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 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
395 Default is @code{downward}.
398 If a signal of stream rises above this level it will affect the gain
400 By default it is 0.125. Range is between 0.00097563 and 1.
403 Set a ratio by which the signal is reduced. 1:2 means that if the level
404 rose 4dB above the threshold, it will be only 2dB above after the reduction.
405 Default is 2. Range is between 1 and 20.
408 Amount of milliseconds the signal has to rise above the threshold before gain
409 reduction starts. Default is 20. Range is between 0.01 and 2000.
412 Amount of milliseconds the signal has to fall below the threshold before
413 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
416 Set the amount by how much signal will be amplified after processing.
417 Default is 1. Range is from 1 to 64.
420 Curve the sharp knee around the threshold to enter gain reduction more softly.
421 Default is 2.82843. Range is between 1 and 8.
424 Choose if the @code{average} level between all channels of input stream
425 or the louder(@code{maximum}) channel of input stream affects the
426 reduction. Default is @code{average}.
429 Should the exact signal be taken in case of @code{peak} or an RMS one in case
430 of @code{rms}. Default is @code{rms} which is mostly smoother.
433 How much to use compressed signal in output. Default is 1.
434 Range is between 0 and 1.
438 Simple audio dynamic range compression/expansion filter.
440 The filter accepts the following options:
444 Set contrast. Default is 33. Allowed range is between 0 and 100.
449 Copy the input audio source unchanged to the output. This is mainly useful for
454 Apply cross fade from one input audio stream to another input audio stream.
455 The cross fade is applied for specified duration near the end of first stream.
457 The filter accepts the following options:
461 Specify the number of samples for which the cross fade effect has to last.
462 At the end of the cross fade effect the first input audio will be completely
463 silent. Default is 44100.
466 Specify the duration of the cross fade effect. See
467 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
468 for the accepted syntax.
469 By default the duration is determined by @var{nb_samples}.
470 If set this option is used instead of @var{nb_samples}.
473 Should first stream end overlap with second stream start. Default is enabled.
476 Set curve for cross fade transition for first stream.
479 Set curve for cross fade transition for second stream.
481 For description of available curve types see @ref{afade} filter description.
488 Cross fade from one input to another:
490 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
494 Cross fade from one input to another but without overlapping:
496 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
501 Split audio stream into several bands.
503 This filter splits audio stream into two or more frequency ranges.
504 Summing all streams back will give flat output.
506 The filter accepts the following options:
510 Set split frequencies. Those must be positive and increasing.
513 Set filter order, can be @var{2nd}, @var{4th} or @var{8th}.
514 Default is @var{4th}.
519 Reduce audio bit resolution.
521 This filter is bit crusher with enhanced functionality. A bit crusher
522 is used to audibly reduce number of bits an audio signal is sampled
523 with. This doesn't change the bit depth at all, it just produces the
524 effect. Material reduced in bit depth sounds more harsh and "digital".
525 This filter is able to even round to continuous values instead of discrete
527 Additionally it has a D/C offset which results in different crushing of
528 the lower and the upper half of the signal.
529 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
531 Another feature of this filter is the logarithmic mode.
532 This setting switches from linear distances between bits to logarithmic ones.
533 The result is a much more "natural" sounding crusher which doesn't gate low
534 signals for example. The human ear has a logarithmic perception,
535 so this kind of crushing is much more pleasant.
536 Logarithmic crushing is also able to get anti-aliased.
538 The filter accepts the following options:
554 Can be linear: @code{lin} or logarithmic: @code{log}.
563 Set sample reduction.
566 Enable LFO. By default disabled.
577 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
581 Remove impulsive noise from input audio.
583 Samples detected as impulsive noise are replaced by interpolated samples using
584 autoregressive modelling.
588 Set window size, in milliseconds. Allowed range is from @code{10} to
589 @code{100}. Default value is @code{55} milliseconds.
590 This sets size of window which will be processed at once.
593 Set window overlap, in percentage of window size. Allowed range is from
594 @code{50} to @code{95}. Default value is @code{75} percent.
595 Setting this to a very high value increases impulsive noise removal but makes
596 whole process much slower.
599 Set autoregression order, in percentage of window size. Allowed range is from
600 @code{0} to @code{25}. Default value is @code{2} percent. This option also
601 controls quality of interpolated samples using neighbour good samples.
604 Set threshold value. Allowed range is from @code{1} to @code{100}.
605 Default value is @code{2}.
606 This controls the strength of impulsive noise which is going to be removed.
607 The lower value, the more samples will be detected as impulsive noise.
610 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
611 @code{10}. Default value is @code{2}.
612 If any two samples detected as noise are spaced less than this value then any
613 sample between those two samples will be also detected as noise.
618 It accepts the following values:
621 Select overlap-add method. Even not interpolated samples are slightly
622 changed with this method.
625 Select overlap-save method. Not interpolated samples remain unchanged.
628 Default value is @code{a}.
632 Remove clipped samples from input audio.
634 Samples detected as clipped are replaced by interpolated samples using
635 autoregressive modelling.
639 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
640 Default value is @code{55} milliseconds.
641 This sets size of window which will be processed at once.
644 Set window overlap, in percentage of window size. Allowed range is from @code{50}
645 to @code{95}. Default value is @code{75} percent.
648 Set autoregression order, in percentage of window size. Allowed range is from
649 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
650 quality of interpolated samples using neighbour good samples.
653 Set threshold value. Allowed range is from @code{1} to @code{100}.
654 Default value is @code{10}. Higher values make clip detection less aggressive.
657 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
658 Default value is @code{1000}. Higher values make clip detection less aggressive.
663 It accepts the following values:
666 Select overlap-add method. Even not interpolated samples are slightly changed
670 Select overlap-save method. Not interpolated samples remain unchanged.
673 Default value is @code{a}.
678 Delay one or more audio channels.
680 Samples in delayed channel are filled with silence.
682 The filter accepts the following option:
686 Set list of delays in milliseconds for each channel separated by '|'.
687 Unused delays will be silently ignored. If number of given delays is
688 smaller than number of channels all remaining channels will not be delayed.
689 If you want to delay exact number of samples, append 'S' to number.
690 If you want instead to delay in seconds, append 's' to number.
697 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
698 the second channel (and any other channels that may be present) unchanged.
704 Delay second channel by 500 samples, the third channel by 700 samples and leave
705 the first channel (and any other channels that may be present) unchanged.
711 @section aderivative, aintegral
713 Compute derivative/integral of audio stream.
715 Applying both filters one after another produces original audio.
719 Apply echoing to the input audio.
721 Echoes are reflected sound and can occur naturally amongst mountains
722 (and sometimes large buildings) when talking or shouting; digital echo
723 effects emulate this behaviour and are often used to help fill out the
724 sound of a single instrument or vocal. The time difference between the
725 original signal and the reflection is the @code{delay}, and the
726 loudness of the reflected signal is the @code{decay}.
727 Multiple echoes can have different delays and decays.
729 A description of the accepted parameters follows.
733 Set input gain of reflected signal. Default is @code{0.6}.
736 Set output gain of reflected signal. Default is @code{0.3}.
739 Set list of time intervals in milliseconds between original signal and reflections
740 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
741 Default is @code{1000}.
744 Set list of loudness of reflected signals separated by '|'.
745 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
746 Default is @code{0.5}.
753 Make it sound as if there are twice as many instruments as are actually playing:
755 aecho=0.8:0.88:60:0.4
759 If delay is very short, then it sound like a (metallic) robot playing music:
765 A longer delay will sound like an open air concert in the mountains:
767 aecho=0.8:0.9:1000:0.3
771 Same as above but with one more mountain:
773 aecho=0.8:0.9:1000|1800:0.3|0.25
778 Audio emphasis filter creates or restores material directly taken from LPs or
779 emphased CDs with different filter curves. E.g. to store music on vinyl the
780 signal has to be altered by a filter first to even out the disadvantages of
781 this recording medium.
782 Once the material is played back the inverse filter has to be applied to
783 restore the distortion of the frequency response.
785 The filter accepts the following options:
795 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
796 use @code{production} mode. Default is @code{reproduction} mode.
799 Set filter type. Selects medium. Can be one of the following:
811 select Compact Disc (CD).
817 select 50µs (FM-KF).
819 select 75µs (FM-KF).
825 Modify an audio signal according to the specified expressions.
827 This filter accepts one or more expressions (one for each channel),
828 which are evaluated and used to modify a corresponding audio signal.
830 It accepts the following parameters:
834 Set the '|'-separated expressions list for each separate channel. If
835 the number of input channels is greater than the number of
836 expressions, the last specified expression is used for the remaining
839 @item channel_layout, c
840 Set output channel layout. If not specified, the channel layout is
841 specified by the number of expressions. If set to @samp{same}, it will
842 use by default the same input channel layout.
845 Each expression in @var{exprs} can contain the following constants and functions:
849 channel number of the current expression
852 number of the evaluated sample, starting from 0
858 time of the evaluated sample expressed in seconds
861 @item nb_out_channels
862 input and output number of channels
865 the value of input channel with number @var{CH}
868 Note: this filter is slow. For faster processing you should use a
877 aeval=val(ch)/2:c=same
881 Invert phase of the second channel:
890 Apply fade-in/out effect to input audio.
892 A description of the accepted parameters follows.
896 Specify the effect type, can be either @code{in} for fade-in, or
897 @code{out} for a fade-out effect. Default is @code{in}.
899 @item start_sample, ss
900 Specify the number of the start sample for starting to apply the fade
901 effect. Default is 0.
904 Specify the number of samples for which the fade effect has to last. At
905 the end of the fade-in effect the output audio will have the same
906 volume as the input audio, at the end of the fade-out transition
907 the output audio will be silence. Default is 44100.
910 Specify the start time of the fade effect. Default is 0.
911 The value must be specified as a time duration; see
912 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
913 for the accepted syntax.
914 If set this option is used instead of @var{start_sample}.
917 Specify the duration of the fade effect. See
918 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
919 for the accepted syntax.
920 At the end of the fade-in effect the output audio will have the same
921 volume as the input audio, at the end of the fade-out transition
922 the output audio will be silence.
923 By default the duration is determined by @var{nb_samples}.
924 If set this option is used instead of @var{nb_samples}.
927 Set curve for fade transition.
929 It accepts the following values:
932 select triangular, linear slope (default)
934 select quarter of sine wave
936 select half of sine wave
938 select exponential sine wave
942 select inverted parabola
956 select inverted quarter of sine wave
958 select inverted half of sine wave
960 select double-exponential seat
962 select double-exponential sigmoid
964 select logistic sigmoid
974 Fade in first 15 seconds of audio:
980 Fade out last 25 seconds of a 900 seconds audio:
982 afade=t=out:st=875:d=25
987 Denoise audio samples with FFT.
989 A description of the accepted parameters follows.
993 Set the noise reduction in dB, allowed range is 0.01 to 97.
994 Default value is 12 dB.
997 Set the noise floor in dB, allowed range is -80 to -20.
998 Default value is -50 dB.
1003 It accepts the following values:
1012 Select shellac noise.
1015 Select custom noise, defined in @code{bn} option.
1017 Default value is white noise.
1021 Set custom band noise for every one of 15 bands.
1022 Bands are separated by ' ' or '|'.
1025 Set the residual floor in dB, allowed range is -80 to -20.
1026 Default value is -38 dB.
1029 Enable noise tracking. By default is disabled.
1030 With this enabled, noise floor is automatically adjusted.
1033 Enable residual tracking. By default is disabled.
1036 Set the output mode.
1038 It accepts the following values:
1041 Pass input unchanged.
1044 Pass noise filtered out.
1049 Default value is @var{o}.
1053 @subsection Commands
1055 This filter supports the following commands:
1057 @item sample_noise, sn
1058 Start or stop measuring noise profile.
1059 Syntax for the command is : "start" or "stop" string.
1060 After measuring noise profile is stopped it will be
1061 automatically applied in filtering.
1063 @item noise_reduction, nr
1064 Change noise reduction. Argument is single float number.
1065 Syntax for the command is : "@var{noise_reduction}"
1067 @item noise_floor, nf
1068 Change noise floor. Argument is single float number.
1069 Syntax for the command is : "@var{noise_floor}"
1071 @item output_mode, om
1072 Change output mode operation.
1073 Syntax for the command is : "i", "o" or "n" string.
1077 Apply arbitrary expressions to samples in frequency domain.
1081 Set frequency domain real expression for each separate channel separated
1082 by '|'. Default is "re".
1083 If the number of input channels is greater than the number of
1084 expressions, the last specified expression is used for the remaining
1088 Set frequency domain imaginary expression for each separate channel
1089 separated by '|'. Default is "im".
1091 Each expression in @var{real} and @var{imag} can contain the following
1092 constants and functions:
1099 current frequency bin number
1102 number of available bins
1105 channel number of the current expression
1114 current real part of frequency bin of current channel
1117 current imaginary part of frequency bin of current channel
1120 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1123 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1127 Set window size. Allowed range is from 16 to 131072.
1128 Default is @code{4096}
1131 Set window function. Default is @code{hann}.
1134 Set window overlap. If set to 1, the recommended overlap for selected
1135 window function will be picked. Default is @code{0.75}.
1138 @subsection Examples
1142 Leave almost only low frequencies in audio:
1144 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1151 Apply an arbitrary Frequency Impulse Response filter.
1153 This filter is designed for applying long FIR filters,
1154 up to 60 seconds long.
1156 It can be used as component for digital crossover filters,
1157 room equalization, cross talk cancellation, wavefield synthesis,
1158 auralization, ambiophonics, ambisonics and spatialization.
1160 This filter uses second stream as FIR coefficients.
1161 If second stream holds single channel, it will be used
1162 for all input channels in first stream, otherwise
1163 number of channels in second stream must be same as
1164 number of channels in first stream.
1166 It accepts the following parameters:
1170 Set dry gain. This sets input gain.
1173 Set wet gain. This sets final output gain.
1176 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1179 Enable applying gain measured from power of IR.
1181 Set which approach to use for auto gain measurement.
1185 Do not apply any gain.
1188 select peak gain, very conservative approach. This is default value.
1191 select DC gain, limited application.
1194 select gain to noise approach, this is most popular one.
1198 Set gain to be applied to IR coefficients before filtering.
1199 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1202 Set format of IR stream. Can be @code{mono} or @code{input}.
1203 Default is @code{input}.
1206 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1207 Allowed range is 0.1 to 60 seconds.
1210 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1211 By default it is disabled.
1214 Set for which IR channel to display frequency response. By default is first channel
1215 displayed. This option is used only when @var{response} is enabled.
1218 Set video stream size. This option is used only when @var{response} is enabled.
1221 Set video stream frame rate. This option is used only when @var{response} is enabled.
1224 Set minimal partition size used for convolution. Default is @var{8192}.
1225 Allowed range is from @var{8} to @var{32768}.
1226 Lower values decreases latency at cost of higher CPU usage.
1229 Set maximal partition size used for convolution. Default is @var{8192}.
1230 Allowed range is from @var{8} to @var{32768}.
1231 Lower values may increase CPU usage.
1234 @subsection Examples
1238 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1240 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1247 Set output format constraints for the input audio. The framework will
1248 negotiate the most appropriate format to minimize conversions.
1250 It accepts the following parameters:
1254 A '|'-separated list of requested sample formats.
1257 A '|'-separated list of requested sample rates.
1259 @item channel_layouts
1260 A '|'-separated list of requested channel layouts.
1262 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1263 for the required syntax.
1266 If a parameter is omitted, all values are allowed.
1268 Force the output to either unsigned 8-bit or signed 16-bit stereo
1270 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1275 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1276 processing reduces disturbing noise between useful signals.
1278 Gating is done by detecting the volume below a chosen level @var{threshold}
1279 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1280 floor is set via @var{range}. Because an exact manipulation of the signal
1281 would cause distortion of the waveform the reduction can be levelled over
1282 time. This is done by setting @var{attack} and @var{release}.
1284 @var{attack} determines how long the signal has to fall below the threshold
1285 before any reduction will occur and @var{release} sets the time the signal
1286 has to rise above the threshold to reduce the reduction again.
1287 Shorter signals than the chosen attack time will be left untouched.
1291 Set input level before filtering.
1292 Default is 1. Allowed range is from 0.015625 to 64.
1295 Set the mode of operation. Can be @code{upward} or @code{downward}.
1296 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1297 will be amplified, expanding dynamic range in upward direction.
1298 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1301 Set the level of gain reduction when the signal is below the threshold.
1302 Default is 0.06125. Allowed range is from 0 to 1.
1303 Setting this to 0 disables reduction and then filter behaves like expander.
1306 If a signal rises above this level the gain reduction is released.
1307 Default is 0.125. Allowed range is from 0 to 1.
1310 Set a ratio by which the signal is reduced.
1311 Default is 2. Allowed range is from 1 to 9000.
1314 Amount of milliseconds the signal has to rise above the threshold before gain
1316 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1319 Amount of milliseconds the signal has to fall below the threshold before the
1320 reduction is increased again. Default is 250 milliseconds.
1321 Allowed range is from 0.01 to 9000.
1324 Set amount of amplification of signal after processing.
1325 Default is 1. Allowed range is from 1 to 64.
1328 Curve the sharp knee around the threshold to enter gain reduction more softly.
1329 Default is 2.828427125. Allowed range is from 1 to 8.
1332 Choose if exact signal should be taken for detection or an RMS like one.
1333 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1336 Choose if the average level between all channels or the louder channel affects
1338 Default is @code{average}. Can be @code{average} or @code{maximum}.
1343 Apply an arbitrary Infinite Impulse Response filter.
1345 It accepts the following parameters:
1349 Set numerator/zeros coefficients.
1352 Set denominator/poles coefficients.
1364 Set coefficients format.
1370 Z-plane zeros/poles, cartesian (default)
1372 Z-plane zeros/poles, polar radians
1374 Z-plane zeros/poles, polar degrees
1378 Set kind of processing.
1379 Can be @code{d} - direct or @code{s} - serial cascading. Default is @code{s}.
1382 Set filtering precision.
1386 double-precision floating-point (default)
1388 single-precision floating-point
1396 How much to use filtered signal in output. Default is 1.
1397 Range is between 0 and 1.
1400 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1401 By default it is disabled.
1404 Set for which IR channel to display frequency response. By default is first channel
1405 displayed. This option is used only when @var{response} is enabled.
1408 Set video stream size. This option is used only when @var{response} is enabled.
1411 Coefficients in @code{tf} format are separated by spaces and are in ascending
1414 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1415 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1418 Different coefficients and gains can be provided for every channel, in such case
1419 use '|' to separate coefficients or gains. Last provided coefficients will be
1420 used for all remaining channels.
1422 @subsection Examples
1426 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1428 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
1432 Same as above but in @code{zp} format:
1434 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
1440 The limiter prevents an input signal from rising over a desired threshold.
1441 This limiter uses lookahead technology to prevent your signal from distorting.
1442 It means that there is a small delay after the signal is processed. Keep in mind
1443 that the delay it produces is the attack time you set.
1445 The filter accepts the following options:
1449 Set input gain. Default is 1.
1452 Set output gain. Default is 1.
1455 Don't let signals above this level pass the limiter. Default is 1.
1458 The limiter will reach its attenuation level in this amount of time in
1459 milliseconds. Default is 5 milliseconds.
1462 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1463 Default is 50 milliseconds.
1466 When gain reduction is always needed ASC takes care of releasing to an
1467 average reduction level rather than reaching a reduction of 0 in the release
1471 Select how much the release time is affected by ASC, 0 means nearly no changes
1472 in release time while 1 produces higher release times.
1475 Auto level output signal. Default is enabled.
1476 This normalizes audio back to 0dB if enabled.
1479 Depending on picked setting it is recommended to upsample input 2x or 4x times
1480 with @ref{aresample} before applying this filter.
1484 Apply a two-pole all-pass filter with central frequency (in Hz)
1485 @var{frequency}, and filter-width @var{width}.
1486 An all-pass filter changes the audio's frequency to phase relationship
1487 without changing its frequency to amplitude relationship.
1489 The filter accepts the following options:
1493 Set frequency in Hz.
1496 Set method to specify band-width of filter.
1511 Specify the band-width of a filter in width_type units.
1514 How much to use filtered signal in output. Default is 1.
1515 Range is between 0 and 1.
1518 Specify which channels to filter, by default all available are filtered.
1521 @subsection Commands
1523 This filter supports the following commands:
1526 Change allpass frequency.
1527 Syntax for the command is : "@var{frequency}"
1530 Change allpass width_type.
1531 Syntax for the command is : "@var{width_type}"
1534 Change allpass width.
1535 Syntax for the command is : "@var{width}"
1539 Syntax for the command is : "@var{mix}"
1546 The filter accepts the following options:
1550 Set the number of loops. Setting this value to -1 will result in infinite loops.
1554 Set maximal number of samples. Default is 0.
1557 Set first sample of loop. Default is 0.
1563 Merge two or more audio streams into a single multi-channel stream.
1565 The filter accepts the following options:
1570 Set the number of inputs. Default is 2.
1574 If the channel layouts of the inputs are disjoint, and therefore compatible,
1575 the channel layout of the output will be set accordingly and the channels
1576 will be reordered as necessary. If the channel layouts of the inputs are not
1577 disjoint, the output will have all the channels of the first input then all
1578 the channels of the second input, in that order, and the channel layout of
1579 the output will be the default value corresponding to the total number of
1582 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1583 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1584 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1585 first input, b1 is the first channel of the second input).
1587 On the other hand, if both input are in stereo, the output channels will be
1588 in the default order: a1, a2, b1, b2, and the channel layout will be
1589 arbitrarily set to 4.0, which may or may not be the expected value.
1591 All inputs must have the same sample rate, and format.
1593 If inputs do not have the same duration, the output will stop with the
1596 @subsection Examples
1600 Merge two mono files into a stereo stream:
1602 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1606 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1608 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
1614 Mixes multiple audio inputs into a single output.
1616 Note that this filter only supports float samples (the @var{amerge}
1617 and @var{pan} audio filters support many formats). If the @var{amix}
1618 input has integer samples then @ref{aresample} will be automatically
1619 inserted to perform the conversion to float samples.
1623 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1625 will mix 3 input audio streams to a single output with the same duration as the
1626 first input and a dropout transition time of 3 seconds.
1628 It accepts the following parameters:
1632 The number of inputs. If unspecified, it defaults to 2.
1635 How to determine the end-of-stream.
1639 The duration of the longest input. (default)
1642 The duration of the shortest input.
1645 The duration of the first input.
1649 @item dropout_transition
1650 The transition time, in seconds, for volume renormalization when an input
1651 stream ends. The default value is 2 seconds.
1654 Specify weight of each input audio stream as sequence.
1655 Each weight is separated by space. By default all inputs have same weight.
1660 Multiply first audio stream with second audio stream and store result
1661 in output audio stream. Multiplication is done by multiplying each
1662 sample from first stream with sample at same position from second stream.
1664 With this element-wise multiplication one can create amplitude fades and
1665 amplitude modulations.
1667 @section anequalizer
1669 High-order parametric multiband equalizer for each channel.
1671 It accepts the following parameters:
1675 This option string is in format:
1676 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1677 Each equalizer band is separated by '|'.
1681 Set channel number to which equalization will be applied.
1682 If input doesn't have that channel the entry is ignored.
1685 Set central frequency for band.
1686 If input doesn't have that frequency the entry is ignored.
1689 Set band width in hertz.
1692 Set band gain in dB.
1695 Set filter type for band, optional, can be:
1699 Butterworth, this is default.
1710 With this option activated frequency response of anequalizer is displayed
1714 Set video stream size. Only useful if curves option is activated.
1717 Set max gain that will be displayed. Only useful if curves option is activated.
1718 Setting this to a reasonable value makes it possible to display gain which is derived from
1719 neighbour bands which are too close to each other and thus produce higher gain
1720 when both are activated.
1723 Set frequency scale used to draw frequency response in video output.
1724 Can be linear or logarithmic. Default is logarithmic.
1727 Set color for each channel curve which is going to be displayed in video stream.
1728 This is list of color names separated by space or by '|'.
1729 Unrecognised or missing colors will be replaced by white color.
1732 @subsection Examples
1736 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1737 for first 2 channels using Chebyshev type 1 filter:
1739 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1743 @subsection Commands
1745 This filter supports the following commands:
1748 Alter existing filter parameters.
1749 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1751 @var{fN} is existing filter number, starting from 0, if no such filter is available
1753 @var{freq} set new frequency parameter.
1754 @var{width} set new width parameter in herz.
1755 @var{gain} set new gain parameter in dB.
1757 Full filter invocation with asendcmd may look like this:
1758 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1763 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1765 Each sample is adjusted by looking for other samples with similar contexts. This
1766 context similarity is defined by comparing their surrounding patches of size
1767 @option{p}. Patches are searched in an area of @option{r} around the sample.
1769 The filter accepts the following options.
1773 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
1776 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
1777 Default value is 2 milliseconds.
1780 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
1781 Default value is 6 milliseconds.
1784 Set the output mode.
1786 It accepts the following values:
1789 Pass input unchanged.
1792 Pass noise filtered out.
1797 Default value is @var{o}.
1801 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
1804 @subsection Commands
1806 This filter supports the following commands:
1809 Change denoise strength. Argument is single float number.
1810 Syntax for the command is : "@var{s}"
1814 Syntax for the command is : "i", "o" or "n" string.
1819 Pass the audio source unchanged to the output.
1823 Pad the end of an audio stream with silence.
1825 This can be used together with @command{ffmpeg} @option{-shortest} to
1826 extend audio streams to the same length as the video stream.
1828 A description of the accepted options follows.
1832 Set silence packet size. Default value is 4096.
1835 Set the number of samples of silence to add to the end. After the
1836 value is reached, the stream is terminated. This option is mutually
1837 exclusive with @option{whole_len}.
1840 Set the minimum total number of samples in the output audio stream. If
1841 the value is longer than the input audio length, silence is added to
1842 the end, until the value is reached. This option is mutually exclusive
1843 with @option{pad_len}.
1846 Specify the duration of samples of silence to add. See
1847 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1848 for the accepted syntax. Used only if set to non-zero value.
1851 Specify the minimum total duration in the output audio stream. See
1852 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1853 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
1854 the input audio length, silence is added to the end, until the value is reached.
1855 This option is mutually exclusive with @option{pad_dur}
1858 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
1859 nor @option{whole_dur} option is set, the filter will add silence to the end of
1860 the input stream indefinitely.
1862 @subsection Examples
1866 Add 1024 samples of silence to the end of the input:
1872 Make sure the audio output will contain at least 10000 samples, pad
1873 the input with silence if required:
1875 apad=whole_len=10000
1879 Use @command{ffmpeg} to pad the audio input with silence, so that the
1880 video stream will always result the shortest and will be converted
1881 until the end in the output file when using the @option{shortest}
1884 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
1889 Add a phasing effect to the input audio.
1891 A phaser filter creates series of peaks and troughs in the frequency spectrum.
1892 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
1894 A description of the accepted parameters follows.
1898 Set input gain. Default is 0.4.
1901 Set output gain. Default is 0.74
1904 Set delay in milliseconds. Default is 3.0.
1907 Set decay. Default is 0.4.
1910 Set modulation speed in Hz. Default is 0.5.
1913 Set modulation type. Default is triangular.
1915 It accepts the following values:
1924 Audio pulsator is something between an autopanner and a tremolo.
1925 But it can produce funny stereo effects as well. Pulsator changes the volume
1926 of the left and right channel based on a LFO (low frequency oscillator) with
1927 different waveforms and shifted phases.
1928 This filter have the ability to define an offset between left and right
1929 channel. An offset of 0 means that both LFO shapes match each other.
1930 The left and right channel are altered equally - a conventional tremolo.
1931 An offset of 50% means that the shape of the right channel is exactly shifted
1932 in phase (or moved backwards about half of the frequency) - pulsator acts as
1933 an autopanner. At 1 both curves match again. Every setting in between moves the
1934 phase shift gapless between all stages and produces some "bypassing" sounds with
1935 sine and triangle waveforms. The more you set the offset near 1 (starting from
1936 the 0.5) the faster the signal passes from the left to the right speaker.
1938 The filter accepts the following options:
1942 Set input gain. By default it is 1. Range is [0.015625 - 64].
1945 Set output gain. By default it is 1. Range is [0.015625 - 64].
1948 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
1949 sawup or sawdown. Default is sine.
1952 Set modulation. Define how much of original signal is affected by the LFO.
1955 Set left channel offset. Default is 0. Allowed range is [0 - 1].
1958 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
1961 Set pulse width. Default is 1. Allowed range is [0 - 2].
1964 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
1967 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
1971 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
1975 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
1976 if timing is set to hz.
1982 Resample the input audio to the specified parameters, using the
1983 libswresample library. If none are specified then the filter will
1984 automatically convert between its input and output.
1986 This filter is also able to stretch/squeeze the audio data to make it match
1987 the timestamps or to inject silence / cut out audio to make it match the
1988 timestamps, do a combination of both or do neither.
1990 The filter accepts the syntax
1991 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
1992 expresses a sample rate and @var{resampler_options} is a list of
1993 @var{key}=@var{value} pairs, separated by ":". See the
1994 @ref{Resampler Options,,"Resampler Options" section in the
1995 ffmpeg-resampler(1) manual,ffmpeg-resampler}
1996 for the complete list of supported options.
1998 @subsection Examples
2002 Resample the input audio to 44100Hz:
2008 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2009 samples per second compensation:
2011 aresample=async=1000
2017 Reverse an audio clip.
2019 Warning: This filter requires memory to buffer the entire clip, so trimming
2022 @subsection Examples
2026 Take the first 5 seconds of a clip, and reverse it.
2028 atrim=end=5,areverse
2032 @section asetnsamples
2034 Set the number of samples per each output audio frame.
2036 The last output packet may contain a different number of samples, as
2037 the filter will flush all the remaining samples when the input audio
2040 The filter accepts the following options:
2044 @item nb_out_samples, n
2045 Set the number of frames per each output audio frame. The number is
2046 intended as the number of samples @emph{per each channel}.
2047 Default value is 1024.
2050 If set to 1, the filter will pad the last audio frame with zeroes, so
2051 that the last frame will contain the same number of samples as the
2052 previous ones. Default value is 1.
2055 For example, to set the number of per-frame samples to 1234 and
2056 disable padding for the last frame, use:
2058 asetnsamples=n=1234:p=0
2063 Set the sample rate without altering the PCM data.
2064 This will result in a change of speed and pitch.
2066 The filter accepts the following options:
2069 @item sample_rate, r
2070 Set the output sample rate. Default is 44100 Hz.
2075 Show a line containing various information for each input audio frame.
2076 The input audio is not modified.
2078 The shown line contains a sequence of key/value pairs of the form
2079 @var{key}:@var{value}.
2081 The following values are shown in the output:
2085 The (sequential) number of the input frame, starting from 0.
2088 The presentation timestamp of the input frame, in time base units; the time base
2089 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2092 The presentation timestamp of the input frame in seconds.
2095 position of the frame in the input stream, -1 if this information in
2096 unavailable and/or meaningless (for example in case of synthetic audio)
2105 The sample rate for the audio frame.
2108 The number of samples (per channel) in the frame.
2111 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2112 audio, the data is treated as if all the planes were concatenated.
2114 @item plane_checksums
2115 A list of Adler-32 checksums for each data plane.
2119 Apply audio soft clipping.
2121 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2122 along a smooth curve, rather than the abrupt shape of hard-clipping.
2124 This filter accepts the following options:
2128 Set type of soft-clipping.
2130 It accepts the following values:
2142 Set additional parameter which controls sigmoid function.
2146 Automatic Speech Recognition
2148 This filter uses PocketSphinx for speech recognition. To enable
2149 compilation of this filter, you need to configure FFmpeg with
2150 @code{--enable-pocketsphinx}.
2152 It accepts the following options:
2156 Set sampling rate of input audio. Defaults is @code{16000}.
2157 This need to match speech models, otherwise one will get poor results.
2160 Set dictionary containing acoustic model files.
2163 Set pronunciation dictionary.
2166 Set language model file.
2169 Set language model set.
2172 Set which language model to use.
2175 Set output for log messages.
2178 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2183 Display time domain statistical information about the audio channels.
2184 Statistics are calculated and displayed for each audio channel and,
2185 where applicable, an overall figure is also given.
2187 It accepts the following option:
2190 Short window length in seconds, used for peak and trough RMS measurement.
2191 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2195 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2196 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2199 Available keys for each channel are:
2241 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2242 this @code{lavfi.astats.Overall.Peak_count}.
2244 For description what each key means read below.
2247 Set number of frame after which stats are going to be recalculated.
2248 Default is disabled.
2250 @item measure_perchannel
2251 Select the entries which need to be measured per channel. The metadata keys can
2252 be used as flags, default is @option{all} which measures everything.
2253 @option{none} disables all per channel measurement.
2255 @item measure_overall
2256 Select the entries which need to be measured overall. The metadata keys can
2257 be used as flags, default is @option{all} which measures everything.
2258 @option{none} disables all overall measurement.
2262 A description of each shown parameter follows:
2266 Mean amplitude displacement from zero.
2269 Minimal sample level.
2272 Maximal sample level.
2274 @item Min difference
2275 Minimal difference between two consecutive samples.
2277 @item Max difference
2278 Maximal difference between two consecutive samples.
2280 @item Mean difference
2281 Mean difference between two consecutive samples.
2282 The average of each difference between two consecutive samples.
2284 @item RMS difference
2285 Root Mean Square difference between two consecutive samples.
2289 Standard peak and RMS level measured in dBFS.
2293 Peak and trough values for RMS level measured over a short window.
2296 Standard ratio of peak to RMS level (note: not in dB).
2299 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2300 (i.e. either @var{Min level} or @var{Max level}).
2303 Number of occasions (not the number of samples) that the signal attained either
2304 @var{Min level} or @var{Max level}.
2307 Overall bit depth of audio. Number of bits used for each sample.
2310 Measured dynamic range of audio in dB.
2312 @item Zero crossings
2313 Number of points where the waveform crosses the zero level axis.
2315 @item Zero crossings rate
2316 Rate of Zero crossings and number of audio samples.
2323 The filter accepts exactly one parameter, the audio tempo. If not
2324 specified then the filter will assume nominal 1.0 tempo. Tempo must
2325 be in the [0.5, 100.0] range.
2327 Note that tempo greater than 2 will skip some samples rather than
2328 blend them in. If for any reason this is a concern it is always
2329 possible to daisy-chain several instances of atempo to achieve the
2330 desired product tempo.
2332 @subsection Examples
2336 Slow down audio to 80% tempo:
2342 To speed up audio to 300% tempo:
2348 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2350 atempo=sqrt(3),atempo=sqrt(3)
2356 Trim the input so that the output contains one continuous subpart of the input.
2358 It accepts the following parameters:
2361 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2362 sample with the timestamp @var{start} will be the first sample in the output.
2365 Specify time of the first audio sample that will be dropped, i.e. the
2366 audio sample immediately preceding the one with the timestamp @var{end} will be
2367 the last sample in the output.
2370 Same as @var{start}, except this option sets the start timestamp in samples
2374 Same as @var{end}, except this option sets the end timestamp in samples instead
2378 The maximum duration of the output in seconds.
2381 The number of the first sample that should be output.
2384 The number of the first sample that should be dropped.
2387 @option{start}, @option{end}, and @option{duration} are expressed as time
2388 duration specifications; see
2389 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2391 Note that the first two sets of the start/end options and the @option{duration}
2392 option look at the frame timestamp, while the _sample options simply count the
2393 samples that pass through the filter. So start/end_pts and start/end_sample will
2394 give different results when the timestamps are wrong, inexact or do not start at
2395 zero. Also note that this filter does not modify the timestamps. If you wish
2396 to have the output timestamps start at zero, insert the asetpts filter after the
2399 If multiple start or end options are set, this filter tries to be greedy and
2400 keep all samples that match at least one of the specified constraints. To keep
2401 only the part that matches all the constraints at once, chain multiple atrim
2404 The defaults are such that all the input is kept. So it is possible to set e.g.
2405 just the end values to keep everything before the specified time.
2410 Drop everything except the second minute of input:
2412 ffmpeg -i INPUT -af atrim=60:120
2416 Keep only the first 1000 samples:
2418 ffmpeg -i INPUT -af atrim=end_sample=1000
2425 Apply a two-pole Butterworth band-pass filter with central
2426 frequency @var{frequency}, and (3dB-point) band-width width.
2427 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2428 instead of the default: constant 0dB peak gain.
2429 The filter roll off at 6dB per octave (20dB per decade).
2431 The filter accepts the following options:
2435 Set the filter's central frequency. Default is @code{3000}.
2438 Constant skirt gain if set to 1. Defaults to 0.
2441 Set method to specify band-width of filter.
2456 Specify the band-width of a filter in width_type units.
2459 How much to use filtered signal in output. Default is 1.
2460 Range is between 0 and 1.
2463 Specify which channels to filter, by default all available are filtered.
2466 @subsection Commands
2468 This filter supports the following commands:
2471 Change bandpass frequency.
2472 Syntax for the command is : "@var{frequency}"
2475 Change bandpass width_type.
2476 Syntax for the command is : "@var{width_type}"
2479 Change bandpass width.
2480 Syntax for the command is : "@var{width}"
2483 Change bandpass mix.
2484 Syntax for the command is : "@var{mix}"
2489 Apply a two-pole Butterworth band-reject filter with central
2490 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2491 The filter roll off at 6dB per octave (20dB per decade).
2493 The filter accepts the following options:
2497 Set the filter's central frequency. Default is @code{3000}.
2500 Set method to specify band-width of filter.
2515 Specify the band-width of a filter in width_type units.
2518 How much to use filtered signal in output. Default is 1.
2519 Range is between 0 and 1.
2522 Specify which channels to filter, by default all available are filtered.
2525 @subsection Commands
2527 This filter supports the following commands:
2530 Change bandreject frequency.
2531 Syntax for the command is : "@var{frequency}"
2534 Change bandreject width_type.
2535 Syntax for the command is : "@var{width_type}"
2538 Change bandreject width.
2539 Syntax for the command is : "@var{width}"
2542 Change bandreject mix.
2543 Syntax for the command is : "@var{mix}"
2546 @section bass, lowshelf
2548 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2549 shelving filter with a response similar to that of a standard
2550 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2552 The filter accepts the following options:
2556 Give the gain at 0 Hz. Its useful range is about -20
2557 (for a large cut) to +20 (for a large boost).
2558 Beware of clipping when using a positive gain.
2561 Set the filter's central frequency and so can be used
2562 to extend or reduce the frequency range to be boosted or cut.
2563 The default value is @code{100} Hz.
2566 Set method to specify band-width of filter.
2581 Determine how steep is the filter's shelf transition.
2584 How much to use filtered signal in output. Default is 1.
2585 Range is between 0 and 1.
2588 Specify which channels to filter, by default all available are filtered.
2591 @subsection Commands
2593 This filter supports the following commands:
2596 Change bass frequency.
2597 Syntax for the command is : "@var{frequency}"
2600 Change bass width_type.
2601 Syntax for the command is : "@var{width_type}"
2605 Syntax for the command is : "@var{width}"
2609 Syntax for the command is : "@var{gain}"
2613 Syntax for the command is : "@var{mix}"
2618 Apply a biquad IIR filter with the given coefficients.
2619 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2620 are the numerator and denominator coefficients respectively.
2621 and @var{channels}, @var{c} specify which channels to filter, by default all
2622 available are filtered.
2624 @subsection Commands
2626 This filter supports the following commands:
2634 Change biquad parameter.
2635 Syntax for the command is : "@var{value}"
2638 How much to use filtered signal in output. Default is 1.
2639 Range is between 0 and 1.
2643 Bauer stereo to binaural transformation, which improves headphone listening of
2644 stereo audio records.
2646 To enable compilation of this filter you need to configure FFmpeg with
2647 @code{--enable-libbs2b}.
2649 It accepts the following parameters:
2653 Pre-defined crossfeed level.
2657 Default level (fcut=700, feed=50).
2660 Chu Moy circuit (fcut=700, feed=60).
2663 Jan Meier circuit (fcut=650, feed=95).
2668 Cut frequency (in Hz).
2677 Remap input channels to new locations.
2679 It accepts the following parameters:
2682 Map channels from input to output. The argument is a '|'-separated list of
2683 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2684 @var{in_channel} form. @var{in_channel} can be either the name of the input
2685 channel (e.g. FL for front left) or its index in the input channel layout.
2686 @var{out_channel} is the name of the output channel or its index in the output
2687 channel layout. If @var{out_channel} is not given then it is implicitly an
2688 index, starting with zero and increasing by one for each mapping.
2690 @item channel_layout
2691 The channel layout of the output stream.
2694 If no mapping is present, the filter will implicitly map input channels to
2695 output channels, preserving indices.
2697 @subsection Examples
2701 For example, assuming a 5.1+downmix input MOV file,
2703 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2705 will create an output WAV file tagged as stereo from the downmix channels of
2709 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2711 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
2715 @section channelsplit
2717 Split each channel from an input audio stream into a separate output stream.
2719 It accepts the following parameters:
2721 @item channel_layout
2722 The channel layout of the input stream. The default is "stereo".
2724 A channel layout describing the channels to be extracted as separate output streams
2725 or "all" to extract each input channel as a separate stream. The default is "all".
2727 Choosing channels not present in channel layout in the input will result in an error.
2730 @subsection Examples
2734 For example, assuming a stereo input MP3 file,
2736 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
2738 will create an output Matroska file with two audio streams, one containing only
2739 the left channel and the other the right channel.
2742 Split a 5.1 WAV file into per-channel files:
2744 ffmpeg -i in.wav -filter_complex
2745 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
2746 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
2747 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
2752 Extract only LFE from a 5.1 WAV file:
2754 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
2755 -map '[LFE]' lfe.wav
2760 Add a chorus effect to the audio.
2762 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
2764 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
2765 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
2766 The modulation depth defines the range the modulated delay is played before or after
2767 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
2768 sound tuned around the original one, like in a chorus where some vocals are slightly
2771 It accepts the following parameters:
2774 Set input gain. Default is 0.4.
2777 Set output gain. Default is 0.4.
2780 Set delays. A typical delay is around 40ms to 60ms.
2792 @subsection Examples
2798 chorus=0.7:0.9:55:0.4:0.25:2
2804 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
2808 Fuller sounding chorus with three delays:
2810 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
2815 Compress or expand the audio's dynamic range.
2817 It accepts the following parameters:
2823 A list of times in seconds for each channel over which the instantaneous level
2824 of the input signal is averaged to determine its volume. @var{attacks} refers to
2825 increase of volume and @var{decays} refers to decrease of volume. For most
2826 situations, the attack time (response to the audio getting louder) should be
2827 shorter than the decay time, because the human ear is more sensitive to sudden
2828 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
2829 a typical value for decay is 0.8 seconds.
2830 If specified number of attacks & decays is lower than number of channels, the last
2831 set attack/decay will be used for all remaining channels.
2834 A list of points for the transfer function, specified in dB relative to the
2835 maximum possible signal amplitude. Each key points list must be defined using
2836 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
2837 @code{x0/y0 x1/y1 x2/y2 ....}
2839 The input values must be in strictly increasing order but the transfer function
2840 does not have to be monotonically rising. The point @code{0/0} is assumed but
2841 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
2842 function are @code{-70/-70|-60/-20|1/0}.
2845 Set the curve radius in dB for all joints. It defaults to 0.01.
2848 Set the additional gain in dB to be applied at all points on the transfer
2849 function. This allows for easy adjustment of the overall gain.
2853 Set an initial volume, in dB, to be assumed for each channel when filtering
2854 starts. This permits the user to supply a nominal level initially, so that, for
2855 example, a very large gain is not applied to initial signal levels before the
2856 companding has begun to operate. A typical value for audio which is initially
2857 quiet is -90 dB. It defaults to 0.
2860 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
2861 delayed before being fed to the volume adjuster. Specifying a delay
2862 approximately equal to the attack/decay times allows the filter to effectively
2863 operate in predictive rather than reactive mode. It defaults to 0.
2867 @subsection Examples
2871 Make music with both quiet and loud passages suitable for listening to in a
2874 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
2877 Another example for audio with whisper and explosion parts:
2879 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
2883 A noise gate for when the noise is at a lower level than the signal:
2885 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
2889 Here is another noise gate, this time for when the noise is at a higher level
2890 than the signal (making it, in some ways, similar to squelch):
2892 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
2896 2:1 compression starting at -6dB:
2898 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
2902 2:1 compression starting at -9dB:
2904 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
2908 2:1 compression starting at -12dB:
2910 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
2914 2:1 compression starting at -18dB:
2916 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
2920 3:1 compression starting at -15dB:
2922 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
2928 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
2934 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
2938 Hard limiter at -6dB:
2940 compand=attacks=0:points=-80/-80|-6/-6|20/-6
2944 Hard limiter at -12dB:
2946 compand=attacks=0:points=-80/-80|-12/-12|20/-12
2950 Hard noise gate at -35 dB:
2952 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
2958 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
2962 @section compensationdelay
2964 Compensation Delay Line is a metric based delay to compensate differing
2965 positions of microphones or speakers.
2967 For example, you have recorded guitar with two microphones placed in
2968 different location. Because the front of sound wave has fixed speed in
2969 normal conditions, the phasing of microphones can vary and depends on
2970 their location and interposition. The best sound mix can be achieved when
2971 these microphones are in phase (synchronized). Note that distance of
2972 ~30 cm between microphones makes one microphone to capture signal in
2973 antiphase to another microphone. That makes the final mix sounding moody.
2974 This filter helps to solve phasing problems by adding different delays
2975 to each microphone track and make them synchronized.
2977 The best result can be reached when you take one track as base and
2978 synchronize other tracks one by one with it.
2979 Remember that synchronization/delay tolerance depends on sample rate, too.
2980 Higher sample rates will give more tolerance.
2982 It accepts the following parameters:
2986 Set millimeters distance. This is compensation distance for fine tuning.
2990 Set cm distance. This is compensation distance for tightening distance setup.
2994 Set meters distance. This is compensation distance for hard distance setup.
2998 Set dry amount. Amount of unprocessed (dry) signal.
3002 Set wet amount. Amount of processed (wet) signal.
3006 Set temperature degree in Celsius. This is the temperature of the environment.
3011 Apply headphone crossfeed filter.
3013 Crossfeed is the process of blending the left and right channels of stereo
3015 It is mainly used to reduce extreme stereo separation of low frequencies.
3017 The intent is to produce more speaker like sound to the listener.
3019 The filter accepts the following options:
3023 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3024 This sets gain of low shelf filter for side part of stereo image.
3025 Default is -6dB. Max allowed is -30db when strength is set to 1.
3028 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3029 This sets cut off frequency of low shelf filter. Default is cut off near
3030 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3033 Set input gain. Default is 0.9.
3036 Set output gain. Default is 1.
3039 @section crystalizer
3040 Simple algorithm to expand audio dynamic range.
3042 The filter accepts the following options:
3046 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3047 (unchanged sound) to 10.0 (maximum effect).
3050 Enable clipping. By default is enabled.
3054 Apply a DC shift to the audio.
3056 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3057 in the recording chain) from the audio. The effect of a DC offset is reduced
3058 headroom and hence volume. The @ref{astats} filter can be used to determine if
3059 a signal has a DC offset.
3063 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3067 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3068 used to prevent clipping.
3073 Apply de-essing to the audio samples.
3077 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3081 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3085 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3089 Set the output mode.
3091 It accepts the following values:
3094 Pass input unchanged.
3097 Pass ess filtered out.
3102 Default value is @var{o}.
3108 Measure audio dynamic range.
3110 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3111 is found in transition material. And anything less that 8 have very poor dynamics
3112 and is very compressed.
3114 The filter accepts the following options:
3118 Set window length in seconds used to split audio into segments of equal length.
3119 Default is 3 seconds.
3123 Dynamic Audio Normalizer.
3125 This filter applies a certain amount of gain to the input audio in order
3126 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3127 contrast to more "simple" normalization algorithms, the Dynamic Audio
3128 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3129 This allows for applying extra gain to the "quiet" sections of the audio
3130 while avoiding distortions or clipping the "loud" sections. In other words:
3131 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3132 sections, in the sense that the volume of each section is brought to the
3133 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3134 this goal *without* applying "dynamic range compressing". It will retain 100%
3135 of the dynamic range *within* each section of the audio file.
3139 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3140 Default is 500 milliseconds.
3141 The Dynamic Audio Normalizer processes the input audio in small chunks,
3142 referred to as frames. This is required, because a peak magnitude has no
3143 meaning for just a single sample value. Instead, we need to determine the
3144 peak magnitude for a contiguous sequence of sample values. While a "standard"
3145 normalizer would simply use the peak magnitude of the complete file, the
3146 Dynamic Audio Normalizer determines the peak magnitude individually for each
3147 frame. The length of a frame is specified in milliseconds. By default, the
3148 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3149 been found to give good results with most files.
3150 Note that the exact frame length, in number of samples, will be determined
3151 automatically, based on the sampling rate of the individual input audio file.
3154 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3155 number. Default is 31.
3156 Probably the most important parameter of the Dynamic Audio Normalizer is the
3157 @code{window size} of the Gaussian smoothing filter. The filter's window size
3158 is specified in frames, centered around the current frame. For the sake of
3159 simplicity, this must be an odd number. Consequently, the default value of 31
3160 takes into account the current frame, as well as the 15 preceding frames and
3161 the 15 subsequent frames. Using a larger window results in a stronger
3162 smoothing effect and thus in less gain variation, i.e. slower gain
3163 adaptation. Conversely, using a smaller window results in a weaker smoothing
3164 effect and thus in more gain variation, i.e. faster gain adaptation.
3165 In other words, the more you increase this value, the more the Dynamic Audio
3166 Normalizer will behave like a "traditional" normalization filter. On the
3167 contrary, the more you decrease this value, the more the Dynamic Audio
3168 Normalizer will behave like a dynamic range compressor.
3171 Set the target peak value. This specifies the highest permissible magnitude
3172 level for the normalized audio input. This filter will try to approach the
3173 target peak magnitude as closely as possible, but at the same time it also
3174 makes sure that the normalized signal will never exceed the peak magnitude.
3175 A frame's maximum local gain factor is imposed directly by the target peak
3176 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3177 It is not recommended to go above this value.
3180 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3181 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3182 factor for each input frame, i.e. the maximum gain factor that does not
3183 result in clipping or distortion. The maximum gain factor is determined by
3184 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3185 additionally bounds the frame's maximum gain factor by a predetermined
3186 (global) maximum gain factor. This is done in order to avoid excessive gain
3187 factors in "silent" or almost silent frames. By default, the maximum gain
3188 factor is 10.0, For most inputs the default value should be sufficient and
3189 it usually is not recommended to increase this value. Though, for input
3190 with an extremely low overall volume level, it may be necessary to allow even
3191 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3192 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3193 Instead, a "sigmoid" threshold function will be applied. This way, the
3194 gain factors will smoothly approach the threshold value, but never exceed that
3198 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3199 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3200 This means that the maximum local gain factor for each frame is defined
3201 (only) by the frame's highest magnitude sample. This way, the samples can
3202 be amplified as much as possible without exceeding the maximum signal
3203 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3204 Normalizer can also take into account the frame's root mean square,
3205 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3206 determine the power of a time-varying signal. It is therefore considered
3207 that the RMS is a better approximation of the "perceived loudness" than
3208 just looking at the signal's peak magnitude. Consequently, by adjusting all
3209 frames to a constant RMS value, a uniform "perceived loudness" can be
3210 established. If a target RMS value has been specified, a frame's local gain
3211 factor is defined as the factor that would result in exactly that RMS value.
3212 Note, however, that the maximum local gain factor is still restricted by the
3213 frame's highest magnitude sample, in order to prevent clipping.
3216 Enable channels coupling. By default is enabled.
3217 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3218 amount. This means the same gain factor will be applied to all channels, i.e.
3219 the maximum possible gain factor is determined by the "loudest" channel.
3220 However, in some recordings, it may happen that the volume of the different
3221 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3222 In this case, this option can be used to disable the channel coupling. This way,
3223 the gain factor will be determined independently for each channel, depending
3224 only on the individual channel's highest magnitude sample. This allows for
3225 harmonizing the volume of the different channels.
3228 Enable DC bias correction. By default is disabled.
3229 An audio signal (in the time domain) is a sequence of sample values.
3230 In the Dynamic Audio Normalizer these sample values are represented in the
3231 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3232 audio signal, or "waveform", should be centered around the zero point.
3233 That means if we calculate the mean value of all samples in a file, or in a
3234 single frame, then the result should be 0.0 or at least very close to that
3235 value. If, however, there is a significant deviation of the mean value from
3236 0.0, in either positive or negative direction, this is referred to as a
3237 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3238 Audio Normalizer provides optional DC bias correction.
3239 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3240 the mean value, or "DC correction" offset, of each input frame and subtract
3241 that value from all of the frame's sample values which ensures those samples
3242 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3243 boundaries, the DC correction offset values will be interpolated smoothly
3244 between neighbouring frames.
