1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Options for filters with several inputs (framesync)
317 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
319 Some filters with several inputs support a common set of options.
320 These options can only be set by name, not with the short notation.
324 The action to take when EOF is encountered on the secondary input; it accepts
325 one of the following values:
329 Repeat the last frame (the default).
333 Pass the main input through.
337 If set to 1, force the output to terminate when the shortest input
338 terminates. Default value is 0.
341 If set to 1, force the filter to extend the last frame of secondary streams
342 until the end of the primary stream. A value of 0 disables this behavior.
346 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
348 @chapter Audio Filters
349 @c man begin AUDIO FILTERS
351 When you configure your FFmpeg build, you can disable any of the
352 existing filters using @code{--disable-filters}.
353 The configure output will show the audio filters included in your
356 Below is a description of the currently available audio filters.
360 A compressor is mainly used to reduce the dynamic range of a signal.
361 Especially modern music is mostly compressed at a high ratio to
362 improve the overall loudness. It's done to get the highest attention
363 of a listener, "fatten" the sound and bring more "power" to the track.
364 If a signal is compressed too much it may sound dull or "dead"
365 afterwards or it may start to "pump" (which could be a powerful effect
366 but can also destroy a track completely).
367 The right compression is the key to reach a professional sound and is
368 the high art of mixing and mastering. Because of its complex settings
369 it may take a long time to get the right feeling for this kind of effect.
371 Compression is done by detecting the volume above a chosen level
372 @code{threshold} and dividing it by the factor set with @code{ratio}.
373 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
374 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
375 the signal would cause distortion of the waveform the reduction can be
376 levelled over the time. This is done by setting "Attack" and "Release".
377 @code{attack} determines how long the signal has to rise above the threshold
378 before any reduction will occur and @code{release} sets the time the signal
379 has to fall below the threshold to reduce the reduction again. Shorter signals
380 than the chosen attack time will be left untouched.
381 The overall reduction of the signal can be made up afterwards with the
382 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
383 raising the makeup to this level results in a signal twice as loud than the
384 source. To gain a softer entry in the compression the @code{knee} flattens the
385 hard edge at the threshold in the range of the chosen decibels.
387 The filter accepts the following options:
391 Set input gain. Default is 1. Range is between 0.015625 and 64.
394 If a signal of stream rises above this level it will affect the gain
396 By default it is 0.125. Range is between 0.00097563 and 1.
399 Set a ratio by which the signal is reduced. 1:2 means that if the level
400 rose 4dB above the threshold, it will be only 2dB above after the reduction.
401 Default is 2. Range is between 1 and 20.
404 Amount of milliseconds the signal has to rise above the threshold before gain
405 reduction starts. Default is 20. Range is between 0.01 and 2000.
408 Amount of milliseconds the signal has to fall below the threshold before
409 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
412 Set the amount by how much signal will be amplified after processing.
413 Default is 1. Range is from 1 to 64.
416 Curve the sharp knee around the threshold to enter gain reduction more softly.
417 Default is 2.82843. Range is between 1 and 8.
420 Choose if the @code{average} level between all channels of input stream
421 or the louder(@code{maximum}) channel of input stream affects the
422 reduction. Default is @code{average}.
425 Should the exact signal be taken in case of @code{peak} or an RMS one in case
426 of @code{rms}. Default is @code{rms} which is mostly smoother.
429 How much to use compressed signal in output. Default is 1.
430 Range is between 0 and 1.
434 Simple audio dynamic range compression/expansion filter.
436 The filter accepts the following options:
440 Set contrast. Default is 33. Allowed range is between 0 and 100.
445 Copy the input audio source unchanged to the output. This is mainly useful for
450 Apply cross fade from one input audio stream to another input audio stream.
451 The cross fade is applied for specified duration near the end of first stream.
453 The filter accepts the following options:
457 Specify the number of samples for which the cross fade effect has to last.
458 At the end of the cross fade effect the first input audio will be completely
459 silent. Default is 44100.
462 Specify the duration of the cross fade effect. See
463 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
464 for the accepted syntax.
465 By default the duration is determined by @var{nb_samples}.
466 If set this option is used instead of @var{nb_samples}.
469 Should first stream end overlap with second stream start. Default is enabled.
472 Set curve for cross fade transition for first stream.
475 Set curve for cross fade transition for second stream.
477 For description of available curve types see @ref{afade} filter description.
484 Cross fade from one input to another:
486 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
490 Cross fade from one input to another but without overlapping:
492 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
497 Split audio stream into several bands.
499 This filter splits audio stream into two or more frequency ranges.
500 Summing all streams back will give flat output.
502 The filter accepts the following options:
506 Set split frequencies. Those must be positive and increasing.
509 Set filter order, can be @var{2nd}, @var{4th} or @var{8th}.
510 Default is @var{4th}.
515 Reduce audio bit resolution.
517 This filter is bit crusher with enhanced functionality. A bit crusher
518 is used to audibly reduce number of bits an audio signal is sampled
519 with. This doesn't change the bit depth at all, it just produces the
520 effect. Material reduced in bit depth sounds more harsh and "digital".
521 This filter is able to even round to continuous values instead of discrete
523 Additionally it has a D/C offset which results in different crushing of
524 the lower and the upper half of the signal.
525 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
527 Another feature of this filter is the logarithmic mode.
528 This setting switches from linear distances between bits to logarithmic ones.
529 The result is a much more "natural" sounding crusher which doesn't gate low
530 signals for example. The human ear has a logarithmic perception,
531 so this kind of crushing is much more pleasant.
532 Logarithmic crushing is also able to get anti-aliased.
534 The filter accepts the following options:
550 Can be linear: @code{lin} or logarithmic: @code{log}.
559 Set sample reduction.
562 Enable LFO. By default disabled.
573 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
577 Remove impulsive noise from input audio.
579 Samples detected as impulsive noise are replaced by interpolated samples using
580 autoregressive modelling.
584 Set window size, in milliseconds. Allowed range is from @code{10} to
585 @code{100}. Default value is @code{55} milliseconds.
586 This sets size of window which will be processed at once.
589 Set window overlap, in percentage of window size. Allowed range is from
590 @code{50} to @code{95}. Default value is @code{75} percent.
591 Setting this to a very high value increases impulsive noise removal but makes
592 whole process much slower.
595 Set autoregression order, in percentage of window size. Allowed range is from
596 @code{0} to @code{25}. Default value is @code{2} percent. This option also
597 controls quality of interpolated samples using neighbour good samples.
600 Set threshold value. Allowed range is from @code{1} to @code{100}.
601 Default value is @code{2}.
602 This controls the strength of impulsive noise which is going to be removed.
603 The lower value, the more samples will be detected as impulsive noise.
606 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
607 @code{10}. Default value is @code{2}.
608 If any two samples detected as noise are spaced less than this value then any
609 sample between those two samples will be also detected as noise.
614 It accepts the following values:
617 Select overlap-add method. Even not interpolated samples are slightly
618 changed with this method.
621 Select overlap-save method. Not interpolated samples remain unchanged.
624 Default value is @code{a}.
628 Remove clipped samples from input audio.
630 Samples detected as clipped are replaced by interpolated samples using
631 autoregressive modelling.
635 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
636 Default value is @code{55} milliseconds.
637 This sets size of window which will be processed at once.
640 Set window overlap, in percentage of window size. Allowed range is from @code{50}
641 to @code{95}. Default value is @code{75} percent.
644 Set autoregression order, in percentage of window size. Allowed range is from
645 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
646 quality of interpolated samples using neighbour good samples.
649 Set threshold value. Allowed range is from @code{1} to @code{100}.
650 Default value is @code{10}. Higher values make clip detection less aggressive.
653 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
654 Default value is @code{1000}. Higher values make clip detection less aggressive.
659 It accepts the following values:
662 Select overlap-add method. Even not interpolated samples are slightly changed
666 Select overlap-save method. Not interpolated samples remain unchanged.
669 Default value is @code{a}.
674 Delay one or more audio channels.
676 Samples in delayed channel are filled with silence.
678 The filter accepts the following option:
682 Set list of delays in milliseconds for each channel separated by '|'.
683 Unused delays will be silently ignored. If number of given delays is
684 smaller than number of channels all remaining channels will not be delayed.
685 If you want to delay exact number of samples, append 'S' to number.
686 If you want instead to delay in seconds, append 's' to number.
693 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
694 the second channel (and any other channels that may be present) unchanged.
700 Delay second channel by 500 samples, the third channel by 700 samples and leave
701 the first channel (and any other channels that may be present) unchanged.
707 @section aderivative, aintegral
709 Compute derivative/integral of audio stream.
711 Applying both filters one after another produces original audio.
715 Apply echoing to the input audio.
717 Echoes are reflected sound and can occur naturally amongst mountains
718 (and sometimes large buildings) when talking or shouting; digital echo
719 effects emulate this behaviour and are often used to help fill out the
720 sound of a single instrument or vocal. The time difference between the
721 original signal and the reflection is the @code{delay}, and the
722 loudness of the reflected signal is the @code{decay}.
723 Multiple echoes can have different delays and decays.
725 A description of the accepted parameters follows.
729 Set input gain of reflected signal. Default is @code{0.6}.
732 Set output gain of reflected signal. Default is @code{0.3}.
735 Set list of time intervals in milliseconds between original signal and reflections
736 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
737 Default is @code{1000}.
740 Set list of loudness of reflected signals separated by '|'.
741 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
742 Default is @code{0.5}.
749 Make it sound as if there are twice as many instruments as are actually playing:
751 aecho=0.8:0.88:60:0.4
755 If delay is very short, then it sound like a (metallic) robot playing music:
761 A longer delay will sound like an open air concert in the mountains:
763 aecho=0.8:0.9:1000:0.3
767 Same as above but with one more mountain:
769 aecho=0.8:0.9:1000|1800:0.3|0.25
774 Audio emphasis filter creates or restores material directly taken from LPs or
775 emphased CDs with different filter curves. E.g. to store music on vinyl the
776 signal has to be altered by a filter first to even out the disadvantages of
777 this recording medium.
778 Once the material is played back the inverse filter has to be applied to
779 restore the distortion of the frequency response.
781 The filter accepts the following options:
791 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
792 use @code{production} mode. Default is @code{reproduction} mode.
795 Set filter type. Selects medium. Can be one of the following:
807 select Compact Disc (CD).
813 select 50µs (FM-KF).
815 select 75µs (FM-KF).
821 Modify an audio signal according to the specified expressions.
823 This filter accepts one or more expressions (one for each channel),
824 which are evaluated and used to modify a corresponding audio signal.
826 It accepts the following parameters:
830 Set the '|'-separated expressions list for each separate channel. If
831 the number of input channels is greater than the number of
832 expressions, the last specified expression is used for the remaining
835 @item channel_layout, c
836 Set output channel layout. If not specified, the channel layout is
837 specified by the number of expressions. If set to @samp{same}, it will
838 use by default the same input channel layout.
841 Each expression in @var{exprs} can contain the following constants and functions:
845 channel number of the current expression
848 number of the evaluated sample, starting from 0
854 time of the evaluated sample expressed in seconds
857 @item nb_out_channels
858 input and output number of channels
861 the value of input channel with number @var{CH}
864 Note: this filter is slow. For faster processing you should use a
873 aeval=val(ch)/2:c=same
877 Invert phase of the second channel:
886 Apply fade-in/out effect to input audio.
888 A description of the accepted parameters follows.
892 Specify the effect type, can be either @code{in} for fade-in, or
893 @code{out} for a fade-out effect. Default is @code{in}.
895 @item start_sample, ss
896 Specify the number of the start sample for starting to apply the fade
897 effect. Default is 0.
900 Specify the number of samples for which the fade effect has to last. At
901 the end of the fade-in effect the output audio will have the same
902 volume as the input audio, at the end of the fade-out transition
903 the output audio will be silence. Default is 44100.
906 Specify the start time of the fade effect. Default is 0.
907 The value must be specified as a time duration; see
908 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
909 for the accepted syntax.
910 If set this option is used instead of @var{start_sample}.
913 Specify the duration of the fade effect. See
914 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
915 for the accepted syntax.
916 At the end of the fade-in effect the output audio will have the same
917 volume as the input audio, at the end of the fade-out transition
918 the output audio will be silence.
919 By default the duration is determined by @var{nb_samples}.
920 If set this option is used instead of @var{nb_samples}.
923 Set curve for fade transition.
925 It accepts the following values:
928 select triangular, linear slope (default)
930 select quarter of sine wave
932 select half of sine wave
934 select exponential sine wave
938 select inverted parabola
952 select inverted quarter of sine wave
954 select inverted half of sine wave
956 select double-exponential seat
958 select double-exponential sigmoid
960 select logistic sigmoid
970 Fade in first 15 seconds of audio:
976 Fade out last 25 seconds of a 900 seconds audio:
978 afade=t=out:st=875:d=25
983 Denoise audio samples with FFT.
985 A description of the accepted parameters follows.
989 Set the noise reduction in dB, allowed range is 0.01 to 97.
990 Default value is 12 dB.
993 Set the noise floor in dB, allowed range is -80 to -20.
994 Default value is -50 dB.
999 It accepts the following values:
1008 Select shellac noise.
1011 Select custom noise, defined in @code{bn} option.
1013 Default value is white noise.
1017 Set custom band noise for every one of 15 bands.
1018 Bands are separated by ' ' or '|'.
1021 Set the residual floor in dB, allowed range is -80 to -20.
1022 Default value is -38 dB.
1025 Enable noise tracking. By default is disabled.
1026 With this enabled, noise floor is automatically adjusted.
1029 Enable residual tracking. By default is disabled.
1032 Set the output mode.
1034 It accepts the following values:
1037 Pass input unchanged.
1040 Pass noise filtered out.
1045 Default value is @var{o}.
1049 @subsection Commands
1051 This filter supports the following commands:
1053 @item sample_noise, sn
1054 Start or stop measuring noise profile.
1055 Syntax for the command is : "start" or "stop" string.
1056 After measuring noise profile is stopped it will be
1057 automatically applied in filtering.
1059 @item noise_reduction, nr
1060 Change noise reduction. Argument is single float number.
1061 Syntax for the command is : "@var{noise_reduction}"
1063 @item noise_floor, nf
1064 Change noise floor. Argument is single float number.
1065 Syntax for the command is : "@var{noise_floor}"
1067 @item output_mode, om
1068 Change output mode operation.
1069 Syntax for the command is : "i", "o" or "n" string.
1073 Apply arbitrary expressions to samples in frequency domain.
1077 Set frequency domain real expression for each separate channel separated
1078 by '|'. Default is "re".
1079 If the number of input channels is greater than the number of
1080 expressions, the last specified expression is used for the remaining
1084 Set frequency domain imaginary expression for each separate channel
1085 separated by '|'. Default is "im".
1087 Each expression in @var{real} and @var{imag} can contain the following
1088 constants and functions:
1095 current frequency bin number
1098 number of available bins
1101 channel number of the current expression
1110 current real part of frequency bin of current channel
1113 current imaginary part of frequency bin of current channel
1116 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1119 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1125 It accepts the following values:
1141 Default is @code{w4096}
1144 Set window function. Default is @code{hann}.
1147 Set window overlap. If set to 1, the recommended overlap for selected
1148 window function will be picked. Default is @code{0.75}.
1151 @subsection Examples
1155 Leave almost only low frequencies in audio:
1157 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1164 Apply an arbitrary Frequency Impulse Response filter.
1166 This filter is designed for applying long FIR filters,
1167 up to 60 seconds long.
1169 It can be used as component for digital crossover filters,
1170 room equalization, cross talk cancellation, wavefield synthesis,
1171 auralization, ambiophonics, ambisonics and spatialization.
1173 This filter uses second stream as FIR coefficients.
1174 If second stream holds single channel, it will be used
1175 for all input channels in first stream, otherwise
1176 number of channels in second stream must be same as
1177 number of channels in first stream.
1179 It accepts the following parameters:
1183 Set dry gain. This sets input gain.
1186 Set wet gain. This sets final output gain.
1189 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1192 Enable applying gain measured from power of IR.
1194 Set which approach to use for auto gain measurement.
1198 Do not apply any gain.
1201 select peak gain, very conservative approach. This is default value.
1204 select DC gain, limited application.
1207 select gain to noise approach, this is most popular one.
1211 Set gain to be applied to IR coefficients before filtering.
1212 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1215 Set format of IR stream. Can be @code{mono} or @code{input}.
1216 Default is @code{input}.
1219 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1220 Allowed range is 0.1 to 60 seconds.
1223 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1224 By default it is disabled.
1227 Set for which IR channel to display frequency response. By default is first channel
1228 displayed. This option is used only when @var{response} is enabled.
1231 Set video stream size. This option is used only when @var{response} is enabled.
1234 Set video stream frame rate. This option is used only when @var{response} is enabled.
1237 Set minimal partition size used for convolution. Default is @var{8192}.
1238 Allowed range is from @var{8} to @var{32768}.
1239 Lower values decreases latency at cost of higher CPU usage.
1242 Set maximal partition size used for convolution. Default is @var{8192}.
1243 Allowed range is from @var{8} to @var{32768}.
1244 Lower values may increase CPU usage.
1247 @subsection Examples
1251 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1253 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1260 Set output format constraints for the input audio. The framework will
1261 negotiate the most appropriate format to minimize conversions.
1263 It accepts the following parameters:
1267 A '|'-separated list of requested sample formats.
1270 A '|'-separated list of requested sample rates.
1272 @item channel_layouts
1273 A '|'-separated list of requested channel layouts.
1275 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1276 for the required syntax.
1279 If a parameter is omitted, all values are allowed.
1281 Force the output to either unsigned 8-bit or signed 16-bit stereo
1283 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1288 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1289 processing reduces disturbing noise between useful signals.
1291 Gating is done by detecting the volume below a chosen level @var{threshold}
1292 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1293 floor is set via @var{range}. Because an exact manipulation of the signal
1294 would cause distortion of the waveform the reduction can be levelled over
1295 time. This is done by setting @var{attack} and @var{release}.
1297 @var{attack} determines how long the signal has to fall below the threshold
1298 before any reduction will occur and @var{release} sets the time the signal
1299 has to rise above the threshold to reduce the reduction again.
1300 Shorter signals than the chosen attack time will be left untouched.
1304 Set input level before filtering.
1305 Default is 1. Allowed range is from 0.015625 to 64.
1308 Set the level of gain reduction when the signal is below the threshold.
1309 Default is 0.06125. Allowed range is from 0 to 1.
1312 If a signal rises above this level the gain reduction is released.
1313 Default is 0.125. Allowed range is from 0 to 1.
1316 Set a ratio by which the signal is reduced.
1317 Default is 2. Allowed range is from 1 to 9000.
1320 Amount of milliseconds the signal has to rise above the threshold before gain
1322 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1325 Amount of milliseconds the signal has to fall below the threshold before the
1326 reduction is increased again. Default is 250 milliseconds.
1327 Allowed range is from 0.01 to 9000.
1330 Set amount of amplification of signal after processing.
1331 Default is 1. Allowed range is from 1 to 64.
1334 Curve the sharp knee around the threshold to enter gain reduction more softly.
1335 Default is 2.828427125. Allowed range is from 1 to 8.
1338 Choose if exact signal should be taken for detection or an RMS like one.
1339 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1342 Choose if the average level between all channels or the louder channel affects
1344 Default is @code{average}. Can be @code{average} or @code{maximum}.
1349 Apply an arbitrary Infinite Impulse Response filter.
1351 It accepts the following parameters:
1355 Set numerator/zeros coefficients.
1358 Set denominator/poles coefficients.
1370 Set coefficients format.
1376 Z-plane zeros/poles, cartesian (default)
1378 Z-plane zeros/poles, polar radians
1380 Z-plane zeros/poles, polar degrees
1384 Set kind of processing.
1385 Can be @code{d} - direct or @code{s} - serial cascading. Default is @code{s}.
1388 Set filtering precision.
1392 double-precision floating-point (default)
1394 single-precision floating-point
1402 Show IR frequency response, magnitude and phase in additional video stream.
1403 By default it is disabled.
1406 Set for which IR channel to display frequency response. By default is first channel
1407 displayed. This option is used only when @var{response} is enabled.
1410 Set video stream size. This option is used only when @var{response} is enabled.
1413 Coefficients in @code{tf} format are separated by spaces and are in ascending
1416 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1417 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1420 Different coefficients and gains can be provided for every channel, in such case
1421 use '|' to separate coefficients or gains. Last provided coefficients will be
1422 used for all remaining channels.
1424 @subsection Examples
1428 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1430 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
1434 Same as above but in @code{zp} format:
1436 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
1442 The limiter prevents an input signal from rising over a desired threshold.
1443 This limiter uses lookahead technology to prevent your signal from distorting.
1444 It means that there is a small delay after the signal is processed. Keep in mind
1445 that the delay it produces is the attack time you set.
1447 The filter accepts the following options:
1451 Set input gain. Default is 1.
1454 Set output gain. Default is 1.
1457 Don't let signals above this level pass the limiter. Default is 1.
1460 The limiter will reach its attenuation level in this amount of time in
1461 milliseconds. Default is 5 milliseconds.
1464 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1465 Default is 50 milliseconds.
1468 When gain reduction is always needed ASC takes care of releasing to an
1469 average reduction level rather than reaching a reduction of 0 in the release
1473 Select how much the release time is affected by ASC, 0 means nearly no changes
1474 in release time while 1 produces higher release times.
1477 Auto level output signal. Default is enabled.
1478 This normalizes audio back to 0dB if enabled.
1481 Depending on picked setting it is recommended to upsample input 2x or 4x times
1482 with @ref{aresample} before applying this filter.
1486 Apply a two-pole all-pass filter with central frequency (in Hz)
1487 @var{frequency}, and filter-width @var{width}.
1488 An all-pass filter changes the audio's frequency to phase relationship
1489 without changing its frequency to amplitude relationship.
1491 The filter accepts the following options:
1495 Set frequency in Hz.
1498 Set method to specify band-width of filter.
1513 Specify the band-width of a filter in width_type units.
1516 Specify which channels to filter, by default all available are filtered.
1519 @subsection Commands
1521 This filter supports the following commands:
1524 Change allpass frequency.
1525 Syntax for the command is : "@var{frequency}"
1528 Change allpass width_type.
1529 Syntax for the command is : "@var{width_type}"
1532 Change allpass width.
1533 Syntax for the command is : "@var{width}"
1540 The filter accepts the following options:
1544 Set the number of loops. Setting this value to -1 will result in infinite loops.
1548 Set maximal number of samples. Default is 0.
1551 Set first sample of loop. Default is 0.
1557 Merge two or more audio streams into a single multi-channel stream.
1559 The filter accepts the following options:
1564 Set the number of inputs. Default is 2.
1568 If the channel layouts of the inputs are disjoint, and therefore compatible,
1569 the channel layout of the output will be set accordingly and the channels
1570 will be reordered as necessary. If the channel layouts of the inputs are not
1571 disjoint, the output will have all the channels of the first input then all
1572 the channels of the second input, in that order, and the channel layout of
1573 the output will be the default value corresponding to the total number of
1576 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1577 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1578 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1579 first input, b1 is the first channel of the second input).
1581 On the other hand, if both input are in stereo, the output channels will be
1582 in the default order: a1, a2, b1, b2, and the channel layout will be
1583 arbitrarily set to 4.0, which may or may not be the expected value.
1585 All inputs must have the same sample rate, and format.
1587 If inputs do not have the same duration, the output will stop with the
1590 @subsection Examples
1594 Merge two mono files into a stereo stream:
1596 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1600 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1602 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
1608 Mixes multiple audio inputs into a single output.
1610 Note that this filter only supports float samples (the @var{amerge}
1611 and @var{pan} audio filters support many formats). If the @var{amix}
1612 input has integer samples then @ref{aresample} will be automatically
1613 inserted to perform the conversion to float samples.
1617 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1619 will mix 3 input audio streams to a single output with the same duration as the
1620 first input and a dropout transition time of 3 seconds.
1622 It accepts the following parameters:
1626 The number of inputs. If unspecified, it defaults to 2.
1629 How to determine the end-of-stream.
1633 The duration of the longest input. (default)
1636 The duration of the shortest input.
1639 The duration of the first input.
1643 @item dropout_transition
1644 The transition time, in seconds, for volume renormalization when an input
1645 stream ends. The default value is 2 seconds.
1648 Specify weight of each input audio stream as sequence.
1649 Each weight is separated by space. By default all inputs have same weight.
1654 Multiply first audio stream with second audio stream and store result
1655 in output audio stream. Multiplication is done by multiplying each
1656 sample from first stream with sample at same position from second stream.
1658 With this element-wise multiplication one can create amplitude fades and
1659 amplitude modulations.
1661 @section anequalizer
1663 High-order parametric multiband equalizer for each channel.
1665 It accepts the following parameters:
1669 This option string is in format:
1670 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1671 Each equalizer band is separated by '|'.
1675 Set channel number to which equalization will be applied.
1676 If input doesn't have that channel the entry is ignored.
1679 Set central frequency for band.
1680 If input doesn't have that frequency the entry is ignored.
1683 Set band width in hertz.
1686 Set band gain in dB.
1689 Set filter type for band, optional, can be:
1693 Butterworth, this is default.
1704 With this option activated frequency response of anequalizer is displayed
1708 Set video stream size. Only useful if curves option is activated.
1711 Set max gain that will be displayed. Only useful if curves option is activated.
1712 Setting this to a reasonable value makes it possible to display gain which is derived from
1713 neighbour bands which are too close to each other and thus produce higher gain
1714 when both are activated.
1717 Set frequency scale used to draw frequency response in video output.
1718 Can be linear or logarithmic. Default is logarithmic.
1721 Set color for each channel curve which is going to be displayed in video stream.
1722 This is list of color names separated by space or by '|'.
1723 Unrecognised or missing colors will be replaced by white color.
1726 @subsection Examples
1730 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1731 for first 2 channels using Chebyshev type 1 filter:
1733 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1737 @subsection Commands
1739 This filter supports the following commands:
1742 Alter existing filter parameters.
1743 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1745 @var{fN} is existing filter number, starting from 0, if no such filter is available
1747 @var{freq} set new frequency parameter.
1748 @var{width} set new width parameter in herz.
1749 @var{gain} set new gain parameter in dB.
1751 Full filter invocation with asendcmd may look like this:
1752 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1757 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1759 Each sample is adjusted by looking for other samples with similar contexts. This
1760 context similarity is defined by comparing their surrounding patches of size
1761 @option{p}. Patches are searched in an area of @option{r} around the sample.
1763 The filter accepts the following options.
1767 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
1770 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
1771 Default value is 2 milliseconds.
1774 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
1775 Default value is 6 milliseconds.
1780 Pass the audio source unchanged to the output.
1784 Pad the end of an audio stream with silence.
1786 This can be used together with @command{ffmpeg} @option{-shortest} to
1787 extend audio streams to the same length as the video stream.
1789 A description of the accepted options follows.
1793 Set silence packet size. Default value is 4096.
1796 Set the number of samples of silence to add to the end. After the
1797 value is reached, the stream is terminated. This option is mutually
1798 exclusive with @option{whole_len}.
1801 Set the minimum total number of samples in the output audio stream. If
1802 the value is longer than the input audio length, silence is added to
1803 the end, until the value is reached. This option is mutually exclusive
1804 with @option{pad_len}.
1807 Specify the duration of samples of silence to add. See
1808 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1809 for the accepted syntax. Used only if set to non-zero value.
1812 Specify the minimum total duration in the output audio stream. See
1813 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1814 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
1815 the input audio length, silence is added to the end, until the value is reached.
1816 This option is mutually exclusive with @option{pad_dur}
1819 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
1820 nor @option{whole_dur} option is set, the filter will add silence to the end of
1821 the input stream indefinitely.
1823 @subsection Examples
1827 Add 1024 samples of silence to the end of the input:
1833 Make sure the audio output will contain at least 10000 samples, pad
1834 the input with silence if required:
1836 apad=whole_len=10000
1840 Use @command{ffmpeg} to pad the audio input with silence, so that the
1841 video stream will always result the shortest and will be converted
1842 until the end in the output file when using the @option{shortest}
1845 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
1850 Add a phasing effect to the input audio.
1852 A phaser filter creates series of peaks and troughs in the frequency spectrum.
1853 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
1855 A description of the accepted parameters follows.
1859 Set input gain. Default is 0.4.
1862 Set output gain. Default is 0.74
1865 Set delay in milliseconds. Default is 3.0.
1868 Set decay. Default is 0.4.
1871 Set modulation speed in Hz. Default is 0.5.
1874 Set modulation type. Default is triangular.
1876 It accepts the following values:
1885 Audio pulsator is something between an autopanner and a tremolo.
1886 But it can produce funny stereo effects as well. Pulsator changes the volume
1887 of the left and right channel based on a LFO (low frequency oscillator) with
1888 different waveforms and shifted phases.
1889 This filter have the ability to define an offset between left and right
1890 channel. An offset of 0 means that both LFO shapes match each other.
1891 The left and right channel are altered equally - a conventional tremolo.
1892 An offset of 50% means that the shape of the right channel is exactly shifted
1893 in phase (or moved backwards about half of the frequency) - pulsator acts as
1894 an autopanner. At 1 both curves match again. Every setting in between moves the
1895 phase shift gapless between all stages and produces some "bypassing" sounds with
1896 sine and triangle waveforms. The more you set the offset near 1 (starting from
1897 the 0.5) the faster the signal passes from the left to the right speaker.
1899 The filter accepts the following options:
1903 Set input gain. By default it is 1. Range is [0.015625 - 64].
1906 Set output gain. By default it is 1. Range is [0.015625 - 64].
1909 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
1910 sawup or sawdown. Default is sine.
1913 Set modulation. Define how much of original signal is affected by the LFO.
1916 Set left channel offset. Default is 0. Allowed range is [0 - 1].
1919 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
1922 Set pulse width. Default is 1. Allowed range is [0 - 2].
1925 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
1928 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
1932 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
1936 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
1937 if timing is set to hz.
1943 Resample the input audio to the specified parameters, using the
1944 libswresample library. If none are specified then the filter will
1945 automatically convert between its input and output.
1947 This filter is also able to stretch/squeeze the audio data to make it match
1948 the timestamps or to inject silence / cut out audio to make it match the
1949 timestamps, do a combination of both or do neither.
1951 The filter accepts the syntax
1952 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
1953 expresses a sample rate and @var{resampler_options} is a list of
1954 @var{key}=@var{value} pairs, separated by ":". See the
1955 @ref{Resampler Options,,"Resampler Options" section in the
1956 ffmpeg-resampler(1) manual,ffmpeg-resampler}
1957 for the complete list of supported options.
1959 @subsection Examples
1963 Resample the input audio to 44100Hz:
1969 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
1970 samples per second compensation:
1972 aresample=async=1000
1978 Reverse an audio clip.
1980 Warning: This filter requires memory to buffer the entire clip, so trimming
1983 @subsection Examples
1987 Take the first 5 seconds of a clip, and reverse it.
1989 atrim=end=5,areverse
1993 @section asetnsamples
1995 Set the number of samples per each output audio frame.
1997 The last output packet may contain a different number of samples, as
1998 the filter will flush all the remaining samples when the input audio
2001 The filter accepts the following options:
2005 @item nb_out_samples, n
2006 Set the number of frames per each output audio frame. The number is
2007 intended as the number of samples @emph{per each channel}.
2008 Default value is 1024.
2011 If set to 1, the filter will pad the last audio frame with zeroes, so
2012 that the last frame will contain the same number of samples as the
2013 previous ones. Default value is 1.
2016 For example, to set the number of per-frame samples to 1234 and
2017 disable padding for the last frame, use:
2019 asetnsamples=n=1234:p=0
2024 Set the sample rate without altering the PCM data.
2025 This will result in a change of speed and pitch.
2027 The filter accepts the following options:
2030 @item sample_rate, r
2031 Set the output sample rate. Default is 44100 Hz.
2036 Show a line containing various information for each input audio frame.
2037 The input audio is not modified.
2039 The shown line contains a sequence of key/value pairs of the form
2040 @var{key}:@var{value}.
2042 The following values are shown in the output:
2046 The (sequential) number of the input frame, starting from 0.
2049 The presentation timestamp of the input frame, in time base units; the time base
2050 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2053 The presentation timestamp of the input frame in seconds.
2056 position of the frame in the input stream, -1 if this information in
2057 unavailable and/or meaningless (for example in case of synthetic audio)
2066 The sample rate for the audio frame.
2069 The number of samples (per channel) in the frame.
2072 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2073 audio, the data is treated as if all the planes were concatenated.
2075 @item plane_checksums
2076 A list of Adler-32 checksums for each data plane.
2082 Display time domain statistical information about the audio channels.
2083 Statistics are calculated and displayed for each audio channel and,
2084 where applicable, an overall figure is also given.
2086 It accepts the following option:
2089 Short window length in seconds, used for peak and trough RMS measurement.
2090 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2094 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2095 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2098 Available keys for each channel are:
2134 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2135 this @code{lavfi.astats.Overall.Peak_count}.
2137 For description what each key means read below.
2140 Set number of frame after which stats are going to be recalculated.
2141 Default is disabled.
2144 A description of each shown parameter follows:
2148 Mean amplitude displacement from zero.
2151 Minimal sample level.
2154 Maximal sample level.
2156 @item Min difference
2157 Minimal difference between two consecutive samples.
2159 @item Max difference
2160 Maximal difference between two consecutive samples.
2162 @item Mean difference
2163 Mean difference between two consecutive samples.
2164 The average of each difference between two consecutive samples.
2166 @item RMS difference
2167 Root Mean Square difference between two consecutive samples.
2171 Standard peak and RMS level measured in dBFS.
2175 Peak and trough values for RMS level measured over a short window.
2178 Standard ratio of peak to RMS level (note: not in dB).
2181 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2182 (i.e. either @var{Min level} or @var{Max level}).
2185 Number of occasions (not the number of samples) that the signal attained either
2186 @var{Min level} or @var{Max level}.
2189 Overall bit depth of audio. Number of bits used for each sample.
2192 Measured dynamic range of audio in dB.
2194 @item Zero crossings
2195 Number of points where the waveform crosses the zero level axis.
2197 @item Zero crossings rate
2198 Rate of Zero crossings and number of audio samples.
2205 The filter accepts exactly one parameter, the audio tempo. If not
2206 specified then the filter will assume nominal 1.0 tempo. Tempo must
2207 be in the [0.5, 100.0] range.
2209 Note that tempo greater than 2 will skip some samples rather than
2210 blend them in. If for any reason this is a concern it is always
2211 possible to daisy-chain several instances of atempo to achieve the
2212 desired product tempo.
2214 @subsection Examples
2218 Slow down audio to 80% tempo:
2224 To speed up audio to 300% tempo:
2230 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2232 atempo=sqrt(3),atempo=sqrt(3)
2238 Trim the input so that the output contains one continuous subpart of the input.
2240 It accepts the following parameters:
2243 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2244 sample with the timestamp @var{start} will be the first sample in the output.
2247 Specify time of the first audio sample that will be dropped, i.e. the
2248 audio sample immediately preceding the one with the timestamp @var{end} will be
2249 the last sample in the output.
2252 Same as @var{start}, except this option sets the start timestamp in samples
2256 Same as @var{end}, except this option sets the end timestamp in samples instead
2260 The maximum duration of the output in seconds.
2263 The number of the first sample that should be output.
2266 The number of the first sample that should be dropped.
2269 @option{start}, @option{end}, and @option{duration} are expressed as time
2270 duration specifications; see
2271 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2273 Note that the first two sets of the start/end options and the @option{duration}
2274 option look at the frame timestamp, while the _sample options simply count the
2275 samples that pass through the filter. So start/end_pts and start/end_sample will
2276 give different results when the timestamps are wrong, inexact or do not start at
2277 zero. Also note that this filter does not modify the timestamps. If you wish
2278 to have the output timestamps start at zero, insert the asetpts filter after the
2281 If multiple start or end options are set, this filter tries to be greedy and
2282 keep all samples that match at least one of the specified constraints. To keep
2283 only the part that matches all the constraints at once, chain multiple atrim
2286 The defaults are such that all the input is kept. So it is possible to set e.g.
2287 just the end values to keep everything before the specified time.
2292 Drop everything except the second minute of input:
2294 ffmpeg -i INPUT -af atrim=60:120
2298 Keep only the first 1000 samples:
2300 ffmpeg -i INPUT -af atrim=end_sample=1000
2307 Apply a two-pole Butterworth band-pass filter with central
2308 frequency @var{frequency}, and (3dB-point) band-width width.
2309 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2310 instead of the default: constant 0dB peak gain.
2311 The filter roll off at 6dB per octave (20dB per decade).
2313 The filter accepts the following options:
2317 Set the filter's central frequency. Default is @code{3000}.
2320 Constant skirt gain if set to 1. Defaults to 0.
2323 Set method to specify band-width of filter.
2338 Specify the band-width of a filter in width_type units.
2341 Specify which channels to filter, by default all available are filtered.
2344 @subsection Commands
2346 This filter supports the following commands:
2349 Change bandpass frequency.
2350 Syntax for the command is : "@var{frequency}"
2353 Change bandpass width_type.
2354 Syntax for the command is : "@var{width_type}"
2357 Change bandpass width.
2358 Syntax for the command is : "@var{width}"
2363 Apply a two-pole Butterworth band-reject filter with central
2364 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2365 The filter roll off at 6dB per octave (20dB per decade).
2367 The filter accepts the following options:
2371 Set the filter's central frequency. Default is @code{3000}.
2374 Set method to specify band-width of filter.
2389 Specify the band-width of a filter in width_type units.
2392 Specify which channels to filter, by default all available are filtered.
2395 @subsection Commands
2397 This filter supports the following commands:
2400 Change bandreject frequency.
2401 Syntax for the command is : "@var{frequency}"
2404 Change bandreject width_type.
2405 Syntax for the command is : "@var{width_type}"
2408 Change bandreject width.
2409 Syntax for the command is : "@var{width}"
2412 @section bass, lowshelf
2414 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2415 shelving filter with a response similar to that of a standard
2416 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2418 The filter accepts the following options:
2422 Give the gain at 0 Hz. Its useful range is about -20
2423 (for a large cut) to +20 (for a large boost).
2424 Beware of clipping when using a positive gain.
2427 Set the filter's central frequency and so can be used
2428 to extend or reduce the frequency range to be boosted or cut.
2429 The default value is @code{100} Hz.
2432 Set method to specify band-width of filter.
2447 Determine how steep is the filter's shelf transition.
2450 Specify which channels to filter, by default all available are filtered.
2453 @subsection Commands
2455 This filter supports the following commands:
2458 Change bass frequency.
2459 Syntax for the command is : "@var{frequency}"
2462 Change bass width_type.
2463 Syntax for the command is : "@var{width_type}"
2467 Syntax for the command is : "@var{width}"
2471 Syntax for the command is : "@var{gain}"
2476 Apply a biquad IIR filter with the given coefficients.
2477 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2478 are the numerator and denominator coefficients respectively.
2479 and @var{channels}, @var{c} specify which channels to filter, by default all
2480 available are filtered.
2482 @subsection Commands
2484 This filter supports the following commands:
2492 Change biquad parameter.
2493 Syntax for the command is : "@var{value}"
2497 Bauer stereo to binaural transformation, which improves headphone listening of
2498 stereo audio records.
2500 To enable compilation of this filter you need to configure FFmpeg with
2501 @code{--enable-libbs2b}.
2503 It accepts the following parameters:
2507 Pre-defined crossfeed level.
2511 Default level (fcut=700, feed=50).
2514 Chu Moy circuit (fcut=700, feed=60).
2517 Jan Meier circuit (fcut=650, feed=95).
2522 Cut frequency (in Hz).
2531 Remap input channels to new locations.
2533 It accepts the following parameters:
2536 Map channels from input to output. The argument is a '|'-separated list of
2537 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2538 @var{in_channel} form. @var{in_channel} can be either the name of the input
2539 channel (e.g. FL for front left) or its index in the input channel layout.
2540 @var{out_channel} is the name of the output channel or its index in the output
2541 channel layout. If @var{out_channel} is not given then it is implicitly an
2542 index, starting with zero and increasing by one for each mapping.
2544 @item channel_layout
2545 The channel layout of the output stream.
2548 If no mapping is present, the filter will implicitly map input channels to
2549 output channels, preserving indices.
2551 @subsection Examples
2555 For example, assuming a 5.1+downmix input MOV file,
2557 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2559 will create an output WAV file tagged as stereo from the downmix channels of
2563 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2565 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
2569 @section channelsplit
2571 Split each channel from an input audio stream into a separate output stream.
2573 It accepts the following parameters:
2575 @item channel_layout
2576 The channel layout of the input stream. The default is "stereo".
2578 A channel layout describing the channels to be extracted as separate output streams
2579 or "all" to extract each input channel as a separate stream. The default is "all".
2581 Choosing channels not present in channel layout in the input will result in an error.
2584 @subsection Examples
2588 For example, assuming a stereo input MP3 file,
2590 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
2592 will create an output Matroska file with two audio streams, one containing only
2593 the left channel and the other the right channel.
2596 Split a 5.1 WAV file into per-channel files:
2598 ffmpeg -i in.wav -filter_complex
2599 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
2600 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
2601 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
2606 Extract only LFE from a 5.1 WAV file:
2608 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
2609 -map '[LFE]' lfe.wav
2614 Add a chorus effect to the audio.
2616 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
2618 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
2619 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
2620 The modulation depth defines the range the modulated delay is played before or after
2621 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
2622 sound tuned around the original one, like in a chorus where some vocals are slightly
2625 It accepts the following parameters:
2628 Set input gain. Default is 0.4.
2631 Set output gain. Default is 0.4.
2634 Set delays. A typical delay is around 40ms to 60ms.
2646 @subsection Examples
2652 chorus=0.7:0.9:55:0.4:0.25:2
2658 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
2662 Fuller sounding chorus with three delays:
2664 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
2669 Compress or expand the audio's dynamic range.
2671 It accepts the following parameters:
2677 A list of times in seconds for each channel over which the instantaneous level
2678 of the input signal is averaged to determine its volume. @var{attacks} refers to
2679 increase of volume and @var{decays} refers to decrease of volume. For most
2680 situations, the attack time (response to the audio getting louder) should be
2681 shorter than the decay time, because the human ear is more sensitive to sudden
2682 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
2683 a typical value for decay is 0.8 seconds.
2684 If specified number of attacks & decays is lower than number of channels, the last
2685 set attack/decay will be used for all remaining channels.
2688 A list of points for the transfer function, specified in dB relative to the
2689 maximum possible signal amplitude. Each key points list must be defined using
2690 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
2691 @code{x0/y0 x1/y1 x2/y2 ....}
2693 The input values must be in strictly increasing order but the transfer function
2694 does not have to be monotonically rising. The point @code{0/0} is assumed but
2695 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
2696 function are @code{-70/-70|-60/-20|1/0}.
2699 Set the curve radius in dB for all joints. It defaults to 0.01.
2702 Set the additional gain in dB to be applied at all points on the transfer
2703 function. This allows for easy adjustment of the overall gain.
2707 Set an initial volume, in dB, to be assumed for each channel when filtering
2708 starts. This permits the user to supply a nominal level initially, so that, for
2709 example, a very large gain is not applied to initial signal levels before the
2710 companding has begun to operate. A typical value for audio which is initially
2711 quiet is -90 dB. It defaults to 0.
2714 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
2715 delayed before being fed to the volume adjuster. Specifying a delay
2716 approximately equal to the attack/decay times allows the filter to effectively
2717 operate in predictive rather than reactive mode. It defaults to 0.
2721 @subsection Examples
2725 Make music with both quiet and loud passages suitable for listening to in a
2728 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
2731 Another example for audio with whisper and explosion parts:
2733 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
2737 A noise gate for when the noise is at a lower level than the signal:
2739 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
2743 Here is another noise gate, this time for when the noise is at a higher level
2744 than the signal (making it, in some ways, similar to squelch):
2746 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
2750 2:1 compression starting at -6dB:
2752 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
2756 2:1 compression starting at -9dB:
2758 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
2762 2:1 compression starting at -12dB:
2764 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
2768 2:1 compression starting at -18dB:
2770 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
2774 3:1 compression starting at -15dB:
2776 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
2782 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
2788 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
2792 Hard limiter at -6dB:
2794 compand=attacks=0:points=-80/-80|-6/-6|20/-6
2798 Hard limiter at -12dB:
2800 compand=attacks=0:points=-80/-80|-12/-12|20/-12
2804 Hard noise gate at -35 dB:
2806 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
2812 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
2816 @section compensationdelay
2818 Compensation Delay Line is a metric based delay to compensate differing
2819 positions of microphones or speakers.
2821 For example, you have recorded guitar with two microphones placed in
2822 different location. Because the front of sound wave has fixed speed in
2823 normal conditions, the phasing of microphones can vary and depends on
2824 their location and interposition. The best sound mix can be achieved when
2825 these microphones are in phase (synchronized). Note that distance of
2826 ~30 cm between microphones makes one microphone to capture signal in
2827 antiphase to another microphone. That makes the final mix sounding moody.
2828 This filter helps to solve phasing problems by adding different delays
2829 to each microphone track and make them synchronized.
2831 The best result can be reached when you take one track as base and
2832 synchronize other tracks one by one with it.
2833 Remember that synchronization/delay tolerance depends on sample rate, too.
2834 Higher sample rates will give more tolerance.
2836 It accepts the following parameters:
2840 Set millimeters distance. This is compensation distance for fine tuning.
2844 Set cm distance. This is compensation distance for tightening distance setup.
2848 Set meters distance. This is compensation distance for hard distance setup.
2852 Set dry amount. Amount of unprocessed (dry) signal.
2856 Set wet amount. Amount of processed (wet) signal.
2860 Set temperature degree in Celsius. This is the temperature of the environment.
2865 Apply headphone crossfeed filter.
2867 Crossfeed is the process of blending the left and right channels of stereo
2869 It is mainly used to reduce extreme stereo separation of low frequencies.
2871 The intent is to produce more speaker like sound to the listener.
2873 The filter accepts the following options:
2877 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
2878 This sets gain of low shelf filter for side part of stereo image.
2879 Default is -6dB. Max allowed is -30db when strength is set to 1.
2882 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
2883 This sets cut off frequency of low shelf filter. Default is cut off near
2884 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
2887 Set input gain. Default is 0.9.
2890 Set output gain. Default is 1.
2893 @section crystalizer
2894 Simple algorithm to expand audio dynamic range.
2896 The filter accepts the following options:
2900 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
2901 (unchanged sound) to 10.0 (maximum effect).
2904 Enable clipping. By default is enabled.
2908 Apply a DC shift to the audio.
2910 This can be useful to remove a DC offset (caused perhaps by a hardware problem
2911 in the recording chain) from the audio. The effect of a DC offset is reduced
2912 headroom and hence volume. The @ref{astats} filter can be used to determine if
2913 a signal has a DC offset.
2917 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
2921 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
2922 used to prevent clipping.
2926 Measure audio dynamic range.
2928 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
2929 is found in transition material. And anything less that 8 have very poor dynamics
2930 and is very compressed.
2932 The filter accepts the following options:
2936 Set window length in seconds used to split audio into segments of equal length.
2937 Default is 3 seconds.
2941 Dynamic Audio Normalizer.
2943 This filter applies a certain amount of gain to the input audio in order
2944 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
2945 contrast to more "simple" normalization algorithms, the Dynamic Audio
2946 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
2947 This allows for applying extra gain to the "quiet" sections of the audio
2948 while avoiding distortions or clipping the "loud" sections. In other words:
2949 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
2950 sections, in the sense that the volume of each section is brought to the
2951 same target level. Note, however, that the Dynamic Audio Normalizer achieves
2952 this goal *without* applying "dynamic range compressing". It will retain 100%
2953 of the dynamic range *within* each section of the audio file.
2957 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
2958 Default is 500 milliseconds.
2959 The Dynamic Audio Normalizer processes the input audio in small chunks,
2960 referred to as frames. This is required, because a peak magnitude has no
2961 meaning for just a single sample value. Instead, we need to determine the
2962 peak magnitude for a contiguous sequence of sample values. While a "standard"
2963 normalizer would simply use the peak magnitude of the complete file, the
2964 Dynamic Audio Normalizer determines the peak magnitude individually for each
2965 frame. The length of a frame is specified in milliseconds. By default, the
2966 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
2967 been found to give good results with most files.
2968 Note that the exact frame length, in number of samples, will be determined
2969 automatically, based on the sampling rate of the individual input audio file.
2972 Set the Gaussian filter window size. In range from 3 to 301, must be odd
2973 number. Default is 31.
2974 Probably the most important parameter of the Dynamic Audio Normalizer is the
2975 @code{window size} of the Gaussian smoothing filter. The filter's window size
2976 is specified in frames, centered around the current frame. For the sake of
2977 simplicity, this must be an odd number. Consequently, the default value of 31
2978 takes into account the current frame, as well as the 15 preceding frames and
2979 the 15 subsequent frames. Using a larger window results in a stronger
2980 smoothing effect and thus in less gain variation, i.e. slower gain
2981 adaptation. Conversely, using a smaller window results in a weaker smoothing
2982 effect and thus in more gain variation, i.e. faster gain adaptation.
2983 In other words, the more you increase this value, the more the Dynamic Audio
2984 Normalizer will behave like a "traditional" normalization filter. On the
2985 contrary, the more you decrease this value, the more the Dynamic Audio
2986 Normalizer will behave like a dynamic range compressor.
2989 Set the target peak value. This specifies the highest permissible magnitude
2990 level for the normalized audio input. This filter will try to approach the
2991 target peak magnitude as closely as possible, but at the same time it also
2992 makes sure that the normalized signal will never exceed the peak magnitude.
2993 A frame's maximum local gain factor is imposed directly by the target peak
2994 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
2995 It is not recommended to go above this value.
2998 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
2999 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3000 factor for each input frame, i.e. the maximum gain factor that does not
3001 result in clipping or distortion. The maximum gain factor is determined by
3002 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3003 additionally bounds the frame's maximum gain factor by a predetermined
3004 (global) maximum gain factor. This is done in order to avoid excessive gain
3005 factors in "silent" or almost silent frames. By default, the maximum gain
3006 factor is 10.0, For most inputs the default value should be sufficient and
3007 it usually is not recommended to increase this value. Though, for input
3008 with an extremely low overall volume level, it may be necessary to allow even
3009 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3010 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3011 Instead, a "sigmoid" threshold function will be applied. This way, the
3012 gain factors will smoothly approach the threshold value, but never exceed that
3016 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3017 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3018 This means that the maximum local gain factor for each frame is defined
3019 (only) by the frame's highest magnitude sample. This way, the samples can
3020 be amplified as much as possible without exceeding the maximum signal
3021 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3022 Normalizer can also take into account the frame's root mean square,
3023 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3024 determine the power of a time-varying signal. It is therefore considered
3025 that the RMS is a better approximation of the "perceived loudness" than
3026 just looking at the signal's peak magnitude. Consequently, by adjusting all
3027 frames to a constant RMS value, a uniform "perceived loudness" can be
3028 established. If a target RMS value has been specified, a frame's local gain
3029 factor is defined as the factor that would result in exactly that RMS value.
3030 Note, however, that the maximum local gain factor is still restricted by the
3031 frame's highest magnitude sample, in order to prevent clipping.
3034 Enable channels coupling. By default is enabled.
3035 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3036 amount. This means the same gain factor will be applied to all channels, i.e.
3037 the maximum possible gain factor is determined by the "loudest" channel.
3038 However, in some recordings, it may happen that the volume of the different
3039 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3040 In this case, this option can be used to disable the channel coupling. This way,
3041 the gain factor will be determined independently for each channel, depending
3042 only on the individual channel's highest magnitude sample. This allows for
3043 harmonizing the volume of the different channels.
3046 Enable DC bias correction. By default is disabled.
3047 An audio signal (in the time domain) is a sequence of sample values.
3048 In the Dynamic Audio Normalizer these sample values are represented in the
3049 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3050 audio signal, or "waveform", should be centered around the zero point.
