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 Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Commands
1091 This filter supports the all above options as @ref{commands}.
1093 @subsection Examples
1097 Fade in first 15 seconds of audio:
1099 afade=t=in:ss=0:d=15
1103 Fade out last 25 seconds of a 900 seconds audio:
1105 afade=t=out:st=875:d=25
1110 Denoise audio samples with FFT.
1112 A description of the accepted parameters follows.
1116 Set the noise reduction in dB, allowed range is 0.01 to 97.
1117 Default value is 12 dB.
1120 Set the noise floor in dB, allowed range is -80 to -20.
1121 Default value is -50 dB.
1126 It accepts the following values:
1135 Select shellac noise.
1138 Select custom noise, defined in @code{bn} option.
1140 Default value is white noise.
1144 Set custom band noise for every one of 15 bands.
1145 Bands are separated by ' ' or '|'.
1148 Set the residual floor in dB, allowed range is -80 to -20.
1149 Default value is -38 dB.
1152 Enable noise tracking. By default is disabled.
1153 With this enabled, noise floor is automatically adjusted.
1156 Enable residual tracking. By default is disabled.
1159 Set the output mode.
1161 It accepts the following values:
1164 Pass input unchanged.
1167 Pass noise filtered out.
1172 Default value is @var{o}.
1176 @subsection Commands
1178 This filter supports the following commands:
1180 @item sample_noise, sn
1181 Start or stop measuring noise profile.
1182 Syntax for the command is : "start" or "stop" string.
1183 After measuring noise profile is stopped it will be
1184 automatically applied in filtering.
1186 @item noise_reduction, nr
1187 Change noise reduction. Argument is single float number.
1188 Syntax for the command is : "@var{noise_reduction}"
1190 @item noise_floor, nf
1191 Change noise floor. Argument is single float number.
1192 Syntax for the command is : "@var{noise_floor}"
1194 @item output_mode, om
1195 Change output mode operation.
1196 Syntax for the command is : "i", "o" or "n" string.
1200 Apply arbitrary expressions to samples in frequency domain.
1204 Set frequency domain real expression for each separate channel separated
1205 by '|'. Default is "re".
1206 If the number of input channels is greater than the number of
1207 expressions, the last specified expression is used for the remaining
1211 Set frequency domain imaginary expression for each separate channel
1212 separated by '|'. Default is "im".
1214 Each expression in @var{real} and @var{imag} can contain the following
1215 constants and functions:
1222 current frequency bin number
1225 number of available bins
1228 channel number of the current expression
1237 current real part of frequency bin of current channel
1240 current imaginary part of frequency bin of current channel
1243 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1246 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1250 Set window size. Allowed range is from 16 to 131072.
1251 Default is @code{4096}
1254 Set window function. Default is @code{hann}.
1257 Set window overlap. If set to 1, the recommended overlap for selected
1258 window function will be picked. Default is @code{0.75}.
1261 @subsection Examples
1265 Leave almost only low frequencies in audio:
1267 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1271 Apply robotize effect:
1273 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1277 Apply whisper effect:
1279 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
1286 Apply an arbitrary Finite Impulse Response filter.
1288 This filter is designed for applying long FIR filters,
1289 up to 60 seconds long.
1291 It can be used as component for digital crossover filters,
1292 room equalization, cross talk cancellation, wavefield synthesis,
1293 auralization, ambiophonics, ambisonics and spatialization.
1295 This filter uses the streams higher than first one as FIR coefficients.
1296 If the non-first stream holds a single channel, it will be used
1297 for all input channels in the first stream, otherwise
1298 the number of channels in the non-first stream must be same as
1299 the number of channels in the first stream.
1301 It accepts the following parameters:
1305 Set dry gain. This sets input gain.
1308 Set wet gain. This sets final output gain.
1311 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1314 Enable applying gain measured from power of IR.
1316 Set which approach to use for auto gain measurement.
1320 Do not apply any gain.
1323 select peak gain, very conservative approach. This is default value.
1326 select DC gain, limited application.
1329 select gain to noise approach, this is most popular one.
1333 Set gain to be applied to IR coefficients before filtering.
1334 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1337 Set format of IR stream. Can be @code{mono} or @code{input}.
1338 Default is @code{input}.
1341 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1342 Allowed range is 0.1 to 60 seconds.
1345 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1346 By default it is disabled.
1349 Set for which IR channel to display frequency response. By default is first channel
1350 displayed. This option is used only when @var{response} is enabled.
1353 Set video stream size. This option is used only when @var{response} is enabled.
1356 Set video stream frame rate. This option is used only when @var{response} is enabled.
1359 Set minimal partition size used for convolution. Default is @var{8192}.
1360 Allowed range is from @var{1} to @var{32768}.
1361 Lower values decreases latency at cost of higher CPU usage.
1364 Set maximal partition size used for convolution. Default is @var{8192}.
1365 Allowed range is from @var{8} to @var{32768}.
1366 Lower values may increase CPU usage.
1369 Set number of input impulse responses streams which will be switchable at runtime.
1370 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1373 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1374 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1375 This option can be changed at runtime via @ref{commands}.
1378 @subsection Examples
1382 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1384 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1391 Set output format constraints for the input audio. The framework will
1392 negotiate the most appropriate format to minimize conversions.
1394 It accepts the following parameters:
1397 @item sample_fmts, f
1398 A '|'-separated list of requested sample formats.
1400 @item sample_rates, r
1401 A '|'-separated list of requested sample rates.
1403 @item channel_layouts, cl
1404 A '|'-separated list of requested channel layouts.
1406 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1407 for the required syntax.
1410 If a parameter is omitted, all values are allowed.
1412 Force the output to either unsigned 8-bit or signed 16-bit stereo
1414 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1418 Apply frequency shift to input audio samples.
1420 The filter accepts the following options:
1424 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1425 Default value is 0.0.
1428 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1429 Default value is 1.0.
1432 @subsection Commands
1434 This filter supports the all above options as @ref{commands}.
1438 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1439 processing reduces disturbing noise between useful signals.
1441 Gating is done by detecting the volume below a chosen level @var{threshold}
1442 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1443 floor is set via @var{range}. Because an exact manipulation of the signal
1444 would cause distortion of the waveform the reduction can be levelled over
1445 time. This is done by setting @var{attack} and @var{release}.
1447 @var{attack} determines how long the signal has to fall below the threshold
1448 before any reduction will occur and @var{release} sets the time the signal
1449 has to rise above the threshold to reduce the reduction again.
1450 Shorter signals than the chosen attack time will be left untouched.
1454 Set input level before filtering.
1455 Default is 1. Allowed range is from 0.015625 to 64.
1458 Set the mode of operation. Can be @code{upward} or @code{downward}.
1459 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1460 will be amplified, expanding dynamic range in upward direction.
1461 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1464 Set the level of gain reduction when the signal is below the threshold.
1465 Default is 0.06125. Allowed range is from 0 to 1.
1466 Setting this to 0 disables reduction and then filter behaves like expander.
1469 If a signal rises above this level the gain reduction is released.
1470 Default is 0.125. Allowed range is from 0 to 1.
1473 Set a ratio by which the signal is reduced.
1474 Default is 2. Allowed range is from 1 to 9000.
1477 Amount of milliseconds the signal has to rise above the threshold before gain
1479 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1482 Amount of milliseconds the signal has to fall below the threshold before the
1483 reduction is increased again. Default is 250 milliseconds.
1484 Allowed range is from 0.01 to 9000.
1487 Set amount of amplification of signal after processing.
1488 Default is 1. Allowed range is from 1 to 64.
1491 Curve the sharp knee around the threshold to enter gain reduction more softly.
1492 Default is 2.828427125. Allowed range is from 1 to 8.
1495 Choose if exact signal should be taken for detection or an RMS like one.
1496 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1499 Choose if the average level between all channels or the louder channel affects
1501 Default is @code{average}. Can be @code{average} or @code{maximum}.
1504 @subsection Commands
1506 This filter supports the all above options as @ref{commands}.
1510 Apply an arbitrary Infinite Impulse Response filter.
1512 It accepts the following parameters:
1516 Set B/numerator/zeros/reflection coefficients.
1519 Set A/denominator/poles/ladder coefficients.
1531 Set coefficients format.
1535 lattice-ladder function
1537 analog transfer function
1539 digital transfer function
1541 Z-plane zeros/poles, cartesian (default)
1543 Z-plane zeros/poles, polar radians
1545 Z-plane zeros/poles, polar degrees
1551 Set type of processing.
1563 Set filtering precision.
1567 double-precision floating-point (default)
1569 single-precision floating-point
1577 Normalize filter coefficients, by default is enabled.
1578 Enabling it will normalize magnitude response at DC to 0dB.
1581 How much to use filtered signal in output. Default is 1.
1582 Range is between 0 and 1.
1585 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1586 By default it is disabled.
1589 Set for which IR channel to display frequency response. By default is first channel
1590 displayed. This option is used only when @var{response} is enabled.
1593 Set video stream size. This option is used only when @var{response} is enabled.
1596 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1599 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1600 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1603 Different coefficients and gains can be provided for every channel, in such case
1604 use '|' to separate coefficients or gains. Last provided coefficients will be
1605 used for all remaining channels.
1607 @subsection Examples
1611 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1613 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
1617 Same as above but in @code{zp} format:
1619 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
1623 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1625 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1631 The limiter prevents an input signal from rising over a desired threshold.
1632 This limiter uses lookahead technology to prevent your signal from distorting.
1633 It means that there is a small delay after the signal is processed. Keep in mind
1634 that the delay it produces is the attack time you set.
1636 The filter accepts the following options:
1640 Set input gain. Default is 1.
1643 Set output gain. Default is 1.
1646 Don't let signals above this level pass the limiter. Default is 1.
1649 The limiter will reach its attenuation level in this amount of time in
1650 milliseconds. Default is 5 milliseconds.
1653 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1654 Default is 50 milliseconds.
1657 When gain reduction is always needed ASC takes care of releasing to an
1658 average reduction level rather than reaching a reduction of 0 in the release
1662 Select how much the release time is affected by ASC, 0 means nearly no changes
1663 in release time while 1 produces higher release times.
1666 Auto level output signal. Default is enabled.
1667 This normalizes audio back to 0dB if enabled.
1670 Depending on picked setting it is recommended to upsample input 2x or 4x times
1671 with @ref{aresample} before applying this filter.
1675 Apply a two-pole all-pass filter with central frequency (in Hz)
1676 @var{frequency}, and filter-width @var{width}.
1677 An all-pass filter changes the audio's frequency to phase relationship
1678 without changing its frequency to amplitude relationship.
1680 The filter accepts the following options:
1684 Set frequency in Hz.
1687 Set method to specify band-width of filter.
1702 Specify the band-width of a filter in width_type units.
1705 How much to use filtered signal in output. Default is 1.
1706 Range is between 0 and 1.
1709 Specify which channels to filter, by default all available are filtered.
1712 Normalize biquad coefficients, by default is disabled.
1713 Enabling it will normalize magnitude response at DC to 0dB.
1716 Set the filter order, can be 1 or 2. Default is 2.
1719 Set transform type of IIR filter.
1728 Set precison of filtering.
1731 Pick automatic sample format depending on surround filters.
1733 Always use signed 16-bit.
1735 Always use signed 32-bit.
1737 Always use float 32-bit.
1739 Always use float 64-bit.
1743 @subsection Commands
1745 This filter supports the following commands:
1748 Change allpass frequency.
1749 Syntax for the command is : "@var{frequency}"
1752 Change allpass width_type.
1753 Syntax for the command is : "@var{width_type}"
1756 Change allpass width.
1757 Syntax for the command is : "@var{width}"
1761 Syntax for the command is : "@var{mix}"
1768 The filter accepts the following options:
1772 Set the number of loops. Setting this value to -1 will result in infinite loops.
1776 Set maximal number of samples. Default is 0.
1779 Set first sample of loop. Default is 0.
1785 Merge two or more audio streams into a single multi-channel stream.
1787 The filter accepts the following options:
1792 Set the number of inputs. Default is 2.
1796 If the channel layouts of the inputs are disjoint, and therefore compatible,
1797 the channel layout of the output will be set accordingly and the channels
1798 will be reordered as necessary. If the channel layouts of the inputs are not
1799 disjoint, the output will have all the channels of the first input then all
1800 the channels of the second input, in that order, and the channel layout of
1801 the output will be the default value corresponding to the total number of
1804 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1805 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1806 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1807 first input, b1 is the first channel of the second input).
1809 On the other hand, if both input are in stereo, the output channels will be
1810 in the default order: a1, a2, b1, b2, and the channel layout will be
1811 arbitrarily set to 4.0, which may or may not be the expected value.
1813 All inputs must have the same sample rate, and format.
1815 If inputs do not have the same duration, the output will stop with the
1818 @subsection Examples
1822 Merge two mono files into a stereo stream:
1824 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1828 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1830 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
1836 Mixes multiple audio inputs into a single output.
1838 Note that this filter only supports float samples (the @var{amerge}
1839 and @var{pan} audio filters support many formats). If the @var{amix}
1840 input has integer samples then @ref{aresample} will be automatically
1841 inserted to perform the conversion to float samples.
1845 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1847 will mix 3 input audio streams to a single output with the same duration as the
1848 first input and a dropout transition time of 3 seconds.
1850 It accepts the following parameters:
1854 The number of inputs. If unspecified, it defaults to 2.
1857 How to determine the end-of-stream.
1861 The duration of the longest input. (default)
1864 The duration of the shortest input.
1867 The duration of the first input.
1871 @item dropout_transition
1872 The transition time, in seconds, for volume renormalization when an input
1873 stream ends. The default value is 2 seconds.
1876 Specify weight of each input audio stream as sequence.
1877 Each weight is separated by space. By default all inputs have same weight.
1880 @subsection Commands
1882 This filter supports the following commands:
1885 Syntax is same as option with same name.
1890 Multiply first audio stream with second audio stream and store result
1891 in output audio stream. Multiplication is done by multiplying each
1892 sample from first stream with sample at same position from second stream.
1894 With this element-wise multiplication one can create amplitude fades and
1895 amplitude modulations.
1897 @section anequalizer
1899 High-order parametric multiband equalizer for each channel.
1901 It accepts the following parameters:
1905 This option string is in format:
1906 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1907 Each equalizer band is separated by '|'.
1911 Set channel number to which equalization will be applied.
1912 If input doesn't have that channel the entry is ignored.
1915 Set central frequency for band.
1916 If input doesn't have that frequency the entry is ignored.
1919 Set band width in Hertz.
1922 Set band gain in dB.
1925 Set filter type for band, optional, can be:
1929 Butterworth, this is default.
1940 With this option activated frequency response of anequalizer is displayed
1944 Set video stream size. Only useful if curves option is activated.
1947 Set max gain that will be displayed. Only useful if curves option is activated.
1948 Setting this to a reasonable value makes it possible to display gain which is derived from
1949 neighbour bands which are too close to each other and thus produce higher gain
1950 when both are activated.
1953 Set frequency scale used to draw frequency response in video output.
1954 Can be linear or logarithmic. Default is logarithmic.
1957 Set color for each channel curve which is going to be displayed in video stream.
1958 This is list of color names separated by space or by '|'.
1959 Unrecognised or missing colors will be replaced by white color.
1962 @subsection Examples
1966 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1967 for first 2 channels using Chebyshev type 1 filter:
1969 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1973 @subsection Commands
1975 This filter supports the following commands:
1978 Alter existing filter parameters.
1979 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1981 @var{fN} is existing filter number, starting from 0, if no such filter is available
1983 @var{freq} set new frequency parameter.
1984 @var{width} set new width parameter in Hertz.
1985 @var{gain} set new gain parameter in dB.
1987 Full filter invocation with asendcmd may look like this:
1988 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1993 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1995 Each sample is adjusted by looking for other samples with similar contexts. This
1996 context similarity is defined by comparing their surrounding patches of size
1997 @option{p}. Patches are searched in an area of @option{r} around the sample.
1999 The filter accepts the following options:
2003 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2006 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2007 Default value is 2 milliseconds.
2010 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2011 Default value is 6 milliseconds.
2014 Set the output mode.
2016 It accepts the following values:
2019 Pass input unchanged.
2022 Pass noise filtered out.
2027 Default value is @var{o}.
2031 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2034 @subsection Commands
2036 This filter supports the all above options as @ref{commands}.
2039 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2041 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2042 relate to producing the least mean square of the error signal (difference between the desired,
2043 2nd input audio stream and the actual signal, the 1st input audio stream).
2045 A description of the accepted options follows.
2058 Set the filter leakage.
2061 It accepts the following values:
2070 Pass filtered samples.
2073 Pass difference between desired and filtered samples.
2075 Default value is @var{o}.
2079 @subsection Examples
2083 One of many usages of this filter is noise reduction, input audio is filtered
2084 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2086 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2090 @subsection Commands
2092 This filter supports the same commands as options, excluding option @code{order}.
2096 Pass the audio source unchanged to the output.
2100 Pad the end of an audio stream with silence.
2102 This can be used together with @command{ffmpeg} @option{-shortest} to
2103 extend audio streams to the same length as the video stream.
2105 A description of the accepted options follows.
2109 Set silence packet size. Default value is 4096.
2112 Set the number of samples of silence to add to the end. After the
2113 value is reached, the stream is terminated. This option is mutually
2114 exclusive with @option{whole_len}.
2117 Set the minimum total number of samples in the output audio stream. If
2118 the value is longer than the input audio length, silence is added to
2119 the end, until the value is reached. This option is mutually exclusive
2120 with @option{pad_len}.
2123 Specify the duration of samples of silence to add. See
2124 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2125 for the accepted syntax. Used only if set to non-zero value.
2128 Specify the minimum total duration in the output audio stream. See
2129 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2130 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2131 the input audio length, silence is added to the end, until the value is reached.
2132 This option is mutually exclusive with @option{pad_dur}
2135 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2136 nor @option{whole_dur} option is set, the filter will add silence to the end of
2137 the input stream indefinitely.
2139 @subsection Examples
2143 Add 1024 samples of silence to the end of the input:
2149 Make sure the audio output will contain at least 10000 samples, pad
2150 the input with silence if required:
2152 apad=whole_len=10000
2156 Use @command{ffmpeg} to pad the audio input with silence, so that the
2157 video stream will always result the shortest and will be converted
2158 until the end in the output file when using the @option{shortest}
2161 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2166 Add a phasing effect to the input audio.
2168 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2169 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2171 A description of the accepted parameters follows.
2175 Set input gain. Default is 0.4.
2178 Set output gain. Default is 0.74
2181 Set delay in milliseconds. Default is 3.0.
2184 Set decay. Default is 0.4.
2187 Set modulation speed in Hz. Default is 0.5.
2190 Set modulation type. Default is triangular.
2192 It accepts the following values:
2199 @section aphaseshift
2200 Apply phase shift to input audio samples.
2202 The filter accepts the following options:
2206 Specify phase shift. Allowed range is from -1.0 to 1.0.
2207 Default value is 0.0.
2210 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2211 Default value is 1.0.
2214 @subsection Commands
2216 This filter supports the all above options as @ref{commands}.
2220 Audio pulsator is something between an autopanner and a tremolo.
2221 But it can produce funny stereo effects as well. Pulsator changes the volume
2222 of the left and right channel based on a LFO (low frequency oscillator) with
2223 different waveforms and shifted phases.
2224 This filter have the ability to define an offset between left and right
2225 channel. An offset of 0 means that both LFO shapes match each other.
2226 The left and right channel are altered equally - a conventional tremolo.
2227 An offset of 50% means that the shape of the right channel is exactly shifted
2228 in phase (or moved backwards about half of the frequency) - pulsator acts as
2229 an autopanner. At 1 both curves match again. Every setting in between moves the
2230 phase shift gapless between all stages and produces some "bypassing" sounds with
2231 sine and triangle waveforms. The more you set the offset near 1 (starting from
2232 the 0.5) the faster the signal passes from the left to the right speaker.
2234 The filter accepts the following options:
2238 Set input gain. By default it is 1. Range is [0.015625 - 64].
2241 Set output gain. By default it is 1. Range is [0.015625 - 64].
2244 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2245 sawup or sawdown. Default is sine.
2248 Set modulation. Define how much of original signal is affected by the LFO.
2251 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2254 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2257 Set pulse width. Default is 1. Allowed range is [0 - 2].
2260 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2263 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2267 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2271 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2272 if timing is set to hz.
2278 Resample the input audio to the specified parameters, using the
2279 libswresample library. If none are specified then the filter will
2280 automatically convert between its input and output.
2282 This filter is also able to stretch/squeeze the audio data to make it match
2283 the timestamps or to inject silence / cut out audio to make it match the
2284 timestamps, do a combination of both or do neither.
2286 The filter accepts the syntax
2287 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2288 expresses a sample rate and @var{resampler_options} is a list of
2289 @var{key}=@var{value} pairs, separated by ":". See the
2290 @ref{Resampler Options,,"Resampler Options" section in the
2291 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2292 for the complete list of supported options.
2294 @subsection Examples
2298 Resample the input audio to 44100Hz:
2304 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2305 samples per second compensation:
2307 aresample=async=1000
2313 Reverse an audio clip.
2315 Warning: This filter requires memory to buffer the entire clip, so trimming
2318 @subsection Examples
2322 Take the first 5 seconds of a clip, and reverse it.
2324 atrim=end=5,areverse
2330 Reduce noise from speech using Recurrent Neural Networks.
2332 This filter accepts the following options:
2336 Set train model file to load. This option is always required.
2339 Set how much to mix filtered samples into final output.
2340 Allowed range is from -1 to 1. Default value is 1.
2341 Negative values are special, they set how much to keep filtered noise
2342 in the final filter output. Set this option to -1 to hear actual
2343 noise removed from input signal.
2346 @section asetnsamples
2348 Set the number of samples per each output audio frame.
2350 The last output packet may contain a different number of samples, as
2351 the filter will flush all the remaining samples when the input audio
2354 The filter accepts the following options:
2358 @item nb_out_samples, n
2359 Set the number of frames per each output audio frame. The number is
2360 intended as the number of samples @emph{per each channel}.
2361 Default value is 1024.
2364 If set to 1, the filter will pad the last audio frame with zeroes, so
2365 that the last frame will contain the same number of samples as the
2366 previous ones. Default value is 1.
2369 For example, to set the number of per-frame samples to 1234 and
2370 disable padding for the last frame, use:
2372 asetnsamples=n=1234:p=0
2377 Set the sample rate without altering the PCM data.
2378 This will result in a change of speed and pitch.
2380 The filter accepts the following options:
2383 @item sample_rate, r
2384 Set the output sample rate. Default is 44100 Hz.
2389 Show a line containing various information for each input audio frame.
2390 The input audio is not modified.
2392 The shown line contains a sequence of key/value pairs of the form
2393 @var{key}:@var{value}.
2395 The following values are shown in the output:
2399 The (sequential) number of the input frame, starting from 0.
2402 The presentation timestamp of the input frame, in time base units; the time base
2403 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2406 The presentation timestamp of the input frame in seconds.
2409 position of the frame in the input stream, -1 if this information in
2410 unavailable and/or meaningless (for example in case of synthetic audio)
2419 The sample rate for the audio frame.
2422 The number of samples (per channel) in the frame.
2425 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2426 audio, the data is treated as if all the planes were concatenated.
2428 @item plane_checksums
2429 A list of Adler-32 checksums for each data plane.
2433 Apply audio soft clipping.
2435 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2436 along a smooth curve, rather than the abrupt shape of hard-clipping.
2438 This filter accepts the following options:
2442 Set type of soft-clipping.
2444 It accepts the following values:
2458 Set threshold from where to start clipping. Default value is 0dB or 1.
2461 Set gain applied to output. Default value is 0dB or 1.
2464 Set additional parameter which controls sigmoid function.
2467 Set oversampling factor.
2470 @subsection Commands
2472 This filter supports the all above options as @ref{commands}.
2475 Automatic Speech Recognition
2477 This filter uses PocketSphinx for speech recognition. To enable
2478 compilation of this filter, you need to configure FFmpeg with
2479 @code{--enable-pocketsphinx}.