3247 Enable alternative boundary mode. By default is disabled.
3248 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3249 around each frame. This includes the preceding frames as well as the
3250 subsequent frames. However, for the "boundary" frames, located at the very
3251 beginning and at the very end of the audio file, not all neighbouring
3252 frames are available. In particular, for the first few frames in the audio
3253 file, the preceding frames are not known. And, similarly, for the last few
3254 frames in the audio file, the subsequent frames are not known. Thus, the
3255 question arises which gain factors should be assumed for the missing frames
3256 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3257 to deal with this situation. The default boundary mode assumes a gain factor
3258 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3259 "fade out" at the beginning and at the end of the input, respectively.
3262 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3263 By default, the Dynamic Audio Normalizer does not apply "traditional"
3264 compression. This means that signal peaks will not be pruned and thus the
3265 full dynamic range will be retained within each local neighbourhood. However,
3266 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3267 normalization algorithm with a more "traditional" compression.
3268 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3269 (thresholding) function. If (and only if) the compression feature is enabled,
3270 all input frames will be processed by a soft knee thresholding function prior
3271 to the actual normalization process. Put simply, the thresholding function is
3272 going to prune all samples whose magnitude exceeds a certain threshold value.
3273 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3274 value. Instead, the threshold value will be adjusted for each individual
3276 In general, smaller parameters result in stronger compression, and vice versa.
3277 Values below 3.0 are not recommended, because audible distortion may appear.
3282 Make audio easier to listen to on headphones.
3284 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3285 so that when listened to on headphones the stereo image is moved from
3286 inside your head (standard for headphones) to outside and in front of
3287 the listener (standard for speakers).
3293 Apply a two-pole peaking equalisation (EQ) filter. With this
3294 filter, the signal-level at and around a selected frequency can
3295 be increased or decreased, whilst (unlike bandpass and bandreject
3296 filters) that at all other frequencies is unchanged.
3298 In order to produce complex equalisation curves, this filter can
3299 be given several times, each with a different central frequency.
3301 The filter accepts the following options:
3305 Set the filter's central frequency in Hz.
3308 Set method to specify band-width of filter.
3323 Specify the band-width of a filter in width_type units.
3326 Set the required gain or attenuation in dB.
3327 Beware of clipping when using a positive gain.
3330 How much to use filtered signal in output. Default is 1.
3331 Range is between 0 and 1.
3334 Specify which channels to filter, by default all available are filtered.
3337 @subsection Examples
3340 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3342 equalizer=f=1000:t=h:width=200:g=-10
3346 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3348 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3352 @subsection Commands
3354 This filter supports the following commands:
3357 Change equalizer frequency.
3358 Syntax for the command is : "@var{frequency}"
3361 Change equalizer width_type.
3362 Syntax for the command is : "@var{width_type}"
3365 Change equalizer width.
3366 Syntax for the command is : "@var{width}"
3369 Change equalizer gain.
3370 Syntax for the command is : "@var{gain}"
3373 Change equalizer mix.
3374 Syntax for the command is : "@var{mix}"
3377 @section extrastereo
3379 Linearly increases the difference between left and right channels which
3380 adds some sort of "live" effect to playback.
3382 The filter accepts the following options:
3386 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3387 (average of both channels), with 1.0 sound will be unchanged, with
3388 -1.0 left and right channels will be swapped.
3391 Enable clipping. By default is enabled.
3394 @section firequalizer
3395 Apply FIR Equalization using arbitrary frequency response.
3397 The filter accepts the following option:
3401 Set gain curve equation (in dB). The expression can contain variables:
3404 the evaluated frequency
3408 channel number, set to 0 when multichannels evaluation is disabled
3410 channel id, see libavutil/channel_layout.h, set to the first channel id when
3411 multichannels evaluation is disabled
3415 channel_layout, see libavutil/channel_layout.h
3420 @item gain_interpolate(f)
3421 interpolate gain on frequency f based on gain_entry
3422 @item cubic_interpolate(f)
3423 same as gain_interpolate, but smoother
3425 This option is also available as command. Default is @code{gain_interpolate(f)}.
3428 Set gain entry for gain_interpolate function. The expression can
3432 store gain entry at frequency f with value g
3434 This option is also available as command.
3437 Set filter delay in seconds. Higher value means more accurate.
3438 Default is @code{0.01}.
3441 Set filter accuracy in Hz. Lower value means more accurate.
3442 Default is @code{5}.
3445 Set window function. Acceptable values are:
3448 rectangular window, useful when gain curve is already smooth
3450 hann window (default)
3456 3-terms continuous 1st derivative nuttall window
3458 minimum 3-terms discontinuous nuttall window
3460 4-terms continuous 1st derivative nuttall window
3462 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3464 blackman-harris window
3470 If enabled, use fixed number of audio samples. This improves speed when
3471 filtering with large delay. Default is disabled.
3474 Enable multichannels evaluation on gain. Default is disabled.
3477 Enable zero phase mode by subtracting timestamp to compensate delay.
3478 Default is disabled.
3481 Set scale used by gain. Acceptable values are:
3484 linear frequency, linear gain
3486 linear frequency, logarithmic (in dB) gain (default)
3488 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3490 logarithmic frequency, logarithmic gain
3494 Set file for dumping, suitable for gnuplot.
3497 Set scale for dumpfile. Acceptable values are same with scale option.
3501 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3502 Default is disabled.
3505 Enable minimum phase impulse response. Default is disabled.
3508 @subsection Examples
3513 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3516 lowpass at 1000 Hz with gain_entry:
3518 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3521 custom equalization:
3523 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3526 higher delay with zero phase to compensate delay:
3528 firequalizer=delay=0.1:fixed=on:zero_phase=on
3531 lowpass on left channel, highpass on right channel:
3533 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
3534 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
3539 Apply a flanging effect to the audio.
3541 The filter accepts the following options:
3545 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
3548 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
3551 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
3555 Set percentage of delayed signal mixed with original. Range from 0 to 100.
3556 Default value is 71.
3559 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
3562 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
3563 Default value is @var{sinusoidal}.
3566 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
3567 Default value is 25.
3570 Set delay-line interpolation, @var{linear} or @var{quadratic}.
3571 Default is @var{linear}.
3575 Apply Haas effect to audio.
3577 Note that this makes most sense to apply on mono signals.
3578 With this filter applied to mono signals it give some directionality and
3579 stretches its stereo image.
3581 The filter accepts the following options:
3585 Set input level. By default is @var{1}, or 0dB
3588 Set output level. By default is @var{1}, or 0dB.
3591 Set gain applied to side part of signal. By default is @var{1}.
3594 Set kind of middle source. Can be one of the following:
3604 Pick middle part signal of stereo image.
3607 Pick side part signal of stereo image.
3611 Change middle phase. By default is disabled.
3614 Set left channel delay. By default is @var{2.05} milliseconds.
3617 Set left channel balance. By default is @var{-1}.
3620 Set left channel gain. By default is @var{1}.
3623 Change left phase. By default is disabled.
3626 Set right channel delay. By defaults is @var{2.12} milliseconds.
3629 Set right channel balance. By default is @var{1}.
3632 Set right channel gain. By default is @var{1}.
3635 Change right phase. By default is enabled.
3640 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3641 embedded HDCD codes is expanded into a 20-bit PCM stream.
3643 The filter supports the Peak Extend and Low-level Gain Adjustment features
3644 of HDCD, and detects the Transient Filter flag.
3647 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3650 When using the filter with wav, note the default encoding for wav is 16-bit,
3651 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3652 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3654 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3655 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3658 The filter accepts the following options:
3661 @item disable_autoconvert
3662 Disable any automatic format conversion or resampling in the filter graph.
3664 @item process_stereo
3665 Process the stereo channels together. If target_gain does not match between
3666 channels, consider it invalid and use the last valid target_gain.
3669 Set the code detect timer period in ms.
3672 Always extend peaks above -3dBFS even if PE isn't signaled.
3675 Replace audio with a solid tone and adjust the amplitude to signal some
3676 specific aspect of the decoding process. The output file can be loaded in
3677 an audio editor alongside the original to aid analysis.
3679 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
3686 Gain adjustment level at each sample
3688 Samples where peak extend occurs
3690 Samples where the code detect timer is active
3692 Samples where the target gain does not match between channels
3698 Apply head-related transfer functions (HRTFs) to create virtual
3699 loudspeakers around the user for binaural listening via headphones.
3700 The HRIRs are provided via additional streams, for each channel
3701 one stereo input stream is needed.
3703 The filter accepts the following options:
3707 Set mapping of input streams for convolution.
3708 The argument is a '|'-separated list of channel names in order as they
3709 are given as additional stream inputs for filter.
3710 This also specify number of input streams. Number of input streams
3711 must be not less than number of channels in first stream plus one.
3714 Set gain applied to audio. Value is in dB. Default is 0.
3717 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
3718 processing audio in time domain which is slow.
3719 @var{freq} is processing audio in frequency domain which is fast.
3720 Default is @var{freq}.
3723 Set custom gain for LFE channels. Value is in dB. Default is 0.
3726 Set size of frame in number of samples which will be processed at once.
3727 Default value is @var{1024}. Allowed range is from 1024 to 96000.
3730 Set format of hrir stream.
3731 Default value is @var{stereo}. Alternative value is @var{multich}.
3732 If value is set to @var{stereo}, number of additional streams should
3733 be greater or equal to number of input channels in first input stream.
3734 Also each additional stream should have stereo number of channels.
3735 If value is set to @var{multich}, number of additional streams should
3736 be exactly one. Also number of input channels of additional stream
3737 should be equal or greater than twice number of channels of first input
3741 @subsection Examples
3745 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3746 each amovie filter use stereo file with IR coefficients as input.
3747 The files give coefficients for each position of virtual loudspeaker:
3750 -filter_complex "amovie=azi_270_ele_0_DFC.wav[sr];amovie=azi_90_ele_0_DFC.wav[sl];amovie=azi_225_ele_0_DFC.wav[br];amovie=azi_135_ele_0_DFC.wav[bl];amovie=azi_0_ele_0_DFC.wav,asplit[fc][lfe];amovie=azi_35_ele_0_DFC.wav[fl];amovie=azi_325_ele_0_DFC.wav[fr];[0:a][fl][fr][fc][lfe][bl][br][sl][sr]headphone=FL|FR|FC|LFE|BL|BR|SL|SR"
3755 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3756 but now in @var{multich} @var{hrir} format.
3758 ffmpeg -i input.wav -filter_complex "amovie=minp.wav[hrirs];[0:a][hrirs]headphone=map=FL|FR|FC|LFE|BL|BR|SL|SR:hrir=multich"
3765 Apply a high-pass filter with 3dB point frequency.
3766 The filter can be either single-pole, or double-pole (the default).
3767 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3769 The filter accepts the following options:
3773 Set frequency in Hz. Default is 3000.
3776 Set number of poles. Default is 2.
3779 Set method to specify band-width of filter.
3794 Specify the band-width of a filter in width_type units.
3795 Applies only to double-pole filter.
3796 The default is 0.707q and gives a Butterworth response.
3799 How much to use filtered signal in output. Default is 1.
3800 Range is between 0 and 1.
3803 Specify which channels to filter, by default all available are filtered.
3806 @subsection Commands
3808 This filter supports the following commands:
3811 Change highpass frequency.
3812 Syntax for the command is : "@var{frequency}"
3815 Change highpass width_type.
3816 Syntax for the command is : "@var{width_type}"
3819 Change highpass width.
3820 Syntax for the command is : "@var{width}"
3823 Change highpass mix.
3824 Syntax for the command is : "@var{mix}"
3829 Join multiple input streams into one multi-channel stream.
3831 It accepts the following parameters:
3835 The number of input streams. It defaults to 2.
3837 @item channel_layout
3838 The desired output channel layout. It defaults to stereo.
3841 Map channels from inputs to output. The argument is a '|'-separated list of
3842 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
3843 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
3844 can be either the name of the input channel (e.g. FL for front left) or its
3845 index in the specified input stream. @var{out_channel} is the name of the output
3849 The filter will attempt to guess the mappings when they are not specified
3850 explicitly. It does so by first trying to find an unused matching input channel
3851 and if that fails it picks the first unused input channel.
3853 Join 3 inputs (with properly set channel layouts):
3855 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
3858 Build a 5.1 output from 6 single-channel streams:
3860 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
3861 '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'
3867 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
3869 To enable compilation of this filter you need to configure FFmpeg with
3870 @code{--enable-ladspa}.
3874 Specifies the name of LADSPA plugin library to load. If the environment
3875 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
3876 each one of the directories specified by the colon separated list in
3877 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
3878 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
3879 @file{/usr/lib/ladspa/}.
3882 Specifies the plugin within the library. Some libraries contain only
3883 one plugin, but others contain many of them. If this is not set filter
3884 will list all available plugins within the specified library.
3887 Set the '|' separated list of controls which are zero or more floating point
3888 values that determine the behavior of the loaded plugin (for example delay,
3890 Controls need to be defined using the following syntax:
3891 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
3892 @var{valuei} is the value set on the @var{i}-th control.
3893 Alternatively they can be also defined using the following syntax:
3894 @var{value0}|@var{value1}|@var{value2}|..., where
3895 @var{valuei} is the value set on the @var{i}-th control.
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.
3919 @subsection Examples
3923 List all available plugins within amp (LADSPA example plugin) library:
3929 List all available controls and their valid ranges for @code{vcf_notch}
3930 plugin from @code{VCF} library:
3932 ladspa=f=vcf:p=vcf_notch:c=help
3936 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
3939 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
3943 Add reverberation to the audio using TAP-plugins
3944 (Tom's Audio Processing plugins):
3946 ladspa=file=tap_reverb:tap_reverb
3950 Generate white noise, with 0.2 amplitude:
3952 ladspa=file=cmt:noise_source_white:c=c0=.2
3956 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
3957 @code{C* Audio Plugin Suite} (CAPS) library:
3959 ladspa=file=caps:Click:c=c1=20'
3963 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
3965 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
3969 Increase volume by 20dB using fast lookahead limiter from Steve Harris
3970 @code{SWH Plugins} collection:
3972 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
3976 Attenuate low frequencies using Multiband EQ from Steve Harris
3977 @code{SWH Plugins} collection:
3979 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
3983 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
3986 ladspa=caps:Narrower
3990 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
3992 ladspa=caps:White:.2
3996 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
3998 ladspa=caps:Fractal:c=c1=1
4002 Dynamic volume normalization using @code{VLevel} plugin:
4004 ladspa=vlevel-ladspa:vlevel_mono
4008 @subsection Commands
4010 This filter supports the following commands:
4013 Modify the @var{N}-th control value.
4015 If the specified value is not valid, it is ignored and prior one is kept.
4020 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4021 Support for both single pass (livestreams, files) and double pass (files) modes.
4022 This algorithm can target IL, LRA, and maximum true peak. To accurately detect true peaks,
4023 the audio stream will be upsampled to 192 kHz unless the normalization mode is linear.
4024 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4026 The filter accepts the following options:
4030 Set integrated loudness target.
4031 Range is -70.0 - -5.0. Default value is -24.0.
4034 Set loudness range target.
4035 Range is 1.0 - 20.0. Default value is 7.0.
4038 Set maximum true peak.
4039 Range is -9.0 - +0.0. Default value is -2.0.
4041 @item measured_I, measured_i
4042 Measured IL of input file.
4043 Range is -99.0 - +0.0.
4045 @item measured_LRA, measured_lra
4046 Measured LRA of input file.
4047 Range is 0.0 - 99.0.
4049 @item measured_TP, measured_tp
4050 Measured true peak of input file.
4051 Range is -99.0 - +99.0.
4053 @item measured_thresh
4054 Measured threshold of input file.
4055 Range is -99.0 - +0.0.
4058 Set offset gain. Gain is applied before the true-peak limiter.
4059 Range is -99.0 - +99.0. Default is +0.0.
4062 Normalize linearly if possible.
4063 measured_I, measured_LRA, measured_TP, and measured_thresh must also
4064 to be specified in order to use this mode.
4065 Options are true or false. Default is true.
4068 Treat mono input files as "dual-mono". If a mono file is intended for playback
4069 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4070 If set to @code{true}, this option will compensate for this effect.
4071 Multi-channel input files are not affected by this option.
4072 Options are true or false. Default is false.
4075 Set print format for stats. Options are summary, json, or none.
4076 Default value is none.
4081 Apply a low-pass filter with 3dB point frequency.
4082 The filter can be either single-pole or double-pole (the default).
4083 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4085 The filter accepts the following options:
4089 Set frequency in Hz. Default is 500.
4092 Set number of poles. Default is 2.
4095 Set method to specify band-width of filter.
4110 Specify the band-width of a filter in width_type units.
4111 Applies only to double-pole filter.
4112 The default is 0.707q and gives a Butterworth response.
4115 How much to use filtered signal in output. Default is 1.
4116 Range is between 0 and 1.
4119 Specify which channels to filter, by default all available are filtered.
4122 @subsection Examples
4125 Lowpass only LFE channel, it LFE is not present it does nothing:
4131 @subsection Commands
4133 This filter supports the following commands:
4136 Change lowpass frequency.
4137 Syntax for the command is : "@var{frequency}"
4140 Change lowpass width_type.
4141 Syntax for the command is : "@var{width_type}"
4144 Change lowpass width.
4145 Syntax for the command is : "@var{width}"
4149 Syntax for the command is : "@var{mix}"
4154 Load a LV2 (LADSPA Version 2) plugin.
4156 To enable compilation of this filter you need to configure FFmpeg with
4157 @code{--enable-lv2}.
4161 Specifies the plugin URI. You may need to escape ':'.
4164 Set the '|' separated list of controls which are zero or more floating point
4165 values that determine the behavior of the loaded plugin (for example delay,
4167 If @option{controls} is set to @code{help}, all available controls and
4168 their valid ranges are printed.
4170 @item sample_rate, s
4171 Specify the sample rate, default to 44100. Only used if plugin have
4175 Set the number of samples per channel per each output frame, default
4176 is 1024. Only used if plugin have zero inputs.
4179 Set the minimum duration of the sourced audio. See
4180 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4181 for the accepted syntax.
4182 Note that the resulting duration may be greater than the specified duration,
4183 as the generated audio is always cut at the end of a complete frame.
4184 If not specified, or the expressed duration is negative, the audio is
4185 supposed to be generated forever.
4186 Only used if plugin have zero inputs.
4189 @subsection Examples
4193 Apply bass enhancer plugin from Calf:
4195 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4199 Apply vinyl plugin from Calf:
4201 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4205 Apply bit crusher plugin from ArtyFX:
4207 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4212 Multiband Compress or expand the audio's dynamic range.
4214 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4215 This is akin to the crossover of a loudspeaker, and results in flat frequency
4216 response when absent compander action.
4218 It accepts the following parameters:
4222 This option syntax is:
4223 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4224 For explanation of each item refer to compand filter documentation.
4230 Mix channels with specific gain levels. The filter accepts the output
4231 channel layout followed by a set of channels definitions.
4233 This filter is also designed to efficiently remap the channels of an audio
4236 The filter accepts parameters of the form:
4237 "@var{l}|@var{outdef}|@var{outdef}|..."
4241 output channel layout or number of channels
4244 output channel specification, of the form:
4245 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4248 output channel to define, either a channel name (FL, FR, etc.) or a channel
4249 number (c0, c1, etc.)
4252 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4255 input channel to use, see out_name for details; it is not possible to mix
4256 named and numbered input channels
4259 If the `=' in a channel specification is replaced by `<', then the gains for
4260 that specification will be renormalized so that the total is 1, thus
4261 avoiding clipping noise.
4263 @subsection Mixing examples
4265 For example, if you want to down-mix from stereo to mono, but with a bigger
4266 factor for the left channel:
4268 pan=1c|c0=0.9*c0+0.1*c1
4271 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4272 7-channels surround:
4274 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4277 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4278 that should be preferred (see "-ac" option) unless you have very specific
4281 @subsection Remapping examples
4283 The channel remapping will be effective if, and only if:
4286 @item gain coefficients are zeroes or ones,
4287 @item only one input per channel output,
4290 If all these conditions are satisfied, the filter will notify the user ("Pure
4291 channel mapping detected"), and use an optimized and lossless method to do the
4294 For example, if you have a 5.1 source and want a stereo audio stream by
4295 dropping the extra channels:
4297 pan="stereo| c0=FL | c1=FR"
4300 Given the same source, you can also switch front left and front right channels
4301 and keep the input channel layout:
4303 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4306 If the input is a stereo audio stream, you can mute the front left channel (and
4307 still keep the stereo channel layout) with:
4312 Still with a stereo audio stream input, you can copy the right channel in both
4313 front left and right:
4315 pan="stereo| c0=FR | c1=FR"
4320 ReplayGain scanner filter. This filter takes an audio stream as an input and
4321 outputs it unchanged.
4322 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4326 Convert the audio sample format, sample rate and channel layout. It is
4327 not meant to be used directly.
4330 Apply time-stretching and pitch-shifting with librubberband.
4332 To enable compilation of this filter, you need to configure FFmpeg with
4333 @code{--enable-librubberband}.
4335 The filter accepts the following options:
4339 Set tempo scale factor.
4342 Set pitch scale factor.
4345 Set transients detector.
4346 Possible values are:
4355 Possible values are:
4364 Possible values are:
4371 Set processing window size.
4372 Possible values are:
4381 Possible values are:
4388 Enable formant preservation when shift pitching.
4389 Possible values are:
4397 Possible values are:
4406 Possible values are:
4413 @section sidechaincompress
4415 This filter acts like normal compressor but has the ability to compress
4416 detected signal using second input signal.
4417 It needs two input streams and returns one output stream.
4418 First input stream will be processed depending on second stream signal.
4419 The filtered signal then can be filtered with other filters in later stages of
4420 processing. See @ref{pan} and @ref{amerge} filter.
4422 The filter accepts the following options:
4426 Set input gain. Default is 1. Range is between 0.015625 and 64.
4429 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4430 Default is @code{downward}.
4433 If a signal of second stream raises above this level it will affect the gain
4434 reduction of first stream.
4435 By default is 0.125. Range is between 0.00097563 and 1.
4438 Set a ratio about which the signal is reduced. 1:2 means that if the level
4439 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4440 Default is 2. Range is between 1 and 20.
4443 Amount of milliseconds the signal has to rise above the threshold before gain
4444 reduction starts. Default is 20. Range is between 0.01 and 2000.
4447 Amount of milliseconds the signal has to fall below the threshold before
4448 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4451 Set the amount by how much signal will be amplified after processing.
4452 Default is 1. Range is from 1 to 64.
4455 Curve the sharp knee around the threshold to enter gain reduction more softly.
4456 Default is 2.82843. Range is between 1 and 8.
4459 Choose if the @code{average} level between all channels of side-chain stream
4460 or the louder(@code{maximum}) channel of side-chain stream affects the
4461 reduction. Default is @code{average}.
4464 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4465 of @code{rms}. Default is @code{rms} which is mainly smoother.
4468 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4471 How much to use compressed signal in output. Default is 1.
4472 Range is between 0 and 1.
4475 @subsection Examples
4479 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
4480 depending on the signal of 2nd input and later compressed signal to be
4481 merged with 2nd input:
4483 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
4487 @section sidechaingate
4489 A sidechain gate acts like a normal (wideband) gate but has the ability to
4490 filter the detected signal before sending it to the gain reduction stage.
4491 Normally a gate uses the full range signal to detect a level above the
4493 For example: If you cut all lower frequencies from your sidechain signal
4494 the gate will decrease the volume of your track only if not enough highs
4495 appear. With this technique you are able to reduce the resonation of a
4496 natural drum or remove "rumbling" of muted strokes from a heavily distorted
4498 It needs two input streams and returns one output stream.
4499 First input stream will be processed depending on second stream signal.
4501 The filter accepts the following options:
4505 Set input level before filtering.
4506 Default is 1. Allowed range is from 0.015625 to 64.
4509 Set the mode of operation. Can be @code{upward} or @code{downward}.
4510 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
4511 will be amplified, expanding dynamic range in upward direction.
4512 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
4515 Set the level of gain reduction when the signal is below the threshold.
4516 Default is 0.06125. Allowed range is from 0 to 1.
4517 Setting this to 0 disables reduction and then filter behaves like expander.
4520 If a signal rises above this level the gain reduction is released.
4521 Default is 0.125. Allowed range is from 0 to 1.
4524 Set a ratio about which the signal is reduced.
4525 Default is 2. Allowed range is from 1 to 9000.
4528 Amount of milliseconds the signal has to rise above the threshold before gain
4530 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
4533 Amount of milliseconds the signal has to fall below the threshold before the
4534 reduction is increased again. Default is 250 milliseconds.
4535 Allowed range is from 0.01 to 9000.
4538 Set amount of amplification of signal after processing.
4539 Default is 1. Allowed range is from 1 to 64.
4542 Curve the sharp knee around the threshold to enter gain reduction more softly.
4543 Default is 2.828427125. Allowed range is from 1 to 8.
4546 Choose if exact signal should be taken for detection or an RMS like one.
4547 Default is rms. Can be peak or rms.
4550 Choose if the average level between all channels or the louder channel affects
4552 Default is average. Can be average or maximum.
4555 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
4558 @section silencedetect
4560 Detect silence in an audio stream.
4562 This filter logs a message when it detects that the input audio volume is less
4563 or equal to a noise tolerance value for a duration greater or equal to the
4564 minimum detected noise duration.
4566 The printed times and duration are expressed in seconds.
4568 The filter accepts the following options:
4572 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4573 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4576 Set silence duration until notification (default is 2 seconds).
4579 Process each channel separately, instead of combined. By default is disabled.
4582 @subsection Examples
4586 Detect 5 seconds of silence with -50dB noise tolerance:
4588 silencedetect=n=-50dB:d=5
4592 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4593 tolerance in @file{silence.mp3}:
4595 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4599 @section silenceremove
4601 Remove silence from the beginning, middle or end of the audio.
4603 The filter accepts the following options:
4607 This value is used to indicate if audio should be trimmed at beginning of
4608 the audio. A value of zero indicates no silence should be trimmed from the
4609 beginning. When specifying a non-zero value, it trims audio up until it
4610 finds non-silence. Normally, when trimming silence from beginning of audio
4611 the @var{start_periods} will be @code{1} but it can be increased to higher
4612 values to trim all audio up to specific count of non-silence periods.
4613 Default value is @code{0}.
4615 @item start_duration
4616 Specify the amount of time that non-silence must be detected before it stops
4617 trimming audio. By increasing the duration, bursts of noises can be treated
4618 as silence and trimmed off. Default value is @code{0}.
4620 @item start_threshold
4621 This indicates what sample value should be treated as silence. For digital
4622 audio, a value of @code{0} may be fine but for audio recorded from analog,
4623 you may wish to increase the value to account for background noise.
4624 Can be specified in dB (in case "dB" is appended to the specified value)
4625 or amplitude ratio. Default value is @code{0}.
4628 Specify max duration of silence at beginning that will be kept after
4629 trimming. Default is 0, which is equal to trimming all samples detected
4633 Specify mode of detection of silence end in start of multi-channel audio.
4634 Can be @var{any} or @var{all}. Default is @var{any}.
4635 With @var{any}, any sample that is detected as non-silence will cause
4636 stopped trimming of silence.
4637 With @var{all}, only if all channels are detected as non-silence will cause
4638 stopped trimming of silence.
4641 Set the count for trimming silence from the end of audio.
4642 To remove silence from the middle of a file, specify a @var{stop_periods}
4643 that is negative. This value is then treated as a positive value and is
4644 used to indicate the effect should restart processing as specified by
4645 @var{start_periods}, making it suitable for removing periods of silence
4646 in the middle of the audio.
4647 Default value is @code{0}.
4650 Specify a duration of silence that must exist before audio is not copied any
4651 more. By specifying a higher duration, silence that is wanted can be left in
4653 Default value is @code{0}.
4655 @item stop_threshold
4656 This is the same as @option{start_threshold} but for trimming silence from
4658 Can be specified in dB (in case "dB" is appended to the specified value)
4659 or amplitude ratio. Default value is @code{0}.
4662 Specify max duration of silence at end that will be kept after
4663 trimming. Default is 0, which is equal to trimming all samples detected
4667 Specify mode of detection of silence start in end of multi-channel audio.
4668 Can be @var{any} or @var{all}. Default is @var{any}.
4669 With @var{any}, any sample that is detected as non-silence will cause
4670 stopped trimming of silence.
4671 With @var{all}, only if all channels are detected as non-silence will cause
4672 stopped trimming of silence.
4675 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4676 and works better with digital silence which is exactly 0.
4677 Default value is @code{rms}.
4680 Set duration in number of seconds used to calculate size of window in number
4681 of samples for detecting silence.
4682 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4685 @subsection Examples
4689 The following example shows how this filter can be used to start a recording
4690 that does not contain the delay at the start which usually occurs between
4691 pressing the record button and the start of the performance:
4693 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
4697 Trim all silence encountered from beginning to end where there is more than 1
4698 second of silence in audio:
4700 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
4706 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4707 loudspeakers around the user for binaural listening via headphones (audio
4708 formats up to 9 channels supported).
4709 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4710 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4711 Austrian Academy of Sciences.
4713 To enable compilation of this filter you need to configure FFmpeg with
4714 @code{--enable-libmysofa}.
4716 The filter accepts the following options:
4720 Set the SOFA file used for rendering.
4723 Set gain applied to audio. Value is in dB. Default is 0.
4726 Set rotation of virtual loudspeakers in deg. Default is 0.
4729 Set elevation of virtual speakers in deg. Default is 0.
4732 Set distance in meters between loudspeakers and the listener with near-field
4733 HRTFs. Default is 1.
4736 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4737 processing audio in time domain which is slow.
4738 @var{freq} is processing audio in frequency domain which is fast.
4739 Default is @var{freq}.
4742 Set custom positions of virtual loudspeakers. Syntax for this option is:
4743 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
4744 Each virtual loudspeaker is described with short channel name following with
4745 azimuth and elevation in degrees.
4746 Each virtual loudspeaker description is separated by '|'.
4747 For example to override front left and front right channel positions use:
4748 'speakers=FL 45 15|FR 345 15'.
4749 Descriptions with unrecognised channel names are ignored.
4752 Set custom gain for LFE channels. Value is in dB. Default is 0.
4755 Set custom frame size in number of samples. Default is 1024.
4756 Allowed range is from 1024 to 96000. Only used if option @samp{type}
4757 is set to @var{freq}.
4760 Should all IRs be normalized upon importing SOFA file.
4761 By default is enabled.
4764 Should nearest IRs be interpolated with neighbor IRs if exact position
4765 does not match. By default is disabled.
4768 Minphase all IRs upon loading of SOFA file. By default is disabled.
4771 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
4774 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
4777 @subsection Examples
4781 Using ClubFritz6 sofa file:
4783 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
4787 Using ClubFritz12 sofa file and bigger radius with small rotation:
4789 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
4793 Similar as above but with custom speaker positions for front left, front right, back left and back right
4794 and also with custom gain:
4796 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
4800 @section stereotools
4802 This filter has some handy utilities to manage stereo signals, for converting
4803 M/S stereo recordings to L/R signal while having control over the parameters
4804 or spreading the stereo image of master track.
4806 The filter accepts the following options:
4810 Set input level before filtering for both channels. Defaults is 1.
4811 Allowed range is from 0.015625 to 64.
4814 Set output level after filtering for both channels. Defaults is 1.
4815 Allowed range is from 0.015625 to 64.
4818 Set input balance between both channels. Default is 0.
4819 Allowed range is from -1 to 1.
4822 Set output balance between both channels. Default is 0.
4823 Allowed range is from -1 to 1.
4826 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
4827 clipping. Disabled by default.
4830 Mute the left channel. Disabled by default.
4833 Mute the right channel. Disabled by default.
4836 Change the phase of the left channel. Disabled by default.
4839 Change the phase of the right channel. Disabled by default.
4842 Set stereo mode. Available values are:
4846 Left/Right to Left/Right, this is default.
4849 Left/Right to Mid/Side.
4852 Mid/Side to Left/Right.
4855 Left/Right to Left/Left.
4858 Left/Right to Right/Right.
4861 Left/Right to Left + Right.
4864 Left/Right to Right/Left.
4867 Mid/Side to Left/Left.
4870 Mid/Side to Right/Right.
4874 Set level of side signal. Default is 1.
4875 Allowed range is from 0.015625 to 64.
4878 Set balance of side signal. Default is 0.
4879 Allowed range is from -1 to 1.
4882 Set level of the middle signal. Default is 1.
4883 Allowed range is from 0.015625 to 64.
4886 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
4889 Set stereo base between mono and inversed channels. Default is 0.
4890 Allowed range is from -1 to 1.
4893 Set delay in milliseconds how much to delay left from right channel and
4894 vice versa. Default is 0. Allowed range is from -20 to 20.
4897 Set S/C level. Default is 1. Allowed range is from 1 to 100.
4900 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
4902 @item bmode_in, bmode_out
4903 Set balance mode for balance_in/balance_out option.
4905 Can be one of the following:
4909 Classic balance mode. Attenuate one channel at time.
4910 Gain is raised up to 1.
4913 Similar as classic mode above but gain is raised up to 2.
4916 Equal power distribution, from -6dB to +6dB range.
4920 @subsection Examples
4924 Apply karaoke like effect:
4926 stereotools=mlev=0.015625
4930 Convert M/S signal to L/R:
4932 "stereotools=mode=ms>lr"
4936 @section stereowiden
4938 This filter enhance the stereo effect by suppressing signal common to both
4939 channels and by delaying the signal of left into right and vice versa,
4940 thereby widening the stereo effect.
4942 The filter accepts the following options:
4946 Time in milliseconds of the delay of left signal into right and vice versa.
4947 Default is 20 milliseconds.
4950 Amount of gain in delayed signal into right and vice versa. Gives a delay
4951 effect of left signal in right output and vice versa which gives widening
4952 effect. Default is 0.3.
4955 Cross feed of left into right with inverted phase. This helps in suppressing
4956 the mono. If the value is 1 it will cancel all the signal common to both
4957 channels. Default is 0.3.
4960 Set level of input signal of original channel. Default is 0.8.
4963 @section superequalizer
4964 Apply 18 band equalizer.
4966 The filter accepts the following options:
4973 Set 131Hz band gain.
4975 Set 185Hz band gain.
4977 Set 262Hz band gain.
4979 Set 370Hz band gain.
4981 Set 523Hz band gain.
4983 Set 740Hz band gain.
4985 Set 1047Hz band gain.
4987 Set 1480Hz band gain.
4989 Set 2093Hz band gain.
4991 Set 2960Hz band gain.
4993 Set 4186Hz band gain.
4995 Set 5920Hz band gain.
4997 Set 8372Hz band gain.
4999 Set 11840Hz band gain.
5001 Set 16744Hz band gain.
5003 Set 20000Hz band gain.
5007 Apply audio surround upmix filter.
5009 This filter allows to produce multichannel output from audio stream.
5011 The filter accepts the following options:
5015 Set output channel layout. By default, this is @var{5.1}.
5017 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5018 for the required syntax.
5021 Set input channel layout. By default, this is @var{stereo}.
5023 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5024 for the required syntax.
5027 Set input volume level. By default, this is @var{1}.
5030 Set output volume level. By default, this is @var{1}.
5033 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5036 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5039 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5042 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5043 In @var{add} mode, LFE channel is created from input audio and added to output.
5044 In @var{sub} mode, LFE channel is created from input audio and added to output but
5045 also all non-LFE output channels are subtracted with output LFE channel.
5048 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5049 Default is @var{90}.
5052 Set front center input volume. By default, this is @var{1}.
5055 Set front center output volume. By default, this is @var{1}.
5058 Set front left input volume. By default, this is @var{1}.
5061 Set front left output volume. By default, this is @var{1}.
5064 Set front right input volume. By default, this is @var{1}.
5067 Set front right output volume. By default, this is @var{1}.
5070 Set side left input volume. By default, this is @var{1}.
5073 Set side left output volume. By default, this is @var{1}.
5076 Set side right input volume. By default, this is @var{1}.
5079 Set side right output volume. By default, this is @var{1}.
5082 Set back left input volume. By default, this is @var{1}.
5085 Set back left output volume. By default, this is @var{1}.
5088 Set back right input volume. By default, this is @var{1}.
5091 Set back right output volume. By default, this is @var{1}.
5094 Set back center input volume. By default, this is @var{1}.
5097 Set back center output volume. By default, this is @var{1}.
5100 Set LFE input volume. By default, this is @var{1}.
5103 Set LFE output volume. By default, this is @var{1}.
5106 Set spread usage of stereo image across X axis for all channels.
5109 Set spread usage of stereo image across Y axis for all channels.
5111 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5112 Set spread usage of stereo image across X axis for each channel.
5114 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5115 Set spread usage of stereo image across Y axis for each channel.
5118 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5121 Set window function.
5123 It accepts the following values:
5146 Default is @code{hann}.
5149 Set window overlap. If set to 1, the recommended overlap for selected
5150 window function will be picked. Default is @code{0.5}.
5153 @section treble, highshelf
5155 Boost or cut treble (upper) frequencies of the audio using a two-pole
5156 shelving filter with a response similar to that of a standard
5157 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5159 The filter accepts the following options:
5163 Give the gain at whichever is the lower of ~22 kHz and the
5164 Nyquist frequency. Its useful range is about -20 (for a large cut)
5165 to +20 (for a large boost). Beware of clipping when using a positive gain.
5168 Set the filter's central frequency and so can be used
5169 to extend or reduce the frequency range to be boosted or cut.
5170 The default value is @code{3000} Hz.
5173 Set method to specify band-width of filter.
5188 Determine how steep is the filter's shelf transition.
5191 How much to use filtered signal in output. Default is 1.
5192 Range is between 0 and 1.
5195 Specify which channels to filter, by default all available are filtered.
5198 @subsection Commands
5200 This filter supports the following commands:
5203 Change treble frequency.
5204 Syntax for the command is : "@var{frequency}"
5207 Change treble width_type.
5208 Syntax for the command is : "@var{width_type}"
5211 Change treble width.
5212 Syntax for the command is : "@var{width}"
5216 Syntax for the command is : "@var{gain}"
5220 Syntax for the command is : "@var{mix}"
5225 Sinusoidal amplitude modulation.
5227 The filter accepts the following options:
5231 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5232 (20 Hz or lower) will result in a tremolo effect.
5233 This filter may also be used as a ring modulator by specifying
5234 a modulation frequency higher than 20 Hz.
5235 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5238 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5239 Default value is 0.5.
5244 Sinusoidal phase modulation.
5246 The filter accepts the following options:
5250 Modulation frequency in Hertz.
5251 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5254 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5255 Default value is 0.5.
5260 Adjust the input audio volume.
5262 It accepts the following parameters:
5266 Set audio volume expression.
5268 Output values are clipped to the maximum value.
5270 The output audio volume is given by the relation:
5272 @var{output_volume} = @var{volume} * @var{input_volume}
5275 The default value for @var{volume} is "1.0".
5278 This parameter represents the mathematical precision.
5280 It determines which input sample formats will be allowed, which affects the
5281 precision of the volume scaling.
5285 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5287 32-bit floating-point; this limits input sample format to FLT. (default)
5289 64-bit floating-point; this limits input sample format to DBL.
5293 Choose the behaviour on encountering ReplayGain side data in input frames.
5297 Remove ReplayGain side data, ignoring its contents (the default).
5300 Ignore ReplayGain side data, but leave it in the frame.
5303 Prefer the track gain, if present.
5306 Prefer the album gain, if present.
5309 @item replaygain_preamp
5310 Pre-amplification gain in dB to apply to the selected replaygain gain.
5312 Default value for @var{replaygain_preamp} is 0.0.
5315 Set when the volume expression is evaluated.
5317 It accepts the following values:
5320 only evaluate expression once during the filter initialization, or
5321 when the @samp{volume} command is sent
5324 evaluate expression for each incoming frame
5327 Default value is @samp{once}.
5330 The volume expression can contain the following parameters.
5334 frame number (starting at zero)
5337 @item nb_consumed_samples
5338 number of samples consumed by the filter
5340 number of samples in the current frame
5342 original frame position in the file
5348 PTS at start of stream
5350 time at start of stream
5356 last set volume value
5359 Note that when @option{eval} is set to @samp{once} only the
5360 @var{sample_rate} and @var{tb} variables are available, all other
5361 variables will evaluate to NAN.
5363 @subsection Commands
5365 This filter supports the following commands:
5368 Modify the volume expression.
5369 The command accepts the same syntax of the corresponding option.
5371 If the specified expression is not valid, it is kept at its current
5373 @item replaygain_noclip
5374 Prevent clipping by limiting the gain applied.
5376 Default value for @var{replaygain_noclip} is 1.
5380 @subsection Examples
5384 Halve the input audio volume:
5388 volume=volume=-6.0206dB
5391 In all the above example the named key for @option{volume} can be
5392 omitted, for example like in:
5398 Increase input audio power by 6 decibels using fixed-point precision:
5400 volume=volume=6dB:precision=fixed
5404 Fade volume after time 10 with an annihilation period of 5 seconds:
5406 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
5410 @section volumedetect
5412 Detect the volume of the input video.
5414 The filter has no parameters. The input is not modified. Statistics about
5415 the volume will be printed in the log when the input stream end is reached.
5417 In particular it will show the mean volume (root mean square), maximum
5418 volume (on a per-sample basis), and the beginning of a histogram of the
5419 registered volume values (from the maximum value to a cumulated 1/1000 of
5422 All volumes are in decibels relative to the maximum PCM value.
5424 @subsection Examples
5426 Here is an excerpt of the output:
5428 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
5429 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
5430 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
5431 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5432 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5433 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5434 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5435 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5436 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5442 The mean square energy is approximately -27 dB, or 10^-2.7.
5444 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5446 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5449 In other words, raising the volume by +4 dB does not cause any clipping,
5450 raising it by +5 dB causes clipping for 6 samples, etc.
5452 @c man end AUDIO FILTERS
5454 @chapter Audio Sources
5455 @c man begin AUDIO SOURCES
5457 Below is a description of the currently available audio sources.
5461 Buffer audio frames, and make them available to the filter chain.
5463 This source is mainly intended for a programmatic use, in particular
5464 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
5466 It accepts the following parameters:
5470 The timebase which will be used for timestamps of submitted frames. It must be
5471 either a floating-point number or in @var{numerator}/@var{denominator} form.
5474 The sample rate of the incoming audio buffers.
5477 The sample format of the incoming audio buffers.
5478 Either a sample format name or its corresponding integer representation from
5479 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5481 @item channel_layout
5482 The channel layout of the incoming audio buffers.
5483 Either a channel layout name from channel_layout_map in
5484 @file{libavutil/channel_layout.c} or its corresponding integer representation
5485 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5488 The number of channels of the incoming audio buffers.
5489 If both @var{channels} and @var{channel_layout} are specified, then they
5494 @subsection Examples
5497 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5500 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5501 Since the sample format with name "s16p" corresponds to the number
5502 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5505 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5510 Generate an audio signal specified by an expression.
5512 This source accepts in input one or more expressions (one for each
5513 channel), which are evaluated and used to generate a corresponding
5516 This source accepts the following options:
5520 Set the '|'-separated expressions list for each separate channel. In case the
5521 @option{channel_layout} option is not specified, the selected channel layout
5522 depends on the number of provided expressions. Otherwise the last
5523 specified expression is applied to the remaining output channels.
5525 @item channel_layout, c
5526 Set the channel layout. The number of channels in the specified layout
5527 must be equal to the number of specified expressions.
5530 Set the minimum duration of the sourced audio. See
5531 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5532 for the accepted syntax.
5533 Note that the resulting duration may be greater than the specified
5534 duration, as the generated audio is always cut at the end of a
5537 If not specified, or the expressed duration is negative, the audio is
5538 supposed to be generated forever.
5541 Set the number of samples per channel per each output frame,
5544 @item sample_rate, s
5545 Specify the sample rate, default to 44100.
5548 Each expression in @var{exprs} can contain the following constants:
5552 number of the evaluated sample, starting from 0
5555 time of the evaluated sample expressed in seconds, starting from 0
5562 @subsection Examples
5572 Generate a sin signal with frequency of 440 Hz, set sample rate to
5575 aevalsrc="sin(440*2*PI*t):s=8000"
5579 Generate a two channels signal, specify the channel layout (Front
5580 Center + Back Center) explicitly:
5582 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5586 Generate white noise:
5588 aevalsrc="-2+random(0)"
5592 Generate an amplitude modulated signal:
5594 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
5598 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
5600 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
5607 The null audio source, return unprocessed audio frames. It is mainly useful
5608 as a template and to be employed in analysis / debugging tools, or as
5609 the source for filters which ignore the input data (for example the sox
5612 This source accepts the following options:
5616 @item channel_layout, cl
5618 Specifies the channel layout, and can be either an integer or a string
5619 representing a channel layout. The default value of @var{channel_layout}
5622 Check the channel_layout_map definition in
5623 @file{libavutil/channel_layout.c} for the mapping between strings and
5624 channel layout values.
5626 @item sample_rate, r
5627 Specifies the sample rate, and defaults to 44100.
5630 Set the number of samples per requested frames.
5634 @subsection Examples
5638 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
5640 anullsrc=r=48000:cl=4
5644 Do the same operation with a more obvious syntax:
5646 anullsrc=r=48000:cl=mono
5650 All the parameters need to be explicitly defined.
5654 Synthesize a voice utterance using the libflite library.
5656 To enable compilation of this filter you need to configure FFmpeg with
5657 @code{--enable-libflite}.
5659 Note that versions of the flite library prior to 2.0 are not thread-safe.
5661 The filter accepts the following options:
5666 If set to 1, list the names of the available voices and exit
5667 immediately. Default value is 0.
5670 Set the maximum number of samples per frame. Default value is 512.
5673 Set the filename containing the text to speak.
5676 Set the text to speak.
5679 Set the voice to use for the speech synthesis. Default value is
5680 @code{kal}. See also the @var{list_voices} option.
5683 @subsection Examples
5687 Read from file @file{speech.txt}, and synthesize the text using the
5688 standard flite voice:
5690 flite=textfile=speech.txt
5694 Read the specified text selecting the @code{slt} voice:
5696 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5700 Input text to ffmpeg:
5702 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5706 Make @file{ffplay} speak the specified text, using @code{flite} and
5707 the @code{lavfi} device:
5709 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
5713 For more information about libflite, check:
5714 @url{http://www.festvox.org/flite/}
5718 Generate a noise audio signal.
5720 The filter accepts the following options:
5723 @item sample_rate, r
5724 Specify the sample rate. Default value is 48000 Hz.
5727 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
5731 Specify the duration of the generated audio stream. Not specifying this option
5732 results in noise with an infinite length.
5734 @item color, colour, c
5735 Specify the color of noise. Available noise colors are white, pink, brown,
5736 blue and violet. Default color is white.
5739 Specify a value used to seed the PRNG.
5742 Set the number of samples per each output frame, default is 1024.
5745 @subsection Examples
5750 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
5752 anoisesrc=d=60:c=pink:r=44100:a=0.5
5758 Generate odd-tap Hilbert transform FIR coefficients.
5760 The resulting stream can be used with @ref{afir} filter for phase-shifting
5761 the signal by 90 degrees.
5763 This is used in many matrix coding schemes and for analytic signal generation.
5764 The process is often written as a multiplication by i (or j), the imaginary unit.
5766 The filter accepts the following options:
5770 @item sample_rate, s
5771 Set sample rate, default is 44100.
5774 Set length of FIR filter, default is 22051.
5777 Set number of samples per each frame.
5780 Set window function to be used when generating FIR coefficients.
5785 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
5787 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
5789 The filter accepts the following options:
5792 @item sample_rate, r
5793 Set sample rate, default is 44100.
5796 Set number of samples per each frame. Default is 1024.
5799 Set high-pass frequency. Default is 0.
5802 Set low-pass frequency. Default is 0.
5803 If high-pass frequency is lower than low-pass frequency and low-pass frequency
5804 is higher than 0 then filter will create band-pass filter coefficients,
5805 otherwise band-reject filter coefficients.
5808 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
5811 Set Kaiser window beta.
5814 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
5817 Enable rounding, by default is disabled.
5820 Set number of taps for high-pass filter.
5823 Set number of taps for low-pass filter.
5828 Generate an audio signal made of a sine wave with amplitude 1/8.
5830 The audio signal is bit-exact.
5832 The filter accepts the following options:
5837 Set the carrier frequency. Default is 440 Hz.
5839 @item beep_factor, b
5840 Enable a periodic beep every second with frequency @var{beep_factor} times
5841 the carrier frequency. Default is 0, meaning the beep is disabled.
5843 @item sample_rate, r
5844 Specify the sample rate, default is 44100.
5847 Specify the duration of the generated audio stream.
5849 @item samples_per_frame
5850 Set the number of samples per output frame.
5852 The expression can contain the following constants:
5856 The (sequential) number of the output audio frame, starting from 0.
5859 The PTS (Presentation TimeStamp) of the output audio frame,
5860 expressed in @var{TB} units.
5863 The PTS of the output audio frame, expressed in seconds.
5866 The timebase of the output audio frames.
5869 Default is @code{1024}.
5872 @subsection Examples
5877 Generate a simple 440 Hz sine wave:
5883 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
5887 sine=frequency=220:beep_factor=4:duration=5
5891 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
5894 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
5898 @c man end AUDIO SOURCES
5900 @chapter Audio Sinks
5901 @c man begin AUDIO SINKS
5903 Below is a description of the currently available audio sinks.
5905 @section abuffersink
5907 Buffer audio frames, and make them available to the end of filter chain.
5909 This sink is mainly intended for programmatic use, in particular
5910 through the interface defined in @file{libavfilter/buffersink.h}
5911 or the options system.
5913 It accepts a pointer to an AVABufferSinkContext structure, which
5914 defines the incoming buffers' formats, to be passed as the opaque
5915 parameter to @code{avfilter_init_filter} for initialization.
5918 Null audio sink; do absolutely nothing with the input audio. It is
5919 mainly useful as a template and for use in analysis / debugging
5922 @c man end AUDIO SINKS
5924 @chapter Video Filters
5925 @c man begin VIDEO FILTERS
5927 When you configure your FFmpeg build, you can disable any of the
5928 existing filters using @code{--disable-filters}.
5929 The configure output will show the video filters included in your
5932 Below is a description of the currently available video filters.
5934 @section alphaextract
5936 Extract the alpha component from the input as a grayscale video. This
5937 is especially useful with the @var{alphamerge} filter.
5941 Add or replace the alpha component of the primary input with the
5942 grayscale value of a second input. This is intended for use with
5943 @var{alphaextract} to allow the transmission or storage of frame
5944 sequences that have alpha in a format that doesn't support an alpha
5947 For example, to reconstruct full frames from a normal YUV-encoded video
5948 and a separate video created with @var{alphaextract}, you might use:
5950 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
5953 Since this filter is designed for reconstruction, it operates on frame
5954 sequences without considering timestamps, and terminates when either
5955 input reaches end of stream. This will cause problems if your encoding
5956 pipeline drops frames. If you're trying to apply an image as an
5957 overlay to a video stream, consider the @var{overlay} filter instead.
5961 Amplify differences between current pixel and pixels of adjacent frames in
5962 same pixel location.
5964 This filter accepts the following options:
5968 Set frame radius. Default is 2. Allowed range is from 1 to 63.
5969 For example radius of 3 will instruct filter to calculate average of 7 frames.
5972 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
5975 Set threshold for difference amplification. Any difference greater or equal to
5976 this value will not alter source pixel. Default is 10.
5977 Allowed range is from 0 to 65535.
5980 Set tolerance for difference amplification. Any difference lower to
5981 this value will not alter source pixel. Default is 0.
5982 Allowed range is from 0 to 65535.
5985 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
5986 This option controls maximum possible value that will decrease source pixel value.
5989 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
5990 This option controls maximum possible value that will increase source pixel value.
5993 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
5998 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
5999 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6000 Substation Alpha) subtitles files.
6002 This filter accepts the following option in addition to the common options from
6003 the @ref{subtitles} filter:
6007 Set the shaping engine
6009 Available values are:
6012 The default libass shaping engine, which is the best available.
6014 Fast, font-agnostic shaper that can do only substitutions
6016 Slower shaper using OpenType for substitutions and positioning
6019 The default is @code{auto}.
6023 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6025 The filter accepts the following options:
6029 Set threshold A for 1st plane. Default is 0.02.
6030 Valid range is 0 to 0.3.
6033 Set threshold B for 1st plane. Default is 0.04.
6034 Valid range is 0 to 5.
6037 Set threshold A for 2nd plane. Default is 0.02.
6038 Valid range is 0 to 0.3.
6041 Set threshold B for 2nd plane. Default is 0.04.
6042 Valid range is 0 to 5.
6045 Set threshold A for 3rd plane. Default is 0.02.
6046 Valid range is 0 to 0.3.
6049 Set threshold B for 3rd plane. Default is 0.04.
6050 Valid range is 0 to 5.
6052 Threshold A is designed to react on abrupt changes in the input signal and
6053 threshold B is designed to react on continuous changes in the input signal.
6056 Set number of frames filter will use for averaging. Default is 9. Must be odd
6057 number in range [5, 129].
6060 Set what planes of frame filter will use for averaging. Default is all.
6065 Apply average blur filter.
6067 The filter accepts the following options:
6071 Set horizontal radius size.
6074 Set which planes to filter. By default all planes are filtered.
6077 Set vertical radius size, if zero it will be same as @code{sizeX}.
6078 Default is @code{0}.
6083 Compute the bounding box for the non-black pixels in the input frame
6086 This filter computes the bounding box containing all the pixels with a
6087 luminance value greater than the minimum allowed value.
