value of 10 means strong shakiness. Default value is 5.
30194 accuracy
30196 Set the accuracy of the detection process. It must be a value in
30197 the range 1-15. A value of 1 means low accuracy, a value of 15
30198 means high accuracy. Default value is 15.
30199 stepsize
30201 Set stepsize of the search process. The region around minimum is
30202 scanned with 1 pixel resolution. Default value is 6.
30203 mincontrast
30205 Set minimum contrast. Below this value a local measurement field is
30206 discarded. Must be a floating point value in the range 0-1. Default
30207 value is 0.3.
30208 tripod
30210 Set reference frame number for tripod mode.
30211 If enabled, the motion of the frames is compared to a reference
30213 frame in the filtered stream, identified by the specified number.
30214 The idea is to compensate all movements in a more-or-less static
30215 scene and keep the camera view absolutely still.
30216 If set to 0, it is disabled. The frames are counted starting from
30218 1.
30219 show
30221 Show fields and transforms in the resulting frames. It accepts an
30222 integer in the range 0-2. Default value is 0, which disables any
30223 visualization.
30224 Examples
30226 • Use default values:
30228 vidstabdetect
30230 • Analyze strongly shaky movie and put the results in file
30232 mytransforms.trf:
30233 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
30235 • Visualize the result of internal transformations in the resulting
30237 video:
30238 vidstabdetect=show=1
30240 • Analyze a video with medium shakiness using ffmpeg:
30242 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
30244 vidstabtransform
30246 Video stabilization/deshaking: pass 2 of 2, see vidstabdetect for pass
30247 1.
30248 Read a file with transform information for each frame and
30250 apply/compensate them. Together with the vidstabdetect filter this can
30251 be used to deshake videos. See also
30252 <http://public.hronopik.de/vid.stab>. It is important to also use the
30253 unsharp filter, see below.
30254 To enable compilation of this filter you need to configure FFmpeg with
30256 "--enable-libvidstab".
30257 Options
30259 input
30261 Set path to the file used to read the transforms. Default value is
30262 transforms.trf.
30263 smoothing
30265 Set the number of frames (value*2 + 1) used for lowpass filtering
30266 the camera movements. Default value is 10.
30267 For example a number of 10 means that 21 frames are used (10 in the
30269 past and 10 in the future) to smoothen the motion in the video. A
30270 larger value leads to a smoother video, but limits the acceleration
30271 of the camera (pan/tilt movements). 0 is a special case where a
30272 static camera is simulated.
30273 optalgo
30275 Set the camera path optimization algorithm.
30276 Accepted values are:
30278 gauss
30280 gaussian kernel low-pass filter on camera motion (default)
30281 avg averaging on transformations
30283 maxshift
30285 Set maximal number of pixels to translate frames. Default value is
30286 -1, meaning no limit.
30287 maxangle
30289 Set maximal angle in radians (degree*PI/180) to rotate frames.
30290 Default value is -1, meaning no limit.
30291 crop
30293 Specify how to deal with borders that may be visible due to
30294 movement compensation.
30295 Available values are:
30297 keep
30299 keep image information from previous frame (default)
30300 black
30302 fill the border black
30303 invert
30305 Invert transforms if set to 1. Default value is 0.
30306 relative
30308 Consider transforms as relative to previous frame if set to 1,
30309 absolute if set to 0. Default value is 0.
30310 zoom
30312 Set percentage to zoom. A positive value will result in a zoom-in
30313 effect, a negative value in a zoom-out effect. Default value is 0
30314 (no zoom).
30315 optzoom
30317 Set optimal zooming to avoid borders.
30318 Accepted values are:
30320 0 disabled
30322 1 optimal static zoom value is determined (only very strong
30324 movements will lead to visible borders) (default)
30325 2 optimal adaptive zoom value is determined (no borders will be
30327 visible), see zoomspeed
30328 Note that the value given at zoom is added to the one calculated
30330 here.
30331 zoomspeed
30333 Set percent to zoom maximally each frame (enabled when optzoom is
30334 set to 2). Range is from 0 to 5, default value is 0.25.
30335 interpol
30337 Specify type of interpolation.
30338 Available values are:
30340 no no interpolation
30342 linear
30344 linear only horizontal
30345 bilinear
30347 linear in both directions (default)
30348 bicubic
30350 cubic in both directions (slow)
30351 tripod
30353 Enable virtual tripod mode if set to 1, which is equivalent to
30354 "relative=0:smoothing=0". Default value is 0.
30355 Use also "tripod" option of vidstabdetect.
30357 debug
30359 Increase log verbosity if set to 1. Also the detected global
30360 motions are written to the temporary file global_motions.trf.
30361 Default value is 0.
30362 Examples
30364 • Use ffmpeg for a typical stabilization with default values:
30366 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
30368 Note the use of the unsharp filter which is always recommended.
30370 • Zoom in a bit more and load transform data from a given file:
30372 vidstabtransform=zoom=5:input="mytransforms.trf"
30374 • Smoothen the video even more:
30376 vidstabtransform=smoothing=30
30378 vflip
30380 Flip the input video vertically.
30381 For example, to vertically flip a video with ffmpeg:
30383 ffmpeg -i in.avi -vf "vflip" out.avi
30385 vfrdet
30387 Detect variable frame rate video.
30388 This filter tries to detect if the input is variable or constant frame
30390 rate.
30391 At end it will output number of frames detected as having variable
30393 delta pts, and ones with constant delta pts. If there was frames with
30394 variable delta, than it will also show min, max and average delta
30395 encountered.
30396 vibrance
30398 Boost or alter saturation.
30399 The filter accepts the following options:
30401 intensity
30403 Set strength of boost if positive value or strength of alter if
30404 negative value. Default is 0. Allowed range is from -2 to 2.
30405 rbal
30407 Set the red balance. Default is 1. Allowed range is from -10 to 10.
30408 gbal
30410 Set the green balance. Default is 1. Allowed range is from -10 to
30411 10.
30412 bbal
30414 Set the blue balance. Default is 1. Allowed range is from -10 to
30415 10.
30416 rlum
30418 Set the red luma coefficient.
30419 glum
30421 Set the green luma coefficient.
30422 blum
30424 Set the blue luma coefficient.
30425 alternate
30427 If "intensity" is negative and this is set to 1, colors will
30428 change, otherwise colors will be less saturated, more towards gray.
30429 Commands
30431 This filter supports the all above options as commands.
30433 vif
30435 Obtain the average VIF (Visual Information Fidelity) between two input
30436 videos.
30437 This filter takes two input videos.
30439 Both input videos must have the same resolution and pixel format for
30441 this filter to work correctly. Also it assumes that both inputs have
30442 the same number of frames, which are compared one by one.
30443 The obtained average VIF score is printed through the logging system.
30445 The filter stores the calculated VIF score of each frame.
30447 In the below example the input file main.mpg being processed is
30449 compared with the reference file ref.mpg.
30450 ffmpeg -i main.mpg -i ref.mpg -lavfi vif -f null -
30452 vignette
30454 Make or reverse a natural vignetting effect.
30455 The filter accepts the following options:
30457 angle, a
30459 Set lens angle expression as a number of radians.
30460 The value is clipped in the "[0,PI/2]" range.
30462 Default value: "PI/5"
30464 x0
30466 y0 Set center coordinates expressions. Respectively "w/2" and "h/2" by
30467 default.
30468 mode
30470 Set forward/backward mode.
30471 Available modes are:
30473 forward
30475 The larger the distance from the central point, the darker the
30476 image becomes.
30477 backward
30479 The larger the distance from the central point, the brighter
30480 the image becomes. This can be used to reverse a vignette
30481 effect, though there is no automatic detection to extract the
30482 lens angle and other settings (yet). It can also be used to
30483 create a burning effect.
30484 Default value is forward.
30486 eval
30488 Set evaluation mode for the expressions (angle, x0, y0).
30489 It accepts the following values:
30491 init
30493 Evaluate expressions only once during the filter
30494 initialization.
30495 frame
30497 Evaluate expressions for each incoming frame. This is way
30498 slower than the init mode since it requires all the scalers to
30499 be re-computed, but it allows advanced dynamic expressions.
30500 Default value is init.
30502 dither
30504 Set dithering to reduce the circular banding effects. Default is 1
30505 (enabled).
30506 aspect
30508 Set vignette aspect. This setting allows one to adjust the shape of
30509 the vignette. Setting this value to the SAR of the input will make
30510 a rectangular vignetting following the dimensions of the video.
30511 Default is "1/1".
30513 Expressions
30515 The alpha, x0 and y0 expressions can contain the following parameters.
30517 w
30519 h input width and height
30520 n the number of input frame, starting from 0
30522 pts the PTS (Presentation TimeStamp) time of the filtered video frame,
30524 expressed in TB units, NAN if undefined
30525 r frame rate of the input video, NAN if the input frame rate is
30527 unknown
30528 t the PTS (Presentation TimeStamp) of the filtered video frame,
30530 expressed in seconds, NAN if undefined
30531 tb time base of the input video
30533 Examples
30535 • Apply simple strong vignetting effect:
30537 vignette=PI/4
30539 • Make a flickering vignetting:
30541 vignette='PI/4+random(1)*PI/50':eval=frame
30543 vmafmotion
30545 Obtain the average VMAF motion score of a video. It is one of the
30546 component metrics of VMAF.
30547 The obtained average motion score is printed through the logging
30549 system.
30550 The filter accepts the following options:
30552 stats_file
30554 If specified, the filter will use the named file to save the motion
30555 score of each frame with respect to the previous frame. When
30556 filename equals "-" the data is sent to standard output.
30557 Example:
30559 ffmpeg -i ref.mpg -vf vmafmotion -f null -
30561 vstack
30563 Stack input videos vertically.
30564 All streams must be of same pixel format and of same width.
30566 Note that this filter is faster than using overlay and pad filter to
30568 create same output.
30569 The filter accepts the following options:
30571 inputs
30573 Set number of input streams. Default is 2.
30574 shortest
30576 If set to 1, force the output to terminate when the shortest input
30577 terminates. Default value is 0.
30578 w3fdif
30580 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
30581 Deinterlacing Filter").
30582 Based on the process described by Martin Weston for BBC R&D, and
30584 implemented based on the de-interlace algorithm written by Jim
30585 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter uses
30586 filter coefficients calculated by BBC R&D.
30587 This filter uses field-dominance information in frame to decide which
30589 of each pair of fields to place first in the output. If it gets it
30590 wrong use setfield filter before "w3fdif" filter.
30591 There are two sets of filter coefficients, so called "simple" and
30593 "complex". Which set of filter coefficients is used can be set by
30594 passing an optional parameter:
30595 filter
30597 Set the interlacing filter coefficients. Accepts one of the
30598 following values:
30599 simple
30601 Simple filter coefficient set.
30602 complex
30604 More-complex filter coefficient set.
30605 Default value is complex.
30607 mode
30609 The interlacing mode to adopt. It accepts one of the following
30610 values:
30611 frame
30613 Output one frame for each frame.
30614 field
30616 Output one frame for each field.
30617 The default value is "field".
30619 parity
30621 The picture field parity assumed for the input interlaced video. It
30622 accepts one of the following values:
30623 tff Assume the top field is first.
30625 bff Assume the bottom field is first.
30627 auto
30629 Enable automatic detection of field parity.
30630 The default value is "auto". If the interlacing is unknown or the
30632 decoder does not export this information, top field first will be
30633 assumed.
30634 deint
30636 Specify which frames to deinterlace. Accepts one of the following
30637 values:
30638 all Deinterlace all frames,
30640 interlaced
30642 Only deinterlace frames marked as interlaced.
30643 Default value is all.
30645 Commands
30647 This filter supports same commands as options.
30649 waveform
30651 Video waveform monitor.
30652 The waveform monitor plots color component intensity. By default
30654 luminance only. Each column of the waveform corresponds to a column of
30655 pixels in the source video.
30656 It accepts the following options:
30658 mode, m
30660 Can be either "row", or "column". Default is "column". In row
30661 mode, the graph on the left side represents color component value 0
30662 and the right side represents value = 255. In column mode, the top
30663 side represents color component value = 0 and bottom side
30664 represents value = 255.
30665 intensity, i
30667 Set intensity. Smaller values are useful to find out how many
30668 values of the same luminance are distributed across input
30669 rows/columns. Default value is 0.04. Allowed range is [0, 1].
30670 mirror, r
30672 Set mirroring mode. 0 means unmirrored, 1 means mirrored. In
30673 mirrored mode, higher values will be represented on the left side
30674 for "row" mode and at the top for "column" mode. Default is 1
30675 (mirrored).
30676 display, d
30678 Set display mode. It accepts the following values:
30679 overlay
30681 Presents information identical to that in the "parade", except
30682 that the graphs representing color components are superimposed
30683 directly over one another.
30684 This display mode makes it easier to spot relative differences
30686 or similarities in overlapping areas of the color components
30687 that are supposed to be identical, such as neutral whites,
30688 grays, or blacks.
30689 stack
30691 Display separate graph for the color components side by side in
30692 "row" mode or one below the other in "column" mode.
30693 parade
30695 Display separate graph for the color components side by side in
30696 "column" mode or one below the other in "row" mode.
30697 Using this display mode makes it easy to spot color casts in
30699 the highlights and shadows of an image, by comparing the
30700 contours of the top and the bottom graphs of each waveform.
30701 Since whites, grays, and blacks are characterized by exactly
30702 equal amounts of red, green, and blue, neutral areas of the
30703 picture should display three waveforms of roughly equal
30704 width/height. If not, the correction is easy to perform by
30705 making level adjustments the three waveforms.
30706 Default is "stack".
30708 components, c
30710 Set which color components to display. Default is 1, which means
30711 only luminance or red color component if input is in RGB
30712 colorspace. If is set for example to 7 it will display all 3 (if)
30713 available color components.
30714 envelope, e
30716 none
30717 No envelope, this is default.
30718 instant
30720 Instant envelope, minimum and maximum values presented in graph
30721 will be easily visible even with small "step" value.
30722 peak
30724 Hold minimum and maximum values presented in graph across time.
30725 This way you can still spot out of range values without
30726 constantly looking at waveforms.
30727 peak+instant
30729 Peak and instant envelope combined together.
30730 filter, f
30732 lowpass
30733 No filtering, this is default.
30734 flat
30736 Luma and chroma combined together.
30737 aflat
30739 Similar as above, but shows difference between blue and red
30740 chroma.
30741 xflat
30743 Similar as above, but use different colors.
30744 yflat
30746 Similar as above, but again with different colors.
30747 chroma
30749 Displays only chroma.
30750 color
30752 Displays actual color value on waveform.
30753 acolor
30755 Similar as above, but with luma showing frequency of chroma
30756 values.
30757 graticule, g
30759 Set which graticule to display.
30760 none
30762 Do not display graticule.
30763 green
30765 Display green graticule showing legal broadcast ranges.
30766 orange
30768 Display orange graticule showing legal broadcast ranges.
30769 invert
30771 Display invert graticule showing legal broadcast ranges.
30772 opacity, o
30774 Set graticule opacity.
30775 flags, fl
30777 Set graticule flags.
30778 numbers
30780 Draw numbers above lines. By default enabled.
30781 dots
30783 Draw dots instead of lines.
30784 scale, s
30786 Set scale used for displaying graticule.
30787 digital
30789 millivolts
30790 ire
30791 Default is digital.
30793 bgopacity, b
30795 Set background opacity.
30796 tint0, t0
30798 tint1, t1
30799 Set tint for output. Only used with lowpass filter and when
30800 display is not overlay and input pixel formats are not RGB.
30801 fitmode, fm
30803 Set sample aspect ratio of video output frames. Can be used to
30804 configure waveform so it is not streched too much in one of
30805 directions.
30806 none
30808 Set sample aspect ration to 1/1.
30809 size
30811 Set sample aspect ratio to match input size of video
30812 Default is none.
30814 weave, doubleweave
30816 The "weave" takes a field-based video input and join each two
30817 sequential fields into single frame, producing a new double height clip
30818 with half the frame rate and half the frame count.
30819 The "doubleweave" works same as "weave" but without halving frame rate
30821 and frame count.
30822 It accepts the following option:
30824 first_field
30826 Set first field. Available values are:
30827 top, t
30829 Set the frame as top-field-first.
30830 bottom, b
30832 Set the frame as bottom-field-first.
30833 Examples
30835 • Interlace video using select and separatefields filter:
30837 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
30839 xbr
30841 Apply the xBR high-quality magnification filter which is designed for
30842 pixel art. It follows a set of edge-detection rules, see
30843 <https://forums.libretro.com/t/xbr-algorithm-tutorial/123>.
30844 It accepts the following option:
30846 n Set the scaling dimension: 2 for "2xBR", 3 for "3xBR" and 4 for
30848 "4xBR". Default is 3.
30849 xcorrelate
30851 Apply normalized cross-correlation between first and second input video
30852 stream.
30853 Second input video stream dimensions must be lower than first input
30855 video stream.
30856 The filter accepts the following options:
30858 planes
30860 Set which planes to process.
30861 secondary
30863 Set which secondary video frames will be processed from second
30864 input video stream, can be first or all. Default is all.
30865 The "xcorrelate" filter also supports the framesync options.
30867 xfade
30869 Apply cross fade from one input video stream to another input video
30870 stream. The cross fade is applied for specified duration.
30871 Both inputs must be constant frame-rate and have the same resolution,
30873 pixel format, frame rate and timebase.
30874 The filter accepts the following options:
30876 transition
30878 Set one of available transition effects:
30879 custom
30881 fade
30882 wipeleft
30883 wiperight
30884 wipeup
30885 wipedown
30886 slideleft
30887 slideright
30888 slideup
30889 slidedown
30890 circlecrop
30891 rectcrop
30892 distance
30893 fadeblack
30894 fadewhite
30895 radial
30896 smoothleft
30897 smoothright
30898 smoothup
30899 smoothdown
30900 circleopen
30901 circleclose
30902 vertopen
30903 vertclose
30904 horzopen
30905 horzclose
30906 dissolve
30907 pixelize
30908 diagtl
30909 diagtr
30910 diagbl
30911 diagbr
30912 hlslice
30913 hrslice
30914 vuslice
30915 vdslice
30916 hblur
30917 fadegrays
30918 wipetl
30919 wipetr
30920 wipebl
30921 wipebr
30922 squeezeh
30923 squeezev
30924 zoomin
30925 fadefast
30926 fadeslow
30927 Default transition effect is fade.
30929 duration
30931 Set cross fade duration in seconds. Range is 0 to 60 seconds.
30932 Default duration is 1 second.
30933 offset
30935 Set cross fade start relative to first input stream in seconds.
30936 Default offset is 0.
30937 expr
30939 Set expression for custom transition effect.
30940 The expressions can use the following variables and functions:
30942 X
30944 Y The coordinates of the current sample.
30945 W
30947 H The width and height of the image.
30948 P Progress of transition effect.
30950 PLANE
30952 Currently processed plane.
30953 A Return value of first input at current location and plane.
30955 B Return value of second input at current location and plane.
30957 a0(x, y)
30959 a1(x, y)
30960 a2(x, y)
30961 a3(x, y)
30962 Return the value of the pixel at location (x,y) of the
30963 first/second/third/fourth component of first input.
30964 b0(x, y)
30966 b1(x, y)
30967 b2(x, y)
30968 b3(x, y)
30969 Return the value of the pixel at location (x,y) of the
30970 first/second/third/fourth component of second input.
30971 Examples
30973 • Cross fade from one input video to another input video, with fade
30975 transition and duration of transition of 2 seconds starting at
30976 offset of 5 seconds:
30977 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
30979 xmedian
30981 Pick median pixels from several input videos.
30982 The filter accepts the following options:
30984 inputs
30986 Set number of inputs. Default is 3. Allowed range is from 3 to
30987 255. If number of inputs is even number, than result will be mean
30988 value between two median values.
30989 planes
30991 Set which planes to filter. Default value is 15, by which all
30992 planes are processed.
30993 percentile
30995 Set median percentile. Default value is 0.5. Default value of 0.5
30996 will pick always median values, while 0 will pick minimum values,
30997 and 1 maximum values.
30998 Commands
31000 This filter supports all above options as commands, excluding option
31002 "inputs".
31003 xstack
31005 Stack video inputs into custom layout.
31006 All streams must be of same pixel format.
31008 The filter accepts the following options:
31010 inputs
31012 Set number of input streams. Default is 2.
31013 layout
31015 Specify layout of inputs. This option requires the desired layout
31016 configuration to be explicitly set by the user. This sets position
31017 of each video input in output. Each input is separated by '|'. The
31018 first number represents the column, and the second number
31019 represents the row. Numbers start at 0 and are separated by '_'.
31020 Optionally one can use wX and hX, where X is video input from which
31021 to take width or height. Multiple values can be used when
31022 separated by '+'. In such case values are summed together.
31023 Note that if inputs are of different sizes gaps may appear, as not
31025 all of the output video frame will be filled. Similarly, videos can
31026 overlap each other if their position doesn't leave enough space for
31027 the full frame of adjoining videos.
31028 For 2 inputs, a default layout of "0_0|w0_0" (equivalent to
31030 "grid=2x1") is set. In all other cases, a layout or a grid must be
31031 set by the user. Either "grid" or "layout" can be specified at a
31032 time. Specifying both will result in an error.
31033 grid
31035 Specify a fixed size grid of inputs. This option is used to create
31036 a fixed size grid of the input streams. Set the grid size in the
31037 form "COLUMNSxROWS". There must be "ROWS * COLUMNS" input streams
31038 and they will be arranged as a grid with "ROWS" rows and "COLUMNS"
31039 columns. When using this option, each input stream within a row
31040 must have the same height and all the rows must have the same
31041 width.
31042 If "grid" is set, then "inputs" option is ignored and is implicitly
31044 set to "ROWS * COLUMNS".
31045 For 2 inputs, a default grid of "2x1" (equivalent to
31047 "layout=0_0|w0_0") is set. In all other cases, a layout or a grid
31048 must be set by the user. Either "grid" or "layout" can be specified
31049 at a time. Specifying both will result in an error.
31050 shortest
31052 If set to 1, force the output to terminate when the shortest input
31053 terminates. Default value is 0.
31054 fill
31056 If set to valid color, all unused pixels will be filled with that
31057 color. By default fill is set to none, so it is disabled.
31058 Examples
31060 • Display 4 inputs into 2x2 grid.
31062 Layout:
31064 input1(0, 0) | input3(w0, 0)
31066 input2(0, h0) | input4(w0, h0)
31067 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
31071 Note that if inputs are of different sizes, gaps or overlaps may
31073 occur.
31074 • Display 4 inputs into 1x4 grid.
31076 Layout:
31078 input1(0, 0)
31080 input2(0, h0)
31081 input3(0, h0+h1)
31082 input4(0, h0+h1+h2)
31083 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
31087 Note that if inputs are of different widths, unused space will
31089 appear.
31090 • Display 9 inputs into 3x3 grid.
31092 Layout:
31094 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
31096 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
31097 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
31098 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
31102 Note that if inputs are of different sizes, gaps or overlaps may
31104 occur.
31105 • Display 16 inputs into 4x4 grid.
31107 Layout:
31109 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
31111 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
31112 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
31113 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
31114 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|
31118 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
31119 Note that if inputs are of different sizes, gaps or overlaps may
31121 occur.
31122 yadif
31124 Deinterlace the input video ("yadif" means "yet another deinterlacing
31125 filter").
31126 It accepts the following parameters:
31128 mode
31130 The interlacing mode to adopt. It accepts one of the following
31131 values:
31132 0, send_frame
31134 Output one frame for each frame.
31135 1, send_field
31137 Output one frame for each field.
31138 2, send_frame_nospatial
31140 Like "send_frame", but it skips the spatial interlacing check.
31141 3, send_field_nospatial
31143 Like "send_field", but it skips the spatial interlacing check.
31144 The default value is "send_frame".
31146 parity
31148 The picture field parity assumed for the input interlaced video. It
31149 accepts one of the following values:
31150 0, tff
31152 Assume the top field is first.
31153 1, bff
31155 Assume the bottom field is first.
31156 -1, auto
31158 Enable automatic detection of field parity.
31159 The default value is "auto". If the interlacing is unknown or the
31161 decoder does not export this information, top field first will be
31162 assumed.
31163 deint
31165 Specify which frames to deinterlace. Accepts one of the following
31166 values:
31167 0, all
31169 Deinterlace all frames.
31170 1, interlaced
31172 Only deinterlace frames marked as interlaced.
31173 The default value is "all".
31175 yadif_cuda
31177 Deinterlace the input video using the yadif algorithm, but implemented
31178 in CUDA so that it can work as part of a GPU accelerated pipeline with
31179 nvdec and/or nvenc.
31180 It accepts the following parameters:
31182 mode
31184 The interlacing mode to adopt. It accepts one of the following
31185 values:
31186 0, send_frame
31188 Output one frame for each frame.
31189 1, send_field
31191 Output one frame for each field.
31192 2, send_frame_nospatial
31194 Like "send_frame", but it skips the spatial interlacing check.
31195 3, send_field_nospatial
31197 Like "send_field", but it skips the spatial interlacing check.
31198 The default value is "send_frame".
31200 parity
31202 The picture field parity assumed for the input interlaced video. It
31203 accepts one of the following values:
31204 0, tff
31206 Assume the top field is first.
31207 1, bff
31209 Assume the bottom field is first.
31210 -1, auto
31212 Enable automatic detection of field parity.
31213 The default value is "auto". If the interlacing is unknown or the
31215 decoder does not export this information, top field first will be
31216 assumed.
31217 deint
31219 Specify which frames to deinterlace. Accepts one of the following
31220 values:
31221 0, all
31223 Deinterlace all frames.
31224 1, interlaced
31226 Only deinterlace frames marked as interlaced.
31227 The default value is "all".
31229 yaepblur
31231 Apply blur filter while preserving edges ("yaepblur" means "yet another
31232 edge preserving blur filter"). The algorithm is described in "J. S.
31233 Lee, Digital image enhancement and noise filtering by use of local
31234 statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
31235 It accepts the following parameters:
31237 radius, r
31239 Set the window radius. Default value is 3.
31240 planes, p
31242 Set which planes to filter. Default is only the first plane.
31243 sigma, s
31245 Set blur strength. Default value is 128.
31246 Commands
31248 This filter supports same commands as options.
31250 zoompan
31252 Apply Zoom & Pan effect.
31253 This filter accepts the following options:
31255 zoom, z
31257 Set the zoom expression. Range is 1-10. Default is 1.
31258 x
31260 y Set the x and y expression. Default is 0.
31261 d Set the duration expression in number of frames. This sets for how
31263 many number of frames effect will last for single input image.
31264 Default is 90.
31265 s Set the output image size, default is 'hd720'.
31267 fps Set the output frame rate, default is '25'.
31269 Each expression can contain the following constants:
31271 in_w, iw
31273 Input width.
31274 in_h, ih
31276 Input height.
31277 out_w, ow
31279 Output width.
31280 out_h, oh
31282 Output height.
31283 in Input frame count.