2481 It accepts the following options:
2485 Set sampling rate of input audio. Defaults is @code{16000}.
2486 This need to match speech models, otherwise one will get poor results.
2489 Set dictionary containing acoustic model files.
2492 Set pronunciation dictionary.
2495 Set language model file.
2498 Set language model set.
2501 Set which language model to use.
2504 Set output for log messages.
2507 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2512 Display time domain statistical information about the audio channels.
2513 Statistics are calculated and displayed for each audio channel and,
2514 where applicable, an overall figure is also given.
2516 It accepts the following option:
2519 Short window length in seconds, used for peak and trough RMS measurement.
2520 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2524 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2525 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2528 Available keys for each channel are:
2574 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2575 this @code{lavfi.astats.Overall.Peak_count}.
2577 For description what each key means read below.
2580 Set number of frame after which stats are going to be recalculated.
2581 Default is disabled.
2583 @item measure_perchannel
2584 Select the entries which need to be measured per channel. The metadata keys can
2585 be used as flags, default is @option{all} which measures everything.
2586 @option{none} disables all per channel measurement.
2588 @item measure_overall
2589 Select the entries which need to be measured overall. The metadata keys can
2590 be used as flags, default is @option{all} which measures everything.
2591 @option{none} disables all overall measurement.
2595 A description of each shown parameter follows:
2599 Mean amplitude displacement from zero.
2602 Minimal sample level.
2605 Maximal sample level.
2607 @item Min difference
2608 Minimal difference between two consecutive samples.
2610 @item Max difference
2611 Maximal difference between two consecutive samples.
2613 @item Mean difference
2614 Mean difference between two consecutive samples.
2615 The average of each difference between two consecutive samples.
2617 @item RMS difference
2618 Root Mean Square difference between two consecutive samples.
2622 Standard peak and RMS level measured in dBFS.
2626 Peak and trough values for RMS level measured over a short window.
2629 Standard ratio of peak to RMS level (note: not in dB).
2632 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2633 (i.e. either @var{Min level} or @var{Max level}).
2636 Number of occasions (not the number of samples) that the signal attained either
2637 @var{Min level} or @var{Max level}.
2639 @item Noise floor dB
2640 Minimum local peak measured in dBFS over a short window.
2642 @item Noise floor count
2643 Number of occasions (not the number of samples) that the signal attained
2647 Overall bit depth of audio. Number of bits used for each sample.
2650 Measured dynamic range of audio in dB.
2652 @item Zero crossings
2653 Number of points where the waveform crosses the zero level axis.
2655 @item Zero crossings rate
2656 Rate of Zero crossings and number of audio samples.
2660 Boost subwoofer frequencies.
2662 The filter accepts the following options:
2666 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2667 Default value is 0.7.
2670 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2671 Default value is 0.7.
2674 Set delay line decay gain value. Allowed range is from 0 to 1.
2675 Default value is 0.7.
2678 Set delay line feedback gain value. Allowed range is from 0 to 1.
2679 Default value is 0.9.
2682 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2683 Default value is 100.
2686 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2687 Default value is 0.5.
2690 Set delay. Allowed range is from 1 to 100.
2691 Default value is 20.
2694 @subsection Commands
2696 This filter supports the all above options as @ref{commands}.
2699 Cut subwoofer frequencies.
2701 This filter allows to set custom, steeper
2702 roll off than highpass filter, and thus is able to more attenuate
2703 frequency content in stop-band.
2705 The filter accepts the following options:
2709 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2710 Default value is 20.
2713 Set filter order. Available values are from 3 to 20.
2714 Default value is 10.
2717 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2720 @subsection Commands
2722 This filter supports the all above options as @ref{commands}.
2725 Cut super frequencies.
2727 The filter accepts the following options:
2731 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2732 Default value is 20000.
2735 Set filter order. Available values are from 3 to 20.
2736 Default value is 10.
2739 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2742 @subsection Commands
2744 This filter supports the all above options as @ref{commands}.
2747 Apply high order Butterworth band-pass filter.
2749 The filter accepts the following options:
2753 Set center frequency in Hertz. Allowed range is 2 to 999999.
2754 Default value is 1000.
2757 Set filter order. Available values are from 4 to 20.
2761 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2764 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2767 @subsection Commands
2769 This filter supports the all above options as @ref{commands}.
2772 Apply high order Butterworth band-stop filter.
2774 The filter accepts the following options:
2778 Set center frequency in Hertz. Allowed range is 2 to 999999.
2779 Default value is 1000.
2782 Set filter order. Available values are from 4 to 20.
2786 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2789 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2792 @subsection Commands
2794 This filter supports the all above options as @ref{commands}.
2800 The filter accepts exactly one parameter, the audio tempo. If not
2801 specified then the filter will assume nominal 1.0 tempo. Tempo must
2802 be in the [0.5, 100.0] range.
2804 Note that tempo greater than 2 will skip some samples rather than
2805 blend them in. If for any reason this is a concern it is always
2806 possible to daisy-chain several instances of atempo to achieve the
2807 desired product tempo.
2809 @subsection Examples
2813 Slow down audio to 80% tempo:
2819 To speed up audio to 300% tempo:
2825 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2827 atempo=sqrt(3),atempo=sqrt(3)
2831 @subsection Commands
2833 This filter supports the following commands:
2836 Change filter tempo scale factor.
2837 Syntax for the command is : "@var{tempo}"
2842 Trim the input so that the output contains one continuous subpart of the input.
2844 It accepts the following parameters:
2847 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2848 sample with the timestamp @var{start} will be the first sample in the output.
2851 Specify time of the first audio sample that will be dropped, i.e. the
2852 audio sample immediately preceding the one with the timestamp @var{end} will be
2853 the last sample in the output.
2856 Same as @var{start}, except this option sets the start timestamp in samples
2860 Same as @var{end}, except this option sets the end timestamp in samples instead
2864 The maximum duration of the output in seconds.
2867 The number of the first sample that should be output.
2870 The number of the first sample that should be dropped.
2873 @option{start}, @option{end}, and @option{duration} are expressed as time
2874 duration specifications; see
2875 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2877 Note that the first two sets of the start/end options and the @option{duration}
2878 option look at the frame timestamp, while the _sample options simply count the
2879 samples that pass through the filter. So start/end_pts and start/end_sample will
2880 give different results when the timestamps are wrong, inexact or do not start at
2881 zero. Also note that this filter does not modify the timestamps. If you wish
2882 to have the output timestamps start at zero, insert the asetpts filter after the
2885 If multiple start or end options are set, this filter tries to be greedy and
2886 keep all samples that match at least one of the specified constraints. To keep
2887 only the part that matches all the constraints at once, chain multiple atrim
2890 The defaults are such that all the input is kept. So it is possible to set e.g.
2891 just the end values to keep everything before the specified time.
2896 Drop everything except the second minute of input:
2898 ffmpeg -i INPUT -af atrim=60:120
2902 Keep only the first 1000 samples:
2904 ffmpeg -i INPUT -af atrim=end_sample=1000
2909 @section axcorrelate
2910 Calculate normalized cross-correlation between two input audio streams.
2912 Resulted samples are always between -1 and 1 inclusive.
2913 If result is 1 it means two input samples are highly correlated in that selected segment.
2914 Result 0 means they are not correlated at all.
2915 If result is -1 it means two input samples are out of phase, which means they cancel each
2918 The filter accepts the following options:
2922 Set size of segment over which cross-correlation is calculated.
2923 Default is 256. Allowed range is from 2 to 131072.
2926 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2927 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2928 are always zero and thus need much less calculations to make.
2929 This is generally not true, but is valid for typical audio streams.
2932 @subsection Examples
2936 Calculate correlation between channels in stereo audio stream:
2938 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2944 Apply a two-pole Butterworth band-pass filter with central
2945 frequency @var{frequency}, and (3dB-point) band-width width.
2946 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2947 instead of the default: constant 0dB peak gain.
2948 The filter roll off at 6dB per octave (20dB per decade).
2950 The filter accepts the following options:
2954 Set the filter's central frequency. Default is @code{3000}.
2957 Constant skirt gain if set to 1. Defaults to 0.
2960 Set method to specify band-width of filter.
2975 Specify the band-width of a filter in width_type units.
2978 How much to use filtered signal in output. Default is 1.
2979 Range is between 0 and 1.
2982 Specify which channels to filter, by default all available are filtered.
2985 Normalize biquad coefficients, by default is disabled.
2986 Enabling it will normalize magnitude response at DC to 0dB.
2989 Set transform type of IIR filter.
2998 Set precison of filtering.
3001 Pick automatic sample format depending on surround filters.
3003 Always use signed 16-bit.
3005 Always use signed 32-bit.
3007 Always use float 32-bit.
3009 Always use float 64-bit.
3013 @subsection Commands
3015 This filter supports the following commands:
3018 Change bandpass frequency.
3019 Syntax for the command is : "@var{frequency}"
3022 Change bandpass width_type.
3023 Syntax for the command is : "@var{width_type}"
3026 Change bandpass width.
3027 Syntax for the command is : "@var{width}"
3030 Change bandpass mix.
3031 Syntax for the command is : "@var{mix}"
3036 Apply a two-pole Butterworth band-reject filter with central
3037 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
3038 The filter roll off at 6dB per octave (20dB per decade).
3040 The filter accepts the following options:
3044 Set the filter's central frequency. Default is @code{3000}.
3047 Set method to specify band-width of filter.
3062 Specify the band-width of a filter in width_type units.
3065 How much to use filtered signal in output. Default is 1.
3066 Range is between 0 and 1.
3069 Specify which channels to filter, by default all available are filtered.
3072 Normalize biquad coefficients, by default is disabled.
3073 Enabling it will normalize magnitude response at DC to 0dB.
3076 Set transform type of IIR filter.
3085 Set precison of filtering.
3088 Pick automatic sample format depending on surround filters.
3090 Always use signed 16-bit.
3092 Always use signed 32-bit.
3094 Always use float 32-bit.
3096 Always use float 64-bit.
3100 @subsection Commands
3102 This filter supports the following commands:
3105 Change bandreject frequency.
3106 Syntax for the command is : "@var{frequency}"
3109 Change bandreject width_type.
3110 Syntax for the command is : "@var{width_type}"
3113 Change bandreject width.
3114 Syntax for the command is : "@var{width}"
3117 Change bandreject mix.
3118 Syntax for the command is : "@var{mix}"
3121 @section bass, lowshelf
3123 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3124 shelving filter with a response similar to that of a standard
3125 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3127 The filter accepts the following options:
3131 Give the gain at 0 Hz. Its useful range is about -20
3132 (for a large cut) to +20 (for a large boost).
3133 Beware of clipping when using a positive gain.
3136 Set the filter's central frequency and so can be used
3137 to extend or reduce the frequency range to be boosted or cut.
3138 The default value is @code{100} Hz.
3141 Set method to specify band-width of filter.
3156 Determine how steep is the filter's shelf transition.
3159 Set number of poles. Default is 2.
3162 How much to use filtered signal in output. Default is 1.
3163 Range is between 0 and 1.
3166 Specify which channels to filter, by default all available are filtered.
3169 Normalize biquad coefficients, by default is disabled.
3170 Enabling it will normalize magnitude response at DC to 0dB.
3173 Set transform type of IIR filter.
3182 Set precison of filtering.
3185 Pick automatic sample format depending on surround filters.
3187 Always use signed 16-bit.
3189 Always use signed 32-bit.
3191 Always use float 32-bit.
3193 Always use float 64-bit.
3197 @subsection Commands
3199 This filter supports the following commands:
3202 Change bass frequency.
3203 Syntax for the command is : "@var{frequency}"
3206 Change bass width_type.
3207 Syntax for the command is : "@var{width_type}"
3211 Syntax for the command is : "@var{width}"
3215 Syntax for the command is : "@var{gain}"
3219 Syntax for the command is : "@var{mix}"
3224 Apply a biquad IIR filter with the given coefficients.
3225 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3226 are the numerator and denominator coefficients respectively.
3227 and @var{channels}, @var{c} specify which channels to filter, by default all
3228 available are filtered.
3230 @subsection Commands
3232 This filter supports the following commands:
3240 Change biquad parameter.
3241 Syntax for the command is : "@var{value}"
3244 How much to use filtered signal in output. Default is 1.
3245 Range is between 0 and 1.
3248 Specify which channels to filter, by default all available are filtered.
3251 Normalize biquad coefficients, by default is disabled.
3252 Enabling it will normalize magnitude response at DC to 0dB.
3255 Set transform type of IIR filter.
3264 Set precison of filtering.
3267 Pick automatic sample format depending on surround filters.
3269 Always use signed 16-bit.
3271 Always use signed 32-bit.
3273 Always use float 32-bit.
3275 Always use float 64-bit.
3280 Bauer stereo to binaural transformation, which improves headphone listening of
3281 stereo audio records.
3283 To enable compilation of this filter you need to configure FFmpeg with
3284 @code{--enable-libbs2b}.
3286 It accepts the following parameters:
3290 Pre-defined crossfeed level.
3294 Default level (fcut=700, feed=50).
3297 Chu Moy circuit (fcut=700, feed=60).
3300 Jan Meier circuit (fcut=650, feed=95).
3305 Cut frequency (in Hz).
3314 Remap input channels to new locations.
3316 It accepts the following parameters:
3319 Map channels from input to output. The argument is a '|'-separated list of
3320 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3321 @var{in_channel} form. @var{in_channel} can be either the name of the input
3322 channel (e.g. FL for front left) or its index in the input channel layout.
3323 @var{out_channel} is the name of the output channel or its index in the output
3324 channel layout. If @var{out_channel} is not given then it is implicitly an
3325 index, starting with zero and increasing by one for each mapping.
3327 @item channel_layout
3328 The channel layout of the output stream.
3331 If no mapping is present, the filter will implicitly map input channels to
3332 output channels, preserving indices.
3334 @subsection Examples
3338 For example, assuming a 5.1+downmix input MOV file,
3340 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3342 will create an output WAV file tagged as stereo from the downmix channels of
3346 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3348 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3352 @section channelsplit
3354 Split each channel from an input audio stream into a separate output stream.
3356 It accepts the following parameters:
3358 @item channel_layout
3359 The channel layout of the input stream. The default is "stereo".
3361 A channel layout describing the channels to be extracted as separate output streams
3362 or "all" to extract each input channel as a separate stream. The default is "all".
3364 Choosing channels not present in channel layout in the input will result in an error.
3367 @subsection Examples
3371 For example, assuming a stereo input MP3 file,
3373 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3375 will create an output Matroska file with two audio streams, one containing only
3376 the left channel and the other the right channel.
3379 Split a 5.1 WAV file into per-channel files:
3381 ffmpeg -i in.wav -filter_complex
3382 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3383 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3384 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3389 Extract only LFE from a 5.1 WAV file:
3391 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3392 -map '[LFE]' lfe.wav
3397 Add a chorus effect to the audio.
3399 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3401 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3402 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3403 The modulation depth defines the range the modulated delay is played before or after
3404 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3405 sound tuned around the original one, like in a chorus where some vocals are slightly
3408 It accepts the following parameters:
3411 Set input gain. Default is 0.4.
3414 Set output gain. Default is 0.4.
3417 Set delays. A typical delay is around 40ms to 60ms.
3429 @subsection Examples
3435 chorus=0.7:0.9:55:0.4:0.25:2
3441 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3445 Fuller sounding chorus with three delays:
3447 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
3452 Compress or expand the audio's dynamic range.
3454 It accepts the following parameters:
3460 A list of times in seconds for each channel over which the instantaneous level
3461 of the input signal is averaged to determine its volume. @var{attacks} refers to
3462 increase of volume and @var{decays} refers to decrease of volume. For most
3463 situations, the attack time (response to the audio getting louder) should be
3464 shorter than the decay time, because the human ear is more sensitive to sudden
3465 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3466 a typical value for decay is 0.8 seconds.
3467 If specified number of attacks & decays is lower than number of channels, the last
3468 set attack/decay will be used for all remaining channels.
3471 A list of points for the transfer function, specified in dB relative to the
3472 maximum possible signal amplitude. Each key points list must be defined using
3473 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3474 @code{x0/y0 x1/y1 x2/y2 ....}
3476 The input values must be in strictly increasing order but the transfer function
3477 does not have to be monotonically rising. The point @code{0/0} is assumed but
3478 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3479 function are @code{-70/-70|-60/-20|1/0}.
3482 Set the curve radius in dB for all joints. It defaults to 0.01.
3485 Set the additional gain in dB to be applied at all points on the transfer
3486 function. This allows for easy adjustment of the overall gain.
3490 Set an initial volume, in dB, to be assumed for each channel when filtering
3491 starts. This permits the user to supply a nominal level initially, so that, for
3492 example, a very large gain is not applied to initial signal levels before the
3493 companding has begun to operate. A typical value for audio which is initially
3494 quiet is -90 dB. It defaults to 0.
3497 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3498 delayed before being fed to the volume adjuster. Specifying a delay
3499 approximately equal to the attack/decay times allows the filter to effectively
3500 operate in predictive rather than reactive mode. It defaults to 0.
3504 @subsection Examples
3508 Make music with both quiet and loud passages suitable for listening to in a
3511 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3514 Another example for audio with whisper and explosion parts:
3516 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3520 A noise gate for when the noise is at a lower level than the signal:
3522 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3526 Here is another noise gate, this time for when the noise is at a higher level
3527 than the signal (making it, in some ways, similar to squelch):
3529 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3533 2:1 compression starting at -6dB:
3535 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3539 2:1 compression starting at -9dB:
3541 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3545 2:1 compression starting at -12dB:
3547 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3551 2:1 compression starting at -18dB:
3553 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3557 3:1 compression starting at -15dB:
3559 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3565 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3571 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
3575 Hard limiter at -6dB:
3577 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3581 Hard limiter at -12dB:
3583 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3587 Hard noise gate at -35 dB:
3589 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3595 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3599 @section compensationdelay
3601 Compensation Delay Line is a metric based delay to compensate differing
3602 positions of microphones or speakers.
3604 For example, you have recorded guitar with two microphones placed in
3605 different locations. Because the front of sound wave has fixed speed in
3606 normal conditions, the phasing of microphones can vary and depends on
3607 their location and interposition. The best sound mix can be achieved when
3608 these microphones are in phase (synchronized). Note that a distance of
3609 ~30 cm between microphones makes one microphone capture the signal in
3610 antiphase to the other microphone. That makes the final mix sound moody.
3611 This filter helps to solve phasing problems by adding different delays
3612 to each microphone track and make them synchronized.
3614 The best result can be reached when you take one track as base and
3615 synchronize other tracks one by one with it.
3616 Remember that synchronization/delay tolerance depends on sample rate, too.
3617 Higher sample rates will give more tolerance.
3619 The filter accepts the following parameters:
3623 Set millimeters distance. This is compensation distance for fine tuning.
3627 Set cm distance. This is compensation distance for tightening distance setup.
3631 Set meters distance. This is compensation distance for hard distance setup.
3635 Set dry amount. Amount of unprocessed (dry) signal.
3639 Set wet amount. Amount of processed (wet) signal.
3643 Set temperature in degrees Celsius. This is the temperature of the environment.
3648 Apply headphone crossfeed filter.
3650 Crossfeed is the process of blending the left and right channels of stereo
3652 It is mainly used to reduce extreme stereo separation of low frequencies.
3654 The intent is to produce more speaker like sound to the listener.
3656 The filter accepts the following options:
3660 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3661 This sets gain of low shelf filter for side part of stereo image.
3662 Default is -6dB. Max allowed is -30db when strength is set to 1.
3665 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3666 This sets cut off frequency of low shelf filter. Default is cut off near
3667 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3670 Set curve slope of low shelf filter. Default is 0.5.
3671 Allowed range is from 0.01 to 1.
3674 Set input gain. Default is 0.9.
3677 Set output gain. Default is 1.
3680 @subsection Commands
3682 This filter supports the all above options as @ref{commands}.
3684 @section crystalizer
3685 Simple algorithm for audio noise sharpening.
3687 This filter linearly increases differences betweeen each audio sample.
3689 The filter accepts the following options:
3693 Sets the intensity of effect (default: 2.0). Must be in range between -10.0 to 0
3694 (unchanged sound) to 10.0 (maximum effect).
3695 To inverse filtering use negative value.
3698 Enable clipping. By default is enabled.
3701 @subsection Commands
3703 This filter supports the all above options as @ref{commands}.
3706 Apply a DC shift to the audio.
3708 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3709 in the recording chain) from the audio. The effect of a DC offset is reduced
3710 headroom and hence volume. The @ref{astats} filter can be used to determine if
3711 a signal has a DC offset.
3715 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3719 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3720 used to prevent clipping.
3725 Apply de-essing to the audio samples.
3729 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3733 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3737 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3741 Set the output mode.
3743 It accepts the following values:
3746 Pass input unchanged.
3749 Pass ess filtered out.
3754 Default value is @var{o}.
3760 Measure audio dynamic range.
3762 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3763 is found in transition material. And anything less that 8 have very poor dynamics
3764 and is very compressed.
3766 The filter accepts the following options:
3770 Set window length in seconds used to split audio into segments of equal length.
3771 Default is 3 seconds.
3775 Dynamic Audio Normalizer.
3777 This filter applies a certain amount of gain to the input audio in order
3778 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3779 contrast to more "simple" normalization algorithms, the Dynamic Audio
3780 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3781 This allows for applying extra gain to the "quiet" sections of the audio
3782 while avoiding distortions or clipping the "loud" sections. In other words:
3783 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3784 sections, in the sense that the volume of each section is brought to the
3785 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3786 this goal *without* applying "dynamic range compressing". It will retain 100%
3787 of the dynamic range *within* each section of the audio file.
3791 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3792 Default is 500 milliseconds.
3793 The Dynamic Audio Normalizer processes the input audio in small chunks,
3794 referred to as frames. This is required, because a peak magnitude has no
3795 meaning for just a single sample value. Instead, we need to determine the
3796 peak magnitude for a contiguous sequence of sample values. While a "standard"
3797 normalizer would simply use the peak magnitude of the complete file, the
3798 Dynamic Audio Normalizer determines the peak magnitude individually for each
3799 frame. The length of a frame is specified in milliseconds. By default, the
3800 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3801 been found to give good results with most files.
3802 Note that the exact frame length, in number of samples, will be determined
3803 automatically, based on the sampling rate of the individual input audio file.
3806 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3807 number. Default is 31.
3808 Probably the most important parameter of the Dynamic Audio Normalizer is the
3809 @code{window size} of the Gaussian smoothing filter. The filter's window size
3810 is specified in frames, centered around the current frame. For the sake of
3811 simplicity, this must be an odd number. Consequently, the default value of 31
3812 takes into account the current frame, as well as the 15 preceding frames and
3813 the 15 subsequent frames. Using a larger window results in a stronger
3814 smoothing effect and thus in less gain variation, i.e. slower gain
3815 adaptation. Conversely, using a smaller window results in a weaker smoothing
3816 effect and thus in more gain variation, i.e. faster gain adaptation.
3817 In other words, the more you increase this value, the more the Dynamic Audio
3818 Normalizer will behave like a "traditional" normalization filter. On the
3819 contrary, the more you decrease this value, the more the Dynamic Audio
3820 Normalizer will behave like a dynamic range compressor.
3823 Set the target peak value. This specifies the highest permissible magnitude
3824 level for the normalized audio input. This filter will try to approach the
3825 target peak magnitude as closely as possible, but at the same time it also
3826 makes sure that the normalized signal will never exceed the peak magnitude.
3827 A frame's maximum local gain factor is imposed directly by the target peak
3828 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3829 It is not recommended to go above this value.