6088 The parameters describing the bounding box are printed on the filter
6091 The filter accepts the following option:
6095 Set the minimal luminance value. Default is @code{16}.
6098 @section bitplanenoise
6100 Show and measure bit plane noise.
6102 The filter accepts the following options:
6106 Set which plane to analyze. Default is @code{1}.
6109 Filter out noisy pixels from @code{bitplane} set above.
6110 Default is disabled.
6113 @section blackdetect
6115 Detect video intervals that are (almost) completely black. Can be
6116 useful to detect chapter transitions, commercials, or invalid
6117 recordings. Output lines contains the time for the start, end and
6118 duration of the detected black interval expressed in seconds.
6120 In order to display the output lines, you need to set the loglevel at
6121 least to the AV_LOG_INFO value.
6123 The filter accepts the following options:
6126 @item black_min_duration, d
6127 Set the minimum detected black duration expressed in seconds. It must
6128 be a non-negative floating point number.
6130 Default value is 2.0.
6132 @item picture_black_ratio_th, pic_th
6133 Set the threshold for considering a picture "black".
6134 Express the minimum value for the ratio:
6136 @var{nb_black_pixels} / @var{nb_pixels}
6139 for which a picture is considered black.
6140 Default value is 0.98.
6142 @item pixel_black_th, pix_th
6143 Set the threshold for considering a pixel "black".
6145 The threshold expresses the maximum pixel luminance value for which a
6146 pixel is considered "black". The provided value is scaled according to
6147 the following equation:
6149 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6152 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6153 the input video format, the range is [0-255] for YUV full-range
6154 formats and [16-235] for YUV non full-range formats.
6156 Default value is 0.10.
6159 The following example sets the maximum pixel threshold to the minimum
6160 value, and detects only black intervals of 2 or more seconds:
6162 blackdetect=d=2:pix_th=0.00
6167 Detect frames that are (almost) completely black. Can be useful to
6168 detect chapter transitions or commercials. Output lines consist of
6169 the frame number of the detected frame, the percentage of blackness,
6170 the position in the file if known or -1 and the timestamp in seconds.
6172 In order to display the output lines, you need to set the loglevel at
6173 least to the AV_LOG_INFO value.
6175 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
6176 The value represents the percentage of pixels in the picture that
6177 are below the threshold value.
6179 It accepts the following parameters:
6184 The percentage of the pixels that have to be below the threshold; it defaults to
6187 @item threshold, thresh
6188 The threshold below which a pixel value is considered black; it defaults to
6193 @section blend, tblend
6195 Blend two video frames into each other.
6197 The @code{blend} filter takes two input streams and outputs one
6198 stream, the first input is the "top" layer and second input is
6199 "bottom" layer. By default, the output terminates when the longest input terminates.
6201 The @code{tblend} (time blend) filter takes two consecutive frames
6202 from one single stream, and outputs the result obtained by blending
6203 the new frame on top of the old frame.
6205 A description of the accepted options follows.
6213 Set blend mode for specific pixel component or all pixel components in case
6214 of @var{all_mode}. Default value is @code{normal}.
6216 Available values for component modes are:
6258 Set blend opacity for specific pixel component or all pixel components in case
6259 of @var{all_opacity}. Only used in combination with pixel component blend modes.
6266 Set blend expression for specific pixel component or all pixel components in case
6267 of @var{all_expr}. Note that related mode options will be ignored if those are set.
6269 The expressions can use the following variables:
6273 The sequential number of the filtered frame, starting from @code{0}.
6277 the coordinates of the current sample
6281 the width and height of currently filtered plane
6285 Width and height scale for the plane being filtered. It is the
6286 ratio between the dimensions of the current plane to the luma plane,
6287 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
6288 the luma plane and @code{0.5,0.5} for the chroma planes.
6291 Time of the current frame, expressed in seconds.
6294 Value of pixel component at current location for first video frame (top layer).
6297 Value of pixel component at current location for second video frame (bottom layer).
6301 The @code{blend} filter also supports the @ref{framesync} options.
6303 @subsection Examples
6307 Apply transition from bottom layer to top layer in first 10 seconds:
6309 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
6313 Apply linear horizontal transition from top layer to bottom layer:
6315 blend=all_expr='A*(X/W)+B*(1-X/W)'
6319 Apply 1x1 checkerboard effect:
6321 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
6325 Apply uncover left effect:
6327 blend=all_expr='if(gte(N*SW+X,W),A,B)'
6331 Apply uncover down effect:
6333 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
6337 Apply uncover up-left effect:
6339 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
6343 Split diagonally video and shows top and bottom layer on each side:
6345 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
6349 Display differences between the current and the previous frame:
6351 tblend=all_mode=grainextract
6357 Denoise frames using Block-Matching 3D algorithm.
6359 The filter accepts the following options.
6363 Set denoising strength. Default value is 1.
6364 Allowed range is from 0 to 999.9.
6365 The denoising algorithm is very sensitive to sigma, so adjust it
6366 according to the source.
6369 Set local patch size. This sets dimensions in 2D.
6372 Set sliding step for processing blocks. Default value is 4.
6373 Allowed range is from 1 to 64.
6374 Smaller values allows processing more reference blocks and is slower.
6377 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
6378 When set to 1, no block matching is done. Larger values allows more blocks
6380 Allowed range is from 1 to 256.
6383 Set radius for search block matching. Default is 9.
6384 Allowed range is from 1 to INT32_MAX.
6387 Set step between two search locations for block matching. Default is 1.
6388 Allowed range is from 1 to 64. Smaller is slower.
6391 Set threshold of mean square error for block matching. Valid range is 0 to
6395 Set thresholding parameter for hard thresholding in 3D transformed domain.
6396 Larger values results in stronger hard-thresholding filtering in frequency
6400 Set filtering estimation mode. Can be @code{basic} or @code{final}.
6401 Default is @code{basic}.
6404 If enabled, filter will use 2nd stream for block matching.
6405 Default is disabled for @code{basic} value of @var{estim} option,
6406 and always enabled if value of @var{estim} is @code{final}.
6409 Set planes to filter. Default is all available except alpha.
6412 @subsection Examples
6416 Basic filtering with bm3d:
6418 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
6422 Same as above, but filtering only luma:
6424 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
6428 Same as above, but with both estimation modes:
6430 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
6434 Same as above, but prefilter with @ref{nlmeans} filter instead:
6436 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
6442 Apply a boxblur algorithm to the input video.
6444 It accepts the following parameters:
6448 @item luma_radius, lr
6449 @item luma_power, lp
6450 @item chroma_radius, cr
6451 @item chroma_power, cp
6452 @item alpha_radius, ar
6453 @item alpha_power, ap
6457 A description of the accepted options follows.
6460 @item luma_radius, lr
6461 @item chroma_radius, cr
6462 @item alpha_radius, ar
6463 Set an expression for the box radius in pixels used for blurring the
6464 corresponding input plane.
6466 The radius value must be a non-negative number, and must not be
6467 greater than the value of the expression @code{min(w,h)/2} for the
6468 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
6471 Default value for @option{luma_radius} is "2". If not specified,
6472 @option{chroma_radius} and @option{alpha_radius} default to the
6473 corresponding value set for @option{luma_radius}.
6475 The expressions can contain the following constants:
6479 The input width and height in pixels.
6483 The input chroma image width and height in pixels.
6487 The horizontal and vertical chroma subsample values. For example, for the
6488 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
6491 @item luma_power, lp
6492 @item chroma_power, cp
6493 @item alpha_power, ap
6494 Specify how many times the boxblur filter is applied to the
6495 corresponding plane.
6497 Default value for @option{luma_power} is 2. If not specified,
6498 @option{chroma_power} and @option{alpha_power} default to the
6499 corresponding value set for @option{luma_power}.
6501 A value of 0 will disable the effect.
6504 @subsection Examples
6508 Apply a boxblur filter with the luma, chroma, and alpha radii
6511 boxblur=luma_radius=2:luma_power=1
6516 Set the luma radius to 2, and alpha and chroma radius to 0:
6518 boxblur=2:1:cr=0:ar=0
6522 Set the luma and chroma radii to a fraction of the video dimension:
6524 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
6530 Deinterlace the input video ("bwdif" stands for "Bob Weaver
6531 Deinterlacing Filter").
6533 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
6534 interpolation algorithms.
6535 It accepts the following parameters:
6539 The interlacing mode to adopt. It accepts one of the following values:
6543 Output one frame for each frame.
6545 Output one frame for each field.
6548 The default value is @code{send_field}.
6551 The picture field parity assumed for the input interlaced video. It accepts one
6552 of the following values:
6556 Assume the top field is first.
6558 Assume the bottom field is first.
6560 Enable automatic detection of field parity.
6563 The default value is @code{auto}.
6564 If the interlacing is unknown or the decoder does not export this information,
6565 top field first will be assumed.
6568 Specify which frames to deinterlace. Accept one of the following
6573 Deinterlace all frames.
6575 Only deinterlace frames marked as interlaced.
6578 The default value is @code{all}.
6582 Remove all color information for all colors except for certain one.
6584 The filter accepts the following options:
6588 The color which will not be replaced with neutral chroma.
6591 Similarity percentage with the above color.
6592 0.01 matches only the exact key color, while 1.0 matches everything.
6596 0.0 makes pixels either fully gray, or not gray at all.
6597 Higher values result in more preserved color.
6600 Signals that the color passed is already in YUV instead of RGB.
6602 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6603 This can be used to pass exact YUV values as hexadecimal numbers.
6607 YUV colorspace color/chroma keying.
6609 The filter accepts the following options:
6613 The color which will be replaced with transparency.
6616 Similarity percentage with the key color.
6618 0.01 matches only the exact key color, while 1.0 matches everything.
6623 0.0 makes pixels either fully transparent, or not transparent at all.
6625 Higher values result in semi-transparent pixels, with a higher transparency
6626 the more similar the pixels color is to the key color.
6629 Signals that the color passed is already in YUV instead of RGB.
6631 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6632 This can be used to pass exact YUV values as hexadecimal numbers.
6635 @subsection Examples
6639 Make every green pixel in the input image transparent:
6641 ffmpeg -i input.png -vf chromakey=green out.png
6645 Overlay a greenscreen-video on top of a static black background.
6647 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
6651 @section chromashift
6652 Shift chroma pixels horizontally and/or vertically.
6654 The filter accepts the following options:
6657 Set amount to shift chroma-blue horizontally.
6659 Set amount to shift chroma-blue vertically.
6661 Set amount to shift chroma-red horizontally.
6663 Set amount to shift chroma-red vertically.
6665 Set edge mode, can be @var{smear}, default, or @var{warp}.
6670 Display CIE color diagram with pixels overlaid onto it.
6672 The filter accepts the following options:
6687 @item uhdtv, rec2020
6700 Set what gamuts to draw.
6702 See @code{system} option for available values.
6705 Set ciescope size, by default set to 512.
6708 Set intensity used to map input pixel values to CIE diagram.
6711 Set contrast used to draw tongue colors that are out of active color system gamut.
6714 Correct gamma displayed on scope, by default enabled.
6717 Show white point on CIE diagram, by default disabled.
6720 Set input gamma. Used only with XYZ input color space.
6725 Visualize information exported by some codecs.
6727 Some codecs can export information through frames using side-data or other
6728 means. For example, some MPEG based codecs export motion vectors through the
6729 @var{export_mvs} flag in the codec @option{flags2} option.
6731 The filter accepts the following option:
6735 Set motion vectors to visualize.
6737 Available flags for @var{mv} are:
6741 forward predicted MVs of P-frames
6743 forward predicted MVs of B-frames
6745 backward predicted MVs of B-frames
6749 Display quantization parameters using the chroma planes.
6752 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
6754 Available flags for @var{mv_type} are:
6758 forward predicted MVs
6760 backward predicted MVs
6763 @item frame_type, ft
6764 Set frame type to visualize motion vectors of.
6766 Available flags for @var{frame_type} are:
6770 intra-coded frames (I-frames)
6772 predicted frames (P-frames)
6774 bi-directionally predicted frames (B-frames)
6778 @subsection Examples
6782 Visualize forward predicted MVs of all frames using @command{ffplay}:
6784 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
6788 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
6790 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
6794 @section colorbalance
6795 Modify intensity of primary colors (red, green and blue) of input frames.
6797 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
6798 regions for the red-cyan, green-magenta or blue-yellow balance.
6800 A positive adjustment value shifts the balance towards the primary color, a negative
6801 value towards the complementary color.
6803 The filter accepts the following options:
6809 Adjust red, green and blue shadows (darkest pixels).
6814 Adjust red, green and blue midtones (medium pixels).
6819 Adjust red, green and blue highlights (brightest pixels).
6821 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
6824 @subsection Examples
6828 Add red color cast to shadows:
6835 RGB colorspace color keying.
6837 The filter accepts the following options:
6841 The color which will be replaced with transparency.
6844 Similarity percentage with the key color.
6846 0.01 matches only the exact key color, while 1.0 matches everything.
6851 0.0 makes pixels either fully transparent, or not transparent at all.
6853 Higher values result in semi-transparent pixels, with a higher transparency
6854 the more similar the pixels color is to the key color.
6857 @subsection Examples
6861 Make every green pixel in the input image transparent:
6863 ffmpeg -i input.png -vf colorkey=green out.png
6867 Overlay a greenscreen-video on top of a static background image.
6869 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
6874 Remove all color information for all RGB colors except for certain one.
6876 The filter accepts the following options:
6880 The color which will not be replaced with neutral gray.
6883 Similarity percentage with the above color.
6884 0.01 matches only the exact key color, while 1.0 matches everything.
6887 Blend percentage. 0.0 makes pixels fully gray.
6888 Higher values result in more preserved color.
6891 @section colorlevels
6893 Adjust video input frames using levels.
6895 The filter accepts the following options:
6902 Adjust red, green, blue and alpha input black point.
6903 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
6909 Adjust red, green, blue and alpha input white point.
6910 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
6912 Input levels are used to lighten highlights (bright tones), darken shadows
6913 (dark tones), change the balance of bright and dark tones.
6919 Adjust red, green, blue and alpha output black point.
6920 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
6926 Adjust red, green, blue and alpha output white point.
6927 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
6929 Output levels allows manual selection of a constrained output level range.
6932 @subsection Examples
6936 Make video output darker:
6938 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
6944 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
6948 Make video output lighter:
6950 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
6954 Increase brightness:
6956 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
6960 @section colorchannelmixer
6962 Adjust video input frames by re-mixing color channels.
6964 This filter modifies a color channel by adding the values associated to
6965 the other channels of the same pixels. For example if the value to
6966 modify is red, the output value will be:
6968 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
6971 The filter accepts the following options:
6978 Adjust contribution of input red, green, blue and alpha channels for output red channel.
6979 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
6985 Adjust contribution of input red, green, blue and alpha channels for output green channel.
6986 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
6992 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
6993 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
6999 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7000 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7002 Allowed ranges for options are @code{[-2.0, 2.0]}.
7005 @subsection Examples
7009 Convert source to grayscale:
7011 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7014 Simulate sepia tones:
7016 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7020 @section colormatrix
7022 Convert color matrix.
7024 The filter accepts the following options:
7029 Specify the source and destination color matrix. Both values must be
7032 The accepted values are:
7060 For example to convert from BT.601 to SMPTE-240M, use the command:
7062 colormatrix=bt601:smpte240m
7067 Convert colorspace, transfer characteristics or color primaries.
7068 Input video needs to have an even size.
7070 The filter accepts the following options:
7075 Specify all color properties at once.
7077 The accepted values are:
7107 Specify output colorspace.
7109 The accepted values are:
7118 BT.470BG or BT.601-6 625
7121 SMPTE-170M or BT.601-6 525
7130 BT.2020 with non-constant luminance
7136 Specify output transfer characteristics.
7138 The accepted values are:
7150 Constant gamma of 2.2
7153 Constant gamma of 2.8
7156 SMPTE-170M, BT.601-6 625 or BT.601-6 525
7174 BT.2020 for 10-bits content
7177 BT.2020 for 12-bits content
7183 Specify output color primaries.
7185 The accepted values are:
7194 BT.470BG or BT.601-6 625
7197 SMPTE-170M or BT.601-6 525
7221 Specify output color range.
7223 The accepted values are:
7226 TV (restricted) range
7229 MPEG (restricted) range
7240 Specify output color format.
7242 The accepted values are:
7245 YUV 4:2:0 planar 8-bits
7248 YUV 4:2:0 planar 10-bits
7251 YUV 4:2:0 planar 12-bits
7254 YUV 4:2:2 planar 8-bits
7257 YUV 4:2:2 planar 10-bits
7260 YUV 4:2:2 planar 12-bits
7263 YUV 4:4:4 planar 8-bits
7266 YUV 4:4:4 planar 10-bits
7269 YUV 4:4:4 planar 12-bits
7274 Do a fast conversion, which skips gamma/primary correction. This will take
7275 significantly less CPU, but will be mathematically incorrect. To get output
7276 compatible with that produced by the colormatrix filter, use fast=1.
7279 Specify dithering mode.
7281 The accepted values are:
7287 Floyd-Steinberg dithering
7291 Whitepoint adaptation mode.
7293 The accepted values are:
7296 Bradford whitepoint adaptation
7299 von Kries whitepoint adaptation
7302 identity whitepoint adaptation (i.e. no whitepoint adaptation)
7306 Override all input properties at once. Same accepted values as @ref{all}.
7309 Override input colorspace. Same accepted values as @ref{space}.
7312 Override input color primaries. Same accepted values as @ref{primaries}.
7315 Override input transfer characteristics. Same accepted values as @ref{trc}.
7318 Override input color range. Same accepted values as @ref{range}.
7322 The filter converts the transfer characteristics, color space and color
7323 primaries to the specified user values. The output value, if not specified,
7324 is set to a default value based on the "all" property. If that property is
7325 also not specified, the filter will log an error. The output color range and
7326 format default to the same value as the input color range and format. The
7327 input transfer characteristics, color space, color primaries and color range
7328 should be set on the input data. If any of these are missing, the filter will
7329 log an error and no conversion will take place.
7331 For example to convert the input to SMPTE-240M, use the command:
7333 colorspace=smpte240m
7336 @section convolution
7338 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
7340 The filter accepts the following options:
7347 Set matrix for each plane.
7348 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
7349 and from 1 to 49 odd number of signed integers in @var{row} mode.
7355 Set multiplier for calculated value for each plane.
7356 If unset or 0, it will be sum of all matrix elements.
7362 Set bias for each plane. This value is added to the result of the multiplication.
7363 Useful for making the overall image brighter or darker. Default is 0.0.
7369 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
7370 Default is @var{square}.
7373 @subsection Examples
7379 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"
7385 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"
7391 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"
7397 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"
7401 Apply laplacian edge detector which includes diagonals:
7403 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"
7409 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"
7415 Apply 2D convolution of video stream in frequency domain using second stream
7418 The filter accepts the following options:
7422 Set which planes to process.
7425 Set which impulse video frames will be processed, can be @var{first}
7426 or @var{all}. Default is @var{all}.
7429 The @code{convolve} filter also supports the @ref{framesync} options.
7433 Copy the input video source unchanged to the output. This is mainly useful for
7438 Video filtering on GPU using Apple's CoreImage API on OSX.
7440 Hardware acceleration is based on an OpenGL context. Usually, this means it is
7441 processed by video hardware. However, software-based OpenGL implementations
7442 exist which means there is no guarantee for hardware processing. It depends on
7445 There are many filters and image generators provided by Apple that come with a
7446 large variety of options. The filter has to be referenced by its name along
7449 The coreimage filter accepts the following options:
7452 List all available filters and generators along with all their respective
7453 options as well as possible minimum and maximum values along with the default
7460 Specify all filters by their respective name and options.
7461 Use @var{list_filters} to determine all valid filter names and options.
7462 Numerical options are specified by a float value and are automatically clamped
7463 to their respective value range. Vector and color options have to be specified
7464 by a list of space separated float values. Character escaping has to be done.
7465 A special option name @code{default} is available to use default options for a
7468 It is required to specify either @code{default} or at least one of the filter options.
7469 All omitted options are used with their default values.
7470 The syntax of the filter string is as follows:
7472 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
7476 Specify a rectangle where the output of the filter chain is copied into the
7477 input image. It is given by a list of space separated float values:
7479 output_rect=x\ y\ width\ height
7481 If not given, the output rectangle equals the dimensions of the input image.
7482 The output rectangle is automatically cropped at the borders of the input
7483 image. Negative values are valid for each component.
7485 output_rect=25\ 25\ 100\ 100
7489 Several filters can be chained for successive processing without GPU-HOST
7490 transfers allowing for fast processing of complex filter chains.
7491 Currently, only filters with zero (generators) or exactly one (filters) input
7492 image and one output image are supported. Also, transition filters are not yet
7495 Some filters generate output images with additional padding depending on the
7496 respective filter kernel. The padding is automatically removed to ensure the
7497 filter output has the same size as the input image.
7499 For image generators, the size of the output image is determined by the
7500 previous output image of the filter chain or the input image of the whole
7501 filterchain, respectively. The generators do not use the pixel information of
7502 this image to generate their output. However, the generated output is
7503 blended onto this image, resulting in partial or complete coverage of the
7506 The @ref{coreimagesrc} video source can be used for generating input images
7507 which are directly fed into the filter chain. By using it, providing input
7508 images by another video source or an input video is not required.
7510 @subsection Examples
7515 List all filters available:
7517 coreimage=list_filters=true
7521 Use the CIBoxBlur filter with default options to blur an image:
7523 coreimage=filter=CIBoxBlur@@default
7527 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
7528 its center at 100x100 and a radius of 50 pixels:
7530 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
7534 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
7535 given as complete and escaped command-line for Apple's standard bash shell:
7537 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
7543 Crop the input video to given dimensions.
7545 It accepts the following parameters:
7549 The width of the output video. It defaults to @code{iw}.
7550 This expression is evaluated only once during the filter
7551 configuration, or when the @samp{w} or @samp{out_w} command is sent.
7554 The height of the output video. It defaults to @code{ih}.
7555 This expression is evaluated only once during the filter
7556 configuration, or when the @samp{h} or @samp{out_h} command is sent.
7559 The horizontal position, in the input video, of the left edge of the output
7560 video. It defaults to @code{(in_w-out_w)/2}.
7561 This expression is evaluated per-frame.
7564 The vertical position, in the input video, of the top edge of the output video.
7565 It defaults to @code{(in_h-out_h)/2}.
7566 This expression is evaluated per-frame.
7569 If set to 1 will force the output display aspect ratio
7570 to be the same of the input, by changing the output sample aspect
7571 ratio. It defaults to 0.
7574 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
7575 width/height/x/y as specified and will not be rounded to nearest smaller value.
7579 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
7580 expressions containing the following constants:
7585 The computed values for @var{x} and @var{y}. They are evaluated for
7590 The input width and height.
7594 These are the same as @var{in_w} and @var{in_h}.
7598 The output (cropped) width and height.
7602 These are the same as @var{out_w} and @var{out_h}.
7605 same as @var{iw} / @var{ih}
7608 input sample aspect ratio
7611 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
7615 horizontal and vertical chroma subsample values. For example for the
7616 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
7619 The number of the input frame, starting from 0.
7622 the position in the file of the input frame, NAN if unknown
7625 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
7629 The expression for @var{out_w} may depend on the value of @var{out_h},
7630 and the expression for @var{out_h} may depend on @var{out_w}, but they
7631 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
7632 evaluated after @var{out_w} and @var{out_h}.
7634 The @var{x} and @var{y} parameters specify the expressions for the
7635 position of the top-left corner of the output (non-cropped) area. They
7636 are evaluated for each frame. If the evaluated value is not valid, it
7637 is approximated to the nearest valid value.
7639 The expression for @var{x} may depend on @var{y}, and the expression
7640 for @var{y} may depend on @var{x}.
7642 @subsection Examples
7646 Crop area with size 100x100 at position (12,34).
7651 Using named options, the example above becomes:
7653 crop=w=100:h=100:x=12:y=34
7657 Crop the central input area with size 100x100:
7663 Crop the central input area with size 2/3 of the input video:
7665 crop=2/3*in_w:2/3*in_h
7669 Crop the input video central square:
7676 Delimit the rectangle with the top-left corner placed at position
7677 100:100 and the right-bottom corner corresponding to the right-bottom
7678 corner of the input image.
7680 crop=in_w-100:in_h-100:100:100
7684 Crop 10 pixels from the left and right borders, and 20 pixels from
7685 the top and bottom borders
7687 crop=in_w-2*10:in_h-2*20
7691 Keep only the bottom right quarter of the input image:
7693 crop=in_w/2:in_h/2:in_w/2:in_h/2
7697 Crop height for getting Greek harmony:
7699 crop=in_w:1/PHI*in_w
7703 Apply trembling effect:
7705 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)
7709 Apply erratic camera effect depending on timestamp:
7711 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)"
7715 Set x depending on the value of y:
7717 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
7721 @subsection Commands
7723 This filter supports the following commands:
7729 Set width/height of the output video and the horizontal/vertical position
7731 The command accepts the same syntax of the corresponding option.
7733 If the specified expression is not valid, it is kept at its current
7739 Auto-detect the crop size.
7741 It calculates the necessary cropping parameters and prints the
7742 recommended parameters via the logging system. The detected dimensions
7743 correspond to the non-black area of the input video.
7745 It accepts the following parameters:
7750 Set higher black value threshold, which can be optionally specified
7751 from nothing (0) to everything (255 for 8-bit based formats). An intensity
7752 value greater to the set value is considered non-black. It defaults to 24.
7753 You can also specify a value between 0.0 and 1.0 which will be scaled depending
7754 on the bitdepth of the pixel format.
7757 The value which the width/height should be divisible by. It defaults to
7758 16. The offset is automatically adjusted to center the video. Use 2 to
7759 get only even dimensions (needed for 4:2:2 video). 16 is best when
7760 encoding to most video codecs.
7762 @item reset_count, reset
7763 Set the counter that determines after how many frames cropdetect will
7764 reset the previously detected largest video area and start over to
7765 detect the current optimal crop area. Default value is 0.
7767 This can be useful when channel logos distort the video area. 0
7768 indicates 'never reset', and returns the largest area encountered during
7775 Delay video filtering until a given wallclock timestamp. The filter first
7776 passes on @option{preroll} amount of frames, then it buffers at most
7777 @option{buffer} amount of frames and waits for the cue. After reaching the cue
7778 it forwards the buffered frames and also any subsequent frames coming in its
7781 The filter can be used synchronize the output of multiple ffmpeg processes for
7782 realtime output devices like decklink. By putting the delay in the filtering
7783 chain and pre-buffering frames the process can pass on data to output almost
7784 immediately after the target wallclock timestamp is reached.
7786 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
7792 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
7795 The duration of content to pass on as preroll expressed in seconds. Default is 0.
7798 The maximum duration of content to buffer before waiting for the cue expressed
7799 in seconds. Default is 0.
7806 Apply color adjustments using curves.
7808 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
7809 component (red, green and blue) has its values defined by @var{N} key points
7810 tied from each other using a smooth curve. The x-axis represents the pixel
7811 values from the input frame, and the y-axis the new pixel values to be set for
7814 By default, a component curve is defined by the two points @var{(0;0)} and
7815 @var{(1;1)}. This creates a straight line where each original pixel value is
7816 "adjusted" to its own value, which means no change to the image.
7818 The filter allows you to redefine these two points and add some more. A new
7819 curve (using a natural cubic spline interpolation) will be define to pass
7820 smoothly through all these new coordinates. The new defined points needs to be
7821 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
7822 be in the @var{[0;1]} interval. If the computed curves happened to go outside
7823 the vector spaces, the values will be clipped accordingly.
7825 The filter accepts the following options:
7829 Select one of the available color presets. This option can be used in addition
7830 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
7831 options takes priority on the preset values.
7832 Available presets are:
7835 @item color_negative
7838 @item increase_contrast
7840 @item linear_contrast
7841 @item medium_contrast
7843 @item strong_contrast
7846 Default is @code{none}.
7848 Set the master key points. These points will define a second pass mapping. It
7849 is sometimes called a "luminance" or "value" mapping. It can be used with
7850 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
7851 post-processing LUT.
7853 Set the key points for the red component.
7855 Set the key points for the green component.
7857 Set the key points for the blue component.
7859 Set the key points for all components (not including master).
7860 Can be used in addition to the other key points component
7861 options. In this case, the unset component(s) will fallback on this
7862 @option{all} setting.
7864 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
7866 Save Gnuplot script of the curves in specified file.
7869 To avoid some filtergraph syntax conflicts, each key points list need to be
7870 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
7872 @subsection Examples
7876 Increase slightly the middle level of blue:
7878 curves=blue='0/0 0.5/0.58 1/1'
7884 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'
7886 Here we obtain the following coordinates for each components:
7889 @code{(0;0.11) (0.42;0.51) (1;0.95)}
7891 @code{(0;0) (0.50;0.48) (1;1)}
7893 @code{(0;0.22) (0.49;0.44) (1;0.80)}
7897 The previous example can also be achieved with the associated built-in preset:
7899 curves=preset=vintage
7909 Use a Photoshop preset and redefine the points of the green component:
7911 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
7915 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
7916 and @command{gnuplot}:
7918 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
7919 gnuplot -p /tmp/curves.plt
7925 Video data analysis filter.
7927 This filter shows hexadecimal pixel values of part of video.
7929 The filter accepts the following options:
7933 Set output video size.
7936 Set x offset from where to pick pixels.
7939 Set y offset from where to pick pixels.
7942 Set scope mode, can be one of the following:
7945 Draw hexadecimal pixel values with white color on black background.
7948 Draw hexadecimal pixel values with input video pixel color on black
7952 Draw hexadecimal pixel values on color background picked from input video,
7953 the text color is picked in such way so its always visible.
7957 Draw rows and columns numbers on left and top of video.
7960 Set background opacity.
7965 Denoise frames using 2D DCT (frequency domain filtering).
7967 This filter is not designed for real time.
7969 The filter accepts the following options:
7973 Set the noise sigma constant.
7975 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
7976 coefficient (absolute value) below this threshold with be dropped.
7978 If you need a more advanced filtering, see @option{expr}.
7980 Default is @code{0}.
7983 Set number overlapping pixels for each block. Since the filter can be slow, you
7984 may want to reduce this value, at the cost of a less effective filter and the
7985 risk of various artefacts.
7987 If the overlapping value doesn't permit processing the whole input width or
7988 height, a warning will be displayed and according borders won't be denoised.
7990 Default value is @var{blocksize}-1, which is the best possible setting.
7993 Set the coefficient factor expression.
7995 For each coefficient of a DCT block, this expression will be evaluated as a
7996 multiplier value for the coefficient.
7998 If this is option is set, the @option{sigma} option will be ignored.
8000 The absolute value of the coefficient can be accessed through the @var{c}
8004 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8005 @var{blocksize}, which is the width and height of the processed blocks.
8007 The default value is @var{3} (8x8) and can be raised to @var{4} for a
8008 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
8009 on the speed processing. Also, a larger block size does not necessarily means a
8013 @subsection Examples
8015 Apply a denoise with a @option{sigma} of @code{4.5}:
8020 The same operation can be achieved using the expression system:
8022 dctdnoiz=e='gte(c, 4.5*3)'
8025 Violent denoise using a block size of @code{16x16}:
8032 Remove banding artifacts from input video.
8033 It works by replacing banded pixels with average value of referenced pixels.
8035 The filter accepts the following options:
8042 Set banding detection threshold for each plane. Default is 0.02.
8043 Valid range is 0.00003 to 0.5.
8044 If difference between current pixel and reference pixel is less than threshold,
8045 it will be considered as banded.
8048 Banding detection range in pixels. Default is 16. If positive, random number
8049 in range 0 to set value will be used. If negative, exact absolute value
8051 The range defines square of four pixels around current pixel.
8054 Set direction in radians from which four pixel will be compared. If positive,
8055 random direction from 0 to set direction will be picked. If negative, exact of
8056 absolute value will be picked. For example direction 0, -PI or -2*PI radians
8057 will pick only pixels on same row and -PI/2 will pick only pixels on same
8061 If enabled, current pixel is compared with average value of all four
8062 surrounding pixels. The default is enabled. If disabled current pixel is
8063 compared with all four surrounding pixels. The pixel is considered banded
8064 if only all four differences with surrounding pixels are less than threshold.
8067 If enabled, current pixel is changed if and only if all pixel components are banded,
8068 e.g. banding detection threshold is triggered for all color components.
8069 The default is disabled.
8074 Remove blocking artifacts from input video.
8076 The filter accepts the following options:
8080 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
8081 This controls what kind of deblocking is applied.
8084 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
8090 Set blocking detection thresholds. Allowed range is 0 to 1.
8091 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
8092 Using higher threshold gives more deblocking strength.
8093 Setting @var{alpha} controls threshold detection at exact edge of block.
8094 Remaining options controls threshold detection near the edge. Each one for
8095 below/above or left/right. Setting any of those to @var{0} disables
8099 Set planes to filter. Default is to filter all available planes.
8102 @subsection Examples
8106 Deblock using weak filter and block size of 4 pixels.
8108 deblock=filter=weak:block=4
8112 Deblock using strong filter, block size of 4 pixels and custom thresholds for
8113 deblocking more edges.
8115 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
8119 Similar as above, but filter only first plane.
8121 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
8125 Similar as above, but filter only second and third plane.
8127 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
8134 Drop duplicated frames at regular intervals.
8136 The filter accepts the following options:
8140 Set the number of frames from which one will be dropped. Setting this to
8141 @var{N} means one frame in every batch of @var{N} frames will be dropped.
8142 Default is @code{5}.
8145 Set the threshold for duplicate detection. If the difference metric for a frame
8146 is less than or equal to this value, then it is declared as duplicate. Default
8150 Set scene change threshold. Default is @code{15}.
8154 Set the size of the x and y-axis blocks used during metric calculations.
8155 Larger blocks give better noise suppression, but also give worse detection of
8156 small movements. Must be a power of two. Default is @code{32}.
8159 Mark main input as a pre-processed input and activate clean source input
8160 stream. This allows the input to be pre-processed with various filters to help
8161 the metrics calculation while keeping the frame selection lossless. When set to
8162 @code{1}, the first stream is for the pre-processed input, and the second
8163 stream is the clean source from where the kept frames are chosen. Default is
8167 Set whether or not chroma is considered in the metric calculations. Default is
8173 Apply 2D deconvolution of video stream in frequency domain using second stream
8176 The filter accepts the following options:
8180 Set which planes to process.
8183 Set which impulse video frames will be processed, can be @var{first}
8184 or @var{all}. Default is @var{all}.
8187 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
8188 and height are not same and not power of 2 or if stream prior to convolving
8192 The @code{deconvolve} filter also supports the @ref{framesync} options.
8196 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
8198 It accepts the following options:
8202 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
8203 @var{rainbows} for cross-color reduction.
8206 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
8209 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
8212 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
8215 Set temporal chroma threshold. Lower values increases reduction of cross-color.
8220 Apply deflate effect to the video.
8222 This filter replaces the pixel by the local(3x3) average by taking into account
8223 only values lower than the pixel.
8225 It accepts the following options:
8232 Limit the maximum change for each plane, default is 65535.
8233 If 0, plane will remain unchanged.
8238 Remove temporal frame luminance variations.
8240 It accepts the following options:
8244 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
8247 Set averaging mode to smooth temporal luminance variations.
8249 Available values are:
8274 Do not actually modify frame. Useful when one only wants metadata.
8279 Remove judder produced by partially interlaced telecined content.
8281 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
8282 source was partially telecined content then the output of @code{pullup,dejudder}
8283 will have a variable frame rate. May change the recorded frame rate of the
8284 container. Aside from that change, this filter will not affect constant frame
8287 The option available in this filter is:
8291 Specify the length of the window over which the judder repeats.
8293 Accepts any integer greater than 1. Useful values are:
8297 If the original was telecined from 24 to 30 fps (Film to NTSC).
8300 If the original was telecined from 25 to 30 fps (PAL to NTSC).
8303 If a mixture of the two.
8306 The default is @samp{4}.
8311 Suppress a TV station logo by a simple interpolation of the surrounding
8312 pixels. Just set a rectangle covering the logo and watch it disappear
8313 (and sometimes something even uglier appear - your mileage may vary).
8315 It accepts the following parameters:
8320 Specify the top left corner coordinates of the logo. They must be
8325 Specify the width and height of the logo to clear. They must be
8329 Specify the thickness of the fuzzy edge of the rectangle (added to
8330 @var{w} and @var{h}). The default value is 1. This option is
8331 deprecated, setting higher values should no longer be necessary and
8335 When set to 1, a green rectangle is drawn on the screen to simplify
8336 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
8337 The default value is 0.
8339 The rectangle is drawn on the outermost pixels which will be (partly)
8340 replaced with interpolated values. The values of the next pixels
8341 immediately outside this rectangle in each direction will be used to
8342 compute the interpolated pixel values inside the rectangle.
8346 @subsection Examples
8350 Set a rectangle covering the area with top left corner coordinates 0,0
8351 and size 100x77, and a band of size 10:
8353 delogo=x=0:y=0:w=100:h=77:band=10
8360 Remove the rain in the input image/video by applying the derain methods based on
8361 convolutional neural networks. Supported models:
8365 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
8366 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
8369 Training scripts as well as scripts for model generation are provided in
8370 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
8372 The filter accepts the following options:
8376 Specify which DNN backend to use for model loading and execution. This option accepts
8377 the following values:
8381 Native implementation of DNN loading and execution.
8383 Default value is @samp{native}.
8386 Set path to model file specifying network architecture and its parameters.
8387 Note that different backends use different file formats. TensorFlow backend
8388 can load files for both formats, while native backend can load files for only
8394 Attempt to fix small changes in horizontal and/or vertical shift. This
8395 filter helps remove camera shake from hand-holding a camera, bumping a
8396 tripod, moving on a vehicle, etc.
8398 The filter accepts the following options:
8406 Specify a rectangular area where to limit the search for motion
8408 If desired the search for motion vectors can be limited to a
8409 rectangular area of the frame defined by its top left corner, width
8410 and height. These parameters have the same meaning as the drawbox
8411 filter which can be used to visualise the position of the bounding
8414 This is useful when simultaneous movement of subjects within the frame
8415 might be confused for camera motion by the motion vector search.
8417 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
8418 then the full frame is used. This allows later options to be set
8419 without specifying the bounding box for the motion vector search.
8421 Default - search the whole frame.
8425 Specify the maximum extent of movement in x and y directions in the
8426 range 0-64 pixels. Default 16.
8429 Specify how to generate pixels to fill blanks at the edge of the
8430 frame. Available values are:
8433 Fill zeroes at blank locations
8435 Original image at blank locations
8437 Extruded edge value at blank locations
8439 Mirrored edge at blank locations
8441 Default value is @samp{mirror}.
8444 Specify the blocksize to use for motion search. Range 4-128 pixels,
8448 Specify the contrast threshold for blocks. Only blocks with more than
8449 the specified contrast (difference between darkest and lightest
8450 pixels) will be considered. Range 1-255, default 125.
8453 Specify the search strategy. Available values are:
8456 Set exhaustive search
8458 Set less exhaustive search.
8460 Default value is @samp{exhaustive}.
8463 If set then a detailed log of the motion search is written to the
8470 Remove unwanted contamination of foreground colors, caused by reflected color of
8471 greenscreen or bluescreen.
8473 This filter accepts the following options:
8477 Set what type of despill to use.
8480 Set how spillmap will be generated.
8483 Set how much to get rid of still remaining spill.
8486 Controls amount of red in spill area.
8489 Controls amount of green in spill area.
8490 Should be -1 for greenscreen.
8493 Controls amount of blue in spill area.
8494 Should be -1 for bluescreen.
8497 Controls brightness of spill area, preserving colors.
8500 Modify alpha from generated spillmap.
8505 Apply an exact inverse of the telecine operation. It requires a predefined
8506 pattern specified using the pattern option which must be the same as that passed
8507 to the telecine filter.
8509 This filter accepts the following options:
8518 The default value is @code{top}.
8522 A string of numbers representing the pulldown pattern you wish to apply.
8523 The default value is @code{23}.
8526 A number representing position of the first frame with respect to the telecine
8527 pattern. This is to be used if the stream is cut. The default value is @code{0}.
8532 Apply dilation effect to the video.
8534 This filter replaces the pixel by the local(3x3) maximum.
8536 It accepts the following options:
8543 Limit the maximum change for each plane, default is 65535.
8544 If 0, plane will remain unchanged.
8547 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
8550 Flags to local 3x3 coordinates maps like this:
8559 Displace pixels as indicated by second and third input stream.
8561 It takes three input streams and outputs one stream, the first input is the
8562 source, and second and third input are displacement maps.
8564 The second input specifies how much to displace pixels along the
8565 x-axis, while the third input specifies how much to displace pixels
8567 If one of displacement map streams terminates, last frame from that
8568 displacement map will be used.
8570 Note that once generated, displacements maps can be reused over and over again.
8572 A description of the accepted options follows.
8576 Set displace behavior for pixels that are out of range.
8578 Available values are:
8581 Missing pixels are replaced by black pixels.
8584 Adjacent pixels will spread out to replace missing pixels.
8587 Out of range pixels are wrapped so they point to pixels of other side.
8590 Out of range pixels will be replaced with mirrored pixels.
8592 Default is @samp{smear}.
8596 @subsection Examples
8600 Add ripple effect to rgb input of video size hd720:
8602 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
8606 Add wave effect to rgb input of video size hd720:
8608 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
8614 Draw a colored box on the input image.
8616 It accepts the following parameters:
8621 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
8625 The expressions which specify the width and height of the box; if 0 they are interpreted as
8626 the input width and height. It defaults to 0.
8629 Specify the color of the box to write. For the general syntax of this option,
8630 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8631 value @code{invert} is used, the box edge color is the same as the
8632 video with inverted luma.
8635 The expression which sets the thickness of the box edge.
8636 A value of @code{fill} will create a filled box. Default value is @code{3}.
8638 See below for the list of accepted constants.
8641 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
8642 will overwrite the video's color and alpha pixels.
8643 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
8646 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8647 following constants:
8651 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8655 horizontal and vertical chroma subsample values. For example for the
8656 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8660 The input width and height.
8663 The input sample aspect ratio.
8667 The x and y offset coordinates where the box is drawn.
8671 The width and height of the drawn box.
8674 The thickness of the drawn box.
8676 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8677 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
8681 @subsection Examples
8685 Draw a black box around the edge of the input image:
8691 Draw a box with color red and an opacity of 50%:
8693 drawbox=10:20:200:60:red@@0.5
8696 The previous example can be specified as:
8698 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
8702 Fill the box with pink color:
8704 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
8708 Draw a 2-pixel red 2.40:1 mask:
8710 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
8716 Draw a grid on the input image.
8718 It accepts the following parameters:
8723 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
8727 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
8728 input width and height, respectively, minus @code{thickness}, so image gets
8729 framed. Default to 0.
8732 Specify the color of the grid. For the general syntax of this option,
8733 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8734 value @code{invert} is used, the grid color is the same as the
8735 video with inverted luma.
8738 The expression which sets the thickness of the grid line. Default value is @code{1}.
8740 See below for the list of accepted constants.
8743 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
8744 will overwrite the video's color and alpha pixels.
8745 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
8748 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8749 following constants:
8753 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8757 horizontal and vertical chroma subsample values. For example for the
8758 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8762 The input grid cell width and height.
8765 The input sample aspect ratio.
8769 The x and y coordinates of some point of grid intersection (meant to configure offset).
8773 The width and height of the drawn cell.
8776 The thickness of the drawn cell.
8778 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8779 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
8783 @subsection Examples
8787 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
8789 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
8793 Draw a white 3x3 grid with an opacity of 50%:
8795 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
8802 Draw a text string or text from a specified file on top of a video, using the
8803 libfreetype library.
8805 To enable compilation of this filter, you need to configure FFmpeg with
8806 @code{--enable-libfreetype}.
8807 To enable default font fallback and the @var{font} option you need to
8808 configure FFmpeg with @code{--enable-libfontconfig}.
8809 To enable the @var{text_shaping} option, you need to configure FFmpeg with
8810 @code{--enable-libfribidi}.
8814 It accepts the following parameters:
8819 Used to draw a box around text using the background color.
8820 The value must be either 1 (enable) or 0 (disable).
8821 The default value of @var{box} is 0.
8824 Set the width of the border to be drawn around the box using @var{boxcolor}.
8825 The default value of @var{boxborderw} is 0.
8828 The color to be used for drawing box around text. For the syntax of this
8829 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8831 The default value of @var{boxcolor} is "white".
8834 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
8835 The default value of @var{line_spacing} is 0.
8838 Set the width of the border to be drawn around the text using @var{bordercolor}.
8839 The default value of @var{borderw} is 0.
8842 Set the color to be used for drawing border around text. For the syntax of this
8843 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8845 The default value of @var{bordercolor} is "black".
8848 Select how the @var{text} is expanded. Can be either @code{none},
8849 @code{strftime} (deprecated) or
8850 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
8854 Set a start time for the count. Value is in microseconds. Only applied
8855 in the deprecated strftime expansion mode. To emulate in normal expansion
8856 mode use the @code{pts} function, supplying the start time (in seconds)
8857 as the second argument.
8860 If true, check and fix text coords to avoid clipping.
8863 The color to be used for drawing fonts. For the syntax of this option, check
8864 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8866 The default value of @var{fontcolor} is "black".
8868 @item fontcolor_expr
8869 String which is expanded the same way as @var{text} to obtain dynamic
8870 @var{fontcolor} value. By default this option has empty value and is not
8871 processed. When this option is set, it overrides @var{fontcolor} option.
8874 The font family to be used for drawing text. By default Sans.
8877 The font file to be used for drawing text. The path must be included.
8878 This parameter is mandatory if the fontconfig support is disabled.
8881 Draw the text applying alpha blending. The value can
8882 be a number between 0.0 and 1.0.
8883 The expression accepts the same variables @var{x, y} as well.
8884 The default value is 1.
8885 Please see @var{fontcolor_expr}.
8888 The font size to be used for drawing text.
8889 The default value of @var{fontsize} is 16.
8892 If set to 1, attempt to shape the text (for example, reverse the order of
8893 right-to-left text and join Arabic characters) before drawing it.
8894 Otherwise, just draw the text exactly as given.
8895 By default 1 (if supported).
8898 The flags to be used for loading the fonts.
8900 The flags map the corresponding flags supported by libfreetype, and are
8901 a combination of the following values:
8908 @item vertical_layout
8909 @item force_autohint
8912 @item ignore_global_advance_width
8914 @item ignore_transform
8920 Default value is "default".
8922 For more information consult the documentation for the FT_LOAD_*
8926 The color to be used for drawing a shadow behind the drawn text. For the
8927 syntax of this option, check the @ref{color syntax,,"Color" section in the
8928 ffmpeg-utils manual,ffmpeg-utils}.
8930 The default value of @var{shadowcolor} is "black".
8934 The x and y offsets for the text shadow position with respect to the
8935 position of the text. They can be either positive or negative
8936 values. The default value for both is "0".
8939 The starting frame number for the n/frame_num variable. The default value
8943 The size in number of spaces to use for rendering the tab.
8947 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
8948 format. It can be used with or without text parameter. @var{timecode_rate}
8949 option must be specified.
8951 @item timecode_rate, rate, r
8952 Set the timecode frame rate (timecode only). Value will be rounded to nearest
8953 integer. Minimum value is "1".
8954 Drop-frame timecode is supported for frame rates 30 & 60.
8957 If set to 1, the output of the timecode option will wrap around at 24 hours.
8958 Default is 0 (disabled).
8961 The text string to be drawn. The text must be a sequence of UTF-8
8963 This parameter is mandatory if no file is specified with the parameter
8967 A text file containing text to be drawn. The text must be a sequence
8968 of UTF-8 encoded characters.
8970 This parameter is mandatory if no text string is specified with the
8971 parameter @var{text}.
8973 If both @var{text} and @var{textfile} are specified, an error is thrown.
8976 If set to 1, the @var{textfile} will be reloaded before each frame.
8977 Be sure to update it atomically, or it may be read partially, or even fail.
8981 The expressions which specify the offsets where text will be drawn
8982 within the video frame. They are relative to the top/left border of the
8985 The default value of @var{x} and @var{y} is "0".
8987 See below for the list of accepted constants and functions.
8990 The parameters for @var{x} and @var{y} are expressions containing the
8991 following constants and functions:
8995 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
8999 horizontal and vertical chroma subsample values. For example for the
9000 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9003 the height of each text line
9011 @item max_glyph_a, ascent
9012 the maximum distance from the baseline to the highest/upper grid
9013 coordinate used to place a glyph outline point, for all the rendered
9015 It is a positive value, due to the grid's orientation with the Y axis
9018 @item max_glyph_d, descent
9019 the maximum distance from the baseline to the lowest grid coordinate
9020 used to place a glyph outline point, for all the rendered glyphs.
9021 This is a negative value, due to the grid's orientation, with the Y axis
9025 maximum glyph height, that is the maximum height for all the glyphs
9026 contained in the rendered text, it is equivalent to @var{ascent} -
9030 maximum glyph width, that is the maximum width for all the glyphs
9031 contained in the rendered text
9034 the number of input frame, starting from 0
9036 @item rand(min, max)
9037 return a random number included between @var{min} and @var{max}
9040 The input sample aspect ratio.
9043 timestamp expressed in seconds, NAN if the input timestamp is unknown
9046 the height of the rendered text
9049 the width of the rendered text
9053 the x and y offset coordinates where the text is drawn.
9055 These parameters allow the @var{x} and @var{y} expressions to refer
9056 to each other, so you can for example specify @code{y=x/dar}.
9059 A one character description of the current frame's picture type.
9062 The current packet's position in the input file or stream
9063 (in bytes, from the start of the input). A value of -1 indicates
9064 this info is not available.
9067 The current packet's duration, in seconds.
9070 The current packet's size (in bytes).
9073 @anchor{drawtext_expansion}
9074 @subsection Text expansion
9076 If @option{expansion} is set to @code{strftime},
9077 the filter recognizes strftime() sequences in the provided text and
9078 expands them accordingly. Check the documentation of strftime(). This
9079 feature is deprecated.
9081 If @option{expansion} is set to @code{none}, the text is printed verbatim.
9083 If @option{expansion} is set to @code{normal} (which is the default),
9084 the following expansion mechanism is used.
9086 The backslash character @samp{\}, followed by any character, always expands to
9087 the second character.
9089 Sequences of the form @code{%@{...@}} are expanded. The text between the
9090 braces is a function name, possibly followed by arguments separated by ':'.
9091 If the arguments contain special characters or delimiters (':' or '@}'),
9092 they should be escaped.
9094 Note that they probably must also be escaped as the value for the
9095 @option{text} option in the filter argument string and as the filter
9096 argument in the filtergraph description, and possibly also for the shell,
9097 that makes up to four levels of escaping; using a text file avoids these
9100 The following functions are available:
9105 The expression evaluation result.
9107 It must take one argument specifying the expression to be evaluated,
9108 which accepts the same constants and functions as the @var{x} and
9109 @var{y} values. Note that not all constants should be used, for
9110 example the text size is not known when evaluating the expression, so
9111 the constants @var{text_w} and @var{text_h} will have an undefined
9114 @item expr_int_format, eif
9115 Evaluate the expression's value and output as formatted integer.
9117 The first argument is the expression to be evaluated, just as for the @var{expr} function.
9118 The second argument specifies the output format. Allowed values are @samp{x},
9119 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
9120 @code{printf} function.
9121 The third parameter is optional and sets the number of positions taken by the output.
9122 It can be used to add padding with zeros from the left.
9125 The time at which the filter is running, expressed in UTC.
9126 It can accept an argument: a strftime() format string.
9129 The time at which the filter is running, expressed in the local time zone.
9130 It can accept an argument: a strftime() format string.
9133 Frame metadata. Takes one or two arguments.
9135 The first argument is mandatory and specifies the metadata key.
9137 The second argument is optional and specifies a default value, used when the
9138 metadata key is not found or empty.
9140 Available metadata can be identified by inspecting entries
9141 starting with TAG included within each frame section
9142 printed by running @code{ffprobe -show_frames}.
9144 String metadata generated in filters leading to
9145 the drawtext filter are also available.
9148 The frame number, starting from 0.
9151 A one character description of the current picture type.
9154 The timestamp of the current frame.
9155 It can take up to three arguments.
9157 The first argument is the format of the timestamp; it defaults to @code{flt}
9158 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
9159 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
9160 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
9161 @code{localtime} stands for the timestamp of the frame formatted as
9162 local time zone time.
9164 The second argument is an offset added to the timestamp.
9166 If the format is set to @code{hms}, a third argument @code{24HH} may be
9167 supplied to present the hour part of the formatted timestamp in 24h format
9170 If the format is set to @code{localtime} or @code{gmtime},
9171 a third argument may be supplied: a strftime() format string.
9172 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
9175 @subsection Commands
9177 This filter supports altering parameters via commands:
9180 Alter existing filter parameters.
9182 Syntax for the argument is the same as for filter invocation, e.g.
9185 fontsize=56:fontcolor=green:text='Hello World'
9188 Full filter invocation with sendcmd would look like this:
9191 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
9195 If the entire argument can't be parsed or applied as valid values then the filter will
9196 continue with its existing parameters.
