1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Options for filters with several inputs (framesync)
317 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
319 Some filters with several inputs support a common set of options.
320 These options can only be set by name, not with the short notation.
324 The action to take when EOF is encountered on the secondary input; it accepts
325 one of the following values:
329 Repeat the last frame (the default).
333 Pass the main input through.
337 If set to 1, force the output to terminate when the shortest input
338 terminates. Default value is 0.
341 If set to 1, force the filter to extend the last frame of secondary streams
342 until the end of the primary stream. A value of 0 disables this behavior.
346 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
348 @chapter Audio Filters
349 @c man begin AUDIO FILTERS
351 When you configure your FFmpeg build, you can disable any of the
352 existing filters using @code{--disable-filters}.
353 The configure output will show the audio filters included in your
356 Below is a description of the currently available audio filters.
360 A compressor is mainly used to reduce the dynamic range of a signal.
361 Especially modern music is mostly compressed at a high ratio to
362 improve the overall loudness. It's done to get the highest attention
363 of a listener, "fatten" the sound and bring more "power" to the track.
364 If a signal is compressed too much it may sound dull or "dead"
365 afterwards or it may start to "pump" (which could be a powerful effect
366 but can also destroy a track completely).
367 The right compression is the key to reach a professional sound and is
368 the high art of mixing and mastering. Because of its complex settings
369 it may take a long time to get the right feeling for this kind of effect.
371 Compression is done by detecting the volume above a chosen level
372 @code{threshold} and dividing it by the factor set with @code{ratio}.
373 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
374 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
375 the signal would cause distortion of the waveform the reduction can be
376 levelled over the time. This is done by setting "Attack" and "Release".
377 @code{attack} determines how long the signal has to rise above the threshold
378 before any reduction will occur and @code{release} sets the time the signal
379 has to fall below the threshold to reduce the reduction again. Shorter signals
380 than the chosen attack time will be left untouched.
381 The overall reduction of the signal can be made up afterwards with the
382 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
383 raising the makeup to this level results in a signal twice as loud than the
384 source. To gain a softer entry in the compression the @code{knee} flattens the
385 hard edge at the threshold in the range of the chosen decibels.
387 The filter accepts the following options:
391 Set input gain. Default is 1. Range is between 0.015625 and 64.
394 If a signal of stream rises above this level it will affect the gain
396 By default it is 0.125. Range is between 0.00097563 and 1.
399 Set a ratio by which the signal is reduced. 1:2 means that if the level
400 rose 4dB above the threshold, it will be only 2dB above after the reduction.
401 Default is 2. Range is between 1 and 20.
404 Amount of milliseconds the signal has to rise above the threshold before gain
405 reduction starts. Default is 20. Range is between 0.01 and 2000.
408 Amount of milliseconds the signal has to fall below the threshold before
409 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
412 Set the amount by how much signal will be amplified after processing.
413 Default is 1. Range is from 1 to 64.
416 Curve the sharp knee around the threshold to enter gain reduction more softly.
417 Default is 2.82843. Range is between 1 and 8.
420 Choose if the @code{average} level between all channels of input stream
421 or the louder(@code{maximum}) channel of input stream affects the
422 reduction. Default is @code{average}.
425 Should the exact signal be taken in case of @code{peak} or an RMS one in case
426 of @code{rms}. Default is @code{rms} which is mostly smoother.
429 How much to use compressed signal in output. Default is 1.
430 Range is between 0 and 1.
434 Simple audio dynamic range commpression/expansion filter.
436 The filter accepts the following options:
440 Set contrast. Default is 33. Allowed range is between 0 and 100.
445 Copy the input audio source unchanged to the output. This is mainly useful for
450 Apply cross fade from one input audio stream to another input audio stream.
451 The cross fade is applied for specified duration near the end of first stream.
453 The filter accepts the following options:
457 Specify the number of samples for which the cross fade effect has to last.
458 At the end of the cross fade effect the first input audio will be completely
459 silent. Default is 44100.
462 Specify the duration of the cross fade effect. See
463 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
464 for the accepted syntax.
465 By default the duration is determined by @var{nb_samples}.
466 If set this option is used instead of @var{nb_samples}.
469 Should first stream end overlap with second stream start. Default is enabled.
472 Set curve for cross fade transition for first stream.
475 Set curve for cross fade transition for second stream.
477 For description of available curve types see @ref{afade} filter description.
484 Cross fade from one input to another:
486 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
490 Cross fade from one input to another but without overlapping:
492 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
497 Split audio stream into several bands.
499 This filter splits audio stream into two or more frequency ranges.
500 Summing all streams back will give flat output.
502 The filter accepts the following options:
506 Set split frequencies. Those must be positive and increasing.
509 Set filter order, can be @var{2nd}, @var{4th} or @var{8th}.
510 Default is @var{4th}.
515 Reduce audio bit resolution.
517 This filter is bit crusher with enhanced functionality. A bit crusher
518 is used to audibly reduce number of bits an audio signal is sampled
519 with. This doesn't change the bit depth at all, it just produces the
520 effect. Material reduced in bit depth sounds more harsh and "digital".
521 This filter is able to even round to continuous values instead of discrete
523 Additionally it has a D/C offset which results in different crushing of
524 the lower and the upper half of the signal.
525 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
527 Another feature of this filter is the logarithmic mode.
528 This setting switches from linear distances between bits to logarithmic ones.
529 The result is a much more "natural" sounding crusher which doesn't gate low
530 signals for example. The human ear has a logarithmic perception,
531 so this kind of crushing is much more pleasant.
532 Logarithmic crushing is also able to get anti-aliased.
534 The filter accepts the following options:
550 Can be linear: @code{lin} or logarithmic: @code{log}.
559 Set sample reduction.
562 Enable LFO. By default disabled.
573 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
577 Remove impulsive noise from input audio.
579 Samples detected as impulsive noise are replaced by interpolated samples using
580 autoregressive modelling.
584 Set window size, in milliseconds. Allowed range is from @code{10} to
585 @code{100}. Default value is @code{55} milliseconds.
586 This sets size of window which will be processed at once.
589 Set window overlap, in percentage of window size. Allowed range is from
590 @code{50} to @code{95}. Default value is @code{75} percent.
591 Setting this to a very high value increases impulsive noise removal but makes
592 whole process much slower.
595 Set autoregression order, in percentage of window size. Allowed range is from
596 @code{0} to @code{25}. Default value is @code{2} percent. This option also
597 controls quality of interpolated samples using neighbour good samples.
600 Set threshold value. Allowed range is from @code{1} to @code{100}.
601 Default value is @code{2}.
602 This controls the strength of impulsive noise which is going to be removed.
603 The lower value, the more samples will be detected as impulsive noise.
606 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
607 @code{10}. Default value is @code{2}.
608 If any two samples deteced as noise are spaced less than this value then any
609 sample inbetween those two samples will be also detected as noise.
614 It accepts the following values:
617 Select overlap-add method. Even not interpolated samples are slightly
618 changed with this method.
621 Select overlap-save method. Not interpolated samples remain unchanged.
624 Default value is @code{a}.
628 Remove clipped samples from input audio.
630 Samples detected as clipped are replaced by interpolated samples using
631 autoregressive modelling.
635 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
636 Default value is @code{55} milliseconds.
637 This sets size of window which will be processed at once.
640 Set window overlap, in percentage of window size. Allowed range is from @code{50}
641 to @code{95}. Default value is @code{75} percent.
644 Set autoregression order, in percentage of window size. Allowed range is from
645 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
646 quality of interpolated samples using neighbour good samples.
649 Set threshold value. Allowed range is from @code{1} to @code{100}.
650 Default value is @code{10}. Higher values make clip detection less aggressive.
653 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
654 Default value is @code{1000}. Higher values make clip detection less aggressive.
659 It accepts the following values:
662 Select overlap-add method. Even not interpolated samples are slightly changed
666 Select overlap-save method. Not interpolated samples remain unchanged.
669 Default value is @code{a}.
674 Delay one or more audio channels.
676 Samples in delayed channel are filled with silence.
678 The filter accepts the following option:
682 Set list of delays in milliseconds for each channel separated by '|'.
683 Unused delays will be silently ignored. If number of given delays is
684 smaller than number of channels all remaining channels will not be delayed.
685 If you want to delay exact number of samples, append 'S' to number.
692 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
693 the second channel (and any other channels that may be present) unchanged.
699 Delay second channel by 500 samples, the third channel by 700 samples and leave
700 the first channel (and any other channels that may be present) unchanged.
706 @section aderivative, aintegral
708 Compute derivative/integral of audio stream.
710 Applying both filters one after another produces original audio.
714 Apply echoing to the input audio.
716 Echoes are reflected sound and can occur naturally amongst mountains
717 (and sometimes large buildings) when talking or shouting; digital echo
718 effects emulate this behaviour and are often used to help fill out the
719 sound of a single instrument or vocal. The time difference between the
720 original signal and the reflection is the @code{delay}, and the
721 loudness of the reflected signal is the @code{decay}.
722 Multiple echoes can have different delays and decays.
724 A description of the accepted parameters follows.
728 Set input gain of reflected signal. Default is @code{0.6}.
731 Set output gain of reflected signal. Default is @code{0.3}.
734 Set list of time intervals in milliseconds between original signal and reflections
735 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
736 Default is @code{1000}.
739 Set list of loudness of reflected signals separated by '|'.
740 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
741 Default is @code{0.5}.
748 Make it sound as if there are twice as many instruments as are actually playing:
750 aecho=0.8:0.88:60:0.4
754 If delay is very short, then it sound like a (metallic) robot playing music:
760 A longer delay will sound like an open air concert in the mountains:
762 aecho=0.8:0.9:1000:0.3
766 Same as above but with one more mountain:
768 aecho=0.8:0.9:1000|1800:0.3|0.25
773 Audio emphasis filter creates or restores material directly taken from LPs or
774 emphased CDs with different filter curves. E.g. to store music on vinyl the
775 signal has to be altered by a filter first to even out the disadvantages of
776 this recording medium.
777 Once the material is played back the inverse filter has to be applied to
778 restore the distortion of the frequency response.
780 The filter accepts the following options:
790 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
791 use @code{production} mode. Default is @code{reproduction} mode.
794 Set filter type. Selects medium. Can be one of the following:
806 select Compact Disc (CD).
812 select 50µs (FM-KF).
814 select 75µs (FM-KF).
820 Modify an audio signal according to the specified expressions.
822 This filter accepts one or more expressions (one for each channel),
823 which are evaluated and used to modify a corresponding audio signal.
825 It accepts the following parameters:
829 Set the '|'-separated expressions list for each separate channel. If
830 the number of input channels is greater than the number of
831 expressions, the last specified expression is used for the remaining
834 @item channel_layout, c
835 Set output channel layout. If not specified, the channel layout is
836 specified by the number of expressions. If set to @samp{same}, it will
837 use by default the same input channel layout.
840 Each expression in @var{exprs} can contain the following constants and functions:
844 channel number of the current expression
847 number of the evaluated sample, starting from 0
853 time of the evaluated sample expressed in seconds
856 @item nb_out_channels
857 input and output number of channels
860 the value of input channel with number @var{CH}
863 Note: this filter is slow. For faster processing you should use a
872 aeval=val(ch)/2:c=same
876 Invert phase of the second channel:
885 Apply fade-in/out effect to input audio.
887 A description of the accepted parameters follows.
891 Specify the effect type, can be either @code{in} for fade-in, or
892 @code{out} for a fade-out effect. Default is @code{in}.
894 @item start_sample, ss
895 Specify the number of the start sample for starting to apply the fade
896 effect. Default is 0.
899 Specify the number of samples for which the fade effect has to last. At
900 the end of the fade-in effect the output audio will have the same
901 volume as the input audio, at the end of the fade-out transition
902 the output audio will be silence. Default is 44100.
905 Specify the start time of the fade effect. Default is 0.
906 The value must be specified as a time duration; see
907 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
908 for the accepted syntax.
909 If set this option is used instead of @var{start_sample}.
912 Specify the duration of the fade effect. See
913 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
914 for the accepted syntax.
915 At the end of the fade-in effect the output audio will have the same
916 volume as the input audio, at the end of the fade-out transition
917 the output audio will be silence.
918 By default the duration is determined by @var{nb_samples}.
919 If set this option is used instead of @var{nb_samples}.
922 Set curve for fade transition.
924 It accepts the following values:
927 select triangular, linear slope (default)
929 select quarter of sine wave
931 select half of sine wave
933 select exponential sine wave
937 select inverted parabola
951 select inverted quarter of sine wave
953 select inverted half of sine wave
955 select double-exponential seat
957 select double-exponential sigmoid
959 select logistic sigmoid
967 Fade in first 15 seconds of audio:
973 Fade out last 25 seconds of a 900 seconds audio:
975 afade=t=out:st=875:d=25
980 Denoise audio samples with FFT.
982 A description of the accepted parameters follows.
986 Set the noise reduction in dB, allowed range is 0.01 to 97.
987 Default value is 12 dB.
990 Set the noise floor in dB, allowed range is -80 to -20.
991 Default value is -50 dB.
996 It accepts the following values:
1005 Select shellac noise.
1008 Select custom noise, defined in @code{bn} option.
1010 Default value is white noise.
1014 Set custom band noise for every one of 15 bands.
1015 Bands are separated by ' ' or '|'.
1018 Set the residual floor in dB, allowed range is -80 to -20.
1019 Default value is -38 dB.
1022 Enable noise tracking. By default is disabled.
1023 With this enabled, noise floor is automatically adjusted.
1026 Enable residual tracking. By default is disabled.
1029 Set the output mode.
1031 It accepts the following values:
1034 Pass input unchanged.
1037 Pass noise filtered out.
1042 Default value is @var{o}.
1046 @subsection Commands
1048 This filter supports the following commands:
1050 @item sample_noise, sn
1051 Start or stop measuring noise profile.
1052 Syntax for the command is : "start" or "stop" string.
1053 After measuring noise profile is stopped it will be
1054 automatically applied in filtering.
1056 @item noise_reduction, nr
1057 Change noise reduction. Argument is single float number.
1058 Syntax for the command is : "@var{noise_reduction}"
1060 @item noise_floor, nf
1061 Change noise floor. Argument is single float number.
1062 Syntax for the command is : "@var{noise_floor}"
1064 @item output_mode, om
1065 Change output mode operation.
1066 Syntax for the command is : "i", "o" or "n" string.
1070 Apply arbitrary expressions to samples in frequency domain.
1074 Set frequency domain real expression for each separate channel separated
1075 by '|'. Default is "1".
1076 If the number of input channels is greater than the number of
1077 expressions, the last specified expression is used for the remaining
1081 Set frequency domain imaginary expression for each separate channel
1082 separated by '|'. If not set, @var{real} option is used.
1084 Each expression in @var{real} and @var{imag} can contain the following
1092 current frequency bin number
1095 number of available bins
1098 channel number of the current expression
1110 It accepts the following values:
1126 Default is @code{w4096}
1129 Set window function. Default is @code{hann}.
1132 Set window overlap. If set to 1, the recommended overlap for selected
1133 window function will be picked. Default is @code{0.75}.
1136 @subsection Examples
1140 Leave almost only low frequencies in audio:
1142 afftfilt="1-clip((b/nb)*b,0,1)"
1149 Apply an arbitrary Frequency Impulse Response filter.
1151 This filter is designed for applying long FIR filters,
1152 up to 60 seconds long.
1154 It can be used as component for digital crossover filters,
1155 room equalization, cross talk cancellation, wavefield synthesis,
1156 auralization, ambiophonics and ambisonics.
1158 This filter uses second stream as FIR coefficients.
1159 If second stream holds single channel, it will be used
1160 for all input channels in first stream, otherwise
1161 number of channels in second stream must be same as
1162 number of channels in first stream.
1164 It accepts the following parameters:
1168 Set dry gain. This sets input gain.
1171 Set wet gain. This sets final output gain.
1174 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1177 Enable applying gain measured from power of IR. For approach to use for measuring power
1178 of IR see next option.
1181 Set which approach to use for auto gain measurement.
1185 select peak gain, very conservative approach. This is default value.
1188 select DC gain, limited application.
1191 select gain to noise approach, this is most popular one.
1195 Set gain to be applied to IR coefficients before filtering.
1196 Allowed range is 0 to 1. This can be set even with @var{again} used.
1199 Set format of IR stream. Can be @code{mono} or @code{input}.
1200 Default is @code{input}.
1203 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1204 Allowed range is 0.1 to 60 seconds.
1207 Show IR frequency reponse, magnitude and phase in additional video stream.
1208 By default it is disabled.
1211 Set for which IR channel to display frequency response. By default is first channel
1212 displayed. This option is used only when @var{response} is enabled.
1215 Set video stream size. This option is used only when @var{response} is enabled.
1218 @subsection Examples
1222 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1224 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1231 Set output format constraints for the input audio. The framework will
1232 negotiate the most appropriate format to minimize conversions.
1234 It accepts the following parameters:
1238 A '|'-separated list of requested sample formats.
1241 A '|'-separated list of requested sample rates.
1243 @item channel_layouts
1244 A '|'-separated list of requested channel layouts.
1246 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1247 for the required syntax.
1250 If a parameter is omitted, all values are allowed.
1252 Force the output to either unsigned 8-bit or signed 16-bit stereo
1254 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1259 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1260 processing reduces disturbing noise between useful signals.
1262 Gating is done by detecting the volume below a chosen level @var{threshold}
1263 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1264 floor is set via @var{range}. Because an exact manipulation of the signal
1265 would cause distortion of the waveform the reduction can be levelled over
1266 time. This is done by setting @var{attack} and @var{release}.
1268 @var{attack} determines how long the signal has to fall below the threshold
1269 before any reduction will occur and @var{release} sets the time the signal
1270 has to rise above the threshold to reduce the reduction again.
1271 Shorter signals than the chosen attack time will be left untouched.
1275 Set input level before filtering.
1276 Default is 1. Allowed range is from 0.015625 to 64.
1279 Set the level of gain reduction when the signal is below the threshold.
1280 Default is 0.06125. Allowed range is from 0 to 1.
1283 If a signal rises above this level the gain reduction is released.
1284 Default is 0.125. Allowed range is from 0 to 1.
1287 Set a ratio by which the signal is reduced.
1288 Default is 2. Allowed range is from 1 to 9000.
1291 Amount of milliseconds the signal has to rise above the threshold before gain
1293 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1296 Amount of milliseconds the signal has to fall below the threshold before the
1297 reduction is increased again. Default is 250 milliseconds.
1298 Allowed range is from 0.01 to 9000.
1301 Set amount of amplification of signal after processing.
1302 Default is 1. Allowed range is from 1 to 64.
1305 Curve the sharp knee around the threshold to enter gain reduction more softly.
1306 Default is 2.828427125. Allowed range is from 1 to 8.
1309 Choose if exact signal should be taken for detection or an RMS like one.
1310 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1313 Choose if the average level between all channels or the louder channel affects
1315 Default is @code{average}. Can be @code{average} or @code{maximum}.
1320 Apply an arbitrary Infinite Impulse Response filter.
1322 It accepts the following parameters:
1326 Set numerator/zeros coefficients.
1329 Set denominator/poles coefficients.
1341 Set coefficients format.
1347 Z-plane zeros/poles, cartesian (default)
1349 Z-plane zeros/poles, polar radians
1351 Z-plane zeros/poles, polar degrees
1355 Set kind of processing.
1356 Can be @code{d} - direct or @code{s} - serial cascading. Defauls is @code{s}.
1359 Set filtering precision.
1363 double-precision floating-point (default)
1365 single-precision floating-point
1373 Show IR frequency reponse, magnitude and phase in additional video stream.
1374 By default it is disabled.
1377 Set for which IR channel to display frequency response. By default is first channel
1378 displayed. This option is used only when @var{response} is enabled.
1381 Set video stream size. This option is used only when @var{response} is enabled.
1384 Coefficients in @code{tf} format are separated by spaces and are in ascending
1387 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1388 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1391 Different coefficients and gains can be provided for every channel, in such case
1392 use '|' to separate coefficients or gains. Last provided coefficients will be
1393 used for all remaining channels.
1395 @subsection Examples
1399 Apply 2 pole elliptic notch at arround 5000Hz for 48000 Hz sample rate:
1401 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
1405 Same as above but in @code{zp} format:
1407 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
1413 The limiter prevents an input signal from rising over a desired threshold.
1414 This limiter uses lookahead technology to prevent your signal from distorting.
1415 It means that there is a small delay after the signal is processed. Keep in mind
1416 that the delay it produces is the attack time you set.
1418 The filter accepts the following options:
1422 Set input gain. Default is 1.
1425 Set output gain. Default is 1.
1428 Don't let signals above this level pass the limiter. Default is 1.
1431 The limiter will reach its attenuation level in this amount of time in
1432 milliseconds. Default is 5 milliseconds.
1435 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1436 Default is 50 milliseconds.
1439 When gain reduction is always needed ASC takes care of releasing to an
1440 average reduction level rather than reaching a reduction of 0 in the release
1444 Select how much the release time is affected by ASC, 0 means nearly no changes
1445 in release time while 1 produces higher release times.
1448 Auto level output signal. Default is enabled.
1449 This normalizes audio back to 0dB if enabled.
1452 Depending on picked setting it is recommended to upsample input 2x or 4x times
1453 with @ref{aresample} before applying this filter.
1457 Apply a two-pole all-pass filter with central frequency (in Hz)
1458 @var{frequency}, and filter-width @var{width}.
1459 An all-pass filter changes the audio's frequency to phase relationship
1460 without changing its frequency to amplitude relationship.
1462 The filter accepts the following options:
1466 Set frequency in Hz.
1469 Set method to specify band-width of filter.
1484 Specify the band-width of a filter in width_type units.
1487 Specify which channels to filter, by default all available are filtered.
1490 @subsection Commands
1492 This filter supports the following commands:
1495 Change allpass frequency.
1496 Syntax for the command is : "@var{frequency}"
1499 Change allpass width_type.
1500 Syntax for the command is : "@var{width_type}"
1503 Change allpass width.
1504 Syntax for the command is : "@var{width}"
1511 The filter accepts the following options:
1515 Set the number of loops. Setting this value to -1 will result in infinite loops.
1519 Set maximal number of samples. Default is 0.
1522 Set first sample of loop. Default is 0.
1528 Merge two or more audio streams into a single multi-channel stream.
1530 The filter accepts the following options:
1535 Set the number of inputs. Default is 2.
1539 If the channel layouts of the inputs are disjoint, and therefore compatible,
1540 the channel layout of the output will be set accordingly and the channels
1541 will be reordered as necessary. If the channel layouts of the inputs are not
1542 disjoint, the output will have all the channels of the first input then all
1543 the channels of the second input, in that order, and the channel layout of
1544 the output will be the default value corresponding to the total number of
1547 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1548 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1549 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1550 first input, b1 is the first channel of the second input).
1552 On the other hand, if both input are in stereo, the output channels will be
1553 in the default order: a1, a2, b1, b2, and the channel layout will be
1554 arbitrarily set to 4.0, which may or may not be the expected value.
1556 All inputs must have the same sample rate, and format.
1558 If inputs do not have the same duration, the output will stop with the
1561 @subsection Examples
1565 Merge two mono files into a stereo stream:
1567 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1571 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1573 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
1579 Mixes multiple audio inputs into a single output.
1581 Note that this filter only supports float samples (the @var{amerge}
1582 and @var{pan} audio filters support many formats). If the @var{amix}
1583 input has integer samples then @ref{aresample} will be automatically
1584 inserted to perform the conversion to float samples.
1588 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1590 will mix 3 input audio streams to a single output with the same duration as the
1591 first input and a dropout transition time of 3 seconds.
1593 It accepts the following parameters:
1597 The number of inputs. If unspecified, it defaults to 2.
1600 How to determine the end-of-stream.
1604 The duration of the longest input. (default)
1607 The duration of the shortest input.
1610 The duration of the first input.
1614 @item dropout_transition
1615 The transition time, in seconds, for volume renormalization when an input
1616 stream ends. The default value is 2 seconds.
1619 Specify weight of each input audio stream as sequence.
1620 Each weight is separated by space. By default all inputs have same weight.
1625 Multiply first audio stream with second audio stream and store result
1626 in output audio stream. Multiplication is done by multiplying each
1627 sample from first stream with sample at same position from second stream.
1629 With this element-wise multiplication one can create amplitude fades and
1630 amplitude modulations.
1632 @section anequalizer
1634 High-order parametric multiband equalizer for each channel.
1636 It accepts the following parameters:
1640 This option string is in format:
1641 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1642 Each equalizer band is separated by '|'.
1646 Set channel number to which equalization will be applied.
1647 If input doesn't have that channel the entry is ignored.
1650 Set central frequency for band.
1651 If input doesn't have that frequency the entry is ignored.
1654 Set band width in hertz.
1657 Set band gain in dB.
1660 Set filter type for band, optional, can be:
1664 Butterworth, this is default.
1675 With this option activated frequency response of anequalizer is displayed
1679 Set video stream size. Only useful if curves option is activated.
1682 Set max gain that will be displayed. Only useful if curves option is activated.
1683 Setting this to a reasonable value makes it possible to display gain which is derived from
1684 neighbour bands which are too close to each other and thus produce higher gain
1685 when both are activated.
1688 Set frequency scale used to draw frequency response in video output.
1689 Can be linear or logarithmic. Default is logarithmic.
1692 Set color for each channel curve which is going to be displayed in video stream.
1693 This is list of color names separated by space or by '|'.
1694 Unrecognised or missing colors will be replaced by white color.
1697 @subsection Examples
1701 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1702 for first 2 channels using Chebyshev type 1 filter:
1704 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1708 @subsection Commands
1710 This filter supports the following commands:
1713 Alter existing filter parameters.
1714 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1716 @var{fN} is existing filter number, starting from 0, if no such filter is available
1718 @var{freq} set new frequency parameter.
1719 @var{width} set new width parameter in herz.
1720 @var{gain} set new gain parameter in dB.
1722 Full filter invocation with asendcmd may look like this:
1723 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1728 Pass the audio source unchanged to the output.
1732 Pad the end of an audio stream with silence.
1734 This can be used together with @command{ffmpeg} @option{-shortest} to
1735 extend audio streams to the same length as the video stream.
1737 A description of the accepted options follows.
1741 Set silence packet size. Default value is 4096.
1744 Set the number of samples of silence to add to the end. After the
1745 value is reached, the stream is terminated. This option is mutually
1746 exclusive with @option{whole_len}.
1749 Set the minimum total number of samples in the output audio stream. If
1750 the value is longer than the input audio length, silence is added to
1751 the end, until the value is reached. This option is mutually exclusive
1752 with @option{pad_len}.
1755 If neither the @option{pad_len} nor the @option{whole_len} option is
1756 set, the filter will add silence to the end of the input stream
1759 @subsection Examples
1763 Add 1024 samples of silence to the end of the input:
1769 Make sure the audio output will contain at least 10000 samples, pad
1770 the input with silence if required:
1772 apad=whole_len=10000
1776 Use @command{ffmpeg} to pad the audio input with silence, so that the
1777 video stream will always result the shortest and will be converted
1778 until the end in the output file when using the @option{shortest}
1781 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
1786 Add a phasing effect to the input audio.
1788 A phaser filter creates series of peaks and troughs in the frequency spectrum.
1789 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
1791 A description of the accepted parameters follows.
1795 Set input gain. Default is 0.4.
1798 Set output gain. Default is 0.74
1801 Set delay in milliseconds. Default is 3.0.
1804 Set decay. Default is 0.4.
1807 Set modulation speed in Hz. Default is 0.5.
1810 Set modulation type. Default is triangular.
1812 It accepts the following values:
1821 Audio pulsator is something between an autopanner and a tremolo.
1822 But it can produce funny stereo effects as well. Pulsator changes the volume
1823 of the left and right channel based on a LFO (low frequency oscillator) with
1824 different waveforms and shifted phases.
1825 This filter have the ability to define an offset between left and right
1826 channel. An offset of 0 means that both LFO shapes match each other.
1827 The left and right channel are altered equally - a conventional tremolo.
1828 An offset of 50% means that the shape of the right channel is exactly shifted
1829 in phase (or moved backwards about half of the frequency) - pulsator acts as
1830 an autopanner. At 1 both curves match again. Every setting in between moves the
1831 phase shift gapless between all stages and produces some "bypassing" sounds with
1832 sine and triangle waveforms. The more you set the offset near 1 (starting from
1833 the 0.5) the faster the signal passes from the left to the right speaker.
1835 The filter accepts the following options:
1839 Set input gain. By default it is 1. Range is [0.015625 - 64].
1842 Set output gain. By default it is 1. Range is [0.015625 - 64].
1845 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
1846 sawup or sawdown. Default is sine.
1849 Set modulation. Define how much of original signal is affected by the LFO.
1852 Set left channel offset. Default is 0. Allowed range is [0 - 1].
1855 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
1858 Set pulse width. Default is 1. Allowed range is [0 - 2].
1861 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
1864 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
1868 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
1872 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
1873 if timing is set to hz.
1879 Resample the input audio to the specified parameters, using the
1880 libswresample library. If none are specified then the filter will
1881 automatically convert between its input and output.
1883 This filter is also able to stretch/squeeze the audio data to make it match
1884 the timestamps or to inject silence / cut out audio to make it match the
1885 timestamps, do a combination of both or do neither.
1887 The filter accepts the syntax
1888 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
1889 expresses a sample rate and @var{resampler_options} is a list of
1890 @var{key}=@var{value} pairs, separated by ":". See the
1891 @ref{Resampler Options,,"Resampler Options" section in the
1892 ffmpeg-resampler(1) manual,ffmpeg-resampler}
1893 for the complete list of supported options.
1895 @subsection Examples
1899 Resample the input audio to 44100Hz:
1905 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
1906 samples per second compensation:
1908 aresample=async=1000
1914 Reverse an audio clip.
1916 Warning: This filter requires memory to buffer the entire clip, so trimming
1919 @subsection Examples
1923 Take the first 5 seconds of a clip, and reverse it.
1925 atrim=end=5,areverse
1929 @section asetnsamples
1931 Set the number of samples per each output audio frame.
1933 The last output packet may contain a different number of samples, as
1934 the filter will flush all the remaining samples when the input audio
1937 The filter accepts the following options:
1941 @item nb_out_samples, n
1942 Set the number of frames per each output audio frame. The number is
1943 intended as the number of samples @emph{per each channel}.
1944 Default value is 1024.
1947 If set to 1, the filter will pad the last audio frame with zeroes, so
1948 that the last frame will contain the same number of samples as the
1949 previous ones. Default value is 1.
1952 For example, to set the number of per-frame samples to 1234 and
1953 disable padding for the last frame, use:
1955 asetnsamples=n=1234:p=0
1960 Set the sample rate without altering the PCM data.
1961 This will result in a change of speed and pitch.
1963 The filter accepts the following options:
1966 @item sample_rate, r
1967 Set the output sample rate. Default is 44100 Hz.
1972 Show a line containing various information for each input audio frame.
1973 The input audio is not modified.
1975 The shown line contains a sequence of key/value pairs of the form
1976 @var{key}:@var{value}.
1978 The following values are shown in the output:
1982 The (sequential) number of the input frame, starting from 0.
1985 The presentation timestamp of the input frame, in time base units; the time base
1986 depends on the filter input pad, and is usually 1/@var{sample_rate}.
1989 The presentation timestamp of the input frame in seconds.
1992 position of the frame in the input stream, -1 if this information in
1993 unavailable and/or meaningless (for example in case of synthetic audio)
2002 The sample rate for the audio frame.
2005 The number of samples (per channel) in the frame.
2008 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2009 audio, the data is treated as if all the planes were concatenated.
2011 @item plane_checksums
2012 A list of Adler-32 checksums for each data plane.
2018 Display time domain statistical information about the audio channels.
2019 Statistics are calculated and displayed for each audio channel and,
2020 where applicable, an overall figure is also given.
2022 It accepts the following option:
2025 Short window length in seconds, used for peak and trough RMS measurement.
2026 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2030 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2031 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2034 Available keys for each channel are:
2070 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2071 this @code{lavfi.astats.Overall.Peak_count}.
2073 For description what each key means read below.
2076 Set number of frame after which stats are going to be recalculated.
2077 Default is disabled.
2080 A description of each shown parameter follows:
2084 Mean amplitude displacement from zero.
2087 Minimal sample level.
2090 Maximal sample level.
2092 @item Min difference
2093 Minimal difference between two consecutive samples.
2095 @item Max difference
2096 Maximal difference between two consecutive samples.
2098 @item Mean difference
2099 Mean difference between two consecutive samples.
2100 The average of each difference between two consecutive samples.
2102 @item RMS difference
2103 Root Mean Square difference between two consecutive samples.
2107 Standard peak and RMS level measured in dBFS.
2111 Peak and trough values for RMS level measured over a short window.
2114 Standard ratio of peak to RMS level (note: not in dB).
2117 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2118 (i.e. either @var{Min level} or @var{Max level}).
2121 Number of occasions (not the number of samples) that the signal attained either
2122 @var{Min level} or @var{Max level}.
2125 Overall bit depth of audio. Number of bits used for each sample.
2128 Measured dynamic range of audio in dB.
2130 @item Zero crossings
2131 Number of points where the waveform crosses the zero level axis.
2133 @item Zero crossings rate
2134 Rate of Zero crossings and number of audio samples.
2141 The filter accepts exactly one parameter, the audio tempo. If not
2142 specified then the filter will assume nominal 1.0 tempo. Tempo must
2143 be in the [0.5, 100.0] range.
2145 Note that tempo greater than 2 will skip some samples rather than
2146 blend them in. If for any reason this is a concern it is always
2147 possible to daisy-chain several instances of atempo to achieve the
2148 desired product tempo.
2150 @subsection Examples
2154 Slow down audio to 80% tempo:
2160 To speed up audio to 300% tempo:
2166 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2168 atempo=sqrt(3),atempo=sqrt(3)
2174 Trim the input so that the output contains one continuous subpart of the input.
2176 It accepts the following parameters:
2179 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2180 sample with the timestamp @var{start} will be the first sample in the output.
2183 Specify time of the first audio sample that will be dropped, i.e. the
2184 audio sample immediately preceding the one with the timestamp @var{end} will be
2185 the last sample in the output.
2188 Same as @var{start}, except this option sets the start timestamp in samples
2192 Same as @var{end}, except this option sets the end timestamp in samples instead
2196 The maximum duration of the output in seconds.
2199 The number of the first sample that should be output.
2202 The number of the first sample that should be dropped.
2205 @option{start}, @option{end}, and @option{duration} are expressed as time
2206 duration specifications; see
2207 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2209 Note that the first two sets of the start/end options and the @option{duration}
2210 option look at the frame timestamp, while the _sample options simply count the
2211 samples that pass through the filter. So start/end_pts and start/end_sample will
2212 give different results when the timestamps are wrong, inexact or do not start at
2213 zero. Also note that this filter does not modify the timestamps. If you wish
2214 to have the output timestamps start at zero, insert the asetpts filter after the
2217 If multiple start or end options are set, this filter tries to be greedy and
2218 keep all samples that match at least one of the specified constraints. To keep
2219 only the part that matches all the constraints at once, chain multiple atrim
2222 The defaults are such that all the input is kept. So it is possible to set e.g.
2223 just the end values to keep everything before the specified time.
2228 Drop everything except the second minute of input:
2230 ffmpeg -i INPUT -af atrim=60:120
2234 Keep only the first 1000 samples:
2236 ffmpeg -i INPUT -af atrim=end_sample=1000
2243 Apply a two-pole Butterworth band-pass filter with central
2244 frequency @var{frequency}, and (3dB-point) band-width width.
2245 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2246 instead of the default: constant 0dB peak gain.
2247 The filter roll off at 6dB per octave (20dB per decade).
2249 The filter accepts the following options:
2253 Set the filter's central frequency. Default is @code{3000}.
2256 Constant skirt gain if set to 1. Defaults to 0.
2259 Set method to specify band-width of filter.
2274 Specify the band-width of a filter in width_type units.
2277 Specify which channels to filter, by default all available are filtered.
2280 @subsection Commands
2282 This filter supports the following commands:
2285 Change bandpass frequency.
2286 Syntax for the command is : "@var{frequency}"
2289 Change bandpass width_type.
2290 Syntax for the command is : "@var{width_type}"
2293 Change bandpass width.
2294 Syntax for the command is : "@var{width}"
2299 Apply a two-pole Butterworth band-reject filter with central
2300 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2301 The filter roll off at 6dB per octave (20dB per decade).
2303 The filter accepts the following options:
2307 Set the filter's central frequency. Default is @code{3000}.
2310 Set method to specify band-width of filter.
2325 Specify the band-width of a filter in width_type units.
2328 Specify which channels to filter, by default all available are filtered.
2331 @subsection Commands
2333 This filter supports the following commands:
2336 Change bandreject frequency.
2337 Syntax for the command is : "@var{frequency}"
2340 Change bandreject width_type.
2341 Syntax for the command is : "@var{width_type}"
2344 Change bandreject width.
2345 Syntax for the command is : "@var{width}"
2348 @section bass, lowshelf
2350 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2351 shelving filter with a response similar to that of a standard
2352 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2354 The filter accepts the following options:
2358 Give the gain at 0 Hz. Its useful range is about -20
2359 (for a large cut) to +20 (for a large boost).
2360 Beware of clipping when using a positive gain.
2363 Set the filter's central frequency and so can be used
2364 to extend or reduce the frequency range to be boosted or cut.
2365 The default value is @code{100} Hz.
2368 Set method to specify band-width of filter.
2383 Determine how steep is the filter's shelf transition.
2386 Specify which channels to filter, by default all available are filtered.
2389 @subsection Commands
2391 This filter supports the following commands:
2394 Change bass frequency.
2395 Syntax for the command is : "@var{frequency}"
2398 Change bass width_type.
2399 Syntax for the command is : "@var{width_type}"
2403 Syntax for the command is : "@var{width}"
2407 Syntax for the command is : "@var{gain}"
2412 Apply a biquad IIR filter with the given coefficients.
2413 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2414 are the numerator and denominator coefficients respectively.
2415 and @var{channels}, @var{c} specify which channels to filter, by default all
2416 available are filtered.
2418 @subsection Commands
2420 This filter supports the following commands:
2428 Change biquad parameter.
2429 Syntax for the command is : "@var{value}"
2433 Bauer stereo to binaural transformation, which improves headphone listening of
2434 stereo audio records.
2436 To enable compilation of this filter you need to configure FFmpeg with
2437 @code{--enable-libbs2b}.
2439 It accepts the following parameters:
2443 Pre-defined crossfeed level.
2447 Default level (fcut=700, feed=50).
2450 Chu Moy circuit (fcut=700, feed=60).
2453 Jan Meier circuit (fcut=650, feed=95).
2458 Cut frequency (in Hz).
2467 Remap input channels to new locations.
2469 It accepts the following parameters:
2472 Map channels from input to output. The argument is a '|'-separated list of
2473 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2474 @var{in_channel} form. @var{in_channel} can be either the name of the input
2475 channel (e.g. FL for front left) or its index in the input channel layout.
2476 @var{out_channel} is the name of the output channel or its index in the output
2477 channel layout. If @var{out_channel} is not given then it is implicitly an
2478 index, starting with zero and increasing by one for each mapping.
2480 @item channel_layout
2481 The channel layout of the output stream.
2484 If no mapping is present, the filter will implicitly map input channels to
2485 output channels, preserving indices.
2487 @subsection Examples
2491 For example, assuming a 5.1+downmix input MOV file,
2493 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2495 will create an output WAV file tagged as stereo from the downmix channels of
2499 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2501 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
2505 @section channelsplit
2507 Split each channel from an input audio stream into a separate output stream.
2509 It accepts the following parameters:
2511 @item channel_layout
2512 The channel layout of the input stream. The default is "stereo".
2514 A channel layout describing the channels to be extracted as separate output streams
2515 or "all" to extract each input channel as a separate stream. The default is "all".
2517 Choosing channels not present in channel layout in the input will result in an error.
2520 @subsection Examples
2524 For example, assuming a stereo input MP3 file,
2526 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
2528 will create an output Matroska file with two audio streams, one containing only
2529 the left channel and the other the right channel.
2532 Split a 5.1 WAV file into per-channel files:
2534 ffmpeg -i in.wav -filter_complex
2535 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
2536 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
2537 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
2542 Extract only LFE from a 5.1 WAV file:
2544 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
2545 -map '[LFE]' lfe.wav
2550 Add a chorus effect to the audio.
2552 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
2554 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
2555 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
2556 The modulation depth defines the range the modulated delay is played before or after
2557 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
2558 sound tuned around the original one, like in a chorus where some vocals are slightly
2561 It accepts the following parameters:
2564 Set input gain. Default is 0.4.
2567 Set output gain. Default is 0.4.
2570 Set delays. A typical delay is around 40ms to 60ms.
2582 @subsection Examples
2588 chorus=0.7:0.9:55:0.4:0.25:2
2594 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
2598 Fuller sounding chorus with three delays:
2600 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
2605 Compress or expand the audio's dynamic range.
2607 It accepts the following parameters:
2613 A list of times in seconds for each channel over which the instantaneous level
2614 of the input signal is averaged to determine its volume. @var{attacks} refers to
2615 increase of volume and @var{decays} refers to decrease of volume. For most
2616 situations, the attack time (response to the audio getting louder) should be
2617 shorter than the decay time, because the human ear is more sensitive to sudden
2618 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
2619 a typical value for decay is 0.8 seconds.
2620 If specified number of attacks & decays is lower than number of channels, the last
2621 set attack/decay will be used for all remaining channels.
2624 A list of points for the transfer function, specified in dB relative to the
2625 maximum possible signal amplitude. Each key points list must be defined using
2626 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
2627 @code{x0/y0 x1/y1 x2/y2 ....}
2629 The input values must be in strictly increasing order but the transfer function
2630 does not have to be monotonically rising. The point @code{0/0} is assumed but
2631 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
2632 function are @code{-70/-70|-60/-20|1/0}.
2635 Set the curve radius in dB for all joints. It defaults to 0.01.
2638 Set the additional gain in dB to be applied at all points on the transfer
2639 function. This allows for easy adjustment of the overall gain.
2643 Set an initial volume, in dB, to be assumed for each channel when filtering
2644 starts. This permits the user to supply a nominal level initially, so that, for
2645 example, a very large gain is not applied to initial signal levels before the
2646 companding has begun to operate. A typical value for audio which is initially
2647 quiet is -90 dB. It defaults to 0.
2650 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
2651 delayed before being fed to the volume adjuster. Specifying a delay
2652 approximately equal to the attack/decay times allows the filter to effectively
2653 operate in predictive rather than reactive mode. It defaults to 0.
2657 @subsection Examples
2661 Make music with both quiet and loud passages suitable for listening to in a
2664 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
2667 Another example for audio with whisper and explosion parts:
2669 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
2673 A noise gate for when the noise is at a lower level than the signal:
2675 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
2679 Here is another noise gate, this time for when the noise is at a higher level
2680 than the signal (making it, in some ways, similar to squelch):
2682 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
2686 2:1 compression starting at -6dB:
2688 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
2692 2:1 compression starting at -9dB:
2694 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
2698 2:1 compression starting at -12dB:
2700 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
2704 2:1 compression starting at -18dB:
2706 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
2710 3:1 compression starting at -15dB:
2712 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
2718 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
2724 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
2728 Hard limiter at -6dB:
2730 compand=attacks=0:points=-80/-80|-6/-6|20/-6
2734 Hard limiter at -12dB:
2736 compand=attacks=0:points=-80/-80|-12/-12|20/-12
2740 Hard noise gate at -35 dB:
2742 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
2748 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
2752 @section compensationdelay
2754 Compensation Delay Line is a metric based delay to compensate differing
2755 positions of microphones or speakers.
2757 For example, you have recorded guitar with two microphones placed in
2758 different location. Because the front of sound wave has fixed speed in
2759 normal conditions, the phasing of microphones can vary and depends on
2760 their location and interposition. The best sound mix can be achieved when
2761 these microphones are in phase (synchronized). Note that distance of
2762 ~30 cm between microphones makes one microphone to capture signal in
2763 antiphase to another microphone. That makes the final mix sounding moody.
2764 This filter helps to solve phasing problems by adding different delays
2765 to each microphone track and make them synchronized.
2767 The best result can be reached when you take one track as base and
2768 synchronize other tracks one by one with it.
2769 Remember that synchronization/delay tolerance depends on sample rate, too.
2770 Higher sample rates will give more tolerance.
2772 It accepts the following parameters:
2776 Set millimeters distance. This is compensation distance for fine tuning.
2780 Set cm distance. This is compensation distance for tightening distance setup.
2784 Set meters distance. This is compensation distance for hard distance setup.
2788 Set dry amount. Amount of unprocessed (dry) signal.
2792 Set wet amount. Amount of processed (wet) signal.
2796 Set temperature degree in Celsius. This is the temperature of the environment.
2801 Apply headphone crossfeed filter.
2803 Crossfeed is the process of blending the left and right channels of stereo
2805 It is mainly used to reduce extreme stereo separation of low frequencies.
2807 The intent is to produce more speaker like sound to the listener.
2809 The filter accepts the following options:
2813 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
2814 This sets gain of low shelf filter for side part of stereo image.
2815 Default is -6dB. Max allowed is -30db when strength is set to 1.
2818 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
2819 This sets cut off frequency of low shelf filter. Default is cut off near
2820 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
2823 Set input gain. Default is 0.9.
2826 Set output gain. Default is 1.
2829 @section crystalizer
2830 Simple algorithm to expand audio dynamic range.
2832 The filter accepts the following options:
2836 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
2837 (unchanged sound) to 10.0 (maximum effect).
2840 Enable clipping. By default is enabled.
2844 Apply a DC shift to the audio.
2846 This can be useful to remove a DC offset (caused perhaps by a hardware problem
2847 in the recording chain) from the audio. The effect of a DC offset is reduced
2848 headroom and hence volume. The @ref{astats} filter can be used to determine if
2849 a signal has a DC offset.
2853 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
2857 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
2858 used to prevent clipping.
2862 Measure audio dynamic range.
2864 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
2865 is found in transition material. And anything less that 8 have very poor dynamics
2866 and is very compressed.
2868 The filter accepts the following options:
2872 Set window length in seconds used to split audio into segments of equal length.
2873 Default is 3 seconds.
2877 Dynamic Audio Normalizer.
2879 This filter applies a certain amount of gain to the input audio in order
2880 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
2881 contrast to more "simple" normalization algorithms, the Dynamic Audio
2882 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
2883 This allows for applying extra gain to the "quiet" sections of the audio
2884 while avoiding distortions or clipping the "loud" sections. In other words:
2885 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
2886 sections, in the sense that the volume of each section is brought to the
2887 same target level. Note, however, that the Dynamic Audio Normalizer achieves
2888 this goal *without* applying "dynamic range compressing". It will retain 100%
2889 of the dynamic range *within* each section of the audio file.
2893 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
2894 Default is 500 milliseconds.
2895 The Dynamic Audio Normalizer processes the input audio in small chunks,
2896 referred to as frames. This is required, because a peak magnitude has no
2897 meaning for just a single sample value. Instead, we need to determine the
2898 peak magnitude for a contiguous sequence of sample values. While a "standard"
2899 normalizer would simply use the peak magnitude of the complete file, the
2900 Dynamic Audio Normalizer determines the peak magnitude individually for each
2901 frame. The length of a frame is specified in milliseconds. By default, the
2902 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
2903 been found to give good results with most files.
2904 Note that the exact frame length, in number of samples, will be determined
2905 automatically, based on the sampling rate of the individual input audio file.
2908 Set the Gaussian filter window size. In range from 3 to 301, must be odd
2909 number. Default is 31.
2910 Probably the most important parameter of the Dynamic Audio Normalizer is the
2911 @code{window size} of the Gaussian smoothing filter. The filter's window size
2912 is specified in frames, centered around the current frame. For the sake of
2913 simplicity, this must be an odd number. Consequently, the default value of 31
2914 takes into account the current frame, as well as the 15 preceding frames and
2915 the 15 subsequent frames. Using a larger window results in a stronger
2916 smoothing effect and thus in less gain variation, i.e. slower gain
2917 adaptation. Conversely, using a smaller window results in a weaker smoothing
2918 effect and thus in more gain variation, i.e. faster gain adaptation.
2919 In other words, the more you increase this value, the more the Dynamic Audio
2920 Normalizer will behave like a "traditional" normalization filter. On the
2921 contrary, the more you decrease this value, the more the Dynamic Audio
2922 Normalizer will behave like a dynamic range compressor.
2925 Set the target peak value. This specifies the highest permissible magnitude
2926 level for the normalized audio input. This filter will try to approach the
2927 target peak magnitude as closely as possible, but at the same time it also
2928 makes sure that the normalized signal will never exceed the peak magnitude.