3832 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3833 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3834 factor for each input frame, i.e. the maximum gain factor that does not
3835 result in clipping or distortion. The maximum gain factor is determined by
3836 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3837 additionally bounds the frame's maximum gain factor by a predetermined
3838 (global) maximum gain factor. This is done in order to avoid excessive gain
3839 factors in "silent" or almost silent frames. By default, the maximum gain
3840 factor is 10.0, For most inputs the default value should be sufficient and
3841 it usually is not recommended to increase this value. Though, for input
3842 with an extremely low overall volume level, it may be necessary to allow even
3843 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3844 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3845 Instead, a "sigmoid" threshold function will be applied. This way, the
3846 gain factors will smoothly approach the threshold value, but never exceed that
3850 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3851 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3852 This means that the maximum local gain factor for each frame is defined
3853 (only) by the frame's highest magnitude sample. This way, the samples can
3854 be amplified as much as possible without exceeding the maximum signal
3855 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3856 Normalizer can also take into account the frame's root mean square,
3857 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3858 determine the power of a time-varying signal. It is therefore considered
3859 that the RMS is a better approximation of the "perceived loudness" than
3860 just looking at the signal's peak magnitude. Consequently, by adjusting all
3861 frames to a constant RMS value, a uniform "perceived loudness" can be
3862 established. If a target RMS value has been specified, a frame's local gain
3863 factor is defined as the factor that would result in exactly that RMS value.
3864 Note, however, that the maximum local gain factor is still restricted by the
3865 frame's highest magnitude sample, in order to prevent clipping.
3868 Enable channels coupling. By default is enabled.
3869 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3870 amount. This means the same gain factor will be applied to all channels, i.e.
3871 the maximum possible gain factor is determined by the "loudest" channel.
3872 However, in some recordings, it may happen that the volume of the different
3873 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3874 In this case, this option can be used to disable the channel coupling. This way,
3875 the gain factor will be determined independently for each channel, depending
3876 only on the individual channel's highest magnitude sample. This allows for
3877 harmonizing the volume of the different channels.
3880 Enable DC bias correction. By default is disabled.
3881 An audio signal (in the time domain) is a sequence of sample values.
3882 In the Dynamic Audio Normalizer these sample values are represented in the
3883 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3884 audio signal, or "waveform", should be centered around the zero point.
3885 That means if we calculate the mean value of all samples in a file, or in a
3886 single frame, then the result should be 0.0 or at least very close to that
3887 value. If, however, there is a significant deviation of the mean value from
3888 0.0, in either positive or negative direction, this is referred to as a
3889 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3890 Audio Normalizer provides optional DC bias correction.
3891 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3892 the mean value, or "DC correction" offset, of each input frame and subtract
3893 that value from all of the frame's sample values which ensures those samples
3894 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3895 boundaries, the DC correction offset values will be interpolated smoothly
3896 between neighbouring frames.
3898 @item altboundary, b
3899 Enable alternative boundary mode. By default is disabled.
3900 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3901 around each frame. This includes the preceding frames as well as the
3902 subsequent frames. However, for the "boundary" frames, located at the very
3903 beginning and at the very end of the audio file, not all neighbouring
3904 frames are available. In particular, for the first few frames in the audio
3905 file, the preceding frames are not known. And, similarly, for the last few
3906 frames in the audio file, the subsequent frames are not known. Thus, the
3907 question arises which gain factors should be assumed for the missing frames
3908 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3909 to deal with this situation. The default boundary mode assumes a gain factor
3910 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3911 "fade out" at the beginning and at the end of the input, respectively.
3914 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3915 By default, the Dynamic Audio Normalizer does not apply "traditional"
3916 compression. This means that signal peaks will not be pruned and thus the
3917 full dynamic range will be retained within each local neighbourhood. However,
3918 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3919 normalization algorithm with a more "traditional" compression.
3920 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3921 (thresholding) function. If (and only if) the compression feature is enabled,
3922 all input frames will be processed by a soft knee thresholding function prior
3923 to the actual normalization process. Put simply, the thresholding function is
3924 going to prune all samples whose magnitude exceeds a certain threshold value.
3925 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3926 value. Instead, the threshold value will be adjusted for each individual
3928 In general, smaller parameters result in stronger compression, and vice versa.
3929 Values below 3.0 are not recommended, because audible distortion may appear.
3932 Set the target threshold value. This specifies the lowest permissible
3933 magnitude level for the audio input which will be normalized.
3934 If input frame volume is above this value frame will be normalized.
3935 Otherwise frame may not be normalized at all. The default value is set
3936 to 0, which means all input frames will be normalized.
3937 This option is mostly useful if digital noise is not wanted to be amplified.
3940 @subsection Commands
3942 This filter supports the all above options as @ref{commands}.
3946 Make audio easier to listen to on headphones.
3948 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3949 so that when listened to on headphones the stereo image is moved from
3950 inside your head (standard for headphones) to outside and in front of
3951 the listener (standard for speakers).
3957 Apply a two-pole peaking equalisation (EQ) filter. With this
3958 filter, the signal-level at and around a selected frequency can
3959 be increased or decreased, whilst (unlike bandpass and bandreject
3960 filters) that at all other frequencies is unchanged.
3962 In order to produce complex equalisation curves, this filter can
3963 be given several times, each with a different central frequency.
3965 The filter accepts the following options:
3969 Set the filter's central frequency in Hz.
3972 Set method to specify band-width of filter.
3987 Specify the band-width of a filter in width_type units.
3990 Set the required gain or attenuation in dB.
3991 Beware of clipping when using a positive gain.
3994 How much to use filtered signal in output. Default is 1.
3995 Range is between 0 and 1.
3998 Specify which channels to filter, by default all available are filtered.
4001 Normalize biquad coefficients, by default is disabled.
4002 Enabling it will normalize magnitude response at DC to 0dB.
4005 Set transform type of IIR filter.
4014 Set precison of filtering.
4017 Pick automatic sample format depending on surround filters.
4019 Always use signed 16-bit.
4021 Always use signed 32-bit.
4023 Always use float 32-bit.
4025 Always use float 64-bit.
4029 @subsection Examples
4032 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4034 equalizer=f=1000:t=h:width=200:g=-10
4038 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4040 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4044 @subsection Commands
4046 This filter supports the following commands:
4049 Change equalizer frequency.
4050 Syntax for the command is : "@var{frequency}"
4053 Change equalizer width_type.
4054 Syntax for the command is : "@var{width_type}"
4057 Change equalizer width.
4058 Syntax for the command is : "@var{width}"
4061 Change equalizer gain.
4062 Syntax for the command is : "@var{gain}"
4065 Change equalizer mix.
4066 Syntax for the command is : "@var{mix}"
4069 @section extrastereo
4071 Linearly increases the difference between left and right channels which
4072 adds some sort of "live" effect to playback.
4074 The filter accepts the following options:
4078 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4079 (average of both channels), with 1.0 sound will be unchanged, with
4080 -1.0 left and right channels will be swapped.
4083 Enable clipping. By default is enabled.
4086 @subsection Commands
4088 This filter supports the all above options as @ref{commands}.
4090 @section firequalizer
4091 Apply FIR Equalization using arbitrary frequency response.
4093 The filter accepts the following option:
4097 Set gain curve equation (in dB). The expression can contain variables:
4100 the evaluated frequency
4104 channel number, set to 0 when multichannels evaluation is disabled
4106 channel id, see libavutil/channel_layout.h, set to the first channel id when
4107 multichannels evaluation is disabled
4111 channel_layout, see libavutil/channel_layout.h
4116 @item gain_interpolate(f)
4117 interpolate gain on frequency f based on gain_entry
4118 @item cubic_interpolate(f)
4119 same as gain_interpolate, but smoother
4121 This option is also available as command. Default is @code{gain_interpolate(f)}.
4124 Set gain entry for gain_interpolate function. The expression can
4128 store gain entry at frequency f with value g
4130 This option is also available as command.
4133 Set filter delay in seconds. Higher value means more accurate.
4134 Default is @code{0.01}.
4137 Set filter accuracy in Hz. Lower value means more accurate.
4138 Default is @code{5}.
4141 Set window function. Acceptable values are:
4144 rectangular window, useful when gain curve is already smooth
4146 hann window (default)
4152 3-terms continuous 1st derivative nuttall window
4154 minimum 3-terms discontinuous nuttall window
4156 4-terms continuous 1st derivative nuttall window
4158 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4160 blackman-harris window
4166 If enabled, use fixed number of audio samples. This improves speed when
4167 filtering with large delay. Default is disabled.
4170 Enable multichannels evaluation on gain. Default is disabled.
4173 Enable zero phase mode by subtracting timestamp to compensate delay.
4174 Default is disabled.
4177 Set scale used by gain. Acceptable values are:
4180 linear frequency, linear gain
4182 linear frequency, logarithmic (in dB) gain (default)
4184 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4186 logarithmic frequency, logarithmic gain
4190 Set file for dumping, suitable for gnuplot.
4193 Set scale for dumpfile. Acceptable values are same with scale option.
4197 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4198 Default is disabled.
4201 Enable minimum phase impulse response. Default is disabled.
4204 @subsection Examples
4209 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4212 lowpass at 1000 Hz with gain_entry:
4214 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4217 custom equalization:
4219 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4222 higher delay with zero phase to compensate delay:
4224 firequalizer=delay=0.1:fixed=on:zero_phase=on
4227 lowpass on left channel, highpass on right channel:
4229 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4230 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4235 Apply a flanging effect to the audio.
4237 The filter accepts the following options:
4241 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4244 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4247 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4251 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4252 Default value is 71.
4255 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4258 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4259 Default value is @var{sinusoidal}.
4262 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4263 Default value is 25.
4266 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4267 Default is @var{linear}.
4271 Apply Haas effect to audio.
4273 Note that this makes most sense to apply on mono signals.
4274 With this filter applied to mono signals it give some directionality and
4275 stretches its stereo image.
4277 The filter accepts the following options:
4281 Set input level. By default is @var{1}, or 0dB
4284 Set output level. By default is @var{1}, or 0dB.
4287 Set gain applied to side part of signal. By default is @var{1}.
4290 Set kind of middle source. Can be one of the following:
4300 Pick middle part signal of stereo image.
4303 Pick side part signal of stereo image.
4307 Change middle phase. By default is disabled.
4310 Set left channel delay. By default is @var{2.05} milliseconds.
4313 Set left channel balance. By default is @var{-1}.
4316 Set left channel gain. By default is @var{1}.
4319 Change left phase. By default is disabled.
4322 Set right channel delay. By defaults is @var{2.12} milliseconds.
4325 Set right channel balance. By default is @var{1}.
4328 Set right channel gain. By default is @var{1}.
4331 Change right phase. By default is enabled.
4336 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4337 embedded HDCD codes is expanded into a 20-bit PCM stream.
4339 The filter supports the Peak Extend and Low-level Gain Adjustment features
4340 of HDCD, and detects the Transient Filter flag.
4343 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4346 When using the filter with wav, note the default encoding for wav is 16-bit,
4347 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4348 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4350 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4351 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4354 The filter accepts the following options:
4357 @item disable_autoconvert
4358 Disable any automatic format conversion or resampling in the filter graph.
4360 @item process_stereo
4361 Process the stereo channels together. If target_gain does not match between
4362 channels, consider it invalid and use the last valid target_gain.
4365 Set the code detect timer period in ms.
4368 Always extend peaks above -3dBFS even if PE isn't signaled.
4371 Replace audio with a solid tone and adjust the amplitude to signal some
4372 specific aspect of the decoding process. The output file can be loaded in
4373 an audio editor alongside the original to aid analysis.
4375 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4382 Gain adjustment level at each sample
4384 Samples where peak extend occurs
4386 Samples where the code detect timer is active
4388 Samples where the target gain does not match between channels
4394 Apply head-related transfer functions (HRTFs) to create virtual
4395 loudspeakers around the user for binaural listening via headphones.
4396 The HRIRs are provided via additional streams, for each channel
4397 one stereo input stream is needed.
4399 The filter accepts the following options:
4403 Set mapping of input streams for convolution.
4404 The argument is a '|'-separated list of channel names in order as they
4405 are given as additional stream inputs for filter.
4406 This also specify number of input streams. Number of input streams
4407 must be not less than number of channels in first stream plus one.
4410 Set gain applied to audio. Value is in dB. Default is 0.
4413 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4414 processing audio in time domain which is slow.
4415 @var{freq} is processing audio in frequency domain which is fast.
4416 Default is @var{freq}.
4419 Set custom gain for LFE channels. Value is in dB. Default is 0.
4422 Set size of frame in number of samples which will be processed at once.
4423 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4426 Set format of hrir stream.
4427 Default value is @var{stereo}. Alternative value is @var{multich}.
4428 If value is set to @var{stereo}, number of additional streams should
4429 be greater or equal to number of input channels in first input stream.
4430 Also each additional stream should have stereo number of channels.
4431 If value is set to @var{multich}, number of additional streams should
4432 be exactly one. Also number of input channels of additional stream
4433 should be equal or greater than twice number of channels of first input
4437 @subsection Examples
4441 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4442 each amovie filter use stereo file with IR coefficients as input.
4443 The files give coefficients for each position of virtual loudspeaker:
4446 -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"
4451 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4452 but now in @var{multich} @var{hrir} format.
4454 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"
4461 Apply a high-pass filter with 3dB point frequency.
4462 The filter can be either single-pole, or double-pole (the default).
4463 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4465 The filter accepts the following options:
4469 Set frequency in Hz. Default is 3000.
4472 Set number of poles. Default is 2.
4475 Set method to specify band-width of filter.
4490 Specify the band-width of a filter in width_type units.
4491 Applies only to double-pole filter.
4492 The default is 0.707q and gives a Butterworth response.
4495 How much to use filtered signal in output. Default is 1.
4496 Range is between 0 and 1.
4499 Specify which channels to filter, by default all available are filtered.
4502 Normalize biquad coefficients, by default is disabled.
4503 Enabling it will normalize magnitude response at DC to 0dB.
4506 Set transform type of IIR filter.
4515 Set precison of filtering.
4518 Pick automatic sample format depending on surround filters.
4520 Always use signed 16-bit.
4522 Always use signed 32-bit.
4524 Always use float 32-bit.
4526 Always use float 64-bit.
4530 @subsection Commands
4532 This filter supports the following commands:
4535 Change highpass frequency.
4536 Syntax for the command is : "@var{frequency}"
4539 Change highpass width_type.
4540 Syntax for the command is : "@var{width_type}"
4543 Change highpass width.
4544 Syntax for the command is : "@var{width}"
4547 Change highpass mix.
4548 Syntax for the command is : "@var{mix}"
4553 Join multiple input streams into one multi-channel stream.
4555 It accepts the following parameters:
4559 The number of input streams. It defaults to 2.
4561 @item channel_layout
4562 The desired output channel layout. It defaults to stereo.
4565 Map channels from inputs to output. The argument is a '|'-separated list of
4566 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4567 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4568 can be either the name of the input channel (e.g. FL for front left) or its
4569 index in the specified input stream. @var{out_channel} is the name of the output
4573 The filter will attempt to guess the mappings when they are not specified
4574 explicitly. It does so by first trying to find an unused matching input channel
4575 and if that fails it picks the first unused input channel.
4577 Join 3 inputs (with properly set channel layouts):
4579 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4582 Build a 5.1 output from 6 single-channel streams:
4584 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4585 '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'
4591 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4593 To enable compilation of this filter you need to configure FFmpeg with
4594 @code{--enable-ladspa}.
4598 Specifies the name of LADSPA plugin library to load. If the environment
4599 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4600 each one of the directories specified by the colon separated list in
4601 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4602 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4603 @file{/usr/lib/ladspa/}.
4606 Specifies the plugin within the library. Some libraries contain only
4607 one plugin, but others contain many of them. If this is not set filter
4608 will list all available plugins within the specified library.
4611 Set the '|' separated list of controls which are zero or more floating point
4612 values that determine the behavior of the loaded plugin (for example delay,
4614 Controls need to be defined using the following syntax:
4615 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4616 @var{valuei} is the value set on the @var{i}-th control.
4617 Alternatively they can be also defined using the following syntax:
4618 @var{value0}|@var{value1}|@var{value2}|..., where
4619 @var{valuei} is the value set on the @var{i}-th control.
4620 If @option{controls} is set to @code{help}, all available controls and
4621 their valid ranges are printed.
4623 @item sample_rate, s
4624 Specify the sample rate, default to 44100. Only used if plugin have
4628 Set the number of samples per channel per each output frame, default
4629 is 1024. Only used if plugin have zero inputs.
4632 Set the minimum duration of the sourced audio. See
4633 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4634 for the accepted syntax.
4635 Note that the resulting duration may be greater than the specified duration,
4636 as the generated audio is always cut at the end of a complete frame.
4637 If not specified, or the expressed duration is negative, the audio is
4638 supposed to be generated forever.
4639 Only used if plugin have zero inputs.
4642 Enable latency compensation, by default is disabled.
4643 Only used if plugin have inputs.
4646 @subsection Examples
4650 List all available plugins within amp (LADSPA example plugin) library:
4656 List all available controls and their valid ranges for @code{vcf_notch}
4657 plugin from @code{VCF} library:
4659 ladspa=f=vcf:p=vcf_notch:c=help
4663 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4666 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4670 Add reverberation to the audio using TAP-plugins
4671 (Tom's Audio Processing plugins):
4673 ladspa=file=tap_reverb:tap_reverb
4677 Generate white noise, with 0.2 amplitude:
4679 ladspa=file=cmt:noise_source_white:c=c0=.2
4683 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4684 @code{C* Audio Plugin Suite} (CAPS) library:
4686 ladspa=file=caps:Click:c=c1=20'
4690 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4692 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4696 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4697 @code{SWH Plugins} collection:
4699 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4703 Attenuate low frequencies using Multiband EQ from Steve Harris
4704 @code{SWH Plugins} collection:
4706 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4710 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4713 ladspa=caps:Narrower
4717 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4719 ladspa=caps:White:.2
4723 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4725 ladspa=caps:Fractal:c=c1=1
4729 Dynamic volume normalization using @code{VLevel} plugin:
4731 ladspa=vlevel-ladspa:vlevel_mono
4735 @subsection Commands
4737 This filter supports the following commands:
4740 Modify the @var{N}-th control value.
4742 If the specified value is not valid, it is ignored and prior one is kept.
4747 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4748 Support for both single pass (livestreams, files) and double pass (files) modes.
4749 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4750 detect true peaks, the audio stream will be upsampled to 192 kHz.
4751 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4753 The filter accepts the following options:
4757 Set integrated loudness target.
4758 Range is -70.0 - -5.0. Default value is -24.0.
4761 Set loudness range target.
4762 Range is 1.0 - 20.0. Default value is 7.0.
4765 Set maximum true peak.
4766 Range is -9.0 - +0.0. Default value is -2.0.
4768 @item measured_I, measured_i
4769 Measured IL of input file.
4770 Range is -99.0 - +0.0.
4772 @item measured_LRA, measured_lra
4773 Measured LRA of input file.
4774 Range is 0.0 - 99.0.
4776 @item measured_TP, measured_tp
4777 Measured true peak of input file.
4778 Range is -99.0 - +99.0.
4780 @item measured_thresh
4781 Measured threshold of input file.
4782 Range is -99.0 - +0.0.
4785 Set offset gain. Gain is applied before the true-peak limiter.
4786 Range is -99.0 - +99.0. Default is +0.0.
4789 Normalize by linearly scaling the source audio.
4790 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4791 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4792 be lower than source LRA and the change in integrated loudness shouldn't
4793 result in a true peak which exceeds the target TP. If any of these
4794 conditions aren't met, normalization mode will revert to @var{dynamic}.
4795 Options are @code{true} or @code{false}. Default is @code{true}.
4798 Treat mono input files as "dual-mono". If a mono file is intended for playback
4799 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4800 If set to @code{true}, this option will compensate for this effect.
4801 Multi-channel input files are not affected by this option.
4802 Options are true or false. Default is false.
4805 Set print format for stats. Options are summary, json, or none.
4806 Default value is none.
4811 Apply a low-pass filter with 3dB point frequency.
4812 The filter can be either single-pole or double-pole (the default).
4813 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4815 The filter accepts the following options:
4819 Set frequency in Hz. Default is 500.
4822 Set number of poles. Default is 2.
4825 Set method to specify band-width of filter.
4840 Specify the band-width of a filter in width_type units.
4841 Applies only to double-pole filter.
4842 The default is 0.707q and gives a Butterworth response.
4845 How much to use filtered signal in output. Default is 1.
4846 Range is between 0 and 1.
4849 Specify which channels to filter, by default all available are filtered.
4852 Normalize biquad coefficients, by default is disabled.
4853 Enabling it will normalize magnitude response at DC to 0dB.
4856 Set transform type of IIR filter.
4865 Set precison of filtering.
4868 Pick automatic sample format depending on surround filters.
4870 Always use signed 16-bit.
4872 Always use signed 32-bit.
4874 Always use float 32-bit.
4876 Always use float 64-bit.
4880 @subsection Examples
4883 Lowpass only LFE channel, it LFE is not present it does nothing:
4889 @subsection Commands
4891 This filter supports the following commands:
4894 Change lowpass frequency.
4895 Syntax for the command is : "@var{frequency}"
4898 Change lowpass width_type.
4899 Syntax for the command is : "@var{width_type}"
4902 Change lowpass width.
4903 Syntax for the command is : "@var{width}"
4907 Syntax for the command is : "@var{mix}"
4912 Load a LV2 (LADSPA Version 2) plugin.
4914 To enable compilation of this filter you need to configure FFmpeg with
4915 @code{--enable-lv2}.
4919 Specifies the plugin URI. You may need to escape ':'.
4922 Set the '|' separated list of controls which are zero or more floating point
4923 values that determine the behavior of the loaded plugin (for example delay,
4925 If @option{controls} is set to @code{help}, all available controls and
4926 their valid ranges are printed.
4928 @item sample_rate, s
4929 Specify the sample rate, default to 44100. Only used if plugin have
4933 Set the number of samples per channel per each output frame, default
4934 is 1024. Only used if plugin have zero inputs.
4937 Set the minimum duration of the sourced audio. See
4938 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4939 for the accepted syntax.
4940 Note that the resulting duration may be greater than the specified duration,
4941 as the generated audio is always cut at the end of a complete frame.
4942 If not specified, or the expressed duration is negative, the audio is
4943 supposed to be generated forever.
4944 Only used if plugin have zero inputs.
4947 @subsection Examples
4951 Apply bass enhancer plugin from Calf:
4953 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4957 Apply vinyl plugin from Calf:
4959 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4963 Apply bit crusher plugin from ArtyFX:
4965 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4970 Multiband Compress or expand the audio's dynamic range.
4972 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4973 This is akin to the crossover of a loudspeaker, and results in flat frequency
4974 response when absent compander action.
4976 It accepts the following parameters:
4980 This option syntax is:
4981 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4982 For explanation of each item refer to compand filter documentation.
4988 Mix channels with specific gain levels. The filter accepts the output
4989 channel layout followed by a set of channels definitions.
4991 This filter is also designed to efficiently remap the channels of an audio
4994 The filter accepts parameters of the form:
4995 "@var{l}|@var{outdef}|@var{outdef}|..."
4999 output channel layout or number of channels
5002 output channel specification, of the form:
5003 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
5006 output channel to define, either a channel name (FL, FR, etc.) or a channel
5007 number (c0, c1, etc.)
5010 multiplicative coefficient for the channel, 1 leaving the volume unchanged
5013 input channel to use, see out_name for details; it is not possible to mix
5014 named and numbered input channels
5017 If the `=' in a channel specification is replaced by `<', then the gains for
5018 that specification will be renormalized so that the total is 1, thus
5019 avoiding clipping noise.
5021 @subsection Mixing examples
5023 For example, if you want to down-mix from stereo to mono, but with a bigger
5024 factor for the left channel:
5026 pan=1c|c0=0.9*c0+0.1*c1
5029 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5030 7-channels surround:
5032 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5035 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5036 that should be preferred (see "-ac" option) unless you have very specific
5039 @subsection Remapping examples
5041 The channel remapping will be effective if, and only if:
5044 @item gain coefficients are zeroes or ones,
5045 @item only one input per channel output,
5048 If all these conditions are satisfied, the filter will notify the user ("Pure
5049 channel mapping detected"), and use an optimized and lossless method to do the
5052 For example, if you have a 5.1 source and want a stereo audio stream by
5053 dropping the extra channels:
5055 pan="stereo| c0=FL | c1=FR"
5058 Given the same source, you can also switch front left and front right channels
5059 and keep the input channel layout:
5061 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5064 If the input is a stereo audio stream, you can mute the front left channel (and
5065 still keep the stereo channel layout) with:
5070 Still with a stereo audio stream input, you can copy the right channel in both
5071 front left and right:
5073 pan="stereo| c0=FR | c1=FR"
5078 ReplayGain scanner filter. This filter takes an audio stream as an input and
5079 outputs it unchanged.