9198 @subsection Examples
9202 Draw "Test Text" with font FreeSerif, using the default values for the
9203 optional parameters.
9206 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
9210 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
9211 and y=50 (counting from the top-left corner of the screen), text is
9212 yellow with a red box around it. Both the text and the box have an
9216 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
9217 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
9220 Note that the double quotes are not necessary if spaces are not used
9221 within the parameter list.
9224 Show the text at the center of the video frame:
9226 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
9230 Show the text at a random position, switching to a new position every 30 seconds:
9232 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)"
9236 Show a text line sliding from right to left in the last row of the video
9237 frame. The file @file{LONG_LINE} is assumed to contain a single line
9240 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
9244 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
9246 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
9250 Draw a single green letter "g", at the center of the input video.
9251 The glyph baseline is placed at half screen height.
9253 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
9257 Show text for 1 second every 3 seconds:
9259 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
9263 Use fontconfig to set the font. Note that the colons need to be escaped.
9265 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
9269 Print the date of a real-time encoding (see strftime(3)):
9271 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
9275 Show text fading in and out (appearing/disappearing):
9278 DS=1.0 # display start
9279 DE=10.0 # display end
9280 FID=1.5 # fade in duration
9281 FOD=5 # fade out duration
9282 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 @}"
9286 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
9287 and the @option{fontsize} value are included in the @option{y} offset.
9289 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
9290 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
9295 For more information about libfreetype, check:
9296 @url{http://www.freetype.org/}.
9298 For more information about fontconfig, check:
9299 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
9301 For more information about libfribidi, check:
9302 @url{http://fribidi.org/}.
9306 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
9308 The filter accepts the following options:
9313 Set low and high threshold values used by the Canny thresholding
9316 The high threshold selects the "strong" edge pixels, which are then
9317 connected through 8-connectivity with the "weak" edge pixels selected
9318 by the low threshold.
9320 @var{low} and @var{high} threshold values must be chosen in the range
9321 [0,1], and @var{low} should be lesser or equal to @var{high}.
9323 Default value for @var{low} is @code{20/255}, and default value for @var{high}
9327 Define the drawing mode.
9331 Draw white/gray wires on black background.
9334 Mix the colors to create a paint/cartoon effect.
9337 Apply Canny edge detector on all selected planes.
9339 Default value is @var{wires}.
9342 Select planes for filtering. By default all available planes are filtered.
9345 @subsection Examples
9349 Standard edge detection with custom values for the hysteresis thresholding:
9351 edgedetect=low=0.1:high=0.4
9355 Painting effect without thresholding:
9357 edgedetect=mode=colormix:high=0
9362 Set brightness, contrast, saturation and approximate gamma adjustment.
9364 The filter accepts the following options:
9368 Set the contrast expression. The value must be a float value in range
9369 @code{-2.0} to @code{2.0}. The default value is "1".
9372 Set the brightness expression. The value must be a float value in
9373 range @code{-1.0} to @code{1.0}. The default value is "0".
9376 Set the saturation expression. The value must be a float in
9377 range @code{0.0} to @code{3.0}. The default value is "1".
9380 Set the gamma expression. The value must be a float in range
9381 @code{0.1} to @code{10.0}. The default value is "1".
9384 Set the gamma expression for red. The value must be a float in
9385 range @code{0.1} to @code{10.0}. The default value is "1".
9388 Set the gamma expression for green. The value must be a float in range
9389 @code{0.1} to @code{10.0}. The default value is "1".
9392 Set the gamma expression for blue. The value must be a float in range
9393 @code{0.1} to @code{10.0}. The default value is "1".
9396 Set the gamma weight expression. It can be used to reduce the effect
9397 of a high gamma value on bright image areas, e.g. keep them from
9398 getting overamplified and just plain white. The value must be a float
9399 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
9400 gamma correction all the way down while @code{1.0} leaves it at its
9401 full strength. Default is "1".
9404 Set when the expressions for brightness, contrast, saturation and
9405 gamma expressions are evaluated.
9407 It accepts the following values:
9410 only evaluate expressions once during the filter initialization or
9411 when a command is processed
9414 evaluate expressions for each incoming frame
9417 Default value is @samp{init}.
9420 The expressions accept the following parameters:
9423 frame count of the input frame starting from 0
9426 byte position of the corresponding packet in the input file, NAN if
9430 frame rate of the input video, NAN if the input frame rate is unknown
9433 timestamp expressed in seconds, NAN if the input timestamp is unknown
9436 @subsection Commands
9437 The filter supports the following commands:
9441 Set the contrast expression.
9444 Set the brightness expression.
9447 Set the saturation expression.
9450 Set the gamma expression.
9453 Set the gamma_r expression.
9456 Set gamma_g expression.
9459 Set gamma_b expression.
9462 Set gamma_weight expression.
9464 The command accepts the same syntax of the corresponding option.
9466 If the specified expression is not valid, it is kept at its current
9473 Apply erosion effect to the video.
9475 This filter replaces the pixel by the local(3x3) minimum.
9477 It accepts the following options:
9484 Limit the maximum change for each plane, default is 65535.
9485 If 0, plane will remain unchanged.
9488 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9491 Flags to local 3x3 coordinates maps like this:
9498 @section extractplanes
9500 Extract color channel components from input video stream into
9501 separate grayscale video streams.
9503 The filter accepts the following option:
9507 Set plane(s) to extract.
9509 Available values for planes are:
9520 Choosing planes not available in the input will result in an error.
9521 That means you cannot select @code{r}, @code{g}, @code{b} planes
9522 with @code{y}, @code{u}, @code{v} planes at same time.
9525 @subsection Examples
9529 Extract luma, u and v color channel component from input video frame
9530 into 3 grayscale outputs:
9532 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
9538 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
9540 For each input image, the filter will compute the optimal mapping from
9541 the input to the output given the codebook length, that is the number
9542 of distinct output colors.
9544 This filter accepts the following options.
9547 @item codebook_length, l
9548 Set codebook length. The value must be a positive integer, and
9549 represents the number of distinct output colors. Default value is 256.
9552 Set the maximum number of iterations to apply for computing the optimal
9553 mapping. The higher the value the better the result and the higher the
9554 computation time. Default value is 1.
9557 Set a random seed, must be an integer included between 0 and
9558 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
9559 will try to use a good random seed on a best effort basis.
9562 Set pal8 output pixel format. This option does not work with codebook
9563 length greater than 256.
9568 Measure graylevel entropy in histogram of color channels of video frames.
9570 It accepts the following parameters:
9574 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
9576 @var{diff} mode measures entropy of histogram delta values, absolute differences
9577 between neighbour histogram values.
9582 Apply a fade-in/out effect to the input video.
9584 It accepts the following parameters:
9588 The effect type can be either "in" for a fade-in, or "out" for a fade-out
9590 Default is @code{in}.
9592 @item start_frame, s
9593 Specify the number of the frame to start applying the fade
9594 effect at. Default is 0.
9597 The number of frames that the fade effect lasts. At the end of the
9598 fade-in effect, the output video will have the same intensity as the input video.
9599 At the end of the fade-out transition, the output video will be filled with the
9600 selected @option{color}.
9604 If set to 1, fade only alpha channel, if one exists on the input.
9607 @item start_time, st
9608 Specify the timestamp (in seconds) of the frame to start to apply the fade
9609 effect. If both start_frame and start_time are specified, the fade will start at
9610 whichever comes last. Default is 0.
9613 The number of seconds for which the fade effect has to last. At the end of the
9614 fade-in effect the output video will have the same intensity as the input video,
9615 at the end of the fade-out transition the output video will be filled with the
9616 selected @option{color}.
9617 If both duration and nb_frames are specified, duration is used. Default is 0
9618 (nb_frames is used by default).
9621 Specify the color of the fade. Default is "black".
9624 @subsection Examples
9628 Fade in the first 30 frames of video:
9633 The command above is equivalent to:
9639 Fade out the last 45 frames of a 200-frame video:
9642 fade=type=out:start_frame=155:nb_frames=45
9646 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
9648 fade=in:0:25, fade=out:975:25
9652 Make the first 5 frames yellow, then fade in from frame 5-24:
9654 fade=in:5:20:color=yellow
9658 Fade in alpha over first 25 frames of video:
9660 fade=in:0:25:alpha=1
9664 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
9666 fade=t=in:st=5.5:d=0.5
9672 Apply arbitrary expressions to samples in frequency domain
9676 Adjust the dc value (gain) of the luma plane of the image. The filter
9677 accepts an integer value in range @code{0} to @code{1000}. The default
9678 value is set to @code{0}.
9681 Adjust the dc value (gain) of the 1st chroma plane of the image. The
9682 filter accepts an integer value in range @code{0} to @code{1000}. The
9683 default value is set to @code{0}.
9686 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
9687 filter accepts an integer value in range @code{0} to @code{1000}. The
9688 default value is set to @code{0}.
9691 Set the frequency domain weight expression for the luma plane.
9694 Set the frequency domain weight expression for the 1st chroma plane.
9697 Set the frequency domain weight expression for the 2nd chroma plane.
9700 Set when the expressions are evaluated.
9702 It accepts the following values:
9705 Only evaluate expressions once during the filter initialization.
9708 Evaluate expressions for each incoming frame.
9711 Default value is @samp{init}.
9713 The filter accepts the following variables:
9716 The coordinates of the current sample.
9720 The width and height of the image.
9723 The number of input frame, starting from 0.
9726 @subsection Examples
9732 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
9738 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
9744 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
9750 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
9756 Denoise frames using 3D FFT (frequency domain filtering).
9758 The filter accepts the following options:
9762 Set the noise sigma constant. This sets denoising strength.
9763 Default value is 1. Allowed range is from 0 to 30.
9764 Using very high sigma with low overlap may give blocking artifacts.
9767 Set amount of denoising. By default all detected noise is reduced.
9768 Default value is 1. Allowed range is from 0 to 1.
9771 Set size of block, Default is 4, can be 3, 4, 5 or 6.
9772 Actual size of block in pixels is 2 to power of @var{block}, so by default
9773 block size in pixels is 2^4 which is 16.
9776 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
9779 Set number of previous frames to use for denoising. By default is set to 0.
9782 Set number of next frames to to use for denoising. By default is set to 0.
9785 Set planes which will be filtered, by default are all available filtered
9791 Extract a single field from an interlaced image using stride
9792 arithmetic to avoid wasting CPU time. The output frames are marked as
9795 The filter accepts the following options:
9799 Specify whether to extract the top (if the value is @code{0} or
9800 @code{top}) or the bottom field (if the value is @code{1} or
9806 Create new frames by copying the top and bottom fields from surrounding frames
9807 supplied as numbers by the hint file.
9811 Set file containing hints: absolute/relative frame numbers.
9813 There must be one line for each frame in a clip. Each line must contain two
9814 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
9815 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
9816 is current frame number for @code{absolute} mode or out of [-1, 1] range
9817 for @code{relative} mode. First number tells from which frame to pick up top
9818 field and second number tells from which frame to pick up bottom field.
9820 If optionally followed by @code{+} output frame will be marked as interlaced,
9821 else if followed by @code{-} output frame will be marked as progressive, else
9822 it will be marked same as input frame.
9823 If line starts with @code{#} or @code{;} that line is skipped.
9826 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
9829 Example of first several lines of @code{hint} file for @code{relative} mode:
9832 1,0 - # second frame, use third's frame top field and second's frame bottom field
9833 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
9850 Field matching filter for inverse telecine. It is meant to reconstruct the
9851 progressive frames from a telecined stream. The filter does not drop duplicated
9852 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
9853 followed by a decimation filter such as @ref{decimate} in the filtergraph.
9855 The separation of the field matching and the decimation is notably motivated by
9856 the possibility of inserting a de-interlacing filter fallback between the two.
9857 If the source has mixed telecined and real interlaced content,
9858 @code{fieldmatch} will not be able to match fields for the interlaced parts.
9859 But these remaining combed frames will be marked as interlaced, and thus can be
9860 de-interlaced by a later filter such as @ref{yadif} before decimation.
9862 In addition to the various configuration options, @code{fieldmatch} can take an
9863 optional second stream, activated through the @option{ppsrc} option. If
9864 enabled, the frames reconstruction will be based on the fields and frames from
9865 this second stream. This allows the first input to be pre-processed in order to
9866 help the various algorithms of the filter, while keeping the output lossless
9867 (assuming the fields are matched properly). Typically, a field-aware denoiser,
9868 or brightness/contrast adjustments can help.
9870 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
9871 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
9872 which @code{fieldmatch} is based on. While the semantic and usage are very
9873 close, some behaviour and options names can differ.
9875 The @ref{decimate} filter currently only works for constant frame rate input.
9876 If your input has mixed telecined (30fps) and progressive content with a lower
9877 framerate like 24fps use the following filterchain to produce the necessary cfr
9878 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
9880 The filter accepts the following options:
9884 Specify the assumed field order of the input stream. Available values are:
9888 Auto detect parity (use FFmpeg's internal parity value).
9890 Assume bottom field first.
9892 Assume top field first.
9895 Note that it is sometimes recommended not to trust the parity announced by the
9898 Default value is @var{auto}.
9901 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
9902 sense that it won't risk creating jerkiness due to duplicate frames when
9903 possible, but if there are bad edits or blended fields it will end up
9904 outputting combed frames when a good match might actually exist. On the other
9905 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
9906 but will almost always find a good frame if there is one. The other values are
9907 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
9908 jerkiness and creating duplicate frames versus finding good matches in sections
9909 with bad edits, orphaned fields, blended fields, etc.
9911 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
9913 Available values are:
9917 2-way matching (p/c)
9919 2-way matching, and trying 3rd match if still combed (p/c + n)
9921 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
9923 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
9924 still combed (p/c + n + u/b)
9926 3-way matching (p/c/n)
9928 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
9929 detected as combed (p/c/n + u/b)
9932 The parenthesis at the end indicate the matches that would be used for that
9933 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
9936 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
9939 Default value is @var{pc_n}.
9942 Mark the main input stream as a pre-processed input, and enable the secondary
9943 input stream as the clean source to pick the fields from. See the filter
9944 introduction for more details. It is similar to the @option{clip2} feature from
9947 Default value is @code{0} (disabled).
9950 Set the field to match from. It is recommended to set this to the same value as
9951 @option{order} unless you experience matching failures with that setting. In
9952 certain circumstances changing the field that is used to match from can have a
9953 large impact on matching performance. Available values are:
9957 Automatic (same value as @option{order}).
9959 Match from the bottom field.
9961 Match from the top field.
9964 Default value is @var{auto}.
9967 Set whether or not chroma is included during the match comparisons. In most
9968 cases it is recommended to leave this enabled. You should set this to @code{0}
9969 only if your clip has bad chroma problems such as heavy rainbowing or other
9970 artifacts. Setting this to @code{0} could also be used to speed things up at
9971 the cost of some accuracy.
9973 Default value is @code{1}.
9977 These define an exclusion band which excludes the lines between @option{y0} and
9978 @option{y1} from being included in the field matching decision. An exclusion
9979 band can be used to ignore subtitles, a logo, or other things that may
9980 interfere with the matching. @option{y0} sets the starting scan line and
9981 @option{y1} sets the ending line; all lines in between @option{y0} and
9982 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
9983 @option{y0} and @option{y1} to the same value will disable the feature.
9984 @option{y0} and @option{y1} defaults to @code{0}.
9987 Set the scene change detection threshold as a percentage of maximum change on
9988 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
9989 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
9990 @option{scthresh} is @code{[0.0, 100.0]}.
9992 Default value is @code{12.0}.
9995 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
9996 account the combed scores of matches when deciding what match to use as the
9997 final match. Available values are:
10001 No final matching based on combed scores.
10003 Combed scores are only used when a scene change is detected.
10005 Use combed scores all the time.
10008 Default is @var{sc}.
10011 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
10012 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
10013 Available values are:
10017 No forced calculation.
10019 Force p/c/n calculations.
10021 Force p/c/n/u/b calculations.
10024 Default value is @var{none}.
10027 This is the area combing threshold used for combed frame detection. This
10028 essentially controls how "strong" or "visible" combing must be to be detected.
10029 Larger values mean combing must be more visible and smaller values mean combing
10030 can be less visible or strong and still be detected. Valid settings are from
10031 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
10032 be detected as combed). This is basically a pixel difference value. A good
10033 range is @code{[8, 12]}.
10035 Default value is @code{9}.
10038 Sets whether or not chroma is considered in the combed frame decision. Only
10039 disable this if your source has chroma problems (rainbowing, etc.) that are
10040 causing problems for the combed frame detection with chroma enabled. Actually,
10041 using @option{chroma}=@var{0} is usually more reliable, except for the case
10042 where there is chroma only combing in the source.
10044 Default value is @code{0}.
10048 Respectively set the x-axis and y-axis size of the window used during combed
10049 frame detection. This has to do with the size of the area in which
10050 @option{combpel} pixels are required to be detected as combed for a frame to be
10051 declared combed. See the @option{combpel} parameter description for more info.
10052 Possible values are any number that is a power of 2 starting at 4 and going up
10055 Default value is @code{16}.
10058 The number of combed pixels inside any of the @option{blocky} by
10059 @option{blockx} size blocks on the frame for the frame to be detected as
10060 combed. While @option{cthresh} controls how "visible" the combing must be, this
10061 setting controls "how much" combing there must be in any localized area (a
10062 window defined by the @option{blockx} and @option{blocky} settings) on the
10063 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
10064 which point no frames will ever be detected as combed). This setting is known
10065 as @option{MI} in TFM/VFM vocabulary.
10067 Default value is @code{80}.
10070 @anchor{p/c/n/u/b meaning}
10071 @subsection p/c/n/u/b meaning
10073 @subsubsection p/c/n
10075 We assume the following telecined stream:
10078 Top fields: 1 2 2 3 4
10079 Bottom fields: 1 2 3 4 4
10082 The numbers correspond to the progressive frame the fields relate to. Here, the
10083 first two frames are progressive, the 3rd and 4th are combed, and so on.
10085 When @code{fieldmatch} is configured to run a matching from bottom
10086 (@option{field}=@var{bottom}) this is how this input stream get transformed:
10091 B 1 2 3 4 4 <-- matching reference
10100 As a result of the field matching, we can see that some frames get duplicated.
10101 To perform a complete inverse telecine, you need to rely on a decimation filter
10102 after this operation. See for instance the @ref{decimate} filter.
10104 The same operation now matching from top fields (@option{field}=@var{top})
10109 T 1 2 2 3 4 <-- matching reference
10119 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
10120 basically, they refer to the frame and field of the opposite parity:
10123 @item @var{p} matches the field of the opposite parity in the previous frame
10124 @item @var{c} matches the field of the opposite parity in the current frame
10125 @item @var{n} matches the field of the opposite parity in the next frame
10130 The @var{u} and @var{b} matching are a bit special in the sense that they match
10131 from the opposite parity flag. In the following examples, we assume that we are
10132 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
10133 'x' is placed above and below each matched fields.
10135 With bottom matching (@option{field}=@var{bottom}):
10140 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10141 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10149 With top matching (@option{field}=@var{top}):
10154 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10155 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10163 @subsection Examples
10165 Simple IVTC of a top field first telecined stream:
10167 fieldmatch=order=tff:combmatch=none, decimate
10170 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
10172 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
10175 @section fieldorder
10177 Transform the field order of the input video.
10179 It accepts the following parameters:
10184 The output field order. Valid values are @var{tff} for top field first or @var{bff}
10185 for bottom field first.
10188 The default value is @samp{tff}.
10190 The transformation is done by shifting the picture content up or down
10191 by one line, and filling the remaining line with appropriate picture content.
10192 This method is consistent with most broadcast field order converters.
10194 If the input video is not flagged as being interlaced, or it is already
10195 flagged as being of the required output field order, then this filter does
10196 not alter the incoming video.
10198 It is very useful when converting to or from PAL DV material,
10199 which is bottom field first.
10203 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
10206 @section fifo, afifo
10208 Buffer input images and send them when they are requested.
10210 It is mainly useful when auto-inserted by the libavfilter
10213 It does not take parameters.
10215 @section fillborders
10217 Fill borders of the input video, without changing video stream dimensions.
10218 Sometimes video can have garbage at the four edges and you may not want to
10219 crop video input to keep size multiple of some number.
10221 This filter accepts the following options:
10225 Number of pixels to fill from left border.
10228 Number of pixels to fill from right border.
10231 Number of pixels to fill from top border.
10234 Number of pixels to fill from bottom border.
10239 It accepts the following values:
10242 fill pixels using outermost pixels
10245 fill pixels using mirroring
10248 fill pixels with constant value
10251 Default is @var{smear}.
10254 Set color for pixels in fixed mode. Default is @var{black}.
10259 Find a rectangular object
10261 It accepts the following options:
10265 Filepath of the object image, needs to be in gray8.
10268 Detection threshold, default is 0.5.
10271 Number of mipmaps, default is 3.
10273 @item xmin, ymin, xmax, ymax
10274 Specifies the rectangle in which to search.
10277 @subsection Examples
10281 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
10283 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
10287 @section cover_rect
10289 Cover a rectangular object
10291 It accepts the following options:
10295 Filepath of the optional cover image, needs to be in yuv420.
10300 It accepts the following values:
10303 cover it by the supplied image
10305 cover it by interpolating the surrounding pixels
10308 Default value is @var{blur}.
10311 @subsection Examples
10315 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
10317 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
10323 Flood area with values of same pixel components with another values.
10325 It accepts the following options:
10328 Set pixel x coordinate.
10331 Set pixel y coordinate.
10334 Set source #0 component value.
10337 Set source #1 component value.
10340 Set source #2 component value.
10343 Set source #3 component value.
10346 Set destination #0 component value.
10349 Set destination #1 component value.
10352 Set destination #2 component value.
10355 Set destination #3 component value.
10361 Convert the input video to one of the specified pixel formats.
10362 Libavfilter will try to pick one that is suitable as input to
10365 It accepts the following parameters:
10369 A '|'-separated list of pixel format names, such as
10370 "pix_fmts=yuv420p|monow|rgb24".
10374 @subsection Examples
10378 Convert the input video to the @var{yuv420p} format
10380 format=pix_fmts=yuv420p
10383 Convert the input video to any of the formats in the list
10385 format=pix_fmts=yuv420p|yuv444p|yuv410p
10392 Convert the video to specified constant frame rate by duplicating or dropping
10393 frames as necessary.
10395 It accepts the following parameters:
10399 The desired output frame rate. The default is @code{25}.
10402 Assume the first PTS should be the given value, in seconds. This allows for
10403 padding/trimming at the start of stream. By default, no assumption is made
10404 about the first frame's expected PTS, so no padding or trimming is done.
10405 For example, this could be set to 0 to pad the beginning with duplicates of
10406 the first frame if a video stream starts after the audio stream or to trim any
10407 frames with a negative PTS.
10410 Timestamp (PTS) rounding method.
10412 Possible values are:
10419 round towards -infinity
10421 round towards +infinity
10425 The default is @code{near}.
10428 Action performed when reading the last frame.
10430 Possible values are:
10433 Use same timestamp rounding method as used for other frames.
10435 Pass through last frame if input duration has not been reached yet.
10437 The default is @code{round}.
10441 Alternatively, the options can be specified as a flat string:
10442 @var{fps}[:@var{start_time}[:@var{round}]].
10444 See also the @ref{setpts} filter.
10446 @subsection Examples
10450 A typical usage in order to set the fps to 25:
10456 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
10458 fps=fps=film:round=near
10464 Pack two different video streams into a stereoscopic video, setting proper
10465 metadata on supported codecs. The two views should have the same size and
10466 framerate and processing will stop when the shorter video ends. Please note
10467 that you may conveniently adjust view properties with the @ref{scale} and
10470 It accepts the following parameters:
10474 The desired packing format. Supported values are:
10479 The views are next to each other (default).
10482 The views are on top of each other.
10485 The views are packed by line.
10488 The views are packed by column.
10491 The views are temporally interleaved.
10500 # Convert left and right views into a frame-sequential video
10501 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
10503 # Convert views into a side-by-side video with the same output resolution as the input
10504 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
10509 Change the frame rate by interpolating new video output frames from the source
10512 This filter is not designed to function correctly with interlaced media. If
10513 you wish to change the frame rate of interlaced media then you are required
10514 to deinterlace before this filter and re-interlace after this filter.
10516 A description of the accepted options follows.
10520 Specify the output frames per second. This option can also be specified
10521 as a value alone. The default is @code{50}.
10524 Specify the start of a range where the output frame will be created as a
10525 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10526 the default is @code{15}.
10529 Specify the end of a range where the output frame will be created as a
10530 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10531 the default is @code{240}.
10534 Specify the level at which a scene change is detected as a value between
10535 0 and 100 to indicate a new scene; a low value reflects a low
10536 probability for the current frame to introduce a new scene, while a higher
10537 value means the current frame is more likely to be one.
10538 The default is @code{8.2}.
10541 Specify flags influencing the filter process.
10543 Available value for @var{flags} is:
10546 @item scene_change_detect, scd
10547 Enable scene change detection using the value of the option @var{scene}.
10548 This flag is enabled by default.
10554 Select one frame every N-th frame.
10556 This filter accepts the following option:
10559 Select frame after every @code{step} frames.
10560 Allowed values are positive integers higher than 0. Default value is @code{1}.
10563 @section freezedetect
10565 Detect frozen video.
10567 This filter logs a message and sets frame metadata when it detects that the
10568 input video has no significant change in content during a specified duration.
10569 Video freeze detection calculates the mean average absolute difference of all
10570 the components of video frames and compares it to a noise floor.
10572 The printed times and duration are expressed in seconds. The
10573 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
10574 whose timestamp equals or exceeds the detection duration and it contains the
10575 timestamp of the first frame of the freeze. The
10576 @code{lavfi.freezedetect.freeze_duration} and
10577 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
10580 The filter accepts the following options:
10584 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
10585 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
10589 Set freeze duration until notification (default is 2 seconds).
10595 Apply a frei0r effect to the input video.
10597 To enable the compilation of this filter, you need to install the frei0r
10598 header and configure FFmpeg with @code{--enable-frei0r}.
10600 It accepts the following parameters:
10605 The name of the frei0r effect to load. If the environment variable
10606 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
10607 directories specified by the colon-separated list in @env{FREI0R_PATH}.
10608 Otherwise, the standard frei0r paths are searched, in this order:
10609 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
10610 @file{/usr/lib/frei0r-1/}.
10612 @item filter_params
10613 A '|'-separated list of parameters to pass to the frei0r effect.
10617 A frei0r effect parameter can be a boolean (its value is either
10618 "y" or "n"), a double, a color (specified as
10619 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
10620 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
10621 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
10622 a position (specified as @var{X}/@var{Y}, where
10623 @var{X} and @var{Y} are floating point numbers) and/or a string.
10625 The number and types of parameters depend on the loaded effect. If an
10626 effect parameter is not specified, the default value is set.
10628 @subsection Examples
10632 Apply the distort0r effect, setting the first two double parameters:
10634 frei0r=filter_name=distort0r:filter_params=0.5|0.01
10638 Apply the colordistance effect, taking a color as the first parameter:
10640 frei0r=colordistance:0.2/0.3/0.4
10641 frei0r=colordistance:violet
10642 frei0r=colordistance:0x112233
10646 Apply the perspective effect, specifying the top left and top right image
10649 frei0r=perspective:0.2/0.2|0.8/0.2
10653 For more information, see
10654 @url{http://frei0r.dyne.org}
10658 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
10660 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
10661 processing filter, one of them is performed once per block, not per pixel.
10662 This allows for much higher speed.
10664 The filter accepts the following options:
10668 Set quality. This option defines the number of levels for averaging. It accepts
10669 an integer in the range 4-5. Default value is @code{4}.
10672 Force a constant quantization parameter. It accepts an integer in range 0-63.
10673 If not set, the filter will use the QP from the video stream (if available).
10676 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
10677 more details but also more artifacts, while higher values make the image smoother
10678 but also blurrier. Default value is @code{0} − PSNR optimal.
10680 @item use_bframe_qp
10681 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
10682 option may cause flicker since the B-Frames have often larger QP. Default is
10683 @code{0} (not enabled).
10689 Apply Gaussian blur filter.
10691 The filter accepts the following options:
10695 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
10698 Set number of steps for Gaussian approximation. Default is @code{1}.
10701 Set which planes to filter. By default all planes are filtered.
10704 Set vertical sigma, if negative it will be same as @code{sigma}.
10705 Default is @code{-1}.
10710 Apply generic equation to each pixel.
10712 The filter accepts the following options:
10715 @item lum_expr, lum
10716 Set the luminance expression.
10718 Set the chrominance blue expression.
10720 Set the chrominance red expression.
10721 @item alpha_expr, a
10722 Set the alpha expression.
10724 Set the red expression.
10725 @item green_expr, g
10726 Set the green expression.
10728 Set the blue expression.
10731 The colorspace is selected according to the specified options. If one
10732 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
10733 options is specified, the filter will automatically select a YCbCr
10734 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
10735 @option{blue_expr} options is specified, it will select an RGB
10738 If one of the chrominance expression is not defined, it falls back on the other
10739 one. If no alpha expression is specified it will evaluate to opaque value.
10740 If none of chrominance expressions are specified, they will evaluate
10741 to the luminance expression.
10743 The expressions can use the following variables and functions:
10747 The sequential number of the filtered frame, starting from @code{0}.
10751 The coordinates of the current sample.
10755 The width and height of the image.
10759 Width and height scale depending on the currently filtered plane. It is the
10760 ratio between the corresponding luma plane number of pixels and the current
10761 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
10762 @code{0.5,0.5} for chroma planes.
10765 Time of the current frame, expressed in seconds.
10768 Return the value of the pixel at location (@var{x},@var{y}) of the current
10772 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
10776 Return the value of the pixel at location (@var{x},@var{y}) of the
10777 blue-difference chroma plane. Return 0 if there is no such plane.
10780 Return the value of the pixel at location (@var{x},@var{y}) of the
10781 red-difference chroma plane. Return 0 if there is no such plane.
10786 Return the value of the pixel at location (@var{x},@var{y}) of the
10787 red/green/blue component. Return 0 if there is no such component.
10790 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
10791 plane. Return 0 if there is no such plane.
10794 For functions, if @var{x} and @var{y} are outside the area, the value will be
10795 automatically clipped to the closer edge.
10797 @subsection Examples
10801 Flip the image horizontally:
10807 Generate a bidimensional sine wave, with angle @code{PI/3} and a
10808 wavelength of 100 pixels:
10810 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
10814 Generate a fancy enigmatic moving light:
10816 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
10820 Generate a quick emboss effect:
10822 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
10826 Modify RGB components depending on pixel position:
10828 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
10832 Create a radial gradient that is the same size as the input (also see
10833 the @ref{vignette} filter):
10835 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
10841 Fix the banding artifacts that are sometimes introduced into nearly flat
10842 regions by truncation to 8-bit color depth.
10843 Interpolate the gradients that should go where the bands are, and
10846 It is designed for playback only. Do not use it prior to
10847 lossy compression, because compression tends to lose the dither and
10848 bring back the bands.
10850 It accepts the following parameters:
10855 The maximum amount by which the filter will change any one pixel. This is also
10856 the threshold for detecting nearly flat regions. Acceptable values range from
10857 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
10861 The neighborhood to fit the gradient to. A larger radius makes for smoother
10862 gradients, but also prevents the filter from modifying the pixels near detailed
10863 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
10864 values will be clipped to the valid range.
10868 Alternatively, the options can be specified as a flat string:
10869 @var{strength}[:@var{radius}]
10871 @subsection Examples
10875 Apply the filter with a @code{3.5} strength and radius of @code{8}:
10881 Specify radius, omitting the strength (which will fall-back to the default
10889 @section graphmonitor, agraphmonitor
10890 Show various filtergraph stats.
10892 With this filter one can debug complete filtergraph.
10893 Especially issues with links filling with queued frames.
10895 The filter accepts the following options:
10899 Set video output size. Default is @var{hd720}.
10902 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
10905 Set output mode, can be @var{fulll} or @var{compact}.
10906 In @var{compact} mode only filters with some queued frames have displayed stats.
10909 Set flags which enable which stats are shown in video.
10911 Available values for flags are:
10914 Display number of queued frames in each link.
10916 @item frame_count_in
10917 Display number of frames taken from filter.
10919 @item frame_count_out
10920 Display number of frames given out from filter.
10923 Display current filtered frame pts.
10926 Display current filtered frame time.
10929 Display time base for filter link.
10932 Display used format for filter link.
10935 Display video size or number of audio channels in case of audio used by filter link.
10938 Display video frame rate or sample rate in case of audio used by filter link.
10942 Set upper limit for video rate of output stream, Default value is @var{25}.
10943 This guarantee that output video frame rate will not be higher than this value.
10947 A color constancy variation filter which estimates scene illumination via grey edge algorithm
10948 and corrects the scene colors accordingly.
10950 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
10952 The filter accepts the following options:
10956 The order of differentiation to be applied on the scene. Must be chosen in the range
10957 [0,2] and default value is 1.
10960 The Minkowski parameter to be used for calculating the Minkowski distance. Must
10961 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
10962 max value instead of calculating Minkowski distance.
10965 The standard deviation of Gaussian blur to be applied on the scene. Must be
10966 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
10967 can't be equal to 0 if @var{difford} is greater than 0.
10970 @subsection Examples
10976 greyedge=difford=1:minknorm=5:sigma=2
10982 greyedge=difford=1:minknorm=0:sigma=2
10990 Apply a Hald CLUT to a video stream.
10992 First input is the video stream to process, and second one is the Hald CLUT.
10993 The Hald CLUT input can be a simple picture or a complete video stream.
10995 The filter accepts the following options:
10999 Force termination when the shortest input terminates. Default is @code{0}.
11001 Continue applying the last CLUT after the end of the stream. A value of
11002 @code{0} disable the filter after the last frame of the CLUT is reached.
11003 Default is @code{1}.
11006 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
11007 filters share the same internals).
11009 This filter also supports the @ref{framesync} options.
11011 More information about the Hald CLUT can be found on Eskil Steenberg's website
11012 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
11014 @subsection Workflow examples
11016 @subsubsection Hald CLUT video stream
11018 Generate an identity Hald CLUT stream altered with various effects:
11020 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
11023 Note: make sure you use a lossless codec.
11025 Then use it with @code{haldclut} to apply it on some random stream:
11027 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
11030 The Hald CLUT will be applied to the 10 first seconds (duration of
11031 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
11032 to the remaining frames of the @code{mandelbrot} stream.
11034 @subsubsection Hald CLUT with preview
11036 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
11037 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
11038 biggest possible square starting at the top left of the picture. The remaining
11039 padding pixels (bottom or right) will be ignored. This area can be used to add
11040 a preview of the Hald CLUT.
11042 Typically, the following generated Hald CLUT will be supported by the
11043 @code{haldclut} filter:
11046 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
11047 pad=iw+320 [padded_clut];
11048 smptebars=s=320x256, split [a][b];
11049 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
11050 [main][b] overlay=W-320" -frames:v 1 clut.png
11053 It contains the original and a preview of the effect of the CLUT: SMPTE color
11054 bars are displayed on the right-top, and below the same color bars processed by
11057 Then, the effect of this Hald CLUT can be visualized with:
11059 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
11064 Flip the input video horizontally.
11066 For example, to horizontally flip the input video with @command{ffmpeg}:
11068 ffmpeg -i in.avi -vf "hflip" out.avi
11072 This filter applies a global color histogram equalization on a
11075 It can be used to correct video that has a compressed range of pixel
11076 intensities. The filter redistributes the pixel intensities to
11077 equalize their distribution across the intensity range. It may be
11078 viewed as an "automatically adjusting contrast filter". This filter is
11079 useful only for correcting degraded or poorly captured source
11082 The filter accepts the following options:
11086 Determine the amount of equalization to be applied. As the strength
11087 is reduced, the distribution of pixel intensities more-and-more
11088 approaches that of the input frame. The value must be a float number
11089 in the range [0,1] and defaults to 0.200.
11092 Set the maximum intensity that can generated and scale the output
11093 values appropriately. The strength should be set as desired and then
11094 the intensity can be limited if needed to avoid washing-out. The value
11095 must be a float number in the range [0,1] and defaults to 0.210.
11098 Set the antibanding level. If enabled the filter will randomly vary
11099 the luminance of output pixels by a small amount to avoid banding of
11100 the histogram. Possible values are @code{none}, @code{weak} or
11101 @code{strong}. It defaults to @code{none}.
11106 Compute and draw a color distribution histogram for the input video.
11108 The computed histogram is a representation of the color component
11109 distribution in an image.
11111 Standard histogram displays the color components distribution in an image.
11112 Displays color graph for each color component. Shows distribution of
11113 the Y, U, V, A or R, G, B components, depending on input format, in the
11114 current frame. Below each graph a color component scale meter is shown.
11116 The filter accepts the following options:
11120 Set height of level. Default value is @code{200}.
11121 Allowed range is [50, 2048].
11124 Set height of color scale. Default value is @code{12}.
11125 Allowed range is [0, 40].
11129 It accepts the following values:
11132 Per color component graphs are placed below each other.
11135 Per color component graphs are placed side by side.
11138 Presents information identical to that in the @code{parade}, except
11139 that the graphs representing color components are superimposed directly
11142 Default is @code{stack}.
11145 Set mode. Can be either @code{linear}, or @code{logarithmic}.
11146 Default is @code{linear}.
11149 Set what color components to display.
11150 Default is @code{7}.
11153 Set foreground opacity. Default is @code{0.7}.
11156 Set background opacity. Default is @code{0.5}.
11159 @subsection Examples
11164 Calculate and draw histogram:
11166 ffplay -i input -vf histogram
11174 This is a high precision/quality 3d denoise filter. It aims to reduce
11175 image noise, producing smooth images and making still images really
11176 still. It should enhance compressibility.
11178 It accepts the following optional parameters:
11182 A non-negative floating point number which specifies spatial luma strength.
11183 It defaults to 4.0.
11185 @item chroma_spatial
11186 A non-negative floating point number which specifies spatial chroma strength.
11187 It defaults to 3.0*@var{luma_spatial}/4.0.
11190 A floating point number which specifies luma temporal strength. It defaults to
11191 6.0*@var{luma_spatial}/4.0.
11194 A floating point number which specifies chroma temporal strength. It defaults to
11195 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
11198 @anchor{hwdownload}
11199 @section hwdownload
11201 Download hardware frames to system memory.
11203 The input must be in hardware frames, and the output a non-hardware format.
11204 Not all formats will be supported on the output - it may be necessary to insert
11205 an additional @option{format} filter immediately following in the graph to get
11206 the output in a supported format.
11210 Map hardware frames to system memory or to another device.
11212 This filter has several different modes of operation; which one is used depends
11213 on the input and output formats:
11216 Hardware frame input, normal frame output
11218 Map the input frames to system memory and pass them to the output. If the
11219 original hardware frame is later required (for example, after overlaying
11220 something else on part of it), the @option{hwmap} filter can be used again
11221 in the next mode to retrieve it.
11223 Normal frame input, hardware frame output
11225 If the input is actually a software-mapped hardware frame, then unmap it -
11226 that is, return the original hardware frame.
11228 Otherwise, a device must be provided. Create new hardware surfaces on that
11229 device for the output, then map them back to the software format at the input
11230 and give those frames to the preceding filter. This will then act like the
11231 @option{hwupload} filter, but may be able to avoid an additional copy when
11232 the input is already in a compatible format.
11234 Hardware frame input and output
11236 A device must be supplied for the output, either directly or with the
11237 @option{derive_device} option. The input and output devices must be of
11238 different types and compatible - the exact meaning of this is
11239 system-dependent, but typically it means that they must refer to the same
11240 underlying hardware context (for example, refer to the same graphics card).
11242 If the input frames were originally created on the output device, then unmap
11243 to retrieve the original frames.
11245 Otherwise, map the frames to the output device - create new hardware frames
11246 on the output corresponding to the frames on the input.
11249 The following additional parameters are accepted:
11253 Set the frame mapping mode. Some combination of:
11256 The mapped frame should be readable.
11258 The mapped frame should be writeable.
11260 The mapping will always overwrite the entire frame.
11262 This may improve performance in some cases, as the original contents of the
11263 frame need not be loaded.
11265 The mapping must not involve any copying.
11267 Indirect mappings to copies of frames are created in some cases where either
11268 direct mapping is not possible or it would have unexpected properties.
11269 Setting this flag ensures that the mapping is direct and will fail if that is
11272 Defaults to @var{read+write} if not specified.
11274 @item derive_device @var{type}
11275 Rather than using the device supplied at initialisation, instead derive a new
11276 device of type @var{type} from the device the input frames exist on.
11279 In a hardware to hardware mapping, map in reverse - create frames in the sink
11280 and map them back to the source. This may be necessary in some cases where
11281 a mapping in one direction is required but only the opposite direction is
11282 supported by the devices being used.
11284 This option is dangerous - it may break the preceding filter in undefined
11285 ways if there are any additional constraints on that filter's output.
11286 Do not use it without fully understanding the implications of its use.
11292 Upload system memory frames to hardware surfaces.
11294 The device to upload to must be supplied when the filter is initialised. If
11295 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
11298 @anchor{hwupload_cuda}
11299 @section hwupload_cuda
11301 Upload system memory frames to a CUDA device.
11303 It accepts the following optional parameters:
11307 The number of the CUDA device to use
11312 Apply a high-quality magnification filter designed for pixel art. This filter
11313 was originally created by Maxim Stepin.
11315 It accepts the following option:
11319 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
11320 @code{hq3x} and @code{4} for @code{hq4x}.
11321 Default is @code{3}.
11325 Stack input videos horizontally.
11327 All streams must be of same pixel format and of same height.
11329 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
11330 to create same output.
11332 The filter accept the following option:
11336 Set number of input streams. Default is 2.
11339 If set to 1, force the output to terminate when the shortest input
11340 terminates. Default value is 0.
11345 Modify the hue and/or the saturation of the input.
11347 It accepts the following parameters:
11351 Specify the hue angle as a number of degrees. It accepts an expression,
11352 and defaults to "0".
11355 Specify the saturation in the [-10,10] range. It accepts an expression and
11359 Specify the hue angle as a number of radians. It accepts an
11360 expression, and defaults to "0".
11363 Specify the brightness in the [-10,10] range. It accepts an expression and
11367 @option{h} and @option{H} are mutually exclusive, and can't be
11368 specified at the same time.
11370 The @option{b}, @option{h}, @option{H} and @option{s} option values are
11371 expressions containing the following constants:
11375 frame count of the input frame starting from 0
11378 presentation timestamp of the input frame expressed in time base units
11381 frame rate of the input video, NAN if the input frame rate is unknown
11384 timestamp expressed in seconds, NAN if the input timestamp is unknown
11387 time base of the input video
11390 @subsection Examples
11394 Set the hue to 90 degrees and the saturation to 1.0:
11400 Same command but expressing the hue in radians:
11406 Rotate hue and make the saturation swing between 0
11407 and 2 over a period of 1 second:
11409 hue="H=2*PI*t: s=sin(2*PI*t)+1"
11413 Apply a 3 seconds saturation fade-in effect starting at 0:
11415 hue="s=min(t/3\,1)"
11418 The general fade-in expression can be written as:
11420 hue="s=min(0\, max((t-START)/DURATION\, 1))"
11424 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
11426 hue="s=max(0\, min(1\, (8-t)/3))"
11429 The general fade-out expression can be written as:
11431 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
11436 @subsection Commands
11438 This filter supports the following commands:
11444 Modify the hue and/or the saturation and/or brightness of the input video.
11445 The command accepts the same syntax of the corresponding option.
11447 If the specified expression is not valid, it is kept at its current
11451 @section hysteresis
11453 Grow first stream into second stream by connecting components.
11454 This makes it possible to build more robust edge masks.
11456 This filter accepts the following options:
11460 Set which planes will be processed as bitmap, unprocessed planes will be
11461 copied from first stream.
11462 By default value 0xf, all planes will be processed.
11465 Set threshold which is used in filtering. If pixel component value is higher than
11466 this value filter algorithm for connecting components is activated.
11467 By default value is 0.
11472 Detect video interlacing type.
11474 This filter tries to detect if the input frames are interlaced, progressive,
11475 top or bottom field first. It will also try to detect fields that are
11476 repeated between adjacent frames (a sign of telecine).
11478 Single frame detection considers only immediately adjacent frames when classifying each frame.
11479 Multiple frame detection incorporates the classification history of previous frames.
11481 The filter will log these metadata values:
11484 @item single.current_frame
11485 Detected type of current frame using single-frame detection. One of:
11486 ``tff'' (top field first), ``bff'' (bottom field first),
11487 ``progressive'', or ``undetermined''
11490 Cumulative number of frames detected as top field first using single-frame detection.
11493 Cumulative number of frames detected as top field first using multiple-frame detection.
11496 Cumulative number of frames detected as bottom field first using single-frame detection.
11498 @item multiple.current_frame
11499 Detected type of current frame using multiple-frame detection. One of:
11500 ``tff'' (top field first), ``bff'' (bottom field first),
11501 ``progressive'', or ``undetermined''
11504 Cumulative number of frames detected as bottom field first using multiple-frame detection.
11506 @item single.progressive
11507 Cumulative number of frames detected as progressive using single-frame detection.
11509 @item multiple.progressive
11510 Cumulative number of frames detected as progressive using multiple-frame detection.
11512 @item single.undetermined
11513 Cumulative number of frames that could not be classified using single-frame detection.
11515 @item multiple.undetermined
11516 Cumulative number of frames that could not be classified using multiple-frame detection.
11518 @item repeated.current_frame
11519 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
11521 @item repeated.neither
11522 Cumulative number of frames with no repeated field.
11525 Cumulative number of frames with the top field repeated from the previous frame's top field.
11527 @item repeated.bottom
11528 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
11531 The filter accepts the following options:
11535 Set interlacing threshold.
11537 Set progressive threshold.
11539 Threshold for repeated field detection.
11541 Number of frames after which a given frame's contribution to the
11542 statistics is halved (i.e., it contributes only 0.5 to its
11543 classification). The default of 0 means that all frames seen are given
11544 full weight of 1.0 forever.
11545 @item analyze_interlaced_flag
11546 When this is not 0 then idet will use the specified number of frames to determine
11547 if the interlaced flag is accurate, it will not count undetermined frames.
11548 If the flag is found to be accurate it will be used without any further
11549 computations, if it is found to be inaccurate it will be cleared without any
11550 further computations. This allows inserting the idet filter as a low computational
11551 method to clean up the interlaced flag
11556 Deinterleave or interleave fields.
11558 This filter allows one to process interlaced images fields without
11559 deinterlacing them. Deinterleaving splits the input frame into 2
11560 fields (so called half pictures). Odd lines are moved to the top
11561 half of the output image, even lines to the bottom half.
11562 You can process (filter) them independently and then re-interleave them.
11564 The filter accepts the following options:
11568 @item chroma_mode, c
11569 @item alpha_mode, a
11570 Available values for @var{luma_mode}, @var{chroma_mode} and
11571 @var{alpha_mode} are:
11577 @item deinterleave, d
11578 Deinterleave fields, placing one above the other.
11580 @item interleave, i
11581 Interleave fields. Reverse the effect of deinterleaving.
11583 Default value is @code{none}.
11585 @item luma_swap, ls
11586 @item chroma_swap, cs
11587 @item alpha_swap, as
11588 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
11593 Apply inflate effect to the video.
11595 This filter replaces the pixel by the local(3x3) average by taking into account
11596 only values higher than the pixel.
11598 It accepts the following options:
11605 Limit the maximum change for each plane, default is 65535.
11606 If 0, plane will remain unchanged.
11611 Simple interlacing filter from progressive contents. This interleaves upper (or
11612 lower) lines from odd frames with lower (or upper) lines from even frames,
11613 halving the frame rate and preserving image height.
11616 Original Original New Frame
11617 Frame 'j' Frame 'j+1' (tff)
11618 ========== =========== ==================
11619 Line 0 --------------------> Frame 'j' Line 0
11620 Line 1 Line 1 ----> Frame 'j+1' Line 1
11621 Line 2 ---------------------> Frame 'j' Line 2
11622 Line 3 Line 3 ----> Frame 'j+1' Line 3
11624 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
11627 It accepts the following optional parameters:
11631 This determines whether the interlaced frame is taken from the even
11632 (tff - default) or odd (bff) lines of the progressive frame.
11635 Vertical lowpass filter to avoid twitter interlacing and
11636 reduce moire patterns.
11640 Disable vertical lowpass filter
11643 Enable linear filter (default)
11646 Enable complex filter. This will slightly less reduce twitter and moire
11647 but better retain detail and subjective sharpness impression.
11654 Deinterlace input video by applying Donald Graft's adaptive kernel
11655 deinterling. Work on interlaced parts of a video to produce
11656 progressive frames.
11658 The description of the accepted parameters follows.
11662 Set the threshold which affects the filter's tolerance when
11663 determining if a pixel line must be processed. It must be an integer
11664 in the range [0,255] and defaults to 10. A value of 0 will result in
11665 applying the process on every pixels.
11668 Paint pixels exceeding the threshold value to white if set to 1.
11672 Set the fields order. Swap fields if set to 1, leave fields alone if
11676 Enable additional sharpening if set to 1. Default is 0.
11679 Enable twoway sharpening if set to 1. Default is 0.
11682 @subsection Examples
11686 Apply default values:
11688 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
11692 Enable additional sharpening:
11698 Paint processed pixels in white:
11706 Slowly update darker pixels.
11708 This filter makes short flashes of light appear longer.
11709 This filter accepts the following options:
11713 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
11716 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
11719 @section lenscorrection
11721 Correct radial lens distortion
11723 This filter can be used to correct for radial distortion as can result from the use
11724 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
11725 one can use tools available for example as part of opencv or simply trial-and-error.
11726 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
11727 and extract the k1 and k2 coefficients from the resulting matrix.
11729 Note that effectively the same filter is available in the open-source tools Krita and
11730 Digikam from the KDE project.
11732 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
11733 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
11734 brightness distribution, so you may want to use both filters together in certain
11735 cases, though you will have to take care of ordering, i.e. whether vignetting should
11736 be applied before or after lens correction.
11738 @subsection Options
11740 The filter accepts the following options:
11744 Relative x-coordinate of the focal point of the image, and thereby the center of the
11745 distortion. This value has a range [0,1] and is expressed as fractions of the image
11746 width. Default is 0.5.
11748 Relative y-coordinate of the focal point of the image, and thereby the center of the
11749 distortion. This value has a range [0,1] and is expressed as fractions of the image
11750 height. Default is 0.5.
11752 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
11753 no correction. Default is 0.
11755 Coefficient of the double quadratic correction term. This value has a range [-1,1].
11756 0 means no correction. Default is 0.
11759 The formula that generates the correction is:
11761 @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)
11763 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
11764 distances from the focal point in the source and target images, respectively.
11768 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
11770 The @code{lensfun} filter requires the camera make, camera model, and lens model
11771 to apply the lens correction. The filter will load the lensfun database and
11772 query it to find the corresponding camera and lens entries in the database. As
11773 long as these entries can be found with the given options, the filter can
11774 perform corrections on frames. Note that incomplete strings will result in the
11775 filter choosing the best match with the given options, and the filter will
11776 output the chosen camera and lens models (logged with level "info"). You must
11777 provide the make, camera model, and lens model as they are required.
11779 The filter accepts the following options:
11783 The make of the camera (for example, "Canon"). This option is required.
11786 The model of the camera (for example, "Canon EOS 100D"). This option is
11790 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
11791 option is required.
11794 The type of correction to apply. The following values are valid options:
11798 Enables fixing lens vignetting.
11801 Enables fixing lens geometry. This is the default.
11804 Enables fixing chromatic aberrations.
11807 Enables fixing lens vignetting and lens geometry.
11810 Enables fixing lens vignetting and chromatic aberrations.
11813 Enables fixing both lens geometry and chromatic aberrations.
11816 Enables all possible corrections.
11820 The focal length of the image/video (zoom; expected constant for video). For
11821 example, a 18--55mm lens has focal length range of [18--55], so a value in that
11822 range should be chosen when using that lens. Default 18.
11825 The aperture of the image/video (expected constant for video). Note that
11826 aperture is only used for vignetting correction. Default 3.5.
11828 @item focus_distance
11829 The focus distance of the image/video (expected constant for video). Note that
11830 focus distance is only used for vignetting and only slightly affects the
11831 vignetting correction process. If unknown, leave it at the default value (which
11835 The scale factor which is applied after transformation. After correction the
11836 video is no longer necessarily rectangular. This parameter controls how much of
11837 the resulting image is visible. The value 0 means that a value will be chosen
11838 automatically such that there is little or no unmapped area in the output
11839 image. 1.0 means that no additional scaling is done. Lower values may result
11840 in more of the corrected image being visible, while higher values may avoid
11841 unmapped areas in the output.