3051 That means if we calculate the mean value of all samples in a file, or in a
3052 single frame, then the result should be 0.0 or at least very close to that
3053 value. If, however, there is a significant deviation of the mean value from
3054 0.0, in either positive or negative direction, this is referred to as a
3055 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3056 Audio Normalizer provides optional DC bias correction.
3057 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3058 the mean value, or "DC correction" offset, of each input frame and subtract
3059 that value from all of the frame's sample values which ensures those samples
3060 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3061 boundaries, the DC correction offset values will be interpolated smoothly
3062 between neighbouring frames.
3065 Enable alternative boundary mode. By default is disabled.
3066 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3067 around each frame. This includes the preceding frames as well as the
3068 subsequent frames. However, for the "boundary" frames, located at the very
3069 beginning and at the very end of the audio file, not all neighbouring
3070 frames are available. In particular, for the first few frames in the audio
3071 file, the preceding frames are not known. And, similarly, for the last few
3072 frames in the audio file, the subsequent frames are not known. Thus, the
3073 question arises which gain factors should be assumed for the missing frames
3074 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3075 to deal with this situation. The default boundary mode assumes a gain factor
3076 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3077 "fade out" at the beginning and at the end of the input, respectively.
3080 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3081 By default, the Dynamic Audio Normalizer does not apply "traditional"
3082 compression. This means that signal peaks will not be pruned and thus the
3083 full dynamic range will be retained within each local neighbourhood. However,
3084 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3085 normalization algorithm with a more "traditional" compression.
3086 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3087 (thresholding) function. If (and only if) the compression feature is enabled,
3088 all input frames will be processed by a soft knee thresholding function prior
3089 to the actual normalization process. Put simply, the thresholding function is
3090 going to prune all samples whose magnitude exceeds a certain threshold value.
3091 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3092 value. Instead, the threshold value will be adjusted for each individual
3094 In general, smaller parameters result in stronger compression, and vice versa.
3095 Values below 3.0 are not recommended, because audible distortion may appear.
3100 Make audio easier to listen to on headphones.
3102 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3103 so that when listened to on headphones the stereo image is moved from
3104 inside your head (standard for headphones) to outside and in front of
3105 the listener (standard for speakers).
3111 Apply a two-pole peaking equalisation (EQ) filter. With this
3112 filter, the signal-level at and around a selected frequency can
3113 be increased or decreased, whilst (unlike bandpass and bandreject
3114 filters) that at all other frequencies is unchanged.
3116 In order to produce complex equalisation curves, this filter can
3117 be given several times, each with a different central frequency.
3119 The filter accepts the following options:
3123 Set the filter's central frequency in Hz.
3126 Set method to specify band-width of filter.
3141 Specify the band-width of a filter in width_type units.
3144 Set the required gain or attenuation in dB.
3145 Beware of clipping when using a positive gain.
3148 Specify which channels to filter, by default all available are filtered.
3151 @subsection Examples
3154 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3156 equalizer=f=1000:t=h:width=200:g=-10
3160 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3162 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3166 @subsection Commands
3168 This filter supports the following commands:
3171 Change equalizer frequency.
3172 Syntax for the command is : "@var{frequency}"
3175 Change equalizer width_type.
3176 Syntax for the command is : "@var{width_type}"
3179 Change equalizer width.
3180 Syntax for the command is : "@var{width}"
3183 Change equalizer gain.
3184 Syntax for the command is : "@var{gain}"
3187 @section extrastereo
3189 Linearly increases the difference between left and right channels which
3190 adds some sort of "live" effect to playback.
3192 The filter accepts the following options:
3196 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3197 (average of both channels), with 1.0 sound will be unchanged, with
3198 -1.0 left and right channels will be swapped.
3201 Enable clipping. By default is enabled.
3204 @section firequalizer
3205 Apply FIR Equalization using arbitrary frequency response.
3207 The filter accepts the following option:
3211 Set gain curve equation (in dB). The expression can contain variables:
3214 the evaluated frequency
3218 channel number, set to 0 when multichannels evaluation is disabled
3220 channel id, see libavutil/channel_layout.h, set to the first channel id when
3221 multichannels evaluation is disabled
3225 channel_layout, see libavutil/channel_layout.h
3230 @item gain_interpolate(f)
3231 interpolate gain on frequency f based on gain_entry
3232 @item cubic_interpolate(f)
3233 same as gain_interpolate, but smoother
3235 This option is also available as command. Default is @code{gain_interpolate(f)}.
3238 Set gain entry for gain_interpolate function. The expression can
3242 store gain entry at frequency f with value g
3244 This option is also available as command.
3247 Set filter delay in seconds. Higher value means more accurate.
3248 Default is @code{0.01}.
3251 Set filter accuracy in Hz. Lower value means more accurate.
3252 Default is @code{5}.
3255 Set window function. Acceptable values are:
3258 rectangular window, useful when gain curve is already smooth
3260 hann window (default)
3266 3-terms continuous 1st derivative nuttall window
3268 minimum 3-terms discontinuous nuttall window
3270 4-terms continuous 1st derivative nuttall window
3272 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3274 blackman-harris window
3280 If enabled, use fixed number of audio samples. This improves speed when
3281 filtering with large delay. Default is disabled.
3284 Enable multichannels evaluation on gain. Default is disabled.
3287 Enable zero phase mode by subtracting timestamp to compensate delay.
3288 Default is disabled.
3291 Set scale used by gain. Acceptable values are:
3294 linear frequency, linear gain
3296 linear frequency, logarithmic (in dB) gain (default)
3298 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3300 logarithmic frequency, logarithmic gain
3304 Set file for dumping, suitable for gnuplot.
3307 Set scale for dumpfile. Acceptable values are same with scale option.
3311 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3312 Default is disabled.
3315 Enable minimum phase impulse response. Default is disabled.
3318 @subsection Examples
3323 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3326 lowpass at 1000 Hz with gain_entry:
3328 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3331 custom equalization:
3333 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3336 higher delay with zero phase to compensate delay:
3338 firequalizer=delay=0.1:fixed=on:zero_phase=on
3341 lowpass on left channel, highpass on right channel:
3343 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
3344 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
3349 Apply a flanging effect to the audio.
3351 The filter accepts the following options:
3355 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
3358 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
3361 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
3365 Set percentage of delayed signal mixed with original. Range from 0 to 100.
3366 Default value is 71.
3369 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
3372 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
3373 Default value is @var{sinusoidal}.
3376 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
3377 Default value is 25.
3380 Set delay-line interpolation, @var{linear} or @var{quadratic}.
3381 Default is @var{linear}.
3385 Apply Haas effect to audio.
3387 Note that this makes most sense to apply on mono signals.
3388 With this filter applied to mono signals it give some directionality and
3389 stretches its stereo image.
3391 The filter accepts the following options:
3395 Set input level. By default is @var{1}, or 0dB
3398 Set output level. By default is @var{1}, or 0dB.
3401 Set gain applied to side part of signal. By default is @var{1}.
3404 Set kind of middle source. Can be one of the following:
3414 Pick middle part signal of stereo image.
3417 Pick side part signal of stereo image.
3421 Change middle phase. By default is disabled.
3424 Set left channel delay. By default is @var{2.05} milliseconds.
3427 Set left channel balance. By default is @var{-1}.
3430 Set left channel gain. By default is @var{1}.
3433 Change left phase. By default is disabled.
3436 Set right channel delay. By defaults is @var{2.12} milliseconds.
3439 Set right channel balance. By default is @var{1}.
3442 Set right channel gain. By default is @var{1}.
3445 Change right phase. By default is enabled.
3450 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3451 embedded HDCD codes is expanded into a 20-bit PCM stream.
3453 The filter supports the Peak Extend and Low-level Gain Adjustment features
3454 of HDCD, and detects the Transient Filter flag.
3457 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3460 When using the filter with wav, note the default encoding for wav is 16-bit,
3461 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3462 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3464 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3465 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3468 The filter accepts the following options:
3471 @item disable_autoconvert
3472 Disable any automatic format conversion or resampling in the filter graph.
3474 @item process_stereo
3475 Process the stereo channels together. If target_gain does not match between
3476 channels, consider it invalid and use the last valid target_gain.
3479 Set the code detect timer period in ms.
3482 Always extend peaks above -3dBFS even if PE isn't signaled.
3485 Replace audio with a solid tone and adjust the amplitude to signal some
3486 specific aspect of the decoding process. The output file can be loaded in
3487 an audio editor alongside the original to aid analysis.
3489 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
3496 Gain adjustment level at each sample
3498 Samples where peak extend occurs
3500 Samples where the code detect timer is active
3502 Samples where the target gain does not match between channels
3508 Apply head-related transfer functions (HRTFs) to create virtual
3509 loudspeakers around the user for binaural listening via headphones.
3510 The HRIRs are provided via additional streams, for each channel
3511 one stereo input stream is needed.
3513 The filter accepts the following options:
3517 Set mapping of input streams for convolution.
3518 The argument is a '|'-separated list of channel names in order as they
3519 are given as additional stream inputs for filter.
3520 This also specify number of input streams. Number of input streams
3521 must be not less than number of channels in first stream plus one.
3524 Set gain applied to audio. Value is in dB. Default is 0.
3527 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
3528 processing audio in time domain which is slow.
3529 @var{freq} is processing audio in frequency domain which is fast.
3530 Default is @var{freq}.
3533 Set custom gain for LFE channels. Value is in dB. Default is 0.
3536 Set size of frame in number of samples which will be processed at once.
3537 Default value is @var{1024}. Allowed range is from 1024 to 96000.
3540 Set format of hrir stream.
3541 Default value is @var{stereo}. Alternative value is @var{multich}.
3542 If value is set to @var{stereo}, number of additional streams should
3543 be greater or equal to number of input channels in first input stream.
3544 Also each additional stream should have stereo number of channels.
3545 If value is set to @var{multich}, number of additional streams should
3546 be exactly one. Also number of input channels of additional stream
3547 should be equal or greater than twice number of channels of first input
3551 @subsection Examples
3555 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3556 each amovie filter use stereo file with IR coefficients as input.
3557 The files give coefficients for each position of virtual loudspeaker:
3560 -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"
3565 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3566 but now in @var{multich} @var{hrir} format.
3568 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"
3575 Apply a high-pass filter with 3dB point frequency.
3576 The filter can be either single-pole, or double-pole (the default).
3577 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3579 The filter accepts the following options:
3583 Set frequency in Hz. Default is 3000.
3586 Set number of poles. Default is 2.
3589 Set method to specify band-width of filter.
3604 Specify the band-width of a filter in width_type units.
3605 Applies only to double-pole filter.
3606 The default is 0.707q and gives a Butterworth response.
3609 Specify which channels to filter, by default all available are filtered.
3612 @subsection Commands
3614 This filter supports the following commands:
3617 Change highpass frequency.
3618 Syntax for the command is : "@var{frequency}"
3621 Change highpass width_type.
3622 Syntax for the command is : "@var{width_type}"
3625 Change highpass width.
3626 Syntax for the command is : "@var{width}"
3631 Join multiple input streams into one multi-channel stream.
3633 It accepts the following parameters:
3637 The number of input streams. It defaults to 2.
3639 @item channel_layout
3640 The desired output channel layout. It defaults to stereo.
3643 Map channels from inputs to output. The argument is a '|'-separated list of
3644 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
3645 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
3646 can be either the name of the input channel (e.g. FL for front left) or its
3647 index in the specified input stream. @var{out_channel} is the name of the output
3651 The filter will attempt to guess the mappings when they are not specified
3652 explicitly. It does so by first trying to find an unused matching input channel
3653 and if that fails it picks the first unused input channel.
3655 Join 3 inputs (with properly set channel layouts):
3657 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
3660 Build a 5.1 output from 6 single-channel streams:
3662 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
3663 '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'
3669 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
3671 To enable compilation of this filter you need to configure FFmpeg with
3672 @code{--enable-ladspa}.
3676 Specifies the name of LADSPA plugin library to load. If the environment
3677 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
3678 each one of the directories specified by the colon separated list in
3679 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
3680 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
3681 @file{/usr/lib/ladspa/}.
3684 Specifies the plugin within the library. Some libraries contain only
3685 one plugin, but others contain many of them. If this is not set filter
3686 will list all available plugins within the specified library.
3689 Set the '|' separated list of controls which are zero or more floating point
3690 values that determine the behavior of the loaded plugin (for example delay,
3692 Controls need to be defined using the following syntax:
3693 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
3694 @var{valuei} is the value set on the @var{i}-th control.
3695 Alternatively they can be also defined using the following syntax:
3696 @var{value0}|@var{value1}|@var{value2}|..., where
3697 @var{valuei} is the value set on the @var{i}-th control.
3698 If @option{controls} is set to @code{help}, all available controls and
3699 their valid ranges are printed.
3701 @item sample_rate, s
3702 Specify the sample rate, default to 44100. Only used if plugin have
3706 Set the number of samples per channel per each output frame, default
3707 is 1024. Only used if plugin have zero inputs.
3710 Set the minimum duration of the sourced audio. See
3711 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
3712 for the accepted syntax.
3713 Note that the resulting duration may be greater than the specified duration,
3714 as the generated audio is always cut at the end of a complete frame.
3715 If not specified, or the expressed duration is negative, the audio is
3716 supposed to be generated forever.
3717 Only used if plugin have zero inputs.
3721 @subsection Examples
3725 List all available plugins within amp (LADSPA example plugin) library:
3731 List all available controls and their valid ranges for @code{vcf_notch}
3732 plugin from @code{VCF} library:
3734 ladspa=f=vcf:p=vcf_notch:c=help
3738 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
3741 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
3745 Add reverberation to the audio using TAP-plugins
3746 (Tom's Audio Processing plugins):
3748 ladspa=file=tap_reverb:tap_reverb
3752 Generate white noise, with 0.2 amplitude:
3754 ladspa=file=cmt:noise_source_white:c=c0=.2
3758 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
3759 @code{C* Audio Plugin Suite} (CAPS) library:
3761 ladspa=file=caps:Click:c=c1=20'
3765 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
3767 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
3771 Increase volume by 20dB using fast lookahead limiter from Steve Harris
3772 @code{SWH Plugins} collection:
3774 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
3778 Attenuate low frequencies using Multiband EQ from Steve Harris
3779 @code{SWH Plugins} collection:
3781 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
3785 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
3788 ladspa=caps:Narrower
3792 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
3794 ladspa=caps:White:.2
3798 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
3800 ladspa=caps:Fractal:c=c1=1
3804 Dynamic volume normalization using @code{VLevel} plugin:
3806 ladspa=vlevel-ladspa:vlevel_mono
3810 @subsection Commands
3812 This filter supports the following commands:
3815 Modify the @var{N}-th control value.
3817 If the specified value is not valid, it is ignored and prior one is kept.
3822 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
3823 Support for both single pass (livestreams, files) and double pass (files) modes.
3824 This algorithm can target IL, LRA, and maximum true peak. To accurately detect true peaks,
3825 the audio stream will be upsampled to 192 kHz unless the normalization mode is linear.
3826 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
3828 The filter accepts the following options:
3832 Set integrated loudness target.
3833 Range is -70.0 - -5.0. Default value is -24.0.
3836 Set loudness range target.
3837 Range is 1.0 - 20.0. Default value is 7.0.
3840 Set maximum true peak.
3841 Range is -9.0 - +0.0. Default value is -2.0.
3843 @item measured_I, measured_i
3844 Measured IL of input file.
3845 Range is -99.0 - +0.0.
3847 @item measured_LRA, measured_lra
3848 Measured LRA of input file.
3849 Range is 0.0 - 99.0.
3851 @item measured_TP, measured_tp
3852 Measured true peak of input file.
3853 Range is -99.0 - +99.0.
3855 @item measured_thresh
3856 Measured threshold of input file.
3857 Range is -99.0 - +0.0.
3860 Set offset gain. Gain is applied before the true-peak limiter.
3861 Range is -99.0 - +99.0. Default is +0.0.
3864 Normalize linearly if possible.
3865 measured_I, measured_LRA, measured_TP, and measured_thresh must also
3866 to be specified in order to use this mode.
3867 Options are true or false. Default is true.
3870 Treat mono input files as "dual-mono". If a mono file is intended for playback
3871 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
3872 If set to @code{true}, this option will compensate for this effect.
3873 Multi-channel input files are not affected by this option.
3874 Options are true or false. Default is false.
3877 Set print format for stats. Options are summary, json, or none.
3878 Default value is none.
3883 Apply a low-pass filter with 3dB point frequency.
3884 The filter can be either single-pole or double-pole (the default).
3885 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3887 The filter accepts the following options:
3891 Set frequency in Hz. Default is 500.
3894 Set number of poles. Default is 2.
3897 Set method to specify band-width of filter.
3912 Specify the band-width of a filter in width_type units.
3913 Applies only to double-pole filter.
3914 The default is 0.707q and gives a Butterworth response.
3917 Specify which channels to filter, by default all available are filtered.
3920 @subsection Examples
3923 Lowpass only LFE channel, it LFE is not present it does nothing:
3929 @subsection Commands
3931 This filter supports the following commands:
3934 Change lowpass frequency.
3935 Syntax for the command is : "@var{frequency}"
3938 Change lowpass width_type.
3939 Syntax for the command is : "@var{width_type}"
3942 Change lowpass width.
3943 Syntax for the command is : "@var{width}"
3948 Load a LV2 (LADSPA Version 2) plugin.
3950 To enable compilation of this filter you need to configure FFmpeg with
3951 @code{--enable-lv2}.
3955 Specifies the plugin URI. You may need to escape ':'.
3958 Set the '|' separated list of controls which are zero or more floating point
3959 values that determine the behavior of the loaded plugin (for example delay,
3961 If @option{controls} is set to @code{help}, all available controls and
3962 their valid ranges are printed.
3964 @item sample_rate, s
3965 Specify the sample rate, default to 44100. Only used if plugin have
3969 Set the number of samples per channel per each output frame, default
3970 is 1024. Only used if plugin have zero inputs.
3973 Set the minimum duration of the sourced audio. See
3974 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
3975 for the accepted syntax.
3976 Note that the resulting duration may be greater than the specified duration,
3977 as the generated audio is always cut at the end of a complete frame.
3978 If not specified, or the expressed duration is negative, the audio is
3979 supposed to be generated forever.
3980 Only used if plugin have zero inputs.
3983 @subsection Examples
3987 Apply bass enhancer plugin from Calf:
3989 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
3993 Apply vinyl plugin from Calf:
3995 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
3999 Apply bit crusher plugin from ArtyFX:
4001 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4006 Multiband Compress or expand the audio's dynamic range.
4008 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4009 This is akin to the crossover of a loudspeaker, and results in flat frequency
4010 response when absent compander action.
4012 It accepts the following parameters:
4016 This option syntax is:
4017 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4018 For explanation of each item refer to compand filter documentation.
4024 Mix channels with specific gain levels. The filter accepts the output
4025 channel layout followed by a set of channels definitions.
4027 This filter is also designed to efficiently remap the channels of an audio
4030 The filter accepts parameters of the form:
4031 "@var{l}|@var{outdef}|@var{outdef}|..."
4035 output channel layout or number of channels
4038 output channel specification, of the form:
4039 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4042 output channel to define, either a channel name (FL, FR, etc.) or a channel
4043 number (c0, c1, etc.)
4046 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4049 input channel to use, see out_name for details; it is not possible to mix
4050 named and numbered input channels
4053 If the `=' in a channel specification is replaced by `<', then the gains for
4054 that specification will be renormalized so that the total is 1, thus
4055 avoiding clipping noise.
4057 @subsection Mixing examples
4059 For example, if you want to down-mix from stereo to mono, but with a bigger
4060 factor for the left channel:
4062 pan=1c|c0=0.9*c0+0.1*c1
4065 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4066 7-channels surround:
4068 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4071 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4072 that should be preferred (see "-ac" option) unless you have very specific
4075 @subsection Remapping examples
4077 The channel remapping will be effective if, and only if:
4080 @item gain coefficients are zeroes or ones,
4081 @item only one input per channel output,
4084 If all these conditions are satisfied, the filter will notify the user ("Pure
4085 channel mapping detected"), and use an optimized and lossless method to do the
4088 For example, if you have a 5.1 source and want a stereo audio stream by
4089 dropping the extra channels:
4091 pan="stereo| c0=FL | c1=FR"
4094 Given the same source, you can also switch front left and front right channels
4095 and keep the input channel layout:
4097 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4100 If the input is a stereo audio stream, you can mute the front left channel (and
4101 still keep the stereo channel layout) with:
4106 Still with a stereo audio stream input, you can copy the right channel in both
4107 front left and right:
4109 pan="stereo| c0=FR | c1=FR"
4114 ReplayGain scanner filter. This filter takes an audio stream as an input and
4115 outputs it unchanged.
4116 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4120 Convert the audio sample format, sample rate and channel layout. It is
4121 not meant to be used directly.
4124 Apply time-stretching and pitch-shifting with librubberband.
4126 To enable compilation of this filter, you need to configure FFmpeg with
4127 @code{--enable-librubberband}.
4129 The filter accepts the following options:
4133 Set tempo scale factor.
4136 Set pitch scale factor.
4139 Set transients detector.
4140 Possible values are:
4149 Possible values are:
4158 Possible values are:
4165 Set processing window size.
4166 Possible values are:
4175 Possible values are:
4182 Enable formant preservation when shift pitching.
4183 Possible values are:
4191 Possible values are:
4200 Possible values are:
4207 @section sidechaincompress
4209 This filter acts like normal compressor but has the ability to compress
4210 detected signal using second input signal.
4211 It needs two input streams and returns one output stream.
4212 First input stream will be processed depending on second stream signal.
4213 The filtered signal then can be filtered with other filters in later stages of
4214 processing. See @ref{pan} and @ref{amerge} filter.
4216 The filter accepts the following options:
4220 Set input gain. Default is 1. Range is between 0.015625 and 64.
4223 If a signal of second stream raises above this level it will affect the gain
4224 reduction of first stream.
4225 By default is 0.125. Range is between 0.00097563 and 1.
4228 Set a ratio about which the signal is reduced. 1:2 means that if the level
4229 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4230 Default is 2. Range is between 1 and 20.
4233 Amount of milliseconds the signal has to rise above the threshold before gain
4234 reduction starts. Default is 20. Range is between 0.01 and 2000.
4237 Amount of milliseconds the signal has to fall below the threshold before
4238 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4241 Set the amount by how much signal will be amplified after processing.
4242 Default is 1. Range is from 1 to 64.
4245 Curve the sharp knee around the threshold to enter gain reduction more softly.
4246 Default is 2.82843. Range is between 1 and 8.
4249 Choose if the @code{average} level between all channels of side-chain stream
4250 or the louder(@code{maximum}) channel of side-chain stream affects the
4251 reduction. Default is @code{average}.
4254 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4255 of @code{rms}. Default is @code{rms} which is mainly smoother.
4258 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4261 How much to use compressed signal in output. Default is 1.
4262 Range is between 0 and 1.
4265 @subsection Examples
4269 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
4270 depending on the signal of 2nd input and later compressed signal to be
4271 merged with 2nd input:
4273 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
4277 @section sidechaingate
4279 A sidechain gate acts like a normal (wideband) gate but has the ability to
4280 filter the detected signal before sending it to the gain reduction stage.
4281 Normally a gate uses the full range signal to detect a level above the
4283 For example: If you cut all lower frequencies from your sidechain signal
4284 the gate will decrease the volume of your track only if not enough highs
4285 appear. With this technique you are able to reduce the resonation of a
4286 natural drum or remove "rumbling" of muted strokes from a heavily distorted
4288 It needs two input streams and returns one output stream.
4289 First input stream will be processed depending on second stream signal.
4291 The filter accepts the following options:
4295 Set input level before filtering.
4296 Default is 1. Allowed range is from 0.015625 to 64.
4299 Set the level of gain reduction when the signal is below the threshold.
4300 Default is 0.06125. Allowed range is from 0 to 1.
4303 If a signal rises above this level the gain reduction is released.
4304 Default is 0.125. Allowed range is from 0 to 1.
4307 Set a ratio about which the signal is reduced.
4308 Default is 2. Allowed range is from 1 to 9000.
4311 Amount of milliseconds the signal has to rise above the threshold before gain
4313 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
4316 Amount of milliseconds the signal has to fall below the threshold before the
4317 reduction is increased again. Default is 250 milliseconds.
4318 Allowed range is from 0.01 to 9000.
4321 Set amount of amplification of signal after processing.
4322 Default is 1. Allowed range is from 1 to 64.
4325 Curve the sharp knee around the threshold to enter gain reduction more softly.
4326 Default is 2.828427125. Allowed range is from 1 to 8.
4329 Choose if exact signal should be taken for detection or an RMS like one.
4330 Default is rms. Can be peak or rms.
4333 Choose if the average level between all channels or the louder channel affects
4335 Default is average. Can be average or maximum.
4338 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
4341 @section silencedetect
4343 Detect silence in an audio stream.
4345 This filter logs a message when it detects that the input audio volume is less
4346 or equal to a noise tolerance value for a duration greater or equal to the
4347 minimum detected noise duration.
4349 The printed times and duration are expressed in seconds.
4351 The filter accepts the following options:
4355 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4356 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4359 Set silence duration until notification (default is 2 seconds).
4362 Process each channel separately, instead of combined. By default is disabled.
4365 @subsection Examples
4369 Detect 5 seconds of silence with -50dB noise tolerance:
4371 silencedetect=n=-50dB:d=5
4375 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4376 tolerance in @file{silence.mp3}:
4378 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4382 @section silenceremove
4384 Remove silence from the beginning, middle or end of the audio.
4386 The filter accepts the following options:
4390 This value is used to indicate if audio should be trimmed at beginning of
4391 the audio. A value of zero indicates no silence should be trimmed from the
4392 beginning. When specifying a non-zero value, it trims audio up until it
4393 finds non-silence. Normally, when trimming silence from beginning of audio
4394 the @var{start_periods} will be @code{1} but it can be increased to higher
4395 values to trim all audio up to specific count of non-silence periods.
4396 Default value is @code{0}.
4398 @item start_duration
4399 Specify the amount of time that non-silence must be detected before it stops
4400 trimming audio. By increasing the duration, bursts of noises can be treated
4401 as silence and trimmed off. Default value is @code{0}.
4403 @item start_threshold
4404 This indicates what sample value should be treated as silence. For digital
4405 audio, a value of @code{0} may be fine but for audio recorded from analog,
4406 you may wish to increase the value to account for background noise.
4407 Can be specified in dB (in case "dB" is appended to the specified value)
4408 or amplitude ratio. Default value is @code{0}.
4411 Specify max duration of silence at beginning that will be kept after
4412 trimming. Default is 0, which is equal to trimming all samples detected
4416 Specify mode of detection of silence end in start of multi-channel audio.
4417 Can be @var{any} or @var{all}. Default is @var{any}.
4418 With @var{any}, any sample that is detected as non-silence will cause
4419 stopped trimming of silence.
4420 With @var{all}, only if all channels are detected as non-silence will cause
4421 stopped trimming of silence.
4424 Set the count for trimming silence from the end of audio.
4425 To remove silence from the middle of a file, specify a @var{stop_periods}
4426 that is negative. This value is then treated as a positive value and is
4427 used to indicate the effect should restart processing as specified by
4428 @var{start_periods}, making it suitable for removing periods of silence
4429 in the middle of the audio.
4430 Default value is @code{0}.
4433 Specify a duration of silence that must exist before audio is not copied any
4434 more. By specifying a higher duration, silence that is wanted can be left in
4436 Default value is @code{0}.
4438 @item stop_threshold
4439 This is the same as @option{start_threshold} but for trimming silence from
4441 Can be specified in dB (in case "dB" is appended to the specified value)
4442 or amplitude ratio. Default value is @code{0}.
4445 Specify max duration of silence at end that will be kept after
4446 trimming. Default is 0, which is equal to trimming all samples detected
4450 Specify mode of detection of silence start in end of multi-channel audio.
4451 Can be @var{any} or @var{all}. Default is @var{any}.
4452 With @var{any}, any sample that is detected as non-silence will cause
4453 stopped trimming of silence.
4454 With @var{all}, only if all channels are detected as non-silence will cause
4455 stopped trimming of silence.
4458 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4459 and works better with digital silence which is exactly 0.
4460 Default value is @code{rms}.
4463 Set duration in number of seconds used to calculate size of window in number
4464 of samples for detecting silence.
4465 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4468 @subsection Examples
4472 The following example shows how this filter can be used to start a recording
4473 that does not contain the delay at the start which usually occurs between
4474 pressing the record button and the start of the performance:
4476 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
4480 Trim all silence encountered from beginning to end where there is more than 1
4481 second of silence in audio:
4483 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
4489 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4490 loudspeakers around the user for binaural listening via headphones (audio
4491 formats up to 9 channels supported).
4492 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4493 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4494 Austrian Academy of Sciences.
4496 To enable compilation of this filter you need to configure FFmpeg with
4497 @code{--enable-libmysofa}.
4499 The filter accepts the following options:
4503 Set the SOFA file used for rendering.
4506 Set gain applied to audio. Value is in dB. Default is 0.
4509 Set rotation of virtual loudspeakers in deg. Default is 0.
4512 Set elevation of virtual speakers in deg. Default is 0.
4515 Set distance in meters between loudspeakers and the listener with near-field
4516 HRTFs. Default is 1.
4519 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4520 processing audio in time domain which is slow.
4521 @var{freq} is processing audio in frequency domain which is fast.
4522 Default is @var{freq}.
4525 Set custom positions of virtual loudspeakers. Syntax for this option is:
4526 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
4527 Each virtual loudspeaker is described with short channel name following with
4528 azimuth and elevation in degrees.
4529 Each virtual loudspeaker description is separated by '|'.
4530 For example to override front left and front right channel positions use:
4531 'speakers=FL 45 15|FR 345 15'.
4532 Descriptions with unrecognised channel names are ignored.
4535 Set custom gain for LFE channels. Value is in dB. Default is 0.
4538 Set custom frame size in number of samples. Default is 1024.
4539 Allowed range is from 1024 to 96000. Only used if option @samp{type}
4540 is set to @var{freq}.
4543 Should all IRs be normalized upon importing SOFA file.
4544 By default is enabled.
4547 Should nearest IRs be interpolated with neighbor IRs if exact position
4548 does not match. By default is disabled.
4551 Minphase all IRs upon loading of SOFA file. By default is disabled.
4554 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
4557 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
4560 @subsection Examples
4564 Using ClubFritz6 sofa file:
4566 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
4570 Using ClubFritz12 sofa file and bigger radius with small rotation:
4572 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
4576 Similar as above but with custom speaker positions for front left, front right, back left and back right
4577 and also with custom gain:
4579 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
4583 @section stereotools
4585 This filter has some handy utilities to manage stereo signals, for converting
4586 M/S stereo recordings to L/R signal while having control over the parameters
4587 or spreading the stereo image of master track.
4589 The filter accepts the following options:
4593 Set input level before filtering for both channels. Defaults is 1.
4594 Allowed range is from 0.015625 to 64.
4597 Set output level after filtering for both channels. Defaults is 1.
4598 Allowed range is from 0.015625 to 64.
4601 Set input balance between both channels. Default is 0.
4602 Allowed range is from -1 to 1.
4605 Set output balance between both channels. Default is 0.
4606 Allowed range is from -1 to 1.
4609 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
4610 clipping. Disabled by default.
4613 Mute the left channel. Disabled by default.
4616 Mute the right channel. Disabled by default.
4619 Change the phase of the left channel. Disabled by default.
4622 Change the phase of the right channel. Disabled by default.
4625 Set stereo mode. Available values are:
4629 Left/Right to Left/Right, this is default.
4632 Left/Right to Mid/Side.
4635 Mid/Side to Left/Right.
4638 Left/Right to Left/Left.
4641 Left/Right to Right/Right.
4644 Left/Right to Left + Right.
4647 Left/Right to Right/Left.
4650 Mid/Side to Left/Left.
4653 Mid/Side to Right/Right.
4657 Set level of side signal. Default is 1.
4658 Allowed range is from 0.015625 to 64.
4661 Set balance of side signal. Default is 0.
4662 Allowed range is from -1 to 1.
4665 Set level of the middle signal. Default is 1.
4666 Allowed range is from 0.015625 to 64.
4669 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
4672 Set stereo base between mono and inversed channels. Default is 0.
4673 Allowed range is from -1 to 1.
4676 Set delay in milliseconds how much to delay left from right channel and
4677 vice versa. Default is 0. Allowed range is from -20 to 20.
4680 Set S/C level. Default is 1. Allowed range is from 1 to 100.
4683 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
4685 @item bmode_in, bmode_out
4686 Set balance mode for balance_in/balance_out option.
4688 Can be one of the following:
4692 Classic balance mode. Attenuate one channel at time.
4693 Gain is raised up to 1.
4696 Similar as classic mode above but gain is raised up to 2.
4699 Equal power distribution, from -6dB to +6dB range.
4703 @subsection Examples
4707 Apply karaoke like effect:
4709 stereotools=mlev=0.015625
4713 Convert M/S signal to L/R:
4715 "stereotools=mode=ms>lr"
4719 @section stereowiden
4721 This filter enhance the stereo effect by suppressing signal common to both
4722 channels and by delaying the signal of left into right and vice versa,
4723 thereby widening the stereo effect.
4725 The filter accepts the following options:
4729 Time in milliseconds of the delay of left signal into right and vice versa.
4730 Default is 20 milliseconds.
4733 Amount of gain in delayed signal into right and vice versa. Gives a delay
4734 effect of left signal in right output and vice versa which gives widening
4735 effect. Default is 0.3.
4738 Cross feed of left into right with inverted phase. This helps in suppressing
4739 the mono. If the value is 1 it will cancel all the signal common to both
4740 channels. Default is 0.3.
4743 Set level of input signal of original channel. Default is 0.8.
4746 @section superequalizer
4747 Apply 18 band equalizer.
4749 The filter accepts the following options:
4756 Set 131Hz band gain.
4758 Set 185Hz band gain.
4760 Set 262Hz band gain.
4762 Set 370Hz band gain.
4764 Set 523Hz band gain.
4766 Set 740Hz band gain.
4768 Set 1047Hz band gain.
4770 Set 1480Hz band gain.
4772 Set 2093Hz band gain.
4774 Set 2960Hz band gain.
4776 Set 4186Hz band gain.
4778 Set 5920Hz band gain.
4780 Set 8372Hz band gain.
4782 Set 11840Hz band gain.
4784 Set 16744Hz band gain.
4786 Set 20000Hz band gain.
4790 Apply audio surround upmix filter.
4792 This filter allows to produce multichannel output from audio stream.
4794 The filter accepts the following options:
4798 Set output channel layout. By default, this is @var{5.1}.
4800 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4801 for the required syntax.
4804 Set input channel layout. By default, this is @var{stereo}.
4806 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4807 for the required syntax.
4810 Set input volume level. By default, this is @var{1}.
4813 Set output volume level. By default, this is @var{1}.
4816 Enable LFE channel output if output channel layout has it. By default, this is enabled.
4819 Set LFE low cut off frequency. By default, this is @var{128} Hz.
4822 Set LFE high cut off frequency. By default, this is @var{256} Hz.
4825 Set front center input volume. By default, this is @var{1}.
4828 Set front center output volume. By default, this is @var{1}.
4831 Set LFE input volume. By default, this is @var{1}.
4834 Set LFE output volume. By default, this is @var{1}.
4837 @section treble, highshelf
4839 Boost or cut treble (upper) frequencies of the audio using a two-pole
4840 shelving filter with a response similar to that of a standard
4841 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
4843 The filter accepts the following options:
4847 Give the gain at whichever is the lower of ~22 kHz and the
4848 Nyquist frequency. Its useful range is about -20 (for a large cut)
4849 to +20 (for a large boost). Beware of clipping when using a positive gain.
4852 Set the filter's central frequency and so can be used
4853 to extend or reduce the frequency range to be boosted or cut.
4854 The default value is @code{3000} Hz.
4857 Set method to specify band-width of filter.
4872 Determine how steep is the filter's shelf transition.
4875 Specify which channels to filter, by default all available are filtered.
4878 @subsection Commands
4880 This filter supports the following commands:
4883 Change treble frequency.
4884 Syntax for the command is : "@var{frequency}"
4887 Change treble width_type.
4888 Syntax for the command is : "@var{width_type}"
4891 Change treble width.
4892 Syntax for the command is : "@var{width}"
4896 Syntax for the command is : "@var{gain}"
4901 Sinusoidal amplitude modulation.
4903 The filter accepts the following options:
4907 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
4908 (20 Hz or lower) will result in a tremolo effect.
4909 This filter may also be used as a ring modulator by specifying
4910 a modulation frequency higher than 20 Hz.
4911 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
4914 Depth of modulation as a percentage. Range is 0.0 - 1.0.
4915 Default value is 0.5.
4920 Sinusoidal phase modulation.
4922 The filter accepts the following options:
4926 Modulation frequency in Hertz.
4927 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
4930 Depth of modulation as a percentage. Range is 0.0 - 1.0.
4931 Default value is 0.5.
4936 Adjust the input audio volume.
4938 It accepts the following parameters:
4942 Set audio volume expression.
4944 Output values are clipped to the maximum value.
4946 The output audio volume is given by the relation:
4948 @var{output_volume} = @var{volume} * @var{input_volume}
4951 The default value for @var{volume} is "1.0".
4954 This parameter represents the mathematical precision.
4956 It determines which input sample formats will be allowed, which affects the
4957 precision of the volume scaling.
4961 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
4963 32-bit floating-point; this limits input sample format to FLT. (default)
4965 64-bit floating-point; this limits input sample format to DBL.
4969 Choose the behaviour on encountering ReplayGain side data in input frames.
4973 Remove ReplayGain side data, ignoring its contents (the default).
4976 Ignore ReplayGain side data, but leave it in the frame.
4979 Prefer the track gain, if present.
4982 Prefer the album gain, if present.
4985 @item replaygain_preamp
4986 Pre-amplification gain in dB to apply to the selected replaygain gain.
4988 Default value for @var{replaygain_preamp} is 0.0.
4991 Set when the volume expression is evaluated.
4993 It accepts the following values:
4996 only evaluate expression once during the filter initialization, or
4997 when the @samp{volume} command is sent
5000 evaluate expression for each incoming frame
5003 Default value is @samp{once}.
5006 The volume expression can contain the following parameters.
5010 frame number (starting at zero)
5013 @item nb_consumed_samples
5014 number of samples consumed by the filter
5016 number of samples in the current frame
5018 original frame position in the file
5024 PTS at start of stream
5026 time at start of stream
5032 last set volume value
5035 Note that when @option{eval} is set to @samp{once} only the
5036 @var{sample_rate} and @var{tb} variables are available, all other
5037 variables will evaluate to NAN.
5039 @subsection Commands
5041 This filter supports the following commands:
5044 Modify the volume expression.
5045 The command accepts the same syntax of the corresponding option.
5047 If the specified expression is not valid, it is kept at its current
5049 @item replaygain_noclip
5050 Prevent clipping by limiting the gain applied.
5052 Default value for @var{replaygain_noclip} is 1.
5056 @subsection Examples
5060 Halve the input audio volume:
5064 volume=volume=-6.0206dB
5067 In all the above example the named key for @option{volume} can be
5068 omitted, for example like in:
5074 Increase input audio power by 6 decibels using fixed-point precision:
5076 volume=volume=6dB:precision=fixed
5080 Fade volume after time 10 with an annihilation period of 5 seconds:
5082 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
5086 @section volumedetect
5088 Detect the volume of the input video.
5090 The filter has no parameters. The input is not modified. Statistics about
5091 the volume will be printed in the log when the input stream end is reached.
5093 In particular it will show the mean volume (root mean square), maximum
5094 volume (on a per-sample basis), and the beginning of a histogram of the
5095 registered volume values (from the maximum value to a cumulated 1/1000 of
5098 All volumes are in decibels relative to the maximum PCM value.
5100 @subsection Examples
5102 Here is an excerpt of the output:
5104 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
5105 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
5106 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
5107 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5108 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5109 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5110 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5111 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5112 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5118 The mean square energy is approximately -27 dB, or 10^-2.7.
5120 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5122 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5125 In other words, raising the volume by +4 dB does not cause any clipping,
5126 raising it by +5 dB causes clipping for 6 samples, etc.
5128 @c man end AUDIO FILTERS
5130 @chapter Audio Sources
5131 @c man begin AUDIO SOURCES
5133 Below is a description of the currently available audio sources.
5137 Buffer audio frames, and make them available to the filter chain.
5139 This source is mainly intended for a programmatic use, in particular
5140 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
5142 It accepts the following parameters:
5146 The timebase which will be used for timestamps of submitted frames. It must be
5147 either a floating-point number or in @var{numerator}/@var{denominator} form.
5150 The sample rate of the incoming audio buffers.
5153 The sample format of the incoming audio buffers.
5154 Either a sample format name or its corresponding integer representation from
5155 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5157 @item channel_layout
5158 The channel layout of the incoming audio buffers.
5159 Either a channel layout name from channel_layout_map in
5160 @file{libavutil/channel_layout.c} or its corresponding integer representation
5161 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5164 The number of channels of the incoming audio buffers.
5165 If both @var{channels} and @var{channel_layout} are specified, then they
5170 @subsection Examples
5173 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5176 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5177 Since the sample format with name "s16p" corresponds to the number
5178 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5181 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5186 Generate an audio signal specified by an expression.
5188 This source accepts in input one or more expressions (one for each
5189 channel), which are evaluated and used to generate a corresponding
5192 This source accepts the following options:
5196 Set the '|'-separated expressions list for each separate channel. In case the
5197 @option{channel_layout} option is not specified, the selected channel layout
5198 depends on the number of provided expressions. Otherwise the last
5199 specified expression is applied to the remaining output channels.
5201 @item channel_layout, c
5202 Set the channel layout. The number of channels in the specified layout
5203 must be equal to the number of specified expressions.
5206 Set the minimum duration of the sourced audio. See
5207 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5208 for the accepted syntax.
5209 Note that the resulting duration may be greater than the specified
5210 duration, as the generated audio is always cut at the end of a
5213 If not specified, or the expressed duration is negative, the audio is
5214 supposed to be generated forever.
5217 Set the number of samples per channel per each output frame,
5220 @item sample_rate, s
5221 Specify the sample rate, default to 44100.
5224 Each expression in @var{exprs} can contain the following constants:
5228 number of the evaluated sample, starting from 0
5231 time of the evaluated sample expressed in seconds, starting from 0
5238 @subsection Examples
5248 Generate a sin signal with frequency of 440 Hz, set sample rate to
5251 aevalsrc="sin(440*2*PI*t):s=8000"
5255 Generate a two channels signal, specify the channel layout (Front
5256 Center + Back Center) explicitly:
5258 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5262 Generate white noise:
5264 aevalsrc="-2+random(0)"
5268 Generate an amplitude modulated signal:
5270 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
5274 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
5276 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
5283 The null audio source, return unprocessed audio frames. It is mainly useful
5284 as a template and to be employed in analysis / debugging tools, or as
5285 the source for filters which ignore the input data (for example the sox
5288 This source accepts the following options:
5292 @item channel_layout, cl
5294 Specifies the channel layout, and can be either an integer or a string
5295 representing a channel layout. The default value of @var{channel_layout}
5298 Check the channel_layout_map definition in
5299 @file{libavutil/channel_layout.c} for the mapping between strings and
5300 channel layout values.
5302 @item sample_rate, r
5303 Specifies the sample rate, and defaults to 44100.
5306 Set the number of samples per requested frames.
5310 @subsection Examples
5314 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
5316 anullsrc=r=48000:cl=4
5320 Do the same operation with a more obvious syntax:
5322 anullsrc=r=48000:cl=mono
5326 All the parameters need to be explicitly defined.
5330 Synthesize a voice utterance using the libflite library.
5332 To enable compilation of this filter you need to configure FFmpeg with
5333 @code{--enable-libflite}.
5335 Note that versions of the flite library prior to 2.0 are not thread-safe.
5337 The filter accepts the following options:
5342 If set to 1, list the names of the available voices and exit
5343 immediately. Default value is 0.
5346 Set the maximum number of samples per frame. Default value is 512.
5349 Set the filename containing the text to speak.
5352 Set the text to speak.
5355 Set the voice to use for the speech synthesis. Default value is
5356 @code{kal}. See also the @var{list_voices} option.
5359 @subsection Examples
5363 Read from file @file{speech.txt}, and synthesize the text using the
5364 standard flite voice:
5366 flite=textfile=speech.txt
5370 Read the specified text selecting the @code{slt} voice:
5372 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5376 Input text to ffmpeg:
5378 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5382 Make @file{ffplay} speak the specified text, using @code{flite} and
5383 the @code{lavfi} device:
5385 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
5389 For more information about libflite, check:
5390 @url{http://www.festvox.org/flite/}
5394 Generate a noise audio signal.
5396 The filter accepts the following options:
5399 @item sample_rate, r
5400 Specify the sample rate. Default value is 48000 Hz.
5403 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
5407 Specify the duration of the generated audio stream. Not specifying this option
5408 results in noise with an infinite length.
5410 @item color, colour, c
5411 Specify the color of noise. Available noise colors are white, pink, brown,
5412 blue and violet. Default color is white.
5415 Specify a value used to seed the PRNG.
5418 Set the number of samples per each output frame, default is 1024.
5421 @subsection Examples
5426 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
5428 anoisesrc=d=60:c=pink:r=44100:a=0.5
5434 Generate odd-tap Hilbert transform FIR coefficients.
5436 The resulting stream can be used with @ref{afir} filter for phase-shifting
5437 the signal by 90 degrees.
5439 This is used in many matrix coding schemes and for analytic signal generation.
5440 The process is often written as a multiplication by i (or j), the imaginary unit.
5442 The filter accepts the following options:
5446 @item sample_rate, s
5447 Set sample rate, default is 44100.
5450 Set length of FIR filter, default is 22051.
5453 Set number of samples per each frame.
5456 Set window function to be used when generating FIR coefficients.
5461 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
5463 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
5465 The filter accepts the following options:
5468 @item sample_rate, r
5469 Set sample rate, default is 44100.
5472 Set number of samples per each frame. Default is 1024.
5475 Set high-pass frequency. Default is 0.
5478 Set low-pass frequency. Default is 0.
5479 If high-pass frequency is lower than low-pass frequency and low-pass frequency
5480 is higher than 0 then filter will create band-pass filter coefficients,
5481 otherwise band-reject filter coefficients.
5484 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
5487 Set Kaiser window beta.
5490 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
5493 Enable rounding, by default is disabled.
5496 Set number of taps for high-pass filter.
5499 Set number of taps for low-pass filter.
5504 Generate an audio signal made of a sine wave with amplitude 1/8.
5506 The audio signal is bit-exact.
5508 The filter accepts the following options:
5513 Set the carrier frequency. Default is 440 Hz.
5515 @item beep_factor, b
5516 Enable a periodic beep every second with frequency @var{beep_factor} times
5517 the carrier frequency. Default is 0, meaning the beep is disabled.
5519 @item sample_rate, r
5520 Specify the sample rate, default is 44100.
5523 Specify the duration of the generated audio stream.
5525 @item samples_per_frame
5526 Set the number of samples per output frame.
5528 The expression can contain the following constants:
5532 The (sequential) number of the output audio frame, starting from 0.
5535 The PTS (Presentation TimeStamp) of the output audio frame,
5536 expressed in @var{TB} units.
5539 The PTS of the output audio frame, expressed in seconds.
5542 The timebase of the output audio frames.
5545 Default is @code{1024}.
5548 @subsection Examples
5553 Generate a simple 440 Hz sine wave:
5559 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
5563 sine=frequency=220:beep_factor=4:duration=5
5567 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
5570 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
5574 @c man end AUDIO SOURCES
5576 @chapter Audio Sinks
5577 @c man begin AUDIO SINKS
5579 Below is a description of the currently available audio sinks.
5581 @section abuffersink
5583 Buffer audio frames, and make them available to the end of filter chain.
5585 This sink is mainly intended for programmatic use, in particular
5586 through the interface defined in @file{libavfilter/buffersink.h}
5587 or the options system.
5589 It accepts a pointer to an AVABufferSinkContext structure, which
5590 defines the incoming buffers' formats, to be passed as the opaque
5591 parameter to @code{avfilter_init_filter} for initialization.
5594 Null audio sink; do absolutely nothing with the input audio. It is
5595 mainly useful as a template and for use in analysis / debugging
5598 @c man end AUDIO SINKS
5600 @chapter Video Filters
5601 @c man begin VIDEO FILTERS
5603 When you configure your FFmpeg build, you can disable any of the
5604 existing filters using @code{--disable-filters}.
5605 The configure output will show the video filters included in your
5608 Below is a description of the currently available video filters.
5610 @section alphaextract
5612 Extract the alpha component from the input as a grayscale video. This
5613 is especially useful with the @var{alphamerge} filter.
5617 Add or replace the alpha component of the primary input with the
5618 grayscale value of a second input. This is intended for use with
5619 @var{alphaextract} to allow the transmission or storage of frame
5620 sequences that have alpha in a format that doesn't support an alpha
5623 For example, to reconstruct full frames from a normal YUV-encoded video
5624 and a separate video created with @var{alphaextract}, you might use:
5626 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
5629 Since this filter is designed for reconstruction, it operates on frame
5630 sequences without considering timestamps, and terminates when either
5631 input reaches end of stream. This will cause problems if your encoding
5632 pipeline drops frames. If you're trying to apply an image as an
5633 overlay to a video stream, consider the @var{overlay} filter instead.
5637 Amplify differences between current pixel and pixels of adjacent frames in
5638 same pixel location.
5640 This filter accepts the following options:
5644 Set frame radius. Default is 2. Allowed range is from 1 to 63.
5645 For example radius of 3 will instruct filter to calculate average of 7 frames.
5648 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
5651 Set threshold for difference amplification. Any difference greater or equal to
5652 this value will not alter source pixel. Default is 10.
5653 Allowed range is from 0 to 65535.
5656 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
5657 This option controls maximum possible value that will decrease source pixel value.
5660 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
5661 This option controls maximum possible value that will increase source pixel value.
5664 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
5669 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
5670 and libavformat to work. On the other hand, it is limited to ASS (Advanced
5671 Substation Alpha) subtitles files.
5673 This filter accepts the following option in addition to the common options from
5674 the @ref{subtitles} filter:
5678 Set the shaping engine
5680 Available values are:
5683 The default libass shaping engine, which is the best available.
5685 Fast, font-agnostic shaper that can do only substitutions
5687 Slower shaper using OpenType for substitutions and positioning
5690 The default is @code{auto}.
5694 Apply an Adaptive Temporal Averaging Denoiser to the video input.
5696 The filter accepts the following options:
5700 Set threshold A for 1st plane. Default is 0.02.
5701 Valid range is 0 to 0.3.
5704 Set threshold B for 1st plane. Default is 0.04.
5705 Valid range is 0 to 5.
5708 Set threshold A for 2nd plane. Default is 0.02.
5709 Valid range is 0 to 0.3.
5712 Set threshold B for 2nd plane. Default is 0.04.
5713 Valid range is 0 to 5.
5716 Set threshold A for 3rd plane. Default is 0.02.
5717 Valid range is 0 to 0.3.
5720 Set threshold B for 3rd plane. Default is 0.04.
5721 Valid range is 0 to 5.
5723 Threshold A is designed to react on abrupt changes in the input signal and
5724 threshold B is designed to react on continuous changes in the input signal.
5727 Set number of frames filter will use for averaging. Default is 9. Must be odd
5728 number in range [5, 129].
5731 Set what planes of frame filter will use for averaging. Default is all.
5736 Apply average blur filter.
5738 The filter accepts the following options:
5742 Set horizontal radius size.
5745 Set which planes to filter. By default all planes are filtered.
5748 Set vertical radius size, if zero it will be same as @code{sizeX}.
5749 Default is @code{0}.
5754 Compute the bounding box for the non-black pixels in the input frame
5757 This filter computes the bounding box containing all the pixels with a
5758 luminance value greater than the minimum allowed value.