31285 on Output frame count.
31287 in_time, it
31289 The input timestamp expressed in seconds. It's NAN if the input
31290 timestamp is unknown.
31291 out_time, time, ot
31293 The output timestamp expressed in seconds.
31294 x
31296 y Last calculated 'x' and 'y' position from 'x' and 'y' expression
31297 for current input frame.
31298 px
31300 py 'x' and 'y' of last output frame of previous input frame or 0 when
31301 there was not yet such frame (first input frame).
31302 zoom
31304 Last calculated zoom from 'z' expression for current input frame.
31305 pzoom
31307 Last calculated zoom of last output frame of previous input frame.
31308 duration
31310 Number of output frames for current input frame. Calculated from
31311 'd' expression for each input frame.
31312 pduration
31314 number of output frames created for previous input frame
31315 a Rational number: input width / input height
31317 sar sample aspect ratio
31319 dar display aspect ratio
31321 Examples
31323 • Zoom in up to 1.5x and pan at same time to some spot near center of
31325 picture:
31326 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
31328 • Zoom in up to 1.5x and pan always at center of picture:
31330 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
31332 • Same as above but without pausing:
31334 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
31336 • Zoom in 2x into center of picture only for the first second of the
31338 input video:
31339 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
31341 zscale
31343 Scale (resize) the input video, using the z.lib library:
31344 <https://github.com/sekrit-twc/zimg>. To enable compilation of this
31345 filter, you need to configure FFmpeg with "--enable-libzimg".
31346 The zscale filter forces the output display aspect ratio to be the same
31348 as the input, by changing the output sample aspect ratio.
31349 If the input image format is different from the format requested by the
31351 next filter, the zscale filter will convert the input to the requested
31352 format.
31353 Options
31355 The filter accepts the following options.
31357 width, w
31359 height, h
31360 Set the output video dimension expression. Default value is the
31361 input dimension.
31362 If the width or w value is 0, the input width is used for the
31364 output. If the height or h value is 0, the input height is used for
31365 the output.
31366 If one and only one of the values is -n with n >= 1, the zscale
31368 filter will use a value that maintains the aspect ratio of the
31369 input image, calculated from the other specified dimension. After
31370 that it will, however, make sure that the calculated dimension is
31371 divisible by n and adjust the value if necessary.
31372 If both values are -n with n >= 1, the behavior will be identical
31374 to both values being set to 0 as previously detailed.
31375 See below for the list of accepted constants for use in the
31377 dimension expression.
31378 size, s
31380 Set the video size. For the syntax of this option, check the "Video
31381 size" section in the ffmpeg-utils manual.
31382 dither, d
31384 Set the dither type.
31385 Possible values are:
31387 none
31389 ordered
31390 random
31391 error_diffusion
31392 Default is none.
31394 filter, f
31396 Set the resize filter type.
31397 Possible values are:
31399 point
31401 bilinear
31402 bicubic
31403 spline16
31404 spline36
31405 lanczos
31406 Default is bilinear.
31408 range, r
31410 Set the color range.
31411 Possible values are:
31413 input
31415 limited
31416 full
31417 Default is same as input.
31419 primaries, p
31421 Set the color primaries.
31422 Possible values are:
31424 input
31426 709
31427 unspecified
31428 170m
31429 240m
31430 2020
31431 Default is same as input.
31433 transfer, t
31435 Set the transfer characteristics.
31436 Possible values are:
31438 input
31440 709
31441 unspecified
31442 601
31443 linear
31444 2020_10
31445 2020_12
31446 smpte2084
31447 iec61966-2-1
31448 arib-std-b67
31449 Default is same as input.
31451 matrix, m
31453 Set the colorspace matrix.
31454 Possible value are:
31456 input
31458 709
31459 unspecified
31460 470bg
31461 170m
31462 2020_ncl
31463 2020_cl
31464 Default is same as input.
31466 rangein, rin
31468 Set the input color range.
31469 Possible values are:
31471 input
31473 limited
31474 full
31475 Default is same as input.
31477 primariesin, pin
31479 Set the input color primaries.
31480 Possible values are:
31482 input
31484 709
31485 unspecified
31486 170m
31487 240m
31488 2020
31489 Default is same as input.
31491 transferin, tin
31493 Set the input transfer characteristics.
31494 Possible values are:
31496 input
31498 709
31499 unspecified
31500 601
31501 linear
31502 2020_10
31503 2020_12
31504 Default is same as input.
31506 matrixin, min
31508 Set the input colorspace matrix.
31509 Possible value are:
31511 input
31513 709
31514 unspecified
31515 470bg
31516 170m
31517 2020_ncl
31518 2020_cl
31519 chromal, c
31520 Set the output chroma location.
31521 Possible values are:
31523 input
31525 left
31526 center
31527 topleft
31528 top
31529 bottomleft
31530 bottom
31531 chromalin, cin
31532 Set the input chroma location.
31533 Possible values are:
31535 input
31537 left
31538 center
31539 topleft
31540 top
31541 bottomleft
31542 bottom
31543 npl Set the nominal peak luminance.
31544 param_a
31546 Parameter A for scaling filters. Parameter "b" for bicubic, and the
31547 number of filter taps for lanczos.
31548 param_b
31550 Parameter B for scaling filters. Parameter "c" for bicubic.
31551 The values of the w and h options are expressions containing the
31553 following constants:
31554 in_w
31556 in_h
31557 The input width and height
31558 iw
31560 ih These are the same as in_w and in_h.
31561 out_w
31563 out_h
31564 The output (scaled) width and height
31565 ow
31567 oh These are the same as out_w and out_h
31568 a The same as iw / ih
31570 sar input sample aspect ratio
31572 dar The input display aspect ratio. Calculated from "(iw / ih) * sar".
31574 hsub
31576 vsub
31577 horizontal and vertical input chroma subsample values. For example
31578 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
31579 ohsub
31581 ovsub
31582 horizontal and vertical output chroma subsample values. For example
31583 for the pixel format "yuv422p" hsub is 2 and vsub is 1.
31584 Commands
31586 This filter supports the following commands:
31588 width, w
31590 height, h
31591 Set the output video dimension expression. The command accepts the
31592 same syntax of the corresponding option.
31593 If the specified expression is not valid, it is kept at its current
31595 value.
31596
31598 Below is a description of the currently available OpenCL video filters.
31599 To enable compilation of these filters you need to configure FFmpeg
31601 with "--enable-opencl".
31602 Running OpenCL filters requires you to initialize a hardware device and
31604 to pass that device to all filters in any filter graph.
31605 -init_hw_device opencl[=name][:device[,key=value...]]
31607 Initialise a new hardware device of type opencl called name, using
31608 the given device parameters.
31609 -filter_hw_device name
31611 Pass the hardware device called name to all filters in any filter
31612 graph.
31613 For more detailed information see
31615 <https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options>
31616 • Example of choosing the first device on the second platform and
31618 running avgblur_opencl filter with default parameters on it.
31619 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
31621 Since OpenCL filters are not able to access frame data in normal
31623 memory, all frame data needs to be uploaded(hwupload) to hardware
31624 surfaces connected to the appropriate device before being used and then
31625 downloaded(hwdownload) back to normal memory. Note that hwupload will
31626 upload to a surface with the same layout as the software frame, so it
31627 may be necessary to add a format filter immediately before to get the
31628 input into the right format and hwdownload does not support all formats
31629 on the output - it may be necessary to insert an additional format
31630 filter immediately following in the graph to get the output in a
31631 supported format.
31632 avgblur_opencl
31634 Apply average blur filter.
31635 The filter accepts the following options:
31637 sizeX
31639 Set horizontal radius size. Range is "[1, 1024]" and default value
31640 is 1.
31641 planes
31643 Set which planes to filter. Default value is 0xf, by which all
31644 planes are processed.
31645 sizeY
31647 Set vertical radius size. Range is "[1, 1024]" and default value is
31648 0. If zero, "sizeX" value will be used.
31649 Example
31651 • Apply average blur filter with horizontal and vertical size of 3,
31653 setting each pixel of the output to the average value of the 7x7
31654 region centered on it in the input. For pixels on the edges of the
31655 image, the region does not extend beyond the image boundaries, and
31656 so out-of-range coordinates are not used in the calculations.
31657 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
31659 boxblur_opencl
31661 Apply a boxblur algorithm to the input video.
31662 It accepts the following parameters:
31664 luma_radius, lr
31666 luma_power, lp
31667 chroma_radius, cr
31668 chroma_power, cp
31669 alpha_radius, ar
31670 alpha_power, ap
31671 A description of the accepted options follows.
31673 luma_radius, lr
31675 chroma_radius, cr
31676 alpha_radius, ar
31677 Set an expression for the box radius in pixels used for blurring
31678 the corresponding input plane.
31679 The radius value must be a non-negative number, and must not be
31681 greater than the value of the expression "min(w,h)/2" for the luma
31682 and alpha planes, and of "min(cw,ch)/2" for the chroma planes.
31683 Default value for luma_radius is "2". If not specified,
31685 chroma_radius and alpha_radius default to the corresponding value
31686 set for luma_radius.
31687 The expressions can contain the following constants:
31689 w
31691 h The input width and height in pixels.
31692 cw
31694 ch The input chroma image width and height in pixels.
31695 hsub
31697 vsub
31698 The horizontal and vertical chroma subsample values. For
31699 example, for the pixel format "yuv422p", hsub is 2 and vsub is
31700 1.
31701 luma_power, lp
31703 chroma_power, cp
31704 alpha_power, ap
31705 Specify how many times the boxblur filter is applied to the
31706 corresponding plane.
31707 Default value for luma_power is 2. If not specified, chroma_power
31709 and alpha_power default to the corresponding value set for
31710 luma_power.
31711 A value of 0 will disable the effect.
31713 Examples
31715 Apply boxblur filter, setting each pixel of the output to the average
31717 value of box-radiuses luma_radius, chroma_radius, alpha_radius for each
31718 plane respectively. The filter will apply luma_power, chroma_power,
31719 alpha_power times onto the corresponding plane. For pixels on the edges
31720 of the image, the radius does not extend beyond the image boundaries,
31721 and so out-of-range coordinates are not used in the calculations.
31722 • Apply a boxblur filter with the luma, chroma, and alpha radius set
31724 to 2 and luma, chroma, and alpha power set to 3. The filter will
31725 run 3 times with box-radius set to 2 for every plane of the image.
31726 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
31728 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
31729 • Apply a boxblur filter with luma radius set to 2, luma_power to 1,
31731 chroma_radius to 4, chroma_power to 5, alpha_radius to 3 and
31732 alpha_power to 7.
31733 For the luma plane, a 2x2 box radius will be run once.
31735 For the chroma plane, a 4x4 box radius will be run 5 times.
31737 For the alpha plane, a 3x3 box radius will be run 7 times.
31739 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
31741 colorkey_opencl
31743 RGB colorspace color keying.
31744 The filter accepts the following options:
31746 color
31748 The color which will be replaced with transparency.
31749 similarity
31751 Similarity percentage with the key color.
31752 0.01 matches only the exact key color, while 1.0 matches
31754 everything.
31755 blend
31757 Blend percentage.
31758 0.0 makes pixels either fully transparent, or not transparent at
31760 all.
31761 Higher values result in semi-transparent pixels, with a higher
31763 transparency the more similar the pixels color is to the key color.
31764 Examples
31766 • Make every semi-green pixel in the input transparent with some
31768 slight blending:
31769 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
31771 convolution_opencl
31773 Apply convolution of 3x3, 5x5, 7x7 matrix.
31774 The filter accepts the following options:
31776 0m
31778 1m
31779 2m
31780 3m Set matrix for each plane. Matrix is sequence of 9, 25 or 49
31781 signed numbers. Default value for each plane is "0 0 0 0 1 0 0 0
31782 0".
31783 0rdiv
31785 1rdiv
31786 2rdiv
31787 3rdiv
31788 Set multiplier for calculated value for each plane. If unset or 0,
31789 it will be sum of all matrix elements. The option value must be a
31790 float number greater or equal to 0.0. Default value is 1.0.
31791 0bias
31793 1bias
31794 2bias
31795 3bias
31796 Set bias for each plane. This value is added to the result of the
31797 multiplication. Useful for making the overall image brighter or
31798 darker. The option value must be a float number greater or equal
31799 to 0.0. Default value is 0.0.
31800 Examples
31802 • Apply sharpen:
31804 -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
31806 • Apply blur:
31808 -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
31810 • Apply edge enhance:
31812 -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
31814 • Apply edge detect:
31816 -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
31818 • Apply laplacian edge detector which includes diagonals:
31820 -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
31822 • Apply emboss:
31824 -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
31826 erosion_opencl
31828 Apply erosion effect to the video.
31829 This filter replaces the pixel by the local(3x3) minimum.
31831 It accepts the following options:
31833 threshold0
31835 threshold1
31836 threshold2
31837 threshold3
31838 Limit the maximum change for each plane. Range is "[0, 65535]" and
31839 default value is 65535. If 0, plane will remain unchanged.
31840 coordinates
31842 Flag which specifies the pixel to refer to. Range is "[0, 255]"
31843 and default value is 255, i.e. all eight pixels are used.
31844 Flags to local 3x3 coordinates region centered on "x":
31846 1 2 3
31848 4 x 5
31850 6 7 8
31852 Example
31854 • Apply erosion filter with threshold0 set to 30, threshold1 set 40,
31856 threshold2 set to 50 and coordinates set to 231, setting each pixel
31857 of the output to the local minimum between pixels: 1, 2, 3, 6, 7, 8
31858 of the 3x3 region centered on it in the input. If the difference
31859 between input pixel and local minimum is more then threshold of the
31860 corresponding plane, output pixel will be set to input pixel -
31861 threshold of corresponding plane.
31862 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
31864 deshake_opencl
31866 Feature-point based video stabilization filter.
31867 The filter accepts the following options:
31869 tripod
31871 Simulates a tripod by preventing any camera movement whatsoever
31872 from the original frame. Defaults to 0.
31873 debug
31875 Whether or not additional debug info should be displayed, both in
31876 the processed output and in the console.
31877 Note that in order to see console debug output you will also need
31879 to pass "-v verbose" to ffmpeg.
31880 Viewing point matches in the output video is only supported for RGB
31882 input.
31883 Defaults to 0.
31885 adaptive_crop
31887 Whether or not to do a tiny bit of cropping at the borders to cut
31888 down on the amount of mirrored pixels.
31889 Defaults to 1.
31891 refine_features
31893 Whether or not feature points should be refined at a sub-pixel
31894 level.
31895 This can be turned off for a slight performance gain at the cost of
31897 precision.
31898 Defaults to 1.
31900 smooth_strength
31902 The strength of the smoothing applied to the camera path from 0.0
31903 to 1.0.
31904 1.0 is the maximum smoothing strength while values less than that
31906 result in less smoothing.
31907 0.0 causes the filter to adaptively choose a smoothing strength on
31909 a per-frame basis.
31910 Defaults to 0.0.
31912 smooth_window_multiplier
31914 Controls the size of the smoothing window (the number of frames
31915 buffered to determine motion information from).
31916 The size of the smoothing window is determined by multiplying the
31918 framerate of the video by this number.
31919 Acceptable values range from 0.1 to 10.0.
31921 Larger values increase the amount of motion data available for
31923 determining how to smooth the camera path, potentially improving
31924 smoothness, but also increase latency and memory usage.
31925 Defaults to 2.0.
31927 Examples
31929 • Stabilize a video with a fixed, medium smoothing strength:
31931 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
31933 • Stabilize a video with debugging (both in console and in rendered
31935 video):
31936 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
31938 dilation_opencl
31940 Apply dilation effect to the video.
31941 This filter replaces the pixel by the local(3x3) maximum.
31943 It accepts the following options:
31945 threshold0
31947 threshold1
31948 threshold2
31949 threshold3
31950 Limit the maximum change for each plane. Range is "[0, 65535]" and
31951 default value is 65535. If 0, plane will remain unchanged.
31952 coordinates
31954 Flag which specifies the pixel to refer to. Range is "[0, 255]"
31955 and default value is 255, i.e. all eight pixels are used.
31956 Flags to local 3x3 coordinates region centered on "x":
31958 1 2 3
31960 4 x 5
31962 6 7 8
31964 Example
31966 • Apply dilation filter with threshold0 set to 30, threshold1 set 40,
31968 threshold2 set to 50 and coordinates set to 231, setting each pixel
31969 of the output to the local maximum between pixels: 1, 2, 3, 6, 7, 8
31970 of the 3x3 region centered on it in the input. If the difference
31971 between input pixel and local maximum is more then threshold of the
31972 corresponding plane, output pixel will be set to input pixel +
31973 threshold of corresponding plane.
31974 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
31976 nlmeans_opencl
31978 Non-local Means denoise filter through OpenCL, this filter accepts same
31979 options as nlmeans.
31980 overlay_opencl
31982 Overlay one video on top of another.
31983 It takes two inputs and has one output. The first input is the "main"
31985 video on which the second input is overlaid. This filter requires same
31986 memory layout for all the inputs. So, format conversion may be needed.
31987 The filter accepts the following options:
31989 x Set the x coordinate of the overlaid video on the main video.
31991 Default value is 0.
31992 y Set the y coordinate of the overlaid video on the main video.
31994 Default value is 0.
31995 Examples
31997 • Overlay an image LOGO at the top-left corner of the INPUT video.
31999 Both inputs are yuv420p format.
32000 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
32002 • The inputs have same memory layout for color channels , the overlay
32004 has additional alpha plane, like INPUT is yuv420p, and the LOGO is
32005 yuva420p.
32006 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
32008 pad_opencl
32010 Add paddings to the input image, and place the original input at the
32011 provided x, y coordinates.
32012 It accepts the following options:
32014 width, w
32016 height, h
32017 Specify an expression for the size of the output image with the
32018 paddings added. If the value for width or height is 0, the
32019 corresponding input size is used for the output.
32020 The width expression can reference the value set by the height
32022 expression, and vice versa.
32023 The default value of width and height is 0.
32025 x
32027 y Specify the offsets to place the input image at within the padded
32028 area, with respect to the top/left border of the output image.
32029 The x expression can reference the value set by the y expression,
32031 and vice versa.
32032 The default value of x and y is 0.
32034 If x or y evaluate to a negative number, they'll be changed so the
32036 input image is centered on the padded area.
32037 color
32039 Specify the color of the padded area. For the syntax of this
32040 option, check the "Color" section in the ffmpeg-utils manual.
32041 aspect
32043 Pad to an aspect instead to a resolution.
32044 The value for the width, height, x, and y options are expressions
32046 containing the following constants:
32047 in_w
32049 in_h
32050 The input video width and height.
32051 iw
32053 ih These are the same as in_w and in_h.
32054 out_w
32056 out_h
32057 The output width and height (the size of the padded area), as
32058 specified by the width and height expressions.
32059 ow
32061 oh These are the same as out_w and out_h.
32062 x
32064 y The x and y offsets as specified by the x and y expressions, or NAN
32065 if not yet specified.
32066 a same as iw / ih
32068 sar input sample aspect ratio
32070 dar input display aspect ratio, it is the same as (iw / ih) * sar
32072 prewitt_opencl
32074 Apply the Prewitt operator
32075 (<https://en.wikipedia.org/wiki/Prewitt_operator>) to input video
32076 stream.
32077 The filter accepts the following option:
32079 planes
32081 Set which planes to filter. Default value is 0xf, by which all
32082 planes are processed.
32083 scale
32085 Set value which will be multiplied with filtered result. Range is
32086 "[0.0, 65535]" and default value is 1.0.
32087 delta
32089 Set value which will be added to filtered result. Range is
32090 "[-65535, 65535]" and default value is 0.0.
32091 Example
32093 • Apply the Prewitt operator with scale set to 2 and delta set to 10.
32095 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
32097 program_opencl
32099 Filter video using an OpenCL program.
32100 source
32102 OpenCL program source file.
32103 kernel
32105 Kernel name in program.
32106 inputs
32108 Number of inputs to the filter. Defaults to 1.
32109 size, s
32111 Size of output frames. Defaults to the same as the first input.
32112 The "program_opencl" filter also supports the framesync options.
32114 The program source file must contain a kernel function with the given
32116 name, which will be run once for each plane of the output. Each run on
32117 a plane gets enqueued as a separate 2D global NDRange with one work-
32118 item for each pixel to be generated. The global ID offset for each
32119 work-item is therefore the coordinates of a pixel in the destination
32120 image.
32121 The kernel function needs to take the following arguments:
32123 • Destination image, __write_only image2d_t.
32125 This image will become the output; the kernel should write all of
32127 it.
32128 • Frame index, unsigned int.
32130 This is a counter starting from zero and increasing by one for each
32132 frame.
32133 • Source images, __read_only image2d_t.
32135 These are the most recent images on each input. The kernel may
32137 read from them to generate the output, but they can't be written
32138 to.
32139 Example programs:
32141 • Copy the input to the output (output must be the same size as the
32143 input).
32144 __kernel void copy(__write_only image2d_t destination,
32146 unsigned int index,
32147 __read_only image2d_t source)
32148 {
32149 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
32150 int2 location = (int2)(get_global_id(0), get_global_id(1));
32152 float4 value = read_imagef(source, sampler, location);
32154 write_imagef(destination, location, value);
32156 }
32157 • Apply a simple transformation, rotating the input by an amount
32159 increasing with the index counter. Pixel values are linearly
32160 interpolated by the sampler, and the output need not have the same
32161 dimensions as the input.
32162 __kernel void rotate_image(__write_only image2d_t dst,
32164 unsigned int index,
32165 __read_only image2d_t src)
32166 {
32167 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
32168 CLK_FILTER_LINEAR);
32169 float angle = (float)index / 100.0f;
32171 float2 dst_dim = convert_float2(get_image_dim(dst));
32173 float2 src_dim = convert_float2(get_image_dim(src));
32174 float2 dst_cen = dst_dim / 2.0f;
32176 float2 src_cen = src_dim / 2.0f;
32177 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
32179 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
32181 float2 src_pos = {
32182 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
32183 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
32184 };
32185 src_pos = src_pos * src_dim / dst_dim;
32186 float2 src_loc = src_pos + src_cen;
32188 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
32190 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
32191 write_imagef(dst, dst_loc, 0.5f);
32192 else
32193 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
32194 }
32195 • Blend two inputs together, with the amount of each input used
32197 varying with the index counter.
32198 __kernel void blend_images(__write_only image2d_t dst,
32200 unsigned int index,
32201 __read_only image2d_t src1,
32202 __read_only image2d_t src2)
32203 {
32204 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
32205 CLK_FILTER_LINEAR);
32206 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
32208 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
32210 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
32211 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
32212 float4 val1 = read_imagef(src1, sampler, src1_loc);
32214 float4 val2 = read_imagef(src2, sampler, src2_loc);
32215 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
32217 }
32218 remap_opencl
32220 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
32221 Destination pixel at position (X, Y) will be picked from source (x, y)
32223 position where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are
32224 out of range, zero value for pixel will be used for destination pixel.
32225 Xmap and Ymap input video streams must be of same dimensions. Output
32227 video stream will have Xmap/Ymap video stream dimensions. Xmap and
32228 Ymap input video streams are 32bit float pixel format, single channel.
32229 interp
32231 Specify interpolation used for remapping of pixels. Allowed values
32232 are "near" and "linear". Default value is "linear".
32233 fill
32235 Specify the color of the unmapped pixels. For the syntax of this
32236 option, check the "Color" section in the ffmpeg-utils manual.
32237 Default color is "black".
32238 roberts_opencl
32240 Apply the Roberts cross operator
32241 (<https://en.wikipedia.org/wiki/Roberts_cross>) to input video stream.
32242 The filter accepts the following option:
32244 planes
32246 Set which planes to filter. Default value is 0xf, by which all
32247 planes are processed.
32248 scale
32250 Set value which will be multiplied with filtered result. Range is
32251 "[0.0, 65535]" and default value is 1.0.
32252 delta
32254 Set value which will be added to filtered result. Range is
32255 "[-65535, 65535]" and default value is 0.0.
32256 Example
32258 • Apply the Roberts cross operator with scale set to 2 and delta set
32260 to 10
32261 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
32263 sobel_opencl
32265 Apply the Sobel operator
32266 (<https://en.wikipedia.org/wiki/Sobel_operator>) to input video stream.
32267 The filter accepts the following option:
32269 planes
32271 Set which planes to filter. Default value is 0xf, by which all
32272 planes are processed.
32273 scale
32275 Set value which will be multiplied with filtered result. Range is
32276 "[0.0, 65535]" and default value is 1.0.
32277 delta
32279 Set value which will be added to filtered result. Range is
32280 "[-65535, 65535]" and default value is 0.0.
32281 Example
32283 • Apply sobel operator with scale set to 2 and delta set to 10
32285 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
32287 tonemap_opencl
32289 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
32290 It accepts the following parameters:
32292 tonemap
32294 Specify the tone-mapping operator to be used. Same as tonemap
32295 option in tonemap.
32296 param
32298 Tune the tone mapping algorithm. same as param option in tonemap.
32299 desat
32301 Apply desaturation for highlights that exceed this level of
32302 brightness. The higher the parameter, the more color information
32303 will be preserved. This setting helps prevent unnaturally blown-out
32304 colors for super-highlights, by (smoothly) turning into white
32305 instead. This makes images feel more natural, at the cost of
32306 reducing information about out-of-range colors.
32307 The default value is 0.5, and the algorithm here is a little
32309 different from the cpu version tonemap currently. A setting of 0.0
32310 disables this option.
32311 threshold
32313 The tonemapping algorithm parameters is fine-tuned per each scene.
32314 And a threshold is used to detect whether the scene has changed or
32315 not. If the distance between the current frame average brightness
32316 and the current running average exceeds a threshold value, we would
32317 re-calculate scene average and peak brightness. The default value
32318 is 0.2.
32319 format
32321 Specify the output pixel format.
32322 Currently supported formats are:
32324 p010
32326 nv12
32327 range, r
32328 Set the output color range.
32329 Possible values are:
32331 tv/mpeg
32333 pc/jpeg
32334 Default is same as input.
32336 primaries, p
32338 Set the output color primaries.
32339 Possible values are:
32341 bt709
32343 bt2020
32344 Default is same as input.
32346 transfer, t
32348 Set the output transfer characteristics.
32349 Possible values are:
32351 bt709
32353 bt2020
32354 Default is bt709.
32356 matrix, m
32358 Set the output colorspace matrix.
32359 Possible value are:
32361 bt709
32363 bt2020
32364 Default is same as input.
32366 Example
32368 • Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010
32370 format using linear operator.
32371 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
32373 unsharp_opencl
32375 Sharpen or blur the input video.
32376 It accepts the following parameters:
32378 luma_msize_x, lx
32380 Set the luma matrix horizontal size. Range is "[1, 23]" and
32381 default value is 5.