5080 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5084 Convert the audio sample format, sample rate and channel layout. It is
5085 not meant to be used directly.
5088 Apply time-stretching and pitch-shifting with librubberband.
5090 To enable compilation of this filter, you need to configure FFmpeg with
5091 @code{--enable-librubberband}.
5093 The filter accepts the following options:
5097 Set tempo scale factor.
5100 Set pitch scale factor.
5103 Set transients detector.
5104 Possible values are:
5113 Possible values are:
5122 Possible values are:
5129 Set processing window size.
5130 Possible values are:
5139 Possible values are:
5146 Enable formant preservation when shift pitching.
5147 Possible values are:
5155 Possible values are:
5164 Possible values are:
5171 @subsection Commands
5173 This filter supports the following commands:
5176 Change filter tempo scale factor.
5177 Syntax for the command is : "@var{tempo}"
5180 Change filter pitch scale factor.
5181 Syntax for the command is : "@var{pitch}"
5184 @section sidechaincompress
5186 This filter acts like normal compressor but has the ability to compress
5187 detected signal using second input signal.
5188 It needs two input streams and returns one output stream.
5189 First input stream will be processed depending on second stream signal.
5190 The filtered signal then can be filtered with other filters in later stages of
5191 processing. See @ref{pan} and @ref{amerge} filter.
5193 The filter accepts the following options:
5197 Set input gain. Default is 1. Range is between 0.015625 and 64.
5200 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5201 Default is @code{downward}.
5204 If a signal of second stream raises above this level it will affect the gain
5205 reduction of first stream.
5206 By default is 0.125. Range is between 0.00097563 and 1.
5209 Set a ratio about which the signal is reduced. 1:2 means that if the level
5210 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5211 Default is 2. Range is between 1 and 20.
5214 Amount of milliseconds the signal has to rise above the threshold before gain
5215 reduction starts. Default is 20. Range is between 0.01 and 2000.
5218 Amount of milliseconds the signal has to fall below the threshold before
5219 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5222 Set the amount by how much signal will be amplified after processing.
5223 Default is 1. Range is from 1 to 64.
5226 Curve the sharp knee around the threshold to enter gain reduction more softly.
5227 Default is 2.82843. Range is between 1 and 8.
5230 Choose if the @code{average} level between all channels of side-chain stream
5231 or the louder(@code{maximum}) channel of side-chain stream affects the
5232 reduction. Default is @code{average}.
5235 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5236 of @code{rms}. Default is @code{rms} which is mainly smoother.
5239 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5242 How much to use compressed signal in output. Default is 1.
5243 Range is between 0 and 1.
5246 @subsection Commands
5248 This filter supports the all above options as @ref{commands}.
5250 @subsection Examples
5254 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5255 depending on the signal of 2nd input and later compressed signal to be
5256 merged with 2nd input:
5258 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5262 @section sidechaingate
5264 A sidechain gate acts like a normal (wideband) gate but has the ability to
5265 filter the detected signal before sending it to the gain reduction stage.
5266 Normally a gate uses the full range signal to detect a level above the
5268 For example: If you cut all lower frequencies from your sidechain signal
5269 the gate will decrease the volume of your track only if not enough highs
5270 appear. With this technique you are able to reduce the resonation of a
5271 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5273 It needs two input streams and returns one output stream.
5274 First input stream will be processed depending on second stream signal.
5276 The filter accepts the following options:
5280 Set input level before filtering.
5281 Default is 1. Allowed range is from 0.015625 to 64.
5284 Set the mode of operation. Can be @code{upward} or @code{downward}.
5285 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5286 will be amplified, expanding dynamic range in upward direction.
5287 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5290 Set the level of gain reduction when the signal is below the threshold.
5291 Default is 0.06125. Allowed range is from 0 to 1.
5292 Setting this to 0 disables reduction and then filter behaves like expander.
5295 If a signal rises above this level the gain reduction is released.
5296 Default is 0.125. Allowed range is from 0 to 1.
5299 Set a ratio about which the signal is reduced.
5300 Default is 2. Allowed range is from 1 to 9000.
5303 Amount of milliseconds the signal has to rise above the threshold before gain
5305 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5308 Amount of milliseconds the signal has to fall below the threshold before the
5309 reduction is increased again. Default is 250 milliseconds.
5310 Allowed range is from 0.01 to 9000.
5313 Set amount of amplification of signal after processing.
5314 Default is 1. Allowed range is from 1 to 64.
5317 Curve the sharp knee around the threshold to enter gain reduction more softly.
5318 Default is 2.828427125. Allowed range is from 1 to 8.
5321 Choose if exact signal should be taken for detection or an RMS like one.
5322 Default is rms. Can be peak or rms.
5325 Choose if the average level between all channels or the louder channel affects
5327 Default is average. Can be average or maximum.
5330 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5333 @subsection Commands
5335 This filter supports the all above options as @ref{commands}.
5337 @section silencedetect
5339 Detect silence in an audio stream.
5341 This filter logs a message when it detects that the input audio volume is less
5342 or equal to a noise tolerance value for a duration greater or equal to the
5343 minimum detected noise duration.
5345 The printed times and duration are expressed in seconds. The
5346 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5347 is set on the first frame whose timestamp equals or exceeds the detection
5348 duration and it contains the timestamp of the first frame of the silence.
5350 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5351 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5352 keys are set on the first frame after the silence. If @option{mono} is
5353 enabled, and each channel is evaluated separately, the @code{.X}
5354 suffixed keys are used, and @code{X} corresponds to the channel number.
5356 The filter accepts the following options:
5360 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5361 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5364 Set silence duration until notification (default is 2 seconds). See
5365 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5366 for the accepted syntax.
5369 Process each channel separately, instead of combined. By default is disabled.
5372 @subsection Examples
5376 Detect 5 seconds of silence with -50dB noise tolerance:
5378 silencedetect=n=-50dB:d=5
5382 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5383 tolerance in @file{silence.mp3}:
5385 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5389 @section silenceremove
5391 Remove silence from the beginning, middle or end of the audio.
5393 The filter accepts the following options:
5397 This value is used to indicate if audio should be trimmed at beginning of
5398 the audio. A value of zero indicates no silence should be trimmed from the
5399 beginning. When specifying a non-zero value, it trims audio up until it
5400 finds non-silence. Normally, when trimming silence from beginning of audio
5401 the @var{start_periods} will be @code{1} but it can be increased to higher
5402 values to trim all audio up to specific count of non-silence periods.
5403 Default value is @code{0}.
5405 @item start_duration
5406 Specify the amount of time that non-silence must be detected before it stops
5407 trimming audio. By increasing the duration, bursts of noises can be treated
5408 as silence and trimmed off. Default value is @code{0}.
5410 @item start_threshold
5411 This indicates what sample value should be treated as silence. For digital
5412 audio, a value of @code{0} may be fine but for audio recorded from analog,
5413 you may wish to increase the value to account for background noise.
5414 Can be specified in dB (in case "dB" is appended to the specified value)
5415 or amplitude ratio. Default value is @code{0}.
5418 Specify max duration of silence at beginning that will be kept after
5419 trimming. Default is 0, which is equal to trimming all samples detected
5423 Specify mode of detection of silence end in start of multi-channel audio.
5424 Can be @var{any} or @var{all}. Default is @var{any}.
5425 With @var{any}, any sample that is detected as non-silence will cause
5426 stopped trimming of silence.
5427 With @var{all}, only if all channels are detected as non-silence will cause
5428 stopped trimming of silence.
5431 Set the count for trimming silence from the end of audio.
5432 To remove silence from the middle of a file, specify a @var{stop_periods}
5433 that is negative. This value is then treated as a positive value and is
5434 used to indicate the effect should restart processing as specified by
5435 @var{start_periods}, making it suitable for removing periods of silence
5436 in the middle of the audio.
5437 Default value is @code{0}.
5440 Specify a duration of silence that must exist before audio is not copied any
5441 more. By specifying a higher duration, silence that is wanted can be left in
5443 Default value is @code{0}.
5445 @item stop_threshold
5446 This is the same as @option{start_threshold} but for trimming silence from
5448 Can be specified in dB (in case "dB" is appended to the specified value)
5449 or amplitude ratio. Default value is @code{0}.
5452 Specify max duration of silence at end that will be kept after
5453 trimming. Default is 0, which is equal to trimming all samples detected
5457 Specify mode of detection of silence start in end of multi-channel audio.
5458 Can be @var{any} or @var{all}. Default is @var{any}.
5459 With @var{any}, any sample that is detected as non-silence will cause
5460 stopped trimming of silence.
5461 With @var{all}, only if all channels are detected as non-silence will cause
5462 stopped trimming of silence.
5465 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5466 and works better with digital silence which is exactly 0.
5467 Default value is @code{rms}.
5470 Set duration in number of seconds used to calculate size of window in number
5471 of samples for detecting silence.
5472 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5475 @subsection Examples
5479 The following example shows how this filter can be used to start a recording
5480 that does not contain the delay at the start which usually occurs between
5481 pressing the record button and the start of the performance:
5483 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5487 Trim all silence encountered from beginning to end where there is more than 1
5488 second of silence in audio:
5490 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5494 Trim all digital silence samples, using peak detection, from beginning to end
5495 where there is more than 0 samples of digital silence in audio and digital
5496 silence is detected in all channels at same positions in stream:
5498 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5504 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5505 loudspeakers around the user for binaural listening via headphones (audio
5506 formats up to 9 channels supported).
5507 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5508 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5509 Austrian Academy of Sciences.
5511 To enable compilation of this filter you need to configure FFmpeg with
5512 @code{--enable-libmysofa}.
5514 The filter accepts the following options:
5518 Set the SOFA file used for rendering.
5521 Set gain applied to audio. Value is in dB. Default is 0.
5524 Set rotation of virtual loudspeakers in deg. Default is 0.
5527 Set elevation of virtual speakers in deg. Default is 0.
5530 Set distance in meters between loudspeakers and the listener with near-field
5531 HRTFs. Default is 1.
5534 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5535 processing audio in time domain which is slow.
5536 @var{freq} is processing audio in frequency domain which is fast.
5537 Default is @var{freq}.
5540 Set custom positions of virtual loudspeakers. Syntax for this option is:
5541 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5542 Each virtual loudspeaker is described with short channel name following with
5543 azimuth and elevation in degrees.
5544 Each virtual loudspeaker description is separated by '|'.
5545 For example to override front left and front right channel positions use:
5546 'speakers=FL 45 15|FR 345 15'.
5547 Descriptions with unrecognised channel names are ignored.
5550 Set custom gain for LFE channels. Value is in dB. Default is 0.
5553 Set custom frame size in number of samples. Default is 1024.
5554 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5555 is set to @var{freq}.
5558 Should all IRs be normalized upon importing SOFA file.
5559 By default is enabled.
5562 Should nearest IRs be interpolated with neighbor IRs if exact position
5563 does not match. By default is disabled.
5566 Minphase all IRs upon loading of SOFA file. By default is disabled.
5569 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5572 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5575 @subsection Examples
5579 Using ClubFritz6 sofa file:
5581 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5585 Using ClubFritz12 sofa file and bigger radius with small rotation:
5587 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5591 Similar as above but with custom speaker positions for front left, front right, back left and back right
5592 and also with custom gain:
5594 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5601 This filter expands or compresses each half-cycle of audio samples
5602 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5603 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5605 The filter accepts the following options:
5609 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5610 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5613 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5614 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5615 would be such that local peak value reaches target peak value but never to surpass it and that
5616 ratio between new and previous peak value does not surpass this option value.
5618 @item compression, c
5619 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5620 This option controls maximum local half-cycle of samples compression. This option is used
5621 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5622 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5623 that peak's half-cycle will be compressed by current compression factor.
5626 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5627 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5628 Any half-cycle samples with their local peak value below or same as this option value will be
5629 compressed by current compression factor, otherwise, if greater than threshold value they will be
5630 expanded with expansion factor so that it could reach peak target value but never surpass it.
5633 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5634 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5635 each new half-cycle until it reaches @option{expansion} value.
5636 Setting this options too high may lead to distortions.
5639 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5640 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5641 each new half-cycle until it reaches @option{compression} value.
5644 Specify which channels to filter, by default all available channels are filtered.
5647 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5648 option. When enabled any half-cycle of samples with their local peak value below or same as
5649 @option{threshold} option will be expanded otherwise it will be compressed.
5652 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5653 When disabled each filtered channel gain calculation is independent, otherwise when this option
5654 is enabled the minimum of all possible gains for each filtered channel is used.
5657 @subsection Commands
5659 This filter supports the all above options as @ref{commands}.
5661 @section stereotools
5663 This filter has some handy utilities to manage stereo signals, for converting
5664 M/S stereo recordings to L/R signal while having control over the parameters
5665 or spreading the stereo image of master track.
5667 The filter accepts the following options:
5671 Set input level before filtering for both channels. Defaults is 1.
5672 Allowed range is from 0.015625 to 64.
5675 Set output level after filtering for both channels. Defaults is 1.
5676 Allowed range is from 0.015625 to 64.
5679 Set input balance between both channels. Default is 0.
5680 Allowed range is from -1 to 1.
5683 Set output balance between both channels. Default is 0.
5684 Allowed range is from -1 to 1.
5687 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5688 clipping. Disabled by default.
5691 Mute the left channel. Disabled by default.
5694 Mute the right channel. Disabled by default.
5697 Change the phase of the left channel. Disabled by default.
5700 Change the phase of the right channel. Disabled by default.
5703 Set stereo mode. Available values are:
5707 Left/Right to Left/Right, this is default.
5710 Left/Right to Mid/Side.
5713 Mid/Side to Left/Right.
5716 Left/Right to Left/Left.
5719 Left/Right to Right/Right.
5722 Left/Right to Left + Right.
5725 Left/Right to Right/Left.
5728 Mid/Side to Left/Left.
5731 Mid/Side to Right/Right.
5734 Mid/Side to Right/Left.
5737 Left/Right to Left - Right.
5741 Set level of side signal. Default is 1.
5742 Allowed range is from 0.015625 to 64.
5745 Set balance of side signal. Default is 0.
5746 Allowed range is from -1 to 1.
5749 Set level of the middle signal. Default is 1.
5750 Allowed range is from 0.015625 to 64.
5753 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5756 Set stereo base between mono and inversed channels. Default is 0.
5757 Allowed range is from -1 to 1.
5760 Set delay in milliseconds how much to delay left from right channel and
5761 vice versa. Default is 0. Allowed range is from -20 to 20.
5764 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5767 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5769 @item bmode_in, bmode_out
5770 Set balance mode for balance_in/balance_out option.
5772 Can be one of the following:
5776 Classic balance mode. Attenuate one channel at time.
5777 Gain is raised up to 1.
5780 Similar as classic mode above but gain is raised up to 2.
5783 Equal power distribution, from -6dB to +6dB range.
5787 @subsection Commands
5789 This filter supports the all above options as @ref{commands}.
5791 @subsection Examples
5795 Apply karaoke like effect:
5797 stereotools=mlev=0.015625
5801 Convert M/S signal to L/R:
5803 "stereotools=mode=ms>lr"
5807 @section stereowiden
5809 This filter enhance the stereo effect by suppressing signal common to both
5810 channels and by delaying the signal of left into right and vice versa,
5811 thereby widening the stereo effect.
5813 The filter accepts the following options:
5817 Time in milliseconds of the delay of left signal into right and vice versa.
5818 Default is 20 milliseconds.
5821 Amount of gain in delayed signal into right and vice versa. Gives a delay
5822 effect of left signal in right output and vice versa which gives widening
5823 effect. Default is 0.3.
5826 Cross feed of left into right with inverted phase. This helps in suppressing
5827 the mono. If the value is 1 it will cancel all the signal common to both
5828 channels. Default is 0.3.
5831 Set level of input signal of original channel. Default is 0.8.
5834 @subsection Commands
5836 This filter supports the all above options except @code{delay} as @ref{commands}.
5838 @section superequalizer
5839 Apply 18 band equalizer.
5841 The filter accepts the following options:
5848 Set 131Hz band gain.
5850 Set 185Hz band gain.
5852 Set 262Hz band gain.
5854 Set 370Hz band gain.
5856 Set 523Hz band gain.
5858 Set 740Hz band gain.
5860 Set 1047Hz band gain.
5862 Set 1480Hz band gain.
5864 Set 2093Hz band gain.
5866 Set 2960Hz band gain.
5868 Set 4186Hz band gain.
5870 Set 5920Hz band gain.
5872 Set 8372Hz band gain.
5874 Set 11840Hz band gain.
5876 Set 16744Hz band gain.
5878 Set 20000Hz band gain.
5882 Apply audio surround upmix filter.
5884 This filter allows to produce multichannel output from audio stream.
5886 The filter accepts the following options:
5890 Set output channel layout. By default, this is @var{5.1}.
5892 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5893 for the required syntax.
5896 Set input channel layout. By default, this is @var{stereo}.
5898 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5899 for the required syntax.
5902 Set input volume level. By default, this is @var{1}.
5905 Set output volume level. By default, this is @var{1}.
5908 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5911 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5914 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5917 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5918 In @var{add} mode, LFE channel is created from input audio and added to output.
5919 In @var{sub} mode, LFE channel is created from input audio and added to output but
5920 also all non-LFE output channels are subtracted with output LFE channel.
5923 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5924 Default is @var{90}.
5927 Set front center input volume. By default, this is @var{1}.
5930 Set front center output volume. By default, this is @var{1}.
5933 Set front left input volume. By default, this is @var{1}.
5936 Set front left output volume. By default, this is @var{1}.
5939 Set front right input volume. By default, this is @var{1}.
5942 Set front right output volume. By default, this is @var{1}.
5945 Set side left input volume. By default, this is @var{1}.
5948 Set side left output volume. By default, this is @var{1}.
5951 Set side right input volume. By default, this is @var{1}.
5954 Set side right output volume. By default, this is @var{1}.
5957 Set back left input volume. By default, this is @var{1}.
5960 Set back left output volume. By default, this is @var{1}.
5963 Set back right input volume. By default, this is @var{1}.
5966 Set back right output volume. By default, this is @var{1}.
5969 Set back center input volume. By default, this is @var{1}.
5972 Set back center output volume. By default, this is @var{1}.
5975 Set LFE input volume. By default, this is @var{1}.
5978 Set LFE output volume. By default, this is @var{1}.
5981 Set spread usage of stereo image across X axis for all channels.
5984 Set spread usage of stereo image across Y axis for all channels.
5986 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5987 Set spread usage of stereo image across X axis for each channel.
5989 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5990 Set spread usage of stereo image across Y axis for each channel.
5993 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5996 Set window function.
5998 It accepts the following values:
6021 Default is @code{hann}.
6024 Set window overlap. If set to 1, the recommended overlap for selected
6025 window function will be picked. Default is @code{0.5}.
6028 @section treble, highshelf
6030 Boost or cut treble (upper) frequencies of the audio using a two-pole
6031 shelving filter with a response similar to that of a standard
6032 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6034 The filter accepts the following options:
6038 Give the gain at whichever is the lower of ~22 kHz and the
6039 Nyquist frequency. Its useful range is about -20 (for a large cut)
6040 to +20 (for a large boost). Beware of clipping when using a positive gain.
6043 Set the filter's central frequency and so can be used
6044 to extend or reduce the frequency range to be boosted or cut.
6045 The default value is @code{3000} Hz.
6048 Set method to specify band-width of filter.
6063 Determine how steep is the filter's shelf transition.
6066 Set number of poles. Default is 2.
6069 How much to use filtered signal in output. Default is 1.
6070 Range is between 0 and 1.
6073 Specify which channels to filter, by default all available are filtered.
6076 Normalize biquad coefficients, by default is disabled.
6077 Enabling it will normalize magnitude response at DC to 0dB.
6080 Set transform type of IIR filter.
6089 Set precison of filtering.
6092 Pick automatic sample format depending on surround filters.
6094 Always use signed 16-bit.
6096 Always use signed 32-bit.
6098 Always use float 32-bit.
6100 Always use float 64-bit.
6104 @subsection Commands
6106 This filter supports the following commands:
6109 Change treble frequency.
6110 Syntax for the command is : "@var{frequency}"
6113 Change treble width_type.
6114 Syntax for the command is : "@var{width_type}"
6117 Change treble width.
6118 Syntax for the command is : "@var{width}"
6122 Syntax for the command is : "@var{gain}"
6126 Syntax for the command is : "@var{mix}"
6131 Sinusoidal amplitude modulation.
6133 The filter accepts the following options:
6137 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6138 (20 Hz or lower) will result in a tremolo effect.
6139 This filter may also be used as a ring modulator by specifying
6140 a modulation frequency higher than 20 Hz.
6141 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6144 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6145 Default value is 0.5.
6150 Sinusoidal phase modulation.
6152 The filter accepts the following options:
6156 Modulation frequency in Hertz.
6157 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6160 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6161 Default value is 0.5.
6166 Adjust the input audio volume.
6168 It accepts the following parameters:
6172 Set audio volume expression.
6174 Output values are clipped to the maximum value.
6176 The output audio volume is given by the relation:
6178 @var{output_volume} = @var{volume} * @var{input_volume}
6181 The default value for @var{volume} is "1.0".
6184 This parameter represents the mathematical precision.
6186 It determines which input sample formats will be allowed, which affects the
6187 precision of the volume scaling.
6191 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6193 32-bit floating-point; this limits input sample format to FLT. (default)
6195 64-bit floating-point; this limits input sample format to DBL.
6199 Choose the behaviour on encountering ReplayGain side data in input frames.
6203 Remove ReplayGain side data, ignoring its contents (the default).
6206 Ignore ReplayGain side data, but leave it in the frame.
6209 Prefer the track gain, if present.
6212 Prefer the album gain, if present.
6215 @item replaygain_preamp
6216 Pre-amplification gain in dB to apply to the selected replaygain gain.
6218 Default value for @var{replaygain_preamp} is 0.0.
6220 @item replaygain_noclip
6221 Prevent clipping by limiting the gain applied.
6223 Default value for @var{replaygain_noclip} is 1.
6226 Set when the volume expression is evaluated.
6228 It accepts the following values:
6231 only evaluate expression once during the filter initialization, or
6232 when the @samp{volume} command is sent
6235 evaluate expression for each incoming frame
6238 Default value is @samp{once}.
6241 The volume expression can contain the following parameters.
6245 frame number (starting at zero)
6248 @item nb_consumed_samples
6249 number of samples consumed by the filter
6251 number of samples in the current frame
6253 original frame position in the file
6259 PTS at start of stream
6261 time at start of stream
6267 last set volume value
6270 Note that when @option{eval} is set to @samp{once} only the
6271 @var{sample_rate} and @var{tb} variables are available, all other
6272 variables will evaluate to NAN.
6274 @subsection Commands
6276 This filter supports the following commands:
6279 Modify the volume expression.
6280 The command accepts the same syntax of the corresponding option.
6282 If the specified expression is not valid, it is kept at its current
6286 @subsection Examples
6290 Halve the input audio volume:
6294 volume=volume=-6.0206dB
6297 In all the above example the named key for @option{volume} can be
6298 omitted, for example like in:
6304 Increase input audio power by 6 decibels using fixed-point precision:
6306 volume=volume=6dB:precision=fixed
6310 Fade volume after time 10 with an annihilation period of 5 seconds:
6312 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6316 @section volumedetect
6318 Detect the volume of the input video.
6320 The filter has no parameters. The input is not modified. Statistics about
6321 the volume will be printed in the log when the input stream end is reached.
6323 In particular it will show the mean volume (root mean square), maximum
6324 volume (on a per-sample basis), and the beginning of a histogram of the
6325 registered volume values (from the maximum value to a cumulated 1/1000 of
6328 All volumes are in decibels relative to the maximum PCM value.
6330 @subsection Examples
6332 Here is an excerpt of the output:
6334 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6335 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6336 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6337 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6338 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6339 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6340 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6341 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6342 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6348 The mean square energy is approximately -27 dB, or 10^-2.7.