11843 @item target_geometry
11844 The target geometry of the output image/video. The following values are valid
11848 @item rectilinear (default)
11851 @item equirectangular
11852 @item fisheye_orthographic
11853 @item fisheye_stereographic
11854 @item fisheye_equisolid
11855 @item fisheye_thoby
11858 Apply the reverse of image correction (instead of correcting distortion, apply
11861 @item interpolation
11862 The type of interpolation used when correcting distortion. The following values
11867 @item linear (default)
11872 @subsection Examples
11876 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
11877 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
11881 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
11885 Apply the same as before, but only for the first 5 seconds of video.
11888 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
11895 Obtain the VMAF (Video Multi-Method Assessment Fusion)
11896 score between two input videos.
11898 The obtained VMAF score is printed through the logging system.
11900 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
11901 After installing the library it can be enabled using:
11902 @code{./configure --enable-libvmaf --enable-version3}.
11903 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
11905 The filter has following options:
11909 Set the model path which is to be used for SVM.
11910 Default value: @code{"vmaf_v0.6.1.pkl"}
11913 Set the file path to be used to store logs.
11916 Set the format of the log file (xml or json).
11918 @item enable_transform
11919 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
11920 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
11921 Default value: @code{false}
11924 Invokes the phone model which will generate VMAF scores higher than in the
11925 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
11928 Enables computing psnr along with vmaf.
11931 Enables computing ssim along with vmaf.
11934 Enables computing ms_ssim along with vmaf.
11937 Set the pool method (mean, min or harmonic mean) to be used for computing vmaf.
11940 Set number of threads to be used when computing vmaf.
11943 Set interval for frame subsampling used when computing vmaf.
11945 @item enable_conf_interval
11946 Enables confidence interval.
11949 This filter also supports the @ref{framesync} options.
11951 On the below examples the input file @file{main.mpg} being processed is
11952 compared with the reference file @file{ref.mpg}.
11955 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
11958 Example with options:
11960 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
11965 Limits the pixel components values to the specified range [min, max].
11967 The filter accepts the following options:
11971 Lower bound. Defaults to the lowest allowed value for the input.
11974 Upper bound. Defaults to the highest allowed value for the input.
11977 Specify which planes will be processed. Defaults to all available.
11984 The filter accepts the following options:
11988 Set the number of loops. Setting this value to -1 will result in infinite loops.
11992 Set maximal size in number of frames. Default is 0.
11995 Set first frame of loop. Default is 0.
11998 @subsection Examples
12002 Loop single first frame infinitely:
12004 loop=loop=-1:size=1:start=0
12008 Loop single first frame 10 times:
12010 loop=loop=10:size=1:start=0
12014 Loop 10 first frames 5 times:
12016 loop=loop=5:size=10:start=0
12022 Apply a 1D LUT to an input video.
12024 The filter accepts the following options:
12028 Set the 1D LUT file name.
12030 Currently supported formats:
12039 Select interpolation mode.
12041 Available values are:
12045 Use values from the nearest defined point.
12047 Interpolate values using the linear interpolation.
12049 Interpolate values using the cosine interpolation.
12051 Interpolate values using the cubic interpolation.
12053 Interpolate values using the spline interpolation.
12060 Apply a 3D LUT to an input video.
12062 The filter accepts the following options:
12066 Set the 3D LUT file name.
12068 Currently supported formats:
12082 Select interpolation mode.
12084 Available values are:
12088 Use values from the nearest defined point.
12090 Interpolate values using the 8 points defining a cube.
12092 Interpolate values using a tetrahedron.
12098 Turn certain luma values into transparency.
12100 The filter accepts the following options:
12104 Set the luma which will be used as base for transparency.
12105 Default value is @code{0}.
12108 Set the range of luma values to be keyed out.
12109 Default value is @code{0}.
12112 Set the range of softness. Default value is @code{0}.
12113 Use this to control gradual transition from zero to full transparency.
12116 @section lut, lutrgb, lutyuv
12118 Compute a look-up table for binding each pixel component input value
12119 to an output value, and apply it to the input video.
12121 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
12122 to an RGB input video.
12124 These filters accept the following parameters:
12127 set first pixel component expression
12129 set second pixel component expression
12131 set third pixel component expression
12133 set fourth pixel component expression, corresponds to the alpha component
12136 set red component expression
12138 set green component expression
12140 set blue component expression
12142 alpha component expression
12145 set Y/luminance component expression
12147 set U/Cb component expression
12149 set V/Cr component expression
12152 Each of them specifies the expression to use for computing the lookup table for
12153 the corresponding pixel component values.
12155 The exact component associated to each of the @var{c*} options depends on the
12158 The @var{lut} filter requires either YUV or RGB pixel formats in input,
12159 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
12161 The expressions can contain the following constants and functions:
12166 The input width and height.
12169 The input value for the pixel component.
12172 The input value, clipped to the @var{minval}-@var{maxval} range.
12175 The maximum value for the pixel component.
12178 The minimum value for the pixel component.
12181 The negated value for the pixel component value, clipped to the
12182 @var{minval}-@var{maxval} range; it corresponds to the expression
12183 "maxval-clipval+minval".
12186 The computed value in @var{val}, clipped to the
12187 @var{minval}-@var{maxval} range.
12189 @item gammaval(gamma)
12190 The computed gamma correction value of the pixel component value,
12191 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
12193 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
12197 All expressions default to "val".
12199 @subsection Examples
12203 Negate input video:
12205 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
12206 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
12209 The above is the same as:
12211 lutrgb="r=negval:g=negval:b=negval"
12212 lutyuv="y=negval:u=negval:v=negval"
12222 Remove chroma components, turning the video into a graytone image:
12224 lutyuv="u=128:v=128"
12228 Apply a luma burning effect:
12234 Remove green and blue components:
12240 Set a constant alpha channel value on input:
12242 format=rgba,lutrgb=a="maxval-minval/2"
12246 Correct luminance gamma by a factor of 0.5:
12248 lutyuv=y=gammaval(0.5)
12252 Discard least significant bits of luma:
12254 lutyuv=y='bitand(val, 128+64+32)'
12258 Technicolor like effect:
12260 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
12264 @section lut2, tlut2
12266 The @code{lut2} filter takes two input streams and outputs one
12269 The @code{tlut2} (time lut2) filter takes two consecutive frames
12270 from one single stream.
12272 This filter accepts the following parameters:
12275 set first pixel component expression
12277 set second pixel component expression
12279 set third pixel component expression
12281 set fourth pixel component expression, corresponds to the alpha component
12284 set output bit depth, only available for @code{lut2} filter. By default is 0,
12285 which means bit depth is automatically picked from first input format.
12288 Each of them specifies the expression to use for computing the lookup table for
12289 the corresponding pixel component values.
12291 The exact component associated to each of the @var{c*} options depends on the
12294 The expressions can contain the following constants:
12299 The input width and height.
12302 The first input value for the pixel component.
12305 The second input value for the pixel component.
12308 The first input video bit depth.
12311 The second input video bit depth.
12314 All expressions default to "x".
12316 @subsection Examples
12320 Highlight differences between two RGB video streams:
12322 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)'
12326 Highlight differences between two YUV video streams:
12328 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)'
12332 Show max difference between two video streams:
12334 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)))'
12338 @section maskedclamp
12340 Clamp the first input stream with the second input and third input stream.
12342 Returns the value of first stream to be between second input
12343 stream - @code{undershoot} and third input stream + @code{overshoot}.
12345 This filter accepts the following options:
12348 Default value is @code{0}.
12351 Default value is @code{0}.
12354 Set which planes will be processed as bitmap, unprocessed planes will be
12355 copied from first stream.
12356 By default value 0xf, all planes will be processed.
12359 @section maskedmerge
12361 Merge the first input stream with the second input stream using per pixel
12362 weights in the third input stream.
12364 A value of 0 in the third stream pixel component means that pixel component
12365 from first stream is returned unchanged, while maximum value (eg. 255 for
12366 8-bit videos) means that pixel component from second stream is returned
12367 unchanged. Intermediate values define the amount of merging between both
12368 input stream's pixel components.
12370 This filter accepts the following options:
12373 Set which planes will be processed as bitmap, unprocessed planes will be
12374 copied from first stream.
12375 By default value 0xf, all planes will be processed.
12379 Create mask from input video.
12381 For example it is useful to create motion masks after @code{tblend} filter.
12383 This filter accepts the following options:
12387 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
12390 Set high threshold. Any pixel component higher than this value will be set to max value
12391 allowed for current pixel format.
12394 Set planes to filter, by default all available planes are filtered.
12397 Fill all frame pixels with this value.
12400 Set max average pixel value for frame. If sum of all pixel components is higher that this
12401 average, output frame will be completely filled with value set by @var{fill} option.
12402 Typically useful for scene changes when used in combination with @code{tblend} filter.
12407 Apply motion-compensation deinterlacing.
12409 It needs one field per frame as input and must thus be used together
12410 with yadif=1/3 or equivalent.
12412 This filter accepts the following options:
12415 Set the deinterlacing mode.
12417 It accepts one of the following values:
12422 use iterative motion estimation
12424 like @samp{slow}, but use multiple reference frames.
12426 Default value is @samp{fast}.
12429 Set the picture field parity assumed for the input video. It must be
12430 one of the following values:
12434 assume top field first
12436 assume bottom field first
12439 Default value is @samp{bff}.
12442 Set per-block quantization parameter (QP) used by the internal
12445 Higher values should result in a smoother motion vector field but less
12446 optimal individual vectors. Default value is 1.
12449 @section mergeplanes
12451 Merge color channel components from several video streams.
12453 The filter accepts up to 4 input streams, and merge selected input
12454 planes to the output video.
12456 This filter accepts the following options:
12459 Set input to output plane mapping. Default is @code{0}.
12461 The mappings is specified as a bitmap. It should be specified as a
12462 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
12463 mapping for the first plane of the output stream. 'A' sets the number of
12464 the input stream to use (from 0 to 3), and 'a' the plane number of the
12465 corresponding input to use (from 0 to 3). The rest of the mappings is
12466 similar, 'Bb' describes the mapping for the output stream second
12467 plane, 'Cc' describes the mapping for the output stream third plane and
12468 'Dd' describes the mapping for the output stream fourth plane.
12471 Set output pixel format. Default is @code{yuva444p}.
12474 @subsection Examples
12478 Merge three gray video streams of same width and height into single video stream:
12480 [a0][a1][a2]mergeplanes=0x001020:yuv444p
12484 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
12486 [a0][a1]mergeplanes=0x00010210:yuva444p
12490 Swap Y and A plane in yuva444p stream:
12492 format=yuva444p,mergeplanes=0x03010200:yuva444p
12496 Swap U and V plane in yuv420p stream:
12498 format=yuv420p,mergeplanes=0x000201:yuv420p
12502 Cast a rgb24 clip to yuv444p:
12504 format=rgb24,mergeplanes=0x000102:yuv444p
12510 Estimate and export motion vectors using block matching algorithms.
12511 Motion vectors are stored in frame side data to be used by other filters.
12513 This filter accepts the following options:
12516 Specify the motion estimation method. Accepts one of the following values:
12520 Exhaustive search algorithm.
12522 Three step search algorithm.
12524 Two dimensional logarithmic search algorithm.
12526 New three step search algorithm.
12528 Four step search algorithm.
12530 Diamond search algorithm.
12532 Hexagon-based search algorithm.
12534 Enhanced predictive zonal search algorithm.
12536 Uneven multi-hexagon search algorithm.
12538 Default value is @samp{esa}.
12541 Macroblock size. Default @code{16}.
12544 Search parameter. Default @code{7}.
12547 @section midequalizer
12549 Apply Midway Image Equalization effect using two video streams.
12551 Midway Image Equalization adjusts a pair of images to have the same
12552 histogram, while maintaining their dynamics as much as possible. It's
12553 useful for e.g. matching exposures from a pair of stereo cameras.
12555 This filter has two inputs and one output, which must be of same pixel format, but
12556 may be of different sizes. The output of filter is first input adjusted with
12557 midway histogram of both inputs.
12559 This filter accepts the following option:
12563 Set which planes to process. Default is @code{15}, which is all available planes.
12566 @section minterpolate
12568 Convert the video to specified frame rate using motion interpolation.
12570 This filter accepts the following options:
12573 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}.
12576 Motion interpolation mode. Following values are accepted:
12579 Duplicate previous or next frame for interpolating new ones.
12581 Blend source frames. Interpolated frame is mean of previous and next frames.
12583 Motion compensated interpolation. Following options are effective when this mode is selected:
12587 Motion compensation mode. Following values are accepted:
12590 Overlapped block motion compensation.
12592 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
12594 Default mode is @samp{obmc}.
12597 Motion estimation mode. Following values are accepted:
12600 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
12602 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
12604 Default mode is @samp{bilat}.
12607 The algorithm to be used for motion estimation. Following values are accepted:
12610 Exhaustive search algorithm.
12612 Three step search algorithm.
12614 Two dimensional logarithmic search algorithm.
12616 New three step search algorithm.
12618 Four step search algorithm.
12620 Diamond search algorithm.
12622 Hexagon-based search algorithm.
12624 Enhanced predictive zonal search algorithm.
12626 Uneven multi-hexagon search algorithm.
12628 Default algorithm is @samp{epzs}.
12631 Macroblock size. Default @code{16}.
12634 Motion estimation search parameter. Default @code{32}.
12637 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).
12642 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:
12645 Disable scene change detection.
12647 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
12649 Default method is @samp{fdiff}.
12651 @item scd_threshold
12652 Scene change detection threshold. Default is @code{5.0}.
12657 Mix several video input streams into one video stream.
12659 A description of the accepted options follows.
12663 The number of inputs. If unspecified, it defaults to 2.
12666 Specify weight of each input video stream as sequence.
12667 Each weight is separated by space. If number of weights
12668 is smaller than number of @var{frames} last specified
12669 weight will be used for all remaining unset weights.
12672 Specify scale, if it is set it will be multiplied with sum
12673 of each weight multiplied with pixel values to give final destination
12674 pixel value. By default @var{scale} is auto scaled to sum of weights.
12677 Specify how end of stream is determined.
12680 The duration of the longest input. (default)
12683 The duration of the shortest input.
12686 The duration of the first input.
12690 @section mpdecimate
12692 Drop frames that do not differ greatly from the previous frame in
12693 order to reduce frame rate.
12695 The main use of this filter is for very-low-bitrate encoding
12696 (e.g. streaming over dialup modem), but it could in theory be used for
12697 fixing movies that were inverse-telecined incorrectly.
12699 A description of the accepted options follows.
12703 Set the maximum number of consecutive frames which can be dropped (if
12704 positive), or the minimum interval between dropped frames (if
12705 negative). If the value is 0, the frame is dropped disregarding the
12706 number of previous sequentially dropped frames.
12708 Default value is 0.
12713 Set the dropping threshold values.
12715 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
12716 represent actual pixel value differences, so a threshold of 64
12717 corresponds to 1 unit of difference for each pixel, or the same spread
12718 out differently over the block.
12720 A frame is a candidate for dropping if no 8x8 blocks differ by more
12721 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
12722 meaning the whole image) differ by more than a threshold of @option{lo}.
12724 Default value for @option{hi} is 64*12, default value for @option{lo} is
12725 64*5, and default value for @option{frac} is 0.33.
12731 Negate (invert) the input video.
12733 It accepts the following option:
12738 With value 1, it negates the alpha component, if present. Default value is 0.
12744 Denoise frames using Non-Local Means algorithm.
12746 Each pixel is adjusted by looking for other pixels with similar contexts. This
12747 context similarity is defined by comparing their surrounding patches of size
12748 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
12751 Note that the research area defines centers for patches, which means some
12752 patches will be made of pixels outside that research area.
12754 The filter accepts the following options.
12758 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
12761 Set patch size. Default is 7. Must be odd number in range [0, 99].
12764 Same as @option{p} but for chroma planes.
12766 The default value is @var{0} and means automatic.
12769 Set research size. Default is 15. Must be odd number in range [0, 99].
12772 Same as @option{r} but for chroma planes.
12774 The default value is @var{0} and means automatic.
12779 Deinterlace video using neural network edge directed interpolation.
12781 This filter accepts the following options:
12785 Mandatory option, without binary file filter can not work.
12786 Currently file can be found here:
12787 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
12790 Set which frames to deinterlace, by default it is @code{all}.
12791 Can be @code{all} or @code{interlaced}.
12794 Set mode of operation.
12796 Can be one of the following:
12800 Use frame flags, both fields.
12802 Use frame flags, single field.
12804 Use top field only.
12806 Use bottom field only.
12808 Use both fields, top first.
12810 Use both fields, bottom first.
12814 Set which planes to process, by default filter process all frames.
12817 Set size of local neighborhood around each pixel, used by the predictor neural
12820 Can be one of the following:
12833 Set the number of neurons in predictor neural network.
12834 Can be one of the following:
12845 Controls the number of different neural network predictions that are blended
12846 together to compute the final output value. Can be @code{fast}, default or
12850 Set which set of weights to use in the predictor.
12851 Can be one of the following:
12855 weights trained to minimize absolute error
12857 weights trained to minimize squared error
12861 Controls whether or not the prescreener neural network is used to decide
12862 which pixels should be processed by the predictor neural network and which
12863 can be handled by simple cubic interpolation.
12864 The prescreener is trained to know whether cubic interpolation will be
12865 sufficient for a pixel or whether it should be predicted by the predictor nn.
12866 The computational complexity of the prescreener nn is much less than that of
12867 the predictor nn. Since most pixels can be handled by cubic interpolation,
12868 using the prescreener generally results in much faster processing.
12869 The prescreener is pretty accurate, so the difference between using it and not
12870 using it is almost always unnoticeable.
12872 Can be one of the following:
12880 Default is @code{new}.
12883 Set various debugging flags.
12888 Force libavfilter not to use any of the specified pixel formats for the
12889 input to the next filter.
12891 It accepts the following parameters:
12895 A '|'-separated list of pixel format names, such as
12896 pix_fmts=yuv420p|monow|rgb24".
12900 @subsection Examples
12904 Force libavfilter to use a format different from @var{yuv420p} for the
12905 input to the vflip filter:
12907 noformat=pix_fmts=yuv420p,vflip
12911 Convert the input video to any of the formats not contained in the list:
12913 noformat=yuv420p|yuv444p|yuv410p
12919 Add noise on video input frame.
12921 The filter accepts the following options:
12929 Set noise seed for specific pixel component or all pixel components in case
12930 of @var{all_seed}. Default value is @code{123457}.
12932 @item all_strength, alls
12933 @item c0_strength, c0s
12934 @item c1_strength, c1s
12935 @item c2_strength, c2s
12936 @item c3_strength, c3s
12937 Set noise strength for specific pixel component or all pixel components in case
12938 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
12940 @item all_flags, allf
12941 @item c0_flags, c0f
12942 @item c1_flags, c1f
12943 @item c2_flags, c2f
12944 @item c3_flags, c3f
12945 Set pixel component flags or set flags for all components if @var{all_flags}.
12946 Available values for component flags are:
12949 averaged temporal noise (smoother)
12951 mix random noise with a (semi)regular pattern
12953 temporal noise (noise pattern changes between frames)
12955 uniform noise (gaussian otherwise)
12959 @subsection Examples
12961 Add temporal and uniform noise to input video:
12963 noise=alls=20:allf=t+u
12968 Normalize RGB video (aka histogram stretching, contrast stretching).
12969 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
12971 For each channel of each frame, the filter computes the input range and maps
12972 it linearly to the user-specified output range. The output range defaults
12973 to the full dynamic range from pure black to pure white.
12975 Temporal smoothing can be used on the input range to reduce flickering (rapid
12976 changes in brightness) caused when small dark or bright objects enter or leave
12977 the scene. This is similar to the auto-exposure (automatic gain control) on a
12978 video camera, and, like a video camera, it may cause a period of over- or
12979 under-exposure of the video.
12981 The R,G,B channels can be normalized independently, which may cause some
12982 color shifting, or linked together as a single channel, which prevents
12983 color shifting. Linked normalization preserves hue. Independent normalization
12984 does not, so it can be used to remove some color casts. Independent and linked
12985 normalization can be combined in any ratio.
12987 The normalize filter accepts the following options:
12992 Colors which define the output range. The minimum input value is mapped to
12993 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
12994 The defaults are black and white respectively. Specifying white for
12995 @var{blackpt} and black for @var{whitept} will give color-inverted,
12996 normalized video. Shades of grey can be used to reduce the dynamic range
12997 (contrast). Specifying saturated colors here can create some interesting
13001 The number of previous frames to use for temporal smoothing. The input range
13002 of each channel is smoothed using a rolling average over the current frame
13003 and the @var{smoothing} previous frames. The default is 0 (no temporal
13007 Controls the ratio of independent (color shifting) channel normalization to
13008 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
13009 independent. Defaults to 1.0 (fully independent).
13012 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
13013 expensive no-op. Defaults to 1.0 (full strength).
13017 @subsection Examples
13019 Stretch video contrast to use the full dynamic range, with no temporal
13020 smoothing; may flicker depending on the source content:
13022 normalize=blackpt=black:whitept=white:smoothing=0
13025 As above, but with 50 frames of temporal smoothing; flicker should be
13026 reduced, depending on the source content:
13028 normalize=blackpt=black:whitept=white:smoothing=50
13031 As above, but with hue-preserving linked channel normalization:
13033 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
13036 As above, but with half strength:
13038 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
13041 Map the darkest input color to red, the brightest input color to cyan:
13043 normalize=blackpt=red:whitept=cyan
13048 Pass the video source unchanged to the output.
13051 Optical Character Recognition
13053 This filter uses Tesseract for optical character recognition. To enable
13054 compilation of this filter, you need to configure FFmpeg with
13055 @code{--enable-libtesseract}.
13057 It accepts the following options:
13061 Set datapath to tesseract data. Default is to use whatever was
13062 set at installation.
13065 Set language, default is "eng".
13068 Set character whitelist.
13071 Set character blacklist.
13074 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
13075 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
13079 Apply a video transform using libopencv.
13081 To enable this filter, install the libopencv library and headers and
13082 configure FFmpeg with @code{--enable-libopencv}.
13084 It accepts the following parameters:
13089 The name of the libopencv filter to apply.
13091 @item filter_params
13092 The parameters to pass to the libopencv filter. If not specified, the default
13093 values are assumed.
13097 Refer to the official libopencv documentation for more precise
13099 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
13101 Several libopencv filters are supported; see the following subsections.
13106 Dilate an image by using a specific structuring element.
13107 It corresponds to the libopencv function @code{cvDilate}.
13109 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
13111 @var{struct_el} represents a structuring element, and has the syntax:
13112 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
13114 @var{cols} and @var{rows} represent the number of columns and rows of
13115 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
13116 point, and @var{shape} the shape for the structuring element. @var{shape}
13117 must be "rect", "cross", "ellipse", or "custom".
13119 If the value for @var{shape} is "custom", it must be followed by a
13120 string of the form "=@var{filename}". The file with name
13121 @var{filename} is assumed to represent a binary image, with each
13122 printable character corresponding to a bright pixel. When a custom
13123 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
13124 or columns and rows of the read file are assumed instead.
13126 The default value for @var{struct_el} is "3x3+0x0/rect".
13128 @var{nb_iterations} specifies the number of times the transform is
13129 applied to the image, and defaults to 1.
13133 # Use the default values
13136 # Dilate using a structuring element with a 5x5 cross, iterating two times
13137 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
13139 # Read the shape from the file diamond.shape, iterating two times.
13140 # The file diamond.shape may contain a pattern of characters like this
13146 # The specified columns and rows are ignored
13147 # but the anchor point coordinates are not
13148 ocv=dilate:0x0+2x2/custom=diamond.shape|2
13153 Erode an image by using a specific structuring element.
13154 It corresponds to the libopencv function @code{cvErode}.
13156 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
13157 with the same syntax and semantics as the @ref{dilate} filter.
13161 Smooth the input video.
13163 The filter takes the following parameters:
13164 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
13166 @var{type} is the type of smooth filter to apply, and must be one of
13167 the following values: "blur", "blur_no_scale", "median", "gaussian",
13168 or "bilateral". The default value is "gaussian".
13170 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
13171 depend on the smooth type. @var{param1} and
13172 @var{param2} accept integer positive values or 0. @var{param3} and
13173 @var{param4} accept floating point values.
13175 The default value for @var{param1} is 3. The default value for the
13176 other parameters is 0.
13178 These parameters correspond to the parameters assigned to the
13179 libopencv function @code{cvSmooth}.
13181 @section oscilloscope
13183 2D Video Oscilloscope.
13185 Useful to measure spatial impulse, step responses, chroma delays, etc.
13187 It accepts the following parameters:
13191 Set scope center x position.
13194 Set scope center y position.
13197 Set scope size, relative to frame diagonal.
13200 Set scope tilt/rotation.
13206 Set trace center x position.
13209 Set trace center y position.
13212 Set trace width, relative to width of frame.
13215 Set trace height, relative to height of frame.
13218 Set which components to trace. By default it traces first three components.
13221 Draw trace grid. By default is enabled.
13224 Draw some statistics. By default is enabled.
13227 Draw scope. By default is enabled.
13230 @subsection Examples
13234 Inspect full first row of video frame.
13236 oscilloscope=x=0.5:y=0:s=1
13240 Inspect full last row of video frame.
13242 oscilloscope=x=0.5:y=1:s=1
13246 Inspect full 5th line of video frame of height 1080.
13248 oscilloscope=x=0.5:y=5/1080:s=1
13252 Inspect full last column of video frame.
13254 oscilloscope=x=1:y=0.5:s=1:t=1
13262 Overlay one video on top of another.
13264 It takes two inputs and has one output. The first input is the "main"
13265 video on which the second input is overlaid.
13267 It accepts the following parameters:
13269 A description of the accepted options follows.
13274 Set the expression for the x and y coordinates of the overlaid video
13275 on the main video. Default value is "0" for both expressions. In case
13276 the expression is invalid, it is set to a huge value (meaning that the
13277 overlay will not be displayed within the output visible area).
13280 See @ref{framesync}.
13283 Set when the expressions for @option{x}, and @option{y} are evaluated.
13285 It accepts the following values:
13288 only evaluate expressions once during the filter initialization or
13289 when a command is processed
13292 evaluate expressions for each incoming frame
13295 Default value is @samp{frame}.
13298 See @ref{framesync}.
13301 Set the format for the output video.
13303 It accepts the following values:
13306 force YUV420 output
13309 force YUV422 output
13312 force YUV444 output
13315 force packed RGB output
13318 force planar RGB output
13321 automatically pick format
13324 Default value is @samp{yuv420}.
13327 See @ref{framesync}.
13330 Set format of alpha of the overlaid video, it can be @var{straight} or
13331 @var{premultiplied}. Default is @var{straight}.
13334 The @option{x}, and @option{y} expressions can contain the following
13340 The main input width and height.
13344 The overlay input width and height.
13348 The computed values for @var{x} and @var{y}. They are evaluated for
13353 horizontal and vertical chroma subsample values of the output
13354 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
13358 the number of input frame, starting from 0
13361 the position in the file of the input frame, NAN if unknown
13364 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
13368 This filter also supports the @ref{framesync} options.
13370 Note that the @var{n}, @var{pos}, @var{t} variables are available only
13371 when evaluation is done @emph{per frame}, and will evaluate to NAN
13372 when @option{eval} is set to @samp{init}.
13374 Be aware that frames are taken from each input video in timestamp
13375 order, hence, if their initial timestamps differ, it is a good idea
13376 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
13377 have them begin in the same zero timestamp, as the example for
13378 the @var{movie} filter does.
13380 You can chain together more overlays but you should test the
13381 efficiency of such approach.
13383 @subsection Commands
13385 This filter supports the following commands:
13389 Modify the x and y of the overlay input.
13390 The command accepts the same syntax of the corresponding option.
13392 If the specified expression is not valid, it is kept at its current
13396 @subsection Examples
13400 Draw the overlay at 10 pixels from the bottom right corner of the main
13403 overlay=main_w-overlay_w-10:main_h-overlay_h-10
13406 Using named options the example above becomes:
13408 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
13412 Insert a transparent PNG logo in the bottom left corner of the input,
13413 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
13415 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
13419 Insert 2 different transparent PNG logos (second logo on bottom
13420 right corner) using the @command{ffmpeg} tool:
13422 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
13426 Add a transparent color layer on top of the main video; @code{WxH}
13427 must specify the size of the main input to the overlay filter:
13429 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
13433 Play an original video and a filtered version (here with the deshake
13434 filter) side by side using the @command{ffplay} tool:
13436 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
13439 The above command is the same as:
13441 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
13445 Make a sliding overlay appearing from the left to the right top part of the
13446 screen starting since time 2:
13448 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
13452 Compose output by putting two input videos side to side:
13454 ffmpeg -i left.avi -i right.avi -filter_complex "
13455 nullsrc=size=200x100 [background];
13456 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
13457 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
13458 [background][left] overlay=shortest=1 [background+left];
13459 [background+left][right] overlay=shortest=1:x=100 [left+right]
13464 Mask 10-20 seconds of a video by applying the delogo filter to a section
13466 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
13467 -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]'
13472 Chain several overlays in cascade:
13474 nullsrc=s=200x200 [bg];
13475 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
13476 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
13477 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
13478 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
13479 [in3] null, [mid2] overlay=100:100 [out0]
13486 Apply Overcomplete Wavelet denoiser.
13488 The filter accepts the following options:
13494 Larger depth values will denoise lower frequency components more, but
13495 slow down filtering.
13497 Must be an int in the range 8-16, default is @code{8}.
13499 @item luma_strength, ls
13502 Must be a double value in the range 0-1000, default is @code{1.0}.
13504 @item chroma_strength, cs
13505 Set chroma strength.
13507 Must be a double value in the range 0-1000, default is @code{1.0}.
13513 Add paddings to the input image, and place the original input at the
13514 provided @var{x}, @var{y} coordinates.
13516 It accepts the following parameters:
13521 Specify an expression for the size of the output image with the
13522 paddings added. If the value for @var{width} or @var{height} is 0, the
13523 corresponding input size is used for the output.
13525 The @var{width} expression can reference the value set by the
13526 @var{height} expression, and vice versa.
13528 The default value of @var{width} and @var{height} is 0.
13532 Specify the offsets to place the input image at within the padded area,
13533 with respect to the top/left border of the output image.
13535 The @var{x} expression can reference the value set by the @var{y}
13536 expression, and vice versa.
13538 The default value of @var{x} and @var{y} is 0.
13540 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
13541 so the input image is centered on the padded area.
13544 Specify the color of the padded area. For the syntax of this option,
13545 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
13546 manual,ffmpeg-utils}.
13548 The default value of @var{color} is "black".
13551 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
13553 It accepts the following values:
13557 Only evaluate expressions once during the filter initialization or when
13558 a command is processed.
13561 Evaluate expressions for each incoming frame.
13565 Default value is @samp{init}.
13568 Pad to aspect instead to a resolution.
13572 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
13573 options are expressions containing the following constants:
13578 The input video width and height.
13582 These are the same as @var{in_w} and @var{in_h}.
13586 The output width and height (the size of the padded area), as
13587 specified by the @var{width} and @var{height} expressions.
13591 These are the same as @var{out_w} and @var{out_h}.
13595 The x and y offsets as specified by the @var{x} and @var{y}
13596 expressions, or NAN if not yet specified.
13599 same as @var{iw} / @var{ih}
13602 input sample aspect ratio
13605 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
13609 The horizontal and vertical chroma subsample values. For example for the
13610 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
13613 @subsection Examples
13617 Add paddings with the color "violet" to the input video. The output video
13618 size is 640x480, and the top-left corner of the input video is placed at
13621 pad=640:480:0:40:violet
13624 The example above is equivalent to the following command:
13626 pad=width=640:height=480:x=0:y=40:color=violet
13630 Pad the input to get an output with dimensions increased by 3/2,
13631 and put the input video at the center of the padded area:
13633 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
13637 Pad the input to get a squared output with size equal to the maximum
13638 value between the input width and height, and put the input video at
13639 the center of the padded area:
13641 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
13645 Pad the input to get a final w/h ratio of 16:9:
13647 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
13651 In case of anamorphic video, in order to set the output display aspect
13652 correctly, it is necessary to use @var{sar} in the expression,
13653 according to the relation:
13655 (ih * X / ih) * sar = output_dar
13656 X = output_dar / sar
13659 Thus the previous example needs to be modified to:
13661 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
13665 Double the output size and put the input video in the bottom-right
13666 corner of the output padded area:
13668 pad="2*iw:2*ih:ow-iw:oh-ih"
13672 @anchor{palettegen}
13673 @section palettegen
13675 Generate one palette for a whole video stream.
13677 It accepts the following options:
13681 Set the maximum number of colors to quantize in the palette.
13682 Note: the palette will still contain 256 colors; the unused palette entries
13685 @item reserve_transparent
13686 Create a palette of 255 colors maximum and reserve the last one for
13687 transparency. Reserving the transparency color is useful for GIF optimization.
13688 If not set, the maximum of colors in the palette will be 256. You probably want
13689 to disable this option for a standalone image.
13692 @item transparency_color
13693 Set the color that will be used as background for transparency.
13696 Set statistics mode.
13698 It accepts the following values:
13701 Compute full frame histograms.
13703 Compute histograms only for the part that differs from previous frame. This
13704 might be relevant to give more importance to the moving part of your input if
13705 the background is static.
13707 Compute new histogram for each frame.
13710 Default value is @var{full}.
13713 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
13714 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
13715 color quantization of the palette. This information is also visible at
13716 @var{info} logging level.
13718 @subsection Examples
13722 Generate a representative palette of a given video using @command{ffmpeg}:
13724 ffmpeg -i input.mkv -vf palettegen palette.png
13728 @section paletteuse
13730 Use a palette to downsample an input video stream.
13732 The filter takes two inputs: one video stream and a palette. The palette must
13733 be a 256 pixels image.
13735 It accepts the following options:
13739 Select dithering mode. Available algorithms are:
13742 Ordered 8x8 bayer dithering (deterministic)
13744 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
13745 Note: this dithering is sometimes considered "wrong" and is included as a
13747 @item floyd_steinberg
13748 Floyd and Steingberg dithering (error diffusion)
13750 Frankie Sierra dithering v2 (error diffusion)
13752 Frankie Sierra dithering v2 "Lite" (error diffusion)
13755 Default is @var{sierra2_4a}.
13758 When @var{bayer} dithering is selected, this option defines the scale of the
13759 pattern (how much the crosshatch pattern is visible). A low value means more
13760 visible pattern for less banding, and higher value means less visible pattern
13761 at the cost of more banding.
13763 The option must be an integer value in the range [0,5]. Default is @var{2}.
13766 If set, define the zone to process
13770 Only the changing rectangle will be reprocessed. This is similar to GIF
13771 cropping/offsetting compression mechanism. This option can be useful for speed
13772 if only a part of the image is changing, and has use cases such as limiting the
13773 scope of the error diffusal @option{dither} to the rectangle that bounds the
13774 moving scene (it leads to more deterministic output if the scene doesn't change
13775 much, and as a result less moving noise and better GIF compression).
13778 Default is @var{none}.
13781 Take new palette for each output frame.
13783 @item alpha_threshold
13784 Sets the alpha threshold for transparency. Alpha values above this threshold
13785 will be treated as completely opaque, and values below this threshold will be
13786 treated as completely transparent.
13788 The option must be an integer value in the range [0,255]. Default is @var{128}.
13791 @subsection Examples
13795 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
13796 using @command{ffmpeg}:
13798 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
13802 @section perspective
13804 Correct perspective of video not recorded perpendicular to the screen.
13806 A description of the accepted parameters follows.
13817 Set coordinates expression for top left, top right, bottom left and bottom right corners.
13818 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
13819 If the @code{sense} option is set to @code{source}, then the specified points will be sent
13820 to the corners of the destination. If the @code{sense} option is set to @code{destination},
13821 then the corners of the source will be sent to the specified coordinates.
13823 The expressions can use the following variables:
13828 the width and height of video frame.
13832 Output frame count.
13835 @item interpolation
13836 Set interpolation for perspective correction.
13838 It accepts the following values:
13844 Default value is @samp{linear}.
13847 Set interpretation of coordinate options.
13849 It accepts the following values:
13853 Send point in the source specified by the given coordinates to
13854 the corners of the destination.
13856 @item 1, destination
13858 Send the corners of the source to the point in the destination specified
13859 by the given coordinates.
13861 Default value is @samp{source}.
13865 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
13867 It accepts the following values:
13870 only evaluate expressions once during the filter initialization or
13871 when a command is processed
13874 evaluate expressions for each incoming frame
13877 Default value is @samp{init}.
13882 Delay interlaced video by one field time so that the field order changes.
13884 The intended use is to fix PAL movies that have been captured with the
13885 opposite field order to the film-to-video transfer.
13887 A description of the accepted parameters follows.
13893 It accepts the following values:
13896 Capture field order top-first, transfer bottom-first.
13897 Filter will delay the bottom field.
13900 Capture field order bottom-first, transfer top-first.
13901 Filter will delay the top field.
13904 Capture and transfer with the same field order. This mode only exists
13905 for the documentation of the other options to refer to, but if you
13906 actually select it, the filter will faithfully do nothing.
13909 Capture field order determined automatically by field flags, transfer
13911 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
13912 basis using field flags. If no field information is available,
13913 then this works just like @samp{u}.
13916 Capture unknown or varying, transfer opposite.
13917 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
13918 analyzing the images and selecting the alternative that produces best
13919 match between the fields.
13922 Capture top-first, transfer unknown or varying.
13923 Filter selects among @samp{t} and @samp{p} using image analysis.
13926 Capture bottom-first, transfer unknown or varying.
13927 Filter selects among @samp{b} and @samp{p} using image analysis.
13930 Capture determined by field flags, transfer unknown or varying.
13931 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
13932 image analysis. If no field information is available, then this works just
13933 like @samp{U}. This is the default mode.
13936 Both capture and transfer unknown or varying.
13937 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
13941 @section pixdesctest
13943 Pixel format descriptor test filter, mainly useful for internal
13944 testing. The output video should be equal to the input video.
13948 format=monow, pixdesctest
13951 can be used to test the monowhite pixel format descriptor definition.
13955 Display sample values of color channels. Mainly useful for checking color
13956 and levels. Minimum supported resolution is 640x480.
13958 The filters accept the following options:
13962 Set scope X position, relative offset on X axis.
13965 Set scope Y position, relative offset on Y axis.
13974 Set window opacity. This window also holds statistics about pixel area.
13977 Set window X position, relative offset on X axis.
13980 Set window Y position, relative offset on Y axis.
13985 Enable the specified chain of postprocessing subfilters using libpostproc. This
13986 library should be automatically selected with a GPL build (@code{--enable-gpl}).
13987 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
13988 Each subfilter and some options have a short and a long name that can be used
13989 interchangeably, i.e. dr/dering are the same.
13991 The filters accept the following options:
13995 Set postprocessing subfilters string.
13998 All subfilters share common options to determine their scope:
14002 Honor the quality commands for this subfilter.
14005 Do chrominance filtering, too (default).
14008 Do luminance filtering only (no chrominance).
14011 Do chrominance filtering only (no luminance).
14014 These options can be appended after the subfilter name, separated by a '|'.
14016 Available subfilters are:
14019 @item hb/hdeblock[|difference[|flatness]]
14020 Horizontal deblocking filter
14023 Difference factor where higher values mean more deblocking (default: @code{32}).
14025 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14028 @item vb/vdeblock[|difference[|flatness]]
14029 Vertical deblocking filter
14032 Difference factor where higher values mean more deblocking (default: @code{32}).
14034 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14037 @item ha/hadeblock[|difference[|flatness]]
14038 Accurate horizontal deblocking filter
14041 Difference factor where higher values mean more deblocking (default: @code{32}).
14043 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14046 @item va/vadeblock[|difference[|flatness]]
14047 Accurate vertical deblocking filter
14050 Difference factor where higher values mean more deblocking (default: @code{32}).
14052 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14056 The horizontal and vertical deblocking filters share the difference and
14057 flatness values so you cannot set different horizontal and vertical
14061 @item h1/x1hdeblock
14062 Experimental horizontal deblocking filter
14064 @item v1/x1vdeblock
14065 Experimental vertical deblocking filter
14070 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
14073 larger -> stronger filtering
14075 larger -> stronger filtering
14077 larger -> stronger filtering
14080 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
14083 Stretch luminance to @code{0-255}.
14086 @item lb/linblenddeint
14087 Linear blend deinterlacing filter that deinterlaces the given block by
14088 filtering all lines with a @code{(1 2 1)} filter.
14090 @item li/linipoldeint
14091 Linear interpolating deinterlacing filter that deinterlaces the given block by
14092 linearly interpolating every second line.
14094 @item ci/cubicipoldeint
14095 Cubic interpolating deinterlacing filter deinterlaces the given block by
14096 cubically interpolating every second line.
14098 @item md/mediandeint
14099 Median deinterlacing filter that deinterlaces the given block by applying a
14100 median filter to every second line.
14102 @item fd/ffmpegdeint
14103 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
14104 second line with a @code{(-1 4 2 4 -1)} filter.
14107 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
14108 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
14110 @item fq/forceQuant[|quantizer]
14111 Overrides the quantizer table from the input with the constant quantizer you
14119 Default pp filter combination (@code{hb|a,vb|a,dr|a})
14122 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
14125 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
14128 @subsection Examples
14132 Apply horizontal and vertical deblocking, deringing and automatic
14133 brightness/contrast:
14139 Apply default filters without brightness/contrast correction:
14145 Apply default filters and temporal denoiser:
14147 pp=default/tmpnoise|1|2|3
14151 Apply deblocking on luminance only, and switch vertical deblocking on or off
14152 automatically depending on available CPU time:
14159 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
14160 similar to spp = 6 with 7 point DCT, where only the center sample is
14163 The filter accepts the following options:
14167 Force a constant quantization parameter. It accepts an integer in range
14168 0 to 63. If not set, the filter will use the QP from the video stream
14172 Set thresholding mode. Available modes are:
14176 Set hard thresholding.
14178 Set soft thresholding (better de-ringing effect, but likely blurrier).
14180 Set medium thresholding (good results, default).
14184 @section premultiply
14185 Apply alpha premultiply effect to input video stream using first plane
14186 of second stream as alpha.
14188 Both streams must have same dimensions and same pixel format.
14190 The filter accepts the following option:
14194 Set which planes will be processed, unprocessed planes will be copied.
14195 By default value 0xf, all planes will be processed.
14198 Do not require 2nd input for processing, instead use alpha plane from input stream.
14202 Apply prewitt operator to input video stream.
14204 The filter accepts the following option:
14208 Set which planes will be processed, unprocessed planes will be copied.
14209 By default value 0xf, all planes will be processed.
14212 Set value which will be multiplied with filtered result.
14215 Set value which will be added to filtered result.
14218 @anchor{program_opencl}
14219 @section program_opencl
14221 Filter video using an OpenCL program.
14226 OpenCL program source file.
14229 Kernel name in program.
14232 Number of inputs to the filter. Defaults to 1.
14235 Size of output frames. Defaults to the same as the first input.
14239 The program source file must contain a kernel function with the given name,
14240 which will be run once for each plane of the output. Each run on a plane
14241 gets enqueued as a separate 2D global NDRange with one work-item for each
14242 pixel to be generated. The global ID offset for each work-item is therefore
14243 the coordinates of a pixel in the destination image.
14245 The kernel function needs to take the following arguments:
14248 Destination image, @var{__write_only image2d_t}.
14250 This image will become the output; the kernel should write all of it.
14252 Frame index, @var{unsigned int}.
14254 This is a counter starting from zero and increasing by one for each frame.
14256 Source images, @var{__read_only image2d_t}.
14258 These are the most recent images on each input. The kernel may read from
14259 them to generate the output, but they can't be written to.
14266 Copy the input to the output (output must be the same size as the input).
14268 __kernel void copy(__write_only image2d_t destination,
14269 unsigned int index,
14270 __read_only image2d_t source)
14272 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
14274 int2 location = (int2)(get_global_id(0), get_global_id(1));
14276 float4 value = read_imagef(source, sampler, location);
14278 write_imagef(destination, location, value);
14283 Apply a simple transformation, rotating the input by an amount increasing
14284 with the index counter. Pixel values are linearly interpolated by the
14285 sampler, and the output need not have the same dimensions as the input.
14287 __kernel void rotate_image(__write_only image2d_t dst,
14288 unsigned int index,
14289 __read_only image2d_t src)
14291 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
14292 CLK_FILTER_LINEAR);
14294 float angle = (float)index / 100.0f;
14296 float2 dst_dim = convert_float2(get_image_dim(dst));
14297 float2 src_dim = convert_float2(get_image_dim(src));
14299 float2 dst_cen = dst_dim / 2.0f;
14300 float2 src_cen = src_dim / 2.0f;
14302 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
14304 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
14306 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
14307 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
14309 src_pos = src_pos * src_dim / dst_dim;
14311 float2 src_loc = src_pos + src_cen;
14313 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
14314 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
14315 write_imagef(dst, dst_loc, 0.5f);
14317 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
14322 Blend two inputs together, with the amount of each input used varying
14323 with the index counter.
14325 __kernel void blend_images(__write_only image2d_t dst,
14326 unsigned int index,
14327 __read_only image2d_t src1,
14328 __read_only image2d_t src2)
14330 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
14331 CLK_FILTER_LINEAR);
14333 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
14335 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
14336 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
14337 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
14339 float4 val1 = read_imagef(src1, sampler, src1_loc);
14340 float4 val2 = read_imagef(src2, sampler, src2_loc);
14342 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
14348 @section pseudocolor
14350 Alter frame colors in video with pseudocolors.
14352 This filter accept the following options:
14356 set pixel first component expression
14359 set pixel second component expression
14362 set pixel third component expression
14365 set pixel fourth component expression, corresponds to the alpha component
14368 set component to use as base for altering colors
14371 Each of them specifies the expression to use for computing the lookup table for
14372 the corresponding pixel component values.
14374 The expressions can contain the following constants and functions:
14379 The input width and height.
14382 The input value for the pixel component.
14384 @item ymin, umin, vmin, amin
14385 The minimum allowed component value.
14387 @item ymax, umax, vmax, amax
14388 The maximum allowed component value.
14391 All expressions default to "val".
14393 @subsection Examples
14397 Change too high luma values to gradient:
14399 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'"
14405 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
14406 Ratio) between two input videos.
14408 This filter takes in input two input videos, the first input is
14409 considered the "main" source and is passed unchanged to the
14410 output. The second input is used as a "reference" video for computing
14413 Both video inputs must have the same resolution and pixel format for
14414 this filter to work correctly. Also it assumes that both inputs
14415 have the same number of frames, which are compared one by one.
14417 The obtained average PSNR is printed through the logging system.
14419 The filter stores the accumulated MSE (mean squared error) of each
14420 frame, and at the end of the processing it is averaged across all frames
14421 equally, and the following formula is applied to obtain the PSNR:
14424 PSNR = 10*log10(MAX^2/MSE)
14427 Where MAX is the average of the maximum values of each component of the
14430 The description of the accepted parameters follows.
14433 @item stats_file, f
14434 If specified the filter will use the named file to save the PSNR of
14435 each individual frame. When filename equals "-" the data is sent to
14438 @item stats_version
14439 Specifies which version of the stats file format to use. Details of
14440 each format are written below.
14441 Default value is 1.
14443 @item stats_add_max
14444 Determines whether the max value is output to the stats log.
14445 Default value is 0.
14446 Requires stats_version >= 2. If this is set and stats_version < 2,
14447 the filter will return an error.
14450 This filter also supports the @ref{framesync} options.
14452 The file printed if @var{stats_file} is selected, contains a sequence of
14453 key/value pairs of the form @var{key}:@var{value} for each compared
14456 If a @var{stats_version} greater than 1 is specified, a header line precedes
14457 the list of per-frame-pair stats, with key value pairs following the frame
14458 format with the following parameters:
14461 @item psnr_log_version
14462 The version of the log file format. Will match @var{stats_version}.
14465 A comma separated list of the per-frame-pair parameters included in
14469 A description of each shown per-frame-pair parameter follows:
14473 sequential number of the input frame, starting from 1
14476 Mean Square Error pixel-by-pixel average difference of the compared
14477 frames, averaged over all the image components.
14479 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
14480 Mean Square Error pixel-by-pixel average difference of the compared
14481 frames for the component specified by the suffix.
14483 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
14484 Peak Signal to Noise ratio of the compared frames for the component
14485 specified by the suffix.
14487 @item max_avg, max_y, max_u, max_v
14488 Maximum allowed value for each channel, and average over all
14494 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
14495 [main][ref] psnr="stats_file=stats.log" [out]
14498 On this example the input file being processed is compared with the
14499 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
14500 is stored in @file{stats.log}.
14505 Pulldown reversal (inverse telecine) filter, capable of handling mixed
14506 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
14509 The pullup filter is designed to take advantage of future context in making
14510 its decisions. This filter is stateless in the sense that it does not lock
14511 onto a pattern to follow, but it instead looks forward to the following
14512 fields in order to identify matches and rebuild progressive frames.
14514 To produce content with an even framerate, insert the fps filter after
14515 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
14516 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
14518 The filter accepts the following options:
14525 These options set the amount of "junk" to ignore at the left, right, top, and
14526 bottom of the image, respectively. Left and right are in units of 8 pixels,
14527 while top and bottom are in units of 2 lines.
14528 The default is 8 pixels on each side.
14531 Set the strict breaks. Setting this option to 1 will reduce the chances of
14532 filter generating an occasional mismatched frame, but it may also cause an
14533 excessive number of frames to be dropped during high motion sequences.
14534 Conversely, setting it to -1 will make filter match fields more easily.
14535 This may help processing of video where there is slight blurring between
14536 the fields, but may also cause there to be interlaced frames in the output.
14537 Default value is @code{0}.
14540 Set the metric plane to use. It accepts the following values:
14546 Use chroma blue plane.
14549 Use chroma red plane.
14552 This option may be set to use chroma plane instead of the default luma plane
14553 for doing filter's computations. This may improve accuracy on very clean
14554 source material, but more likely will decrease accuracy, especially if there
14555 is chroma noise (rainbow effect) or any grayscale video.
14556 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
14557 load and make pullup usable in realtime on slow machines.
14560 For best results (without duplicated frames in the output file) it is
14561 necessary to change the output frame rate. For example, to inverse
14562 telecine NTSC input:
14564 ffmpeg -i input -vf pullup -r 24000/1001 ...
14569 Change video quantization parameters (QP).
14571 The filter accepts the following option:
14575 Set expression for quantization parameter.
14578 The expression is evaluated through the eval API and can contain, among others,
14579 the following constants:
14583 1 if index is not 129, 0 otherwise.
14586 Sequential index starting from -129 to 128.
14589 @subsection Examples
14593 Some equation like:
14601 Flush video frames from internal cache of frames into a random order.
14602 No frame is discarded.
14603 Inspired by @ref{frei0r} nervous filter.
14607 Set size in number of frames of internal cache, in range from @code{2} to
14608 @code{512}. Default is @code{30}.
14611 Set seed for random number generator, must be an integer included between
14612 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
14613 less than @code{0}, the filter will try to use a good random seed on a
14617 @section readeia608
14619 Read closed captioning (EIA-608) information from the top lines of a video frame.
14621 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
14622 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
14623 with EIA-608 data (starting from 0). A description of each metadata value follows:
14626 @item lavfi.readeia608.X.cc
14627 The two bytes stored as EIA-608 data (printed in hexadecimal).
14629 @item lavfi.readeia608.X.line
14630 The number of the line on which the EIA-608 data was identified and read.
14633 This filter accepts the following options:
14637 Set the line to start scanning for EIA-608 data. Default is @code{0}.
14640 Set the line to end scanning for EIA-608 data. Default is @code{29}.
14643 Set minimal acceptable amplitude change for sync codes detection.
14644 Default is @code{0.2}. Allowed range is @code{[0.001 - 1]}.
14647 Set the ratio of width reserved for sync code detection.
14648 Default is @code{0.27}. Allowed range is @code{[0.01 - 0.7]}.
14651 Set the max peaks height difference for sync code detection.
14652 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
14655 Set max peaks period difference for sync code detection.
14656 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
14659 Set the first two max start code bits differences.
14660 Default is @code{0.02}. Allowed range is @code{[0.0 - 0.5]}.
14663 Set the minimum ratio of bits height compared to 3rd start code bit.
14664 Default is @code{0.75}. Allowed range is @code{[0.01 - 1]}.
14667 Set the white color threshold. Default is @code{0.35}. Allowed range is @code{[0.1 - 1]}.
14670 Set the black color threshold. Default is @code{0.15}. Allowed range is @code{[0.0 - 0.5]}.
14673 Enable checking the parity bit. In the event of a parity error, the filter will output
14674 @code{0x00} for that character. Default is false.
14677 Lowpass lines prior further proccessing. Default is disabled.
14680 @subsection Examples
14684 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
14686 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
14692 Read vertical interval timecode (VITC) information from the top lines of a
14695 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
14696 timecode value, if a valid timecode has been detected. Further metadata key
14697 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
14698 timecode data has been found or not.
14700 This filter accepts the following options:
14704 Set the maximum number of lines to scan for VITC data. If the value is set to
14705 @code{-1} the full video frame is scanned. Default is @code{45}.
14708 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
14709 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
14712 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
14713 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
14716 @subsection Examples
14720 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
14721 draw @code{--:--:--:--} as a placeholder:
14723 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
14729 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
14731 Destination pixel at position (X, Y) will be picked from source (x, y) position
14732 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
14733 value for pixel will be used for destination pixel.