2929 A frame's maximum local gain factor is imposed directly by the target peak
2930 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
2931 It is not recommended to go above this value.
2934 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
2935 The Dynamic Audio Normalizer determines the maximum possible (local) gain
2936 factor for each input frame, i.e. the maximum gain factor that does not
2937 result in clipping or distortion. The maximum gain factor is determined by
2938 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
2939 additionally bounds the frame's maximum gain factor by a predetermined
2940 (global) maximum gain factor. This is done in order to avoid excessive gain
2941 factors in "silent" or almost silent frames. By default, the maximum gain
2942 factor is 10.0, For most inputs the default value should be sufficient and
2943 it usually is not recommended to increase this value. Though, for input
2944 with an extremely low overall volume level, it may be necessary to allow even
2945 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
2946 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
2947 Instead, a "sigmoid" threshold function will be applied. This way, the
2948 gain factors will smoothly approach the threshold value, but never exceed that
2952 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
2953 By default, the Dynamic Audio Normalizer performs "peak" normalization.
2954 This means that the maximum local gain factor for each frame is defined
2955 (only) by the frame's highest magnitude sample. This way, the samples can
2956 be amplified as much as possible without exceeding the maximum signal
2957 level, i.e. without clipping. Optionally, however, the Dynamic Audio
2958 Normalizer can also take into account the frame's root mean square,
2959 abbreviated RMS. In electrical engineering, the RMS is commonly used to
2960 determine the power of a time-varying signal. It is therefore considered
2961 that the RMS is a better approximation of the "perceived loudness" than
2962 just looking at the signal's peak magnitude. Consequently, by adjusting all
2963 frames to a constant RMS value, a uniform "perceived loudness" can be
2964 established. If a target RMS value has been specified, a frame's local gain
2965 factor is defined as the factor that would result in exactly that RMS value.
2966 Note, however, that the maximum local gain factor is still restricted by the
2967 frame's highest magnitude sample, in order to prevent clipping.
2970 Enable channels coupling. By default is enabled.
2971 By default, the Dynamic Audio Normalizer will amplify all channels by the same
2972 amount. This means the same gain factor will be applied to all channels, i.e.
2973 the maximum possible gain factor is determined by the "loudest" channel.
2974 However, in some recordings, it may happen that the volume of the different
2975 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
2976 In this case, this option can be used to disable the channel coupling. This way,
2977 the gain factor will be determined independently for each channel, depending
2978 only on the individual channel's highest magnitude sample. This allows for
2979 harmonizing the volume of the different channels.
2982 Enable DC bias correction. By default is disabled.
2983 An audio signal (in the time domain) is a sequence of sample values.
2984 In the Dynamic Audio Normalizer these sample values are represented in the
2985 -1.0 to 1.0 range, regardless of the original input format. Normally, the
2986 audio signal, or "waveform", should be centered around the zero point.
2987 That means if we calculate the mean value of all samples in a file, or in a
2988 single frame, then the result should be 0.0 or at least very close to that
2989 value. If, however, there is a significant deviation of the mean value from
2990 0.0, in either positive or negative direction, this is referred to as a
2991 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
2992 Audio Normalizer provides optional DC bias correction.
2993 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
2994 the mean value, or "DC correction" offset, of each input frame and subtract
2995 that value from all of the frame's sample values which ensures those samples
2996 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
2997 boundaries, the DC correction offset values will be interpolated smoothly
2998 between neighbouring frames.
3001 Enable alternative boundary mode. By default is disabled.
3002 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3003 around each frame. This includes the preceding frames as well as the
3004 subsequent frames. However, for the "boundary" frames, located at the very
3005 beginning and at the very end of the audio file, not all neighbouring
3006 frames are available. In particular, for the first few frames in the audio
3007 file, the preceding frames are not known. And, similarly, for the last few
3008 frames in the audio file, the subsequent frames are not known. Thus, the
3009 question arises which gain factors should be assumed for the missing frames
3010 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3011 to deal with this situation. The default boundary mode assumes a gain factor
3012 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3013 "fade out" at the beginning and at the end of the input, respectively.
3016 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3017 By default, the Dynamic Audio Normalizer does not apply "traditional"
3018 compression. This means that signal peaks will not be pruned and thus the
3019 full dynamic range will be retained within each local neighbourhood. However,
3020 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3021 normalization algorithm with a more "traditional" compression.
3022 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3023 (thresholding) function. If (and only if) the compression feature is enabled,
3024 all input frames will be processed by a soft knee thresholding function prior
3025 to the actual normalization process. Put simply, the thresholding function is
3026 going to prune all samples whose magnitude exceeds a certain threshold value.
3027 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3028 value. Instead, the threshold value will be adjusted for each individual
3030 In general, smaller parameters result in stronger compression, and vice versa.
3031 Values below 3.0 are not recommended, because audible distortion may appear.
3036 Make audio easier to listen to on headphones.
3038 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3039 so that when listened to on headphones the stereo image is moved from
3040 inside your head (standard for headphones) to outside and in front of
3041 the listener (standard for speakers).
3047 Apply a two-pole peaking equalisation (EQ) filter. With this
3048 filter, the signal-level at and around a selected frequency can
3049 be increased or decreased, whilst (unlike bandpass and bandreject
3050 filters) that at all other frequencies is unchanged.
3052 In order to produce complex equalisation curves, this filter can
3053 be given several times, each with a different central frequency.
3055 The filter accepts the following options:
3059 Set the filter's central frequency in Hz.
3062 Set method to specify band-width of filter.
3077 Specify the band-width of a filter in width_type units.
3080 Set the required gain or attenuation in dB.
3081 Beware of clipping when using a positive gain.
3084 Specify which channels to filter, by default all available are filtered.
3087 @subsection Examples
3090 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3092 equalizer=f=1000:t=h:width=200:g=-10
3096 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3098 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3102 @subsection Commands
3104 This filter supports the following commands:
3107 Change equalizer frequency.
3108 Syntax for the command is : "@var{frequency}"
3111 Change equalizer width_type.
3112 Syntax for the command is : "@var{width_type}"
3115 Change equalizer width.
3116 Syntax for the command is : "@var{width}"
3119 Change equalizer gain.
3120 Syntax for the command is : "@var{gain}"
3123 @section extrastereo
3125 Linearly increases the difference between left and right channels which
3126 adds some sort of "live" effect to playback.
3128 The filter accepts the following options:
3132 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3133 (average of both channels), with 1.0 sound will be unchanged, with
3134 -1.0 left and right channels will be swapped.
3137 Enable clipping. By default is enabled.
3140 @section firequalizer
3141 Apply FIR Equalization using arbitrary frequency response.
3143 The filter accepts the following option:
3147 Set gain curve equation (in dB). The expression can contain variables:
3150 the evaluated frequency
3154 channel number, set to 0 when multichannels evaluation is disabled
3156 channel id, see libavutil/channel_layout.h, set to the first channel id when
3157 multichannels evaluation is disabled
3161 channel_layout, see libavutil/channel_layout.h
3166 @item gain_interpolate(f)
3167 interpolate gain on frequency f based on gain_entry
3168 @item cubic_interpolate(f)
3169 same as gain_interpolate, but smoother
3171 This option is also available as command. Default is @code{gain_interpolate(f)}.
3174 Set gain entry for gain_interpolate function. The expression can
3178 store gain entry at frequency f with value g
3180 This option is also available as command.
3183 Set filter delay in seconds. Higher value means more accurate.
3184 Default is @code{0.01}.
3187 Set filter accuracy in Hz. Lower value means more accurate.
3188 Default is @code{5}.
3191 Set window function. Acceptable values are:
3194 rectangular window, useful when gain curve is already smooth
3196 hann window (default)
3202 3-terms continuous 1st derivative nuttall window
3204 minimum 3-terms discontinuous nuttall window
3206 4-terms continuous 1st derivative nuttall window
3208 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3210 blackman-harris window
3216 If enabled, use fixed number of audio samples. This improves speed when
3217 filtering with large delay. Default is disabled.
3220 Enable multichannels evaluation on gain. Default is disabled.
3223 Enable zero phase mode by subtracting timestamp to compensate delay.
3224 Default is disabled.
3227 Set scale used by gain. Acceptable values are:
3230 linear frequency, linear gain
3232 linear frequency, logarithmic (in dB) gain (default)
3234 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3236 logarithmic frequency, logarithmic gain
3240 Set file for dumping, suitable for gnuplot.
3243 Set scale for dumpfile. Acceptable values are same with scale option.
3247 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3248 Default is disabled.
3251 Enable minimum phase impulse response. Default is disabled.
3254 @subsection Examples
3259 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3262 lowpass at 1000 Hz with gain_entry:
3264 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3267 custom equalization:
3269 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3272 higher delay with zero phase to compensate delay:
3274 firequalizer=delay=0.1:fixed=on:zero_phase=on
3277 lowpass on left channel, highpass on right channel:
3279 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
3280 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
3285 Apply a flanging effect to the audio.
3287 The filter accepts the following options:
3291 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
3294 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
3297 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
3301 Set percentage of delayed signal mixed with original. Range from 0 to 100.
3302 Default value is 71.
3305 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
3308 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
3309 Default value is @var{sinusoidal}.
3312 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
3313 Default value is 25.
3316 Set delay-line interpolation, @var{linear} or @var{quadratic}.
3317 Default is @var{linear}.
3321 Apply Haas effect to audio.
3323 Note that this makes most sense to apply on mono signals.
3324 With this filter applied to mono signals it give some directionality and
3325 stretches its stereo image.
3327 The filter accepts the following options:
3331 Set input level. By default is @var{1}, or 0dB
3334 Set output level. By default is @var{1}, or 0dB.
3337 Set gain applied to side part of signal. By default is @var{1}.
3340 Set kind of middle source. Can be one of the following:
3350 Pick middle part signal of stereo image.
3353 Pick side part signal of stereo image.
3357 Change middle phase. By default is disabled.
3360 Set left channel delay. By default is @var{2.05} milliseconds.
3363 Set left channel balance. By default is @var{-1}.
3366 Set left channel gain. By default is @var{1}.
3369 Change left phase. By default is disabled.
3372 Set right channel delay. By defaults is @var{2.12} milliseconds.
3375 Set right channel balance. By default is @var{1}.
3378 Set right channel gain. By default is @var{1}.
3381 Change right phase. By default is enabled.
3386 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3387 embedded HDCD codes is expanded into a 20-bit PCM stream.
3389 The filter supports the Peak Extend and Low-level Gain Adjustment features
3390 of HDCD, and detects the Transient Filter flag.
3393 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3396 When using the filter with wav, note the default encoding for wav is 16-bit,
3397 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3398 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3400 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3401 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3404 The filter accepts the following options:
3407 @item disable_autoconvert
3408 Disable any automatic format conversion or resampling in the filter graph.
3410 @item process_stereo
3411 Process the stereo channels together. If target_gain does not match between
3412 channels, consider it invalid and use the last valid target_gain.
3415 Set the code detect timer period in ms.
3418 Always extend peaks above -3dBFS even if PE isn't signaled.
3421 Replace audio with a solid tone and adjust the amplitude to signal some
3422 specific aspect of the decoding process. The output file can be loaded in
3423 an audio editor alongside the original to aid analysis.
3425 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
3432 Gain adjustment level at each sample
3434 Samples where peak extend occurs
3436 Samples where the code detect timer is active
3438 Samples where the target gain does not match between channels
3444 Apply head-related transfer functions (HRTFs) to create virtual
3445 loudspeakers around the user for binaural listening via headphones.
3446 The HRIRs are provided via additional streams, for each channel
3447 one stereo input stream is needed.
3449 The filter accepts the following options:
3453 Set mapping of input streams for convolution.
3454 The argument is a '|'-separated list of channel names in order as they
3455 are given as additional stream inputs for filter.
3456 This also specify number of input streams. Number of input streams
3457 must be not less than number of channels in first stream plus one.
3460 Set gain applied to audio. Value is in dB. Default is 0.
3463 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
3464 processing audio in time domain which is slow.
3465 @var{freq} is processing audio in frequency domain which is fast.
3466 Default is @var{freq}.
3469 Set custom gain for LFE channels. Value is in dB. Default is 0.
3472 Set size of frame in number of samples which will be processed at once.
3473 Default value is @var{1024}. Allowed range is from 1024 to 96000.
3476 Set format of hrir stream.
3477 Default value is @var{stereo}. Alternative value is @var{multich}.
3478 If value is set to @var{stereo}, number of additional streams should
3479 be greater or equal to number of input channels in first input stream.
3480 Also each additional stream should have stereo number of channels.
3481 If value is set to @var{multich}, number of additional streams should
3482 be exactly one. Also number of input channels of additional stream
3483 should be equal or greater than twice number of channels of first input
3487 @subsection Examples
3491 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3492 each amovie filter use stereo file with IR coefficients as input.
3493 The files give coefficients for each position of virtual loudspeaker:
3495 ffmpeg -i input.wav -lavfi-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],[a:0][fl][fr][fc][lfe][bl][br][sl][sr]headphone=FL|FR|FC|LFE|BL|BR|SL|SR"
3500 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3501 but now in @var{multich} @var{hrir} format.
3503 ffmpeg -i input.wav -lavfi-complex "amovie=minp.wav[hrirs],[a:0][hrirs]headphone=map=FL|FR|FC|LFE|BL|BR|SL|SR:hrir=multich"
3510 Apply a high-pass filter with 3dB point frequency.
3511 The filter can be either single-pole, or double-pole (the default).
3512 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3514 The filter accepts the following options:
3518 Set frequency in Hz. Default is 3000.
3521 Set number of poles. Default is 2.
3524 Set method to specify band-width of filter.
3539 Specify the band-width of a filter in width_type units.
3540 Applies only to double-pole filter.
3541 The default is 0.707q and gives a Butterworth response.
3544 Specify which channels to filter, by default all available are filtered.
3547 @subsection Commands
3549 This filter supports the following commands:
3552 Change highpass frequency.
3553 Syntax for the command is : "@var{frequency}"
3556 Change highpass width_type.
3557 Syntax for the command is : "@var{width_type}"
3560 Change highpass width.
3561 Syntax for the command is : "@var{width}"
3566 Join multiple input streams into one multi-channel stream.
3568 It accepts the following parameters:
3572 The number of input streams. It defaults to 2.
3574 @item channel_layout
3575 The desired output channel layout. It defaults to stereo.
3578 Map channels from inputs to output. The argument is a '|'-separated list of
3579 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
3580 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
3581 can be either the name of the input channel (e.g. FL for front left) or its
3582 index in the specified input stream. @var{out_channel} is the name of the output
3586 The filter will attempt to guess the mappings when they are not specified
3587 explicitly. It does so by first trying to find an unused matching input channel
3588 and if that fails it picks the first unused input channel.
3590 Join 3 inputs (with properly set channel layouts):
3592 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
3595 Build a 5.1 output from 6 single-channel streams:
3597 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
3598 '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'
3604 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
3606 To enable compilation of this filter you need to configure FFmpeg with
3607 @code{--enable-ladspa}.
3611 Specifies the name of LADSPA plugin library to load. If the environment
3612 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
3613 each one of the directories specified by the colon separated list in
3614 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
3615 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
3616 @file{/usr/lib/ladspa/}.
3619 Specifies the plugin within the library. Some libraries contain only
3620 one plugin, but others contain many of them. If this is not set filter
3621 will list all available plugins within the specified library.
3624 Set the '|' separated list of controls which are zero or more floating point
3625 values that determine the behavior of the loaded plugin (for example delay,
3627 Controls need to be defined using the following syntax:
3628 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
3629 @var{valuei} is the value set on the @var{i}-th control.
3630 Alternatively they can be also defined using the following syntax:
3631 @var{value0}|@var{value1}|@var{value2}|..., where
3632 @var{valuei} is the value set on the @var{i}-th control.
3633 If @option{controls} is set to @code{help}, all available controls and
3634 their valid ranges are printed.
3636 @item sample_rate, s
3637 Specify the sample rate, default to 44100. Only used if plugin have
3641 Set the number of samples per channel per each output frame, default
3642 is 1024. Only used if plugin have zero inputs.
3645 Set the minimum duration of the sourced audio. See
3646 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
3647 for the accepted syntax.
3648 Note that the resulting duration may be greater than the specified duration,
3649 as the generated audio is always cut at the end of a complete frame.
3650 If not specified, or the expressed duration is negative, the audio is
3651 supposed to be generated forever.
3652 Only used if plugin have zero inputs.
3656 @subsection Examples
3660 List all available plugins within amp (LADSPA example plugin) library:
3666 List all available controls and their valid ranges for @code{vcf_notch}
3667 plugin from @code{VCF} library:
3669 ladspa=f=vcf:p=vcf_notch:c=help
3673 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
3676 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
3680 Add reverberation to the audio using TAP-plugins
3681 (Tom's Audio Processing plugins):
3683 ladspa=file=tap_reverb:tap_reverb
3687 Generate white noise, with 0.2 amplitude:
3689 ladspa=file=cmt:noise_source_white:c=c0=.2
3693 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
3694 @code{C* Audio Plugin Suite} (CAPS) library:
3696 ladspa=file=caps:Click:c=c1=20'
3700 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
3702 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
3706 Increase volume by 20dB using fast lookahead limiter from Steve Harris
3707 @code{SWH Plugins} collection:
3709 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
3713 Attenuate low frequencies using Multiband EQ from Steve Harris
3714 @code{SWH Plugins} collection:
3716 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
3720 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
3723 ladspa=caps:Narrower
3727 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
3729 ladspa=caps:White:.2
3733 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
3735 ladspa=caps:Fractal:c=c1=1
3739 Dynamic volume normalization using @code{VLevel} plugin:
3741 ladspa=vlevel-ladspa:vlevel_mono
3745 @subsection Commands
3747 This filter supports the following commands:
3750 Modify the @var{N}-th control value.
3752 If the specified value is not valid, it is ignored and prior one is kept.
3757 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
3758 Support for both single pass (livestreams, files) and double pass (files) modes.
3759 This algorithm can target IL, LRA, and maximum true peak. To accurately detect true peaks,
3760 the audio stream will be upsampled to 192 kHz unless the normalization mode is linear.
3761 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
3763 The filter accepts the following options:
3767 Set integrated loudness target.
3768 Range is -70.0 - -5.0. Default value is -24.0.
3771 Set loudness range target.
3772 Range is 1.0 - 20.0. Default value is 7.0.
3775 Set maximum true peak.
3776 Range is -9.0 - +0.0. Default value is -2.0.
3778 @item measured_I, measured_i
3779 Measured IL of input file.
3780 Range is -99.0 - +0.0.
3782 @item measured_LRA, measured_lra
3783 Measured LRA of input file.
3784 Range is 0.0 - 99.0.
3786 @item measured_TP, measured_tp
3787 Measured true peak of input file.
3788 Range is -99.0 - +99.0.
3790 @item measured_thresh
3791 Measured threshold of input file.
3792 Range is -99.0 - +0.0.
3795 Set offset gain. Gain is applied before the true-peak limiter.
3796 Range is -99.0 - +99.0. Default is +0.0.
3799 Normalize linearly if possible.
3800 measured_I, measured_LRA, measured_TP, and measured_thresh must also
3801 to be specified in order to use this mode.
3802 Options are true or false. Default is true.
3805 Treat mono input files as "dual-mono". If a mono file is intended for playback
3806 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
3807 If set to @code{true}, this option will compensate for this effect.
3808 Multi-channel input files are not affected by this option.
3809 Options are true or false. Default is false.
3812 Set print format for stats. Options are summary, json, or none.
3813 Default value is none.
3818 Apply a low-pass filter with 3dB point frequency.
3819 The filter can be either single-pole or double-pole (the default).
3820 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3822 The filter accepts the following options:
3826 Set frequency in Hz. Default is 500.
3829 Set number of poles. Default is 2.
3832 Set method to specify band-width of filter.
3847 Specify the band-width of a filter in width_type units.
3848 Applies only to double-pole filter.
3849 The default is 0.707q and gives a Butterworth response.
3852 Specify which channels to filter, by default all available are filtered.
3855 @subsection Examples
3858 Lowpass only LFE channel, it LFE is not present it does nothing:
3864 @subsection Commands
3866 This filter supports the following commands:
3869 Change lowpass frequency.
3870 Syntax for the command is : "@var{frequency}"
3873 Change lowpass width_type.
3874 Syntax for the command is : "@var{width_type}"
3877 Change lowpass width.
3878 Syntax for the command is : "@var{width}"
3883 Load a LV2 (LADSPA Version 2) plugin.
3885 To enable compilation of this filter you need to configure FFmpeg with
3886 @code{--enable-lv2}.
3890 Specifies the plugin URI. You may need to escape ':'.
3893 Set the '|' separated list of controls which are zero or more floating point
3894 values that determine the behavior of the loaded plugin (for example delay,
3896 If @option{controls} is set to @code{help}, all available controls and
3897 their valid ranges are printed.
3899 @item sample_rate, s
3900 Specify the sample rate, default to 44100. Only used if plugin have
3904 Set the number of samples per channel per each output frame, default
3905 is 1024. Only used if plugin have zero inputs.
3908 Set the minimum duration of the sourced audio. See
3909 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
3910 for the accepted syntax.
3911 Note that the resulting duration may be greater than the specified duration,
3912 as the generated audio is always cut at the end of a complete frame.
3913 If not specified, or the expressed duration is negative, the audio is
3914 supposed to be generated forever.
3915 Only used if plugin have zero inputs.
3918 @subsection Examples
3922 Apply bass enhancer plugin from Calf:
3924 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
3928 Apply vinyl plugin from Calf:
3930 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
3934 Apply bit crusher plugin from ArtyFX:
3936 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
3941 Multiband Compress or expand the audio's dynamic range.
3943 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
3944 This is akin to the crossover of a loudspeaker, and results in flat frequency
3945 response when absent compander action.
3947 It accepts the following parameters:
3951 This option syntax is:
3952 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
3953 For explanation of each item refer to compand filter documentation.
3959 Mix channels with specific gain levels. The filter accepts the output
3960 channel layout followed by a set of channels definitions.
3962 This filter is also designed to efficiently remap the channels of an audio
3965 The filter accepts parameters of the form:
3966 "@var{l}|@var{outdef}|@var{outdef}|..."
3970 output channel layout or number of channels
3973 output channel specification, of the form:
3974 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
3977 output channel to define, either a channel name (FL, FR, etc.) or a channel
3978 number (c0, c1, etc.)
3981 multiplicative coefficient for the channel, 1 leaving the volume unchanged
3984 input channel to use, see out_name for details; it is not possible to mix
3985 named and numbered input channels
3988 If the `=' in a channel specification is replaced by `<', then the gains for
3989 that specification will be renormalized so that the total is 1, thus
3990 avoiding clipping noise.
3992 @subsection Mixing examples
3994 For example, if you want to down-mix from stereo to mono, but with a bigger
3995 factor for the left channel:
3997 pan=1c|c0=0.9*c0+0.1*c1
4000 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4001 7-channels surround:
4003 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4006 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4007 that should be preferred (see "-ac" option) unless you have very specific
4010 @subsection Remapping examples
4012 The channel remapping will be effective if, and only if:
4015 @item gain coefficients are zeroes or ones,
4016 @item only one input per channel output,
4019 If all these conditions are satisfied, the filter will notify the user ("Pure
4020 channel mapping detected"), and use an optimized and lossless method to do the
4023 For example, if you have a 5.1 source and want a stereo audio stream by
4024 dropping the extra channels:
4026 pan="stereo| c0=FL | c1=FR"
4029 Given the same source, you can also switch front left and front right channels
4030 and keep the input channel layout:
4032 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4035 If the input is a stereo audio stream, you can mute the front left channel (and
4036 still keep the stereo channel layout) with:
4041 Still with a stereo audio stream input, you can copy the right channel in both
4042 front left and right:
4044 pan="stereo| c0=FR | c1=FR"
4049 ReplayGain scanner filter. This filter takes an audio stream as an input and
4050 outputs it unchanged.
4051 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4055 Convert the audio sample format, sample rate and channel layout. It is
4056 not meant to be used directly.
4059 Apply time-stretching and pitch-shifting with librubberband.
4061 To enable compilation of this filter, you need to configure FFmpeg with
4062 @code{--enable-librubberband}.
4064 The filter accepts the following options:
4068 Set tempo scale factor.
4071 Set pitch scale factor.
4074 Set transients detector.
4075 Possible values are:
4084 Possible values are:
4093 Possible values are:
4100 Set processing window size.
4101 Possible values are:
4110 Possible values are:
4117 Enable formant preservation when shift pitching.
4118 Possible values are:
4126 Possible values are:
4135 Possible values are:
4142 @section sidechaincompress
4144 This filter acts like normal compressor but has the ability to compress
4145 detected signal using second input signal.
4146 It needs two input streams and returns one output stream.
4147 First input stream will be processed depending on second stream signal.
4148 The filtered signal then can be filtered with other filters in later stages of
4149 processing. See @ref{pan} and @ref{amerge} filter.
4151 The filter accepts the following options:
4155 Set input gain. Default is 1. Range is between 0.015625 and 64.
4158 If a signal of second stream raises above this level it will affect the gain
4159 reduction of first stream.
4160 By default is 0.125. Range is between 0.00097563 and 1.
4163 Set a ratio about which the signal is reduced. 1:2 means that if the level
4164 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4165 Default is 2. Range is between 1 and 20.
4168 Amount of milliseconds the signal has to rise above the threshold before gain
4169 reduction starts. Default is 20. Range is between 0.01 and 2000.
4172 Amount of milliseconds the signal has to fall below the threshold before
4173 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4176 Set the amount by how much signal will be amplified after processing.
4177 Default is 1. Range is from 1 to 64.
4180 Curve the sharp knee around the threshold to enter gain reduction more softly.
4181 Default is 2.82843. Range is between 1 and 8.
4184 Choose if the @code{average} level between all channels of side-chain stream
4185 or the louder(@code{maximum}) channel of side-chain stream affects the
4186 reduction. Default is @code{average}.
4189 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4190 of @code{rms}. Default is @code{rms} which is mainly smoother.
4193 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4196 How much to use compressed signal in output. Default is 1.
4197 Range is between 0 and 1.
4200 @subsection Examples
4204 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
4205 depending on the signal of 2nd input and later compressed signal to be
4206 merged with 2nd input:
4208 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
4212 @section sidechaingate
4214 A sidechain gate acts like a normal (wideband) gate but has the ability to
4215 filter the detected signal before sending it to the gain reduction stage.
4216 Normally a gate uses the full range signal to detect a level above the
4218 For example: If you cut all lower frequencies from your sidechain signal
4219 the gate will decrease the volume of your track only if not enough highs
4220 appear. With this technique you are able to reduce the resonation of a
4221 natural drum or remove "rumbling" of muted strokes from a heavily distorted
4223 It needs two input streams and returns one output stream.
4224 First input stream will be processed depending on second stream signal.
4226 The filter accepts the following options:
4230 Set input level before filtering.
4231 Default is 1. Allowed range is from 0.015625 to 64.
4234 Set the level of gain reduction when the signal is below the threshold.
4235 Default is 0.06125. Allowed range is from 0 to 1.
4238 If a signal rises above this level the gain reduction is released.
4239 Default is 0.125. Allowed range is from 0 to 1.
4242 Set a ratio about which the signal is reduced.
4243 Default is 2. Allowed range is from 1 to 9000.
4246 Amount of milliseconds the signal has to rise above the threshold before gain
4248 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
4251 Amount of milliseconds the signal has to fall below the threshold before the
4252 reduction is increased again. Default is 250 milliseconds.
4253 Allowed range is from 0.01 to 9000.
4256 Set amount of amplification of signal after processing.
4257 Default is 1. Allowed range is from 1 to 64.
4260 Curve the sharp knee around the threshold to enter gain reduction more softly.
4261 Default is 2.828427125. Allowed range is from 1 to 8.
4264 Choose if exact signal should be taken for detection or an RMS like one.
4265 Default is rms. Can be peak or rms.
4268 Choose if the average level between all channels or the louder channel affects
4270 Default is average. Can be average or maximum.
4273 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
4276 @section silencedetect
4278 Detect silence in an audio stream.
4280 This filter logs a message when it detects that the input audio volume is less
4281 or equal to a noise tolerance value for a duration greater or equal to the
4282 minimum detected noise duration.
4284 The printed times and duration are expressed in seconds.
4286 The filter accepts the following options:
4290 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4291 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4294 Set silence duration until notification (default is 2 seconds).
4297 Process each channel separately, instead of combined. By default is disabled.
4300 @subsection Examples
4304 Detect 5 seconds of silence with -50dB noise tolerance:
4306 silencedetect=n=-50dB:d=5
4310 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4311 tolerance in @file{silence.mp3}:
4313 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4317 @section silenceremove
4319 Remove silence from the beginning, middle or end of the audio.
4321 The filter accepts the following options:
4325 This value is used to indicate if audio should be trimmed at beginning of
4326 the audio. A value of zero indicates no silence should be trimmed from the
4327 beginning. When specifying a non-zero value, it trims audio up until it
4328 finds non-silence. Normally, when trimming silence from beginning of audio
4329 the @var{start_periods} will be @code{1} but it can be increased to higher
4330 values to trim all audio up to specific count of non-silence periods.
4331 Default value is @code{0}.
4333 @item start_duration
4334 Specify the amount of time that non-silence must be detected before it stops
4335 trimming audio. By increasing the duration, bursts of noises can be treated
4336 as silence and trimmed off. Default value is @code{0}.
4338 @item start_threshold
4339 This indicates what sample value should be treated as silence. For digital
4340 audio, a value of @code{0} may be fine but for audio recorded from analog,
4341 you may wish to increase the value to account for background noise.
4342 Can be specified in dB (in case "dB" is appended to the specified value)
4343 or amplitude ratio. Default value is @code{0}.
4346 Set the count for trimming silence from the end of audio.
4347 To remove silence from the middle of a file, specify a @var{stop_periods}
4348 that is negative. This value is then treated as a positive value and is
4349 used to indicate the effect should restart processing as specified by
4350 @var{start_periods}, making it suitable for removing periods of silence
4351 in the middle of the audio.
4352 Default value is @code{0}.
4355 Specify a duration of silence that must exist before audio is not copied any
4356 more. By specifying a higher duration, silence that is wanted can be left in
4358 Default value is @code{0}.
4360 @item stop_threshold
4361 This is the same as @option{start_threshold} but for trimming silence from
4363 Can be specified in dB (in case "dB" is appended to the specified value)
4364 or amplitude ratio. Default value is @code{0}.
4367 This indicates that @var{stop_duration} length of audio should be left intact
4368 at the beginning of each period of silence.
4369 For example, if you want to remove long pauses between words but do not want
4370 to remove the pauses completely. Default value is @code{0}.
4373 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4374 and works better with digital silence which is exactly 0.
4375 Default value is @code{rms}.
4378 Set ratio used to calculate size of window for detecting silence.
4379 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4382 @subsection Examples
4386 The following example shows how this filter can be used to start a recording
4387 that does not contain the delay at the start which usually occurs between
4388 pressing the record button and the start of the performance:
4390 silenceremove=1:5:0.02
4394 Trim all silence encountered from beginning to end where there is more than 1
4395 second of silence in audio:
4397 silenceremove=0:0:0:-1:1:-90dB
4403 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4404 loudspeakers around the user for binaural listening via headphones (audio
4405 formats up to 9 channels supported).
4406 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4407 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4408 Austrian Academy of Sciences.
4410 To enable compilation of this filter you need to configure FFmpeg with
4411 @code{--enable-libmysofa}.
4413 The filter accepts the following options:
4417 Set the SOFA file used for rendering.
4420 Set gain applied to audio. Value is in dB. Default is 0.
4423 Set rotation of virtual loudspeakers in deg. Default is 0.
4426 Set elevation of virtual speakers in deg. Default is 0.
4429 Set distance in meters between loudspeakers and the listener with near-field
4430 HRTFs. Default is 1.
4433 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4434 processing audio in time domain which is slow.
4435 @var{freq} is processing audio in frequency domain which is fast.
4436 Default is @var{freq}.
4439 Set custom positions of virtual loudspeakers. Syntax for this option is:
4440 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
4441 Each virtual loudspeaker is described with short channel name following with
4442 azimuth and elevation in degrees.
4443 Each virtual loudspeaker description is separated by '|'.
4444 For example to override front left and front right channel positions use:
4445 'speakers=FL 45 15|FR 345 15'.
4446 Descriptions with unrecognised channel names are ignored.
4449 Set custom gain for LFE channels. Value is in dB. Default is 0.
4452 @subsection Examples
4456 Using ClubFritz6 sofa file:
4458 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
4462 Using ClubFritz12 sofa file and bigger radius with small rotation:
4464 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
4468 Similar as above but with custom speaker positions for front left, front right, back left and back right
4469 and also with custom gain:
4471 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
4475 @section stereotools
4477 This filter has some handy utilities to manage stereo signals, for converting
4478 M/S stereo recordings to L/R signal while having control over the parameters
4479 or spreading the stereo image of master track.
4481 The filter accepts the following options:
4485 Set input level before filtering for both channels. Defaults is 1.
4486 Allowed range is from 0.015625 to 64.
4489 Set output level after filtering for both channels. Defaults is 1.
4490 Allowed range is from 0.015625 to 64.
4493 Set input balance between both channels. Default is 0.
4494 Allowed range is from -1 to 1.
4497 Set output balance between both channels. Default is 0.
4498 Allowed range is from -1 to 1.
4501 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
4502 clipping. Disabled by default.
4505 Mute the left channel. Disabled by default.
4508 Mute the right channel. Disabled by default.
4511 Change the phase of the left channel. Disabled by default.
4514 Change the phase of the right channel. Disabled by default.
4517 Set stereo mode. Available values are:
4521 Left/Right to Left/Right, this is default.
4524 Left/Right to Mid/Side.
4527 Mid/Side to Left/Right.
4530 Left/Right to Left/Left.
4533 Left/Right to Right/Right.
4536 Left/Right to Left + Right.
4539 Left/Right to Right/Left.
4542 Mid/Side to Left/Left.
4545 Mid/Side to Right/Right.
4549 Set level of side signal. Default is 1.
4550 Allowed range is from 0.015625 to 64.
4553 Set balance of side signal. Default is 0.
4554 Allowed range is from -1 to 1.
4557 Set level of the middle signal. Default is 1.
4558 Allowed range is from 0.015625 to 64.
4561 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
4564 Set stereo base between mono and inversed channels. Default is 0.
4565 Allowed range is from -1 to 1.
4568 Set delay in milliseconds how much to delay left from right channel and
4569 vice versa. Default is 0. Allowed range is from -20 to 20.
4572 Set S/C level. Default is 1. Allowed range is from 1 to 100.
4575 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
4577 @item bmode_in, bmode_out
4578 Set balance mode for balance_in/balance_out option.
4580 Can be one of the following:
4584 Classic balance mode. Attenuate one channel at time.
4585 Gain is raised up to 1.
4588 Similar as classic mode above but gain is raised up to 2.
4591 Equal power distribution, from -6dB to +6dB range.
4595 @subsection Examples
4599 Apply karaoke like effect:
4601 stereotools=mlev=0.015625
4605 Convert M/S signal to L/R:
4607 "stereotools=mode=ms>lr"
4611 @section stereowiden
4613 This filter enhance the stereo effect by suppressing signal common to both
4614 channels and by delaying the signal of left into right and vice versa,
4615 thereby widening the stereo effect.
4617 The filter accepts the following options:
4621 Time in milliseconds of the delay of left signal into right and vice versa.
4622 Default is 20 milliseconds.
4625 Amount of gain in delayed signal into right and vice versa. Gives a delay
4626 effect of left signal in right output and vice versa which gives widening
4627 effect. Default is 0.3.
4630 Cross feed of left into right with inverted phase. This helps in suppressing
4631 the mono. If the value is 1 it will cancel all the signal common to both
4632 channels. Default is 0.3.
4635 Set level of input signal of original channel. Default is 0.8.
4638 @section superequalizer
4639 Apply 18 band equalizer.
4641 The filter accepts the following options:
4648 Set 131Hz band gain.
4650 Set 185Hz band gain.
4652 Set 262Hz band gain.
4654 Set 370Hz band gain.
4656 Set 523Hz band gain.
4658 Set 740Hz band gain.
4660 Set 1047Hz band gain.
4662 Set 1480Hz band gain.
4664 Set 2093Hz band gain.
4666 Set 2960Hz band gain.
4668 Set 4186Hz band gain.
4670 Set 5920Hz band gain.
4672 Set 8372Hz band gain.
4674 Set 11840Hz band gain.
4676 Set 16744Hz band gain.
4678 Set 20000Hz band gain.
4682 Apply audio surround upmix filter.
4684 This filter allows to produce multichannel output from audio stream.
4686 The filter accepts the following options:
4690 Set output channel layout. By default, this is @var{5.1}.
4692 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4693 for the required syntax.
4696 Set input channel layout. By default, this is @var{stereo}.
4698 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4699 for the required syntax.
4702 Set input volume level. By default, this is @var{1}.
4705 Set output volume level. By default, this is @var{1}.
4708 Enable LFE channel output if output channel layout has it. By default, this is enabled.
4711 Set LFE low cut off frequency. By default, this is @var{128} Hz.
4714 Set LFE high cut off frequency. By default, this is @var{256} Hz.
4717 Set front center input volume. By default, this is @var{1}.
4720 Set front center output volume. By default, this is @var{1}.
4723 Set LFE input volume. By default, this is @var{1}.
4726 Set LFE output volume. By default, this is @var{1}.
4729 @section treble, highshelf
4731 Boost or cut treble (upper) frequencies of the audio using a two-pole
4732 shelving filter with a response similar to that of a standard
4733 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
4735 The filter accepts the following options:
4739 Give the gain at whichever is the lower of ~22 kHz and the
4740 Nyquist frequency. Its useful range is about -20 (for a large cut)
4741 to +20 (for a large boost). Beware of clipping when using a positive gain.
4744 Set the filter's central frequency and so can be used
4745 to extend or reduce the frequency range to be boosted or cut.
4746 The default value is @code{3000} Hz.
4749 Set method to specify band-width of filter.
4764 Determine how steep is the filter's shelf transition.
4767 Specify which channels to filter, by default all available are filtered.
4770 @subsection Commands
4772 This filter supports the following commands:
4775 Change treble frequency.
4776 Syntax for the command is : "@var{frequency}"
4779 Change treble width_type.
4780 Syntax for the command is : "@var{width_type}"
4783 Change treble width.
4784 Syntax for the command is : "@var{width}"
4788 Syntax for the command is : "@var{gain}"
4793 Sinusoidal amplitude modulation.
4795 The filter accepts the following options:
4799 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
4800 (20 Hz or lower) will result in a tremolo effect.
4801 This filter may also be used as a ring modulator by specifying
4802 a modulation frequency higher than 20 Hz.
4803 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
4806 Depth of modulation as a percentage. Range is 0.0 - 1.0.
4807 Default value is 0.5.
4812 Sinusoidal phase modulation.
4814 The filter accepts the following options:
4818 Modulation frequency in Hertz.
4819 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
4822 Depth of modulation as a percentage. Range is 0.0 - 1.0.
4823 Default value is 0.5.
4828 Adjust the input audio volume.
4830 It accepts the following parameters:
4834 Set audio volume expression.
4836 Output values are clipped to the maximum value.
4838 The output audio volume is given by the relation:
4840 @var{output_volume} = @var{volume} * @var{input_volume}
4843 The default value for @var{volume} is "1.0".
4846 This parameter represents the mathematical precision.
4848 It determines which input sample formats will be allowed, which affects the
4849 precision of the volume scaling.
4853 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
4855 32-bit floating-point; this limits input sample format to FLT. (default)
4857 64-bit floating-point; this limits input sample format to DBL.
4861 Choose the behaviour on encountering ReplayGain side data in input frames.
4865 Remove ReplayGain side data, ignoring its contents (the default).
4868 Ignore ReplayGain side data, but leave it in the frame.
4871 Prefer the track gain, if present.
4874 Prefer the album gain, if present.
4877 @item replaygain_preamp
4878 Pre-amplification gain in dB to apply to the selected replaygain gain.
4880 Default value for @var{replaygain_preamp} is 0.0.
4883 Set when the volume expression is evaluated.
4885 It accepts the following values:
4888 only evaluate expression once during the filter initialization, or
4889 when the @samp{volume} command is sent
4892 evaluate expression for each incoming frame
4895 Default value is @samp{once}.
4898 The volume expression can contain the following parameters.
4902 frame number (starting at zero)
4905 @item nb_consumed_samples
4906 number of samples consumed by the filter
4908 number of samples in the current frame
4910 original frame position in the file
4916 PTS at start of stream
4918 time at start of stream
4924 last set volume value
4927 Note that when @option{eval} is set to @samp{once} only the
4928 @var{sample_rate} and @var{tb} variables are available, all other
4929 variables will evaluate to NAN.
4931 @subsection Commands
4933 This filter supports the following commands:
4936 Modify the volume expression.
4937 The command accepts the same syntax of the corresponding option.
4939 If the specified expression is not valid, it is kept at its current
4941 @item replaygain_noclip
4942 Prevent clipping by limiting the gain applied.
4944 Default value for @var{replaygain_noclip} is 1.
4948 @subsection Examples
4952 Halve the input audio volume:
4956 volume=volume=-6.0206dB
4959 In all the above example the named key for @option{volume} can be
4960 omitted, for example like in:
4966 Increase input audio power by 6 decibels using fixed-point precision:
4968 volume=volume=6dB:precision=fixed
4972 Fade volume after time 10 with an annihilation period of 5 seconds:
4974 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
4978 @section volumedetect
4980 Detect the volume of the input video.
4982 The filter has no parameters. The input is not modified. Statistics about
4983 the volume will be printed in the log when the input stream end is reached.
4985 In particular it will show the mean volume (root mean square), maximum
4986 volume (on a per-sample basis), and the beginning of a histogram of the
4987 registered volume values (from the maximum value to a cumulated 1/1000 of
4990 All volumes are in decibels relative to the maximum PCM value.
4992 @subsection Examples
4994 Here is an excerpt of the output:
4996 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
4997 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
4998 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
4999 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5000 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5001 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5002 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5003 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5004 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5010 The mean square energy is approximately -27 dB, or 10^-2.7.
5012 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5014 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5017 In other words, raising the volume by +4 dB does not cause any clipping,
5018 raising it by +5 dB causes clipping for 6 samples, etc.
5020 @c man end AUDIO FILTERS
5022 @chapter Audio Sources
5023 @c man begin AUDIO SOURCES
5025 Below is a description of the currently available audio sources.
5029 Buffer audio frames, and make them available to the filter chain.
5031 This source is mainly intended for a programmatic use, in particular
5032 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
5034 It accepts the following parameters:
5038 The timebase which will be used for timestamps of submitted frames. It must be
5039 either a floating-point number or in @var{numerator}/@var{denominator} form.
5042 The sample rate of the incoming audio buffers.
5045 The sample format of the incoming audio buffers.
5046 Either a sample format name or its corresponding integer representation from
5047 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5049 @item channel_layout
5050 The channel layout of the incoming audio buffers.
5051 Either a channel layout name from channel_layout_map in
5052 @file{libavutil/channel_layout.c} or its corresponding integer representation
5053 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5056 The number of channels of the incoming audio buffers.
5057 If both @var{channels} and @var{channel_layout} are specified, then they
5062 @subsection Examples
5065 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5068 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5069 Since the sample format with name "s16p" corresponds to the number
5070 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5073 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5078 Generate an audio signal specified by an expression.
5080 This source accepts in input one or more expressions (one for each
5081 channel), which are evaluated and used to generate a corresponding
5084 This source accepts the following options:
5088 Set the '|'-separated expressions list for each separate channel. In case the
5089 @option{channel_layout} option is not specified, the selected channel layout
5090 depends on the number of provided expressions. Otherwise the last
5091 specified expression is applied to the remaining output channels.
5093 @item channel_layout, c
5094 Set the channel layout. The number of channels in the specified layout
5095 must be equal to the number of specified expressions.
5098 Set the minimum duration of the sourced audio. See
5099 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5100 for the accepted syntax.
5101 Note that the resulting duration may be greater than the specified
5102 duration, as the generated audio is always cut at the end of a
5105 If not specified, or the expressed duration is negative, the audio is
5106 supposed to be generated forever.
5109 Set the number of samples per channel per each output frame,
5112 @item sample_rate, s
5113 Specify the sample rate, default to 44100.
5116 Each expression in @var{exprs} can contain the following constants:
5120 number of the evaluated sample, starting from 0
5123 time of the evaluated sample expressed in seconds, starting from 0
5130 @subsection Examples
5140 Generate a sin signal with frequency of 440 Hz, set sample rate to
5143 aevalsrc="sin(440*2*PI*t):s=8000"
5147 Generate a two channels signal, specify the channel layout (Front
5148 Center + Back Center) explicitly:
5150 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5154 Generate white noise:
5156 aevalsrc="-2+random(0)"
5160 Generate an amplitude modulated signal:
5162 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
5166 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
5168 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
5175 The null audio source, return unprocessed audio frames. It is mainly useful
5176 as a template and to be employed in analysis / debugging tools, or as
5177 the source for filters which ignore the input data (for example the sox
5180 This source accepts the following options:
5184 @item channel_layout, cl
5186 Specifies the channel layout, and can be either an integer or a string
5187 representing a channel layout. The default value of @var{channel_layout}
5190 Check the channel_layout_map definition in
5191 @file{libavutil/channel_layout.c} for the mapping between strings and
5192 channel layout values.
5194 @item sample_rate, r
5195 Specifies the sample rate, and defaults to 44100.
5198 Set the number of samples per requested frames.
5202 @subsection Examples
5206 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
5208 anullsrc=r=48000:cl=4
5212 Do the same operation with a more obvious syntax:
5214 anullsrc=r=48000:cl=mono
5218 All the parameters need to be explicitly defined.
5222 Synthesize a voice utterance using the libflite library.
5224 To enable compilation of this filter you need to configure FFmpeg with
5225 @code{--enable-libflite}.
5227 Note that versions of the flite library prior to 2.0 are not thread-safe.
5229 The filter accepts the following options:
5234 If set to 1, list the names of the available voices and exit
5235 immediately. Default value is 0.
5238 Set the maximum number of samples per frame. Default value is 512.
5241 Set the filename containing the text to speak.
5244 Set the text to speak.
5247 Set the voice to use for the speech synthesis. Default value is
5248 @code{kal}. See also the @var{list_voices} option.
5251 @subsection Examples
5255 Read from file @file{speech.txt}, and synthesize the text using the
5256 standard flite voice:
5258 flite=textfile=speech.txt
5262 Read the specified text selecting the @code{slt} voice:
5264 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5268 Input text to ffmpeg:
5270 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5274 Make @file{ffplay} speak the specified text, using @code{flite} and
5275 the @code{lavfi} device:
5277 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
5281 For more information about libflite, check:
5282 @url{http://www.festvox.org/flite/}
5286 Generate a noise audio signal.
5288 The filter accepts the following options:
5291 @item sample_rate, r
5292 Specify the sample rate. Default value is 48000 Hz.
5295 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
5299 Specify the duration of the generated audio stream. Not specifying this option
5300 results in noise with an infinite length.
5302 @item color, colour, c
5303 Specify the color of noise. Available noise colors are white, pink, brown,
5304 blue and violet. Default color is white.
5307 Specify a value used to seed the PRNG.
5310 Set the number of samples per each output frame, default is 1024.
5313 @subsection Examples
5318 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
5320 anoisesrc=d=60:c=pink:r=44100:a=0.5
5326 Generate odd-tap Hilbert transform FIR coefficients.
5328 The resulting stream can be used with @ref{afir} filter for phase-shifting
5329 the signal by 90 degrees.
5331 This is used in many matrix coding schemes and for analytic signal generation.
5332 The process is often written as a multiplication by i (or j), the imaginary unit.
5334 The filter accepts the following options:
5338 @item sample_rate, s
5339 Set sample rate, default is 44100.
5342 Set length of FIR filter, default is 22051.
5345 Set number of samples per each frame.
5348 Set window function to be used when generating FIR coefficients.
5353 Generate an audio signal made of a sine wave with amplitude 1/8.
5355 The audio signal is bit-exact.
5357 The filter accepts the following options:
5362 Set the carrier frequency. Default is 440 Hz.
5364 @item beep_factor, b
5365 Enable a periodic beep every second with frequency @var{beep_factor} times
5366 the carrier frequency. Default is 0, meaning the beep is disabled.
5368 @item sample_rate, r
5369 Specify the sample rate, default is 44100.
5372 Specify the duration of the generated audio stream.
5374 @item samples_per_frame
5375 Set the number of samples per output frame.
5377 The expression can contain the following constants:
5381 The (sequential) number of the output audio frame, starting from 0.
5384 The PTS (Presentation TimeStamp) of the output audio frame,
5385 expressed in @var{TB} units.
5388 The PTS of the output audio frame, expressed in seconds.
5391 The timebase of the output audio frames.
5394 Default is @code{1024}.
5397 @subsection Examples
5402 Generate a simple 440 Hz sine wave:
5408 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
5412 sine=frequency=220:beep_factor=4:duration=5
5416 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
5419 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
5423 @c man end AUDIO SOURCES
5425 @chapter Audio Sinks
5426 @c man begin AUDIO SINKS
5428 Below is a description of the currently available audio sinks.
5430 @section abuffersink
5432 Buffer audio frames, and make them available to the end of filter chain.
5434 This sink is mainly intended for programmatic use, in particular
5435 through the interface defined in @file{libavfilter/buffersink.h}
5436 or the options system.
5438 It accepts a pointer to an AVABufferSinkContext structure, which
5439 defines the incoming buffers' formats, to be passed as the opaque
5440 parameter to @code{avfilter_init_filter} for initialization.
5443 Null audio sink; do absolutely nothing with the input audio. It is
5444 mainly useful as a template and for use in analysis / debugging
5447 @c man end AUDIO SINKS
5449 @chapter Video Filters
5450 @c man begin VIDEO FILTERS
5452 When you configure your FFmpeg build, you can disable any of the
5453 existing filters using @code{--disable-filters}.
5454 The configure output will show the video filters included in your
5457 Below is a description of the currently available video filters.
5459 @section alphaextract
5461 Extract the alpha component from the input as a grayscale video. This
5462 is especially useful with the @var{alphamerge} filter.