32382 luma_msize_y, ly
32384 Set the luma matrix vertical size. Range is "[1, 23]" and default
32385 value is 5.
32386 luma_amount, la
32388 Set the luma effect strength. Range is "[-10, 10]" and default
32389 value is 1.0.
32390 Negative values will blur the input video, while positive values
32392 will sharpen it, a value of zero will disable the effect.
32393 chroma_msize_x, cx
32395 Set the chroma matrix horizontal size. Range is "[1, 23]" and
32396 default value is 5.
32397 chroma_msize_y, cy
32399 Set the chroma matrix vertical size. Range is "[1, 23]" and
32400 default value is 5.
32401 chroma_amount, ca
32403 Set the chroma effect strength. Range is "[-10, 10]" and default
32404 value is 0.0.
32405 Negative values will blur the input video, while positive values
32407 will sharpen it, a value of zero will disable the effect.
32408 All parameters are optional and default to the equivalent of the string
32410 '5:5:1.0:5:5:0.0'.
32411 Examples
32413 • Apply strong luma sharpen effect:
32415 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
32417 • Apply a strong blur of both luma and chroma parameters:
32419 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
32421 xfade_opencl
32423 Cross fade two videos with custom transition effect by using OpenCL.
32424 It accepts the following options:
32426 transition
32428 Set one of possible transition effects.
32429 custom
32431 Select custom transition effect, the actual transition
32432 description will be picked from source and kernel options.
32433 fade
32435 wipeleft
32436 wiperight
32437 wipeup
32438 wipedown
32439 slideleft
32440 slideright
32441 slideup
32442 slidedown
32443 Default transition is fade.
32444 source
32446 OpenCL program source file for custom transition.
32447 kernel
32449 Set name of kernel to use for custom transition from program source
32450 file.
32451 duration
32453 Set duration of video transition.
32454 offset
32456 Set time of start of transition relative to first video.
32457 The program source file must contain a kernel function with the given
32459 name, which will be run once for each plane of the output. Each run on
32460 a plane gets enqueued as a separate 2D global NDRange with one work-
32461 item for each pixel to be generated. The global ID offset for each
32462 work-item is therefore the coordinates of a pixel in the destination
32463 image.
32464 The kernel function needs to take the following arguments:
32466 • Destination image, __write_only image2d_t.
32468 This image will become the output; the kernel should write all of
32470 it.
32471 • First Source image, __read_only image2d_t. Second Source image,
32473 __read_only image2d_t.
32474 These are the most recent images on each input. The kernel may
32476 read from them to generate the output, but they can't be written
32477 to.
32478 • Transition progress, float. This value is always between 0 and 1
32480 inclusive.
32481 Example programs:
32483 • Apply dots curtain transition effect:
32485 __kernel void blend_images(__write_only image2d_t dst,
32487 __read_only image2d_t src1,
32488 __read_only image2d_t src2,
32489 float progress)
32490 {
32491 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
32492 CLK_FILTER_LINEAR);
32493 int2 p = (int2)(get_global_id(0), get_global_id(1));
32494 float2 rp = (float2)(get_global_id(0), get_global_id(1));
32495 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
32496 rp = rp / dim;
32497 float2 dots = (float2)(20.0, 20.0);
32499 float2 center = (float2)(0,0);
32500 float2 unused;
32501 float4 val1 = read_imagef(src1, sampler, p);
32503 float4 val2 = read_imagef(src2, sampler, p);
32504 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
32505 write_imagef(dst, p, next ? val1 : val2);
32507 }
32508
32510 VAAPI Video filters are usually used with VAAPI decoder and VAAPI
32511 encoder. Below is a description of VAAPI video filters.
32512 To enable compilation of these filters you need to configure FFmpeg
32514 with "--enable-vaapi".
32515 To use vaapi filters, you need to setup the vaapi device correctly. For
32517 more information, please read
32518 <https://trac.ffmpeg.org/wiki/Hardware/VAAPI>
32519 overlay_vaapi
32521 Overlay one video on the top of another.
32522 It takes two inputs and has one output. The first input is the "main"
32524 video on which the second input is overlaid. This filter requires same
32525 memory layout for all the inputs. So, format conversion may be needed.
32526 The filter accepts the following options:
32528 x Set the x coordinate of the overlaid video on the main video.
32530 Default value is 0.
32531 y Set the y coordinate of the overlaid video on the main video.
32533 Default value is 0.
32534 w Set the width of the overlaid video on the main video. Default
32536 value is the width of input overlay video.
32537 h Set the height of the overlaid video on the main video. Default
32539 value is the height of input overlay video.
32540 alpha
32542 Set blocking detection thresholds. Allowed range is 0.0 to 1.0, it
32543 requires an input video with alpha channel. Default value is 0.0.
32544 Examples
32546 • Overlay an image LOGO at the top-left corner of the INPUT video.
32548 Both inputs for this filter are yuv420p format.
32549 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_vaapi" OUTPUT
32551 • Overlay an image LOGO at the offset (200, 100) from the top-left
32553 corner of the INPUT video. The inputs have same memory layout for
32554 color channels, the overlay has additional alpha plane, like INPUT
32555 is yuv420p, and the LOGO is yuva420p.
32556 -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
32558 tonemap_vaapi
32560 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range)
32561 conversion with tone-mapping. It maps the dynamic range of HDR10
32562 content to the SDR content. It currently only accepts HDR10 as input.
32563 It accepts the following parameters:
32565 format
32567 Specify the output pixel format.
32568 Currently supported formats are:
32570 p010
32572 nv12
32573 Default is nv12.
32575 primaries, p
32577 Set the output color primaries.
32578 Default is same as input.
32580 transfer, t
32582 Set the output transfer characteristics.
32583 Default is bt709.
32585 matrix, m
32587 Set the output colorspace matrix.
32588 Default is same as input.
32590 Example
32592 • Convert HDR(HDR10) video to bt2020-transfer-characteristic p010
32594 format
32595 tonemap_vaapi=format=p010:t=bt2020-10
32597
32599 Below is a description of the currently available video sources.
32600 buffer
32602 Buffer video frames, and make them available to the filter chain.
32603 This source is mainly intended for a programmatic use, in particular
32605 through the interface defined in libavfilter/buffersrc.h.
32606 It accepts the following parameters:
32608 video_size
32610 Specify the size (width and height) of the buffered video frames.
32611 For the syntax of this option, check the "Video size" section in
32612 the ffmpeg-utils manual.
32613 width
32615 The input video width.
32616 height
32618 The input video height.
32619 pix_fmt
32621 A string representing the pixel format of the buffered video
32622 frames. It may be a number corresponding to a pixel format, or a
32623 pixel format name.
32624 time_base
32626 Specify the timebase assumed by the timestamps of the buffered
32627 frames.
32628 frame_rate
32630 Specify the frame rate expected for the video stream.
32631 pixel_aspect, sar
32633 The sample (pixel) aspect ratio of the input video.
32634 sws_param
32636 This option is deprecated and ignored. Prepend "sws_flags=flags;"
32637 to the filtergraph description to specify swscale flags for
32638 automatically inserted scalers. See Filtergraph syntax.
32639 hw_frames_ctx
32641 When using a hardware pixel format, this should be a reference to
32642 an AVHWFramesContext describing input frames.
32643 For example:
32645 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
32647 will instruct the source to accept video frames with size 320x240 and
32649 with format "yuv410p", assuming 1/24 as the timestamps timebase and
32650 square pixels (1:1 sample aspect ratio). Since the pixel format with
32651 name "yuv410p" corresponds to the number 6 (check the enum
32652 AVPixelFormat definition in libavutil/pixfmt.h), this example
32653 corresponds to:
32654 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
32656 Alternatively, the options can be specified as a flat string, but this
32658 syntax is deprecated:
32659 width:height:pix_fmt:time_base.num:time_base.den:pixel_aspect.num:pixel_aspect.den
32661 cellauto
32663 Create a pattern generated by an elementary cellular automaton.
32664 The initial state of the cellular automaton can be defined through the
32666 filename and pattern options. If such options are not specified an
32667 initial state is created randomly.
32668 At each new frame a new row in the video is filled with the result of
32670 the cellular automaton next generation. The behavior when the whole
32671 frame is filled is defined by the scroll option.
32672 This source accepts the following options:
32674 filename, f
32676 Read the initial cellular automaton state, i.e. the starting row,
32677 from the specified file. In the file, each non-whitespace
32678 character is considered an alive cell, a newline will terminate the
32679 row, and further characters in the file will be ignored.
32680 pattern, p
32682 Read the initial cellular automaton state, i.e. the starting row,
32683 from the specified string.
32684 Each non-whitespace character in the string is considered an alive
32686 cell, a newline will terminate the row, and further characters in
32687 the string will be ignored.
32688 rate, r
32690 Set the video rate, that is the number of frames generated per
32691 second. Default is 25.
32692 random_fill_ratio, ratio
32694 Set the random fill ratio for the initial cellular automaton row.
32695 It is a floating point number value ranging from 0 to 1, defaults
32696 to 1/PHI.
32697 This option is ignored when a file or a pattern is specified.
32699 random_seed, seed
32701 Set the seed for filling randomly the initial row, must be an
32702 integer included between 0 and UINT32_MAX. If not specified, or if
32703 explicitly set to -1, the filter will try to use a good random seed
32704 on a best effort basis.
32705 rule
32707 Set the cellular automaton rule, it is a number ranging from 0 to
32708 255. Default value is 110.
32709 size, s
32711 Set the size of the output video. For the syntax of this option,
32712 check the "Video size" section in the ffmpeg-utils manual.
32713 If filename or pattern is specified, the size is set by default to
32715 the width of the specified initial state row, and the height is set
32716 to width * PHI.
32717 If size is set, it must contain the width of the specified pattern
32719 string, and the specified pattern will be centered in the larger
32720 row.
32721 If a filename or a pattern string is not specified, the size value
32723 defaults to "320x518" (used for a randomly generated initial
32724 state).
32725 scroll
32727 If set to 1, scroll the output upward when all the rows in the
32728 output have been already filled. If set to 0, the new generated row
32729 will be written over the top row just after the bottom row is
32730 filled. Defaults to 1.
32731 start_full, full
32733 If set to 1, completely fill the output with generated rows before
32734 outputting the first frame. This is the default behavior, for
32735 disabling set the value to 0.
32736 stitch
32738 If set to 1, stitch the left and right row edges together. This is
32739 the default behavior, for disabling set the value to 0.
32740 Examples
32742 • Read the initial state from pattern, and specify an output of size
32744 200x400.
32745 cellauto=f=pattern:s=200x400
32747 • Generate a random initial row with a width of 200 cells, with a
32749 fill ratio of 2/3:
32750 cellauto=ratio=2/3:s=200x200
32752 • Create a pattern generated by rule 18 starting by a single alive
32754 cell centered on an initial row with width 100:
32755 cellauto=p=@s=100x400:full=0:rule=18
32757 • Specify a more elaborated initial pattern:
32759 cellauto=p='@@ @ @@':s=100x400:full=0:rule=18
32761 coreimagesrc
32763 Video source generated on GPU using Apple's CoreImage API on OSX.
32764 This video source is a specialized version of the coreimage video
32766 filter. Use a core image generator at the beginning of the applied
32767 filterchain to generate the content.
32768 The coreimagesrc video source accepts the following options:
32770 list_generators
32772 List all available generators along with all their respective
32773 options as well as possible minimum and maximum values along with
32774 the default values.
32775 list_generators=true
32777 size, s
32779 Specify the size of the sourced video. For the syntax of this
32780 option, check the "Video size" section in the ffmpeg-utils manual.
32781 The default value is "320x240".
32782 rate, r
32784 Specify the frame rate of the sourced video, as the number of
32785 frames generated per second. It has to be a string in the format
32786 frame_rate_num/frame_rate_den, an integer number, a floating point
32787 number or a valid video frame rate abbreviation. The default value
32788 is "25".
32789 sar Set the sample aspect ratio of the sourced video.
32791 duration, d
32793 Set the duration of the sourced video. See the Time duration
32794 section in the ffmpeg-utils(1) manual for the accepted syntax.
32795 If not specified, or the expressed duration is negative, the video
32797 is supposed to be generated forever.
32798 Additionally, all options of the coreimage video filter are accepted.
32800 A complete filterchain can be used for further processing of the
32801 generated input without CPU-HOST transfer. See coreimage documentation
32802 and examples for details.
32803 Examples
32805 • Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
32807 given as complete and escaped command-line for Apple's standard
32808 bash shell:
32809 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@inputMessage=https\\\\\://FFmpeg.org/@inputCorrectionLevel=H -frames:v 1 QRCode.png
32811 This example is equivalent to the QRCode example of coreimage
32813 without the need for a nullsrc video source.
32814 gradients
32816 Generate several gradients.
32817 size, s
32819 Set frame size. For the syntax of this option, check the "Video
32820 size" section in the ffmpeg-utils manual. Default value is
32821 "640x480".
32822 rate, r
32824 Set frame rate, expressed as number of frames per second. Default
32825 value is "25".
32826 c0, c1, c2, c3, c4, c5, c6, c7
32828 Set 8 colors. Default values for colors is to pick random one.
32829 x0, y0, y0, y1
32831 Set gradient line source and destination points. If negative or out
32832 of range, random ones are picked.
32833 nb_colors, n
32835 Set number of colors to use at once. Allowed range is from 2 to 8.
32836 Default value is 2.
32837 seed
32839 Set seed for picking gradient line points.
32840 duration, d
32842 Set the duration of the sourced video. See the Time duration
32843 section in the ffmpeg-utils(1) manual for the accepted syntax.
32844 If not specified, or the expressed duration is negative, the video
32846 is supposed to be generated forever.
32847 speed
32849 Set speed of gradients rotation.
32850 type, t
32852 Set type of gradients, can be "linear" or "radial" or "circular" or
32853 "spiral".
32854 mandelbrot
32856 Generate a Mandelbrot set fractal, and progressively zoom towards the
32857 point specified with start_x and start_y.
32858 This source accepts the following options:
32860 end_pts
32862 Set the terminal pts value. Default value is 400.
32863 end_scale
32865 Set the terminal scale value. Must be a floating point value.
32866 Default value is 0.3.
32867 inner
32869 Set the inner coloring mode, that is the algorithm used to draw the
32870 Mandelbrot fractal internal region.
32871 It shall assume one of the following values:
32873 black
32875 Set black mode.
32876 convergence
32878 Show time until convergence.
32879 mincol
32881 Set color based on point closest to the origin of the
32882 iterations.
32883 period
32885 Set period mode.
32886 Default value is mincol.
32888 bailout
32890 Set the bailout value. Default value is 10.0.
32891 maxiter
32893 Set the maximum of iterations performed by the rendering algorithm.
32894 Default value is 7189.
32895 outer
32897 Set outer coloring mode. It shall assume one of following values:
32898 iteration_count
32900 Set iteration count mode.
32901 normalized_iteration_count
32903 set normalized iteration count mode.
32904 Default value is normalized_iteration_count.
32906 rate, r
32908 Set frame rate, expressed as number of frames per second. Default
32909 value is "25".
32910 size, s
32912 Set frame size. For the syntax of this option, check the "Video
32913 size" section in the ffmpeg-utils manual. Default value is
32914 "640x480".
32915 start_scale
32917 Set the initial scale value. Default value is 3.0.
32918 start_x
32920 Set the initial x position. Must be a floating point value between
32921 -100 and 100. Default value is
32922 -0.743643887037158704752191506114774.
32923 start_y
32925 Set the initial y position. Must be a floating point value between
32926 -100 and 100. Default value is
32927 -0.131825904205311970493132056385139.
32928 mptestsrc
32930 Generate various test patterns, as generated by the MPlayer test
32931 filter.
32932 The size of the generated video is fixed, and is 256x256. This source
32934 is useful in particular for testing encoding features.
32935 This source accepts the following options:
32937 rate, r
32939 Specify the frame rate of the sourced video, as the number of
32940 frames generated per second. It has to be a string in the format
32941 frame_rate_num/frame_rate_den, an integer number, a floating point
32942 number or a valid video frame rate abbreviation. The default value
32943 is "25".
32944 duration, d
32946 Set the duration of the sourced video. See the Time duration
32947 section in the ffmpeg-utils(1) manual for the accepted syntax.
32948 If not specified, or the expressed duration is negative, the video
32950 is supposed to be generated forever.
32951 test, t
32953 Set the number or the name of the test to perform. Supported tests
32954 are:
32955 dc_luma
32957 dc_chroma
32958 freq_luma
32959 freq_chroma
32960 amp_luma
32961 amp_chroma
32962 cbp
32963 mv
32964 ring1
32965 ring2
32966 all
32967 max_frames, m
32968 Set the maximum number of frames generated for each test,
32969 default value is 30.
32970 Default value is "all", which will cycle through the list of all
32972 tests.
32973 Some examples:
32975 mptestsrc=t=dc_luma
32977 will generate a "dc_luma" test pattern.
32979 frei0r_src
32981 Provide a frei0r source.
32982 To enable compilation of this filter you need to install the frei0r
32984 header and configure FFmpeg with "--enable-frei0r".
32985 This source accepts the following parameters:
32987 size
32989 The size of the video to generate. For the syntax of this option,
32990 check the "Video size" section in the ffmpeg-utils manual.
32991 framerate
32993 The framerate of the generated video. It may be a string of the
32994 form num/den or a frame rate abbreviation.
32995 filter_name
32997 The name to the frei0r source to load. For more information
32998 regarding frei0r and how to set the parameters, read the frei0r
32999 section in the video filters documentation.
33000 filter_params
33002 A '|'-separated list of parameters to pass to the frei0r source.
33003 For example, to generate a frei0r partik0l source with size 200x200 and
33005 frame rate 10 which is overlaid on the overlay filter main input:
33006 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
33008 life
33010 Generate a life pattern.
33011 This source is based on a generalization of John Conway's life game.
33013 The sourced input represents a life grid, each pixel represents a cell
33015 which can be in one of two possible states, alive or dead. Every cell
33016 interacts with its eight neighbours, which are the cells that are
33017 horizontally, vertically, or diagonally adjacent.
33018 At each interaction the grid evolves according to the adopted rule,
33020 which specifies the number of neighbor alive cells which will make a
33021 cell stay alive or born. The rule option allows one to specify the rule
33022 to adopt.
33023 This source accepts the following options:
33025 filename, f
33027 Set the file from which to read the initial grid state. In the
33028 file, each non-whitespace character is considered an alive cell,
33029 and newline is used to delimit the end of each row.
33030 If this option is not specified, the initial grid is generated
33032 randomly.
33033 rate, r
33035 Set the video rate, that is the number of frames generated per
33036 second. Default is 25.
33037 random_fill_ratio, ratio
33039 Set the random fill ratio for the initial random grid. It is a
33040 floating point number value ranging from 0 to 1, defaults to 1/PHI.
33041 It is ignored when a file is specified.
33042 random_seed, seed
33044 Set the seed for filling the initial random grid, must be an
33045 integer included between 0 and UINT32_MAX. If not specified, or if
33046 explicitly set to -1, the filter will try to use a good random seed
33047 on a best effort basis.
33048 rule
33050 Set the life rule.
33051 A rule can be specified with a code of the kind "SNS/BNB", where NS
33053 and NB are sequences of numbers in the range 0-8, NS specifies the
33054 number of alive neighbor cells which make a live cell stay alive,
33055 and NB the number of alive neighbor cells which make a dead cell to
33056 become alive (i.e. to "born"). "s" and "b" can be used in place of
33057 "S" and "B", respectively.
33058 Alternatively a rule can be specified by an 18-bits integer. The 9
33060 high order bits are used to encode the next cell state if it is
33061 alive for each number of neighbor alive cells, the low order bits
33062 specify the rule for "borning" new cells. Higher order bits encode
33063 for an higher number of neighbor cells. For example the number
33064 6153 = "(12<<9)+9" specifies a stay alive rule of 12 and a born
33065 rule of 9, which corresponds to "S23/B03".
33066 Default value is "S23/B3", which is the original Conway's game of
33068 life rule, and will keep a cell alive if it has 2 or 3 neighbor
33069 alive cells, and will born a new cell if there are three alive
33070 cells around a dead cell.
33071 size, s
33073 Set the size of the output video. For the syntax of this option,
33074 check the "Video size" section in the ffmpeg-utils manual.
33075 If filename is specified, the size is set by default to the same
33077 size of the input file. If size is set, it must contain the size
33078 specified in the input file, and the initial grid defined in that
33079 file is centered in the larger resulting area.
33080 If a filename is not specified, the size value defaults to
33082 "320x240" (used for a randomly generated initial grid).
33083 stitch
33085 If set to 1, stitch the left and right grid edges together, and the
33086 top and bottom edges also. Defaults to 1.
33087 mold
33089 Set cell mold speed. If set, a dead cell will go from death_color
33090 to mold_color with a step of mold. mold can have a value from 0 to
33091 255.
33092 life_color
33094 Set the color of living (or new born) cells.
33095 death_color
33097 Set the color of dead cells. If mold is set, this is the first
33098 color used to represent a dead cell.
33099 mold_color
33101 Set mold color, for definitely dead and moldy cells.
33102 For the syntax of these 3 color options, check the "Color" section
33104 in the ffmpeg-utils manual.
33105 Examples
33107 • Read a grid from pattern, and center it on a grid of size 300x300
33109 pixels:
33110 life=f=pattern:s=300x300
33112 • Generate a random grid of size 200x200, with a fill ratio of 2/3:
33114 life=ratio=2/3:s=200x200
33116 • Specify a custom rule for evolving a randomly generated grid:
33118 life=rule=S14/B34
33120 • Full example with slow death effect (mold) using ffplay:
33122 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
33124 allrgb, allyuv, color, colorchart, colorspectrum, haldclutsrc, nullsrc,
33126 pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc,
33127 testsrc2, yuvtestsrc
33128 The "allrgb" source returns frames of size 4096x4096 of all rgb colors.
33129 The "allyuv" source returns frames of size 4096x4096 of all yuv colors.
33131 The "color" source provides an uniformly colored input.
33133 The "colorchart" source provides a colors checker chart.
33135 The "colorspectrum" source provides a color spectrum input.
33137 The "haldclutsrc" source provides an identity Hald CLUT. See also
33139 haldclut filter.
33140 The "nullsrc" source returns unprocessed video frames. It is mainly
33142 useful to be employed in analysis / debugging tools, or as the source
33143 for filters which ignore the input data.
33144 The "pal75bars" source generates a color bars pattern, based on EBU PAL
33146 recommendations with 75% color levels.
33147 The "pal100bars" source generates a color bars pattern, based on EBU
33149 PAL recommendations with 100% color levels.
33150 The "rgbtestsrc" source generates an RGB test pattern useful for
33152 detecting RGB vs BGR issues. You should see a red, green and blue
33153 stripe from top to bottom.
33154 The "smptebars" source generates a color bars pattern, based on the
33156 SMPTE Engineering Guideline EG 1-1990.
33157 The "smptehdbars" source generates a color bars pattern, based on the
33159 SMPTE RP 219-2002.
33160 The "testsrc" source generates a test video pattern, showing a color
33162 pattern, a scrolling gradient and a timestamp. This is mainly intended
33163 for testing purposes.
33164 The "testsrc2" source is similar to testsrc, but supports more pixel
33166 formats instead of just "rgb24". This allows using it as an input for
33167 other tests without requiring a format conversion.
33168 The "yuvtestsrc" source generates an YUV test pattern. You should see a
33170 y, cb and cr stripe from top to bottom.
33171 The sources accept the following parameters:
33173 level
33175 Specify the level of the Hald CLUT, only available in the
33176 "haldclutsrc" source. A level of "N" generates a picture of "N*N*N"
33177 by "N*N*N" pixels to be used as identity matrix for 3D lookup
33178 tables. Each component is coded on a "1/(N*N)" scale.
33179 color, c
33181 Specify the color of the source, only available in the "color"
33182 source. For the syntax of this option, check the "Color" section in
33183 the ffmpeg-utils manual.
33184 size, s
33186 Specify the size of the sourced video. For the syntax of this
33187 option, check the "Video size" section in the ffmpeg-utils manual.
33188 The default value is "320x240".
33189 This option is not available with the "allrgb", "allyuv", and
33191 "haldclutsrc" filters.
33192 rate, r
33194 Specify the frame rate of the sourced video, as the number of
33195 frames generated per second. It has to be a string in the format
33196 frame_rate_num/frame_rate_den, an integer number, a floating point
33197 number or a valid video frame rate abbreviation. The default value
33198 is "25".
33199 duration, d
33201 Set the duration of the sourced video. See the Time duration
33202 section in the ffmpeg-utils(1) manual for the accepted syntax.
33203 If not specified, or the expressed duration is negative, the video
33205 is supposed to be generated forever.
33206 Since the frame rate is used as time base, all frames including the
33208 last one will have their full duration. If the specified duration
33209 is not a multiple of the frame duration, it will be rounded up.
33210 sar Set the sample aspect ratio of the sourced video.
33212 alpha
33214 Specify the alpha (opacity) of the background, only available in
33215 the "testsrc2" source. The value must be between 0 (fully
33216 transparent) and 255 (fully opaque, the default).
33217 decimals, n
33219 Set the number of decimals to show in the timestamp, only available
33220 in the "testsrc" source.
33221 The displayed timestamp value will correspond to the original
33223 timestamp value multiplied by the power of 10 of the specified
33224 value. Default value is 0.
33225 type
33227 Set the type of the color spectrum, only available in the
33228 "colorspectrum" source. Can be one of the following:
33229 black
33231 white
33232 all
33233 patch_size
33234 Set patch size of single color patch, only available in the
33235 "colorchart" source. Default is "64x64".
33236 preset
33238 Set colorchecker colors preset, only available in the "colorchart"
33239 source.
33240 Available values are:
33242 reference
33244 skintones
33245 Default value is "reference".
33247 Examples
33249 • Generate a video with a duration of 5.3 seconds, with size 176x144
33251 and a frame rate of 10 frames per second:
33252 testsrc=duration=5.3:size=qcif:rate=10
33254 • The following graph description will generate a red source with an
33256 opacity of 0.2, with size "qcif" and a frame rate of 10 frames per
33257 second:
33258 color=c=red@0.2:s=qcif:r=10
33260 • If the input content is to be ignored, "nullsrc" can be used. The
33262 following command generates noise in the luminance plane by
33263 employing the "geq" filter:
33264 nullsrc=s=256x256, geq=random(1)*255:128:128
33266 Commands
33268 The "color" source supports the following commands:
33270 c, color
33272 Set the color of the created image. Accepts the same syntax of the
33273 corresponding color option.
33274 openclsrc
33276 Generate video using an OpenCL program.
33277 source
33279 OpenCL program source file.
33280 kernel
33282 Kernel name in program.
33283 size, s
33285 Size of frames to generate. This must be set.
33286 format
33288 Pixel format to use for the generated frames. This must be set.
33289 rate, r
33291 Number of frames generated every second. Default value is '25'.
33292 For details of how the program loading works, see the program_opencl
33294 filter.