14735 Xmap and Ymap input video streams must be of same dimensions. Output video stream
14736 will have Xmap/Ymap video stream dimensions.
14737 Xmap and Ymap input video streams are 16bit depth, single channel.
14741 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
14742 Default is @code{color}.
14745 @section removegrain
14747 The removegrain filter is a spatial denoiser for progressive video.
14751 Set mode for the first plane.
14754 Set mode for the second plane.
14757 Set mode for the third plane.
14760 Set mode for the fourth plane.
14763 Range of mode is from 0 to 24. Description of each mode follows:
14767 Leave input plane unchanged. Default.
14770 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
14773 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
14776 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
14779 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
14780 This is equivalent to a median filter.
14783 Line-sensitive clipping giving the minimal change.
14786 Line-sensitive clipping, intermediate.
14789 Line-sensitive clipping, intermediate.
14792 Line-sensitive clipping, intermediate.
14795 Line-sensitive clipping on a line where the neighbours pixels are the closest.
14798 Replaces the target pixel with the closest neighbour.
14801 [1 2 1] horizontal and vertical kernel blur.
14807 Bob mode, interpolates top field from the line where the neighbours
14808 pixels are the closest.
14811 Bob mode, interpolates bottom field from the line where the neighbours
14812 pixels are the closest.
14815 Bob mode, interpolates top field. Same as 13 but with a more complicated
14816 interpolation formula.
14819 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
14820 interpolation formula.
14823 Clips the pixel with the minimum and maximum of respectively the maximum and
14824 minimum of each pair of opposite neighbour pixels.
14827 Line-sensitive clipping using opposite neighbours whose greatest distance from
14828 the current pixel is minimal.
14831 Replaces the pixel with the average of its 8 neighbours.
14834 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
14837 Clips pixels using the averages of opposite neighbour.
14840 Same as mode 21 but simpler and faster.
14843 Small edge and halo removal, but reputed useless.
14849 @section removelogo
14851 Suppress a TV station logo, using an image file to determine which
14852 pixels comprise the logo. It works by filling in the pixels that
14853 comprise the logo with neighboring pixels.
14855 The filter accepts the following options:
14859 Set the filter bitmap file, which can be any image format supported by
14860 libavformat. The width and height of the image file must match those of the
14861 video stream being processed.
14864 Pixels in the provided bitmap image with a value of zero are not
14865 considered part of the logo, non-zero pixels are considered part of
14866 the logo. If you use white (255) for the logo and black (0) for the
14867 rest, you will be safe. For making the filter bitmap, it is
14868 recommended to take a screen capture of a black frame with the logo
14869 visible, and then using a threshold filter followed by the erode
14870 filter once or twice.
14872 If needed, little splotches can be fixed manually. Remember that if
14873 logo pixels are not covered, the filter quality will be much
14874 reduced. Marking too many pixels as part of the logo does not hurt as
14875 much, but it will increase the amount of blurring needed to cover over
14876 the image and will destroy more information than necessary, and extra
14877 pixels will slow things down on a large logo.
14879 @section repeatfields
14881 This filter uses the repeat_field flag from the Video ES headers and hard repeats
14882 fields based on its value.
14886 Reverse a video clip.
14888 Warning: This filter requires memory to buffer the entire clip, so trimming
14891 @subsection Examples
14895 Take the first 5 seconds of a clip, and reverse it.
14902 Shift R/G/B/A pixels horizontally and/or vertically.
14904 The filter accepts the following options:
14907 Set amount to shift red horizontally.
14909 Set amount to shift red vertically.
14911 Set amount to shift green horizontally.
14913 Set amount to shift green vertically.
14915 Set amount to shift blue horizontally.
14917 Set amount to shift blue vertically.
14919 Set amount to shift alpha horizontally.
14921 Set amount to shift alpha vertically.
14923 Set edge mode, can be @var{smear}, default, or @var{warp}.
14927 Apply roberts cross operator to input video stream.
14929 The filter accepts the following option:
14933 Set which planes will be processed, unprocessed planes will be copied.
14934 By default value 0xf, all planes will be processed.
14937 Set value which will be multiplied with filtered result.
14940 Set value which will be added to filtered result.
14945 Rotate video by an arbitrary angle expressed in radians.
14947 The filter accepts the following options:
14949 A description of the optional parameters follows.
14952 Set an expression for the angle by which to rotate the input video
14953 clockwise, expressed as a number of radians. A negative value will
14954 result in a counter-clockwise rotation. By default it is set to "0".
14956 This expression is evaluated for each frame.
14959 Set the output width expression, default value is "iw".
14960 This expression is evaluated just once during configuration.
14963 Set the output height expression, default value is "ih".
14964 This expression is evaluated just once during configuration.
14967 Enable bilinear interpolation if set to 1, a value of 0 disables
14968 it. Default value is 1.
14971 Set the color used to fill the output area not covered by the rotated
14972 image. For the general syntax of this option, check the
14973 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
14974 If the special value "none" is selected then no
14975 background is printed (useful for example if the background is never shown).
14977 Default value is "black".
14980 The expressions for the angle and the output size can contain the
14981 following constants and functions:
14985 sequential number of the input frame, starting from 0. It is always NAN
14986 before the first frame is filtered.
14989 time in seconds of the input frame, it is set to 0 when the filter is
14990 configured. It is always NAN before the first frame is filtered.
14994 horizontal and vertical chroma subsample values. For example for the
14995 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14999 the input video width and height
15003 the output width and height, that is the size of the padded area as
15004 specified by the @var{width} and @var{height} expressions
15008 the minimal width/height required for completely containing the input
15009 video rotated by @var{a} radians.
15011 These are only available when computing the @option{out_w} and
15012 @option{out_h} expressions.
15015 @subsection Examples
15019 Rotate the input by PI/6 radians clockwise:
15025 Rotate the input by PI/6 radians counter-clockwise:
15031 Rotate the input by 45 degrees clockwise:
15037 Apply a constant rotation with period T, starting from an angle of PI/3:
15039 rotate=PI/3+2*PI*t/T
15043 Make the input video rotation oscillating with a period of T
15044 seconds and an amplitude of A radians:
15046 rotate=A*sin(2*PI/T*t)
15050 Rotate the video, output size is chosen so that the whole rotating
15051 input video is always completely contained in the output:
15053 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
15057 Rotate the video, reduce the output size so that no background is ever
15060 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
15064 @subsection Commands
15066 The filter supports the following commands:
15070 Set the angle expression.
15071 The command accepts the same syntax of the corresponding option.
15073 If the specified expression is not valid, it is kept at its current
15079 Apply Shape Adaptive Blur.
15081 The filter accepts the following options:
15084 @item luma_radius, lr
15085 Set luma blur filter strength, must be a value in range 0.1-4.0, default
15086 value is 1.0. A greater value will result in a more blurred image, and
15087 in slower processing.
15089 @item luma_pre_filter_radius, lpfr
15090 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
15093 @item luma_strength, ls
15094 Set luma maximum difference between pixels to still be considered, must
15095 be a value in the 0.1-100.0 range, default value is 1.0.
15097 @item chroma_radius, cr
15098 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
15099 greater value will result in a more blurred image, and in slower
15102 @item chroma_pre_filter_radius, cpfr
15103 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
15105 @item chroma_strength, cs
15106 Set chroma maximum difference between pixels to still be considered,
15107 must be a value in the -0.9-100.0 range.
15110 Each chroma option value, if not explicitly specified, is set to the
15111 corresponding luma option value.
15116 Scale (resize) the input video, using the libswscale library.
15118 The scale filter forces the output display aspect ratio to be the same
15119 of the input, by changing the output sample aspect ratio.
15121 If the input image format is different from the format requested by
15122 the next filter, the scale filter will convert the input to the
15125 @subsection Options
15126 The filter accepts the following options, or any of the options
15127 supported by the libswscale scaler.
15129 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
15130 the complete list of scaler options.
15135 Set the output video dimension expression. Default value is the input
15138 If the @var{width} or @var{w} value is 0, the input width is used for
15139 the output. If the @var{height} or @var{h} value is 0, the input height
15140 is used for the output.
15142 If one and only one of the values is -n with n >= 1, the scale filter
15143 will use a value that maintains the aspect ratio of the input image,
15144 calculated from the other specified dimension. After that it will,
15145 however, make sure that the calculated dimension is divisible by n and
15146 adjust the value if necessary.
15148 If both values are -n with n >= 1, the behavior will be identical to
15149 both values being set to 0 as previously detailed.
15151 See below for the list of accepted constants for use in the dimension
15155 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
15159 Only evaluate expressions once during the filter initialization or when a command is processed.
15162 Evaluate expressions for each incoming frame.
15166 Default value is @samp{init}.
15170 Set the interlacing mode. It accepts the following values:
15174 Force interlaced aware scaling.
15177 Do not apply interlaced scaling.
15180 Select interlaced aware scaling depending on whether the source frames
15181 are flagged as interlaced or not.
15184 Default value is @samp{0}.
15187 Set libswscale scaling flags. See
15188 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15189 complete list of values. If not explicitly specified the filter applies
15193 @item param0, param1
15194 Set libswscale input parameters for scaling algorithms that need them. See
15195 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15196 complete documentation. If not explicitly specified the filter applies
15202 Set the video size. For the syntax of this option, check the
15203 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
15205 @item in_color_matrix
15206 @item out_color_matrix
15207 Set in/output YCbCr color space type.
15209 This allows the autodetected value to be overridden as well as allows forcing
15210 a specific value used for the output and encoder.
15212 If not specified, the color space type depends on the pixel format.
15218 Choose automatically.
15221 Format conforming to International Telecommunication Union (ITU)
15222 Recommendation BT.709.
15225 Set color space conforming to the United States Federal Communications
15226 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
15231 Set color space conforming to:
15235 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
15238 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
15241 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
15246 Set color space conforming to SMPTE ST 240:1999.
15249 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
15254 Set in/output YCbCr sample range.
15256 This allows the autodetected value to be overridden as well as allows forcing
15257 a specific value used for the output and encoder. If not specified, the
15258 range depends on the pixel format. Possible values:
15262 Choose automatically.
15265 Set full range (0-255 in case of 8-bit luma).
15267 @item mpeg/limited/tv
15268 Set "MPEG" range (16-235 in case of 8-bit luma).
15271 @item force_original_aspect_ratio
15272 Enable decreasing or increasing output video width or height if necessary to
15273 keep the original aspect ratio. Possible values:
15277 Scale the video as specified and disable this feature.
15280 The output video dimensions will automatically be decreased if needed.
15283 The output video dimensions will automatically be increased if needed.
15287 One useful instance of this option is that when you know a specific device's
15288 maximum allowed resolution, you can use this to limit the output video to
15289 that, while retaining the aspect ratio. For example, device A allows
15290 1280x720 playback, and your video is 1920x800. Using this option (set it to
15291 decrease) and specifying 1280x720 to the command line makes the output
15294 Please note that this is a different thing than specifying -1 for @option{w}
15295 or @option{h}, you still need to specify the output resolution for this option
15300 The values of the @option{w} and @option{h} options are expressions
15301 containing the following constants:
15306 The input width and height
15310 These are the same as @var{in_w} and @var{in_h}.
15314 The output (scaled) width and height
15318 These are the same as @var{out_w} and @var{out_h}
15321 The same as @var{iw} / @var{ih}
15324 input sample aspect ratio
15327 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
15331 horizontal and vertical input chroma subsample values. For example for the
15332 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15336 horizontal and vertical output chroma subsample values. For example for the
15337 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15340 @subsection Examples
15344 Scale the input video to a size of 200x100
15349 This is equivalent to:
15360 Specify a size abbreviation for the output size:
15365 which can also be written as:
15371 Scale the input to 2x:
15373 scale=w=2*iw:h=2*ih
15377 The above is the same as:
15379 scale=2*in_w:2*in_h
15383 Scale the input to 2x with forced interlaced scaling:
15385 scale=2*iw:2*ih:interl=1
15389 Scale the input to half size:
15391 scale=w=iw/2:h=ih/2
15395 Increase the width, and set the height to the same size:
15401 Seek Greek harmony:
15408 Increase the height, and set the width to 3/2 of the height:
15410 scale=w=3/2*oh:h=3/5*ih
15414 Increase the size, making the size a multiple of the chroma
15417 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
15421 Increase the width to a maximum of 500 pixels,
15422 keeping the same aspect ratio as the input:
15424 scale=w='min(500\, iw*3/2):h=-1'
15428 Make pixels square by combining scale and setsar:
15430 scale='trunc(ih*dar):ih',setsar=1/1
15434 Make pixels square by combining scale and setsar,
15435 making sure the resulting resolution is even (required by some codecs):
15437 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
15441 @subsection Commands
15443 This filter supports the following commands:
15447 Set the output video dimension expression.
15448 The command accepts the same syntax of the corresponding option.
15450 If the specified expression is not valid, it is kept at its current
15456 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
15457 format conversion on CUDA video frames. Setting the output width and height
15458 works in the same way as for the @var{scale} filter.
15460 The following additional options are accepted:
15463 The pixel format of the output CUDA frames. If set to the string "same" (the
15464 default), the input format will be kept. Note that automatic format negotiation
15465 and conversion is not yet supported for hardware frames
15468 The interpolation algorithm used for resizing. One of the following:
15475 @item cubic2p_bspline
15476 2-parameter cubic (B=1, C=0)
15478 @item cubic2p_catmullrom
15479 2-parameter cubic (B=0, C=1/2)
15481 @item cubic2p_b05c03
15482 2-parameter cubic (B=1/2, C=3/10)
15494 Scale (resize) the input video, based on a reference video.
15496 See the scale filter for available options, scale2ref supports the same but
15497 uses the reference video instead of the main input as basis. scale2ref also
15498 supports the following additional constants for the @option{w} and
15499 @option{h} options:
15504 The main input video's width and height
15507 The same as @var{main_w} / @var{main_h}
15510 The main input video's sample aspect ratio
15512 @item main_dar, mdar
15513 The main input video's display aspect ratio. Calculated from
15514 @code{(main_w / main_h) * main_sar}.
15518 The main input video's horizontal and vertical chroma subsample values.
15519 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
15523 @subsection Examples
15527 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
15529 'scale2ref[b][a];[a][b]overlay'
15533 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
15535 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
15539 @anchor{selectivecolor}
15540 @section selectivecolor
15542 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
15543 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
15544 by the "purity" of the color (that is, how saturated it already is).
15546 This filter is similar to the Adobe Photoshop Selective Color tool.
15548 The filter accepts the following options:
15551 @item correction_method
15552 Select color correction method.
15554 Available values are:
15557 Specified adjustments are applied "as-is" (added/subtracted to original pixel
15560 Specified adjustments are relative to the original component value.
15562 Default is @code{absolute}.
15564 Adjustments for red pixels (pixels where the red component is the maximum)
15566 Adjustments for yellow pixels (pixels where the blue component is the minimum)
15568 Adjustments for green pixels (pixels where the green component is the maximum)
15570 Adjustments for cyan pixels (pixels where the red component is the minimum)
15572 Adjustments for blue pixels (pixels where the blue component is the maximum)
15574 Adjustments for magenta pixels (pixels where the green component is the minimum)
15576 Adjustments for white pixels (pixels where all components are greater than 128)
15578 Adjustments for all pixels except pure black and pure white
15580 Adjustments for black pixels (pixels where all components are lesser than 128)
15582 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
15585 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
15586 4 space separated floating point adjustment values in the [-1,1] range,
15587 respectively to adjust the amount of cyan, magenta, yellow and black for the
15588 pixels of its range.
15590 @subsection Examples
15594 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
15595 increase magenta by 27% in blue areas:
15597 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
15601 Use a Photoshop selective color preset:
15603 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
15607 @anchor{separatefields}
15608 @section separatefields
15610 The @code{separatefields} takes a frame-based video input and splits
15611 each frame into its components fields, producing a new half height clip
15612 with twice the frame rate and twice the frame count.
15614 This filter use field-dominance information in frame to decide which
15615 of each pair of fields to place first in the output.
15616 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
15618 @section setdar, setsar
15620 The @code{setdar} filter sets the Display Aspect Ratio for the filter
15623 This is done by changing the specified Sample (aka Pixel) Aspect
15624 Ratio, according to the following equation:
15626 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
15629 Keep in mind that the @code{setdar} filter does not modify the pixel
15630 dimensions of the video frame. Also, the display aspect ratio set by
15631 this filter may be changed by later filters in the filterchain,
15632 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
15635 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
15636 the filter output video.
15638 Note that as a consequence of the application of this filter, the
15639 output display aspect ratio will change according to the equation
15642 Keep in mind that the sample aspect ratio set by the @code{setsar}
15643 filter may be changed by later filters in the filterchain, e.g. if
15644 another "setsar" or a "setdar" filter is applied.
15646 It accepts the following parameters:
15649 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
15650 Set the aspect ratio used by the filter.
15652 The parameter can be a floating point number string, an expression, or
15653 a string of the form @var{num}:@var{den}, where @var{num} and
15654 @var{den} are the numerator and denominator of the aspect ratio. If
15655 the parameter is not specified, it is assumed the value "0".
15656 In case the form "@var{num}:@var{den}" is used, the @code{:} character
15660 Set the maximum integer value to use for expressing numerator and
15661 denominator when reducing the expressed aspect ratio to a rational.
15662 Default value is @code{100}.
15666 The parameter @var{sar} is an expression containing
15667 the following constants:
15671 These are approximated values for the mathematical constants e
15672 (Euler's number), pi (Greek pi), and phi (the golden ratio).
15675 The input width and height.
15678 These are the same as @var{w} / @var{h}.
15681 The input sample aspect ratio.
15684 The input display aspect ratio. It is the same as
15685 (@var{w} / @var{h}) * @var{sar}.
15688 Horizontal and vertical chroma subsample values. For example, for the
15689 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15692 @subsection Examples
15697 To change the display aspect ratio to 16:9, specify one of the following:
15704 To change the sample aspect ratio to 10:11, specify:
15710 To set a display aspect ratio of 16:9, and specify a maximum integer value of
15711 1000 in the aspect ratio reduction, use the command:
15713 setdar=ratio=16/9:max=1000
15721 Force field for the output video frame.
15723 The @code{setfield} filter marks the interlace type field for the
15724 output frames. It does not change the input frame, but only sets the
15725 corresponding property, which affects how the frame is treated by
15726 following filters (e.g. @code{fieldorder} or @code{yadif}).
15728 The filter accepts the following options:
15733 Available values are:
15737 Keep the same field property.
15740 Mark the frame as bottom-field-first.
15743 Mark the frame as top-field-first.
15746 Mark the frame as progressive.
15753 Force frame parameter for the output video frame.
15755 The @code{setparams} filter marks interlace and color range for the
15756 output frames. It does not change the input frame, but only sets the
15757 corresponding property, which affects how the frame is treated by
15762 Available values are:
15766 Keep the same field property (default).
15769 Mark the frame as bottom-field-first.
15772 Mark the frame as top-field-first.
15775 Mark the frame as progressive.
15779 Available values are:
15783 Keep the same color range property (default).
15785 @item unspecified, unknown
15786 Mark the frame as unspecified color range.
15788 @item limited, tv, mpeg
15789 Mark the frame as limited range.
15791 @item full, pc, jpeg
15792 Mark the frame as full range.
15795 @item color_primaries
15796 Set the color primaries.
15797 Available values are:
15801 Keep the same color primaries property (default).
15818 Set the color transfer.
15819 Available values are:
15823 Keep the same color trc property (default).
15845 Set the colorspace.
15846 Available values are:
15850 Keep the same colorspace property (default).
15863 @item chroma-derived-nc
15864 @item chroma-derived-c
15871 Show a line containing various information for each input video frame.
15872 The input video is not modified.
15874 This filter supports the following options:
15878 Calculate checksums of each plane. By default enabled.
15881 The shown line contains a sequence of key/value pairs of the form
15882 @var{key}:@var{value}.
15884 The following values are shown in the output:
15888 The (sequential) number of the input frame, starting from 0.
15891 The Presentation TimeStamp of the input frame, expressed as a number of
15892 time base units. The time base unit depends on the filter input pad.
15895 The Presentation TimeStamp of the input frame, expressed as a number of
15899 The position of the frame in the input stream, or -1 if this information is
15900 unavailable and/or meaningless (for example in case of synthetic video).
15903 The pixel format name.
15906 The sample aspect ratio of the input frame, expressed in the form
15907 @var{num}/@var{den}.
15910 The size of the input frame. For the syntax of this option, check the
15911 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
15914 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
15915 for bottom field first).
15918 This is 1 if the frame is a key frame, 0 otherwise.
15921 The picture type of the input frame ("I" for an I-frame, "P" for a
15922 P-frame, "B" for a B-frame, or "?" for an unknown type).
15923 Also refer to the documentation of the @code{AVPictureType} enum and of
15924 the @code{av_get_picture_type_char} function defined in
15925 @file{libavutil/avutil.h}.
15928 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
15930 @item plane_checksum
15931 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
15932 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
15935 @section showpalette
15937 Displays the 256 colors palette of each frame. This filter is only relevant for
15938 @var{pal8} pixel format frames.
15940 It accepts the following option:
15944 Set the size of the box used to represent one palette color entry. Default is
15945 @code{30} (for a @code{30x30} pixel box).
15948 @section shuffleframes
15950 Reorder and/or duplicate and/or drop video frames.
15952 It accepts the following parameters:
15956 Set the destination indexes of input frames.
15957 This is space or '|' separated list of indexes that maps input frames to output
15958 frames. Number of indexes also sets maximal value that each index may have.
15959 '-1' index have special meaning and that is to drop frame.
15962 The first frame has the index 0. The default is to keep the input unchanged.
15964 @subsection Examples
15968 Swap second and third frame of every three frames of the input:
15970 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
15974 Swap 10th and 1st frame of every ten frames of the input:
15976 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
15980 @section shuffleplanes
15982 Reorder and/or duplicate video planes.
15984 It accepts the following parameters:
15989 The index of the input plane to be used as the first output plane.
15992 The index of the input plane to be used as the second output plane.
15995 The index of the input plane to be used as the third output plane.
15998 The index of the input plane to be used as the fourth output plane.
16002 The first plane has the index 0. The default is to keep the input unchanged.
16004 @subsection Examples
16008 Swap the second and third planes of the input:
16010 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
16014 @anchor{signalstats}
16015 @section signalstats
16016 Evaluate various visual metrics that assist in determining issues associated
16017 with the digitization of analog video media.
16019 By default the filter will log these metadata values:
16023 Display the minimal Y value contained within the input frame. Expressed in
16027 Display the Y value at the 10% percentile within the input frame. Expressed in
16031 Display the average Y value within the input frame. Expressed in range of
16035 Display the Y value at the 90% percentile within the input frame. Expressed in
16039 Display the maximum Y value contained within the input frame. Expressed in
16043 Display the minimal U value contained within the input frame. Expressed in
16047 Display the U value at the 10% percentile within the input frame. Expressed in
16051 Display the average U value within the input frame. Expressed in range of
16055 Display the U value at the 90% percentile within the input frame. Expressed in
16059 Display the maximum U value contained within the input frame. Expressed in
16063 Display the minimal V value contained within the input frame. Expressed in
16067 Display the V value at the 10% percentile within the input frame. Expressed in
16071 Display the average V value within the input frame. Expressed in range of
16075 Display the V value at the 90% percentile within the input frame. Expressed in
16079 Display the maximum V value contained within the input frame. Expressed in
16083 Display the minimal saturation value contained within the input frame.
16084 Expressed in range of [0-~181.02].
16087 Display the saturation value at the 10% percentile within the input frame.
16088 Expressed in range of [0-~181.02].
16091 Display the average saturation value within the input frame. Expressed in range
16095 Display the saturation value at the 90% percentile within the input frame.
16096 Expressed in range of [0-~181.02].
16099 Display the maximum saturation value contained within the input frame.
16100 Expressed in range of [0-~181.02].
16103 Display the median value for hue within the input frame. Expressed in range of
16107 Display the average value for hue within the input frame. Expressed in range of
16111 Display the average of sample value difference between all values of the Y
16112 plane in the current frame and corresponding values of the previous input frame.
16113 Expressed in range of [0-255].
16116 Display the average of sample value difference between all values of the U
16117 plane in the current frame and corresponding values of the previous input frame.
16118 Expressed in range of [0-255].
16121 Display the average of sample value difference between all values of the V
16122 plane in the current frame and corresponding values of the previous input frame.
16123 Expressed in range of [0-255].
16126 Display bit depth of Y plane in current frame.
16127 Expressed in range of [0-16].
16130 Display bit depth of U plane in current frame.
16131 Expressed in range of [0-16].
16134 Display bit depth of V plane in current frame.
16135 Expressed in range of [0-16].
16138 The filter accepts the following options:
16144 @option{stat} specify an additional form of image analysis.
16145 @option{out} output video with the specified type of pixel highlighted.
16147 Both options accept the following values:
16151 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
16152 unlike the neighboring pixels of the same field. Examples of temporal outliers
16153 include the results of video dropouts, head clogs, or tape tracking issues.
16156 Identify @var{vertical line repetition}. Vertical line repetition includes
16157 similar rows of pixels within a frame. In born-digital video vertical line
16158 repetition is common, but this pattern is uncommon in video digitized from an
16159 analog source. When it occurs in video that results from the digitization of an
16160 analog source it can indicate concealment from a dropout compensator.
16163 Identify pixels that fall outside of legal broadcast range.
16167 Set the highlight color for the @option{out} option. The default color is
16171 @subsection Examples
16175 Output data of various video metrics:
16177 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
16181 Output specific data about the minimum and maximum values of the Y plane per frame:
16183 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
16187 Playback video while highlighting pixels that are outside of broadcast range in red.
16189 ffplay example.mov -vf signalstats="out=brng:color=red"
16193 Playback video with signalstats metadata drawn over the frame.
16195 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
16198 The contents of signalstat_drawtext.txt used in the command are:
16201 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
16202 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
16203 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
16204 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
16212 Calculates the MPEG-7 Video Signature. The filter can handle more than one
16213 input. In this case the matching between the inputs can be calculated additionally.
16214 The filter always passes through the first input. The signature of each stream can
16215 be written into a file.
16217 It accepts the following options:
16221 Enable or disable the matching process.
16223 Available values are:
16227 Disable the calculation of a matching (default).
16229 Calculate the matching for the whole video and output whether the whole video
16230 matches or only parts.
16232 Calculate only until a matching is found or the video ends. Should be faster in
16237 Set the number of inputs. The option value must be a non negative integer.
16238 Default value is 1.
16241 Set the path to which the output is written. If there is more than one input,
16242 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
16243 integer), that will be replaced with the input number. If no filename is
16244 specified, no output will be written. This is the default.
16247 Choose the output format.
16249 Available values are:
16253 Use the specified binary representation (default).
16255 Use the specified xml representation.
16259 Set threshold to detect one word as similar. The option value must be an integer
16260 greater than zero. The default value is 9000.
16263 Set threshold to detect all words as similar. The option value must be an integer
16264 greater than zero. The default value is 60000.
16267 Set threshold to detect frames as similar. The option value must be an integer
16268 greater than zero. The default value is 116.
16271 Set the minimum length of a sequence in frames to recognize it as matching
16272 sequence. The option value must be a non negative integer value.
16273 The default value is 0.
16276 Set the minimum relation, that matching frames to all frames must have.
16277 The option value must be a double value between 0 and 1. The default value is 0.5.
16280 @subsection Examples
16284 To calculate the signature of an input video and store it in signature.bin:
16286 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
16290 To detect whether two videos match and store the signatures in XML format in
16291 signature0.xml and signature1.xml:
16293 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 -
16301 Blur the input video without impacting the outlines.
16303 It accepts the following options:
16306 @item luma_radius, lr
16307 Set the luma radius. The option value must be a float number in
16308 the range [0.1,5.0] that specifies the variance of the gaussian filter
16309 used to blur the image (slower if larger). Default value is 1.0.
16311 @item luma_strength, ls
16312 Set the luma strength. The option value must be a float number
16313 in the range [-1.0,1.0] that configures the blurring. A value included
16314 in [0.0,1.0] will blur the image whereas a value included in
16315 [-1.0,0.0] will sharpen the image. Default value is 1.0.
16317 @item luma_threshold, lt
16318 Set the luma threshold used as a coefficient to determine
16319 whether a pixel should be blurred or not. The option value must be an
16320 integer in the range [-30,30]. A value of 0 will filter all the image,
16321 a value included in [0,30] will filter flat areas and a value included
16322 in [-30,0] will filter edges. Default value is 0.
16324 @item chroma_radius, cr
16325 Set the chroma radius. The option value must be a float number in
16326 the range [0.1,5.0] that specifies the variance of the gaussian filter
16327 used to blur the image (slower if larger). Default value is @option{luma_radius}.
16329 @item chroma_strength, cs
16330 Set the chroma strength. The option value must be a float number
16331 in the range [-1.0,1.0] that configures the blurring. A value included
16332 in [0.0,1.0] will blur the image whereas a value included in
16333 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
16335 @item chroma_threshold, ct
16336 Set the chroma threshold used as a coefficient to determine
16337 whether a pixel should be blurred or not. The option value must be an
16338 integer in the range [-30,30]. A value of 0 will filter all the image,
16339 a value included in [0,30] will filter flat areas and a value included
16340 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
16343 If a chroma option is not explicitly set, the corresponding luma value
16348 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
16350 This filter takes in input two input videos, the first input is
16351 considered the "main" source and is passed unchanged to the
16352 output. The second input is used as a "reference" video for computing
16355 Both video inputs must have the same resolution and pixel format for
16356 this filter to work correctly. Also it assumes that both inputs
16357 have the same number of frames, which are compared one by one.
16359 The filter stores the calculated SSIM of each frame.
16361 The description of the accepted parameters follows.
16364 @item stats_file, f
16365 If specified the filter will use the named file to save the SSIM of
16366 each individual frame. When filename equals "-" the data is sent to
16370 The file printed if @var{stats_file} is selected, contains a sequence of
16371 key/value pairs of the form @var{key}:@var{value} for each compared
16374 A description of each shown parameter follows:
16378 sequential number of the input frame, starting from 1
16380 @item Y, U, V, R, G, B
16381 SSIM of the compared frames for the component specified by the suffix.
16384 SSIM of the compared frames for the whole frame.
16387 Same as above but in dB representation.
16390 This filter also supports the @ref{framesync} options.
16394 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16395 [main][ref] ssim="stats_file=stats.log" [out]
16398 On this example the input file being processed is compared with the
16399 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
16400 is stored in @file{stats.log}.
16402 Another example with both psnr and ssim at same time:
16404 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
16409 Convert between different stereoscopic image formats.
16411 The filters accept the following options:
16415 Set stereoscopic image format of input.
16417 Available values for input image formats are:
16420 side by side parallel (left eye left, right eye right)
16423 side by side crosseye (right eye left, left eye right)
16426 side by side parallel with half width resolution
16427 (left eye left, right eye right)
16430 side by side crosseye with half width resolution
16431 (right eye left, left eye right)
16434 above-below (left eye above, right eye below)
16437 above-below (right eye above, left eye below)
16440 above-below with half height resolution
16441 (left eye above, right eye below)
16444 above-below with half height resolution
16445 (right eye above, left eye below)
16448 alternating frames (left eye first, right eye second)
16451 alternating frames (right eye first, left eye second)
16454 interleaved rows (left eye has top row, right eye starts on next row)
16457 interleaved rows (right eye has top row, left eye starts on next row)
16460 interleaved columns, left eye first
16463 interleaved columns, right eye first
16465 Default value is @samp{sbsl}.
16469 Set stereoscopic image format of output.
16473 side by side parallel (left eye left, right eye right)
16476 side by side crosseye (right eye left, left eye right)
16479 side by side parallel with half width resolution
16480 (left eye left, right eye right)
16483 side by side crosseye with half width resolution
16484 (right eye left, left eye right)
16487 above-below (left eye above, right eye below)
16490 above-below (right eye above, left eye below)
16493 above-below with half height resolution
16494 (left eye above, right eye below)
16497 above-below with half height resolution
16498 (right eye above, left eye below)
16501 alternating frames (left eye first, right eye second)
16504 alternating frames (right eye first, left eye second)
16507 interleaved rows (left eye has top row, right eye starts on next row)
16510 interleaved rows (right eye has top row, left eye starts on next row)
16513 anaglyph red/blue gray
16514 (red filter on left eye, blue filter on right eye)
16517 anaglyph red/green gray
16518 (red filter on left eye, green filter on right eye)
16521 anaglyph red/cyan gray
16522 (red filter on left eye, cyan filter on right eye)
16525 anaglyph red/cyan half colored
16526 (red filter on left eye, cyan filter on right eye)
16529 anaglyph red/cyan color
16530 (red filter on left eye, cyan filter on right eye)
16533 anaglyph red/cyan color optimized with the least squares projection of dubois
16534 (red filter on left eye, cyan filter on right eye)
16537 anaglyph green/magenta gray
16538 (green filter on left eye, magenta filter on right eye)
16541 anaglyph green/magenta half colored
16542 (green filter on left eye, magenta filter on right eye)
16545 anaglyph green/magenta colored
16546 (green filter on left eye, magenta filter on right eye)
16549 anaglyph green/magenta color optimized with the least squares projection of dubois
16550 (green filter on left eye, magenta filter on right eye)
16553 anaglyph yellow/blue gray
16554 (yellow filter on left eye, blue filter on right eye)
16557 anaglyph yellow/blue half colored
16558 (yellow filter on left eye, blue filter on right eye)
16561 anaglyph yellow/blue colored
16562 (yellow filter on left eye, blue filter on right eye)
16565 anaglyph yellow/blue color optimized with the least squares projection of dubois
16566 (yellow filter on left eye, blue filter on right eye)
16569 mono output (left eye only)
16572 mono output (right eye only)
16575 checkerboard, left eye first
16578 checkerboard, right eye first
16581 interleaved columns, left eye first
16584 interleaved columns, right eye first
16590 Default value is @samp{arcd}.
16593 @subsection Examples
16597 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
16603 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
16609 @section streamselect, astreamselect
16610 Select video or audio streams.
16612 The filter accepts the following options:
16616 Set number of inputs. Default is 2.
16619 Set input indexes to remap to outputs.
16622 @subsection Commands
16624 The @code{streamselect} and @code{astreamselect} filter supports the following
16629 Set input indexes to remap to outputs.
16632 @subsection Examples
16636 Select first 5 seconds 1st stream and rest of time 2nd stream:
16638 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
16642 Same as above, but for audio:
16644 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
16649 Apply sobel operator to input video stream.
16651 The filter accepts the following option:
16655 Set which planes will be processed, unprocessed planes will be copied.
16656 By default value 0xf, all planes will be processed.
16659 Set value which will be multiplied with filtered result.
16662 Set value which will be added to filtered result.
16668 Apply a simple postprocessing filter that compresses and decompresses the image
16669 at several (or - in the case of @option{quality} level @code{6} - all) shifts
16670 and average the results.
16672 The filter accepts the following options:
16676 Set quality. This option defines the number of levels for averaging. It accepts
16677 an integer in the range 0-6. If set to @code{0}, the filter will have no
16678 effect. A value of @code{6} means the higher quality. For each increment of
16679 that value the speed drops by a factor of approximately 2. Default value is
16683 Force a constant quantization parameter. If not set, the filter will use the QP
16684 from the video stream (if available).
16687 Set thresholding mode. Available modes are:
16691 Set hard thresholding (default).
16693 Set soft thresholding (better de-ringing effect, but likely blurrier).
16696 @item use_bframe_qp
16697 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
16698 option may cause flicker since the B-Frames have often larger QP. Default is
16699 @code{0} (not enabled).
16704 Scale the input by applying one of the super-resolution methods based on
16705 convolutional neural networks. Supported models:
16709 Super-Resolution Convolutional Neural Network model (SRCNN).
16710 See @url{https://arxiv.org/abs/1501.00092}.
16713 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
16714 See @url{https://arxiv.org/abs/1609.05158}.
16717 Training scripts as well as scripts for model file (.pb) saving can be found at
16718 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
16719 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
16721 Native model files (.model) can be generated from TensorFlow model
16722 files (.pb) by using tools/python/convert.py
16724 The filter accepts the following options:
16728 Specify which DNN backend to use for model loading and execution. This option accepts
16729 the following values:
16733 Native implementation of DNN loading and execution.
16736 TensorFlow backend. To enable this backend you
16737 need to install the TensorFlow for C library (see
16738 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
16739 @code{--enable-libtensorflow}
16742 Default value is @samp{native}.
16745 Set path to model file specifying network architecture and its parameters.
16746 Note that different backends use different file formats. TensorFlow backend
16747 can load files for both formats, while native backend can load files for only
16751 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
16752 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
16753 input upscaled using bicubic upscaling with proper scale factor.
16759 Draw subtitles on top of input video using the libass library.
16761 To enable compilation of this filter you need to configure FFmpeg with
16762 @code{--enable-libass}. This filter also requires a build with libavcodec and
16763 libavformat to convert the passed subtitles file to ASS (Advanced Substation
16764 Alpha) subtitles format.
16766 The filter accepts the following options:
16770 Set the filename of the subtitle file to read. It must be specified.
16772 @item original_size
16773 Specify the size of the original video, the video for which the ASS file
16774 was composed. For the syntax of this option, check the
16775 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16776 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
16777 correctly scale the fonts if the aspect ratio has been changed.
16780 Set a directory path containing fonts that can be used by the filter.
16781 These fonts will be used in addition to whatever the font provider uses.
16784 Process alpha channel, by default alpha channel is untouched.
16787 Set subtitles input character encoding. @code{subtitles} filter only. Only
16788 useful if not UTF-8.
16790 @item stream_index, si
16791 Set subtitles stream index. @code{subtitles} filter only.
16794 Override default style or script info parameters of the subtitles. It accepts a
16795 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
16798 If the first key is not specified, it is assumed that the first value
16799 specifies the @option{filename}.
16801 For example, to render the file @file{sub.srt} on top of the input
16802 video, use the command:
16807 which is equivalent to:
16809 subtitles=filename=sub.srt
16812 To render the default subtitles stream from file @file{video.mkv}, use:
16814 subtitles=video.mkv
16817 To render the second subtitles stream from that file, use:
16819 subtitles=video.mkv:si=1
16822 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
16823 @code{DejaVu Serif}, use:
16825 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HCCFF0000'
16828 @section super2xsai
16830 Scale the input by 2x and smooth using the Super2xSaI (Scale and
16831 Interpolate) pixel art scaling algorithm.
16833 Useful for enlarging pixel art images without reducing sharpness.
16837 Swap two rectangular objects in video.
16839 This filter accepts the following options:
16849 Set 1st rect x coordinate.
16852 Set 1st rect y coordinate.
16855 Set 2nd rect x coordinate.
16858 Set 2nd rect y coordinate.
16860 All expressions are evaluated once for each frame.
16863 The all options are expressions containing the following constants:
16868 The input width and height.
16871 same as @var{w} / @var{h}
16874 input sample aspect ratio
16877 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
16880 The number of the input frame, starting from 0.
16883 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
16886 the position in the file of the input frame, NAN if unknown
16894 Apply telecine process to the video.
16896 This filter accepts the following options:
16905 The default value is @code{top}.
16909 A string of numbers representing the pulldown pattern you wish to apply.
16910 The default value is @code{23}.
16914 Some typical patterns:
16919 24p: 2332 (preferred)
16926 24p: 222222222223 ("Euro pulldown")
16933 Apply threshold effect to video stream.
16935 This filter needs four video streams to perform thresholding.
16936 First stream is stream we are filtering.
16937 Second stream is holding threshold values, third stream is holding min values,
16938 and last, fourth stream is holding max values.
16940 The filter accepts the following option:
16944 Set which planes will be processed, unprocessed planes will be copied.
16945 By default value 0xf, all planes will be processed.
16948 For example if first stream pixel's component value is less then threshold value
16949 of pixel component from 2nd threshold stream, third stream value will picked,
16950 otherwise fourth stream pixel component value will be picked.
16952 Using color source filter one can perform various types of thresholding:
16954 @subsection Examples
16958 Binary threshold, using gray color as threshold:
16960 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
16964 Inverted binary threshold, using gray color as threshold:
16966 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
16970 Truncate binary threshold, using gray color as threshold:
16972 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
16976 Threshold to zero, using gray color as threshold:
16978 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
16982 Inverted threshold to zero, using gray color as threshold:
16984 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
16989 Select the most representative frame in a given sequence of consecutive frames.
16991 The filter accepts the following options:
16995 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
16996 will pick one of them, and then handle the next batch of @var{n} frames until
16997 the end. Default is @code{100}.
17000 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
17001 value will result in a higher memory usage, so a high value is not recommended.
17003 @subsection Examples
17007 Extract one picture each 50 frames:
17013 Complete example of a thumbnail creation with @command{ffmpeg}:
17015 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
17021 Tile several successive frames together.
17023 The filter accepts the following options:
17028 Set the grid size (i.e. the number of lines and columns). For the syntax of
17029 this option, check the
17030 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17033 Set the maximum number of frames to render in the given area. It must be less
17034 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
17035 the area will be used.
17038 Set the outer border margin in pixels.
17041 Set the inner border thickness (i.e. the number of pixels between frames). For
17042 more advanced padding options (such as having different values for the edges),
17043 refer to the pad video filter.
17046 Specify the color of the unused area. For the syntax of this option, check the
17047 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
17048 The default value of @var{color} is "black".
17051 Set the number of frames to overlap when tiling several successive frames together.
17052 The value must be between @code{0} and @var{nb_frames - 1}.
17055 Set the number of frames to initially be empty before displaying first output frame.
17056 This controls how soon will one get first output frame.
17057 The value must be between @code{0} and @var{nb_frames - 1}.
17060 @subsection Examples
17064 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
17066 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
17068 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
17069 duplicating each output frame to accommodate the originally detected frame
17073 Display @code{5} pictures in an area of @code{3x2} frames,
17074 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
17075 mixed flat and named options:
17077 tile=3x2:nb_frames=5:padding=7:margin=2
17081 @section tinterlace
17083 Perform various types of temporal field interlacing.
17085 Frames are counted starting from 1, so the first input frame is
17088 The filter accepts the following options:
17093 Specify the mode of the interlacing. This option can also be specified
17094 as a value alone. See below for a list of values for this option.
17096 Available values are:
17100 Move odd frames into the upper field, even into the lower field,
17101 generating a double height frame at half frame rate.
17105 Frame 1 Frame 2 Frame 3 Frame 4
17107 11111 22222 33333 44444
17108 11111 22222 33333 44444
17109 11111 22222 33333 44444
17110 11111 22222 33333 44444
17124 Only output odd frames, even frames are dropped, generating a frame with
17125 unchanged height at half frame rate.
17130 Frame 1 Frame 2 Frame 3 Frame 4
17132 11111 22222 33333 44444
17133 11111 22222 33333 44444
17134 11111 22222 33333 44444
17135 11111 22222 33333 44444
17145 Only output even frames, odd frames are dropped, generating a frame with
17146 unchanged height at half frame rate.
17151 Frame 1 Frame 2 Frame 3 Frame 4
17153 11111 22222 33333 44444
17154 11111 22222 33333 44444
17155 11111 22222 33333 44444
17156 11111 22222 33333 44444
17166 Expand each frame to full height, but pad alternate lines with black,
17167 generating a frame with double height at the same input frame rate.
17172 Frame 1 Frame 2 Frame 3 Frame 4
17174 11111 22222 33333 44444
17175 11111 22222 33333 44444
17176 11111 22222 33333 44444
17177 11111 22222 33333 44444
17180 11111 ..... 33333 .....
17181 ..... 22222 ..... 44444
17182 11111 ..... 33333 .....
17183 ..... 22222 ..... 44444
17184 11111 ..... 33333 .....
17185 ..... 22222 ..... 44444
17186 11111 ..... 33333 .....
17187 ..... 22222 ..... 44444
17191 @item interleave_top, 4
17192 Interleave the upper field from odd frames with the lower field from
17193 even frames, generating a frame with unchanged height at half frame rate.
17198 Frame 1 Frame 2 Frame 3 Frame 4
17200 11111<- 22222 33333<- 44444
17201 11111 22222<- 33333 44444<-
17202 11111<- 22222 33333<- 44444
17203 11111 22222<- 33333 44444<-
17213 @item interleave_bottom, 5
17214 Interleave the lower field from odd frames with the upper field from
17215 even frames, generating a frame with unchanged height at half frame rate.
17220 Frame 1 Frame 2 Frame 3 Frame 4
17222 11111 22222<- 33333 44444<-
17223 11111<- 22222 33333<- 44444
17224 11111 22222<- 33333 44444<-
17225 11111<- 22222 33333<- 44444
17235 @item interlacex2, 6
17236 Double frame rate with unchanged height. Frames are inserted each
17237 containing the second temporal field from the previous input frame and
17238 the first temporal field from the next input frame. This mode relies on
17239 the top_field_first flag. Useful for interlaced video displays with no
17240 field synchronisation.
17245 Frame 1 Frame 2 Frame 3 Frame 4
17247 11111 22222 33333 44444
17248 11111 22222 33333 44444
17249 11111 22222 33333 44444
17250 11111 22222 33333 44444
17253 11111 22222 22222 33333 33333 44444 44444
17254 11111 11111 22222 22222 33333 33333 44444
17255 11111 22222 22222 33333 33333 44444 44444
17256 11111 11111 22222 22222 33333 33333 44444
17261 Move odd frames into the upper field, even into the lower field,
17262 generating a double height frame at same frame rate.
17267 Frame 1 Frame 2 Frame 3 Frame 4
17269 11111 22222 33333 44444
17270 11111 22222 33333 44444
17271 11111 22222 33333 44444
17272 11111 22222 33333 44444
17275 11111 33333 33333 55555
17276 22222 22222 44444 44444
17277 11111 33333 33333 55555
17278 22222 22222 44444 44444
17279 11111 33333 33333 55555
17280 22222 22222 44444 44444
17281 11111 33333 33333 55555
17282 22222 22222 44444 44444
17287 Numeric values are deprecated but are accepted for backward
17288 compatibility reasons.
17290 Default mode is @code{merge}.
17293 Specify flags influencing the filter process.
17295 Available value for @var{flags} is:
17298 @item low_pass_filter, vlpf
17299 Enable linear vertical low-pass filtering in the filter.
17300 Vertical low-pass filtering is required when creating an interlaced
17301 destination from a progressive source which contains high-frequency
17302 vertical detail. Filtering will reduce interlace 'twitter' and Moire
17305 @item complex_filter, cvlpf
17306 Enable complex vertical low-pass filtering.
17307 This will slightly less reduce interlace 'twitter' and Moire
17308 patterning but better retain detail and subjective sharpness impression.
17312 Vertical low-pass filtering can only be enabled for @option{mode}
17313 @var{interleave_top} and @var{interleave_bottom}.
17319 Mix successive video frames.
17321 A description of the accepted options follows.
17325 The number of successive frames to mix. If unspecified, it defaults to 3.
17328 Specify weight of each input video frame.
17329 Each weight is separated by space. If number of weights is smaller than
17330 number of @var{frames} last specified weight will be used for all remaining
17334 Specify scale, if it is set it will be multiplied with sum
17335 of each weight multiplied with pixel values to give final destination
17336 pixel value. By default @var{scale} is auto scaled to sum of weights.
17339 @subsection Examples
17343 Average 7 successive frames:
17345 tmix=frames=7:weights="1 1 1 1 1 1 1"
17349 Apply simple temporal convolution:
17351 tmix=frames=3:weights="-1 3 -1"
17355 Similar as above but only showing temporal differences:
17357 tmix=frames=3:weights="-1 2 -1":scale=1
17363 Tone map colors from different dynamic ranges.
17365 This filter expects data in single precision floating point, as it needs to
17366 operate on (and can output) out-of-range values. Another filter, such as
17367 @ref{zscale}, is needed to convert the resulting frame to a usable format.
17369 The tonemapping algorithms implemented only work on linear light, so input
17370 data should be linearized beforehand (and possibly correctly tagged).
17373 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
17376 @subsection Options
17377 The filter accepts the following options.
17381 Set the tone map algorithm to use.
17383 Possible values are:
17386 Do not apply any tone map, only desaturate overbright pixels.
17389 Hard-clip any out-of-range values. Use it for perfect color accuracy for
17390 in-range values, while distorting out-of-range values.
17393 Stretch the entire reference gamut to a linear multiple of the display.
17396 Fit a logarithmic transfer between the tone curves.
17399 Preserve overall image brightness with a simple curve, using nonlinear
17400 contrast, which results in flattening details and degrading color accuracy.
17403 Preserve both dark and bright details better than @var{reinhard}, at the cost
17404 of slightly darkening everything. Use it when detail preservation is more
17405 important than color and brightness accuracy.
17408 Smoothly map out-of-range values, while retaining contrast and colors for
17409 in-range material as much as possible. Use it when color accuracy is more
17410 important than detail preservation.
17416 Tune the tone mapping algorithm.
17418 This affects the following algorithms:
17424 Specifies the scale factor to use while stretching.
17428 Specifies the exponent of the function.
17432 Specify an extra linear coefficient to multiply into the signal before clipping.
17436 Specify the local contrast coefficient at the display peak.
17437 Default to 0.5, which means that in-gamut values will be about half as bright
17444 Specify the transition point from linear to mobius transform. Every value
17445 below this point is guaranteed to be mapped 1:1. The higher the value, the
17446 more accurate the result will be, at the cost of losing bright details.
17447 Default to 0.3, which due to the steep initial slope still preserves in-range
17448 colors fairly accurately.
17452 Apply desaturation for highlights that exceed this level of brightness. The
17453 higher the parameter, the more color information will be preserved. This
17454 setting helps prevent unnaturally blown-out colors for super-highlights, by
17455 (smoothly) turning into white instead. This makes images feel more natural,
17456 at the cost of reducing information about out-of-range colors.
17458 The default of 2.0 is somewhat conservative and will mostly just apply to
17459 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
17461 This option works only if the input frame has a supported color tag.
17464 Override signal/nominal/reference peak with this value. Useful when the
17465 embedded peak information in display metadata is not reliable or when tone
17466 mapping from a lower range to a higher range.
17471 Temporarily pad video frames.
17473 The filter accepts the following options:
17477 Specify number of delay frames before input video stream.
17480 Specify number of padding frames after input video stream.
17481 Set to -1 to pad indefinitely.
17484 Set kind of frames added to beginning of stream.
17485 Can be either @var{add} or @var{clone}.
17486 With @var{add} frames of solid-color are added.
17487 With @var{clone} frames are clones of first frame.
17490 Set kind of frames added to end of stream.
17491 Can be either @var{add} or @var{clone}.
17492 With @var{add} frames of solid-color are added.
17493 With @var{clone} frames are clones of last frame.
17495 @item start_duration, stop_duration
17496 Specify the duration of the start/stop delay. See
17497 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17498 for the accepted syntax.
17499 These options override @var{start} and @var{stop}.
17502 Specify the color of the padded area. For the syntax of this option,
17503 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
17504 manual,ffmpeg-utils}.
17506 The default value of @var{color} is "black".
17512 Transpose rows with columns in the input video and optionally flip it.
17514 It accepts the following parameters:
17519 Specify the transposition direction.
17521 Can assume the following values:
17523 @item 0, 4, cclock_flip
17524 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
17532 Rotate by 90 degrees clockwise, that is:
17540 Rotate by 90 degrees counterclockwise, that is:
17547 @item 3, 7, clock_flip
17548 Rotate by 90 degrees clockwise and vertically flip, that is:
17556 For values between 4-7, the transposition is only done if the input
17557 video geometry is portrait and not landscape. These values are
17558 deprecated, the @code{passthrough} option should be used instead.
17560 Numerical values are deprecated, and should be dropped in favor of
17561 symbolic constants.
17564 Do not apply the transposition if the input geometry matches the one
17565 specified by the specified value. It accepts the following values:
17568 Always apply transposition.
17570 Preserve portrait geometry (when @var{height} >= @var{width}).
17572 Preserve landscape geometry (when @var{width} >= @var{height}).
17575 Default value is @code{none}.
17578 For example to rotate by 90 degrees clockwise and preserve portrait
17581 transpose=dir=1:passthrough=portrait
17584 The command above can also be specified as:
17586 transpose=1:portrait
17589 @section transpose_npp
17591 Transpose rows with columns in the input video and optionally flip it.
17592 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
17594 It accepts the following parameters:
17599 Specify the transposition direction.
17601 Can assume the following values:
17604 Rotate by 90 degrees counterclockwise and vertically flip. (default)
17607 Rotate by 90 degrees clockwise.
17610 Rotate by 90 degrees counterclockwise.
17613 Rotate by 90 degrees clockwise and vertically flip.
17617 Do not apply the transposition if the input geometry matches the one
17618 specified by the specified value. It accepts the following values:
17621 Always apply transposition. (default)
17623 Preserve portrait geometry (when @var{height} >= @var{width}).
17625 Preserve landscape geometry (when @var{width} >= @var{height}).
17631 Trim the input so that the output contains one continuous subpart of the input.
17633 It accepts the following parameters:
17636 Specify the time of the start of the kept section, i.e. the frame with the
17637 timestamp @var{start} will be the first frame in the output.
17640 Specify the time of the first frame that will be dropped, i.e. the frame
17641 immediately preceding the one with the timestamp @var{end} will be the last
17642 frame in the output.