5466 Add or replace the alpha component of the primary input with the
5467 grayscale value of a second input. This is intended for use with
5468 @var{alphaextract} to allow the transmission or storage of frame
5469 sequences that have alpha in a format that doesn't support an alpha
5472 For example, to reconstruct full frames from a normal YUV-encoded video
5473 and a separate video created with @var{alphaextract}, you might use:
5475 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
5478 Since this filter is designed for reconstruction, it operates on frame
5479 sequences without considering timestamps, and terminates when either
5480 input reaches end of stream. This will cause problems if your encoding
5481 pipeline drops frames. If you're trying to apply an image as an
5482 overlay to a video stream, consider the @var{overlay} filter instead.
5486 Amplify differences between current pixel and pixels of adjacent frames in
5487 same pixel location.
5489 This filter accepts the following options:
5493 Set frame radius. Default is 2. Allowed range is from 1 to 63.
5494 For example radius of 3 will instruct filter to calculate average of 7 frames.
5497 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
5500 Set threshold for difference amplification. Any differrence greater or equal to
5501 this value will not alter source pixel. Default is 10.
5502 Allowed range is from 0 to 65535.
5505 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
5506 This option controls maximum possible value that will decrease source pixel value.
5509 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
5510 This option controls maximum possible value that will increase source pixel value.
5513 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
5518 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
5519 and libavformat to work. On the other hand, it is limited to ASS (Advanced
5520 Substation Alpha) subtitles files.
5522 This filter accepts the following option in addition to the common options from
5523 the @ref{subtitles} filter:
5527 Set the shaping engine
5529 Available values are:
5532 The default libass shaping engine, which is the best available.
5534 Fast, font-agnostic shaper that can do only substitutions
5536 Slower shaper using OpenType for substitutions and positioning
5539 The default is @code{auto}.
5543 Apply an Adaptive Temporal Averaging Denoiser to the video input.
5545 The filter accepts the following options:
5549 Set threshold A for 1st plane. Default is 0.02.
5550 Valid range is 0 to 0.3.
5553 Set threshold B for 1st plane. Default is 0.04.
5554 Valid range is 0 to 5.
5557 Set threshold A for 2nd plane. Default is 0.02.
5558 Valid range is 0 to 0.3.
5561 Set threshold B for 2nd plane. Default is 0.04.
5562 Valid range is 0 to 5.
5565 Set threshold A for 3rd plane. Default is 0.02.
5566 Valid range is 0 to 0.3.
5569 Set threshold B for 3rd plane. Default is 0.04.
5570 Valid range is 0 to 5.
5572 Threshold A is designed to react on abrupt changes in the input signal and
5573 threshold B is designed to react on continuous changes in the input signal.
5576 Set number of frames filter will use for averaging. Default is 9. Must be odd
5577 number in range [5, 129].
5580 Set what planes of frame filter will use for averaging. Default is all.
5585 Apply average blur filter.
5587 The filter accepts the following options:
5591 Set horizontal radius size.
5594 Set which planes to filter. By default all planes are filtered.
5597 Set vertical radius size, if zero it will be same as @code{sizeX}.
5598 Default is @code{0}.
5603 Compute the bounding box for the non-black pixels in the input frame
5606 This filter computes the bounding box containing all the pixels with a
5607 luminance value greater than the minimum allowed value.
5608 The parameters describing the bounding box are printed on the filter
5611 The filter accepts the following option:
5615 Set the minimal luminance value. Default is @code{16}.
5618 @section bitplanenoise
5620 Show and measure bit plane noise.
5622 The filter accepts the following options:
5626 Set which plane to analyze. Default is @code{1}.
5629 Filter out noisy pixels from @code{bitplane} set above.
5630 Default is disabled.
5633 @section blackdetect
5635 Detect video intervals that are (almost) completely black. Can be
5636 useful to detect chapter transitions, commercials, or invalid
5637 recordings. Output lines contains the time for the start, end and
5638 duration of the detected black interval expressed in seconds.
5640 In order to display the output lines, you need to set the loglevel at
5641 least to the AV_LOG_INFO value.
5643 The filter accepts the following options:
5646 @item black_min_duration, d
5647 Set the minimum detected black duration expressed in seconds. It must
5648 be a non-negative floating point number.
5650 Default value is 2.0.
5652 @item picture_black_ratio_th, pic_th
5653 Set the threshold for considering a picture "black".
5654 Express the minimum value for the ratio:
5656 @var{nb_black_pixels} / @var{nb_pixels}
5659 for which a picture is considered black.
5660 Default value is 0.98.
5662 @item pixel_black_th, pix_th
5663 Set the threshold for considering a pixel "black".
5665 The threshold expresses the maximum pixel luminance value for which a
5666 pixel is considered "black". The provided value is scaled according to
5667 the following equation:
5669 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
5672 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
5673 the input video format, the range is [0-255] for YUV full-range
5674 formats and [16-235] for YUV non full-range formats.
5676 Default value is 0.10.
5679 The following example sets the maximum pixel threshold to the minimum
5680 value, and detects only black intervals of 2 or more seconds:
5682 blackdetect=d=2:pix_th=0.00
5687 Detect frames that are (almost) completely black. Can be useful to
5688 detect chapter transitions or commercials. Output lines consist of
5689 the frame number of the detected frame, the percentage of blackness,
5690 the position in the file if known or -1 and the timestamp in seconds.
5692 In order to display the output lines, you need to set the loglevel at
5693 least to the AV_LOG_INFO value.
5695 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
5696 The value represents the percentage of pixels in the picture that
5697 are below the threshold value.
5699 It accepts the following parameters:
5704 The percentage of the pixels that have to be below the threshold; it defaults to
5707 @item threshold, thresh
5708 The threshold below which a pixel value is considered black; it defaults to
5713 @section blend, tblend
5715 Blend two video frames into each other.
5717 The @code{blend} filter takes two input streams and outputs one
5718 stream, the first input is the "top" layer and second input is
5719 "bottom" layer. By default, the output terminates when the longest input terminates.
5721 The @code{tblend} (time blend) filter takes two consecutive frames
5722 from one single stream, and outputs the result obtained by blending
5723 the new frame on top of the old frame.
5725 A description of the accepted options follows.
5733 Set blend mode for specific pixel component or all pixel components in case
5734 of @var{all_mode}. Default value is @code{normal}.
5736 Available values for component modes are:
5778 Set blend opacity for specific pixel component or all pixel components in case
5779 of @var{all_opacity}. Only used in combination with pixel component blend modes.
5786 Set blend expression for specific pixel component or all pixel components in case
5787 of @var{all_expr}. Note that related mode options will be ignored if those are set.
5789 The expressions can use the following variables:
5793 The sequential number of the filtered frame, starting from @code{0}.
5797 the coordinates of the current sample
5801 the width and height of currently filtered plane
5805 Width and height scale for the plane being filtered. It is the
5806 ratio between the dimensions of the current plane to the luma plane,
5807 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
5808 the luma plane and @code{0.5,0.5} for the chroma planes.
5811 Time of the current frame, expressed in seconds.
5814 Value of pixel component at current location for first video frame (top layer).
5817 Value of pixel component at current location for second video frame (bottom layer).
5821 The @code{blend} filter also supports the @ref{framesync} options.
5823 @subsection Examples
5827 Apply transition from bottom layer to top layer in first 10 seconds:
5829 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
5833 Apply linear horizontal transition from top layer to bottom layer:
5835 blend=all_expr='A*(X/W)+B*(1-X/W)'
5839 Apply 1x1 checkerboard effect:
5841 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
5845 Apply uncover left effect:
5847 blend=all_expr='if(gte(N*SW+X,W),A,B)'
5851 Apply uncover down effect:
5853 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
5857 Apply uncover up-left effect:
5859 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
5863 Split diagonally video and shows top and bottom layer on each side:
5865 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
5869 Display differences between the current and the previous frame:
5871 tblend=all_mode=grainextract
5877 Denoise frames using Block-Matching 3D algorithm.
5879 The filter accepts the following options.
5883 Set denoising strength. Default value is 1.
5884 Allowed range is from 0 to 999.9.
5885 The denoising algorith is very sensitive to sigma, so adjust it
5886 according to the source.
5889 Set local patch size. This sets dimensions in 2D.
5892 Set sliding step for processing blocks. Default value is 4.
5893 Allowed range is from 1 to 64.
5894 Smaller values allows processing more reference blocks and is slower.
5897 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
5898 When set to 1, no block matching is done. Larger values allows more blocks
5900 Allowed range is from 1 to 256.
5903 Set radius for search block matching. Default is 9.
5904 Allowed range is from 1 to INT32_MAX.
5907 Set step between two search locations for block matching. Default is 1.
5908 Allowed range is from 1 to 64. Smaller is slower.
5911 Set threshold of mean square error for block matching. Valid range is 0 to
5915 Set thresholding parameter for hard thresholding in 3D transformed domain.
5916 Larger values results in stronger hard-thresholding filtering in frequency
5920 Set filtering estimation mode. Can be @code{basic} or @code{final}.
5921 Default is @code{basic}.
5924 If enabled, filter will use 2nd stream for block matching.
5925 Default is disabled for @code{basic} value of @var{estim} option,
5926 and always enabled if value of @var{estim} is @code{final}.
5929 Set planes to filter. Default is all available except alpha.
5932 @subsection Examples
5936 Basic filtering with bm3d:
5938 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
5942 Same as above, but filtering only luma:
5944 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
5948 Same as above, but with both estimation modes:
5950 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
5954 Same as above, but prefilter with @ref{nlmeans} filter instead:
5956 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
5962 Apply a boxblur algorithm to the input video.
5964 It accepts the following parameters:
5968 @item luma_radius, lr
5969 @item luma_power, lp
5970 @item chroma_radius, cr
5971 @item chroma_power, cp
5972 @item alpha_radius, ar
5973 @item alpha_power, ap
5977 A description of the accepted options follows.
5980 @item luma_radius, lr
5981 @item chroma_radius, cr
5982 @item alpha_radius, ar
5983 Set an expression for the box radius in pixels used for blurring the
5984 corresponding input plane.
5986 The radius value must be a non-negative number, and must not be
5987 greater than the value of the expression @code{min(w,h)/2} for the
5988 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
5991 Default value for @option{luma_radius} is "2". If not specified,
5992 @option{chroma_radius} and @option{alpha_radius} default to the
5993 corresponding value set for @option{luma_radius}.
5995 The expressions can contain the following constants:
5999 The input width and height in pixels.
6003 The input chroma image width and height in pixels.
6007 The horizontal and vertical chroma subsample values. For example, for the
6008 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
6011 @item luma_power, lp
6012 @item chroma_power, cp
6013 @item alpha_power, ap
6014 Specify how many times the boxblur filter is applied to the
6015 corresponding plane.
6017 Default value for @option{luma_power} is 2. If not specified,
6018 @option{chroma_power} and @option{alpha_power} default to the
6019 corresponding value set for @option{luma_power}.
6021 A value of 0 will disable the effect.
6024 @subsection Examples
6028 Apply a boxblur filter with the luma, chroma, and alpha radii
6031 boxblur=luma_radius=2:luma_power=1
6036 Set the luma radius to 2, and alpha and chroma radius to 0:
6038 boxblur=2:1:cr=0:ar=0
6042 Set the luma and chroma radii to a fraction of the video dimension:
6044 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
6050 Deinterlace the input video ("bwdif" stands for "Bob Weaver
6051 Deinterlacing Filter").
6053 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
6054 interpolation algorithms.
6055 It accepts the following parameters:
6059 The interlacing mode to adopt. It accepts one of the following values:
6063 Output one frame for each frame.
6065 Output one frame for each field.
6068 The default value is @code{send_field}.
6071 The picture field parity assumed for the input interlaced video. It accepts one
6072 of the following values:
6076 Assume the top field is first.
6078 Assume the bottom field is first.
6080 Enable automatic detection of field parity.
6083 The default value is @code{auto}.
6084 If the interlacing is unknown or the decoder does not export this information,
6085 top field first will be assumed.
6088 Specify which frames to deinterlace. Accept one of the following
6093 Deinterlace all frames.
6095 Only deinterlace frames marked as interlaced.
6098 The default value is @code{all}.
6102 YUV colorspace color/chroma keying.
6104 The filter accepts the following options:
6108 The color which will be replaced with transparency.
6111 Similarity percentage with the key color.
6113 0.01 matches only the exact key color, while 1.0 matches everything.
6118 0.0 makes pixels either fully transparent, or not transparent at all.
6120 Higher values result in semi-transparent pixels, with a higher transparency
6121 the more similar the pixels color is to the key color.
6124 Signals that the color passed is already in YUV instead of RGB.
6126 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6127 This can be used to pass exact YUV values as hexadecimal numbers.
6130 @subsection Examples
6134 Make every green pixel in the input image transparent:
6136 ffmpeg -i input.png -vf chromakey=green out.png
6140 Overlay a greenscreen-video on top of a static black background.
6142 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
6148 Display CIE color diagram with pixels overlaid onto it.
6150 The filter accepts the following options:
6165 @item uhdtv, rec2020
6178 Set what gamuts to draw.
6180 See @code{system} option for available values.
6183 Set ciescope size, by default set to 512.
6186 Set intensity used to map input pixel values to CIE diagram.
6189 Set contrast used to draw tongue colors that are out of active color system gamut.
6192 Correct gamma displayed on scope, by default enabled.
6195 Show white point on CIE diagram, by default disabled.
6198 Set input gamma. Used only with XYZ input color space.
6203 Visualize information exported by some codecs.
6205 Some codecs can export information through frames using side-data or other
6206 means. For example, some MPEG based codecs export motion vectors through the
6207 @var{export_mvs} flag in the codec @option{flags2} option.
6209 The filter accepts the following option:
6213 Set motion vectors to visualize.
6215 Available flags for @var{mv} are:
6219 forward predicted MVs of P-frames
6221 forward predicted MVs of B-frames
6223 backward predicted MVs of B-frames
6227 Display quantization parameters using the chroma planes.
6230 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
6232 Available flags for @var{mv_type} are:
6236 forward predicted MVs
6238 backward predicted MVs
6241 @item frame_type, ft
6242 Set frame type to visualize motion vectors of.
6244 Available flags for @var{frame_type} are:
6248 intra-coded frames (I-frames)
6250 predicted frames (P-frames)
6252 bi-directionally predicted frames (B-frames)
6256 @subsection Examples
6260 Visualize forward predicted MVs of all frames using @command{ffplay}:
6262 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
6266 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
6268 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
6272 @section colorbalance
6273 Modify intensity of primary colors (red, green and blue) of input frames.
6275 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
6276 regions for the red-cyan, green-magenta or blue-yellow balance.
6278 A positive adjustment value shifts the balance towards the primary color, a negative
6279 value towards the complementary color.
6281 The filter accepts the following options:
6287 Adjust red, green and blue shadows (darkest pixels).
6292 Adjust red, green and blue midtones (medium pixels).
6297 Adjust red, green and blue highlights (brightest pixels).
6299 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
6302 @subsection Examples
6306 Add red color cast to shadows:
6313 RGB colorspace color keying.
6315 The filter accepts the following options:
6319 The color which will be replaced with transparency.
6322 Similarity percentage with the key color.
6324 0.01 matches only the exact key color, while 1.0 matches everything.
6329 0.0 makes pixels either fully transparent, or not transparent at all.
6331 Higher values result in semi-transparent pixels, with a higher transparency
6332 the more similar the pixels color is to the key color.
6335 @subsection Examples
6339 Make every green pixel in the input image transparent:
6341 ffmpeg -i input.png -vf colorkey=green out.png
6345 Overlay a greenscreen-video on top of a static background image.
6347 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
6351 @section colorlevels
6353 Adjust video input frames using levels.
6355 The filter accepts the following options:
6362 Adjust red, green, blue and alpha input black point.
6363 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
6369 Adjust red, green, blue and alpha input white point.
6370 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
6372 Input levels are used to lighten highlights (bright tones), darken shadows
6373 (dark tones), change the balance of bright and dark tones.
6379 Adjust red, green, blue and alpha output black point.
6380 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
6386 Adjust red, green, blue and alpha output white point.
6387 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
6389 Output levels allows manual selection of a constrained output level range.
6392 @subsection Examples
6396 Make video output darker:
6398 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
6404 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
6408 Make video output lighter:
6410 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
6414 Increase brightness:
6416 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
6420 @section colorchannelmixer
6422 Adjust video input frames by re-mixing color channels.
6424 This filter modifies a color channel by adding the values associated to
6425 the other channels of the same pixels. For example if the value to
6426 modify is red, the output value will be:
6428 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
6431 The filter accepts the following options:
6438 Adjust contribution of input red, green, blue and alpha channels for output red channel.
6439 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
6445 Adjust contribution of input red, green, blue and alpha channels for output green channel.
6446 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
6452 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
6453 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
6459 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
6460 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
6462 Allowed ranges for options are @code{[-2.0, 2.0]}.
6465 @subsection Examples
6469 Convert source to grayscale:
6471 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
6474 Simulate sepia tones:
6476 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
6480 @section colormatrix
6482 Convert color matrix.
6484 The filter accepts the following options:
6489 Specify the source and destination color matrix. Both values must be
6492 The accepted values are:
6520 For example to convert from BT.601 to SMPTE-240M, use the command:
6522 colormatrix=bt601:smpte240m
6527 Convert colorspace, transfer characteristics or color primaries.
6528 Input video needs to have an even size.
6530 The filter accepts the following options:
6535 Specify all color properties at once.
6537 The accepted values are:
6567 Specify output colorspace.
6569 The accepted values are:
6578 BT.470BG or BT.601-6 625
6581 SMPTE-170M or BT.601-6 525
6590 BT.2020 with non-constant luminance
6596 Specify output transfer characteristics.
6598 The accepted values are:
6610 Constant gamma of 2.2
6613 Constant gamma of 2.8
6616 SMPTE-170M, BT.601-6 625 or BT.601-6 525
6634 BT.2020 for 10-bits content
6637 BT.2020 for 12-bits content
6643 Specify output color primaries.
6645 The accepted values are:
6654 BT.470BG or BT.601-6 625
6657 SMPTE-170M or BT.601-6 525
6681 Specify output color range.
6683 The accepted values are:
6686 TV (restricted) range
6689 MPEG (restricted) range
6700 Specify output color format.
6702 The accepted values are:
6705 YUV 4:2:0 planar 8-bits
6708 YUV 4:2:0 planar 10-bits
6711 YUV 4:2:0 planar 12-bits
6714 YUV 4:2:2 planar 8-bits
6717 YUV 4:2:2 planar 10-bits
6720 YUV 4:2:2 planar 12-bits
6723 YUV 4:4:4 planar 8-bits
6726 YUV 4:4:4 planar 10-bits
6729 YUV 4:4:4 planar 12-bits
6734 Do a fast conversion, which skips gamma/primary correction. This will take
6735 significantly less CPU, but will be mathematically incorrect. To get output
6736 compatible with that produced by the colormatrix filter, use fast=1.
6739 Specify dithering mode.
6741 The accepted values are:
6747 Floyd-Steinberg dithering
6751 Whitepoint adaptation mode.
6753 The accepted values are:
6756 Bradford whitepoint adaptation
6759 von Kries whitepoint adaptation
6762 identity whitepoint adaptation (i.e. no whitepoint adaptation)
6766 Override all input properties at once. Same accepted values as @ref{all}.
6769 Override input colorspace. Same accepted values as @ref{space}.
6772 Override input color primaries. Same accepted values as @ref{primaries}.
6775 Override input transfer characteristics. Same accepted values as @ref{trc}.
6778 Override input color range. Same accepted values as @ref{range}.
6782 The filter converts the transfer characteristics, color space and color
6783 primaries to the specified user values. The output value, if not specified,
6784 is set to a default value based on the "all" property. If that property is
6785 also not specified, the filter will log an error. The output color range and
6786 format default to the same value as the input color range and format. The
6787 input transfer characteristics, color space, color primaries and color range
6788 should be set on the input data. If any of these are missing, the filter will
6789 log an error and no conversion will take place.
6791 For example to convert the input to SMPTE-240M, use the command:
6793 colorspace=smpte240m
6796 @section convolution
6798 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
6800 The filter accepts the following options:
6807 Set matrix for each plane.
6808 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
6809 and from 1 to 49 odd number of signed integers in @var{row} mode.
6815 Set multiplier for calculated value for each plane.
6816 If unset or 0, it will be sum of all matrix elements.
6822 Set bias for each plane. This value is added to the result of the multiplication.
6823 Useful for making the overall image brighter or darker. Default is 0.0.
6829 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
6830 Default is @var{square}.
6833 @subsection Examples
6839 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"
6845 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"
6851 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"
6857 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"
6861 Apply laplacian edge detector which includes diagonals:
6863 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"
6869 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"
6875 Apply 2D convolution of video stream in frequency domain using second stream
6878 The filter accepts the following options:
6882 Set which planes to process.
6885 Set which impulse video frames will be processed, can be @var{first}
6886 or @var{all}. Default is @var{all}.
6889 The @code{convolve} filter also supports the @ref{framesync} options.
6893 Copy the input video source unchanged to the output. This is mainly useful for
6898 Video filtering on GPU using Apple's CoreImage API on OSX.
6900 Hardware acceleration is based on an OpenGL context. Usually, this means it is
6901 processed by video hardware. However, software-based OpenGL implementations
6902 exist which means there is no guarantee for hardware processing. It depends on
6905 There are many filters and image generators provided by Apple that come with a
6906 large variety of options. The filter has to be referenced by its name along
6909 The coreimage filter accepts the following options:
6912 List all available filters and generators along with all their respective
6913 options as well as possible minimum and maximum values along with the default
6920 Specify all filters by their respective name and options.
6921 Use @var{list_filters} to determine all valid filter names and options.
6922 Numerical options are specified by a float value and are automatically clamped
6923 to their respective value range. Vector and color options have to be specified
6924 by a list of space separated float values. Character escaping has to be done.
6925 A special option name @code{default} is available to use default options for a
6928 It is required to specify either @code{default} or at least one of the filter options.
6929 All omitted options are used with their default values.
6930 The syntax of the filter string is as follows:
6932 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
6936 Specify a rectangle where the output of the filter chain is copied into the
6937 input image. It is given by a list of space separated float values:
6939 output_rect=x\ y\ width\ height
6941 If not given, the output rectangle equals the dimensions of the input image.
6942 The output rectangle is automatically cropped at the borders of the input
6943 image. Negative values are valid for each component.
6945 output_rect=25\ 25\ 100\ 100
6949 Several filters can be chained for successive processing without GPU-HOST
6950 transfers allowing for fast processing of complex filter chains.
6951 Currently, only filters with zero (generators) or exactly one (filters) input
6952 image and one output image are supported. Also, transition filters are not yet
6955 Some filters generate output images with additional padding depending on the
6956 respective filter kernel. The padding is automatically removed to ensure the
6957 filter output has the same size as the input image.
6959 For image generators, the size of the output image is determined by the
6960 previous output image of the filter chain or the input image of the whole
6961 filterchain, respectively. The generators do not use the pixel information of
6962 this image to generate their output. However, the generated output is
6963 blended onto this image, resulting in partial or complete coverage of the
6966 The @ref{coreimagesrc} video source can be used for generating input images
6967 which are directly fed into the filter chain. By using it, providing input
6968 images by another video source or an input video is not required.
6970 @subsection Examples
6975 List all filters available:
6977 coreimage=list_filters=true
6981 Use the CIBoxBlur filter with default options to blur an image:
6983 coreimage=filter=CIBoxBlur@@default
6987 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
6988 its center at 100x100 and a radius of 50 pixels:
6990 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
6994 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
6995 given as complete and escaped command-line for Apple's standard bash shell:
6997 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
7003 Crop the input video to given dimensions.
7005 It accepts the following parameters:
7009 The width of the output video. It defaults to @code{iw}.
7010 This expression is evaluated only once during the filter
7011 configuration, or when the @samp{w} or @samp{out_w} command is sent.
7014 The height of the output video. It defaults to @code{ih}.
7015 This expression is evaluated only once during the filter
7016 configuration, or when the @samp{h} or @samp{out_h} command is sent.
7019 The horizontal position, in the input video, of the left edge of the output
7020 video. It defaults to @code{(in_w-out_w)/2}.
7021 This expression is evaluated per-frame.
7024 The vertical position, in the input video, of the top edge of the output video.
7025 It defaults to @code{(in_h-out_h)/2}.
7026 This expression is evaluated per-frame.
7029 If set to 1 will force the output display aspect ratio
7030 to be the same of the input, by changing the output sample aspect
7031 ratio. It defaults to 0.
7034 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
7035 width/height/x/y as specified and will not be rounded to nearest smaller value.
7039 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
7040 expressions containing the following constants:
7045 The computed values for @var{x} and @var{y}. They are evaluated for
7050 The input width and height.
7054 These are the same as @var{in_w} and @var{in_h}.
7058 The output (cropped) width and height.
7062 These are the same as @var{out_w} and @var{out_h}.
7065 same as @var{iw} / @var{ih}
7068 input sample aspect ratio
7071 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
7075 horizontal and vertical chroma subsample values. For example for the
7076 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
7079 The number of the input frame, starting from 0.
7082 the position in the file of the input frame, NAN if unknown
7085 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
7089 The expression for @var{out_w} may depend on the value of @var{out_h},
7090 and the expression for @var{out_h} may depend on @var{out_w}, but they
7091 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
7092 evaluated after @var{out_w} and @var{out_h}.
7094 The @var{x} and @var{y} parameters specify the expressions for the
7095 position of the top-left corner of the output (non-cropped) area. They
7096 are evaluated for each frame. If the evaluated value is not valid, it
7097 is approximated to the nearest valid value.
7099 The expression for @var{x} may depend on @var{y}, and the expression
7100 for @var{y} may depend on @var{x}.
7102 @subsection Examples
7106 Crop area with size 100x100 at position (12,34).
7111 Using named options, the example above becomes:
7113 crop=w=100:h=100:x=12:y=34
7117 Crop the central input area with size 100x100:
7123 Crop the central input area with size 2/3 of the input video:
7125 crop=2/3*in_w:2/3*in_h
7129 Crop the input video central square:
7136 Delimit the rectangle with the top-left corner placed at position
7137 100:100 and the right-bottom corner corresponding to the right-bottom
7138 corner of the input image.
7140 crop=in_w-100:in_h-100:100:100
7144 Crop 10 pixels from the left and right borders, and 20 pixels from
7145 the top and bottom borders
7147 crop=in_w-2*10:in_h-2*20
7151 Keep only the bottom right quarter of the input image:
7153 crop=in_w/2:in_h/2:in_w/2:in_h/2
7157 Crop height for getting Greek harmony:
7159 crop=in_w:1/PHI*in_w
7163 Apply trembling effect:
7165 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)
7169 Apply erratic camera effect depending on timestamp:
7171 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)"
7175 Set x depending on the value of y:
7177 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
7181 @subsection Commands
7183 This filter supports the following commands:
7189 Set width/height of the output video and the horizontal/vertical position
7191 The command accepts the same syntax of the corresponding option.
7193 If the specified expression is not valid, it is kept at its current
7199 Auto-detect the crop size.
7201 It calculates the necessary cropping parameters and prints the
7202 recommended parameters via the logging system. The detected dimensions
7203 correspond to the non-black area of the input video.
7205 It accepts the following parameters:
7210 Set higher black value threshold, which can be optionally specified
7211 from nothing (0) to everything (255 for 8-bit based formats). An intensity
7212 value greater to the set value is considered non-black. It defaults to 24.
7213 You can also specify a value between 0.0 and 1.0 which will be scaled depending
7214 on the bitdepth of the pixel format.
7217 The value which the width/height should be divisible by. It defaults to
7218 16. The offset is automatically adjusted to center the video. Use 2 to
7219 get only even dimensions (needed for 4:2:2 video). 16 is best when
7220 encoding to most video codecs.
7222 @item reset_count, reset
7223 Set the counter that determines after how many frames cropdetect will
7224 reset the previously detected largest video area and start over to
7225 detect the current optimal crop area. Default value is 0.
7227 This can be useful when channel logos distort the video area. 0
7228 indicates 'never reset', and returns the largest area encountered during
7235 Delay video filtering until a given wallclock timestamp. The filter first
7236 passes on @option{preroll} amount of frames, then it buffers at most
7237 @option{buffer} amount of frames and waits for the cue. After reaching the cue
7238 it forwards the buffered frames and also any subsequent frames coming in its
7241 The filter can be used synchronize the output of multiple ffmpeg processes for
7242 realtime output devices like decklink. By putting the delay in the filtering
7243 chain and pre-buffering frames the process can pass on data to output almost
7244 immediately after the target wallclock timestamp is reached.
7246 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
7252 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
7255 The duration of content to pass on as preroll expressed in seconds. Default is 0.
7258 The maximum duration of content to buffer before waiting for the cue expressed
7259 in seconds. Default is 0.
7266 Apply color adjustments using curves.
7268 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
7269 component (red, green and blue) has its values defined by @var{N} key points
7270 tied from each other using a smooth curve. The x-axis represents the pixel
7271 values from the input frame, and the y-axis the new pixel values to be set for
7274 By default, a component curve is defined by the two points @var{(0;0)} and
7275 @var{(1;1)}. This creates a straight line where each original pixel value is
7276 "adjusted" to its own value, which means no change to the image.
7278 The filter allows you to redefine these two points and add some more. A new
7279 curve (using a natural cubic spline interpolation) will be define to pass
7280 smoothly through all these new coordinates. The new defined points needs to be
7281 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
7282 be in the @var{[0;1]} interval. If the computed curves happened to go outside
7283 the vector spaces, the values will be clipped accordingly.
7285 The filter accepts the following options:
7289 Select one of the available color presets. This option can be used in addition
7290 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
7291 options takes priority on the preset values.
7292 Available presets are:
7295 @item color_negative
7298 @item increase_contrast
7300 @item linear_contrast
7301 @item medium_contrast
7303 @item strong_contrast
7306 Default is @code{none}.
7308 Set the master key points. These points will define a second pass mapping. It
7309 is sometimes called a "luminance" or "value" mapping. It can be used with
7310 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
7311 post-processing LUT.
7313 Set the key points for the red component.
7315 Set the key points for the green component.
7317 Set the key points for the blue component.
7319 Set the key points for all components (not including master).
7320 Can be used in addition to the other key points component
7321 options. In this case, the unset component(s) will fallback on this
7322 @option{all} setting.
7324 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
7326 Save Gnuplot script of the curves in specified file.
7329 To avoid some filtergraph syntax conflicts, each key points list need to be
7330 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
7332 @subsection Examples
7336 Increase slightly the middle level of blue:
7338 curves=blue='0/0 0.5/0.58 1/1'
7344 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'
7346 Here we obtain the following coordinates for each components:
7349 @code{(0;0.11) (0.42;0.51) (1;0.95)}
7351 @code{(0;0) (0.50;0.48) (1;1)}
7353 @code{(0;0.22) (0.49;0.44) (1;0.80)}
7357 The previous example can also be achieved with the associated built-in preset:
7359 curves=preset=vintage
7369 Use a Photoshop preset and redefine the points of the green component:
7371 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
7375 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
7376 and @command{gnuplot}:
7378 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
7379 gnuplot -p /tmp/curves.plt
7385 Video data analysis filter.
7387 This filter shows hexadecimal pixel values of part of video.
7389 The filter accepts the following options:
7393 Set output video size.
7396 Set x offset from where to pick pixels.
7399 Set y offset from where to pick pixels.
7402 Set scope mode, can be one of the following:
7405 Draw hexadecimal pixel values with white color on black background.
7408 Draw hexadecimal pixel values with input video pixel color on black
7412 Draw hexadecimal pixel values on color background picked from input video,
7413 the text color is picked in such way so its always visible.
7417 Draw rows and columns numbers on left and top of video.
7420 Set background opacity.
7425 Denoise frames using 2D DCT (frequency domain filtering).
7427 This filter is not designed for real time.
7429 The filter accepts the following options:
7433 Set the noise sigma constant.
7435 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
7436 coefficient (absolute value) below this threshold with be dropped.
7438 If you need a more advanced filtering, see @option{expr}.
7440 Default is @code{0}.
7443 Set number overlapping pixels for each block. Since the filter can be slow, you
7444 may want to reduce this value, at the cost of a less effective filter and the
7445 risk of various artefacts.
7447 If the overlapping value doesn't permit processing the whole input width or
7448 height, a warning will be displayed and according borders won't be denoised.
7450 Default value is @var{blocksize}-1, which is the best possible setting.
7453 Set the coefficient factor expression.
7455 For each coefficient of a DCT block, this expression will be evaluated as a
7456 multiplier value for the coefficient.
7458 If this is option is set, the @option{sigma} option will be ignored.
7460 The absolute value of the coefficient can be accessed through the @var{c}
7464 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
7465 @var{blocksize}, which is the width and height of the processed blocks.
7467 The default value is @var{3} (8x8) and can be raised to @var{4} for a
7468 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
7469 on the speed processing. Also, a larger block size does not necessarily means a
7473 @subsection Examples
7475 Apply a denoise with a @option{sigma} of @code{4.5}:
7480 The same operation can be achieved using the expression system:
7482 dctdnoiz=e='gte(c, 4.5*3)'
7485 Violent denoise using a block size of @code{16x16}:
7492 Remove banding artifacts from input video.
7493 It works by replacing banded pixels with average value of referenced pixels.
7495 The filter accepts the following options:
7502 Set banding detection threshold for each plane. Default is 0.02.
7503 Valid range is 0.00003 to 0.5.
7504 If difference between current pixel and reference pixel is less than threshold,
7505 it will be considered as banded.
7508 Banding detection range in pixels. Default is 16. If positive, random number
7509 in range 0 to set value will be used. If negative, exact absolute value
7511 The range defines square of four pixels around current pixel.
7514 Set direction in radians from which four pixel will be compared. If positive,
7515 random direction from 0 to set direction will be picked. If negative, exact of
7516 absolute value will be picked. For example direction 0, -PI or -2*PI radians
7517 will pick only pixels on same row and -PI/2 will pick only pixels on same
7521 If enabled, current pixel is compared with average value of all four
7522 surrounding pixels. The default is enabled. If disabled current pixel is
7523 compared with all four surrounding pixels. The pixel is considered banded
7524 if only all four differences with surrounding pixels are less than threshold.
7527 If enabled, current pixel is changed if and only if all pixel components are banded,
7528 e.g. banding detection threshold is triggered for all color components.
7529 The default is disabled.
7534 Remove blocking artifacts from input video.
7536 The filter accepts the following options:
7540 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
7541 This controls what kind of deblocking is applied.
7544 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
7550 Set blocking detection thresholds. Allowed range is 0 to 1.
7551 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
7552 Using higher threshold gives more deblocking strength.
7553 Setting @var{alpha} controls threshold detection at exact edge of block.
7554 Remaining options controls threshold detection near the edge. Each one for
7555 below/above or left/right. Setting any of those to @var{0} disables
7559 Set planes to filter. Default is to filter all available planes.
7562 @subsection Examples
7566 Deblock using weak filter and block size of 4 pixels.
7568 deblock=filter=weak:block=4
7572 Deblock using strong filter, block size of 4 pixels and custom thresholds for
7573 deblocking more edges.
7575 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
7579 Similar as above, but filter only first plane.
7581 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
7585 Similar as above, but filter only second and third plane.
7587 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
7594 Drop duplicated frames at regular intervals.
7596 The filter accepts the following options:
7600 Set the number of frames from which one will be dropped. Setting this to
7601 @var{N} means one frame in every batch of @var{N} frames will be dropped.
7602 Default is @code{5}.
7605 Set the threshold for duplicate detection. If the difference metric for a frame
7606 is less than or equal to this value, then it is declared as duplicate. Default
7610 Set scene change threshold. Default is @code{15}.
7614 Set the size of the x and y-axis blocks used during metric calculations.
7615 Larger blocks give better noise suppression, but also give worse detection of
7616 small movements. Must be a power of two. Default is @code{32}.
7619 Mark main input as a pre-processed input and activate clean source input
7620 stream. This allows the input to be pre-processed with various filters to help
7621 the metrics calculation while keeping the frame selection lossless. When set to
7622 @code{1}, the first stream is for the pre-processed input, and the second
7623 stream is the clean source from where the kept frames are chosen. Default is
7627 Set whether or not chroma is considered in the metric calculations. Default is
7633 Apply 2D deconvolution of video stream in frequency domain using second stream
7636 The filter accepts the following options:
7640 Set which planes to process.
7643 Set which impulse video frames will be processed, can be @var{first}
7644 or @var{all}. Default is @var{all}.
7647 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
7648 and height are not same and not power of 2 or if stream prior to convolving
7652 The @code{deconvolve} filter also supports the @ref{framesync} options.
7656 Apply deflate effect to the video.
7658 This filter replaces the pixel by the local(3x3) average by taking into account
7659 only values lower than the pixel.
7661 It accepts the following options:
7668 Limit the maximum change for each plane, default is 65535.
7669 If 0, plane will remain unchanged.
7674 Remove temporal frame luminance variations.
7676 It accepts the following options:
7680 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
7683 Set averaging mode to smooth temporal luminance variations.
7685 Available values are:
7710 Do not actually modify frame. Useful when one only wants metadata.
7715 Remove judder produced by partially interlaced telecined content.
7717 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
7718 source was partially telecined content then the output of @code{pullup,dejudder}
7719 will have a variable frame rate. May change the recorded frame rate of the
7720 container. Aside from that change, this filter will not affect constant frame
7723 The option available in this filter is:
7727 Specify the length of the window over which the judder repeats.
7729 Accepts any integer greater than 1. Useful values are:
7733 If the original was telecined from 24 to 30 fps (Film to NTSC).
7736 If the original was telecined from 25 to 30 fps (PAL to NTSC).
7739 If a mixture of the two.
7742 The default is @samp{4}.
7747 Suppress a TV station logo by a simple interpolation of the surrounding
7748 pixels. Just set a rectangle covering the logo and watch it disappear
7749 (and sometimes something even uglier appear - your mileage may vary).
7751 It accepts the following parameters:
7756 Specify the top left corner coordinates of the logo. They must be
7761 Specify the width and height of the logo to clear. They must be
7765 Specify the thickness of the fuzzy edge of the rectangle (added to
7766 @var{w} and @var{h}). The default value is 1. This option is
7767 deprecated, setting higher values should no longer be necessary and
7771 When set to 1, a green rectangle is drawn on the screen to simplify
7772 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
7773 The default value is 0.
7775 The rectangle is drawn on the outermost pixels which will be (partly)
7776 replaced with interpolated values. The values of the next pixels
7777 immediately outside this rectangle in each direction will be used to
7778 compute the interpolated pixel values inside the rectangle.
7782 @subsection Examples
7786 Set a rectangle covering the area with top left corner coordinates 0,0
7787 and size 100x77, and a band of size 10:
7789 delogo=x=0:y=0:w=100:h=77:band=10
7796 Attempt to fix small changes in horizontal and/or vertical shift. This
7797 filter helps remove camera shake from hand-holding a camera, bumping a
7798 tripod, moving on a vehicle, etc.
7800 The filter accepts the following options:
7808 Specify a rectangular area where to limit the search for motion
7810 If desired the search for motion vectors can be limited to a
7811 rectangular area of the frame defined by its top left corner, width
7812 and height. These parameters have the same meaning as the drawbox
7813 filter which can be used to visualise the position of the bounding
7816 This is useful when simultaneous movement of subjects within the frame
7817 might be confused for camera motion by the motion vector search.
7819 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
7820 then the full frame is used. This allows later options to be set
7821 without specifying the bounding box for the motion vector search.
7823 Default - search the whole frame.
7827 Specify the maximum extent of movement in x and y directions in the
7828 range 0-64 pixels. Default 16.
7831 Specify how to generate pixels to fill blanks at the edge of the
7832 frame. Available values are:
7835 Fill zeroes at blank locations
7837 Original image at blank locations
7839 Extruded edge value at blank locations
7841 Mirrored edge at blank locations
7843 Default value is @samp{mirror}.
7846 Specify the blocksize to use for motion search. Range 4-128 pixels,
7850 Specify the contrast threshold for blocks. Only blocks with more than
7851 the specified contrast (difference between darkest and lightest
7852 pixels) will be considered. Range 1-255, default 125.
7855 Specify the search strategy. Available values are:
7858 Set exhaustive search
7860 Set less exhaustive search.
7862 Default value is @samp{exhaustive}.
7865 If set then a detailed log of the motion search is written to the
7872 Remove unwanted contamination of foreground colors, caused by reflected color of
7873 greenscreen or bluescreen.
7875 This filter accepts the following options:
7879 Set what type of despill to use.
7882 Set how spillmap will be generated.
7885 Set how much to get rid of still remaining spill.
7888 Controls amount of red in spill area.
7891 Controls amount of green in spill area.
7892 Should be -1 for greenscreen.
7895 Controls amount of blue in spill area.
7896 Should be -1 for bluescreen.
7899 Controls brightness of spill area, preserving colors.
7902 Modify alpha from generated spillmap.
7907 Apply an exact inverse of the telecine operation. It requires a predefined
7908 pattern specified using the pattern option which must be the same as that passed
7909 to the telecine filter.
7911 This filter accepts the following options:
7920 The default value is @code{top}.
7924 A string of numbers representing the pulldown pattern you wish to apply.
7925 The default value is @code{23}.
7928 A number representing position of the first frame with respect to the telecine
7929 pattern. This is to be used if the stream is cut. The default value is @code{0}.
7934 Apply dilation effect to the video.
7936 This filter replaces the pixel by the local(3x3) maximum.
7938 It accepts the following options:
7945 Limit the maximum change for each plane, default is 65535.
7946 If 0, plane will remain unchanged.
7949 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
7952 Flags to local 3x3 coordinates maps like this:
7961 Displace pixels as indicated by second and third input stream.
7963 It takes three input streams and outputs one stream, the first input is the
7964 source, and second and third input are displacement maps.
7966 The second input specifies how much to displace pixels along the
7967 x-axis, while the third input specifies how much to displace pixels
7969 If one of displacement map streams terminates, last frame from that
7970 displacement map will be used.
7972 Note that once generated, displacements maps can be reused over and over again.
7974 A description of the accepted options follows.
7978 Set displace behavior for pixels that are out of range.
7980 Available values are:
7983 Missing pixels are replaced by black pixels.
7986 Adjacent pixels will spread out to replace missing pixels.
7989 Out of range pixels are wrapped so they point to pixels of other side.
7992 Out of range pixels will be replaced with mirrored pixels.
7994 Default is @samp{smear}.
7998 @subsection Examples
8002 Add ripple effect to rgb input of video size hd720:
8004 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
8008 Add wave effect to rgb input of video size hd720:
8010 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
8016 Draw a colored box on the input image.
8018 It accepts the following parameters:
8023 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
8027 The expressions which specify the width and height of the box; if 0 they are interpreted as
8028 the input width and height. It defaults to 0.
8031 Specify the color of the box to write. For the general syntax of this option,
8032 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8033 value @code{invert} is used, the box edge color is the same as the
8034 video with inverted luma.
8037 The expression which sets the thickness of the box edge.
8038 A value of @code{fill} will create a filled box. Default value is @code{3}.
8040 See below for the list of accepted constants.
8043 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
8044 will overwrite the video's color and alpha pixels.
8045 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
8048 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8049 following constants:
8053 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8057 horizontal and vertical chroma subsample values. For example for the
8058 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8062 The input width and height.
8065 The input sample aspect ratio.
8069 The x and y offset coordinates where the box is drawn.
8073 The width and height of the drawn box.
8076 The thickness of the drawn box.
8078 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8079 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
8083 @subsection Examples
8087 Draw a black box around the edge of the input image:
8093 Draw a box with color red and an opacity of 50%:
8095 drawbox=10:20:200:60:red@@0.5
8098 The previous example can be specified as:
8100 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
8104 Fill the box with pink color:
8106 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
8110 Draw a 2-pixel red 2.40:1 mask:
8112 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
8118 Draw a grid on the input image.
8120 It accepts the following parameters:
8125 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
8129 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
8130 input width and height, respectively, minus @code{thickness}, so image gets
8131 framed. Default to 0.
8134 Specify the color of the grid. For the general syntax of this option,
8135 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8136 value @code{invert} is used, the grid color is the same as the
8137 video with inverted luma.
8140 The expression which sets the thickness of the grid line. Default value is @code{1}.
8142 See below for the list of accepted constants.
8145 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
8146 will overwrite the video's color and alpha pixels.
8147 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
8150 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8151 following constants:
8155 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8159 horizontal and vertical chroma subsample values. For example for the
8160 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8164 The input grid cell width and height.
8167 The input sample aspect ratio.
8171 The x and y coordinates of some point of grid intersection (meant to configure offset).
8175 The width and height of the drawn cell.
8178 The thickness of the drawn cell.
8180 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8181 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
8185 @subsection Examples
8189 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
8191 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
8195 Draw a white 3x3 grid with an opacity of 50%:
8197 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
8204 Draw a text string or text from a specified file on top of a video, using the
8205 libfreetype library.
8207 To enable compilation of this filter, you need to configure FFmpeg with
8208 @code{--enable-libfreetype}.
8209 To enable default font fallback and the @var{font} option you need to
8210 configure FFmpeg with @code{--enable-libfontconfig}.
8211 To enable the @var{text_shaping} option, you need to configure FFmpeg with
8212 @code{--enable-libfribidi}.
8216 It accepts the following parameters:
8221 Used to draw a box around text using the background color.
8222 The value must be either 1 (enable) or 0 (disable).
8223 The default value of @var{box} is 0.
8226 Set the width of the border to be drawn around the box using @var{boxcolor}.
8227 The default value of @var{boxborderw} is 0.
8230 The color to be used for drawing box around text. For the syntax of this
8231 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8233 The default value of @var{boxcolor} is "white".
8236 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
8237 The default value of @var{line_spacing} is 0.
8240 Set the width of the border to be drawn around the text using @var{bordercolor}.
8241 The default value of @var{borderw} is 0.
8244 Set the color to be used for drawing border around text. For the syntax of this
8245 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8247 The default value of @var{bordercolor} is "black".
8250 Select how the @var{text} is expanded. Can be either @code{none},
8251 @code{strftime} (deprecated) or
8252 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
8256 Set a start time for the count. Value is in microseconds. Only applied
8257 in the deprecated strftime expansion mode. To emulate in normal expansion
8258 mode use the @code{pts} function, supplying the start time (in seconds)
8259 as the second argument.
8262 If true, check and fix text coords to avoid clipping.
8265 The color to be used for drawing fonts. For the syntax of this option, check
8266 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8268 The default value of @var{fontcolor} is "black".
8270 @item fontcolor_expr
8271 String which is expanded the same way as @var{text} to obtain dynamic
8272 @var{fontcolor} value. By default this option has empty value and is not
8273 processed. When this option is set, it overrides @var{fontcolor} option.
8276 The font family to be used for drawing text. By default Sans.
8279 The font file to be used for drawing text. The path must be included.
8280 This parameter is mandatory if the fontconfig support is disabled.
8283 Draw the text applying alpha blending. The value can
8284 be a number between 0.0 and 1.0.
8285 The expression accepts the same variables @var{x, y} as well.
8286 The default value is 1.
8287 Please see @var{fontcolor_expr}.
8290 The font size to be used for drawing text.
8291 The default value of @var{fontsize} is 16.
8294 If set to 1, attempt to shape the text (for example, reverse the order of
8295 right-to-left text and join Arabic characters) before drawing it.
8296 Otherwise, just draw the text exactly as given.
8297 By default 1 (if supported).
8300 The flags to be used for loading the fonts.
8302 The flags map the corresponding flags supported by libfreetype, and are
8303 a combination of the following values:
8310 @item vertical_layout
8311 @item force_autohint
8314 @item ignore_global_advance_width
8316 @item ignore_transform
8322 Default value is "default".
8324 For more information consult the documentation for the FT_LOAD_*
8328 The color to be used for drawing a shadow behind the drawn text. For the
8329 syntax of this option, check the @ref{color syntax,,"Color" section in the
8330 ffmpeg-utils manual,ffmpeg-utils}.
8332 The default value of @var{shadowcolor} is "black".
8336 The x and y offsets for the text shadow position with respect to the
8337 position of the text. They can be either positive or negative
8338 values. The default value for both is "0".
8341 The starting frame number for the n/frame_num variable. The default value
8345 The size in number of spaces to use for rendering the tab.
8349 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
8350 format. It can be used with or without text parameter. @var{timecode_rate}
8351 option must be specified.
8353 @item timecode_rate, rate, r
8354 Set the timecode frame rate (timecode only). Value will be rounded to nearest
8355 integer. Minimum value is "1".
8356 Drop-frame timecode is supported for frame rates 30 & 60.
8359 If set to 1, the output of the timecode option will wrap around at 24 hours.
8360 Default is 0 (disabled).
8363 The text string to be drawn. The text must be a sequence of UTF-8
8365 This parameter is mandatory if no file is specified with the parameter
8369 A text file containing text to be drawn. The text must be a sequence
8370 of UTF-8 encoded characters.
8372 This parameter is mandatory if no text string is specified with the
8373 parameter @var{text}.