6350 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6352 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6355 In other words, raising the volume by +4 dB does not cause any clipping,
6356 raising it by +5 dB causes clipping for 6 samples, etc.
6358 @c man end AUDIO FILTERS
6360 @chapter Audio Sources
6361 @c man begin AUDIO SOURCES
6363 Below is a description of the currently available audio sources.
6367 Buffer audio frames, and make them available to the filter chain.
6369 This source is mainly intended for a programmatic use, in particular
6370 through the interface defined in @file{libavfilter/buffersrc.h}.
6372 It accepts the following parameters:
6376 The timebase which will be used for timestamps of submitted frames. It must be
6377 either a floating-point number or in @var{numerator}/@var{denominator} form.
6380 The sample rate of the incoming audio buffers.
6383 The sample format of the incoming audio buffers.
6384 Either a sample format name or its corresponding integer representation from
6385 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6387 @item channel_layout
6388 The channel layout of the incoming audio buffers.
6389 Either a channel layout name from channel_layout_map in
6390 @file{libavutil/channel_layout.c} or its corresponding integer representation
6391 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6394 The number of channels of the incoming audio buffers.
6395 If both @var{channels} and @var{channel_layout} are specified, then they
6400 @subsection Examples
6403 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6406 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6407 Since the sample format with name "s16p" corresponds to the number
6408 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6411 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6416 Generate an audio signal specified by an expression.
6418 This source accepts in input one or more expressions (one for each
6419 channel), which are evaluated and used to generate a corresponding
6422 This source accepts the following options:
6426 Set the '|'-separated expressions list for each separate channel. In case the
6427 @option{channel_layout} option is not specified, the selected channel layout
6428 depends on the number of provided expressions. Otherwise the last
6429 specified expression is applied to the remaining output channels.
6431 @item channel_layout, c
6432 Set the channel layout. The number of channels in the specified layout
6433 must be equal to the number of specified expressions.
6436 Set the minimum duration of the sourced audio. See
6437 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6438 for the accepted syntax.
6439 Note that the resulting duration may be greater than the specified
6440 duration, as the generated audio is always cut at the end of a
6443 If not specified, or the expressed duration is negative, the audio is
6444 supposed to be generated forever.
6447 Set the number of samples per channel per each output frame,
6450 @item sample_rate, s
6451 Specify the sample rate, default to 44100.
6454 Each expression in @var{exprs} can contain the following constants:
6458 number of the evaluated sample, starting from 0
6461 time of the evaluated sample expressed in seconds, starting from 0
6468 @subsection Examples
6478 Generate a sin signal with frequency of 440 Hz, set sample rate to
6481 aevalsrc="sin(440*2*PI*t):s=8000"
6485 Generate a two channels signal, specify the channel layout (Front
6486 Center + Back Center) explicitly:
6488 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6492 Generate white noise:
6494 aevalsrc="-2+random(0)"
6498 Generate an amplitude modulated signal:
6500 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6504 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6506 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6513 Generate a FIR coefficients using frequency sampling method.
6515 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6517 The filter accepts the following options:
6521 Set number of filter coefficents in output audio stream.
6522 Default value is 1025.
6525 Set frequency points from where magnitude and phase are set.
6526 This must be in non decreasing order, and first element must be 0, while last element
6527 must be 1. Elements are separated by white spaces.
6530 Set magnitude value for every frequency point set by @option{frequency}.
6531 Number of values must be same as number of frequency points.
6532 Values are separated by white spaces.
6535 Set phase value for every frequency point set by @option{frequency}.
6536 Number of values must be same as number of frequency points.
6537 Values are separated by white spaces.
6539 @item sample_rate, r
6540 Set sample rate, default is 44100.
6543 Set number of samples per each frame. Default is 1024.
6546 Set window function. Default is blackman.
6551 The null audio source, return unprocessed audio frames. It is mainly useful
6552 as a template and to be employed in analysis / debugging tools, or as
6553 the source for filters which ignore the input data (for example the sox
6556 This source accepts the following options:
6560 @item channel_layout, cl
6562 Specifies the channel layout, and can be either an integer or a string
6563 representing a channel layout. The default value of @var{channel_layout}
6566 Check the channel_layout_map definition in
6567 @file{libavutil/channel_layout.c} for the mapping between strings and
6568 channel layout values.
6570 @item sample_rate, r
6571 Specifies the sample rate, and defaults to 44100.
6574 Set the number of samples per requested frames.
6577 Set the duration of the sourced audio. See
6578 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6579 for the accepted syntax.
6581 If not specified, or the expressed duration is negative, the audio is
6582 supposed to be generated forever.
6585 @subsection Examples
6589 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6591 anullsrc=r=48000:cl=4
6595 Do the same operation with a more obvious syntax:
6597 anullsrc=r=48000:cl=mono
6601 All the parameters need to be explicitly defined.
6605 Synthesize a voice utterance using the libflite library.
6607 To enable compilation of this filter you need to configure FFmpeg with
6608 @code{--enable-libflite}.
6610 Note that versions of the flite library prior to 2.0 are not thread-safe.
6612 The filter accepts the following options:
6617 If set to 1, list the names of the available voices and exit
6618 immediately. Default value is 0.
6621 Set the maximum number of samples per frame. Default value is 512.
6624 Set the filename containing the text to speak.
6627 Set the text to speak.
6630 Set the voice to use for the speech synthesis. Default value is
6631 @code{kal}. See also the @var{list_voices} option.
6634 @subsection Examples
6638 Read from file @file{speech.txt}, and synthesize the text using the
6639 standard flite voice:
6641 flite=textfile=speech.txt
6645 Read the specified text selecting the @code{slt} voice:
6647 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6651 Input text to ffmpeg:
6653 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6657 Make @file{ffplay} speak the specified text, using @code{flite} and
6658 the @code{lavfi} device:
6660 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6664 For more information about libflite, check:
6665 @url{http://www.festvox.org/flite/}
6669 Generate a noise audio signal.
6671 The filter accepts the following options:
6674 @item sample_rate, r
6675 Specify the sample rate. Default value is 48000 Hz.
6678 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6682 Specify the duration of the generated audio stream. Not specifying this option
6683 results in noise with an infinite length.
6685 @item color, colour, c
6686 Specify the color of noise. Available noise colors are white, pink, brown,
6687 blue, violet and velvet. Default color is white.
6690 Specify a value used to seed the PRNG.
6693 Set the number of samples per each output frame, default is 1024.
6696 @subsection Examples
6701 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6703 anoisesrc=d=60:c=pink:r=44100:a=0.5
6709 Generate odd-tap Hilbert transform FIR coefficients.
6711 The resulting stream can be used with @ref{afir} filter for phase-shifting
6712 the signal by 90 degrees.
6714 This is used in many matrix coding schemes and for analytic signal generation.
6715 The process is often written as a multiplication by i (or j), the imaginary unit.
6717 The filter accepts the following options:
6721 @item sample_rate, s
6722 Set sample rate, default is 44100.
6725 Set length of FIR filter, default is 22051.
6728 Set number of samples per each frame.
6731 Set window function to be used when generating FIR coefficients.
6736 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6738 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6740 The filter accepts the following options:
6743 @item sample_rate, r
6744 Set sample rate, default is 44100.
6747 Set number of samples per each frame. Default is 1024.
6750 Set high-pass frequency. Default is 0.
6753 Set low-pass frequency. Default is 0.
6754 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6755 is higher than 0 then filter will create band-pass filter coefficients,
6756 otherwise band-reject filter coefficients.
6759 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6762 Set Kaiser window beta.
6765 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6768 Enable rounding, by default is disabled.
6771 Set number of taps for high-pass filter.
6774 Set number of taps for low-pass filter.
6779 Generate an audio signal made of a sine wave with amplitude 1/8.
6781 The audio signal is bit-exact.
6783 The filter accepts the following options:
6788 Set the carrier frequency. Default is 440 Hz.
6790 @item beep_factor, b
6791 Enable a periodic beep every second with frequency @var{beep_factor} times
6792 the carrier frequency. Default is 0, meaning the beep is disabled.
6794 @item sample_rate, r
6795 Specify the sample rate, default is 44100.
6798 Specify the duration of the generated audio stream.
6800 @item samples_per_frame
6801 Set the number of samples per output frame.
6803 The expression can contain the following constants:
6807 The (sequential) number of the output audio frame, starting from 0.
6810 The PTS (Presentation TimeStamp) of the output audio frame,
6811 expressed in @var{TB} units.
6814 The PTS of the output audio frame, expressed in seconds.
6817 The timebase of the output audio frames.
6820 Default is @code{1024}.
6823 @subsection Examples
6828 Generate a simple 440 Hz sine wave:
6834 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6838 sine=frequency=220:beep_factor=4:duration=5
6842 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6845 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6849 @c man end AUDIO SOURCES
6851 @chapter Audio Sinks
6852 @c man begin AUDIO SINKS
6854 Below is a description of the currently available audio sinks.
6856 @section abuffersink
6858 Buffer audio frames, and make them available to the end of filter chain.
6860 This sink is mainly intended for programmatic use, in particular
6861 through the interface defined in @file{libavfilter/buffersink.h}
6862 or the options system.
6864 It accepts a pointer to an AVABufferSinkContext structure, which
6865 defines the incoming buffers' formats, to be passed as the opaque
6866 parameter to @code{avfilter_init_filter} for initialization.
6869 Null audio sink; do absolutely nothing with the input audio. It is
6870 mainly useful as a template and for use in analysis / debugging
6873 @c man end AUDIO SINKS
6875 @chapter Video Filters
6876 @c man begin VIDEO FILTERS
6878 When you configure your FFmpeg build, you can disable any of the
6879 existing filters using @code{--disable-filters}.
6880 The configure output will show the video filters included in your
6883 Below is a description of the currently available video filters.
6887 Mark a region of interest in a video frame.
6889 The frame data is passed through unchanged, but metadata is attached
6890 to the frame indicating regions of interest which can affect the
6891 behaviour of later encoding. Multiple regions can be marked by
6892 applying the filter multiple times.
6896 Region distance in pixels from the left edge of the frame.
6898 Region distance in pixels from the top edge of the frame.
6900 Region width in pixels.
6902 Region height in pixels.
6904 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6905 and may contain the following variables:
6908 Width of the input frame.
6910 Height of the input frame.
6914 Quantisation offset to apply within the region.
6916 This must be a real value in the range -1 to +1. A value of zero
6917 indicates no quality change. A negative value asks for better quality
6918 (less quantisation), while a positive value asks for worse quality
6919 (greater quantisation).
6921 The range is calibrated so that the extreme values indicate the
6922 largest possible offset - if the rest of the frame is encoded with the
6923 worst possible quality, an offset of -1 indicates that this region
6924 should be encoded with the best possible quality anyway. Intermediate
6925 values are then interpolated in some codec-dependent way.
6927 For example, in 10-bit H.264 the quantisation parameter varies between
6928 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6929 this region should be encoded with a QP around one-tenth of the full
6930 range better than the rest of the frame. So, if most of the frame
6931 were to be encoded with a QP of around 30, this region would get a QP
6932 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6933 An extreme value of -1 would indicate that this region should be
6934 encoded with the best possible quality regardless of the treatment of
6935 the rest of the frame - that is, should be encoded at a QP of -12.
6937 If set to true, remove any existing regions of interest marked on the
6938 frame before adding the new one.
6941 @subsection Examples
6945 Mark the centre quarter of the frame as interesting.
6947 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6950 Mark the 100-pixel-wide region on the left edge of the frame as very
6951 uninteresting (to be encoded at much lower quality than the rest of
6954 addroi=0:0:100:ih:+1/5
6958 @section alphaextract
6960 Extract the alpha component from the input as a grayscale video. This
6961 is especially useful with the @var{alphamerge} filter.
6965 Add or replace the alpha component of the primary input with the
6966 grayscale value of a second input. This is intended for use with
6967 @var{alphaextract} to allow the transmission or storage of frame
6968 sequences that have alpha in a format that doesn't support an alpha
6971 For example, to reconstruct full frames from a normal YUV-encoded video
6972 and a separate video created with @var{alphaextract}, you might use:
6974 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6979 Amplify differences between current pixel and pixels of adjacent frames in
6980 same pixel location.
6982 This filter accepts the following options:
6986 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6987 For example radius of 3 will instruct filter to calculate average of 7 frames.
6990 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6993 Set threshold for difference amplification. Any difference greater or equal to
6994 this value will not alter source pixel. Default is 10.
6995 Allowed range is from 0 to 65535.
6998 Set tolerance for difference amplification. Any difference lower to
6999 this value will not alter source pixel. Default is 0.
7000 Allowed range is from 0 to 65535.
7003 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7004 This option controls maximum possible value that will decrease source pixel value.
7007 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7008 This option controls maximum possible value that will increase source pixel value.
7011 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
7014 @subsection Commands
7016 This filter supports the following @ref{commands} that corresponds to option of same name:
7028 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7029 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7030 Substation Alpha) subtitles files.
7032 This filter accepts the following option in addition to the common options from
7033 the @ref{subtitles} filter:
7037 Set the shaping engine
7039 Available values are:
7042 The default libass shaping engine, which is the best available.
7044 Fast, font-agnostic shaper that can do only substitutions
7046 Slower shaper using OpenType for substitutions and positioning
7049 The default is @code{auto}.
7053 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7055 The filter accepts the following options:
7059 Set threshold A for 1st plane. Default is 0.02.
7060 Valid range is 0 to 0.3.
7063 Set threshold B for 1st plane. Default is 0.04.
7064 Valid range is 0 to 5.
7067 Set threshold A for 2nd plane. Default is 0.02.
7068 Valid range is 0 to 0.3.
7071 Set threshold B for 2nd plane. Default is 0.04.
7072 Valid range is 0 to 5.
7075 Set threshold A for 3rd plane. Default is 0.02.
7076 Valid range is 0 to 0.3.
7079 Set threshold B for 3rd plane. Default is 0.04.
7080 Valid range is 0 to 5.
7082 Threshold A is designed to react on abrupt changes in the input signal and
7083 threshold B is designed to react on continuous changes in the input signal.
7086 Set number of frames filter will use for averaging. Default is 9. Must be odd
7087 number in range [5, 129].
7090 Set what planes of frame filter will use for averaging. Default is all.
7093 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7094 Alternatively can be set to @code{s} serial.
7096 Parallel can be faster then serial, while other way around is never true.
7097 Parallel will abort early on first change being greater then thresholds, while serial
7098 will continue processing other side of frames if they are equal or below thresholds.
7103 Set sigma for 1st plane, 2nd plane or 3rd plane. Default is 32767.
7104 Valid range is from 0 to 32767.
7105 This options controls weight for each pixel in radius defined by size.
7106 Default value means every pixel have same weight.
7107 Setting this option to 0 effectively disables filtering.
7110 @subsection Commands
7111 This filter supports same @ref{commands} as options except option @code{s}.
7112 The command accepts the same syntax of the corresponding option.
7116 Apply average blur filter.
7118 The filter accepts the following options:
7122 Set horizontal radius size.
7125 Set which planes to filter. By default all planes are filtered.
7128 Set vertical radius size, if zero it will be same as @code{sizeX}.
7129 Default is @code{0}.
7132 @subsection Commands
7133 This filter supports same commands as options.
7134 The command accepts the same syntax of the corresponding option.
7136 If the specified expression is not valid, it is kept at its current
7141 Compute the bounding box for the non-black pixels in the input frame
7144 This filter computes the bounding box containing all the pixels with a
7145 luminance value greater than the minimum allowed value.
7146 The parameters describing the bounding box are printed on the filter
7149 The filter accepts the following option:
7153 Set the minimal luminance value. Default is @code{16}.
7156 @subsection Commands
7158 This filter supports the all above options as @ref{commands}.
7161 Apply bilateral filter, spatial smoothing while preserving edges.
7163 The filter accepts the following options:
7166 Set sigma of gaussian function to calculate spatial weight.
7167 Allowed range is 0 to 512. Default is 0.1.
7170 Set sigma of gaussian function to calculate range weight.
7171 Allowed range is 0 to 1. Default is 0.1.
7174 Set planes to filter. Default is first only.
7177 @subsection Commands
7179 This filter supports the all above options as @ref{commands}.
7181 @section bitplanenoise
7183 Show and measure bit plane noise.
7185 The filter accepts the following options:
7189 Set which plane to analyze. Default is @code{1}.
7192 Filter out noisy pixels from @code{bitplane} set above.
7193 Default is disabled.
7196 @section blackdetect
7198 Detect video intervals that are (almost) completely black. Can be
7199 useful to detect chapter transitions, commercials, or invalid
7202 The filter outputs its detection analysis to both the log as well as
7203 frame metadata. If a black segment of at least the specified minimum
7204 duration is found, a line with the start and end timestamps as well
7205 as duration is printed to the log with level @code{info}. In addition,
7206 a log line with level @code{debug} is printed per frame showing the
7207 black amount detected for that frame.
7209 The filter also attaches metadata to the first frame of a black
7210 segment with key @code{lavfi.black_start} and to the first frame
7211 after the black segment ends with key @code{lavfi.black_end}. The
7212 value is the frame's timestamp. This metadata is added regardless
7213 of the minimum duration specified.
7215 The filter accepts the following options:
7218 @item black_min_duration, d
7219 Set the minimum detected black duration expressed in seconds. It must
7220 be a non-negative floating point number.
7222 Default value is 2.0.
7224 @item picture_black_ratio_th, pic_th
7225 Set the threshold for considering a picture "black".
7226 Express the minimum value for the ratio:
7228 @var{nb_black_pixels} / @var{nb_pixels}
7231 for which a picture is considered black.
7232 Default value is 0.98.
7234 @item pixel_black_th, pix_th
7235 Set the threshold for considering a pixel "black".
7237 The threshold expresses the maximum pixel luminance value for which a
7238 pixel is considered "black". The provided value is scaled according to
7239 the following equation:
7241 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7244 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7245 the input video format, the range is [0-255] for YUV full-range
7246 formats and [16-235] for YUV non full-range formats.
7248 Default value is 0.10.
7251 The following example sets the maximum pixel threshold to the minimum
7252 value, and detects only black intervals of 2 or more seconds:
7254 blackdetect=d=2:pix_th=0.00
7259 Detect frames that are (almost) completely black. Can be useful to
7260 detect chapter transitions or commercials. Output lines consist of
7261 the frame number of the detected frame, the percentage of blackness,
7262 the position in the file if known or -1 and the timestamp in seconds.
7264 In order to display the output lines, you need to set the loglevel at
7265 least to the AV_LOG_INFO value.
7267 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7268 The value represents the percentage of pixels in the picture that
7269 are below the threshold value.
7271 It accepts the following parameters:
7276 The percentage of the pixels that have to be below the threshold; it defaults to
7279 @item threshold, thresh
7280 The threshold below which a pixel value is considered black; it defaults to
7288 Blend two video frames into each other.
7290 The @code{blend} filter takes two input streams and outputs one
7291 stream, the first input is the "top" layer and second input is
7292 "bottom" layer. By default, the output terminates when the longest input terminates.
7294 The @code{tblend} (time blend) filter takes two consecutive frames
7295 from one single stream, and outputs the result obtained by blending
7296 the new frame on top of the old frame.
7298 A description of the accepted options follows.
7306 Set blend mode for specific pixel component or all pixel components in case
7307 of @var{all_mode}. Default value is @code{normal}.
7309 Available values for component modes are:
7351 Set blend opacity for specific pixel component or all pixel components in case
7352 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7359 Set blend expression for specific pixel component or all pixel components in case
7360 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7362 The expressions can use the following variables:
7366 The sequential number of the filtered frame, starting from @code{0}.
7370 the coordinates of the current sample
7374 the width and height of currently filtered plane
7378 Width and height scale for the plane being filtered. It is the
7379 ratio between the dimensions of the current plane to the luma plane,
7380 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7381 the luma plane and @code{0.5,0.5} for the chroma planes.
7384 Time of the current frame, expressed in seconds.
7387 Value of pixel component at current location for first video frame (top layer).
7390 Value of pixel component at current location for second video frame (bottom layer).
7394 The @code{blend} filter also supports the @ref{framesync} options.
7396 @subsection Examples
7400 Apply transition from bottom layer to top layer in first 10 seconds:
7402 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7406 Apply linear horizontal transition from top layer to bottom layer:
7408 blend=all_expr='A*(X/W)+B*(1-X/W)'
7412 Apply 1x1 checkerboard effect:
7414 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7418 Apply uncover left effect:
7420 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7424 Apply uncover down effect:
7426 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7430 Apply uncover up-left effect:
7432 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7436 Split diagonally video and shows top and bottom layer on each side:
7438 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7442 Display differences between the current and the previous frame:
7444 tblend=all_mode=grainextract
7450 Denoise frames using Block-Matching 3D algorithm.
7452 The filter accepts the following options.
7456 Set denoising strength. Default value is 1.
7457 Allowed range is from 0 to 999.9.
7458 The denoising algorithm is very sensitive to sigma, so adjust it
7459 according to the source.
7462 Set local patch size. This sets dimensions in 2D.
7465 Set sliding step for processing blocks. Default value is 4.
7466 Allowed range is from 1 to 64.
7467 Smaller values allows processing more reference blocks and is slower.
7470 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7471 When set to 1, no block matching is done. Larger values allows more blocks
7473 Allowed range is from 1 to 256.
7476 Set radius for search block matching. Default is 9.
7477 Allowed range is from 1 to INT32_MAX.
7480 Set step between two search locations for block matching. Default is 1.
7481 Allowed range is from 1 to 64. Smaller is slower.
7484 Set threshold of mean square error for block matching. Valid range is 0 to
7488 Set thresholding parameter for hard thresholding in 3D transformed domain.
7489 Larger values results in stronger hard-thresholding filtering in frequency
7493 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7494 Default is @code{basic}.
7497 If enabled, filter will use 2nd stream for block matching.
7498 Default is disabled for @code{basic} value of @var{estim} option,
7499 and always enabled if value of @var{estim} is @code{final}.
7502 Set planes to filter. Default is all available except alpha.
7505 @subsection Examples
7509 Basic filtering with bm3d:
7511 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7515 Same as above, but filtering only luma:
7517 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7521 Same as above, but with both estimation modes:
7523 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
7527 Same as above, but prefilter with @ref{nlmeans} filter instead:
7529 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
7535 Apply a boxblur algorithm to the input video.
7537 It accepts the following parameters:
7541 @item luma_radius, lr
7542 @item luma_power, lp
7543 @item chroma_radius, cr
7544 @item chroma_power, cp
7545 @item alpha_radius, ar
7546 @item alpha_power, ap
7550 A description of the accepted options follows.
7553 @item luma_radius, lr
7554 @item chroma_radius, cr
7555 @item alpha_radius, ar
7556 Set an expression for the box radius in pixels used for blurring the
7557 corresponding input plane.
7559 The radius value must be a non-negative number, and must not be
7560 greater than the value of the expression @code{min(w,h)/2} for the
7561 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7564 Default value for @option{luma_radius} is "2". If not specified,
7565 @option{chroma_radius} and @option{alpha_radius} default to the
7566 corresponding value set for @option{luma_radius}.
7568 The expressions can contain the following constants:
7572 The input width and height in pixels.
7576 The input chroma image width and height in pixels.
7580 The horizontal and vertical chroma subsample values. For example, for the
7581 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7584 @item luma_power, lp
7585 @item chroma_power, cp
7586 @item alpha_power, ap
7587 Specify how many times the boxblur filter is applied to the
7588 corresponding plane.
7590 Default value for @option{luma_power} is 2. If not specified,
7591 @option{chroma_power} and @option{alpha_power} default to the
7592 corresponding value set for @option{luma_power}.
7594 A value of 0 will disable the effect.
7597 @subsection Examples
7601 Apply a boxblur filter with the luma, chroma, and alpha radii
7604 boxblur=luma_radius=2:luma_power=1
7609 Set the luma radius to 2, and alpha and chroma radius to 0:
7611 boxblur=2:1:cr=0:ar=0
7615 Set the luma and chroma radii to a fraction of the video dimension:
7617 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7623 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7624 Deinterlacing Filter").
7626 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7627 interpolation algorithms.