17645 This is the same as @var{start}, except this option sets the start timestamp
17646 in timebase units instead of seconds.
17649 This is the same as @var{end}, except this option sets the end timestamp
17650 in timebase units instead of seconds.
17653 The maximum duration of the output in seconds.
17656 The number of the first frame that should be passed to the output.
17659 The number of the first frame that should be dropped.
17662 @option{start}, @option{end}, and @option{duration} are expressed as time
17663 duration specifications; see
17664 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17665 for the accepted syntax.
17667 Note that the first two sets of the start/end options and the @option{duration}
17668 option look at the frame timestamp, while the _frame variants simply count the
17669 frames that pass through the filter. Also note that this filter does not modify
17670 the timestamps. If you wish for the output timestamps to start at zero, insert a
17671 setpts filter after the trim filter.
17673 If multiple start or end options are set, this filter tries to be greedy and
17674 keep all the frames that match at least one of the specified constraints. To keep
17675 only the part that matches all the constraints at once, chain multiple trim
17678 The defaults are such that all the input is kept. So it is possible to set e.g.
17679 just the end values to keep everything before the specified time.
17684 Drop everything except the second minute of input:
17686 ffmpeg -i INPUT -vf trim=60:120
17690 Keep only the first second:
17692 ffmpeg -i INPUT -vf trim=duration=1
17697 @section unpremultiply
17698 Apply alpha unpremultiply effect to input video stream using first plane
17699 of second stream as alpha.
17701 Both streams must have same dimensions and same pixel format.
17703 The filter accepts the following option:
17707 Set which planes will be processed, unprocessed planes will be copied.
17708 By default value 0xf, all planes will be processed.
17710 If the format has 1 or 2 components, then luma is bit 0.
17711 If the format has 3 or 4 components:
17712 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
17713 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
17714 If present, the alpha channel is always the last bit.
17717 Do not require 2nd input for processing, instead use alpha plane from input stream.
17723 Sharpen or blur the input video.
17725 It accepts the following parameters:
17728 @item luma_msize_x, lx
17729 Set the luma matrix horizontal size. It must be an odd integer between
17730 3 and 23. The default value is 5.
17732 @item luma_msize_y, ly
17733 Set the luma matrix vertical size. It must be an odd integer between 3
17734 and 23. The default value is 5.
17736 @item luma_amount, la
17737 Set the luma effect strength. It must be a floating point number, reasonable
17738 values lay between -1.5 and 1.5.
17740 Negative values will blur the input video, while positive values will
17741 sharpen it, a value of zero will disable the effect.
17743 Default value is 1.0.
17745 @item chroma_msize_x, cx
17746 Set the chroma matrix horizontal size. It must be an odd integer
17747 between 3 and 23. The default value is 5.
17749 @item chroma_msize_y, cy
17750 Set the chroma matrix vertical size. It must be an odd integer
17751 between 3 and 23. The default value is 5.
17753 @item chroma_amount, ca
17754 Set the chroma effect strength. It must be a floating point number, reasonable
17755 values lay between -1.5 and 1.5.
17757 Negative values will blur the input video, while positive values will
17758 sharpen it, a value of zero will disable the effect.
17760 Default value is 0.0.
17764 All parameters are optional and default to the equivalent of the
17765 string '5:5:1.0:5:5:0.0'.
17767 @subsection Examples
17771 Apply strong luma sharpen effect:
17773 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
17777 Apply a strong blur of both luma and chroma parameters:
17779 unsharp=7:7:-2:7:7:-2
17785 Apply ultra slow/simple postprocessing filter that compresses and decompresses
17786 the image at several (or - in the case of @option{quality} level @code{8} - all)
17787 shifts and average the results.
17789 The way this differs from the behavior of spp is that uspp actually encodes &
17790 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
17791 DCT similar to MJPEG.
17793 The filter accepts the following options:
17797 Set quality. This option defines the number of levels for averaging. It accepts
17798 an integer in the range 0-8. If set to @code{0}, the filter will have no
17799 effect. A value of @code{8} means the higher quality. For each increment of
17800 that value the speed drops by a factor of approximately 2. Default value is
17804 Force a constant quantization parameter. If not set, the filter will use the QP
17805 from the video stream (if available).
17808 @section vaguedenoiser
17810 Apply a wavelet based denoiser.
17812 It transforms each frame from the video input into the wavelet domain,
17813 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
17814 the obtained coefficients. It does an inverse wavelet transform after.
17815 Due to wavelet properties, it should give a nice smoothed result, and
17816 reduced noise, without blurring picture features.
17818 This filter accepts the following options:
17822 The filtering strength. The higher, the more filtered the video will be.
17823 Hard thresholding can use a higher threshold than soft thresholding
17824 before the video looks overfiltered. Default value is 2.
17827 The filtering method the filter will use.
17829 It accepts the following values:
17832 All values under the threshold will be zeroed.
17835 All values under the threshold will be zeroed. All values above will be
17836 reduced by the threshold.
17839 Scales or nullifies coefficients - intermediary between (more) soft and
17840 (less) hard thresholding.
17843 Default is garrote.
17846 Number of times, the wavelet will decompose the picture. Picture can't
17847 be decomposed beyond a particular point (typically, 8 for a 640x480
17848 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
17851 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
17854 A list of the planes to process. By default all planes are processed.
17857 @section vectorscope
17859 Display 2 color component values in the two dimensional graph (which is called
17862 This filter accepts the following options:
17866 Set vectorscope mode.
17868 It accepts the following values:
17871 Gray values are displayed on graph, higher brightness means more pixels have
17872 same component color value on location in graph. This is the default mode.
17875 Gray values are displayed on graph. Surrounding pixels values which are not
17876 present in video frame are drawn in gradient of 2 color components which are
17877 set by option @code{x} and @code{y}. The 3rd color component is static.
17880 Actual color components values present in video frame are displayed on graph.
17883 Similar as color2 but higher frequency of same values @code{x} and @code{y}
17884 on graph increases value of another color component, which is luminance by
17885 default values of @code{x} and @code{y}.
17888 Actual colors present in video frame are displayed on graph. If two different
17889 colors map to same position on graph then color with higher value of component
17890 not present in graph is picked.
17893 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
17894 component picked from radial gradient.
17898 Set which color component will be represented on X-axis. Default is @code{1}.
17901 Set which color component will be represented on Y-axis. Default is @code{2}.
17904 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
17905 of color component which represents frequency of (X, Y) location in graph.
17910 No envelope, this is default.
17913 Instant envelope, even darkest single pixel will be clearly highlighted.
17916 Hold maximum and minimum values presented in graph over time. This way you
17917 can still spot out of range values without constantly looking at vectorscope.
17920 Peak and instant envelope combined together.
17924 Set what kind of graticule to draw.
17932 Set graticule opacity.
17935 Set graticule flags.
17939 Draw graticule for white point.
17942 Draw graticule for black point.
17945 Draw color points short names.
17949 Set background opacity.
17951 @item lthreshold, l
17952 Set low threshold for color component not represented on X or Y axis.
17953 Values lower than this value will be ignored. Default is 0.
17954 Note this value is multiplied with actual max possible value one pixel component
17955 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
17958 @item hthreshold, h
17959 Set high threshold for color component not represented on X or Y axis.
17960 Values higher than this value will be ignored. Default is 1.
17961 Note this value is multiplied with actual max possible value one pixel component
17962 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
17963 is 0.9 * 255 = 230.
17965 @item colorspace, c
17966 Set what kind of colorspace to use when drawing graticule.
17975 @anchor{vidstabdetect}
17976 @section vidstabdetect
17978 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
17979 @ref{vidstabtransform} for pass 2.
17981 This filter generates a file with relative translation and rotation
17982 transform information about subsequent frames, which is then used by
17983 the @ref{vidstabtransform} filter.
17985 To enable compilation of this filter you need to configure FFmpeg with
17986 @code{--enable-libvidstab}.
17988 This filter accepts the following options:
17992 Set the path to the file used to write the transforms information.
17993 Default value is @file{transforms.trf}.
17996 Set how shaky the video is and how quick the camera is. It accepts an
17997 integer in the range 1-10, a value of 1 means little shakiness, a
17998 value of 10 means strong shakiness. Default value is 5.
18001 Set the accuracy of the detection process. It must be a value in the
18002 range 1-15. A value of 1 means low accuracy, a value of 15 means high
18003 accuracy. Default value is 15.
18006 Set stepsize of the search process. The region around minimum is
18007 scanned with 1 pixel resolution. Default value is 6.
18010 Set minimum contrast. Below this value a local measurement field is
18011 discarded. Must be a floating point value in the range 0-1. Default
18015 Set reference frame number for tripod mode.
18017 If enabled, the motion of the frames is compared to a reference frame
18018 in the filtered stream, identified by the specified number. The idea
18019 is to compensate all movements in a more-or-less static scene and keep
18020 the camera view absolutely still.
18022 If set to 0, it is disabled. The frames are counted starting from 1.
18025 Show fields and transforms in the resulting frames. It accepts an
18026 integer in the range 0-2. Default value is 0, which disables any
18030 @subsection Examples
18034 Use default values:
18040 Analyze strongly shaky movie and put the results in file
18041 @file{mytransforms.trf}:
18043 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
18047 Visualize the result of internal transformations in the resulting
18050 vidstabdetect=show=1
18054 Analyze a video with medium shakiness using @command{ffmpeg}:
18056 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
18060 @anchor{vidstabtransform}
18061 @section vidstabtransform
18063 Video stabilization/deshaking: pass 2 of 2,
18064 see @ref{vidstabdetect} for pass 1.
18066 Read a file with transform information for each frame and
18067 apply/compensate them. Together with the @ref{vidstabdetect}
18068 filter this can be used to deshake videos. See also
18069 @url{http://public.hronopik.de/vid.stab}. It is important to also use
18070 the @ref{unsharp} filter, see below.
18072 To enable compilation of this filter you need to configure FFmpeg with
18073 @code{--enable-libvidstab}.
18075 @subsection Options
18079 Set path to the file used to read the transforms. Default value is
18080 @file{transforms.trf}.
18083 Set the number of frames (value*2 + 1) used for lowpass filtering the
18084 camera movements. Default value is 10.
18086 For example a number of 10 means that 21 frames are used (10 in the
18087 past and 10 in the future) to smoothen the motion in the video. A
18088 larger value leads to a smoother video, but limits the acceleration of
18089 the camera (pan/tilt movements). 0 is a special case where a static
18090 camera is simulated.
18093 Set the camera path optimization algorithm.
18095 Accepted values are:
18098 gaussian kernel low-pass filter on camera motion (default)
18100 averaging on transformations
18104 Set maximal number of pixels to translate frames. Default value is -1,
18108 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
18109 value is -1, meaning no limit.
18112 Specify how to deal with borders that may be visible due to movement
18115 Available values are:
18118 keep image information from previous frame (default)
18120 fill the border black
18124 Invert transforms if set to 1. Default value is 0.
18127 Consider transforms as relative to previous frame if set to 1,
18128 absolute if set to 0. Default value is 0.
18131 Set percentage to zoom. A positive value will result in a zoom-in
18132 effect, a negative value in a zoom-out effect. Default value is 0 (no
18136 Set optimal zooming to avoid borders.
18138 Accepted values are:
18143 optimal static zoom value is determined (only very strong movements
18144 will lead to visible borders) (default)
18146 optimal adaptive zoom value is determined (no borders will be
18147 visible), see @option{zoomspeed}
18150 Note that the value given at zoom is added to the one calculated here.
18153 Set percent to zoom maximally each frame (enabled when
18154 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
18158 Specify type of interpolation.
18160 Available values are:
18165 linear only horizontal
18167 linear in both directions (default)
18169 cubic in both directions (slow)
18173 Enable virtual tripod mode if set to 1, which is equivalent to
18174 @code{relative=0:smoothing=0}. Default value is 0.
18176 Use also @code{tripod} option of @ref{vidstabdetect}.
18179 Increase log verbosity if set to 1. Also the detected global motions
18180 are written to the temporary file @file{global_motions.trf}. Default
18184 @subsection Examples
18188 Use @command{ffmpeg} for a typical stabilization with default values:
18190 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
18193 Note the use of the @ref{unsharp} filter which is always recommended.
18196 Zoom in a bit more and load transform data from a given file:
18198 vidstabtransform=zoom=5:input="mytransforms.trf"
18202 Smoothen the video even more:
18204 vidstabtransform=smoothing=30
18210 Flip the input video vertically.
18212 For example, to vertically flip a video with @command{ffmpeg}:
18214 ffmpeg -i in.avi -vf "vflip" out.avi
18219 Detect variable frame rate video.
18221 This filter tries to detect if the input is variable or constant frame rate.
18223 At end it will output number of frames detected as having variable delta pts,
18224 and ones with constant delta pts.
18225 If there was frames with variable delta, than it will also show min and max delta
18230 Boost or alter saturation.
18232 The filter accepts the following options:
18235 Set strength of boost if positive value or strength of alter if negative value.
18236 Default is 0. Allowed range is from -2 to 2.
18239 Set the red balance. Default is 1. Allowed range is from -10 to 10.
18242 Set the green balance. Default is 1. Allowed range is from -10 to 10.
18245 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
18248 Set the red luma coefficient.
18251 Set the green luma coefficient.
18254 Set the blue luma coefficient.
18257 If @code{intensity} is negative and this is set to 1, colors will change,
18258 otherwise colors will be less saturated, more towards gray.
18264 Make or reverse a natural vignetting effect.
18266 The filter accepts the following options:
18270 Set lens angle expression as a number of radians.
18272 The value is clipped in the @code{[0,PI/2]} range.
18274 Default value: @code{"PI/5"}
18278 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
18282 Set forward/backward mode.
18284 Available modes are:
18287 The larger the distance from the central point, the darker the image becomes.
18290 The larger the distance from the central point, the brighter the image becomes.
18291 This can be used to reverse a vignette effect, though there is no automatic
18292 detection to extract the lens @option{angle} and other settings (yet). It can
18293 also be used to create a burning effect.
18296 Default value is @samp{forward}.
18299 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
18301 It accepts the following values:
18304 Evaluate expressions only once during the filter initialization.
18307 Evaluate expressions for each incoming frame. This is way slower than the
18308 @samp{init} mode since it requires all the scalers to be re-computed, but it
18309 allows advanced dynamic expressions.
18312 Default value is @samp{init}.
18315 Set dithering to reduce the circular banding effects. Default is @code{1}
18319 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
18320 Setting this value to the SAR of the input will make a rectangular vignetting
18321 following the dimensions of the video.
18323 Default is @code{1/1}.
18326 @subsection Expressions
18328 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
18329 following parameters.
18334 input width and height
18337 the number of input frame, starting from 0
18340 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
18341 @var{TB} units, NAN if undefined
18344 frame rate of the input video, NAN if the input frame rate is unknown
18347 the PTS (Presentation TimeStamp) of the filtered video frame,
18348 expressed in seconds, NAN if undefined
18351 time base of the input video
18355 @subsection Examples
18359 Apply simple strong vignetting effect:
18365 Make a flickering vignetting:
18367 vignette='PI/4+random(1)*PI/50':eval=frame
18372 @section vmafmotion
18374 Obtain the average vmaf motion score of a video.
18375 It is one of the component filters of VMAF.
18377 The obtained average motion score is printed through the logging system.
18379 In the below example the input file @file{ref.mpg} is being processed and score
18383 ffmpeg -i ref.mpg -lavfi vmafmotion -f null -
18387 Stack input videos vertically.
18389 All streams must be of same pixel format and of same width.
18391 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
18392 to create same output.
18394 The filter accept the following option:
18398 Set number of input streams. Default is 2.
18401 If set to 1, force the output to terminate when the shortest input
18402 terminates. Default value is 0.
18407 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
18408 Deinterlacing Filter").
18410 Based on the process described by Martin Weston for BBC R&D, and
18411 implemented based on the de-interlace algorithm written by Jim
18412 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
18413 uses filter coefficients calculated by BBC R&D.
18415 This filter use field-dominance information in frame to decide which
18416 of each pair of fields to place first in the output.
18417 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
18419 There are two sets of filter coefficients, so called "simple":
18420 and "complex". Which set of filter coefficients is used can
18421 be set by passing an optional parameter:
18425 Set the interlacing filter coefficients. Accepts one of the following values:
18429 Simple filter coefficient set.
18431 More-complex filter coefficient set.
18433 Default value is @samp{complex}.
18436 Specify which frames to deinterlace. Accept one of the following values:
18440 Deinterlace all frames,
18442 Only deinterlace frames marked as interlaced.
18445 Default value is @samp{all}.
18449 Video waveform monitor.
18451 The waveform monitor plots color component intensity. By default luminance
18452 only. Each column of the waveform corresponds to a column of pixels in the
18455 It accepts the following options:
18459 Can be either @code{row}, or @code{column}. Default is @code{column}.
18460 In row mode, the graph on the left side represents color component value 0 and
18461 the right side represents value = 255. In column mode, the top side represents
18462 color component value = 0 and bottom side represents value = 255.
18465 Set intensity. Smaller values are useful to find out how many values of the same
18466 luminance are distributed across input rows/columns.
18467 Default value is @code{0.04}. Allowed range is [0, 1].
18470 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
18471 In mirrored mode, higher values will be represented on the left
18472 side for @code{row} mode and at the top for @code{column} mode. Default is
18473 @code{1} (mirrored).
18477 It accepts the following values:
18480 Presents information identical to that in the @code{parade}, except
18481 that the graphs representing color components are superimposed directly
18484 This display mode makes it easier to spot relative differences or similarities
18485 in overlapping areas of the color components that are supposed to be identical,
18486 such as neutral whites, grays, or blacks.
18489 Display separate graph for the color components side by side in
18490 @code{row} mode or one below the other in @code{column} mode.
18493 Display separate graph for the color components side by side in
18494 @code{column} mode or one below the other in @code{row} mode.
18496 Using this display mode makes it easy to spot color casts in the highlights
18497 and shadows of an image, by comparing the contours of the top and the bottom
18498 graphs of each waveform. Since whites, grays, and blacks are characterized
18499 by exactly equal amounts of red, green, and blue, neutral areas of the picture
18500 should display three waveforms of roughly equal width/height. If not, the
18501 correction is easy to perform by making level adjustments the three waveforms.
18503 Default is @code{stack}.
18505 @item components, c
18506 Set which color components to display. Default is 1, which means only luminance
18507 or red color component if input is in RGB colorspace. If is set for example to
18508 7 it will display all 3 (if) available color components.
18513 No envelope, this is default.
18516 Instant envelope, minimum and maximum values presented in graph will be easily
18517 visible even with small @code{step} value.
18520 Hold minimum and maximum values presented in graph across time. This way you
18521 can still spot out of range values without constantly looking at waveforms.
18524 Peak and instant envelope combined together.
18530 No filtering, this is default.
18533 Luma and chroma combined together.
18536 Similar as above, but shows difference between blue and red chroma.
18539 Similar as above, but use different colors.
18542 Displays only chroma.
18545 Displays actual color value on waveform.
18548 Similar as above, but with luma showing frequency of chroma values.
18552 Set which graticule to display.
18556 Do not display graticule.
18559 Display green graticule showing legal broadcast ranges.
18562 Display orange graticule showing legal broadcast ranges.
18566 Set graticule opacity.
18569 Set graticule flags.
18573 Draw numbers above lines. By default enabled.
18576 Draw dots instead of lines.
18580 Set scale used for displaying graticule.
18587 Default is digital.
18590 Set background opacity.
18593 @section weave, doubleweave
18595 The @code{weave} takes a field-based video input and join
18596 each two sequential fields into single frame, producing a new double
18597 height clip with half the frame rate and half the frame count.
18599 The @code{doubleweave} works same as @code{weave} but without
18600 halving frame rate and frame count.
18602 It accepts the following option:
18606 Set first field. Available values are:
18610 Set the frame as top-field-first.
18613 Set the frame as bottom-field-first.
18617 @subsection Examples
18621 Interlace video using @ref{select} and @ref{separatefields} filter:
18623 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
18628 Apply the xBR high-quality magnification filter which is designed for pixel
18629 art. It follows a set of edge-detection rules, see
18630 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
18632 It accepts the following option:
18636 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
18637 @code{3xBR} and @code{4} for @code{4xBR}.
18638 Default is @code{3}.
18642 Pick median pixels from several input videos.
18644 The filter accept the following options:
18648 Set number of inputs.
18649 Default is 3. Allowed range is from 3 to 255.
18650 If number of inputs is even number, than result will be mean value between two median values.
18653 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
18657 Stack video inputs into custom layout.
18659 All streams must be of same pixel format.
18661 The filter accept the following option:
18665 Set number of input streams. Default is 2.
18668 Specify layout of inputs.
18669 This option requires the desired layout configuration to be explicitly set by the user.
18670 This sets position of each video input in output. Each input
18671 is separated by '|'.
18672 The first number represents the column, and the second number represents the row.
18673 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
18674 where X is video input from which to take width or height.
18675 Multiple values can be used when separated by '+'. In such
18676 case values are summed together.
18678 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
18679 a layout must be set by the user.
18682 If set to 1, force the output to terminate when the shortest input
18683 terminates. Default value is 0.
18686 @subsection Examples
18690 Display 4 inputs into 2x2 grid,
18691 note that if inputs are of different sizes unused gaps might appear,
18692 as not all of output video is used.
18694 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
18698 Display 4 inputs into 1x4 grid,
18699 note that if inputs are of different sizes unused gaps might appear,
18700 as not all of output video is used.
18702 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
18706 Display 9 inputs into 3x3 grid,
18707 note that if inputs are of different sizes unused gaps might appear,
18708 as not all of output video is used.
18710 xstack=inputs=9:layout=w3_0|w3_h0+h2|w3_h0|0_h4|0_0|w3+w1_0|0_h1+h2|w3+w1_h0|w3+w1_h1+h2
18717 Deinterlace the input video ("yadif" means "yet another deinterlacing
18720 It accepts the following parameters:
18726 The interlacing mode to adopt. It accepts one of the following values:
18729 @item 0, send_frame
18730 Output one frame for each frame.
18731 @item 1, send_field
18732 Output one frame for each field.
18733 @item 2, send_frame_nospatial
18734 Like @code{send_frame}, but it skips the spatial interlacing check.
18735 @item 3, send_field_nospatial
18736 Like @code{send_field}, but it skips the spatial interlacing check.
18739 The default value is @code{send_frame}.
18742 The picture field parity assumed for the input interlaced video. It accepts one
18743 of the following values:
18747 Assume the top field is first.
18749 Assume the bottom field is first.
18751 Enable automatic detection of field parity.
18754 The default value is @code{auto}.
18755 If the interlacing is unknown or the decoder does not export this information,
18756 top field first will be assumed.
18759 Specify which frames to deinterlace. Accept one of the following
18764 Deinterlace all frames.
18765 @item 1, interlaced
18766 Only deinterlace frames marked as interlaced.
18769 The default value is @code{all}.
18772 @section yadif_cuda
18774 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
18775 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
18778 It accepts the following parameters:
18784 The interlacing mode to adopt. It accepts one of the following values:
18787 @item 0, send_frame
18788 Output one frame for each frame.
18789 @item 1, send_field
18790 Output one frame for each field.
18791 @item 2, send_frame_nospatial
18792 Like @code{send_frame}, but it skips the spatial interlacing check.
18793 @item 3, send_field_nospatial
18794 Like @code{send_field}, but it skips the spatial interlacing check.
18797 The default value is @code{send_frame}.
18800 The picture field parity assumed for the input interlaced video. It accepts one
18801 of the following values:
18805 Assume the top field is first.
18807 Assume the bottom field is first.
18809 Enable automatic detection of field parity.
18812 The default value is @code{auto}.
18813 If the interlacing is unknown or the decoder does not export this information,
18814 top field first will be assumed.
18817 Specify which frames to deinterlace. Accept one of the following
18822 Deinterlace all frames.
18823 @item 1, interlaced
18824 Only deinterlace frames marked as interlaced.
18827 The default value is @code{all}.
18832 Apply Zoom & Pan effect.
18834 This filter accepts the following options:
18838 Set the zoom expression. Range is 1-10. Default is 1.
18842 Set the x and y expression. Default is 0.
18845 Set the duration expression in number of frames.
18846 This sets for how many number of frames effect will last for
18847 single input image.
18850 Set the output image size, default is 'hd720'.
18853 Set the output frame rate, default is '25'.
18856 Each expression can contain the following constants:
18875 Output frame count.
18879 Last calculated 'x' and 'y' position from 'x' and 'y' expression
18880 for current input frame.
18884 'x' and 'y' of last output frame of previous input frame or 0 when there was
18885 not yet such frame (first input frame).
18888 Last calculated zoom from 'z' expression for current input frame.
18891 Last calculated zoom of last output frame of previous input frame.
18894 Number of output frames for current input frame. Calculated from 'd' expression
18895 for each input frame.
18898 number of output frames created for previous input frame
18901 Rational number: input width / input height
18904 sample aspect ratio
18907 display aspect ratio
18911 @subsection Examples
18915 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
18917 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
18921 Zoom-in up to 1.5 and pan always at center of picture:
18923 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
18927 Same as above but without pausing:
18929 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
18935 Scale (resize) the input video, using the z.lib library:
18936 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
18937 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
18939 The zscale filter forces the output display aspect ratio to be the same
18940 as the input, by changing the output sample aspect ratio.
18942 If the input image format is different from the format requested by
18943 the next filter, the zscale filter will convert the input to the
18946 @subsection Options
18947 The filter accepts the following options.
18952 Set the output video dimension expression. Default value is the input
18955 If the @var{width} or @var{w} value is 0, the input width is used for
18956 the output. If the @var{height} or @var{h} value is 0, the input height
18957 is used for the output.
18959 If one and only one of the values is -n with n >= 1, the zscale filter
18960 will use a value that maintains the aspect ratio of the input image,
18961 calculated from the other specified dimension. After that it will,
18962 however, make sure that the calculated dimension is divisible by n and
18963 adjust the value if necessary.
18965 If both values are -n with n >= 1, the behavior will be identical to
18966 both values being set to 0 as previously detailed.
18968 See below for the list of accepted constants for use in the dimension
18972 Set the video size. For the syntax of this option, check the
18973 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18976 Set the dither type.
18978 Possible values are:
18983 @item error_diffusion
18989 Set the resize filter type.
18991 Possible values are:
19001 Default is bilinear.
19004 Set the color range.
19006 Possible values are:
19013 Default is same as input.
19016 Set the color primaries.
19018 Possible values are:
19028 Default is same as input.
19031 Set the transfer characteristics.
19033 Possible values are:
19047 Default is same as input.
19050 Set the colorspace matrix.
19052 Possible value are:
19063 Default is same as input.
19066 Set the input color range.
19068 Possible values are:
19075 Default is same as input.
19077 @item primariesin, pin
19078 Set the input color primaries.
19080 Possible values are:
19090 Default is same as input.
19092 @item transferin, tin
19093 Set the input transfer characteristics.
19095 Possible values are:
19106 Default is same as input.
19108 @item matrixin, min
19109 Set the input colorspace matrix.
19111 Possible value are:
19123 Set the output chroma location.
19125 Possible values are:
19136 @item chromalin, cin
19137 Set the input chroma location.
19139 Possible values are:
19151 Set the nominal peak luminance.
19154 The values of the @option{w} and @option{h} options are expressions
19155 containing the following constants:
19160 The input width and height
19164 These are the same as @var{in_w} and @var{in_h}.
19168 The output (scaled) width and height
19172 These are the same as @var{out_w} and @var{out_h}
19175 The same as @var{iw} / @var{ih}
19178 input sample aspect ratio
19181 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
19185 horizontal and vertical input chroma subsample values. For example for the
19186 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
19190 horizontal and vertical output chroma subsample values. For example for the
19191 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
19197 @c man end VIDEO FILTERS
19199 @chapter OpenCL Video Filters
19200 @c man begin OPENCL VIDEO FILTERS
19202 Below is a description of the currently available OpenCL video filters.
19204 To enable compilation of these filters you need to configure FFmpeg with
19205 @code{--enable-opencl}.
19207 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
19210 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
19211 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
19212 given device parameters.
19214 @item -filter_hw_device @var{name}
19215 Pass the hardware device called @var{name} to all filters in any filter graph.
19219 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
19223 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
19225 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
19229 Since OpenCL filters are not able to access frame data in normal memory, all frame data needs to be uploaded(@ref{hwupload}) to hardware surfaces connected to the appropriate device before being used and then downloaded(@ref{hwdownload}) back to normal memory. Note that @ref{hwupload} will upload to a surface with the same layout as the software frame, so it may be necessary to add a @ref{format} filter immediately before to get the input into the right format and @ref{hwdownload} does not support all formats on the output - it may be necessary to insert an additional @ref{format} filter immediately following in the graph to get the output in a supported format.
19231 @section avgblur_opencl
19233 Apply average blur filter.
19235 The filter accepts the following options:
19239 Set horizontal radius size.
19240 Range is @code{[1, 1024]} and default value is @code{1}.
19243 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19246 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
19249 @subsection Example
19253 Apply average blur filter with horizontal and vertical size of 3, setting each pixel of the output to the average value of the 7x7 region centered on it in the input. For pixels on the edges of the image, the region does not extend beyond the image boundaries, and so out-of-range coordinates are not used in the calculations.
19255 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
19259 @section boxblur_opencl
19261 Apply a boxblur algorithm to the input video.
19263 It accepts the following parameters:
19267 @item luma_radius, lr
19268 @item luma_power, lp
19269 @item chroma_radius, cr
19270 @item chroma_power, cp
19271 @item alpha_radius, ar
19272 @item alpha_power, ap
19276 A description of the accepted options follows.
19279 @item luma_radius, lr
19280 @item chroma_radius, cr
19281 @item alpha_radius, ar
19282 Set an expression for the box radius in pixels used for blurring the
19283 corresponding input plane.
19285 The radius value must be a non-negative number, and must not be
19286 greater than the value of the expression @code{min(w,h)/2} for the
19287 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
19290 Default value for @option{luma_radius} is "2". If not specified,
19291 @option{chroma_radius} and @option{alpha_radius} default to the
19292 corresponding value set for @option{luma_radius}.
19294 The expressions can contain the following constants:
19298 The input width and height in pixels.
19302 The input chroma image width and height in pixels.
19306 The horizontal and vertical chroma subsample values. For example, for the
19307 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
19310 @item luma_power, lp
19311 @item chroma_power, cp
19312 @item alpha_power, ap
19313 Specify how many times the boxblur filter is applied to the
19314 corresponding plane.
19316 Default value for @option{luma_power} is 2. If not specified,
19317 @option{chroma_power} and @option{alpha_power} default to the
19318 corresponding value set for @option{luma_power}.
19320 A value of 0 will disable the effect.
19323 @subsection Examples
19325 Apply boxblur filter, setting each pixel of the output to the average value of box-radiuses @var{luma_radius}, @var{chroma_radius}, @var{alpha_radius} for each plane respectively. The filter will apply @var{luma_power}, @var{chroma_power}, @var{alpha_power} times onto the corresponding plane. For pixels on the edges of the image, the radius does not extend beyond the image boundaries, and so out-of-range coordinates are not used in the calculations.
19329 Apply a boxblur filter with the luma, chroma, and alpha radius
19330 set to 2 and luma, chroma, and alpha power set to 3. The filter will run 3 times with box-radius set to 2 for every plane of the image.
19332 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
19333 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
19337 Apply a boxblur filter with luma radius set to 2, luma_power to 1, chroma_radius to 4, chroma_power to 5, alpha_radius to 3 and alpha_power to 7.
19339 For the luma plane, a 2x2 box radius will be run once.
19341 For the chroma plane, a 4x4 box radius will be run 5 times.
19343 For the alpha plane, a 3x3 box radius will be run 7 times.
19345 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
19349 @section convolution_opencl
19351 Apply convolution of 3x3, 5x5, 7x7 matrix.
19353 The filter accepts the following options:
19360 Set matrix for each plane.
19361 Matrix is sequence of 9, 25 or 49 signed numbers.
19362 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
19368 Set multiplier for calculated value for each plane.
19369 If unset or 0, it will be sum of all matrix elements.
19370 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
19376 Set bias for each plane. This value is added to the result of the multiplication.
19377 Useful for making the overall image brighter or darker.
19378 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
19382 @subsection Examples
19388 -i INPUT -vf "hwupload, convolution_opencl=0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0:0 -1 0 -1 5 -1 0 -1 0, hwdownload" OUTPUT
19394 -i INPUT -vf "hwupload, convolution_opencl=1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1 1 1 1 1 1 1 1 1:1/9:1/9:1/9:1/9, hwdownload" OUTPUT
19398 Apply edge enhance:
19400 -i INPUT -vf "hwupload, convolution_opencl=0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:0 0 0 -1 1 0 0 0 0:5:1:1:1:0:128:128:128, hwdownload" OUTPUT
19406 -i INPUT -vf "hwupload, convolution_opencl=0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:0 1 0 1 -4 1 0 1 0:5:5:5:1:0:128:128:128, hwdownload" OUTPUT
19410 Apply laplacian edge detector which includes diagonals:
19412 -i INPUT -vf "hwupload, convolution_opencl=1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:1 1 1 1 -8 1 1 1 1:5:5:5:1:0:128:128:0, hwdownload" OUTPUT
19418 -i INPUT -vf "hwupload, convolution_opencl=-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2:-2 -1 0 -1 1 1 0 1 2, hwdownload" OUTPUT
19422 @section dilation_opencl
19424 Apply dilation effect to the video.
19426 This filter replaces the pixel by the local(3x3) maximum.
19428 It accepts the following options:
19435 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
19436 If @code{0}, plane will remain unchanged.
19439 Flag which specifies the pixel to refer to.
19440 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
19442 Flags to local 3x3 coordinates region centered on @code{x}:
19451 @subsection Example
19455 Apply dilation filter with threshold0 set to 30, threshold1 set 40, threshold2 set to 50 and coordinates set to 231, setting each pixel of the output to the local maximum between pixels: 1, 2, 3, 6, 7, 8 of the 3x3 region centered on it in the input. If the difference between input pixel and local maximum is more then threshold of the corresponding plane, output pixel will be set to input pixel + threshold of corresponding plane.
19457 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
19461 @section erosion_opencl
19463 Apply erosion effect to the video.
19465 This filter replaces the pixel by the local(3x3) minimum.
19467 It accepts the following options:
19474 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
19475 If @code{0}, plane will remain unchanged.
19478 Flag which specifies the pixel to refer to.
19479 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
19481 Flags to local 3x3 coordinates region centered on @code{x}:
19490 @subsection Example
19494 Apply erosion filter with threshold0 set to 30, threshold1 set 40, threshold2 set to 50 and coordinates set to 231, setting each pixel of the output to the local minimum between pixels: 1, 2, 3, 6, 7, 8 of the 3x3 region centered on it in the input. If the difference between input pixel and local minimum is more then threshold of the corresponding plane, output pixel will be set to input pixel - threshold of corresponding plane.
19496 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
19500 @section colorkey_opencl
19501 RGB colorspace color keying.
19503 The filter accepts the following options:
19507 The color which will be replaced with transparency.
19510 Similarity percentage with the key color.
19512 0.01 matches only the exact key color, while 1.0 matches everything.
19517 0.0 makes pixels either fully transparent, or not transparent at all.
19519 Higher values result in semi-transparent pixels, with a higher transparency
19520 the more similar the pixels color is to the key color.
19523 @subsection Examples
19527 Make every semi-green pixel in the input transparent with some slight blending:
19529 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
19533 @section nlmeans_opencl
19535 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
19537 @section overlay_opencl
19539 Overlay one video on top of another.
19541 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
19542 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
19544 The filter accepts the following options:
19549 Set the x coordinate of the overlaid video on the main video.
19550 Default value is @code{0}.
19553 Set the x coordinate of the overlaid video on the main video.
19554 Default value is @code{0}.
19558 @subsection Examples
19562 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
19564 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
19567 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
19569 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
19574 @section prewitt_opencl
19576 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
19578 The filter accepts the following option:
19582 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19585 Set value which will be multiplied with filtered result.
19586 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
19589 Set value which will be added to filtered result.
19590 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
19593 @subsection Example
19597 Apply the Prewitt operator with scale set to 2 and delta set to 10.
19599 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
19603 @section roberts_opencl
19604 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
19606 The filter accepts the following option:
19610 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19613 Set value which will be multiplied with filtered result.
19614 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
19617 Set value which will be added to filtered result.
19618 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
19621 @subsection Example
19625 Apply the Roberts cross operator with scale set to 2 and delta set to 10
19627 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
19631 @section sobel_opencl
19633 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
19635 The filter accepts the following option:
19639 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19642 Set value which will be multiplied with filtered result.
19643 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
19646 Set value which will be added to filtered result.
19647 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
19650 @subsection Example
19654 Apply sobel operator with scale set to 2 and delta set to 10
19656 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
19660 @section tonemap_opencl
19662 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
19664 It accepts the following parameters:
19668 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
19671 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
19674 Apply desaturation for highlights that exceed this level of brightness. The
19675 higher the parameter, the more color information will be preserved. This
19676 setting helps prevent unnaturally blown-out colors for super-highlights, by
19677 (smoothly) turning into white instead. This makes images feel more natural,
19678 at the cost of reducing information about out-of-range colors.
19680 The default value is 0.5, and the algorithm here is a little different from
19681 the cpu version tonemap currently. A setting of 0.0 disables this option.
19684 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
19685 is used to detect whether the scene has changed or not. If the distance between
19686 the current frame average brightness and the current running average exceeds
19687 a threshold value, we would re-calculate scene average and peak brightness.
19688 The default value is 0.2.
19691 Specify the output pixel format.
19693 Currently supported formats are:
19700 Set the output color range.
19702 Possible values are:
19708 Default is same as input.
19711 Set the output color primaries.
19713 Possible values are:
19719 Default is same as input.
19722 Set the output transfer characteristics.
19724 Possible values are:
19733 Set the output colorspace matrix.
19735 Possible value are:
19741 Default is same as input.
19745 @subsection Example
19749 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
19751 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
19755 @section unsharp_opencl
19757 Sharpen or blur the input video.
19759 It accepts the following parameters:
19762 @item luma_msize_x, lx
19763 Set the luma matrix horizontal size.
19764 Range is @code{[1, 23]} and default value is @code{5}.
19766 @item luma_msize_y, ly
19767 Set the luma matrix vertical size.
19768 Range is @code{[1, 23]} and default value is @code{5}.
19770 @item luma_amount, la
19771 Set the luma effect strength.
19772 Range is @code{[-10, 10]} and default value is @code{1.0}.
19774 Negative values will blur the input video, while positive values will
19775 sharpen it, a value of zero will disable the effect.
19777 @item chroma_msize_x, cx
19778 Set the chroma matrix horizontal size.
19779 Range is @code{[1, 23]} and default value is @code{5}.
19781 @item chroma_msize_y, cy
19782 Set the chroma matrix vertical size.
19783 Range is @code{[1, 23]} and default value is @code{5}.
19785 @item chroma_amount, ca
19786 Set the chroma effect strength.
19787 Range is @code{[-10, 10]} and default value is @code{0.0}.
19789 Negative values will blur the input video, while positive values will
19790 sharpen it, a value of zero will disable the effect.
19794 All parameters are optional and default to the equivalent of the
19795 string '5:5:1.0:5:5:0.0'.
19797 @subsection Examples
19801 Apply strong luma sharpen effect:
19803 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
19807 Apply a strong blur of both luma and chroma parameters:
19809 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
19813 @c man end OPENCL VIDEO FILTERS
19815 @chapter Video Sources
19816 @c man begin VIDEO SOURCES
19818 Below is a description of the currently available video sources.
19822 Buffer video frames, and make them available to the filter chain.
19824 This source is mainly intended for a programmatic use, in particular
19825 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
19827 It accepts the following parameters:
19832 Specify the size (width and height) of the buffered video frames. For the
19833 syntax of this option, check the
19834 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19837 The input video width.
19840 The input video height.
19843 A string representing the pixel format of the buffered video frames.
19844 It may be a number corresponding to a pixel format, or a pixel format
19848 Specify the timebase assumed by the timestamps of the buffered frames.
19851 Specify the frame rate expected for the video stream.
19853 @item pixel_aspect, sar
19854 The sample (pixel) aspect ratio of the input video.
19857 Specify the optional parameters to be used for the scale filter which
19858 is automatically inserted when an input change is detected in the
19859 input size or format.
19861 @item hw_frames_ctx
19862 When using a hardware pixel format, this should be a reference to an
19863 AVHWFramesContext describing input frames.
19868 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
19871 will instruct the source to accept video frames with size 320x240 and
19872 with format "yuv410p", assuming 1/24 as the timestamps timebase and
19873 square pixels (1:1 sample aspect ratio).
19874 Since the pixel format with name "yuv410p" corresponds to the number 6
19875 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
19876 this example corresponds to:
19878 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
19881 Alternatively, the options can be specified as a flat string, but this
19882 syntax is deprecated:
19884 @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}]
19888 Create a pattern generated by an elementary cellular automaton.
19890 The initial state of the cellular automaton can be defined through the
19891 @option{filename} and @option{pattern} options. If such options are
19892 not specified an initial state is created randomly.
19894 At each new frame a new row in the video is filled with the result of
19895 the cellular automaton next generation. The behavior when the whole
19896 frame is filled is defined by the @option{scroll} option.
19898 This source accepts the following options:
19902 Read the initial cellular automaton state, i.e. the starting row, from
19903 the specified file.
19904 In the file, each non-whitespace character is considered an alive
19905 cell, a newline will terminate the row, and further characters in the
19906 file will be ignored.
19909 Read the initial cellular automaton state, i.e. the starting row, from
19910 the specified string.
19912 Each non-whitespace character in the string is considered an alive
19913 cell, a newline will terminate the row, and further characters in the
19914 string will be ignored.
19917 Set the video rate, that is the number of frames generated per second.
19920 @item random_fill_ratio, ratio
19921 Set the random fill ratio for the initial cellular automaton row. It
19922 is a floating point number value ranging from 0 to 1, defaults to
19925 This option is ignored when a file or a pattern is specified.
19927 @item random_seed, seed
19928 Set the seed for filling randomly the initial row, must be an integer
19929 included between 0 and UINT32_MAX. If not specified, or if explicitly
19930 set to -1, the filter will try to use a good random seed on a best
19934 Set the cellular automaton rule, it is a number ranging from 0 to 255.
19935 Default value is 110.
19938 Set the size of the output video. For the syntax of this option, check the
19939 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19941 If @option{filename} or @option{pattern} is specified, the size is set
19942 by default to the width of the specified initial state row, and the
19943 height is set to @var{width} * PHI.
19945 If @option{size} is set, it must contain the width of the specified
19946 pattern string, and the specified pattern will be centered in the
19949 If a filename or a pattern string is not specified, the size value
19950 defaults to "320x518" (used for a randomly generated initial state).
19953 If set to 1, scroll the output upward when all the rows in the output
19954 have been already filled. If set to 0, the new generated row will be
19955 written over the top row just after the bottom row is filled.
19958 @item start_full, full
19959 If set to 1, completely fill the output with generated rows before
19960 outputting the first frame.
19961 This is the default behavior, for disabling set the value to 0.
19964 If set to 1, stitch the left and right row edges together.
19965 This is the default behavior, for disabling set the value to 0.
19968 @subsection Examples
19972 Read the initial state from @file{pattern}, and specify an output of
19975 cellauto=f=pattern:s=200x400
19979 Generate a random initial row with a width of 200 cells, with a fill
19982 cellauto=ratio=2/3:s=200x200
19986 Create a pattern generated by rule 18 starting by a single alive cell
19987 centered on an initial row with width 100:
19989 cellauto=p=@@:s=100x400:full=0:rule=18
19993 Specify a more elaborated initial pattern:
19995 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
20000 @anchor{coreimagesrc}
20001 @section coreimagesrc
20002 Video source generated on GPU using Apple's CoreImage API on OSX.
20004 This video source is a specialized version of the @ref{coreimage} video filter.
20005 Use a core image generator at the beginning of the applied filterchain to
20006 generate the content.
20008 The coreimagesrc video source accepts the following options:
20010 @item list_generators
20011 List all available generators along with all their respective options as well as
20012 possible minimum and maximum values along with the default values.
20014 list_generators=true
20018 Specify the size of the sourced video. For the syntax of this option, check the
20019 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20020 The default value is @code{320x240}.
20023 Specify the frame rate of the sourced video, as the number of frames
20024 generated per second. It has to be a string in the format
20025 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
20026 number or a valid video frame rate abbreviation. The default value is
20030 Set the sample aspect ratio of the sourced video.
20033 Set the duration of the sourced video. See
20034 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20035 for the accepted syntax.
20037 If not specified, or the expressed duration is negative, the video is
20038 supposed to be generated forever.
20041 Additionally, all options of the @ref{coreimage} video filter are accepted.
20042 A complete filterchain can be used for further processing of the
20043 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
20044 and examples for details.
20046 @subsection Examples
20051 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
20052 given as complete and escaped command-line for Apple's standard bash shell:
20054 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
20056 This example is equivalent to the QRCode example of @ref{coreimage} without the
20057 need for a nullsrc video source.
20061 @section mandelbrot
20063 Generate a Mandelbrot set fractal, and progressively zoom towards the
20064 point specified with @var{start_x} and @var{start_y}.
20066 This source accepts the following options:
20071 Set the terminal pts value. Default value is 400.
20074 Set the terminal scale value.
20075 Must be a floating point value. Default value is 0.3.
20078 Set the inner coloring mode, that is the algorithm used to draw the
20079 Mandelbrot fractal internal region.
20081 It shall assume one of the following values:
20086 Show time until convergence.
20088 Set color based on point closest to the origin of the iterations.
20093 Default value is @var{mincol}.
20096 Set the bailout value. Default value is 10.0.
20099 Set the maximum of iterations performed by the rendering
20100 algorithm. Default value is 7189.
20103 Set outer coloring mode.
20104 It shall assume one of following values:
20106 @item iteration_count
20107 Set iteration count mode.
20108 @item normalized_iteration_count
20109 set normalized iteration count mode.
20111 Default value is @var{normalized_iteration_count}.
20114 Set frame rate, expressed as number of frames per second. Default
20118 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
20119 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
20122 Set the initial scale value. Default value is 3.0.
20125 Set the initial x position. Must be a floating point value between
20126 -100 and 100. Default value is -0.743643887037158704752191506114774.
20129 Set the initial y position. Must be a floating point value between
20130 -100 and 100. Default value is -0.131825904205311970493132056385139.
20135 Generate various test patterns, as generated by the MPlayer test filter.
20137 The size of the generated video is fixed, and is 256x256.
20138 This source is useful in particular for testing encoding features.
20140 This source accepts the following options:
20145 Specify the frame rate of the sourced video, as the number of frames
20146 generated per second. It has to be a string in the format
20147 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
20148 number or a valid video frame rate abbreviation. The default value is
20152 Set the duration of the sourced video. See
20153 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20154 for the accepted syntax.
20156 If not specified, or the expressed duration is negative, the video is
20157 supposed to be generated forever.
20161 Set the number or the name of the test to perform. Supported tests are:
20177 Default value is "all", which will cycle through the list of all tests.
20182 mptestsrc=t=dc_luma
20185 will generate a "dc_luma" test pattern.
20187 @section frei0r_src
20189 Provide a frei0r source.
20191 To enable compilation of this filter you need to install the frei0r
20192 header and configure FFmpeg with @code{--enable-frei0r}.
20194 This source accepts the following parameters:
20199 The size of the video to generate. For the syntax of this option, check the
20200 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20203 The framerate of the generated video. It may be a string of the form
20204 @var{num}/@var{den} or a frame rate abbreviation.
20207 The name to the frei0r source to load. For more information regarding frei0r and
20208 how to set the parameters, read the @ref{frei0r} section in the video filters
20211 @item filter_params
20212 A '|'-separated list of parameters to pass to the frei0r source.
20216 For example, to generate a frei0r partik0l source with size 200x200
20217 and frame rate 10 which is overlaid on the overlay filter main input:
20219 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
20224 Generate a life pattern.
20226 This source is based on a generalization of John Conway's life game.
20228 The sourced input represents a life grid, each pixel represents a cell
20229 which can be in one of two possible states, alive or dead. Every cell
20230 interacts with its eight neighbours, which are the cells that are
20231 horizontally, vertically, or diagonally adjacent.
20233 At each interaction the grid evolves according to the adopted rule,
20234 which specifies the number of neighbor alive cells which will make a
20235 cell stay alive or born. The @option{rule} option allows one to specify
20238 This source accepts the following options:
20242 Set the file from which to read the initial grid state. In the file,
20243 each non-whitespace character is considered an alive cell, and newline
20244 is used to delimit the end of each row.
20246 If this option is not specified, the initial grid is generated
20250 Set the video rate, that is the number of frames generated per second.
20253 @item random_fill_ratio, ratio
20254 Set the random fill ratio for the initial random grid. It is a
20255 floating point number value ranging from 0 to 1, defaults to 1/PHI.
20256 It is ignored when a file is specified.
20258 @item random_seed, seed
20259 Set the seed for filling the initial random grid, must be an integer
20260 included between 0 and UINT32_MAX. If not specified, or if explicitly
20261 set to -1, the filter will try to use a good random seed on a best
20267 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
20268 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
20269 @var{NS} specifies the number of alive neighbor cells which make a
20270 live cell stay alive, and @var{NB} the number of alive neighbor cells
20271 which make a dead cell to become alive (i.e. to "born").
20272 "s" and "b" can be used in place of "S" and "B", respectively.
20274 Alternatively a rule can be specified by an 18-bits integer. The 9
20275 high order bits are used to encode the next cell state if it is alive
20276 for each number of neighbor alive cells, the low order bits specify
20277 the rule for "borning" new cells. Higher order bits encode for an
20278 higher number of neighbor cells.
20279 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
20280 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
20282 Default value is "S23/B3", which is the original Conway's game of life
20283 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
20284 cells, and will born a new cell if there are three alive cells around
20288 Set the size of the output video. For the syntax of this option, check the
20289 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20291 If @option{filename} is specified, the size is set by default to the
20292 same size of the input file. If @option{size} is set, it must contain
20293 the size specified in the input file, and the initial grid defined in
20294 that file is centered in the larger resulting area.
20296 If a filename is not specified, the size value defaults to "320x240"
20297 (used for a randomly generated initial grid).
20300 If set to 1, stitch the left and right grid edges together, and the
20301 top and bottom edges also. Defaults to 1.
20304 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
20305 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
20306 value from 0 to 255.
20309 Set the color of living (or new born) cells.
20312 Set the color of dead cells. If @option{mold} is set, this is the first color
20313 used to represent a dead cell.
20316 Set mold color, for definitely dead and moldy cells.
20318 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
20319 ffmpeg-utils manual,ffmpeg-utils}.
20322 @subsection Examples
20326 Read a grid from @file{pattern}, and center it on a grid of size
20329 life=f=pattern:s=300x300
20333 Generate a random grid of size 200x200, with a fill ratio of 2/3:
20335 life=ratio=2/3:s=200x200
20339 Specify a custom rule for evolving a randomly generated grid:
20345 Full example with slow death effect (mold) using @command{ffplay}:
20347 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
20354 @anchor{haldclutsrc}
20357 @anchor{pal100bars}
20358 @anchor{rgbtestsrc}
20360 @anchor{smptehdbars}
20363 @anchor{yuvtestsrc}
20364 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
20366 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
20368 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
20370 The @code{color} source provides an uniformly colored input.
20372 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
20373 @ref{haldclut} filter.
20375 The @code{nullsrc} source returns unprocessed video frames. It is
20376 mainly useful to be employed in analysis / debugging tools, or as the
20377 source for filters which ignore the input data.
20379 The @code{pal75bars} source generates a color bars pattern, based on
20380 EBU PAL recommendations with 75% color levels.
20382 The @code{pal100bars} source generates a color bars pattern, based on
20383 EBU PAL recommendations with 100% color levels.
20385 The @code{rgbtestsrc} source generates an RGB test pattern useful for
20386 detecting RGB vs BGR issues. You should see a red, green and blue
20387 stripe from top to bottom.
20389 The @code{smptebars} source generates a color bars pattern, based on
20390 the SMPTE Engineering Guideline EG 1-1990.
20392 The @code{smptehdbars} source generates a color bars pattern, based on
20393 the SMPTE RP 219-2002.
20395 The @code{testsrc} source generates a test video pattern, showing a
20396 color pattern, a scrolling gradient and a timestamp. This is mainly
20397 intended for testing purposes.
20399 The @code{testsrc2} source is similar to testsrc, but supports more
20400 pixel formats instead of just @code{rgb24}. This allows using it as an
20401 input for other tests without requiring a format conversion.
20403 The @code{yuvtestsrc} source generates an YUV test pattern. You should
20404 see a y, cb and cr stripe from top to bottom.
20406 The sources accept the following parameters:
20411 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
20412 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
20413 pixels to be used as identity matrix for 3D lookup tables. Each component is
20414 coded on a @code{1/(N*N)} scale.
20417 Specify the color of the source, only available in the @code{color}
20418 source. For the syntax of this option, check the
20419 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
20422 Specify the size of the sourced video. For the syntax of this option, check the
20423 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20424 The default value is @code{320x240}.