8375 If both @var{text} and @var{textfile} are specified, an error is thrown.
8378 If set to 1, the @var{textfile} will be reloaded before each frame.
8379 Be sure to update it atomically, or it may be read partially, or even fail.
8383 The expressions which specify the offsets where text will be drawn
8384 within the video frame. They are relative to the top/left border of the
8387 The default value of @var{x} and @var{y} is "0".
8389 See below for the list of accepted constants and functions.
8392 The parameters for @var{x} and @var{y} are expressions containing the
8393 following constants and functions:
8397 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
8401 horizontal and vertical chroma subsample values. For example for the
8402 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8405 the height of each text line
8413 @item max_glyph_a, ascent
8414 the maximum distance from the baseline to the highest/upper grid
8415 coordinate used to place a glyph outline point, for all the rendered
8417 It is a positive value, due to the grid's orientation with the Y axis
8420 @item max_glyph_d, descent
8421 the maximum distance from the baseline to the lowest grid coordinate
8422 used to place a glyph outline point, for all the rendered glyphs.
8423 This is a negative value, due to the grid's orientation, with the Y axis
8427 maximum glyph height, that is the maximum height for all the glyphs
8428 contained in the rendered text, it is equivalent to @var{ascent} -
8432 maximum glyph width, that is the maximum width for all the glyphs
8433 contained in the rendered text
8436 the number of input frame, starting from 0
8438 @item rand(min, max)
8439 return a random number included between @var{min} and @var{max}
8442 The input sample aspect ratio.
8445 timestamp expressed in seconds, NAN if the input timestamp is unknown
8448 the height of the rendered text
8451 the width of the rendered text
8455 the x and y offset coordinates where the text is drawn.
8457 These parameters allow the @var{x} and @var{y} expressions to refer
8458 each other, so you can for example specify @code{y=x/dar}.
8461 @anchor{drawtext_expansion}
8462 @subsection Text expansion
8464 If @option{expansion} is set to @code{strftime},
8465 the filter recognizes strftime() sequences in the provided text and
8466 expands them accordingly. Check the documentation of strftime(). This
8467 feature is deprecated.
8469 If @option{expansion} is set to @code{none}, the text is printed verbatim.
8471 If @option{expansion} is set to @code{normal} (which is the default),
8472 the following expansion mechanism is used.
8474 The backslash character @samp{\}, followed by any character, always expands to
8475 the second character.
8477 Sequences of the form @code{%@{...@}} are expanded. The text between the
8478 braces is a function name, possibly followed by arguments separated by ':'.
8479 If the arguments contain special characters or delimiters (':' or '@}'),
8480 they should be escaped.
8482 Note that they probably must also be escaped as the value for the
8483 @option{text} option in the filter argument string and as the filter
8484 argument in the filtergraph description, and possibly also for the shell,
8485 that makes up to four levels of escaping; using a text file avoids these
8488 The following functions are available:
8493 The expression evaluation result.
8495 It must take one argument specifying the expression to be evaluated,
8496 which accepts the same constants and functions as the @var{x} and
8497 @var{y} values. Note that not all constants should be used, for
8498 example the text size is not known when evaluating the expression, so
8499 the constants @var{text_w} and @var{text_h} will have an undefined
8502 @item expr_int_format, eif
8503 Evaluate the expression's value and output as formatted integer.
8505 The first argument is the expression to be evaluated, just as for the @var{expr} function.
8506 The second argument specifies the output format. Allowed values are @samp{x},
8507 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
8508 @code{printf} function.
8509 The third parameter is optional and sets the number of positions taken by the output.
8510 It can be used to add padding with zeros from the left.
8513 The time at which the filter is running, expressed in UTC.
8514 It can accept an argument: a strftime() format string.
8517 The time at which the filter is running, expressed in the local time zone.
8518 It can accept an argument: a strftime() format string.
8521 Frame metadata. Takes one or two arguments.
8523 The first argument is mandatory and specifies the metadata key.
8525 The second argument is optional and specifies a default value, used when the
8526 metadata key is not found or empty.
8529 The frame number, starting from 0.
8532 A 1 character description of the current picture type.
8535 The timestamp of the current frame.
8536 It can take up to three arguments.
8538 The first argument is the format of the timestamp; it defaults to @code{flt}
8539 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
8540 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
8541 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
8542 @code{localtime} stands for the timestamp of the frame formatted as
8543 local time zone time.
8545 The second argument is an offset added to the timestamp.
8547 If the format is set to @code{hms}, a third argument @code{24HH} may be
8548 supplied to present the hour part of the formatted timestamp in 24h format
8551 If the format is set to @code{localtime} or @code{gmtime},
8552 a third argument may be supplied: a strftime() format string.
8553 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
8556 @subsection Examples
8560 Draw "Test Text" with font FreeSerif, using the default values for the
8561 optional parameters.
8564 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
8568 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
8569 and y=50 (counting from the top-left corner of the screen), text is
8570 yellow with a red box around it. Both the text and the box have an
8574 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
8575 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
8578 Note that the double quotes are not necessary if spaces are not used
8579 within the parameter list.
8582 Show the text at the center of the video frame:
8584 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
8588 Show the text at a random position, switching to a new position every 30 seconds:
8590 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)"
8594 Show a text line sliding from right to left in the last row of the video
8595 frame. The file @file{LONG_LINE} is assumed to contain a single line
8598 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
8602 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
8604 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
8608 Draw a single green letter "g", at the center of the input video.
8609 The glyph baseline is placed at half screen height.
8611 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
8615 Show text for 1 second every 3 seconds:
8617 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
8621 Use fontconfig to set the font. Note that the colons need to be escaped.
8623 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
8627 Print the date of a real-time encoding (see strftime(3)):
8629 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
8633 Show text fading in and out (appearing/disappearing):
8636 DS=1.0 # display start
8637 DE=10.0 # display end
8638 FID=1.5 # fade in duration
8639 FOD=5 # fade out duration
8640 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 @}"
8644 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
8645 and the @option{fontsize} value are included in the @option{y} offset.
8647 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
8648 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
8653 For more information about libfreetype, check:
8654 @url{http://www.freetype.org/}.
8656 For more information about fontconfig, check:
8657 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
8659 For more information about libfribidi, check:
8660 @url{http://fribidi.org/}.
8664 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
8666 The filter accepts the following options:
8671 Set low and high threshold values used by the Canny thresholding
8674 The high threshold selects the "strong" edge pixels, which are then
8675 connected through 8-connectivity with the "weak" edge pixels selected
8676 by the low threshold.
8678 @var{low} and @var{high} threshold values must be chosen in the range
8679 [0,1], and @var{low} should be lesser or equal to @var{high}.
8681 Default value for @var{low} is @code{20/255}, and default value for @var{high}
8685 Define the drawing mode.
8689 Draw white/gray wires on black background.
8692 Mix the colors to create a paint/cartoon effect.
8695 Apply Canny edge detector on all selected planes.
8697 Default value is @var{wires}.
8700 Select planes for filtering. By default all available planes are filtered.
8703 @subsection Examples
8707 Standard edge detection with custom values for the hysteresis thresholding:
8709 edgedetect=low=0.1:high=0.4
8713 Painting effect without thresholding:
8715 edgedetect=mode=colormix:high=0
8720 Set brightness, contrast, saturation and approximate gamma adjustment.
8722 The filter accepts the following options:
8726 Set the contrast expression. The value must be a float value in range
8727 @code{-2.0} to @code{2.0}. The default value is "1".
8730 Set the brightness expression. The value must be a float value in
8731 range @code{-1.0} to @code{1.0}. The default value is "0".
8734 Set the saturation expression. The value must be a float in
8735 range @code{0.0} to @code{3.0}. The default value is "1".
8738 Set the gamma expression. The value must be a float in range
8739 @code{0.1} to @code{10.0}. The default value is "1".
8742 Set the gamma expression for red. The value must be a float in
8743 range @code{0.1} to @code{10.0}. The default value is "1".
8746 Set the gamma expression for green. The value must be a float in range
8747 @code{0.1} to @code{10.0}. The default value is "1".
8750 Set the gamma expression for blue. The value must be a float in range
8751 @code{0.1} to @code{10.0}. The default value is "1".
8754 Set the gamma weight expression. It can be used to reduce the effect
8755 of a high gamma value on bright image areas, e.g. keep them from
8756 getting overamplified and just plain white. The value must be a float
8757 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
8758 gamma correction all the way down while @code{1.0} leaves it at its
8759 full strength. Default is "1".
8762 Set when the expressions for brightness, contrast, saturation and
8763 gamma expressions are evaluated.
8765 It accepts the following values:
8768 only evaluate expressions once during the filter initialization or
8769 when a command is processed
8772 evaluate expressions for each incoming frame
8775 Default value is @samp{init}.
8778 The expressions accept the following parameters:
8781 frame count of the input frame starting from 0
8784 byte position of the corresponding packet in the input file, NAN if
8788 frame rate of the input video, NAN if the input frame rate is unknown
8791 timestamp expressed in seconds, NAN if the input timestamp is unknown
8794 @subsection Commands
8795 The filter supports the following commands:
8799 Set the contrast expression.
8802 Set the brightness expression.
8805 Set the saturation expression.
8808 Set the gamma expression.
8811 Set the gamma_r expression.
8814 Set gamma_g expression.
8817 Set gamma_b expression.
8820 Set gamma_weight expression.
8822 The command accepts the same syntax of the corresponding option.
8824 If the specified expression is not valid, it is kept at its current
8831 Apply erosion effect to the video.
8833 This filter replaces the pixel by the local(3x3) minimum.
8835 It accepts the following options:
8842 Limit the maximum change for each plane, default is 65535.
8843 If 0, plane will remain unchanged.
8846 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
8849 Flags to local 3x3 coordinates maps like this:
8856 @section extractplanes
8858 Extract color channel components from input video stream into
8859 separate grayscale video streams.
8861 The filter accepts the following option:
8865 Set plane(s) to extract.
8867 Available values for planes are:
8878 Choosing planes not available in the input will result in an error.
8879 That means you cannot select @code{r}, @code{g}, @code{b} planes
8880 with @code{y}, @code{u}, @code{v} planes at same time.
8883 @subsection Examples
8887 Extract luma, u and v color channel component from input video frame
8888 into 3 grayscale outputs:
8890 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
8896 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
8898 For each input image, the filter will compute the optimal mapping from
8899 the input to the output given the codebook length, that is the number
8900 of distinct output colors.
8902 This filter accepts the following options.
8905 @item codebook_length, l
8906 Set codebook length. The value must be a positive integer, and
8907 represents the number of distinct output colors. Default value is 256.
8910 Set the maximum number of iterations to apply for computing the optimal
8911 mapping. The higher the value the better the result and the higher the
8912 computation time. Default value is 1.
8915 Set a random seed, must be an integer included between 0 and
8916 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
8917 will try to use a good random seed on a best effort basis.
8920 Set pal8 output pixel format. This option does not work with codebook
8921 length greater than 256.
8926 Measure graylevel entropy in histogram of color channels of video frames.
8928 It accepts the following parameters:
8932 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
8934 @var{diff} mode measures entropy of histogram delta values, absolute differences
8935 between neighbour histogram values.
8940 Apply a fade-in/out effect to the input video.
8942 It accepts the following parameters:
8946 The effect type can be either "in" for a fade-in, or "out" for a fade-out
8948 Default is @code{in}.
8950 @item start_frame, s
8951 Specify the number of the frame to start applying the fade
8952 effect at. Default is 0.
8955 The number of frames that the fade effect lasts. At the end of the
8956 fade-in effect, the output video will have the same intensity as the input video.
8957 At the end of the fade-out transition, the output video will be filled with the
8958 selected @option{color}.
8962 If set to 1, fade only alpha channel, if one exists on the input.
8965 @item start_time, st
8966 Specify the timestamp (in seconds) of the frame to start to apply the fade
8967 effect. If both start_frame and start_time are specified, the fade will start at
8968 whichever comes last. Default is 0.
8971 The number of seconds for which the fade effect has to last. At the end of the
8972 fade-in effect the output video will have the same intensity as the input video,
8973 at the end of the fade-out transition the output video will be filled with the
8974 selected @option{color}.
8975 If both duration and nb_frames are specified, duration is used. Default is 0
8976 (nb_frames is used by default).
8979 Specify the color of the fade. Default is "black".
8982 @subsection Examples
8986 Fade in the first 30 frames of video:
8991 The command above is equivalent to:
8997 Fade out the last 45 frames of a 200-frame video:
9000 fade=type=out:start_frame=155:nb_frames=45
9004 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
9006 fade=in:0:25, fade=out:975:25
9010 Make the first 5 frames yellow, then fade in from frame 5-24:
9012 fade=in:5:20:color=yellow
9016 Fade in alpha over first 25 frames of video:
9018 fade=in:0:25:alpha=1
9022 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
9024 fade=t=in:st=5.5:d=0.5
9030 Apply arbitrary expressions to samples in frequency domain
9034 Adjust the dc value (gain) of the luma plane of the image. The filter
9035 accepts an integer value in range @code{0} to @code{1000}. The default
9036 value is set to @code{0}.
9039 Adjust the dc value (gain) of the 1st chroma plane of the image. The
9040 filter accepts an integer value in range @code{0} to @code{1000}. The
9041 default value is set to @code{0}.
9044 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
9045 filter accepts an integer value in range @code{0} to @code{1000}. The
9046 default value is set to @code{0}.
9049 Set the frequency domain weight expression for the luma plane.
9052 Set the frequency domain weight expression for the 1st chroma plane.
9055 Set the frequency domain weight expression for the 2nd chroma plane.
9058 Set when the expressions are evaluated.
9060 It accepts the following values:
9063 Only evaluate expressions once during the filter initialization.
9066 Evaluate expressions for each incoming frame.
9069 Default value is @samp{init}.
9071 The filter accepts the following variables:
9074 The coordinates of the current sample.
9078 The width and height of the image.
9081 The number of input frame, starting from 0.
9084 @subsection Examples
9090 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
9096 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
9102 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
9108 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
9114 Denoise frames using 3D FFT (frequency domain filtering).
9116 The filter accepts the following options:
9120 Set the noise sigma constant. This sets denoising strength.
9121 Default value is 1. Allowed range is from 0 to 30.
9122 Using very high sigma with low overlap may give blocking artifacts.
9125 Set amount of denoising. By default all detected noise is reduced.
9126 Default value is 1. Allowed range is from 0 to 1.
9129 Set size of block, Default is 4, can be 3, 4, 5 or 6.
9130 Actual size of block in pixels is 2 to power of @var{block}, so by default
9131 block size in pixels is 2^4 which is 16.
9134 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
9137 Set number of previous frames to use for denoising. By default is set to 0.
9140 Set number of next frames to to use for denoising. By default is set to 0.
9143 Set planes which will be filtered, by default are all available filtered
9149 Extract a single field from an interlaced image using stride
9150 arithmetic to avoid wasting CPU time. The output frames are marked as
9153 The filter accepts the following options:
9157 Specify whether to extract the top (if the value is @code{0} or
9158 @code{top}) or the bottom field (if the value is @code{1} or
9164 Create new frames by copying the top and bottom fields from surrounding frames
9165 supplied as numbers by the hint file.
9169 Set file containing hints: absolute/relative frame numbers.
9171 There must be one line for each frame in a clip. Each line must contain two
9172 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
9173 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
9174 is current frame number for @code{absolute} mode or out of [-1, 1] range
9175 for @code{relative} mode. First number tells from which frame to pick up top
9176 field and second number tells from which frame to pick up bottom field.
9178 If optionally followed by @code{+} output frame will be marked as interlaced,
9179 else if followed by @code{-} output frame will be marked as progressive, else
9180 it will be marked same as input frame.
9181 If line starts with @code{#} or @code{;} that line is skipped.
9184 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
9187 Example of first several lines of @code{hint} file for @code{relative} mode:
9190 1,0 - # second frame, use third's frame top field and second's frame bottom field
9191 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
9208 Field matching filter for inverse telecine. It is meant to reconstruct the
9209 progressive frames from a telecined stream. The filter does not drop duplicated
9210 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
9211 followed by a decimation filter such as @ref{decimate} in the filtergraph.
9213 The separation of the field matching and the decimation is notably motivated by
9214 the possibility of inserting a de-interlacing filter fallback between the two.
9215 If the source has mixed telecined and real interlaced content,
9216 @code{fieldmatch} will not be able to match fields for the interlaced parts.
9217 But these remaining combed frames will be marked as interlaced, and thus can be
9218 de-interlaced by a later filter such as @ref{yadif} before decimation.
9220 In addition to the various configuration options, @code{fieldmatch} can take an
9221 optional second stream, activated through the @option{ppsrc} option. If
9222 enabled, the frames reconstruction will be based on the fields and frames from
9223 this second stream. This allows the first input to be pre-processed in order to
9224 help the various algorithms of the filter, while keeping the output lossless
9225 (assuming the fields are matched properly). Typically, a field-aware denoiser,
9226 or brightness/contrast adjustments can help.
9228 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
9229 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
9230 which @code{fieldmatch} is based on. While the semantic and usage are very
9231 close, some behaviour and options names can differ.
9233 The @ref{decimate} filter currently only works for constant frame rate input.
9234 If your input has mixed telecined (30fps) and progressive content with a lower
9235 framerate like 24fps use the following filterchain to produce the necessary cfr
9236 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
9238 The filter accepts the following options:
9242 Specify the assumed field order of the input stream. Available values are:
9246 Auto detect parity (use FFmpeg's internal parity value).
9248 Assume bottom field first.
9250 Assume top field first.
9253 Note that it is sometimes recommended not to trust the parity announced by the
9256 Default value is @var{auto}.
9259 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
9260 sense that it won't risk creating jerkiness due to duplicate frames when
9261 possible, but if there are bad edits or blended fields it will end up
9262 outputting combed frames when a good match might actually exist. On the other
9263 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
9264 but will almost always find a good frame if there is one. The other values are
9265 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
9266 jerkiness and creating duplicate frames versus finding good matches in sections
9267 with bad edits, orphaned fields, blended fields, etc.
9269 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
9271 Available values are:
9275 2-way matching (p/c)
9277 2-way matching, and trying 3rd match if still combed (p/c + n)
9279 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
9281 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
9282 still combed (p/c + n + u/b)
9284 3-way matching (p/c/n)
9286 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
9287 detected as combed (p/c/n + u/b)
9290 The parenthesis at the end indicate the matches that would be used for that
9291 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
9294 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
9297 Default value is @var{pc_n}.
9300 Mark the main input stream as a pre-processed input, and enable the secondary
9301 input stream as the clean source to pick the fields from. See the filter
9302 introduction for more details. It is similar to the @option{clip2} feature from
9305 Default value is @code{0} (disabled).
9308 Set the field to match from. It is recommended to set this to the same value as
9309 @option{order} unless you experience matching failures with that setting. In
9310 certain circumstances changing the field that is used to match from can have a
9311 large impact on matching performance. Available values are:
9315 Automatic (same value as @option{order}).
9317 Match from the bottom field.
9319 Match from the top field.
9322 Default value is @var{auto}.
9325 Set whether or not chroma is included during the match comparisons. In most
9326 cases it is recommended to leave this enabled. You should set this to @code{0}
9327 only if your clip has bad chroma problems such as heavy rainbowing or other
9328 artifacts. Setting this to @code{0} could also be used to speed things up at
9329 the cost of some accuracy.
9331 Default value is @code{1}.
9335 These define an exclusion band which excludes the lines between @option{y0} and
9336 @option{y1} from being included in the field matching decision. An exclusion
9337 band can be used to ignore subtitles, a logo, or other things that may
9338 interfere with the matching. @option{y0} sets the starting scan line and
9339 @option{y1} sets the ending line; all lines in between @option{y0} and
9340 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
9341 @option{y0} and @option{y1} to the same value will disable the feature.
9342 @option{y0} and @option{y1} defaults to @code{0}.
9345 Set the scene change detection threshold as a percentage of maximum change on
9346 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
9347 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
9348 @option{scthresh} is @code{[0.0, 100.0]}.
9350 Default value is @code{12.0}.
9353 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
9354 account the combed scores of matches when deciding what match to use as the
9355 final match. Available values are:
9359 No final matching based on combed scores.
9361 Combed scores are only used when a scene change is detected.
9363 Use combed scores all the time.
9366 Default is @var{sc}.
9369 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
9370 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
9371 Available values are:
9375 No forced calculation.
9377 Force p/c/n calculations.
9379 Force p/c/n/u/b calculations.
9382 Default value is @var{none}.
9385 This is the area combing threshold used for combed frame detection. This
9386 essentially controls how "strong" or "visible" combing must be to be detected.
9387 Larger values mean combing must be more visible and smaller values mean combing
9388 can be less visible or strong and still be detected. Valid settings are from
9389 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
9390 be detected as combed). This is basically a pixel difference value. A good
9391 range is @code{[8, 12]}.
9393 Default value is @code{9}.
9396 Sets whether or not chroma is considered in the combed frame decision. Only
9397 disable this if your source has chroma problems (rainbowing, etc.) that are
9398 causing problems for the combed frame detection with chroma enabled. Actually,
9399 using @option{chroma}=@var{0} is usually more reliable, except for the case
9400 where there is chroma only combing in the source.
9402 Default value is @code{0}.
9406 Respectively set the x-axis and y-axis size of the window used during combed
9407 frame detection. This has to do with the size of the area in which
9408 @option{combpel} pixels are required to be detected as combed for a frame to be
9409 declared combed. See the @option{combpel} parameter description for more info.
9410 Possible values are any number that is a power of 2 starting at 4 and going up
9413 Default value is @code{16}.
9416 The number of combed pixels inside any of the @option{blocky} by
9417 @option{blockx} size blocks on the frame for the frame to be detected as
9418 combed. While @option{cthresh} controls how "visible" the combing must be, this
9419 setting controls "how much" combing there must be in any localized area (a
9420 window defined by the @option{blockx} and @option{blocky} settings) on the
9421 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
9422 which point no frames will ever be detected as combed). This setting is known
9423 as @option{MI} in TFM/VFM vocabulary.
9425 Default value is @code{80}.
9428 @anchor{p/c/n/u/b meaning}
9429 @subsection p/c/n/u/b meaning
9431 @subsubsection p/c/n
9433 We assume the following telecined stream:
9436 Top fields: 1 2 2 3 4
9437 Bottom fields: 1 2 3 4 4
9440 The numbers correspond to the progressive frame the fields relate to. Here, the
9441 first two frames are progressive, the 3rd and 4th are combed, and so on.
9443 When @code{fieldmatch} is configured to run a matching from bottom
9444 (@option{field}=@var{bottom}) this is how this input stream get transformed:
9449 B 1 2 3 4 4 <-- matching reference
9458 As a result of the field matching, we can see that some frames get duplicated.
9459 To perform a complete inverse telecine, you need to rely on a decimation filter
9460 after this operation. See for instance the @ref{decimate} filter.
9462 The same operation now matching from top fields (@option{field}=@var{top})
9467 T 1 2 2 3 4 <-- matching reference
9477 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
9478 basically, they refer to the frame and field of the opposite parity:
9481 @item @var{p} matches the field of the opposite parity in the previous frame
9482 @item @var{c} matches the field of the opposite parity in the current frame
9483 @item @var{n} matches the field of the opposite parity in the next frame
9488 The @var{u} and @var{b} matching are a bit special in the sense that they match
9489 from the opposite parity flag. In the following examples, we assume that we are
9490 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
9491 'x' is placed above and below each matched fields.
9493 With bottom matching (@option{field}=@var{bottom}):
9498 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
9499 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
9507 With top matching (@option{field}=@var{top}):
9512 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
9513 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
9521 @subsection Examples
9523 Simple IVTC of a top field first telecined stream:
9525 fieldmatch=order=tff:combmatch=none, decimate
9528 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
9530 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
9535 Transform the field order of the input video.
9537 It accepts the following parameters:
9542 The output field order. Valid values are @var{tff} for top field first or @var{bff}
9543 for bottom field first.
9546 The default value is @samp{tff}.
9548 The transformation is done by shifting the picture content up or down
9549 by one line, and filling the remaining line with appropriate picture content.
9550 This method is consistent with most broadcast field order converters.
9552 If the input video is not flagged as being interlaced, or it is already
9553 flagged as being of the required output field order, then this filter does
9554 not alter the incoming video.
9556 It is very useful when converting to or from PAL DV material,
9557 which is bottom field first.
9561 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
9564 @section fifo, afifo
9566 Buffer input images and send them when they are requested.
9568 It is mainly useful when auto-inserted by the libavfilter
9571 It does not take parameters.
9573 @section fillborders
9575 Fill borders of the input video, without changing video stream dimensions.
9576 Sometimes video can have garbage at the four edges and you may not want to
9577 crop video input to keep size multiple of some number.
9579 This filter accepts the following options:
9583 Number of pixels to fill from left border.
9586 Number of pixels to fill from right border.
9589 Number of pixels to fill from top border.
9592 Number of pixels to fill from bottom border.
9597 It accepts the following values:
9600 fill pixels using outermost pixels
9603 fill pixels using mirroring
9606 fill pixels with constant value
9609 Default is @var{smear}.
9612 Set color for pixels in fixed mode. Default is @var{black}.
9617 Find a rectangular object
9619 It accepts the following options:
9623 Filepath of the object image, needs to be in gray8.
9626 Detection threshold, default is 0.5.
9629 Number of mipmaps, default is 3.
9631 @item xmin, ymin, xmax, ymax
9632 Specifies the rectangle in which to search.
9635 @subsection Examples
9639 Generate a representative palette of a given video using @command{ffmpeg}:
9641 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
9647 Cover a rectangular object
9649 It accepts the following options:
9653 Filepath of the optional cover image, needs to be in yuv420.
9658 It accepts the following values:
9661 cover it by the supplied image
9663 cover it by interpolating the surrounding pixels
9666 Default value is @var{blur}.
9669 @subsection Examples
9673 Generate a representative palette of a given video using @command{ffmpeg}:
9675 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
9681 Flood area with values of same pixel components with another values.
9683 It accepts the following options:
9686 Set pixel x coordinate.
9689 Set pixel y coordinate.
9692 Set source #0 component value.
9695 Set source #1 component value.
9698 Set source #2 component value.
9701 Set source #3 component value.
9704 Set destination #0 component value.
9707 Set destination #1 component value.
9710 Set destination #2 component value.
9713 Set destination #3 component value.
9719 Convert the input video to one of the specified pixel formats.
9720 Libavfilter will try to pick one that is suitable as input to
9723 It accepts the following parameters:
9727 A '|'-separated list of pixel format names, such as
9728 "pix_fmts=yuv420p|monow|rgb24".
9732 @subsection Examples
9736 Convert the input video to the @var{yuv420p} format
9738 format=pix_fmts=yuv420p
9741 Convert the input video to any of the formats in the list
9743 format=pix_fmts=yuv420p|yuv444p|yuv410p
9750 Convert the video to specified constant frame rate by duplicating or dropping
9751 frames as necessary.
9753 It accepts the following parameters:
9757 The desired output frame rate. The default is @code{25}.
9760 Assume the first PTS should be the given value, in seconds. This allows for
9761 padding/trimming at the start of stream. By default, no assumption is made
9762 about the first frame's expected PTS, so no padding or trimming is done.
9763 For example, this could be set to 0 to pad the beginning with duplicates of
9764 the first frame if a video stream starts after the audio stream or to trim any
9765 frames with a negative PTS.
9768 Timestamp (PTS) rounding method.
9770 Possible values are:
9777 round towards -infinity
9779 round towards +infinity
9783 The default is @code{near}.
9786 Action performed when reading the last frame.
9788 Possible values are:
9791 Use same timestamp rounding method as used for other frames.
9793 Pass through last frame if input duration has not been reached yet.
9795 The default is @code{round}.
9799 Alternatively, the options can be specified as a flat string:
9800 @var{fps}[:@var{start_time}[:@var{round}]].
9802 See also the @ref{setpts} filter.
9804 @subsection Examples
9808 A typical usage in order to set the fps to 25:
9814 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
9816 fps=fps=film:round=near
9822 Pack two different video streams into a stereoscopic video, setting proper
9823 metadata on supported codecs. The two views should have the same size and
9824 framerate and processing will stop when the shorter video ends. Please note
9825 that you may conveniently adjust view properties with the @ref{scale} and
9828 It accepts the following parameters:
9832 The desired packing format. Supported values are:
9837 The views are next to each other (default).
9840 The views are on top of each other.
9843 The views are packed by line.
9846 The views are packed by column.
9849 The views are temporally interleaved.
9858 # Convert left and right views into a frame-sequential video
9859 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
9861 # Convert views into a side-by-side video with the same output resolution as the input
9862 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
9867 Change the frame rate by interpolating new video output frames from the source
9870 This filter is not designed to function correctly with interlaced media. If
9871 you wish to change the frame rate of interlaced media then you are required
9872 to deinterlace before this filter and re-interlace after this filter.
9874 A description of the accepted options follows.
9878 Specify the output frames per second. This option can also be specified
9879 as a value alone. The default is @code{50}.
9882 Specify the start of a range where the output frame will be created as a
9883 linear interpolation of two frames. The range is [@code{0}-@code{255}],
9884 the default is @code{15}.
9887 Specify the end of a range where the output frame will be created as a
9888 linear interpolation of two frames. The range is [@code{0}-@code{255}],
9889 the default is @code{240}.
9892 Specify the level at which a scene change is detected as a value between
9893 0 and 100 to indicate a new scene; a low value reflects a low
9894 probability for the current frame to introduce a new scene, while a higher
9895 value means the current frame is more likely to be one.
9896 The default is @code{8.2}.
9899 Specify flags influencing the filter process.
9901 Available value for @var{flags} is:
9904 @item scene_change_detect, scd
9905 Enable scene change detection using the value of the option @var{scene}.
9906 This flag is enabled by default.
9912 Select one frame every N-th frame.
9914 This filter accepts the following option:
9917 Select frame after every @code{step} frames.
9918 Allowed values are positive integers higher than 0. Default value is @code{1}.
9924 Apply a frei0r effect to the input video.
9926 To enable the compilation of this filter, you need to install the frei0r
9927 header and configure FFmpeg with @code{--enable-frei0r}.
9929 It accepts the following parameters:
9934 The name of the frei0r effect to load. If the environment variable
9935 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
9936 directories specified by the colon-separated list in @env{FREI0R_PATH}.
9937 Otherwise, the standard frei0r paths are searched, in this order:
9938 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
9939 @file{/usr/lib/frei0r-1/}.
9942 A '|'-separated list of parameters to pass to the frei0r effect.
9946 A frei0r effect parameter can be a boolean (its value is either
9947 "y" or "n"), a double, a color (specified as
9948 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
9949 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
9950 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
9951 a position (specified as @var{X}/@var{Y}, where
9952 @var{X} and @var{Y} are floating point numbers) and/or a string.
9954 The number and types of parameters depend on the loaded effect. If an
9955 effect parameter is not specified, the default value is set.
9957 @subsection Examples
9961 Apply the distort0r effect, setting the first two double parameters:
9963 frei0r=filter_name=distort0r:filter_params=0.5|0.01
9967 Apply the colordistance effect, taking a color as the first parameter:
9969 frei0r=colordistance:0.2/0.3/0.4
9970 frei0r=colordistance:violet
9971 frei0r=colordistance:0x112233
9975 Apply the perspective effect, specifying the top left and top right image
9978 frei0r=perspective:0.2/0.2|0.8/0.2
9982 For more information, see
9983 @url{http://frei0r.dyne.org}
9987 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
9989 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
9990 processing filter, one of them is performed once per block, not per pixel.
9991 This allows for much higher speed.
9993 The filter accepts the following options:
9997 Set quality. This option defines the number of levels for averaging. It accepts
9998 an integer in the range 4-5. Default value is @code{4}.
10001 Force a constant quantization parameter. It accepts an integer in range 0-63.
10002 If not set, the filter will use the QP from the video stream (if available).
10005 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
10006 more details but also more artifacts, while higher values make the image smoother
10007 but also blurrier. Default value is @code{0} − PSNR optimal.
10009 @item use_bframe_qp
10010 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
10011 option may cause flicker since the B-Frames have often larger QP. Default is
10012 @code{0} (not enabled).
10018 Apply Gaussian blur filter.
10020 The filter accepts the following options:
10024 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
10027 Set number of steps for Gaussian approximation. Defauls is @code{1}.
10030 Set which planes to filter. By default all planes are filtered.
10033 Set vertical sigma, if negative it will be same as @code{sigma}.
10034 Default is @code{-1}.
10039 The filter accepts the following options:
10042 @item lum_expr, lum
10043 Set the luminance expression.
10045 Set the chrominance blue expression.
10047 Set the chrominance red expression.
10048 @item alpha_expr, a
10049 Set the alpha expression.
10051 Set the red expression.
10052 @item green_expr, g
10053 Set the green expression.
10055 Set the blue expression.
10058 The colorspace is selected according to the specified options. If one
10059 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
10060 options is specified, the filter will automatically select a YCbCr
10061 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
10062 @option{blue_expr} options is specified, it will select an RGB
10065 If one of the chrominance expression is not defined, it falls back on the other
10066 one. If no alpha expression is specified it will evaluate to opaque value.
10067 If none of chrominance expressions are specified, they will evaluate
10068 to the luminance expression.
10070 The expressions can use the following variables and functions:
10074 The sequential number of the filtered frame, starting from @code{0}.
10078 The coordinates of the current sample.
10082 The width and height of the image.
10086 Width and height scale depending on the currently filtered plane. It is the
10087 ratio between the corresponding luma plane number of pixels and the current
10088 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
10089 @code{0.5,0.5} for chroma planes.
10092 Time of the current frame, expressed in seconds.
10095 Return the value of the pixel at location (@var{x},@var{y}) of the current
10099 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
10103 Return the value of the pixel at location (@var{x},@var{y}) of the
10104 blue-difference chroma plane. Return 0 if there is no such plane.
10107 Return the value of the pixel at location (@var{x},@var{y}) of the
10108 red-difference chroma plane. Return 0 if there is no such plane.
10113 Return the value of the pixel at location (@var{x},@var{y}) of the
10114 red/green/blue component. Return 0 if there is no such component.
10117 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
10118 plane. Return 0 if there is no such plane.
10121 For functions, if @var{x} and @var{y} are outside the area, the value will be
10122 automatically clipped to the closer edge.
10124 @subsection Examples
10128 Flip the image horizontally:
10134 Generate a bidimensional sine wave, with angle @code{PI/3} and a
10135 wavelength of 100 pixels:
10137 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
10141 Generate a fancy enigmatic moving light:
10143 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
10147 Generate a quick emboss effect:
10149 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
10153 Modify RGB components depending on pixel position:
10155 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
10159 Create a radial gradient that is the same size as the input (also see
10160 the @ref{vignette} filter):
10162 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
10168 Fix the banding artifacts that are sometimes introduced into nearly flat
10169 regions by truncation to 8-bit color depth.
10170 Interpolate the gradients that should go where the bands are, and
10173 It is designed for playback only. Do not use it prior to
10174 lossy compression, because compression tends to lose the dither and
10175 bring back the bands.
10177 It accepts the following parameters:
10182 The maximum amount by which the filter will change any one pixel. This is also
10183 the threshold for detecting nearly flat regions. Acceptable values range from
10184 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
10188 The neighborhood to fit the gradient to. A larger radius makes for smoother
10189 gradients, but also prevents the filter from modifying the pixels near detailed
10190 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
10191 values will be clipped to the valid range.
10195 Alternatively, the options can be specified as a flat string:
10196 @var{strength}[:@var{radius}]
10198 @subsection Examples
10202 Apply the filter with a @code{3.5} strength and radius of @code{8}:
10208 Specify radius, omitting the strength (which will fall-back to the default
10217 A color constancy variation filter which estimates scene illumination via grey edge algorithm
10218 and corrects the scene colors accordingly.
10220 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
10222 The filter accepts the following options:
10226 The order of differentiation to be applied on the scene. Must be chosen in the range
10227 [0,2] and default value is 1.
10230 The Minkowski parameter to be used for calculating the Minkowski distance. Must
10231 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
10232 max value instead of calculating Minkowski distance.
10235 The standard deviation of Gaussian blur to be applied on the scene. Must be
10236 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
10237 can't be euqal to 0 if @var{difford} is greater than 0.
10240 @subsection Examples
10246 greyedge=difford=1:minknorm=5:sigma=2
10252 greyedge=difford=1:minknorm=0:sigma=2
10260 Apply a Hald CLUT to a video stream.
10262 First input is the video stream to process, and second one is the Hald CLUT.
10263 The Hald CLUT input can be a simple picture or a complete video stream.
10265 The filter accepts the following options:
10269 Force termination when the shortest input terminates. Default is @code{0}.
10271 Continue applying the last CLUT after the end of the stream. A value of
10272 @code{0} disable the filter after the last frame of the CLUT is reached.
10273 Default is @code{1}.
10276 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
10277 filters share the same internals).
10279 More information about the Hald CLUT can be found on Eskil Steenberg's website
10280 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
10282 @subsection Workflow examples
10284 @subsubsection Hald CLUT video stream
10286 Generate an identity Hald CLUT stream altered with various effects:
10288 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "hue=H=2*PI*t:s=sin(2*PI*t)+1, curves=cross_process" -t 10 -c:v ffv1 clut.nut
10291 Note: make sure you use a lossless codec.
10293 Then use it with @code{haldclut} to apply it on some random stream:
10295 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
10298 The Hald CLUT will be applied to the 10 first seconds (duration of
10299 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
10300 to the remaining frames of the @code{mandelbrot} stream.
10302 @subsubsection Hald CLUT with preview
10304 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
10305 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
10306 biggest possible square starting at the top left of the picture. The remaining
10307 padding pixels (bottom or right) will be ignored. This area can be used to add
10308 a preview of the Hald CLUT.
10310 Typically, the following generated Hald CLUT will be supported by the
10311 @code{haldclut} filter:
10314 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
10315 pad=iw+320 [padded_clut];
10316 smptebars=s=320x256, split [a][b];
10317 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
10318 [main][b] overlay=W-320" -frames:v 1 clut.png
10321 It contains the original and a preview of the effect of the CLUT: SMPTE color
10322 bars are displayed on the right-top, and below the same color bars processed by
10325 Then, the effect of this Hald CLUT can be visualized with:
10327 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
10332 Flip the input video horizontally.
10334 For example, to horizontally flip the input video with @command{ffmpeg}:
10336 ffmpeg -i in.avi -vf "hflip" out.avi
10340 This filter applies a global color histogram equalization on a
10343 It can be used to correct video that has a compressed range of pixel
10344 intensities. The filter redistributes the pixel intensities to
10345 equalize their distribution across the intensity range. It may be
10346 viewed as an "automatically adjusting contrast filter". This filter is
10347 useful only for correcting degraded or poorly captured source
10350 The filter accepts the following options:
10354 Determine the amount of equalization to be applied. As the strength
10355 is reduced, the distribution of pixel intensities more-and-more
10356 approaches that of the input frame. The value must be a float number
10357 in the range [0,1] and defaults to 0.200.
10360 Set the maximum intensity that can generated and scale the output
10361 values appropriately. The strength should be set as desired and then
10362 the intensity can be limited if needed to avoid washing-out. The value
10363 must be a float number in the range [0,1] and defaults to 0.210.
10366 Set the antibanding level. If enabled the filter will randomly vary
10367 the luminance of output pixels by a small amount to avoid banding of
10368 the histogram. Possible values are @code{none}, @code{weak} or
10369 @code{strong}. It defaults to @code{none}.
10374 Compute and draw a color distribution histogram for the input video.
10376 The computed histogram is a representation of the color component
10377 distribution in an image.
10379 Standard histogram displays the color components distribution in an image.
10380 Displays color graph for each color component. Shows distribution of
10381 the Y, U, V, A or R, G, B components, depending on input format, in the
10382 current frame. Below each graph a color component scale meter is shown.
10384 The filter accepts the following options:
10388 Set height of level. Default value is @code{200}.
10389 Allowed range is [50, 2048].
10392 Set height of color scale. Default value is @code{12}.
10393 Allowed range is [0, 40].
10397 It accepts the following values:
10400 Per color component graphs are placed below each other.
10403 Per color component graphs are placed side by side.
10406 Presents information identical to that in the @code{parade}, except
10407 that the graphs representing color components are superimposed directly
10410 Default is @code{stack}.
10413 Set mode. Can be either @code{linear}, or @code{logarithmic}.
10414 Default is @code{linear}.
10417 Set what color components to display.
10418 Default is @code{7}.
10421 Set foreground opacity. Default is @code{0.7}.
10424 Set background opacity. Default is @code{0.5}.
10427 @subsection Examples
10432 Calculate and draw histogram:
10434 ffplay -i input -vf histogram
10442 This is a high precision/quality 3d denoise filter. It aims to reduce
10443 image noise, producing smooth images and making still images really
10444 still. It should enhance compressibility.
10446 It accepts the following optional parameters:
10450 A non-negative floating point number which specifies spatial luma strength.
10451 It defaults to 4.0.
10453 @item chroma_spatial
10454 A non-negative floating point number which specifies spatial chroma strength.
10455 It defaults to 3.0*@var{luma_spatial}/4.0.
10458 A floating point number which specifies luma temporal strength. It defaults to
10459 6.0*@var{luma_spatial}/4.0.
10462 A floating point number which specifies chroma temporal strength. It defaults to
10463 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
10466 @section hwdownload
10468 Download hardware frames to system memory.
10470 The input must be in hardware frames, and the output a non-hardware format.
10471 Not all formats will be supported on the output - it may be necessary to insert
10472 an additional @option{format} filter immediately following in the graph to get
10473 the output in a supported format.
10477 Map hardware frames to system memory or to another device.
10479 This filter has several different modes of operation; which one is used depends
10480 on the input and output formats:
10483 Hardware frame input, normal frame output
10485 Map the input frames to system memory and pass them to the output. If the
10486 original hardware frame is later required (for example, after overlaying
10487 something else on part of it), the @option{hwmap} filter can be used again
10488 in the next mode to retrieve it.
10490 Normal frame input, hardware frame output
10492 If the input is actually a software-mapped hardware frame, then unmap it -
10493 that is, return the original hardware frame.
10495 Otherwise, a device must be provided. Create new hardware surfaces on that
10496 device for the output, then map them back to the software format at the input
10497 and give those frames to the preceding filter. This will then act like the
10498 @option{hwupload} filter, but may be able to avoid an additional copy when
10499 the input is already in a compatible format.
10501 Hardware frame input and output
10503 A device must be supplied for the output, either directly or with the
10504 @option{derive_device} option. The input and output devices must be of
10505 different types and compatible - the exact meaning of this is
10506 system-dependent, but typically it means that they must refer to the same
10507 underlying hardware context (for example, refer to the same graphics card).
10509 If the input frames were originally created on the output device, then unmap
10510 to retrieve the original frames.
10512 Otherwise, map the frames to the output device - create new hardware frames
10513 on the output corresponding to the frames on the input.
10516 The following additional parameters are accepted:
10520 Set the frame mapping mode. Some combination of:
10523 The mapped frame should be readable.
10525 The mapped frame should be writeable.
10527 The mapping will always overwrite the entire frame.
10529 This may improve performance in some cases, as the original contents of the
10530 frame need not be loaded.
10532 The mapping must not involve any copying.
10534 Indirect mappings to copies of frames are created in some cases where either
10535 direct mapping is not possible or it would have unexpected properties.
10536 Setting this flag ensures that the mapping is direct and will fail if that is
10539 Defaults to @var{read+write} if not specified.
10541 @item derive_device @var{type}
10542 Rather than using the device supplied at initialisation, instead derive a new
10543 device of type @var{type} from the device the input frames exist on.
10546 In a hardware to hardware mapping, map in reverse - create frames in the sink
10547 and map them back to the source. This may be necessary in some cases where
10548 a mapping in one direction is required but only the opposite direction is
10549 supported by the devices being used.
10551 This option is dangerous - it may break the preceding filter in undefined
10552 ways if there are any additional constraints on that filter's output.
10553 Do not use it without fully understanding the implications of its use.
10558 Upload system memory frames to hardware surfaces.
10560 The device to upload to must be supplied when the filter is initialised. If
10561 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
10564 @anchor{hwupload_cuda}
10565 @section hwupload_cuda
10567 Upload system memory frames to a CUDA device.
10569 It accepts the following optional parameters:
10573 The number of the CUDA device to use
10578 Apply a high-quality magnification filter designed for pixel art. This filter
10579 was originally created by Maxim Stepin.
10581 It accepts the following option:
10585 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
10586 @code{hq3x} and @code{4} for @code{hq4x}.
10587 Default is @code{3}.
10591 Stack input videos horizontally.
10593 All streams must be of same pixel format and of same height.
10595 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
10596 to create same output.
10598 The filter accept the following option:
10602 Set number of input streams. Default is 2.
10605 If set to 1, force the output to terminate when the shortest input
10606 terminates. Default value is 0.
10611 Modify the hue and/or the saturation of the input.
10613 It accepts the following parameters:
10617 Specify the hue angle as a number of degrees. It accepts an expression,
10618 and defaults to "0".
10621 Specify the saturation in the [-10,10] range. It accepts an expression and
10625 Specify the hue angle as a number of radians. It accepts an
10626 expression, and defaults to "0".
10629 Specify the brightness in the [-10,10] range. It accepts an expression and
10633 @option{h} and @option{H} are mutually exclusive, and can't be
10634 specified at the same time.
10636 The @option{b}, @option{h}, @option{H} and @option{s} option values are
10637 expressions containing the following constants:
10641 frame count of the input frame starting from 0
10644 presentation timestamp of the input frame expressed in time base units
10647 frame rate of the input video, NAN if the input frame rate is unknown
10650 timestamp expressed in seconds, NAN if the input timestamp is unknown
10653 time base of the input video
10656 @subsection Examples
10660 Set the hue to 90 degrees and the saturation to 1.0:
10666 Same command but expressing the hue in radians:
10672 Rotate hue and make the saturation swing between 0
10673 and 2 over a period of 1 second:
10675 hue="H=2*PI*t: s=sin(2*PI*t)+1"
10679 Apply a 3 seconds saturation fade-in effect starting at 0:
10681 hue="s=min(t/3\,1)"
10684 The general fade-in expression can be written as:
10686 hue="s=min(0\, max((t-START)/DURATION\, 1))"
10690 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
10692 hue="s=max(0\, min(1\, (8-t)/3))"
10695 The general fade-out expression can be written as:
10697 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
10702 @subsection Commands
10704 This filter supports the following commands:
10710 Modify the hue and/or the saturation and/or brightness of the input video.
10711 The command accepts the same syntax of the corresponding option.
10713 If the specified expression is not valid, it is kept at its current
10717 @section hysteresis
10719 Grow first stream into second stream by connecting components.
10720 This makes it possible to build more robust edge masks.
10722 This filter accepts the following options:
10726 Set which planes will be processed as bitmap, unprocessed planes will be
10727 copied from first stream.
10728 By default value 0xf, all planes will be processed.
10731 Set threshold which is used in filtering. If pixel component value is higher than
10732 this value filter algorithm for connecting components is activated.
10733 By default value is 0.
10738 Detect video interlacing type.
10740 This filter tries to detect if the input frames are interlaced, progressive,
10741 top or bottom field first. It will also try to detect fields that are
10742 repeated between adjacent frames (a sign of telecine).
10744 Single frame detection considers only immediately adjacent frames when classifying each frame.
10745 Multiple frame detection incorporates the classification history of previous frames.
10747 The filter will log these metadata values:
10750 @item single.current_frame
10751 Detected type of current frame using single-frame detection. One of:
10752 ``tff'' (top field first), ``bff'' (bottom field first),
10753 ``progressive'', or ``undetermined''
10756 Cumulative number of frames detected as top field first using single-frame detection.
10759 Cumulative number of frames detected as top field first using multiple-frame detection.
10762 Cumulative number of frames detected as bottom field first using single-frame detection.
10764 @item multiple.current_frame
10765 Detected type of current frame using multiple-frame detection. One of:
10766 ``tff'' (top field first), ``bff'' (bottom field first),
10767 ``progressive'', or ``undetermined''
10770 Cumulative number of frames detected as bottom field first using multiple-frame detection.
10772 @item single.progressive
10773 Cumulative number of frames detected as progressive using single-frame detection.
10775 @item multiple.progressive
10776 Cumulative number of frames detected as progressive using multiple-frame detection.
10778 @item single.undetermined
10779 Cumulative number of frames that could not be classified using single-frame detection.
10781 @item multiple.undetermined
10782 Cumulative number of frames that could not be classified using multiple-frame detection.
10784 @item repeated.current_frame
10785 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
10787 @item repeated.neither
10788 Cumulative number of frames with no repeated field.
10791 Cumulative number of frames with the top field repeated from the previous frame's top field.
10793 @item repeated.bottom
10794 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
10797 The filter accepts the following options:
10801 Set interlacing threshold.
10803 Set progressive threshold.
10805 Threshold for repeated field detection.
10807 Number of frames after which a given frame's contribution to the
10808 statistics is halved (i.e., it contributes only 0.5 to its
10809 classification). The default of 0 means that all frames seen are given
10810 full weight of 1.0 forever.
10811 @item analyze_interlaced_flag
10812 When this is not 0 then idet will use the specified number of frames to determine
10813 if the interlaced flag is accurate, it will not count undetermined frames.
10814 If the flag is found to be accurate it will be used without any further
10815 computations, if it is found to be inaccurate it will be cleared without any
10816 further computations. This allows inserting the idet filter as a low computational
10817 method to clean up the interlaced flag
10822 Deinterleave or interleave fields.
10824 This filter allows one to process interlaced images fields without
10825 deinterlacing them. Deinterleaving splits the input frame into 2
10826 fields (so called half pictures). Odd lines are moved to the top
10827 half of the output image, even lines to the bottom half.