33295 Example programs:
33297 • Generate a colour ramp by setting pixel values from the position of
33299 the pixel in the output image. (Note that this will work with all
33300 pixel formats, but the generated output will not be the same.)
33301 __kernel void ramp(__write_only image2d_t dst,
33303 unsigned int index)
33304 {
33305 int2 loc = (int2)(get_global_id(0), get_global_id(1));
33306 float4 val;
33308 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
33309 write_imagef(dst, loc, val);
33311 }
33312 • Generate a Sierpinski carpet pattern, panning by a single pixel
33314 each frame.
33315 __kernel void sierpinski_carpet(__write_only image2d_t dst,
33317 unsigned int index)
33318 {
33319 int2 loc = (int2)(get_global_id(0), get_global_id(1));
33320 float4 value = 0.0f;
33322 int x = loc.x + index;
33323 int y = loc.y + index;
33324 while (x > 0 || y > 0) {
33325 if (x % 3 == 1 && y % 3 == 1) {
33326 value = 1.0f;
33327 break;
33328 }
33329 x /= 3;
33330 y /= 3;
33331 }
33332 write_imagef(dst, loc, value);
33334 }
33335 sierpinski
33337 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
33338 This source accepts the following options:
33340 size, s
33342 Set frame size. For the syntax of this option, check the "Video
33343 size" section in the ffmpeg-utils manual. Default value is
33344 "640x480".
33345 rate, r
33347 Set frame rate, expressed as number of frames per second. Default
33348 value is "25".
33349 seed
33351 Set seed which is used for random panning.
33352 jump
33354 Set max jump for single pan destination. Allowed range is from 1 to
33355 10000.
33356 type
33358 Set fractal type, can be default "carpet" or "triangle".
33359
33361 Below is a description of the currently available video sinks.
33362 buffersink
33364 Buffer video frames, and make them available to the end of the filter
33365 graph.
33366 This sink is mainly intended for programmatic use, in particular
33368 through the interface defined in libavfilter/buffersink.h or the
33369 options system.
33370 It accepts a pointer to an AVBufferSinkContext structure, which defines
33372 the incoming buffers' formats, to be passed as the opaque parameter to
33373 "avfilter_init_filter" for initialization.
33374 nullsink
33376 Null video sink: do absolutely nothing with the input video. It is
33377 mainly useful as a template and for use in analysis / debugging tools.
33378
33380 Below is a description of the currently available multimedia filters.
33381 abitscope
33383 Convert input audio to a video output, displaying the audio bit scope.
33384 The filter accepts the following options:
33386 rate, r
33388 Set frame rate, expressed as number of frames per second. Default
33389 value is "25".
33390 size, s
33392 Specify the video size for the output. For the syntax of this
33393 option, check the "Video size" section in the ffmpeg-utils manual.
33394 Default value is "1024x256".
33395 colors
33397 Specify list of colors separated by space or by '|' which will be
33398 used to draw channels. Unrecognized or missing colors will be
33399 replaced by white color.
33400 mode, m
33402 Set output mode. Can be "bars" or "trace". Default is "bars".
33403 adrawgraph
33405 Draw a graph using input audio metadata.
33406 See drawgraph
33408 agraphmonitor
33410 See graphmonitor.
33411 ahistogram
33413 Convert input audio to a video output, displaying the volume histogram.
33414 The filter accepts the following options:
33416 dmode
33418 Specify how histogram is calculated.
33419 It accepts the following values:
33421 single
33423 Use single histogram for all channels.
33424 separate
33426 Use separate histogram for each channel.
33427 Default is "single".
33429 rate, r
33431 Set frame rate, expressed as number of frames per second. Default
33432 value is "25".
33433 size, s
33435 Specify the video size for the output. For the syntax of this
33436 option, check the "Video size" section in the ffmpeg-utils manual.
33437 Default value is "hd720".
33438 scale
33440 Set display scale.
33441 It accepts the following values:
33443 log logarithmic
33445 sqrt
33447 square root
33448 cbrt
33450 cubic root
33451 lin linear
33453 rlog
33455 reverse logarithmic
33456 Default is "log".
33458 ascale
33460 Set amplitude scale.
33461 It accepts the following values:
33463 log logarithmic
33465 lin linear
33467 Default is "log".
33469 acount
33471 Set how much frames to accumulate in histogram. Default is 1.
33472 Setting this to -1 accumulates all frames.
33473 rheight
33475 Set histogram ratio of window height.
33476 slide
33478 Set sonogram sliding.
33479 It accepts the following values:
33481 replace
33483 replace old rows with new ones.
33484 scroll
33486 scroll from top to bottom.
33487 Default is "replace".
33489 hmode
33491 Set histogram mode.
33492 It accepts the following values:
33494 abs Use absolute values of samples.
33496 sign
33498 Use untouched values of samples.
33499 Default is "abs".
33501 aphasemeter
33503 Measures phase of input audio, which is exported as metadata
33504 "lavfi.aphasemeter.phase", representing mean phase of current audio
33505 frame. A video output can also be produced and is enabled by default.
33506 The audio is passed through as first output.
33507 Audio will be rematrixed to stereo if it has a different channel
33509 layout. Phase value is in range "[-1, 1]" where "-1" means left and
33510 right channels are completely out of phase and 1 means channels are in
33511 phase.
33512 The filter accepts the following options, all related to its video
33514 output:
33515 rate, r
33517 Set the output frame rate. Default value is 25.
33518 size, s
33520 Set the video size for the output. For the syntax of this option,
33521 check the "Video size" section in the ffmpeg-utils manual. Default
33522 value is "800x400".
33523 rc
33525 gc
33526 bc Specify the red, green, blue contrast. Default values are 2, 7 and
33527 1. Allowed range is "[0, 255]".
33528 mpc Set color which will be used for drawing median phase. If color is
33530 "none" which is default, no median phase value will be drawn.
33531 video
33533 Enable video output. Default is enabled.
33534 phasing detection
33536 The filter also detects out of phase and mono sequences in stereo
33538 streams. It logs the sequence start, end and duration when it lasts
33539 longer or as long as the minimum set.
33540 The filter accepts the following options for this detection:
33542 phasing
33544 Enable mono and out of phase detection. Default is disabled.
33545 tolerance, t
33547 Set phase tolerance for mono detection, in amplitude ratio. Default
33548 is 0. Allowed range is "[0, 1]".
33549 angle, a
33551 Set angle threshold for out of phase detection, in degree. Default
33552 is 170. Allowed range is "[90, 180]".
33553 duration, d
33555 Set mono or out of phase duration until notification, expressed in
33556 seconds. Default is 2.
33557 Examples
33559 • Complete example with ffmpeg to detect 1 second of mono with 0.001
33561 phase tolerance:
33562 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
33564 avectorscope
33566 Convert input audio to a video output, representing the audio vector
33567 scope.
33568 The filter is used to measure the difference between channels of stereo
33570 audio stream. A monaural signal, consisting of identical left and right
33571 signal, results in straight vertical line. Any stereo separation is
33572 visible as a deviation from this line, creating a Lissajous figure. If
33573 the straight (or deviation from it) but horizontal line appears this
33574 indicates that the left and right channels are out of phase.
33575 The filter accepts the following options:
33577 mode, m
33579 Set the vectorscope mode.
33580 Available values are:
33582 lissajous
33584 Lissajous rotated by 45 degrees.
33585 lissajous_xy
33587 Same as above but not rotated.
33588 polar
33590 Shape resembling half of circle.
33591 Default value is lissajous.
33593 size, s
33595 Set the video size for the output. For the syntax of this option,
33596 check the "Video size" section in the ffmpeg-utils manual. Default
33597 value is "400x400".
33598 rate, r
33600 Set the output frame rate. Default value is 25.
33601 rc
33603 gc
33604 bc
33605 ac Specify the red, green, blue and alpha contrast. Default values are
33606 40, 160, 80 and 255. Allowed range is "[0, 255]".
33607 rf
33609 gf
33610 bf
33611 af Specify the red, green, blue and alpha fade. Default values are 15,
33612 10, 5 and 5. Allowed range is "[0, 255]".
33613 zoom
33615 Set the zoom factor. Default value is 1. Allowed range is "[0,
33616 10]". Values lower than 1 will auto adjust zoom factor to maximal
33617 possible value.
33618 draw
33620 Set the vectorscope drawing mode.
33621 Available values are:
33623 dot Draw dot for each sample.
33625 line
33627 Draw line between previous and current sample.
33628 Default value is dot.
33630 scale
33632 Specify amplitude scale of audio samples.
33633 Available values are:
33635 lin Linear.
33637 sqrt
33639 Square root.
33640 cbrt
33642 Cubic root.
33643 log Logarithmic.
33645 swap
33647 Swap left channel axis with right channel axis.
33648 mirror
33650 Mirror axis.
33651 none
33653 No mirror.
33654 x Mirror only x axis.
33656 y Mirror only y axis.
33658 xy Mirror both axis.
33660 Examples
33662 • Complete example using ffplay:
33664 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
33666 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
33667 Commands
33669 This filter supports the all above options as commands except options
33671 "size" and "rate".
33672 bench, abench
33674 Benchmark part of a filtergraph.
33675 The filter accepts the following options:
33677 action
33679 Start or stop a timer.
33680 Available values are:
33682 start
33684 Get the current time, set it as frame metadata (using the key
33685 "lavfi.bench.start_time"), and forward the frame to the next
33686 filter.
33687 stop
33689 Get the current time and fetch the "lavfi.bench.start_time"
33690 metadata from the input frame metadata to get the time
33691 difference. Time difference, average, maximum and minimum time
33692 (respectively "t", "avg", "max" and "min") are then printed.
33693 The timestamps are expressed in seconds.
33694 Examples
33696 • Benchmark selectivecolor filter:
33698 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
33700 concat
33702 Concatenate audio and video streams, joining them together one after
33703 the other.
33704 The filter works on segments of synchronized video and audio streams.
33706 All segments must have the same number of streams of each type, and
33707 that will also be the number of streams at output.
33708 The filter accepts the following options:
33710 n Set the number of segments. Default is 2.
33712 v Set the number of output video streams, that is also the number of
33714 video streams in each segment. Default is 1.
33715 a Set the number of output audio streams, that is also the number of
33717 audio streams in each segment. Default is 0.
33718 unsafe
33720 Activate unsafe mode: do not fail if segments have a different
33721 format.
33722 The filter has v+a outputs: first v video outputs, then a audio
33724 outputs.
33725 There are nx(v+a) inputs: first the inputs for the first segment, in
33727 the same order as the outputs, then the inputs for the second segment,
33728 etc.
33729 Related streams do not always have exactly the same duration, for
33731 various reasons including codec frame size or sloppy authoring. For
33732 that reason, related synchronized streams (e.g. a video and its audio
33733 track) should be concatenated at once. The concat filter will use the
33734 duration of the longest stream in each segment (except the last one),
33735 and if necessary pad shorter audio streams with silence.
33736 For this filter to work correctly, all segments must start at timestamp
33738 0.
33739 All corresponding streams must have the same parameters in all
33741 segments; the filtering system will automatically select a common pixel
33742 format for video streams, and a common sample format, sample rate and
33743 channel layout for audio streams, but other settings, such as
33744 resolution, must be converted explicitly by the user.
33745 Different frame rates are acceptable but will result in variable frame
33747 rate at output; be sure to configure the output file to handle it.
33748 Examples
33750 • Concatenate an opening, an episode and an ending, all in bilingual
33752 version (video in stream 0, audio in streams 1 and 2):
33753 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
33755 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
33756 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
33757 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
33758 • Concatenate two parts, handling audio and video separately, using
33760 the (a)movie sources, and adjusting the resolution:
33761 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
33763 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
33764 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
33765 Note that a desync will happen at the stitch if the audio and video
33767 streams do not have exactly the same duration in the first file.
33768 Commands
33770 This filter supports the following commands:
33772 next
33774 Close the current segment and step to the next one
33775 ebur128
33777 EBU R128 scanner filter. This filter takes an audio stream and analyzes
33778 its loudness level. By default, it logs a message at a frequency of
33779 10Hz with the Momentary loudness (identified by "M"), Short-term
33780 loudness ("S"), Integrated loudness ("I") and Loudness Range ("LRA").
33781 The filter can only analyze streams which have sample format is double-
33783 precision floating point. The input stream will be converted to this
33784 specification, if needed. Users may need to insert aformat and/or
33785 aresample filters after this filter to obtain the original parameters.
33786 The filter also has a video output (see the video option) with a real
33788 time graph to observe the loudness evolution. The graphic contains the
33789 logged message mentioned above, so it is not printed anymore when this
33790 option is set, unless the verbose logging is set. The main graphing
33791 area contains the short-term loudness (3 seconds of analysis), and the
33792 gauge on the right is for the momentary loudness (400 milliseconds),
33793 but can optionally be configured to instead display short-term loudness
33794 (see gauge).
33795 The green area marks a +/- 1LU target range around the target loudness
33797 (-23LUFS by default, unless modified through target).
33798 More information about the Loudness Recommendation EBU R128 on
33800 <http://tech.ebu.ch/loudness>.
33801 The filter accepts the following options:
33803 video
33805 Activate the video output. The audio stream is passed unchanged
33806 whether this option is set or no. The video stream will be the
33807 first output stream if activated. Default is 0.
33808 size
33810 Set the video size. This option is for video only. For the syntax
33811 of this option, check the "Video size" section in the ffmpeg-utils
33812 manual. Default and minimum resolution is "640x480".
33813 meter
33815 Set the EBU scale meter. Default is 9. Common values are 9 and 18,
33816 respectively for EBU scale meter +9 and EBU scale meter +18. Any
33817 other integer value between this range is allowed.
33818 metadata
33820 Set metadata injection. If set to 1, the audio input will be
33821 segmented into 100ms output frames, each of them containing various
33822 loudness information in metadata. All the metadata keys are
33823 prefixed with "lavfi.r128.".
33824 Default is 0.
33826 framelog
33828 Force the frame logging level.
33829 Available values are:
33831 info
33833 information logging level
33834 verbose
33836 verbose logging level
33837 By default, the logging level is set to info. If the video or the
33839 metadata options are set, it switches to verbose.
33840 peak
33842 Set peak mode(s).
33843 Available modes can be cumulated (the option is a "flag" type).
33845 Possible values are:
33846 none
33848 Disable any peak mode (default).
33849 sample
33851 Enable sample-peak mode.
33852 Simple peak mode looking for the higher sample value. It logs a
33854 message for sample-peak (identified by "SPK").
33855 true
33857 Enable true-peak mode.
33858 If enabled, the peak lookup is done on an over-sampled version
33860 of the input stream for better peak accuracy. It logs a message
33861 for true-peak. (identified by "TPK") and true-peak per frame
33862 (identified by "FTPK"). This mode requires a build with
33863 "libswresample".
33864 dualmono
33866 Treat mono input files as "dual mono". If a mono file is intended
33867 for playback on a stereo system, its EBU R128 measurement will be
33868 perceptually incorrect. If set to "true", this option will
33869 compensate for this effect. Multi-channel input files are not
33870 affected by this option.
33871 panlaw
33873 Set a specific pan law to be used for the measurement of dual mono
33874 files. This parameter is optional, and has a default value of
33875 -3.01dB.
33876 target
33878 Set a specific target level (in LUFS) used as relative zero in the
33879 visualization. This parameter is optional and has a default value
33880 of -23LUFS as specified by EBU R128. However, material published
33881 online may prefer a level of -16LUFS (e.g. for use with podcasts or
33882 video platforms).
33883 gauge
33885 Set the value displayed by the gauge. Valid values are "momentary"
33886 and s "shortterm". By default the momentary value will be used, but
33887 in certain scenarios it may be more useful to observe the short
33888 term value instead (e.g. live mixing).
33889 scale
33891 Sets the display scale for the loudness. Valid parameters are
33892 "absolute" (in LUFS) or "relative" (LU) relative to the target.
33893 This only affects the video output, not the summary or continuous
33894 log output.
33895 Examples
33897 • Real-time graph using ffplay, with a EBU scale meter +18:
33899 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
33901 • Run an analysis with ffmpeg:
33903 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
33905 interleave, ainterleave
33907 Temporally interleave frames from several inputs.
33908 "interleave" works with video inputs, "ainterleave" with audio.
33910 These filters read frames from several inputs and send the oldest
33912 queued frame to the output.
33913 Input streams must have well defined, monotonically increasing frame
33915 timestamp values.
33916 In order to submit one frame to output, these filters need to enqueue
33918 at least one frame for each input, so they cannot work in case one
33919 input is not yet terminated and will not receive incoming frames.
33920 For example consider the case when one input is a "select" filter which
33922 always drops input frames. The "interleave" filter will keep reading
33923 from that input, but it will never be able to send new frames to output
33924 until the input sends an end-of-stream signal.
33925 Also, depending on inputs synchronization, the filters will drop frames
33927 in case one input receives more frames than the other ones, and the
33928 queue is already filled.
33929 These filters accept the following options:
33931 nb_inputs, n
33933 Set the number of different inputs, it is 2 by default.
33934 duration
33936 How to determine the end-of-stream.
33937 longest
33939 The duration of the longest input. (default)
33940 shortest
33942 The duration of the shortest input.
33943 first
33945 The duration of the first input.
33946 Examples
33948 • Interleave frames belonging to different streams using ffmpeg:
33950 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
33952 • Add flickering blur effect:
33954 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
33956 latency, alatency
33958 Measure filtering latency.
33959 Report previous filter filtering latency, delay in number of audio
33961 samples for audio filters or number of video frames for video filters.
33962 On end of input stream, filter will report min and max measured latency
33964 for previous running filter in filtergraph.
33965 metadata, ametadata
33967 Manipulate frame metadata.
33968 This filter accepts the following options:
33970 mode
33972 Set mode of operation of the filter.
33973 Can be one of the following:
33975 select
33977 If both "value" and "key" is set, select frames which have such
33978 metadata. If only "key" is set, select every frame that has
33979 such key in metadata.
33980 add Add new metadata "key" and "value". If key is already available
33982 do nothing.
33983 modify
33985 Modify value of already present key.
33986 delete
33988 If "value" is set, delete only keys that have such value.
33989 Otherwise, delete key. If "key" is not set, delete all metadata
33990 values in the frame.
33991 print
33993 Print key and its value if metadata was found. If "key" is not
33994 set print all metadata values available in frame.
33995 key Set key used with all modes. Must be set for all modes except
33997 "print" and "delete".
33998 value
34000 Set metadata value which will be used. This option is mandatory for
34001 "modify" and "add" mode.
34002 function
34004 Which function to use when comparing metadata value and "value".
34005 Can be one of following:
34007 same_str
34009 Values are interpreted as strings, returns true if metadata
34010 value is same as "value".
34011 starts_with
34013 Values are interpreted as strings, returns true if metadata
34014 value starts with the "value" option string.
34015 less
34017 Values are interpreted as floats, returns true if metadata
34018 value is less than "value".
34019 equal
34021 Values are interpreted as floats, returns true if "value" is
34022 equal with metadata value.
34023 greater
34025 Values are interpreted as floats, returns true if metadata
34026 value is greater than "value".
34027 expr
34029 Values are interpreted as floats, returns true if expression
34030 from option "expr" evaluates to true.
34031 ends_with
34033 Values are interpreted as strings, returns true if metadata
34034 value ends with the "value" option string.
34035 expr
34037 Set expression which is used when "function" is set to "expr". The
34038 expression is evaluated through the eval API and can contain the
34039 following constants:
34040 VALUE1, FRAMEVAL
34042 Float representation of "value" from metadata key.
34043 VALUE2, USERVAL
34045 Float representation of "value" as supplied by user in "value"
34046 option.
34047 file
34049 If specified in "print" mode, output is written to the named file.
34050 Instead of plain filename any writable url can be specified.
34051 Filename ``-'' is a shorthand for standard output. If "file" option
34052 is not set, output is written to the log with AV_LOG_INFO loglevel.
34053 direct
34055 Reduces buffering in print mode when output is written to a URL set
34056 using file.
34057 Examples
34059 • Print all metadata values for frames with key
34061 "lavfi.signalstats.YDIF" with values between 0 and 1.
34062 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
34064 • Print silencedetect output to file metadata.txt.
34066 silencedetect,ametadata=mode=print:file=metadata.txt
34068 • Direct all metadata to a pipe with file descriptor 4.
34070 metadata=mode=print:file='pipe\:4'
34072 perms, aperms
34074 Set read/write permissions for the output frames.
34075 These filters are mainly aimed at developers to test direct path in the
34077 following filter in the filtergraph.
34078 The filters accept the following options:
34080 mode
34082 Select the permissions mode.
34083 It accepts the following values:
34085 none
34087 Do nothing. This is the default.
34088 ro Set all the output frames read-only.
34090 rw Set all the output frames directly writable.
34092 toggle
34094 Make the frame read-only if writable, and writable if read-
34095 only.
34096 random
34098 Set each output frame read-only or writable randomly.
34099 seed
34101 Set the seed for the random mode, must be an integer included
34102 between 0 and "UINT32_MAX". If not specified, or if explicitly set
34103 to "-1", the filter will try to use a good random seed on a best
34104 effort basis.
34105 Note: in case of auto-inserted filter between the permission filter and
34107 the following one, the permission might not be received as expected in
34108 that following filter. Inserting a format or aformat filter before the
34109 perms/aperms filter can avoid this problem.
34110 realtime, arealtime
34112 Slow down filtering to match real time approximately.
34113 These filters will pause the filtering for a variable amount of time to
34115 match the output rate with the input timestamps. They are similar to
34116 the re option to "ffmpeg".
34117 They accept the following options:
34119 limit
34121 Time limit for the pauses. Any pause longer than that will be
34122 considered a timestamp discontinuity and reset the timer. Default
34123 is 2 seconds.
34124 speed
34126 Speed factor for processing. The value must be a float larger than
34127 zero. Values larger than 1.0 will result in faster than realtime
34128 processing, smaller will slow processing down. The limit is
34129 automatically adapted accordingly. Default is 1.0.
34130 A processing speed faster than what is possible without these
34132 filters cannot be achieved.
34133 Commands
34135 Both filters supports the all above options as commands.
34137 segment, asegment
34139 Split single input stream into multiple streams.
34140 This filter does opposite of concat filters.
34142 "segment" works on video frames, "asegment" on audio samples.
34144 This filter accepts the following options:
34146 timestamps
34148 Timestamps of output segments separated by '|'. The first segment
34149 will run from the beginning of the input stream. The last segment
34150 will run until the end of the input stream
34151 frames, samples
34153 Exact frame/sample count to split the segments.
34154 In all cases, prefixing an each segment with '+' will make it relative
34156 to the previous segment.
34157 Examples
34159 • Split input audio stream into three output audio streams, starting
34161 at start of input audio stream and storing that in 1st output audio
34162 stream, then following at 60th second and storing than in 2nd
34163 output audio stream, and last after 150th second of input audio
34164 stream store in 3rd output audio stream:
34165 asegment=timestamps="60|150"
34167 select, aselect
34169 Select frames to pass in output.
34170 This filter accepts the following options:
34172 expr, e
34174 Set expression, which is evaluated for each input frame.
34175 If the expression is evaluated to zero, the frame is discarded.
34177 If the evaluation result is negative or NaN, the frame is sent to
34179 the first output; otherwise it is sent to the output with index
34180 "ceil(val)-1", assuming that the input index starts from 0.
34181 For example a value of 1.2 corresponds to the output with index
34183 "ceil(1.2)-1 = 2-1 = 1", that is the second output.
34184 outputs, n
34186 Set the number of outputs. The output to which to send the selected
34187 frame is based on the result of the evaluation. Default value is 1.
34188 The expression can contain the following constants:
34190 n The (sequential) number of the filtered frame, starting from 0.
34192 selected_n
34194 The (sequential) number of the selected frame, starting from 0.
34195 prev_selected_n
34197 The sequential number of the last selected frame. It's NAN if
34198 undefined.
34199 TB The timebase of the input timestamps.
34201 pts The PTS (Presentation TimeStamp) of the filtered frame, expressed
34203 in TB units. It's NAN if undefined.
34204 t The PTS of the filtered frame, expressed in seconds. It's NAN if
34206 undefined.
34207 prev_pts
34209 The PTS of the previously filtered frame. It's NAN if undefined.
34210 prev_selected_pts
34212 The PTS of the last previously filtered frame. It's NAN if
34213 undefined.
34214 prev_selected_t
34216 The PTS of the last previously selected frame, expressed in
34217 seconds. It's NAN if undefined.
34218 start_pts
34220 The first PTS in the stream which is not NAN. It remains NAN if not
34221 found.
34222 start_t
34224 The first PTS, in seconds, in the stream which is not NAN. It
34225 remains NAN if not found.
34226 pict_type (video only)
34228 The type of the filtered frame. It can assume one of the following
34229 values:
34230 I
34232 P
34233 B
34234 S
34235 SI
34236 SP
34237 BI
34238 interlace_type (video only)
34239 The frame interlace type. It can assume one of the following
34240 values:
34241 PROGRESSIVE
34243 The frame is progressive (not interlaced).
34244 TOPFIRST
34246 The frame is top-field-first.
34247 BOTTOMFIRST
34249 The frame is bottom-field-first.
34250 consumed_sample_n (audio only)
34252 the number of selected samples before the current frame
34253 samples_n (audio only)
34255 the number of samples in the current frame
34256 sample_rate (audio only)
34258 the input sample rate
34259 key This is 1 if the filtered frame is a key-frame, 0 otherwise.
34261 pos the position in the file of the filtered frame, -1 if the
34263 information is not available (e.g. for synthetic video)
34264 scene (video only)
34266 value between 0 and 1 to indicate a new scene; a low value reflects
34267 a low probability for the current frame to introduce a new scene,
34268 while a higher value means the current frame is more likely to be
34269 one (see the example below)
34270 concatdec_select
34272 The concat demuxer can select only part of a concat input file by
34273 setting an inpoint and an outpoint, but the output packets may not
34274 be entirely contained in the selected interval. By using this
34275 variable, it is possible to skip frames generated by the concat
34276 demuxer which are not exactly contained in the selected interval.
34277 This works by comparing the frame pts against the
34279 lavf.concat.start_time and the lavf.concat.duration packet metadata
34280 values which are also present in the decoded frames.
34281 The concatdec_select variable is -1 if the frame pts is at least
34283 start_time and either the duration metadata is missing or the frame
34284 pts is less than start_time + duration, 0 otherwise, and NaN if the
34285 start_time metadata is missing.
34286 That basically means that an input frame is selected if its pts is
34288 within the interval set by the concat demuxer.