5759 The parameters describing the bounding box are printed on the filter
5762 The filter accepts the following option:
5766 Set the minimal luminance value. Default is @code{16}.
5769 @section bitplanenoise
5771 Show and measure bit plane noise.
5773 The filter accepts the following options:
5777 Set which plane to analyze. Default is @code{1}.
5780 Filter out noisy pixels from @code{bitplane} set above.
5781 Default is disabled.
5784 @section blackdetect
5786 Detect video intervals that are (almost) completely black. Can be
5787 useful to detect chapter transitions, commercials, or invalid
5788 recordings. Output lines contains the time for the start, end and
5789 duration of the detected black interval expressed in seconds.
5791 In order to display the output lines, you need to set the loglevel at
5792 least to the AV_LOG_INFO value.
5794 The filter accepts the following options:
5797 @item black_min_duration, d
5798 Set the minimum detected black duration expressed in seconds. It must
5799 be a non-negative floating point number.
5801 Default value is 2.0.
5803 @item picture_black_ratio_th, pic_th
5804 Set the threshold for considering a picture "black".
5805 Express the minimum value for the ratio:
5807 @var{nb_black_pixels} / @var{nb_pixels}
5810 for which a picture is considered black.
5811 Default value is 0.98.
5813 @item pixel_black_th, pix_th
5814 Set the threshold for considering a pixel "black".
5816 The threshold expresses the maximum pixel luminance value for which a
5817 pixel is considered "black". The provided value is scaled according to
5818 the following equation:
5820 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
5823 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
5824 the input video format, the range is [0-255] for YUV full-range
5825 formats and [16-235] for YUV non full-range formats.
5827 Default value is 0.10.
5830 The following example sets the maximum pixel threshold to the minimum
5831 value, and detects only black intervals of 2 or more seconds:
5833 blackdetect=d=2:pix_th=0.00
5838 Detect frames that are (almost) completely black. Can be useful to
5839 detect chapter transitions or commercials. Output lines consist of
5840 the frame number of the detected frame, the percentage of blackness,
5841 the position in the file if known or -1 and the timestamp in seconds.
5843 In order to display the output lines, you need to set the loglevel at
5844 least to the AV_LOG_INFO value.
5846 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
5847 The value represents the percentage of pixels in the picture that
5848 are below the threshold value.
5850 It accepts the following parameters:
5855 The percentage of the pixels that have to be below the threshold; it defaults to
5858 @item threshold, thresh
5859 The threshold below which a pixel value is considered black; it defaults to
5864 @section blend, tblend
5866 Blend two video frames into each other.
5868 The @code{blend} filter takes two input streams and outputs one
5869 stream, the first input is the "top" layer and second input is
5870 "bottom" layer. By default, the output terminates when the longest input terminates.
5872 The @code{tblend} (time blend) filter takes two consecutive frames
5873 from one single stream, and outputs the result obtained by blending
5874 the new frame on top of the old frame.
5876 A description of the accepted options follows.
5884 Set blend mode for specific pixel component or all pixel components in case
5885 of @var{all_mode}. Default value is @code{normal}.
5887 Available values for component modes are:
5929 Set blend opacity for specific pixel component or all pixel components in case
5930 of @var{all_opacity}. Only used in combination with pixel component blend modes.
5937 Set blend expression for specific pixel component or all pixel components in case
5938 of @var{all_expr}. Note that related mode options will be ignored if those are set.
5940 The expressions can use the following variables:
5944 The sequential number of the filtered frame, starting from @code{0}.
5948 the coordinates of the current sample
5952 the width and height of currently filtered plane
5956 Width and height scale for the plane being filtered. It is the
5957 ratio between the dimensions of the current plane to the luma plane,
5958 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
5959 the luma plane and @code{0.5,0.5} for the chroma planes.
5962 Time of the current frame, expressed in seconds.
5965 Value of pixel component at current location for first video frame (top layer).
5968 Value of pixel component at current location for second video frame (bottom layer).
5972 The @code{blend} filter also supports the @ref{framesync} options.
5974 @subsection Examples
5978 Apply transition from bottom layer to top layer in first 10 seconds:
5980 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
5984 Apply linear horizontal transition from top layer to bottom layer:
5986 blend=all_expr='A*(X/W)+B*(1-X/W)'
5990 Apply 1x1 checkerboard effect:
5992 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
5996 Apply uncover left effect:
5998 blend=all_expr='if(gte(N*SW+X,W),A,B)'
6002 Apply uncover down effect:
6004 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
6008 Apply uncover up-left effect:
6010 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
6014 Split diagonally video and shows top and bottom layer on each side:
6016 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
6020 Display differences between the current and the previous frame:
6022 tblend=all_mode=grainextract
6028 Denoise frames using Block-Matching 3D algorithm.
6030 The filter accepts the following options.
6034 Set denoising strength. Default value is 1.
6035 Allowed range is from 0 to 999.9.
6036 The denoising algorithm is very sensitive to sigma, so adjust it
6037 according to the source.
6040 Set local patch size. This sets dimensions in 2D.
6043 Set sliding step for processing blocks. Default value is 4.
6044 Allowed range is from 1 to 64.
6045 Smaller values allows processing more reference blocks and is slower.
6048 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
6049 When set to 1, no block matching is done. Larger values allows more blocks
6051 Allowed range is from 1 to 256.
6054 Set radius for search block matching. Default is 9.
6055 Allowed range is from 1 to INT32_MAX.
6058 Set step between two search locations for block matching. Default is 1.
6059 Allowed range is from 1 to 64. Smaller is slower.
6062 Set threshold of mean square error for block matching. Valid range is 0 to
6066 Set thresholding parameter for hard thresholding in 3D transformed domain.
6067 Larger values results in stronger hard-thresholding filtering in frequency
6071 Set filtering estimation mode. Can be @code{basic} or @code{final}.
6072 Default is @code{basic}.
6075 If enabled, filter will use 2nd stream for block matching.
6076 Default is disabled for @code{basic} value of @var{estim} option,
6077 and always enabled if value of @var{estim} is @code{final}.
6080 Set planes to filter. Default is all available except alpha.
6083 @subsection Examples
6087 Basic filtering with bm3d:
6089 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
6093 Same as above, but filtering only luma:
6095 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
6099 Same as above, but with both estimation modes:
6101 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
6105 Same as above, but prefilter with @ref{nlmeans} filter instead:
6107 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
6113 Apply a boxblur algorithm to the input video.
6115 It accepts the following parameters:
6119 @item luma_radius, lr
6120 @item luma_power, lp
6121 @item chroma_radius, cr
6122 @item chroma_power, cp
6123 @item alpha_radius, ar
6124 @item alpha_power, ap
6128 A description of the accepted options follows.
6131 @item luma_radius, lr
6132 @item chroma_radius, cr
6133 @item alpha_radius, ar
6134 Set an expression for the box radius in pixels used for blurring the
6135 corresponding input plane.
6137 The radius value must be a non-negative number, and must not be
6138 greater than the value of the expression @code{min(w,h)/2} for the
6139 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
6142 Default value for @option{luma_radius} is "2". If not specified,
6143 @option{chroma_radius} and @option{alpha_radius} default to the
6144 corresponding value set for @option{luma_radius}.
6146 The expressions can contain the following constants:
6150 The input width and height in pixels.
6154 The input chroma image width and height in pixels.
6158 The horizontal and vertical chroma subsample values. For example, for the
6159 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
6162 @item luma_power, lp
6163 @item chroma_power, cp
6164 @item alpha_power, ap
6165 Specify how many times the boxblur filter is applied to the
6166 corresponding plane.
6168 Default value for @option{luma_power} is 2. If not specified,
6169 @option{chroma_power} and @option{alpha_power} default to the
6170 corresponding value set for @option{luma_power}.
6172 A value of 0 will disable the effect.
6175 @subsection Examples
6179 Apply a boxblur filter with the luma, chroma, and alpha radii
6182 boxblur=luma_radius=2:luma_power=1
6187 Set the luma radius to 2, and alpha and chroma radius to 0:
6189 boxblur=2:1:cr=0:ar=0
6193 Set the luma and chroma radii to a fraction of the video dimension:
6195 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
6201 Deinterlace the input video ("bwdif" stands for "Bob Weaver
6202 Deinterlacing Filter").
6204 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
6205 interpolation algorithms.
6206 It accepts the following parameters:
6210 The interlacing mode to adopt. It accepts one of the following values:
6214 Output one frame for each frame.
6216 Output one frame for each field.
6219 The default value is @code{send_field}.
6222 The picture field parity assumed for the input interlaced video. It accepts one
6223 of the following values:
6227 Assume the top field is first.
6229 Assume the bottom field is first.
6231 Enable automatic detection of field parity.
6234 The default value is @code{auto}.
6235 If the interlacing is unknown or the decoder does not export this information,
6236 top field first will be assumed.
6239 Specify which frames to deinterlace. Accept one of the following
6244 Deinterlace all frames.
6246 Only deinterlace frames marked as interlaced.
6249 The default value is @code{all}.
6253 Remove all color information for all colors except for certain one.
6255 The filter accepts the following options:
6259 The color which will not be replaced with neutral chroma.
6262 Similarity percentage with the above color.
6263 0.01 matches only the exact key color, while 1.0 matches everything.
6266 Signals that the color passed is already in YUV instead of RGB.
6268 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6269 This can be used to pass exact YUV values as hexadecimal numbers.
6273 YUV colorspace color/chroma keying.
6275 The filter accepts the following options:
6279 The color which will be replaced with transparency.
6282 Similarity percentage with the key color.
6284 0.01 matches only the exact key color, while 1.0 matches everything.
6289 0.0 makes pixels either fully transparent, or not transparent at all.
6291 Higher values result in semi-transparent pixels, with a higher transparency
6292 the more similar the pixels color is to the key color.
6295 Signals that the color passed is already in YUV instead of RGB.
6297 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6298 This can be used to pass exact YUV values as hexadecimal numbers.
6301 @subsection Examples
6305 Make every green pixel in the input image transparent:
6307 ffmpeg -i input.png -vf chromakey=green out.png
6311 Overlay a greenscreen-video on top of a static black background.
6313 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
6317 @section chromashift
6318 Shift chroma pixels horizontally and/or vertically.
6320 The filter accepts the following options:
6323 Set amount to shift chroma-blue horizontally.
6325 Set amount to shift chroma-blue vertically.
6327 Set amount to shift chroma-red horizontally.
6329 Set amount to shift chroma-red vertically.
6331 Set edge mode, can be @var{smear}, default, or @var{warp}.
6336 Display CIE color diagram with pixels overlaid onto it.
6338 The filter accepts the following options:
6353 @item uhdtv, rec2020
6366 Set what gamuts to draw.
6368 See @code{system} option for available values.
6371 Set ciescope size, by default set to 512.
6374 Set intensity used to map input pixel values to CIE diagram.
6377 Set contrast used to draw tongue colors that are out of active color system gamut.
6380 Correct gamma displayed on scope, by default enabled.
6383 Show white point on CIE diagram, by default disabled.
6386 Set input gamma. Used only with XYZ input color space.
6391 Visualize information exported by some codecs.
6393 Some codecs can export information through frames using side-data or other
6394 means. For example, some MPEG based codecs export motion vectors through the
6395 @var{export_mvs} flag in the codec @option{flags2} option.
6397 The filter accepts the following option:
6401 Set motion vectors to visualize.
6403 Available flags for @var{mv} are:
6407 forward predicted MVs of P-frames
6409 forward predicted MVs of B-frames
6411 backward predicted MVs of B-frames
6415 Display quantization parameters using the chroma planes.
6418 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
6420 Available flags for @var{mv_type} are:
6424 forward predicted MVs
6426 backward predicted MVs
6429 @item frame_type, ft
6430 Set frame type to visualize motion vectors of.
6432 Available flags for @var{frame_type} are:
6436 intra-coded frames (I-frames)
6438 predicted frames (P-frames)
6440 bi-directionally predicted frames (B-frames)
6444 @subsection Examples
6448 Visualize forward predicted MVs of all frames using @command{ffplay}:
6450 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
6454 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
6456 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
6460 @section colorbalance
6461 Modify intensity of primary colors (red, green and blue) of input frames.
6463 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
6464 regions for the red-cyan, green-magenta or blue-yellow balance.
6466 A positive adjustment value shifts the balance towards the primary color, a negative
6467 value towards the complementary color.
6469 The filter accepts the following options:
6475 Adjust red, green and blue shadows (darkest pixels).
6480 Adjust red, green and blue midtones (medium pixels).
6485 Adjust red, green and blue highlights (brightest pixels).
6487 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
6490 @subsection Examples
6494 Add red color cast to shadows:
6501 RGB colorspace color keying.
6503 The filter accepts the following options:
6507 The color which will be replaced with transparency.
6510 Similarity percentage with the key color.
6512 0.01 matches only the exact key color, while 1.0 matches everything.
6517 0.0 makes pixels either fully transparent, or not transparent at all.
6519 Higher values result in semi-transparent pixels, with a higher transparency
6520 the more similar the pixels color is to the key color.
6523 @subsection Examples
6527 Make every green pixel in the input image transparent:
6529 ffmpeg -i input.png -vf colorkey=green out.png
6533 Overlay a greenscreen-video on top of a static background image.
6535 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
6539 @section colorlevels
6541 Adjust video input frames using levels.
6543 The filter accepts the following options:
6550 Adjust red, green, blue and alpha input black point.
6551 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
6557 Adjust red, green, blue and alpha input white point.
6558 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
6560 Input levels are used to lighten highlights (bright tones), darken shadows
6561 (dark tones), change the balance of bright and dark tones.
6567 Adjust red, green, blue and alpha output black point.
6568 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
6574 Adjust red, green, blue and alpha output white point.
6575 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
6577 Output levels allows manual selection of a constrained output level range.
6580 @subsection Examples
6584 Make video output darker:
6586 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
6592 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
6596 Make video output lighter:
6598 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
6602 Increase brightness:
6604 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
6608 @section colorchannelmixer
6610 Adjust video input frames by re-mixing color channels.
6612 This filter modifies a color channel by adding the values associated to
6613 the other channels of the same pixels. For example if the value to
6614 modify is red, the output value will be:
6616 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
6619 The filter accepts the following options:
6626 Adjust contribution of input red, green, blue and alpha channels for output red channel.
6627 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
6633 Adjust contribution of input red, green, blue and alpha channels for output green channel.
6634 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
6640 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
6641 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
6647 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
6648 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
6650 Allowed ranges for options are @code{[-2.0, 2.0]}.
6653 @subsection Examples
6657 Convert source to grayscale:
6659 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
6662 Simulate sepia tones:
6664 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
6668 @section colormatrix
6670 Convert color matrix.
6672 The filter accepts the following options:
6677 Specify the source and destination color matrix. Both values must be
6680 The accepted values are:
6708 For example to convert from BT.601 to SMPTE-240M, use the command:
6710 colormatrix=bt601:smpte240m
6715 Convert colorspace, transfer characteristics or color primaries.
6716 Input video needs to have an even size.
6718 The filter accepts the following options:
6723 Specify all color properties at once.
6725 The accepted values are:
6755 Specify output colorspace.
6757 The accepted values are:
6766 BT.470BG or BT.601-6 625
6769 SMPTE-170M or BT.601-6 525
6778 BT.2020 with non-constant luminance
6784 Specify output transfer characteristics.
6786 The accepted values are:
6798 Constant gamma of 2.2
6801 Constant gamma of 2.8
6804 SMPTE-170M, BT.601-6 625 or BT.601-6 525
6822 BT.2020 for 10-bits content
6825 BT.2020 for 12-bits content
6831 Specify output color primaries.
6833 The accepted values are:
6842 BT.470BG or BT.601-6 625
6845 SMPTE-170M or BT.601-6 525
6869 Specify output color range.
6871 The accepted values are:
6874 TV (restricted) range
6877 MPEG (restricted) range
6888 Specify output color format.
6890 The accepted values are:
6893 YUV 4:2:0 planar 8-bits
6896 YUV 4:2:0 planar 10-bits
6899 YUV 4:2:0 planar 12-bits
6902 YUV 4:2:2 planar 8-bits
6905 YUV 4:2:2 planar 10-bits
6908 YUV 4:2:2 planar 12-bits
6911 YUV 4:4:4 planar 8-bits
6914 YUV 4:4:4 planar 10-bits
6917 YUV 4:4:4 planar 12-bits
6922 Do a fast conversion, which skips gamma/primary correction. This will take
6923 significantly less CPU, but will be mathematically incorrect. To get output
6924 compatible with that produced by the colormatrix filter, use fast=1.
6927 Specify dithering mode.
6929 The accepted values are:
6935 Floyd-Steinberg dithering
6939 Whitepoint adaptation mode.
6941 The accepted values are:
6944 Bradford whitepoint adaptation
6947 von Kries whitepoint adaptation
6950 identity whitepoint adaptation (i.e. no whitepoint adaptation)
6954 Override all input properties at once. Same accepted values as @ref{all}.
6957 Override input colorspace. Same accepted values as @ref{space}.
6960 Override input color primaries. Same accepted values as @ref{primaries}.
6963 Override input transfer characteristics. Same accepted values as @ref{trc}.
6966 Override input color range. Same accepted values as @ref{range}.
6970 The filter converts the transfer characteristics, color space and color
6971 primaries to the specified user values. The output value, if not specified,
6972 is set to a default value based on the "all" property. If that property is
6973 also not specified, the filter will log an error. The output color range and
6974 format default to the same value as the input color range and format. The
6975 input transfer characteristics, color space, color primaries and color range
6976 should be set on the input data. If any of these are missing, the filter will
6977 log an error and no conversion will take place.
6979 For example to convert the input to SMPTE-240M, use the command:
6981 colorspace=smpte240m
6984 @section convolution
6986 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
6988 The filter accepts the following options:
6995 Set matrix for each plane.
6996 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
6997 and from 1 to 49 odd number of signed integers in @var{row} mode.
7003 Set multiplier for calculated value for each plane.
7004 If unset or 0, it will be sum of all matrix elements.
7010 Set bias for each plane. This value is added to the result of the multiplication.
7011 Useful for making the overall image brighter or darker. Default is 0.0.
7017 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
7018 Default is @var{square}.
7021 @subsection Examples
7027 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"
7033 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"
7039 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"
7045 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"
7049 Apply laplacian edge detector which includes diagonals:
7051 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"
7057 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"
7063 Apply 2D convolution of video stream in frequency domain using second stream
7066 The filter accepts the following options:
7070 Set which planes to process.
7073 Set which impulse video frames will be processed, can be @var{first}
7074 or @var{all}. Default is @var{all}.
7077 The @code{convolve} filter also supports the @ref{framesync} options.
7081 Copy the input video source unchanged to the output. This is mainly useful for
7086 Video filtering on GPU using Apple's CoreImage API on OSX.
7088 Hardware acceleration is based on an OpenGL context. Usually, this means it is
7089 processed by video hardware. However, software-based OpenGL implementations
7090 exist which means there is no guarantee for hardware processing. It depends on
7093 There are many filters and image generators provided by Apple that come with a
7094 large variety of options. The filter has to be referenced by its name along
7097 The coreimage filter accepts the following options:
7100 List all available filters and generators along with all their respective
7101 options as well as possible minimum and maximum values along with the default
7108 Specify all filters by their respective name and options.
7109 Use @var{list_filters} to determine all valid filter names and options.
7110 Numerical options are specified by a float value and are automatically clamped
7111 to their respective value range. Vector and color options have to be specified
7112 by a list of space separated float values. Character escaping has to be done.
7113 A special option name @code{default} is available to use default options for a
7116 It is required to specify either @code{default} or at least one of the filter options.
7117 All omitted options are used with their default values.
7118 The syntax of the filter string is as follows:
7120 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
7124 Specify a rectangle where the output of the filter chain is copied into the
7125 input image. It is given by a list of space separated float values:
7127 output_rect=x\ y\ width\ height
7129 If not given, the output rectangle equals the dimensions of the input image.
7130 The output rectangle is automatically cropped at the borders of the input
7131 image. Negative values are valid for each component.
7133 output_rect=25\ 25\ 100\ 100
7137 Several filters can be chained for successive processing without GPU-HOST
7138 transfers allowing for fast processing of complex filter chains.
7139 Currently, only filters with zero (generators) or exactly one (filters) input
7140 image and one output image are supported. Also, transition filters are not yet
7143 Some filters generate output images with additional padding depending on the
7144 respective filter kernel. The padding is automatically removed to ensure the
7145 filter output has the same size as the input image.
7147 For image generators, the size of the output image is determined by the
7148 previous output image of the filter chain or the input image of the whole
7149 filterchain, respectively. The generators do not use the pixel information of
7150 this image to generate their output. However, the generated output is
7151 blended onto this image, resulting in partial or complete coverage of the
7154 The @ref{coreimagesrc} video source can be used for generating input images
7155 which are directly fed into the filter chain. By using it, providing input
7156 images by another video source or an input video is not required.
7158 @subsection Examples
7163 List all filters available:
7165 coreimage=list_filters=true
7169 Use the CIBoxBlur filter with default options to blur an image:
7171 coreimage=filter=CIBoxBlur@@default
7175 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
7176 its center at 100x100 and a radius of 50 pixels:
7178 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
7182 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
7183 given as complete and escaped command-line for Apple's standard bash shell:
7185 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
7191 Crop the input video to given dimensions.
7193 It accepts the following parameters:
7197 The width of the output video. It defaults to @code{iw}.
7198 This expression is evaluated only once during the filter
7199 configuration, or when the @samp{w} or @samp{out_w} command is sent.
7202 The height of the output video. It defaults to @code{ih}.
7203 This expression is evaluated only once during the filter
7204 configuration, or when the @samp{h} or @samp{out_h} command is sent.
7207 The horizontal position, in the input video, of the left edge of the output
7208 video. It defaults to @code{(in_w-out_w)/2}.
7209 This expression is evaluated per-frame.
7212 The vertical position, in the input video, of the top edge of the output video.
7213 It defaults to @code{(in_h-out_h)/2}.
7214 This expression is evaluated per-frame.
7217 If set to 1 will force the output display aspect ratio
7218 to be the same of the input, by changing the output sample aspect
7219 ratio. It defaults to 0.
7222 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
7223 width/height/x/y as specified and will not be rounded to nearest smaller value.
7227 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
7228 expressions containing the following constants:
7233 The computed values for @var{x} and @var{y}. They are evaluated for
7238 The input width and height.
7242 These are the same as @var{in_w} and @var{in_h}.
7246 The output (cropped) width and height.
7250 These are the same as @var{out_w} and @var{out_h}.
7253 same as @var{iw} / @var{ih}
7256 input sample aspect ratio
7259 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
7263 horizontal and vertical chroma subsample values. For example for the
7264 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
7267 The number of the input frame, starting from 0.
7270 the position in the file of the input frame, NAN if unknown
7273 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
7277 The expression for @var{out_w} may depend on the value of @var{out_h},
7278 and the expression for @var{out_h} may depend on @var{out_w}, but they
7279 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
7280 evaluated after @var{out_w} and @var{out_h}.
7282 The @var{x} and @var{y} parameters specify the expressions for the
7283 position of the top-left corner of the output (non-cropped) area. They
7284 are evaluated for each frame. If the evaluated value is not valid, it
7285 is approximated to the nearest valid value.
7287 The expression for @var{x} may depend on @var{y}, and the expression
7288 for @var{y} may depend on @var{x}.
7290 @subsection Examples
7294 Crop area with size 100x100 at position (12,34).
7299 Using named options, the example above becomes:
7301 crop=w=100:h=100:x=12:y=34
7305 Crop the central input area with size 100x100:
7311 Crop the central input area with size 2/3 of the input video:
7313 crop=2/3*in_w:2/3*in_h
7317 Crop the input video central square:
7324 Delimit the rectangle with the top-left corner placed at position
7325 100:100 and the right-bottom corner corresponding to the right-bottom
7326 corner of the input image.
7328 crop=in_w-100:in_h-100:100:100
7332 Crop 10 pixels from the left and right borders, and 20 pixels from
7333 the top and bottom borders
7335 crop=in_w-2*10:in_h-2*20
7339 Keep only the bottom right quarter of the input image:
7341 crop=in_w/2:in_h/2:in_w/2:in_h/2
7345 Crop height for getting Greek harmony:
7347 crop=in_w:1/PHI*in_w
7351 Apply trembling effect:
7353 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)
7357 Apply erratic camera effect depending on timestamp:
7359 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)"
7363 Set x depending on the value of y:
7365 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
7369 @subsection Commands
7371 This filter supports the following commands:
7377 Set width/height of the output video and the horizontal/vertical position
7379 The command accepts the same syntax of the corresponding option.
7381 If the specified expression is not valid, it is kept at its current
7387 Auto-detect the crop size.
7389 It calculates the necessary cropping parameters and prints the
7390 recommended parameters via the logging system. The detected dimensions
7391 correspond to the non-black area of the input video.
7393 It accepts the following parameters:
7398 Set higher black value threshold, which can be optionally specified
7399 from nothing (0) to everything (255 for 8-bit based formats). An intensity
7400 value greater to the set value is considered non-black. It defaults to 24.
7401 You can also specify a value between 0.0 and 1.0 which will be scaled depending
7402 on the bitdepth of the pixel format.
7405 The value which the width/height should be divisible by. It defaults to
7406 16. The offset is automatically adjusted to center the video. Use 2 to
7407 get only even dimensions (needed for 4:2:2 video). 16 is best when
7408 encoding to most video codecs.
7410 @item reset_count, reset
7411 Set the counter that determines after how many frames cropdetect will
7412 reset the previously detected largest video area and start over to
7413 detect the current optimal crop area. Default value is 0.
7415 This can be useful when channel logos distort the video area. 0
7416 indicates 'never reset', and returns the largest area encountered during
7423 Delay video filtering until a given wallclock timestamp. The filter first
7424 passes on @option{preroll} amount of frames, then it buffers at most
7425 @option{buffer} amount of frames and waits for the cue. After reaching the cue
7426 it forwards the buffered frames and also any subsequent frames coming in its
7429 The filter can be used synchronize the output of multiple ffmpeg processes for
7430 realtime output devices like decklink. By putting the delay in the filtering
7431 chain and pre-buffering frames the process can pass on data to output almost
7432 immediately after the target wallclock timestamp is reached.
7434 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
7440 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
7443 The duration of content to pass on as preroll expressed in seconds. Default is 0.
7446 The maximum duration of content to buffer before waiting for the cue expressed
7447 in seconds. Default is 0.
7454 Apply color adjustments using curves.
7456 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
7457 component (red, green and blue) has its values defined by @var{N} key points
7458 tied from each other using a smooth curve. The x-axis represents the pixel
7459 values from the input frame, and the y-axis the new pixel values to be set for
7462 By default, a component curve is defined by the two points @var{(0;0)} and
7463 @var{(1;1)}. This creates a straight line where each original pixel value is
7464 "adjusted" to its own value, which means no change to the image.
7466 The filter allows you to redefine these two points and add some more. A new
7467 curve (using a natural cubic spline interpolation) will be define to pass
7468 smoothly through all these new coordinates. The new defined points needs to be
7469 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
7470 be in the @var{[0;1]} interval. If the computed curves happened to go outside
7471 the vector spaces, the values will be clipped accordingly.
7473 The filter accepts the following options:
7477 Select one of the available color presets. This option can be used in addition
7478 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
7479 options takes priority on the preset values.
7480 Available presets are:
7483 @item color_negative
7486 @item increase_contrast
7488 @item linear_contrast
7489 @item medium_contrast
7491 @item strong_contrast
7494 Default is @code{none}.
7496 Set the master key points. These points will define a second pass mapping. It
7497 is sometimes called a "luminance" or "value" mapping. It can be used with
7498 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
7499 post-processing LUT.
7501 Set the key points for the red component.
7503 Set the key points for the green component.
7505 Set the key points for the blue component.
7507 Set the key points for all components (not including master).
7508 Can be used in addition to the other key points component
7509 options. In this case, the unset component(s) will fallback on this
7510 @option{all} setting.
7512 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
7514 Save Gnuplot script of the curves in specified file.
7517 To avoid some filtergraph syntax conflicts, each key points list need to be
7518 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
7520 @subsection Examples
7524 Increase slightly the middle level of blue:
7526 curves=blue='0/0 0.5/0.58 1/1'
7532 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'
7534 Here we obtain the following coordinates for each components:
7537 @code{(0;0.11) (0.42;0.51) (1;0.95)}
7539 @code{(0;0) (0.50;0.48) (1;1)}
7541 @code{(0;0.22) (0.49;0.44) (1;0.80)}
7545 The previous example can also be achieved with the associated built-in preset:
7547 curves=preset=vintage
7557 Use a Photoshop preset and redefine the points of the green component:
7559 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
7563 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
7564 and @command{gnuplot}:
7566 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
7567 gnuplot -p /tmp/curves.plt
7573 Video data analysis filter.
7575 This filter shows hexadecimal pixel values of part of video.
7577 The filter accepts the following options:
7581 Set output video size.
7584 Set x offset from where to pick pixels.
7587 Set y offset from where to pick pixels.
7590 Set scope mode, can be one of the following:
7593 Draw hexadecimal pixel values with white color on black background.
7596 Draw hexadecimal pixel values with input video pixel color on black
7600 Draw hexadecimal pixel values on color background picked from input video,
7601 the text color is picked in such way so its always visible.
7605 Draw rows and columns numbers on left and top of video.
7608 Set background opacity.
7613 Denoise frames using 2D DCT (frequency domain filtering).
7615 This filter is not designed for real time.
7617 The filter accepts the following options:
7621 Set the noise sigma constant.
7623 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
7624 coefficient (absolute value) below this threshold with be dropped.
7626 If you need a more advanced filtering, see @option{expr}.
7628 Default is @code{0}.
7631 Set number overlapping pixels for each block. Since the filter can be slow, you
7632 may want to reduce this value, at the cost of a less effective filter and the
7633 risk of various artefacts.
7635 If the overlapping value doesn't permit processing the whole input width or
7636 height, a warning will be displayed and according borders won't be denoised.
7638 Default value is @var{blocksize}-1, which is the best possible setting.
7641 Set the coefficient factor expression.
7643 For each coefficient of a DCT block, this expression will be evaluated as a
7644 multiplier value for the coefficient.
7646 If this is option is set, the @option{sigma} option will be ignored.
7648 The absolute value of the coefficient can be accessed through the @var{c}
7652 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
7653 @var{blocksize}, which is the width and height of the processed blocks.
7655 The default value is @var{3} (8x8) and can be raised to @var{4} for a
7656 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
7657 on the speed processing. Also, a larger block size does not necessarily means a
7661 @subsection Examples
7663 Apply a denoise with a @option{sigma} of @code{4.5}:
7668 The same operation can be achieved using the expression system:
7670 dctdnoiz=e='gte(c, 4.5*3)'
7673 Violent denoise using a block size of @code{16x16}:
7680 Remove banding artifacts from input video.
7681 It works by replacing banded pixels with average value of referenced pixels.
7683 The filter accepts the following options:
7690 Set banding detection threshold for each plane. Default is 0.02.
7691 Valid range is 0.00003 to 0.5.
7692 If difference between current pixel and reference pixel is less than threshold,
7693 it will be considered as banded.
7696 Banding detection range in pixels. Default is 16. If positive, random number
7697 in range 0 to set value will be used. If negative, exact absolute value
7699 The range defines square of four pixels around current pixel.
7702 Set direction in radians from which four pixel will be compared. If positive,
7703 random direction from 0 to set direction will be picked. If negative, exact of
7704 absolute value will be picked. For example direction 0, -PI or -2*PI radians
7705 will pick only pixels on same row and -PI/2 will pick only pixels on same
7709 If enabled, current pixel is compared with average value of all four
7710 surrounding pixels. The default is enabled. If disabled current pixel is
7711 compared with all four surrounding pixels. The pixel is considered banded
7712 if only all four differences with surrounding pixels are less than threshold.
7715 If enabled, current pixel is changed if and only if all pixel components are banded,
7716 e.g. banding detection threshold is triggered for all color components.
7717 The default is disabled.
7722 Remove blocking artifacts from input video.
7724 The filter accepts the following options:
7728 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
7729 This controls what kind of deblocking is applied.
7732 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
7738 Set blocking detection thresholds. Allowed range is 0 to 1.
7739 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
7740 Using higher threshold gives more deblocking strength.
7741 Setting @var{alpha} controls threshold detection at exact edge of block.
7742 Remaining options controls threshold detection near the edge. Each one for
7743 below/above or left/right. Setting any of those to @var{0} disables
7747 Set planes to filter. Default is to filter all available planes.
7750 @subsection Examples
7754 Deblock using weak filter and block size of 4 pixels.
7756 deblock=filter=weak:block=4
7760 Deblock using strong filter, block size of 4 pixels and custom thresholds for
7761 deblocking more edges.
7763 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
7767 Similar as above, but filter only first plane.
7769 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
7773 Similar as above, but filter only second and third plane.
7775 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
7782 Drop duplicated frames at regular intervals.
7784 The filter accepts the following options:
7788 Set the number of frames from which one will be dropped. Setting this to
7789 @var{N} means one frame in every batch of @var{N} frames will be dropped.
7790 Default is @code{5}.
7793 Set the threshold for duplicate detection. If the difference metric for a frame
7794 is less than or equal to this value, then it is declared as duplicate. Default
7798 Set scene change threshold. Default is @code{15}.
7802 Set the size of the x and y-axis blocks used during metric calculations.
7803 Larger blocks give better noise suppression, but also give worse detection of
7804 small movements. Must be a power of two. Default is @code{32}.
7807 Mark main input as a pre-processed input and activate clean source input
7808 stream. This allows the input to be pre-processed with various filters to help
7809 the metrics calculation while keeping the frame selection lossless. When set to
7810 @code{1}, the first stream is for the pre-processed input, and the second
7811 stream is the clean source from where the kept frames are chosen. Default is
7815 Set whether or not chroma is considered in the metric calculations. Default is
7821 Apply 2D deconvolution of video stream in frequency domain using second stream
7824 The filter accepts the following options:
7828 Set which planes to process.
7831 Set which impulse video frames will be processed, can be @var{first}
7832 or @var{all}. Default is @var{all}.
7835 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
7836 and height are not same and not power of 2 or if stream prior to convolving
7840 The @code{deconvolve} filter also supports the @ref{framesync} options.
7844 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
7846 It accepts the following options:
7850 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
7851 @var{rainbows} for cross-color reduction.
7854 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
7857 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
7860 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
7863 Set temporal chroma threshold. Lower values increases reduction of cross-color.
7868 Apply deflate effect to the video.
7870 This filter replaces the pixel by the local(3x3) average by taking into account
7871 only values lower than the pixel.
7873 It accepts the following options:
7880 Limit the maximum change for each plane, default is 65535.
7881 If 0, plane will remain unchanged.
7886 Remove temporal frame luminance variations.
7888 It accepts the following options:
7892 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
7895 Set averaging mode to smooth temporal luminance variations.
7897 Available values are:
7922 Do not actually modify frame. Useful when one only wants metadata.
7927 Remove judder produced by partially interlaced telecined content.
7929 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
7930 source was partially telecined content then the output of @code{pullup,dejudder}
7931 will have a variable frame rate. May change the recorded frame rate of the
7932 container. Aside from that change, this filter will not affect constant frame
7935 The option available in this filter is:
7939 Specify the length of the window over which the judder repeats.
7941 Accepts any integer greater than 1. Useful values are:
7945 If the original was telecined from 24 to 30 fps (Film to NTSC).
7948 If the original was telecined from 25 to 30 fps (PAL to NTSC).
7951 If a mixture of the two.
7954 The default is @samp{4}.
7959 Suppress a TV station logo by a simple interpolation of the surrounding
7960 pixels. Just set a rectangle covering the logo and watch it disappear
7961 (and sometimes something even uglier appear - your mileage may vary).
7963 It accepts the following parameters:
7968 Specify the top left corner coordinates of the logo. They must be
7973 Specify the width and height of the logo to clear. They must be
7977 Specify the thickness of the fuzzy edge of the rectangle (added to
7978 @var{w} and @var{h}). The default value is 1. This option is
7979 deprecated, setting higher values should no longer be necessary and
7983 When set to 1, a green rectangle is drawn on the screen to simplify
7984 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
7985 The default value is 0.
7987 The rectangle is drawn on the outermost pixels which will be (partly)
7988 replaced with interpolated values. The values of the next pixels
7989 immediately outside this rectangle in each direction will be used to
7990 compute the interpolated pixel values inside the rectangle.
7994 @subsection Examples
7998 Set a rectangle covering the area with top left corner coordinates 0,0
7999 and size 100x77, and a band of size 10:
8001 delogo=x=0:y=0:w=100:h=77:band=10
8008 Attempt to fix small changes in horizontal and/or vertical shift. This
8009 filter helps remove camera shake from hand-holding a camera, bumping a
8010 tripod, moving on a vehicle, etc.
8012 The filter accepts the following options:
8020 Specify a rectangular area where to limit the search for motion
8022 If desired the search for motion vectors can be limited to a
8023 rectangular area of the frame defined by its top left corner, width
8024 and height. These parameters have the same meaning as the drawbox
8025 filter which can be used to visualise the position of the bounding
8028 This is useful when simultaneous movement of subjects within the frame
8029 might be confused for camera motion by the motion vector search.
8031 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
8032 then the full frame is used. This allows later options to be set
8033 without specifying the bounding box for the motion vector search.
8035 Default - search the whole frame.
8039 Specify the maximum extent of movement in x and y directions in the
8040 range 0-64 pixels. Default 16.
8043 Specify how to generate pixels to fill blanks at the edge of the
8044 frame. Available values are:
8047 Fill zeroes at blank locations
8049 Original image at blank locations
8051 Extruded edge value at blank locations
8053 Mirrored edge at blank locations
8055 Default value is @samp{mirror}.
8058 Specify the blocksize to use for motion search. Range 4-128 pixels,
8062 Specify the contrast threshold for blocks. Only blocks with more than
8063 the specified contrast (difference between darkest and lightest
8064 pixels) will be considered. Range 1-255, default 125.
8067 Specify the search strategy. Available values are:
8070 Set exhaustive search
8072 Set less exhaustive search.
8074 Default value is @samp{exhaustive}.
8077 If set then a detailed log of the motion search is written to the
8084 Remove unwanted contamination of foreground colors, caused by reflected color of
8085 greenscreen or bluescreen.
8087 This filter accepts the following options:
8091 Set what type of despill to use.
8094 Set how spillmap will be generated.
8097 Set how much to get rid of still remaining spill.
8100 Controls amount of red in spill area.
8103 Controls amount of green in spill area.
8104 Should be -1 for greenscreen.
8107 Controls amount of blue in spill area.
8108 Should be -1 for bluescreen.
8111 Controls brightness of spill area, preserving colors.
8114 Modify alpha from generated spillmap.
8119 Apply an exact inverse of the telecine operation. It requires a predefined
8120 pattern specified using the pattern option which must be the same as that passed
8121 to the telecine filter.
8123 This filter accepts the following options:
8132 The default value is @code{top}.
8136 A string of numbers representing the pulldown pattern you wish to apply.
8137 The default value is @code{23}.
8140 A number representing position of the first frame with respect to the telecine
8141 pattern. This is to be used if the stream is cut. The default value is @code{0}.
8146 Apply dilation effect to the video.
8148 This filter replaces the pixel by the local(3x3) maximum.
8150 It accepts the following options:
8157 Limit the maximum change for each plane, default is 65535.
8158 If 0, plane will remain unchanged.
8161 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
8164 Flags to local 3x3 coordinates maps like this:
8173 Displace pixels as indicated by second and third input stream.
8175 It takes three input streams and outputs one stream, the first input is the
8176 source, and second and third input are displacement maps.
8178 The second input specifies how much to displace pixels along the
8179 x-axis, while the third input specifies how much to displace pixels
8181 If one of displacement map streams terminates, last frame from that
8182 displacement map will be used.
8184 Note that once generated, displacements maps can be reused over and over again.
8186 A description of the accepted options follows.
8190 Set displace behavior for pixels that are out of range.
8192 Available values are:
8195 Missing pixels are replaced by black pixels.
8198 Adjacent pixels will spread out to replace missing pixels.
8201 Out of range pixels are wrapped so they point to pixels of other side.
8204 Out of range pixels will be replaced with mirrored pixels.
8206 Default is @samp{smear}.
8210 @subsection Examples
8214 Add ripple effect to rgb input of video size hd720:
8216 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
8220 Add wave effect to rgb input of video size hd720:
8222 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
8228 Draw a colored box on the input image.
8230 It accepts the following parameters:
8235 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
8239 The expressions which specify the width and height of the box; if 0 they are interpreted as
8240 the input width and height. It defaults to 0.
8243 Specify the color of the box to write. For the general syntax of this option,
8244 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8245 value @code{invert} is used, the box edge color is the same as the
8246 video with inverted luma.
8249 The expression which sets the thickness of the box edge.
8250 A value of @code{fill} will create a filled box. Default value is @code{3}.
8252 See below for the list of accepted constants.
8255 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
8256 will overwrite the video's color and alpha pixels.
8257 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
8260 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8261 following constants:
8265 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8269 horizontal and vertical chroma subsample values. For example for the
8270 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8274 The input width and height.
8277 The input sample aspect ratio.
8281 The x and y offset coordinates where the box is drawn.
8285 The width and height of the drawn box.
8288 The thickness of the drawn box.
8290 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8291 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
8295 @subsection Examples
8299 Draw a black box around the edge of the input image:
8305 Draw a box with color red and an opacity of 50%:
8307 drawbox=10:20:200:60:red@@0.5
8310 The previous example can be specified as:
8312 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
8316 Fill the box with pink color:
8318 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
8322 Draw a 2-pixel red 2.40:1 mask:
8324 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
8330 Draw a grid on the input image.
8332 It accepts the following parameters:
8337 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
8341 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
8342 input width and height, respectively, minus @code{thickness}, so image gets
8343 framed. Default to 0.
8346 Specify the color of the grid. For the general syntax of this option,
8347 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
8348 value @code{invert} is used, the grid color is the same as the
8349 video with inverted luma.
8352 The expression which sets the thickness of the grid line. Default value is @code{1}.
8354 See below for the list of accepted constants.
8357 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
8358 will overwrite the video's color and alpha pixels.
8359 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
8362 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
8363 following constants:
8367 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
8371 horizontal and vertical chroma subsample values. For example for the
8372 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8376 The input grid cell width and height.
8379 The input sample aspect ratio.
8383 The x and y coordinates of some point of grid intersection (meant to configure offset).
8387 The width and height of the drawn cell.
8390 The thickness of the drawn cell.
8392 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
8393 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
8397 @subsection Examples
8401 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
8403 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
8407 Draw a white 3x3 grid with an opacity of 50%:
8409 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
8416 Draw a text string or text from a specified file on top of a video, using the
8417 libfreetype library.
8419 To enable compilation of this filter, you need to configure FFmpeg with
8420 @code{--enable-libfreetype}.
8421 To enable default font fallback and the @var{font} option you need to
8422 configure FFmpeg with @code{--enable-libfontconfig}.
8423 To enable the @var{text_shaping} option, you need to configure FFmpeg with
8424 @code{--enable-libfribidi}.
8428 It accepts the following parameters:
8433 Used to draw a box around text using the background color.
8434 The value must be either 1 (enable) or 0 (disable).
8435 The default value of @var{box} is 0.
8438 Set the width of the border to be drawn around the box using @var{boxcolor}.
8439 The default value of @var{boxborderw} is 0.
8442 The color to be used for drawing box around text. For the syntax of this
8443 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8445 The default value of @var{boxcolor} is "white".
8448 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
8449 The default value of @var{line_spacing} is 0.
8452 Set the width of the border to be drawn around the text using @var{bordercolor}.
8453 The default value of @var{borderw} is 0.
8456 Set the color to be used for drawing border around text. For the syntax of this
8457 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8459 The default value of @var{bordercolor} is "black".
8462 Select how the @var{text} is expanded. Can be either @code{none},
8463 @code{strftime} (deprecated) or
8464 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
8468 Set a start time for the count. Value is in microseconds. Only applied
8469 in the deprecated strftime expansion mode. To emulate in normal expansion
8470 mode use the @code{pts} function, supplying the start time (in seconds)
8471 as the second argument.
8474 If true, check and fix text coords to avoid clipping.
8477 The color to be used for drawing fonts. For the syntax of this option, check
8478 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
8480 The default value of @var{fontcolor} is "black".
8482 @item fontcolor_expr
8483 String which is expanded the same way as @var{text} to obtain dynamic
8484 @var{fontcolor} value. By default this option has empty value and is not
8485 processed. When this option is set, it overrides @var{fontcolor} option.
8488 The font family to be used for drawing text. By default Sans.
8491 The font file to be used for drawing text. The path must be included.
8492 This parameter is mandatory if the fontconfig support is disabled.
8495 Draw the text applying alpha blending. The value can
8496 be a number between 0.0 and 1.0.
8497 The expression accepts the same variables @var{x, y} as well.
8498 The default value is 1.
8499 Please see @var{fontcolor_expr}.
8502 The font size to be used for drawing text.
8503 The default value of @var{fontsize} is 16.
8506 If set to 1, attempt to shape the text (for example, reverse the order of
8507 right-to-left text and join Arabic characters) before drawing it.
8508 Otherwise, just draw the text exactly as given.
8509 By default 1 (if supported).
8512 The flags to be used for loading the fonts.
8514 The flags map the corresponding flags supported by libfreetype, and are
8515 a combination of the following values:
8522 @item vertical_layout
8523 @item force_autohint
8526 @item ignore_global_advance_width
8528 @item ignore_transform
8534 Default value is "default".
8536 For more information consult the documentation for the FT_LOAD_*
8540 The color to be used for drawing a shadow behind the drawn text. For the
8541 syntax of this option, check the @ref{color syntax,,"Color" section in the
8542 ffmpeg-utils manual,ffmpeg-utils}.
8544 The default value of @var{shadowcolor} is "black".
8548 The x and y offsets for the text shadow position with respect to the
8549 position of the text. They can be either positive or negative
8550 values. The default value for both is "0".
8553 The starting frame number for the n/frame_num variable. The default value
8557 The size in number of spaces to use for rendering the tab.
8561 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
8562 format. It can be used with or without text parameter. @var{timecode_rate}
8563 option must be specified.
8565 @item timecode_rate, rate, r
8566 Set the timecode frame rate (timecode only). Value will be rounded to nearest
8567 integer. Minimum value is "1".
8568 Drop-frame timecode is supported for frame rates 30 & 60.
8571 If set to 1, the output of the timecode option will wrap around at 24 hours.
8572 Default is 0 (disabled).
8575 The text string to be drawn. The text must be a sequence of UTF-8
8577 This parameter is mandatory if no file is specified with the parameter
8581 A text file containing text to be drawn. The text must be a sequence
8582 of UTF-8 encoded characters.
8584 This parameter is mandatory if no text string is specified with the
8585 parameter @var{text}.
8587 If both @var{text} and @var{textfile} are specified, an error is thrown.
8590 If set to 1, the @var{textfile} will be reloaded before each frame.
8591 Be sure to update it atomically, or it may be read partially, or even fail.
8595 The expressions which specify the offsets where text will be drawn
8596 within the video frame. They are relative to the top/left border of the
8599 The default value of @var{x} and @var{y} is "0".
8601 See below for the list of accepted constants and functions.
8604 The parameters for @var{x} and @var{y} are expressions containing the
8605 following constants and functions:
8609 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
8613 horizontal and vertical chroma subsample values. For example for the
8614 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8617 the height of each text line
8625 @item max_glyph_a, ascent
8626 the maximum distance from the baseline to the highest/upper grid
8627 coordinate used to place a glyph outline point, for all the rendered
8629 It is a positive value, due to the grid's orientation with the Y axis
8632 @item max_glyph_d, descent
8633 the maximum distance from the baseline to the lowest grid coordinate
8634 used to place a glyph outline point, for all the rendered glyphs.
8635 This is a negative value, due to the grid's orientation, with the Y axis
8639 maximum glyph height, that is the maximum height for all the glyphs
8640 contained in the rendered text, it is equivalent to @var{ascent} -
8644 maximum glyph width, that is the maximum width for all the glyphs
8645 contained in the rendered text
8648 the number of input frame, starting from 0
8650 @item rand(min, max)
8651 return a random number included between @var{min} and @var{max}
8654 The input sample aspect ratio.
8657 timestamp expressed in seconds, NAN if the input timestamp is unknown
8660 the height of the rendered text
8663 the width of the rendered text
8667 the x and y offset coordinates where the text is drawn.
8669 These parameters allow the @var{x} and @var{y} expressions to refer
8670 each other, so you can for example specify @code{y=x/dar}.
8673 @anchor{drawtext_expansion}
8674 @subsection Text expansion
8676 If @option{expansion} is set to @code{strftime},
8677 the filter recognizes strftime() sequences in the provided text and
8678 expands them accordingly. Check the documentation of strftime(). This
8679 feature is deprecated.
8681 If @option{expansion} is set to @code{none}, the text is printed verbatim.
8683 If @option{expansion} is set to @code{normal} (which is the default),
8684 the following expansion mechanism is used.
8686 The backslash character @samp{\}, followed by any character, always expands to
8687 the second character.
8689 Sequences of the form @code{%@{...@}} are expanded. The text between the
8690 braces is a function name, possibly followed by arguments separated by ':'.
8691 If the arguments contain special characters or delimiters (':' or '@}'),
8692 they should be escaped.
8694 Note that they probably must also be escaped as the value for the
8695 @option{text} option in the filter argument string and as the filter
8696 argument in the filtergraph description, and possibly also for the shell,
8697 that makes up to four levels of escaping; using a text file avoids these
8700 The following functions are available:
8705 The expression evaluation result.
8707 It must take one argument specifying the expression to be evaluated,
8708 which accepts the same constants and functions as the @var{x} and
8709 @var{y} values. Note that not all constants should be used, for
8710 example the text size is not known when evaluating the expression, so
8711 the constants @var{text_w} and @var{text_h} will have an undefined
8714 @item expr_int_format, eif
8715 Evaluate the expression's value and output as formatted integer.
8717 The first argument is the expression to be evaluated, just as for the @var{expr} function.
8718 The second argument specifies the output format. Allowed values are @samp{x},
8719 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
8720 @code{printf} function.
8721 The third parameter is optional and sets the number of positions taken by the output.
8722 It can be used to add padding with zeros from the left.
8725 The time at which the filter is running, expressed in UTC.
8726 It can accept an argument: a strftime() format string.
8729 The time at which the filter is running, expressed in the local time zone.
8730 It can accept an argument: a strftime() format string.
8733 Frame metadata. Takes one or two arguments.
8735 The first argument is mandatory and specifies the metadata key.
8737 The second argument is optional and specifies a default value, used when the
8738 metadata key is not found or empty.
8741 The frame number, starting from 0.
8744 A 1 character description of the current picture type.
8747 The timestamp of the current frame.
8748 It can take up to three arguments.
8750 The first argument is the format of the timestamp; it defaults to @code{flt}
8751 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
8752 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
8753 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
8754 @code{localtime} stands for the timestamp of the frame formatted as
8755 local time zone time.
8757 The second argument is an offset added to the timestamp.
8759 If the format is set to @code{hms}, a third argument @code{24HH} may be
8760 supplied to present the hour part of the formatted timestamp in 24h format
8763 If the format is set to @code{localtime} or @code{gmtime},
8764 a third argument may be supplied: a strftime() format string.
8765 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
8768 @subsection Examples
8772 Draw "Test Text" with font FreeSerif, using the default values for the
8773 optional parameters.
8776 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
8780 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
8781 and y=50 (counting from the top-left corner of the screen), text is
8782 yellow with a red box around it. Both the text and the box have an
8786 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
8787 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
8790 Note that the double quotes are not necessary if spaces are not used
8791 within the parameter list.