7628 It accepts the following parameters:
7632 The interlacing mode to adopt. It accepts one of the following values:
7636 Output one frame for each frame.
7638 Output one frame for each field.
7641 The default value is @code{send_field}.
7644 The picture field parity assumed for the input interlaced video. It accepts one
7645 of the following values:
7649 Assume the top field is first.
7651 Assume the bottom field is first.
7653 Enable automatic detection of field parity.
7656 The default value is @code{auto}.
7657 If the interlacing is unknown or the decoder does not export this information,
7658 top field first will be assumed.
7661 Specify which frames to deinterlace. Accepts one of the following
7666 Deinterlace all frames.
7668 Only deinterlace frames marked as interlaced.
7671 The default value is @code{all}.
7676 Apply Contrast Adaptive Sharpen filter to video stream.
7678 The filter accepts the following options:
7682 Set the sharpening strength. Default value is 0.
7685 Set planes to filter. Default value is to filter all
7686 planes except alpha plane.
7689 @subsection Commands
7690 This filter supports same @ref{commands} as options.
7693 Remove all color information for all colors except for certain one.
7695 The filter accepts the following options:
7699 The color which will not be replaced with neutral chroma.
7702 Similarity percentage with the above color.
7703 0.01 matches only the exact key color, while 1.0 matches everything.
7707 0.0 makes pixels either fully gray, or not gray at all.
7708 Higher values result in more preserved color.
7711 Signals that the color passed is already in YUV instead of RGB.
7713 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7714 This can be used to pass exact YUV values as hexadecimal numbers.
7717 @subsection Commands
7718 This filter supports same @ref{commands} as options.
7719 The command accepts the same syntax of the corresponding option.
7721 If the specified expression is not valid, it is kept at its current
7725 YUV colorspace color/chroma keying.
7727 The filter accepts the following options:
7731 The color which will be replaced with transparency.
7734 Similarity percentage with the key color.
7736 0.01 matches only the exact key color, while 1.0 matches everything.
7741 0.0 makes pixels either fully transparent, or not transparent at all.
7743 Higher values result in semi-transparent pixels, with a higher transparency
7744 the more similar the pixels color is to the key color.
7747 Signals that the color passed is already in YUV instead of RGB.
7749 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7750 This can be used to pass exact YUV values as hexadecimal numbers.
7753 @subsection Commands
7754 This filter supports same @ref{commands} as options.
7755 The command accepts the same syntax of the corresponding option.
7757 If the specified expression is not valid, it is kept at its current
7760 @subsection Examples
7764 Make every green pixel in the input image transparent:
7766 ffmpeg -i input.png -vf chromakey=green out.png
7770 Overlay a greenscreen-video on top of a static black background.
7772 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
7777 Reduce chrominance noise.
7779 The filter accepts the following options:
7783 Set threshold for averaging chrominance values.
7784 Sum of absolute difference of Y, U and V pixel components of current
7785 pixel and neighbour pixels lower than this threshold will be used in
7786 averaging. Luma component is left unchanged and is copied to output.
7787 Default value is 30. Allowed range is from 1 to 200.
7790 Set horizontal radius of rectangle used for averaging.
7791 Allowed range is from 1 to 100. Default value is 5.
7794 Set vertical radius of rectangle used for averaging.
7795 Allowed range is from 1 to 100. Default value is 5.
7798 Set horizontal step when averaging. Default value is 1.
7799 Allowed range is from 1 to 50.
7800 Mostly useful to speed-up filtering.
7803 Set vertical step when averaging. Default value is 1.
7804 Allowed range is from 1 to 50.
7805 Mostly useful to speed-up filtering.
7808 Set Y threshold for averaging chrominance values.
7809 Set finer control for max allowed difference between Y components
7810 of current pixel and neigbour pixels.
7811 Default value is 200. Allowed range is from 1 to 200.
7814 Set U threshold for averaging chrominance values.
7815 Set finer control for max allowed difference between U components
7816 of current pixel and neigbour pixels.
7817 Default value is 200. Allowed range is from 1 to 200.
7820 Set V threshold for averaging chrominance values.
7821 Set finer control for max allowed difference between V components
7822 of current pixel and neigbour pixels.
7823 Default value is 200. Allowed range is from 1 to 200.
7826 @subsection Commands
7827 This filter supports same @ref{commands} as options.
7828 The command accepts the same syntax of the corresponding option.
7830 @section chromashift
7831 Shift chroma pixels horizontally and/or vertically.
7833 The filter accepts the following options:
7836 Set amount to shift chroma-blue horizontally.
7838 Set amount to shift chroma-blue vertically.
7840 Set amount to shift chroma-red horizontally.
7842 Set amount to shift chroma-red vertically.
7844 Set edge mode, can be @var{smear}, default, or @var{warp}.
7847 @subsection Commands
7849 This filter supports the all above options as @ref{commands}.
7853 Display CIE color diagram with pixels overlaid onto it.
7855 The filter accepts the following options:
7870 @item uhdtv, rec2020
7884 Set what gamuts to draw.
7886 See @code{system} option for available values.
7889 Set ciescope size, by default set to 512.
7892 Set intensity used to map input pixel values to CIE diagram.
7895 Set contrast used to draw tongue colors that are out of active color system gamut.
7898 Correct gamma displayed on scope, by default enabled.
7901 Show white point on CIE diagram, by default disabled.
7904 Set input gamma. Used only with XYZ input color space.
7909 Visualize information exported by some codecs.
7911 Some codecs can export information through frames using side-data or other
7912 means. For example, some MPEG based codecs export motion vectors through the
7913 @var{export_mvs} flag in the codec @option{flags2} option.
7915 The filter accepts the following option:
7919 Set motion vectors to visualize.
7921 Available flags for @var{mv} are:
7925 forward predicted MVs of P-frames
7927 forward predicted MVs of B-frames
7929 backward predicted MVs of B-frames
7933 Display quantization parameters using the chroma planes.
7936 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7938 Available flags for @var{mv_type} are:
7942 forward predicted MVs
7944 backward predicted MVs
7947 @item frame_type, ft
7948 Set frame type to visualize motion vectors of.
7950 Available flags for @var{frame_type} are:
7954 intra-coded frames (I-frames)
7956 predicted frames (P-frames)
7958 bi-directionally predicted frames (B-frames)
7962 @subsection Examples
7966 Visualize forward predicted MVs of all frames using @command{ffplay}:
7968 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7972 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7974 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7978 @section colorbalance
7979 Modify intensity of primary colors (red, green and blue) of input frames.
7981 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7982 regions for the red-cyan, green-magenta or blue-yellow balance.
7984 A positive adjustment value shifts the balance towards the primary color, a negative
7985 value towards the complementary color.
7987 The filter accepts the following options:
7993 Adjust red, green and blue shadows (darkest pixels).
7998 Adjust red, green and blue midtones (medium pixels).
8003 Adjust red, green and blue highlights (brightest pixels).
8005 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8008 Preserve lightness when changing color balance. Default is disabled.
8011 @subsection Examples
8015 Add red color cast to shadows:
8021 @subsection Commands
8023 This filter supports the all above options as @ref{commands}.
8025 @section colorchannelmixer
8027 Adjust video input frames by re-mixing color channels.
8029 This filter modifies a color channel by adding the values associated to
8030 the other channels of the same pixels. For example if the value to
8031 modify is red, the output value will be:
8033 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8036 The filter accepts the following options:
8043 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8044 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8050 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8051 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8057 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8058 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8064 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8065 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8067 Allowed ranges for options are @code{[-2.0, 2.0]}.
8070 Preserve lightness when changing colors. Allowed range is from @code{[0.0, 1.0]}.
8071 Default is @code{0.0}, thus disabled.
8074 @subsection Examples
8078 Convert source to grayscale:
8080 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8083 Simulate sepia tones:
8085 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8089 @subsection Commands
8091 This filter supports the all above options as @ref{commands}.
8094 RGB colorspace color keying.
8096 The filter accepts the following options:
8100 The color which will be replaced with transparency.
8103 Similarity percentage with the key color.
8105 0.01 matches only the exact key color, while 1.0 matches everything.
8110 0.0 makes pixels either fully transparent, or not transparent at all.
8112 Higher values result in semi-transparent pixels, with a higher transparency
8113 the more similar the pixels color is to the key color.
8116 @subsection Examples
8120 Make every green pixel in the input image transparent:
8122 ffmpeg -i input.png -vf colorkey=green out.png
8126 Overlay a greenscreen-video on top of a static background image.
8128 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
8132 @subsection Commands
8133 This filter supports same @ref{commands} as options.
8134 The command accepts the same syntax of the corresponding option.
8136 If the specified expression is not valid, it is kept at its current
8140 Remove all color information for all RGB colors except for certain one.
8142 The filter accepts the following options:
8146 The color which will not be replaced with neutral gray.
8149 Similarity percentage with the above color.
8150 0.01 matches only the exact key color, while 1.0 matches everything.
8153 Blend percentage. 0.0 makes pixels fully gray.
8154 Higher values result in more preserved color.
8157 @subsection Commands
8158 This filter supports same @ref{commands} as options.
8159 The command accepts the same syntax of the corresponding option.
8161 If the specified expression is not valid, it is kept at its current
8164 @section colorlevels
8166 Adjust video input frames using levels.
8168 The filter accepts the following options:
8175 Adjust red, green, blue and alpha input black point.
8176 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8182 Adjust red, green, blue and alpha input white point.
8183 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8185 Input levels are used to lighten highlights (bright tones), darken shadows
8186 (dark tones), change the balance of bright and dark tones.
8192 Adjust red, green, blue and alpha output black point.
8193 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8199 Adjust red, green, blue and alpha output white point.
8200 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8202 Output levels allows manual selection of a constrained output level range.
8205 @subsection Examples
8209 Make video output darker:
8211 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8217 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8221 Make video output lighter:
8223 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8227 Increase brightness:
8229 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8233 @subsection Commands
8235 This filter supports the all above options as @ref{commands}.
8237 @section colormatrix
8239 Convert color matrix.
8241 The filter accepts the following options:
8246 Specify the source and destination color matrix. Both values must be
8249 The accepted values are:
8277 For example to convert from BT.601 to SMPTE-240M, use the command:
8279 colormatrix=bt601:smpte240m
8284 Convert colorspace, transfer characteristics or color primaries.
8285 Input video needs to have an even size.
8287 The filter accepts the following options:
8292 Specify all color properties at once.
8294 The accepted values are:
8324 Specify output colorspace.
8326 The accepted values are:
8335 BT.470BG or BT.601-6 625
8338 SMPTE-170M or BT.601-6 525
8347 BT.2020 with non-constant luminance
8353 Specify output transfer characteristics.
8355 The accepted values are:
8367 Constant gamma of 2.2
8370 Constant gamma of 2.8
8373 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8391 BT.2020 for 10-bits content
8394 BT.2020 for 12-bits content
8400 Specify output color primaries.
8402 The accepted values are:
8411 BT.470BG or BT.601-6 625
8414 SMPTE-170M or BT.601-6 525
8438 Specify output color range.
8440 The accepted values are:
8443 TV (restricted) range
8446 MPEG (restricted) range
8457 Specify output color format.
8459 The accepted values are:
8462 YUV 4:2:0 planar 8-bits
8465 YUV 4:2:0 planar 10-bits
8468 YUV 4:2:0 planar 12-bits
8471 YUV 4:2:2 planar 8-bits
8474 YUV 4:2:2 planar 10-bits
8477 YUV 4:2:2 planar 12-bits
8480 YUV 4:4:4 planar 8-bits
8483 YUV 4:4:4 planar 10-bits
8486 YUV 4:4:4 planar 12-bits
8491 Do a fast conversion, which skips gamma/primary correction. This will take
8492 significantly less CPU, but will be mathematically incorrect. To get output
8493 compatible with that produced by the colormatrix filter, use fast=1.
8496 Specify dithering mode.
8498 The accepted values are:
8504 Floyd-Steinberg dithering
8508 Whitepoint adaptation mode.
8510 The accepted values are:
8513 Bradford whitepoint adaptation
8516 von Kries whitepoint adaptation
8519 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8523 Override all input properties at once. Same accepted values as @ref{all}.
8526 Override input colorspace. Same accepted values as @ref{space}.
8529 Override input color primaries. Same accepted values as @ref{primaries}.
8532 Override input transfer characteristics. Same accepted values as @ref{trc}.
8535 Override input color range. Same accepted values as @ref{range}.
8539 The filter converts the transfer characteristics, color space and color
8540 primaries to the specified user values. The output value, if not specified,
8541 is set to a default value based on the "all" property. If that property is
8542 also not specified, the filter will log an error. The output color range and
8543 format default to the same value as the input color range and format. The
8544 input transfer characteristics, color space, color primaries and color range
8545 should be set on the input data. If any of these are missing, the filter will
8546 log an error and no conversion will take place.
8548 For example to convert the input to SMPTE-240M, use the command:
8550 colorspace=smpte240m
8553 @section colortemperature
8554 Adjust color temperature in video to simulate variations in ambient color temperature.
8556 The filter accepts the following options:
8560 Set the temperature in Kelvin. Allowed range is from 1000 to 40000.
8561 Default value is 6500 K.
8564 Set mixing with filtered output. Allowed range is from 0 to 1.
8568 Set the amount of preserving lightness. Allowed range is from 0 to 1.
8572 @subsection Commands
8573 This filter supports same @ref{commands} as options.
8575 @section convolution
8577 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8579 The filter accepts the following options:
8586 Set matrix for each plane.
8587 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8588 and from 1 to 49 odd number of signed integers in @var{row} mode.
8594 Set multiplier for calculated value for each plane.
8595 If unset or 0, it will be sum of all matrix elements.
8601 Set bias for each plane. This value is added to the result of the multiplication.
8602 Useful for making the overall image brighter or darker. Default is 0.0.
8608 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8609 Default is @var{square}.
8612 @subsection Commands
8614 This filter supports the all above options as @ref{commands}.
8616 @subsection Examples
8622 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"
8628 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"
8634 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"
8640 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"
8644 Apply laplacian edge detector which includes diagonals:
8646 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"
8652 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"
8658 Apply 2D convolution of video stream in frequency domain using second stream
8661 The filter accepts the following options:
8665 Set which planes to process.
8668 Set which impulse video frames will be processed, can be @var{first}
8669 or @var{all}. Default is @var{all}.
8672 The @code{convolve} filter also supports the @ref{framesync} options.
8676 Copy the input video source unchanged to the output. This is mainly useful for
8681 Video filtering on GPU using Apple's CoreImage API on OSX.
8683 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8684 processed by video hardware. However, software-based OpenGL implementations
8685 exist which means there is no guarantee for hardware processing. It depends on
8688 There are many filters and image generators provided by Apple that come with a
8689 large variety of options. The filter has to be referenced by its name along
8692 The coreimage filter accepts the following options:
8695 List all available filters and generators along with all their respective
8696 options as well as possible minimum and maximum values along with the default
8703 Specify all filters by their respective name and options.
8704 Use @var{list_filters} to determine all valid filter names and options.
8705 Numerical options are specified by a float value and are automatically clamped
8706 to their respective value range. Vector and color options have to be specified
8707 by a list of space separated float values. Character escaping has to be done.
8708 A special option name @code{default} is available to use default options for a
8711 It is required to specify either @code{default} or at least one of the filter options.
8712 All omitted options are used with their default values.
8713 The syntax of the filter string is as follows:
8715 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8719 Specify a rectangle where the output of the filter chain is copied into the
8720 input image. It is given by a list of space separated float values:
8722 output_rect=x\ y\ width\ height
8724 If not given, the output rectangle equals the dimensions of the input image.
8725 The output rectangle is automatically cropped at the borders of the input
8726 image. Negative values are valid for each component.
8728 output_rect=25\ 25\ 100\ 100
8732 Several filters can be chained for successive processing without GPU-HOST
8733 transfers allowing for fast processing of complex filter chains.
8734 Currently, only filters with zero (generators) or exactly one (filters) input
8735 image and one output image are supported. Also, transition filters are not yet
8738 Some filters generate output images with additional padding depending on the
8739 respective filter kernel. The padding is automatically removed to ensure the
8740 filter output has the same size as the input image.
8742 For image generators, the size of the output image is determined by the
8743 previous output image of the filter chain or the input image of the whole
8744 filterchain, respectively. The generators do not use the pixel information of
8745 this image to generate their output. However, the generated output is
8746 blended onto this image, resulting in partial or complete coverage of the
8749 The @ref{coreimagesrc} video source can be used for generating input images
8750 which are directly fed into the filter chain. By using it, providing input
8751 images by another video source or an input video is not required.
8753 @subsection Examples
8758 List all filters available:
8760 coreimage=list_filters=true
8764 Use the CIBoxBlur filter with default options to blur an image:
8766 coreimage=filter=CIBoxBlur@@default
8770 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8771 its center at 100x100 and a radius of 50 pixels:
8773 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8777 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8778 given as complete and escaped command-line for Apple's standard bash shell:
8780 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8786 Cover a rectangular object
8788 It accepts the following options:
8792 Filepath of the optional cover image, needs to be in yuv420.
8797 It accepts the following values:
8800 cover it by the supplied image
8802 cover it by interpolating the surrounding pixels
8805 Default value is @var{blur}.
8808 @subsection Examples
8812 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8814 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8820 Crop the input video to given dimensions.
8822 It accepts the following parameters:
8826 The width of the output video. It defaults to @code{iw}.
8827 This expression is evaluated only once during the filter
8828 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8831 The height of the output video. It defaults to @code{ih}.
8832 This expression is evaluated only once during the filter
8833 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8836 The horizontal position, in the input video, of the left edge of the output
8837 video. It defaults to @code{(in_w-out_w)/2}.
8838 This expression is evaluated per-frame.
8841 The vertical position, in the input video, of the top edge of the output video.
8842 It defaults to @code{(in_h-out_h)/2}.
8843 This expression is evaluated per-frame.
8846 If set to 1 will force the output display aspect ratio
8847 to be the same of the input, by changing the output sample aspect
8848 ratio. It defaults to 0.
8851 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8852 width/height/x/y as specified and will not be rounded to nearest smaller value.
8856 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8857 expressions containing the following constants:
8862 The computed values for @var{x} and @var{y}. They are evaluated for
8867 The input width and height.
8871 These are the same as @var{in_w} and @var{in_h}.
8875 The output (cropped) width and height.
8879 These are the same as @var{out_w} and @var{out_h}.
8882 same as @var{iw} / @var{ih}
8885 input sample aspect ratio
8888 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8892 horizontal and vertical chroma subsample values. For example for the
8893 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8896 The number of the input frame, starting from 0.
8899 the position in the file of the input frame, NAN if unknown
8902 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8906 The expression for @var{out_w} may depend on the value of @var{out_h},
8907 and the expression for @var{out_h} may depend on @var{out_w}, but they
8908 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8909 evaluated after @var{out_w} and @var{out_h}.
8911 The @var{x} and @var{y} parameters specify the expressions for the
8912 position of the top-left corner of the output (non-cropped) area. They
8913 are evaluated for each frame. If the evaluated value is not valid, it
8914 is approximated to the nearest valid value.
8916 The expression for @var{x} may depend on @var{y}, and the expression
8917 for @var{y} may depend on @var{x}.
8919 @subsection Examples
8923 Crop area with size 100x100 at position (12,34).
8928 Using named options, the example above becomes:
8930 crop=w=100:h=100:x=12:y=34
8934 Crop the central input area with size 100x100:
8940 Crop the central input area with size 2/3 of the input video:
8942 crop=2/3*in_w:2/3*in_h
8946 Crop the input video central square:
8953 Delimit the rectangle with the top-left corner placed at position
8954 100:100 and the right-bottom corner corresponding to the right-bottom
8955 corner of the input image.
8957 crop=in_w-100:in_h-100:100:100
8961 Crop 10 pixels from the left and right borders, and 20 pixels from
8962 the top and bottom borders
8964 crop=in_w-2*10:in_h-2*20
8968 Keep only the bottom right quarter of the input image:
8970 crop=in_w/2:in_h/2:in_w/2:in_h/2
8974 Crop height for getting Greek harmony:
8976 crop=in_w:1/PHI*in_w
8980 Apply trembling effect:
8982 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)
8986 Apply erratic camera effect depending on timestamp:
8988 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)"
8992 Set x depending on the value of y:
8994 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8998 @subsection Commands
9000 This filter supports the following commands:
9006 Set width/height of the output video and the horizontal/vertical position
9008 The command accepts the same syntax of the corresponding option.
9010 If the specified expression is not valid, it is kept at its current
9016 Auto-detect the crop size.
9018 It calculates the necessary cropping parameters and prints the
9019 recommended parameters via the logging system. The detected dimensions
9020 correspond to the non-black area of the input video.
9022 It accepts the following parameters:
9027 Set higher black value threshold, which can be optionally specified
9028 from nothing (0) to everything (255 for 8-bit based formats). An intensity
9029 value greater to the set value is considered non-black. It defaults to 24.
9030 You can also specify a value between 0.0 and 1.0 which will be scaled depending
9031 on the bitdepth of the pixel format.
9034 The value which the width/height should be divisible by. It defaults to
9035 16. The offset is automatically adjusted to center the video. Use 2 to
9036 get only even dimensions (needed for 4:2:2 video). 16 is best when
9037 encoding to most video codecs.
9040 Set the number of initial frames for which evaluation is skipped.
9041 Default is 2. Range is 0 to INT_MAX.
9043 @item reset_count, reset
9044 Set the counter that determines after how many frames cropdetect will
9045 reset the previously detected largest video area and start over to
9046 detect the current optimal crop area. Default value is 0.
9048 This can be useful when channel logos distort the video area. 0
9049 indicates 'never reset', and returns the largest area encountered during
9056 Delay video filtering until a given wallclock timestamp. The filter first
9057 passes on @option{preroll} amount of frames, then it buffers at most
9058 @option{buffer} amount of frames and waits for the cue. After reaching the cue
9059 it forwards the buffered frames and also any subsequent frames coming in its
9062 The filter can be used synchronize the output of multiple ffmpeg processes for
9063 realtime output devices like decklink. By putting the delay in the filtering
9064 chain and pre-buffering frames the process can pass on data to output almost
9065 immediately after the target wallclock timestamp is reached.
9067 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9073 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9076 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9079 The maximum duration of content to buffer before waiting for the cue expressed
9080 in seconds. Default is 0.
9087 Apply color adjustments using curves.
9089 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9090 component (red, green and blue) has its values defined by @var{N} key points
9091 tied from each other using a smooth curve. The x-axis represents the pixel
9092 values from the input frame, and the y-axis the new pixel values to be set for
9095 By default, a component curve is defined by the two points @var{(0;0)} and
9096 @var{(1;1)}. This creates a straight line where each original pixel value is
9097 "adjusted" to its own value, which means no change to the image.
9099 The filter allows you to redefine these two points and add some more. A new
9100 curve (using a natural cubic spline interpolation) will be define to pass
9101 smoothly through all these new coordinates. The new defined points needs to be
9102 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9103 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9104 the vector spaces, the values will be clipped accordingly.
9106 The filter accepts the following options:
9110 Select one of the available color presets. This option can be used in addition
9111 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9112 options takes priority on the preset values.
9113 Available presets are:
9116 @item color_negative
9119 @item increase_contrast
9121 @item linear_contrast
9122 @item medium_contrast
9124 @item strong_contrast
9127 Default is @code{none}.
9129 Set the master key points. These points will define a second pass mapping. It
9130 is sometimes called a "luminance" or "value" mapping. It can be used with
9131 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9132 post-processing LUT.
9134 Set the key points for the red component.
9136 Set the key points for the green component.
9138 Set the key points for the blue component.
9140 Set the key points for all components (not including master).
9141 Can be used in addition to the other key points component
9142 options. In this case, the unset component(s) will fallback on this
9143 @option{all} setting.
9145 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9147 Save Gnuplot script of the curves in specified file.