10828 You can process (filter) them independently and then re-interleave them.
10830 The filter accepts the following options:
10834 @item chroma_mode, c
10835 @item alpha_mode, a
10836 Available values for @var{luma_mode}, @var{chroma_mode} and
10837 @var{alpha_mode} are:
10843 @item deinterleave, d
10844 Deinterleave fields, placing one above the other.
10846 @item interleave, i
10847 Interleave fields. Reverse the effect of deinterleaving.
10849 Default value is @code{none}.
10851 @item luma_swap, ls
10852 @item chroma_swap, cs
10853 @item alpha_swap, as
10854 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
10859 Apply inflate effect to the video.
10861 This filter replaces the pixel by the local(3x3) average by taking into account
10862 only values higher than the pixel.
10864 It accepts the following options:
10871 Limit the maximum change for each plane, default is 65535.
10872 If 0, plane will remain unchanged.
10877 Simple interlacing filter from progressive contents. This interleaves upper (or
10878 lower) lines from odd frames with lower (or upper) lines from even frames,
10879 halving the frame rate and preserving image height.
10882 Original Original New Frame
10883 Frame 'j' Frame 'j+1' (tff)
10884 ========== =========== ==================
10885 Line 0 --------------------> Frame 'j' Line 0
10886 Line 1 Line 1 ----> Frame 'j+1' Line 1
10887 Line 2 ---------------------> Frame 'j' Line 2
10888 Line 3 Line 3 ----> Frame 'j+1' Line 3
10890 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
10893 It accepts the following optional parameters:
10897 This determines whether the interlaced frame is taken from the even
10898 (tff - default) or odd (bff) lines of the progressive frame.
10901 Vertical lowpass filter to avoid twitter interlacing and
10902 reduce moire patterns.
10906 Disable vertical lowpass filter
10909 Enable linear filter (default)
10912 Enable complex filter. This will slightly less reduce twitter and moire
10913 but better retain detail and subjective sharpness impression.
10920 Deinterlace input video by applying Donald Graft's adaptive kernel
10921 deinterling. Work on interlaced parts of a video to produce
10922 progressive frames.
10924 The description of the accepted parameters follows.
10928 Set the threshold which affects the filter's tolerance when
10929 determining if a pixel line must be processed. It must be an integer
10930 in the range [0,255] and defaults to 10. A value of 0 will result in
10931 applying the process on every pixels.
10934 Paint pixels exceeding the threshold value to white if set to 1.
10938 Set the fields order. Swap fields if set to 1, leave fields alone if
10942 Enable additional sharpening if set to 1. Default is 0.
10945 Enable twoway sharpening if set to 1. Default is 0.
10948 @subsection Examples
10952 Apply default values:
10954 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
10958 Enable additional sharpening:
10964 Paint processed pixels in white:
10970 @section lenscorrection
10972 Correct radial lens distortion
10974 This filter can be used to correct for radial distortion as can result from the use
10975 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
10976 one can use tools available for example as part of opencv or simply trial-and-error.
10977 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
10978 and extract the k1 and k2 coefficients from the resulting matrix.
10980 Note that effectively the same filter is available in the open-source tools Krita and
10981 Digikam from the KDE project.
10983 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
10984 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
10985 brightness distribution, so you may want to use both filters together in certain
10986 cases, though you will have to take care of ordering, i.e. whether vignetting should
10987 be applied before or after lens correction.
10989 @subsection Options
10991 The filter accepts the following options:
10995 Relative x-coordinate of the focal point of the image, and thereby the center of the
10996 distortion. This value has a range [0,1] and is expressed as fractions of the image
10997 width. Default is 0.5.
10999 Relative y-coordinate of the focal point of the image, and thereby the center of the
11000 distortion. This value has a range [0,1] and is expressed as fractions of the image
11001 height. Default is 0.5.
11003 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
11004 no correction. Default is 0.
11006 Coefficient of the double quadratic correction term. This value has a range [-1,1].
11007 0 means no correction. Default is 0.
11010 The formula that generates the correction is:
11012 @var{r_src} = @var{r_tgt} * (1 + @var{k1} * (@var{r_tgt} / @var{r_0})^2 + @var{k2} * (@var{r_tgt} / @var{r_0})^4)
11014 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
11015 distances from the focal point in the source and target images, respectively.
11019 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
11021 The @code{lensfun} filter requires the camera make, camera model, and lens model
11022 to apply the lens correction. The filter will load the lensfun database and
11023 query it to find the corresponding camera and lens entries in the database. As
11024 long as these entries can be found with the given options, the filter can
11025 perform corrections on frames. Note that incomplete strings will result in the
11026 filter choosing the best match with the given options, and the filter will
11027 output the chosen camera and lens models (logged with level "info"). You must
11028 provide the make, camera model, and lens model as they are required.
11030 The filter accepts the following options:
11034 The make of the camera (for example, "Canon"). This option is required.
11037 The model of the camera (for example, "Canon EOS 100D"). This option is
11041 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
11042 option is required.
11045 The type of correction to apply. The following values are valid options:
11049 Enables fixing lens vignetting.
11052 Enables fixing lens geometry. This is the default.
11055 Enables fixing chromatic aberrations.
11058 Enables fixing lens vignetting and lens geometry.
11061 Enables fixing lens vignetting and chromatic aberrations.
11064 Enables fixing both lens geometry and chromatic aberrations.
11067 Enables all possible corrections.
11071 The focal length of the image/video (zoom; expected constant for video). For
11072 example, a 18--55mm lens has focal length range of [18--55], so a value in that
11073 range should be chosen when using that lens. Default 18.
11076 The aperture of the image/video (expected constant for video). Note that
11077 aperture is only used for vignetting correction. Default 3.5.
11079 @item focus_distance
11080 The focus distance of the image/video (expected constant for video). Note that
11081 focus distance is only used for vignetting and only slightly affects the
11082 vignetting correction process. If unknown, leave it at the default value (which
11085 @item target_geometry
11086 The target geometry of the output image/video. The following values are valid
11090 @item rectilinear (default)
11093 @item equirectangular
11094 @item fisheye_orthographic
11095 @item fisheye_stereographic
11096 @item fisheye_equisolid
11097 @item fisheye_thoby
11100 Apply the reverse of image correction (instead of correcting distortion, apply
11103 @item interpolation
11104 The type of interpolation used when correcting distortion. The following values
11109 @item linear (default)
11114 @subsection Examples
11118 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
11119 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
11123 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
11127 Apply the same as before, but only for the first 5 seconds of video.
11130 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
11137 Obtain the VMAF (Video Multi-Method Assessment Fusion)
11138 score between two input videos.
11140 The obtained VMAF score is printed through the logging system.
11142 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
11143 After installing the library it can be enabled using:
11144 @code{./configure --enable-libvmaf --enable-version3}.
11145 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
11147 The filter has following options:
11151 Set the model path which is to be used for SVM.
11152 Default value: @code{"vmaf_v0.6.1.pkl"}
11155 Set the file path to be used to store logs.
11158 Set the format of the log file (xml or json).
11160 @item enable_transform
11161 Enables transform for computing vmaf.
11164 Invokes the phone model which will generate VMAF scores higher than in the
11165 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
11168 Enables computing psnr along with vmaf.
11171 Enables computing ssim along with vmaf.
11174 Enables computing ms_ssim along with vmaf.
11177 Set the pool method (mean, min or harmonic mean) to be used for computing vmaf.
11180 Set number of threads to be used when computing vmaf.
11183 Set interval for frame subsampling used when computing vmaf.
11185 @item enable_conf_interval
11186 Enables confidence interval.
11189 This filter also supports the @ref{framesync} options.
11191 On the below examples the input file @file{main.mpg} being processed is
11192 compared with the reference file @file{ref.mpg}.
11195 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
11198 Example with options:
11200 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:enable-transform=1" -f null -
11205 Limits the pixel components values to the specified range [min, max].
11207 The filter accepts the following options:
11211 Lower bound. Defaults to the lowest allowed value for the input.
11214 Upper bound. Defaults to the highest allowed value for the input.
11217 Specify which planes will be processed. Defaults to all available.
11224 The filter accepts the following options:
11228 Set the number of loops. Setting this value to -1 will result in infinite loops.
11232 Set maximal size in number of frames. Default is 0.
11235 Set first frame of loop. Default is 0.
11240 Apply a 1D LUT to an input video.
11242 The filter accepts the following options:
11246 Set the 1D LUT file name.
11248 Currently supported formats:
11255 Select interpolation mode.
11257 Available values are:
11261 Use values from the nearest defined point.
11263 Interpolate values using the linear interpolation.
11265 Interpolate values using the cubic interpolation.
11272 Apply a 3D LUT to an input video.
11274 The filter accepts the following options:
11278 Set the 3D LUT file name.
11280 Currently supported formats:
11292 Select interpolation mode.
11294 Available values are:
11298 Use values from the nearest defined point.
11300 Interpolate values using the 8 points defining a cube.
11302 Interpolate values using a tetrahedron.
11306 This filter also supports the @ref{framesync} options.
11310 Turn certain luma values into transparency.
11312 The filter accepts the following options:
11316 Set the luma which will be used as base for transparency.
11317 Default value is @code{0}.
11320 Set the range of luma values to be keyed out.
11321 Default value is @code{0}.
11324 Set the range of softness. Default value is @code{0}.
11325 Use this to control gradual transition from zero to full transparency.
11328 @section lut, lutrgb, lutyuv
11330 Compute a look-up table for binding each pixel component input value
11331 to an output value, and apply it to the input video.
11333 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
11334 to an RGB input video.
11336 These filters accept the following parameters:
11339 set first pixel component expression
11341 set second pixel component expression
11343 set third pixel component expression
11345 set fourth pixel component expression, corresponds to the alpha component
11348 set red component expression
11350 set green component expression
11352 set blue component expression
11354 alpha component expression
11357 set Y/luminance component expression
11359 set U/Cb component expression
11361 set V/Cr component expression
11364 Each of them specifies the expression to use for computing the lookup table for
11365 the corresponding pixel component values.
11367 The exact component associated to each of the @var{c*} options depends on the
11370 The @var{lut} filter requires either YUV or RGB pixel formats in input,
11371 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
11373 The expressions can contain the following constants and functions:
11378 The input width and height.
11381 The input value for the pixel component.
11384 The input value, clipped to the @var{minval}-@var{maxval} range.
11387 The maximum value for the pixel component.
11390 The minimum value for the pixel component.
11393 The negated value for the pixel component value, clipped to the
11394 @var{minval}-@var{maxval} range; it corresponds to the expression
11395 "maxval-clipval+minval".
11398 The computed value in @var{val}, clipped to the
11399 @var{minval}-@var{maxval} range.
11401 @item gammaval(gamma)
11402 The computed gamma correction value of the pixel component value,
11403 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
11405 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
11409 All expressions default to "val".
11411 @subsection Examples
11415 Negate input video:
11417 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
11418 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
11421 The above is the same as:
11423 lutrgb="r=negval:g=negval:b=negval"
11424 lutyuv="y=negval:u=negval:v=negval"
11434 Remove chroma components, turning the video into a graytone image:
11436 lutyuv="u=128:v=128"
11440 Apply a luma burning effect:
11446 Remove green and blue components:
11452 Set a constant alpha channel value on input:
11454 format=rgba,lutrgb=a="maxval-minval/2"
11458 Correct luminance gamma by a factor of 0.5:
11460 lutyuv=y=gammaval(0.5)
11464 Discard least significant bits of luma:
11466 lutyuv=y='bitand(val, 128+64+32)'
11470 Technicolor like effect:
11472 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
11476 @section lut2, tlut2
11478 The @code{lut2} filter takes two input streams and outputs one
11481 The @code{tlut2} (time lut2) filter takes two consecutive frames
11482 from one single stream.
11484 This filter accepts the following parameters:
11487 set first pixel component expression
11489 set second pixel component expression
11491 set third pixel component expression
11493 set fourth pixel component expression, corresponds to the alpha component
11496 Each of them specifies the expression to use for computing the lookup table for
11497 the corresponding pixel component values.
11499 The exact component associated to each of the @var{c*} options depends on the
11502 The expressions can contain the following constants:
11507 The input width and height.
11510 The first input value for the pixel component.
11513 The second input value for the pixel component.
11516 The first input video bit depth.
11519 The second input video bit depth.
11522 All expressions default to "x".
11524 @subsection Examples
11528 Highlight differences between two RGB video streams:
11530 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)'
11534 Highlight differences between two YUV video streams:
11536 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)'
11540 Show max difference between two video streams:
11542 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)))'
11546 @section maskedclamp
11548 Clamp the first input stream with the second input and third input stream.
11550 Returns the value of first stream to be between second input
11551 stream - @code{undershoot} and third input stream + @code{overshoot}.
11553 This filter accepts the following options:
11556 Default value is @code{0}.
11559 Default value is @code{0}.
11562 Set which planes will be processed as bitmap, unprocessed planes will be
11563 copied from first stream.
11564 By default value 0xf, all planes will be processed.
11567 @section maskedmerge
11569 Merge the first input stream with the second input stream using per pixel
11570 weights in the third input stream.
11572 A value of 0 in the third stream pixel component means that pixel component
11573 from first stream is returned unchanged, while maximum value (eg. 255 for
11574 8-bit videos) means that pixel component from second stream is returned
11575 unchanged. Intermediate values define the amount of merging between both
11576 input stream's pixel components.
11578 This filter accepts the following options:
11581 Set which planes will be processed as bitmap, unprocessed planes will be
11582 copied from first stream.
11583 By default value 0xf, all planes will be processed.
11588 Apply motion-compensation deinterlacing.
11590 It needs one field per frame as input and must thus be used together
11591 with yadif=1/3 or equivalent.
11593 This filter accepts the following options:
11596 Set the deinterlacing mode.
11598 It accepts one of the following values:
11603 use iterative motion estimation
11605 like @samp{slow}, but use multiple reference frames.
11607 Default value is @samp{fast}.
11610 Set the picture field parity assumed for the input video. It must be
11611 one of the following values:
11615 assume top field first
11617 assume bottom field first
11620 Default value is @samp{bff}.
11623 Set per-block quantization parameter (QP) used by the internal
11626 Higher values should result in a smoother motion vector field but less
11627 optimal individual vectors. Default value is 1.
11630 @section mergeplanes
11632 Merge color channel components from several video streams.
11634 The filter accepts up to 4 input streams, and merge selected input
11635 planes to the output video.
11637 This filter accepts the following options:
11640 Set input to output plane mapping. Default is @code{0}.
11642 The mappings is specified as a bitmap. It should be specified as a
11643 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
11644 mapping for the first plane of the output stream. 'A' sets the number of
11645 the input stream to use (from 0 to 3), and 'a' the plane number of the
11646 corresponding input to use (from 0 to 3). The rest of the mappings is
11647 similar, 'Bb' describes the mapping for the output stream second
11648 plane, 'Cc' describes the mapping for the output stream third plane and
11649 'Dd' describes the mapping for the output stream fourth plane.
11652 Set output pixel format. Default is @code{yuva444p}.
11655 @subsection Examples
11659 Merge three gray video streams of same width and height into single video stream:
11661 [a0][a1][a2]mergeplanes=0x001020:yuv444p
11665 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
11667 [a0][a1]mergeplanes=0x00010210:yuva444p
11671 Swap Y and A plane in yuva444p stream:
11673 format=yuva444p,mergeplanes=0x03010200:yuva444p
11677 Swap U and V plane in yuv420p stream:
11679 format=yuv420p,mergeplanes=0x000201:yuv420p
11683 Cast a rgb24 clip to yuv444p:
11685 format=rgb24,mergeplanes=0x000102:yuv444p
11691 Estimate and export motion vectors using block matching algorithms.
11692 Motion vectors are stored in frame side data to be used by other filters.
11694 This filter accepts the following options:
11697 Specify the motion estimation method. Accepts one of the following values:
11701 Exhaustive search algorithm.
11703 Three step search algorithm.
11705 Two dimensional logarithmic search algorithm.
11707 New three step search algorithm.
11709 Four step search algorithm.
11711 Diamond search algorithm.
11713 Hexagon-based search algorithm.
11715 Enhanced predictive zonal search algorithm.
11717 Uneven multi-hexagon search algorithm.
11719 Default value is @samp{esa}.
11722 Macroblock size. Default @code{16}.
11725 Search parameter. Default @code{7}.
11728 @section midequalizer
11730 Apply Midway Image Equalization effect using two video streams.
11732 Midway Image Equalization adjusts a pair of images to have the same
11733 histogram, while maintaining their dynamics as much as possible. It's
11734 useful for e.g. matching exposures from a pair of stereo cameras.
11736 This filter has two inputs and one output, which must be of same pixel format, but
11737 may be of different sizes. The output of filter is first input adjusted with
11738 midway histogram of both inputs.
11740 This filter accepts the following option:
11744 Set which planes to process. Default is @code{15}, which is all available planes.
11747 @section minterpolate
11749 Convert the video to specified frame rate using motion interpolation.
11751 This filter accepts the following options:
11754 Specify the output frame rate. This can be rational e.g. @code{60000/1001}. Frames are dropped if @var{fps} is lower than source fps. Default @code{60}.
11757 Motion interpolation mode. Following values are accepted:
11760 Duplicate previous or next frame for interpolating new ones.
11762 Blend source frames. Interpolated frame is mean of previous and next frames.
11764 Motion compensated interpolation. Following options are effective when this mode is selected:
11768 Motion compensation mode. Following values are accepted:
11771 Overlapped block motion compensation.
11773 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
11775 Default mode is @samp{obmc}.
11778 Motion estimation mode. Following values are accepted:
11781 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
11783 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
11785 Default mode is @samp{bilat}.
11788 The algorithm to be used for motion estimation. Following values are accepted:
11791 Exhaustive search algorithm.
11793 Three step search algorithm.
11795 Two dimensional logarithmic search algorithm.
11797 New three step search algorithm.
11799 Four step search algorithm.
11801 Diamond search algorithm.
11803 Hexagon-based search algorithm.
11805 Enhanced predictive zonal search algorithm.
11807 Uneven multi-hexagon search algorithm.
11809 Default algorithm is @samp{epzs}.
11812 Macroblock size. Default @code{16}.
11815 Motion estimation search parameter. Default @code{32}.
11818 Enable variable-size block motion compensation. Motion estimation is applied with smaller block sizes at object boundaries in order to make the them less blur. Default is @code{0} (disabled).
11823 Scene change detection method. Scene change leads motion vectors to be in random direction. Scene change detection replace interpolated frames by duplicate ones. May not be needed for other modes. Following values are accepted:
11826 Disable scene change detection.
11828 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
11830 Default method is @samp{fdiff}.
11832 @item scd_threshold
11833 Scene change detection threshold. Default is @code{5.0}.
11838 Mix several video input streams into one video stream.
11840 A description of the accepted options follows.
11844 The number of inputs. If unspecified, it defaults to 2.
11847 Specify weight of each input video stream as sequence.
11848 Each weight is separated by space. If number of weights
11849 is smaller than number of @var{frames} last specified
11850 weight will be used for all remaining unset weights.
11853 Specify scale, if it is set it will be multiplied with sum
11854 of each weight multiplied with pixel values to give final destination
11855 pixel value. By default @var{scale} is auto scaled to sum of weights.
11858 Specify how end of stream is determined.
11861 The duration of the longest input. (default)
11864 The duration of the shortest input.
11867 The duration of the first input.
11871 @section mpdecimate
11873 Drop frames that do not differ greatly from the previous frame in
11874 order to reduce frame rate.
11876 The main use of this filter is for very-low-bitrate encoding
11877 (e.g. streaming over dialup modem), but it could in theory be used for
11878 fixing movies that were inverse-telecined incorrectly.
11880 A description of the accepted options follows.
11884 Set the maximum number of consecutive frames which can be dropped (if
11885 positive), or the minimum interval between dropped frames (if
11886 negative). If the value is 0, the frame is dropped disregarding the
11887 number of previous sequentially dropped frames.
11889 Default value is 0.
11894 Set the dropping threshold values.
11896 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
11897 represent actual pixel value differences, so a threshold of 64
11898 corresponds to 1 unit of difference for each pixel, or the same spread
11899 out differently over the block.
11901 A frame is a candidate for dropping if no 8x8 blocks differ by more
11902 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
11903 meaning the whole image) differ by more than a threshold of @option{lo}.
11905 Default value for @option{hi} is 64*12, default value for @option{lo} is
11906 64*5, and default value for @option{frac} is 0.33.
11912 Negate (invert) the input video.
11914 It accepts the following option:
11919 With value 1, it negates the alpha component, if present. Default value is 0.
11925 Denoise frames using Non-Local Means algorithm.
11927 Each pixel is adjusted by looking for other pixels with similar contexts. This
11928 context similarity is defined by comparing their surrounding patches of size
11929 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
11932 Note that the research area defines centers for patches, which means some
11933 patches will be made of pixels outside that research area.
11935 The filter accepts the following options.
11939 Set denoising strength.
11945 Same as @option{p} but for chroma planes.
11947 The default value is @var{0} and means automatic.
11953 Same as @option{r} but for chroma planes.
11955 The default value is @var{0} and means automatic.
11960 Deinterlace video using neural network edge directed interpolation.
11962 This filter accepts the following options:
11966 Mandatory option, without binary file filter can not work.
11967 Currently file can be found here:
11968 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
11971 Set which frames to deinterlace, by default it is @code{all}.
11972 Can be @code{all} or @code{interlaced}.
11975 Set mode of operation.
11977 Can be one of the following:
11981 Use frame flags, both fields.
11983 Use frame flags, single field.
11985 Use top field only.
11987 Use bottom field only.
11989 Use both fields, top first.
11991 Use both fields, bottom first.
11995 Set which planes to process, by default filter process all frames.
11998 Set size of local neighborhood around each pixel, used by the predictor neural
12001 Can be one of the following:
12014 Set the number of neurons in predictor neural network.
12015 Can be one of the following:
12026 Controls the number of different neural network predictions that are blended
12027 together to compute the final output value. Can be @code{fast}, default or
12031 Set which set of weights to use in the predictor.
12032 Can be one of the following:
12036 weights trained to minimize absolute error
12038 weights trained to minimize squared error
12042 Controls whether or not the prescreener neural network is used to decide
12043 which pixels should be processed by the predictor neural network and which
12044 can be handled by simple cubic interpolation.
12045 The prescreener is trained to know whether cubic interpolation will be
12046 sufficient for a pixel or whether it should be predicted by the predictor nn.
12047 The computational complexity of the prescreener nn is much less than that of
12048 the predictor nn. Since most pixels can be handled by cubic interpolation,
12049 using the prescreener generally results in much faster processing.
12050 The prescreener is pretty accurate, so the difference between using it and not
12051 using it is almost always unnoticeable.
12053 Can be one of the following:
12061 Default is @code{new}.
12064 Set various debugging flags.
12069 Force libavfilter not to use any of the specified pixel formats for the
12070 input to the next filter.
12072 It accepts the following parameters:
12076 A '|'-separated list of pixel format names, such as
12077 pix_fmts=yuv420p|monow|rgb24".
12081 @subsection Examples
12085 Force libavfilter to use a format different from @var{yuv420p} for the
12086 input to the vflip filter:
12088 noformat=pix_fmts=yuv420p,vflip
12092 Convert the input video to any of the formats not contained in the list:
12094 noformat=yuv420p|yuv444p|yuv410p
12100 Add noise on video input frame.
12102 The filter accepts the following options:
12110 Set noise seed for specific pixel component or all pixel components in case
12111 of @var{all_seed}. Default value is @code{123457}.
12113 @item all_strength, alls
12114 @item c0_strength, c0s
12115 @item c1_strength, c1s
12116 @item c2_strength, c2s
12117 @item c3_strength, c3s
12118 Set noise strength for specific pixel component or all pixel components in case
12119 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
12121 @item all_flags, allf
12122 @item c0_flags, c0f
12123 @item c1_flags, c1f
12124 @item c2_flags, c2f
12125 @item c3_flags, c3f
12126 Set pixel component flags or set flags for all components if @var{all_flags}.
12127 Available values for component flags are:
12130 averaged temporal noise (smoother)
12132 mix random noise with a (semi)regular pattern
12134 temporal noise (noise pattern changes between frames)
12136 uniform noise (gaussian otherwise)
12140 @subsection Examples
12142 Add temporal and uniform noise to input video:
12144 noise=alls=20:allf=t+u
12149 Normalize RGB video (aka histogram stretching, contrast stretching).
12150 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
12152 For each channel of each frame, the filter computes the input range and maps
12153 it linearly to the user-specified output range. The output range defaults
12154 to the full dynamic range from pure black to pure white.
12156 Temporal smoothing can be used on the input range to reduce flickering (rapid
12157 changes in brightness) caused when small dark or bright objects enter or leave
12158 the scene. This is similar to the auto-exposure (automatic gain control) on a
12159 video camera, and, like a video camera, it may cause a period of over- or
12160 under-exposure of the video.
12162 The R,G,B channels can be normalized independently, which may cause some
12163 color shifting, or linked together as a single channel, which prevents
12164 color shifting. Linked normalization preserves hue. Independent normalization
12165 does not, so it can be used to remove some color casts. Independent and linked
12166 normalization can be combined in any ratio.
12168 The normalize filter accepts the following options:
12173 Colors which define the output range. The minimum input value is mapped to
12174 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
12175 The defaults are black and white respectively. Specifying white for
12176 @var{blackpt} and black for @var{whitept} will give color-inverted,
12177 normalized video. Shades of grey can be used to reduce the dynamic range
12178 (contrast). Specifying saturated colors here can create some interesting
12182 The number of previous frames to use for temporal smoothing. The input range
12183 of each channel is smoothed using a rolling average over the current frame
12184 and the @var{smoothing} previous frames. The default is 0 (no temporal
12188 Controls the ratio of independent (color shifting) channel normalization to
12189 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
12190 independent. Defaults to 1.0 (fully independent).
12193 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
12194 expensive no-op. Defaults to 1.0 (full strength).
12198 @subsection Examples
12200 Stretch video contrast to use the full dynamic range, with no temporal
12201 smoothing; may flicker depending on the source content:
12203 normalize=blackpt=black:whitept=white:smoothing=0
12206 As above, but with 50 frames of temporal smoothing; flicker should be
12207 reduced, depending on the source content:
12209 normalize=blackpt=black:whitept=white:smoothing=50
12212 As above, but with hue-preserving linked channel normalization:
12214 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
12217 As above, but with half strength:
12219 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
12222 Map the darkest input color to red, the brightest input color to cyan:
12224 normalize=blackpt=red:whitept=cyan
12229 Pass the video source unchanged to the output.
12232 Optical Character Recognition
12234 This filter uses Tesseract for optical character recognition. To enable
12235 compilation of this filter, you need to configure FFmpeg with
12236 @code{--enable-libtesseract}.
12238 It accepts the following options:
12242 Set datapath to tesseract data. Default is to use whatever was
12243 set at installation.
12246 Set language, default is "eng".
12249 Set character whitelist.
12252 Set character blacklist.
12255 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
12259 Apply a video transform using libopencv.
12261 To enable this filter, install the libopencv library and headers and
12262 configure FFmpeg with @code{--enable-libopencv}.
12264 It accepts the following parameters:
12269 The name of the libopencv filter to apply.
12271 @item filter_params
12272 The parameters to pass to the libopencv filter. If not specified, the default
12273 values are assumed.
12277 Refer to the official libopencv documentation for more precise
12279 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
12281 Several libopencv filters are supported; see the following subsections.
12286 Dilate an image by using a specific structuring element.
12287 It corresponds to the libopencv function @code{cvDilate}.
12289 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
12291 @var{struct_el} represents a structuring element, and has the syntax:
12292 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
12294 @var{cols} and @var{rows} represent the number of columns and rows of
12295 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
12296 point, and @var{shape} the shape for the structuring element. @var{shape}
12297 must be "rect", "cross", "ellipse", or "custom".
12299 If the value for @var{shape} is "custom", it must be followed by a
12300 string of the form "=@var{filename}". The file with name
12301 @var{filename} is assumed to represent a binary image, with each
12302 printable character corresponding to a bright pixel. When a custom
12303 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
12304 or columns and rows of the read file are assumed instead.
12306 The default value for @var{struct_el} is "3x3+0x0/rect".
12308 @var{nb_iterations} specifies the number of times the transform is
12309 applied to the image, and defaults to 1.
12313 # Use the default values
12316 # Dilate using a structuring element with a 5x5 cross, iterating two times
12317 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
12319 # Read the shape from the file diamond.shape, iterating two times.
12320 # The file diamond.shape may contain a pattern of characters like this
12326 # The specified columns and rows are ignored
12327 # but the anchor point coordinates are not
12328 ocv=dilate:0x0+2x2/custom=diamond.shape|2
12333 Erode an image by using a specific structuring element.
12334 It corresponds to the libopencv function @code{cvErode}.
12336 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
12337 with the same syntax and semantics as the @ref{dilate} filter.
12341 Smooth the input video.
12343 The filter takes the following parameters:
12344 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
12346 @var{type} is the type of smooth filter to apply, and must be one of
12347 the following values: "blur", "blur_no_scale", "median", "gaussian",
12348 or "bilateral". The default value is "gaussian".
12350 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
12351 depend on the smooth type. @var{param1} and
12352 @var{param2} accept integer positive values or 0. @var{param3} and
12353 @var{param4} accept floating point values.
12355 The default value for @var{param1} is 3. The default value for the
12356 other parameters is 0.
12358 These parameters correspond to the parameters assigned to the
12359 libopencv function @code{cvSmooth}.
12361 @section oscilloscope
12363 2D Video Oscilloscope.
12365 Useful to measure spatial impulse, step responses, chroma delays, etc.
12367 It accepts the following parameters:
12371 Set scope center x position.
12374 Set scope center y position.
12377 Set scope size, relative to frame diagonal.
12380 Set scope tilt/rotation.
12386 Set trace center x position.
12389 Set trace center y position.
12392 Set trace width, relative to width of frame.
12395 Set trace height, relative to height of frame.
12398 Set which components to trace. By default it traces first three components.
12401 Draw trace grid. By default is enabled.
12404 Draw some statistics. By default is enabled.
12407 Draw scope. By default is enabled.
12410 @subsection Examples
12414 Inspect full first row of video frame.
12416 oscilloscope=x=0.5:y=0:s=1
12420 Inspect full last row of video frame.
12422 oscilloscope=x=0.5:y=1:s=1
12426 Inspect full 5th line of video frame of height 1080.
12428 oscilloscope=x=0.5:y=5/1080:s=1
12432 Inspect full last column of video frame.
12434 oscilloscope=x=1:y=0.5:s=1:t=1
12442 Overlay one video on top of another.
12444 It takes two inputs and has one output. The first input is the "main"
12445 video on which the second input is overlaid.
12447 It accepts the following parameters:
12449 A description of the accepted options follows.
12454 Set the expression for the x and y coordinates of the overlaid video
12455 on the main video. Default value is "0" for both expressions. In case
12456 the expression is invalid, it is set to a huge value (meaning that the
12457 overlay will not be displayed within the output visible area).
12460 See @ref{framesync}.
12463 Set when the expressions for @option{x}, and @option{y} are evaluated.
12465 It accepts the following values:
12468 only evaluate expressions once during the filter initialization or
12469 when a command is processed
12472 evaluate expressions for each incoming frame
12475 Default value is @samp{frame}.
12478 See @ref{framesync}.
12481 Set the format for the output video.
12483 It accepts the following values:
12486 force YUV420 output
12489 force YUV422 output
12492 force YUV444 output
12495 force packed RGB output
12498 force planar RGB output
12501 automatically pick format
12504 Default value is @samp{yuv420}.
12507 See @ref{framesync}.
12510 Set format of alpha of the overlaid video, it can be @var{straight} or
12511 @var{premultiplied}. Default is @var{straight}.
12514 The @option{x}, and @option{y} expressions can contain the following
12520 The main input width and height.
12524 The overlay input width and height.
12528 The computed values for @var{x} and @var{y}. They are evaluated for
12533 horizontal and vertical chroma subsample values of the output
12534 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
12538 the number of input frame, starting from 0
12541 the position in the file of the input frame, NAN if unknown
12544 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
12548 This filter also supports the @ref{framesync} options.
12550 Note that the @var{n}, @var{pos}, @var{t} variables are available only
12551 when evaluation is done @emph{per frame}, and will evaluate to NAN
12552 when @option{eval} is set to @samp{init}.
12554 Be aware that frames are taken from each input video in timestamp
12555 order, hence, if their initial timestamps differ, it is a good idea
12556 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
12557 have them begin in the same zero timestamp, as the example for
12558 the @var{movie} filter does.
12560 You can chain together more overlays but you should test the
12561 efficiency of such approach.
12563 @subsection Commands
12565 This filter supports the following commands:
12569 Modify the x and y of the overlay input.
12570 The command accepts the same syntax of the corresponding option.
12572 If the specified expression is not valid, it is kept at its current
12576 @subsection Examples
12580 Draw the overlay at 10 pixels from the bottom right corner of the main
12583 overlay=main_w-overlay_w-10:main_h-overlay_h-10
12586 Using named options the example above becomes:
12588 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
12592 Insert a transparent PNG logo in the bottom left corner of the input,
12593 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
12595 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
12599 Insert 2 different transparent PNG logos (second logo on bottom
12600 right corner) using the @command{ffmpeg} tool:
12602 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
12606 Add a transparent color layer on top of the main video; @code{WxH}
12607 must specify the size of the main input to the overlay filter:
12609 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
12613 Play an original video and a filtered version (here with the deshake
12614 filter) side by side using the @command{ffplay} tool:
12616 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
12619 The above command is the same as:
12621 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
12625 Make a sliding overlay appearing from the left to the right top part of the
12626 screen starting since time 2:
12628 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
12632 Compose output by putting two input videos side to side:
12634 ffmpeg -i left.avi -i right.avi -filter_complex "
12635 nullsrc=size=200x100 [background];
12636 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
12637 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
12638 [background][left] overlay=shortest=1 [background+left];
12639 [background+left][right] overlay=shortest=1:x=100 [left+right]
12644 Mask 10-20 seconds of a video by applying the delogo filter to a section
12646 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
12647 -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]'
12652 Chain several overlays in cascade:
12654 nullsrc=s=200x200 [bg];
12655 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
12656 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
12657 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
12658 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
12659 [in3] null, [mid2] overlay=100:100 [out0]
12666 Apply Overcomplete Wavelet denoiser.
12668 The filter accepts the following options:
12674 Larger depth values will denoise lower frequency components more, but
12675 slow down filtering.
12677 Must be an int in the range 8-16, default is @code{8}.
12679 @item luma_strength, ls
12682 Must be a double value in the range 0-1000, default is @code{1.0}.
12684 @item chroma_strength, cs
12685 Set chroma strength.
12687 Must be a double value in the range 0-1000, default is @code{1.0}.
12693 Add paddings to the input image, and place the original input at the
12694 provided @var{x}, @var{y} coordinates.
12696 It accepts the following parameters:
12701 Specify an expression for the size of the output image with the
12702 paddings added. If the value for @var{width} or @var{height} is 0, the
12703 corresponding input size is used for the output.
12705 The @var{width} expression can reference the value set by the
12706 @var{height} expression, and vice versa.
12708 The default value of @var{width} and @var{height} is 0.
12712 Specify the offsets to place the input image at within the padded area,
12713 with respect to the top/left border of the output image.
12715 The @var{x} expression can reference the value set by the @var{y}
12716 expression, and vice versa.
12718 The default value of @var{x} and @var{y} is 0.
12720 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
12721 so the input image is centered on the padded area.
12724 Specify the color of the padded area. For the syntax of this option,
12725 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
12726 manual,ffmpeg-utils}.
12728 The default value of @var{color} is "black".
12731 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
12733 It accepts the following values:
12737 Only evaluate expressions once during the filter initialization or when
12738 a command is processed.
12741 Evaluate expressions for each incoming frame.
12745 Default value is @samp{init}.
12748 Pad to aspect instead to a resolution.
12752 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
12753 options are expressions containing the following constants:
12758 The input video width and height.
12762 These are the same as @var{in_w} and @var{in_h}.
12766 The output width and height (the size of the padded area), as
12767 specified by the @var{width} and @var{height} expressions.
12771 These are the same as @var{out_w} and @var{out_h}.
12775 The x and y offsets as specified by the @var{x} and @var{y}
12776 expressions, or NAN if not yet specified.
12779 same as @var{iw} / @var{ih}
12782 input sample aspect ratio
12785 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
12789 The horizontal and vertical chroma subsample values. For example for the
12790 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
12793 @subsection Examples
12797 Add paddings with the color "violet" to the input video. The output video
12798 size is 640x480, and the top-left corner of the input video is placed at
12801 pad=640:480:0:40:violet
12804 The example above is equivalent to the following command:
12806 pad=width=640:height=480:x=0:y=40:color=violet
12810 Pad the input to get an output with dimensions increased by 3/2,
12811 and put the input video at the center of the padded area:
12813 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
12817 Pad the input to get a squared output with size equal to the maximum
12818 value between the input width and height, and put the input video at
12819 the center of the padded area:
12821 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
12825 Pad the input to get a final w/h ratio of 16:9:
12827 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
12831 In case of anamorphic video, in order to set the output display aspect
12832 correctly, it is necessary to use @var{sar} in the expression,
12833 according to the relation:
12835 (ih * X / ih) * sar = output_dar
12836 X = output_dar / sar
12839 Thus the previous example needs to be modified to:
12841 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
12845 Double the output size and put the input video in the bottom-right
12846 corner of the output padded area:
12848 pad="2*iw:2*ih:ow-iw:oh-ih"
12852 @anchor{palettegen}
12853 @section palettegen
12855 Generate one palette for a whole video stream.
12857 It accepts the following options:
12861 Set the maximum number of colors to quantize in the palette.
12862 Note: the palette will still contain 256 colors; the unused palette entries
12865 @item reserve_transparent
12866 Create a palette of 255 colors maximum and reserve the last one for
12867 transparency. Reserving the transparency color is useful for GIF optimization.
12868 If not set, the maximum of colors in the palette will be 256. You probably want
12869 to disable this option for a standalone image.
12872 @item transparency_color
12873 Set the color that will be used as background for transparency.
12876 Set statistics mode.
12878 It accepts the following values:
12881 Compute full frame histograms.
12883 Compute histograms only for the part that differs from previous frame. This
12884 might be relevant to give more importance to the moving part of your input if
12885 the background is static.
12887 Compute new histogram for each frame.
12890 Default value is @var{full}.
12893 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
12894 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
12895 color quantization of the palette. This information is also visible at
12896 @var{info} logging level.
12898 @subsection Examples
12902 Generate a representative palette of a given video using @command{ffmpeg}:
12904 ffmpeg -i input.mkv -vf palettegen palette.png
12908 @section paletteuse
12910 Use a palette to downsample an input video stream.
12912 The filter takes two inputs: one video stream and a palette. The palette must
12913 be a 256 pixels image.
12915 It accepts the following options:
12919 Select dithering mode. Available algorithms are:
12922 Ordered 8x8 bayer dithering (deterministic)
12924 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
12925 Note: this dithering is sometimes considered "wrong" and is included as a
12927 @item floyd_steinberg
12928 Floyd and Steingberg dithering (error diffusion)
12930 Frankie Sierra dithering v2 (error diffusion)
12932 Frankie Sierra dithering v2 "Lite" (error diffusion)
12935 Default is @var{sierra2_4a}.
12938 When @var{bayer} dithering is selected, this option defines the scale of the
12939 pattern (how much the crosshatch pattern is visible). A low value means more
12940 visible pattern for less banding, and higher value means less visible pattern
12941 at the cost of more banding.
12943 The option must be an integer value in the range [0,5]. Default is @var{2}.
12946 If set, define the zone to process
12950 Only the changing rectangle will be reprocessed. This is similar to GIF
12951 cropping/offsetting compression mechanism. This option can be useful for speed
12952 if only a part of the image is changing, and has use cases such as limiting the
12953 scope of the error diffusal @option{dither} to the rectangle that bounds the
12954 moving scene (it leads to more deterministic output if the scene doesn't change
12955 much, and as a result less moving noise and better GIF compression).
12958 Default is @var{none}.
12961 Take new palette for each output frame.
12963 @item alpha_threshold
12964 Sets the alpha threshold for transparency. Alpha values above this threshold
12965 will be treated as completely opaque, and values below this threshold will be
12966 treated as completely transparent.
12968 The option must be an integer value in the range [0,255]. Default is @var{128}.
12971 @subsection Examples
12975 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
12976 using @command{ffmpeg}:
12978 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
12982 @section perspective
12984 Correct perspective of video not recorded perpendicular to the screen.
12986 A description of the accepted parameters follows.
12997 Set coordinates expression for top left, top right, bottom left and bottom right corners.
12998 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
12999 If the @code{sense} option is set to @code{source}, then the specified points will be sent
13000 to the corners of the destination. If the @code{sense} option is set to @code{destination},
13001 then the corners of the source will be sent to the specified coordinates.
13003 The expressions can use the following variables:
13008 the width and height of video frame.
13012 Output frame count.
13015 @item interpolation
13016 Set interpolation for perspective correction.
13018 It accepts the following values:
13024 Default value is @samp{linear}.
13027 Set interpretation of coordinate options.
13029 It accepts the following values:
13033 Send point in the source specified by the given coordinates to
13034 the corners of the destination.
13036 @item 1, destination
13038 Send the corners of the source to the point in the destination specified
13039 by the given coordinates.
13041 Default value is @samp{source}.
13045 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
13047 It accepts the following values:
13050 only evaluate expressions once during the filter initialization or
13051 when a command is processed
13054 evaluate expressions for each incoming frame
13057 Default value is @samp{init}.
13062 Delay interlaced video by one field time so that the field order changes.
13064 The intended use is to fix PAL movies that have been captured with the
13065 opposite field order to the film-to-video transfer.
13067 A description of the accepted parameters follows.
13073 It accepts the following values:
13076 Capture field order top-first, transfer bottom-first.
13077 Filter will delay the bottom field.
13080 Capture field order bottom-first, transfer top-first.
13081 Filter will delay the top field.
13084 Capture and transfer with the same field order. This mode only exists
13085 for the documentation of the other options to refer to, but if you
13086 actually select it, the filter will faithfully do nothing.
13089 Capture field order determined automatically by field flags, transfer
13091 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
13092 basis using field flags. If no field information is available,
13093 then this works just like @samp{u}.
13096 Capture unknown or varying, transfer opposite.
13097 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
13098 analyzing the images and selecting the alternative that produces best
13099 match between the fields.
13102 Capture top-first, transfer unknown or varying.
13103 Filter selects among @samp{t} and @samp{p} using image analysis.
13106 Capture bottom-first, transfer unknown or varying.
13107 Filter selects among @samp{b} and @samp{p} using image analysis.
13110 Capture determined by field flags, transfer unknown or varying.
13111 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
13112 image analysis. If no field information is available, then this works just
13113 like @samp{U}. This is the default mode.
13116 Both capture and transfer unknown or varying.
13117 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
13121 @section pixdesctest
13123 Pixel format descriptor test filter, mainly useful for internal
13124 testing. The output video should be equal to the input video.
13128 format=monow, pixdesctest
13131 can be used to test the monowhite pixel format descriptor definition.
13135 Display sample values of color channels. Mainly useful for checking color
13136 and levels. Minimum supported resolution is 640x480.
13138 The filters accept the following options:
13142 Set scope X position, relative offset on X axis.
13145 Set scope Y position, relative offset on Y axis.
13154 Set window opacity. This window also holds statistics about pixel area.
13157 Set window X position, relative offset on X axis.
13160 Set window Y position, relative offset on Y axis.
13165 Enable the specified chain of postprocessing subfilters using libpostproc. This
13166 library should be automatically selected with a GPL build (@code{--enable-gpl}).
13167 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
13168 Each subfilter and some options have a short and a long name that can be used
13169 interchangeably, i.e. dr/dering are the same.
13171 The filters accept the following options:
13175 Set postprocessing subfilters string.
13178 All subfilters share common options to determine their scope:
13182 Honor the quality commands for this subfilter.
13185 Do chrominance filtering, too (default).
13188 Do luminance filtering only (no chrominance).
13191 Do chrominance filtering only (no luminance).
13194 These options can be appended after the subfilter name, separated by a '|'.
13196 Available subfilters are:
13199 @item hb/hdeblock[|difference[|flatness]]
13200 Horizontal deblocking filter
13203 Difference factor where higher values mean more deblocking (default: @code{32}).
13205 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13208 @item vb/vdeblock[|difference[|flatness]]
13209 Vertical deblocking filter
13212 Difference factor where higher values mean more deblocking (default: @code{32}).
13214 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13217 @item ha/hadeblock[|difference[|flatness]]
13218 Accurate horizontal deblocking filter
13221 Difference factor where higher values mean more deblocking (default: @code{32}).
13223 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13226 @item va/vadeblock[|difference[|flatness]]
13227 Accurate vertical deblocking filter
13230 Difference factor where higher values mean more deblocking (default: @code{32}).
13232 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13236 The horizontal and vertical deblocking filters share the difference and
13237 flatness values so you cannot set different horizontal and vertical
13241 @item h1/x1hdeblock
13242 Experimental horizontal deblocking filter
13244 @item v1/x1vdeblock
13245 Experimental vertical deblocking filter
13250 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
13253 larger -> stronger filtering
13255 larger -> stronger filtering
13257 larger -> stronger filtering
13260 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
13263 Stretch luminance to @code{0-255}.
13266 @item lb/linblenddeint
13267 Linear blend deinterlacing filter that deinterlaces the given block by
13268 filtering all lines with a @code{(1 2 1)} filter.
13270 @item li/linipoldeint
13271 Linear interpolating deinterlacing filter that deinterlaces the given block by
13272 linearly interpolating every second line.
13274 @item ci/cubicipoldeint
13275 Cubic interpolating deinterlacing filter deinterlaces the given block by
13276 cubically interpolating every second line.
13278 @item md/mediandeint
13279 Median deinterlacing filter that deinterlaces the given block by applying a
13280 median filter to every second line.
13282 @item fd/ffmpegdeint
13283 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
13284 second line with a @code{(-1 4 2 4 -1)} filter.
13287 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
13288 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
13290 @item fq/forceQuant[|quantizer]
13291 Overrides the quantizer table from the input with the constant quantizer you
13299 Default pp filter combination (@code{hb|a,vb|a,dr|a})
13302 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
13305 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
13308 @subsection Examples
13312 Apply horizontal and vertical deblocking, deringing and automatic
13313 brightness/contrast:
13319 Apply default filters without brightness/contrast correction:
13325 Apply default filters and temporal denoiser:
13327 pp=default/tmpnoise|1|2|3
13331 Apply deblocking on luminance only, and switch vertical deblocking on or off
13332 automatically depending on available CPU time:
13339 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
13340 similar to spp = 6 with 7 point DCT, where only the center sample is
13343 The filter accepts the following options:
13347 Force a constant quantization parameter. It accepts an integer in range
13348 0 to 63. If not set, the filter will use the QP from the video stream
13352 Set thresholding mode. Available modes are:
13356 Set hard thresholding.
13358 Set soft thresholding (better de-ringing effect, but likely blurrier).
13360 Set medium thresholding (good results, default).
13364 @section premultiply
13365 Apply alpha premultiply effect to input video stream using first plane
13366 of second stream as alpha.
13368 Both streams must have same dimensions and same pixel format.
13370 The filter accepts the following option:
13374 Set which planes will be processed, unprocessed planes will be copied.
13375 By default value 0xf, all planes will be processed.
13378 Do not require 2nd input for processing, instead use alpha plane from input stream.
13382 Apply prewitt operator to input video stream.
13384 The filter accepts the following option:
13388 Set which planes will be processed, unprocessed planes will be copied.
13389 By default value 0xf, all planes will be processed.
13392 Set value which will be multiplied with filtered result.
13395 Set value which will be added to filtered result.
13398 @anchor{program_opencl}
13399 @section program_opencl
13401 Filter video using an OpenCL program.
13406 OpenCL program source file.
13409 Kernel name in program.
13412 Number of inputs to the filter. Defaults to 1.
13415 Size of output frames. Defaults to the same as the first input.
13419 The program source file must contain a kernel function with the given name,
13420 which will be run once for each plane of the output. Each run on a plane
13421 gets enqueued as a separate 2D global NDRange with one work-item for each
13422 pixel to be generated. The global ID offset for each work-item is therefore
13423 the coordinates of a pixel in the destination image.
13425 The kernel function needs to take the following arguments:
13428 Destination image, @var{__write_only image2d_t}.
13430 This image will become the output; the kernel should write all of it.
13432 Frame index, @var{unsigned int}.
13434 This is a counter starting from zero and increasing by one for each frame.
13436 Source images, @var{__read_only image2d_t}.
13438 These are the most recent images on each input. The kernel may read from
13439 them to generate the output, but they can't be written to.
13446 Copy the input to the output (output must be the same size as the input).
13448 __kernel void copy(__write_only image2d_t destination,
13449 unsigned int index,
13450 __read_only image2d_t source)
13452 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
13454 int2 location = (int2)(get_global_id(0), get_global_id(1));
13456 float4 value = read_imagef(source, sampler, location);
13458 write_imagef(destination, location, value);
13463 Apply a simple transformation, rotating the input by an amount increasing
13464 with the index counter. Pixel values are linearly interpolated by the
13465 sampler, and the output need not have the same dimensions as the input.