8794 Show the text at the center of the video frame:
8796 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
8800 Show the text at a random position, switching to a new position every 30 seconds:
8802 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)"
8806 Show a text line sliding from right to left in the last row of the video
8807 frame. The file @file{LONG_LINE} is assumed to contain a single line
8810 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
8814 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
8816 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
8820 Draw a single green letter "g", at the center of the input video.
8821 The glyph baseline is placed at half screen height.
8823 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
8827 Show text for 1 second every 3 seconds:
8829 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
8833 Use fontconfig to set the font. Note that the colons need to be escaped.
8835 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
8839 Print the date of a real-time encoding (see strftime(3)):
8841 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
8845 Show text fading in and out (appearing/disappearing):
8848 DS=1.0 # display start
8849 DE=10.0 # display end
8850 FID=1.5 # fade in duration
8851 FOD=5 # fade out duration
8852 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 @}"
8856 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
8857 and the @option{fontsize} value are included in the @option{y} offset.
8859 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
8860 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
8865 For more information about libfreetype, check:
8866 @url{http://www.freetype.org/}.
8868 For more information about fontconfig, check:
8869 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
8871 For more information about libfribidi, check:
8872 @url{http://fribidi.org/}.
8876 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
8878 The filter accepts the following options:
8883 Set low and high threshold values used by the Canny thresholding
8886 The high threshold selects the "strong" edge pixels, which are then
8887 connected through 8-connectivity with the "weak" edge pixels selected
8888 by the low threshold.
8890 @var{low} and @var{high} threshold values must be chosen in the range
8891 [0,1], and @var{low} should be lesser or equal to @var{high}.
8893 Default value for @var{low} is @code{20/255}, and default value for @var{high}
8897 Define the drawing mode.
8901 Draw white/gray wires on black background.
8904 Mix the colors to create a paint/cartoon effect.
8907 Apply Canny edge detector on all selected planes.
8909 Default value is @var{wires}.
8912 Select planes for filtering. By default all available planes are filtered.
8915 @subsection Examples
8919 Standard edge detection with custom values for the hysteresis thresholding:
8921 edgedetect=low=0.1:high=0.4
8925 Painting effect without thresholding:
8927 edgedetect=mode=colormix:high=0
8932 Set brightness, contrast, saturation and approximate gamma adjustment.
8934 The filter accepts the following options:
8938 Set the contrast expression. The value must be a float value in range
8939 @code{-2.0} to @code{2.0}. The default value is "1".
8942 Set the brightness expression. The value must be a float value in
8943 range @code{-1.0} to @code{1.0}. The default value is "0".
8946 Set the saturation expression. The value must be a float in
8947 range @code{0.0} to @code{3.0}. The default value is "1".
8950 Set the gamma expression. The value must be a float in range
8951 @code{0.1} to @code{10.0}. The default value is "1".
8954 Set the gamma expression for red. The value must be a float in
8955 range @code{0.1} to @code{10.0}. The default value is "1".
8958 Set the gamma expression for green. The value must be a float in range
8959 @code{0.1} to @code{10.0}. The default value is "1".
8962 Set the gamma expression for blue. The value must be a float in range
8963 @code{0.1} to @code{10.0}. The default value is "1".
8966 Set the gamma weight expression. It can be used to reduce the effect
8967 of a high gamma value on bright image areas, e.g. keep them from
8968 getting overamplified and just plain white. The value must be a float
8969 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
8970 gamma correction all the way down while @code{1.0} leaves it at its
8971 full strength. Default is "1".
8974 Set when the expressions for brightness, contrast, saturation and
8975 gamma expressions are evaluated.
8977 It accepts the following values:
8980 only evaluate expressions once during the filter initialization or
8981 when a command is processed
8984 evaluate expressions for each incoming frame
8987 Default value is @samp{init}.
8990 The expressions accept the following parameters:
8993 frame count of the input frame starting from 0
8996 byte position of the corresponding packet in the input file, NAN if
9000 frame rate of the input video, NAN if the input frame rate is unknown
9003 timestamp expressed in seconds, NAN if the input timestamp is unknown
9006 @subsection Commands
9007 The filter supports the following commands:
9011 Set the contrast expression.
9014 Set the brightness expression.
9017 Set the saturation expression.
9020 Set the gamma expression.
9023 Set the gamma_r expression.
9026 Set gamma_g expression.
9029 Set gamma_b expression.
9032 Set gamma_weight expression.
9034 The command accepts the same syntax of the corresponding option.
9036 If the specified expression is not valid, it is kept at its current
9043 Apply erosion effect to the video.
9045 This filter replaces the pixel by the local(3x3) minimum.
9047 It accepts the following options:
9054 Limit the maximum change for each plane, default is 65535.
9055 If 0, plane will remain unchanged.
9058 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9061 Flags to local 3x3 coordinates maps like this:
9068 @section extractplanes
9070 Extract color channel components from input video stream into
9071 separate grayscale video streams.
9073 The filter accepts the following option:
9077 Set plane(s) to extract.
9079 Available values for planes are:
9090 Choosing planes not available in the input will result in an error.
9091 That means you cannot select @code{r}, @code{g}, @code{b} planes
9092 with @code{y}, @code{u}, @code{v} planes at same time.
9095 @subsection Examples
9099 Extract luma, u and v color channel component from input video frame
9100 into 3 grayscale outputs:
9102 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
9108 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
9110 For each input image, the filter will compute the optimal mapping from
9111 the input to the output given the codebook length, that is the number
9112 of distinct output colors.
9114 This filter accepts the following options.
9117 @item codebook_length, l
9118 Set codebook length. The value must be a positive integer, and
9119 represents the number of distinct output colors. Default value is 256.
9122 Set the maximum number of iterations to apply for computing the optimal
9123 mapping. The higher the value the better the result and the higher the
9124 computation time. Default value is 1.
9127 Set a random seed, must be an integer included between 0 and
9128 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
9129 will try to use a good random seed on a best effort basis.
9132 Set pal8 output pixel format. This option does not work with codebook
9133 length greater than 256.
9138 Measure graylevel entropy in histogram of color channels of video frames.
9140 It accepts the following parameters:
9144 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
9146 @var{diff} mode measures entropy of histogram delta values, absolute differences
9147 between neighbour histogram values.
9152 Apply a fade-in/out effect to the input video.
9154 It accepts the following parameters:
9158 The effect type can be either "in" for a fade-in, or "out" for a fade-out
9160 Default is @code{in}.
9162 @item start_frame, s
9163 Specify the number of the frame to start applying the fade
9164 effect at. Default is 0.
9167 The number of frames that the fade effect lasts. At the end of the
9168 fade-in effect, the output video will have the same intensity as the input video.
9169 At the end of the fade-out transition, the output video will be filled with the
9170 selected @option{color}.
9174 If set to 1, fade only alpha channel, if one exists on the input.
9177 @item start_time, st
9178 Specify the timestamp (in seconds) of the frame to start to apply the fade
9179 effect. If both start_frame and start_time are specified, the fade will start at
9180 whichever comes last. Default is 0.
9183 The number of seconds for which the fade effect has to last. At the end of the
9184 fade-in effect the output video will have the same intensity as the input video,
9185 at the end of the fade-out transition the output video will be filled with the
9186 selected @option{color}.
9187 If both duration and nb_frames are specified, duration is used. Default is 0
9188 (nb_frames is used by default).
9191 Specify the color of the fade. Default is "black".
9194 @subsection Examples
9198 Fade in the first 30 frames of video:
9203 The command above is equivalent to:
9209 Fade out the last 45 frames of a 200-frame video:
9212 fade=type=out:start_frame=155:nb_frames=45
9216 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
9218 fade=in:0:25, fade=out:975:25
9222 Make the first 5 frames yellow, then fade in from frame 5-24:
9224 fade=in:5:20:color=yellow
9228 Fade in alpha over first 25 frames of video:
9230 fade=in:0:25:alpha=1
9234 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
9236 fade=t=in:st=5.5:d=0.5
9242 Apply arbitrary expressions to samples in frequency domain
9246 Adjust the dc value (gain) of the luma plane of the image. The filter
9247 accepts an integer value in range @code{0} to @code{1000}. The default
9248 value is set to @code{0}.
9251 Adjust the dc value (gain) of the 1st chroma plane of the image. The
9252 filter accepts an integer value in range @code{0} to @code{1000}. The
9253 default value is set to @code{0}.
9256 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
9257 filter accepts an integer value in range @code{0} to @code{1000}. The
9258 default value is set to @code{0}.
9261 Set the frequency domain weight expression for the luma plane.
9264 Set the frequency domain weight expression for the 1st chroma plane.
9267 Set the frequency domain weight expression for the 2nd chroma plane.
9270 Set when the expressions are evaluated.
9272 It accepts the following values:
9275 Only evaluate expressions once during the filter initialization.
9278 Evaluate expressions for each incoming frame.
9281 Default value is @samp{init}.
9283 The filter accepts the following variables:
9286 The coordinates of the current sample.
9290 The width and height of the image.
9293 The number of input frame, starting from 0.
9296 @subsection Examples
9302 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
9308 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
9314 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
9320 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
9326 Denoise frames using 3D FFT (frequency domain filtering).
9328 The filter accepts the following options:
9332 Set the noise sigma constant. This sets denoising strength.
9333 Default value is 1. Allowed range is from 0 to 30.
9334 Using very high sigma with low overlap may give blocking artifacts.
9337 Set amount of denoising. By default all detected noise is reduced.
9338 Default value is 1. Allowed range is from 0 to 1.
9341 Set size of block, Default is 4, can be 3, 4, 5 or 6.
9342 Actual size of block in pixels is 2 to power of @var{block}, so by default
9343 block size in pixels is 2^4 which is 16.
9346 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
9349 Set number of previous frames to use for denoising. By default is set to 0.
9352 Set number of next frames to to use for denoising. By default is set to 0.
9355 Set planes which will be filtered, by default are all available filtered
9361 Extract a single field from an interlaced image using stride
9362 arithmetic to avoid wasting CPU time. The output frames are marked as
9365 The filter accepts the following options:
9369 Specify whether to extract the top (if the value is @code{0} or
9370 @code{top}) or the bottom field (if the value is @code{1} or
9376 Create new frames by copying the top and bottom fields from surrounding frames
9377 supplied as numbers by the hint file.
9381 Set file containing hints: absolute/relative frame numbers.
9383 There must be one line for each frame in a clip. Each line must contain two
9384 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
9385 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
9386 is current frame number for @code{absolute} mode or out of [-1, 1] range
9387 for @code{relative} mode. First number tells from which frame to pick up top
9388 field and second number tells from which frame to pick up bottom field.
9390 If optionally followed by @code{+} output frame will be marked as interlaced,
9391 else if followed by @code{-} output frame will be marked as progressive, else
9392 it will be marked same as input frame.
9393 If line starts with @code{#} or @code{;} that line is skipped.
9396 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
9399 Example of first several lines of @code{hint} file for @code{relative} mode:
9402 1,0 - # second frame, use third's frame top field and second's frame bottom field
9403 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
9420 Field matching filter for inverse telecine. It is meant to reconstruct the
9421 progressive frames from a telecined stream. The filter does not drop duplicated
9422 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
9423 followed by a decimation filter such as @ref{decimate} in the filtergraph.
9425 The separation of the field matching and the decimation is notably motivated by
9426 the possibility of inserting a de-interlacing filter fallback between the two.
9427 If the source has mixed telecined and real interlaced content,
9428 @code{fieldmatch} will not be able to match fields for the interlaced parts.
9429 But these remaining combed frames will be marked as interlaced, and thus can be
9430 de-interlaced by a later filter such as @ref{yadif} before decimation.
9432 In addition to the various configuration options, @code{fieldmatch} can take an
9433 optional second stream, activated through the @option{ppsrc} option. If
9434 enabled, the frames reconstruction will be based on the fields and frames from
9435 this second stream. This allows the first input to be pre-processed in order to
9436 help the various algorithms of the filter, while keeping the output lossless
9437 (assuming the fields are matched properly). Typically, a field-aware denoiser,
9438 or brightness/contrast adjustments can help.
9440 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
9441 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
9442 which @code{fieldmatch} is based on. While the semantic and usage are very
9443 close, some behaviour and options names can differ.
9445 The @ref{decimate} filter currently only works for constant frame rate input.
9446 If your input has mixed telecined (30fps) and progressive content with a lower
9447 framerate like 24fps use the following filterchain to produce the necessary cfr
9448 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
9450 The filter accepts the following options:
9454 Specify the assumed field order of the input stream. Available values are:
9458 Auto detect parity (use FFmpeg's internal parity value).
9460 Assume bottom field first.
9462 Assume top field first.
9465 Note that it is sometimes recommended not to trust the parity announced by the
9468 Default value is @var{auto}.
9471 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
9472 sense that it won't risk creating jerkiness due to duplicate frames when
9473 possible, but if there are bad edits or blended fields it will end up
9474 outputting combed frames when a good match might actually exist. On the other
9475 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
9476 but will almost always find a good frame if there is one. The other values are
9477 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
9478 jerkiness and creating duplicate frames versus finding good matches in sections
9479 with bad edits, orphaned fields, blended fields, etc.
9481 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
9483 Available values are:
9487 2-way matching (p/c)
9489 2-way matching, and trying 3rd match if still combed (p/c + n)
9491 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
9493 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
9494 still combed (p/c + n + u/b)
9496 3-way matching (p/c/n)
9498 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
9499 detected as combed (p/c/n + u/b)
9502 The parenthesis at the end indicate the matches that would be used for that
9503 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
9506 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
9509 Default value is @var{pc_n}.
9512 Mark the main input stream as a pre-processed input, and enable the secondary
9513 input stream as the clean source to pick the fields from. See the filter
9514 introduction for more details. It is similar to the @option{clip2} feature from
9517 Default value is @code{0} (disabled).
9520 Set the field to match from. It is recommended to set this to the same value as
9521 @option{order} unless you experience matching failures with that setting. In
9522 certain circumstances changing the field that is used to match from can have a
9523 large impact on matching performance. Available values are:
9527 Automatic (same value as @option{order}).
9529 Match from the bottom field.
9531 Match from the top field.
9534 Default value is @var{auto}.
9537 Set whether or not chroma is included during the match comparisons. In most
9538 cases it is recommended to leave this enabled. You should set this to @code{0}
9539 only if your clip has bad chroma problems such as heavy rainbowing or other
9540 artifacts. Setting this to @code{0} could also be used to speed things up at
9541 the cost of some accuracy.
9543 Default value is @code{1}.
9547 These define an exclusion band which excludes the lines between @option{y0} and
9548 @option{y1} from being included in the field matching decision. An exclusion
9549 band can be used to ignore subtitles, a logo, or other things that may
9550 interfere with the matching. @option{y0} sets the starting scan line and
9551 @option{y1} sets the ending line; all lines in between @option{y0} and
9552 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
9553 @option{y0} and @option{y1} to the same value will disable the feature.
9554 @option{y0} and @option{y1} defaults to @code{0}.
9557 Set the scene change detection threshold as a percentage of maximum change on
9558 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
9559 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
9560 @option{scthresh} is @code{[0.0, 100.0]}.
9562 Default value is @code{12.0}.
9565 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
9566 account the combed scores of matches when deciding what match to use as the
9567 final match. Available values are:
9571 No final matching based on combed scores.
9573 Combed scores are only used when a scene change is detected.
9575 Use combed scores all the time.
9578 Default is @var{sc}.
9581 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
9582 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
9583 Available values are:
9587 No forced calculation.
9589 Force p/c/n calculations.
9591 Force p/c/n/u/b calculations.
9594 Default value is @var{none}.
9597 This is the area combing threshold used for combed frame detection. This
9598 essentially controls how "strong" or "visible" combing must be to be detected.
9599 Larger values mean combing must be more visible and smaller values mean combing
9600 can be less visible or strong and still be detected. Valid settings are from
9601 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
9602 be detected as combed). This is basically a pixel difference value. A good
9603 range is @code{[8, 12]}.
9605 Default value is @code{9}.
9608 Sets whether or not chroma is considered in the combed frame decision. Only
9609 disable this if your source has chroma problems (rainbowing, etc.) that are
9610 causing problems for the combed frame detection with chroma enabled. Actually,
9611 using @option{chroma}=@var{0} is usually more reliable, except for the case
9612 where there is chroma only combing in the source.
9614 Default value is @code{0}.
9618 Respectively set the x-axis and y-axis size of the window used during combed
9619 frame detection. This has to do with the size of the area in which
9620 @option{combpel} pixels are required to be detected as combed for a frame to be
9621 declared combed. See the @option{combpel} parameter description for more info.
9622 Possible values are any number that is a power of 2 starting at 4 and going up
9625 Default value is @code{16}.
9628 The number of combed pixels inside any of the @option{blocky} by
9629 @option{blockx} size blocks on the frame for the frame to be detected as
9630 combed. While @option{cthresh} controls how "visible" the combing must be, this
9631 setting controls "how much" combing there must be in any localized area (a
9632 window defined by the @option{blockx} and @option{blocky} settings) on the
9633 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
9634 which point no frames will ever be detected as combed). This setting is known
9635 as @option{MI} in TFM/VFM vocabulary.
9637 Default value is @code{80}.
9640 @anchor{p/c/n/u/b meaning}
9641 @subsection p/c/n/u/b meaning
9643 @subsubsection p/c/n
9645 We assume the following telecined stream:
9648 Top fields: 1 2 2 3 4
9649 Bottom fields: 1 2 3 4 4
9652 The numbers correspond to the progressive frame the fields relate to. Here, the
9653 first two frames are progressive, the 3rd and 4th are combed, and so on.
9655 When @code{fieldmatch} is configured to run a matching from bottom
9656 (@option{field}=@var{bottom}) this is how this input stream get transformed:
9661 B 1 2 3 4 4 <-- matching reference
9670 As a result of the field matching, we can see that some frames get duplicated.
9671 To perform a complete inverse telecine, you need to rely on a decimation filter
9672 after this operation. See for instance the @ref{decimate} filter.
9674 The same operation now matching from top fields (@option{field}=@var{top})
9679 T 1 2 2 3 4 <-- matching reference
9689 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
9690 basically, they refer to the frame and field of the opposite parity:
9693 @item @var{p} matches the field of the opposite parity in the previous frame
9694 @item @var{c} matches the field of the opposite parity in the current frame
9695 @item @var{n} matches the field of the opposite parity in the next frame
9700 The @var{u} and @var{b} matching are a bit special in the sense that they match
9701 from the opposite parity flag. In the following examples, we assume that we are
9702 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
9703 'x' is placed above and below each matched fields.
9705 With bottom matching (@option{field}=@var{bottom}):
9710 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
9711 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
9719 With top matching (@option{field}=@var{top}):
9724 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
9725 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
9733 @subsection Examples
9735 Simple IVTC of a top field first telecined stream:
9737 fieldmatch=order=tff:combmatch=none, decimate
9740 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
9742 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
9747 Transform the field order of the input video.
9749 It accepts the following parameters:
9754 The output field order. Valid values are @var{tff} for top field first or @var{bff}
9755 for bottom field first.
9758 The default value is @samp{tff}.
9760 The transformation is done by shifting the picture content up or down
9761 by one line, and filling the remaining line with appropriate picture content.
9762 This method is consistent with most broadcast field order converters.
9764 If the input video is not flagged as being interlaced, or it is already
9765 flagged as being of the required output field order, then this filter does
9766 not alter the incoming video.
9768 It is very useful when converting to or from PAL DV material,
9769 which is bottom field first.
9773 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
9776 @section fifo, afifo
9778 Buffer input images and send them when they are requested.
9780 It is mainly useful when auto-inserted by the libavfilter
9783 It does not take parameters.
9785 @section fillborders
9787 Fill borders of the input video, without changing video stream dimensions.
9788 Sometimes video can have garbage at the four edges and you may not want to
9789 crop video input to keep size multiple of some number.
9791 This filter accepts the following options:
9795 Number of pixels to fill from left border.
9798 Number of pixels to fill from right border.
9801 Number of pixels to fill from top border.
9804 Number of pixels to fill from bottom border.
9809 It accepts the following values:
9812 fill pixels using outermost pixels
9815 fill pixels using mirroring
9818 fill pixels with constant value
9821 Default is @var{smear}.
9824 Set color for pixels in fixed mode. Default is @var{black}.
9829 Find a rectangular object
9831 It accepts the following options:
9835 Filepath of the object image, needs to be in gray8.
9838 Detection threshold, default is 0.5.
9841 Number of mipmaps, default is 3.
9843 @item xmin, ymin, xmax, ymax
9844 Specifies the rectangle in which to search.
9847 @subsection Examples
9851 Generate a representative palette of a given video using @command{ffmpeg}:
9853 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
9859 Cover a rectangular object
9861 It accepts the following options:
9865 Filepath of the optional cover image, needs to be in yuv420.
9870 It accepts the following values:
9873 cover it by the supplied image
9875 cover it by interpolating the surrounding pixels
9878 Default value is @var{blur}.
9881 @subsection Examples
9885 Generate a representative palette of a given video using @command{ffmpeg}:
9887 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
9893 Flood area with values of same pixel components with another values.
9895 It accepts the following options:
9898 Set pixel x coordinate.
9901 Set pixel y coordinate.
9904 Set source #0 component value.
9907 Set source #1 component value.
9910 Set source #2 component value.
9913 Set source #3 component value.
9916 Set destination #0 component value.
9919 Set destination #1 component value.
9922 Set destination #2 component value.
9925 Set destination #3 component value.
9931 Convert the input video to one of the specified pixel formats.
9932 Libavfilter will try to pick one that is suitable as input to
9935 It accepts the following parameters:
9939 A '|'-separated list of pixel format names, such as
9940 "pix_fmts=yuv420p|monow|rgb24".
9944 @subsection Examples
9948 Convert the input video to the @var{yuv420p} format
9950 format=pix_fmts=yuv420p
9953 Convert the input video to any of the formats in the list
9955 format=pix_fmts=yuv420p|yuv444p|yuv410p
9962 Convert the video to specified constant frame rate by duplicating or dropping
9963 frames as necessary.
9965 It accepts the following parameters:
9969 The desired output frame rate. The default is @code{25}.
9972 Assume the first PTS should be the given value, in seconds. This allows for
9973 padding/trimming at the start of stream. By default, no assumption is made
9974 about the first frame's expected PTS, so no padding or trimming is done.
9975 For example, this could be set to 0 to pad the beginning with duplicates of
9976 the first frame if a video stream starts after the audio stream or to trim any
9977 frames with a negative PTS.
9980 Timestamp (PTS) rounding method.
9982 Possible values are:
9989 round towards -infinity
9991 round towards +infinity
9995 The default is @code{near}.
9998 Action performed when reading the last frame.
10000 Possible values are:
10003 Use same timestamp rounding method as used for other frames.
10005 Pass through last frame if input duration has not been reached yet.
10007 The default is @code{round}.
10011 Alternatively, the options can be specified as a flat string:
10012 @var{fps}[:@var{start_time}[:@var{round}]].
10014 See also the @ref{setpts} filter.
10016 @subsection Examples
10020 A typical usage in order to set the fps to 25:
10026 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
10028 fps=fps=film:round=near
10034 Pack two different video streams into a stereoscopic video, setting proper
10035 metadata on supported codecs. The two views should have the same size and
10036 framerate and processing will stop when the shorter video ends. Please note
10037 that you may conveniently adjust view properties with the @ref{scale} and
10040 It accepts the following parameters:
10044 The desired packing format. Supported values are:
10049 The views are next to each other (default).
10052 The views are on top of each other.
10055 The views are packed by line.
10058 The views are packed by column.
10061 The views are temporally interleaved.
10070 # Convert left and right views into a frame-sequential video
10071 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
10073 # Convert views into a side-by-side video with the same output resolution as the input
10074 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
10079 Change the frame rate by interpolating new video output frames from the source
10082 This filter is not designed to function correctly with interlaced media. If
10083 you wish to change the frame rate of interlaced media then you are required
10084 to deinterlace before this filter and re-interlace after this filter.
10086 A description of the accepted options follows.
10090 Specify the output frames per second. This option can also be specified
10091 as a value alone. The default is @code{50}.
10094 Specify the start of a range where the output frame will be created as a
10095 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10096 the default is @code{15}.
10099 Specify the end of a range where the output frame will be created as a
10100 linear interpolation of two frames. The range is [@code{0}-@code{255}],
10101 the default is @code{240}.
10104 Specify the level at which a scene change is detected as a value between
10105 0 and 100 to indicate a new scene; a low value reflects a low
10106 probability for the current frame to introduce a new scene, while a higher
10107 value means the current frame is more likely to be one.
10108 The default is @code{8.2}.
10111 Specify flags influencing the filter process.
10113 Available value for @var{flags} is:
10116 @item scene_change_detect, scd
10117 Enable scene change detection using the value of the option @var{scene}.
10118 This flag is enabled by default.
10124 Select one frame every N-th frame.
10126 This filter accepts the following option:
10129 Select frame after every @code{step} frames.
10130 Allowed values are positive integers higher than 0. Default value is @code{1}.
10133 @section freezedetect
10135 Detect frozen video.
10137 This filter logs a message and sets frame metadata when it detects that the
10138 input video has no significant change in content during a specified duration.
10139 Video freeze detection calculates the mean average absolute difference of all
10140 the components of video frames and compares it to a noise floor.
10142 The printed times and duration are expressed in seconds. The
10143 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
10144 whose timestamp equals or exceeds the detection duration and it contains the
10145 timestamp of the first frame of the freeze. The
10146 @code{lavfi.freezedetect.freeze_duration} and
10147 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
10150 The filter accepts the following options:
10154 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
10155 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
10159 Set freeze duration until notification (default is 2 seconds).
10165 Apply a frei0r effect to the input video.
10167 To enable the compilation of this filter, you need to install the frei0r
10168 header and configure FFmpeg with @code{--enable-frei0r}.
10170 It accepts the following parameters:
10175 The name of the frei0r effect to load. If the environment variable
10176 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
10177 directories specified by the colon-separated list in @env{FREI0R_PATH}.
10178 Otherwise, the standard frei0r paths are searched, in this order:
10179 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
10180 @file{/usr/lib/frei0r-1/}.
10182 @item filter_params
10183 A '|'-separated list of parameters to pass to the frei0r effect.
10187 A frei0r effect parameter can be a boolean (its value is either
10188 "y" or "n"), a double, a color (specified as
10189 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
10190 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
10191 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
10192 a position (specified as @var{X}/@var{Y}, where
10193 @var{X} and @var{Y} are floating point numbers) and/or a string.
10195 The number and types of parameters depend on the loaded effect. If an
10196 effect parameter is not specified, the default value is set.
10198 @subsection Examples
10202 Apply the distort0r effect, setting the first two double parameters:
10204 frei0r=filter_name=distort0r:filter_params=0.5|0.01
10208 Apply the colordistance effect, taking a color as the first parameter:
10210 frei0r=colordistance:0.2/0.3/0.4
10211 frei0r=colordistance:violet
10212 frei0r=colordistance:0x112233
10216 Apply the perspective effect, specifying the top left and top right image
10219 frei0r=perspective:0.2/0.2|0.8/0.2
10223 For more information, see
10224 @url{http://frei0r.dyne.org}
10228 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
10230 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
10231 processing filter, one of them is performed once per block, not per pixel.
10232 This allows for much higher speed.
10234 The filter accepts the following options:
10238 Set quality. This option defines the number of levels for averaging. It accepts
10239 an integer in the range 4-5. Default value is @code{4}.
10242 Force a constant quantization parameter. It accepts an integer in range 0-63.
10243 If not set, the filter will use the QP from the video stream (if available).
10246 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
10247 more details but also more artifacts, while higher values make the image smoother
10248 but also blurrier. Default value is @code{0} − PSNR optimal.
10250 @item use_bframe_qp
10251 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
10252 option may cause flicker since the B-Frames have often larger QP. Default is
10253 @code{0} (not enabled).
10259 Apply Gaussian blur filter.
10261 The filter accepts the following options:
10265 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
10268 Set number of steps for Gaussian approximation. Default is @code{1}.
10271 Set which planes to filter. By default all planes are filtered.
10274 Set vertical sigma, if negative it will be same as @code{sigma}.
10275 Default is @code{-1}.
10280 Apply generic equation to each pixel.
10282 The filter accepts the following options:
10285 @item lum_expr, lum
10286 Set the luminance expression.
10288 Set the chrominance blue expression.
10290 Set the chrominance red expression.
10291 @item alpha_expr, a
10292 Set the alpha expression.
10294 Set the red expression.
10295 @item green_expr, g
10296 Set the green expression.
10298 Set the blue expression.
10301 The colorspace is selected according to the specified options. If one
10302 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
10303 options is specified, the filter will automatically select a YCbCr
10304 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
10305 @option{blue_expr} options is specified, it will select an RGB
10308 If one of the chrominance expression is not defined, it falls back on the other
10309 one. If no alpha expression is specified it will evaluate to opaque value.
10310 If none of chrominance expressions are specified, they will evaluate
10311 to the luminance expression.
10313 The expressions can use the following variables and functions:
10317 The sequential number of the filtered frame, starting from @code{0}.
10321 The coordinates of the current sample.
10325 The width and height of the image.
10329 Width and height scale depending on the currently filtered plane. It is the
10330 ratio between the corresponding luma plane number of pixels and the current
10331 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
10332 @code{0.5,0.5} for chroma planes.
10335 Time of the current frame, expressed in seconds.
10338 Return the value of the pixel at location (@var{x},@var{y}) of the current
10342 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
10346 Return the value of the pixel at location (@var{x},@var{y}) of the
10347 blue-difference chroma plane. Return 0 if there is no such plane.
10350 Return the value of the pixel at location (@var{x},@var{y}) of the
10351 red-difference chroma plane. Return 0 if there is no such plane.
10356 Return the value of the pixel at location (@var{x},@var{y}) of the
10357 red/green/blue component. Return 0 if there is no such component.
10360 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
10361 plane. Return 0 if there is no such plane.
10364 For functions, if @var{x} and @var{y} are outside the area, the value will be
10365 automatically clipped to the closer edge.
10367 @subsection Examples
10371 Flip the image horizontally:
10377 Generate a bidimensional sine wave, with angle @code{PI/3} and a
10378 wavelength of 100 pixels:
10380 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
10384 Generate a fancy enigmatic moving light:
10386 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
10390 Generate a quick emboss effect:
10392 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
10396 Modify RGB components depending on pixel position:
10398 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
10402 Create a radial gradient that is the same size as the input (also see
10403 the @ref{vignette} filter):
10405 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
10411 Fix the banding artifacts that are sometimes introduced into nearly flat
10412 regions by truncation to 8-bit color depth.
10413 Interpolate the gradients that should go where the bands are, and
10416 It is designed for playback only. Do not use it prior to
10417 lossy compression, because compression tends to lose the dither and
10418 bring back the bands.
10420 It accepts the following parameters:
10425 The maximum amount by which the filter will change any one pixel. This is also
10426 the threshold for detecting nearly flat regions. Acceptable values range from
10427 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
10431 The neighborhood to fit the gradient to. A larger radius makes for smoother
10432 gradients, but also prevents the filter from modifying the pixels near detailed
10433 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
10434 values will be clipped to the valid range.
10438 Alternatively, the options can be specified as a flat string:
10439 @var{strength}[:@var{radius}]
10441 @subsection Examples
10445 Apply the filter with a @code{3.5} strength and radius of @code{8}:
10451 Specify radius, omitting the strength (which will fall-back to the default
10459 @section graphmonitor, agraphmonitor
10460 Show various filtergraph stats.
10462 With this filter one can debug complete filtergraph.
10463 Especially issues with links filling with queued frames.
10465 The filter accepts the following options:
10469 Set video output size. Default is @var{hd720}.
10472 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
10475 Set output mode, can be @var{fulll} or @var{compact}.
10476 In @var{compact} mode only filters with some queued frames have displayed stats.
10479 Set flags which enable which stats are shown in video.
10481 Available values for flags are:
10484 Display number of queued frames in each link.
10486 @item frame_count_in
10487 Display number of frames taken from filter.
10489 @item frame_count_out
10490 Display number of frames given out from filter.
10493 Display current filtered frame pts.
10496 Display current filtered frame time.
10499 Display time base for filter link.
10502 Display used format for filter link.
10505 Display video size or number of audio channels in case of audio used by filter link.
10508 Display video frame rate or sample rate in case of audio used by filter link.
10512 Set upper limit for video rate of output stream, Default value is @var{25}.
10513 This guarantee that output video frame rate will not be higher than this value.
10517 A color constancy variation filter which estimates scene illumination via grey edge algorithm
10518 and corrects the scene colors accordingly.
10520 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
10522 The filter accepts the following options:
10526 The order of differentiation to be applied on the scene. Must be chosen in the range
10527 [0,2] and default value is 1.
10530 The Minkowski parameter to be used for calculating the Minkowski distance. Must
10531 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
10532 max value instead of calculating Minkowski distance.
10535 The standard deviation of Gaussian blur to be applied on the scene. Must be
10536 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
10537 can't be equal to 0 if @var{difford} is greater than 0.
10540 @subsection Examples
10546 greyedge=difford=1:minknorm=5:sigma=2
10552 greyedge=difford=1:minknorm=0:sigma=2
10560 Apply a Hald CLUT to a video stream.
10562 First input is the video stream to process, and second one is the Hald CLUT.
10563 The Hald CLUT input can be a simple picture or a complete video stream.
10565 The filter accepts the following options:
10569 Force termination when the shortest input terminates. Default is @code{0}.
10571 Continue applying the last CLUT after the end of the stream. A value of
10572 @code{0} disable the filter after the last frame of the CLUT is reached.
10573 Default is @code{1}.
10576 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
10577 filters share the same internals).
10579 More information about the Hald CLUT can be found on Eskil Steenberg's website
10580 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
10582 @subsection Workflow examples
10584 @subsubsection Hald CLUT video stream
10586 Generate an identity Hald CLUT stream altered with various effects:
10588 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
10591 Note: make sure you use a lossless codec.
10593 Then use it with @code{haldclut} to apply it on some random stream:
10595 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
10598 The Hald CLUT will be applied to the 10 first seconds (duration of
10599 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
10600 to the remaining frames of the @code{mandelbrot} stream.
10602 @subsubsection Hald CLUT with preview
10604 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
10605 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
10606 biggest possible square starting at the top left of the picture. The remaining
10607 padding pixels (bottom or right) will be ignored. This area can be used to add
10608 a preview of the Hald CLUT.
10610 Typically, the following generated Hald CLUT will be supported by the
10611 @code{haldclut} filter:
10614 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
10615 pad=iw+320 [padded_clut];
10616 smptebars=s=320x256, split [a][b];
10617 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
10618 [main][b] overlay=W-320" -frames:v 1 clut.png
10621 It contains the original and a preview of the effect of the CLUT: SMPTE color
10622 bars are displayed on the right-top, and below the same color bars processed by
10625 Then, the effect of this Hald CLUT can be visualized with:
10627 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
10632 Flip the input video horizontally.
10634 For example, to horizontally flip the input video with @command{ffmpeg}:
10636 ffmpeg -i in.avi -vf "hflip" out.avi
10640 This filter applies a global color histogram equalization on a
10643 It can be used to correct video that has a compressed range of pixel
10644 intensities. The filter redistributes the pixel intensities to
10645 equalize their distribution across the intensity range. It may be
10646 viewed as an "automatically adjusting contrast filter". This filter is
10647 useful only for correcting degraded or poorly captured source
10650 The filter accepts the following options:
10654 Determine the amount of equalization to be applied. As the strength
10655 is reduced, the distribution of pixel intensities more-and-more
10656 approaches that of the input frame. The value must be a float number
10657 in the range [0,1] and defaults to 0.200.
10660 Set the maximum intensity that can generated and scale the output
10661 values appropriately. The strength should be set as desired and then
10662 the intensity can be limited if needed to avoid washing-out. The value
10663 must be a float number in the range [0,1] and defaults to 0.210.
10666 Set the antibanding level. If enabled the filter will randomly vary
10667 the luminance of output pixels by a small amount to avoid banding of
10668 the histogram. Possible values are @code{none}, @code{weak} or
10669 @code{strong}. It defaults to @code{none}.
10674 Compute and draw a color distribution histogram for the input video.
10676 The computed histogram is a representation of the color component
10677 distribution in an image.
10679 Standard histogram displays the color components distribution in an image.
10680 Displays color graph for each color component. Shows distribution of
10681 the Y, U, V, A or R, G, B components, depending on input format, in the
10682 current frame. Below each graph a color component scale meter is shown.
10684 The filter accepts the following options:
10688 Set height of level. Default value is @code{200}.
10689 Allowed range is [50, 2048].
10692 Set height of color scale. Default value is @code{12}.
10693 Allowed range is [0, 40].
10697 It accepts the following values:
10700 Per color component graphs are placed below each other.
10703 Per color component graphs are placed side by side.
10706 Presents information identical to that in the @code{parade}, except
10707 that the graphs representing color components are superimposed directly
10710 Default is @code{stack}.
10713 Set mode. Can be either @code{linear}, or @code{logarithmic}.
10714 Default is @code{linear}.
10717 Set what color components to display.
10718 Default is @code{7}.
10721 Set foreground opacity. Default is @code{0.7}.
10724 Set background opacity. Default is @code{0.5}.
10727 @subsection Examples
10732 Calculate and draw histogram:
10734 ffplay -i input -vf histogram
10742 This is a high precision/quality 3d denoise filter. It aims to reduce
10743 image noise, producing smooth images and making still images really
10744 still. It should enhance compressibility.
10746 It accepts the following optional parameters:
10750 A non-negative floating point number which specifies spatial luma strength.
10751 It defaults to 4.0.
10753 @item chroma_spatial
10754 A non-negative floating point number which specifies spatial chroma strength.
10755 It defaults to 3.0*@var{luma_spatial}/4.0.
10758 A floating point number which specifies luma temporal strength. It defaults to
10759 6.0*@var{luma_spatial}/4.0.
10762 A floating point number which specifies chroma temporal strength. It defaults to
10763 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
10766 @anchor{hwdownload}
10767 @section hwdownload
10769 Download hardware frames to system memory.
10771 The input must be in hardware frames, and the output a non-hardware format.
10772 Not all formats will be supported on the output - it may be necessary to insert
10773 an additional @option{format} filter immediately following in the graph to get
10774 the output in a supported format.
10778 Map hardware frames to system memory or to another device.
10780 This filter has several different modes of operation; which one is used depends
10781 on the input and output formats:
10784 Hardware frame input, normal frame output
10786 Map the input frames to system memory and pass them to the output. If the
10787 original hardware frame is later required (for example, after overlaying
10788 something else on part of it), the @option{hwmap} filter can be used again
10789 in the next mode to retrieve it.
10791 Normal frame input, hardware frame output
10793 If the input is actually a software-mapped hardware frame, then unmap it -
10794 that is, return the original hardware frame.
10796 Otherwise, a device must be provided. Create new hardware surfaces on that
10797 device for the output, then map them back to the software format at the input
10798 and give those frames to the preceding filter. This will then act like the
10799 @option{hwupload} filter, but may be able to avoid an additional copy when
10800 the input is already in a compatible format.
10802 Hardware frame input and output
10804 A device must be supplied for the output, either directly or with the
10805 @option{derive_device} option. The input and output devices must be of
10806 different types and compatible - the exact meaning of this is
10807 system-dependent, but typically it means that they must refer to the same
10808 underlying hardware context (for example, refer to the same graphics card).
10810 If the input frames were originally created on the output device, then unmap
10811 to retrieve the original frames.
10813 Otherwise, map the frames to the output device - create new hardware frames
10814 on the output corresponding to the frames on the input.
10817 The following additional parameters are accepted:
10821 Set the frame mapping mode. Some combination of:
10824 The mapped frame should be readable.
10826 The mapped frame should be writeable.
10828 The mapping will always overwrite the entire frame.
10830 This may improve performance in some cases, as the original contents of the
10831 frame need not be loaded.
10833 The mapping must not involve any copying.
10835 Indirect mappings to copies of frames are created in some cases where either
10836 direct mapping is not possible or it would have unexpected properties.
10837 Setting this flag ensures that the mapping is direct and will fail if that is
10840 Defaults to @var{read+write} if not specified.
10842 @item derive_device @var{type}
10843 Rather than using the device supplied at initialisation, instead derive a new
10844 device of type @var{type} from the device the input frames exist on.
10847 In a hardware to hardware mapping, map in reverse - create frames in the sink
10848 and map them back to the source. This may be necessary in some cases where
10849 a mapping in one direction is required but only the opposite direction is
10850 supported by the devices being used.
10852 This option is dangerous - it may break the preceding filter in undefined
10853 ways if there are any additional constraints on that filter's output.
10854 Do not use it without fully understanding the implications of its use.
10860 Upload system memory frames to hardware surfaces.
10862 The device to upload to must be supplied when the filter is initialised. If
10863 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
10866 @anchor{hwupload_cuda}
10867 @section hwupload_cuda
10869 Upload system memory frames to a CUDA device.
10871 It accepts the following optional parameters:
10875 The number of the CUDA device to use
10880 Apply a high-quality magnification filter designed for pixel art. This filter
10881 was originally created by Maxim Stepin.
10883 It accepts the following option:
10887 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
10888 @code{hq3x} and @code{4} for @code{hq4x}.
10889 Default is @code{3}.
10893 Stack input videos horizontally.
10895 All streams must be of same pixel format and of same height.
10897 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
10898 to create same output.
10900 The filter accept the following option:
10904 Set number of input streams. Default is 2.
10907 If set to 1, force the output to terminate when the shortest input
10908 terminates. Default value is 0.
10913 Modify the hue and/or the saturation of the input.
10915 It accepts the following parameters:
10919 Specify the hue angle as a number of degrees. It accepts an expression,
10920 and defaults to "0".
10923 Specify the saturation in the [-10,10] range. It accepts an expression and
10927 Specify the hue angle as a number of radians. It accepts an
10928 expression, and defaults to "0".
10931 Specify the brightness in the [-10,10] range. It accepts an expression and
10935 @option{h} and @option{H} are mutually exclusive, and can't be
10936 specified at the same time.
10938 The @option{b}, @option{h}, @option{H} and @option{s} option values are
10939 expressions containing the following constants:
10943 frame count of the input frame starting from 0
10946 presentation timestamp of the input frame expressed in time base units
10949 frame rate of the input video, NAN if the input frame rate is unknown
10952 timestamp expressed in seconds, NAN if the input timestamp is unknown
10955 time base of the input video
10958 @subsection Examples
10962 Set the hue to 90 degrees and the saturation to 1.0:
10968 Same command but expressing the hue in radians:
10974 Rotate hue and make the saturation swing between 0
10975 and 2 over a period of 1 second:
10977 hue="H=2*PI*t: s=sin(2*PI*t)+1"
10981 Apply a 3 seconds saturation fade-in effect starting at 0:
10983 hue="s=min(t/3\,1)"
10986 The general fade-in expression can be written as:
10988 hue="s=min(0\, max((t-START)/DURATION\, 1))"
10992 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
10994 hue="s=max(0\, min(1\, (8-t)/3))"
10997 The general fade-out expression can be written as:
10999 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
11004 @subsection Commands
11006 This filter supports the following commands:
11012 Modify the hue and/or the saturation and/or brightness of the input video.
11013 The command accepts the same syntax of the corresponding option.
11015 If the specified expression is not valid, it is kept at its current
11019 @section hysteresis
11021 Grow first stream into second stream by connecting components.
11022 This makes it possible to build more robust edge masks.
11024 This filter accepts the following options:
11028 Set which planes will be processed as bitmap, unprocessed planes will be
11029 copied from first stream.
11030 By default value 0xf, all planes will be processed.
11033 Set threshold which is used in filtering. If pixel component value is higher than
11034 this value filter algorithm for connecting components is activated.
11035 By default value is 0.
11040 Detect video interlacing type.
11042 This filter tries to detect if the input frames are interlaced, progressive,
11043 top or bottom field first. It will also try to detect fields that are
11044 repeated between adjacent frames (a sign of telecine).
11046 Single frame detection considers only immediately adjacent frames when classifying each frame.
11047 Multiple frame detection incorporates the classification history of previous frames.
11049 The filter will log these metadata values:
11052 @item single.current_frame
11053 Detected type of current frame using single-frame detection. One of:
11054 ``tff'' (top field first), ``bff'' (bottom field first),
11055 ``progressive'', or ``undetermined''
11058 Cumulative number of frames detected as top field first using single-frame detection.
11061 Cumulative number of frames detected as top field first using multiple-frame detection.
11064 Cumulative number of frames detected as bottom field first using single-frame detection.
11066 @item multiple.current_frame
11067 Detected type of current frame using multiple-frame detection. One of:
11068 ``tff'' (top field first), ``bff'' (bottom field first),
11069 ``progressive'', or ``undetermined''
11072 Cumulative number of frames detected as bottom field first using multiple-frame detection.
11074 @item single.progressive
11075 Cumulative number of frames detected as progressive using single-frame detection.
11077 @item multiple.progressive
11078 Cumulative number of frames detected as progressive using multiple-frame detection.
11080 @item single.undetermined
11081 Cumulative number of frames that could not be classified using single-frame detection.
11083 @item multiple.undetermined
11084 Cumulative number of frames that could not be classified using multiple-frame detection.
11086 @item repeated.current_frame
11087 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
11089 @item repeated.neither
11090 Cumulative number of frames with no repeated field.
11093 Cumulative number of frames with the top field repeated from the previous frame's top field.
11095 @item repeated.bottom
11096 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
11099 The filter accepts the following options:
11103 Set interlacing threshold.
11105 Set progressive threshold.
11107 Threshold for repeated field detection.
11109 Number of frames after which a given frame's contribution to the
11110 statistics is halved (i.e., it contributes only 0.5 to its
11111 classification). The default of 0 means that all frames seen are given
11112 full weight of 1.0 forever.
11113 @item analyze_interlaced_flag
11114 When this is not 0 then idet will use the specified number of frames to determine
11115 if the interlaced flag is accurate, it will not count undetermined frames.
11116 If the flag is found to be accurate it will be used without any further
11117 computations, if it is found to be inaccurate it will be cleared without any
11118 further computations. This allows inserting the idet filter as a low computational
11119 method to clean up the interlaced flag
11124 Deinterleave or interleave fields.
11126 This filter allows one to process interlaced images fields without
11127 deinterlacing them. Deinterleaving splits the input frame into 2
11128 fields (so called half pictures). Odd lines are moved to the top
11129 half of the output image, even lines to the bottom half.
11130 You can process (filter) them independently and then re-interleave them.
11132 The filter accepts the following options:
11136 @item chroma_mode, c
11137 @item alpha_mode, a
11138 Available values for @var{luma_mode}, @var{chroma_mode} and
11139 @var{alpha_mode} are:
11145 @item deinterleave, d
11146 Deinterleave fields, placing one above the other.
11148 @item interleave, i
11149 Interleave fields. Reverse the effect of deinterleaving.
11151 Default value is @code{none}.
11153 @item luma_swap, ls
11154 @item chroma_swap, cs
11155 @item alpha_swap, as
11156 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
11161 Apply inflate effect to the video.
11163 This filter replaces the pixel by the local(3x3) average by taking into account
11164 only values higher than the pixel.
11166 It accepts the following options:
11173 Limit the maximum change for each plane, default is 65535.
11174 If 0, plane will remain unchanged.
11179 Simple interlacing filter from progressive contents. This interleaves upper (or
11180 lower) lines from odd frames with lower (or upper) lines from even frames,
11181 halving the frame rate and preserving image height.
11184 Original Original New Frame
11185 Frame 'j' Frame 'j+1' (tff)
11186 ========== =========== ==================
11187 Line 0 --------------------> Frame 'j' Line 0
11188 Line 1 Line 1 ----> Frame 'j+1' Line 1
11189 Line 2 ---------------------> Frame 'j' Line 2
11190 Line 3 Line 3 ----> Frame 'j+1' Line 3
11192 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
11195 It accepts the following optional parameters:
11199 This determines whether the interlaced frame is taken from the even
11200 (tff - default) or odd (bff) lines of the progressive frame.
11203 Vertical lowpass filter to avoid twitter interlacing and
11204 reduce moire patterns.
11208 Disable vertical lowpass filter
11211 Enable linear filter (default)
11214 Enable complex filter. This will slightly less reduce twitter and moire
11215 but better retain detail and subjective sharpness impression.
11222 Deinterlace input video by applying Donald Graft's adaptive kernel
11223 deinterling. Work on interlaced parts of a video to produce
11224 progressive frames.
11226 The description of the accepted parameters follows.
11230 Set the threshold which affects the filter's tolerance when
11231 determining if a pixel line must be processed. It must be an integer
11232 in the range [0,255] and defaults to 10. A value of 0 will result in
11233 applying the process on every pixels.
11236 Paint pixels exceeding the threshold value to white if set to 1.
11240 Set the fields order. Swap fields if set to 1, leave fields alone if
11244 Enable additional sharpening if set to 1. Default is 0.
11247 Enable twoway sharpening if set to 1. Default is 0.
11250 @subsection Examples
11254 Apply default values:
11256 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
11260 Enable additional sharpening:
11266 Paint processed pixels in white:
11272 @section lenscorrection
11274 Correct radial lens distortion
11276 This filter can be used to correct for radial distortion as can result from the use
11277 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
11278 one can use tools available for example as part of opencv or simply trial-and-error.
11279 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
11280 and extract the k1 and k2 coefficients from the resulting matrix.
11282 Note that effectively the same filter is available in the open-source tools Krita and
11283 Digikam from the KDE project.
11285 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
11286 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
11287 brightness distribution, so you may want to use both filters together in certain
11288 cases, though you will have to take care of ordering, i.e. whether vignetting should
11289 be applied before or after lens correction.
11291 @subsection Options
11293 The filter accepts the following options:
11297 Relative x-coordinate of the focal point of the image, and thereby the center of the
11298 distortion. This value has a range [0,1] and is expressed as fractions of the image
11299 width. Default is 0.5.
11301 Relative y-coordinate of the focal point of the image, and thereby the center of the
11302 distortion. This value has a range [0,1] and is expressed as fractions of the image
11303 height. Default is 0.5.
11305 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
11306 no correction. Default is 0.
11308 Coefficient of the double quadratic correction term. This value has a range [-1,1].
11309 0 means no correction. Default is 0.
11312 The formula that generates the correction is:
11314 @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)
11316 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
11317 distances from the focal point in the source and target images, respectively.
11321 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
11323 The @code{lensfun} filter requires the camera make, camera model, and lens model
11324 to apply the lens correction. The filter will load the lensfun database and
11325 query it to find the corresponding camera and lens entries in the database. As
11326 long as these entries can be found with the given options, the filter can
11327 perform corrections on frames. Note that incomplete strings will result in the
11328 filter choosing the best match with the given options, and the filter will
11329 output the chosen camera and lens models (logged with level "info"). You must
11330 provide the make, camera model, and lens model as they are required.
11332 The filter accepts the following options:
11336 The make of the camera (for example, "Canon"). This option is required.
11339 The model of the camera (for example, "Canon EOS 100D"). This option is
11343 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
11344 option is required.
11347 The type of correction to apply. The following values are valid options:
11351 Enables fixing lens vignetting.
11354 Enables fixing lens geometry. This is the default.
11357 Enables fixing chromatic aberrations.
11360 Enables fixing lens vignetting and lens geometry.
11363 Enables fixing lens vignetting and chromatic aberrations.
11366 Enables fixing both lens geometry and chromatic aberrations.
11369 Enables all possible corrections.
11373 The focal length of the image/video (zoom; expected constant for video). For
11374 example, a 18--55mm lens has focal length range of [18--55], so a value in that
11375 range should be chosen when using that lens. Default 18.
11378 The aperture of the image/video (expected constant for video). Note that
11379 aperture is only used for vignetting correction. Default 3.5.