34289 The default value of the select expression is "1".
34291 Examples
34293 • Select all frames in input:
34295 select
34297 The example above is the same as:
34299 select=1
34301 • Skip all frames:
34303 select=0
34305 • Select only I-frames:
34307 select='eq(pict_type\,I)'
34309 • Select one frame every 100:
34311 select='not(mod(n\,100))'
34313 • Select only frames contained in the 10-20 time interval:
34315 select=between(t\,10\,20)
34317 • Select only I-frames contained in the 10-20 time interval:
34319 select=between(t\,10\,20)*eq(pict_type\,I)
34321 • Select frames with a minimum distance of 10 seconds:
34323 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
34325 • Use aselect to select only audio frames with samples number > 100:
34327 aselect='gt(samples_n\,100)'
34329 • Create a mosaic of the first scenes:
34331 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
34333 Comparing scene against a value between 0.3 and 0.5 is generally a
34335 sane choice.
34336 • Send even and odd frames to separate outputs, and compose them:
34338 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
34340 • Select useful frames from an ffconcat file which is using inpoints
34342 and outpoints but where the source files are not intra frame only.
34343 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
34345 sendcmd, asendcmd
34347 Send commands to filters in the filtergraph.
34348 These filters read commands to be sent to other filters in the
34350 filtergraph.
34351 "sendcmd" must be inserted between two video filters, "asendcmd" must
34353 be inserted between two audio filters, but apart from that they act the
34354 same way.
34355 The specification of commands can be provided in the filter arguments
34357 with the commands option, or in a file specified by the filename
34358 option.
34359 These filters accept the following options:
34361 commands, c
34363 Set the commands to be read and sent to the other filters.
34364 filename, f
34366 Set the filename of the commands to be read and sent to the other
34367 filters.
34368 Commands syntax
34370 A commands description consists of a sequence of interval
34372 specifications, comprising a list of commands to be executed when a
34373 particular event related to that interval occurs. The occurring event
34374 is typically the current frame time entering or leaving a given time
34375 interval.
34376 An interval is specified by the following syntax:
34378 <START>[-<END>] <COMMANDS>;
34380 The time interval is specified by the START and END times. END is
34382 optional and defaults to the maximum time.
34383 The current frame time is considered within the specified interval if
34385 it is included in the interval [START, END), that is when the time is
34386 greater or equal to START and is lesser than END.
34387 COMMANDS consists of a sequence of one or more command specifications,
34389 separated by ",", relating to that interval. The syntax of a command
34390 specification is given by:
34391 [<FLAGS>] <TARGET> <COMMAND> <ARG>
34393 FLAGS is optional and specifies the type of events relating to the time
34395 interval which enable sending the specified command, and must be a non-
34396 null sequence of identifier flags separated by "+" or "|" and enclosed
34397 between "[" and "]".
34398 The following flags are recognized:
34400 enter
34402 The command is sent when the current frame timestamp enters the
34403 specified interval. In other words, the command is sent when the
34404 previous frame timestamp was not in the given interval, and the
34405 current is.
34406 leave
34408 The command is sent when the current frame timestamp leaves the
34409 specified interval. In other words, the command is sent when the
34410 previous frame timestamp was in the given interval, and the current
34411 is not.
34412 expr
34414 The command ARG is interpreted as expression and result of
34415 expression is passed as ARG.
34416 The expression is evaluated through the eval API and can contain
34418 the following constants:
34419 POS Original position in the file of the frame, or undefined if
34421 undefined for the current frame.
34422 PTS The presentation timestamp in input.
34424 N The count of the input frame for video or audio, starting from
34426 0.
34427 T The time in seconds of the current frame.
34429 TS The start time in seconds of the current command interval.
34431 TE The end time in seconds of the current command interval.
34433 TI The interpolated time of the current command interval, TI = (T
34435 - TS) / (TE - TS).
34436 W The video frame width.
34438 H The video frame height.
34440 If FLAGS is not specified, a default value of "[enter]" is assumed.
34442 TARGET specifies the target of the command, usually the name of the
34444 filter class or a specific filter instance name.
34445 COMMAND specifies the name of the command for the target filter.
34447 ARG is optional and specifies the optional list of argument for the
34449 given COMMAND.
34450 Between one interval specification and another, whitespaces, or
34452 sequences of characters starting with "#" until the end of line, are
34453 ignored and can be used to annotate comments.
34454 A simplified BNF description of the commands specification syntax
34456 follows:
34457 <COMMAND_FLAG> ::= "enter" | "leave"
34459 <COMMAND_FLAGS> ::= <COMMAND_FLAG> [(+|"|")<COMMAND_FLAG>]
34460 <COMMAND> ::= ["[" <COMMAND_FLAGS> "]"] <TARGET> <COMMAND> [<ARG>]
34461 <COMMANDS> ::= <COMMAND> [,<COMMANDS>]
34462 <INTERVAL> ::= <START>[-<END>] <COMMANDS>
34463 <INTERVALS> ::= <INTERVAL>[;<INTERVALS>]
34464 Examples
34466 • Specify audio tempo change at second 4:
34468 asendcmd=c='4.0 atempo tempo 1.5',atempo
34470 • Target a specific filter instance:
34472 asendcmd=c='4.0 atempo@my tempo 1.5',atempo@my
34474 • Specify a list of drawtext and hue commands in a file.
34476 # show text in the interval 5-10
34478 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
34479 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
34480 # desaturate the image in the interval 15-20
34482 15.0-20.0 [enter] hue s 0,
34483 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
34484 [leave] hue s 1,
34485 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
34486 # apply an exponential saturation fade-out effect, starting from time 25
34488 25 [enter] hue s exp(25-t)
34489 A filtergraph allowing to read and process the above command list
34491 stored in a file test.cmd, can be specified with:
34492 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
34494 setpts, asetpts
34496 Change the PTS (presentation timestamp) of the input frames.
34497 "setpts" works on video frames, "asetpts" on audio frames.
34499 This filter accepts the following options:
34501 expr
34503 The expression which is evaluated for each frame to construct its
34504 timestamp.
34505 The expression is evaluated through the eval API and can contain the
34507 following constants:
34508 FRAME_RATE, FR
34510 frame rate, only defined for constant frame-rate video
34511 PTS The presentation timestamp in input
34513 N The count of the input frame for video or the number of consumed
34515 samples, not including the current frame for audio, starting from
34516 0.
34517 NB_CONSUMED_SAMPLES
34519 The number of consumed samples, not including the current frame
34520 (only audio)
34521 NB_SAMPLES, S
34523 The number of samples in the current frame (only audio)
34524 SAMPLE_RATE, SR
34526 The audio sample rate.
34527 STARTPTS
34529 The PTS of the first frame.
34530 STARTT
34532 the time in seconds of the first frame
34533 INTERLACED
34535 State whether the current frame is interlaced.
34536 T the time in seconds of the current frame
34538 POS original position in the file of the frame, or undefined if
34540 undefined for the current frame
34541 PREV_INPTS
34543 The previous input PTS.
34544 PREV_INT
34546 previous input time in seconds
34547 PREV_OUTPTS
34549 The previous output PTS.
34550 PREV_OUTT
34552 previous output time in seconds
34553 RTCTIME
34555 The wallclock (RTC) time in microseconds. This is deprecated, use
34556 time(0) instead.
34557 RTCSTART
34559 The wallclock (RTC) time at the start of the movie in microseconds.
34560 TB The timebase of the input timestamps.
34562 Examples
34564 • Start counting PTS from zero
34566 setpts=PTS-STARTPTS
34568 • Apply fast motion effect:
34570 setpts=0.5*PTS
34572 • Apply slow motion effect:
34574 setpts=2.0*PTS
34576 • Set fixed rate of 25 frames per second:
34578 setpts=N/(25*TB)
34580 • Set fixed rate 25 fps with some jitter:
34582 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
34584 • Apply an offset of 10 seconds to the input PTS:
34586 setpts=PTS+10/TB
34588 • Generate timestamps from a "live source" and rebase onto the
34590 current timebase:
34591 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
34593 • Generate timestamps by counting samples:
34595 asetpts=N/SR/TB
34597 setrange
34599 Force color range for the output video frame.
34600 The "setrange" filter marks the color range property for the output
34602 frames. It does not change the input frame, but only sets the
34603 corresponding property, which affects how the frame is treated by
34604 following filters.
34605 The filter accepts the following options:
34607 range
34609 Available values are:
34610 auto
34612 Keep the same color range property.
34613 unspecified, unknown
34615 Set the color range as unspecified.
34616 limited, tv, mpeg
34618 Set the color range as limited.
34619 full, pc, jpeg
34621 Set the color range as full.
34622 settb, asettb
34624 Set the timebase to use for the output frames timestamps. It is mainly
34625 useful for testing timebase configuration.
34626 It accepts the following parameters:
34628 expr, tb
34630 The expression which is evaluated into the output timebase.
34631 The value for tb is an arithmetic expression representing a rational.
34633 The expression can contain the constants "AVTB" (the default timebase),
34634 "intb" (the input timebase) and "sr" (the sample rate, audio only).
34635 Default value is "intb".
34636 Examples
34638 • Set the timebase to 1/25:
34640 settb=expr=1/25
34642 • Set the timebase to 1/10:
34644 settb=expr=0.1
34646 • Set the timebase to 1001/1000:
34648 settb=1+0.001
34650 • Set the timebase to 2*intb:
34652 settb=2*intb
34654 • Set the default timebase value:
34656 settb=AVTB
34658 showcqt
34660 Convert input audio to a video output representing frequency spectrum
34661 logarithmically using Brown-Puckette constant Q transform algorithm
34662 with direct frequency domain coefficient calculation (but the transform
34663 itself is not really constant Q, instead the Q factor is actually
34664 variable/clamped), with musical tone scale, from E0 to D#10.
34665 The filter accepts the following options:
34667 size, s
34669 Specify the video size for the output. It must be even. For the
34670 syntax of this option, check the "Video size" section in the
34671 ffmpeg-utils manual. Default value is "1920x1080".
34672 fps, rate, r
34674 Set the output frame rate. Default value is 25.
34675 bar_h
34677 Set the bargraph height. It must be even. Default value is "-1"
34678 which computes the bargraph height automatically.
34679 axis_h
34681 Set the axis height. It must be even. Default value is "-1" which
34682 computes the axis height automatically.
34683 sono_h
34685 Set the sonogram height. It must be even. Default value is "-1"
34686 which computes the sonogram height automatically.
34687 fullhd
34689 Set the fullhd resolution. This option is deprecated, use size, s
34690 instead. Default value is 1.
34691 sono_v, volume
34693 Specify the sonogram volume expression. It can contain variables:
34694 bar_v
34696 the bar_v evaluated expression
34697 frequency, freq, f
34699 the frequency where it is evaluated
34700 timeclamp, tc
34702 the value of timeclamp option
34703 and functions:
34705 a_weighting(f)
34707 A-weighting of equal loudness
34708 b_weighting(f)
34710 B-weighting of equal loudness
34711 c_weighting(f)
34713 C-weighting of equal loudness.
34714 Default value is 16.
34716 bar_v, volume2
34718 Specify the bargraph volume expression. It can contain variables:
34719 sono_v
34721 the sono_v evaluated expression
34722 frequency, freq, f
34724 the frequency where it is evaluated
34725 timeclamp, tc
34727 the value of timeclamp option
34728 and functions:
34730 a_weighting(f)
34732 A-weighting of equal loudness
34733 b_weighting(f)
34735 B-weighting of equal loudness
34736 c_weighting(f)
34738 C-weighting of equal loudness.
34739 Default value is "sono_v".
34741 sono_g, gamma
34743 Specify the sonogram gamma. Lower gamma makes the spectrum more
34744 contrast, higher gamma makes the spectrum having more range.
34745 Default value is 3. Acceptable range is "[1, 7]".
34746 bar_g, gamma2
34748 Specify the bargraph gamma. Default value is 1. Acceptable range is
34749 "[1, 7]".
34750 bar_t
34752 Specify the bargraph transparency level. Lower value makes the
34753 bargraph sharper. Default value is 1. Acceptable range is "[0,
34754 1]".
34755 timeclamp, tc
34757 Specify the transform timeclamp. At low frequency, there is trade-
34758 off between accuracy in time domain and frequency domain. If
34759 timeclamp is lower, event in time domain is represented more
34760 accurately (such as fast bass drum), otherwise event in frequency
34761 domain is represented more accurately (such as bass guitar).
34762 Acceptable range is "[0.002, 1]". Default value is 0.17.
34763 attack
34765 Set attack time in seconds. The default is 0 (disabled). Otherwise,
34766 it limits future samples by applying asymmetric windowing in time
34767 domain, useful when low latency is required. Accepted range is "[0,
34768 1]".
34769 basefreq
34771 Specify the transform base frequency. Default value is
34772 20.01523126408007475, which is frequency 50 cents below E0.
34773 Acceptable range is "[10, 100000]".
34774 endfreq
34776 Specify the transform end frequency. Default value is
34777 20495.59681441799654, which is frequency 50 cents above D#10.
34778 Acceptable range is "[10, 100000]".
34779 coeffclamp
34781 This option is deprecated and ignored.
34782 tlength
34784 Specify the transform length in time domain. Use this option to
34785 control accuracy trade-off between time domain and frequency domain
34786 at every frequency sample. It can contain variables:
34787 frequency, freq, f
34789 the frequency where it is evaluated
34790 timeclamp, tc
34792 the value of timeclamp option.
34793 Default value is "384*tc/(384+tc*f)".
34795 count
34797 Specify the transform count for every video frame. Default value is
34798 6. Acceptable range is "[1, 30]".
34799 fcount
34801 Specify the transform count for every single pixel. Default value
34802 is 0, which makes it computed automatically. Acceptable range is
34803 "[0, 10]".
34804 fontfile
34806 Specify font file for use with freetype to draw the axis. If not
34807 specified, use embedded font. Note that drawing with font file or
34808 embedded font is not implemented with custom basefreq and endfreq,
34809 use axisfile option instead.
34810 font
34812 Specify fontconfig pattern. This has lower priority than fontfile.
34813 The ":" in the pattern may be replaced by "|" to avoid unnecessary
34814 escaping.
34815 fontcolor
34817 Specify font color expression. This is arithmetic expression that
34818 should return integer value 0xRRGGBB. It can contain variables:
34819 frequency, freq, f
34821 the frequency where it is evaluated
34822 timeclamp, tc
34824 the value of timeclamp option
34825 and functions:
34827 midi(f)
34829 midi number of frequency f, some midi numbers: E0(16), C1(24),
34830 C2(36), A4(69)
34831 r(x), g(x), b(x)
34833 red, green, and blue value of intensity x.
34834 Default value is "st(0, (midi(f)-59.5)/12); st(1,
34836 if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0)); r(1-ld(1)) +
34837 b(ld(1))".
34838 axisfile
34840 Specify image file to draw the axis. This option override fontfile
34841 and fontcolor option.
34842 axis, text
34844 Enable/disable drawing text to the axis. If it is set to 0, drawing
34845 to the axis is disabled, ignoring fontfile and axisfile option.
34846 Default value is 1.
34847 csp Set colorspace. The accepted values are:
34849 unspecified
34851 Unspecified (default)
34852 bt709
34854 BT.709
34855 fcc FCC
34857 bt470bg
34859 BT.470BG or BT.601-6 625
34860 smpte170m
34862 SMPTE-170M or BT.601-6 525
34863 smpte240m
34865 SMPTE-240M
34866 bt2020ncl
34868 BT.2020 with non-constant luminance
34869 cscheme
34871 Set spectrogram color scheme. This is list of floating point values
34872 with format "left_r|left_g|left_b|right_r|right_g|right_b". The
34873 default is "1|0.5|0|0|0.5|1".
34874 Examples
34876 • Playing audio while showing the spectrum:
34878 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
34880 • Same as above, but with frame rate 30 fps:
34882 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
34884 • Playing at 1280x720:
34886 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
34888 • Disable sonogram display:
34890 sono_h=0
34892 • A1 and its harmonics: A1, A2, (near)E3, A3:
34894 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),
34896 asplit[a][out1]; [a] showcqt [out0]'
34897 • Same as above, but with more accuracy in frequency domain:
34899 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),
34901 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
34902 • Custom volume:
34904 bar_v=10:sono_v=bar_v*a_weighting(f)
34906 • Custom gamma, now spectrum is linear to the amplitude.
34908 bar_g=2:sono_g=2
34910 • Custom tlength equation:
34912 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)))'
34914 • Custom fontcolor and fontfile, C-note is colored green, others are
34916 colored blue:
34917 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
34919 • Custom font using fontconfig:
34921 font='Courier New,Monospace,mono|bold'
34923 • Custom frequency range with custom axis using image file:
34925 axisfile=myaxis.png:basefreq=40:endfreq=10000
34927 showfreqs
34929 Convert input audio to video output representing the audio power
34930 spectrum. Audio amplitude is on Y-axis while frequency is on X-axis.
34931 The filter accepts the following options:
34933 size, s
34935 Specify size of video. For the syntax of this option, check the
34936 "Video size" section in the ffmpeg-utils manual. Default is
34937 "1024x512".
34938 rate, r
34940 Set video rate. Default is 25.
34941 mode
34943 Set display mode. This set how each frequency bin will be
34944 represented.
34945 It accepts the following values:
34947 line
34949 bar
34950 dot
34951 Default is "bar".
34953 ascale
34955 Set amplitude scale.
34956 It accepts the following values:
34958 lin Linear scale.
34960 sqrt
34962 Square root scale.
34963 cbrt
34965 Cubic root scale.
34966 log Logarithmic scale.
34968 Default is "log".
34970 fscale
34972 Set frequency scale.
34973 It accepts the following values:
34975 lin Linear scale.
34977 log Logarithmic scale.
34979 rlog
34981 Reverse logarithmic scale.
34982 Default is "lin".
34984 win_size
34986 Set window size. Allowed range is from 16 to 65536.
34987 Default is 2048
34989 win_func
34991 Set windowing function.
34992 It accepts the following values:
34994 rect
34996 bartlett
34997 hanning
34998 hamming
34999 blackman
35000 welch
35001 flattop
35002 bharris
35003 bnuttall
35004 bhann
35005 sine
35006 nuttall
35007 lanczos
35008 gauss
35009 tukey
35010 dolph
35011 cauchy
35012 parzen
35013 poisson
35014 bohman
35015 Default is "hanning".
35017 overlap
35019 Set window overlap. In range "[0, 1]". Default is 1, which means
35020 optimal overlap for selected window function will be picked.
35021 averaging
35023 Set time averaging. Setting this to 0 will display current maximal
35024 peaks. Default is 1, which means time averaging is disabled.
35025 colors
35027 Specify list of colors separated by space or by '|' which will be
35028 used to draw channel frequencies. Unrecognized or missing colors
35029 will be replaced by white color.
35030 cmode
35032 Set channel display mode.
35033 It accepts the following values:
35035 combined
35037 separate
35038 Default is "combined".
35040 minamp
35042 Set minimum amplitude used in "log" amplitude scaler.
35043 data
35045 Set data display mode.
35046 It accepts the following values:
35048 magnitude
35050 phase
35051 delay
35052 Default is "magnitude".
35054 channels
35056 Set channels to use when processing audio. By default all are
35057 processed.
35058 showspatial
35060 Convert stereo input audio to a video output, representing the spatial
35061 relationship between two channels.
35062 The filter accepts the following options:
35064 size, s
35066 Specify the video size for the output. For the syntax of this
35067 option, check the "Video size" section in the ffmpeg-utils manual.
35068 Default value is "512x512".
35069 win_size
35071 Set window size. Allowed range is from 1024 to 65536. Default size
35072 is 4096.
35073 win_func
35075 Set window function.
35076 It accepts the following values:
35078 rect
35080 bartlett
35081 hann
35082 hanning
35083 hamming
35084 blackman
35085 welch
35086 flattop
35087 bharris
35088 bnuttall
35089 bhann
35090 sine
35091 nuttall
35092 lanczos
35093 gauss
35094 tukey
35095 dolph
35096 cauchy
35097 parzen
35098 poisson
35099 bohman
35100 Default value is "hann".
35102 overlap
35104 Set ratio of overlap window. Default value is 0.5. When value is 1
35105 overlap is set to recommended size for specific window function
35106 currently used.
35107 showspectrum
35109 Convert input audio to a video output, representing the audio frequency
35110 spectrum.
35111 The filter accepts the following options:
35113 size, s
35115 Specify the video size for the output. For the syntax of this
35116 option, check the "Video size" section in the ffmpeg-utils manual.
35117 Default value is "640x512".
35118 slide
35120 Specify how the spectrum should slide along the window.
35121 It accepts the following values:
35123 replace
35125 the samples start again on the left when they reach the right
35126 scroll
35128 the samples scroll from right to left
35129 fullframe
35131 frames are only produced when the samples reach the right
35132 rscroll
35134 the samples scroll from left to right
35135 lreplace
35137 the samples start again on the right when they reach the left
35138 Default value is "replace".
35140 mode
35142 Specify display mode.
35143 It accepts the following values:
35145 combined
35147 all channels are displayed in the same row
35148 separate
35150 all channels are displayed in separate rows
35151 Default value is combined.
35153 color
35155 Specify display color mode.
35156 It accepts the following values:
35158 channel
35160 each channel is displayed in a separate color
35161 intensity
35163 each channel is displayed using the same color scheme
35164 rainbow
35166 each channel is displayed using the rainbow color scheme
35167 moreland
35169 each channel is displayed using the moreland color scheme
35170 nebulae
35172 each channel is displayed using the nebulae color scheme
35173 fire
35175 each channel is displayed using the fire color scheme
35176 fiery
35178 each channel is displayed using the fiery color scheme
35179 fruit
35181 each channel is displayed using the fruit color scheme
35182 cool
35184 each channel is displayed using the cool color scheme
35185 magma
35187 each channel is displayed using the magma color scheme
35188 green
35190 each channel is displayed using the green color scheme
35191 viridis
35193 each channel is displayed using the viridis color scheme
35194 plasma
35196 each channel is displayed using the plasma color scheme
35197 cividis
35199 each channel is displayed using the cividis color scheme
35200 terrain
35202 each channel is displayed using the terrain color scheme
35203 Default value is channel.
35205 scale
35207 Specify scale used for calculating intensity color values.
35208 It accepts the following values:
35210 lin linear
35212 sqrt
35214 square root, default
35215 cbrt
35217 cubic root
35218 log logarithmic
35220 4thrt
35222 4th root
35223 5thrt
35225 5th root
35226 Default value is sqrt.
35228 fscale
35230 Specify frequency scale.
35231 It accepts the following values:
35233 lin linear
35235 log logarithmic
35237 Default value is lin.
35239 saturation
35241 Set saturation modifier for displayed colors. Negative values
35242 provide alternative color scheme. 0 is no saturation at all.
35243 Saturation must be in [-10.0, 10.0] range. Default value is 1.
35244 win_func
35246 Set window function.
35247 It accepts the following values:
35249 rect
35251 bartlett
35252 hann
35253 hanning
35254 hamming
35255 blackman
35256 welch
35257 flattop
35258 bharris
35259 bnuttall
35260 bhann
35261 sine
35262 nuttall
35263 lanczos
35264 gauss
35265 tukey
35266 dolph
35267 cauchy
35268 parzen
35269 poisson
35270 bohman
35271 Default value is "hann".
35273 orientation
35275 Set orientation of time vs frequency axis. Can be "vertical" or
35276 "horizontal". Default is "vertical".
35277 overlap
35279 Set ratio of overlap window. Default value is 0. When value is 1
35280 overlap is set to recommended size for specific window function
35281 currently used.
35282 gain
35284 Set scale gain for calculating intensity color values. Default
35285 value is 1.
35286 data
35288 Set which data to display. Can be "magnitude", default or "phase",
35289 or unwrapped phase: "uphase".
35290 rotation
35292 Set color rotation, must be in [-1.0, 1.0] range. Default value is
35293 0.
35294 start
35296 Set start frequency from which to display spectrogram. Default is
35297 0.
35298 stop
35300 Set stop frequency to which to display spectrogram. Default is 0.
35301 fps Set upper frame rate limit. Default is "auto", unlimited.
35303 legend
35305 Draw time and frequency axes and legends. Default is disabled.
35306 drange
35308 Set dynamic range used to calculate intensity color values. Default
35309 is 120 dBFS. Allowed range is from 10 to 200.
35310 limit
35312 Set upper limit of input audio samples volume in dBFS. Default is 0
35313 dBFS. Allowed range is from -100 to 100.
35314 opacity
35316 Set opacity strength when using pixel format output with alpha
35317 component.
35318 The usage is very similar to the showwaves filter; see the examples in
35320 that section.
35321 Examples
35323 • Large window with logarithmic color scaling:
35325 showspectrum=s=1280x480:scale=log
35327 • Complete example for a colored and sliding spectrum per channel
35329 using ffplay:
35330 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
35332 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
35333 showspectrumpic
35335 Convert input audio to a single video frame, representing the audio
35336 frequency spectrum.
35337 The filter accepts the following options:
35339 size, s
35341 Specify the video size for the output. For the syntax of this
35342 option, check the "Video size" section in the ffmpeg-utils manual.
35343 Default value is "4096x2048".
35344 mode
35346 Specify display mode.
35347 It accepts the following values:
35349 combined
35351 all channels are displayed in the same row
35352 separate
35354 all channels are displayed in separate rows
35355 Default value is combined.
35357 color
35359 Specify display color mode.
35360 It accepts the following values:
35362 channel
35364 each channel is displayed in a separate color
35365 intensity
35367 each channel is displayed using the same color scheme
35368 rainbow
35370 each channel is displayed using the rainbow color scheme
35371 moreland
35373 each channel is displayed using the moreland color scheme
35374 nebulae
35376 each channel is displayed using the nebulae color scheme
35377 fire
35379 each channel is displayed using the fire color scheme
35380 fiery
35382 each channel is displayed using the fiery color scheme
35383 fruit
35385 each channel is displayed using the fruit color scheme
35386 cool
35388 each channel is displayed using the cool color scheme
35389 magma
35391 each channel is displayed using the magma color scheme
35392 green
35394 each channel is displayed using the green color scheme
35395 viridis
35397 each channel is displayed using the viridis color scheme
35398 plasma
35400 each channel is displayed using the plasma color scheme
35401 cividis
35403 each channel is displayed using the cividis color scheme
35404 terrain
35406 each channel is displayed using the terrain color scheme
35407 Default value is intensity.
35409 scale
35411 Specify scale used for calculating intensity color values.
35412 It accepts the following values:
35414 lin linear
35416 sqrt
35418 square root, default
35419 cbrt
35421 cubic root
35422 log logarithmic
35424 4thrt
35426 4th root
35427 5thrt
35429 5th root
35430 Default value is log.
35432 fscale
35434 Specify frequency scale.
35435 It accepts the following values:
35437 lin linear
35439 log logarithmic
35441 Default value is lin.
35443 saturation
35445 Set saturation modifier for displayed colors. Negative values
35446 provide alternative color scheme. 0 is no saturation at all.