13467 __kernel void rotate_image(__write_only image2d_t dst,
13468 unsigned int index,
13469 __read_only image2d_t src)
13471 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
13472 CLK_FILTER_LINEAR);
13474 float angle = (float)index / 100.0f;
13476 float2 dst_dim = convert_float2(get_image_dim(dst));
13477 float2 src_dim = convert_float2(get_image_dim(src));
13479 float2 dst_cen = dst_dim / 2.0f;
13480 float2 src_cen = src_dim / 2.0f;
13482 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
13484 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
13486 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
13487 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
13489 src_pos = src_pos * src_dim / dst_dim;
13491 float2 src_loc = src_pos + src_cen;
13493 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
13494 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
13495 write_imagef(dst, dst_loc, 0.5f);
13497 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
13502 Blend two inputs together, with the amount of each input used varying
13503 with the index counter.
13505 __kernel void blend_images(__write_only image2d_t dst,
13506 unsigned int index,
13507 __read_only image2d_t src1,
13508 __read_only image2d_t src2)
13510 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
13511 CLK_FILTER_LINEAR);
13513 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
13515 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
13516 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
13517 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
13519 float4 val1 = read_imagef(src1, sampler, src1_loc);
13520 float4 val2 = read_imagef(src2, sampler, src2_loc);
13522 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
13528 @section pseudocolor
13530 Alter frame colors in video with pseudocolors.
13532 This filter accept the following options:
13536 set pixel first component expression
13539 set pixel second component expression
13542 set pixel third component expression
13545 set pixel fourth component expression, corresponds to the alpha component
13548 set component to use as base for altering colors
13551 Each of them specifies the expression to use for computing the lookup table for
13552 the corresponding pixel component values.
13554 The expressions can contain the following constants and functions:
13559 The input width and height.
13562 The input value for the pixel component.
13564 @item ymin, umin, vmin, amin
13565 The minimum allowed component value.
13567 @item ymax, umax, vmax, amax
13568 The maximum allowed component value.
13571 All expressions default to "val".
13573 @subsection Examples
13577 Change too high luma values to gradient:
13579 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'"
13585 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
13586 Ratio) between two input videos.
13588 This filter takes in input two input videos, the first input is
13589 considered the "main" source and is passed unchanged to the
13590 output. The second input is used as a "reference" video for computing
13593 Both video inputs must have the same resolution and pixel format for
13594 this filter to work correctly. Also it assumes that both inputs
13595 have the same number of frames, which are compared one by one.
13597 The obtained average PSNR is printed through the logging system.
13599 The filter stores the accumulated MSE (mean squared error) of each
13600 frame, and at the end of the processing it is averaged across all frames
13601 equally, and the following formula is applied to obtain the PSNR:
13604 PSNR = 10*log10(MAX^2/MSE)
13607 Where MAX is the average of the maximum values of each component of the
13610 The description of the accepted parameters follows.
13613 @item stats_file, f
13614 If specified the filter will use the named file to save the PSNR of
13615 each individual frame. When filename equals "-" the data is sent to
13618 @item stats_version
13619 Specifies which version of the stats file format to use. Details of
13620 each format are written below.
13621 Default value is 1.
13623 @item stats_add_max
13624 Determines whether the max value is output to the stats log.
13625 Default value is 0.
13626 Requires stats_version >= 2. If this is set and stats_version < 2,
13627 the filter will return an error.
13630 This filter also supports the @ref{framesync} options.
13632 The file printed if @var{stats_file} is selected, contains a sequence of
13633 key/value pairs of the form @var{key}:@var{value} for each compared
13636 If a @var{stats_version} greater than 1 is specified, a header line precedes
13637 the list of per-frame-pair stats, with key value pairs following the frame
13638 format with the following parameters:
13641 @item psnr_log_version
13642 The version of the log file format. Will match @var{stats_version}.
13645 A comma separated list of the per-frame-pair parameters included in
13649 A description of each shown per-frame-pair parameter follows:
13653 sequential number of the input frame, starting from 1
13656 Mean Square Error pixel-by-pixel average difference of the compared
13657 frames, averaged over all the image components.
13659 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
13660 Mean Square Error pixel-by-pixel average difference of the compared
13661 frames for the component specified by the suffix.
13663 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
13664 Peak Signal to Noise ratio of the compared frames for the component
13665 specified by the suffix.
13667 @item max_avg, max_y, max_u, max_v
13668 Maximum allowed value for each channel, and average over all
13674 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
13675 [main][ref] psnr="stats_file=stats.log" [out]
13678 On this example the input file being processed is compared with the
13679 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
13680 is stored in @file{stats.log}.
13685 Pulldown reversal (inverse telecine) filter, capable of handling mixed
13686 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
13689 The pullup filter is designed to take advantage of future context in making
13690 its decisions. This filter is stateless in the sense that it does not lock
13691 onto a pattern to follow, but it instead looks forward to the following
13692 fields in order to identify matches and rebuild progressive frames.
13694 To produce content with an even framerate, insert the fps filter after
13695 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
13696 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
13698 The filter accepts the following options:
13705 These options set the amount of "junk" to ignore at the left, right, top, and
13706 bottom of the image, respectively. Left and right are in units of 8 pixels,
13707 while top and bottom are in units of 2 lines.
13708 The default is 8 pixels on each side.
13711 Set the strict breaks. Setting this option to 1 will reduce the chances of
13712 filter generating an occasional mismatched frame, but it may also cause an
13713 excessive number of frames to be dropped during high motion sequences.
13714 Conversely, setting it to -1 will make filter match fields more easily.
13715 This may help processing of video where there is slight blurring between
13716 the fields, but may also cause there to be interlaced frames in the output.
13717 Default value is @code{0}.
13720 Set the metric plane to use. It accepts the following values:
13726 Use chroma blue plane.
13729 Use chroma red plane.
13732 This option may be set to use chroma plane instead of the default luma plane
13733 for doing filter's computations. This may improve accuracy on very clean
13734 source material, but more likely will decrease accuracy, especially if there
13735 is chroma noise (rainbow effect) or any grayscale video.
13736 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
13737 load and make pullup usable in realtime on slow machines.
13740 For best results (without duplicated frames in the output file) it is
13741 necessary to change the output frame rate. For example, to inverse
13742 telecine NTSC input:
13744 ffmpeg -i input -vf pullup -r 24000/1001 ...
13749 Change video quantization parameters (QP).
13751 The filter accepts the following option:
13755 Set expression for quantization parameter.
13758 The expression is evaluated through the eval API and can contain, among others,
13759 the following constants:
13763 1 if index is not 129, 0 otherwise.
13766 Sequential index starting from -129 to 128.
13769 @subsection Examples
13773 Some equation like:
13781 Flush video frames from internal cache of frames into a random order.
13782 No frame is discarded.
13783 Inspired by @ref{frei0r} nervous filter.
13787 Set size in number of frames of internal cache, in range from @code{2} to
13788 @code{512}. Default is @code{30}.
13791 Set seed for random number generator, must be an integer included between
13792 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
13793 less than @code{0}, the filter will try to use a good random seed on a
13797 @section readeia608
13799 Read closed captioning (EIA-608) information from the top lines of a video frame.
13801 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
13802 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
13803 with EIA-608 data (starting from 0). A description of each metadata value follows:
13806 @item lavfi.readeia608.X.cc
13807 The two bytes stored as EIA-608 data (printed in hexadecimal).
13809 @item lavfi.readeia608.X.line
13810 The number of the line on which the EIA-608 data was identified and read.
13813 This filter accepts the following options:
13817 Set the line to start scanning for EIA-608 data. Default is @code{0}.
13820 Set the line to end scanning for EIA-608 data. Default is @code{29}.
13823 Set minimal acceptable amplitude change for sync codes detection.
13824 Default is @code{0.2}. Allowed range is @code{[0.001 - 1]}.
13827 Set the ratio of width reserved for sync code detection.
13828 Default is @code{0.27}. Allowed range is @code{[0.01 - 0.7]}.
13831 Set the max peaks height difference for sync code detection.
13832 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
13835 Set max peaks period difference for sync code detection.
13836 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
13839 Set the first two max start code bits differences.
13840 Default is @code{0.02}. Allowed range is @code{[0.0 - 0.5]}.
13843 Set the minimum ratio of bits height compared to 3rd start code bit.
13844 Default is @code{0.75}. Allowed range is @code{[0.01 - 1]}.
13847 Set the white color threshold. Default is @code{0.35}. Allowed range is @code{[0.1 - 1]}.
13850 Set the black color threshold. Default is @code{0.15}. Allowed range is @code{[0.0 - 0.5]}.
13853 Enable checking the parity bit. In the event of a parity error, the filter will output
13854 @code{0x00} for that character. Default is false.
13857 @subsection Examples
13861 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
13863 ffprobe -f lavfi -i movie=captioned_video.mov,readeia608 -show_entries frame=pkt_pts_time:frame_tags=lavfi.readeia608.0.cc,lavfi.readeia608.1.cc -of csv
13869 Read vertical interval timecode (VITC) information from the top lines of a
13872 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
13873 timecode value, if a valid timecode has been detected. Further metadata key
13874 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
13875 timecode data has been found or not.
13877 This filter accepts the following options:
13881 Set the maximum number of lines to scan for VITC data. If the value is set to
13882 @code{-1} the full video frame is scanned. Default is @code{45}.
13885 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
13886 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
13889 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
13890 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
13893 @subsection Examples
13897 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
13898 draw @code{--:--:--:--} as a placeholder:
13900 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
13906 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
13908 Destination pixel at position (X, Y) will be picked from source (x, y) position
13909 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
13910 value for pixel will be used for destination pixel.
13912 Xmap and Ymap input video streams must be of same dimensions. Output video stream
13913 will have Xmap/Ymap video stream dimensions.
13914 Xmap and Ymap input video streams are 16bit depth, single channel.
13916 @section removegrain
13918 The removegrain filter is a spatial denoiser for progressive video.
13922 Set mode for the first plane.
13925 Set mode for the second plane.
13928 Set mode for the third plane.
13931 Set mode for the fourth plane.
13934 Range of mode is from 0 to 24. Description of each mode follows:
13938 Leave input plane unchanged. Default.
13941 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
13944 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
13947 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
13950 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
13951 This is equivalent to a median filter.
13954 Line-sensitive clipping giving the minimal change.
13957 Line-sensitive clipping, intermediate.
13960 Line-sensitive clipping, intermediate.
13963 Line-sensitive clipping, intermediate.
13966 Line-sensitive clipping on a line where the neighbours pixels are the closest.
13969 Replaces the target pixel with the closest neighbour.
13972 [1 2 1] horizontal and vertical kernel blur.
13978 Bob mode, interpolates top field from the line where the neighbours
13979 pixels are the closest.
13982 Bob mode, interpolates bottom field from the line where the neighbours
13983 pixels are the closest.
13986 Bob mode, interpolates top field. Same as 13 but with a more complicated
13987 interpolation formula.
13990 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
13991 interpolation formula.
13994 Clips the pixel with the minimum and maximum of respectively the maximum and
13995 minimum of each pair of opposite neighbour pixels.
13998 Line-sensitive clipping using opposite neighbours whose greatest distance from
13999 the current pixel is minimal.
14002 Replaces the pixel with the average of its 8 neighbours.
14005 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
14008 Clips pixels using the averages of opposite neighbour.
14011 Same as mode 21 but simpler and faster.
14014 Small edge and halo removal, but reputed useless.
14020 @section removelogo
14022 Suppress a TV station logo, using an image file to determine which
14023 pixels comprise the logo. It works by filling in the pixels that
14024 comprise the logo with neighboring pixels.
14026 The filter accepts the following options:
14030 Set the filter bitmap file, which can be any image format supported by
14031 libavformat. The width and height of the image file must match those of the
14032 video stream being processed.
14035 Pixels in the provided bitmap image with a value of zero are not
14036 considered part of the logo, non-zero pixels are considered part of
14037 the logo. If you use white (255) for the logo and black (0) for the
14038 rest, you will be safe. For making the filter bitmap, it is
14039 recommended to take a screen capture of a black frame with the logo
14040 visible, and then using a threshold filter followed by the erode
14041 filter once or twice.
14043 If needed, little splotches can be fixed manually. Remember that if
14044 logo pixels are not covered, the filter quality will be much
14045 reduced. Marking too many pixels as part of the logo does not hurt as
14046 much, but it will increase the amount of blurring needed to cover over
14047 the image and will destroy more information than necessary, and extra
14048 pixels will slow things down on a large logo.
14050 @section repeatfields
14052 This filter uses the repeat_field flag from the Video ES headers and hard repeats
14053 fields based on its value.
14057 Reverse a video clip.
14059 Warning: This filter requires memory to buffer the entire clip, so trimming
14062 @subsection Examples
14066 Take the first 5 seconds of a clip, and reverse it.
14073 Apply roberts cross operator to input video stream.
14075 The filter accepts the following option:
14079 Set which planes will be processed, unprocessed planes will be copied.
14080 By default value 0xf, all planes will be processed.
14083 Set value which will be multiplied with filtered result.
14086 Set value which will be added to filtered result.
14091 Rotate video by an arbitrary angle expressed in radians.
14093 The filter accepts the following options:
14095 A description of the optional parameters follows.
14098 Set an expression for the angle by which to rotate the input video
14099 clockwise, expressed as a number of radians. A negative value will
14100 result in a counter-clockwise rotation. By default it is set to "0".
14102 This expression is evaluated for each frame.
14105 Set the output width expression, default value is "iw".
14106 This expression is evaluated just once during configuration.
14109 Set the output height expression, default value is "ih".
14110 This expression is evaluated just once during configuration.
14113 Enable bilinear interpolation if set to 1, a value of 0 disables
14114 it. Default value is 1.
14117 Set the color used to fill the output area not covered by the rotated
14118 image. For the general syntax of this option, check the
14119 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
14120 If the special value "none" is selected then no
14121 background is printed (useful for example if the background is never shown).
14123 Default value is "black".
14126 The expressions for the angle and the output size can contain the
14127 following constants and functions:
14131 sequential number of the input frame, starting from 0. It is always NAN
14132 before the first frame is filtered.
14135 time in seconds of the input frame, it is set to 0 when the filter is
14136 configured. It is always NAN before the first frame is filtered.
14140 horizontal and vertical chroma subsample values. For example for the
14141 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14145 the input video width and height
14149 the output width and height, that is the size of the padded area as
14150 specified by the @var{width} and @var{height} expressions
14154 the minimal width/height required for completely containing the input
14155 video rotated by @var{a} radians.
14157 These are only available when computing the @option{out_w} and
14158 @option{out_h} expressions.
14161 @subsection Examples
14165 Rotate the input by PI/6 radians clockwise:
14171 Rotate the input by PI/6 radians counter-clockwise:
14177 Rotate the input by 45 degrees clockwise:
14183 Apply a constant rotation with period T, starting from an angle of PI/3:
14185 rotate=PI/3+2*PI*t/T
14189 Make the input video rotation oscillating with a period of T
14190 seconds and an amplitude of A radians:
14192 rotate=A*sin(2*PI/T*t)
14196 Rotate the video, output size is chosen so that the whole rotating
14197 input video is always completely contained in the output:
14199 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
14203 Rotate the video, reduce the output size so that no background is ever
14206 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
14210 @subsection Commands
14212 The filter supports the following commands:
14216 Set the angle expression.
14217 The command accepts the same syntax of the corresponding option.
14219 If the specified expression is not valid, it is kept at its current
14225 Apply Shape Adaptive Blur.
14227 The filter accepts the following options:
14230 @item luma_radius, lr
14231 Set luma blur filter strength, must be a value in range 0.1-4.0, default
14232 value is 1.0. A greater value will result in a more blurred image, and
14233 in slower processing.
14235 @item luma_pre_filter_radius, lpfr
14236 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
14239 @item luma_strength, ls
14240 Set luma maximum difference between pixels to still be considered, must
14241 be a value in the 0.1-100.0 range, default value is 1.0.
14243 @item chroma_radius, cr
14244 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
14245 greater value will result in a more blurred image, and in slower
14248 @item chroma_pre_filter_radius, cpfr
14249 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
14251 @item chroma_strength, cs
14252 Set chroma maximum difference between pixels to still be considered,
14253 must be a value in the -0.9-100.0 range.
14256 Each chroma option value, if not explicitly specified, is set to the
14257 corresponding luma option value.
14262 Scale (resize) the input video, using the libswscale library.
14264 The scale filter forces the output display aspect ratio to be the same
14265 of the input, by changing the output sample aspect ratio.
14267 If the input image format is different from the format requested by
14268 the next filter, the scale filter will convert the input to the
14271 @subsection Options
14272 The filter accepts the following options, or any of the options
14273 supported by the libswscale scaler.
14275 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
14276 the complete list of scaler options.
14281 Set the output video dimension expression. Default value is the input
14284 If the @var{width} or @var{w} value is 0, the input width is used for
14285 the output. If the @var{height} or @var{h} value is 0, the input height
14286 is used for the output.
14288 If one and only one of the values is -n with n >= 1, the scale filter
14289 will use a value that maintains the aspect ratio of the input image,
14290 calculated from the other specified dimension. After that it will,
14291 however, make sure that the calculated dimension is divisible by n and
14292 adjust the value if necessary.
14294 If both values are -n with n >= 1, the behavior will be identical to
14295 both values being set to 0 as previously detailed.
14297 See below for the list of accepted constants for use in the dimension
14301 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
14305 Only evaluate expressions once during the filter initialization or when a command is processed.
14308 Evaluate expressions for each incoming frame.
14312 Default value is @samp{init}.
14316 Set the interlacing mode. It accepts the following values:
14320 Force interlaced aware scaling.
14323 Do not apply interlaced scaling.
14326 Select interlaced aware scaling depending on whether the source frames
14327 are flagged as interlaced or not.
14330 Default value is @samp{0}.
14333 Set libswscale scaling flags. See
14334 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
14335 complete list of values. If not explicitly specified the filter applies
14339 @item param0, param1
14340 Set libswscale input parameters for scaling algorithms that need them. See
14341 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
14342 complete documentation. If not explicitly specified the filter applies
14348 Set the video size. For the syntax of this option, check the
14349 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
14351 @item in_color_matrix
14352 @item out_color_matrix
14353 Set in/output YCbCr color space type.
14355 This allows the autodetected value to be overridden as well as allows forcing
14356 a specific value used for the output and encoder.
14358 If not specified, the color space type depends on the pixel format.
14364 Choose automatically.
14367 Format conforming to International Telecommunication Union (ITU)
14368 Recommendation BT.709.
14371 Set color space conforming to the United States Federal Communications
14372 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
14375 Set color space conforming to:
14379 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
14382 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
14385 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
14390 Set color space conforming to SMPTE ST 240:1999.
14395 Set in/output YCbCr sample range.
14397 This allows the autodetected value to be overridden as well as allows forcing
14398 a specific value used for the output and encoder. If not specified, the
14399 range depends on the pixel format. Possible values:
14403 Choose automatically.
14406 Set full range (0-255 in case of 8-bit luma).
14408 @item mpeg/limited/tv
14409 Set "MPEG" range (16-235 in case of 8-bit luma).
14412 @item force_original_aspect_ratio
14413 Enable decreasing or increasing output video width or height if necessary to
14414 keep the original aspect ratio. Possible values:
14418 Scale the video as specified and disable this feature.
14421 The output video dimensions will automatically be decreased if needed.
14424 The output video dimensions will automatically be increased if needed.
14428 One useful instance of this option is that when you know a specific device's
14429 maximum allowed resolution, you can use this to limit the output video to
14430 that, while retaining the aspect ratio. For example, device A allows
14431 1280x720 playback, and your video is 1920x800. Using this option (set it to
14432 decrease) and specifying 1280x720 to the command line makes the output
14435 Please note that this is a different thing than specifying -1 for @option{w}
14436 or @option{h}, you still need to specify the output resolution for this option
14441 The values of the @option{w} and @option{h} options are expressions
14442 containing the following constants:
14447 The input width and height
14451 These are the same as @var{in_w} and @var{in_h}.
14455 The output (scaled) width and height
14459 These are the same as @var{out_w} and @var{out_h}
14462 The same as @var{iw} / @var{ih}
14465 input sample aspect ratio
14468 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
14472 horizontal and vertical input chroma subsample values. For example for the
14473 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14477 horizontal and vertical output chroma subsample values. For example for the
14478 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14481 @subsection Examples
14485 Scale the input video to a size of 200x100
14490 This is equivalent to:
14501 Specify a size abbreviation for the output size:
14506 which can also be written as:
14512 Scale the input to 2x:
14514 scale=w=2*iw:h=2*ih
14518 The above is the same as:
14520 scale=2*in_w:2*in_h
14524 Scale the input to 2x with forced interlaced scaling:
14526 scale=2*iw:2*ih:interl=1
14530 Scale the input to half size:
14532 scale=w=iw/2:h=ih/2
14536 Increase the width, and set the height to the same size:
14542 Seek Greek harmony:
14549 Increase the height, and set the width to 3/2 of the height:
14551 scale=w=3/2*oh:h=3/5*ih
14555 Increase the size, making the size a multiple of the chroma
14558 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
14562 Increase the width to a maximum of 500 pixels,
14563 keeping the same aspect ratio as the input:
14565 scale=w='min(500\, iw*3/2):h=-1'
14569 Make pixels square by combining scale and setsar:
14571 scale='trunc(ih*dar):ih',setsar=1/1
14575 Make pixels square by combining scale and setsar,
14576 making sure the resulting resolution is even (required by some codecs):
14578 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
14582 @subsection Commands
14584 This filter supports the following commands:
14588 Set the output video dimension expression.
14589 The command accepts the same syntax of the corresponding option.
14591 If the specified expression is not valid, it is kept at its current
14597 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
14598 format conversion on CUDA video frames. Setting the output width and height
14599 works in the same way as for the @var{scale} filter.
14601 The following additional options are accepted:
14604 The pixel format of the output CUDA frames. If set to the string "same" (the
14605 default), the input format will be kept. Note that automatic format negotiation
14606 and conversion is not yet supported for hardware frames
14609 The interpolation algorithm used for resizing. One of the following:
14616 @item cubic2p_bspline
14617 2-parameter cubic (B=1, C=0)
14619 @item cubic2p_catmullrom
14620 2-parameter cubic (B=0, C=1/2)
14622 @item cubic2p_b05c03
14623 2-parameter cubic (B=1/2, C=3/10)
14635 Scale (resize) the input video, based on a reference video.
14637 See the scale filter for available options, scale2ref supports the same but
14638 uses the reference video instead of the main input as basis. scale2ref also
14639 supports the following additional constants for the @option{w} and
14640 @option{h} options:
14645 The main input video's width and height
14648 The same as @var{main_w} / @var{main_h}
14651 The main input video's sample aspect ratio
14653 @item main_dar, mdar
14654 The main input video's display aspect ratio. Calculated from
14655 @code{(main_w / main_h) * main_sar}.
14659 The main input video's horizontal and vertical chroma subsample values.
14660 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
14664 @subsection Examples
14668 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
14670 'scale2ref[b][a];[a][b]overlay'
14674 @anchor{selectivecolor}
14675 @section selectivecolor
14677 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
14678 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
14679 by the "purity" of the color (that is, how saturated it already is).
14681 This filter is similar to the Adobe Photoshop Selective Color tool.
14683 The filter accepts the following options:
14686 @item correction_method
14687 Select color correction method.
14689 Available values are:
14692 Specified adjustments are applied "as-is" (added/subtracted to original pixel
14695 Specified adjustments are relative to the original component value.
14697 Default is @code{absolute}.
14699 Adjustments for red pixels (pixels where the red component is the maximum)
14701 Adjustments for yellow pixels (pixels where the blue component is the minimum)
14703 Adjustments for green pixels (pixels where the green component is the maximum)
14705 Adjustments for cyan pixels (pixels where the red component is the minimum)
14707 Adjustments for blue pixels (pixels where the blue component is the maximum)
14709 Adjustments for magenta pixels (pixels where the green component is the minimum)
14711 Adjustments for white pixels (pixels where all components are greater than 128)
14713 Adjustments for all pixels except pure black and pure white
14715 Adjustments for black pixels (pixels where all components are lesser than 128)
14717 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
14720 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
14721 4 space separated floating point adjustment values in the [-1,1] range,
14722 respectively to adjust the amount of cyan, magenta, yellow and black for the
14723 pixels of its range.
14725 @subsection Examples
14729 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
14730 increase magenta by 27% in blue areas:
14732 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
14736 Use a Photoshop selective color preset:
14738 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
14742 @anchor{separatefields}
14743 @section separatefields
14745 The @code{separatefields} takes a frame-based video input and splits
14746 each frame into its components fields, producing a new half height clip
14747 with twice the frame rate and twice the frame count.
14749 This filter use field-dominance information in frame to decide which
14750 of each pair of fields to place first in the output.
14751 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
14753 @section setdar, setsar
14755 The @code{setdar} filter sets the Display Aspect Ratio for the filter
14758 This is done by changing the specified Sample (aka Pixel) Aspect
14759 Ratio, according to the following equation:
14761 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
14764 Keep in mind that the @code{setdar} filter does not modify the pixel
14765 dimensions of the video frame. Also, the display aspect ratio set by
14766 this filter may be changed by later filters in the filterchain,
14767 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
14770 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
14771 the filter output video.
14773 Note that as a consequence of the application of this filter, the
14774 output display aspect ratio will change according to the equation
14777 Keep in mind that the sample aspect ratio set by the @code{setsar}
14778 filter may be changed by later filters in the filterchain, e.g. if
14779 another "setsar" or a "setdar" filter is applied.
14781 It accepts the following parameters:
14784 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
14785 Set the aspect ratio used by the filter.
14787 The parameter can be a floating point number string, an expression, or
14788 a string of the form @var{num}:@var{den}, where @var{num} and
14789 @var{den} are the numerator and denominator of the aspect ratio. If
14790 the parameter is not specified, it is assumed the value "0".
14791 In case the form "@var{num}:@var{den}" is used, the @code{:} character
14795 Set the maximum integer value to use for expressing numerator and
14796 denominator when reducing the expressed aspect ratio to a rational.
14797 Default value is @code{100}.
14801 The parameter @var{sar} is an expression containing
14802 the following constants:
14806 These are approximated values for the mathematical constants e
14807 (Euler's number), pi (Greek pi), and phi (the golden ratio).
14810 The input width and height.
14813 These are the same as @var{w} / @var{h}.
14816 The input sample aspect ratio.
14819 The input display aspect ratio. It is the same as
14820 (@var{w} / @var{h}) * @var{sar}.
14823 Horizontal and vertical chroma subsample values. For example, for the
14824 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14827 @subsection Examples
14832 To change the display aspect ratio to 16:9, specify one of the following:
14839 To change the sample aspect ratio to 10:11, specify:
14845 To set a display aspect ratio of 16:9, and specify a maximum integer value of
14846 1000 in the aspect ratio reduction, use the command:
14848 setdar=ratio=16/9:max=1000
14856 Force field for the output video frame.
14858 The @code{setfield} filter marks the interlace type field for the
14859 output frames. It does not change the input frame, but only sets the
14860 corresponding property, which affects how the frame is treated by
14861 following filters (e.g. @code{fieldorder} or @code{yadif}).
14863 The filter accepts the following options:
14868 Available values are:
14872 Keep the same field property.
14875 Mark the frame as bottom-field-first.
14878 Mark the frame as top-field-first.
14881 Mark the frame as progressive.
14887 Show a line containing various information for each input video frame.
14888 The input video is not modified.
14890 The shown line contains a sequence of key/value pairs of the form
14891 @var{key}:@var{value}.
14893 The following values are shown in the output:
14897 The (sequential) number of the input frame, starting from 0.
14900 The Presentation TimeStamp of the input frame, expressed as a number of
14901 time base units. The time base unit depends on the filter input pad.
14904 The Presentation TimeStamp of the input frame, expressed as a number of
14908 The position of the frame in the input stream, or -1 if this information is
14909 unavailable and/or meaningless (for example in case of synthetic video).
14912 The pixel format name.
14915 The sample aspect ratio of the input frame, expressed in the form
14916 @var{num}/@var{den}.
14919 The size of the input frame. For the syntax of this option, check the
14920 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
14923 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
14924 for bottom field first).
14927 This is 1 if the frame is a key frame, 0 otherwise.
14930 The picture type of the input frame ("I" for an I-frame, "P" for a
14931 P-frame, "B" for a B-frame, or "?" for an unknown type).
14932 Also refer to the documentation of the @code{AVPictureType} enum and of
14933 the @code{av_get_picture_type_char} function defined in
14934 @file{libavutil/avutil.h}.
14937 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
14939 @item plane_checksum
14940 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
14941 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
14944 @section showpalette
14946 Displays the 256 colors palette of each frame. This filter is only relevant for
14947 @var{pal8} pixel format frames.
14949 It accepts the following option:
14953 Set the size of the box used to represent one palette color entry. Default is
14954 @code{30} (for a @code{30x30} pixel box).
14957 @section shuffleframes
14959 Reorder and/or duplicate and/or drop video frames.
14961 It accepts the following parameters:
14965 Set the destination indexes of input frames.
14966 This is space or '|' separated list of indexes that maps input frames to output
14967 frames. Number of indexes also sets maximal value that each index may have.
14968 '-1' index have special meaning and that is to drop frame.
14971 The first frame has the index 0. The default is to keep the input unchanged.
14973 @subsection Examples
14977 Swap second and third frame of every three frames of the input:
14979 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
14983 Swap 10th and 1st frame of every ten frames of the input:
14985 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
14989 @section shuffleplanes
14991 Reorder and/or duplicate video planes.
14993 It accepts the following parameters:
14998 The index of the input plane to be used as the first output plane.
15001 The index of the input plane to be used as the second output plane.
15004 The index of the input plane to be used as the third output plane.
15007 The index of the input plane to be used as the fourth output plane.
15011 The first plane has the index 0. The default is to keep the input unchanged.
15013 @subsection Examples
15017 Swap the second and third planes of the input:
15019 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
15023 @anchor{signalstats}
15024 @section signalstats
15025 Evaluate various visual metrics that assist in determining issues associated
15026 with the digitization of analog video media.
15028 By default the filter will log these metadata values:
15032 Display the minimal Y value contained within the input frame. Expressed in
15036 Display the Y value at the 10% percentile within the input frame. Expressed in
15040 Display the average Y value within the input frame. Expressed in range of
15044 Display the Y value at the 90% percentile within the input frame. Expressed in
15048 Display the maximum Y value contained within the input frame. Expressed in
15052 Display the minimal U value contained within the input frame. Expressed in
15056 Display the U value at the 10% percentile within the input frame. Expressed in
15060 Display the average U value within the input frame. Expressed in range of
15064 Display the U value at the 90% percentile within the input frame. Expressed in
15068 Display the maximum U value contained within the input frame. Expressed in
15072 Display the minimal V value contained within the input frame. Expressed in
15076 Display the V value at the 10% percentile within the input frame. Expressed in
15080 Display the average V value within the input frame. Expressed in range of
15084 Display the V value at the 90% percentile within the input frame. Expressed in
15088 Display the maximum V value contained within the input frame. Expressed in
15092 Display the minimal saturation value contained within the input frame.
15093 Expressed in range of [0-~181.02].
15096 Display the saturation value at the 10% percentile within the input frame.
15097 Expressed in range of [0-~181.02].
15100 Display the average saturation value within the input frame. Expressed in range
15104 Display the saturation value at the 90% percentile within the input frame.
15105 Expressed in range of [0-~181.02].
15108 Display the maximum saturation value contained within the input frame.
15109 Expressed in range of [0-~181.02].
15112 Display the median value for hue within the input frame. Expressed in range of
15116 Display the average value for hue within the input frame. Expressed in range of
15120 Display the average of sample value difference between all values of the Y
15121 plane in the current frame and corresponding values of the previous input frame.
15122 Expressed in range of [0-255].
15125 Display the average of sample value difference between all values of the U
15126 plane in the current frame and corresponding values of the previous input frame.
15127 Expressed in range of [0-255].
15130 Display the average of sample value difference between all values of the V
15131 plane in the current frame and corresponding values of the previous input frame.
15132 Expressed in range of [0-255].
15135 Display bit depth of Y plane in current frame.
15136 Expressed in range of [0-16].
15139 Display bit depth of U plane in current frame.
15140 Expressed in range of [0-16].
15143 Display bit depth of V plane in current frame.
15144 Expressed in range of [0-16].
15147 The filter accepts the following options:
15153 @option{stat} specify an additional form of image analysis.
15154 @option{out} output video with the specified type of pixel highlighted.
15156 Both options accept the following values:
15160 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
15161 unlike the neighboring pixels of the same field. Examples of temporal outliers
15162 include the results of video dropouts, head clogs, or tape tracking issues.
15165 Identify @var{vertical line repetition}. Vertical line repetition includes
15166 similar rows of pixels within a frame. In born-digital video vertical line
15167 repetition is common, but this pattern is uncommon in video digitized from an
15168 analog source. When it occurs in video that results from the digitization of an
15169 analog source it can indicate concealment from a dropout compensator.
15172 Identify pixels that fall outside of legal broadcast range.
15176 Set the highlight color for the @option{out} option. The default color is
15180 @subsection Examples
15184 Output data of various video metrics:
15186 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
15190 Output specific data about the minimum and maximum values of the Y plane per frame:
15192 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
15196 Playback video while highlighting pixels that are outside of broadcast range in red.
15198 ffplay example.mov -vf signalstats="out=brng:color=red"
15202 Playback video with signalstats metadata drawn over the frame.
15204 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
15207 The contents of signalstat_drawtext.txt used in the command are:
15210 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
15211 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
15212 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
15213 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
15221 Calculates the MPEG-7 Video Signature. The filter can handle more than one
15222 input. In this case the matching between the inputs can be calculated additionally.
15223 The filter always passes through the first input. The signature of each stream can
15224 be written into a file.
15226 It accepts the following options:
15230 Enable or disable the matching process.
15232 Available values are:
15236 Disable the calculation of a matching (default).
15238 Calculate the matching for the whole video and output whether the whole video
15239 matches or only parts.
15241 Calculate only until a matching is found or the video ends. Should be faster in
15246 Set the number of inputs. The option value must be a non negative integer.
15247 Default value is 1.
15250 Set the path to which the output is written. If there is more than one input,
15251 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
15252 integer), that will be replaced with the input number. If no filename is
15253 specified, no output will be written. This is the default.
15256 Choose the output format.
15258 Available values are:
15262 Use the specified binary representation (default).
15264 Use the specified xml representation.
15268 Set threshold to detect one word as similar. The option value must be an integer
15269 greater than zero. The default value is 9000.
15272 Set threshold to detect all words as similar. The option value must be an integer
15273 greater than zero. The default value is 60000.
15276 Set threshold to detect frames as similar. The option value must be an integer
15277 greater than zero. The default value is 116.
15280 Set the minimum length of a sequence in frames to recognize it as matching
15281 sequence. The option value must be a non negative integer value.
15282 The default value is 0.
15285 Set the minimum relation, that matching frames to all frames must have.
15286 The option value must be a double value between 0 and 1. The default value is 0.5.
15289 @subsection Examples
15293 To calculate the signature of an input video and store it in signature.bin:
15295 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
15299 To detect whether two videos match and store the signatures in XML format in
15300 signature0.xml and signature1.xml:
15302 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 -
15310 Blur the input video without impacting the outlines.
15312 It accepts the following options:
15315 @item luma_radius, lr
15316 Set the luma radius. The option value must be a float number in
15317 the range [0.1,5.0] that specifies the variance of the gaussian filter
15318 used to blur the image (slower if larger). Default value is 1.0.
15320 @item luma_strength, ls
15321 Set the luma strength. The option value must be a float number
15322 in the range [-1.0,1.0] that configures the blurring. A value included
15323 in [0.0,1.0] will blur the image whereas a value included in
15324 [-1.0,0.0] will sharpen the image. Default value is 1.0.
15326 @item luma_threshold, lt
15327 Set the luma threshold used as a coefficient to determine
15328 whether a pixel should be blurred or not. The option value must be an
15329 integer in the range [-30,30]. A value of 0 will filter all the image,
15330 a value included in [0,30] will filter flat areas and a value included
15331 in [-30,0] will filter edges. Default value is 0.
15333 @item chroma_radius, cr
15334 Set the chroma radius. The option value must be a float number in
15335 the range [0.1,5.0] that specifies the variance of the gaussian filter
15336 used to blur the image (slower if larger). Default value is @option{luma_radius}.
15338 @item chroma_strength, cs
15339 Set the chroma strength. The option value must be a float number
15340 in the range [-1.0,1.0] that configures the blurring. A value included
15341 in [0.0,1.0] will blur the image whereas a value included in
15342 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
15344 @item chroma_threshold, ct
15345 Set the chroma threshold used as a coefficient to determine
15346 whether a pixel should be blurred or not. The option value must be an
15347 integer in the range [-30,30]. A value of 0 will filter all the image,
15348 a value included in [0,30] will filter flat areas and a value included
15349 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
15352 If a chroma option is not explicitly set, the corresponding luma value
15357 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
15359 This filter takes in input two input videos, the first input is
15360 considered the "main" source and is passed unchanged to the
15361 output. The second input is used as a "reference" video for computing
15364 Both video inputs must have the same resolution and pixel format for
15365 this filter to work correctly. Also it assumes that both inputs
15366 have the same number of frames, which are compared one by one.
15368 The filter stores the calculated SSIM of each frame.
15370 The description of the accepted parameters follows.
15373 @item stats_file, f
15374 If specified the filter will use the named file to save the SSIM of
15375 each individual frame. When filename equals "-" the data is sent to
15379 The file printed if @var{stats_file} is selected, contains a sequence of
15380 key/value pairs of the form @var{key}:@var{value} for each compared
15383 A description of each shown parameter follows:
15387 sequential number of the input frame, starting from 1
15389 @item Y, U, V, R, G, B
15390 SSIM of the compared frames for the component specified by the suffix.
15393 SSIM of the compared frames for the whole frame.
15396 Same as above but in dB representation.
15399 This filter also supports the @ref{framesync} options.
15403 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
15404 [main][ref] ssim="stats_file=stats.log" [out]
15407 On this example the input file being processed is compared with the
15408 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
15409 is stored in @file{stats.log}.
15411 Another example with both psnr and ssim at same time:
15413 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
15418 Convert between different stereoscopic image formats.
15420 The filters accept the following options:
15424 Set stereoscopic image format of input.
15426 Available values for input image formats are:
15429 side by side parallel (left eye left, right eye right)
15432 side by side crosseye (right eye left, left eye right)
15435 side by side parallel with half width resolution
15436 (left eye left, right eye right)
15439 side by side crosseye with half width resolution
15440 (right eye left, left eye right)
15443 above-below (left eye above, right eye below)
15446 above-below (right eye above, left eye below)
15449 above-below with half height resolution
15450 (left eye above, right eye below)
15453 above-below with half height resolution
15454 (right eye above, left eye below)
15457 alternating frames (left eye first, right eye second)
15460 alternating frames (right eye first, left eye second)
15463 interleaved rows (left eye has top row, right eye starts on next row)
15466 interleaved rows (right eye has top row, left eye starts on next row)
15469 interleaved columns, left eye first
15472 interleaved columns, right eye first
15474 Default value is @samp{sbsl}.
15478 Set stereoscopic image format of output.
15482 side by side parallel (left eye left, right eye right)
15485 side by side crosseye (right eye left, left eye right)
15488 side by side parallel with half width resolution
15489 (left eye left, right eye right)
15492 side by side crosseye with half width resolution
15493 (right eye left, left eye right)
15496 above-below (left eye above, right eye below)
15499 above-below (right eye above, left eye below)
15502 above-below with half height resolution
15503 (left eye above, right eye below)
15506 above-below with half height resolution
15507 (right eye above, left eye below)
15510 alternating frames (left eye first, right eye second)
15513 alternating frames (right eye first, left eye second)
15516 interleaved rows (left eye has top row, right eye starts on next row)
15519 interleaved rows (right eye has top row, left eye starts on next row)
15522 anaglyph red/blue gray
15523 (red filter on left eye, blue filter on right eye)
15526 anaglyph red/green gray
15527 (red filter on left eye, green filter on right eye)
15530 anaglyph red/cyan gray
15531 (red filter on left eye, cyan filter on right eye)
15534 anaglyph red/cyan half colored
15535 (red filter on left eye, cyan filter on right eye)
15538 anaglyph red/cyan color
15539 (red filter on left eye, cyan filter on right eye)
15542 anaglyph red/cyan color optimized with the least squares projection of dubois
15543 (red filter on left eye, cyan filter on right eye)
15546 anaglyph green/magenta gray
15547 (green filter on left eye, magenta filter on right eye)
15550 anaglyph green/magenta half colored
15551 (green filter on left eye, magenta filter on right eye)
15554 anaglyph green/magenta colored
15555 (green filter on left eye, magenta filter on right eye)
15558 anaglyph green/magenta color optimized with the least squares projection of dubois
15559 (green filter on left eye, magenta filter on right eye)
15562 anaglyph yellow/blue gray
15563 (yellow filter on left eye, blue filter on right eye)
15566 anaglyph yellow/blue half colored
15567 (yellow filter on left eye, blue filter on right eye)
15570 anaglyph yellow/blue colored
15571 (yellow filter on left eye, blue filter on right eye)
15574 anaglyph yellow/blue color optimized with the least squares projection of dubois
15575 (yellow filter on left eye, blue filter on right eye)
15578 mono output (left eye only)
15581 mono output (right eye only)
15584 checkerboard, left eye first
15587 checkerboard, right eye first
15590 interleaved columns, left eye first
15593 interleaved columns, right eye first
15599 Default value is @samp{arcd}.
15602 @subsection Examples
15606 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
15612 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
15618 @section streamselect, astreamselect
15619 Select video or audio streams.
15621 The filter accepts the following options:
15625 Set number of inputs. Default is 2.
15628 Set input indexes to remap to outputs.
15631 @subsection Commands
15633 The @code{streamselect} and @code{astreamselect} filter supports the following
15638 Set input indexes to remap to outputs.
15641 @subsection Examples
15645 Select first 5 seconds 1st stream and rest of time 2nd stream:
15647 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
15651 Same as above, but for audio:
15653 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
15658 Apply sobel operator to input video stream.
15660 The filter accepts the following option:
15664 Set which planes will be processed, unprocessed planes will be copied.
15665 By default value 0xf, all planes will be processed.
15668 Set value which will be multiplied with filtered result.
15671 Set value which will be added to filtered result.
15677 Apply a simple postprocessing filter that compresses and decompresses the image
15678 at several (or - in the case of @option{quality} level @code{6} - all) shifts
15679 and average the results.
15681 The filter accepts the following options:
15685 Set quality. This option defines the number of levels for averaging. It accepts
15686 an integer in the range 0-6. If set to @code{0}, the filter will have no
15687 effect. A value of @code{6} means the higher quality. For each increment of
15688 that value the speed drops by a factor of approximately 2. Default value is
15692 Force a constant quantization parameter. If not set, the filter will use the QP
15693 from the video stream (if available).
15696 Set thresholding mode. Available modes are:
15700 Set hard thresholding (default).
15702 Set soft thresholding (better de-ringing effect, but likely blurrier).
15705 @item use_bframe_qp
15706 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
15707 option may cause flicker since the B-Frames have often larger QP. Default is
15708 @code{0} (not enabled).
15713 Scale the input by applying one of the super-resolution methods based on
15714 convolutional neural networks. Supported models:
15718 Super-Resolution Convolutional Neural Network model (SRCNN).
15719 See @url{https://arxiv.org/abs/1501.00092}.
15722 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
15723 See @url{https://arxiv.org/abs/1609.05158}.
15726 Training scripts as well as scripts for model generation are provided in
15727 the repository at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
15729 The filter accepts the following options:
15733 Specify which DNN backend to use for model loading and execution. This option accepts
15734 the following values:
15738 Native implementation of DNN loading and execution.
15741 TensorFlow backend. To enable this backend you
15742 need to install the TensorFlow for C library (see
15743 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
15744 @code{--enable-libtensorflow}
15747 Default value is @samp{native}.
15750 Set path to model file specifying network architecture and its parameters.
15751 Note that different backends use different file formats. TensorFlow backend
15752 can load files for both formats, while native backend can load files for only
15756 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
15757 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
15758 input upscaled using bicubic upscaling with proper scale factor.
15764 Draw subtitles on top of input video using the libass library.
15766 To enable compilation of this filter you need to configure FFmpeg with
15767 @code{--enable-libass}. This filter also requires a build with libavcodec and
15768 libavformat to convert the passed subtitles file to ASS (Advanced Substation
15769 Alpha) subtitles format.
15771 The filter accepts the following options:
15775 Set the filename of the subtitle file to read. It must be specified.
15777 @item original_size
15778 Specify the size of the original video, the video for which the ASS file
15779 was composed. For the syntax of this option, check the
15780 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
15781 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
15782 correctly scale the fonts if the aspect ratio has been changed.
15785 Set a directory path containing fonts that can be used by the filter.
15786 These fonts will be used in addition to whatever the font provider uses.
15789 Process alpha channel, by default alpha channel is untouched.
15792 Set subtitles input character encoding. @code{subtitles} filter only. Only
15793 useful if not UTF-8.
15795 @item stream_index, si
15796 Set subtitles stream index. @code{subtitles} filter only.
15799 Override default style or script info parameters of the subtitles. It accepts a
15800 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
15803 If the first key is not specified, it is assumed that the first value
15804 specifies the @option{filename}.
15806 For example, to render the file @file{sub.srt} on top of the input
15807 video, use the command:
15812 which is equivalent to:
15814 subtitles=filename=sub.srt
15817 To render the default subtitles stream from file @file{video.mkv}, use:
15819 subtitles=video.mkv
15822 To render the second subtitles stream from that file, use:
15824 subtitles=video.mkv:si=1
15827 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
15828 @code{DejaVu Serif}, use:
15830 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HCCFF0000'
15833 @section super2xsai
15835 Scale the input by 2x and smooth using the Super2xSaI (Scale and
15836 Interpolate) pixel art scaling algorithm.
15838 Useful for enlarging pixel art images without reducing sharpness.
15842 Swap two rectangular objects in video.
15844 This filter accepts the following options:
15854 Set 1st rect x coordinate.
15857 Set 1st rect y coordinate.
15860 Set 2nd rect x coordinate.
15863 Set 2nd rect y coordinate.
15865 All expressions are evaluated once for each frame.
15868 The all options are expressions containing the following constants:
15873 The input width and height.
15876 same as @var{w} / @var{h}
15879 input sample aspect ratio
15882 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
15885 The number of the input frame, starting from 0.
15888 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
15891 the position in the file of the input frame, NAN if unknown
15899 Apply telecine process to the video.
15901 This filter accepts the following options:
15910 The default value is @code{top}.
15914 A string of numbers representing the pulldown pattern you wish to apply.
15915 The default value is @code{23}.
15919 Some typical patterns:
15924 24p: 2332 (preferred)
15931 24p: 222222222223 ("Euro pulldown")
15938 Apply threshold effect to video stream.
15940 This filter needs four video streams to perform thresholding.
15941 First stream is stream we are filtering.
15942 Second stream is holding threshold values, third stream is holding min values,
15943 and last, fourth stream is holding max values.
15945 The filter accepts the following option:
15949 Set which planes will be processed, unprocessed planes will be copied.
15950 By default value 0xf, all planes will be processed.
15953 For example if first stream pixel's component value is less then threshold value
15954 of pixel component from 2nd threshold stream, third stream value will picked,
15955 otherwise fourth stream pixel component value will be picked.
15957 Using color source filter one can perform various types of thresholding:
15959 @subsection Examples
15963 Binary threshold, using gray color as threshold:
15965 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
15969 Inverted binary threshold, using gray color as threshold:
15971 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
15975 Truncate binary threshold, using gray color as threshold:
15977 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
15981 Threshold to zero, using gray color as threshold:
15983 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
15987 Inverted threshold to zero, using gray color as threshold:
15989 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
15994 Select the most representative frame in a given sequence of consecutive frames.
15996 The filter accepts the following options:
16000 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
16001 will pick one of them, and then handle the next batch of @var{n} frames until
16002 the end. Default is @code{100}.
16005 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
16006 value will result in a higher memory usage, so a high value is not recommended.
16008 @subsection Examples
16012 Extract one picture each 50 frames:
16018 Complete example of a thumbnail creation with @command{ffmpeg}:
16020 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
16026 Tile several successive frames together.
16028 The filter accepts the following options:
16033 Set the grid size (i.e. the number of lines and columns). For the syntax of
16034 this option, check the
16035 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16038 Set the maximum number of frames to render in the given area. It must be less
16039 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
16040 the area will be used.
16043 Set the outer border margin in pixels.
16046 Set the inner border thickness (i.e. the number of pixels between frames). For
16047 more advanced padding options (such as having different values for the edges),
16048 refer to the pad video filter.
16051 Specify the color of the unused area. For the syntax of this option, check the
16052 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16053 The default value of @var{color} is "black".
16056 Set the number of frames to overlap when tiling several successive frames together.
16057 The value must be between @code{0} and @var{nb_frames - 1}.
16060 Set the number of frames to initially be empty before displaying first output frame.
16061 This controls how soon will one get first output frame.
16062 The value must be between @code{0} and @var{nb_frames - 1}.
16065 @subsection Examples
16069 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
16071 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
16073 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
16074 duplicating each output frame to accommodate the originally detected frame
16078 Display @code{5} pictures in an area of @code{3x2} frames,
16079 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
16080 mixed flat and named options:
16082 tile=3x2:nb_frames=5:padding=7:margin=2
16086 @section tinterlace
16088 Perform various types of temporal field interlacing.
16090 Frames are counted starting from 1, so the first input frame is
16093 The filter accepts the following options:
16098 Specify the mode of the interlacing. This option can also be specified
16099 as a value alone. See below for a list of values for this option.
16101 Available values are:
16105 Move odd frames into the upper field, even into the lower field,
16106 generating a double height frame at half frame rate.