11381 @item focus_distance
11382 The focus distance of the image/video (expected constant for video). Note that
11383 focus distance is only used for vignetting and only slightly affects the
11384 vignetting correction process. If unknown, leave it at the default value (which
11387 @item target_geometry
11388 The target geometry of the output image/video. The following values are valid
11392 @item rectilinear (default)
11395 @item equirectangular
11396 @item fisheye_orthographic
11397 @item fisheye_stereographic
11398 @item fisheye_equisolid
11399 @item fisheye_thoby
11402 Apply the reverse of image correction (instead of correcting distortion, apply
11405 @item interpolation
11406 The type of interpolation used when correcting distortion. The following values
11411 @item linear (default)
11416 @subsection Examples
11420 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
11421 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
11425 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
11429 Apply the same as before, but only for the first 5 seconds of video.
11432 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
11439 Obtain the VMAF (Video Multi-Method Assessment Fusion)
11440 score between two input videos.
11442 The obtained VMAF score is printed through the logging system.
11444 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
11445 After installing the library it can be enabled using:
11446 @code{./configure --enable-libvmaf --enable-version3}.
11447 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
11449 The filter has following options:
11453 Set the model path which is to be used for SVM.
11454 Default value: @code{"vmaf_v0.6.1.pkl"}
11457 Set the file path to be used to store logs.
11460 Set the format of the log file (xml or json).
11462 @item enable_transform
11463 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
11464 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
11465 Default value: @code{false}
11468 Invokes the phone model which will generate VMAF scores higher than in the
11469 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
11472 Enables computing psnr along with vmaf.
11475 Enables computing ssim along with vmaf.
11478 Enables computing ms_ssim along with vmaf.
11481 Set the pool method (mean, min or harmonic mean) to be used for computing vmaf.
11484 Set number of threads to be used when computing vmaf.
11487 Set interval for frame subsampling used when computing vmaf.
11489 @item enable_conf_interval
11490 Enables confidence interval.
11493 This filter also supports the @ref{framesync} options.
11495 On the below examples the input file @file{main.mpg} being processed is
11496 compared with the reference file @file{ref.mpg}.
11499 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
11502 Example with options:
11504 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
11509 Limits the pixel components values to the specified range [min, max].
11511 The filter accepts the following options:
11515 Lower bound. Defaults to the lowest allowed value for the input.
11518 Upper bound. Defaults to the highest allowed value for the input.
11521 Specify which planes will be processed. Defaults to all available.
11528 The filter accepts the following options:
11532 Set the number of loops. Setting this value to -1 will result in infinite loops.
11536 Set maximal size in number of frames. Default is 0.
11539 Set first frame of loop. Default is 0.
11542 @subsection Examples
11546 Loop single first frame infinitely:
11548 loop=loop=-1:size=1:start=0
11552 Loop single first frame 10 times:
11554 loop=loop=10:size=1:start=0
11558 Loop 10 first frames 5 times:
11560 loop=loop=5:size=10:start=0
11566 Apply a 1D LUT to an input video.
11568 The filter accepts the following options:
11572 Set the 1D LUT file name.
11574 Currently supported formats:
11581 Select interpolation mode.
11583 Available values are:
11587 Use values from the nearest defined point.
11589 Interpolate values using the linear interpolation.
11591 Interpolate values using the cosine interpolation.
11593 Interpolate values using the cubic interpolation.
11595 Interpolate values using the spline interpolation.
11602 Apply a 3D LUT to an input video.
11604 The filter accepts the following options:
11608 Set the 3D LUT file name.
11610 Currently supported formats:
11622 Select interpolation mode.
11624 Available values are:
11628 Use values from the nearest defined point.
11630 Interpolate values using the 8 points defining a cube.
11632 Interpolate values using a tetrahedron.
11636 This filter also supports the @ref{framesync} options.
11640 Turn certain luma values into transparency.
11642 The filter accepts the following options:
11646 Set the luma which will be used as base for transparency.
11647 Default value is @code{0}.
11650 Set the range of luma values to be keyed out.
11651 Default value is @code{0}.
11654 Set the range of softness. Default value is @code{0}.
11655 Use this to control gradual transition from zero to full transparency.
11658 @section lut, lutrgb, lutyuv
11660 Compute a look-up table for binding each pixel component input value
11661 to an output value, and apply it to the input video.
11663 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
11664 to an RGB input video.
11666 These filters accept the following parameters:
11669 set first pixel component expression
11671 set second pixel component expression
11673 set third pixel component expression
11675 set fourth pixel component expression, corresponds to the alpha component
11678 set red component expression
11680 set green component expression
11682 set blue component expression
11684 alpha component expression
11687 set Y/luminance component expression
11689 set U/Cb component expression
11691 set V/Cr component expression
11694 Each of them specifies the expression to use for computing the lookup table for
11695 the corresponding pixel component values.
11697 The exact component associated to each of the @var{c*} options depends on the
11700 The @var{lut} filter requires either YUV or RGB pixel formats in input,
11701 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
11703 The expressions can contain the following constants and functions:
11708 The input width and height.
11711 The input value for the pixel component.
11714 The input value, clipped to the @var{minval}-@var{maxval} range.
11717 The maximum value for the pixel component.
11720 The minimum value for the pixel component.
11723 The negated value for the pixel component value, clipped to the
11724 @var{minval}-@var{maxval} range; it corresponds to the expression
11725 "maxval-clipval+minval".
11728 The computed value in @var{val}, clipped to the
11729 @var{minval}-@var{maxval} range.
11731 @item gammaval(gamma)
11732 The computed gamma correction value of the pixel component value,
11733 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
11735 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
11739 All expressions default to "val".
11741 @subsection Examples
11745 Negate input video:
11747 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
11748 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
11751 The above is the same as:
11753 lutrgb="r=negval:g=negval:b=negval"
11754 lutyuv="y=negval:u=negval:v=negval"
11764 Remove chroma components, turning the video into a graytone image:
11766 lutyuv="u=128:v=128"
11770 Apply a luma burning effect:
11776 Remove green and blue components:
11782 Set a constant alpha channel value on input:
11784 format=rgba,lutrgb=a="maxval-minval/2"
11788 Correct luminance gamma by a factor of 0.5:
11790 lutyuv=y=gammaval(0.5)
11794 Discard least significant bits of luma:
11796 lutyuv=y='bitand(val, 128+64+32)'
11800 Technicolor like effect:
11802 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
11806 @section lut2, tlut2
11808 The @code{lut2} filter takes two input streams and outputs one
11811 The @code{tlut2} (time lut2) filter takes two consecutive frames
11812 from one single stream.
11814 This filter accepts the following parameters:
11817 set first pixel component expression
11819 set second pixel component expression
11821 set third pixel component expression
11823 set fourth pixel component expression, corresponds to the alpha component
11826 set output bit depth, only available for @code{lut2} filter. By default is 0,
11827 which means bit depth is automatically picked from first input format.
11830 Each of them specifies the expression to use for computing the lookup table for
11831 the corresponding pixel component values.
11833 The exact component associated to each of the @var{c*} options depends on the
11836 The expressions can contain the following constants:
11841 The input width and height.
11844 The first input value for the pixel component.
11847 The second input value for the pixel component.
11850 The first input video bit depth.
11853 The second input video bit depth.
11856 All expressions default to "x".
11858 @subsection Examples
11862 Highlight differences between two RGB video streams:
11864 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)'
11868 Highlight differences between two YUV video streams:
11870 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)'
11874 Show max difference between two video streams:
11876 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)))'
11880 @section maskedclamp
11882 Clamp the first input stream with the second input and third input stream.
11884 Returns the value of first stream to be between second input
11885 stream - @code{undershoot} and third input stream + @code{overshoot}.
11887 This filter accepts the following options:
11890 Default value is @code{0}.
11893 Default value is @code{0}.
11896 Set which planes will be processed as bitmap, unprocessed planes will be
11897 copied from first stream.
11898 By default value 0xf, all planes will be processed.
11901 @section maskedmerge
11903 Merge the first input stream with the second input stream using per pixel
11904 weights in the third input stream.
11906 A value of 0 in the third stream pixel component means that pixel component
11907 from first stream is returned unchanged, while maximum value (eg. 255 for
11908 8-bit videos) means that pixel component from second stream is returned
11909 unchanged. Intermediate values define the amount of merging between both
11910 input stream's pixel components.
11912 This filter accepts the following options:
11915 Set which planes will be processed as bitmap, unprocessed planes will be
11916 copied from first stream.
11917 By default value 0xf, all planes will be processed.
11921 Create mask from input video.
11923 For example it is useful to create motion masks after @code{tblend} filter.
11925 This filter accepts the following options:
11929 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
11932 Set high threshold. Any pixel component higher than this value will be set to max value
11933 allowed for current pixel format.
11936 Set planes to filter, by default all available planes are filtered.
11939 Fill all frame pixels with this value.
11942 Set max average pixel value for frame. If sum of all pixel components is higher that this
11943 average, output frame will be completely filled with value set by @var{fill} option.
11944 Typically useful for scene changes when used in combination with @code{tblend} filter.
11949 Apply motion-compensation deinterlacing.
11951 It needs one field per frame as input and must thus be used together
11952 with yadif=1/3 or equivalent.
11954 This filter accepts the following options:
11957 Set the deinterlacing mode.
11959 It accepts one of the following values:
11964 use iterative motion estimation
11966 like @samp{slow}, but use multiple reference frames.
11968 Default value is @samp{fast}.
11971 Set the picture field parity assumed for the input video. It must be
11972 one of the following values:
11976 assume top field first
11978 assume bottom field first
11981 Default value is @samp{bff}.
11984 Set per-block quantization parameter (QP) used by the internal
11987 Higher values should result in a smoother motion vector field but less
11988 optimal individual vectors. Default value is 1.
11991 @section mergeplanes
11993 Merge color channel components from several video streams.
11995 The filter accepts up to 4 input streams, and merge selected input
11996 planes to the output video.
11998 This filter accepts the following options:
12001 Set input to output plane mapping. Default is @code{0}.
12003 The mappings is specified as a bitmap. It should be specified as a
12004 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
12005 mapping for the first plane of the output stream. 'A' sets the number of
12006 the input stream to use (from 0 to 3), and 'a' the plane number of the
12007 corresponding input to use (from 0 to 3). The rest of the mappings is
12008 similar, 'Bb' describes the mapping for the output stream second
12009 plane, 'Cc' describes the mapping for the output stream third plane and
12010 'Dd' describes the mapping for the output stream fourth plane.
12013 Set output pixel format. Default is @code{yuva444p}.
12016 @subsection Examples
12020 Merge three gray video streams of same width and height into single video stream:
12022 [a0][a1][a2]mergeplanes=0x001020:yuv444p
12026 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
12028 [a0][a1]mergeplanes=0x00010210:yuva444p
12032 Swap Y and A plane in yuva444p stream:
12034 format=yuva444p,mergeplanes=0x03010200:yuva444p
12038 Swap U and V plane in yuv420p stream:
12040 format=yuv420p,mergeplanes=0x000201:yuv420p
12044 Cast a rgb24 clip to yuv444p:
12046 format=rgb24,mergeplanes=0x000102:yuv444p
12052 Estimate and export motion vectors using block matching algorithms.
12053 Motion vectors are stored in frame side data to be used by other filters.
12055 This filter accepts the following options:
12058 Specify the motion estimation method. Accepts one of the following values:
12062 Exhaustive search algorithm.
12064 Three step search algorithm.
12066 Two dimensional logarithmic search algorithm.
12068 New three step search algorithm.
12070 Four step search algorithm.
12072 Diamond search algorithm.
12074 Hexagon-based search algorithm.
12076 Enhanced predictive zonal search algorithm.
12078 Uneven multi-hexagon search algorithm.
12080 Default value is @samp{esa}.
12083 Macroblock size. Default @code{16}.
12086 Search parameter. Default @code{7}.
12089 @section midequalizer
12091 Apply Midway Image Equalization effect using two video streams.
12093 Midway Image Equalization adjusts a pair of images to have the same
12094 histogram, while maintaining their dynamics as much as possible. It's
12095 useful for e.g. matching exposures from a pair of stereo cameras.
12097 This filter has two inputs and one output, which must be of same pixel format, but
12098 may be of different sizes. The output of filter is first input adjusted with
12099 midway histogram of both inputs.
12101 This filter accepts the following option:
12105 Set which planes to process. Default is @code{15}, which is all available planes.
12108 @section minterpolate
12110 Convert the video to specified frame rate using motion interpolation.
12112 This filter accepts the following options:
12115 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}.
12118 Motion interpolation mode. Following values are accepted:
12121 Duplicate previous or next frame for interpolating new ones.
12123 Blend source frames. Interpolated frame is mean of previous and next frames.
12125 Motion compensated interpolation. Following options are effective when this mode is selected:
12129 Motion compensation mode. Following values are accepted:
12132 Overlapped block motion compensation.
12134 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
12136 Default mode is @samp{obmc}.
12139 Motion estimation mode. Following values are accepted:
12142 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
12144 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
12146 Default mode is @samp{bilat}.
12149 The algorithm to be used for motion estimation. Following values are accepted:
12152 Exhaustive search algorithm.
12154 Three step search algorithm.
12156 Two dimensional logarithmic search algorithm.
12158 New three step search algorithm.
12160 Four step search algorithm.
12162 Diamond search algorithm.
12164 Hexagon-based search algorithm.
12166 Enhanced predictive zonal search algorithm.
12168 Uneven multi-hexagon search algorithm.
12170 Default algorithm is @samp{epzs}.
12173 Macroblock size. Default @code{16}.
12176 Motion estimation search parameter. Default @code{32}.
12179 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).
12184 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:
12187 Disable scene change detection.
12189 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
12191 Default method is @samp{fdiff}.
12193 @item scd_threshold
12194 Scene change detection threshold. Default is @code{5.0}.
12199 Mix several video input streams into one video stream.
12201 A description of the accepted options follows.
12205 The number of inputs. If unspecified, it defaults to 2.
12208 Specify weight of each input video stream as sequence.
12209 Each weight is separated by space. If number of weights
12210 is smaller than number of @var{frames} last specified
12211 weight will be used for all remaining unset weights.
12214 Specify scale, if it is set it will be multiplied with sum
12215 of each weight multiplied with pixel values to give final destination
12216 pixel value. By default @var{scale} is auto scaled to sum of weights.
12219 Specify how end of stream is determined.
12222 The duration of the longest input. (default)
12225 The duration of the shortest input.
12228 The duration of the first input.
12232 @section mpdecimate
12234 Drop frames that do not differ greatly from the previous frame in
12235 order to reduce frame rate.
12237 The main use of this filter is for very-low-bitrate encoding
12238 (e.g. streaming over dialup modem), but it could in theory be used for
12239 fixing movies that were inverse-telecined incorrectly.
12241 A description of the accepted options follows.
12245 Set the maximum number of consecutive frames which can be dropped (if
12246 positive), or the minimum interval between dropped frames (if
12247 negative). If the value is 0, the frame is dropped disregarding the
12248 number of previous sequentially dropped frames.
12250 Default value is 0.
12255 Set the dropping threshold values.
12257 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
12258 represent actual pixel value differences, so a threshold of 64
12259 corresponds to 1 unit of difference for each pixel, or the same spread
12260 out differently over the block.
12262 A frame is a candidate for dropping if no 8x8 blocks differ by more
12263 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
12264 meaning the whole image) differ by more than a threshold of @option{lo}.
12266 Default value for @option{hi} is 64*12, default value for @option{lo} is
12267 64*5, and default value for @option{frac} is 0.33.
12273 Negate (invert) the input video.
12275 It accepts the following option:
12280 With value 1, it negates the alpha component, if present. Default value is 0.
12286 Denoise frames using Non-Local Means algorithm.
12288 Each pixel is adjusted by looking for other pixels with similar contexts. This
12289 context similarity is defined by comparing their surrounding patches of size
12290 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
12293 Note that the research area defines centers for patches, which means some
12294 patches will be made of pixels outside that research area.
12296 The filter accepts the following options.
12300 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
12303 Set patch size. Default is 7. Must be odd number in range [0, 99].
12306 Same as @option{p} but for chroma planes.
12308 The default value is @var{0} and means automatic.
12311 Set research size. Default is 15. Must be odd number in range [0, 99].
12314 Same as @option{r} but for chroma planes.
12316 The default value is @var{0} and means automatic.
12321 Deinterlace video using neural network edge directed interpolation.
12323 This filter accepts the following options:
12327 Mandatory option, without binary file filter can not work.
12328 Currently file can be found here:
12329 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
12332 Set which frames to deinterlace, by default it is @code{all}.
12333 Can be @code{all} or @code{interlaced}.
12336 Set mode of operation.
12338 Can be one of the following:
12342 Use frame flags, both fields.
12344 Use frame flags, single field.
12346 Use top field only.
12348 Use bottom field only.
12350 Use both fields, top first.
12352 Use both fields, bottom first.
12356 Set which planes to process, by default filter process all frames.
12359 Set size of local neighborhood around each pixel, used by the predictor neural
12362 Can be one of the following:
12375 Set the number of neurons in predictor neural network.
12376 Can be one of the following:
12387 Controls the number of different neural network predictions that are blended
12388 together to compute the final output value. Can be @code{fast}, default or
12392 Set which set of weights to use in the predictor.
12393 Can be one of the following:
12397 weights trained to minimize absolute error
12399 weights trained to minimize squared error
12403 Controls whether or not the prescreener neural network is used to decide
12404 which pixels should be processed by the predictor neural network and which
12405 can be handled by simple cubic interpolation.
12406 The prescreener is trained to know whether cubic interpolation will be
12407 sufficient for a pixel or whether it should be predicted by the predictor nn.
12408 The computational complexity of the prescreener nn is much less than that of
12409 the predictor nn. Since most pixels can be handled by cubic interpolation,
12410 using the prescreener generally results in much faster processing.
12411 The prescreener is pretty accurate, so the difference between using it and not
12412 using it is almost always unnoticeable.
12414 Can be one of the following:
12422 Default is @code{new}.
12425 Set various debugging flags.
12430 Force libavfilter not to use any of the specified pixel formats for the
12431 input to the next filter.
12433 It accepts the following parameters:
12437 A '|'-separated list of pixel format names, such as
12438 pix_fmts=yuv420p|monow|rgb24".
12442 @subsection Examples
12446 Force libavfilter to use a format different from @var{yuv420p} for the
12447 input to the vflip filter:
12449 noformat=pix_fmts=yuv420p,vflip
12453 Convert the input video to any of the formats not contained in the list:
12455 noformat=yuv420p|yuv444p|yuv410p
12461 Add noise on video input frame.
12463 The filter accepts the following options:
12471 Set noise seed for specific pixel component or all pixel components in case
12472 of @var{all_seed}. Default value is @code{123457}.
12474 @item all_strength, alls
12475 @item c0_strength, c0s
12476 @item c1_strength, c1s
12477 @item c2_strength, c2s
12478 @item c3_strength, c3s
12479 Set noise strength for specific pixel component or all pixel components in case
12480 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
12482 @item all_flags, allf
12483 @item c0_flags, c0f
12484 @item c1_flags, c1f
12485 @item c2_flags, c2f
12486 @item c3_flags, c3f
12487 Set pixel component flags or set flags for all components if @var{all_flags}.
12488 Available values for component flags are:
12491 averaged temporal noise (smoother)
12493 mix random noise with a (semi)regular pattern
12495 temporal noise (noise pattern changes between frames)
12497 uniform noise (gaussian otherwise)
12501 @subsection Examples
12503 Add temporal and uniform noise to input video:
12505 noise=alls=20:allf=t+u
12510 Normalize RGB video (aka histogram stretching, contrast stretching).
12511 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
12513 For each channel of each frame, the filter computes the input range and maps
12514 it linearly to the user-specified output range. The output range defaults
12515 to the full dynamic range from pure black to pure white.
12517 Temporal smoothing can be used on the input range to reduce flickering (rapid
12518 changes in brightness) caused when small dark or bright objects enter or leave
12519 the scene. This is similar to the auto-exposure (automatic gain control) on a
12520 video camera, and, like a video camera, it may cause a period of over- or
12521 under-exposure of the video.
12523 The R,G,B channels can be normalized independently, which may cause some
12524 color shifting, or linked together as a single channel, which prevents
12525 color shifting. Linked normalization preserves hue. Independent normalization
12526 does not, so it can be used to remove some color casts. Independent and linked
12527 normalization can be combined in any ratio.
12529 The normalize filter accepts the following options:
12534 Colors which define the output range. The minimum input value is mapped to
12535 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
12536 The defaults are black and white respectively. Specifying white for
12537 @var{blackpt} and black for @var{whitept} will give color-inverted,
12538 normalized video. Shades of grey can be used to reduce the dynamic range
12539 (contrast). Specifying saturated colors here can create some interesting
12543 The number of previous frames to use for temporal smoothing. The input range
12544 of each channel is smoothed using a rolling average over the current frame
12545 and the @var{smoothing} previous frames. The default is 0 (no temporal
12549 Controls the ratio of independent (color shifting) channel normalization to
12550 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
12551 independent. Defaults to 1.0 (fully independent).
12554 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
12555 expensive no-op. Defaults to 1.0 (full strength).
12559 @subsection Examples
12561 Stretch video contrast to use the full dynamic range, with no temporal
12562 smoothing; may flicker depending on the source content:
12564 normalize=blackpt=black:whitept=white:smoothing=0
12567 As above, but with 50 frames of temporal smoothing; flicker should be
12568 reduced, depending on the source content:
12570 normalize=blackpt=black:whitept=white:smoothing=50
12573 As above, but with hue-preserving linked channel normalization:
12575 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
12578 As above, but with half strength:
12580 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
12583 Map the darkest input color to red, the brightest input color to cyan:
12585 normalize=blackpt=red:whitept=cyan
12590 Pass the video source unchanged to the output.
12593 Optical Character Recognition
12595 This filter uses Tesseract for optical character recognition. To enable
12596 compilation of this filter, you need to configure FFmpeg with
12597 @code{--enable-libtesseract}.
12599 It accepts the following options:
12603 Set datapath to tesseract data. Default is to use whatever was
12604 set at installation.
12607 Set language, default is "eng".
12610 Set character whitelist.
12613 Set character blacklist.
12616 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
12620 Apply a video transform using libopencv.
12622 To enable this filter, install the libopencv library and headers and
12623 configure FFmpeg with @code{--enable-libopencv}.
12625 It accepts the following parameters:
12630 The name of the libopencv filter to apply.
12632 @item filter_params
12633 The parameters to pass to the libopencv filter. If not specified, the default
12634 values are assumed.
12638 Refer to the official libopencv documentation for more precise
12640 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
12642 Several libopencv filters are supported; see the following subsections.
12647 Dilate an image by using a specific structuring element.
12648 It corresponds to the libopencv function @code{cvDilate}.
12650 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
12652 @var{struct_el} represents a structuring element, and has the syntax:
12653 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
12655 @var{cols} and @var{rows} represent the number of columns and rows of
12656 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
12657 point, and @var{shape} the shape for the structuring element. @var{shape}
12658 must be "rect", "cross", "ellipse", or "custom".
12660 If the value for @var{shape} is "custom", it must be followed by a
12661 string of the form "=@var{filename}". The file with name
12662 @var{filename} is assumed to represent a binary image, with each
12663 printable character corresponding to a bright pixel. When a custom
12664 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
12665 or columns and rows of the read file are assumed instead.
12667 The default value for @var{struct_el} is "3x3+0x0/rect".
12669 @var{nb_iterations} specifies the number of times the transform is
12670 applied to the image, and defaults to 1.
12674 # Use the default values
12677 # Dilate using a structuring element with a 5x5 cross, iterating two times
12678 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
12680 # Read the shape from the file diamond.shape, iterating two times.
12681 # The file diamond.shape may contain a pattern of characters like this
12687 # The specified columns and rows are ignored
12688 # but the anchor point coordinates are not
12689 ocv=dilate:0x0+2x2/custom=diamond.shape|2
12694 Erode an image by using a specific structuring element.
12695 It corresponds to the libopencv function @code{cvErode}.
12697 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
12698 with the same syntax and semantics as the @ref{dilate} filter.
12702 Smooth the input video.
12704 The filter takes the following parameters:
12705 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
12707 @var{type} is the type of smooth filter to apply, and must be one of
12708 the following values: "blur", "blur_no_scale", "median", "gaussian",
12709 or "bilateral". The default value is "gaussian".
12711 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
12712 depend on the smooth type. @var{param1} and
12713 @var{param2} accept integer positive values or 0. @var{param3} and
12714 @var{param4} accept floating point values.
12716 The default value for @var{param1} is 3. The default value for the
12717 other parameters is 0.
12719 These parameters correspond to the parameters assigned to the
12720 libopencv function @code{cvSmooth}.
12722 @section oscilloscope
12724 2D Video Oscilloscope.
12726 Useful to measure spatial impulse, step responses, chroma delays, etc.
12728 It accepts the following parameters:
12732 Set scope center x position.
12735 Set scope center y position.
12738 Set scope size, relative to frame diagonal.
12741 Set scope tilt/rotation.
12747 Set trace center x position.
12750 Set trace center y position.
12753 Set trace width, relative to width of frame.
12756 Set trace height, relative to height of frame.
12759 Set which components to trace. By default it traces first three components.
12762 Draw trace grid. By default is enabled.
12765 Draw some statistics. By default is enabled.
12768 Draw scope. By default is enabled.
12771 @subsection Examples
12775 Inspect full first row of video frame.
12777 oscilloscope=x=0.5:y=0:s=1
12781 Inspect full last row of video frame.
12783 oscilloscope=x=0.5:y=1:s=1
12787 Inspect full 5th line of video frame of height 1080.
12789 oscilloscope=x=0.5:y=5/1080:s=1
12793 Inspect full last column of video frame.
12795 oscilloscope=x=1:y=0.5:s=1:t=1
12803 Overlay one video on top of another.
12805 It takes two inputs and has one output. The first input is the "main"
12806 video on which the second input is overlaid.
12808 It accepts the following parameters:
12810 A description of the accepted options follows.
12815 Set the expression for the x and y coordinates of the overlaid video
12816 on the main video. Default value is "0" for both expressions. In case
12817 the expression is invalid, it is set to a huge value (meaning that the
12818 overlay will not be displayed within the output visible area).
12821 See @ref{framesync}.
12824 Set when the expressions for @option{x}, and @option{y} are evaluated.
12826 It accepts the following values:
12829 only evaluate expressions once during the filter initialization or
12830 when a command is processed
12833 evaluate expressions for each incoming frame
12836 Default value is @samp{frame}.
12839 See @ref{framesync}.
12842 Set the format for the output video.
12844 It accepts the following values:
12847 force YUV420 output
12850 force YUV422 output
12853 force YUV444 output
12856 force packed RGB output
12859 force planar RGB output
12862 automatically pick format
12865 Default value is @samp{yuv420}.
12868 See @ref{framesync}.
12871 Set format of alpha of the overlaid video, it can be @var{straight} or
12872 @var{premultiplied}. Default is @var{straight}.
12875 The @option{x}, and @option{y} expressions can contain the following
12881 The main input width and height.
12885 The overlay input width and height.
12889 The computed values for @var{x} and @var{y}. They are evaluated for
12894 horizontal and vertical chroma subsample values of the output
12895 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
12899 the number of input frame, starting from 0
12902 the position in the file of the input frame, NAN if unknown
12905 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
12909 This filter also supports the @ref{framesync} options.
12911 Note that the @var{n}, @var{pos}, @var{t} variables are available only
12912 when evaluation is done @emph{per frame}, and will evaluate to NAN
12913 when @option{eval} is set to @samp{init}.
12915 Be aware that frames are taken from each input video in timestamp
12916 order, hence, if their initial timestamps differ, it is a good idea
12917 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
12918 have them begin in the same zero timestamp, as the example for
12919 the @var{movie} filter does.
12921 You can chain together more overlays but you should test the
12922 efficiency of such approach.
12924 @subsection Commands
12926 This filter supports the following commands:
12930 Modify the x and y of the overlay input.
12931 The command accepts the same syntax of the corresponding option.
12933 If the specified expression is not valid, it is kept at its current
12937 @subsection Examples
12941 Draw the overlay at 10 pixels from the bottom right corner of the main
12944 overlay=main_w-overlay_w-10:main_h-overlay_h-10
12947 Using named options the example above becomes:
12949 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
12953 Insert a transparent PNG logo in the bottom left corner of the input,
12954 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
12956 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
12960 Insert 2 different transparent PNG logos (second logo on bottom
12961 right corner) using the @command{ffmpeg} tool:
12963 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
12967 Add a transparent color layer on top of the main video; @code{WxH}
12968 must specify the size of the main input to the overlay filter:
12970 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
12974 Play an original video and a filtered version (here with the deshake
12975 filter) side by side using the @command{ffplay} tool:
12977 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
12980 The above command is the same as:
12982 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
12986 Make a sliding overlay appearing from the left to the right top part of the
12987 screen starting since time 2:
12989 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
12993 Compose output by putting two input videos side to side:
12995 ffmpeg -i left.avi -i right.avi -filter_complex "
12996 nullsrc=size=200x100 [background];
12997 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
12998 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
12999 [background][left] overlay=shortest=1 [background+left];
13000 [background+left][right] overlay=shortest=1:x=100 [left+right]
13005 Mask 10-20 seconds of a video by applying the delogo filter to a section
13007 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
13008 -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]'
13013 Chain several overlays in cascade:
13015 nullsrc=s=200x200 [bg];
13016 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
13017 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
13018 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
13019 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
13020 [in3] null, [mid2] overlay=100:100 [out0]
13027 Apply Overcomplete Wavelet denoiser.
13029 The filter accepts the following options:
13035 Larger depth values will denoise lower frequency components more, but
13036 slow down filtering.
13038 Must be an int in the range 8-16, default is @code{8}.
13040 @item luma_strength, ls
13043 Must be a double value in the range 0-1000, default is @code{1.0}.
13045 @item chroma_strength, cs
13046 Set chroma strength.
13048 Must be a double value in the range 0-1000, default is @code{1.0}.
13054 Add paddings to the input image, and place the original input at the
13055 provided @var{x}, @var{y} coordinates.
13057 It accepts the following parameters:
13062 Specify an expression for the size of the output image with the
13063 paddings added. If the value for @var{width} or @var{height} is 0, the
13064 corresponding input size is used for the output.
13066 The @var{width} expression can reference the value set by the
13067 @var{height} expression, and vice versa.
13069 The default value of @var{width} and @var{height} is 0.
13073 Specify the offsets to place the input image at within the padded area,
13074 with respect to the top/left border of the output image.
13076 The @var{x} expression can reference the value set by the @var{y}
13077 expression, and vice versa.
13079 The default value of @var{x} and @var{y} is 0.
13081 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
13082 so the input image is centered on the padded area.
13085 Specify the color of the padded area. For the syntax of this option,
13086 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
13087 manual,ffmpeg-utils}.
13089 The default value of @var{color} is "black".
13092 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
13094 It accepts the following values:
13098 Only evaluate expressions once during the filter initialization or when
13099 a command is processed.
13102 Evaluate expressions for each incoming frame.
13106 Default value is @samp{init}.
13109 Pad to aspect instead to a resolution.
13113 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
13114 options are expressions containing the following constants:
13119 The input video width and height.
13123 These are the same as @var{in_w} and @var{in_h}.
13127 The output width and height (the size of the padded area), as
13128 specified by the @var{width} and @var{height} expressions.
13132 These are the same as @var{out_w} and @var{out_h}.
13136 The x and y offsets as specified by the @var{x} and @var{y}
13137 expressions, or NAN if not yet specified.
13140 same as @var{iw} / @var{ih}
13143 input sample aspect ratio
13146 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
13150 The horizontal and vertical chroma subsample values. For example for the
13151 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
13154 @subsection Examples
13158 Add paddings with the color "violet" to the input video. The output video
13159 size is 640x480, and the top-left corner of the input video is placed at
13162 pad=640:480:0:40:violet
13165 The example above is equivalent to the following command:
13167 pad=width=640:height=480:x=0:y=40:color=violet
13171 Pad the input to get an output with dimensions increased by 3/2,
13172 and put the input video at the center of the padded area:
13174 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
13178 Pad the input to get a squared output with size equal to the maximum
13179 value between the input width and height, and put the input video at
13180 the center of the padded area:
13182 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
13186 Pad the input to get a final w/h ratio of 16:9:
13188 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
13192 In case of anamorphic video, in order to set the output display aspect
13193 correctly, it is necessary to use @var{sar} in the expression,
13194 according to the relation:
13196 (ih * X / ih) * sar = output_dar
13197 X = output_dar / sar
13200 Thus the previous example needs to be modified to:
13202 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
13206 Double the output size and put the input video in the bottom-right
13207 corner of the output padded area:
13209 pad="2*iw:2*ih:ow-iw:oh-ih"
13213 @anchor{palettegen}
13214 @section palettegen
13216 Generate one palette for a whole video stream.
13218 It accepts the following options:
13222 Set the maximum number of colors to quantize in the palette.
13223 Note: the palette will still contain 256 colors; the unused palette entries
13226 @item reserve_transparent
13227 Create a palette of 255 colors maximum and reserve the last one for
13228 transparency. Reserving the transparency color is useful for GIF optimization.
13229 If not set, the maximum of colors in the palette will be 256. You probably want
13230 to disable this option for a standalone image.
13233 @item transparency_color
13234 Set the color that will be used as background for transparency.
13237 Set statistics mode.
13239 It accepts the following values:
13242 Compute full frame histograms.
13244 Compute histograms only for the part that differs from previous frame. This
13245 might be relevant to give more importance to the moving part of your input if
13246 the background is static.
13248 Compute new histogram for each frame.
13251 Default value is @var{full}.
13254 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
13255 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
13256 color quantization of the palette. This information is also visible at
13257 @var{info} logging level.
13259 @subsection Examples
13263 Generate a representative palette of a given video using @command{ffmpeg}:
13265 ffmpeg -i input.mkv -vf palettegen palette.png
13269 @section paletteuse
13271 Use a palette to downsample an input video stream.
13273 The filter takes two inputs: one video stream and a palette. The palette must
13274 be a 256 pixels image.
13276 It accepts the following options:
13280 Select dithering mode. Available algorithms are:
13283 Ordered 8x8 bayer dithering (deterministic)
13285 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
13286 Note: this dithering is sometimes considered "wrong" and is included as a
13288 @item floyd_steinberg
13289 Floyd and Steingberg dithering (error diffusion)
13291 Frankie Sierra dithering v2 (error diffusion)
13293 Frankie Sierra dithering v2 "Lite" (error diffusion)
13296 Default is @var{sierra2_4a}.
13299 When @var{bayer} dithering is selected, this option defines the scale of the
13300 pattern (how much the crosshatch pattern is visible). A low value means more
13301 visible pattern for less banding, and higher value means less visible pattern
13302 at the cost of more banding.
13304 The option must be an integer value in the range [0,5]. Default is @var{2}.
13307 If set, define the zone to process
13311 Only the changing rectangle will be reprocessed. This is similar to GIF
13312 cropping/offsetting compression mechanism. This option can be useful for speed
13313 if only a part of the image is changing, and has use cases such as limiting the
13314 scope of the error diffusal @option{dither} to the rectangle that bounds the
13315 moving scene (it leads to more deterministic output if the scene doesn't change
13316 much, and as a result less moving noise and better GIF compression).
13319 Default is @var{none}.
13322 Take new palette for each output frame.
13324 @item alpha_threshold
13325 Sets the alpha threshold for transparency. Alpha values above this threshold
13326 will be treated as completely opaque, and values below this threshold will be
13327 treated as completely transparent.
13329 The option must be an integer value in the range [0,255]. Default is @var{128}.
13332 @subsection Examples
13336 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
13337 using @command{ffmpeg}:
13339 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
13343 @section perspective
13345 Correct perspective of video not recorded perpendicular to the screen.
13347 A description of the accepted parameters follows.
13358 Set coordinates expression for top left, top right, bottom left and bottom right corners.
13359 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
13360 If the @code{sense} option is set to @code{source}, then the specified points will be sent
13361 to the corners of the destination. If the @code{sense} option is set to @code{destination},
13362 then the corners of the source will be sent to the specified coordinates.
13364 The expressions can use the following variables:
13369 the width and height of video frame.
13373 Output frame count.
13376 @item interpolation
13377 Set interpolation for perspective correction.
13379 It accepts the following values:
13385 Default value is @samp{linear}.
13388 Set interpretation of coordinate options.
13390 It accepts the following values:
13394 Send point in the source specified by the given coordinates to
13395 the corners of the destination.
13397 @item 1, destination
13399 Send the corners of the source to the point in the destination specified
13400 by the given coordinates.
13402 Default value is @samp{source}.
13406 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
13408 It accepts the following values:
13411 only evaluate expressions once during the filter initialization or
13412 when a command is processed
13415 evaluate expressions for each incoming frame
13418 Default value is @samp{init}.
13423 Delay interlaced video by one field time so that the field order changes.
13425 The intended use is to fix PAL movies that have been captured with the
13426 opposite field order to the film-to-video transfer.
13428 A description of the accepted parameters follows.
13434 It accepts the following values:
13437 Capture field order top-first, transfer bottom-first.
13438 Filter will delay the bottom field.
13441 Capture field order bottom-first, transfer top-first.
13442 Filter will delay the top field.
13445 Capture and transfer with the same field order. This mode only exists
13446 for the documentation of the other options to refer to, but if you
13447 actually select it, the filter will faithfully do nothing.
13450 Capture field order determined automatically by field flags, transfer
13452 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
13453 basis using field flags. If no field information is available,
13454 then this works just like @samp{u}.
13457 Capture unknown or varying, transfer opposite.
13458 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
13459 analyzing the images and selecting the alternative that produces best
13460 match between the fields.
13463 Capture top-first, transfer unknown or varying.
13464 Filter selects among @samp{t} and @samp{p} using image analysis.
13467 Capture bottom-first, transfer unknown or varying.
13468 Filter selects among @samp{b} and @samp{p} using image analysis.
13471 Capture determined by field flags, transfer unknown or varying.
13472 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
13473 image analysis. If no field information is available, then this works just
13474 like @samp{U}. This is the default mode.
13477 Both capture and transfer unknown or varying.
13478 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
13482 @section pixdesctest
13484 Pixel format descriptor test filter, mainly useful for internal
13485 testing. The output video should be equal to the input video.
13489 format=monow, pixdesctest
13492 can be used to test the monowhite pixel format descriptor definition.
13496 Display sample values of color channels. Mainly useful for checking color
13497 and levels. Minimum supported resolution is 640x480.
13499 The filters accept the following options:
13503 Set scope X position, relative offset on X axis.
13506 Set scope Y position, relative offset on Y axis.
13515 Set window opacity. This window also holds statistics about pixel area.
13518 Set window X position, relative offset on X axis.
13521 Set window Y position, relative offset on Y axis.
13526 Enable the specified chain of postprocessing subfilters using libpostproc. This
13527 library should be automatically selected with a GPL build (@code{--enable-gpl}).
13528 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
13529 Each subfilter and some options have a short and a long name that can be used
13530 interchangeably, i.e. dr/dering are the same.
13532 The filters accept the following options:
13536 Set postprocessing subfilters string.
13539 All subfilters share common options to determine their scope:
13543 Honor the quality commands for this subfilter.
13546 Do chrominance filtering, too (default).
13549 Do luminance filtering only (no chrominance).
13552 Do chrominance filtering only (no luminance).
13555 These options can be appended after the subfilter name, separated by a '|'.
13557 Available subfilters are:
13560 @item hb/hdeblock[|difference[|flatness]]
13561 Horizontal deblocking filter
13564 Difference factor where higher values mean more deblocking (default: @code{32}).
13566 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13569 @item vb/vdeblock[|difference[|flatness]]
13570 Vertical deblocking filter
13573 Difference factor where higher values mean more deblocking (default: @code{32}).
13575 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13578 @item ha/hadeblock[|difference[|flatness]]
13579 Accurate horizontal deblocking filter
13582 Difference factor where higher values mean more deblocking (default: @code{32}).
13584 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13587 @item va/vadeblock[|difference[|flatness]]
13588 Accurate vertical deblocking filter
13591 Difference factor where higher values mean more deblocking (default: @code{32}).
13593 Flatness threshold where lower values mean more deblocking (default: @code{39}).
13597 The horizontal and vertical deblocking filters share the difference and
13598 flatness values so you cannot set different horizontal and vertical
13602 @item h1/x1hdeblock
13603 Experimental horizontal deblocking filter
13605 @item v1/x1vdeblock
13606 Experimental vertical deblocking filter
13611 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
13614 larger -> stronger filtering
13616 larger -> stronger filtering
13618 larger -> stronger filtering
13621 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
13624 Stretch luminance to @code{0-255}.
13627 @item lb/linblenddeint
13628 Linear blend deinterlacing filter that deinterlaces the given block by
13629 filtering all lines with a @code{(1 2 1)} filter.
13631 @item li/linipoldeint
13632 Linear interpolating deinterlacing filter that deinterlaces the given block by
13633 linearly interpolating every second line.
13635 @item ci/cubicipoldeint
13636 Cubic interpolating deinterlacing filter deinterlaces the given block by
13637 cubically interpolating every second line.
13639 @item md/mediandeint
13640 Median deinterlacing filter that deinterlaces the given block by applying a
13641 median filter to every second line.
13643 @item fd/ffmpegdeint
13644 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
13645 second line with a @code{(-1 4 2 4 -1)} filter.
13648 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
13649 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
13651 @item fq/forceQuant[|quantizer]
13652 Overrides the quantizer table from the input with the constant quantizer you
13660 Default pp filter combination (@code{hb|a,vb|a,dr|a})
13663 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
13666 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
13669 @subsection Examples
13673 Apply horizontal and vertical deblocking, deringing and automatic
13674 brightness/contrast:
13680 Apply default filters without brightness/contrast correction:
13686 Apply default filters and temporal denoiser:
13688 pp=default/tmpnoise|1|2|3
13692 Apply deblocking on luminance only, and switch vertical deblocking on or off
13693 automatically depending on available CPU time:
13700 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
13701 similar to spp = 6 with 7 point DCT, where only the center sample is
13704 The filter accepts the following options:
13708 Force a constant quantization parameter. It accepts an integer in range
13709 0 to 63. If not set, the filter will use the QP from the video stream
13713 Set thresholding mode. Available modes are:
13717 Set hard thresholding.
13719 Set soft thresholding (better de-ringing effect, but likely blurrier).
13721 Set medium thresholding (good results, default).
13725 @section premultiply
13726 Apply alpha premultiply effect to input video stream using first plane
13727 of second stream as alpha.
13729 Both streams must have same dimensions and same pixel format.
13731 The filter accepts the following option:
13735 Set which planes will be processed, unprocessed planes will be copied.
13736 By default value 0xf, all planes will be processed.
13739 Do not require 2nd input for processing, instead use alpha plane from input stream.
13743 Apply prewitt operator to input video stream.
13745 The filter accepts the following option:
13749 Set which planes will be processed, unprocessed planes will be copied.
13750 By default value 0xf, all planes will be processed.
13753 Set value which will be multiplied with filtered result.
13756 Set value which will be added to filtered result.
13759 @anchor{program_opencl}
13760 @section program_opencl
13762 Filter video using an OpenCL program.
13767 OpenCL program source file.
13770 Kernel name in program.
13773 Number of inputs to the filter. Defaults to 1.
13776 Size of output frames. Defaults to the same as the first input.
13780 The program source file must contain a kernel function with the given name,
13781 which will be run once for each plane of the output. Each run on a plane
13782 gets enqueued as a separate 2D global NDRange with one work-item for each
13783 pixel to be generated. The global ID offset for each work-item is therefore
13784 the coordinates of a pixel in the destination image.
13786 The kernel function needs to take the following arguments:
13789 Destination image, @var{__write_only image2d_t}.
13791 This image will become the output; the kernel should write all of it.
13793 Frame index, @var{unsigned int}.
13795 This is a counter starting from zero and increasing by one for each frame.
13797 Source images, @var{__read_only image2d_t}.
13799 These are the most recent images on each input. The kernel may read from
13800 them to generate the output, but they can't be written to.
13807 Copy the input to the output (output must be the same size as the input).
13809 __kernel void copy(__write_only image2d_t destination,
13810 unsigned int index,
13811 __read_only image2d_t source)
13813 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
13815 int2 location = (int2)(get_global_id(0), get_global_id(1));
13817 float4 value = read_imagef(source, sampler, location);
13819 write_imagef(destination, location, value);
13824 Apply a simple transformation, rotating the input by an amount increasing
13825 with the index counter. Pixel values are linearly interpolated by the
13826 sampler, and the output need not have the same dimensions as the input.
13828 __kernel void rotate_image(__write_only image2d_t dst,
13829 unsigned int index,
13830 __read_only image2d_t src)
13832 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
13833 CLK_FILTER_LINEAR);
13835 float angle = (float)index / 100.0f;
13837 float2 dst_dim = convert_float2(get_image_dim(dst));
13838 float2 src_dim = convert_float2(get_image_dim(src));
13840 float2 dst_cen = dst_dim / 2.0f;
13841 float2 src_cen = src_dim / 2.0f;
13843 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
13845 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
13847 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
13848 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
13850 src_pos = src_pos * src_dim / dst_dim;
13852 float2 src_loc = src_pos + src_cen;
13854 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
13855 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
13856 write_imagef(dst, dst_loc, 0.5f);
13858 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
13863 Blend two inputs together, with the amount of each input used varying
13864 with the index counter.
13866 __kernel void blend_images(__write_only image2d_t dst,
13867 unsigned int index,
13868 __read_only image2d_t src1,
13869 __read_only image2d_t src2)
13871 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
13872 CLK_FILTER_LINEAR);
13874 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
13876 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
13877 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
13878 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
13880 float4 val1 = read_imagef(src1, sampler, src1_loc);
13881 float4 val2 = read_imagef(src2, sampler, src2_loc);
13883 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
13889 @section pseudocolor
13891 Alter frame colors in video with pseudocolors.
13893 This filter accept the following options:
13897 set pixel first component expression
13900 set pixel second component expression
13903 set pixel third component expression
13906 set pixel fourth component expression, corresponds to the alpha component
13909 set component to use as base for altering colors
13912 Each of them specifies the expression to use for computing the lookup table for
13913 the corresponding pixel component values.
13915 The expressions can contain the following constants and functions:
13920 The input width and height.
13923 The input value for the pixel component.
13925 @item ymin, umin, vmin, amin
13926 The minimum allowed component value.
13928 @item ymax, umax, vmax, amax
13929 The maximum allowed component value.
13932 All expressions default to "val".
13934 @subsection Examples
13938 Change too high luma values to gradient:
13940 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'"
13946 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
13947 Ratio) between two input videos.
13949 This filter takes in input two input videos, the first input is
13950 considered the "main" source and is passed unchanged to the
13951 output. The second input is used as a "reference" video for computing
13954 Both video inputs must have the same resolution and pixel format for
13955 this filter to work correctly. Also it assumes that both inputs
13956 have the same number of frames, which are compared one by one.
13958 The obtained average PSNR is printed through the logging system.
13960 The filter stores the accumulated MSE (mean squared error) of each
13961 frame, and at the end of the processing it is averaged across all frames
13962 equally, and the following formula is applied to obtain the PSNR:
13965 PSNR = 10*log10(MAX^2/MSE)
13968 Where MAX is the average of the maximum values of each component of the
13971 The description of the accepted parameters follows.
13974 @item stats_file, f
13975 If specified the filter will use the named file to save the PSNR of
13976 each individual frame. When filename equals "-" the data is sent to
13979 @item stats_version
13980 Specifies which version of the stats file format to use. Details of
13981 each format are written below.
13982 Default value is 1.
13984 @item stats_add_max
13985 Determines whether the max value is output to the stats log.
13986 Default value is 0.
13987 Requires stats_version >= 2. If this is set and stats_version < 2,
13988 the filter will return an error.
13991 This filter also supports the @ref{framesync} options.
13993 The file printed if @var{stats_file} is selected, contains a sequence of
13994 key/value pairs of the form @var{key}:@var{value} for each compared
13997 If a @var{stats_version} greater than 1 is specified, a header line precedes
13998 the list of per-frame-pair stats, with key value pairs following the frame
13999 format with the following parameters:
14002 @item psnr_log_version
14003 The version of the log file format. Will match @var{stats_version}.
14006 A comma separated list of the per-frame-pair parameters included in
14010 A description of each shown per-frame-pair parameter follows:
14014 sequential number of the input frame, starting from 1
14017 Mean Square Error pixel-by-pixel average difference of the compared
14018 frames, averaged over all the image components.
14020 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
14021 Mean Square Error pixel-by-pixel average difference of the compared
14022 frames for the component specified by the suffix.
14024 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
14025 Peak Signal to Noise ratio of the compared frames for the component
14026 specified by the suffix.
14028 @item max_avg, max_y, max_u, max_v
14029 Maximum allowed value for each channel, and average over all
14035 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
14036 [main][ref] psnr="stats_file=stats.log" [out]
14039 On this example the input file being processed is compared with the
14040 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
14041 is stored in @file{stats.log}.
14046 Pulldown reversal (inverse telecine) filter, capable of handling mixed
14047 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
14050 The pullup filter is designed to take advantage of future context in making
14051 its decisions. This filter is stateless in the sense that it does not lock
14052 onto a pattern to follow, but it instead looks forward to the following
14053 fields in order to identify matches and rebuild progressive frames.
14055 To produce content with an even framerate, insert the fps filter after
14056 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
14057 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
14059 The filter accepts the following options:
14066 These options set the amount of "junk" to ignore at the left, right, top, and
14067 bottom of the image, respectively. Left and right are in units of 8 pixels,
14068 while top and bottom are in units of 2 lines.
14069 The default is 8 pixels on each side.
14072 Set the strict breaks. Setting this option to 1 will reduce the chances of
14073 filter generating an occasional mismatched frame, but it may also cause an
14074 excessive number of frames to be dropped during high motion sequences.
14075 Conversely, setting it to -1 will make filter match fields more easily.
14076 This may help processing of video where there is slight blurring between
14077 the fields, but may also cause there to be interlaced frames in the output.
14078 Default value is @code{0}.
14081 Set the metric plane to use. It accepts the following values:
14087 Use chroma blue plane.
14090 Use chroma red plane.
14093 This option may be set to use chroma plane instead of the default luma plane
14094 for doing filter's computations. This may improve accuracy on very clean
14095 source material, but more likely will decrease accuracy, especially if there
14096 is chroma noise (rainbow effect) or any grayscale video.
14097 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
14098 load and make pullup usable in realtime on slow machines.
14101 For best results (without duplicated frames in the output file) it is
14102 necessary to change the output frame rate. For example, to inverse
14103 telecine NTSC input:
14105 ffmpeg -i input -vf pullup -r 24000/1001 ...
14110 Change video quantization parameters (QP).
14112 The filter accepts the following option:
14116 Set expression for quantization parameter.
14119 The expression is evaluated through the eval API and can contain, among others,
14120 the following constants:
14124 1 if index is not 129, 0 otherwise.
14127 Sequential index starting from -129 to 128.
14130 @subsection Examples
14134 Some equation like:
14142 Flush video frames from internal cache of frames into a random order.
14143 No frame is discarded.
14144 Inspired by @ref{frei0r} nervous filter.
14148 Set size in number of frames of internal cache, in range from @code{2} to
14149 @code{512}. Default is @code{30}.