35447 Saturation must be in [-10.0, 10.0] range. Default value is 1.
35448 win_func
35450 Set window function.
35451 It accepts the following values:
35453 rect
35455 bartlett
35456 hann
35457 hanning
35458 hamming
35459 blackman
35460 welch
35461 flattop
35462 bharris
35463 bnuttall
35464 bhann
35465 sine
35466 nuttall
35467 lanczos
35468 gauss
35469 tukey
35470 dolph
35471 cauchy
35472 parzen
35473 poisson
35474 bohman
35475 Default value is "hann".
35477 orientation
35479 Set orientation of time vs frequency axis. Can be "vertical" or
35480 "horizontal". Default is "vertical".
35481 gain
35483 Set scale gain for calculating intensity color values. Default
35484 value is 1.
35485 legend
35487 Draw time and frequency axes and legends. Default is enabled.
35488 rotation
35490 Set color rotation, must be in [-1.0, 1.0] range. Default value is
35491 0.
35492 start
35494 Set start frequency from which to display spectrogram. Default is
35495 0.
35496 stop
35498 Set stop frequency to which to display spectrogram. Default is 0.
35499 drange
35501 Set dynamic range used to calculate intensity color values. Default
35502 is 120 dBFS. Allowed range is from 10 to 200.
35503 limit
35505 Set upper limit of input audio samples volume in dBFS. Default is 0
35506 dBFS. Allowed range is from -100 to 100.
35507 opacity
35509 Set opacity strength when using pixel format output with alpha
35510 component.
35511 Examples
35513 • Extract an audio spectrogram of a whole audio track in a 1024x1024
35515 picture using ffmpeg:
35516 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
35518 showvolume
35520 Convert input audio volume to a video output.
35521 The filter accepts the following options:
35523 rate, r
35525 Set video rate.
35526 b Set border width, allowed range is [0, 5]. Default is 1.
35528 w Set channel width, allowed range is [80, 8192]. Default is 400.
35530 h Set channel height, allowed range is [1, 900]. Default is 20.
35532 f Set fade, allowed range is [0, 1]. Default is 0.95.
35534 c Set volume color expression.
35536 The expression can use the following variables:
35538 VOLUME
35540 Current max volume of channel in dB.
35541 PEAK
35543 Current peak.
35544 CHANNEL
35546 Current channel number, starting from 0.
35547 t If set, displays channel names. Default is enabled.
35549 v If set, displays volume values. Default is enabled.
35551 o Set orientation, can be horizontal: "h" or vertical: "v", default
35553 is "h".
35554 s Set step size, allowed range is [0, 5]. Default is 0, which means
35556 step is disabled.
35557 p Set background opacity, allowed range is [0, 1]. Default is 0.
35559 m Set metering mode, can be peak: "p" or rms: "r", default is "p".
35561 ds Set display scale, can be linear: "lin" or log: "log", default is
35563 "lin".
35564 dm In second. If set to > 0., display a line for the max level in the
35566 previous seconds. default is disabled: 0.
35567 dmc The color of the max line. Use when "dm" option is set to > 0.
35569 default is: "orange"
35570 showwaves
35572 Convert input audio to a video output, representing the samples waves.
35573 The filter accepts the following options:
35575 size, s
35577 Specify the video size for the output. For the syntax of this
35578 option, check the "Video size" section in the ffmpeg-utils manual.
35579 Default value is "600x240".
35580 mode
35582 Set display mode.
35583 Available values are:
35585 point
35587 Draw a point for each sample.
35588 line
35590 Draw a vertical line for each sample.
35591 p2p Draw a point for each sample and a line between them.
35593 cline
35595 Draw a centered vertical line for each sample.
35596 Default value is "point".
35598 n Set the number of samples which are printed on the same column. A
35600 larger value will decrease the frame rate. Must be a positive
35601 integer. This option can be set only if the value for rate is not
35602 explicitly specified.
35603 rate, r
35605 Set the (approximate) output frame rate. This is done by setting
35606 the option n. Default value is "25".
35607 split_channels
35609 Set if channels should be drawn separately or overlap. Default
35610 value is 0.
35611 colors
35613 Set colors separated by '|' which are going to be used for drawing
35614 of each channel.
35615 scale
35617 Set amplitude scale.
35618 Available values are:
35620 lin Linear.
35622 log Logarithmic.
35624 sqrt
35626 Square root.
35627 cbrt
35629 Cubic root.
35630 Default is linear.
35632 draw
35634 Set the draw mode. This is mostly useful to set for high n.
35635 Available values are:
35637 scale
35639 Scale pixel values for each drawn sample.
35640 full
35642 Draw every sample directly.
35643 Default value is "scale".
35645 Examples
35647 • Output the input file audio and the corresponding video
35649 representation at the same time:
35650 amovie=a.mp3,asplit[out0],showwaves[out1]
35652 • Create a synthetic signal and show it with showwaves, forcing a
35654 frame rate of 30 frames per second:
35655 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
35657 showwavespic
35659 Convert input audio to a single video frame, representing the samples
35660 waves.
35661 The filter accepts the following options:
35663 size, s
35665 Specify the video size for the output. For the syntax of this
35666 option, check the "Video size" section in the ffmpeg-utils manual.
35667 Default value is "600x240".
35668 split_channels
35670 Set if channels should be drawn separately or overlap. Default
35671 value is 0.
35672 colors
35674 Set colors separated by '|' which are going to be used for drawing
35675 of each channel.
35676 scale
35678 Set amplitude scale.
35679 Available values are:
35681 lin Linear.
35683 log Logarithmic.
35685 sqrt
35687 Square root.
35688 cbrt
35690 Cubic root.
35691 Default is linear.
35693 draw
35695 Set the draw mode.
35696 Available values are:
35698 scale
35700 Scale pixel values for each drawn sample.
35701 full
35703 Draw every sample directly.
35704 Default value is "scale".
35706 filter
35708 Set the filter mode.
35709 Available values are:
35711 average
35713 Use average samples values for each drawn sample.
35714 peak
35716 Use peak samples values for each drawn sample.
35717 Default value is "average".
35719 Examples
35721 • Extract a channel split representation of the wave form of a whole
35723 audio track in a 1024x800 picture using ffmpeg:
35724 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
35726 sidedata, asidedata
35728 Delete frame side data, or select frames based on it.
35729 This filter accepts the following options:
35731 mode
35733 Set mode of operation of the filter.
35734 Can be one of the following:
35736 select
35738 Select every frame with side data of "type".
35739 delete
35741 Delete side data of "type". If "type" is not set, delete all
35742 side data in the frame.
35743 type
35745 Set side data type used with all modes. Must be set for "select"
35746 mode. For the list of frame side data types, refer to the
35747 "AVFrameSideDataType" enum in libavutil/frame.h. For example, to
35748 choose "AV_FRAME_DATA_PANSCAN" side data, you must specify
35749 "PANSCAN".
35750 spectrumsynth
35752 Synthesize audio from 2 input video spectrums, first input stream
35753 represents magnitude across time and second represents phase across
35754 time. The filter will transform from frequency domain as displayed in
35755 videos back to time domain as presented in audio output.
35756 This filter is primarily created for reversing processed showspectrum
35758 filter outputs, but can synthesize sound from other spectrograms too.
35759 But in such case results are going to be poor if the phase data is not
35760 available, because in such cases phase data need to be recreated,
35761 usually it's just recreated from random noise. For best results use
35762 gray only output ("channel" color mode in showspectrum filter) and
35763 "log" scale for magnitude video and "lin" scale for phase video. To
35764 produce phase, for 2nd video, use "data" option. Inputs videos should
35765 generally use "fullframe" slide mode as that saves resources needed for
35766 decoding video.
35767 The filter accepts the following options:
35769 sample_rate
35771 Specify sample rate of output audio, the sample rate of audio from
35772 which spectrum was generated may differ.
35773 channels
35775 Set number of channels represented in input video spectrums.
35776 scale
35778 Set scale which was used when generating magnitude input spectrum.
35779 Can be "lin" or "log". Default is "log".
35780 slide
35782 Set slide which was used when generating inputs spectrums. Can be
35783 "replace", "scroll", "fullframe" or "rscroll". Default is
35784 "fullframe".
35785 win_func
35787 Set window function used for resynthesis.
35788 overlap
35790 Set window overlap. In range "[0, 1]". Default is 1, which means
35791 optimal overlap for selected window function will be picked.
35792 orientation
35794 Set orientation of input videos. Can be "vertical" or "horizontal".
35795 Default is "vertical".
35796 Examples
35798 • First create magnitude and phase videos from audio, assuming audio
35800 is stereo with 44100 sample rate, then resynthesize videos back to
35801 audio with spectrumsynth:
35802 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
35804 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
35805 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
35806 split, asplit
35808 Split input into several identical outputs.
35809 "asplit" works with audio input, "split" with video.
35811 The filter accepts a single parameter which specifies the number of
35813 outputs. If unspecified, it defaults to 2.
35814 Examples
35816 • Create two separate outputs from the same input:
35818 [in] split [out0][out1]
35820 • To create 3 or more outputs, you need to specify the number of
35822 outputs, like in:
35823 [in] asplit=3 [out0][out1][out2]
35825 • Create two separate outputs from the same input, one cropped and
35827 one padded:
35828 [in] split [splitout1][splitout2];
35830 [splitout1] crop=100:100:0:0 [cropout];
35831 [splitout2] pad=200:200:100:100 [padout];
35832 • Create 5 copies of the input audio with ffmpeg:
35834 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
35836 zmq, azmq
35838 Receive commands sent through a libzmq client, and forward them to
35839 filters in the filtergraph.
35840 "zmq" and "azmq" work as a pass-through filters. "zmq" must be inserted
35842 between two video filters, "azmq" between two audio filters. Both are
35843 capable to send messages to any filter type.
35844 To enable these filters you need to install the libzmq library and
35846 headers and configure FFmpeg with "--enable-libzmq".
35847 For more information about libzmq see: <http://www.zeromq.org/>
35849 The "zmq" and "azmq" filters work as a libzmq server, which receives
35851 messages sent through a network interface defined by the bind_address
35852 (or the abbreviation "b") option. Default value of this option is
35853 tcp://localhost:5555. You may want to alter this value to your needs,
35854 but do not forget to escape any ':' signs (see filtergraph escaping).
35855 The received message must be in the form:
35857 <TARGET> <COMMAND> [<ARG>]
35859 TARGET specifies the target of the command, usually the name of the
35861 filter class or a specific filter instance name. The default filter
35862 instance name uses the pattern Parsed_<filter_name>_<index>, but you
35863 can override this by using the filter_name@id syntax (see Filtergraph
35864 syntax).
35865 COMMAND specifies the name of the command for the target filter.
35867 ARG is optional and specifies the optional argument list for the given
35869 COMMAND.
35870 Upon reception, the message is processed and the corresponding command
35872 is injected into the filtergraph. Depending on the result, the filter
35873 will send a reply to the client, adopting the format:
35874 <ERROR_CODE> <ERROR_REASON>
35876 <MESSAGE>
35877 MESSAGE is optional.
35879 Examples
35881 Look at tools/zmqsend for an example of a zmq client which can be used
35883 to send commands processed by these filters.
35884 Consider the following filtergraph generated by ffplay. In this
35886 example the last overlay filter has an instance name. All other filters
35887 will have default instance names.
35888 ffplay -dumpgraph 1 -f lavfi "
35890 color=s=100x100:c=red [l];
35891 color=s=100x100:c=blue [r];
35892 nullsrc=s=200x100, zmq [bg];
35893 [bg][l] overlay [bg+l];
35894 [bg+l][r] overlay@my=x=100 "
35895 To change the color of the left side of the video, the following
35897 command can be used:
35898 echo Parsed_color_0 c yellow | tools/zmqsend
35900 To change the right side:
35902 echo Parsed_color_1 c pink | tools/zmqsend
35904 To change the position of the right side:
35906 echo overlay@my x 150 | tools/zmqsend
35908
35910 Below is a description of the currently available multimedia sources.
35911 amovie
35913 This is the same as movie source, except it selects an audio stream by
35914 default.
35915 avsynctest
35917 Generate an Audio/Video Sync Test.
35918 Generated stream periodically shows flash video frame and emits beep in
35920 audio. Useful to inspect A/V sync issues.
35921 It accepts the following options:
35923 size, s
35925 Set output video size. Default value is "hd720".
35926 framerate, fr
35928 Set output video frame rate. Default value is 30.
35929 samplerate, sr
35931 Set output audio sample rate. Default value is 44100.
35932 amplitude, a
35934 Set output audio beep amplitude. Default value is 0.7.
35935 period, p
35937 Set output audio beep period in seconds. Default value is 3.
35938 delay, dl
35940 Set output video flash delay in number of frames. Default value is
35941 0.
35942 cycle, c
35944 Enable cycling of video delays, by default is disabled.
35945 duration, d
35947 Set stream output duration. By default duration is unlimited.
35948 fg, bg, ag
35950 Set foreground/background/additional color.
35951 movie
35953 Read audio and/or video stream(s) from a movie container.
35954 It accepts the following parameters:
35956 filename
35958 The name of the resource to read (not necessarily a file; it can
35959 also be a device or a stream accessed through some protocol).
35960 format_name, f
35962 Specifies the format assumed for the movie to read, and can be
35963 either the name of a container or an input device. If not
35964 specified, the format is guessed from movie_name or by probing.
35965 seek_point, sp
35967 Specifies the seek point in seconds. The frames will be output
35968 starting from this seek point. The parameter is evaluated with
35969 "av_strtod", so the numerical value may be suffixed by an IS
35970 postfix. The default value is "0".
35971 streams, s
35973 Specifies the streams to read. Several streams can be specified,
35974 separated by "+". The source will then have as many outputs, in the
35975 same order. The syntax is explained in the "Stream specifiers"
35976 section in the ffmpeg manual. Two special names, "dv" and "da"
35977 specify respectively the default (best suited) video and audio
35978 stream. Default is "dv", or "da" if the filter is called as
35979 "amovie".
35980 stream_index, si
35982 Specifies the index of the video stream to read. If the value is
35983 -1, the most suitable video stream will be automatically selected.
35984 The default value is "-1". Deprecated. If the filter is called
35985 "amovie", it will select audio instead of video.
35986 loop
35988 Specifies how many times to read the stream in sequence. If the
35989 value is 0, the stream will be looped infinitely. Default value is
35990 "1".
35991 Note that when the movie is looped the source timestamps are not
35993 changed, so it will generate non monotonically increasing
35994 timestamps.
35995 discontinuity
35997 Specifies the time difference between frames above which the point
35998 is considered a timestamp discontinuity which is removed by
35999 adjusting the later timestamps.
36000 dec_threads
36002 Specifies the number of threads for decoding
36003 format_opts
36005 Specify format options for the opened file. Format options can be
36006 specified as a list of key=value pairs separated by ':'. The
36007 following example shows how to add protocol_whitelist and
36008 protocol_blacklist options:
36009 ffplay -f lavfi
36011 "movie=filename='1.sdp':format_opts='protocol_whitelist=file,rtp,udp\:protocol_blacklist=http'"
36012 It allows overlaying a second video on top of the main input of a
36014 filtergraph, as shown in this graph:
36015 input -----------> deltapts0 --> overlay --> output
36017 ^
36018 |
36019 movie --> scale--> deltapts1 -------+
36020 Examples
36022 • Skip 3.2 seconds from the start of the AVI file in.avi, and overlay
36024 it on top of the input labelled "in":
36025 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
36027 [in] setpts=PTS-STARTPTS [main];
36028 [main][over] overlay=16:16 [out]
36029 • Read from a video4linux2 device, and overlay it on top of the input
36031 labelled "in":
36032 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
36034 [in] setpts=PTS-STARTPTS [main];
36035 [main][over] overlay=16:16 [out]
36036 • Read the first video stream and the audio stream with id 0x81 from
36038 dvd.vob; the video is connected to the pad named "video" and the
36039 audio is connected to the pad named "audio":
36040 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
36042 Commands
36044 Both movie and amovie support the following commands:
36046 seek
36048 Perform seek using "av_seek_frame". The syntax is: seek
36049 stream_index|timestamp|flags
36050 • stream_index: If stream_index is -1, a default stream is
36052 selected, and timestamp is automatically converted from
36053 AV_TIME_BASE units to the stream specific time_base.
36054 • timestamp: Timestamp in AVStream.time_base units or, if no
36056 stream is specified, in AV_TIME_BASE units.
36057 • flags: Flags which select direction and seeking mode.
36059 get_duration
36061 Get movie duration in AV_TIME_BASE units.
36062
36064 FFmpeg can be hooked up with a number of external libraries to add
36065 support for more formats. None of them are used by default, their use
36066 has to be explicitly requested by passing the appropriate flags to
36067 ./configure.
36068 Alliance for Open Media (AOM)
36070 FFmpeg can make use of the AOM library for AV1 decoding and encoding.
36071 Go to <http://aomedia.org/> and follow the instructions for installing
36073 the library. Then pass "--enable-libaom" to configure to enable it.
36074 AMD AMF/VCE
36076 FFmpeg can use the AMD Advanced Media Framework library for accelerated
36077 H.264 and HEVC(only windows) encoding on hardware with Video Coding
36078 Engine (VCE).
36079 To enable support you must obtain the AMF framework header
36081 files(version 1.4.9+) from
36082 <https://github.com/GPUOpen-LibrariesAndSDKs/AMF.git>.
36083 Create an "AMF/" directory in the system include path. Copy the
36085 contents of "AMF/amf/public/include/" into that directory. Then
36086 configure FFmpeg with "--enable-amf".
36087 Initialization of amf encoder occurs in this order: 1) trying to
36089 initialize through dx11(only windows) 2) trying to initialize through
36090 dx9(only windows) 3) trying to initialize through vulkan
36091 To use h.264(AMD VCE) encoder on linux amdgru-pro version 19.20+ and
36093 amf-amdgpu-pro package(amdgru-pro contains, but does not install
36094 automatically) are required.
36095 This driver can be installed using amdgpu-pro-install script in
36097 official amd driver archive.
36098 AviSynth
36100 FFmpeg can read AviSynth scripts as input. To enable support, pass
36101 "--enable-avisynth" to configure after installing the headers provided
36102 by <https://github.com/AviSynth/AviSynthPlus>. AviSynth+ can be
36103 configured to install only the headers by either passing
36104 "-DHEADERS_ONLY:bool=on" to the normal CMake-based build system, or by
36105 using the supplied "GNUmakefile".
36106 For Windows, supported AviSynth variants are <http://avisynth.nl> for
36108 32-bit builds and <http://avisynth.nl/index.php/AviSynth+> for 32-bit
36109 and 64-bit builds.
36110 For Linux, macOS, and BSD, the only supported AviSynth variant is
36112 <https://github.com/AviSynth/AviSynthPlus>, starting with version 3.5.
36113 In 2016, AviSynth+ added support for building with GCC. However,
36115 due to the eccentricities of Windows' calling conventions, 32-bit
36116 GCC builds of AviSynth+ are not compatible with typical 32-bit
36117 builds of FFmpeg.
36118 By default, FFmpeg assumes compatibility with 32-bit MSVC builds of
36120 AviSynth+ since that is the most widely-used and entrenched build
36121 configuration. Users can override this and enable support for
36122 32-bit GCC builds of AviSynth+ by passing "-DAVSC_WIN32_GCC32" to
36123 "--extra-cflags" when configuring FFmpeg.
36124 64-bit builds of FFmpeg are not affected, and can use either MSVC
36126 or GCC builds of AviSynth+ without any special flags.
36127 AviSynth(+) is loaded dynamically. Distributors can build FFmpeg
36129 with "--enable-avisynth", and the binaries will work regardless of
36130 the end user having AviSynth installed. If/when an end user would
36131 like to use AviSynth scripts, then they can install AviSynth(+) and
36132 FFmpeg will be able to find and use it to open scripts.
36133 Chromaprint
36135 FFmpeg can make use of the Chromaprint library for generating audio
36136 fingerprints. Pass "--enable-chromaprint" to configure to enable it.
36137 See <https://acoustid.org/chromaprint>.
36138 codec2
36140 FFmpeg can make use of the codec2 library for codec2 decoding and
36141 encoding. There is currently no native decoder, so libcodec2 must be
36142 used for decoding.
36143 Go to <http://freedv.org/>, download "Codec 2 source archive". Build
36145 and install using CMake. Debian users can install the libcodec2-dev
36146 package instead. Once libcodec2 is installed you can pass
36147 "--enable-libcodec2" to configure to enable it.
36148 The easiest way to use codec2 is with .c2 files, since they contain the
36150 mode information required for decoding. To encode such a file, use a
36151 .c2 file extension and give the libcodec2 encoder the -mode option:
36152 "ffmpeg -i input.wav -mode 700C output.c2". Playback is as simple as
36153 "ffplay output.c2". For a list of supported modes, run "ffmpeg -h
36154 encoder=libcodec2". Raw codec2 files are also supported. To make
36155 sense of them the mode in use needs to be specified as a format option:
36156 "ffmpeg -f codec2raw -mode 1300 -i input.raw output.wav".
36157 dav1d
36159 FFmpeg can make use of the dav1d library for AV1 video decoding.
36160 Go to <https://code.videolan.org/videolan/dav1d> and follow the
36162 instructions for installing the library. Then pass "--enable-libdav1d"
36163 to configure to enable it.
36164 davs2
36166 FFmpeg can make use of the davs2 library for AVS2-P2/IEEE1857.4 video
36167 decoding.
36168 Go to <https://github.com/pkuvcl/davs2> and follow the instructions for
36170 installing the library. Then pass "--enable-libdavs2" to configure to
36171 enable it.
36172 libdavs2 is under the GNU Public License Version 2 or later (see
36174 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
36175 details), you must upgrade FFmpeg's license to GPL in order to use
36176 it.
36177 uavs3d
36179 FFmpeg can make use of the uavs3d library for AVS3-P2/IEEE1857.10 video
36180 decoding.
36181 Go to <https://github.com/uavs3/uavs3d> and follow the instructions for
36183 installing the library. Then pass "--enable-libuavs3d" to configure to
36184 enable it.
36185 Game Music Emu
36187 FFmpeg can make use of the Game Music Emu library to read audio from
36188 supported video game music file formats. Pass "--enable-libgme" to
36189 configure to enable it. See
36190 <https://bitbucket.org/mpyne/game-music-emu/overview>.
36191 Intel QuickSync Video
36193 FFmpeg can use Intel QuickSync Video (QSV) for accelerated decoding and
36194 encoding of multiple codecs. To use QSV, FFmpeg must be linked against
36195 the "libmfx" dispatcher, which loads the actual decoding libraries.
36196 The dispatcher is open source and can be downloaded from
36198 <https://github.com/lu-zero/mfx_dispatch.git>. FFmpeg needs to be
36199 configured with the "--enable-libmfx" option and "pkg-config" needs to
36200 be able to locate the dispatcher's ".pc" files.
36201 Kvazaar
36203 FFmpeg can make use of the Kvazaar library for HEVC encoding.
36204 Go to <https://github.com/ultravideo/kvazaar> and follow the
36206 instructions for installing the library. Then pass
36207 "--enable-libkvazaar" to configure to enable it.
36208 LAME
36210 FFmpeg can make use of the LAME library for MP3 encoding.
36211 Go to <http://lame.sourceforge.net/> and follow the instructions for
36213 installing the library. Then pass "--enable-libmp3lame" to configure
36214 to enable it.
36215 libilbc
36217 iLBC is a narrowband speech codec that has been made freely available
36218 by Google as part of the WebRTC project. libilbc is a packaging
36219 friendly copy of the iLBC codec. FFmpeg can make use of the libilbc
36220 library for iLBC decoding and encoding.
36221 Go to <https://github.com/TimothyGu/libilbc> and follow the
36223 instructions for installing the library. Then pass "--enable-libilbc"
36224 to configure to enable it.
36225 libjxl
36227 JPEG XL is an image format intended to fully replace legacy JPEG for an
36228 extended period of life. See <https://jpegxl.info/> for more
36229 information, and see <https://github.com/libjxl/libjxl> for the library
36230 source. You can pass "--enable-libjxl" to configure in order enable the
36231 libjxl wrapper.
36232 libvpx
36234 FFmpeg can make use of the libvpx library for VP8/VP9 decoding and
36235 encoding.
36236 Go to <http://www.webmproject.org/> and follow the instructions for
36238 installing the library. Then pass "--enable-libvpx" to configure to
36239 enable it.
36240 ModPlug
36242 FFmpeg can make use of this library, originating in Modplug-XMMS, to
36243 read from MOD-like music files. See
36244 <https://github.com/Konstanty/libmodplug>. Pass "--enable-libmodplug"
36245 to configure to enable it.
36246 OpenCORE, VisualOn, and Fraunhofer libraries
36248 Spun off Google Android sources, OpenCore, VisualOn and Fraunhofer
36249 libraries provide encoders for a number of audio codecs.
36250 OpenCORE and VisualOn libraries are under the Apache License 2.0
36252 (see <http://www.apache.org/licenses/LICENSE-2.0> for details),
36253 which is incompatible to the LGPL version 2.1 and GPL version 2.
36254 You have to upgrade FFmpeg's license to LGPL version 3 (or if you
36255 have enabled GPL components, GPL version 3) by passing
36256 "--enable-version3" to configure in order to use it.
36257 The license of the Fraunhofer AAC library is incompatible with the
36259 GPL. Therefore, for GPL builds, you have to pass
36260 "--enable-nonfree" to configure in order to use it. To the best of
36261 our knowledge, it is compatible with the LGPL.
36262 OpenCORE AMR
36264 FFmpeg can make use of the OpenCORE libraries for AMR-NB
36266 decoding/encoding and AMR-WB decoding.
36267 Go to <http://sourceforge.net/projects/opencore-amr/> and follow the
36269 instructions for installing the libraries. Then pass
36270 "--enable-libopencore-amrnb" and/or "--enable-libopencore-amrwb" to
36271 configure to enable them.
36272 VisualOn AMR-WB encoder library
36274 FFmpeg can make use of the VisualOn AMR-WBenc library for AMR-WB
36276 encoding.
36277 Go to <http://sourceforge.net/projects/opencore-amr/> and follow the
36279 instructions for installing the library. Then pass
36280 "--enable-libvo-amrwbenc" to configure to enable it.
36281 Fraunhofer AAC library
36283 FFmpeg can make use of the Fraunhofer AAC library for AAC decoding &
36285 encoding.
36286 Go to <http://sourceforge.net/projects/opencore-amr/> and follow the
36288 instructions for installing the library. Then pass
36289 "--enable-libfdk-aac" to configure to enable it.
36290 OpenH264
36292 FFmpeg can make use of the OpenH264 library for H.264 decoding and
36293 encoding.
36294 Go to <http://www.openh264.org/> and follow the instructions for
36296 installing the library. Then pass "--enable-libopenh264" to configure
36297 to enable it.
36298 For decoding, this library is much more limited than the built-in
36300 decoder in libavcodec; currently, this library lacks support for
36301 decoding B-frames and some other main/high profile features. (It
36302 currently only supports constrained baseline profile and CABAC.) Using
36303 it is mostly useful for testing and for taking advantage of Cisco's
36304 patent portfolio license
36305 (<http://www.openh264.org/BINARY_LICENSE.txt>).