9150 To avoid some filtergraph syntax conflicts, each key points list need to be
9151 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9153 @subsection Examples
9157 Increase slightly the middle level of blue:
9159 curves=blue='0/0 0.5/0.58 1/1'
9165 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'
9167 Here we obtain the following coordinates for each components:
9170 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9172 @code{(0;0) (0.50;0.48) (1;1)}
9174 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9178 The previous example can also be achieved with the associated built-in preset:
9180 curves=preset=vintage
9190 Use a Photoshop preset and redefine the points of the green component:
9192 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9196 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9197 and @command{gnuplot}:
9199 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9200 gnuplot -p /tmp/curves.plt
9206 Video data analysis filter.
9208 This filter shows hexadecimal pixel values of part of video.
9210 The filter accepts the following options:
9214 Set output video size.
9217 Set x offset from where to pick pixels.
9220 Set y offset from where to pick pixels.
9223 Set scope mode, can be one of the following:
9226 Draw hexadecimal pixel values with white color on black background.
9229 Draw hexadecimal pixel values with input video pixel color on black
9233 Draw hexadecimal pixel values on color background picked from input video,
9234 the text color is picked in such way so its always visible.
9238 Draw rows and columns numbers on left and top of video.
9241 Set background opacity.
9244 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9247 Set pixel components to display. By default all pixel components are displayed.
9251 Apply Directional blur filter.
9253 The filter accepts the following options:
9257 Set angle of directional blur. Default is @code{45}.
9260 Set radius of directional blur. Default is @code{5}.
9263 Set which planes to filter. By default all planes are filtered.
9266 @subsection Commands
9267 This filter supports same @ref{commands} as options.
9268 The command accepts the same syntax of the corresponding option.
9270 If the specified expression is not valid, it is kept at its current
9275 Denoise frames using 2D DCT (frequency domain filtering).
9277 This filter is not designed for real time.
9279 The filter accepts the following options:
9283 Set the noise sigma constant.
9285 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9286 coefficient (absolute value) below this threshold with be dropped.
9288 If you need a more advanced filtering, see @option{expr}.
9290 Default is @code{0}.
9293 Set number overlapping pixels for each block. Since the filter can be slow, you
9294 may want to reduce this value, at the cost of a less effective filter and the
9295 risk of various artefacts.
9297 If the overlapping value doesn't permit processing the whole input width or
9298 height, a warning will be displayed and according borders won't be denoised.
9300 Default value is @var{blocksize}-1, which is the best possible setting.
9303 Set the coefficient factor expression.
9305 For each coefficient of a DCT block, this expression will be evaluated as a
9306 multiplier value for the coefficient.
9308 If this is option is set, the @option{sigma} option will be ignored.
9310 The absolute value of the coefficient can be accessed through the @var{c}
9314 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9315 @var{blocksize}, which is the width and height of the processed blocks.
9317 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9318 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9319 on the speed processing. Also, a larger block size does not necessarily means a
9323 @subsection Examples
9325 Apply a denoise with a @option{sigma} of @code{4.5}:
9330 The same operation can be achieved using the expression system:
9332 dctdnoiz=e='gte(c, 4.5*3)'
9335 Violent denoise using a block size of @code{16x16}:
9342 Remove banding artifacts from input video.
9343 It works by replacing banded pixels with average value of referenced pixels.
9345 The filter accepts the following options:
9352 Set banding detection threshold for each plane. Default is 0.02.
9353 Valid range is 0.00003 to 0.5.
9354 If difference between current pixel and reference pixel is less than threshold,
9355 it will be considered as banded.
9358 Banding detection range in pixels. Default is 16. If positive, random number
9359 in range 0 to set value will be used. If negative, exact absolute value
9361 The range defines square of four pixels around current pixel.
9364 Set direction in radians from which four pixel will be compared. If positive,
9365 random direction from 0 to set direction will be picked. If negative, exact of
9366 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9367 will pick only pixels on same row and -PI/2 will pick only pixels on same
9371 If enabled, current pixel is compared with average value of all four
9372 surrounding pixels. The default is enabled. If disabled current pixel is
9373 compared with all four surrounding pixels. The pixel is considered banded
9374 if only all four differences with surrounding pixels are less than threshold.
9377 If enabled, current pixel is changed if and only if all pixel components are banded,
9378 e.g. banding detection threshold is triggered for all color components.
9379 The default is disabled.
9384 Remove blocking artifacts from input video.
9386 The filter accepts the following options:
9390 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9391 This controls what kind of deblocking is applied.
9394 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9400 Set blocking detection thresholds. Allowed range is 0 to 1.
9401 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9402 Using higher threshold gives more deblocking strength.
9403 Setting @var{alpha} controls threshold detection at exact edge of block.
9404 Remaining options controls threshold detection near the edge. Each one for
9405 below/above or left/right. Setting any of those to @var{0} disables
9409 Set planes to filter. Default is to filter all available planes.
9412 @subsection Examples
9416 Deblock using weak filter and block size of 4 pixels.
9418 deblock=filter=weak:block=4
9422 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9423 deblocking more edges.
9425 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9429 Similar as above, but filter only first plane.
9431 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9435 Similar as above, but filter only second and third plane.
9437 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9444 Drop duplicated frames at regular intervals.
9446 The filter accepts the following options:
9450 Set the number of frames from which one will be dropped. Setting this to
9451 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9452 Default is @code{5}.
9455 Set the threshold for duplicate detection. If the difference metric for a frame
9456 is less than or equal to this value, then it is declared as duplicate. Default
9460 Set scene change threshold. Default is @code{15}.
9464 Set the size of the x and y-axis blocks used during metric calculations.
9465 Larger blocks give better noise suppression, but also give worse detection of
9466 small movements. Must be a power of two. Default is @code{32}.
9469 Mark main input as a pre-processed input and activate clean source input
9470 stream. This allows the input to be pre-processed with various filters to help
9471 the metrics calculation while keeping the frame selection lossless. When set to
9472 @code{1}, the first stream is for the pre-processed input, and the second
9473 stream is the clean source from where the kept frames are chosen. Default is
9477 Set whether or not chroma is considered in the metric calculations. Default is
9483 Apply 2D deconvolution of video stream in frequency domain using second stream
9486 The filter accepts the following options:
9490 Set which planes to process.
9493 Set which impulse video frames will be processed, can be @var{first}
9494 or @var{all}. Default is @var{all}.
9497 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9498 and height are not same and not power of 2 or if stream prior to convolving
9502 The @code{deconvolve} filter also supports the @ref{framesync} options.
9506 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9508 It accepts the following options:
9512 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9513 @var{rainbows} for cross-color reduction.
9516 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9519 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9522 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9525 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9530 Apply deflate effect to the video.
9532 This filter replaces the pixel by the local(3x3) average by taking into account
9533 only values lower than the pixel.
9535 It accepts the following options:
9542 Limit the maximum change for each plane, default is 65535.
9543 If 0, plane will remain unchanged.
9546 @subsection Commands
9548 This filter supports the all above options as @ref{commands}.
9552 Remove temporal frame luminance variations.
9554 It accepts the following options:
9558 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9561 Set averaging mode to smooth temporal luminance variations.
9563 Available values are:
9588 Do not actually modify frame. Useful when one only wants metadata.
9593 Remove judder produced by partially interlaced telecined content.
9595 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9596 source was partially telecined content then the output of @code{pullup,dejudder}
9597 will have a variable frame rate. May change the recorded frame rate of the
9598 container. Aside from that change, this filter will not affect constant frame
9601 The option available in this filter is:
9605 Specify the length of the window over which the judder repeats.
9607 Accepts any integer greater than 1. Useful values are:
9611 If the original was telecined from 24 to 30 fps (Film to NTSC).
9614 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9617 If a mixture of the two.
9620 The default is @samp{4}.
9625 Suppress a TV station logo by a simple interpolation of the surrounding
9626 pixels. Just set a rectangle covering the logo and watch it disappear
9627 (and sometimes something even uglier appear - your mileage may vary).
9629 It accepts the following parameters:
9634 Specify the top left corner coordinates of the logo. They must be
9639 Specify the width and height of the logo to clear. They must be
9643 Specify the thickness of the fuzzy edge of the rectangle (added to
9644 @var{w} and @var{h}). The default value is 1. This option is
9645 deprecated, setting higher values should no longer be necessary and
9649 When set to 1, a green rectangle is drawn on the screen to simplify
9650 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9651 The default value is 0.
9653 The rectangle is drawn on the outermost pixels which will be (partly)
9654 replaced with interpolated values. The values of the next pixels
9655 immediately outside this rectangle in each direction will be used to
9656 compute the interpolated pixel values inside the rectangle.
9660 @subsection Examples
9664 Set a rectangle covering the area with top left corner coordinates 0,0
9665 and size 100x77, and a band of size 10:
9667 delogo=x=0:y=0:w=100:h=77:band=10
9675 Remove the rain in the input image/video by applying the derain methods based on
9676 convolutional neural networks. Supported models:
9680 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9681 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9684 Training as well as model generation scripts are provided in
9685 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9687 Native model files (.model) can be generated from TensorFlow model
9688 files (.pb) by using tools/python/convert.py
9690 The filter accepts the following options:
9694 Specify which filter to use. This option accepts the following values:
9698 Derain filter. To conduct derain filter, you need to use a derain model.
9701 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9703 Default value is @samp{derain}.
9706 Specify which DNN backend to use for model loading and execution. This option accepts
9707 the following values:
9711 Native implementation of DNN loading and execution.
9714 TensorFlow backend. To enable this backend you
9715 need to install the TensorFlow for C library (see
9716 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9717 @code{--enable-libtensorflow}
9719 Default value is @samp{native}.
9722 Set path to model file specifying network architecture and its parameters.
9723 Note that different backends use different file formats. TensorFlow and native
9724 backend can load files for only its format.
9727 It can also be finished with @ref{dnn_processing} filter.
9731 Attempt to fix small changes in horizontal and/or vertical shift. This
9732 filter helps remove camera shake from hand-holding a camera, bumping a
9733 tripod, moving on a vehicle, etc.
9735 The filter accepts the following options:
9743 Specify a rectangular area where to limit the search for motion
9745 If desired the search for motion vectors can be limited to a
9746 rectangular area of the frame defined by its top left corner, width
9747 and height. These parameters have the same meaning as the drawbox
9748 filter which can be used to visualise the position of the bounding
9751 This is useful when simultaneous movement of subjects within the frame
9752 might be confused for camera motion by the motion vector search.
9754 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9755 then the full frame is used. This allows later options to be set
9756 without specifying the bounding box for the motion vector search.
9758 Default - search the whole frame.
9762 Specify the maximum extent of movement in x and y directions in the
9763 range 0-64 pixels. Default 16.
9766 Specify how to generate pixels to fill blanks at the edge of the
9767 frame. Available values are:
9770 Fill zeroes at blank locations
9772 Original image at blank locations
9774 Extruded edge value at blank locations
9776 Mirrored edge at blank locations
9778 Default value is @samp{mirror}.
9781 Specify the blocksize to use for motion search. Range 4-128 pixels,
9785 Specify the contrast threshold for blocks. Only blocks with more than
9786 the specified contrast (difference between darkest and lightest
9787 pixels) will be considered. Range 1-255, default 125.
9790 Specify the search strategy. Available values are:
9793 Set exhaustive search
9795 Set less exhaustive search.
9797 Default value is @samp{exhaustive}.
9800 If set then a detailed log of the motion search is written to the
9807 Remove unwanted contamination of foreground colors, caused by reflected color of
9808 greenscreen or bluescreen.
9810 This filter accepts the following options:
9814 Set what type of despill to use.
9817 Set how spillmap will be generated.
9820 Set how much to get rid of still remaining spill.
9823 Controls amount of red in spill area.
9826 Controls amount of green in spill area.
9827 Should be -1 for greenscreen.
9830 Controls amount of blue in spill area.
9831 Should be -1 for bluescreen.
9834 Controls brightness of spill area, preserving colors.
9837 Modify alpha from generated spillmap.
9840 @subsection Commands
9842 This filter supports the all above options as @ref{commands}.
9846 Apply an exact inverse of the telecine operation. It requires a predefined
9847 pattern specified using the pattern option which must be the same as that passed
9848 to the telecine filter.
9850 This filter accepts the following options:
9859 The default value is @code{top}.
9863 A string of numbers representing the pulldown pattern you wish to apply.
9864 The default value is @code{23}.
9867 A number representing position of the first frame with respect to the telecine
9868 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9873 Apply dilation effect to the video.
9875 This filter replaces the pixel by the local(3x3) maximum.
9877 It accepts the following options:
9884 Limit the maximum change for each plane, default is 65535.
9885 If 0, plane will remain unchanged.
9888 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9891 Flags to local 3x3 coordinates maps like this:
9898 @subsection Commands
9900 This filter supports the all above options as @ref{commands}.
9904 Displace pixels as indicated by second and third input stream.
9906 It takes three input streams and outputs one stream, the first input is the
9907 source, and second and third input are displacement maps.
9909 The second input specifies how much to displace pixels along the
9910 x-axis, while the third input specifies how much to displace pixels
9912 If one of displacement map streams terminates, last frame from that
9913 displacement map will be used.
9915 Note that once generated, displacements maps can be reused over and over again.
9917 A description of the accepted options follows.
9921 Set displace behavior for pixels that are out of range.
9923 Available values are:
9926 Missing pixels are replaced by black pixels.
9929 Adjacent pixels will spread out to replace missing pixels.
9932 Out of range pixels are wrapped so they point to pixels of other side.
9935 Out of range pixels will be replaced with mirrored pixels.
9937 Default is @samp{smear}.
9941 @subsection Examples
9945 Add ripple effect to rgb input of video size hd720:
9947 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
9951 Add wave effect to rgb input of video size hd720:
9953 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
9957 @anchor{dnn_processing}
9958 @section dnn_processing
9960 Do image processing with deep neural networks. It works together with another filter
9961 which converts the pixel format of the Frame to what the dnn network requires.
9963 The filter accepts the following options:
9967 Specify which DNN backend to use for model loading and execution. This option accepts
9968 the following values:
9972 Native implementation of DNN loading and execution.
9975 TensorFlow backend. To enable this backend you
9976 need to install the TensorFlow for C library (see
9977 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9978 @code{--enable-libtensorflow}
9981 OpenVINO backend. To enable this backend you
9982 need to build and install the OpenVINO for C library (see
9983 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9984 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9985 be needed if the header files and libraries are not installed into system path)
9989 Default value is @samp{native}.
9992 Set path to model file specifying network architecture and its parameters.
9993 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9994 backend can load files for only its format.
9996 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9999 Set the input name of the dnn network.
10002 Set the output name of the dnn network.
10005 use DNN async execution if set (default: set),
10006 roll back to sync execution if the backend does not support async.
10010 @subsection Examples
10014 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
10016 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
10020 Halve the pixel value of the frame with format gray32f:
10022 ffmpeg -i input.jpg -vf format=grayf32,dnn_processing=model=halve_gray_float.model:input=dnn_in:output=dnn_out:dnn_backend=native -y out.native.png
10026 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
10028 ./ffmpeg -i 480p.jpg -vf format=yuv420p,scale=w=iw*2:h=ih*2,dnn_processing=dnn_backend=tensorflow:model=srcnn.pb:input=x:output=y -y srcnn.jpg
10032 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
10034 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
10041 Draw a colored box on the input image.
10043 It accepts the following parameters:
10048 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
10052 The expressions which specify the width and height of the box; if 0 they are interpreted as
10053 the input width and height. It defaults to 0.
10056 Specify the color of the box to write. For the general syntax of this option,
10057 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10058 value @code{invert} is used, the box edge color is the same as the
10059 video with inverted luma.
10062 The expression which sets the thickness of the box edge.
10063 A value of @code{fill} will create a filled box. Default value is @code{3}.
10065 See below for the list of accepted constants.
10068 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10069 will overwrite the video's color and alpha pixels.
10070 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10073 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10074 following constants:
10078 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10082 horizontal and vertical chroma subsample values. For example for the
10083 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10087 The input width and height.
10090 The input sample aspect ratio.
10094 The x and y offset coordinates where the box is drawn.
10098 The width and height of the drawn box.
10101 The thickness of the drawn box.
10103 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10104 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10108 @subsection Examples
10112 Draw a black box around the edge of the input image:
10118 Draw a box with color red and an opacity of 50%:
10120 drawbox=10:20:200:60:red@@0.5
10123 The previous example can be specified as:
10125 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10129 Fill the box with pink color:
10131 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10135 Draw a 2-pixel red 2.40:1 mask:
10137 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
10141 @subsection Commands
10142 This filter supports same commands as options.
10143 The command accepts the same syntax of the corresponding option.
10145 If the specified expression is not valid, it is kept at its current
10150 Draw a graph using input video metadata.
10152 It accepts the following parameters:
10156 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10159 Set 1st foreground color expression.
10162 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10165 Set 2nd foreground color expression.
10168 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10171 Set 3rd foreground color expression.
10174 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10177 Set 4th foreground color expression.
10180 Set minimal value of metadata value.
10183 Set maximal value of metadata value.
10186 Set graph background color. Default is white.
10191 Available values for mode is:
10198 Default is @code{line}.
10203 Available values for slide is:
10206 Draw new frame when right border is reached.
10209 Replace old columns with new ones.
10212 Scroll from right to left.
10215 Scroll from left to right.
10218 Draw single picture.
10221 Default is @code{frame}.
10224 Set size of graph video. For the syntax of this option, check the
10225 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10226 The default value is @code{900x256}.
10229 Set the output frame rate. Default value is @code{25}.
10231 The foreground color expressions can use the following variables:
10234 Minimal value of metadata value.
10237 Maximal value of metadata value.
10240 Current metadata key value.
10243 The color is defined as 0xAABBGGRR.
10246 Example using metadata from @ref{signalstats} filter:
10248 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10251 Example using metadata from @ref{ebur128} filter:
10253 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10258 Draw a grid on the input image.
10260 It accepts the following parameters:
10265 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10269 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10270 input width and height, respectively, minus @code{thickness}, so image gets
10271 framed. Default to 0.
10274 Specify the color of the grid. For the general syntax of this option,
10275 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10276 value @code{invert} is used, the grid color is the same as the
10277 video with inverted luma.
10280 The expression which sets the thickness of the grid line. Default value is @code{1}.
10282 See below for the list of accepted constants.
10285 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10286 will overwrite the video's color and alpha pixels.
10287 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10290 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10291 following constants:
10295 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10299 horizontal and vertical chroma subsample values. For example for the
10300 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10304 The input grid cell width and height.
10307 The input sample aspect ratio.
10311 The x and y coordinates of some point of grid intersection (meant to configure offset).
10315 The width and height of the drawn cell.
10318 The thickness of the drawn cell.
10320 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10321 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10325 @subsection Examples
10329 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10331 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10335 Draw a white 3x3 grid with an opacity of 50%:
10337 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10341 @subsection Commands
10342 This filter supports same commands as options.
10343 The command accepts the same syntax of the corresponding option.
10345 If the specified expression is not valid, it is kept at its current
10351 Draw a text string or text from a specified file on top of a video, using the
10352 libfreetype library.
10354 To enable compilation of this filter, you need to configure FFmpeg with
10355 @code{--enable-libfreetype}.
10356 To enable default font fallback and the @var{font} option you need to
10357 configure FFmpeg with @code{--enable-libfontconfig}.
10358 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10359 @code{--enable-libfribidi}.
10363 It accepts the following parameters:
10368 Used to draw a box around text using the background color.
10369 The value must be either 1 (enable) or 0 (disable).
10370 The default value of @var{box} is 0.
10373 Set the width of the border to be drawn around the box using @var{boxcolor}.
10374 The default value of @var{boxborderw} is 0.
10377 The color to be used for drawing box around text. For the syntax of this
10378 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10380 The default value of @var{boxcolor} is "white".
10383 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10384 The default value of @var{line_spacing} is 0.
10387 Set the width of the border to be drawn around the text using @var{bordercolor}.
10388 The default value of @var{borderw} is 0.
10391 Set the color to be used for drawing border around text. For the syntax of this
10392 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10394 The default value of @var{bordercolor} is "black".
10397 Select how the @var{text} is expanded. Can be either @code{none},
10398 @code{strftime} (deprecated) or
10399 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10403 Set a start time for the count. Value is in microseconds. Only applied
10404 in the deprecated strftime expansion mode. To emulate in normal expansion
10405 mode use the @code{pts} function, supplying the start time (in seconds)
10406 as the second argument.
10409 If true, check and fix text coords to avoid clipping.
10412 The color to be used for drawing fonts. For the syntax of this option, check
10413 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10415 The default value of @var{fontcolor} is "black".
10417 @item fontcolor_expr
10418 String which is expanded the same way as @var{text} to obtain dynamic
10419 @var{fontcolor} value. By default this option has empty value and is not
10420 processed. When this option is set, it overrides @var{fontcolor} option.
10423 The font family to be used for drawing text. By default Sans.
10426 The font file to be used for drawing text. The path must be included.
10427 This parameter is mandatory if the fontconfig support is disabled.
10430 Draw the text applying alpha blending. The value can
10431 be a number between 0.0 and 1.0.
10432 The expression accepts the same variables @var{x, y} as well.
10433 The default value is 1.
10434 Please see @var{fontcolor_expr}.
10437 The font size to be used for drawing text.
10438 The default value of @var{fontsize} is 16.
10441 If set to 1, attempt to shape the text (for example, reverse the order of
10442 right-to-left text and join Arabic characters) before drawing it.
10443 Otherwise, just draw the text exactly as given.
10444 By default 1 (if supported).
10446 @item ft_load_flags
10447 The flags to be used for loading the fonts.
10449 The flags map the corresponding flags supported by libfreetype, and are
10450 a combination of the following values:
10457 @item vertical_layout
10458 @item force_autohint
10461 @item ignore_global_advance_width
10463 @item ignore_transform
10465 @item linear_design
10469 Default value is "default".
10471 For more information consult the documentation for the FT_LOAD_*
10475 The color to be used for drawing a shadow behind the drawn text. For the
10476 syntax of this option, check the @ref{color syntax,,"Color" section in the
10477 ffmpeg-utils manual,ffmpeg-utils}.
10479 The default value of @var{shadowcolor} is "black".
10483 The x and y offsets for the text shadow position with respect to the
10484 position of the text. They can be either positive or negative
10485 values. The default value for both is "0".
10488 The starting frame number for the n/frame_num variable. The default value
10492 The size in number of spaces to use for rendering the tab.
10493 Default value is 4.
10496 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10497 format. It can be used with or without text parameter. @var{timecode_rate}
10498 option must be specified.
10500 @item timecode_rate, rate, r
10501 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10502 integer. Minimum value is "1".
10503 Drop-frame timecode is supported for frame rates 30 & 60.
10506 If set to 1, the output of the timecode option will wrap around at 24 hours.
10507 Default is 0 (disabled).
10510 The text string to be drawn. The text must be a sequence of UTF-8
10511 encoded characters.
10512 This parameter is mandatory if no file is specified with the parameter
10516 A text file containing text to be drawn. The text must be a sequence
10517 of UTF-8 encoded characters.
10519 This parameter is mandatory if no text string is specified with the
10520 parameter @var{text}.
10522 If both @var{text} and @var{textfile} are specified, an error is thrown.
10525 If set to 1, the @var{textfile} will be reloaded before each frame.
10526 Be sure to update it atomically, or it may be read partially, or even fail.
10530 The expressions which specify the offsets where text will be drawn
10531 within the video frame. They are relative to the top/left border of the
10534 The default value of @var{x} and @var{y} is "0".
10536 See below for the list of accepted constants and functions.
10539 The parameters for @var{x} and @var{y} are expressions containing the
10540 following constants and functions:
10544 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10548 horizontal and vertical chroma subsample values. For example for the
10549 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10552 the height of each text line
10560 @item max_glyph_a, ascent
10561 the maximum distance from the baseline to the highest/upper grid
10562 coordinate used to place a glyph outline point, for all the rendered
10564 It is a positive value, due to the grid's orientation with the Y axis
10567 @item max_glyph_d, descent
10568 the maximum distance from the baseline to the lowest grid coordinate
10569 used to place a glyph outline point, for all the rendered glyphs.