14152 Set seed for random number generator, must be an integer included between
14153 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
14154 less than @code{0}, the filter will try to use a good random seed on a
14158 @section readeia608
14160 Read closed captioning (EIA-608) information from the top lines of a video frame.
14162 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
14163 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
14164 with EIA-608 data (starting from 0). A description of each metadata value follows:
14167 @item lavfi.readeia608.X.cc
14168 The two bytes stored as EIA-608 data (printed in hexadecimal).
14170 @item lavfi.readeia608.X.line
14171 The number of the line on which the EIA-608 data was identified and read.
14174 This filter accepts the following options:
14178 Set the line to start scanning for EIA-608 data. Default is @code{0}.
14181 Set the line to end scanning for EIA-608 data. Default is @code{29}.
14184 Set minimal acceptable amplitude change for sync codes detection.
14185 Default is @code{0.2}. Allowed range is @code{[0.001 - 1]}.
14188 Set the ratio of width reserved for sync code detection.
14189 Default is @code{0.27}. Allowed range is @code{[0.01 - 0.7]}.
14192 Set the max peaks height difference for sync code detection.
14193 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
14196 Set max peaks period difference for sync code detection.
14197 Default is @code{0.1}. Allowed range is @code{[0.0 - 0.5]}.
14200 Set the first two max start code bits differences.
14201 Default is @code{0.02}. Allowed range is @code{[0.0 - 0.5]}.
14204 Set the minimum ratio of bits height compared to 3rd start code bit.
14205 Default is @code{0.75}. Allowed range is @code{[0.01 - 1]}.
14208 Set the white color threshold. Default is @code{0.35}. Allowed range is @code{[0.1 - 1]}.
14211 Set the black color threshold. Default is @code{0.15}. Allowed range is @code{[0.0 - 0.5]}.
14214 Enable checking the parity bit. In the event of a parity error, the filter will output
14215 @code{0x00} for that character. Default is false.
14218 @subsection Examples
14222 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
14224 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
14230 Read vertical interval timecode (VITC) information from the top lines of a
14233 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
14234 timecode value, if a valid timecode has been detected. Further metadata key
14235 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
14236 timecode data has been found or not.
14238 This filter accepts the following options:
14242 Set the maximum number of lines to scan for VITC data. If the value is set to
14243 @code{-1} the full video frame is scanned. Default is @code{45}.
14246 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
14247 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
14250 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
14251 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
14254 @subsection Examples
14258 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
14259 draw @code{--:--:--:--} as a placeholder:
14261 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
14267 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
14269 Destination pixel at position (X, Y) will be picked from source (x, y) position
14270 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
14271 value for pixel will be used for destination pixel.
14273 Xmap and Ymap input video streams must be of same dimensions. Output video stream
14274 will have Xmap/Ymap video stream dimensions.
14275 Xmap and Ymap input video streams are 16bit depth, single channel.
14277 @section removegrain
14279 The removegrain filter is a spatial denoiser for progressive video.
14283 Set mode for the first plane.
14286 Set mode for the second plane.
14289 Set mode for the third plane.
14292 Set mode for the fourth plane.
14295 Range of mode is from 0 to 24. Description of each mode follows:
14299 Leave input plane unchanged. Default.
14302 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
14305 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
14308 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
14311 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
14312 This is equivalent to a median filter.
14315 Line-sensitive clipping giving the minimal change.
14318 Line-sensitive clipping, intermediate.
14321 Line-sensitive clipping, intermediate.
14324 Line-sensitive clipping, intermediate.
14327 Line-sensitive clipping on a line where the neighbours pixels are the closest.
14330 Replaces the target pixel with the closest neighbour.
14333 [1 2 1] horizontal and vertical kernel blur.
14339 Bob mode, interpolates top field from the line where the neighbours
14340 pixels are the closest.
14343 Bob mode, interpolates bottom field from the line where the neighbours
14344 pixels are the closest.
14347 Bob mode, interpolates top field. Same as 13 but with a more complicated
14348 interpolation formula.
14351 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
14352 interpolation formula.
14355 Clips the pixel with the minimum and maximum of respectively the maximum and
14356 minimum of each pair of opposite neighbour pixels.
14359 Line-sensitive clipping using opposite neighbours whose greatest distance from
14360 the current pixel is minimal.
14363 Replaces the pixel with the average of its 8 neighbours.
14366 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
14369 Clips pixels using the averages of opposite neighbour.
14372 Same as mode 21 but simpler and faster.
14375 Small edge and halo removal, but reputed useless.
14381 @section removelogo
14383 Suppress a TV station logo, using an image file to determine which
14384 pixels comprise the logo. It works by filling in the pixels that
14385 comprise the logo with neighboring pixels.
14387 The filter accepts the following options:
14391 Set the filter bitmap file, which can be any image format supported by
14392 libavformat. The width and height of the image file must match those of the
14393 video stream being processed.
14396 Pixels in the provided bitmap image with a value of zero are not
14397 considered part of the logo, non-zero pixels are considered part of
14398 the logo. If you use white (255) for the logo and black (0) for the
14399 rest, you will be safe. For making the filter bitmap, it is
14400 recommended to take a screen capture of a black frame with the logo
14401 visible, and then using a threshold filter followed by the erode
14402 filter once or twice.
14404 If needed, little splotches can be fixed manually. Remember that if
14405 logo pixels are not covered, the filter quality will be much
14406 reduced. Marking too many pixels as part of the logo does not hurt as
14407 much, but it will increase the amount of blurring needed to cover over
14408 the image and will destroy more information than necessary, and extra
14409 pixels will slow things down on a large logo.
14411 @section repeatfields
14413 This filter uses the repeat_field flag from the Video ES headers and hard repeats
14414 fields based on its value.
14418 Reverse a video clip.
14420 Warning: This filter requires memory to buffer the entire clip, so trimming
14423 @subsection Examples
14427 Take the first 5 seconds of a clip, and reverse it.
14434 Shift R/G/B/A pixels horizontally and/or vertically.
14436 The filter accepts the following options:
14439 Set amount to shift red horizontally.
14441 Set amount to shift red vertically.
14443 Set amount to shift green horizontally.
14445 Set amount to shift green vertically.
14447 Set amount to shift blue horizontally.
14449 Set amount to shift blue vertically.
14451 Set amount to shift alpha horizontally.
14453 Set amount to shift alpha vertically.
14455 Set edge mode, can be @var{smear}, default, or @var{warp}.
14459 Apply roberts cross operator to input video stream.
14461 The filter accepts the following option:
14465 Set which planes will be processed, unprocessed planes will be copied.
14466 By default value 0xf, all planes will be processed.
14469 Set value which will be multiplied with filtered result.
14472 Set value which will be added to filtered result.
14477 Rotate video by an arbitrary angle expressed in radians.
14479 The filter accepts the following options:
14481 A description of the optional parameters follows.
14484 Set an expression for the angle by which to rotate the input video
14485 clockwise, expressed as a number of radians. A negative value will
14486 result in a counter-clockwise rotation. By default it is set to "0".
14488 This expression is evaluated for each frame.
14491 Set the output width expression, default value is "iw".
14492 This expression is evaluated just once during configuration.
14495 Set the output height expression, default value is "ih".
14496 This expression is evaluated just once during configuration.
14499 Enable bilinear interpolation if set to 1, a value of 0 disables
14500 it. Default value is 1.
14503 Set the color used to fill the output area not covered by the rotated
14504 image. For the general syntax of this option, check the
14505 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
14506 If the special value "none" is selected then no
14507 background is printed (useful for example if the background is never shown).
14509 Default value is "black".
14512 The expressions for the angle and the output size can contain the
14513 following constants and functions:
14517 sequential number of the input frame, starting from 0. It is always NAN
14518 before the first frame is filtered.
14521 time in seconds of the input frame, it is set to 0 when the filter is
14522 configured. It is always NAN before the first frame is filtered.
14526 horizontal and vertical chroma subsample values. For example for the
14527 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14531 the input video width and height
14535 the output width and height, that is the size of the padded area as
14536 specified by the @var{width} and @var{height} expressions
14540 the minimal width/height required for completely containing the input
14541 video rotated by @var{a} radians.
14543 These are only available when computing the @option{out_w} and
14544 @option{out_h} expressions.
14547 @subsection Examples
14551 Rotate the input by PI/6 radians clockwise:
14557 Rotate the input by PI/6 radians counter-clockwise:
14563 Rotate the input by 45 degrees clockwise:
14569 Apply a constant rotation with period T, starting from an angle of PI/3:
14571 rotate=PI/3+2*PI*t/T
14575 Make the input video rotation oscillating with a period of T
14576 seconds and an amplitude of A radians:
14578 rotate=A*sin(2*PI/T*t)
14582 Rotate the video, output size is chosen so that the whole rotating
14583 input video is always completely contained in the output:
14585 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
14589 Rotate the video, reduce the output size so that no background is ever
14592 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
14596 @subsection Commands
14598 The filter supports the following commands:
14602 Set the angle expression.
14603 The command accepts the same syntax of the corresponding option.
14605 If the specified expression is not valid, it is kept at its current
14611 Apply Shape Adaptive Blur.
14613 The filter accepts the following options:
14616 @item luma_radius, lr
14617 Set luma blur filter strength, must be a value in range 0.1-4.0, default
14618 value is 1.0. A greater value will result in a more blurred image, and
14619 in slower processing.
14621 @item luma_pre_filter_radius, lpfr
14622 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
14625 @item luma_strength, ls
14626 Set luma maximum difference between pixels to still be considered, must
14627 be a value in the 0.1-100.0 range, default value is 1.0.
14629 @item chroma_radius, cr
14630 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
14631 greater value will result in a more blurred image, and in slower
14634 @item chroma_pre_filter_radius, cpfr
14635 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
14637 @item chroma_strength, cs
14638 Set chroma maximum difference between pixels to still be considered,
14639 must be a value in the -0.9-100.0 range.
14642 Each chroma option value, if not explicitly specified, is set to the
14643 corresponding luma option value.
14648 Scale (resize) the input video, using the libswscale library.
14650 The scale filter forces the output display aspect ratio to be the same
14651 of the input, by changing the output sample aspect ratio.
14653 If the input image format is different from the format requested by
14654 the next filter, the scale filter will convert the input to the
14657 @subsection Options
14658 The filter accepts the following options, or any of the options
14659 supported by the libswscale scaler.
14661 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
14662 the complete list of scaler options.
14667 Set the output video dimension expression. Default value is the input
14670 If the @var{width} or @var{w} value is 0, the input width is used for
14671 the output. If the @var{height} or @var{h} value is 0, the input height
14672 is used for the output.
14674 If one and only one of the values is -n with n >= 1, the scale filter
14675 will use a value that maintains the aspect ratio of the input image,
14676 calculated from the other specified dimension. After that it will,
14677 however, make sure that the calculated dimension is divisible by n and
14678 adjust the value if necessary.
14680 If both values are -n with n >= 1, the behavior will be identical to
14681 both values being set to 0 as previously detailed.
14683 See below for the list of accepted constants for use in the dimension
14687 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
14691 Only evaluate expressions once during the filter initialization or when a command is processed.
14694 Evaluate expressions for each incoming frame.
14698 Default value is @samp{init}.
14702 Set the interlacing mode. It accepts the following values:
14706 Force interlaced aware scaling.
14709 Do not apply interlaced scaling.
14712 Select interlaced aware scaling depending on whether the source frames
14713 are flagged as interlaced or not.
14716 Default value is @samp{0}.
14719 Set libswscale scaling flags. See
14720 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
14721 complete list of values. If not explicitly specified the filter applies
14725 @item param0, param1
14726 Set libswscale input parameters for scaling algorithms that need them. See
14727 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
14728 complete documentation. If not explicitly specified the filter applies
14734 Set the video size. For the syntax of this option, check the
14735 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
14737 @item in_color_matrix
14738 @item out_color_matrix
14739 Set in/output YCbCr color space type.
14741 This allows the autodetected value to be overridden as well as allows forcing
14742 a specific value used for the output and encoder.
14744 If not specified, the color space type depends on the pixel format.
14750 Choose automatically.
14753 Format conforming to International Telecommunication Union (ITU)
14754 Recommendation BT.709.
14757 Set color space conforming to the United States Federal Communications
14758 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
14761 Set color space conforming to:
14765 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
14768 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
14771 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
14776 Set color space conforming to SMPTE ST 240:1999.
14781 Set in/output YCbCr sample range.
14783 This allows the autodetected value to be overridden as well as allows forcing
14784 a specific value used for the output and encoder. If not specified, the
14785 range depends on the pixel format. Possible values:
14789 Choose automatically.
14792 Set full range (0-255 in case of 8-bit luma).
14794 @item mpeg/limited/tv
14795 Set "MPEG" range (16-235 in case of 8-bit luma).
14798 @item force_original_aspect_ratio
14799 Enable decreasing or increasing output video width or height if necessary to
14800 keep the original aspect ratio. Possible values:
14804 Scale the video as specified and disable this feature.
14807 The output video dimensions will automatically be decreased if needed.
14810 The output video dimensions will automatically be increased if needed.
14814 One useful instance of this option is that when you know a specific device's
14815 maximum allowed resolution, you can use this to limit the output video to
14816 that, while retaining the aspect ratio. For example, device A allows
14817 1280x720 playback, and your video is 1920x800. Using this option (set it to
14818 decrease) and specifying 1280x720 to the command line makes the output
14821 Please note that this is a different thing than specifying -1 for @option{w}
14822 or @option{h}, you still need to specify the output resolution for this option
14827 The values of the @option{w} and @option{h} options are expressions
14828 containing the following constants:
14833 The input width and height
14837 These are the same as @var{in_w} and @var{in_h}.
14841 The output (scaled) width and height
14845 These are the same as @var{out_w} and @var{out_h}
14848 The same as @var{iw} / @var{ih}
14851 input sample aspect ratio
14854 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
14858 horizontal and vertical input chroma subsample values. For example for the
14859 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14863 horizontal and vertical output chroma subsample values. For example for the
14864 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14867 @subsection Examples
14871 Scale the input video to a size of 200x100
14876 This is equivalent to:
14887 Specify a size abbreviation for the output size:
14892 which can also be written as:
14898 Scale the input to 2x:
14900 scale=w=2*iw:h=2*ih
14904 The above is the same as:
14906 scale=2*in_w:2*in_h
14910 Scale the input to 2x with forced interlaced scaling:
14912 scale=2*iw:2*ih:interl=1
14916 Scale the input to half size:
14918 scale=w=iw/2:h=ih/2
14922 Increase the width, and set the height to the same size:
14928 Seek Greek harmony:
14935 Increase the height, and set the width to 3/2 of the height:
14937 scale=w=3/2*oh:h=3/5*ih
14941 Increase the size, making the size a multiple of the chroma
14944 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
14948 Increase the width to a maximum of 500 pixels,
14949 keeping the same aspect ratio as the input:
14951 scale=w='min(500\, iw*3/2):h=-1'
14955 Make pixels square by combining scale and setsar:
14957 scale='trunc(ih*dar):ih',setsar=1/1
14961 Make pixels square by combining scale and setsar,
14962 making sure the resulting resolution is even (required by some codecs):
14964 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
14968 @subsection Commands
14970 This filter supports the following commands:
14974 Set the output video dimension expression.
14975 The command accepts the same syntax of the corresponding option.
14977 If the specified expression is not valid, it is kept at its current
14983 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
14984 format conversion on CUDA video frames. Setting the output width and height
14985 works in the same way as for the @var{scale} filter.
14987 The following additional options are accepted:
14990 The pixel format of the output CUDA frames. If set to the string "same" (the
14991 default), the input format will be kept. Note that automatic format negotiation
14992 and conversion is not yet supported for hardware frames
14995 The interpolation algorithm used for resizing. One of the following:
15002 @item cubic2p_bspline
15003 2-parameter cubic (B=1, C=0)
15005 @item cubic2p_catmullrom
15006 2-parameter cubic (B=0, C=1/2)
15008 @item cubic2p_b05c03
15009 2-parameter cubic (B=1/2, C=3/10)
15021 Scale (resize) the input video, based on a reference video.
15023 See the scale filter for available options, scale2ref supports the same but
15024 uses the reference video instead of the main input as basis. scale2ref also
15025 supports the following additional constants for the @option{w} and
15026 @option{h} options:
15031 The main input video's width and height
15034 The same as @var{main_w} / @var{main_h}
15037 The main input video's sample aspect ratio
15039 @item main_dar, mdar
15040 The main input video's display aspect ratio. Calculated from
15041 @code{(main_w / main_h) * main_sar}.
15045 The main input video's horizontal and vertical chroma subsample values.
15046 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
15050 @subsection Examples
15054 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
15056 'scale2ref[b][a];[a][b]overlay'
15060 @anchor{selectivecolor}
15061 @section selectivecolor
15063 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
15064 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
15065 by the "purity" of the color (that is, how saturated it already is).
15067 This filter is similar to the Adobe Photoshop Selective Color tool.
15069 The filter accepts the following options:
15072 @item correction_method
15073 Select color correction method.
15075 Available values are:
15078 Specified adjustments are applied "as-is" (added/subtracted to original pixel
15081 Specified adjustments are relative to the original component value.
15083 Default is @code{absolute}.
15085 Adjustments for red pixels (pixels where the red component is the maximum)
15087 Adjustments for yellow pixels (pixels where the blue component is the minimum)
15089 Adjustments for green pixels (pixels where the green component is the maximum)
15091 Adjustments for cyan pixels (pixels where the red component is the minimum)
15093 Adjustments for blue pixels (pixels where the blue component is the maximum)
15095 Adjustments for magenta pixels (pixels where the green component is the minimum)
15097 Adjustments for white pixels (pixels where all components are greater than 128)
15099 Adjustments for all pixels except pure black and pure white
15101 Adjustments for black pixels (pixels where all components are lesser than 128)
15103 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
15106 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
15107 4 space separated floating point adjustment values in the [-1,1] range,
15108 respectively to adjust the amount of cyan, magenta, yellow and black for the
15109 pixels of its range.
15111 @subsection Examples
15115 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
15116 increase magenta by 27% in blue areas:
15118 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
15122 Use a Photoshop selective color preset:
15124 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
15128 @anchor{separatefields}
15129 @section separatefields
15131 The @code{separatefields} takes a frame-based video input and splits
15132 each frame into its components fields, producing a new half height clip
15133 with twice the frame rate and twice the frame count.
15135 This filter use field-dominance information in frame to decide which
15136 of each pair of fields to place first in the output.
15137 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
15139 @section setdar, setsar
15141 The @code{setdar} filter sets the Display Aspect Ratio for the filter
15144 This is done by changing the specified Sample (aka Pixel) Aspect
15145 Ratio, according to the following equation:
15147 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
15150 Keep in mind that the @code{setdar} filter does not modify the pixel
15151 dimensions of the video frame. Also, the display aspect ratio set by
15152 this filter may be changed by later filters in the filterchain,
15153 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
15156 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
15157 the filter output video.
15159 Note that as a consequence of the application of this filter, the
15160 output display aspect ratio will change according to the equation
15163 Keep in mind that the sample aspect ratio set by the @code{setsar}
15164 filter may be changed by later filters in the filterchain, e.g. if
15165 another "setsar" or a "setdar" filter is applied.
15167 It accepts the following parameters:
15170 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
15171 Set the aspect ratio used by the filter.
15173 The parameter can be a floating point number string, an expression, or
15174 a string of the form @var{num}:@var{den}, where @var{num} and
15175 @var{den} are the numerator and denominator of the aspect ratio. If
15176 the parameter is not specified, it is assumed the value "0".
15177 In case the form "@var{num}:@var{den}" is used, the @code{:} character
15181 Set the maximum integer value to use for expressing numerator and
15182 denominator when reducing the expressed aspect ratio to a rational.
15183 Default value is @code{100}.
15187 The parameter @var{sar} is an expression containing
15188 the following constants:
15192 These are approximated values for the mathematical constants e
15193 (Euler's number), pi (Greek pi), and phi (the golden ratio).
15196 The input width and height.
15199 These are the same as @var{w} / @var{h}.
15202 The input sample aspect ratio.
15205 The input display aspect ratio. It is the same as
15206 (@var{w} / @var{h}) * @var{sar}.
15209 Horizontal and vertical chroma subsample values. For example, for the
15210 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15213 @subsection Examples
15218 To change the display aspect ratio to 16:9, specify one of the following:
15225 To change the sample aspect ratio to 10:11, specify:
15231 To set a display aspect ratio of 16:9, and specify a maximum integer value of
15232 1000 in the aspect ratio reduction, use the command:
15234 setdar=ratio=16/9:max=1000
15242 Force field for the output video frame.
15244 The @code{setfield} filter marks the interlace type field for the
15245 output frames. It does not change the input frame, but only sets the
15246 corresponding property, which affects how the frame is treated by
15247 following filters (e.g. @code{fieldorder} or @code{yadif}).
15249 The filter accepts the following options:
15254 Available values are:
15258 Keep the same field property.
15261 Mark the frame as bottom-field-first.
15264 Mark the frame as top-field-first.
15267 Mark the frame as progressive.
15274 Force frame parameter for the output video frame.
15276 The @code{setparams} filter marks interlace and color range for the
15277 output frames. It does not change the input frame, but only sets the
15278 corresponding property, which affects how the frame is treated by
15283 Available values are:
15287 Keep the same field property (default).
15290 Mark the frame as bottom-field-first.
15293 Mark the frame as top-field-first.
15296 Mark the frame as progressive.
15300 Available values are:
15304 Keep the same color range property (default).
15306 @item unspecified, unknown
15307 Mark the frame as unspecified color range.
15309 @item limited, tv, mpeg
15310 Mark the frame as limited range.
15312 @item full, pc, jpeg
15313 Mark the frame as full range.
15316 @item color_primaries
15317 Set the color primaries.
15318 Available values are:
15322 Keep the same color primaries property (default).
15339 Set the color transfer.
15340 Available values are:
15344 Keep the same color trc property (default).
15366 Set the colorspace.
15367 Available values are:
15371 Keep the same colorspace property (default).
15384 @item chroma-derived-nc
15385 @item chroma-derived-c
15392 Show a line containing various information for each input video frame.
15393 The input video is not modified.
15395 This filter supports the following options:
15399 Calculate checksums of each plane. By default enabled.
15402 The shown line contains a sequence of key/value pairs of the form
15403 @var{key}:@var{value}.
15405 The following values are shown in the output:
15409 The (sequential) number of the input frame, starting from 0.
15412 The Presentation TimeStamp of the input frame, expressed as a number of
15413 time base units. The time base unit depends on the filter input pad.
15416 The Presentation TimeStamp of the input frame, expressed as a number of
15420 The position of the frame in the input stream, or -1 if this information is
15421 unavailable and/or meaningless (for example in case of synthetic video).
15424 The pixel format name.
15427 The sample aspect ratio of the input frame, expressed in the form
15428 @var{num}/@var{den}.
15431 The size of the input frame. For the syntax of this option, check the
15432 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
15435 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
15436 for bottom field first).
15439 This is 1 if the frame is a key frame, 0 otherwise.
15442 The picture type of the input frame ("I" for an I-frame, "P" for a
15443 P-frame, "B" for a B-frame, or "?" for an unknown type).
15444 Also refer to the documentation of the @code{AVPictureType} enum and of
15445 the @code{av_get_picture_type_char} function defined in
15446 @file{libavutil/avutil.h}.
15449 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
15451 @item plane_checksum
15452 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
15453 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
15456 @section showpalette
15458 Displays the 256 colors palette of each frame. This filter is only relevant for
15459 @var{pal8} pixel format frames.
15461 It accepts the following option:
15465 Set the size of the box used to represent one palette color entry. Default is
15466 @code{30} (for a @code{30x30} pixel box).
15469 @section shuffleframes
15471 Reorder and/or duplicate and/or drop video frames.
15473 It accepts the following parameters:
15477 Set the destination indexes of input frames.
15478 This is space or '|' separated list of indexes that maps input frames to output
15479 frames. Number of indexes also sets maximal value that each index may have.
15480 '-1' index have special meaning and that is to drop frame.
15483 The first frame has the index 0. The default is to keep the input unchanged.
15485 @subsection Examples
15489 Swap second and third frame of every three frames of the input:
15491 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
15495 Swap 10th and 1st frame of every ten frames of the input:
15497 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
15501 @section shuffleplanes
15503 Reorder and/or duplicate video planes.
15505 It accepts the following parameters:
15510 The index of the input plane to be used as the first output plane.
15513 The index of the input plane to be used as the second output plane.
15516 The index of the input plane to be used as the third output plane.
15519 The index of the input plane to be used as the fourth output plane.
15523 The first plane has the index 0. The default is to keep the input unchanged.
15525 @subsection Examples
15529 Swap the second and third planes of the input:
15531 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
15535 @anchor{signalstats}
15536 @section signalstats
15537 Evaluate various visual metrics that assist in determining issues associated
15538 with the digitization of analog video media.
15540 By default the filter will log these metadata values:
15544 Display the minimal Y value contained within the input frame. Expressed in
15548 Display the Y value at the 10% percentile within the input frame. Expressed in
15552 Display the average Y value within the input frame. Expressed in range of
15556 Display the Y value at the 90% percentile within the input frame. Expressed in
15560 Display the maximum Y value contained within the input frame. Expressed in
15564 Display the minimal U value contained within the input frame. Expressed in
15568 Display the U value at the 10% percentile within the input frame. Expressed in
15572 Display the average U value within the input frame. Expressed in range of
15576 Display the U value at the 90% percentile within the input frame. Expressed in
15580 Display the maximum U value contained within the input frame. Expressed in
15584 Display the minimal V value contained within the input frame. Expressed in
15588 Display the V value at the 10% percentile within the input frame. Expressed in
15592 Display the average V value within the input frame. Expressed in range of
15596 Display the V value at the 90% percentile within the input frame. Expressed in
15600 Display the maximum V value contained within the input frame. Expressed in
15604 Display the minimal saturation value contained within the input frame.
15605 Expressed in range of [0-~181.02].
15608 Display the saturation value at the 10% percentile within the input frame.
15609 Expressed in range of [0-~181.02].
15612 Display the average saturation value within the input frame. Expressed in range
15616 Display the saturation value at the 90% percentile within the input frame.
15617 Expressed in range of [0-~181.02].
15620 Display the maximum saturation value contained within the input frame.
15621 Expressed in range of [0-~181.02].
15624 Display the median value for hue within the input frame. Expressed in range of
15628 Display the average value for hue within the input frame. Expressed in range of
15632 Display the average of sample value difference between all values of the Y
15633 plane in the current frame and corresponding values of the previous input frame.
15634 Expressed in range of [0-255].
15637 Display the average of sample value difference between all values of the U
15638 plane in the current frame and corresponding values of the previous input frame.
15639 Expressed in range of [0-255].
15642 Display the average of sample value difference between all values of the V
15643 plane in the current frame and corresponding values of the previous input frame.
15644 Expressed in range of [0-255].
15647 Display bit depth of Y plane in current frame.
15648 Expressed in range of [0-16].
15651 Display bit depth of U plane in current frame.
15652 Expressed in range of [0-16].
15655 Display bit depth of V plane in current frame.
15656 Expressed in range of [0-16].
15659 The filter accepts the following options:
15665 @option{stat} specify an additional form of image analysis.
15666 @option{out} output video with the specified type of pixel highlighted.
15668 Both options accept the following values:
15672 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
15673 unlike the neighboring pixels of the same field. Examples of temporal outliers
15674 include the results of video dropouts, head clogs, or tape tracking issues.
15677 Identify @var{vertical line repetition}. Vertical line repetition includes
15678 similar rows of pixels within a frame. In born-digital video vertical line
15679 repetition is common, but this pattern is uncommon in video digitized from an
15680 analog source. When it occurs in video that results from the digitization of an
15681 analog source it can indicate concealment from a dropout compensator.
15684 Identify pixels that fall outside of legal broadcast range.
15688 Set the highlight color for the @option{out} option. The default color is
15692 @subsection Examples
15696 Output data of various video metrics:
15698 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
15702 Output specific data about the minimum and maximum values of the Y plane per frame:
15704 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
15708 Playback video while highlighting pixels that are outside of broadcast range in red.
15710 ffplay example.mov -vf signalstats="out=brng:color=red"
15714 Playback video with signalstats metadata drawn over the frame.
15716 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
15719 The contents of signalstat_drawtext.txt used in the command are:
15722 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
15723 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
15724 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
15725 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
15733 Calculates the MPEG-7 Video Signature. The filter can handle more than one
15734 input. In this case the matching between the inputs can be calculated additionally.
15735 The filter always passes through the first input. The signature of each stream can
15736 be written into a file.
15738 It accepts the following options:
15742 Enable or disable the matching process.
15744 Available values are:
15748 Disable the calculation of a matching (default).
15750 Calculate the matching for the whole video and output whether the whole video
15751 matches or only parts.
15753 Calculate only until a matching is found or the video ends. Should be faster in
15758 Set the number of inputs. The option value must be a non negative integer.
15759 Default value is 1.
15762 Set the path to which the output is written. If there is more than one input,
15763 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
15764 integer), that will be replaced with the input number. If no filename is
15765 specified, no output will be written. This is the default.
15768 Choose the output format.
15770 Available values are:
15774 Use the specified binary representation (default).
15776 Use the specified xml representation.
15780 Set threshold to detect one word as similar. The option value must be an integer
15781 greater than zero. The default value is 9000.
15784 Set threshold to detect all words as similar. The option value must be an integer
15785 greater than zero. The default value is 60000.
15788 Set threshold to detect frames as similar. The option value must be an integer
15789 greater than zero. The default value is 116.
15792 Set the minimum length of a sequence in frames to recognize it as matching
15793 sequence. The option value must be a non negative integer value.
15794 The default value is 0.
15797 Set the minimum relation, that matching frames to all frames must have.
15798 The option value must be a double value between 0 and 1. The default value is 0.5.
15801 @subsection Examples
15805 To calculate the signature of an input video and store it in signature.bin:
15807 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
15811 To detect whether two videos match and store the signatures in XML format in
15812 signature0.xml and signature1.xml:
15814 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 -
15822 Blur the input video without impacting the outlines.
15824 It accepts the following options:
15827 @item luma_radius, lr
15828 Set the luma radius. The option value must be a float number in
15829 the range [0.1,5.0] that specifies the variance of the gaussian filter
15830 used to blur the image (slower if larger). Default value is 1.0.
15832 @item luma_strength, ls
15833 Set the luma strength. The option value must be a float number
15834 in the range [-1.0,1.0] that configures the blurring. A value included
15835 in [0.0,1.0] will blur the image whereas a value included in
15836 [-1.0,0.0] will sharpen the image. Default value is 1.0.
15838 @item luma_threshold, lt
15839 Set the luma threshold used as a coefficient to determine
15840 whether a pixel should be blurred or not. The option value must be an
15841 integer in the range [-30,30]. A value of 0 will filter all the image,
15842 a value included in [0,30] will filter flat areas and a value included
15843 in [-30,0] will filter edges. Default value is 0.
15845 @item chroma_radius, cr
15846 Set the chroma radius. The option value must be a float number in
15847 the range [0.1,5.0] that specifies the variance of the gaussian filter
15848 used to blur the image (slower if larger). Default value is @option{luma_radius}.
15850 @item chroma_strength, cs
15851 Set the chroma strength. The option value must be a float number
15852 in the range [-1.0,1.0] that configures the blurring. A value included
15853 in [0.0,1.0] will blur the image whereas a value included in
15854 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
15856 @item chroma_threshold, ct
15857 Set the chroma threshold used as a coefficient to determine
15858 whether a pixel should be blurred or not. The option value must be an
15859 integer in the range [-30,30]. A value of 0 will filter all the image,
15860 a value included in [0,30] will filter flat areas and a value included
15861 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
15864 If a chroma option is not explicitly set, the corresponding luma value
15869 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
15871 This filter takes in input two input videos, the first input is
15872 considered the "main" source and is passed unchanged to the
15873 output. The second input is used as a "reference" video for computing
15876 Both video inputs must have the same resolution and pixel format for
15877 this filter to work correctly. Also it assumes that both inputs
15878 have the same number of frames, which are compared one by one.
15880 The filter stores the calculated SSIM of each frame.
15882 The description of the accepted parameters follows.
15885 @item stats_file, f
15886 If specified the filter will use the named file to save the SSIM of
15887 each individual frame. When filename equals "-" the data is sent to
15891 The file printed if @var{stats_file} is selected, contains a sequence of
15892 key/value pairs of the form @var{key}:@var{value} for each compared
15895 A description of each shown parameter follows:
15899 sequential number of the input frame, starting from 1
15901 @item Y, U, V, R, G, B
15902 SSIM of the compared frames for the component specified by the suffix.
15905 SSIM of the compared frames for the whole frame.
15908 Same as above but in dB representation.
15911 This filter also supports the @ref{framesync} options.
15915 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
15916 [main][ref] ssim="stats_file=stats.log" [out]
15919 On this example the input file being processed is compared with the
15920 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
15921 is stored in @file{stats.log}.
15923 Another example with both psnr and ssim at same time:
15925 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
15930 Convert between different stereoscopic image formats.
15932 The filters accept the following options:
15936 Set stereoscopic image format of input.
15938 Available values for input image formats are:
15941 side by side parallel (left eye left, right eye right)
15944 side by side crosseye (right eye left, left eye right)
15947 side by side parallel with half width resolution
15948 (left eye left, right eye right)
15951 side by side crosseye with half width resolution
15952 (right eye left, left eye right)
15955 above-below (left eye above, right eye below)
15958 above-below (right eye above, left eye below)
15961 above-below with half height resolution
15962 (left eye above, right eye below)
15965 above-below with half height resolution
15966 (right eye above, left eye below)
15969 alternating frames (left eye first, right eye second)
15972 alternating frames (right eye first, left eye second)
15975 interleaved rows (left eye has top row, right eye starts on next row)
15978 interleaved rows (right eye has top row, left eye starts on next row)
15981 interleaved columns, left eye first
15984 interleaved columns, right eye first
15986 Default value is @samp{sbsl}.
15990 Set stereoscopic image format of output.
15994 side by side parallel (left eye left, right eye right)
15997 side by side crosseye (right eye left, left eye right)
16000 side by side parallel with half width resolution
16001 (left eye left, right eye right)
16004 side by side crosseye with half width resolution
16005 (right eye left, left eye right)
16008 above-below (left eye above, right eye below)
16011 above-below (right eye above, left eye below)
16014 above-below with half height resolution
16015 (left eye above, right eye below)
16018 above-below with half height resolution
16019 (right eye above, left eye below)
16022 alternating frames (left eye first, right eye second)
16025 alternating frames (right eye first, left eye second)
16028 interleaved rows (left eye has top row, right eye starts on next row)
16031 interleaved rows (right eye has top row, left eye starts on next row)
16034 anaglyph red/blue gray
16035 (red filter on left eye, blue filter on right eye)
16038 anaglyph red/green gray
16039 (red filter on left eye, green filter on right eye)
16042 anaglyph red/cyan gray
16043 (red filter on left eye, cyan filter on right eye)
16046 anaglyph red/cyan half colored
16047 (red filter on left eye, cyan filter on right eye)
16050 anaglyph red/cyan color
16051 (red filter on left eye, cyan filter on right eye)
16054 anaglyph red/cyan color optimized with the least squares projection of dubois
16055 (red filter on left eye, cyan filter on right eye)
16058 anaglyph green/magenta gray
16059 (green filter on left eye, magenta filter on right eye)
16062 anaglyph green/magenta half colored
16063 (green filter on left eye, magenta filter on right eye)
16066 anaglyph green/magenta colored
16067 (green filter on left eye, magenta filter on right eye)
16070 anaglyph green/magenta color optimized with the least squares projection of dubois
16071 (green filter on left eye, magenta filter on right eye)
16074 anaglyph yellow/blue gray
16075 (yellow filter on left eye, blue filter on right eye)
16078 anaglyph yellow/blue half colored
16079 (yellow filter on left eye, blue filter on right eye)
16082 anaglyph yellow/blue colored
16083 (yellow filter on left eye, blue filter on right eye)
16086 anaglyph yellow/blue color optimized with the least squares projection of dubois
16087 (yellow filter on left eye, blue filter on right eye)
16090 mono output (left eye only)
16093 mono output (right eye only)
16096 checkerboard, left eye first
16099 checkerboard, right eye first
16102 interleaved columns, left eye first
16105 interleaved columns, right eye first
16111 Default value is @samp{arcd}.
16114 @subsection Examples
16118 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
16124 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
16130 @section streamselect, astreamselect
16131 Select video or audio streams.
16133 The filter accepts the following options:
16137 Set number of inputs. Default is 2.
16140 Set input indexes to remap to outputs.
16143 @subsection Commands
16145 The @code{streamselect} and @code{astreamselect} filter supports the following
16150 Set input indexes to remap to outputs.
16153 @subsection Examples
16157 Select first 5 seconds 1st stream and rest of time 2nd stream:
16159 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
16163 Same as above, but for audio:
16165 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
16170 Apply sobel operator to input video stream.
16172 The filter accepts the following option:
16176 Set which planes will be processed, unprocessed planes will be copied.
16177 By default value 0xf, all planes will be processed.
16180 Set value which will be multiplied with filtered result.
16183 Set value which will be added to filtered result.
16189 Apply a simple postprocessing filter that compresses and decompresses the image
16190 at several (or - in the case of @option{quality} level @code{6} - all) shifts
16191 and average the results.
16193 The filter accepts the following options:
16197 Set quality. This option defines the number of levels for averaging. It accepts
16198 an integer in the range 0-6. If set to @code{0}, the filter will have no
16199 effect. A value of @code{6} means the higher quality. For each increment of
16200 that value the speed drops by a factor of approximately 2. Default value is
16204 Force a constant quantization parameter. If not set, the filter will use the QP
16205 from the video stream (if available).
16208 Set thresholding mode. Available modes are:
16212 Set hard thresholding (default).
16214 Set soft thresholding (better de-ringing effect, but likely blurrier).
16217 @item use_bframe_qp
16218 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
16219 option may cause flicker since the B-Frames have often larger QP. Default is
16220 @code{0} (not enabled).
16225 Scale the input by applying one of the super-resolution methods based on
16226 convolutional neural networks. Supported models:
16230 Super-Resolution Convolutional Neural Network model (SRCNN).
16231 See @url{https://arxiv.org/abs/1501.00092}.
16234 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
16235 See @url{https://arxiv.org/abs/1609.05158}.
16238 Training scripts as well as scripts for model generation are provided in
16239 the repository at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
16241 The filter accepts the following options:
16245 Specify which DNN backend to use for model loading and execution. This option accepts
16246 the following values:
16250 Native implementation of DNN loading and execution.
16253 TensorFlow backend. To enable this backend you
16254 need to install the TensorFlow for C library (see
16255 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
16256 @code{--enable-libtensorflow}
16259 Default value is @samp{native}.
16262 Set path to model file specifying network architecture and its parameters.
16263 Note that different backends use different file formats. TensorFlow backend
16264 can load files for both formats, while native backend can load files for only
16268 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
16269 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
16270 input upscaled using bicubic upscaling with proper scale factor.
16276 Draw subtitles on top of input video using the libass library.
16278 To enable compilation of this filter you need to configure FFmpeg with
16279 @code{--enable-libass}. This filter also requires a build with libavcodec and
16280 libavformat to convert the passed subtitles file to ASS (Advanced Substation
16281 Alpha) subtitles format.
16283 The filter accepts the following options:
16287 Set the filename of the subtitle file to read. It must be specified.
16289 @item original_size
16290 Specify the size of the original video, the video for which the ASS file
16291 was composed. For the syntax of this option, check the
16292 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16293 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
16294 correctly scale the fonts if the aspect ratio has been changed.
16297 Set a directory path containing fonts that can be used by the filter.
16298 These fonts will be used in addition to whatever the font provider uses.
16301 Process alpha channel, by default alpha channel is untouched.
16304 Set subtitles input character encoding. @code{subtitles} filter only. Only
16305 useful if not UTF-8.
16307 @item stream_index, si
16308 Set subtitles stream index. @code{subtitles} filter only.
16311 Override default style or script info parameters of the subtitles. It accepts a
16312 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
16315 If the first key is not specified, it is assumed that the first value
16316 specifies the @option{filename}.
16318 For example, to render the file @file{sub.srt} on top of the input
16319 video, use the command:
16324 which is equivalent to:
16326 subtitles=filename=sub.srt
16329 To render the default subtitles stream from file @file{video.mkv}, use:
16331 subtitles=video.mkv
16334 To render the second subtitles stream from that file, use:
16336 subtitles=video.mkv:si=1
16339 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
16340 @code{DejaVu Serif}, use:
16342 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HCCFF0000'
16345 @section super2xsai
16347 Scale the input by 2x and smooth using the Super2xSaI (Scale and
16348 Interpolate) pixel art scaling algorithm.
16350 Useful for enlarging pixel art images without reducing sharpness.
16354 Swap two rectangular objects in video.
16356 This filter accepts the following options:
16366 Set 1st rect x coordinate.
16369 Set 1st rect y coordinate.
16372 Set 2nd rect x coordinate.
16375 Set 2nd rect y coordinate.
16377 All expressions are evaluated once for each frame.
16380 The all options are expressions containing the following constants:
16385 The input width and height.
16388 same as @var{w} / @var{h}
16391 input sample aspect ratio
16394 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
16397 The number of the input frame, starting from 0.
16400 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
16403 the position in the file of the input frame, NAN if unknown
16411 Apply telecine process to the video.
16413 This filter accepts the following options:
16422 The default value is @code{top}.
16426 A string of numbers representing the pulldown pattern you wish to apply.
16427 The default value is @code{23}.
16431 Some typical patterns:
16436 24p: 2332 (preferred)
16443 24p: 222222222223 ("Euro pulldown")
16450 Apply threshold effect to video stream.
16452 This filter needs four video streams to perform thresholding.
16453 First stream is stream we are filtering.
16454 Second stream is holding threshold values, third stream is holding min values,
16455 and last, fourth stream is holding max values.
16457 The filter accepts the following option:
16461 Set which planes will be processed, unprocessed planes will be copied.
16462 By default value 0xf, all planes will be processed.
16465 For example if first stream pixel's component value is less then threshold value
16466 of pixel component from 2nd threshold stream, third stream value will picked,
16467 otherwise fourth stream pixel component value will be picked.
16469 Using color source filter one can perform various types of thresholding:
16471 @subsection Examples
16475 Binary threshold, using gray color as threshold:
16477 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
16481 Inverted binary threshold, using gray color as threshold:
16483 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
16487 Truncate binary threshold, using gray color as threshold:
16489 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
16493 Threshold to zero, using gray color as threshold:
16495 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
16499 Inverted threshold to zero, using gray color as threshold:
16501 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
16506 Select the most representative frame in a given sequence of consecutive frames.
16508 The filter accepts the following options:
16512 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
16513 will pick one of them, and then handle the next batch of @var{n} frames until
16514 the end. Default is @code{100}.
16517 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
16518 value will result in a higher memory usage, so a high value is not recommended.
16520 @subsection Examples
16524 Extract one picture each 50 frames:
16530 Complete example of a thumbnail creation with @command{ffmpeg}:
16532 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
16538 Tile several successive frames together.
16540 The filter accepts the following options:
16545 Set the grid size (i.e. the number of lines and columns). For the syntax of
16546 this option, check the
16547 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16550 Set the maximum number of frames to render in the given area. It must be less
16551 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
16552 the area will be used.
16555 Set the outer border margin in pixels.
16558 Set the inner border thickness (i.e. the number of pixels between frames). For
16559 more advanced padding options (such as having different values for the edges),
16560 refer to the pad video filter.
16563 Specify the color of the unused area. For the syntax of this option, check the
16564 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16565 The default value of @var{color} is "black".
16568 Set the number of frames to overlap when tiling several successive frames together.
16569 The value must be between @code{0} and @var{nb_frames - 1}.
16572 Set the number of frames to initially be empty before displaying first output frame.
16573 This controls how soon will one get first output frame.
16574 The value must be between @code{0} and @var{nb_frames - 1}.
16577 @subsection Examples
16581 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
16583 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
16585 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
16586 duplicating each output frame to accommodate the originally detected frame
16590 Display @code{5} pictures in an area of @code{3x2} frames,
16591 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
16592 mixed flat and named options:
16594 tile=3x2:nb_frames=5:padding=7:margin=2
16598 @section tinterlace
16600 Perform various types of temporal field interlacing.
16602 Frames are counted starting from 1, so the first input frame is
16605 The filter accepts the following options:
16610 Specify the mode of the interlacing. This option can also be specified
16611 as a value alone. See below for a list of values for this option.
16613 Available values are:
16617 Move odd frames into the upper field, even into the lower field,
16618 generating a double height frame at half frame rate.
16622 Frame 1 Frame 2 Frame 3 Frame 4
16624 11111 22222 33333 44444
16625 11111 22222 33333 44444
16626 11111 22222 33333 44444
16627 11111 22222 33333 44444
16641 Only output odd frames, even frames are dropped, generating a frame with
16642 unchanged height at half frame rate.
16647 Frame 1 Frame 2 Frame 3 Frame 4
16649 11111 22222 33333 44444
16650 11111 22222 33333 44444
16651 11111 22222 33333 44444
16652 11111 22222 33333 44444
16662 Only output even frames, odd frames are dropped, generating a frame with
16663 unchanged height at half frame rate.
16668 Frame 1 Frame 2 Frame 3 Frame 4
16670 11111 22222 33333 44444
16671 11111 22222 33333 44444
16672 11111 22222 33333 44444
16673 11111 22222 33333 44444
16683 Expand each frame to full height, but pad alternate lines with black,
16684 generating a frame with double height at the same input frame rate.
16689 Frame 1 Frame 2 Frame 3 Frame 4
16691 11111 22222 33333 44444
16692 11111 22222 33333 44444
16693 11111 22222 33333 44444
16694 11111 22222 33333 44444
16697 11111 ..... 33333 .....
16698 ..... 22222 ..... 44444
16699 11111 ..... 33333 .....
16700 ..... 22222 ..... 44444
16701 11111 ..... 33333 .....
16702 ..... 22222 ..... 44444
16703 11111 ..... 33333 .....
16704 ..... 22222 ..... 44444
16708 @item interleave_top, 4
16709 Interleave the upper field from odd frames with the lower field from
16710 even frames, generating a frame with unchanged height at half frame rate.
16715 Frame 1 Frame 2 Frame 3 Frame 4
16717 11111<- 22222 33333<- 44444
16718 11111 22222<- 33333 44444<-
16719 11111<- 22222 33333<- 44444
16720 11111 22222<- 33333 44444<-
16730 @item interleave_bottom, 5
16731 Interleave the lower field from odd frames with the upper field from
16732 even frames, generating a frame with unchanged height at half frame rate.
16737 Frame 1 Frame 2 Frame 3 Frame 4
16739 11111 22222<- 33333 44444<-
16740 11111<- 22222 33333<- 44444
16741 11111 22222<- 33333 44444<-
16742 11111<- 22222 33333<- 44444
16752 @item interlacex2, 6
16753 Double frame rate with unchanged height. Frames are inserted each
16754 containing the second temporal field from the previous input frame and
16755 the first temporal field from the next input frame. This mode relies on
16756 the top_field_first flag. Useful for interlaced video displays with no
16757 field synchronisation.
16762 Frame 1 Frame 2 Frame 3 Frame 4
16764 11111 22222 33333 44444
16765 11111 22222 33333 44444
16766 11111 22222 33333 44444
16767 11111 22222 33333 44444
16770 11111 22222 22222 33333 33333 44444 44444
16771 11111 11111 22222 22222 33333 33333 44444
16772 11111 22222 22222 33333 33333 44444 44444
16773 11111 11111 22222 22222 33333 33333 44444
16778 Move odd frames into the upper field, even into the lower field,
16779 generating a double height frame at same frame rate.
16784 Frame 1 Frame 2 Frame 3 Frame 4
16786 11111 22222 33333 44444
16787 11111 22222 33333 44444
16788 11111 22222 33333 44444
16789 11111 22222 33333 44444
16792 11111 33333 33333 55555
16793 22222 22222 44444 44444
16794 11111 33333 33333 55555
16795 22222 22222 44444 44444
16796 11111 33333 33333 55555
16797 22222 22222 44444 44444
16798 11111 33333 33333 55555
16799 22222 22222 44444 44444
16804 Numeric values are deprecated but are accepted for backward
16805 compatibility reasons.
16807 Default mode is @code{merge}.
16810 Specify flags influencing the filter process.
16812 Available value for @var{flags} is:
16815 @item low_pass_filter, vlfp
16816 Enable linear vertical low-pass filtering in the filter.
16817 Vertical low-pass filtering is required when creating an interlaced
16818 destination from a progressive source which contains high-frequency
16819 vertical detail. Filtering will reduce interlace 'twitter' and Moire
16822 @item complex_filter, cvlfp
16823 Enable complex vertical low-pass filtering.
16824 This will slightly less reduce interlace 'twitter' and Moire
16825 patterning but better retain detail and subjective sharpness impression.
16829 Vertical low-pass filtering can only be enabled for @option{mode}
16830 @var{interleave_top} and @var{interleave_bottom}.
16836 Mix successive video frames.
16838 A description of the accepted options follows.
16842 The number of successive frames to mix. If unspecified, it defaults to 3.
16845 Specify weight of each input video frame.
16846 Each weight is separated by space. If number of weights is smaller than
16847 number of @var{frames} last specified weight will be used for all remaining
16851 Specify scale, if it is set it will be multiplied with sum
16852 of each weight multiplied with pixel values to give final destination
16853 pixel value. By default @var{scale} is auto scaled to sum of weights.
16856 @subsection Examples
16860 Average 7 successive frames:
16862 tmix=frames=7:weights="1 1 1 1 1 1 1"
16866 Apply simple temporal convolution:
16868 tmix=frames=3:weights="-1 3 -1"
16872 Similar as above but only showing temporal differences:
16874 tmix=frames=3:weights="-1 2 -1":scale=1
16880 Tone map colors from different dynamic ranges.
16882 This filter expects data in single precision floating point, as it needs to
16883 operate on (and can output) out-of-range values. Another filter, such as
16884 @ref{zscale}, is needed to convert the resulting frame to a usable format.
16886 The tonemapping algorithms implemented only work on linear light, so input
16887 data should be linearized beforehand (and possibly correctly tagged).
16890 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
16893 @subsection Options
16894 The filter accepts the following options.
16898 Set the tone map algorithm to use.
16900 Possible values are:
16903 Do not apply any tone map, only desaturate overbright pixels.
16906 Hard-clip any out-of-range values. Use it for perfect color accuracy for
16907 in-range values, while distorting out-of-range values.
16910 Stretch the entire reference gamut to a linear multiple of the display.
16913 Fit a logarithmic transfer between the tone curves.
16916 Preserve overall image brightness with a simple curve, using nonlinear
16917 contrast, which results in flattening details and degrading color accuracy.
16920 Preserve both dark and bright details better than @var{reinhard}, at the cost
16921 of slightly darkening everything. Use it when detail preservation is more
16922 important than color and brightness accuracy.
16925 Smoothly map out-of-range values, while retaining contrast and colors for
16926 in-range material as much as possible. Use it when color accuracy is more
16927 important than detail preservation.
16933 Tune the tone mapping algorithm.
16935 This affects the following algorithms:
16941 Specifies the scale factor to use while stretching.
16945 Specifies the exponent of the function.
16949 Specify an extra linear coefficient to multiply into the signal before clipping.
16953 Specify the local contrast coefficient at the display peak.
16954 Default to 0.5, which means that in-gamut values will be about half as bright
16961 Specify the transition point from linear to mobius transform. Every value
16962 below this point is guaranteed to be mapped 1:1. The higher the value, the
16963 more accurate the result will be, at the cost of losing bright details.
16964 Default to 0.3, which due to the steep initial slope still preserves in-range
16965 colors fairly accurately.
16969 Apply desaturation for highlights that exceed this level of brightness. The
16970 higher the parameter, the more color information will be preserved. This
16971 setting helps prevent unnaturally blown-out colors for super-highlights, by
16972 (smoothly) turning into white instead. This makes images feel more natural,
16973 at the cost of reducing information about out-of-range colors.
16975 The default of 2.0 is somewhat conservative and will mostly just apply to
16976 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
16978 This option works only if the input frame has a supported color tag.
16981 Override signal/nominal/reference peak with this value. Useful when the
16982 embedded peak information in display metadata is not reliable or when tone
16983 mapping from a lower range to a higher range.