36306 OpenJPEG
36308 FFmpeg can use the OpenJPEG libraries for decoding/encoding J2K videos.
36309 Go to <http://www.openjpeg.org/> to get the libraries and follow the
36310 installation instructions. To enable using OpenJPEG in FFmpeg, pass
36311 "--enable-libopenjpeg" to ./configure.
36312 rav1e
36314 FFmpeg can make use of rav1e (Rust AV1 Encoder) via its C bindings to
36315 encode videos. Go to <https://github.com/xiph/rav1e/> and follow the
36316 instructions to build the C library. To enable using rav1e in FFmpeg,
36317 pass "--enable-librav1e" to ./configure.
36318 SVT-AV1
36320 FFmpeg can make use of the Scalable Video Technology for AV1 library
36321 for AV1 encoding.
36322 Go to <https://gitlab.com/AOMediaCodec/SVT-AV1/> and follow the
36324 instructions for installing the library. Then pass "--enable-libsvtav1"
36325 to configure to enable it.
36326 TwoLAME
36328 FFmpeg can make use of the TwoLAME library for MP2 encoding.
36329 Go to <http://www.twolame.org/> and follow the instructions for
36331 installing the library. Then pass "--enable-libtwolame" to configure
36332 to enable it.
36333 VapourSynth
36335 FFmpeg can read VapourSynth scripts as input. To enable support, pass
36336 "--enable-vapoursynth" to configure. Vapoursynth is detected via
36337 "pkg-config". Versions 42 or greater supported. See
36338 <http://www.vapoursynth.com/>.
36339 Due to security concerns, Vapoursynth scripts will not be autodetected
36341 so the input format has to be forced. For ff* CLI tools, add "-f
36342 vapoursynth" before the input "-i yourscript.vpy".
36343 x264
36345 FFmpeg can make use of the x264 library for H.264 encoding.
36346 Go to <http://www.videolan.org/developers/x264.html> and follow the
36348 instructions for installing the library. Then pass "--enable-libx264"
36349 to configure to enable it.
36350 x264 is under the GNU Public License Version 2 or later (see
36352 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
36353 details), you must upgrade FFmpeg's license to GPL in order to use
36354 it.
36355 x265
36357 FFmpeg can make use of the x265 library for HEVC encoding.
36358 Go to <http://x265.org/developers.html> and follow the instructions for
36360 installing the library. Then pass "--enable-libx265" to configure to
36361 enable it.
36362 x265 is under the GNU Public License Version 2 or later (see
36364 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
36365 details), you must upgrade FFmpeg's license to GPL in order to use
36366 it.
36367 xavs
36369 FFmpeg can make use of the xavs library for AVS encoding.
36370 Go to <http://xavs.sf.net/> and follow the instructions for installing
36372 the library. Then pass "--enable-libxavs" to configure to enable it.
36373 xavs2
36375 FFmpeg can make use of the xavs2 library for AVS2-P2/IEEE1857.4 video
36376 encoding.
36377 Go to <https://github.com/pkuvcl/xavs2> and follow the instructions for
36379 installing the library. Then pass "--enable-libxavs2" to configure to
36380 enable it.
36381 libxavs2 is under the GNU Public License Version 2 or later (see
36383 <http://www.gnu.org/licenses/old-licenses/gpl-2.0.html> for
36384 details), you must upgrade FFmpeg's license to GPL in order to use
36385 it.
36386 ZVBI
36388 ZVBI is a VBI decoding library which can be used by FFmpeg to decode
36389 DVB teletext pages and DVB teletext subtitles.
36390 Go to <http://sourceforge.net/projects/zapping/> and follow the
36392 instructions for installing the library. Then pass "--enable-libzvbi"
36393 to configure to enable it.
36394
36396 You can use the "-formats" and "-codecs" options to have an exhaustive
36397 list.
36398 File Formats
36400 FFmpeg supports the following file formats through the "libavformat"
36401 library:
36402 Name : Encoding @tab Decoding @tab Comments
36404 3dostr : @tab X
36405 4xm : @tab X
36406 @tab 4X Technologies format, used in some games.
36407 8088flex TMV : @tab X
36409 AAX : @tab X
36410 @tab Audible Enhanced Audio format, used in audiobooks.
36411 AA : @tab X
36413 @tab Audible Format 2, 3, and 4, used in audiobooks.
36414 ACT Voice : @tab X
36416 @tab contains G.729 audio
36417 Adobe Filmstrip : X @tab X
36419 Audio IFF (AIFF) : X @tab X
36420 American Laser Games MM : @tab X
36421 @tab Multimedia format used in games like Mad Dog McCree.
36422 3GPP AMR : X @tab X
36424 Amazing Studio Packed Animation File : @tab X
36425 @tab Multimedia format used in game Heart Of Darkness.
36426 Apple HTTP Live Streaming : @tab X
36428 Artworx Data Format : @tab X
36429 Interplay ACM : @tab X
36430 @tab Audio only format used in some Interplay games.
36431 ADP : @tab X
36433 @tab Audio format used on the Nintendo Gamecube.
36434 AFC : @tab X
36436 @tab Audio format used on the Nintendo Gamecube.
36437 ADS/SS2 : @tab X
36439 @tab Audio format used on the PS2.
36440 APNG : X @tab X
36442 ASF : X @tab X
36443 @tab Advanced / Active Streaming Format.
36444 AST : X @tab X
36446 @tab Audio format used on the Nintendo Wii.
36447 AVI : X @tab X
36449 AviSynth : @tab X
36450 AVR : @tab X
36451 @tab Audio format used on Mac.
36452 AVS : @tab X
36454 @tab Multimedia format used by the Creature Shock game.
36455 Beam Software SIFF : @tab X
36457 @tab Audio and video format used in some games by Beam Software.
36458 Bethesda Softworks VID : @tab X
36460 @tab Used in some games from Bethesda Softworks.
36461 Binary text : @tab X
36463 Bink : @tab X
36464 @tab Multimedia format used by many games.
36465 Bink Audio : @tab X
36467 @tab Audio only multimedia format used by some games.
36468 Bitmap Brothers JV : @tab X
36470 @tab Used in Z and Z95 games.
36471 BRP : @tab X
36473 @tab Argonaut Games format.
36474 Brute Force & Ignorance : @tab X
36476 @tab Used in the game Flash Traffic: City of Angels.
36477 BFSTM : @tab X
36479 @tab Audio format used on the Nintendo WiiU (based on BRSTM).
36480 BRSTM : @tab X
36482 @tab Audio format used on the Nintendo Wii.
36483 BW64 : @tab X
36485 @tab Broadcast Wave 64bit.
36486 BWF : X @tab X
36488 codec2 (raw) : X @tab X
36489 @tab Must be given -mode format option to decode correctly.
36490 codec2 (.c2 files) : X @tab X
36492 @tab Contains header with version and mode info, simplifying playback.
36493 CRI ADX : X @tab X
36495 @tab Audio-only format used in console video games.
36496 CRI AIX : @tab X
36498 CRI HCA : @tab X
36499 @tab Audio-only format used in console video games.
36500 Discworld II BMV : @tab X
36502 Interplay C93 : @tab X
36503 @tab Used in the game Cyberia from Interplay.
36504 Delphine Software International CIN : @tab X
36506 @tab Multimedia format used by Delphine Software games.
36507 Digital Speech Standard (DSS) : @tab X
36509 CD+G : @tab X
36510 @tab Video format used by CD+G karaoke disks
36511 Phantom Cine : @tab X
36513 Commodore CDXL : @tab X
36514 @tab Amiga CD video format
36515 Core Audio Format : X @tab X
36517 @tab Apple Core Audio Format
36518 CRC testing format : X @tab
36520 Creative Voice : X @tab X
36521 @tab Created for the Sound Blaster Pro.
36522 CRYO APC : @tab X
36524 @tab Audio format used in some games by CRYO Interactive Entertainment.
36525 D-Cinema audio : X @tab X
36527 Deluxe Paint Animation : @tab X
36528 DCSTR : @tab X
36529 DFA : @tab X
36530 @tab This format is used in Chronomaster game
36531 DirectDraw Surface : @tab X
36533 DSD Stream File (DSF) : @tab X
36534 DV video : X @tab X
36535 DXA : @tab X
36536 @tab This format is used in the non-Windows version of the Feeble Files
36537 game and different game cutscenes repacked for use with ScummVM.
36538 Electronic Arts cdata : @tab X
36540 Electronic Arts Multimedia : @tab X
36541 @tab Used in various EA games; files have extensions like WVE and UV2.
36542 Ensoniq Paris Audio File : @tab X
36544 FFM (FFserver live feed) : X @tab X
36545 Flash (SWF) : X @tab X
36546 Flash 9 (AVM2) : X @tab X
36547 @tab Only embedded audio is decoded.
36548 FLI/FLC/FLX animation : @tab X
36550 @tab .fli/.flc files
36551 Flash Video (FLV) : X @tab X
36553 @tab Macromedia Flash video files
36554 framecrc testing format : X @tab
36556 FunCom ISS : @tab X
36557 @tab Audio format used in various games from FunCom like The Longest Journey.
36558 G.723.1 : X @tab X
36560 G.726 : @tab X @tab Both left- and right-
36561 justified.
36562 G.729 BIT : X @tab X
36563 G.729 raw : @tab X
36564 GENH : @tab X
36565 @tab Audio format for various games.
36566 GIF Animation : X @tab X
36568 GXF : X @tab X
36569 @tab General eXchange Format SMPTE 360M, used by Thomson Grass Valley
36570 playout servers.
36571 HNM : @tab X
36573 @tab Only version 4 supported, used in some games from Cryo Interactive
36574 iCEDraw File : @tab X
36576 ICO : X @tab X
36577 @tab Microsoft Windows ICO
36578 id Quake II CIN video : @tab X
36580 id RoQ : X @tab X
36581 @tab Used in Quake III, Jedi Knight 2 and other computer games.
36582 IEC61937 encapsulation : X @tab X
36584 IFF : @tab X
36585 @tab Interchange File Format
36586 IFV : @tab X
36588 @tab A format used by some old CCTV DVRs.
36589 iLBC : X @tab X
36591 Interplay MVE : @tab X
36592 @tab Format used in various Interplay computer games.
36593 Iterated Systems ClearVideo : @tab X
36595 @tab I-frames only
36596 IV8 : @tab X
36598 @tab A format generated by IndigoVision 8000 video server.
36599 IVF (On2) : X @tab X
36601 @tab A format used by libvpx
36602 Internet Video Recording : @tab X
36604 IRCAM : X @tab X
36605 LATM : X @tab X
36606 LMLM4 : @tab X
36607 @tab Used by Linux Media Labs MPEG-4 PCI boards
36608 LOAS : @tab X
36610 @tab contains LATM multiplexed AAC audio
36611 LRC : X @tab X
36613 LVF : @tab X
36614 LXF : @tab X
36615 @tab VR native stream format, used by Leitch/Harris' video servers.
36616 Magic Lantern Video (MLV) : @tab X
36618 Matroska : X @tab X
36619 Matroska audio : X @tab
36620 FFmpeg metadata : X @tab X
36621 @tab Metadata in text format.
36622 MAXIS XA : @tab X
36624 @tab Used in Sim City 3000; file extension .xa.
36625 MCA : @tab X
36627 @tab Used in some games from Capcom; file extension .mca.
36628 MD Studio : @tab X
36630 Metal Gear Solid: The Twin Snakes : @tab X
36631 Megalux Frame : @tab X
36632 @tab Used by Megalux Ultimate Paint
36633 Mobotix .mxg : @tab X
36635 Monkey's Audio : @tab X
36636 Motion Pixels MVI : @tab X
36637 MOV/QuickTime/MP4 : X @tab X
36638 @tab 3GP, 3GP2, PSP, iPod variants supported
36639 MP2 : X @tab X
36641 MP3 : X @tab X
36642 MPEG-1 System : X @tab X
36643 @tab muxed audio and video, VCD format supported
36644 MPEG-PS (program stream) : X @tab X
36646 @tab also known as C<VOB> file, SVCD and DVD format supported
36647 MPEG-TS (transport stream) : X @tab X
36649 @tab also known as DVB Transport Stream
36650 MPEG-4 : X @tab X
36652 @tab MPEG-4 is a variant of QuickTime.
36653 MSF : @tab X
36655 @tab Audio format used on the PS3.
36656 Mirillis FIC video : @tab X
36658 @tab No cursor rendering.
36659 MIDI Sample Dump Standard : @tab X
36661 MIME multipart JPEG : X @tab
36662 MSN TCP webcam : @tab X
36663 @tab Used by MSN Messenger webcam streams.
36664 MTV : @tab X
36666 Musepack : @tab X
36667 Musepack SV8 : @tab X
36668 Material eXchange Format (MXF) : X @tab X
36669 @tab SMPTE 377M, used by D-Cinema, broadcast industry.
36670 Material eXchange Format (MXF), D-10 Mapping : X @tab X
36672 @tab SMPTE 386M, D-10/IMX Mapping.
36673 NC camera feed : @tab X
36675 @tab NC (AVIP NC4600) camera streams
36676 NIST SPeech HEader REsources : @tab X
36678 Computerized Speech Lab NSP : @tab X
36679 NTT TwinVQ (VQF) : @tab X
36680 @tab Nippon Telegraph and Telephone Corporation TwinVQ.
36681 Nullsoft Streaming Video : @tab X
36683 NuppelVideo : @tab X
36684 NUT : X @tab X
36685 @tab NUT Open Container Format
36686 Ogg : X @tab X
36688 Playstation Portable PMP : @tab X
36689 Portable Voice Format : @tab X
36690 TechnoTrend PVA : @tab X
36691 @tab Used by TechnoTrend DVB PCI boards.
36692 QCP : @tab X
36694 raw ADTS (AAC) : X @tab X
36695 raw AC-3 : X @tab X
36696 raw AMR-NB : @tab X
36697 raw AMR-WB : @tab X
36698 raw aptX : X @tab X
36699 raw aptX HD : X @tab X
36700 raw Chinese AVS video : X @tab X
36701 raw DFPWM : X @tab X
36702 raw Dirac : X @tab X
36703 raw DNxHD : X @tab X
36704 raw DTS : X @tab X
36705 raw DTS-HD : @tab X
36706 raw E-AC-3 : X @tab X
36707 raw FLAC : X @tab X
36708 raw GSM : @tab X
36709 raw H.261 : X @tab X
36710 raw H.263 : X @tab X
36711 raw H.264 : X @tab X
36712 raw HEVC : X @tab X
36713 raw Ingenient MJPEG : @tab X
36714 raw MJPEG : X @tab X
36715 raw MLP : @tab X
36716 raw MPEG : @tab X
36717 raw MPEG-1 : @tab X
36718 raw MPEG-2 : @tab X
36719 raw MPEG-4 : X @tab X
36720 raw NULL : X @tab
36721 raw video : X @tab X
36722 raw id RoQ : X @tab
36723 raw OBU : X @tab X
36724 raw SBC : X @tab X
36725 raw Shorten : @tab X
36726 raw TAK : @tab X
36727 raw TrueHD : X @tab X
36728 raw VC-1 : X @tab X
36729 raw PCM A-law : X @tab X
36730 raw PCM mu-law : X @tab X
36731 raw PCM Archimedes VIDC : X @tab X
36732 raw PCM signed 8 bit : X @tab X
36733 raw PCM signed 16 bit big-endian : X @tab X
36734 raw PCM signed 16 bit little-endian : X @tab X
36735 raw PCM signed 24 bit big-endian : X @tab X
36736 raw PCM signed 24 bit little-endian : X @tab X
36737 raw PCM signed 32 bit big-endian : X @tab X
36738 raw PCM signed 32 bit little-endian : X @tab X
36739 raw PCM signed 64 bit big-endian : X @tab X
36740 raw PCM signed 64 bit little-endian : X @tab X
36741 raw PCM unsigned 8 bit : X @tab X
36742 raw PCM unsigned 16 bit big-endian : X @tab X
36743 raw PCM unsigned 16 bit little-endian : X @tab X
36744 raw PCM unsigned 24 bit big-endian : X @tab X
36745 raw PCM unsigned 24 bit little-endian : X @tab X
36746 raw PCM unsigned 32 bit big-endian : X @tab X
36747 raw PCM unsigned 32 bit little-endian : X @tab X
36748 raw PCM 16.8 floating point little-endian : @tab X
36749 raw PCM 24.0 floating point little-endian : @tab X
36750 raw PCM floating-point 32 bit big-endian : X @tab X
36751 raw PCM floating-point 32 bit little-endian : X @tab X
36752 raw PCM floating-point 64 bit big-endian : X @tab X
36753 raw PCM floating-point 64 bit little-endian : X @tab X
36754 RDT : @tab X
36755 REDCODE R3D : @tab X
36756 @tab File format used by RED Digital cameras, contains JPEG 2000 frames and PCM audio.
36757 RealMedia : X @tab X
36759 Redirector : @tab X
36760 RedSpark : @tab X
36761 Renderware TeXture Dictionary : @tab X
36762 Resolume DXV : @tab X
36763 RF64 : @tab X
36764 RL2 : @tab X
36765 @tab Audio and video format used in some games by Entertainment Software Partners.
36766 RPL/ARMovie : @tab X
36768 Lego Mindstorms RSO : X @tab X
36769 RSD : @tab X
36770 RTMP : X @tab X
36771 @tab Output is performed by publishing stream to RTMP server
36772 RTP : X @tab X
36774 RTSP : X @tab X
36775 Sample Dump eXchange : @tab X
36776 SAP : X @tab X
36777 SBG : @tab X
36778 SDP : @tab X
36779 SER : @tab X
36780 Sega FILM/CPK : X @tab X
36781 @tab Used in many Sega Saturn console games.
36782 Silicon Graphics Movie : @tab X
36784 Sierra SOL : @tab X
36785 @tab .sol files used in Sierra Online games.
36786 Sierra VMD : @tab X
36788 @tab Used in Sierra CD-ROM games.
36789 Smacker : @tab X
36791 @tab Multimedia format used by many games.
36792 SMJPEG : X @tab X
36794 @tab Used in certain Loki game ports.
36795 SMPTE 337M encapsulation : @tab X
36797 Smush : @tab X
36798 @tab Multimedia format used in some LucasArts games.
36799 Sony OpenMG (OMA) : X @tab X
36801 @tab Audio format used in Sony Sonic Stage and Sony Vegas.
36802 Sony PlayStation STR : @tab X
36804 Sony Wave64 (W64) : X @tab X
36805 SoX native format : X @tab X
36806 SUN AU format : X @tab X
36807 SUP raw PGS subtitles : X @tab X
36808 SVAG : @tab X
36809 @tab Audio format used in Konami PS2 games.
36810 TDSC : @tab X
36812 Text files : @tab X
36813 THP : @tab X
36814 @tab Used on the Nintendo GameCube.
36815 Tiertex Limited SEQ : @tab X
36817 @tab Tiertex .seq files used in the DOS CD-ROM version of the game Flashback.
36818 True Audio : X @tab X
36820 VAG : @tab X
36821 @tab Audio format used in many Sony PS2 games.
36822 VC-1 test bitstream : X @tab X
36824 Vidvox Hap : X @tab X
36825 Vivo : @tab X
36826 VPK : @tab X
36827 @tab Audio format used in Sony PS games.
36828 WAV : X @tab X
36830 WavPack : X @tab X
36831 WebM : X @tab X
36832 Windows Televison (WTV) : X @tab X
36833 Wing Commander III movie : @tab X
36834 @tab Multimedia format used in Origin's Wing Commander III computer game.
36835 Westwood Studios audio : X @tab X
36837 @tab Multimedia format used in Westwood Studios games.
36838 Westwood Studios VQA : @tab X
36840 @tab Multimedia format used in Westwood Studios games.
36841 Wideband Single-bit Data (WSD) : @tab X
36843 WVE : @tab X
36844 XMV : @tab X
36845 @tab Microsoft video container used in Xbox games.
36846 XVAG : @tab X
36848 @tab Audio format used on the PS3.
36849 xWMA : @tab X
36851 @tab Microsoft audio container used by XAudio 2.
36852 eXtended BINary text (XBIN) : @tab X
36854 YUV4MPEG pipe : X @tab X
36855 Psygnosis YOP : @tab X
36856 "X" means that the feature in that column (encoding / decoding) is
36858 supported.
36859 Image Formats
36861 FFmpeg can read and write images for each frame of a video sequence.
36862 The following image formats are supported:
36863 Name : Encoding @tab Decoding @tab Comments
36865 .Y.U.V : X @tab X
36866 @tab one raw file per component
36867 Alias PIX : X @tab X
36869 @tab Alias/Wavefront PIX image format
36870 animated GIF : X @tab X
36872 APNG : X @tab X
36873 @tab Animated Portable Network Graphics
36874 BMP : X @tab X
36876 @tab Microsoft BMP image
36877 BRender PIX : @tab X
36879 @tab Argonaut BRender 3D engine image format.
36880 CRI : @tab X
36882 @tab Cintel RAW
36883 DPX : X @tab X
36885 @tab Digital Picture Exchange
36886 EXR : @tab X
36888 @tab OpenEXR
36889 FITS : X @tab X
36891 @tab Flexible Image Transport System
36892 IMG : @tab X
36894 @tab GEM Raster image
36895 JPEG : X @tab X
36897 @tab Progressive JPEG is not supported.
36898 JPEG 2000 : X @tab X
36900 JPEG-LS : X @tab X
36901 LJPEG : X @tab
36902 @tab Lossless JPEG
36903 MSP : @tab X
36905 @tab Microsoft Paint image
36906 PAM : X @tab X
36908 @tab PAM is a PNM extension with alpha support.
36909 PBM : X @tab X
36911 @tab Portable BitMap image
36912 PCD : @tab X
36914 @tab PhotoCD
36915 PCX : X @tab X
36917 @tab PC Paintbrush
36918 PFM : X @tab X
36920 @tab Portable FloatMap image
36921 PGM : X @tab X
36923 @tab Portable GrayMap image
36924 PGMYUV : X @tab X
36926 @tab PGM with U and V components in YUV 4:2:0
36927 PGX : @tab X
36929 @tab PGX file decoder
36930 PHM : X @tab X
36932 @tab Portable HalfFloatMap image
36933 PIC : @tab X
36935 @tab Pictor/PC Paint
36936 PNG : X @tab X
36938 @tab Portable Network Graphics image
36939 PPM : X @tab X
36941 @tab Portable PixelMap image
36942 PSD : @tab X
36944 @tab Photoshop
36945 PTX : @tab X
36947 @tab V.Flash PTX format
36948 QOI : X @tab X
36950 @tab Quite OK Image format
36951 SGI : X @tab X
36953 @tab SGI RGB image format
36954 Sun Rasterfile : X @tab X
36956 @tab Sun RAS image format
36957 TIFF : X @tab X
36959 @tab YUV, JPEG and some extension is not supported yet.
36960 Truevision Targa : X @tab X
36962 @tab Targa (.TGA) image format
36963 VBN : X @tab X
36965 @tab Vizrt Binary Image format
36966 WebP : E @tab X
36968 @tab WebP image format, encoding supported through external library libwebp
36969 XBM : X @tab X
36971 @tab X BitMap image format
36972 XFace : X @tab X
36974 @tab X-Face image format
36975 XPM : @tab X
36977 @tab X PixMap image format
36978 XWD : X @tab X
36980 @tab X Window Dump image format
36981 "X" means that the feature in that column (encoding / decoding) is
36983 supported.
36984 "E" means that support is provided through an external library.
36986 Video Codecs
36988 Name : Encoding @tab Decoding @tab Comments
36989 4X Movie : @tab X
36990 @tab Used in certain computer games.
36991 8088flex TMV : @tab X
36993 A64 multicolor : X @tab
36994 @tab Creates video suitable to be played on a commodore 64 (multicolor mode).
36995 Amazing Studio PAF Video : @tab X
36997 American Laser Games MM : @tab X
36998 @tab Used in games like Mad Dog McCree.
36999 Amuse Graphics Movie : @tab X
37001 AMV Video : X @tab X
37002 @tab Used in Chinese MP3 players.
37003 ANSI/ASCII art : @tab X
37005 Apple Intermediate Codec : @tab X
37006 Apple MJPEG-B : @tab X
37007 Apple Pixlet : @tab X
37008 Apple ProRes : X @tab X
37009 @tab fourcc: apch,apcn,apcs,apco,ap4h,ap4x
37010 Apple QuickDraw : @tab X
37012 @tab fourcc: qdrw
37013 Argonaut Video : @tab X
37015 @tab Used in some Argonaut games.
37016 Asus v1 : X @tab X
37018 @tab fourcc: ASV1
37019 Asus v2 : X @tab X
37021 @tab fourcc: ASV2
37022 ATI VCR1 : @tab X
37024 @tab fourcc: VCR1
37025 ATI VCR2 : @tab X
37027 @tab fourcc: VCR2
37028 Auravision Aura : @tab X
37030 Auravision Aura 2 : @tab X
37031 Autodesk Animator Flic video : @tab X
37032 Autodesk RLE : @tab X
37033 @tab fourcc: AASC
37034 AV1 : E @tab E
37036 @tab Supported through external libraries libaom, libdav1d, librav1e and libsvtav1
37037 Avid 1:1 10-bit RGB Packer : X @tab X
37039 @tab fourcc: AVrp
37040 AVS (Audio Video Standard) video : @tab X
37042 @tab Video encoding used by the Creature Shock game.
37043 AVS2-P2/IEEE1857.4 : E @tab E
37045 @tab Supported through external libraries libxavs2 and libdavs2
37046 AVS3-P2/IEEE1857.10 : @tab E
37048 @tab Supported through external library libuavs3d
37049 AYUV : X @tab X
37051 @tab Microsoft uncompressed packed 4:4:4:4
37052 Beam Software VB : @tab X
37054 Bethesda VID video : @tab X
37055 @tab Used in some games from Bethesda Softworks.
37056 Bink Video : @tab X
37058 BitJazz SheerVideo : @tab X
37059 Bitmap Brothers JV video : @tab X
37060 y41p Brooktree uncompressed 4:1:1 12-bit : X @tab X
37061 Brooktree ProSumer Video : @tab X
37062 @tab fourcc: BT20
37063 Brute Force & Ignorance : @tab X
37065 @tab Used in the game Flash Traffic: City of Angels.
37066 C93 video : @tab X
37068 @tab Codec used in Cyberia game.
37069 CamStudio : @tab X
37071 @tab fourcc: CSCD
37072 CD+G : @tab X
37074 @tab Video codec for CD+G karaoke disks
37075 CDXL : @tab X
37077 @tab Amiga CD video codec
37078 Chinese AVS video : E @tab X
37080 @tab AVS1-P2, JiZhun profile, encoding through external library libxavs
37081 Delphine Software International CIN video : @tab X
37083 @tab Codec used in Delphine Software International games.