16110 Frame 1 Frame 2 Frame 3 Frame 4
16112 11111 22222 33333 44444
16113 11111 22222 33333 44444
16114 11111 22222 33333 44444
16115 11111 22222 33333 44444
16129 Only output odd frames, even frames are dropped, generating a frame with
16130 unchanged height at half frame rate.
16135 Frame 1 Frame 2 Frame 3 Frame 4
16137 11111 22222 33333 44444
16138 11111 22222 33333 44444
16139 11111 22222 33333 44444
16140 11111 22222 33333 44444
16150 Only output even frames, odd frames are dropped, generating a frame with
16151 unchanged height at half frame rate.
16156 Frame 1 Frame 2 Frame 3 Frame 4
16158 11111 22222 33333 44444
16159 11111 22222 33333 44444
16160 11111 22222 33333 44444
16161 11111 22222 33333 44444
16171 Expand each frame to full height, but pad alternate lines with black,
16172 generating a frame with double height at the same input frame rate.
16177 Frame 1 Frame 2 Frame 3 Frame 4
16179 11111 22222 33333 44444
16180 11111 22222 33333 44444
16181 11111 22222 33333 44444
16182 11111 22222 33333 44444
16185 11111 ..... 33333 .....
16186 ..... 22222 ..... 44444
16187 11111 ..... 33333 .....
16188 ..... 22222 ..... 44444
16189 11111 ..... 33333 .....
16190 ..... 22222 ..... 44444
16191 11111 ..... 33333 .....
16192 ..... 22222 ..... 44444
16196 @item interleave_top, 4
16197 Interleave the upper field from odd frames with the lower field from
16198 even frames, generating a frame with unchanged height at half frame rate.
16203 Frame 1 Frame 2 Frame 3 Frame 4
16205 11111<- 22222 33333<- 44444
16206 11111 22222<- 33333 44444<-
16207 11111<- 22222 33333<- 44444
16208 11111 22222<- 33333 44444<-
16218 @item interleave_bottom, 5
16219 Interleave the lower field from odd frames with the upper field from
16220 even frames, generating a frame with unchanged height at half frame rate.
16225 Frame 1 Frame 2 Frame 3 Frame 4
16227 11111 22222<- 33333 44444<-
16228 11111<- 22222 33333<- 44444
16229 11111 22222<- 33333 44444<-
16230 11111<- 22222 33333<- 44444
16240 @item interlacex2, 6
16241 Double frame rate with unchanged height. Frames are inserted each
16242 containing the second temporal field from the previous input frame and
16243 the first temporal field from the next input frame. This mode relies on
16244 the top_field_first flag. Useful for interlaced video displays with no
16245 field synchronisation.
16250 Frame 1 Frame 2 Frame 3 Frame 4
16252 11111 22222 33333 44444
16253 11111 22222 33333 44444
16254 11111 22222 33333 44444
16255 11111 22222 33333 44444
16258 11111 22222 22222 33333 33333 44444 44444
16259 11111 11111 22222 22222 33333 33333 44444
16260 11111 22222 22222 33333 33333 44444 44444
16261 11111 11111 22222 22222 33333 33333 44444
16266 Move odd frames into the upper field, even into the lower field,
16267 generating a double height frame at same frame rate.
16272 Frame 1 Frame 2 Frame 3 Frame 4
16274 11111 22222 33333 44444
16275 11111 22222 33333 44444
16276 11111 22222 33333 44444
16277 11111 22222 33333 44444
16280 11111 33333 33333 55555
16281 22222 22222 44444 44444
16282 11111 33333 33333 55555
16283 22222 22222 44444 44444
16284 11111 33333 33333 55555
16285 22222 22222 44444 44444
16286 11111 33333 33333 55555
16287 22222 22222 44444 44444
16292 Numeric values are deprecated but are accepted for backward
16293 compatibility reasons.
16295 Default mode is @code{merge}.
16298 Specify flags influencing the filter process.
16300 Available value for @var{flags} is:
16303 @item low_pass_filter, vlfp
16304 Enable linear vertical low-pass filtering in the filter.
16305 Vertical low-pass filtering is required when creating an interlaced
16306 destination from a progressive source which contains high-frequency
16307 vertical detail. Filtering will reduce interlace 'twitter' and Moire
16310 @item complex_filter, cvlfp
16311 Enable complex vertical low-pass filtering.
16312 This will slightly less reduce interlace 'twitter' and Moire
16313 patterning but better retain detail and subjective sharpness impression.
16317 Vertical low-pass filtering can only be enabled for @option{mode}
16318 @var{interleave_top} and @var{interleave_bottom}.
16324 Mix successive video frames.
16326 A description of the accepted options follows.
16330 The number of successive frames to mix. If unspecified, it defaults to 3.
16333 Specify weight of each input video frame.
16334 Each weight is separated by space. If number of weights is smaller than
16335 number of @var{frames} last specified weight will be used for all remaining
16339 Specify scale, if it is set it will be multiplied with sum
16340 of each weight multiplied with pixel values to give final destination
16341 pixel value. By default @var{scale} is auto scaled to sum of weights.
16344 @subsection Examples
16348 Average 7 successive frames:
16350 tmix=frames=7:weights="1 1 1 1 1 1 1"
16354 Apply simple temporal convolution:
16356 tmix=frames=3:weights="-1 3 -1"
16360 Similar as above but only showing temporal differences:
16362 tmix=frames=3:weights="-1 2 -1":scale=1
16367 Tone map colors from different dynamic ranges.
16369 This filter expects data in single precision floating point, as it needs to
16370 operate on (and can output) out-of-range values. Another filter, such as
16371 @ref{zscale}, is needed to convert the resulting frame to a usable format.
16373 The tonemapping algorithms implemented only work on linear light, so input
16374 data should be linearized beforehand (and possibly correctly tagged).
16377 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
16380 @subsection Options
16381 The filter accepts the following options.
16385 Set the tone map algorithm to use.
16387 Possible values are:
16390 Do not apply any tone map, only desaturate overbright pixels.
16393 Hard-clip any out-of-range values. Use it for perfect color accuracy for
16394 in-range values, while distorting out-of-range values.
16397 Stretch the entire reference gamut to a linear multiple of the display.
16400 Fit a logarithmic transfer between the tone curves.
16403 Preserve overall image brightness with a simple curve, using nonlinear
16404 contrast, which results in flattening details and degrading color accuracy.
16407 Preserve both dark and bright details better than @var{reinhard}, at the cost
16408 of slightly darkening everything. Use it when detail preservation is more
16409 important than color and brightness accuracy.
16412 Smoothly map out-of-range values, while retaining contrast and colors for
16413 in-range material as much as possible. Use it when color accuracy is more
16414 important than detail preservation.
16420 Tune the tone mapping algorithm.
16422 This affects the following algorithms:
16428 Specifies the scale factor to use while stretching.
16432 Specifies the exponent of the function.
16436 Specify an extra linear coefficient to multiply into the signal before clipping.
16440 Specify the local contrast coefficient at the display peak.
16441 Default to 0.5, which means that in-gamut values will be about half as bright
16448 Specify the transition point from linear to mobius transform. Every value
16449 below this point is guaranteed to be mapped 1:1. The higher the value, the
16450 more accurate the result will be, at the cost of losing bright details.
16451 Default to 0.3, which due to the steep initial slope still preserves in-range
16452 colors fairly accurately.
16456 Apply desaturation for highlights that exceed this level of brightness. The
16457 higher the parameter, the more color information will be preserved. This
16458 setting helps prevent unnaturally blown-out colors for super-highlights, by
16459 (smoothly) turning into white instead. This makes images feel more natural,
16460 at the cost of reducing information about out-of-range colors.
16462 The default of 2.0 is somewhat conservative and will mostly just apply to
16463 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
16465 This option works only if the input frame has a supported color tag.
16468 Override signal/nominal/reference peak with this value. Useful when the
16469 embedded peak information in display metadata is not reliable or when tone
16470 mapping from a lower range to a higher range.
16476 Transpose rows with columns in the input video and optionally flip it.
16478 It accepts the following parameters:
16483 Specify the transposition direction.
16485 Can assume the following values:
16487 @item 0, 4, cclock_flip
16488 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
16496 Rotate by 90 degrees clockwise, that is:
16504 Rotate by 90 degrees counterclockwise, that is:
16511 @item 3, 7, clock_flip
16512 Rotate by 90 degrees clockwise and vertically flip, that is:
16520 For values between 4-7, the transposition is only done if the input
16521 video geometry is portrait and not landscape. These values are
16522 deprecated, the @code{passthrough} option should be used instead.
16524 Numerical values are deprecated, and should be dropped in favor of
16525 symbolic constants.
16528 Do not apply the transposition if the input geometry matches the one
16529 specified by the specified value. It accepts the following values:
16532 Always apply transposition.
16534 Preserve portrait geometry (when @var{height} >= @var{width}).
16536 Preserve landscape geometry (when @var{width} >= @var{height}).
16539 Default value is @code{none}.
16542 For example to rotate by 90 degrees clockwise and preserve portrait
16545 transpose=dir=1:passthrough=portrait
16548 The command above can also be specified as:
16550 transpose=1:portrait
16553 @section transpose_npp
16555 Transpose rows with columns in the input video and optionally flip it.
16556 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
16558 It accepts the following parameters:
16563 Specify the transposition direction.
16565 Can assume the following values:
16568 Rotate by 90 degrees counterclockwise and vertically flip. (default)
16571 Rotate by 90 degrees clockwise.
16574 Rotate by 90 degrees counterclockwise.
16577 Rotate by 90 degrees clockwise and vertically flip.
16581 Do not apply the transposition if the input geometry matches the one
16582 specified by the specified value. It accepts the following values:
16585 Always apply transposition. (default)
16587 Preserve portrait geometry (when @var{height} >= @var{width}).
16589 Preserve landscape geometry (when @var{width} >= @var{height}).
16595 Trim the input so that the output contains one continuous subpart of the input.
16597 It accepts the following parameters:
16600 Specify the time of the start of the kept section, i.e. the frame with the
16601 timestamp @var{start} will be the first frame in the output.
16604 Specify the time of the first frame that will be dropped, i.e. the frame
16605 immediately preceding the one with the timestamp @var{end} will be the last
16606 frame in the output.
16609 This is the same as @var{start}, except this option sets the start timestamp
16610 in timebase units instead of seconds.
16613 This is the same as @var{end}, except this option sets the end timestamp
16614 in timebase units instead of seconds.
16617 The maximum duration of the output in seconds.
16620 The number of the first frame that should be passed to the output.
16623 The number of the first frame that should be dropped.
16626 @option{start}, @option{end}, and @option{duration} are expressed as time
16627 duration specifications; see
16628 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
16629 for the accepted syntax.
16631 Note that the first two sets of the start/end options and the @option{duration}
16632 option look at the frame timestamp, while the _frame variants simply count the
16633 frames that pass through the filter. Also note that this filter does not modify
16634 the timestamps. If you wish for the output timestamps to start at zero, insert a
16635 setpts filter after the trim filter.
16637 If multiple start or end options are set, this filter tries to be greedy and
16638 keep all the frames that match at least one of the specified constraints. To keep
16639 only the part that matches all the constraints at once, chain multiple trim
16642 The defaults are such that all the input is kept. So it is possible to set e.g.
16643 just the end values to keep everything before the specified time.
16648 Drop everything except the second minute of input:
16650 ffmpeg -i INPUT -vf trim=60:120
16654 Keep only the first second:
16656 ffmpeg -i INPUT -vf trim=duration=1
16661 @section unpremultiply
16662 Apply alpha unpremultiply effect to input video stream using first plane
16663 of second stream as alpha.
16665 Both streams must have same dimensions and same pixel format.
16667 The filter accepts the following option:
16671 Set which planes will be processed, unprocessed planes will be copied.
16672 By default value 0xf, all planes will be processed.
16674 If the format has 1 or 2 components, then luma is bit 0.
16675 If the format has 3 or 4 components:
16676 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
16677 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
16678 If present, the alpha channel is always the last bit.
16681 Do not require 2nd input for processing, instead use alpha plane from input stream.
16687 Sharpen or blur the input video.
16689 It accepts the following parameters:
16692 @item luma_msize_x, lx
16693 Set the luma matrix horizontal size. It must be an odd integer between
16694 3 and 23. The default value is 5.
16696 @item luma_msize_y, ly
16697 Set the luma matrix vertical size. It must be an odd integer between 3
16698 and 23. The default value is 5.
16700 @item luma_amount, la
16701 Set the luma effect strength. It must be a floating point number, reasonable
16702 values lay between -1.5 and 1.5.
16704 Negative values will blur the input video, while positive values will
16705 sharpen it, a value of zero will disable the effect.
16707 Default value is 1.0.
16709 @item chroma_msize_x, cx
16710 Set the chroma matrix horizontal size. It must be an odd integer
16711 between 3 and 23. The default value is 5.
16713 @item chroma_msize_y, cy
16714 Set the chroma matrix vertical size. It must be an odd integer
16715 between 3 and 23. The default value is 5.
16717 @item chroma_amount, ca
16718 Set the chroma effect strength. It must be a floating point number, reasonable
16719 values lay between -1.5 and 1.5.
16721 Negative values will blur the input video, while positive values will
16722 sharpen it, a value of zero will disable the effect.
16724 Default value is 0.0.
16728 All parameters are optional and default to the equivalent of the
16729 string '5:5:1.0:5:5:0.0'.
16731 @subsection Examples
16735 Apply strong luma sharpen effect:
16737 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
16741 Apply a strong blur of both luma and chroma parameters:
16743 unsharp=7:7:-2:7:7:-2
16749 Apply ultra slow/simple postprocessing filter that compresses and decompresses
16750 the image at several (or - in the case of @option{quality} level @code{8} - all)
16751 shifts and average the results.
16753 The way this differs from the behavior of spp is that uspp actually encodes &
16754 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
16755 DCT similar to MJPEG.
16757 The filter accepts the following options:
16761 Set quality. This option defines the number of levels for averaging. It accepts
16762 an integer in the range 0-8. If set to @code{0}, the filter will have no
16763 effect. A value of @code{8} means the higher quality. For each increment of
16764 that value the speed drops by a factor of approximately 2. Default value is
16768 Force a constant quantization parameter. If not set, the filter will use the QP
16769 from the video stream (if available).
16772 @section vaguedenoiser
16774 Apply a wavelet based denoiser.
16776 It transforms each frame from the video input into the wavelet domain,
16777 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
16778 the obtained coefficients. It does an inverse wavelet transform after.
16779 Due to wavelet properties, it should give a nice smoothed result, and
16780 reduced noise, without blurring picture features.
16782 This filter accepts the following options:
16786 The filtering strength. The higher, the more filtered the video will be.
16787 Hard thresholding can use a higher threshold than soft thresholding
16788 before the video looks overfiltered. Default value is 2.
16791 The filtering method the filter will use.
16793 It accepts the following values:
16796 All values under the threshold will be zeroed.
16799 All values under the threshold will be zeroed. All values above will be
16800 reduced by the threshold.
16803 Scales or nullifies coefficients - intermediary between (more) soft and
16804 (less) hard thresholding.
16807 Default is garrote.
16810 Number of times, the wavelet will decompose the picture. Picture can't
16811 be decomposed beyond a particular point (typically, 8 for a 640x480
16812 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
16815 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
16818 A list of the planes to process. By default all planes are processed.
16821 @section vectorscope
16823 Display 2 color component values in the two dimensional graph (which is called
16826 This filter accepts the following options:
16830 Set vectorscope mode.
16832 It accepts the following values:
16835 Gray values are displayed on graph, higher brightness means more pixels have
16836 same component color value on location in graph. This is the default mode.
16839 Gray values are displayed on graph. Surrounding pixels values which are not
16840 present in video frame are drawn in gradient of 2 color components which are
16841 set by option @code{x} and @code{y}. The 3rd color component is static.
16844 Actual color components values present in video frame are displayed on graph.
16847 Similar as color2 but higher frequency of same values @code{x} and @code{y}
16848 on graph increases value of another color component, which is luminance by
16849 default values of @code{x} and @code{y}.
16852 Actual colors present in video frame are displayed on graph. If two different
16853 colors map to same position on graph then color with higher value of component
16854 not present in graph is picked.
16857 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
16858 component picked from radial gradient.
16862 Set which color component will be represented on X-axis. Default is @code{1}.
16865 Set which color component will be represented on Y-axis. Default is @code{2}.
16868 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
16869 of color component which represents frequency of (X, Y) location in graph.
16874 No envelope, this is default.
16877 Instant envelope, even darkest single pixel will be clearly highlighted.
16880 Hold maximum and minimum values presented in graph over time. This way you
16881 can still spot out of range values without constantly looking at vectorscope.
16884 Peak and instant envelope combined together.
16888 Set what kind of graticule to draw.
16896 Set graticule opacity.
16899 Set graticule flags.
16903 Draw graticule for white point.
16906 Draw graticule for black point.
16909 Draw color points short names.
16913 Set background opacity.
16915 @item lthreshold, l
16916 Set low threshold for color component not represented on X or Y axis.
16917 Values lower than this value will be ignored. Default is 0.
16918 Note this value is multiplied with actual max possible value one pixel component
16919 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
16922 @item hthreshold, h
16923 Set high threshold for color component not represented on X or Y axis.
16924 Values higher than this value will be ignored. Default is 1.
16925 Note this value is multiplied with actual max possible value one pixel component
16926 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
16927 is 0.9 * 255 = 230.
16929 @item colorspace, c
16930 Set what kind of colorspace to use when drawing graticule.
16939 @anchor{vidstabdetect}
16940 @section vidstabdetect
16942 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
16943 @ref{vidstabtransform} for pass 2.
16945 This filter generates a file with relative translation and rotation
16946 transform information about subsequent frames, which is then used by
16947 the @ref{vidstabtransform} filter.
16949 To enable compilation of this filter you need to configure FFmpeg with
16950 @code{--enable-libvidstab}.
16952 This filter accepts the following options:
16956 Set the path to the file used to write the transforms information.
16957 Default value is @file{transforms.trf}.
16960 Set how shaky the video is and how quick the camera is. It accepts an
16961 integer in the range 1-10, a value of 1 means little shakiness, a
16962 value of 10 means strong shakiness. Default value is 5.
16965 Set the accuracy of the detection process. It must be a value in the
16966 range 1-15. A value of 1 means low accuracy, a value of 15 means high
16967 accuracy. Default value is 15.
16970 Set stepsize of the search process. The region around minimum is
16971 scanned with 1 pixel resolution. Default value is 6.
16974 Set minimum contrast. Below this value a local measurement field is
16975 discarded. Must be a floating point value in the range 0-1. Default
16979 Set reference frame number for tripod mode.
16981 If enabled, the motion of the frames is compared to a reference frame
16982 in the filtered stream, identified by the specified number. The idea
16983 is to compensate all movements in a more-or-less static scene and keep
16984 the camera view absolutely still.
16986 If set to 0, it is disabled. The frames are counted starting from 1.
16989 Show fields and transforms in the resulting frames. It accepts an
16990 integer in the range 0-2. Default value is 0, which disables any
16994 @subsection Examples
16998 Use default values:
17004 Analyze strongly shaky movie and put the results in file
17005 @file{mytransforms.trf}:
17007 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
17011 Visualize the result of internal transformations in the resulting
17014 vidstabdetect=show=1
17018 Analyze a video with medium shakiness using @command{ffmpeg}:
17020 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
17024 @anchor{vidstabtransform}
17025 @section vidstabtransform
17027 Video stabilization/deshaking: pass 2 of 2,
17028 see @ref{vidstabdetect} for pass 1.
17030 Read a file with transform information for each frame and
17031 apply/compensate them. Together with the @ref{vidstabdetect}
17032 filter this can be used to deshake videos. See also
17033 @url{http://public.hronopik.de/vid.stab}. It is important to also use
17034 the @ref{unsharp} filter, see below.
17036 To enable compilation of this filter you need to configure FFmpeg with
17037 @code{--enable-libvidstab}.
17039 @subsection Options
17043 Set path to the file used to read the transforms. Default value is
17044 @file{transforms.trf}.
17047 Set the number of frames (value*2 + 1) used for lowpass filtering the
17048 camera movements. Default value is 10.
17050 For example a number of 10 means that 21 frames are used (10 in the
17051 past and 10 in the future) to smoothen the motion in the video. A
17052 larger value leads to a smoother video, but limits the acceleration of
17053 the camera (pan/tilt movements). 0 is a special case where a static
17054 camera is simulated.
17057 Set the camera path optimization algorithm.
17059 Accepted values are:
17062 gaussian kernel low-pass filter on camera motion (default)
17064 averaging on transformations
17068 Set maximal number of pixels to translate frames. Default value is -1,
17072 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
17073 value is -1, meaning no limit.
17076 Specify how to deal with borders that may be visible due to movement
17079 Available values are:
17082 keep image information from previous frame (default)
17084 fill the border black
17088 Invert transforms if set to 1. Default value is 0.
17091 Consider transforms as relative to previous frame if set to 1,
17092 absolute if set to 0. Default value is 0.
17095 Set percentage to zoom. A positive value will result in a zoom-in
17096 effect, a negative value in a zoom-out effect. Default value is 0 (no
17100 Set optimal zooming to avoid borders.
17102 Accepted values are:
17107 optimal static zoom value is determined (only very strong movements
17108 will lead to visible borders) (default)
17110 optimal adaptive zoom value is determined (no borders will be
17111 visible), see @option{zoomspeed}
17114 Note that the value given at zoom is added to the one calculated here.
17117 Set percent to zoom maximally each frame (enabled when
17118 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
17122 Specify type of interpolation.
17124 Available values are:
17129 linear only horizontal
17131 linear in both directions (default)
17133 cubic in both directions (slow)
17137 Enable virtual tripod mode if set to 1, which is equivalent to
17138 @code{relative=0:smoothing=0}. Default value is 0.
17140 Use also @code{tripod} option of @ref{vidstabdetect}.
17143 Increase log verbosity if set to 1. Also the detected global motions
17144 are written to the temporary file @file{global_motions.trf}. Default
17148 @subsection Examples
17152 Use @command{ffmpeg} for a typical stabilization with default values:
17154 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
17157 Note the use of the @ref{unsharp} filter which is always recommended.
17160 Zoom in a bit more and load transform data from a given file:
17162 vidstabtransform=zoom=5:input="mytransforms.trf"
17166 Smoothen the video even more:
17168 vidstabtransform=smoothing=30
17174 Flip the input video vertically.
17176 For example, to vertically flip a video with @command{ffmpeg}:
17178 ffmpeg -i in.avi -vf "vflip" out.avi
17183 Detect variable frame rate video.
17185 This filter tries to detect if the input is variable or constant frame rate.
17187 At end it will output number of frames detected as having variable delta pts,
17188 and ones with constant delta pts.
17189 If there was frames with variable delta, than it will also show min and max delta
17195 Make or reverse a natural vignetting effect.
17197 The filter accepts the following options:
17201 Set lens angle expression as a number of radians.
17203 The value is clipped in the @code{[0,PI/2]} range.
17205 Default value: @code{"PI/5"}
17209 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
17213 Set forward/backward mode.
17215 Available modes are:
17218 The larger the distance from the central point, the darker the image becomes.
17221 The larger the distance from the central point, the brighter the image becomes.
17222 This can be used to reverse a vignette effect, though there is no automatic
17223 detection to extract the lens @option{angle} and other settings (yet). It can
17224 also be used to create a burning effect.
17227 Default value is @samp{forward}.
17230 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
17232 It accepts the following values:
17235 Evaluate expressions only once during the filter initialization.
17238 Evaluate expressions for each incoming frame. This is way slower than the
17239 @samp{init} mode since it requires all the scalers to be re-computed, but it
17240 allows advanced dynamic expressions.
17243 Default value is @samp{init}.
17246 Set dithering to reduce the circular banding effects. Default is @code{1}
17250 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
17251 Setting this value to the SAR of the input will make a rectangular vignetting
17252 following the dimensions of the video.
17254 Default is @code{1/1}.
17257 @subsection Expressions
17259 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
17260 following parameters.
17265 input width and height
17268 the number of input frame, starting from 0
17271 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
17272 @var{TB} units, NAN if undefined
17275 frame rate of the input video, NAN if the input frame rate is unknown
17278 the PTS (Presentation TimeStamp) of the filtered video frame,
17279 expressed in seconds, NAN if undefined
17282 time base of the input video
17286 @subsection Examples
17290 Apply simple strong vignetting effect:
17296 Make a flickering vignetting:
17298 vignette='PI/4+random(1)*PI/50':eval=frame
17303 @section vmafmotion
17305 Obtain the average vmaf motion score of a video.
17306 It is one of the component filters of VMAF.
17308 The obtained average motion score is printed through the logging system.
17310 In the below example the input file @file{ref.mpg} is being processed and score
17314 ffmpeg -i ref.mpg -lavfi vmafmotion -f null -
17318 Stack input videos vertically.
17320 All streams must be of same pixel format and of same width.
17322 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
17323 to create same output.
17325 The filter accept the following option:
17329 Set number of input streams. Default is 2.
17332 If set to 1, force the output to terminate when the shortest input
17333 terminates. Default value is 0.
17338 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
17339 Deinterlacing Filter").
17341 Based on the process described by Martin Weston for BBC R&D, and
17342 implemented based on the de-interlace algorithm written by Jim
17343 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
17344 uses filter coefficients calculated by BBC R&D.
17346 There are two sets of filter coefficients, so called "simple":
17347 and "complex". Which set of filter coefficients is used can
17348 be set by passing an optional parameter:
17352 Set the interlacing filter coefficients. Accepts one of the following values:
17356 Simple filter coefficient set.
17358 More-complex filter coefficient set.
17360 Default value is @samp{complex}.
17363 Specify which frames to deinterlace. Accept one of the following values:
17367 Deinterlace all frames,
17369 Only deinterlace frames marked as interlaced.
17372 Default value is @samp{all}.
17376 Video waveform monitor.
17378 The waveform monitor plots color component intensity. By default luminance
17379 only. Each column of the waveform corresponds to a column of pixels in the
17382 It accepts the following options:
17386 Can be either @code{row}, or @code{column}. Default is @code{column}.
17387 In row mode, the graph on the left side represents color component value 0 and
17388 the right side represents value = 255. In column mode, the top side represents
17389 color component value = 0 and bottom side represents value = 255.
17392 Set intensity. Smaller values are useful to find out how many values of the same
17393 luminance are distributed across input rows/columns.
17394 Default value is @code{0.04}. Allowed range is [0, 1].
17397 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
17398 In mirrored mode, higher values will be represented on the left
17399 side for @code{row} mode and at the top for @code{column} mode. Default is
17400 @code{1} (mirrored).
17404 It accepts the following values:
17407 Presents information identical to that in the @code{parade}, except
17408 that the graphs representing color components are superimposed directly
17411 This display mode makes it easier to spot relative differences or similarities
17412 in overlapping areas of the color components that are supposed to be identical,
17413 such as neutral whites, grays, or blacks.
17416 Display separate graph for the color components side by side in
17417 @code{row} mode or one below the other in @code{column} mode.
17420 Display separate graph for the color components side by side in
17421 @code{column} mode or one below the other in @code{row} mode.
17423 Using this display mode makes it easy to spot color casts in the highlights
17424 and shadows of an image, by comparing the contours of the top and the bottom
17425 graphs of each waveform. Since whites, grays, and blacks are characterized
17426 by exactly equal amounts of red, green, and blue, neutral areas of the picture
17427 should display three waveforms of roughly equal width/height. If not, the
17428 correction is easy to perform by making level adjustments the three waveforms.
17430 Default is @code{stack}.
17432 @item components, c
17433 Set which color components to display. Default is 1, which means only luminance
17434 or red color component if input is in RGB colorspace. If is set for example to
17435 7 it will display all 3 (if) available color components.
17440 No envelope, this is default.
17443 Instant envelope, minimum and maximum values presented in graph will be easily
17444 visible even with small @code{step} value.
17447 Hold minimum and maximum values presented in graph across time. This way you
17448 can still spot out of range values without constantly looking at waveforms.
17451 Peak and instant envelope combined together.
17457 No filtering, this is default.
17460 Luma and chroma combined together.
17463 Similar as above, but shows difference between blue and red chroma.
17466 Similar as above, but use different colors.
17469 Displays only chroma.
17472 Displays actual color value on waveform.
17475 Similar as above, but with luma showing frequency of chroma values.
17479 Set which graticule to display.
17483 Do not display graticule.
17486 Display green graticule showing legal broadcast ranges.
17489 Display orange graticule showing legal broadcast ranges.
17493 Set graticule opacity.
17496 Set graticule flags.
17500 Draw numbers above lines. By default enabled.
17503 Draw dots instead of lines.
17507 Set scale used for displaying graticule.
17514 Default is digital.
17517 Set background opacity.
17520 @section weave, doubleweave
17522 The @code{weave} takes a field-based video input and join
17523 each two sequential fields into single frame, producing a new double
17524 height clip with half the frame rate and half the frame count.
17526 The @code{doubleweave} works same as @code{weave} but without
17527 halving frame rate and frame count.
17529 It accepts the following option:
17533 Set first field. Available values are:
17537 Set the frame as top-field-first.
17540 Set the frame as bottom-field-first.
17544 @subsection Examples
17548 Interlace video using @ref{select} and @ref{separatefields} filter:
17550 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
17555 Apply the xBR high-quality magnification filter which is designed for pixel
17556 art. It follows a set of edge-detection rules, see
17557 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
17559 It accepts the following option:
17563 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
17564 @code{3xBR} and @code{4} for @code{4xBR}.
17565 Default is @code{3}.
17571 Deinterlace the input video ("yadif" means "yet another deinterlacing
17574 It accepts the following parameters:
17580 The interlacing mode to adopt. It accepts one of the following values:
17583 @item 0, send_frame
17584 Output one frame for each frame.
17585 @item 1, send_field
17586 Output one frame for each field.
17587 @item 2, send_frame_nospatial
17588 Like @code{send_frame}, but it skips the spatial interlacing check.
17589 @item 3, send_field_nospatial
17590 Like @code{send_field}, but it skips the spatial interlacing check.
17593 The default value is @code{send_frame}.
17596 The picture field parity assumed for the input interlaced video. It accepts one
17597 of the following values:
17601 Assume the top field is first.
17603 Assume the bottom field is first.
17605 Enable automatic detection of field parity.
17608 The default value is @code{auto}.
17609 If the interlacing is unknown or the decoder does not export this information,
17610 top field first will be assumed.
17613 Specify which frames to deinterlace. Accept one of the following
17618 Deinterlace all frames.
17619 @item 1, interlaced
17620 Only deinterlace frames marked as interlaced.
17623 The default value is @code{all}.
17628 Apply Zoom & Pan effect.
17630 This filter accepts the following options:
17634 Set the zoom expression. Default is 1.
17638 Set the x and y expression. Default is 0.
17641 Set the duration expression in number of frames.
17642 This sets for how many number of frames effect will last for
17643 single input image.
17646 Set the output image size, default is 'hd720'.
17649 Set the output frame rate, default is '25'.
17652 Each expression can contain the following constants:
17671 Output frame count.
17675 Last calculated 'x' and 'y' position from 'x' and 'y' expression
17676 for current input frame.
17680 'x' and 'y' of last output frame of previous input frame or 0 when there was
17681 not yet such frame (first input frame).
17684 Last calculated zoom from 'z' expression for current input frame.
17687 Last calculated zoom of last output frame of previous input frame.
17690 Number of output frames for current input frame. Calculated from 'd' expression
17691 for each input frame.
17694 number of output frames created for previous input frame
17697 Rational number: input width / input height
17700 sample aspect ratio
17703 display aspect ratio
17707 @subsection Examples
17711 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
17713 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
17717 Zoom-in up to 1.5 and pan always at center of picture:
17719 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
17723 Same as above but without pausing:
17725 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
17731 Scale (resize) the input video, using the z.lib library:
17732 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
17733 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
17735 The zscale filter forces the output display aspect ratio to be the same
17736 as the input, by changing the output sample aspect ratio.
17738 If the input image format is different from the format requested by
17739 the next filter, the zscale filter will convert the input to the
17742 @subsection Options
17743 The filter accepts the following options.
17748 Set the output video dimension expression. Default value is the input
17751 If the @var{width} or @var{w} value is 0, the input width is used for
17752 the output. If the @var{height} or @var{h} value is 0, the input height
17753 is used for the output.
17755 If one and only one of the values is -n with n >= 1, the zscale filter
17756 will use a value that maintains the aspect ratio of the input image,
17757 calculated from the other specified dimension. After that it will,
17758 however, make sure that the calculated dimension is divisible by n and
17759 adjust the value if necessary.
17761 If both values are -n with n >= 1, the behavior will be identical to
17762 both values being set to 0 as previously detailed.
17764 See below for the list of accepted constants for use in the dimension
17768 Set the video size. For the syntax of this option, check the
17769 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17772 Set the dither type.
17774 Possible values are:
17779 @item error_diffusion
17785 Set the resize filter type.
17787 Possible values are:
17797 Default is bilinear.
17800 Set the color range.
17802 Possible values are:
17809 Default is same as input.
17812 Set the color primaries.
17814 Possible values are:
17824 Default is same as input.
17827 Set the transfer characteristics.
17829 Possible values are:
17843 Default is same as input.
17846 Set the colorspace matrix.
17848 Possible value are:
17859 Default is same as input.
17862 Set the input color range.
17864 Possible values are:
17871 Default is same as input.
17873 @item primariesin, pin
17874 Set the input color primaries.
17876 Possible values are:
17886 Default is same as input.
17888 @item transferin, tin
17889 Set the input transfer characteristics.
17891 Possible values are:
17902 Default is same as input.
17904 @item matrixin, min
17905 Set the input colorspace matrix.
17907 Possible value are:
17919 Set the output chroma location.
17921 Possible values are:
17932 @item chromalin, cin
17933 Set the input chroma location.
17935 Possible values are:
17947 Set the nominal peak luminance.
17950 The values of the @option{w} and @option{h} options are expressions
17951 containing the following constants:
17956 The input width and height
17960 These are the same as @var{in_w} and @var{in_h}.
17964 The output (scaled) width and height
17968 These are the same as @var{out_w} and @var{out_h}
17971 The same as @var{iw} / @var{ih}
17974 input sample aspect ratio
17977 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17981 horizontal and vertical input chroma subsample values. For example for the
17982 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17986 horizontal and vertical output chroma subsample values. For example for the
17987 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17993 @c man end VIDEO FILTERS
17995 @chapter Video Sources
17996 @c man begin VIDEO SOURCES
17998 Below is a description of the currently available video sources.
18002 Buffer video frames, and make them available to the filter chain.
18004 This source is mainly intended for a programmatic use, in particular
18005 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
18007 It accepts the following parameters:
18012 Specify the size (width and height) of the buffered video frames. For the
18013 syntax of this option, check the
18014 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18017 The input video width.
18020 The input video height.
18023 A string representing the pixel format of the buffered video frames.
18024 It may be a number corresponding to a pixel format, or a pixel format
18028 Specify the timebase assumed by the timestamps of the buffered frames.
18031 Specify the frame rate expected for the video stream.
18033 @item pixel_aspect, sar
18034 The sample (pixel) aspect ratio of the input video.
18037 Specify the optional parameters to be used for the scale filter which
18038 is automatically inserted when an input change is detected in the
18039 input size or format.
18041 @item hw_frames_ctx
18042 When using a hardware pixel format, this should be a reference to an
18043 AVHWFramesContext describing input frames.
18048 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
18051 will instruct the source to accept video frames with size 320x240 and
18052 with format "yuv410p", assuming 1/24 as the timestamps timebase and
18053 square pixels (1:1 sample aspect ratio).
18054 Since the pixel format with name "yuv410p" corresponds to the number 6
18055 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
18056 this example corresponds to:
18058 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
18061 Alternatively, the options can be specified as a flat string, but this
18062 syntax is deprecated:
18064 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}[:@var{sws_param}]
18068 Create a pattern generated by an elementary cellular automaton.
18070 The initial state of the cellular automaton can be defined through the
18071 @option{filename} and @option{pattern} options. If such options are
18072 not specified an initial state is created randomly.
18074 At each new frame a new row in the video is filled with the result of
18075 the cellular automaton next generation. The behavior when the whole
18076 frame is filled is defined by the @option{scroll} option.
18078 This source accepts the following options:
18082 Read the initial cellular automaton state, i.e. the starting row, from
18083 the specified file.
18084 In the file, each non-whitespace character is considered an alive
18085 cell, a newline will terminate the row, and further characters in the
18086 file will be ignored.
18089 Read the initial cellular automaton state, i.e. the starting row, from
18090 the specified string.
18092 Each non-whitespace character in the string is considered an alive
18093 cell, a newline will terminate the row, and further characters in the
18094 string will be ignored.
18097 Set the video rate, that is the number of frames generated per second.
18100 @item random_fill_ratio, ratio
18101 Set the random fill ratio for the initial cellular automaton row. It
18102 is a floating point number value ranging from 0 to 1, defaults to
18105 This option is ignored when a file or a pattern is specified.
18107 @item random_seed, seed
18108 Set the seed for filling randomly the initial row, must be an integer
18109 included between 0 and UINT32_MAX. If not specified, or if explicitly
18110 set to -1, the filter will try to use a good random seed on a best
18114 Set the cellular automaton rule, it is a number ranging from 0 to 255.
18115 Default value is 110.
18118 Set the size of the output video. For the syntax of this option, check the
18119 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18121 If @option{filename} or @option{pattern} is specified, the size is set
18122 by default to the width of the specified initial state row, and the
18123 height is set to @var{width} * PHI.
18125 If @option{size} is set, it must contain the width of the specified
18126 pattern string, and the specified pattern will be centered in the
18129 If a filename or a pattern string is not specified, the size value
18130 defaults to "320x518" (used for a randomly generated initial state).
18133 If set to 1, scroll the output upward when all the rows in the output
18134 have been already filled. If set to 0, the new generated row will be
18135 written over the top row just after the bottom row is filled.
18138 @item start_full, full
18139 If set to 1, completely fill the output with generated rows before
18140 outputting the first frame.
18141 This is the default behavior, for disabling set the value to 0.
18144 If set to 1, stitch the left and right row edges together.
18145 This is the default behavior, for disabling set the value to 0.
18148 @subsection Examples
18152 Read the initial state from @file{pattern}, and specify an output of
18155 cellauto=f=pattern:s=200x400
18159 Generate a random initial row with a width of 200 cells, with a fill
18162 cellauto=ratio=2/3:s=200x200
18166 Create a pattern generated by rule 18 starting by a single alive cell
18167 centered on an initial row with width 100:
18169 cellauto=p=@@:s=100x400:full=0:rule=18
18173 Specify a more elaborated initial pattern:
18175 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
18180 @anchor{coreimagesrc}
18181 @section coreimagesrc
18182 Video source generated on GPU using Apple's CoreImage API on OSX.
18184 This video source is a specialized version of the @ref{coreimage} video filter.
18185 Use a core image generator at the beginning of the applied filterchain to
18186 generate the content.
18188 The coreimagesrc video source accepts the following options:
18190 @item list_generators
18191 List all available generators along with all their respective options as well as
18192 possible minimum and maximum values along with the default values.
18194 list_generators=true
18198 Specify the size of the sourced video. For the syntax of this option, check the
18199 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18200 The default value is @code{320x240}.
18203 Specify the frame rate of the sourced video, as the number of frames
18204 generated per second. It has to be a string in the format
18205 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
18206 number or a valid video frame rate abbreviation. The default value is
18210 Set the sample aspect ratio of the sourced video.
18213 Set the duration of the sourced video. See
18214 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18215 for the accepted syntax.
18217 If not specified, or the expressed duration is negative, the video is
18218 supposed to be generated forever.
18221 Additionally, all options of the @ref{coreimage} video filter are accepted.
18222 A complete filterchain can be used for further processing of the
18223 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
18224 and examples for details.
18226 @subsection Examples
18231 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
18232 given as complete and escaped command-line for Apple's standard bash shell:
18234 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
18236 This example is equivalent to the QRCode example of @ref{coreimage} without the
18237 need for a nullsrc video source.
18241 @section mandelbrot
18243 Generate a Mandelbrot set fractal, and progressively zoom towards the
18244 point specified with @var{start_x} and @var{start_y}.
18246 This source accepts the following options:
18251 Set the terminal pts value. Default value is 400.
18254 Set the terminal scale value.
18255 Must be a floating point value. Default value is 0.3.
18258 Set the inner coloring mode, that is the algorithm used to draw the
18259 Mandelbrot fractal internal region.
18261 It shall assume one of the following values:
18266 Show time until convergence.
18268 Set color based on point closest to the origin of the iterations.
18273 Default value is @var{mincol}.
18276 Set the bailout value. Default value is 10.0.
18279 Set the maximum of iterations performed by the rendering
18280 algorithm. Default value is 7189.
18283 Set outer coloring mode.
18284 It shall assume one of following values:
18286 @item iteration_count
18287 Set iteration cound mode.
18288 @item normalized_iteration_count
18289 set normalized iteration count mode.
18291 Default value is @var{normalized_iteration_count}.
18294 Set frame rate, expressed as number of frames per second. Default
18298 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
18299 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
18302 Set the initial scale value. Default value is 3.0.
18305 Set the initial x position. Must be a floating point value between
18306 -100 and 100. Default value is -0.743643887037158704752191506114774.
18309 Set the initial y position. Must be a floating point value between
18310 -100 and 100. Default value is -0.131825904205311970493132056385139.
18315 Generate various test patterns, as generated by the MPlayer test filter.
18317 The size of the generated video is fixed, and is 256x256.
18318 This source is useful in particular for testing encoding features.
18320 This source accepts the following options:
18325 Specify the frame rate of the sourced video, as the number of frames
18326 generated per second. It has to be a string in the format
18327 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
18328 number or a valid video frame rate abbreviation. The default value is
18332 Set the duration of the sourced video. See
18333 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18334 for the accepted syntax.
18336 If not specified, or the expressed duration is negative, the video is
18337 supposed to be generated forever.
18341 Set the number or the name of the test to perform. Supported tests are:
18357 Default value is "all", which will cycle through the list of all tests.
18362 mptestsrc=t=dc_luma
18365 will generate a "dc_luma" test pattern.
18367 @section frei0r_src
18369 Provide a frei0r source.
18371 To enable compilation of this filter you need to install the frei0r
18372 header and configure FFmpeg with @code{--enable-frei0r}.
18374 This source accepts the following parameters:
18379 The size of the video to generate. For the syntax of this option, check the
18380 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18383 The framerate of the generated video. It may be a string of the form
18384 @var{num}/@var{den} or a frame rate abbreviation.
18387 The name to the frei0r source to load. For more information regarding frei0r and
18388 how to set the parameters, read the @ref{frei0r} section in the video filters
18391 @item filter_params
18392 A '|'-separated list of parameters to pass to the frei0r source.
18396 For example, to generate a frei0r partik0l source with size 200x200
18397 and frame rate 10 which is overlaid on the overlay filter main input:
18399 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
18404 Generate a life pattern.
18406 This source is based on a generalization of John Conway's life game.
18408 The sourced input represents a life grid, each pixel represents a cell
18409 which can be in one of two possible states, alive or dead. Every cell
18410 interacts with its eight neighbours, which are the cells that are
18411 horizontally, vertically, or diagonally adjacent.
18413 At each interaction the grid evolves according to the adopted rule,
18414 which specifies the number of neighbor alive cells which will make a
18415 cell stay alive or born. The @option{rule} option allows one to specify
18418 This source accepts the following options:
18422 Set the file from which to read the initial grid state. In the file,
18423 each non-whitespace character is considered an alive cell, and newline
18424 is used to delimit the end of each row.
18426 If this option is not specified, the initial grid is generated
18430 Set the video rate, that is the number of frames generated per second.
18433 @item random_fill_ratio, ratio
18434 Set the random fill ratio for the initial random grid. It is a
18435 floating point number value ranging from 0 to 1, defaults to 1/PHI.
18436 It is ignored when a file is specified.
18438 @item random_seed, seed
18439 Set the seed for filling the initial random grid, must be an integer
18440 included between 0 and UINT32_MAX. If not specified, or if explicitly
18441 set to -1, the filter will try to use a good random seed on a best
18447 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
18448 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
18449 @var{NS} specifies the number of alive neighbor cells which make a
18450 live cell stay alive, and @var{NB} the number of alive neighbor cells
18451 which make a dead cell to become alive (i.e. to "born").
18452 "s" and "b" can be used in place of "S" and "B", respectively.
18454 Alternatively a rule can be specified by an 18-bits integer. The 9
18455 high order bits are used to encode the next cell state if it is alive
18456 for each number of neighbor alive cells, the low order bits specify
18457 the rule for "borning" new cells. Higher order bits encode for an
18458 higher number of neighbor cells.
18459 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
18460 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
18462 Default value is "S23/B3", which is the original Conway's game of life
18463 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
18464 cells, and will born a new cell if there are three alive cells around
18468 Set the size of the output video. For the syntax of this option, check the
18469 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18471 If @option{filename} is specified, the size is set by default to the
18472 same size of the input file. If @option{size} is set, it must contain
18473 the size specified in the input file, and the initial grid defined in
18474 that file is centered in the larger resulting area.
18476 If a filename is not specified, the size value defaults to "320x240"
18477 (used for a randomly generated initial grid).
18480 If set to 1, stitch the left and right grid edges together, and the
18481 top and bottom edges also. Defaults to 1.
18484 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
18485 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
18486 value from 0 to 255.
18489 Set the color of living (or new born) cells.
18492 Set the color of dead cells. If @option{mold} is set, this is the first color
18493 used to represent a dead cell.
18496 Set mold color, for definitely dead and moldy cells.
18498 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
18499 ffmpeg-utils manual,ffmpeg-utils}.
18502 @subsection Examples
18506 Read a grid from @file{pattern}, and center it on a grid of size
18509 life=f=pattern:s=300x300
18513 Generate a random grid of size 200x200, with a fill ratio of 2/3:
18515 life=ratio=2/3:s=200x200
18519 Specify a custom rule for evolving a randomly generated grid:
18525 Full example with slow death effect (mold) using @command{ffplay}:
18527 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
18534 @anchor{haldclutsrc}
18537 @anchor{pal100bars}
18538 @anchor{rgbtestsrc}
18540 @anchor{smptehdbars}
18543 @anchor{yuvtestsrc}
18544 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
18546 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
18548 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
18550 The @code{color} source provides an uniformly colored input.
18552 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
18553 @ref{haldclut} filter.
18555 The @code{nullsrc} source returns unprocessed video frames. It is
18556 mainly useful to be employed in analysis / debugging tools, or as the
18557 source for filters which ignore the input data.
18559 The @code{pal75bars} source generates a color bars pattern, based on
18560 EBU PAL recommendations with 75% color levels.
18562 The @code{pal100bars} source generates a color bars pattern, based on
18563 EBU PAL recommendations with 100% color levels.
18565 The @code{rgbtestsrc} source generates an RGB test pattern useful for
18566 detecting RGB vs BGR issues. You should see a red, green and blue
18567 stripe from top to bottom.
18569 The @code{smptebars} source generates a color bars pattern, based on
18570 the SMPTE Engineering Guideline EG 1-1990.
18572 The @code{smptehdbars} source generates a color bars pattern, based on
18573 the SMPTE RP 219-2002.
18575 The @code{testsrc} source generates a test video pattern, showing a
18576 color pattern, a scrolling gradient and a timestamp. This is mainly
18577 intended for testing purposes.
18579 The @code{testsrc2} source is similar to testsrc, but supports more
18580 pixel formats instead of just @code{rgb24}. This allows using it as an
18581 input for other tests without requiring a format conversion.
18583 The @code{yuvtestsrc} source generates an YUV test pattern. You should
18584 see a y, cb and cr stripe from top to bottom.
18586 The sources accept the following parameters:
18591 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
18592 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
18593 pixels to be used as identity matrix for 3D lookup tables. Each component is
18594 coded on a @code{1/(N*N)} scale.
18597 Specify the color of the source, only available in the @code{color}
18598 source. For the syntax of this option, check the
18599 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18602 Specify the size of the sourced video. For the syntax of this option, check the
18603 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18604 The default value is @code{320x240}.
18606 This option is not available with the @code{allrgb}, @code{allyuv}, and
18607 @code{haldclutsrc} filters.
18610 Specify the frame rate of the sourced video, as the number of frames
18611 generated per second. It has to be a string in the format
18612 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
18613 number or a valid video frame rate abbreviation. The default value is
18617 Set the duration of the sourced video. See
18618 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18619 for the accepted syntax.
18621 If not specified, or the expressed duration is negative, the video is
18622 supposed to be generated forever.
18625 Set the sample aspect ratio of the sourced video.
18628 Specify the alpha (opacity) of the background, only available in the
18629 @code{testsrc2} source. The value must be between 0 (fully transparent) and
18630 255 (fully opaque, the default).
18633 Set the number of decimals to show in the timestamp, only available in the
18634 @code{testsrc} source.
18636 The displayed timestamp value will correspond to the original
18637 timestamp value multiplied by the power of 10 of the specified
18638 value. Default value is 0.
18641 @subsection Examples
18645 Generate a video with a duration of 5.3 seconds, with size
18646 176x144 and a frame rate of 10 frames per second:
18648 testsrc=duration=5.3:size=qcif:rate=10
18652 The following graph description will generate a red source
18653 with an opacity of 0.2, with size "qcif" and a frame rate of 10
18656 color=c=red@@0.2:s=qcif:r=10
18660 If the input content is to be ignored, @code{nullsrc} can be used. The
18661 following command generates noise in the luminance plane by employing
18662 the @code{geq} filter:
18664 nullsrc=s=256x256, geq=random(1)*255:128:128
18668 @subsection Commands
18670 The @code{color} source supports the following commands:
18674 Set the color of the created image. Accepts the same syntax of the
18675 corresponding @option{color} option.