16988 Temporarily pad video frames.
16990 The filter accepts the following options:
16994 Specify number of delay frames before input video stream.
16997 Specify number of padding frames after input video stream.
16998 Set to -1 to pad indefinitely.
17001 Set kind of frames added to beginning of stream.
17002 Can be either @var{add} or @var{clone}.
17003 With @var{add} frames of solid-color are added.
17004 With @var{clone} frames are clones of first frame.
17007 Set kind of frames added to end of stream.
17008 Can be either @var{add} or @var{clone}.
17009 With @var{add} frames of solid-color are added.
17010 With @var{clone} frames are clones of last frame.
17012 @item start_duration, stop_duration
17013 Specify the duration of the start/stop delay. See
17014 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17015 for the accepted syntax.
17016 These options override @var{start} and @var{stop}.
17019 Specify the color of the padded area. For the syntax of this option,
17020 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
17021 manual,ffmpeg-utils}.
17023 The default value of @var{color} is "black".
17029 Transpose rows with columns in the input video and optionally flip it.
17031 It accepts the following parameters:
17036 Specify the transposition direction.
17038 Can assume the following values:
17040 @item 0, 4, cclock_flip
17041 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
17049 Rotate by 90 degrees clockwise, that is:
17057 Rotate by 90 degrees counterclockwise, that is:
17064 @item 3, 7, clock_flip
17065 Rotate by 90 degrees clockwise and vertically flip, that is:
17073 For values between 4-7, the transposition is only done if the input
17074 video geometry is portrait and not landscape. These values are
17075 deprecated, the @code{passthrough} option should be used instead.
17077 Numerical values are deprecated, and should be dropped in favor of
17078 symbolic constants.
17081 Do not apply the transposition if the input geometry matches the one
17082 specified by the specified value. It accepts the following values:
17085 Always apply transposition.
17087 Preserve portrait geometry (when @var{height} >= @var{width}).
17089 Preserve landscape geometry (when @var{width} >= @var{height}).
17092 Default value is @code{none}.
17095 For example to rotate by 90 degrees clockwise and preserve portrait
17098 transpose=dir=1:passthrough=portrait
17101 The command above can also be specified as:
17103 transpose=1:portrait
17106 @section transpose_npp
17108 Transpose rows with columns in the input video and optionally flip it.
17109 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
17111 It accepts the following parameters:
17116 Specify the transposition direction.
17118 Can assume the following values:
17121 Rotate by 90 degrees counterclockwise and vertically flip. (default)
17124 Rotate by 90 degrees clockwise.
17127 Rotate by 90 degrees counterclockwise.
17130 Rotate by 90 degrees clockwise and vertically flip.
17134 Do not apply the transposition if the input geometry matches the one
17135 specified by the specified value. It accepts the following values:
17138 Always apply transposition. (default)
17140 Preserve portrait geometry (when @var{height} >= @var{width}).
17142 Preserve landscape geometry (when @var{width} >= @var{height}).
17148 Trim the input so that the output contains one continuous subpart of the input.
17150 It accepts the following parameters:
17153 Specify the time of the start of the kept section, i.e. the frame with the
17154 timestamp @var{start} will be the first frame in the output.
17157 Specify the time of the first frame that will be dropped, i.e. the frame
17158 immediately preceding the one with the timestamp @var{end} will be the last
17159 frame in the output.
17162 This is the same as @var{start}, except this option sets the start timestamp
17163 in timebase units instead of seconds.
17166 This is the same as @var{end}, except this option sets the end timestamp
17167 in timebase units instead of seconds.
17170 The maximum duration of the output in seconds.
17173 The number of the first frame that should be passed to the output.
17176 The number of the first frame that should be dropped.
17179 @option{start}, @option{end}, and @option{duration} are expressed as time
17180 duration specifications; see
17181 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
17182 for the accepted syntax.
17184 Note that the first two sets of the start/end options and the @option{duration}
17185 option look at the frame timestamp, while the _frame variants simply count the
17186 frames that pass through the filter. Also note that this filter does not modify
17187 the timestamps. If you wish for the output timestamps to start at zero, insert a
17188 setpts filter after the trim filter.
17190 If multiple start or end options are set, this filter tries to be greedy and
17191 keep all the frames that match at least one of the specified constraints. To keep
17192 only the part that matches all the constraints at once, chain multiple trim
17195 The defaults are such that all the input is kept. So it is possible to set e.g.
17196 just the end values to keep everything before the specified time.
17201 Drop everything except the second minute of input:
17203 ffmpeg -i INPUT -vf trim=60:120
17207 Keep only the first second:
17209 ffmpeg -i INPUT -vf trim=duration=1
17214 @section unpremultiply
17215 Apply alpha unpremultiply effect to input video stream using first plane
17216 of second stream as alpha.
17218 Both streams must have same dimensions and same pixel format.
17220 The filter accepts the following option:
17224 Set which planes will be processed, unprocessed planes will be copied.
17225 By default value 0xf, all planes will be processed.
17227 If the format has 1 or 2 components, then luma is bit 0.
17228 If the format has 3 or 4 components:
17229 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
17230 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
17231 If present, the alpha channel is always the last bit.
17234 Do not require 2nd input for processing, instead use alpha plane from input stream.
17240 Sharpen or blur the input video.
17242 It accepts the following parameters:
17245 @item luma_msize_x, lx
17246 Set the luma matrix horizontal size. It must be an odd integer between
17247 3 and 23. The default value is 5.
17249 @item luma_msize_y, ly
17250 Set the luma matrix vertical size. It must be an odd integer between 3
17251 and 23. The default value is 5.
17253 @item luma_amount, la
17254 Set the luma effect strength. It must be a floating point number, reasonable
17255 values lay between -1.5 and 1.5.
17257 Negative values will blur the input video, while positive values will
17258 sharpen it, a value of zero will disable the effect.
17260 Default value is 1.0.
17262 @item chroma_msize_x, cx
17263 Set the chroma matrix horizontal size. It must be an odd integer
17264 between 3 and 23. The default value is 5.
17266 @item chroma_msize_y, cy
17267 Set the chroma matrix vertical size. It must be an odd integer
17268 between 3 and 23. The default value is 5.
17270 @item chroma_amount, ca
17271 Set the chroma effect strength. It must be a floating point number, reasonable
17272 values lay between -1.5 and 1.5.
17274 Negative values will blur the input video, while positive values will
17275 sharpen it, a value of zero will disable the effect.
17277 Default value is 0.0.
17281 All parameters are optional and default to the equivalent of the
17282 string '5:5:1.0:5:5:0.0'.
17284 @subsection Examples
17288 Apply strong luma sharpen effect:
17290 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
17294 Apply a strong blur of both luma and chroma parameters:
17296 unsharp=7:7:-2:7:7:-2
17302 Apply ultra slow/simple postprocessing filter that compresses and decompresses
17303 the image at several (or - in the case of @option{quality} level @code{8} - all)
17304 shifts and average the results.
17306 The way this differs from the behavior of spp is that uspp actually encodes &
17307 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
17308 DCT similar to MJPEG.
17310 The filter accepts the following options:
17314 Set quality. This option defines the number of levels for averaging. It accepts
17315 an integer in the range 0-8. If set to @code{0}, the filter will have no
17316 effect. A value of @code{8} means the higher quality. For each increment of
17317 that value the speed drops by a factor of approximately 2. Default value is
17321 Force a constant quantization parameter. If not set, the filter will use the QP
17322 from the video stream (if available).
17325 @section vaguedenoiser
17327 Apply a wavelet based denoiser.
17329 It transforms each frame from the video input into the wavelet domain,
17330 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
17331 the obtained coefficients. It does an inverse wavelet transform after.
17332 Due to wavelet properties, it should give a nice smoothed result, and
17333 reduced noise, without blurring picture features.
17335 This filter accepts the following options:
17339 The filtering strength. The higher, the more filtered the video will be.
17340 Hard thresholding can use a higher threshold than soft thresholding
17341 before the video looks overfiltered. Default value is 2.
17344 The filtering method the filter will use.
17346 It accepts the following values:
17349 All values under the threshold will be zeroed.
17352 All values under the threshold will be zeroed. All values above will be
17353 reduced by the threshold.
17356 Scales or nullifies coefficients - intermediary between (more) soft and
17357 (less) hard thresholding.
17360 Default is garrote.
17363 Number of times, the wavelet will decompose the picture. Picture can't
17364 be decomposed beyond a particular point (typically, 8 for a 640x480
17365 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
17368 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
17371 A list of the planes to process. By default all planes are processed.
17374 @section vectorscope
17376 Display 2 color component values in the two dimensional graph (which is called
17379 This filter accepts the following options:
17383 Set vectorscope mode.
17385 It accepts the following values:
17388 Gray values are displayed on graph, higher brightness means more pixels have
17389 same component color value on location in graph. This is the default mode.
17392 Gray values are displayed on graph. Surrounding pixels values which are not
17393 present in video frame are drawn in gradient of 2 color components which are
17394 set by option @code{x} and @code{y}. The 3rd color component is static.
17397 Actual color components values present in video frame are displayed on graph.
17400 Similar as color2 but higher frequency of same values @code{x} and @code{y}
17401 on graph increases value of another color component, which is luminance by
17402 default values of @code{x} and @code{y}.
17405 Actual colors present in video frame are displayed on graph. If two different
17406 colors map to same position on graph then color with higher value of component
17407 not present in graph is picked.
17410 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
17411 component picked from radial gradient.
17415 Set which color component will be represented on X-axis. Default is @code{1}.
17418 Set which color component will be represented on Y-axis. Default is @code{2}.
17421 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
17422 of color component which represents frequency of (X, Y) location in graph.
17427 No envelope, this is default.
17430 Instant envelope, even darkest single pixel will be clearly highlighted.
17433 Hold maximum and minimum values presented in graph over time. This way you
17434 can still spot out of range values without constantly looking at vectorscope.
17437 Peak and instant envelope combined together.
17441 Set what kind of graticule to draw.
17449 Set graticule opacity.
17452 Set graticule flags.
17456 Draw graticule for white point.
17459 Draw graticule for black point.
17462 Draw color points short names.
17466 Set background opacity.
17468 @item lthreshold, l
17469 Set low threshold for color component not represented on X or Y axis.
17470 Values lower than this value will be ignored. Default is 0.
17471 Note this value is multiplied with actual max possible value one pixel component
17472 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
17475 @item hthreshold, h
17476 Set high threshold for color component not represented on X or Y axis.
17477 Values higher than this value will be ignored. Default is 1.
17478 Note this value is multiplied with actual max possible value one pixel component
17479 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
17480 is 0.9 * 255 = 230.
17482 @item colorspace, c
17483 Set what kind of colorspace to use when drawing graticule.
17492 @anchor{vidstabdetect}
17493 @section vidstabdetect
17495 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
17496 @ref{vidstabtransform} for pass 2.
17498 This filter generates a file with relative translation and rotation
17499 transform information about subsequent frames, which is then used by
17500 the @ref{vidstabtransform} filter.
17502 To enable compilation of this filter you need to configure FFmpeg with
17503 @code{--enable-libvidstab}.
17505 This filter accepts the following options:
17509 Set the path to the file used to write the transforms information.
17510 Default value is @file{transforms.trf}.
17513 Set how shaky the video is and how quick the camera is. It accepts an
17514 integer in the range 1-10, a value of 1 means little shakiness, a
17515 value of 10 means strong shakiness. Default value is 5.
17518 Set the accuracy of the detection process. It must be a value in the
17519 range 1-15. A value of 1 means low accuracy, a value of 15 means high
17520 accuracy. Default value is 15.
17523 Set stepsize of the search process. The region around minimum is
17524 scanned with 1 pixel resolution. Default value is 6.
17527 Set minimum contrast. Below this value a local measurement field is
17528 discarded. Must be a floating point value in the range 0-1. Default
17532 Set reference frame number for tripod mode.
17534 If enabled, the motion of the frames is compared to a reference frame
17535 in the filtered stream, identified by the specified number. The idea
17536 is to compensate all movements in a more-or-less static scene and keep
17537 the camera view absolutely still.
17539 If set to 0, it is disabled. The frames are counted starting from 1.
17542 Show fields and transforms in the resulting frames. It accepts an
17543 integer in the range 0-2. Default value is 0, which disables any
17547 @subsection Examples
17551 Use default values:
17557 Analyze strongly shaky movie and put the results in file
17558 @file{mytransforms.trf}:
17560 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
17564 Visualize the result of internal transformations in the resulting
17567 vidstabdetect=show=1
17571 Analyze a video with medium shakiness using @command{ffmpeg}:
17573 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
17577 @anchor{vidstabtransform}
17578 @section vidstabtransform
17580 Video stabilization/deshaking: pass 2 of 2,
17581 see @ref{vidstabdetect} for pass 1.
17583 Read a file with transform information for each frame and
17584 apply/compensate them. Together with the @ref{vidstabdetect}
17585 filter this can be used to deshake videos. See also
17586 @url{http://public.hronopik.de/vid.stab}. It is important to also use
17587 the @ref{unsharp} filter, see below.
17589 To enable compilation of this filter you need to configure FFmpeg with
17590 @code{--enable-libvidstab}.
17592 @subsection Options
17596 Set path to the file used to read the transforms. Default value is
17597 @file{transforms.trf}.
17600 Set the number of frames (value*2 + 1) used for lowpass filtering the
17601 camera movements. Default value is 10.
17603 For example a number of 10 means that 21 frames are used (10 in the
17604 past and 10 in the future) to smoothen the motion in the video. A
17605 larger value leads to a smoother video, but limits the acceleration of
17606 the camera (pan/tilt movements). 0 is a special case where a static
17607 camera is simulated.
17610 Set the camera path optimization algorithm.
17612 Accepted values are:
17615 gaussian kernel low-pass filter on camera motion (default)
17617 averaging on transformations
17621 Set maximal number of pixels to translate frames. Default value is -1,
17625 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
17626 value is -1, meaning no limit.
17629 Specify how to deal with borders that may be visible due to movement
17632 Available values are:
17635 keep image information from previous frame (default)
17637 fill the border black
17641 Invert transforms if set to 1. Default value is 0.
17644 Consider transforms as relative to previous frame if set to 1,
17645 absolute if set to 0. Default value is 0.
17648 Set percentage to zoom. A positive value will result in a zoom-in
17649 effect, a negative value in a zoom-out effect. Default value is 0 (no
17653 Set optimal zooming to avoid borders.
17655 Accepted values are:
17660 optimal static zoom value is determined (only very strong movements
17661 will lead to visible borders) (default)
17663 optimal adaptive zoom value is determined (no borders will be
17664 visible), see @option{zoomspeed}
17667 Note that the value given at zoom is added to the one calculated here.
17670 Set percent to zoom maximally each frame (enabled when
17671 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
17675 Specify type of interpolation.
17677 Available values are:
17682 linear only horizontal
17684 linear in both directions (default)
17686 cubic in both directions (slow)
17690 Enable virtual tripod mode if set to 1, which is equivalent to
17691 @code{relative=0:smoothing=0}. Default value is 0.
17693 Use also @code{tripod} option of @ref{vidstabdetect}.
17696 Increase log verbosity if set to 1. Also the detected global motions
17697 are written to the temporary file @file{global_motions.trf}. Default
17701 @subsection Examples
17705 Use @command{ffmpeg} for a typical stabilization with default values:
17707 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
17710 Note the use of the @ref{unsharp} filter which is always recommended.
17713 Zoom in a bit more and load transform data from a given file:
17715 vidstabtransform=zoom=5:input="mytransforms.trf"
17719 Smoothen the video even more:
17721 vidstabtransform=smoothing=30
17727 Flip the input video vertically.
17729 For example, to vertically flip a video with @command{ffmpeg}:
17731 ffmpeg -i in.avi -vf "vflip" out.avi
17736 Detect variable frame rate video.
17738 This filter tries to detect if the input is variable or constant frame rate.
17740 At end it will output number of frames detected as having variable delta pts,
17741 and ones with constant delta pts.
17742 If there was frames with variable delta, than it will also show min and max delta
17747 Boost or alter saturation.
17749 The filter accepts the following options:
17752 Set strength of boost if positive value or strength of alter if negative value.
17753 Default is 0. Allowed range is from -2 to 2.
17756 Set the red balance. Default is 1. Allowed range is from -10 to 10.
17759 Set the green balance. Default is 1. Allowed range is from -10 to 10.
17762 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
17765 Set the red luma coefficient.
17768 Set the green luma coefficient.
17771 Set the blue luma coefficient.
17777 Make or reverse a natural vignetting effect.
17779 The filter accepts the following options:
17783 Set lens angle expression as a number of radians.
17785 The value is clipped in the @code{[0,PI/2]} range.
17787 Default value: @code{"PI/5"}
17791 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
17795 Set forward/backward mode.
17797 Available modes are:
17800 The larger the distance from the central point, the darker the image becomes.
17803 The larger the distance from the central point, the brighter the image becomes.
17804 This can be used to reverse a vignette effect, though there is no automatic
17805 detection to extract the lens @option{angle} and other settings (yet). It can
17806 also be used to create a burning effect.
17809 Default value is @samp{forward}.
17812 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
17814 It accepts the following values:
17817 Evaluate expressions only once during the filter initialization.
17820 Evaluate expressions for each incoming frame. This is way slower than the
17821 @samp{init} mode since it requires all the scalers to be re-computed, but it
17822 allows advanced dynamic expressions.
17825 Default value is @samp{init}.
17828 Set dithering to reduce the circular banding effects. Default is @code{1}
17832 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
17833 Setting this value to the SAR of the input will make a rectangular vignetting
17834 following the dimensions of the video.
17836 Default is @code{1/1}.
17839 @subsection Expressions
17841 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
17842 following parameters.
17847 input width and height
17850 the number of input frame, starting from 0
17853 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
17854 @var{TB} units, NAN if undefined
17857 frame rate of the input video, NAN if the input frame rate is unknown
17860 the PTS (Presentation TimeStamp) of the filtered video frame,
17861 expressed in seconds, NAN if undefined
17864 time base of the input video
17868 @subsection Examples
17872 Apply simple strong vignetting effect:
17878 Make a flickering vignetting:
17880 vignette='PI/4+random(1)*PI/50':eval=frame
17885 @section vmafmotion
17887 Obtain the average vmaf motion score of a video.
17888 It is one of the component filters of VMAF.
17890 The obtained average motion score is printed through the logging system.
17892 In the below example the input file @file{ref.mpg} is being processed and score
17896 ffmpeg -i ref.mpg -lavfi vmafmotion -f null -
17900 Stack input videos vertically.
17902 All streams must be of same pixel format and of same width.
17904 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
17905 to create same output.
17907 The filter accept the following option:
17911 Set number of input streams. Default is 2.
17914 If set to 1, force the output to terminate when the shortest input
17915 terminates. Default value is 0.
17920 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
17921 Deinterlacing Filter").
17923 Based on the process described by Martin Weston for BBC R&D, and
17924 implemented based on the de-interlace algorithm written by Jim
17925 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
17926 uses filter coefficients calculated by BBC R&D.
17928 There are two sets of filter coefficients, so called "simple":
17929 and "complex". Which set of filter coefficients is used can
17930 be set by passing an optional parameter:
17934 Set the interlacing filter coefficients. Accepts one of the following values:
17938 Simple filter coefficient set.
17940 More-complex filter coefficient set.
17942 Default value is @samp{complex}.
17945 Specify which frames to deinterlace. Accept one of the following values:
17949 Deinterlace all frames,
17951 Only deinterlace frames marked as interlaced.
17954 Default value is @samp{all}.
17958 Video waveform monitor.
17960 The waveform monitor plots color component intensity. By default luminance
17961 only. Each column of the waveform corresponds to a column of pixels in the
17964 It accepts the following options:
17968 Can be either @code{row}, or @code{column}. Default is @code{column}.
17969 In row mode, the graph on the left side represents color component value 0 and
17970 the right side represents value = 255. In column mode, the top side represents
17971 color component value = 0 and bottom side represents value = 255.
17974 Set intensity. Smaller values are useful to find out how many values of the same
17975 luminance are distributed across input rows/columns.
17976 Default value is @code{0.04}. Allowed range is [0, 1].
17979 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
17980 In mirrored mode, higher values will be represented on the left
17981 side for @code{row} mode and at the top for @code{column} mode. Default is
17982 @code{1} (mirrored).
17986 It accepts the following values:
17989 Presents information identical to that in the @code{parade}, except
17990 that the graphs representing color components are superimposed directly
17993 This display mode makes it easier to spot relative differences or similarities
17994 in overlapping areas of the color components that are supposed to be identical,
17995 such as neutral whites, grays, or blacks.
17998 Display separate graph for the color components side by side in
17999 @code{row} mode or one below the other in @code{column} mode.
18002 Display separate graph for the color components side by side in
18003 @code{column} mode or one below the other in @code{row} mode.
18005 Using this display mode makes it easy to spot color casts in the highlights
18006 and shadows of an image, by comparing the contours of the top and the bottom
18007 graphs of each waveform. Since whites, grays, and blacks are characterized
18008 by exactly equal amounts of red, green, and blue, neutral areas of the picture
18009 should display three waveforms of roughly equal width/height. If not, the
18010 correction is easy to perform by making level adjustments the three waveforms.
18012 Default is @code{stack}.
18014 @item components, c
18015 Set which color components to display. Default is 1, which means only luminance
18016 or red color component if input is in RGB colorspace. If is set for example to
18017 7 it will display all 3 (if) available color components.
18022 No envelope, this is default.
18025 Instant envelope, minimum and maximum values presented in graph will be easily
18026 visible even with small @code{step} value.
18029 Hold minimum and maximum values presented in graph across time. This way you
18030 can still spot out of range values without constantly looking at waveforms.
18033 Peak and instant envelope combined together.
18039 No filtering, this is default.
18042 Luma and chroma combined together.
18045 Similar as above, but shows difference between blue and red chroma.
18048 Similar as above, but use different colors.
18051 Displays only chroma.
18054 Displays actual color value on waveform.
18057 Similar as above, but with luma showing frequency of chroma values.
18061 Set which graticule to display.
18065 Do not display graticule.
18068 Display green graticule showing legal broadcast ranges.
18071 Display orange graticule showing legal broadcast ranges.
18075 Set graticule opacity.
18078 Set graticule flags.
18082 Draw numbers above lines. By default enabled.
18085 Draw dots instead of lines.
18089 Set scale used for displaying graticule.
18096 Default is digital.
18099 Set background opacity.
18102 @section weave, doubleweave
18104 The @code{weave} takes a field-based video input and join
18105 each two sequential fields into single frame, producing a new double
18106 height clip with half the frame rate and half the frame count.
18108 The @code{doubleweave} works same as @code{weave} but without
18109 halving frame rate and frame count.
18111 It accepts the following option:
18115 Set first field. Available values are:
18119 Set the frame as top-field-first.
18122 Set the frame as bottom-field-first.
18126 @subsection Examples
18130 Interlace video using @ref{select} and @ref{separatefields} filter:
18132 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
18137 Apply the xBR high-quality magnification filter which is designed for pixel
18138 art. It follows a set of edge-detection rules, see
18139 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
18141 It accepts the following option:
18145 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
18146 @code{3xBR} and @code{4} for @code{4xBR}.
18147 Default is @code{3}.
18151 Stack video inputs into custom layout.
18153 All streams must be of same pixel format.
18155 The filter accept the following option:
18159 Set number of input streams. Default is 2.
18162 Specify layout of inputs.
18163 This option requires the desired layout configuration to be explicitly set by the user.
18164 This sets position of each video input in output. Each input
18165 is separated by '|'.
18166 The first number represents the column, and the second number represents the row.
18167 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
18168 where X is video input from which to take width or height.
18169 Multiple values can be used when separated by '+'. In such
18170 case values are summed together.
18173 If set to 1, force the output to terminate when the shortest input
18174 terminates. Default value is 0.
18177 @subsection Examples
18181 Display 4 inputs into 2x2 grid,
18182 note that if inputs are of different sizes unused gaps might appear,
18183 as not all of output video is used.
18185 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
18189 Display 4 inputs into 1x4 grid,
18190 note that if inputs are of different sizes unused gaps might appear,
18191 as not all of output video is used.
18193 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
18197 Display 9 inputs into 3x3 grid,
18198 note that if inputs are of different sizes unused gaps might appear,
18199 as not all of output video is used.
18201 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
18208 Deinterlace the input video ("yadif" means "yet another deinterlacing
18211 It accepts the following parameters:
18217 The interlacing mode to adopt. It accepts one of the following values:
18220 @item 0, send_frame
18221 Output one frame for each frame.
18222 @item 1, send_field
18223 Output one frame for each field.
18224 @item 2, send_frame_nospatial
18225 Like @code{send_frame}, but it skips the spatial interlacing check.
18226 @item 3, send_field_nospatial
18227 Like @code{send_field}, but it skips the spatial interlacing check.
18230 The default value is @code{send_frame}.
18233 The picture field parity assumed for the input interlaced video. It accepts one
18234 of the following values:
18238 Assume the top field is first.
18240 Assume the bottom field is first.
18242 Enable automatic detection of field parity.
18245 The default value is @code{auto}.
18246 If the interlacing is unknown or the decoder does not export this information,
18247 top field first will be assumed.
18250 Specify which frames to deinterlace. Accept one of the following
18255 Deinterlace all frames.
18256 @item 1, interlaced
18257 Only deinterlace frames marked as interlaced.
18260 The default value is @code{all}.
18263 @section yadif_cuda
18265 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
18266 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
18269 It accepts the following parameters:
18275 The interlacing mode to adopt. It accepts one of the following values:
18278 @item 0, send_frame
18279 Output one frame for each frame.
18280 @item 1, send_field
18281 Output one frame for each field.
18282 @item 2, send_frame_nospatial
18283 Like @code{send_frame}, but it skips the spatial interlacing check.
18284 @item 3, send_field_nospatial
18285 Like @code{send_field}, but it skips the spatial interlacing check.
18288 The default value is @code{send_frame}.
18291 The picture field parity assumed for the input interlaced video. It accepts one
18292 of the following values:
18296 Assume the top field is first.
18298 Assume the bottom field is first.
18300 Enable automatic detection of field parity.
18303 The default value is @code{auto}.
18304 If the interlacing is unknown or the decoder does not export this information,
18305 top field first will be assumed.
18308 Specify which frames to deinterlace. Accept one of the following
18313 Deinterlace all frames.
18314 @item 1, interlaced
18315 Only deinterlace frames marked as interlaced.
18318 The default value is @code{all}.
18323 Apply Zoom & Pan effect.
18325 This filter accepts the following options:
18329 Set the zoom expression. Default is 1.
18333 Set the x and y expression. Default is 0.
18336 Set the duration expression in number of frames.
18337 This sets for how many number of frames effect will last for
18338 single input image.
18341 Set the output image size, default is 'hd720'.
18344 Set the output frame rate, default is '25'.
18347 Each expression can contain the following constants:
18366 Output frame count.
18370 Last calculated 'x' and 'y' position from 'x' and 'y' expression
18371 for current input frame.
18375 'x' and 'y' of last output frame of previous input frame or 0 when there was
18376 not yet such frame (first input frame).
18379 Last calculated zoom from 'z' expression for current input frame.
18382 Last calculated zoom of last output frame of previous input frame.
18385 Number of output frames for current input frame. Calculated from 'd' expression
18386 for each input frame.
18389 number of output frames created for previous input frame
18392 Rational number: input width / input height
18395 sample aspect ratio
18398 display aspect ratio
18402 @subsection Examples
18406 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
18408 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
18412 Zoom-in up to 1.5 and pan always at center of picture:
18414 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
18418 Same as above but without pausing:
18420 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
18426 Scale (resize) the input video, using the z.lib library:
18427 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
18428 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
18430 The zscale filter forces the output display aspect ratio to be the same
18431 as the input, by changing the output sample aspect ratio.
18433 If the input image format is different from the format requested by
18434 the next filter, the zscale filter will convert the input to the
18437 @subsection Options
18438 The filter accepts the following options.
18443 Set the output video dimension expression. Default value is the input
18446 If the @var{width} or @var{w} value is 0, the input width is used for
18447 the output. If the @var{height} or @var{h} value is 0, the input height
18448 is used for the output.
18450 If one and only one of the values is -n with n >= 1, the zscale filter
18451 will use a value that maintains the aspect ratio of the input image,
18452 calculated from the other specified dimension. After that it will,
18453 however, make sure that the calculated dimension is divisible by n and
18454 adjust the value if necessary.
18456 If both values are -n with n >= 1, the behavior will be identical to
18457 both values being set to 0 as previously detailed.
18459 See below for the list of accepted constants for use in the dimension
18463 Set the video size. For the syntax of this option, check the
18464 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18467 Set the dither type.
18469 Possible values are:
18474 @item error_diffusion
18480 Set the resize filter type.
18482 Possible values are:
18492 Default is bilinear.
18495 Set the color range.
18497 Possible values are:
18504 Default is same as input.
18507 Set the color primaries.
18509 Possible values are:
18519 Default is same as input.
18522 Set the transfer characteristics.
18524 Possible values are:
18538 Default is same as input.
18541 Set the colorspace matrix.
18543 Possible value are:
18554 Default is same as input.
18557 Set the input color range.
18559 Possible values are:
18566 Default is same as input.
18568 @item primariesin, pin
18569 Set the input color primaries.
18571 Possible values are:
18581 Default is same as input.
18583 @item transferin, tin
18584 Set the input transfer characteristics.
18586 Possible values are:
18597 Default is same as input.
18599 @item matrixin, min
18600 Set the input colorspace matrix.
18602 Possible value are:
18614 Set the output chroma location.
18616 Possible values are:
18627 @item chromalin, cin
18628 Set the input chroma location.
18630 Possible values are:
18642 Set the nominal peak luminance.
18645 The values of the @option{w} and @option{h} options are expressions
18646 containing the following constants:
18651 The input width and height
18655 These are the same as @var{in_w} and @var{in_h}.
18659 The output (scaled) width and height
18663 These are the same as @var{out_w} and @var{out_h}
18666 The same as @var{iw} / @var{ih}
18669 input sample aspect ratio
18672 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
18676 horizontal and vertical input chroma subsample values. For example for the
18677 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
18681 horizontal and vertical output chroma subsample values. For example for the
18682 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
18688 @c man end VIDEO FILTERS
18690 @chapter OpenCL Video Filters
18691 @c man begin OPENCL VIDEO FILTERS
18693 Below is a description of the currently available OpenCL video filters.
18695 To enable compilation of these filters you need to configure FFmpeg with
18696 @code{--enable-opencl}.
18698 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
18701 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
18702 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
18703 given device parameters.
18705 @item -filter_hw_device @var{name}
18706 Pass the hardware device called @var{name} to all filters in any filter graph.
18710 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
18714 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
18716 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
18720 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.
18722 @section avgblur_opencl
18724 Apply average blur filter.
18726 The filter accepts the following options:
18730 Set horizontal radius size.
18731 Range is @code{[1, 1024]} and default value is @code{1}.
18734 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
18737 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
18740 @subsection Example
18744 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.
18746 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
18750 @section boxblur_opencl
18752 Apply a boxblur algorithm to the input video.
18754 It accepts the following parameters:
18758 @item luma_radius, lr
18759 @item luma_power, lp
18760 @item chroma_radius, cr
18761 @item chroma_power, cp
18762 @item alpha_radius, ar
18763 @item alpha_power, ap
18767 A description of the accepted options follows.
18770 @item luma_radius, lr
18771 @item chroma_radius, cr
18772 @item alpha_radius, ar
18773 Set an expression for the box radius in pixels used for blurring the
18774 corresponding input plane.
18776 The radius value must be a non-negative number, and must not be
18777 greater than the value of the expression @code{min(w,h)/2} for the
18778 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
18781 Default value for @option{luma_radius} is "2". If not specified,
18782 @option{chroma_radius} and @option{alpha_radius} default to the
18783 corresponding value set for @option{luma_radius}.
18785 The expressions can contain the following constants:
18789 The input width and height in pixels.
18793 The input chroma image width and height in pixels.
18797 The horizontal and vertical chroma subsample values. For example, for the
18798 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
18801 @item luma_power, lp
18802 @item chroma_power, cp
18803 @item alpha_power, ap
18804 Specify how many times the boxblur filter is applied to the
18805 corresponding plane.
18807 Default value for @option{luma_power} is 2. If not specified,
18808 @option{chroma_power} and @option{alpha_power} default to the
18809 corresponding value set for @option{luma_power}.
18811 A value of 0 will disable the effect.
18814 @subsection Examples
18816 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.
18820 Apply a boxblur filter with the luma, chroma, and alpha radius
18821 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.
18823 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
18824 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
18828 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.
18830 For the luma plane, a 2x2 box radius will be run once.
18832 For the chroma plane, a 4x4 box radius will be run 5 times.
18834 For the alpha plane, a 3x3 box radius will be run 7 times.
18836 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
18840 @section convolution_opencl
18842 Apply convolution of 3x3, 5x5, 7x7 matrix.
18844 The filter accepts the following options:
18851 Set matrix for each plane.
18852 Matrix is sequence of 9, 25 or 49 signed numbers.
18853 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
18859 Set multiplier for calculated value for each plane.
18860 If unset or 0, it will be sum of all matrix elements.
18861 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
18867 Set bias for each plane. This value is added to the result of the multiplication.
18868 Useful for making the overall image brighter or darker.
18869 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
18873 @subsection Examples
18879 -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
18885 -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
18889 Apply edge enhance:
18891 -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
18897 -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
18901 Apply laplacian edge detector which includes diagonals:
18903 -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
18909 -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
18913 @section dilation_opencl
18915 Apply dilation effect to the video.
18917 This filter replaces the pixel by the local(3x3) maximum.
18919 It accepts the following options:
18926 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
18927 If @code{0}, plane will remain unchanged.
18930 Flag which specifies the pixel to refer to.
18931 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
18933 Flags to local 3x3 coordinates region centered on @code{x}:
18942 @subsection Example
18946 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.
18948 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
18952 @section erosion_opencl
18954 Apply erosion effect to the video.
18956 This filter replaces the pixel by the local(3x3) minimum.
18958 It accepts the following options:
18965 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
18966 If @code{0}, plane will remain unchanged.
18969 Flag which specifies the pixel to refer to.
18970 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
18972 Flags to local 3x3 coordinates region centered on @code{x}:
18981 @subsection Example
18985 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.
18987 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
18991 @section overlay_opencl
18993 Overlay one video on top of another.
18995 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
18996 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
18998 The filter accepts the following options:
19003 Set the x coordinate of the overlaid video on the main video.
19004 Default value is @code{0}.
19007 Set the x coordinate of the overlaid video on the main video.
19008 Default value is @code{0}.
19012 @subsection Examples
19016 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
19018 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
19021 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
19023 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
19028 @section prewitt_opencl
19030 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
19032 The filter accepts the following option:
19036 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19039 Set value which will be multiplied with filtered result.
19040 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
19043 Set value which will be added to filtered result.
19044 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
19047 @subsection Example
19051 Apply the Prewitt operator with scale set to 2 and delta set to 10.
19053 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
19057 @section roberts_opencl
19058 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
19060 The filter accepts the following option:
19064 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19067 Set value which will be multiplied with filtered result.
19068 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
19071 Set value which will be added to filtered result.
19072 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
19075 @subsection Example
19079 Apply the Roberts cross operator with scale set to 2 and delta set to 10
19081 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
19085 @section sobel_opencl
19087 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
19089 The filter accepts the following option:
19093 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
19096 Set value which will be multiplied with filtered result.
19097 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
19100 Set value which will be added to filtered result.
19101 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
19104 @subsection Example
19108 Apply sobel operator with scale set to 2 and delta set to 10
19110 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
19114 @section tonemap_opencl
19116 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
19118 It accepts the following parameters:
19122 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
19125 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
19128 Apply desaturation for highlights that exceed this level of brightness. The
19129 higher the parameter, the more color information will be preserved. This
19130 setting helps prevent unnaturally blown-out colors for super-highlights, by
19131 (smoothly) turning into white instead. This makes images feel more natural,
19132 at the cost of reducing information about out-of-range colors.
19134 The default value is 0.5, and the algorithm here is a little different from
19135 the cpu version tonemap currently. A setting of 0.0 disables this option.
19138 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
19139 is used to detect whether the scene has changed or not. If the distance between
19140 the current frame average brightness and the current running average exceeds
19141 a threshold value, we would re-calculate scene average and peak brightness.
19142 The default value is 0.2.
19145 Specify the output pixel format.
19147 Currently supported formats are:
19154 Set the output color range.
19156 Possible values are:
19162 Default is same as input.
19165 Set the output color primaries.
19167 Possible values are:
19173 Default is same as input.
19176 Set the output transfer characteristics.
19178 Possible values are:
19187 Set the output colorspace matrix.
19189 Possible value are:
19195 Default is same as input.
19199 @subsection Example
19203 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
19205 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
19209 @section unsharp_opencl
19211 Sharpen or blur the input video.
19213 It accepts the following parameters:
19216 @item luma_msize_x, lx
19217 Set the luma matrix horizontal size.
19218 Range is @code{[1, 23]} and default value is @code{5}.
19220 @item luma_msize_y, ly
19221 Set the luma matrix vertical size.
19222 Range is @code{[1, 23]} and default value is @code{5}.
19224 @item luma_amount, la
19225 Set the luma effect strength.
19226 Range is @code{[-10, 10]} and default value is @code{1.0}.
19228 Negative values will blur the input video, while positive values will
19229 sharpen it, a value of zero will disable the effect.
19231 @item chroma_msize_x, cx
19232 Set the chroma matrix horizontal size.
19233 Range is @code{[1, 23]} and default value is @code{5}.
19235 @item chroma_msize_y, cy
19236 Set the chroma matrix vertical size.
19237 Range is @code{[1, 23]} and default value is @code{5}.
19239 @item chroma_amount, ca
19240 Set the chroma effect strength.
19241 Range is @code{[-10, 10]} and default value is @code{0.0}.
19243 Negative values will blur the input video, while positive values will
19244 sharpen it, a value of zero will disable the effect.
19248 All parameters are optional and default to the equivalent of the
19249 string '5:5:1.0:5:5:0.0'.
19251 @subsection Examples
19255 Apply strong luma sharpen effect:
19257 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
19261 Apply a strong blur of both luma and chroma parameters:
19263 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
19267 @c man end OPENCL VIDEO FILTERS
19269 @chapter Video Sources
19270 @c man begin VIDEO SOURCES
19272 Below is a description of the currently available video sources.
19276 Buffer video frames, and make them available to the filter chain.
19278 This source is mainly intended for a programmatic use, in particular
19279 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
19281 It accepts the following parameters:
19286 Specify the size (width and height) of the buffered video frames. For the
19287 syntax of this option, check the
19288 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19291 The input video width.
19294 The input video height.
19297 A string representing the pixel format of the buffered video frames.
19298 It may be a number corresponding to a pixel format, or a pixel format
19302 Specify the timebase assumed by the timestamps of the buffered frames.
19305 Specify the frame rate expected for the video stream.
19307 @item pixel_aspect, sar
19308 The sample (pixel) aspect ratio of the input video.
19311 Specify the optional parameters to be used for the scale filter which
19312 is automatically inserted when an input change is detected in the
19313 input size or format.
19315 @item hw_frames_ctx
19316 When using a hardware pixel format, this should be a reference to an
19317 AVHWFramesContext describing input frames.
19322 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
19325 will instruct the source to accept video frames with size 320x240 and
19326 with format "yuv410p", assuming 1/24 as the timestamps timebase and
19327 square pixels (1:1 sample aspect ratio).
19328 Since the pixel format with name "yuv410p" corresponds to the number 6
19329 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
19330 this example corresponds to:
19332 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
19335 Alternatively, the options can be specified as a flat string, but this
19336 syntax is deprecated:
19338 @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}]
19342 Create a pattern generated by an elementary cellular automaton.
19344 The initial state of the cellular automaton can be defined through the
19345 @option{filename} and @option{pattern} options. If such options are
19346 not specified an initial state is created randomly.
19348 At each new frame a new row in the video is filled with the result of
19349 the cellular automaton next generation. The behavior when the whole
19350 frame is filled is defined by the @option{scroll} option.
19352 This source accepts the following options:
19356 Read the initial cellular automaton state, i.e. the starting row, from
19357 the specified file.
19358 In the file, each non-whitespace character is considered an alive
19359 cell, a newline will terminate the row, and further characters in the
19360 file will be ignored.
19363 Read the initial cellular automaton state, i.e. the starting row, from
19364 the specified string.
19366 Each non-whitespace character in the string is considered an alive
19367 cell, a newline will terminate the row, and further characters in the
19368 string will be ignored.
19371 Set the video rate, that is the number of frames generated per second.
19374 @item random_fill_ratio, ratio
19375 Set the random fill ratio for the initial cellular automaton row. It
19376 is a floating point number value ranging from 0 to 1, defaults to
19379 This option is ignored when a file or a pattern is specified.
19381 @item random_seed, seed
19382 Set the seed for filling randomly the initial row, must be an integer
19383 included between 0 and UINT32_MAX. If not specified, or if explicitly
19384 set to -1, the filter will try to use a good random seed on a best
19388 Set the cellular automaton rule, it is a number ranging from 0 to 255.
19389 Default value is 110.
19392 Set the size of the output video. For the syntax of this option, check the
19393 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19395 If @option{filename} or @option{pattern} is specified, the size is set
19396 by default to the width of the specified initial state row, and the
19397 height is set to @var{width} * PHI.
19399 If @option{size} is set, it must contain the width of the specified
19400 pattern string, and the specified pattern will be centered in the
19403 If a filename or a pattern string is not specified, the size value
19404 defaults to "320x518" (used for a randomly generated initial state).
19407 If set to 1, scroll the output upward when all the rows in the output
19408 have been already filled. If set to 0, the new generated row will be
19409 written over the top row just after the bottom row is filled.
19412 @item start_full, full
19413 If set to 1, completely fill the output with generated rows before
19414 outputting the first frame.
19415 This is the default behavior, for disabling set the value to 0.
19418 If set to 1, stitch the left and right row edges together.
19419 This is the default behavior, for disabling set the value to 0.
19422 @subsection Examples
19426 Read the initial state from @file{pattern}, and specify an output of
19429 cellauto=f=pattern:s=200x400
19433 Generate a random initial row with a width of 200 cells, with a fill
19436 cellauto=ratio=2/3:s=200x200
19440 Create a pattern generated by rule 18 starting by a single alive cell
19441 centered on an initial row with width 100:
19443 cellauto=p=@@:s=100x400:full=0:rule=18
19447 Specify a more elaborated initial pattern:
19449 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
19454 @anchor{coreimagesrc}
19455 @section coreimagesrc
19456 Video source generated on GPU using Apple's CoreImage API on OSX.
19458 This video source is a specialized version of the @ref{coreimage} video filter.
19459 Use a core image generator at the beginning of the applied filterchain to
19460 generate the content.
19462 The coreimagesrc video source accepts the following options:
19464 @item list_generators
19465 List all available generators along with all their respective options as well as
19466 possible minimum and maximum values along with the default values.
19468 list_generators=true
19472 Specify the size of the sourced video. For the syntax of this option, check the
19473 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19474 The default value is @code{320x240}.
19477 Specify the frame rate of the sourced video, as the number of frames
19478 generated per second. It has to be a string in the format
19479 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
19480 number or a valid video frame rate abbreviation. The default value is
19484 Set the sample aspect ratio of the sourced video.
19487 Set the duration of the sourced video. See
19488 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19489 for the accepted syntax.
19491 If not specified, or the expressed duration is negative, the video is
19492 supposed to be generated forever.
19495 Additionally, all options of the @ref{coreimage} video filter are accepted.
19496 A complete filterchain can be used for further processing of the
19497 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
19498 and examples for details.
19500 @subsection Examples
19505 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
19506 given as complete and escaped command-line for Apple's standard bash shell:
19508 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
19510 This example is equivalent to the QRCode example of @ref{coreimage} without the
19511 need for a nullsrc video source.
19515 @section mandelbrot
19517 Generate a Mandelbrot set fractal, and progressively zoom towards the
19518 point specified with @var{start_x} and @var{start_y}.
19520 This source accepts the following options:
19525 Set the terminal pts value. Default value is 400.
19528 Set the terminal scale value.
19529 Must be a floating point value. Default value is 0.3.
19532 Set the inner coloring mode, that is the algorithm used to draw the
19533 Mandelbrot fractal internal region.
19535 It shall assume one of the following values:
19540 Show time until convergence.
19542 Set color based on point closest to the origin of the iterations.
19547 Default value is @var{mincol}.
19550 Set the bailout value. Default value is 10.0.
19553 Set the maximum of iterations performed by the rendering
19554 algorithm. Default value is 7189.
19557 Set outer coloring mode.
19558 It shall assume one of following values:
19560 @item iteration_count
19561 Set iteration count mode.
19562 @item normalized_iteration_count
19563 set normalized iteration count mode.
19565 Default value is @var{normalized_iteration_count}.
19568 Set frame rate, expressed as number of frames per second. Default
19572 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
19573 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
19576 Set the initial scale value. Default value is 3.0.
19579 Set the initial x position. Must be a floating point value between
19580 -100 and 100. Default value is -0.743643887037158704752191506114774.
19583 Set the initial y position. Must be a floating point value between
19584 -100 and 100. Default value is -0.131825904205311970493132056385139.
19589 Generate various test patterns, as generated by the MPlayer test filter.
19591 The size of the generated video is fixed, and is 256x256.
19592 This source is useful in particular for testing encoding features.
19594 This source accepts the following options:
19599 Specify the frame rate of the sourced video, as the number of frames
19600 generated per second. It has to be a string in the format
19601 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
19602 number or a valid video frame rate abbreviation. The default value is
19606 Set the duration of the sourced video. See
19607 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19608 for the accepted syntax.
19610 If not specified, or the expressed duration is negative, the video is
19611 supposed to be generated forever.
19615 Set the number or the name of the test to perform. Supported tests are:
19631 Default value is "all", which will cycle through the list of all tests.
19636 mptestsrc=t=dc_luma
19639 will generate a "dc_luma" test pattern.
19641 @section frei0r_src
19643 Provide a frei0r source.
19645 To enable compilation of this filter you need to install the frei0r
19646 header and configure FFmpeg with @code{--enable-frei0r}.
19648 This source accepts the following parameters:
19653 The size of the video to generate. For the syntax of this option, check the
19654 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19657 The framerate of the generated video. It may be a string of the form
19658 @var{num}/@var{den} or a frame rate abbreviation.
19661 The name to the frei0r source to load. For more information regarding frei0r and
19662 how to set the parameters, read the @ref{frei0r} section in the video filters
19665 @item filter_params
19666 A '|'-separated list of parameters to pass to the frei0r source.
19670 For example, to generate a frei0r partik0l source with size 200x200
19671 and frame rate 10 which is overlaid on the overlay filter main input:
19673 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
19678 Generate a life pattern.
19680 This source is based on a generalization of John Conway's life game.
19682 The sourced input represents a life grid, each pixel represents a cell
19683 which can be in one of two possible states, alive or dead. Every cell
19684 interacts with its eight neighbours, which are the cells that are
19685 horizontally, vertically, or diagonally adjacent.
19687 At each interaction the grid evolves according to the adopted rule,
19688 which specifies the number of neighbor alive cells which will make a
19689 cell stay alive or born. The @option{rule} option allows one to specify
19692 This source accepts the following options:
19696 Set the file from which to read the initial grid state. In the file,
19697 each non-whitespace character is considered an alive cell, and newline
19698 is used to delimit the end of each row.
19700 If this option is not specified, the initial grid is generated
19704 Set the video rate, that is the number of frames generated per second.
19707 @item random_fill_ratio, ratio
19708 Set the random fill ratio for the initial random grid. It is a
19709 floating point number value ranging from 0 to 1, defaults to 1/PHI.
19710 It is ignored when a file is specified.
19712 @item random_seed, seed
19713 Set the seed for filling the initial random grid, must be an integer
19714 included between 0 and UINT32_MAX. If not specified, or if explicitly
19715 set to -1, the filter will try to use a good random seed on a best
19721 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
19722 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
19723 @var{NS} specifies the number of alive neighbor cells which make a
19724 live cell stay alive, and @var{NB} the number of alive neighbor cells
19725 which make a dead cell to become alive (i.e. to "born").
19726 "s" and "b" can be used in place of "S" and "B", respectively.
19728 Alternatively a rule can be specified by an 18-bits integer. The 9
19729 high order bits are used to encode the next cell state if it is alive
19730 for each number of neighbor alive cells, the low order bits specify
19731 the rule for "borning" new cells. Higher order bits encode for an
19732 higher number of neighbor cells.
19733 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
19734 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
19736 Default value is "S23/B3", which is the original Conway's game of life
19737 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
19738 cells, and will born a new cell if there are three alive cells around
19742 Set the size of the output video. For the syntax of this option, check the
19743 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19745 If @option{filename} is specified, the size is set by default to the
19746 same size of the input file. If @option{size} is set, it must contain
19747 the size specified in the input file, and the initial grid defined in
19748 that file is centered in the larger resulting area.
19750 If a filename is not specified, the size value defaults to "320x240"
19751 (used for a randomly generated initial grid).
19754 If set to 1, stitch the left and right grid edges together, and the
19755 top and bottom edges also. Defaults to 1.
19758 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
19759 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
19760 value from 0 to 255.
19763 Set the color of living (or new born) cells.
19766 Set the color of dead cells. If @option{mold} is set, this is the first color
19767 used to represent a dead cell.
19770 Set mold color, for definitely dead and moldy cells.
19772 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
19773 ffmpeg-utils manual,ffmpeg-utils}.
19776 @subsection Examples
19780 Read a grid from @file{pattern}, and center it on a grid of size
19783 life=f=pattern:s=300x300
19787 Generate a random grid of size 200x200, with a fill ratio of 2/3:
19789 life=ratio=2/3:s=200x200
19793 Specify a custom rule for evolving a randomly generated grid:
19799 Full example with slow death effect (mold) using @command{ffplay}:
19801 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
19808 @anchor{haldclutsrc}
19811 @anchor{pal100bars}
19812 @anchor{rgbtestsrc}
19814 @anchor{smptehdbars}
19817 @anchor{yuvtestsrc}
19818 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
19820 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
19822 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
19824 The @code{color} source provides an uniformly colored input.
19826 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
19827 @ref{haldclut} filter.
19829 The @code{nullsrc} source returns unprocessed video frames. It is
19830 mainly useful to be employed in analysis / debugging tools, or as the
19831 source for filters which ignore the input data.
19833 The @code{pal75bars} source generates a color bars pattern, based on
19834 EBU PAL recommendations with 75% color levels.
19836 The @code{pal100bars} source generates a color bars pattern, based on
19837 EBU PAL recommendations with 100% color levels.
19839 The @code{rgbtestsrc} source generates an RGB test pattern useful for
19840 detecting RGB vs BGR issues. You should see a red, green and blue
19841 stripe from top to bottom.
19843 The @code{smptebars} source generates a color bars pattern, based on
19844 the SMPTE Engineering Guideline EG 1-1990.
19846 The @code{smptehdbars} source generates a color bars pattern, based on
19847 the SMPTE RP 219-2002.
19849 The @code{testsrc} source generates a test video pattern, showing a
19850 color pattern, a scrolling gradient and a timestamp. This is mainly
19851 intended for testing purposes.
19853 The @code{testsrc2} source is similar to testsrc, but supports more
19854 pixel formats instead of just @code{rgb24}. This allows using it as an
19855 input for other tests without requiring a format conversion.
19857 The @code{yuvtestsrc} source generates an YUV test pattern. You should
19858 see a y, cb and cr stripe from top to bottom.
19860 The sources accept the following parameters:
19865 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
19866 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
19867 pixels to be used as identity matrix for 3D lookup tables. Each component is
19868 coded on a @code{1/(N*N)} scale.
19871 Specify the color of the source, only available in the @code{color}
19872 source. For the syntax of this option, check the
19873 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19876 Specify the size of the sourced video. For the syntax of this option, check the
19877 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19878 The default value is @code{320x240}.
19880 This option is not available with the @code{allrgb}, @code{allyuv}, and
19881 @code{haldclutsrc} filters.