37084 Discworld II BMV Video : @tab X
37086 CineForm HD : X @tab X
37087 Canopus HQ : @tab X
37088 Canopus HQA : @tab X
37089 Canopus HQX : @tab X
37090 Canopus Lossless Codec : @tab X
37091 CDToons : @tab X
37092 @tab Codec used in various Broderbund games.
37093 Cinepak : @tab X
37095 Cirrus Logic AccuPak : X @tab X
37096 @tab fourcc: CLJR
37097 CPiA Video Format : @tab X
37099 Creative YUV (CYUV) : @tab X
37100 DFA : @tab X
37101 @tab Codec used in Chronomaster game.
37102 Dirac : E @tab X
37104 @tab supported though the native vc2 (Dirac Pro) encoder
37105 Deluxe Paint Animation : @tab X
37107 DNxHD : X @tab X
37108 @tab aka SMPTE VC3
37109 Duck TrueMotion 1.0 : @tab X
37111 @tab fourcc: DUCK
37112 Duck TrueMotion 2.0 : @tab X
37114 @tab fourcc: TM20
37115 Duck TrueMotion 2.0 RT : @tab X
37117 @tab fourcc: TR20
37118 DV (Digital Video) : X @tab X
37120 Dxtory capture format : @tab X
37121 Feeble Files/ScummVM DXA : @tab X
37122 @tab Codec originally used in Feeble Files game.
37123 Electronic Arts CMV video : @tab X
37125 @tab Used in NHL 95 game.
37126 Electronic Arts Madcow video : @tab X
37128 Electronic Arts TGV video : @tab X
37129 Electronic Arts TGQ video : @tab X
37130 Electronic Arts TQI video : @tab X
37131 Escape 124 : @tab X
37132 Escape 130 : @tab X
37133 FFmpeg video codec #1 : X @tab X
37134 @tab lossless codec (fourcc: FFV1)
37135 Flash Screen Video v1 : X @tab X
37137 @tab fourcc: FSV1
37138 Flash Screen Video v2 : X @tab X
37140 Flash Video (FLV) : X @tab X
37141 @tab Sorenson H.263 used in Flash
37142 FM Screen Capture Codec : @tab X
37144 Forward Uncompressed : @tab X
37145 Fraps : @tab X
37146 Go2Meeting : @tab X
37147 @tab fourcc: G2M2, G2M3
37148 Go2Webinar : @tab X
37150 @tab fourcc: G2M4
37151 Gremlin Digital Video : @tab X
37153 H.261 : X @tab X
37154 H.263 / H.263-1996 : X @tab X
37155 H.263+ / H.263-1998 / H.263 version 2 : X @tab X
37156 H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10 : E @tab X
37157 @tab encoding supported through external library libx264 and OpenH264
37158 HEVC : X @tab X
37160 @tab encoding supported through external library libx265 and libkvazaar
37161 HNM version 4 : @tab X
37163 HuffYUV : X @tab X
37164 HuffYUV FFmpeg variant : X @tab X
37165 IBM Ultimotion : @tab X
37166 @tab fourcc: ULTI
37167 id Cinematic video : @tab X
37169 @tab Used in Quake II.
37170 id RoQ video : X @tab X
37172 @tab Used in Quake III, Jedi Knight 2, other computer games.
37173 IFF ILBM : @tab X
37175 @tab IFF interleaved bitmap
37176 IFF ByteRun1 : @tab X
37178 @tab IFF run length encoded bitmap
37179 Infinity IMM4 : @tab X
37181 Intel H.263 : @tab X
37182 Intel Indeo 2 : @tab X
37183 Intel Indeo 3 : @tab X
37184 Intel Indeo 4 : @tab X
37185 Intel Indeo 5 : @tab X
37186 Interplay C93 : @tab X
37187 @tab Used in the game Cyberia from Interplay.
37188 Interplay MVE video : @tab X
37190 @tab Used in Interplay .MVE files.
37191 J2K : X @tab X
37193 Karl Morton's video codec : @tab X
37194 @tab Codec used in Worms games.
37195 Kega Game Video (KGV1) : @tab X
37197 @tab Kega emulator screen capture codec.
37198 Lagarith : @tab X
37200 LCL (LossLess Codec Library) MSZH : @tab X
37201 LCL (LossLess Codec Library) ZLIB : E @tab E
37202 LOCO : @tab X
37203 LucasArts SANM/Smush : @tab X
37204 @tab Used in LucasArts games / SMUSH animations.
37205 lossless MJPEG : X @tab X
37207 MagicYUV Video : X @tab X
37208 Mandsoft Screen Capture Codec : @tab X
37209 Microsoft ATC Screen : @tab X
37210 @tab Also known as Microsoft Screen 3.
37211 Microsoft Expression Encoder Screen : @tab X
37213 @tab Also known as Microsoft Titanium Screen 2.
37214 Microsoft RLE : @tab X
37216 Microsoft Screen 1 : @tab X
37217 @tab Also known as Windows Media Video V7 Screen.
37218 Microsoft Screen 2 : @tab X
37220 @tab Also known as Windows Media Video V9 Screen.
37221 Microsoft Video 1 : @tab X
37223 Mimic : @tab X
37224 @tab Used in MSN Messenger Webcam streams.
37225 Miro VideoXL : @tab X
37227 @tab fourcc: VIXL
37228 MJPEG (Motion JPEG) : X @tab X
37230 Mobotix MxPEG video : @tab X
37231 Motion Pixels video : @tab X
37232 MPEG-1 video : X @tab X
37233 MPEG-2 video : X @tab X
37234 MPEG-4 part 2 : X @tab X
37235 @tab libxvidcore can be used alternatively for encoding.
37236 MPEG-4 part 2 Microsoft variant version 1 : @tab X
37238 MPEG-4 part 2 Microsoft variant version 2 : X @tab X
37239 MPEG-4 part 2 Microsoft variant version 3 : X @tab X
37240 Newtek SpeedHQ : X @tab X
37241 Nintendo Gamecube THP video : @tab X
37242 NotchLC : @tab X
37243 NuppelVideo/RTjpeg : @tab X
37244 @tab Video encoding used in NuppelVideo files.
37245 On2 VP3 : @tab X
37247 @tab still experimental
37248 On2 VP4 : @tab X
37250 @tab fourcc: VP40
37251 On2 VP5 : @tab X
37253 @tab fourcc: VP50
37254 On2 VP6 : @tab X
37256 @tab fourcc: VP60,VP61,VP62
37257 On2 VP7 : @tab X
37259 @tab fourcc: VP70,VP71
37260 VP8 : E @tab X
37262 @tab fourcc: VP80, encoding supported through external library libvpx
37263 VP9 : E @tab X
37265 @tab encoding supported through external library libvpx
37266 Pinnacle TARGA CineWave YUV16 : @tab X
37268 @tab fourcc: Y216
37269 Q-team QPEG : @tab X
37271 @tab fourccs: QPEG, Q1.0, Q1.1
37272 QuickTime 8BPS video : @tab X
37274 QuickTime Animation (RLE) video : X @tab X
37275 @tab fourcc: 'rle '
37276 QuickTime Graphics (SMC) : X @tab X
37278 @tab fourcc: 'smc '
37279 QuickTime video (RPZA) : X @tab X
37281 @tab fourcc: rpza
37282 R10K AJA Kona 10-bit RGB Codec : X @tab X
37284 R210 Quicktime Uncompressed RGB 10-bit : X @tab X
37285 Raw Video : X @tab X
37286 RealVideo 1.0 : X @tab X
37287 RealVideo 2.0 : X @tab X
37288 RealVideo 3.0 : @tab X
37289 @tab still far from ideal
37290 RealVideo 4.0 : @tab X
37292 Renderware TXD (TeXture Dictionary) : @tab X
37293 @tab Texture dictionaries used by the Renderware Engine.
37294 RL2 video : @tab X
37296 @tab used in some games by Entertainment Software Partners
37297 ScreenPressor : @tab X
37299 Screenpresso : @tab X
37300 Screen Recorder Gold Codec : @tab X
37301 Sierra VMD video : @tab X
37302 @tab Used in Sierra VMD files.
37303 Silicon Graphics Motion Video Compressor 1 (MVC1) : @tab X
37305 Silicon Graphics Motion Video Compressor 2 (MVC2) : @tab X
37306 Silicon Graphics RLE 8-bit video : @tab X
37307 Smacker video : @tab X
37308 @tab Video encoding used in Smacker.
37309 SMPTE VC-1 : @tab X
37311 Snow : X @tab X
37312 @tab experimental wavelet codec (fourcc: SNOW)
37313 Sony PlayStation MDEC (Motion DECoder) : @tab X
37315 Sorenson Vector Quantizer 1 : X @tab X
37316 @tab fourcc: SVQ1
37317 Sorenson Vector Quantizer 3 : @tab X
37319 @tab fourcc: SVQ3
37320 Sunplus JPEG (SP5X) : @tab X
37322 @tab fourcc: SP5X
37323 TechSmith Screen Capture Codec : @tab X
37325 @tab fourcc: TSCC
37326 TechSmith Screen Capture Codec 2 : @tab X
37328 @tab fourcc: TSC2
37329 Theora : E @tab X
37331 @tab encoding supported through external library libtheora
37332 Tiertex Limited SEQ video : @tab X
37334 @tab Codec used in DOS CD-ROM FlashBack game.
37335 Ut Video : X @tab X
37337 v210 QuickTime uncompressed 4:2:2 10-bit : X @tab X
37338 v308 QuickTime uncompressed 4:4:4 : X @tab X
37339 v408 QuickTime uncompressed 4:4:4:4 : X @tab X
37340 v410 QuickTime uncompressed 4:4:4 10-bit : X @tab X
37341 VBLE Lossless Codec : @tab X
37342 VMware Screen Codec / VMware Video : @tab X
37343 @tab Codec used in videos captured by VMware.
37344 Westwood Studios VQA (Vector Quantized Animation) video : @tab
37346 X
37347 Windows Media Image : @tab X
37348 Windows Media Video 7 : X @tab X
37349 Windows Media Video 8 : X @tab X
37350 Windows Media Video 9 : @tab X
37351 @tab not completely working
37352 Wing Commander III / Xan : @tab X
37354 @tab Used in Wing Commander III .MVE files.
37355 Wing Commander IV / Xan : @tab X
37357 @tab Used in Wing Commander IV.
37358 Winnov WNV1 : @tab X
37360 WMV7 : X @tab X
37361 YAMAHA SMAF : X @tab X
37362 Psygnosis YOP Video : @tab X
37363 yuv4 : X @tab X
37364 @tab libquicktime uncompressed packed 4:2:0
37365 ZeroCodec Lossless Video : @tab X
37367 ZLIB : X @tab X
37368 @tab part of LCL, encoder experimental
37369 Zip Motion Blocks Video : X @tab X
37371 @tab Encoder works only in PAL8.
37372 "X" means that the feature in that column (encoding / decoding) is
37374 supported.
37375 "E" means that support is provided through an external library.
37377 Audio Codecs
37379 Name : Encoding @tab Decoding @tab Comments
37380 8SVX exponential : @tab X
37381 8SVX fibonacci : @tab X
37382 AAC : EX @tab X
37383 @tab encoding supported through internal encoder and external library libfdk-aac
37384 AAC+ : E @tab IX
37386 @tab encoding supported through external library libfdk-aac
37387 AC-3 : IX @tab IX
37389 ACELP.KELVIN : @tab X
37390 ADPCM 4X Movie : @tab X
37391 ADPCM Yamaha AICA : @tab X
37392 ADPCM AmuseGraphics Movie : @tab X
37393 ADPCM Argonaut Games : X @tab X
37394 ADPCM CDROM XA : @tab X
37395 ADPCM Creative Technology : @tab X
37396 @tab 16 -E<gt> 4, 8 -E<gt> 4, 8 -E<gt> 3, 8 -E<gt> 2
37397 ADPCM Electronic Arts : @tab X
37399 @tab Used in various EA titles.
37400 ADPCM Electronic Arts Maxis CDROM XS : @tab X
37402 @tab Used in Sim City 3000.
37403 ADPCM Electronic Arts R1 : @tab X
37405 ADPCM Electronic Arts R2 : @tab X
37406 ADPCM Electronic Arts R3 : @tab X
37407 ADPCM Electronic Arts XAS : @tab X
37408 ADPCM G.722 : X @tab X
37409 ADPCM G.726 : X @tab X
37410 ADPCM IMA Acorn Replay : @tab X
37411 ADPCM IMA AMV : X @tab X
37412 @tab Used in AMV files
37413 ADPCM IMA Cunning Developments : @tab X
37415 ADPCM IMA Electronic Arts EACS : @tab X
37416 ADPCM IMA Electronic Arts SEAD : @tab X
37417 ADPCM IMA Funcom : @tab X
37418 ADPCM IMA High Voltage Software ALP : X @tab X
37419 ADPCM IMA QuickTime : X @tab X
37420 ADPCM IMA Simon & Schuster Interactive : X @tab X
37421 ADPCM IMA Ubisoft APM : X @tab X
37422 ADPCM IMA Loki SDL MJPEG : @tab X
37423 ADPCM IMA WAV : X @tab X
37424 ADPCM IMA Westwood : @tab X
37425 ADPCM ISS IMA : @tab X
37426 @tab Used in FunCom games.
37427 ADPCM IMA Dialogic : @tab X
37429 ADPCM IMA Duck DK3 : @tab X
37430 @tab Used in some Sega Saturn console games.
37431 ADPCM IMA Duck DK4 : @tab X
37433 @tab Used in some Sega Saturn console games.
37434 ADPCM IMA Radical : @tab X
37436 ADPCM Microsoft : X @tab X
37437 ADPCM MS IMA : X @tab X
37438 ADPCM Nintendo Gamecube AFC : @tab X
37439 ADPCM Nintendo Gamecube DTK : @tab X
37440 ADPCM Nintendo THP : @tab X
37441 ADPCM Playstation : @tab X
37442 ADPCM QT IMA : X @tab X
37443 ADPCM SEGA CRI ADX : X @tab X
37444 @tab Used in Sega Dreamcast games.
37445 ADPCM Shockwave Flash : X @tab X
37447 ADPCM Sound Blaster Pro 2-bit : @tab X
37448 ADPCM Sound Blaster Pro 2.6-bit : @tab X
37449 ADPCM Sound Blaster Pro 4-bit : @tab X
37450 ADPCM VIMA : @tab X
37451 @tab Used in LucasArts SMUSH animations.
37452 ADPCM Westwood Studios IMA : X @tab X
37454 @tab Used in Westwood Studios games like Command and Conquer.
37455 ADPCM Yamaha : X @tab X
37457 ADPCM Zork : @tab X
37458 AMR-NB : E @tab X
37459 @tab encoding supported through external library libopencore-amrnb
37460 AMR-WB : E @tab X
37462 @tab encoding supported through external library libvo-amrwbenc
37463 Amazing Studio PAF Audio : @tab X
37465 Apple lossless audio : X @tab X
37466 @tab QuickTime fourcc 'alac'
37467 aptX : X @tab X
37469 @tab Used in Bluetooth A2DP
37470 aptX HD : X @tab X
37472 @tab Used in Bluetooth A2DP
37473 ATRAC1 : @tab X
37475 ATRAC3 : @tab X
37476 ATRAC3+ : @tab X
37477 ATRAC9 : @tab X
37478 Bink Audio : @tab X
37479 @tab Used in Bink and Smacker files in many games.
37480 CELT : @tab E
37482 @tab decoding supported through external library libcelt
37483 codec2 : E @tab E
37485 @tab en/decoding supported through external library libcodec2
37486 CRI HCA : @tab X
37488 Delphine Software International CIN audio : @tab X
37489 @tab Codec used in Delphine Software International games.
37490 DFPWM : X @tab X
37492 Digital Speech Standard - Standard Play mode (DSS SP) : @tab X
37493 Discworld II BMV Audio : @tab X
37494 COOK : @tab X
37495 @tab All versions except 5.1 are supported.
37496 DCA (DTS Coherent Acoustics) : X @tab X
37498 @tab supported extensions: XCh, XXCH, X96, XBR, XLL, LBR (partially)
37499 Dolby E : @tab X
37501 DPCM Gremlin : @tab X
37502 DPCM id RoQ : X @tab X
37503 @tab Used in Quake III, Jedi Knight 2 and other computer games.
37504 DPCM Interplay : @tab X
37506 @tab Used in various Interplay computer games.
37507 DPCM Squareroot-Delta-Exact : @tab X
37509 @tab Used in various games.
37510 DPCM Sierra Online : @tab X
37512 @tab Used in Sierra Online game audio files.
37513 DPCM Sol : @tab X
37515 DPCM Xan : @tab X
37516 @tab Used in Origin's Wing Commander IV AVI files.
37517 DPCM Xilam DERF : @tab X
37519 DSD (Direct Stream Digital), least significant bit first : @tab X
37520 DSD (Direct Stream Digital), most significant bit first : @tab X
37521 DSD (Direct Stream Digital), least significant bit first, planar :
37522 @tab X
37523 DSD (Direct Stream Digital), most significant bit first, planar :
37524 @tab X
37525 DSP Group TrueSpeech : @tab X
37526 DST (Direct Stream Transfer) : @tab X
37527 DV audio : @tab X
37528 Enhanced AC-3 : X @tab X
37529 EVRC (Enhanced Variable Rate Codec) : @tab X
37530 FLAC (Free Lossless Audio Codec) : X @tab IX
37531 G.723.1 : X @tab X
37532 G.729 : @tab X
37533 GSM : E @tab X
37534 @tab encoding supported through external library libgsm
37535 GSM Microsoft variant : E @tab X
37537 @tab encoding supported through external library libgsm
37538 IAC (Indeo Audio Coder) : @tab X
37540 iLBC (Internet Low Bitrate Codec) : E @tab EX
37541 @tab encoding and decoding supported through external library libilbc
37542 IMC (Intel Music Coder) : @tab X
37544 Interplay ACM : @tab X
37545 MACE (Macintosh Audio Compression/Expansion) 3:1 : @tab X
37546 MACE (Macintosh Audio Compression/Expansion) 6:1 : @tab X
37547 MLP (Meridian Lossless Packing) : X @tab X
37548 @tab Used in DVD-Audio discs.
37549 Monkey's Audio : @tab X
37551 MP1 (MPEG audio layer 1) : @tab IX
37552 MP2 (MPEG audio layer 2) : IX @tab IX
37553 @tab encoding supported also through external library TwoLAME
37554 MP3 (MPEG audio layer 3) : E @tab IX
37556 @tab encoding supported through external library LAME, ADU MP3 and MP3onMP4 also supported
37557 MPEG-4 Audio Lossless Coding (ALS) : @tab X
37559 Musepack SV7 : @tab X
37560 Musepack SV8 : @tab X
37561 Nellymoser Asao : X @tab X
37562 On2 AVC (Audio for Video Codec) : @tab X
37563 Opus : E @tab X
37564 @tab encoding supported through external library libopus
37565 PCM A-law : X @tab X
37567 PCM mu-law : X @tab X
37568 PCM Archimedes VIDC : X @tab X
37569 PCM signed 8-bit planar : X @tab X
37570 PCM signed 16-bit big-endian planar : X @tab X
37571 PCM signed 16-bit little-endian planar : X @tab X
37572 PCM signed 24-bit little-endian planar : X @tab X
37573 PCM signed 32-bit little-endian planar : X @tab X
37574 PCM 32-bit floating point big-endian : X @tab X
37575 PCM 32-bit floating point little-endian : X @tab X
37576 PCM 64-bit floating point big-endian : X @tab X
37577 PCM 64-bit floating point little-endian : X @tab X
37578 PCM D-Cinema audio signed 24-bit : X @tab X
37579 PCM signed 8-bit : X @tab X
37580 PCM signed 16-bit big-endian : X @tab X
37581 PCM signed 16-bit little-endian : X @tab X
37582 PCM signed 24-bit big-endian : X @tab X
37583 PCM signed 24-bit little-endian : X @tab X
37584 PCM signed 32-bit big-endian : X @tab X
37585 PCM signed 32-bit little-endian : X @tab X
37586 PCM signed 16/20/24-bit big-endian in MPEG-TS : @tab X
37587 PCM unsigned 8-bit : X @tab X
37588 PCM unsigned 16-bit big-endian : X @tab X
37589 PCM unsigned 16-bit little-endian : X @tab X
37590 PCM unsigned 24-bit big-endian : X @tab X
37591 PCM unsigned 24-bit little-endian : X @tab X
37592 PCM unsigned 32-bit big-endian : X @tab X
37593 PCM unsigned 32-bit little-endian : X @tab X
37594 QCELP / PureVoice : @tab X
37595 QDesign Music Codec 1 : @tab X
37596 QDesign Music Codec 2 : @tab X
37597 @tab There are still some distortions.
37598 RealAudio 1.0 (14.4K) : X @tab X
37600 @tab Real 14400 bit/s codec
37601 RealAudio 2.0 (28.8K) : @tab X
37603 @tab Real 28800 bit/s codec
37604 RealAudio 3.0 (dnet) : IX @tab X
37606 @tab Real low bitrate AC-3 codec
37607 RealAudio Lossless : @tab X
37609 RealAudio SIPR / ACELP.NET : @tab X
37610 SBC (low-complexity subband codec) : X @tab X
37611 @tab Used in Bluetooth A2DP
37612 Shorten : @tab X
37614 Sierra VMD audio : @tab X
37615 @tab Used in Sierra VMD files.
37616 Smacker audio : @tab X
37618 SMPTE 302M AES3 audio : X @tab X
37619 Sonic : X @tab X
37620 @tab experimental codec
37621 Sonic lossless : X @tab X
37623 @tab experimental codec
37624 Speex : E @tab EX
37626 @tab supported through external library libspeex
37627 TAK (Tom's lossless Audio Kompressor) : @tab X
37629 True Audio (TTA) : X @tab X
37630 TrueHD : X @tab X
37631 @tab Used in HD-DVD and Blu-Ray discs.
37632 TwinVQ (VQF flavor) : @tab X
37634 VIMA : @tab X
37635 @tab Used in LucasArts SMUSH animations.
37636 Vorbis : E @tab X
37638 @tab A native but very primitive encoder exists.
37639 Voxware MetaSound : @tab X
37641 WavPack : X @tab X
37642 Westwood Audio (SND1) : @tab X
37643 Windows Media Audio 1 : X @tab X
37644 Windows Media Audio 2 : X @tab X
37645 Windows Media Audio Lossless : @tab X
37646 Windows Media Audio Pro : @tab X
37647 Windows Media Audio Voice : @tab X
37648 Xbox Media Audio 1 : @tab X
37649 Xbox Media Audio 2 : @tab X
37650 "X" means that the feature in that column (encoding / decoding) is
37652 supported.
37653 "E" means that support is provided through an external library.
37655 "I" means that an integer-only version is available, too (ensures high
37657 performance on systems without hardware floating point support).
37658 Subtitle Formats
37660 Name : Muxing @tab Demuxing @tab Encoding @tab Decoding
37661 3GPP Timed Text : @tab @tab X @tab X
37662 AQTitle : @tab X @tab @tab X
37663 DVB : X @tab X @tab X @tab X
37664 DVB teletext : @tab X @tab @tab E
37665 DVD : X @tab X @tab X @tab X
37666 JACOsub : X @tab X @tab @tab X
37667 MicroDVD : X @tab X @tab @tab X
37668 MPL2 : @tab X @tab @tab X
37669 MPsub (MPlayer) : @tab X @tab @tab X
37670 PGS : @tab @tab @tab X
37671 PJS (Phoenix) : @tab X @tab @tab X
37672 RealText : @tab X @tab @tab X
37673 SAMI : @tab X @tab @tab X
37674 Spruce format (STL) : @tab X @tab @tab X
37675 SSA/ASS : X @tab X @tab X @tab X
37676 SubRip (SRT) : X @tab X @tab X @tab X
37677 SubViewer v1 : @tab X @tab @tab X
37678 SubViewer : @tab X @tab @tab X
37679 TED Talks captions : @tab X @tab @tab X
37680 TTML : X @tab @tab X @tab
37681 VobSub (IDX+SUB) : @tab X @tab @tab X
37682 VPlayer : @tab X @tab @tab X
37683 WebVTT : X @tab X @tab X @tab X
37684 XSUB : @tab @tab X @tab X
37685 "X" means that the feature is supported.
37687 "E" means that support is provided through an external library.
37689 Network Protocols
37691 Name : Support
37692 AMQP : E
37693 file : X
37694 FTP : X
37695 Gopher : X
37696 Gophers : X
37697 HLS : X
37698 HTTP : X
37699 HTTPS : X
37700 Icecast : X
37701 MMSH : X
37702 MMST : X
37703 pipe : X
37704 Pro-MPEG FEC : X
37705 RTMP : X
37706 RTMPE : X
37707 RTMPS : X
37708 RTMPT : X
37709 RTMPTE : X
37710 RTMPTS : X
37711 RTP : X
37712 SAMBA : E
37713 SCTP : X
37714 SFTP : E
37715 TCP : X
37716 TLS : X
37717 UDP : X
37718 ZMQ : E
37719 "X" means that the protocol is supported.
37721 "E" means that support is provided through an external library.
37723 Input/Output Devices
37725 Name : Input @tab Output
37726 ALSA : X @tab X
37727 BKTR : X @tab
37728 caca : @tab X
37729 DV1394 : X @tab
37730 Lavfi virtual device : X @tab
37731 Linux framebuffer : X @tab X
37732 JACK : X @tab
37733 LIBCDIO : X
37734 LIBDC1394 : X @tab
37735 OpenAL : X
37736 OpenGL : @tab X
37737 OSS : X @tab X
37738 PulseAudio : X @tab X
37739 SDL : @tab X
37740 Video4Linux2 : X @tab X
37741 VfW capture : X @tab
37742 X11 grabbing : X @tab
37743 Win32 grabbing : X @tab
37744 "X" means that input/output is supported.
37746 Timecode
37748 Codec/format : Read @tab Write
37749 AVI : X @tab X
37750 DV : X @tab X
37751 GXF : X @tab X
37752 MOV : X @tab X
37753 MPEG1/2 : X @tab X
37754 MXF : X @tab X
37755
37757 ffprobe(1), ffmpeg(1), ffplay(1), ffmpeg-utils(1), ffmpeg-scaler(1),
37758 ffmpeg-resampler(1), ffmpeg-codecs(1), ffmpeg-bitstream-filters(1),
37759 ffmpeg-formats(1), ffmpeg-devices(1), ffmpeg-protocols(1),
37760 ffmpeg-filters(1)
37761
37763 The FFmpeg developers.
37764 For details about the authorship, see the Git history of the project
37766 (git://source.ffmpeg.org/ffmpeg), e.g. by typing the command git log in
37767 the FFmpeg source directory, or browsing the online repository at
37768 <http://source.ffmpeg.org>.
37769 Maintainers for the specific components are listed in the file
37771 MAINTAINERS in the source code tree.
37772 FFPROBE-ALL(1)