18680 Generate video using an OpenCL program.
18685 OpenCL program source file.
18688 Kernel name in program.
18691 Size of frames to generate. This must be set.
18694 Pixel format to use for the generated frames. This must be set.
18697 Number of frames generated every second. Default value is '25'.
18701 For details of how the program loading works, see the @ref{program_opencl}
18708 Generate a colour ramp by setting pixel values from the position of the pixel
18709 in the output image. (Note that this will work with all pixel formats, but
18710 the generated output will not be the same.)
18712 __kernel void ramp(__write_only image2d_t dst,
18713 unsigned int index)
18715 int2 loc = (int2)(get_global_id(0), get_global_id(1));
18718 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
18720 write_imagef(dst, loc, val);
18725 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
18727 __kernel void sierpinski_carpet(__write_only image2d_t dst,
18728 unsigned int index)
18730 int2 loc = (int2)(get_global_id(0), get_global_id(1));
18732 float4 value = 0.0f;
18733 int x = loc.x + index;
18734 int y = loc.y + index;
18735 while (x > 0 || y > 0) {
18736 if (x % 3 == 1 && y % 3 == 1) {
18744 write_imagef(dst, loc, value);
18750 @c man end VIDEO SOURCES
18752 @chapter Video Sinks
18753 @c man begin VIDEO SINKS
18755 Below is a description of the currently available video sinks.
18757 @section buffersink
18759 Buffer video frames, and make them available to the end of the filter
18762 This sink is mainly intended for programmatic use, in particular
18763 through the interface defined in @file{libavfilter/buffersink.h}
18764 or the options system.
18766 It accepts a pointer to an AVBufferSinkContext structure, which
18767 defines the incoming buffers' formats, to be passed as the opaque
18768 parameter to @code{avfilter_init_filter} for initialization.
18772 Null video sink: do absolutely nothing with the input video. It is
18773 mainly useful as a template and for use in analysis / debugging
18776 @c man end VIDEO SINKS
18778 @chapter Multimedia Filters
18779 @c man begin MULTIMEDIA FILTERS
18781 Below is a description of the currently available multimedia filters.
18785 Convert input audio to a video output, displaying the audio bit scope.
18787 The filter accepts the following options:
18791 Set frame rate, expressed as number of frames per second. Default
18795 Specify the video size for the output. For the syntax of this option, check the
18796 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18797 Default value is @code{1024x256}.
18800 Specify list of colors separated by space or by '|' which will be used to
18801 draw channels. Unrecognized or missing colors will be replaced
18805 @section ahistogram
18807 Convert input audio to a video output, displaying the volume histogram.
18809 The filter accepts the following options:
18813 Specify how histogram is calculated.
18815 It accepts the following values:
18818 Use single histogram for all channels.
18820 Use separate histogram for each channel.
18822 Default is @code{single}.
18825 Set frame rate, expressed as number of frames per second. Default
18829 Specify the video size for the output. For the syntax of this option, check the
18830 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18831 Default value is @code{hd720}.
18836 It accepts the following values:
18847 reverse logarithmic
18849 Default is @code{log}.
18852 Set amplitude scale.
18854 It accepts the following values:
18861 Default is @code{log}.
18864 Set how much frames to accumulate in histogram.
18865 Defauls is 1. Setting this to -1 accumulates all frames.
18868 Set histogram ratio of window height.
18871 Set sonogram sliding.
18873 It accepts the following values:
18876 replace old rows with new ones.
18878 scroll from top to bottom.
18880 Default is @code{replace}.
18883 @section aphasemeter
18885 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
18886 representing mean phase of current audio frame. A video output can also be produced and is
18887 enabled by default. The audio is passed through as first output.
18889 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
18890 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
18891 and @code{1} means channels are in phase.
18893 The filter accepts the following options, all related to its video output:
18897 Set the output frame rate. Default value is @code{25}.
18900 Set the video size for the output. For the syntax of this option, check the
18901 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18902 Default value is @code{800x400}.
18907 Specify the red, green, blue contrast. Default values are @code{2},
18908 @code{7} and @code{1}.
18909 Allowed range is @code{[0, 255]}.
18912 Set color which will be used for drawing median phase. If color is
18913 @code{none} which is default, no median phase value will be drawn.
18916 Enable video output. Default is enabled.
18919 @section avectorscope
18921 Convert input audio to a video output, representing the audio vector
18924 The filter is used to measure the difference between channels of stereo
18925 audio stream. A monoaural signal, consisting of identical left and right
18926 signal, results in straight vertical line. Any stereo separation is visible
18927 as a deviation from this line, creating a Lissajous figure.
18928 If the straight (or deviation from it) but horizontal line appears this
18929 indicates that the left and right channels are out of phase.
18931 The filter accepts the following options:
18935 Set the vectorscope mode.
18937 Available values are:
18940 Lissajous rotated by 45 degrees.
18943 Same as above but not rotated.
18946 Shape resembling half of circle.
18949 Default value is @samp{lissajous}.
18952 Set the video size for the output. For the syntax of this option, check the
18953 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18954 Default value is @code{400x400}.
18957 Set the output frame rate. Default value is @code{25}.
18963 Specify the red, green, blue and alpha contrast. Default values are @code{40},
18964 @code{160}, @code{80} and @code{255}.
18965 Allowed range is @code{[0, 255]}.
18971 Specify the red, green, blue and alpha fade. Default values are @code{15},
18972 @code{10}, @code{5} and @code{5}.
18973 Allowed range is @code{[0, 255]}.
18976 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
18977 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
18980 Set the vectorscope drawing mode.
18982 Available values are:
18985 Draw dot for each sample.
18988 Draw line between previous and current sample.
18991 Default value is @samp{dot}.
18994 Specify amplitude scale of audio samples.
18996 Available values are:
19012 Swap left channel axis with right channel axis.
19022 Mirror only x axis.
19025 Mirror only y axis.
19033 @subsection Examples
19037 Complete example using @command{ffplay}:
19039 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
19040 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
19044 @section bench, abench
19046 Benchmark part of a filtergraph.
19048 The filter accepts the following options:
19052 Start or stop a timer.
19054 Available values are:
19057 Get the current time, set it as frame metadata (using the key
19058 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
19061 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
19062 the input frame metadata to get the time difference. Time difference, average,
19063 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
19064 @code{min}) are then printed. The timestamps are expressed in seconds.
19068 @subsection Examples
19072 Benchmark @ref{selectivecolor} filter:
19074 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
19080 Concatenate audio and video streams, joining them together one after the
19083 The filter works on segments of synchronized video and audio streams. All
19084 segments must have the same number of streams of each type, and that will
19085 also be the number of streams at output.
19087 The filter accepts the following options:
19092 Set the number of segments. Default is 2.
19095 Set the number of output video streams, that is also the number of video
19096 streams in each segment. Default is 1.
19099 Set the number of output audio streams, that is also the number of audio
19100 streams in each segment. Default is 0.
19103 Activate unsafe mode: do not fail if segments have a different format.
19107 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
19108 @var{a} audio outputs.
19110 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
19111 segment, in the same order as the outputs, then the inputs for the second
19114 Related streams do not always have exactly the same duration, for various
19115 reasons including codec frame size or sloppy authoring. For that reason,
19116 related synchronized streams (e.g. a video and its audio track) should be
19117 concatenated at once. The concat filter will use the duration of the longest
19118 stream in each segment (except the last one), and if necessary pad shorter
19119 audio streams with silence.
19121 For this filter to work correctly, all segments must start at timestamp 0.
19123 All corresponding streams must have the same parameters in all segments; the
19124 filtering system will automatically select a common pixel format for video
19125 streams, and a common sample format, sample rate and channel layout for
19126 audio streams, but other settings, such as resolution, must be converted
19127 explicitly by the user.
19129 Different frame rates are acceptable but will result in variable frame rate
19130 at output; be sure to configure the output file to handle it.
19132 @subsection Examples
19136 Concatenate an opening, an episode and an ending, all in bilingual version
19137 (video in stream 0, audio in streams 1 and 2):
19139 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
19140 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
19141 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
19142 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
19146 Concatenate two parts, handling audio and video separately, using the
19147 (a)movie sources, and adjusting the resolution:
19149 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
19150 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
19151 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
19153 Note that a desync will happen at the stitch if the audio and video streams
19154 do not have exactly the same duration in the first file.
19158 @subsection Commands
19160 This filter supports the following commands:
19163 Close the current segment and step to the next one
19166 @section drawgraph, adrawgraph
19168 Draw a graph using input video or audio metadata.
19170 It accepts the following parameters:
19174 Set 1st frame metadata key from which metadata values will be used to draw a graph.
19177 Set 1st foreground color expression.
19180 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
19183 Set 2nd foreground color expression.
19186 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
19189 Set 3rd foreground color expression.
19192 Set 4th frame metadata key from which metadata values will be used to draw a graph.
19195 Set 4th foreground color expression.
19198 Set minimal value of metadata value.
19201 Set maximal value of metadata value.
19204 Set graph background color. Default is white.
19209 Available values for mode is:
19216 Default is @code{line}.
19221 Available values for slide is:
19224 Draw new frame when right border is reached.
19227 Replace old columns with new ones.
19230 Scroll from right to left.
19233 Scroll from left to right.
19236 Draw single picture.
19239 Default is @code{frame}.
19242 Set size of graph video. For the syntax of this option, check the
19243 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19244 The default value is @code{900x256}.
19246 The foreground color expressions can use the following variables:
19249 Minimal value of metadata value.
19252 Maximal value of metadata value.
19255 Current metadata key value.
19258 The color is defined as 0xAABBGGRR.
19261 Example using metadata from @ref{signalstats} filter:
19263 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
19266 Example using metadata from @ref{ebur128} filter:
19268 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
19274 EBU R128 scanner filter. This filter takes an audio stream as input and outputs
19275 it unchanged. By default, it logs a message at a frequency of 10Hz with the
19276 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
19277 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
19279 The filter also has a video output (see the @var{video} option) with a real
19280 time graph to observe the loudness evolution. The graphic contains the logged
19281 message mentioned above, so it is not printed anymore when this option is set,
19282 unless the verbose logging is set. The main graphing area contains the
19283 short-term loudness (3 seconds of analysis), and the gauge on the right is for
19284 the momentary loudness (400 milliseconds).
19286 More information about the Loudness Recommendation EBU R128 on
19287 @url{http://tech.ebu.ch/loudness}.
19289 The filter accepts the following options:
19294 Activate the video output. The audio stream is passed unchanged whether this
19295 option is set or no. The video stream will be the first output stream if
19296 activated. Default is @code{0}.
19299 Set the video size. This option is for video only. For the syntax of this
19301 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19302 Default and minimum resolution is @code{640x480}.
19305 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
19306 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
19307 other integer value between this range is allowed.
19310 Set metadata injection. If set to @code{1}, the audio input will be segmented
19311 into 100ms output frames, each of them containing various loudness information
19312 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
19314 Default is @code{0}.
19317 Force the frame logging level.
19319 Available values are:
19322 information logging level
19324 verbose logging level
19327 By default, the logging level is set to @var{info}. If the @option{video} or
19328 the @option{metadata} options are set, it switches to @var{verbose}.
19333 Available modes can be cumulated (the option is a @code{flag} type). Possible
19337 Disable any peak mode (default).
19339 Enable sample-peak mode.
19341 Simple peak mode looking for the higher sample value. It logs a message
19342 for sample-peak (identified by @code{SPK}).
19344 Enable true-peak mode.
19346 If enabled, the peak lookup is done on an over-sampled version of the input
19347 stream for better peak accuracy. It logs a message for true-peak.
19348 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
19349 This mode requires a build with @code{libswresample}.
19353 Treat mono input files as "dual mono". If a mono file is intended for playback
19354 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
19355 If set to @code{true}, this option will compensate for this effect.
19356 Multi-channel input files are not affected by this option.
19359 Set a specific pan law to be used for the measurement of dual mono files.
19360 This parameter is optional, and has a default value of -3.01dB.
19363 @subsection Examples
19367 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
19369 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
19373 Run an analysis with @command{ffmpeg}:
19375 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
19379 @section interleave, ainterleave
19381 Temporally interleave frames from several inputs.
19383 @code{interleave} works with video inputs, @code{ainterleave} with audio.
19385 These filters read frames from several inputs and send the oldest
19386 queued frame to the output.
19388 Input streams must have well defined, monotonically increasing frame
19391 In order to submit one frame to output, these filters need to enqueue
19392 at least one frame for each input, so they cannot work in case one
19393 input is not yet terminated and will not receive incoming frames.
19395 For example consider the case when one input is a @code{select} filter
19396 which always drops input frames. The @code{interleave} filter will keep
19397 reading from that input, but it will never be able to send new frames
19398 to output until the input sends an end-of-stream signal.
19400 Also, depending on inputs synchronization, the filters will drop
19401 frames in case one input receives more frames than the other ones, and
19402 the queue is already filled.
19404 These filters accept the following options:
19408 Set the number of different inputs, it is 2 by default.
19411 @subsection Examples
19415 Interleave frames belonging to different streams using @command{ffmpeg}:
19417 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
19421 Add flickering blur effect:
19423 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
19427 @section metadata, ametadata
19429 Manipulate frame metadata.
19431 This filter accepts the following options:
19435 Set mode of operation of the filter.
19437 Can be one of the following:
19441 If both @code{value} and @code{key} is set, select frames
19442 which have such metadata. If only @code{key} is set, select
19443 every frame that has such key in metadata.
19446 Add new metadata @code{key} and @code{value}. If key is already available
19450 Modify value of already present key.
19453 If @code{value} is set, delete only keys that have such value.
19454 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
19458 Print key and its value if metadata was found. If @code{key} is not set print all
19459 metadata values available in frame.
19463 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
19466 Set metadata value which will be used. This option is mandatory for
19467 @code{modify} and @code{add} mode.
19470 Which function to use when comparing metadata value and @code{value}.
19472 Can be one of following:
19476 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
19479 Values are interpreted as strings, returns true if metadata value starts with
19480 the @code{value} option string.
19483 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
19486 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
19489 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
19492 Values are interpreted as floats, returns true if expression from option @code{expr}
19497 Set expression which is used when @code{function} is set to @code{expr}.
19498 The expression is evaluated through the eval API and can contain the following
19503 Float representation of @code{value} from metadata key.
19506 Float representation of @code{value} as supplied by user in @code{value} option.
19510 If specified in @code{print} mode, output is written to the named file. Instead of
19511 plain filename any writable url can be specified. Filename ``-'' is a shorthand
19512 for standard output. If @code{file} option is not set, output is written to the log
19513 with AV_LOG_INFO loglevel.
19517 @subsection Examples
19521 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
19524 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
19527 Print silencedetect output to file @file{metadata.txt}.
19529 silencedetect,ametadata=mode=print:file=metadata.txt
19532 Direct all metadata to a pipe with file descriptor 4.
19534 metadata=mode=print:file='pipe\:4'
19538 @section perms, aperms
19540 Set read/write permissions for the output frames.
19542 These filters are mainly aimed at developers to test direct path in the
19543 following filter in the filtergraph.
19545 The filters accept the following options:
19549 Select the permissions mode.
19551 It accepts the following values:
19554 Do nothing. This is the default.
19556 Set all the output frames read-only.
19558 Set all the output frames directly writable.
19560 Make the frame read-only if writable, and writable if read-only.
19562 Set each output frame read-only or writable randomly.
19566 Set the seed for the @var{random} mode, must be an integer included between
19567 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
19568 @code{-1}, the filter will try to use a good random seed on a best effort
19572 Note: in case of auto-inserted filter between the permission filter and the
19573 following one, the permission might not be received as expected in that
19574 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
19575 perms/aperms filter can avoid this problem.
19577 @section realtime, arealtime
19579 Slow down filtering to match real time approximately.
19581 These filters will pause the filtering for a variable amount of time to
19582 match the output rate with the input timestamps.
19583 They are similar to the @option{re} option to @code{ffmpeg}.
19585 They accept the following options:
19589 Time limit for the pauses. Any pause longer than that will be considered
19590 a timestamp discontinuity and reset the timer. Default is 2 seconds.
19594 @section select, aselect
19596 Select frames to pass in output.
19598 This filter accepts the following options:
19603 Set expression, which is evaluated for each input frame.
19605 If the expression is evaluated to zero, the frame is discarded.
19607 If the evaluation result is negative or NaN, the frame is sent to the
19608 first output; otherwise it is sent to the output with index
19609 @code{ceil(val)-1}, assuming that the input index starts from 0.
19611 For example a value of @code{1.2} corresponds to the output with index
19612 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
19615 Set the number of outputs. The output to which to send the selected
19616 frame is based on the result of the evaluation. Default value is 1.
19619 The expression can contain the following constants:
19623 The (sequential) number of the filtered frame, starting from 0.
19626 The (sequential) number of the selected frame, starting from 0.
19628 @item prev_selected_n
19629 The sequential number of the last selected frame. It's NAN if undefined.
19632 The timebase of the input timestamps.
19635 The PTS (Presentation TimeStamp) of the filtered video frame,
19636 expressed in @var{TB} units. It's NAN if undefined.
19639 The PTS of the filtered video frame,
19640 expressed in seconds. It's NAN if undefined.
19643 The PTS of the previously filtered video frame. It's NAN if undefined.
19645 @item prev_selected_pts
19646 The PTS of the last previously filtered video frame. It's NAN if undefined.
19648 @item prev_selected_t
19649 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
19652 The PTS of the first video frame in the video. It's NAN if undefined.
19655 The time of the first video frame in the video. It's NAN if undefined.
19657 @item pict_type @emph{(video only)}
19658 The type of the filtered frame. It can assume one of the following
19670 @item interlace_type @emph{(video only)}
19671 The frame interlace type. It can assume one of the following values:
19674 The frame is progressive (not interlaced).
19676 The frame is top-field-first.
19678 The frame is bottom-field-first.
19681 @item consumed_sample_n @emph{(audio only)}
19682 the number of selected samples before the current frame
19684 @item samples_n @emph{(audio only)}
19685 the number of samples in the current frame
19687 @item sample_rate @emph{(audio only)}
19688 the input sample rate
19691 This is 1 if the filtered frame is a key-frame, 0 otherwise.
19694 the position in the file of the filtered frame, -1 if the information
19695 is not available (e.g. for synthetic video)
19697 @item scene @emph{(video only)}
19698 value between 0 and 1 to indicate a new scene; a low value reflects a low
19699 probability for the current frame to introduce a new scene, while a higher
19700 value means the current frame is more likely to be one (see the example below)
19702 @item concatdec_select
19703 The concat demuxer can select only part of a concat input file by setting an
19704 inpoint and an outpoint, but the output packets may not be entirely contained
19705 in the selected interval. By using this variable, it is possible to skip frames
19706 generated by the concat demuxer which are not exactly contained in the selected
19709 This works by comparing the frame pts against the @var{lavf.concat.start_time}
19710 and the @var{lavf.concat.duration} packet metadata values which are also
19711 present in the decoded frames.
19713 The @var{concatdec_select} variable is -1 if the frame pts is at least
19714 start_time and either the duration metadata is missing or the frame pts is less
19715 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
19718 That basically means that an input frame is selected if its pts is within the
19719 interval set by the concat demuxer.
19723 The default value of the select expression is "1".
19725 @subsection Examples
19729 Select all frames in input:
19734 The example above is the same as:
19746 Select only I-frames:
19748 select='eq(pict_type\,I)'
19752 Select one frame every 100:
19754 select='not(mod(n\,100))'
19758 Select only frames contained in the 10-20 time interval:
19760 select=between(t\,10\,20)
19764 Select only I-frames contained in the 10-20 time interval:
19766 select=between(t\,10\,20)*eq(pict_type\,I)
19770 Select frames with a minimum distance of 10 seconds:
19772 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
19776 Use aselect to select only audio frames with samples number > 100:
19778 aselect='gt(samples_n\,100)'
19782 Create a mosaic of the first scenes:
19784 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
19787 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
19791 Send even and odd frames to separate outputs, and compose them:
19793 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
19797 Select useful frames from an ffconcat file which is using inpoints and
19798 outpoints but where the source files are not intra frame only.
19800 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
19804 @section sendcmd, asendcmd
19806 Send commands to filters in the filtergraph.
19808 These filters read commands to be sent to other filters in the
19811 @code{sendcmd} must be inserted between two video filters,
19812 @code{asendcmd} must be inserted between two audio filters, but apart
19813 from that they act the same way.
19815 The specification of commands can be provided in the filter arguments
19816 with the @var{commands} option, or in a file specified by the
19817 @var{filename} option.
19819 These filters accept the following options:
19822 Set the commands to be read and sent to the other filters.
19824 Set the filename of the commands to be read and sent to the other
19828 @subsection Commands syntax
19830 A commands description consists of a sequence of interval
19831 specifications, comprising a list of commands to be executed when a
19832 particular event related to that interval occurs. The occurring event
19833 is typically the current frame time entering or leaving a given time
19836 An interval is specified by the following syntax:
19838 @var{START}[-@var{END}] @var{COMMANDS};
19841 The time interval is specified by the @var{START} and @var{END} times.
19842 @var{END} is optional and defaults to the maximum time.
19844 The current frame time is considered within the specified interval if
19845 it is included in the interval [@var{START}, @var{END}), that is when
19846 the time is greater or equal to @var{START} and is lesser than
19849 @var{COMMANDS} consists of a sequence of one or more command
19850 specifications, separated by ",", relating to that interval. The
19851 syntax of a command specification is given by:
19853 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
19856 @var{FLAGS} is optional and specifies the type of events relating to
19857 the time interval which enable sending the specified command, and must
19858 be a non-null sequence of identifier flags separated by "+" or "|" and
19859 enclosed between "[" and "]".
19861 The following flags are recognized:
19864 The command is sent when the current frame timestamp enters the
19865 specified interval. In other words, the command is sent when the
19866 previous frame timestamp was not in the given interval, and the
19870 The command is sent when the current frame timestamp leaves the
19871 specified interval. In other words, the command is sent when the
19872 previous frame timestamp was in the given interval, and the
19876 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
19879 @var{TARGET} specifies the target of the command, usually the name of
19880 the filter class or a specific filter instance name.
19882 @var{COMMAND} specifies the name of the command for the target filter.
19884 @var{ARG} is optional and specifies the optional list of argument for
19885 the given @var{COMMAND}.
19887 Between one interval specification and another, whitespaces, or
19888 sequences of characters starting with @code{#} until the end of line,
19889 are ignored and can be used to annotate comments.
19891 A simplified BNF description of the commands specification syntax
19894 @var{COMMAND_FLAG} ::= "enter" | "leave"
19895 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
19896 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
19897 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
19898 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
19899 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
19902 @subsection Examples
19906 Specify audio tempo change at second 4:
19908 asendcmd=c='4.0 atempo tempo 1.5',atempo
19912 Target a specific filter instance:
19914 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
19918 Specify a list of drawtext and hue commands in a file.
19920 # show text in the interval 5-10
19921 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
19922 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
19924 # desaturate the image in the interval 15-20
19925 15.0-20.0 [enter] hue s 0,
19926 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
19928 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
19930 # apply an exponential saturation fade-out effect, starting from time 25
19931 25 [enter] hue s exp(25-t)
19934 A filtergraph allowing to read and process the above command list
19935 stored in a file @file{test.cmd}, can be specified with:
19937 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
19942 @section setpts, asetpts
19944 Change the PTS (presentation timestamp) of the input frames.
19946 @code{setpts} works on video frames, @code{asetpts} on audio frames.
19948 This filter accepts the following options:
19953 The expression which is evaluated for each frame to construct its timestamp.
19957 The expression is evaluated through the eval API and can contain the following
19961 @item FRAME_RATE, FR
19962 frame rate, only defined for constant frame-rate video
19965 The presentation timestamp in input
19968 The count of the input frame for video or the number of consumed samples,
19969 not including the current frame for audio, starting from 0.
19971 @item NB_CONSUMED_SAMPLES
19972 The number of consumed samples, not including the current frame (only
19975 @item NB_SAMPLES, S
19976 The number of samples in the current frame (only audio)
19978 @item SAMPLE_RATE, SR
19979 The audio sample rate.
19982 The PTS of the first frame.
19985 the time in seconds of the first frame
19988 State whether the current frame is interlaced.
19991 the time in seconds of the current frame
19994 original position in the file of the frame, or undefined if undefined
19995 for the current frame
19998 The previous input PTS.
20001 previous input time in seconds
20004 The previous output PTS.
20007 previous output time in seconds
20010 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
20014 The wallclock (RTC) time at the start of the movie in microseconds.
20017 The timebase of the input timestamps.
20021 @subsection Examples
20025 Start counting PTS from zero
20027 setpts=PTS-STARTPTS
20031 Apply fast motion effect:
20037 Apply slow motion effect:
20043 Set fixed rate of 25 frames per second:
20049 Set fixed rate 25 fps with some jitter:
20051 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
20055 Apply an offset of 10 seconds to the input PTS:
20061 Generate timestamps from a "live source" and rebase onto the current timebase:
20063 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
20067 Generate timestamps by counting samples:
20076 Force color range for the output video frame.
20078 The @code{setrange} filter marks the color range property for the
20079 output frames. It does not change the input frame, but only sets the
20080 corresponding property, which affects how the frame is treated by
20083 The filter accepts the following options:
20088 Available values are:
20092 Keep the same color range property.
20094 @item unspecified, unknown
20095 Set the color range as unspecified.
20097 @item limited, tv, mpeg
20098 Set the color range as limited.
20100 @item full, pc, jpeg
20101 Set the color range as full.
20105 @section settb, asettb
20107 Set the timebase to use for the output frames timestamps.
20108 It is mainly useful for testing timebase configuration.
20110 It accepts the following parameters:
20115 The expression which is evaluated into the output timebase.
20119 The value for @option{tb} is an arithmetic expression representing a
20120 rational. The expression can contain the constants "AVTB" (the default
20121 timebase), "intb" (the input timebase) and "sr" (the sample rate,
20122 audio only). Default value is "intb".
20124 @subsection Examples
20128 Set the timebase to 1/25:
20134 Set the timebase to 1/10:
20140 Set the timebase to 1001/1000:
20146 Set the timebase to 2*intb:
20152 Set the default timebase value:
20159 Convert input audio to a video output representing frequency spectrum
20160 logarithmically using Brown-Puckette constant Q transform algorithm with
20161 direct frequency domain coefficient calculation (but the transform itself
20162 is not really constant Q, instead the Q factor is actually variable/clamped),
20163 with musical tone scale, from E0 to D#10.
20165 The filter accepts the following options:
20169 Specify the video size for the output. It must be even. For the syntax of this option,
20170 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20171 Default value is @code{1920x1080}.
20174 Set the output frame rate. Default value is @code{25}.
20177 Set the bargraph height. It must be even. Default value is @code{-1} which
20178 computes the bargraph height automatically.
20181 Set the axis height. It must be even. Default value is @code{-1} which computes
20182 the axis height automatically.
20185 Set the sonogram height. It must be even. Default value is @code{-1} which
20186 computes the sonogram height automatically.
20189 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
20190 instead. Default value is @code{1}.
20192 @item sono_v, volume
20193 Specify the sonogram volume expression. It can contain variables:
20196 the @var{bar_v} evaluated expression
20197 @item frequency, freq, f
20198 the frequency where it is evaluated
20199 @item timeclamp, tc
20200 the value of @var{timeclamp} option
20204 @item a_weighting(f)
20205 A-weighting of equal loudness
20206 @item b_weighting(f)
20207 B-weighting of equal loudness
20208 @item c_weighting(f)
20209 C-weighting of equal loudness.
20211 Default value is @code{16}.
20213 @item bar_v, volume2
20214 Specify the bargraph volume expression. It can contain variables:
20217 the @var{sono_v} evaluated expression
20218 @item frequency, freq, f
20219 the frequency where it is evaluated
20220 @item timeclamp, tc
20221 the value of @var{timeclamp} option
20225 @item a_weighting(f)
20226 A-weighting of equal loudness
20227 @item b_weighting(f)
20228 B-weighting of equal loudness
20229 @item c_weighting(f)
20230 C-weighting of equal loudness.
20232 Default value is @code{sono_v}.
20234 @item sono_g, gamma
20235 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
20236 higher gamma makes the spectrum having more range. Default value is @code{3}.
20237 Acceptable range is @code{[1, 7]}.
20239 @item bar_g, gamma2
20240 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
20244 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
20245 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
20247 @item timeclamp, tc
20248 Specify the transform timeclamp. At low frequency, there is trade-off between
20249 accuracy in time domain and frequency domain. If timeclamp is lower,
20250 event in time domain is represented more accurately (such as fast bass drum),
20251 otherwise event in frequency domain is represented more accurately
20252 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
20255 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
20256 limits future samples by applying asymmetric windowing in time domain, useful
20257 when low latency is required. Accepted range is @code{[0, 1]}.
20260 Specify the transform base frequency. Default value is @code{20.01523126408007475},
20261 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
20264 Specify the transform end frequency. Default value is @code{20495.59681441799654},
20265 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
20268 This option is deprecated and ignored.
20271 Specify the transform length in time domain. Use this option to control accuracy
20272 trade-off between time domain and frequency domain at every frequency sample.
20273 It can contain variables:
20275 @item frequency, freq, f
20276 the frequency where it is evaluated
20277 @item timeclamp, tc
20278 the value of @var{timeclamp} option.
20280 Default value is @code{384*tc/(384+tc*f)}.
20283 Specify the transform count for every video frame. Default value is @code{6}.
20284 Acceptable range is @code{[1, 30]}.
20287 Specify the transform count for every single pixel. Default value is @code{0},
20288 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
20291 Specify font file for use with freetype to draw the axis. If not specified,
20292 use embedded font. Note that drawing with font file or embedded font is not
20293 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
20297 Specify fontconfig pattern. This has lower priority than @var{fontfile}.
20298 The : in the pattern may be replaced by | to avoid unnecessary escaping.
20301 Specify font color expression. This is arithmetic expression that should return
20302 integer value 0xRRGGBB. It can contain variables:
20304 @item frequency, freq, f
20305 the frequency where it is evaluated
20306 @item timeclamp, tc
20307 the value of @var{timeclamp} option
20312 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
20313 @item r(x), g(x), b(x)
20314 red, green, and blue value of intensity x.
20316 Default value is @code{st(0, (midi(f)-59.5)/12);
20317 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
20318 r(1-ld(1)) + b(ld(1))}.
20321 Specify image file to draw the axis. This option override @var{fontfile} and
20322 @var{fontcolor} option.
20325 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
20326 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
20327 Default value is @code{1}.
20330 Set colorspace. The accepted values are:
20333 Unspecified (default)
20342 BT.470BG or BT.601-6 625
20345 SMPTE-170M or BT.601-6 525
20351 BT.2020 with non-constant luminance
20356 Set spectrogram color scheme. This is list of floating point values with format
20357 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
20358 The default is @code{1|0.5|0|0|0.5|1}.
20362 @subsection Examples
20366 Playing audio while showing the spectrum:
20368 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
20372 Same as above, but with frame rate 30 fps:
20374 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
20378 Playing at 1280x720:
20380 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
20384 Disable sonogram display:
20390 A1 and its harmonics: A1, A2, (near)E3, A3:
20392 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),
20393 asplit[a][out1]; [a] showcqt [out0]'
20397 Same as above, but with more accuracy in frequency domain:
20399 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),
20400 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
20406 bar_v=10:sono_v=bar_v*a_weighting(f)
20410 Custom gamma, now spectrum is linear to the amplitude.
20416 Custom tlength equation:
20418 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)))'
20422 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
20424 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
20428 Custom font using fontconfig:
20430 font='Courier New,Monospace,mono|bold'
20434 Custom frequency range with custom axis using image file:
20436 axisfile=myaxis.png:basefreq=40:endfreq=10000
20442 Convert input audio to video output representing the audio power spectrum.
20443 Audio amplitude is on Y-axis while frequency is on X-axis.
20445 The filter accepts the following options:
20449 Specify size of video. For the syntax of this option, check the
20450 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20451 Default is @code{1024x512}.
20455 This set how each frequency bin will be represented.
20457 It accepts the following values:
20463 Default is @code{bar}.
20466 Set amplitude scale.
20468 It accepts the following values:
20482 Default is @code{log}.
20485 Set frequency scale.
20487 It accepts the following values:
20496 Reverse logarithmic scale.
20498 Default is @code{lin}.
20503 It accepts the following values:
20519 Default is @code{w2048}
20522 Set windowing function.
20524 It accepts the following values:
20546 Default is @code{hanning}.
20549 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
20550 which means optimal overlap for selected window function will be picked.
20553 Set time averaging. Setting this to 0 will display current maximal peaks.
20554 Default is @code{1}, which means time averaging is disabled.
20557 Specify list of colors separated by space or by '|' which will be used to
20558 draw channel frequencies. Unrecognized or missing colors will be replaced
20562 Set channel display mode.
20564 It accepts the following values:
20569 Default is @code{combined}.
20572 Set minimum amplitude used in @code{log} amplitude scaler.
20576 @anchor{showspectrum}
20577 @section showspectrum
20579 Convert input audio to a video output, representing the audio frequency
20582 The filter accepts the following options:
20586 Specify the video size for the output. For the syntax of this option, check the
20587 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20588 Default value is @code{640x512}.
20591 Specify how the spectrum should slide along the window.
20593 It accepts the following values:
20596 the samples start again on the left when they reach the right
20598 the samples scroll from right to left
20600 frames are only produced when the samples reach the right
20602 the samples scroll from left to right
20605 Default value is @code{replace}.
20608 Specify display mode.
20610 It accepts the following values:
20613 all channels are displayed in the same row
20615 all channels are displayed in separate rows
20618 Default value is @samp{combined}.
20621 Specify display color mode.
20623 It accepts the following values:
20626 each channel is displayed in a separate color
20628 each channel is displayed using the same color scheme
20630 each channel is displayed using the rainbow color scheme
20632 each channel is displayed using the moreland color scheme
20634 each channel is displayed using the nebulae color scheme
20636 each channel is displayed using the fire color scheme
20638 each channel is displayed using the fiery color scheme
20640 each channel is displayed using the fruit color scheme
20642 each channel is displayed using the cool color scheme
20644 each channel is displayed using the magma color scheme
20646 each channel is displayed using the green color scheme
20649 Default value is @samp{channel}.
20652 Specify scale used for calculating intensity color values.
20654 It accepts the following values:
20659 square root, default
20670 Default value is @samp{sqrt}.
20673 Set saturation modifier for displayed colors. Negative values provide
20674 alternative color scheme. @code{0} is no saturation at all.
20675 Saturation must be in [-10.0, 10.0] range.
20676 Default value is @code{1}.
20679 Set window function.
20681 It accepts the following values:
20705 Default value is @code{hann}.
20708 Set orientation of time vs frequency axis. Can be @code{vertical} or
20709 @code{horizontal}. Default is @code{vertical}.
20712 Set ratio of overlap window. Default value is @code{0}.
20713 When value is @code{1} overlap is set to recommended size for specific
20714 window function currently used.
20717 Set scale gain for calculating intensity color values.
20718 Default value is @code{1}.
20721 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
20724 Set color rotation, must be in [-1.0, 1.0] range.
20725 Default value is @code{0}.
20728 Set start frequency from which to display spectrogram. Default is @code{0}.
20731 Set stop frequency to which to display spectrogram. Default is @code{0}.
20734 Set upper frame rate limit. Default is @code{auto}, unlimited.
20737 Draw time and frequency axes and legends. Default is disabled.
20740 The usage is very similar to the showwaves filter; see the examples in that
20743 @subsection Examples
20747 Large window with logarithmic color scaling:
20749 showspectrum=s=1280x480:scale=log
20753 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
20755 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
20756 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
20760 @section showspectrumpic
20762 Convert input audio to a single video frame, representing the audio frequency
20765 The filter accepts the following options:
20769 Specify the video size for the output. For the syntax of this option, check the
20770 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20771 Default value is @code{4096x2048}.
20774 Specify display mode.
20776 It accepts the following values:
20779 all channels are displayed in the same row
20781 all channels are displayed in separate rows
20783 Default value is @samp{combined}.
20786 Specify display color mode.
20788 It accepts the following values:
20791 each channel is displayed in a separate color
20793 each channel is displayed using the same color scheme
20795 each channel is displayed using the rainbow color scheme
20797 each channel is displayed using the moreland color scheme
20799 each channel is displayed using the nebulae color scheme
20801 each channel is displayed using the fire color scheme
20803 each channel is displayed using the fiery color scheme
20805 each channel is displayed using the fruit color scheme
20807 each channel is displayed using the cool color scheme
20809 each channel is displayed using the magma color scheme
20811 each channel is displayed using the green color scheme
20813 Default value is @samp{intensity}.
20816 Specify scale used for calculating intensity color values.
20818 It accepts the following values:
20823 square root, default
20833 Default value is @samp{log}.
20836 Set saturation modifier for displayed colors. Negative values provide
20837 alternative color scheme. @code{0} is no saturation at all.
20838 Saturation must be in [-10.0, 10.0] range.
20839 Default value is @code{1}.
20842 Set window function.
20844 It accepts the following values:
20867 Default value is @code{hann}.
20870 Set orientation of time vs frequency axis. Can be @code{vertical} or
20871 @code{horizontal}. Default is @code{vertical}.
20874 Set scale gain for calculating intensity color values.
20875 Default value is @code{1}.
20878 Draw time and frequency axes and legends. Default is enabled.
20881 Set color rotation, must be in [-1.0, 1.0] range.
20882 Default value is @code{0}.
20885 Set start frequency from which to display spectrogram. Default is @code{0}.
20888 Set stop frequency to which to display spectrogram. Default is @code{0}.
20891 @subsection Examples
20895 Extract an audio spectrogram of a whole audio track
20896 in a 1024x1024 picture using @command{ffmpeg}:
20898 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
20902 @section showvolume
20904 Convert input audio volume to a video output.
20906 The filter accepts the following options:
20913 Set border width, allowed range is [0, 5]. Default is 1.
20916 Set channel width, allowed range is [80, 8192]. Default is 400.
20919 Set channel height, allowed range is [1, 900]. Default is 20.
20922 Set fade, allowed range is [0, 1]. Default is 0.95.
20925 Set volume color expression.
20927 The expression can use the following variables:
20931 Current max volume of channel in dB.
20937 Current channel number, starting from 0.
20941 If set, displays channel names. Default is enabled.
20944 If set, displays volume values. Default is enabled.
20947 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
20948 default is @code{h}.
20951 Set step size, allowed range is [0, 5]. Default is 0, which means
20955 Set background opacity, allowed range is [0, 1]. Default is 0.
20958 Set metering mode, can be peak: @code{p} or rms: @code{r},
20959 default is @code{p}.
20962 Set display scale, can be linear: @code{lin} or log: @code{log},
20963 default is @code{lin}.
20967 If set to > 0., display a line for the max level
20968 in the previous seconds.
20969 default is disabled: @code{0.}
20972 The color of the max line. Use when @code{dm} option is set to > 0.
20973 default is: @code{orange}
20978 Convert input audio to a video output, representing the samples waves.
20980 The filter accepts the following options:
20984 Specify the video size for the output. For the syntax of this option, check the
20985 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20986 Default value is @code{600x240}.
20991 Available values are:
20994 Draw a point for each sample.
20997 Draw a vertical line for each sample.
21000 Draw a point for each sample and a line between them.
21003 Draw a centered vertical line for each sample.
21006 Default value is @code{point}.
21009 Set the number of samples which are printed on the same column. A
21010 larger value will decrease the frame rate. Must be a positive
21011 integer. This option can be set only if the value for @var{rate}
21012 is not explicitly specified.
21015 Set the (approximate) output frame rate. This is done by setting the
21016 option @var{n}. Default value is "25".
21018 @item split_channels
21019 Set if channels should be drawn separately or overlap. Default value is 0.
21022 Set colors separated by '|' which are going to be used for drawing of each channel.
21025 Set amplitude scale.
21027 Available values are:
21045 Set the draw mode. This is mostly useful to set for high @var{n}.
21047 Available values are:
21050 Scale pixel values for each drawn sample.
21053 Draw every sample directly.
21056 Default value is @code{scale}.
21059 @subsection Examples
21063 Output the input file audio and the corresponding video representation
21066 amovie=a.mp3,asplit[out0],showwaves[out1]
21070 Create a synthetic signal and show it with showwaves, forcing a
21071 frame rate of 30 frames per second:
21073 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
21077 @section showwavespic
21079 Convert input audio to a single video frame, representing the samples waves.
21081 The filter accepts the following options:
21085 Specify the video size for the output. For the syntax of this option, check the
21086 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21087 Default value is @code{600x240}.
21089 @item split_channels
21090 Set if channels should be drawn separately or overlap. Default value is 0.
21093 Set colors separated by '|' which are going to be used for drawing of each channel.
21096 Set amplitude scale.
21098 Available values are:
21116 @subsection Examples
21120 Extract a channel split representation of the wave form of a whole audio track
21121 in a 1024x800 picture using @command{ffmpeg}:
21123 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
21127 @section sidedata, asidedata
21129 Delete frame side data, or select frames based on it.
21131 This filter accepts the following options:
21135 Set mode of operation of the filter.
21137 Can be one of the following:
21141 Select every frame with side data of @code{type}.
21144 Delete side data of @code{type}. If @code{type} is not set, delete all side
21150 Set side data type used with all modes. Must be set for @code{select} mode. For
21151 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
21152 in @file{libavutil/frame.h}. For example, to choose
21153 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
21157 @section spectrumsynth
21159 Sythesize audio from 2 input video spectrums, first input stream represents
21160 magnitude across time and second represents phase across time.
21161 The filter will transform from frequency domain as displayed in videos back
21162 to time domain as presented in audio output.
21164 This filter is primarily created for reversing processed @ref{showspectrum}
21165 filter outputs, but can synthesize sound from other spectrograms too.
21166 But in such case results are going to be poor if the phase data is not
21167 available, because in such cases phase data need to be recreated, usually
21168 its just recreated from random noise.
21169 For best results use gray only output (@code{channel} color mode in
21170 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
21171 @code{lin} scale for phase video. To produce phase, for 2nd video, use
21172 @code{data} option. Inputs videos should generally use @code{fullframe}
21173 slide mode as that saves resources needed for decoding video.
21175 The filter accepts the following options:
21179 Specify sample rate of output audio, the sample rate of audio from which
21180 spectrum was generated may differ.
21183 Set number of channels represented in input video spectrums.
21186 Set scale which was used when generating magnitude input spectrum.
21187 Can be @code{lin} or @code{log}. Default is @code{log}.
21190 Set slide which was used when generating inputs spectrums.
21191 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
21192 Default is @code{fullframe}.
21195 Set window function used for resynthesis.
21198 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
21199 which means optimal overlap for selected window function will be picked.
21202 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
21203 Default is @code{vertical}.
21206 @subsection Examples
21210 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
21211 then resynthesize videos back to audio with spectrumsynth:
21213 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
21214 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
21215 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
21219 @section split, asplit
21221 Split input into several identical outputs.
21223 @code{asplit} works with audio input, @code{split} with video.
21225 The filter accepts a single parameter which specifies the number of outputs. If
21226 unspecified, it defaults to 2.
21228 @subsection Examples
21232 Create two separate outputs from the same input:
21234 [in] split [out0][out1]
21238 To create 3 or more outputs, you need to specify the number of
21241 [in] asplit=3 [out0][out1][out2]
21245 Create two separate outputs from the same input, one cropped and
21248 [in] split [splitout1][splitout2];
21249 [splitout1] crop=100:100:0:0 [cropout];
21250 [splitout2] pad=200:200:100:100 [padout];
21254 Create 5 copies of the input audio with @command{ffmpeg}:
21256 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
21262 Receive commands sent through a libzmq client, and forward them to
21263 filters in the filtergraph.
21265 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
21266 must be inserted between two video filters, @code{azmq} between two
21267 audio filters. Both are capable to send messages to any filter type.
21269 To enable these filters you need to install the libzmq library and
21270 headers and configure FFmpeg with @code{--enable-libzmq}.
21272 For more information about libzmq see:
21273 @url{http://www.zeromq.org/}
21275 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
21276 receives messages sent through a network interface defined by the
21277 @option{bind_address} (or the abbreviation "@option{b}") option.
21278 Default value of this option is @file{tcp://localhost:5555}. You may
21279 want to alter this value to your needs, but do not forget to escape any
21280 ':' signs (see @ref{filtergraph escaping}).
21282 The received message must be in the form:
21284 @var{TARGET} @var{COMMAND} [@var{ARG}]
21287 @var{TARGET} specifies the target of the command, usually the name of
21288 the filter class or a specific filter instance name. The default
21289 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
21290 but you can override this by using the @samp{filter_name@@id} syntax
21291 (see @ref{Filtergraph syntax}).
21293 @var{COMMAND} specifies the name of the command for the target filter.
21295 @var{ARG} is optional and specifies the optional argument list for the
21296 given @var{COMMAND}.
21298 Upon reception, the message is processed and the corresponding command
21299 is injected into the filtergraph. Depending on the result, the filter
21300 will send a reply to the client, adopting the format:
21302 @var{ERROR_CODE} @var{ERROR_REASON}
21306 @var{MESSAGE} is optional.
21308 @subsection Examples
21310 Look at @file{tools/zmqsend} for an example of a zmq client which can
21311 be used to send commands processed by these filters.
21313 Consider the following filtergraph generated by @command{ffplay}.
21314 In this example the last overlay filter has an instance name. All other
21315 filters will have default instance names.
21318 ffplay -dumpgraph 1 -f lavfi "
21319 color=s=100x100:c=red [l];
21320 color=s=100x100:c=blue [r];
21321 nullsrc=s=200x100, zmq [bg];
21322 [bg][l] overlay [bg+l];
21323 [bg+l][r] overlay@@my=x=100 "
21326 To change the color of the left side of the video, the following
21327 command can be used:
21329 echo Parsed_color_0 c yellow | tools/zmqsend
21332 To change the right side:
21334 echo Parsed_color_1 c pink | tools/zmqsend
21337 To change the position of the right side:
21339 echo overlay@@my x 150 | tools/zmqsend
21343 @c man end MULTIMEDIA FILTERS
21345 @chapter Multimedia Sources
21346 @c man begin MULTIMEDIA SOURCES
21348 Below is a description of the currently available multimedia sources.
21352 This is the same as @ref{movie} source, except it selects an audio
21358 Read audio and/or video stream(s) from a movie container.
21360 It accepts the following parameters:
21364 The name of the resource to read (not necessarily a file; it can also be a
21365 device or a stream accessed through some protocol).
21367 @item format_name, f
21368 Specifies the format assumed for the movie to read, and can be either
21369 the name of a container or an input device. If not specified, the
21370 format is guessed from @var{movie_name} or by probing.
21372 @item seek_point, sp
21373 Specifies the seek point in seconds. The frames will be output
21374 starting from this seek point. The parameter is evaluated with
21375 @code{av_strtod}, so the numerical value may be suffixed by an IS
21376 postfix. The default value is "0".
21379 Specifies the streams to read. Several streams can be specified,
21380 separated by "+". The source will then have as many outputs, in the
21381 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
21382 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
21383 respectively the default (best suited) video and audio stream. Default
21384 is "dv", or "da" if the filter is called as "amovie".
21386 @item stream_index, si
21387 Specifies the index of the video stream to read. If the value is -1,
21388 the most suitable video stream will be automatically selected. The default
21389 value is "-1". Deprecated. If the filter is called "amovie", it will select
21390 audio instead of video.
21393 Specifies how many times to read the stream in sequence.
21394 If the value is 0, the stream will be looped infinitely.
21395 Default value is "1".
21397 Note that when the movie is looped the source timestamps are not
21398 changed, so it will generate non monotonically increasing timestamps.
21400 @item discontinuity
21401 Specifies the time difference between frames above which the point is
21402 considered a timestamp discontinuity which is removed by adjusting the later
21406 It allows overlaying a second video on top of the main input of
21407 a filtergraph, as shown in this graph:
21409 input -----------> deltapts0 --> overlay --> output
21412 movie --> scale--> deltapts1 -------+
21414 @subsection Examples
21418 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
21419 on top of the input labelled "in":
21421 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
21422 [in] setpts=PTS-STARTPTS [main];
21423 [main][over] overlay=16:16 [out]
21427 Read from a video4linux2 device, and overlay it on top of the input
21430 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
21431 [in] setpts=PTS-STARTPTS [main];
21432 [main][over] overlay=16:16 [out]
21436 Read the first video stream and the audio stream with id 0x81 from
21437 dvd.vob; the video is connected to the pad named "video" and the audio is
21438 connected to the pad named "audio":
21440 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
21444 @subsection Commands
21446 Both movie and amovie support the following commands:
21449 Perform seek using "av_seek_frame".
21450 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
21453 @var{stream_index}: If stream_index is -1, a default
21454 stream is selected, and @var{timestamp} is automatically converted
21455 from AV_TIME_BASE units to the stream specific time_base.
21457 @var{timestamp}: Timestamp in AVStream.time_base units
21458 or, if no stream is specified, in AV_TIME_BASE units.
21460 @var{flags}: Flags which select direction and seeking mode.
21464 Get movie duration in AV_TIME_BASE units.
21468 @c man end MULTIMEDIA SOURCES