19884 Specify the frame rate of the sourced video, as the number of frames
19885 generated per second. It has to be a string in the format
19886 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
19887 number or a valid video frame rate abbreviation. The default value is
19891 Set the duration of the sourced video. See
19892 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19893 for the accepted syntax.
19895 If not specified, or the expressed duration is negative, the video is
19896 supposed to be generated forever.
19899 Set the sample aspect ratio of the sourced video.
19902 Specify the alpha (opacity) of the background, only available in the
19903 @code{testsrc2} source. The value must be between 0 (fully transparent) and
19904 255 (fully opaque, the default).
19907 Set the number of decimals to show in the timestamp, only available in the
19908 @code{testsrc} source.
19910 The displayed timestamp value will correspond to the original
19911 timestamp value multiplied by the power of 10 of the specified
19912 value. Default value is 0.
19915 @subsection Examples
19919 Generate a video with a duration of 5.3 seconds, with size
19920 176x144 and a frame rate of 10 frames per second:
19922 testsrc=duration=5.3:size=qcif:rate=10
19926 The following graph description will generate a red source
19927 with an opacity of 0.2, with size "qcif" and a frame rate of 10
19930 color=c=red@@0.2:s=qcif:r=10
19934 If the input content is to be ignored, @code{nullsrc} can be used. The
19935 following command generates noise in the luminance plane by employing
19936 the @code{geq} filter:
19938 nullsrc=s=256x256, geq=random(1)*255:128:128
19942 @subsection Commands
19944 The @code{color} source supports the following commands:
19948 Set the color of the created image. Accepts the same syntax of the
19949 corresponding @option{color} option.
19954 Generate video using an OpenCL program.
19959 OpenCL program source file.
19962 Kernel name in program.
19965 Size of frames to generate. This must be set.
19968 Pixel format to use for the generated frames. This must be set.
19971 Number of frames generated every second. Default value is '25'.
19975 For details of how the program loading works, see the @ref{program_opencl}
19982 Generate a colour ramp by setting pixel values from the position of the pixel
19983 in the output image. (Note that this will work with all pixel formats, but
19984 the generated output will not be the same.)
19986 __kernel void ramp(__write_only image2d_t dst,
19987 unsigned int index)
19989 int2 loc = (int2)(get_global_id(0), get_global_id(1));
19992 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
19994 write_imagef(dst, loc, val);
19999 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
20001 __kernel void sierpinski_carpet(__write_only image2d_t dst,
20002 unsigned int index)
20004 int2 loc = (int2)(get_global_id(0), get_global_id(1));
20006 float4 value = 0.0f;
20007 int x = loc.x + index;
20008 int y = loc.y + index;
20009 while (x > 0 || y > 0) {
20010 if (x % 3 == 1 && y % 3 == 1) {
20018 write_imagef(dst, loc, value);
20024 @c man end VIDEO SOURCES
20026 @chapter Video Sinks
20027 @c man begin VIDEO SINKS
20029 Below is a description of the currently available video sinks.
20031 @section buffersink
20033 Buffer video frames, and make them available to the end of the filter
20036 This sink is mainly intended for programmatic use, in particular
20037 through the interface defined in @file{libavfilter/buffersink.h}
20038 or the options system.
20040 It accepts a pointer to an AVBufferSinkContext structure, which
20041 defines the incoming buffers' formats, to be passed as the opaque
20042 parameter to @code{avfilter_init_filter} for initialization.
20046 Null video sink: do absolutely nothing with the input video. It is
20047 mainly useful as a template and for use in analysis / debugging
20050 @c man end VIDEO SINKS
20052 @chapter Multimedia Filters
20053 @c man begin MULTIMEDIA FILTERS
20055 Below is a description of the currently available multimedia filters.
20059 Convert input audio to a video output, displaying the audio bit scope.
20061 The filter accepts the following options:
20065 Set frame rate, expressed as number of frames per second. Default
20069 Specify the video size for the output. For the syntax of this option, check the
20070 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20071 Default value is @code{1024x256}.
20074 Specify list of colors separated by space or by '|' which will be used to
20075 draw channels. Unrecognized or missing colors will be replaced
20079 @section ahistogram
20081 Convert input audio to a video output, displaying the volume histogram.
20083 The filter accepts the following options:
20087 Specify how histogram is calculated.
20089 It accepts the following values:
20092 Use single histogram for all channels.
20094 Use separate histogram for each channel.
20096 Default is @code{single}.
20099 Set frame rate, expressed as number of frames per second. Default
20103 Specify the video size for the output. For the syntax of this option, check the
20104 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20105 Default value is @code{hd720}.
20110 It accepts the following values:
20121 reverse logarithmic
20123 Default is @code{log}.
20126 Set amplitude scale.
20128 It accepts the following values:
20135 Default is @code{log}.
20138 Set how much frames to accumulate in histogram.
20139 Default is 1. Setting this to -1 accumulates all frames.
20142 Set histogram ratio of window height.
20145 Set sonogram sliding.
20147 It accepts the following values:
20150 replace old rows with new ones.
20152 scroll from top to bottom.
20154 Default is @code{replace}.
20157 @section aphasemeter
20159 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
20160 representing mean phase of current audio frame. A video output can also be produced and is
20161 enabled by default. The audio is passed through as first output.
20163 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
20164 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
20165 and @code{1} means channels are in phase.
20167 The filter accepts the following options, all related to its video output:
20171 Set the output frame rate. Default value is @code{25}.
20174 Set the video size for the output. For the syntax of this option, check the
20175 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20176 Default value is @code{800x400}.
20181 Specify the red, green, blue contrast. Default values are @code{2},
20182 @code{7} and @code{1}.
20183 Allowed range is @code{[0, 255]}.
20186 Set color which will be used for drawing median phase. If color is
20187 @code{none} which is default, no median phase value will be drawn.
20190 Enable video output. Default is enabled.
20193 @section avectorscope
20195 Convert input audio to a video output, representing the audio vector
20198 The filter is used to measure the difference between channels of stereo
20199 audio stream. A monoaural signal, consisting of identical left and right
20200 signal, results in straight vertical line. Any stereo separation is visible
20201 as a deviation from this line, creating a Lissajous figure.
20202 If the straight (or deviation from it) but horizontal line appears this
20203 indicates that the left and right channels are out of phase.
20205 The filter accepts the following options:
20209 Set the vectorscope mode.
20211 Available values are:
20214 Lissajous rotated by 45 degrees.
20217 Same as above but not rotated.
20220 Shape resembling half of circle.
20223 Default value is @samp{lissajous}.
20226 Set the video size for the output. For the syntax of this option, check the
20227 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20228 Default value is @code{400x400}.
20231 Set the output frame rate. Default value is @code{25}.
20237 Specify the red, green, blue and alpha contrast. Default values are @code{40},
20238 @code{160}, @code{80} and @code{255}.
20239 Allowed range is @code{[0, 255]}.
20245 Specify the red, green, blue and alpha fade. Default values are @code{15},
20246 @code{10}, @code{5} and @code{5}.
20247 Allowed range is @code{[0, 255]}.
20250 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
20251 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
20254 Set the vectorscope drawing mode.
20256 Available values are:
20259 Draw dot for each sample.
20262 Draw line between previous and current sample.
20265 Default value is @samp{dot}.
20268 Specify amplitude scale of audio samples.
20270 Available values are:
20286 Swap left channel axis with right channel axis.
20296 Mirror only x axis.
20299 Mirror only y axis.
20307 @subsection Examples
20311 Complete example using @command{ffplay}:
20313 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
20314 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
20318 @section bench, abench
20320 Benchmark part of a filtergraph.
20322 The filter accepts the following options:
20326 Start or stop a timer.
20328 Available values are:
20331 Get the current time, set it as frame metadata (using the key
20332 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
20335 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
20336 the input frame metadata to get the time difference. Time difference, average,
20337 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
20338 @code{min}) are then printed. The timestamps are expressed in seconds.
20342 @subsection Examples
20346 Benchmark @ref{selectivecolor} filter:
20348 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
20354 Concatenate audio and video streams, joining them together one after the
20357 The filter works on segments of synchronized video and audio streams. All
20358 segments must have the same number of streams of each type, and that will
20359 also be the number of streams at output.
20361 The filter accepts the following options:
20366 Set the number of segments. Default is 2.
20369 Set the number of output video streams, that is also the number of video
20370 streams in each segment. Default is 1.
20373 Set the number of output audio streams, that is also the number of audio
20374 streams in each segment. Default is 0.
20377 Activate unsafe mode: do not fail if segments have a different format.
20381 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
20382 @var{a} audio outputs.
20384 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
20385 segment, in the same order as the outputs, then the inputs for the second
20388 Related streams do not always have exactly the same duration, for various
20389 reasons including codec frame size or sloppy authoring. For that reason,
20390 related synchronized streams (e.g. a video and its audio track) should be
20391 concatenated at once. The concat filter will use the duration of the longest
20392 stream in each segment (except the last one), and if necessary pad shorter
20393 audio streams with silence.
20395 For this filter to work correctly, all segments must start at timestamp 0.
20397 All corresponding streams must have the same parameters in all segments; the
20398 filtering system will automatically select a common pixel format for video
20399 streams, and a common sample format, sample rate and channel layout for
20400 audio streams, but other settings, such as resolution, must be converted
20401 explicitly by the user.
20403 Different frame rates are acceptable but will result in variable frame rate
20404 at output; be sure to configure the output file to handle it.
20406 @subsection Examples
20410 Concatenate an opening, an episode and an ending, all in bilingual version
20411 (video in stream 0, audio in streams 1 and 2):
20413 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
20414 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
20415 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
20416 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
20420 Concatenate two parts, handling audio and video separately, using the
20421 (a)movie sources, and adjusting the resolution:
20423 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
20424 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
20425 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
20427 Note that a desync will happen at the stitch if the audio and video streams
20428 do not have exactly the same duration in the first file.
20432 @subsection Commands
20434 This filter supports the following commands:
20437 Close the current segment and step to the next one
20440 @section drawgraph, adrawgraph
20442 Draw a graph using input video or audio metadata.
20444 It accepts the following parameters:
20448 Set 1st frame metadata key from which metadata values will be used to draw a graph.
20451 Set 1st foreground color expression.
20454 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
20457 Set 2nd foreground color expression.
20460 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
20463 Set 3rd foreground color expression.
20466 Set 4th frame metadata key from which metadata values will be used to draw a graph.
20469 Set 4th foreground color expression.
20472 Set minimal value of metadata value.
20475 Set maximal value of metadata value.
20478 Set graph background color. Default is white.
20483 Available values for mode is:
20490 Default is @code{line}.
20495 Available values for slide is:
20498 Draw new frame when right border is reached.
20501 Replace old columns with new ones.
20504 Scroll from right to left.
20507 Scroll from left to right.
20510 Draw single picture.
20513 Default is @code{frame}.
20516 Set size of graph video. For the syntax of this option, check the
20517 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20518 The default value is @code{900x256}.
20520 The foreground color expressions can use the following variables:
20523 Minimal value of metadata value.
20526 Maximal value of metadata value.
20529 Current metadata key value.
20532 The color is defined as 0xAABBGGRR.
20535 Example using metadata from @ref{signalstats} filter:
20537 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
20540 Example using metadata from @ref{ebur128} filter:
20542 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
20548 EBU R128 scanner filter. This filter takes an audio stream as input and outputs
20549 it unchanged. By default, it logs a message at a frequency of 10Hz with the
20550 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
20551 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
20553 The filter also has a video output (see the @var{video} option) with a real
20554 time graph to observe the loudness evolution. The graphic contains the logged
20555 message mentioned above, so it is not printed anymore when this option is set,
20556 unless the verbose logging is set. The main graphing area contains the
20557 short-term loudness (3 seconds of analysis), and the gauge on the right is for
20558 the momentary loudness (400 milliseconds), but can optionally be configured
20559 to instead display short-term loudness (see @var{gauge}).
20561 The green area marks a +/- 1LU target range around the target loudness
20562 (-23LUFS by default, unless modified through @var{target}).
20564 More information about the Loudness Recommendation EBU R128 on
20565 @url{http://tech.ebu.ch/loudness}.
20567 The filter accepts the following options:
20572 Activate the video output. The audio stream is passed unchanged whether this
20573 option is set or no. The video stream will be the first output stream if
20574 activated. Default is @code{0}.
20577 Set the video size. This option is for video only. For the syntax of this
20579 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20580 Default and minimum resolution is @code{640x480}.
20583 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
20584 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
20585 other integer value between this range is allowed.
20588 Set metadata injection. If set to @code{1}, the audio input will be segmented
20589 into 100ms output frames, each of them containing various loudness information
20590 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
20592 Default is @code{0}.
20595 Force the frame logging level.
20597 Available values are:
20600 information logging level
20602 verbose logging level
20605 By default, the logging level is set to @var{info}. If the @option{video} or
20606 the @option{metadata} options are set, it switches to @var{verbose}.
20611 Available modes can be cumulated (the option is a @code{flag} type). Possible
20615 Disable any peak mode (default).
20617 Enable sample-peak mode.
20619 Simple peak mode looking for the higher sample value. It logs a message
20620 for sample-peak (identified by @code{SPK}).
20622 Enable true-peak mode.
20624 If enabled, the peak lookup is done on an over-sampled version of the input
20625 stream for better peak accuracy. It logs a message for true-peak.
20626 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
20627 This mode requires a build with @code{libswresample}.
20631 Treat mono input files as "dual mono". If a mono file is intended for playback
20632 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
20633 If set to @code{true}, this option will compensate for this effect.
20634 Multi-channel input files are not affected by this option.
20637 Set a specific pan law to be used for the measurement of dual mono files.
20638 This parameter is optional, and has a default value of -3.01dB.
20641 Set a specific target level (in LUFS) used as relative zero in the visualization.
20642 This parameter is optional and has a default value of -23LUFS as specified
20643 by EBU R128. However, material published online may prefer a level of -16LUFS
20644 (e.g. for use with podcasts or video platforms).
20647 Set the value displayed by the gauge. Valid values are @code{momentary} and s
20648 @code{shortterm}. By default the momentary value will be used, but in certain
20649 scenarios it may be more useful to observe the short term value instead (e.g.
20653 Sets the display scale for the loudness. Valid parameters are @code{absolute}
20654 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
20655 video output, not the summary or continuous log output.
20658 @subsection Examples
20662 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
20664 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
20668 Run an analysis with @command{ffmpeg}:
20670 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
20674 @section interleave, ainterleave
20676 Temporally interleave frames from several inputs.
20678 @code{interleave} works with video inputs, @code{ainterleave} with audio.
20680 These filters read frames from several inputs and send the oldest
20681 queued frame to the output.
20683 Input streams must have well defined, monotonically increasing frame
20686 In order to submit one frame to output, these filters need to enqueue
20687 at least one frame for each input, so they cannot work in case one
20688 input is not yet terminated and will not receive incoming frames.
20690 For example consider the case when one input is a @code{select} filter
20691 which always drops input frames. The @code{interleave} filter will keep
20692 reading from that input, but it will never be able to send new frames
20693 to output until the input sends an end-of-stream signal.
20695 Also, depending on inputs synchronization, the filters will drop
20696 frames in case one input receives more frames than the other ones, and
20697 the queue is already filled.
20699 These filters accept the following options:
20703 Set the number of different inputs, it is 2 by default.
20706 @subsection Examples
20710 Interleave frames belonging to different streams using @command{ffmpeg}:
20712 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
20716 Add flickering blur effect:
20718 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
20722 @section metadata, ametadata
20724 Manipulate frame metadata.
20726 This filter accepts the following options:
20730 Set mode of operation of the filter.
20732 Can be one of the following:
20736 If both @code{value} and @code{key} is set, select frames
20737 which have such metadata. If only @code{key} is set, select
20738 every frame that has such key in metadata.
20741 Add new metadata @code{key} and @code{value}. If key is already available
20745 Modify value of already present key.
20748 If @code{value} is set, delete only keys that have such value.
20749 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
20753 Print key and its value if metadata was found. If @code{key} is not set print all
20754 metadata values available in frame.
20758 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
20761 Set metadata value which will be used. This option is mandatory for
20762 @code{modify} and @code{add} mode.
20765 Which function to use when comparing metadata value and @code{value}.
20767 Can be one of following:
20771 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
20774 Values are interpreted as strings, returns true if metadata value starts with
20775 the @code{value} option string.
20778 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
20781 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
20784 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
20787 Values are interpreted as floats, returns true if expression from option @code{expr}
20792 Set expression which is used when @code{function} is set to @code{expr}.
20793 The expression is evaluated through the eval API and can contain the following
20798 Float representation of @code{value} from metadata key.
20801 Float representation of @code{value} as supplied by user in @code{value} option.
20805 If specified in @code{print} mode, output is written to the named file. Instead of
20806 plain filename any writable url can be specified. Filename ``-'' is a shorthand
20807 for standard output. If @code{file} option is not set, output is written to the log
20808 with AV_LOG_INFO loglevel.
20812 @subsection Examples
20816 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
20819 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
20822 Print silencedetect output to file @file{metadata.txt}.
20824 silencedetect,ametadata=mode=print:file=metadata.txt
20827 Direct all metadata to a pipe with file descriptor 4.
20829 metadata=mode=print:file='pipe\:4'
20833 @section perms, aperms
20835 Set read/write permissions for the output frames.
20837 These filters are mainly aimed at developers to test direct path in the
20838 following filter in the filtergraph.
20840 The filters accept the following options:
20844 Select the permissions mode.
20846 It accepts the following values:
20849 Do nothing. This is the default.
20851 Set all the output frames read-only.
20853 Set all the output frames directly writable.
20855 Make the frame read-only if writable, and writable if read-only.
20857 Set each output frame read-only or writable randomly.
20861 Set the seed for the @var{random} mode, must be an integer included between
20862 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
20863 @code{-1}, the filter will try to use a good random seed on a best effort
20867 Note: in case of auto-inserted filter between the permission filter and the
20868 following one, the permission might not be received as expected in that
20869 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
20870 perms/aperms filter can avoid this problem.
20872 @section realtime, arealtime
20874 Slow down filtering to match real time approximately.
20876 These filters will pause the filtering for a variable amount of time to
20877 match the output rate with the input timestamps.
20878 They are similar to the @option{re} option to @code{ffmpeg}.
20880 They accept the following options:
20884 Time limit for the pauses. Any pause longer than that will be considered
20885 a timestamp discontinuity and reset the timer. Default is 2 seconds.
20889 @section select, aselect
20891 Select frames to pass in output.
20893 This filter accepts the following options:
20898 Set expression, which is evaluated for each input frame.
20900 If the expression is evaluated to zero, the frame is discarded.
20902 If the evaluation result is negative or NaN, the frame is sent to the
20903 first output; otherwise it is sent to the output with index
20904 @code{ceil(val)-1}, assuming that the input index starts from 0.
20906 For example a value of @code{1.2} corresponds to the output with index
20907 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
20910 Set the number of outputs. The output to which to send the selected
20911 frame is based on the result of the evaluation. Default value is 1.
20914 The expression can contain the following constants:
20918 The (sequential) number of the filtered frame, starting from 0.
20921 The (sequential) number of the selected frame, starting from 0.
20923 @item prev_selected_n
20924 The sequential number of the last selected frame. It's NAN if undefined.
20927 The timebase of the input timestamps.
20930 The PTS (Presentation TimeStamp) of the filtered video frame,
20931 expressed in @var{TB} units. It's NAN if undefined.
20934 The PTS of the filtered video frame,
20935 expressed in seconds. It's NAN if undefined.
20938 The PTS of the previously filtered video frame. It's NAN if undefined.
20940 @item prev_selected_pts
20941 The PTS of the last previously filtered video frame. It's NAN if undefined.
20943 @item prev_selected_t
20944 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
20947 The PTS of the first video frame in the video. It's NAN if undefined.
20950 The time of the first video frame in the video. It's NAN if undefined.
20952 @item pict_type @emph{(video only)}
20953 The type of the filtered frame. It can assume one of the following
20965 @item interlace_type @emph{(video only)}
20966 The frame interlace type. It can assume one of the following values:
20969 The frame is progressive (not interlaced).
20971 The frame is top-field-first.
20973 The frame is bottom-field-first.
20976 @item consumed_sample_n @emph{(audio only)}
20977 the number of selected samples before the current frame
20979 @item samples_n @emph{(audio only)}
20980 the number of samples in the current frame
20982 @item sample_rate @emph{(audio only)}
20983 the input sample rate
20986 This is 1 if the filtered frame is a key-frame, 0 otherwise.
20989 the position in the file of the filtered frame, -1 if the information
20990 is not available (e.g. for synthetic video)
20992 @item scene @emph{(video only)}
20993 value between 0 and 1 to indicate a new scene; a low value reflects a low
20994 probability for the current frame to introduce a new scene, while a higher
20995 value means the current frame is more likely to be one (see the example below)
20997 @item concatdec_select
20998 The concat demuxer can select only part of a concat input file by setting an
20999 inpoint and an outpoint, but the output packets may not be entirely contained
21000 in the selected interval. By using this variable, it is possible to skip frames
21001 generated by the concat demuxer which are not exactly contained in the selected
21004 This works by comparing the frame pts against the @var{lavf.concat.start_time}
21005 and the @var{lavf.concat.duration} packet metadata values which are also
21006 present in the decoded frames.
21008 The @var{concatdec_select} variable is -1 if the frame pts is at least
21009 start_time and either the duration metadata is missing or the frame pts is less
21010 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
21013 That basically means that an input frame is selected if its pts is within the
21014 interval set by the concat demuxer.
21018 The default value of the select expression is "1".
21020 @subsection Examples
21024 Select all frames in input:
21029 The example above is the same as:
21041 Select only I-frames:
21043 select='eq(pict_type\,I)'
21047 Select one frame every 100:
21049 select='not(mod(n\,100))'
21053 Select only frames contained in the 10-20 time interval:
21055 select=between(t\,10\,20)
21059 Select only I-frames contained in the 10-20 time interval:
21061 select=between(t\,10\,20)*eq(pict_type\,I)
21065 Select frames with a minimum distance of 10 seconds:
21067 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
21071 Use aselect to select only audio frames with samples number > 100:
21073 aselect='gt(samples_n\,100)'
21077 Create a mosaic of the first scenes:
21079 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
21082 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
21086 Send even and odd frames to separate outputs, and compose them:
21088 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
21092 Select useful frames from an ffconcat file which is using inpoints and
21093 outpoints but where the source files are not intra frame only.
21095 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
21099 @section sendcmd, asendcmd
21101 Send commands to filters in the filtergraph.
21103 These filters read commands to be sent to other filters in the
21106 @code{sendcmd} must be inserted between two video filters,
21107 @code{asendcmd} must be inserted between two audio filters, but apart
21108 from that they act the same way.
21110 The specification of commands can be provided in the filter arguments
21111 with the @var{commands} option, or in a file specified by the
21112 @var{filename} option.
21114 These filters accept the following options:
21117 Set the commands to be read and sent to the other filters.
21119 Set the filename of the commands to be read and sent to the other
21123 @subsection Commands syntax
21125 A commands description consists of a sequence of interval
21126 specifications, comprising a list of commands to be executed when a
21127 particular event related to that interval occurs. The occurring event
21128 is typically the current frame time entering or leaving a given time
21131 An interval is specified by the following syntax:
21133 @var{START}[-@var{END}] @var{COMMANDS};
21136 The time interval is specified by the @var{START} and @var{END} times.
21137 @var{END} is optional and defaults to the maximum time.
21139 The current frame time is considered within the specified interval if
21140 it is included in the interval [@var{START}, @var{END}), that is when
21141 the time is greater or equal to @var{START} and is lesser than
21144 @var{COMMANDS} consists of a sequence of one or more command
21145 specifications, separated by ",", relating to that interval. The
21146 syntax of a command specification is given by:
21148 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
21151 @var{FLAGS} is optional and specifies the type of events relating to
21152 the time interval which enable sending the specified command, and must
21153 be a non-null sequence of identifier flags separated by "+" or "|" and
21154 enclosed between "[" and "]".
21156 The following flags are recognized:
21159 The command is sent when the current frame timestamp enters the
21160 specified interval. In other words, the command is sent when the
21161 previous frame timestamp was not in the given interval, and the
21165 The command is sent when the current frame timestamp leaves the
21166 specified interval. In other words, the command is sent when the
21167 previous frame timestamp was in the given interval, and the
21171 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
21174 @var{TARGET} specifies the target of the command, usually the name of
21175 the filter class or a specific filter instance name.
21177 @var{COMMAND} specifies the name of the command for the target filter.
21179 @var{ARG} is optional and specifies the optional list of argument for
21180 the given @var{COMMAND}.
21182 Between one interval specification and another, whitespaces, or
21183 sequences of characters starting with @code{#} until the end of line,
21184 are ignored and can be used to annotate comments.
21186 A simplified BNF description of the commands specification syntax
21189 @var{COMMAND_FLAG} ::= "enter" | "leave"
21190 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
21191 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
21192 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
21193 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
21194 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
21197 @subsection Examples
21201 Specify audio tempo change at second 4:
21203 asendcmd=c='4.0 atempo tempo 1.5',atempo
21207 Target a specific filter instance:
21209 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
21213 Specify a list of drawtext and hue commands in a file.
21215 # show text in the interval 5-10
21216 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
21217 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
21219 # desaturate the image in the interval 15-20
21220 15.0-20.0 [enter] hue s 0,
21221 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
21223 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
21225 # apply an exponential saturation fade-out effect, starting from time 25
21226 25 [enter] hue s exp(25-t)
21229 A filtergraph allowing to read and process the above command list
21230 stored in a file @file{test.cmd}, can be specified with:
21232 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
21237 @section setpts, asetpts
21239 Change the PTS (presentation timestamp) of the input frames.
21241 @code{setpts} works on video frames, @code{asetpts} on audio frames.
21243 This filter accepts the following options:
21248 The expression which is evaluated for each frame to construct its timestamp.
21252 The expression is evaluated through the eval API and can contain the following
21256 @item FRAME_RATE, FR
21257 frame rate, only defined for constant frame-rate video
21260 The presentation timestamp in input
21263 The count of the input frame for video or the number of consumed samples,
21264 not including the current frame for audio, starting from 0.
21266 @item NB_CONSUMED_SAMPLES
21267 The number of consumed samples, not including the current frame (only
21270 @item NB_SAMPLES, S
21271 The number of samples in the current frame (only audio)
21273 @item SAMPLE_RATE, SR
21274 The audio sample rate.
21277 The PTS of the first frame.
21280 the time in seconds of the first frame
21283 State whether the current frame is interlaced.
21286 the time in seconds of the current frame
21289 original position in the file of the frame, or undefined if undefined
21290 for the current frame
21293 The previous input PTS.
21296 previous input time in seconds
21299 The previous output PTS.
21302 previous output time in seconds
21305 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
21309 The wallclock (RTC) time at the start of the movie in microseconds.
21312 The timebase of the input timestamps.
21316 @subsection Examples
21320 Start counting PTS from zero
21322 setpts=PTS-STARTPTS
21326 Apply fast motion effect:
21332 Apply slow motion effect:
21338 Set fixed rate of 25 frames per second:
21344 Set fixed rate 25 fps with some jitter:
21346 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
21350 Apply an offset of 10 seconds to the input PTS:
21356 Generate timestamps from a "live source" and rebase onto the current timebase:
21358 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
21362 Generate timestamps by counting samples:
21371 Force color range for the output video frame.
21373 The @code{setrange} filter marks the color range property for the
21374 output frames. It does not change the input frame, but only sets the
21375 corresponding property, which affects how the frame is treated by
21378 The filter accepts the following options:
21383 Available values are:
21387 Keep the same color range property.
21389 @item unspecified, unknown
21390 Set the color range as unspecified.
21392 @item limited, tv, mpeg
21393 Set the color range as limited.
21395 @item full, pc, jpeg
21396 Set the color range as full.
21400 @section settb, asettb
21402 Set the timebase to use for the output frames timestamps.
21403 It is mainly useful for testing timebase configuration.
21405 It accepts the following parameters:
21410 The expression which is evaluated into the output timebase.
21414 The value for @option{tb} is an arithmetic expression representing a
21415 rational. The expression can contain the constants "AVTB" (the default
21416 timebase), "intb" (the input timebase) and "sr" (the sample rate,
21417 audio only). Default value is "intb".
21419 @subsection Examples
21423 Set the timebase to 1/25:
21429 Set the timebase to 1/10:
21435 Set the timebase to 1001/1000:
21441 Set the timebase to 2*intb:
21447 Set the default timebase value:
21454 Convert input audio to a video output representing frequency spectrum
21455 logarithmically using Brown-Puckette constant Q transform algorithm with
21456 direct frequency domain coefficient calculation (but the transform itself
21457 is not really constant Q, instead the Q factor is actually variable/clamped),
21458 with musical tone scale, from E0 to D#10.
21460 The filter accepts the following options:
21464 Specify the video size for the output. It must be even. For the syntax of this option,
21465 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21466 Default value is @code{1920x1080}.
21469 Set the output frame rate. Default value is @code{25}.
21472 Set the bargraph height. It must be even. Default value is @code{-1} which
21473 computes the bargraph height automatically.
21476 Set the axis height. It must be even. Default value is @code{-1} which computes
21477 the axis height automatically.
21480 Set the sonogram height. It must be even. Default value is @code{-1} which
21481 computes the sonogram height automatically.
21484 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
21485 instead. Default value is @code{1}.
21487 @item sono_v, volume
21488 Specify the sonogram volume expression. It can contain variables:
21491 the @var{bar_v} evaluated expression
21492 @item frequency, freq, f
21493 the frequency where it is evaluated
21494 @item timeclamp, tc
21495 the value of @var{timeclamp} option
21499 @item a_weighting(f)
21500 A-weighting of equal loudness
21501 @item b_weighting(f)
21502 B-weighting of equal loudness
21503 @item c_weighting(f)
21504 C-weighting of equal loudness.
21506 Default value is @code{16}.
21508 @item bar_v, volume2
21509 Specify the bargraph volume expression. It can contain variables:
21512 the @var{sono_v} evaluated expression
21513 @item frequency, freq, f
21514 the frequency where it is evaluated
21515 @item timeclamp, tc
21516 the value of @var{timeclamp} option
21520 @item a_weighting(f)
21521 A-weighting of equal loudness
21522 @item b_weighting(f)
21523 B-weighting of equal loudness
21524 @item c_weighting(f)
21525 C-weighting of equal loudness.
21527 Default value is @code{sono_v}.
21529 @item sono_g, gamma
21530 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
21531 higher gamma makes the spectrum having more range. Default value is @code{3}.
21532 Acceptable range is @code{[1, 7]}.
21534 @item bar_g, gamma2
21535 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
21539 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
21540 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
21542 @item timeclamp, tc
21543 Specify the transform timeclamp. At low frequency, there is trade-off between
21544 accuracy in time domain and frequency domain. If timeclamp is lower,
21545 event in time domain is represented more accurately (such as fast bass drum),
21546 otherwise event in frequency domain is represented more accurately
21547 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
21550 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
21551 limits future samples by applying asymmetric windowing in time domain, useful
21552 when low latency is required. Accepted range is @code{[0, 1]}.
21555 Specify the transform base frequency. Default value is @code{20.01523126408007475},
21556 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
21559 Specify the transform end frequency. Default value is @code{20495.59681441799654},
21560 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
21563 This option is deprecated and ignored.
21566 Specify the transform length in time domain. Use this option to control accuracy
21567 trade-off between time domain and frequency domain at every frequency sample.
21568 It can contain variables:
21570 @item frequency, freq, f
21571 the frequency where it is evaluated
21572 @item timeclamp, tc
21573 the value of @var{timeclamp} option.
21575 Default value is @code{384*tc/(384+tc*f)}.
21578 Specify the transform count for every video frame. Default value is @code{6}.
21579 Acceptable range is @code{[1, 30]}.
21582 Specify the transform count for every single pixel. Default value is @code{0},
21583 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
21586 Specify font file for use with freetype to draw the axis. If not specified,
21587 use embedded font. Note that drawing with font file or embedded font is not
21588 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
21592 Specify fontconfig pattern. This has lower priority than @var{fontfile}.
21593 The : in the pattern may be replaced by | to avoid unnecessary escaping.
21596 Specify font color expression. This is arithmetic expression that should return
21597 integer value 0xRRGGBB. It can contain variables:
21599 @item frequency, freq, f
21600 the frequency where it is evaluated
21601 @item timeclamp, tc
21602 the value of @var{timeclamp} option
21607 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
21608 @item r(x), g(x), b(x)
21609 red, green, and blue value of intensity x.
21611 Default value is @code{st(0, (midi(f)-59.5)/12);
21612 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
21613 r(1-ld(1)) + b(ld(1))}.
21616 Specify image file to draw the axis. This option override @var{fontfile} and
21617 @var{fontcolor} option.
21620 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
21621 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
21622 Default value is @code{1}.
21625 Set colorspace. The accepted values are:
21628 Unspecified (default)
21637 BT.470BG or BT.601-6 625
21640 SMPTE-170M or BT.601-6 525
21646 BT.2020 with non-constant luminance
21651 Set spectrogram color scheme. This is list of floating point values with format
21652 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
21653 The default is @code{1|0.5|0|0|0.5|1}.
21657 @subsection Examples
21661 Playing audio while showing the spectrum:
21663 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
21667 Same as above, but with frame rate 30 fps:
21669 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
21673 Playing at 1280x720:
21675 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
21679 Disable sonogram display:
21685 A1 and its harmonics: A1, A2, (near)E3, A3:
21687 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),
21688 asplit[a][out1]; [a] showcqt [out0]'
21692 Same as above, but with more accuracy in frequency domain:
21694 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),
21695 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
21701 bar_v=10:sono_v=bar_v*a_weighting(f)
21705 Custom gamma, now spectrum is linear to the amplitude.
21711 Custom tlength equation:
21713 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)))'
21717 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
21719 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
21723 Custom font using fontconfig:
21725 font='Courier New,Monospace,mono|bold'
21729 Custom frequency range with custom axis using image file:
21731 axisfile=myaxis.png:basefreq=40:endfreq=10000
21737 Convert input audio to video output representing the audio power spectrum.
21738 Audio amplitude is on Y-axis while frequency is on X-axis.
21740 The filter accepts the following options:
21744 Specify size of video. For the syntax of this option, check the
21745 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21746 Default is @code{1024x512}.
21750 This set how each frequency bin will be represented.
21752 It accepts the following values:
21758 Default is @code{bar}.
21761 Set amplitude scale.
21763 It accepts the following values:
21777 Default is @code{log}.
21780 Set frequency scale.
21782 It accepts the following values:
21791 Reverse logarithmic scale.
21793 Default is @code{lin}.
21798 It accepts the following values:
21814 Default is @code{w2048}
21817 Set windowing function.
21819 It accepts the following values:
21842 Default is @code{hanning}.
21845 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
21846 which means optimal overlap for selected window function will be picked.
21849 Set time averaging. Setting this to 0 will display current maximal peaks.
21850 Default is @code{1}, which means time averaging is disabled.
21853 Specify list of colors separated by space or by '|' which will be used to
21854 draw channel frequencies. Unrecognized or missing colors will be replaced
21858 Set channel display mode.
21860 It accepts the following values:
21865 Default is @code{combined}.
21868 Set minimum amplitude used in @code{log} amplitude scaler.
21872 @anchor{showspectrum}
21873 @section showspectrum
21875 Convert input audio to a video output, representing the audio frequency
21878 The filter accepts the following options:
21882 Specify the video size for the output. For the syntax of this option, check the
21883 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21884 Default value is @code{640x512}.
21887 Specify how the spectrum should slide along the window.
21889 It accepts the following values:
21892 the samples start again on the left when they reach the right
21894 the samples scroll from right to left
21896 frames are only produced when the samples reach the right
21898 the samples scroll from left to right
21901 Default value is @code{replace}.
21904 Specify display mode.
21906 It accepts the following values:
21909 all channels are displayed in the same row
21911 all channels are displayed in separate rows
21914 Default value is @samp{combined}.
21917 Specify display color mode.
21919 It accepts the following values:
21922 each channel is displayed in a separate color
21924 each channel is displayed using the same color scheme
21926 each channel is displayed using the rainbow color scheme
21928 each channel is displayed using the moreland color scheme
21930 each channel is displayed using the nebulae color scheme
21932 each channel is displayed using the fire color scheme
21934 each channel is displayed using the fiery color scheme
21936 each channel is displayed using the fruit color scheme
21938 each channel is displayed using the cool color scheme
21940 each channel is displayed using the magma color scheme
21942 each channel is displayed using the green color scheme
21944 each channel is displayed using the viridis color scheme
21946 each channel is displayed using the plasma color scheme
21948 each channel is displayed using the cividis color scheme
21950 each channel is displayed using the terrain color scheme
21953 Default value is @samp{channel}.
21956 Specify scale used for calculating intensity color values.
21958 It accepts the following values:
21963 square root, default
21974 Default value is @samp{sqrt}.
21977 Set saturation modifier for displayed colors. Negative values provide
21978 alternative color scheme. @code{0} is no saturation at all.
21979 Saturation must be in [-10.0, 10.0] range.
21980 Default value is @code{1}.
21983 Set window function.
21985 It accepts the following values:
22010 Default value is @code{hann}.
22013 Set orientation of time vs frequency axis. Can be @code{vertical} or
22014 @code{horizontal}. Default is @code{vertical}.
22017 Set ratio of overlap window. Default value is @code{0}.
22018 When value is @code{1} overlap is set to recommended size for specific
22019 window function currently used.
22022 Set scale gain for calculating intensity color values.
22023 Default value is @code{1}.
22026 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
22029 Set color rotation, must be in [-1.0, 1.0] range.
22030 Default value is @code{0}.
22033 Set start frequency from which to display spectrogram. Default is @code{0}.
22036 Set stop frequency to which to display spectrogram. Default is @code{0}.
22039 Set upper frame rate limit. Default is @code{auto}, unlimited.
22042 Draw time and frequency axes and legends. Default is disabled.
22045 The usage is very similar to the showwaves filter; see the examples in that
22048 @subsection Examples
22052 Large window with logarithmic color scaling:
22054 showspectrum=s=1280x480:scale=log
22058 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
22060 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
22061 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
22065 @section showspectrumpic
22067 Convert input audio to a single video frame, representing the audio frequency
22070 The filter accepts the following options:
22074 Specify the video size for the output. For the syntax of this option, check the
22075 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22076 Default value is @code{4096x2048}.
22079 Specify display mode.
22081 It accepts the following values:
22084 all channels are displayed in the same row
22086 all channels are displayed in separate rows
22088 Default value is @samp{combined}.
22091 Specify display color mode.
22093 It accepts the following values:
22096 each channel is displayed in a separate color
22098 each channel is displayed using the same color scheme
22100 each channel is displayed using the rainbow color scheme
22102 each channel is displayed using the moreland color scheme
22104 each channel is displayed using the nebulae color scheme
22106 each channel is displayed using the fire color scheme
22108 each channel is displayed using the fiery color scheme
22110 each channel is displayed using the fruit color scheme
22112 each channel is displayed using the cool color scheme
22114 each channel is displayed using the magma color scheme
22116 each channel is displayed using the green color scheme
22118 each channel is displayed using the viridis color scheme
22120 each channel is displayed using the plasma color scheme
22122 each channel is displayed using the cividis color scheme
22124 each channel is displayed using the terrain color scheme
22126 Default value is @samp{intensity}.
22129 Specify scale used for calculating intensity color values.
22131 It accepts the following values:
22136 square root, default
22146 Default value is @samp{log}.
22149 Set saturation modifier for displayed colors. Negative values provide
22150 alternative color scheme. @code{0} is no saturation at all.
22151 Saturation must be in [-10.0, 10.0] range.
22152 Default value is @code{1}.
22155 Set window function.
22157 It accepts the following values:
22181 Default value is @code{hann}.
22184 Set orientation of time vs frequency axis. Can be @code{vertical} or
22185 @code{horizontal}. Default is @code{vertical}.
22188 Set scale gain for calculating intensity color values.
22189 Default value is @code{1}.
22192 Draw time and frequency axes and legends. Default is enabled.
22195 Set color rotation, must be in [-1.0, 1.0] range.
22196 Default value is @code{0}.
22199 Set start frequency from which to display spectrogram. Default is @code{0}.
22202 Set stop frequency to which to display spectrogram. Default is @code{0}.
22205 @subsection Examples
22209 Extract an audio spectrogram of a whole audio track
22210 in a 1024x1024 picture using @command{ffmpeg}:
22212 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
22216 @section showvolume
22218 Convert input audio volume to a video output.
22220 The filter accepts the following options:
22227 Set border width, allowed range is [0, 5]. Default is 1.
22230 Set channel width, allowed range is [80, 8192]. Default is 400.
22233 Set channel height, allowed range is [1, 900]. Default is 20.
22236 Set fade, allowed range is [0, 1]. Default is 0.95.
22239 Set volume color expression.
22241 The expression can use the following variables:
22245 Current max volume of channel in dB.
22251 Current channel number, starting from 0.
22255 If set, displays channel names. Default is enabled.
22258 If set, displays volume values. Default is enabled.
22261 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
22262 default is @code{h}.
22265 Set step size, allowed range is [0, 5]. Default is 0, which means
22269 Set background opacity, allowed range is [0, 1]. Default is 0.
22272 Set metering mode, can be peak: @code{p} or rms: @code{r},
22273 default is @code{p}.
22276 Set display scale, can be linear: @code{lin} or log: @code{log},
22277 default is @code{lin}.
22281 If set to > 0., display a line for the max level
22282 in the previous seconds.
22283 default is disabled: @code{0.}
22286 The color of the max line. Use when @code{dm} option is set to > 0.
22287 default is: @code{orange}
22292 Convert input audio to a video output, representing the samples waves.
22294 The filter accepts the following options:
22298 Specify the video size for the output. For the syntax of this option, check the
22299 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22300 Default value is @code{600x240}.
22305 Available values are:
22308 Draw a point for each sample.
22311 Draw a vertical line for each sample.
22314 Draw a point for each sample and a line between them.
22317 Draw a centered vertical line for each sample.
22320 Default value is @code{point}.
22323 Set the number of samples which are printed on the same column. A
22324 larger value will decrease the frame rate. Must be a positive
22325 integer. This option can be set only if the value for @var{rate}
22326 is not explicitly specified.
22329 Set the (approximate) output frame rate. This is done by setting the
22330 option @var{n}. Default value is "25".
22332 @item split_channels
22333 Set if channels should be drawn separately or overlap. Default value is 0.
22336 Set colors separated by '|' which are going to be used for drawing of each channel.
22339 Set amplitude scale.
22341 Available values are:
22359 Set the draw mode. This is mostly useful to set for high @var{n}.
22361 Available values are:
22364 Scale pixel values for each drawn sample.
22367 Draw every sample directly.
22370 Default value is @code{scale}.
22373 @subsection Examples
22377 Output the input file audio and the corresponding video representation
22380 amovie=a.mp3,asplit[out0],showwaves[out1]
22384 Create a synthetic signal and show it with showwaves, forcing a
22385 frame rate of 30 frames per second:
22387 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
22391 @section showwavespic
22393 Convert input audio to a single video frame, representing the samples waves.
22395 The filter accepts the following options:
22399 Specify the video size for the output. For the syntax of this option, check the
22400 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22401 Default value is @code{600x240}.
22403 @item split_channels
22404 Set if channels should be drawn separately or overlap. Default value is 0.
22407 Set colors separated by '|' which are going to be used for drawing of each channel.
22410 Set amplitude scale.
22412 Available values are:
22430 @subsection Examples
22434 Extract a channel split representation of the wave form of a whole audio track
22435 in a 1024x800 picture using @command{ffmpeg}:
22437 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
22441 @section sidedata, asidedata
22443 Delete frame side data, or select frames based on it.
22445 This filter accepts the following options:
22449 Set mode of operation of the filter.
22451 Can be one of the following:
22455 Select every frame with side data of @code{type}.
22458 Delete side data of @code{type}. If @code{type} is not set, delete all side
22464 Set side data type used with all modes. Must be set for @code{select} mode. For
22465 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
22466 in @file{libavutil/frame.h}. For example, to choose
22467 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
22471 @section spectrumsynth
22473 Sythesize audio from 2 input video spectrums, first input stream represents
22474 magnitude across time and second represents phase across time.
22475 The filter will transform from frequency domain as displayed in videos back
22476 to time domain as presented in audio output.
22478 This filter is primarily created for reversing processed @ref{showspectrum}
22479 filter outputs, but can synthesize sound from other spectrograms too.
22480 But in such case results are going to be poor if the phase data is not
22481 available, because in such cases phase data need to be recreated, usually
22482 it's just recreated from random noise.
22483 For best results use gray only output (@code{channel} color mode in
22484 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
22485 @code{lin} scale for phase video. To produce phase, for 2nd video, use
22486 @code{data} option. Inputs videos should generally use @code{fullframe}
22487 slide mode as that saves resources needed for decoding video.
22489 The filter accepts the following options:
22493 Specify sample rate of output audio, the sample rate of audio from which
22494 spectrum was generated may differ.
22497 Set number of channels represented in input video spectrums.
22500 Set scale which was used when generating magnitude input spectrum.
22501 Can be @code{lin} or @code{log}. Default is @code{log}.
22504 Set slide which was used when generating inputs spectrums.
22505 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
22506 Default is @code{fullframe}.
22509 Set window function used for resynthesis.
22512 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
22513 which means optimal overlap for selected window function will be picked.
22516 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
22517 Default is @code{vertical}.
22520 @subsection Examples
22524 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
22525 then resynthesize videos back to audio with spectrumsynth:
22527 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
22528 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
22529 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
22533 @section split, asplit
22535 Split input into several identical outputs.
22537 @code{asplit} works with audio input, @code{split} with video.
22539 The filter accepts a single parameter which specifies the number of outputs. If
22540 unspecified, it defaults to 2.
22542 @subsection Examples
22546 Create two separate outputs from the same input:
22548 [in] split [out0][out1]
22552 To create 3 or more outputs, you need to specify the number of
22555 [in] asplit=3 [out0][out1][out2]
22559 Create two separate outputs from the same input, one cropped and
22562 [in] split [splitout1][splitout2];
22563 [splitout1] crop=100:100:0:0 [cropout];
22564 [splitout2] pad=200:200:100:100 [padout];
22568 Create 5 copies of the input audio with @command{ffmpeg}:
22570 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
22576 Receive commands sent through a libzmq client, and forward them to
22577 filters in the filtergraph.
22579 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
22580 must be inserted between two video filters, @code{azmq} between two
22581 audio filters. Both are capable to send messages to any filter type.
22583 To enable these filters you need to install the libzmq library and
22584 headers and configure FFmpeg with @code{--enable-libzmq}.
22586 For more information about libzmq see:
22587 @url{http://www.zeromq.org/}
22589 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
22590 receives messages sent through a network interface defined by the
22591 @option{bind_address} (or the abbreviation "@option{b}") option.
22592 Default value of this option is @file{tcp://localhost:5555}. You may
22593 want to alter this value to your needs, but do not forget to escape any
22594 ':' signs (see @ref{filtergraph escaping}).
22596 The received message must be in the form:
22598 @var{TARGET} @var{COMMAND} [@var{ARG}]
22601 @var{TARGET} specifies the target of the command, usually the name of
22602 the filter class or a specific filter instance name. The default
22603 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
22604 but you can override this by using the @samp{filter_name@@id} syntax
22605 (see @ref{Filtergraph syntax}).
22607 @var{COMMAND} specifies the name of the command for the target filter.
22609 @var{ARG} is optional and specifies the optional argument list for the
22610 given @var{COMMAND}.
22612 Upon reception, the message is processed and the corresponding command
22613 is injected into the filtergraph. Depending on the result, the filter
22614 will send a reply to the client, adopting the format:
22616 @var{ERROR_CODE} @var{ERROR_REASON}
22620 @var{MESSAGE} is optional.
22622 @subsection Examples
22624 Look at @file{tools/zmqsend} for an example of a zmq client which can
22625 be used to send commands processed by these filters.
22627 Consider the following filtergraph generated by @command{ffplay}.
22628 In this example the last overlay filter has an instance name. All other
22629 filters will have default instance names.
22632 ffplay -dumpgraph 1 -f lavfi "
22633 color=s=100x100:c=red [l];
22634 color=s=100x100:c=blue [r];
22635 nullsrc=s=200x100, zmq [bg];
22636 [bg][l] overlay [bg+l];
22637 [bg+l][r] overlay@@my=x=100 "
22640 To change the color of the left side of the video, the following
22641 command can be used:
22643 echo Parsed_color_0 c yellow | tools/zmqsend
22646 To change the right side:
22648 echo Parsed_color_1 c pink | tools/zmqsend
22651 To change the position of the right side:
22653 echo overlay@@my x 150 | tools/zmqsend
22657 @c man end MULTIMEDIA FILTERS
22659 @chapter Multimedia Sources
22660 @c man begin MULTIMEDIA SOURCES
22662 Below is a description of the currently available multimedia sources.
22666 This is the same as @ref{movie} source, except it selects an audio
22672 Read audio and/or video stream(s) from a movie container.
22674 It accepts the following parameters:
22678 The name of the resource to read (not necessarily a file; it can also be a
22679 device or a stream accessed through some protocol).
22681 @item format_name, f
22682 Specifies the format assumed for the movie to read, and can be either
22683 the name of a container or an input device. If not specified, the
22684 format is guessed from @var{movie_name} or by probing.
22686 @item seek_point, sp
22687 Specifies the seek point in seconds. The frames will be output
22688 starting from this seek point. The parameter is evaluated with
22689 @code{av_strtod}, so the numerical value may be suffixed by an IS
22690 postfix. The default value is "0".
22693 Specifies the streams to read. Several streams can be specified,
22694 separated by "+". The source will then have as many outputs, in the
22695 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
22696 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
22697 respectively the default (best suited) video and audio stream. Default
22698 is "dv", or "da" if the filter is called as "amovie".
22700 @item stream_index, si
22701 Specifies the index of the video stream to read. If the value is -1,
22702 the most suitable video stream will be automatically selected. The default
22703 value is "-1". Deprecated. If the filter is called "amovie", it will select
22704 audio instead of video.
22707 Specifies how many times to read the stream in sequence.
22708 If the value is 0, the stream will be looped infinitely.
22709 Default value is "1".
22711 Note that when the movie is looped the source timestamps are not
22712 changed, so it will generate non monotonically increasing timestamps.
22714 @item discontinuity
22715 Specifies the time difference between frames above which the point is
22716 considered a timestamp discontinuity which is removed by adjusting the later
22720 It allows overlaying a second video on top of the main input of
22721 a filtergraph, as shown in this graph:
22723 input -----------> deltapts0 --> overlay --> output
22726 movie --> scale--> deltapts1 -------+
22728 @subsection Examples
22732 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
22733 on top of the input labelled "in":
22735 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
22736 [in] setpts=PTS-STARTPTS [main];
22737 [main][over] overlay=16:16 [out]
22741 Read from a video4linux2 device, and overlay it on top of the input
22744 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
22745 [in] setpts=PTS-STARTPTS [main];
22746 [main][over] overlay=16:16 [out]
22750 Read the first video stream and the audio stream with id 0x81 from
22751 dvd.vob; the video is connected to the pad named "video" and the audio is
22752 connected to the pad named "audio":
22754 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
22758 @subsection Commands
22760 Both movie and amovie support the following commands:
22763 Perform seek using "av_seek_frame".
22764 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
22767 @var{stream_index}: If stream_index is -1, a default
22768 stream is selected, and @var{timestamp} is automatically converted
22769 from AV_TIME_BASE units to the stream specific time_base.
22771 @var{timestamp}: Timestamp in AVStream.time_base units
22772 or, if no stream is specified, in AV_TIME_BASE units.
22774 @var{flags}: Flags which select direction and seeking mode.
22778 Get movie duration in AV_TIME_BASE units.
22782 @c man end MULTIMEDIA SOURCES