20426 This option is not available with the @code{allrgb}, @code{allyuv}, and
20427 @code{haldclutsrc} filters.
20430 Specify the frame rate of the sourced video, as the number of frames
20431 generated per second. It has to be a string in the format
20432 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
20433 number or a valid video frame rate abbreviation. The default value is
20437 Set the duration of the sourced video. See
20438 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
20439 for the accepted syntax.
20441 If not specified, or the expressed duration is negative, the video is
20442 supposed to be generated forever.
20445 Set the sample aspect ratio of the sourced video.
20448 Specify the alpha (opacity) of the background, only available in the
20449 @code{testsrc2} source. The value must be between 0 (fully transparent) and
20450 255 (fully opaque, the default).
20453 Set the number of decimals to show in the timestamp, only available in the
20454 @code{testsrc} source.
20456 The displayed timestamp value will correspond to the original
20457 timestamp value multiplied by the power of 10 of the specified
20458 value. Default value is 0.
20461 @subsection Examples
20465 Generate a video with a duration of 5.3 seconds, with size
20466 176x144 and a frame rate of 10 frames per second:
20468 testsrc=duration=5.3:size=qcif:rate=10
20472 The following graph description will generate a red source
20473 with an opacity of 0.2, with size "qcif" and a frame rate of 10
20476 color=c=red@@0.2:s=qcif:r=10
20480 If the input content is to be ignored, @code{nullsrc} can be used. The
20481 following command generates noise in the luminance plane by employing
20482 the @code{geq} filter:
20484 nullsrc=s=256x256, geq=random(1)*255:128:128
20488 @subsection Commands
20490 The @code{color} source supports the following commands:
20494 Set the color of the created image. Accepts the same syntax of the
20495 corresponding @option{color} option.
20500 Generate video using an OpenCL program.
20505 OpenCL program source file.
20508 Kernel name in program.
20511 Size of frames to generate. This must be set.
20514 Pixel format to use for the generated frames. This must be set.
20517 Number of frames generated every second. Default value is '25'.
20521 For details of how the program loading works, see the @ref{program_opencl}
20528 Generate a colour ramp by setting pixel values from the position of the pixel
20529 in the output image. (Note that this will work with all pixel formats, but
20530 the generated output will not be the same.)
20532 __kernel void ramp(__write_only image2d_t dst,
20533 unsigned int index)
20535 int2 loc = (int2)(get_global_id(0), get_global_id(1));
20538 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
20540 write_imagef(dst, loc, val);
20545 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
20547 __kernel void sierpinski_carpet(__write_only image2d_t dst,
20548 unsigned int index)
20550 int2 loc = (int2)(get_global_id(0), get_global_id(1));
20552 float4 value = 0.0f;
20553 int x = loc.x + index;
20554 int y = loc.y + index;
20555 while (x > 0 || y > 0) {
20556 if (x % 3 == 1 && y % 3 == 1) {
20564 write_imagef(dst, loc, value);
20570 @c man end VIDEO SOURCES
20572 @chapter Video Sinks
20573 @c man begin VIDEO SINKS
20575 Below is a description of the currently available video sinks.
20577 @section buffersink
20579 Buffer video frames, and make them available to the end of the filter
20582 This sink is mainly intended for programmatic use, in particular
20583 through the interface defined in @file{libavfilter/buffersink.h}
20584 or the options system.
20586 It accepts a pointer to an AVBufferSinkContext structure, which
20587 defines the incoming buffers' formats, to be passed as the opaque
20588 parameter to @code{avfilter_init_filter} for initialization.
20592 Null video sink: do absolutely nothing with the input video. It is
20593 mainly useful as a template and for use in analysis / debugging
20596 @c man end VIDEO SINKS
20598 @chapter Multimedia Filters
20599 @c man begin MULTIMEDIA FILTERS
20601 Below is a description of the currently available multimedia filters.
20605 Convert input audio to a video output, displaying the audio bit scope.
20607 The filter accepts the following options:
20611 Set frame rate, expressed as number of frames per second. Default
20615 Specify the video size for the output. For the syntax of this option, check the
20616 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20617 Default value is @code{1024x256}.
20620 Specify list of colors separated by space or by '|' which will be used to
20621 draw channels. Unrecognized or missing colors will be replaced
20625 @section ahistogram
20627 Convert input audio to a video output, displaying the volume histogram.
20629 The filter accepts the following options:
20633 Specify how histogram is calculated.
20635 It accepts the following values:
20638 Use single histogram for all channels.
20640 Use separate histogram for each channel.
20642 Default is @code{single}.
20645 Set frame rate, expressed as number of frames per second. Default
20649 Specify the video size for the output. For the syntax of this option, check the
20650 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20651 Default value is @code{hd720}.
20656 It accepts the following values:
20667 reverse logarithmic
20669 Default is @code{log}.
20672 Set amplitude scale.
20674 It accepts the following values:
20681 Default is @code{log}.
20684 Set how much frames to accumulate in histogram.
20685 Default is 1. Setting this to -1 accumulates all frames.
20688 Set histogram ratio of window height.
20691 Set sonogram sliding.
20693 It accepts the following values:
20696 replace old rows with new ones.
20698 scroll from top to bottom.
20700 Default is @code{replace}.
20703 @section aphasemeter
20705 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
20706 representing mean phase of current audio frame. A video output can also be produced and is
20707 enabled by default. The audio is passed through as first output.
20709 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
20710 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
20711 and @code{1} means channels are in phase.
20713 The filter accepts the following options, all related to its video output:
20717 Set the output frame rate. Default value is @code{25}.
20720 Set the video size for the output. For the syntax of this option, check the
20721 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20722 Default value is @code{800x400}.
20727 Specify the red, green, blue contrast. Default values are @code{2},
20728 @code{7} and @code{1}.
20729 Allowed range is @code{[0, 255]}.
20732 Set color which will be used for drawing median phase. If color is
20733 @code{none} which is default, no median phase value will be drawn.
20736 Enable video output. Default is enabled.
20739 @section avectorscope
20741 Convert input audio to a video output, representing the audio vector
20744 The filter is used to measure the difference between channels of stereo
20745 audio stream. A monoaural signal, consisting of identical left and right
20746 signal, results in straight vertical line. Any stereo separation is visible
20747 as a deviation from this line, creating a Lissajous figure.
20748 If the straight (or deviation from it) but horizontal line appears this
20749 indicates that the left and right channels are out of phase.
20751 The filter accepts the following options:
20755 Set the vectorscope mode.
20757 Available values are:
20760 Lissajous rotated by 45 degrees.
20763 Same as above but not rotated.
20766 Shape resembling half of circle.
20769 Default value is @samp{lissajous}.
20772 Set the video size for the output. For the syntax of this option, check the
20773 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20774 Default value is @code{400x400}.
20777 Set the output frame rate. Default value is @code{25}.
20783 Specify the red, green, blue and alpha contrast. Default values are @code{40},
20784 @code{160}, @code{80} and @code{255}.
20785 Allowed range is @code{[0, 255]}.
20791 Specify the red, green, blue and alpha fade. Default values are @code{15},
20792 @code{10}, @code{5} and @code{5}.
20793 Allowed range is @code{[0, 255]}.
20796 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
20797 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
20800 Set the vectorscope drawing mode.
20802 Available values are:
20805 Draw dot for each sample.
20808 Draw line between previous and current sample.
20811 Default value is @samp{dot}.
20814 Specify amplitude scale of audio samples.
20816 Available values are:
20832 Swap left channel axis with right channel axis.
20842 Mirror only x axis.
20845 Mirror only y axis.
20853 @subsection Examples
20857 Complete example using @command{ffplay}:
20859 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
20860 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
20864 @section bench, abench
20866 Benchmark part of a filtergraph.
20868 The filter accepts the following options:
20872 Start or stop a timer.
20874 Available values are:
20877 Get the current time, set it as frame metadata (using the key
20878 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
20881 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
20882 the input frame metadata to get the time difference. Time difference, average,
20883 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
20884 @code{min}) are then printed. The timestamps are expressed in seconds.
20888 @subsection Examples
20892 Benchmark @ref{selectivecolor} filter:
20894 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
20900 Concatenate audio and video streams, joining them together one after the
20903 The filter works on segments of synchronized video and audio streams. All
20904 segments must have the same number of streams of each type, and that will
20905 also be the number of streams at output.
20907 The filter accepts the following options:
20912 Set the number of segments. Default is 2.
20915 Set the number of output video streams, that is also the number of video
20916 streams in each segment. Default is 1.
20919 Set the number of output audio streams, that is also the number of audio
20920 streams in each segment. Default is 0.
20923 Activate unsafe mode: do not fail if segments have a different format.
20927 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
20928 @var{a} audio outputs.
20930 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
20931 segment, in the same order as the outputs, then the inputs for the second
20934 Related streams do not always have exactly the same duration, for various
20935 reasons including codec frame size or sloppy authoring. For that reason,
20936 related synchronized streams (e.g. a video and its audio track) should be
20937 concatenated at once. The concat filter will use the duration of the longest
20938 stream in each segment (except the last one), and if necessary pad shorter
20939 audio streams with silence.
20941 For this filter to work correctly, all segments must start at timestamp 0.
20943 All corresponding streams must have the same parameters in all segments; the
20944 filtering system will automatically select a common pixel format for video
20945 streams, and a common sample format, sample rate and channel layout for
20946 audio streams, but other settings, such as resolution, must be converted
20947 explicitly by the user.
20949 Different frame rates are acceptable but will result in variable frame rate
20950 at output; be sure to configure the output file to handle it.
20952 @subsection Examples
20956 Concatenate an opening, an episode and an ending, all in bilingual version
20957 (video in stream 0, audio in streams 1 and 2):
20959 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
20960 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
20961 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
20962 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
20966 Concatenate two parts, handling audio and video separately, using the
20967 (a)movie sources, and adjusting the resolution:
20969 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
20970 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
20971 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
20973 Note that a desync will happen at the stitch if the audio and video streams
20974 do not have exactly the same duration in the first file.
20978 @subsection Commands
20980 This filter supports the following commands:
20983 Close the current segment and step to the next one
20986 @section drawgraph, adrawgraph
20988 Draw a graph using input video or audio metadata.
20990 It accepts the following parameters:
20994 Set 1st frame metadata key from which metadata values will be used to draw a graph.
20997 Set 1st foreground color expression.
21000 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
21003 Set 2nd foreground color expression.
21006 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
21009 Set 3rd foreground color expression.
21012 Set 4th frame metadata key from which metadata values will be used to draw a graph.
21015 Set 4th foreground color expression.
21018 Set minimal value of metadata value.
21021 Set maximal value of metadata value.
21024 Set graph background color. Default is white.
21029 Available values for mode is:
21036 Default is @code{line}.
21041 Available values for slide is:
21044 Draw new frame when right border is reached.
21047 Replace old columns with new ones.
21050 Scroll from right to left.
21053 Scroll from left to right.
21056 Draw single picture.
21059 Default is @code{frame}.
21062 Set size of graph video. For the syntax of this option, check the
21063 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21064 The default value is @code{900x256}.
21066 The foreground color expressions can use the following variables:
21069 Minimal value of metadata value.
21072 Maximal value of metadata value.
21075 Current metadata key value.
21078 The color is defined as 0xAABBGGRR.
21081 Example using metadata from @ref{signalstats} filter:
21083 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
21086 Example using metadata from @ref{ebur128} filter:
21088 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
21094 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
21095 level. By default, it logs a message at a frequency of 10Hz with the
21096 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
21097 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
21099 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
21100 sample format is double-precision floating point. The input stream will be converted to
21101 this specification, if needed. Users may need to insert aformat and/or aresample filters
21102 after this filter to obtain the original parameters.
21104 The filter also has a video output (see the @var{video} option) with a real
21105 time graph to observe the loudness evolution. The graphic contains the logged
21106 message mentioned above, so it is not printed anymore when this option is set,
21107 unless the verbose logging is set. The main graphing area contains the
21108 short-term loudness (3 seconds of analysis), and the gauge on the right is for
21109 the momentary loudness (400 milliseconds), but can optionally be configured
21110 to instead display short-term loudness (see @var{gauge}).
21112 The green area marks a +/- 1LU target range around the target loudness
21113 (-23LUFS by default, unless modified through @var{target}).
21115 More information about the Loudness Recommendation EBU R128 on
21116 @url{http://tech.ebu.ch/loudness}.
21118 The filter accepts the following options:
21123 Activate the video output. The audio stream is passed unchanged whether this
21124 option is set or no. The video stream will be the first output stream if
21125 activated. Default is @code{0}.
21128 Set the video size. This option is for video only. For the syntax of this
21130 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21131 Default and minimum resolution is @code{640x480}.
21134 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
21135 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
21136 other integer value between this range is allowed.
21139 Set metadata injection. If set to @code{1}, the audio input will be segmented
21140 into 100ms output frames, each of them containing various loudness information
21141 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
21143 Default is @code{0}.
21146 Force the frame logging level.
21148 Available values are:
21151 information logging level
21153 verbose logging level
21156 By default, the logging level is set to @var{info}. If the @option{video} or
21157 the @option{metadata} options are set, it switches to @var{verbose}.
21162 Available modes can be cumulated (the option is a @code{flag} type). Possible
21166 Disable any peak mode (default).
21168 Enable sample-peak mode.
21170 Simple peak mode looking for the higher sample value. It logs a message
21171 for sample-peak (identified by @code{SPK}).
21173 Enable true-peak mode.
21175 If enabled, the peak lookup is done on an over-sampled version of the input
21176 stream for better peak accuracy. It logs a message for true-peak.
21177 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
21178 This mode requires a build with @code{libswresample}.
21182 Treat mono input files as "dual mono". If a mono file is intended for playback
21183 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
21184 If set to @code{true}, this option will compensate for this effect.
21185 Multi-channel input files are not affected by this option.
21188 Set a specific pan law to be used for the measurement of dual mono files.
21189 This parameter is optional, and has a default value of -3.01dB.
21192 Set a specific target level (in LUFS) used as relative zero in the visualization.
21193 This parameter is optional and has a default value of -23LUFS as specified
21194 by EBU R128. However, material published online may prefer a level of -16LUFS
21195 (e.g. for use with podcasts or video platforms).
21198 Set the value displayed by the gauge. Valid values are @code{momentary} and s
21199 @code{shortterm}. By default the momentary value will be used, but in certain
21200 scenarios it may be more useful to observe the short term value instead (e.g.
21204 Sets the display scale for the loudness. Valid parameters are @code{absolute}
21205 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
21206 video output, not the summary or continuous log output.
21209 @subsection Examples
21213 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
21215 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
21219 Run an analysis with @command{ffmpeg}:
21221 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
21225 @section interleave, ainterleave
21227 Temporally interleave frames from several inputs.
21229 @code{interleave} works with video inputs, @code{ainterleave} with audio.
21231 These filters read frames from several inputs and send the oldest
21232 queued frame to the output.
21234 Input streams must have well defined, monotonically increasing frame
21237 In order to submit one frame to output, these filters need to enqueue
21238 at least one frame for each input, so they cannot work in case one
21239 input is not yet terminated and will not receive incoming frames.
21241 For example consider the case when one input is a @code{select} filter
21242 which always drops input frames. The @code{interleave} filter will keep
21243 reading from that input, but it will never be able to send new frames
21244 to output until the input sends an end-of-stream signal.
21246 Also, depending on inputs synchronization, the filters will drop
21247 frames in case one input receives more frames than the other ones, and
21248 the queue is already filled.
21250 These filters accept the following options:
21254 Set the number of different inputs, it is 2 by default.
21257 @subsection Examples
21261 Interleave frames belonging to different streams using @command{ffmpeg}:
21263 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
21267 Add flickering blur effect:
21269 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
21273 @section metadata, ametadata
21275 Manipulate frame metadata.
21277 This filter accepts the following options:
21281 Set mode of operation of the filter.
21283 Can be one of the following:
21287 If both @code{value} and @code{key} is set, select frames
21288 which have such metadata. If only @code{key} is set, select
21289 every frame that has such key in metadata.
21292 Add new metadata @code{key} and @code{value}. If key is already available
21296 Modify value of already present key.
21299 If @code{value} is set, delete only keys that have such value.
21300 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
21304 Print key and its value if metadata was found. If @code{key} is not set print all
21305 metadata values available in frame.
21309 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
21312 Set metadata value which will be used. This option is mandatory for
21313 @code{modify} and @code{add} mode.
21316 Which function to use when comparing metadata value and @code{value}.
21318 Can be one of following:
21322 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
21325 Values are interpreted as strings, returns true if metadata value starts with
21326 the @code{value} option string.
21329 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
21332 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
21335 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
21338 Values are interpreted as floats, returns true if expression from option @code{expr}
21343 Set expression which is used when @code{function} is set to @code{expr}.
21344 The expression is evaluated through the eval API and can contain the following
21349 Float representation of @code{value} from metadata key.
21352 Float representation of @code{value} as supplied by user in @code{value} option.
21356 If specified in @code{print} mode, output is written to the named file. Instead of
21357 plain filename any writable url can be specified. Filename ``-'' is a shorthand
21358 for standard output. If @code{file} option is not set, output is written to the log
21359 with AV_LOG_INFO loglevel.
21363 @subsection Examples
21367 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
21370 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
21373 Print silencedetect output to file @file{metadata.txt}.
21375 silencedetect,ametadata=mode=print:file=metadata.txt
21378 Direct all metadata to a pipe with file descriptor 4.
21380 metadata=mode=print:file='pipe\:4'
21384 @section perms, aperms
21386 Set read/write permissions for the output frames.
21388 These filters are mainly aimed at developers to test direct path in the
21389 following filter in the filtergraph.
21391 The filters accept the following options:
21395 Select the permissions mode.
21397 It accepts the following values:
21400 Do nothing. This is the default.
21402 Set all the output frames read-only.
21404 Set all the output frames directly writable.
21406 Make the frame read-only if writable, and writable if read-only.
21408 Set each output frame read-only or writable randomly.
21412 Set the seed for the @var{random} mode, must be an integer included between
21413 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
21414 @code{-1}, the filter will try to use a good random seed on a best effort
21418 Note: in case of auto-inserted filter between the permission filter and the
21419 following one, the permission might not be received as expected in that
21420 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
21421 perms/aperms filter can avoid this problem.
21423 @section realtime, arealtime
21425 Slow down filtering to match real time approximately.
21427 These filters will pause the filtering for a variable amount of time to
21428 match the output rate with the input timestamps.
21429 They are similar to the @option{re} option to @code{ffmpeg}.
21431 They accept the following options:
21435 Time limit for the pauses. Any pause longer than that will be considered
21436 a timestamp discontinuity and reset the timer. Default is 2 seconds.
21438 Speed factor for processing. The value must be a float larger than zero.
21439 Values larger than 1.0 will result in faster than realtime processing,
21440 smaller will slow processing down. The @var{limit} is automatically adapted
21441 accordingly. Default is 1.0.
21443 A processing speed faster than what is possible without these filters cannot
21448 @section select, aselect
21450 Select frames to pass in output.
21452 This filter accepts the following options:
21457 Set expression, which is evaluated for each input frame.
21459 If the expression is evaluated to zero, the frame is discarded.
21461 If the evaluation result is negative or NaN, the frame is sent to the
21462 first output; otherwise it is sent to the output with index
21463 @code{ceil(val)-1}, assuming that the input index starts from 0.
21465 For example a value of @code{1.2} corresponds to the output with index
21466 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
21469 Set the number of outputs. The output to which to send the selected
21470 frame is based on the result of the evaluation. Default value is 1.
21473 The expression can contain the following constants:
21477 The (sequential) number of the filtered frame, starting from 0.
21480 The (sequential) number of the selected frame, starting from 0.
21482 @item prev_selected_n
21483 The sequential number of the last selected frame. It's NAN if undefined.
21486 The timebase of the input timestamps.
21489 The PTS (Presentation TimeStamp) of the filtered video frame,
21490 expressed in @var{TB} units. It's NAN if undefined.
21493 The PTS of the filtered video frame,
21494 expressed in seconds. It's NAN if undefined.
21497 The PTS of the previously filtered video frame. It's NAN if undefined.
21499 @item prev_selected_pts
21500 The PTS of the last previously filtered video frame. It's NAN if undefined.
21502 @item prev_selected_t
21503 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
21506 The PTS of the first video frame in the video. It's NAN if undefined.
21509 The time of the first video frame in the video. It's NAN if undefined.
21511 @item pict_type @emph{(video only)}
21512 The type of the filtered frame. It can assume one of the following
21524 @item interlace_type @emph{(video only)}
21525 The frame interlace type. It can assume one of the following values:
21528 The frame is progressive (not interlaced).
21530 The frame is top-field-first.
21532 The frame is bottom-field-first.
21535 @item consumed_sample_n @emph{(audio only)}
21536 the number of selected samples before the current frame
21538 @item samples_n @emph{(audio only)}
21539 the number of samples in the current frame
21541 @item sample_rate @emph{(audio only)}
21542 the input sample rate
21545 This is 1 if the filtered frame is a key-frame, 0 otherwise.
21548 the position in the file of the filtered frame, -1 if the information
21549 is not available (e.g. for synthetic video)
21551 @item scene @emph{(video only)}
21552 value between 0 and 1 to indicate a new scene; a low value reflects a low
21553 probability for the current frame to introduce a new scene, while a higher
21554 value means the current frame is more likely to be one (see the example below)
21556 @item concatdec_select
21557 The concat demuxer can select only part of a concat input file by setting an
21558 inpoint and an outpoint, but the output packets may not be entirely contained
21559 in the selected interval. By using this variable, it is possible to skip frames
21560 generated by the concat demuxer which are not exactly contained in the selected
21563 This works by comparing the frame pts against the @var{lavf.concat.start_time}
21564 and the @var{lavf.concat.duration} packet metadata values which are also
21565 present in the decoded frames.
21567 The @var{concatdec_select} variable is -1 if the frame pts is at least
21568 start_time and either the duration metadata is missing or the frame pts is less
21569 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
21572 That basically means that an input frame is selected if its pts is within the
21573 interval set by the concat demuxer.
21577 The default value of the select expression is "1".
21579 @subsection Examples
21583 Select all frames in input:
21588 The example above is the same as:
21600 Select only I-frames:
21602 select='eq(pict_type\,I)'
21606 Select one frame every 100:
21608 select='not(mod(n\,100))'
21612 Select only frames contained in the 10-20 time interval:
21614 select=between(t\,10\,20)
21618 Select only I-frames contained in the 10-20 time interval:
21620 select=between(t\,10\,20)*eq(pict_type\,I)
21624 Select frames with a minimum distance of 10 seconds:
21626 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
21630 Use aselect to select only audio frames with samples number > 100:
21632 aselect='gt(samples_n\,100)'
21636 Create a mosaic of the first scenes:
21638 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
21641 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
21645 Send even and odd frames to separate outputs, and compose them:
21647 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
21651 Select useful frames from an ffconcat file which is using inpoints and
21652 outpoints but where the source files are not intra frame only.
21654 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
21658 @section sendcmd, asendcmd
21660 Send commands to filters in the filtergraph.
21662 These filters read commands to be sent to other filters in the
21665 @code{sendcmd} must be inserted between two video filters,
21666 @code{asendcmd} must be inserted between two audio filters, but apart
21667 from that they act the same way.
21669 The specification of commands can be provided in the filter arguments
21670 with the @var{commands} option, or in a file specified by the
21671 @var{filename} option.
21673 These filters accept the following options:
21676 Set the commands to be read and sent to the other filters.
21678 Set the filename of the commands to be read and sent to the other
21682 @subsection Commands syntax
21684 A commands description consists of a sequence of interval
21685 specifications, comprising a list of commands to be executed when a
21686 particular event related to that interval occurs. The occurring event
21687 is typically the current frame time entering or leaving a given time
21690 An interval is specified by the following syntax:
21692 @var{START}[-@var{END}] @var{COMMANDS};
21695 The time interval is specified by the @var{START} and @var{END} times.
21696 @var{END} is optional and defaults to the maximum time.
21698 The current frame time is considered within the specified interval if
21699 it is included in the interval [@var{START}, @var{END}), that is when
21700 the time is greater or equal to @var{START} and is lesser than
21703 @var{COMMANDS} consists of a sequence of one or more command
21704 specifications, separated by ",", relating to that interval. The
21705 syntax of a command specification is given by:
21707 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
21710 @var{FLAGS} is optional and specifies the type of events relating to
21711 the time interval which enable sending the specified command, and must
21712 be a non-null sequence of identifier flags separated by "+" or "|" and
21713 enclosed between "[" and "]".
21715 The following flags are recognized:
21718 The command is sent when the current frame timestamp enters the
21719 specified interval. In other words, the command is sent when the
21720 previous frame timestamp was not in the given interval, and the
21724 The command is sent when the current frame timestamp leaves the
21725 specified interval. In other words, the command is sent when the
21726 previous frame timestamp was in the given interval, and the
21730 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
21733 @var{TARGET} specifies the target of the command, usually the name of
21734 the filter class or a specific filter instance name.
21736 @var{COMMAND} specifies the name of the command for the target filter.
21738 @var{ARG} is optional and specifies the optional list of argument for
21739 the given @var{COMMAND}.
21741 Between one interval specification and another, whitespaces, or
21742 sequences of characters starting with @code{#} until the end of line,
21743 are ignored and can be used to annotate comments.
21745 A simplified BNF description of the commands specification syntax
21748 @var{COMMAND_FLAG} ::= "enter" | "leave"
21749 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
21750 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
21751 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
21752 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
21753 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
21756 @subsection Examples
21760 Specify audio tempo change at second 4:
21762 asendcmd=c='4.0 atempo tempo 1.5',atempo
21766 Target a specific filter instance:
21768 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
21772 Specify a list of drawtext and hue commands in a file.
21774 # show text in the interval 5-10
21775 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
21776 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
21778 # desaturate the image in the interval 15-20
21779 15.0-20.0 [enter] hue s 0,
21780 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
21782 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
21784 # apply an exponential saturation fade-out effect, starting from time 25
21785 25 [enter] hue s exp(25-t)
21788 A filtergraph allowing to read and process the above command list
21789 stored in a file @file{test.cmd}, can be specified with:
21791 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
21796 @section setpts, asetpts
21798 Change the PTS (presentation timestamp) of the input frames.
21800 @code{setpts} works on video frames, @code{asetpts} on audio frames.
21802 This filter accepts the following options:
21807 The expression which is evaluated for each frame to construct its timestamp.
21811 The expression is evaluated through the eval API and can contain the following
21815 @item FRAME_RATE, FR
21816 frame rate, only defined for constant frame-rate video
21819 The presentation timestamp in input
21822 The count of the input frame for video or the number of consumed samples,
21823 not including the current frame for audio, starting from 0.
21825 @item NB_CONSUMED_SAMPLES
21826 The number of consumed samples, not including the current frame (only
21829 @item NB_SAMPLES, S
21830 The number of samples in the current frame (only audio)
21832 @item SAMPLE_RATE, SR
21833 The audio sample rate.
21836 The PTS of the first frame.
21839 the time in seconds of the first frame
21842 State whether the current frame is interlaced.
21845 the time in seconds of the current frame
21848 original position in the file of the frame, or undefined if undefined
21849 for the current frame
21852 The previous input PTS.
21855 previous input time in seconds
21858 The previous output PTS.
21861 previous output time in seconds
21864 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
21868 The wallclock (RTC) time at the start of the movie in microseconds.
21871 The timebase of the input timestamps.
21875 @subsection Examples
21879 Start counting PTS from zero
21881 setpts=PTS-STARTPTS
21885 Apply fast motion effect:
21891 Apply slow motion effect:
21897 Set fixed rate of 25 frames per second:
21903 Set fixed rate 25 fps with some jitter:
21905 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
21909 Apply an offset of 10 seconds to the input PTS:
21915 Generate timestamps from a "live source" and rebase onto the current timebase:
21917 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
21921 Generate timestamps by counting samples:
21930 Force color range for the output video frame.
21932 The @code{setrange} filter marks the color range property for the
21933 output frames. It does not change the input frame, but only sets the
21934 corresponding property, which affects how the frame is treated by
21937 The filter accepts the following options:
21942 Available values are:
21946 Keep the same color range property.
21948 @item unspecified, unknown
21949 Set the color range as unspecified.
21951 @item limited, tv, mpeg
21952 Set the color range as limited.
21954 @item full, pc, jpeg
21955 Set the color range as full.
21959 @section settb, asettb
21961 Set the timebase to use for the output frames timestamps.
21962 It is mainly useful for testing timebase configuration.
21964 It accepts the following parameters:
21969 The expression which is evaluated into the output timebase.
21973 The value for @option{tb} is an arithmetic expression representing a
21974 rational. The expression can contain the constants "AVTB" (the default
21975 timebase), "intb" (the input timebase) and "sr" (the sample rate,
21976 audio only). Default value is "intb".
21978 @subsection Examples
21982 Set the timebase to 1/25:
21988 Set the timebase to 1/10:
21994 Set the timebase to 1001/1000:
22000 Set the timebase to 2*intb:
22006 Set the default timebase value:
22013 Convert input audio to a video output representing frequency spectrum
22014 logarithmically using Brown-Puckette constant Q transform algorithm with
22015 direct frequency domain coefficient calculation (but the transform itself
22016 is not really constant Q, instead the Q factor is actually variable/clamped),
22017 with musical tone scale, from E0 to D#10.
22019 The filter accepts the following options:
22023 Specify the video size for the output. It must be even. For the syntax of this option,
22024 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22025 Default value is @code{1920x1080}.
22028 Set the output frame rate. Default value is @code{25}.
22031 Set the bargraph height. It must be even. Default value is @code{-1} which
22032 computes the bargraph height automatically.
22035 Set the axis height. It must be even. Default value is @code{-1} which computes
22036 the axis height automatically.
22039 Set the sonogram height. It must be even. Default value is @code{-1} which
22040 computes the sonogram height automatically.
22043 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
22044 instead. Default value is @code{1}.
22046 @item sono_v, volume
22047 Specify the sonogram volume expression. It can contain variables:
22050 the @var{bar_v} evaluated expression
22051 @item frequency, freq, f
22052 the frequency where it is evaluated
22053 @item timeclamp, tc
22054 the value of @var{timeclamp} option
22058 @item a_weighting(f)
22059 A-weighting of equal loudness
22060 @item b_weighting(f)
22061 B-weighting of equal loudness
22062 @item c_weighting(f)
22063 C-weighting of equal loudness.
22065 Default value is @code{16}.
22067 @item bar_v, volume2
22068 Specify the bargraph volume expression. It can contain variables:
22071 the @var{sono_v} evaluated expression
22072 @item frequency, freq, f
22073 the frequency where it is evaluated
22074 @item timeclamp, tc
22075 the value of @var{timeclamp} option
22079 @item a_weighting(f)
22080 A-weighting of equal loudness
22081 @item b_weighting(f)
22082 B-weighting of equal loudness
22083 @item c_weighting(f)
22084 C-weighting of equal loudness.
22086 Default value is @code{sono_v}.
22088 @item sono_g, gamma
22089 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
22090 higher gamma makes the spectrum having more range. Default value is @code{3}.
22091 Acceptable range is @code{[1, 7]}.
22093 @item bar_g, gamma2
22094 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
22098 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
22099 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
22101 @item timeclamp, tc
22102 Specify the transform timeclamp. At low frequency, there is trade-off between
22103 accuracy in time domain and frequency domain. If timeclamp is lower,
22104 event in time domain is represented more accurately (such as fast bass drum),
22105 otherwise event in frequency domain is represented more accurately
22106 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
22109 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
22110 limits future samples by applying asymmetric windowing in time domain, useful
22111 when low latency is required. Accepted range is @code{[0, 1]}.
22114 Specify the transform base frequency. Default value is @code{20.01523126408007475},
22115 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
22118 Specify the transform end frequency. Default value is @code{20495.59681441799654},
22119 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
22122 This option is deprecated and ignored.
22125 Specify the transform length in time domain. Use this option to control accuracy
22126 trade-off between time domain and frequency domain at every frequency sample.
22127 It can contain variables:
22129 @item frequency, freq, f
22130 the frequency where it is evaluated
22131 @item timeclamp, tc
22132 the value of @var{timeclamp} option.
22134 Default value is @code{384*tc/(384+tc*f)}.
22137 Specify the transform count for every video frame. Default value is @code{6}.
22138 Acceptable range is @code{[1, 30]}.
22141 Specify the transform count for every single pixel. Default value is @code{0},
22142 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
22145 Specify font file for use with freetype to draw the axis. If not specified,
22146 use embedded font. Note that drawing with font file or embedded font is not
22147 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
22151 Specify fontconfig pattern. This has lower priority than @var{fontfile}.
22152 The : in the pattern may be replaced by | to avoid unnecessary escaping.
22155 Specify font color expression. This is arithmetic expression that should return
22156 integer value 0xRRGGBB. It can contain variables:
22158 @item frequency, freq, f
22159 the frequency where it is evaluated
22160 @item timeclamp, tc
22161 the value of @var{timeclamp} option
22166 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
22167 @item r(x), g(x), b(x)
22168 red, green, and blue value of intensity x.
22170 Default value is @code{st(0, (midi(f)-59.5)/12);
22171 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
22172 r(1-ld(1)) + b(ld(1))}.
22175 Specify image file to draw the axis. This option override @var{fontfile} and
22176 @var{fontcolor} option.
22179 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
22180 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
22181 Default value is @code{1}.
22184 Set colorspace. The accepted values are:
22187 Unspecified (default)
22196 BT.470BG or BT.601-6 625
22199 SMPTE-170M or BT.601-6 525
22205 BT.2020 with non-constant luminance
22210 Set spectrogram color scheme. This is list of floating point values with format
22211 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
22212 The default is @code{1|0.5|0|0|0.5|1}.
22216 @subsection Examples
22220 Playing audio while showing the spectrum:
22222 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
22226 Same as above, but with frame rate 30 fps:
22228 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
22232 Playing at 1280x720:
22234 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
22238 Disable sonogram display:
22244 A1 and its harmonics: A1, A2, (near)E3, A3:
22246 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),
22247 asplit[a][out1]; [a] showcqt [out0]'
22251 Same as above, but with more accuracy in frequency domain:
22253 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),
22254 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
22260 bar_v=10:sono_v=bar_v*a_weighting(f)
22264 Custom gamma, now spectrum is linear to the amplitude.
22270 Custom tlength equation:
22272 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)))'
22276 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
22278 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
22282 Custom font using fontconfig:
22284 font='Courier New,Monospace,mono|bold'
22288 Custom frequency range with custom axis using image file:
22290 axisfile=myaxis.png:basefreq=40:endfreq=10000
22296 Convert input audio to video output representing the audio power spectrum.
22297 Audio amplitude is on Y-axis while frequency is on X-axis.
22299 The filter accepts the following options:
22303 Specify size of video. For the syntax of this option, check the
22304 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22305 Default is @code{1024x512}.
22309 This set how each frequency bin will be represented.
22311 It accepts the following values:
22317 Default is @code{bar}.
22320 Set amplitude scale.
22322 It accepts the following values:
22336 Default is @code{log}.
22339 Set frequency scale.
22341 It accepts the following values:
22350 Reverse logarithmic scale.
22352 Default is @code{lin}.
22355 Set window size. Allowed range is from 16 to 65536.
22357 Default is @code{2048}
22360 Set windowing function.
22362 It accepts the following values:
22385 Default is @code{hanning}.
22388 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
22389 which means optimal overlap for selected window function will be picked.
22392 Set time averaging. Setting this to 0 will display current maximal peaks.
22393 Default is @code{1}, which means time averaging is disabled.
22396 Specify list of colors separated by space or by '|' which will be used to
22397 draw channel frequencies. Unrecognized or missing colors will be replaced
22401 Set channel display mode.
22403 It accepts the following values:
22408 Default is @code{combined}.
22411 Set minimum amplitude used in @code{log} amplitude scaler.
22415 @section showspatial
22417 Convert stereo input audio to a video output, representing the spatial relationship
22418 between two channels.
22420 The filter accepts the following options:
22424 Specify the video size for the output. For the syntax of this option, check the
22425 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22426 Default value is @code{512x512}.
22429 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
22432 Set window function.
22434 It accepts the following values:
22459 Default value is @code{hann}.
22462 Set ratio of overlap window. Default value is @code{0.5}.
22463 When value is @code{1} overlap is set to recommended size for specific
22464 window function currently used.
22467 @anchor{showspectrum}
22468 @section showspectrum
22470 Convert input audio to a video output, representing the audio frequency
22473 The filter accepts the following options:
22477 Specify the video size for the output. For the syntax of this option, check the
22478 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22479 Default value is @code{640x512}.
22482 Specify how the spectrum should slide along the window.
22484 It accepts the following values:
22487 the samples start again on the left when they reach the right
22489 the samples scroll from right to left
22491 frames are only produced when the samples reach the right
22493 the samples scroll from left to right
22496 Default value is @code{replace}.
22499 Specify display mode.
22501 It accepts the following values:
22504 all channels are displayed in the same row
22506 all channels are displayed in separate rows
22509 Default value is @samp{combined}.
22512 Specify display color mode.
22514 It accepts the following values:
22517 each channel is displayed in a separate color
22519 each channel is displayed using the same color scheme
22521 each channel is displayed using the rainbow color scheme
22523 each channel is displayed using the moreland color scheme
22525 each channel is displayed using the nebulae color scheme
22527 each channel is displayed using the fire color scheme
22529 each channel is displayed using the fiery color scheme
22531 each channel is displayed using the fruit color scheme
22533 each channel is displayed using the cool color scheme
22535 each channel is displayed using the magma color scheme
22537 each channel is displayed using the green color scheme
22539 each channel is displayed using the viridis color scheme
22541 each channel is displayed using the plasma color scheme
22543 each channel is displayed using the cividis color scheme
22545 each channel is displayed using the terrain color scheme
22548 Default value is @samp{channel}.
22551 Specify scale used for calculating intensity color values.
22553 It accepts the following values:
22558 square root, default
22569 Default value is @samp{sqrt}.
22572 Specify frequency scale.
22574 It accepts the following values:
22582 Default value is @samp{lin}.
22585 Set saturation modifier for displayed colors. Negative values provide
22586 alternative color scheme. @code{0} is no saturation at all.
22587 Saturation must be in [-10.0, 10.0] range.
22588 Default value is @code{1}.
22591 Set window function.
22593 It accepts the following values:
22618 Default value is @code{hann}.
22621 Set orientation of time vs frequency axis. Can be @code{vertical} or
22622 @code{horizontal}. Default is @code{vertical}.
22625 Set ratio of overlap window. Default value is @code{0}.
22626 When value is @code{1} overlap is set to recommended size for specific
22627 window function currently used.
22630 Set scale gain for calculating intensity color values.
22631 Default value is @code{1}.
22634 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
22637 Set color rotation, must be in [-1.0, 1.0] range.
22638 Default value is @code{0}.
22641 Set start frequency from which to display spectrogram. Default is @code{0}.
22644 Set stop frequency to which to display spectrogram. Default is @code{0}.
22647 Set upper frame rate limit. Default is @code{auto}, unlimited.
22650 Draw time and frequency axes and legends. Default is disabled.
22653 The usage is very similar to the showwaves filter; see the examples in that
22656 @subsection Examples
22660 Large window with logarithmic color scaling:
22662 showspectrum=s=1280x480:scale=log
22666 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
22668 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
22669 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
22673 @section showspectrumpic
22675 Convert input audio to a single video frame, representing the audio frequency
22678 The filter accepts the following options:
22682 Specify the video size for the output. For the syntax of this option, check the
22683 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22684 Default value is @code{4096x2048}.
22687 Specify display mode.
22689 It accepts the following values:
22692 all channels are displayed in the same row
22694 all channels are displayed in separate rows
22696 Default value is @samp{combined}.
22699 Specify display color mode.
22701 It accepts the following values:
22704 each channel is displayed in a separate color
22706 each channel is displayed using the same color scheme
22708 each channel is displayed using the rainbow color scheme
22710 each channel is displayed using the moreland color scheme
22712 each channel is displayed using the nebulae color scheme
22714 each channel is displayed using the fire color scheme
22716 each channel is displayed using the fiery color scheme
22718 each channel is displayed using the fruit color scheme
22720 each channel is displayed using the cool color scheme
22722 each channel is displayed using the magma color scheme
22724 each channel is displayed using the green color scheme
22726 each channel is displayed using the viridis color scheme
22728 each channel is displayed using the plasma color scheme
22730 each channel is displayed using the cividis color scheme
22732 each channel is displayed using the terrain color scheme
22734 Default value is @samp{intensity}.
22737 Specify scale used for calculating intensity color values.
22739 It accepts the following values:
22744 square root, default
22754 Default value is @samp{log}.
22757 Specify frequency scale.
22759 It accepts the following values:
22767 Default value is @samp{lin}.
22770 Set saturation modifier for displayed colors. Negative values provide
22771 alternative color scheme. @code{0} is no saturation at all.
22772 Saturation must be in [-10.0, 10.0] range.
22773 Default value is @code{1}.
22776 Set window function.
22778 It accepts the following values:
22802 Default value is @code{hann}.
22805 Set orientation of time vs frequency axis. Can be @code{vertical} or
22806 @code{horizontal}. Default is @code{vertical}.
22809 Set scale gain for calculating intensity color values.
22810 Default value is @code{1}.
22813 Draw time and frequency axes and legends. Default is enabled.
22816 Set color rotation, must be in [-1.0, 1.0] range.
22817 Default value is @code{0}.
22820 Set start frequency from which to display spectrogram. Default is @code{0}.
22823 Set stop frequency to which to display spectrogram. Default is @code{0}.
22826 @subsection Examples
22830 Extract an audio spectrogram of a whole audio track
22831 in a 1024x1024 picture using @command{ffmpeg}:
22833 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
22837 @section showvolume
22839 Convert input audio volume to a video output.
22841 The filter accepts the following options:
22848 Set border width, allowed range is [0, 5]. Default is 1.
22851 Set channel width, allowed range is [80, 8192]. Default is 400.
22854 Set channel height, allowed range is [1, 900]. Default is 20.
22857 Set fade, allowed range is [0, 1]. Default is 0.95.
22860 Set volume color expression.
22862 The expression can use the following variables:
22866 Current max volume of channel in dB.
22872 Current channel number, starting from 0.
22876 If set, displays channel names. Default is enabled.
22879 If set, displays volume values. Default is enabled.
22882 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
22883 default is @code{h}.
22886 Set step size, allowed range is [0, 5]. Default is 0, which means
22890 Set background opacity, allowed range is [0, 1]. Default is 0.
22893 Set metering mode, can be peak: @code{p} or rms: @code{r},
22894 default is @code{p}.
22897 Set display scale, can be linear: @code{lin} or log: @code{log},
22898 default is @code{lin}.
22902 If set to > 0., display a line for the max level
22903 in the previous seconds.
22904 default is disabled: @code{0.}
22907 The color of the max line. Use when @code{dm} option is set to > 0.
22908 default is: @code{orange}
22913 Convert input audio to a video output, representing the samples waves.
22915 The filter accepts the following options:
22919 Specify the video size for the output. For the syntax of this option, check the
22920 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22921 Default value is @code{600x240}.
22926 Available values are:
22929 Draw a point for each sample.
22932 Draw a vertical line for each sample.
22935 Draw a point for each sample and a line between them.
22938 Draw a centered vertical line for each sample.
22941 Default value is @code{point}.
22944 Set the number of samples which are printed on the same column. A
22945 larger value will decrease the frame rate. Must be a positive
22946 integer. This option can be set only if the value for @var{rate}
22947 is not explicitly specified.
22950 Set the (approximate) output frame rate. This is done by setting the
22951 option @var{n}. Default value is "25".
22953 @item split_channels
22954 Set if channels should be drawn separately or overlap. Default value is 0.
22957 Set colors separated by '|' which are going to be used for drawing of each channel.
22960 Set amplitude scale.
22962 Available values are:
22980 Set the draw mode. This is mostly useful to set for high @var{n}.
22982 Available values are:
22985 Scale pixel values for each drawn sample.
22988 Draw every sample directly.
22991 Default value is @code{scale}.
22994 @subsection Examples
22998 Output the input file audio and the corresponding video representation
23001 amovie=a.mp3,asplit[out0],showwaves[out1]
23005 Create a synthetic signal and show it with showwaves, forcing a
23006 frame rate of 30 frames per second:
23008 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
23012 @section showwavespic
23014 Convert input audio to a single video frame, representing the samples waves.
23016 The filter accepts the following options:
23020 Specify the video size for the output. For the syntax of this option, check the
23021 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23022 Default value is @code{600x240}.
23024 @item split_channels
23025 Set if channels should be drawn separately or overlap. Default value is 0.
23028 Set colors separated by '|' which are going to be used for drawing of each channel.
23031 Set amplitude scale.
23033 Available values are:
23053 Available values are:
23056 Scale pixel values for each drawn sample.
23059 Draw every sample directly.
23062 Default value is @code{scale}.
23065 @subsection Examples
23069 Extract a channel split representation of the wave form of a whole audio track
23070 in a 1024x800 picture using @command{ffmpeg}:
23072 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
23076 @section sidedata, asidedata
23078 Delete frame side data, or select frames based on it.
23080 This filter accepts the following options:
23084 Set mode of operation of the filter.
23086 Can be one of the following:
23090 Select every frame with side data of @code{type}.
23093 Delete side data of @code{type}. If @code{type} is not set, delete all side
23099 Set side data type used with all modes. Must be set for @code{select} mode. For
23100 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
23101 in @file{libavutil/frame.h}. For example, to choose
23102 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
23106 @section spectrumsynth
23108 Sythesize audio from 2 input video spectrums, first input stream represents
23109 magnitude across time and second represents phase across time.
23110 The filter will transform from frequency domain as displayed in videos back
23111 to time domain as presented in audio output.
23113 This filter is primarily created for reversing processed @ref{showspectrum}
23114 filter outputs, but can synthesize sound from other spectrograms too.
23115 But in such case results are going to be poor if the phase data is not
23116 available, because in such cases phase data need to be recreated, usually
23117 it's just recreated from random noise.
23118 For best results use gray only output (@code{channel} color mode in
23119 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
23120 @code{lin} scale for phase video. To produce phase, for 2nd video, use
23121 @code{data} option. Inputs videos should generally use @code{fullframe}
23122 slide mode as that saves resources needed for decoding video.
23124 The filter accepts the following options:
23128 Specify sample rate of output audio, the sample rate of audio from which
23129 spectrum was generated may differ.
23132 Set number of channels represented in input video spectrums.
23135 Set scale which was used when generating magnitude input spectrum.
23136 Can be @code{lin} or @code{log}. Default is @code{log}.
23139 Set slide which was used when generating inputs spectrums.
23140 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
23141 Default is @code{fullframe}.
23144 Set window function used for resynthesis.
23147 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
23148 which means optimal overlap for selected window function will be picked.
23151 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
23152 Default is @code{vertical}.
23155 @subsection Examples
23159 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
23160 then resynthesize videos back to audio with spectrumsynth:
23162 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
23163 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
23164 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
23168 @section split, asplit
23170 Split input into several identical outputs.
23172 @code{asplit} works with audio input, @code{split} with video.
23174 The filter accepts a single parameter which specifies the number of outputs. If
23175 unspecified, it defaults to 2.
23177 @subsection Examples
23181 Create two separate outputs from the same input:
23183 [in] split [out0][out1]
23187 To create 3 or more outputs, you need to specify the number of
23190 [in] asplit=3 [out0][out1][out2]
23194 Create two separate outputs from the same input, one cropped and
23197 [in] split [splitout1][splitout2];
23198 [splitout1] crop=100:100:0:0 [cropout];
23199 [splitout2] pad=200:200:100:100 [padout];
23203 Create 5 copies of the input audio with @command{ffmpeg}:
23205 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
23211 Receive commands sent through a libzmq client, and forward them to
23212 filters in the filtergraph.
23214 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
23215 must be inserted between two video filters, @code{azmq} between two
23216 audio filters. Both are capable to send messages to any filter type.
23218 To enable these filters you need to install the libzmq library and
23219 headers and configure FFmpeg with @code{--enable-libzmq}.
23221 For more information about libzmq see:
23222 @url{http://www.zeromq.org/}
23224 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
23225 receives messages sent through a network interface defined by the
23226 @option{bind_address} (or the abbreviation "@option{b}") option.
23227 Default value of this option is @file{tcp://localhost:5555}. You may
23228 want to alter this value to your needs, but do not forget to escape any
23229 ':' signs (see @ref{filtergraph escaping}).
23231 The received message must be in the form:
23233 @var{TARGET} @var{COMMAND} [@var{ARG}]
23236 @var{TARGET} specifies the target of the command, usually the name of
23237 the filter class or a specific filter instance name. The default
23238 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
23239 but you can override this by using the @samp{filter_name@@id} syntax
23240 (see @ref{Filtergraph syntax}).
23242 @var{COMMAND} specifies the name of the command for the target filter.
23244 @var{ARG} is optional and specifies the optional argument list for the
23245 given @var{COMMAND}.
23247 Upon reception, the message is processed and the corresponding command
23248 is injected into the filtergraph. Depending on the result, the filter
23249 will send a reply to the client, adopting the format:
23251 @var{ERROR_CODE} @var{ERROR_REASON}
23255 @var{MESSAGE} is optional.
23257 @subsection Examples
23259 Look at @file{tools/zmqsend} for an example of a zmq client which can
23260 be used to send commands processed by these filters.
23262 Consider the following filtergraph generated by @command{ffplay}.
23263 In this example the last overlay filter has an instance name. All other
23264 filters will have default instance names.
23267 ffplay -dumpgraph 1 -f lavfi "
23268 color=s=100x100:c=red [l];
23269 color=s=100x100:c=blue [r];
23270 nullsrc=s=200x100, zmq [bg];
23271 [bg][l] overlay [bg+l];
23272 [bg+l][r] overlay@@my=x=100 "
23275 To change the color of the left side of the video, the following
23276 command can be used:
23278 echo Parsed_color_0 c yellow | tools/zmqsend
23281 To change the right side:
23283 echo Parsed_color_1 c pink | tools/zmqsend
23286 To change the position of the right side:
23288 echo overlay@@my x 150 | tools/zmqsend
23292 @c man end MULTIMEDIA FILTERS
23294 @chapter Multimedia Sources
23295 @c man begin MULTIMEDIA SOURCES
23297 Below is a description of the currently available multimedia sources.
23301 This is the same as @ref{movie} source, except it selects an audio
23307 Read audio and/or video stream(s) from a movie container.
23309 It accepts the following parameters:
23313 The name of the resource to read (not necessarily a file; it can also be a
23314 device or a stream accessed through some protocol).
23316 @item format_name, f
23317 Specifies the format assumed for the movie to read, and can be either
23318 the name of a container or an input device. If not specified, the
23319 format is guessed from @var{movie_name} or by probing.
23321 @item seek_point, sp
23322 Specifies the seek point in seconds. The frames will be output
23323 starting from this seek point. The parameter is evaluated with
23324 @code{av_strtod}, so the numerical value may be suffixed by an IS
23325 postfix. The default value is "0".
23328 Specifies the streams to read. Several streams can be specified,
23329 separated by "+". The source will then have as many outputs, in the
23330 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
23331 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
23332 respectively the default (best suited) video and audio stream. Default
23333 is "dv", or "da" if the filter is called as "amovie".
23335 @item stream_index, si
23336 Specifies the index of the video stream to read. If the value is -1,
23337 the most suitable video stream will be automatically selected. The default
23338 value is "-1". Deprecated. If the filter is called "amovie", it will select
23339 audio instead of video.
23342 Specifies how many times to read the stream in sequence.
23343 If the value is 0, the stream will be looped infinitely.
23344 Default value is "1".
23346 Note that when the movie is looped the source timestamps are not
23347 changed, so it will generate non monotonically increasing timestamps.
23349 @item discontinuity
23350 Specifies the time difference between frames above which the point is
23351 considered a timestamp discontinuity which is removed by adjusting the later
23355 It allows overlaying a second video on top of the main input of
23356 a filtergraph, as shown in this graph:
23358 input -----------> deltapts0 --> overlay --> output
23361 movie --> scale--> deltapts1 -------+
23363 @subsection Examples
23367 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
23368 on top of the input labelled "in":
23370 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
23371 [in] setpts=PTS-STARTPTS [main];
23372 [main][over] overlay=16:16 [out]
23376 Read from a video4linux2 device, and overlay it on top of the input
23379 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
23380 [in] setpts=PTS-STARTPTS [main];
23381 [main][over] overlay=16:16 [out]
23385 Read the first video stream and the audio stream with id 0x81 from
23386 dvd.vob; the video is connected to the pad named "video" and the audio is
23387 connected to the pad named "audio":
23389 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
23393 @subsection Commands
23395 Both movie and amovie support the following commands:
23398 Perform seek using "av_seek_frame".
23399 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
23402 @var{stream_index}: If stream_index is -1, a default
23403 stream is selected, and @var{timestamp} is automatically converted
23404 from AV_TIME_BASE units to the stream specific time_base.
23406 @var{timestamp}: Timestamp in AVStream.time_base units
23407 or, if no stream is specified, in AV_TIME_BASE units.
23409 @var{flags}: Flags which select direction and seeking mode.
23413 Get movie duration in AV_TIME_BASE units.
23417 @c man end MULTIMEDIA SOURCES