10570 This is a negative value, due to the grid's orientation, with the Y axis
10574 maximum glyph height, that is the maximum height for all the glyphs
10575 contained in the rendered text, it is equivalent to @var{ascent} -
10579 maximum glyph width, that is the maximum width for all the glyphs
10580 contained in the rendered text
10583 the number of input frame, starting from 0
10585 @item rand(min, max)
10586 return a random number included between @var{min} and @var{max}
10589 The input sample aspect ratio.
10592 timestamp expressed in seconds, NAN if the input timestamp is unknown
10595 the height of the rendered text
10598 the width of the rendered text
10602 the x and y offset coordinates where the text is drawn.
10604 These parameters allow the @var{x} and @var{y} expressions to refer
10605 to each other, so you can for example specify @code{y=x/dar}.
10608 A one character description of the current frame's picture type.
10611 The current packet's position in the input file or stream
10612 (in bytes, from the start of the input). A value of -1 indicates
10613 this info is not available.
10616 The current packet's duration, in seconds.
10619 The current packet's size (in bytes).
10622 @anchor{drawtext_expansion}
10623 @subsection Text expansion
10625 If @option{expansion} is set to @code{strftime},
10626 the filter recognizes strftime() sequences in the provided text and
10627 expands them accordingly. Check the documentation of strftime(). This
10628 feature is deprecated.
10630 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10632 If @option{expansion} is set to @code{normal} (which is the default),
10633 the following expansion mechanism is used.
10635 The backslash character @samp{\}, followed by any character, always expands to
10636 the second character.
10638 Sequences of the form @code{%@{...@}} are expanded. The text between the
10639 braces is a function name, possibly followed by arguments separated by ':'.
10640 If the arguments contain special characters or delimiters (':' or '@}'),
10641 they should be escaped.
10643 Note that they probably must also be escaped as the value for the
10644 @option{text} option in the filter argument string and as the filter
10645 argument in the filtergraph description, and possibly also for the shell,
10646 that makes up to four levels of escaping; using a text file avoids these
10649 The following functions are available:
10654 The expression evaluation result.
10656 It must take one argument specifying the expression to be evaluated,
10657 which accepts the same constants and functions as the @var{x} and
10658 @var{y} values. Note that not all constants should be used, for
10659 example the text size is not known when evaluating the expression, so
10660 the constants @var{text_w} and @var{text_h} will have an undefined
10663 @item expr_int_format, eif
10664 Evaluate the expression's value and output as formatted integer.
10666 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10667 The second argument specifies the output format. Allowed values are @samp{x},
10668 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10669 @code{printf} function.
10670 The third parameter is optional and sets the number of positions taken by the output.
10671 It can be used to add padding with zeros from the left.
10674 The time at which the filter is running, expressed in UTC.
10675 It can accept an argument: a strftime() format string.
10678 The time at which the filter is running, expressed in the local time zone.
10679 It can accept an argument: a strftime() format string.
10682 Frame metadata. Takes one or two arguments.
10684 The first argument is mandatory and specifies the metadata key.
10686 The second argument is optional and specifies a default value, used when the
10687 metadata key is not found or empty.
10689 Available metadata can be identified by inspecting entries
10690 starting with TAG included within each frame section
10691 printed by running @code{ffprobe -show_frames}.
10693 String metadata generated in filters leading to
10694 the drawtext filter are also available.
10697 The frame number, starting from 0.
10700 A one character description of the current picture type.
10703 The timestamp of the current frame.
10704 It can take up to three arguments.
10706 The first argument is the format of the timestamp; it defaults to @code{flt}
10707 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10708 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10709 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10710 @code{localtime} stands for the timestamp of the frame formatted as
10711 local time zone time.
10713 The second argument is an offset added to the timestamp.
10715 If the format is set to @code{hms}, a third argument @code{24HH} may be
10716 supplied to present the hour part of the formatted timestamp in 24h format
10719 If the format is set to @code{localtime} or @code{gmtime},
10720 a third argument may be supplied: a strftime() format string.
10721 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10724 @subsection Commands
10726 This filter supports altering parameters via commands:
10729 Alter existing filter parameters.
10731 Syntax for the argument is the same as for filter invocation, e.g.
10734 fontsize=56:fontcolor=green:text='Hello World'
10737 Full filter invocation with sendcmd would look like this:
10740 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10744 If the entire argument can't be parsed or applied as valid values then the filter will
10745 continue with its existing parameters.
10747 @subsection Examples
10751 Draw "Test Text" with font FreeSerif, using the default values for the
10752 optional parameters.
10755 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10759 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10760 and y=50 (counting from the top-left corner of the screen), text is
10761 yellow with a red box around it. Both the text and the box have an
10765 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10766 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10769 Note that the double quotes are not necessary if spaces are not used
10770 within the parameter list.
10773 Show the text at the center of the video frame:
10775 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10779 Show the text at a random position, switching to a new position every 30 seconds:
10781 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)"
10785 Show a text line sliding from right to left in the last row of the video
10786 frame. The file @file{LONG_LINE} is assumed to contain a single line
10789 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10793 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10795 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10799 Draw a single green letter "g", at the center of the input video.
10800 The glyph baseline is placed at half screen height.
10802 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10806 Show text for 1 second every 3 seconds:
10808 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10812 Use fontconfig to set the font. Note that the colons need to be escaped.
10814 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10818 Draw "Test Text" with font size dependent on height of the video.
10820 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10824 Print the date of a real-time encoding (see strftime(3)):
10826 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10830 Show text fading in and out (appearing/disappearing):
10833 DS=1.0 # display start
10834 DE=10.0 # display end
10835 FID=1.5 # fade in duration
10836 FOD=5 # fade out duration
10837 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 @}"
10841 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10842 and the @option{fontsize} value are included in the @option{y} offset.
10844 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10845 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10849 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10850 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10851 must have option @option{-export_path_metadata 1} for the special metadata fields
10852 to be available for filters.
10854 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10859 For more information about libfreetype, check:
10860 @url{http://www.freetype.org/}.
10862 For more information about fontconfig, check:
10863 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10865 For more information about libfribidi, check:
10866 @url{http://fribidi.org/}.
10868 @section edgedetect
10870 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10872 The filter accepts the following options:
10877 Set low and high threshold values used by the Canny thresholding
10880 The high threshold selects the "strong" edge pixels, which are then
10881 connected through 8-connectivity with the "weak" edge pixels selected
10882 by the low threshold.
10884 @var{low} and @var{high} threshold values must be chosen in the range
10885 [0,1], and @var{low} should be lesser or equal to @var{high}.
10887 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10891 Define the drawing mode.
10895 Draw white/gray wires on black background.
10898 Mix the colors to create a paint/cartoon effect.
10901 Apply Canny edge detector on all selected planes.
10903 Default value is @var{wires}.
10906 Select planes for filtering. By default all available planes are filtered.
10909 @subsection Examples
10913 Standard edge detection with custom values for the hysteresis thresholding:
10915 edgedetect=low=0.1:high=0.4
10919 Painting effect without thresholding:
10921 edgedetect=mode=colormix:high=0
10927 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10929 For each input image, the filter will compute the optimal mapping from
10930 the input to the output given the codebook length, that is the number
10931 of distinct output colors.
10933 This filter accepts the following options.
10936 @item codebook_length, l
10937 Set codebook length. The value must be a positive integer, and
10938 represents the number of distinct output colors. Default value is 256.
10941 Set the maximum number of iterations to apply for computing the optimal
10942 mapping. The higher the value the better the result and the higher the
10943 computation time. Default value is 1.
10946 Set a random seed, must be an integer included between 0 and
10947 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10948 will try to use a good random seed on a best effort basis.
10951 Set pal8 output pixel format. This option does not work with codebook
10952 length greater than 256.
10957 Measure graylevel entropy in histogram of color channels of video frames.
10959 It accepts the following parameters:
10963 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10965 @var{diff} mode measures entropy of histogram delta values, absolute differences
10966 between neighbour histogram values.
10970 Apply the EPX magnification filter which is designed for pixel art.
10972 It accepts the following option:
10976 Set the scaling dimension: @code{2} for @code{2xEPX}, @code{3} for
10978 Default is @code{3}.
10982 Set brightness, contrast, saturation and approximate gamma adjustment.
10984 The filter accepts the following options:
10988 Set the contrast expression. The value must be a float value in range
10989 @code{-1000.0} to @code{1000.0}. The default value is "1".
10992 Set the brightness expression. The value must be a float value in
10993 range @code{-1.0} to @code{1.0}. The default value is "0".
10996 Set the saturation expression. The value must be a float in
10997 range @code{0.0} to @code{3.0}. The default value is "1".
11000 Set the gamma expression. The value must be a float in range
11001 @code{0.1} to @code{10.0}. The default value is "1".
11004 Set the gamma expression for red. The value must be a float in
11005 range @code{0.1} to @code{10.0}. The default value is "1".
11008 Set the gamma expression for green. The value must be a float in range
11009 @code{0.1} to @code{10.0}. The default value is "1".
11012 Set the gamma expression for blue. The value must be a float in range
11013 @code{0.1} to @code{10.0}. The default value is "1".
11016 Set the gamma weight expression. It can be used to reduce the effect
11017 of a high gamma value on bright image areas, e.g. keep them from
11018 getting overamplified and just plain white. The value must be a float
11019 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
11020 gamma correction all the way down while @code{1.0} leaves it at its
11021 full strength. Default is "1".
11024 Set when the expressions for brightness, contrast, saturation and
11025 gamma expressions are evaluated.
11027 It accepts the following values:
11030 only evaluate expressions once during the filter initialization or
11031 when a command is processed
11034 evaluate expressions for each incoming frame
11037 Default value is @samp{init}.
11040 The expressions accept the following parameters:
11043 frame count of the input frame starting from 0
11046 byte position of the corresponding packet in the input file, NAN if
11050 frame rate of the input video, NAN if the input frame rate is unknown
11053 timestamp expressed in seconds, NAN if the input timestamp is unknown
11056 @subsection Commands
11057 The filter supports the following commands:
11061 Set the contrast expression.
11064 Set the brightness expression.
11067 Set the saturation expression.
11070 Set the gamma expression.
11073 Set the gamma_r expression.
11076 Set gamma_g expression.
11079 Set gamma_b expression.
11082 Set gamma_weight expression.
11084 The command accepts the same syntax of the corresponding option.
11086 If the specified expression is not valid, it is kept at its current
11093 Apply erosion effect to the video.
11095 This filter replaces the pixel by the local(3x3) minimum.
11097 It accepts the following options:
11104 Limit the maximum change for each plane, default is 65535.
11105 If 0, plane will remain unchanged.
11108 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11111 Flags to local 3x3 coordinates maps like this:
11118 @subsection Commands
11120 This filter supports the all above options as @ref{commands}.
11124 Deinterlace the input video ("estdif" stands for "Edge Slope
11125 Tracing Deinterlacing Filter").
11127 Spatial only filter that uses edge slope tracing algorithm
11128 to interpolate missing lines.
11129 It accepts the following parameters:
11133 The interlacing mode to adopt. It accepts one of the following values:
11137 Output one frame for each frame.
11139 Output one frame for each field.
11142 The default value is @code{field}.
11145 The picture field parity assumed for the input interlaced video. It accepts one
11146 of the following values:
11150 Assume the top field is first.
11152 Assume the bottom field is first.
11154 Enable automatic detection of field parity.
11157 The default value is @code{auto}.
11158 If the interlacing is unknown or the decoder does not export this information,
11159 top field first will be assumed.
11162 Specify which frames to deinterlace. Accepts one of the following
11167 Deinterlace all frames.
11169 Only deinterlace frames marked as interlaced.
11172 The default value is @code{all}.
11175 Specify the search radius for edge slope tracing. Default value is 1.
11176 Allowed range is from 1 to 15.
11179 Specify the search radius for best edge matching. Default value is 2.
11180 Allowed range is from 0 to 15.
11183 Specify the interpolation used. Default is 4-point interpolation. It accepts one
11184 of the following values:
11188 Two-point interpolation.
11190 Four-point interpolation.
11192 Six-point interpolation.
11196 @subsection Commands
11197 This filter supports same @ref{commands} as options.
11199 @section extractplanes
11201 Extract color channel components from input video stream into
11202 separate grayscale video streams.
11204 The filter accepts the following option:
11208 Set plane(s) to extract.
11210 Available values for planes are:
11221 Choosing planes not available in the input will result in an error.
11222 That means you cannot select @code{r}, @code{g}, @code{b} planes
11223 with @code{y}, @code{u}, @code{v} planes at same time.
11226 @subsection Examples
11230 Extract luma, u and v color channel component from input video frame
11231 into 3 grayscale outputs:
11233 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
11239 Apply a fade-in/out effect to the input video.
11241 It accepts the following parameters:
11245 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11247 Default is @code{in}.
11249 @item start_frame, s
11250 Specify the number of the frame to start applying the fade
11251 effect at. Default is 0.
11254 The number of frames that the fade effect lasts. At the end of the
11255 fade-in effect, the output video will have the same intensity as the input video.
11256 At the end of the fade-out transition, the output video will be filled with the
11257 selected @option{color}.
11261 If set to 1, fade only alpha channel, if one exists on the input.
11262 Default value is 0.
11264 @item start_time, st
11265 Specify the timestamp (in seconds) of the frame to start to apply the fade
11266 effect. If both start_frame and start_time are specified, the fade will start at
11267 whichever comes last. Default is 0.
11270 The number of seconds for which the fade effect has to last. At the end of the
11271 fade-in effect the output video will have the same intensity as the input video,
11272 at the end of the fade-out transition the output video will be filled with the
11273 selected @option{color}.
11274 If both duration and nb_frames are specified, duration is used. Default is 0
11275 (nb_frames is used by default).
11278 Specify the color of the fade. Default is "black".
11281 @subsection Examples
11285 Fade in the first 30 frames of video:
11290 The command above is equivalent to:
11296 Fade out the last 45 frames of a 200-frame video:
11299 fade=type=out:start_frame=155:nb_frames=45
11303 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11305 fade=in:0:25, fade=out:975:25
11309 Make the first 5 frames yellow, then fade in from frame 5-24:
11311 fade=in:5:20:color=yellow
11315 Fade in alpha over first 25 frames of video:
11317 fade=in:0:25:alpha=1
11321 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11323 fade=t=in:st=5.5:d=0.5
11329 Denoise frames using 3D FFT (frequency domain filtering).
11331 The filter accepts the following options:
11335 Set the noise sigma constant. This sets denoising strength.
11336 Default value is 1. Allowed range is from 0 to 30.
11337 Using very high sigma with low overlap may give blocking artifacts.
11340 Set amount of denoising. By default all detected noise is reduced.
11341 Default value is 1. Allowed range is from 0 to 1.
11344 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11345 Actual size of block in pixels is 2 to power of @var{block}, so by default
11346 block size in pixels is 2^4 which is 16.
11349 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11352 Set number of previous frames to use for denoising. By default is set to 0.
11355 Set number of next frames to to use for denoising. By default is set to 0.
11358 Set planes which will be filtered, by default are all available filtered
11363 Apply arbitrary expressions to samples in frequency domain
11367 Adjust the dc value (gain) of the luma plane of the image. The filter
11368 accepts an integer value in range @code{0} to @code{1000}. The default
11369 value is set to @code{0}.
11372 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11373 filter accepts an integer value in range @code{0} to @code{1000}. The
11374 default value is set to @code{0}.
11377 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11378 filter accepts an integer value in range @code{0} to @code{1000}. The
11379 default value is set to @code{0}.
11382 Set the frequency domain weight expression for the luma plane.
11385 Set the frequency domain weight expression for the 1st chroma plane.
11388 Set the frequency domain weight expression for the 2nd chroma plane.
11391 Set when the expressions are evaluated.
11393 It accepts the following values:
11396 Only evaluate expressions once during the filter initialization.
11399 Evaluate expressions for each incoming frame.
11402 Default value is @samp{init}.
11404 The filter accepts the following variables:
11407 The coordinates of the current sample.
11411 The width and height of the image.
11414 The number of input frame, starting from 0.
11417 @subsection Examples
11423 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11429 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11435 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11441 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11448 Extract a single field from an interlaced image using stride
11449 arithmetic to avoid wasting CPU time. The output frames are marked as
11452 The filter accepts the following options:
11456 Specify whether to extract the top (if the value is @code{0} or
11457 @code{top}) or the bottom field (if the value is @code{1} or
11463 Create new frames by copying the top and bottom fields from surrounding frames
11464 supplied as numbers by the hint file.
11468 Set file containing hints: absolute/relative frame numbers.
11470 There must be one line for each frame in a clip. Each line must contain two
11471 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11472 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11473 is current frame number for @code{absolute} mode or out of [-1, 1] range
11474 for @code{relative} mode. First number tells from which frame to pick up top
11475 field and second number tells from which frame to pick up bottom field.
11477 If optionally followed by @code{+} output frame will be marked as interlaced,
11478 else if followed by @code{-} output frame will be marked as progressive, else
11479 it will be marked same as input frame.
11480 If optionally followed by @code{t} output frame will use only top field, or in
11481 case of @code{b} it will use only bottom field.
11482 If line starts with @code{#} or @code{;} that line is skipped.
11485 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11488 Example of first several lines of @code{hint} file for @code{relative} mode:
11490 0,0 - # first frame
11491 1,0 - # second frame, use third's frame top field and second's frame bottom field
11492 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11507 @section fieldmatch
11509 Field matching filter for inverse telecine. It is meant to reconstruct the
11510 progressive frames from a telecined stream. The filter does not drop duplicated
11511 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11512 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11514 The separation of the field matching and the decimation is notably motivated by
11515 the possibility of inserting a de-interlacing filter fallback between the two.
11516 If the source has mixed telecined and real interlaced content,
11517 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11518 But these remaining combed frames will be marked as interlaced, and thus can be
11519 de-interlaced by a later filter such as @ref{yadif} before decimation.
11521 In addition to the various configuration options, @code{fieldmatch} can take an
11522 optional second stream, activated through the @option{ppsrc} option. If
11523 enabled, the frames reconstruction will be based on the fields and frames from
11524 this second stream. This allows the first input to be pre-processed in order to
11525 help the various algorithms of the filter, while keeping the output lossless
11526 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11527 or brightness/contrast adjustments can help.
11529 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11530 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11531 which @code{fieldmatch} is based on. While the semantic and usage are very
11532 close, some behaviour and options names can differ.
11534 The @ref{decimate} filter currently only works for constant frame rate input.
11535 If your input has mixed telecined (30fps) and progressive content with a lower
11536 framerate like 24fps use the following filterchain to produce the necessary cfr
11537 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11539 The filter accepts the following options:
11543 Specify the assumed field order of the input stream. Available values are:
11547 Auto detect parity (use FFmpeg's internal parity value).
11549 Assume bottom field first.
11551 Assume top field first.
11554 Note that it is sometimes recommended not to trust the parity announced by the
11557 Default value is @var{auto}.
11560 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11561 sense that it won't risk creating jerkiness due to duplicate frames when
11562 possible, but if there are bad edits or blended fields it will end up
11563 outputting combed frames when a good match might actually exist. On the other
11564 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11565 but will almost always find a good frame if there is one. The other values are
11566 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11567 jerkiness and creating duplicate frames versus finding good matches in sections
11568 with bad edits, orphaned fields, blended fields, etc.
11570 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11572 Available values are:
11576 2-way matching (p/c)
11578 2-way matching, and trying 3rd match if still combed (p/c + n)
11580 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11582 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11583 still combed (p/c + n + u/b)
11585 3-way matching (p/c/n)
11587 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11588 detected as combed (p/c/n + u/b)
11591 The parenthesis at the end indicate the matches that would be used for that
11592 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11595 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11598 Default value is @var{pc_n}.
11601 Mark the main input stream as a pre-processed input, and enable the secondary
11602 input stream as the clean source to pick the fields from. See the filter
11603 introduction for more details. It is similar to the @option{clip2} feature from
11606 Default value is @code{0} (disabled).
11609 Set the field to match from. It is recommended to set this to the same value as
11610 @option{order} unless you experience matching failures with that setting. In
11611 certain circumstances changing the field that is used to match from can have a
11612 large impact on matching performance. Available values are:
11616 Automatic (same value as @option{order}).
11618 Match from the bottom field.
11620 Match from the top field.
11623 Default value is @var{auto}.
11626 Set whether or not chroma is included during the match comparisons. In most
11627 cases it is recommended to leave this enabled. You should set this to @code{0}
11628 only if your clip has bad chroma problems such as heavy rainbowing or other
11629 artifacts. Setting this to @code{0} could also be used to speed things up at
11630 the cost of some accuracy.
11632 Default value is @code{1}.
11636 These define an exclusion band which excludes the lines between @option{y0} and
11637 @option{y1} from being included in the field matching decision. An exclusion
11638 band can be used to ignore subtitles, a logo, or other things that may
11639 interfere with the matching. @option{y0} sets the starting scan line and
11640 @option{y1} sets the ending line; all lines in between @option{y0} and
11641 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11642 @option{y0} and @option{y1} to the same value will disable the feature.
11643 @option{y0} and @option{y1} defaults to @code{0}.
11646 Set the scene change detection threshold as a percentage of maximum change on
11647 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11648 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11649 @option{scthresh} is @code{[0.0, 100.0]}.
11651 Default value is @code{12.0}.
11654 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11655 account the combed scores of matches when deciding what match to use as the
11656 final match. Available values are:
11660 No final matching based on combed scores.
11662 Combed scores are only used when a scene change is detected.
11664 Use combed scores all the time.
11667 Default is @var{sc}.
11670 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11671 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11672 Available values are:
11676 No forced calculation.
11678 Force p/c/n calculations.
11680 Force p/c/n/u/b calculations.
11683 Default value is @var{none}.
11686 This is the area combing threshold used for combed frame detection. This
11687 essentially controls how "strong" or "visible" combing must be to be detected.
11688 Larger values mean combing must be more visible and smaller values mean combing
11689 can be less visible or strong and still be detected. Valid settings are from
11690 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11691 be detected as combed). This is basically a pixel difference value. A good
11692 range is @code{[8, 12]}.
11694 Default value is @code{9}.
11697 Sets whether or not chroma is considered in the combed frame decision. Only
11698 disable this if your source has chroma problems (rainbowing, etc.) that are
11699 causing problems for the combed frame detection with chroma enabled. Actually,
11700 using @option{chroma}=@var{0} is usually more reliable, except for the case
11701 where there is chroma only combing in the source.
11703 Default value is @code{0}.
11707 Respectively set the x-axis and y-axis size of the window used during combed
11708 frame detection. This has to do with the size of the area in which
11709 @option{combpel} pixels are required to be detected as combed for a frame to be
11710 declared combed. See the @option{combpel} parameter description for more info.
11711 Possible values are any number that is a power of 2 starting at 4 and going up
11714 Default value is @code{16}.
11717 The number of combed pixels inside any of the @option{blocky} by
11718 @option{blockx} size blocks on the frame for the frame to be detected as
11719 combed. While @option{cthresh} controls how "visible" the combing must be, this
11720 setting controls "how much" combing there must be in any localized area (a
11721 window defined by the @option{blockx} and @option{blocky} settings) on the
11722 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11723 which point no frames will ever be detected as combed). This setting is known
11724 as @option{MI} in TFM/VFM vocabulary.
11726 Default value is @code{80}.
11729 @anchor{p/c/n/u/b meaning}
11730 @subsection p/c/n/u/b meaning
11732 @subsubsection p/c/n
11734 We assume the following telecined stream:
11737 Top fields: 1 2 2 3 4
11738 Bottom fields: 1 2 3 4 4
11741 The numbers correspond to the progressive frame the fields relate to. Here, the
11742 first two frames are progressive, the 3rd and 4th are combed, and so on.
11744 When @code{fieldmatch} is configured to run a matching from bottom
11745 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11750 B 1 2 3 4 4 <-- matching reference
11759 As a result of the field matching, we can see that some frames get duplicated.
11760 To perform a complete inverse telecine, you need to rely on a decimation filter
11761 after this operation. See for instance the @ref{decimate} filter.
11763 The same operation now matching from top fields (@option{field}=@var{top})
11768 T 1 2 2 3 4 <-- matching reference
11778 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11779 basically, they refer to the frame and field of the opposite parity:
11782 @item @var{p} matches the field of the opposite parity in the previous frame
11783 @item @var{c} matches the field of the opposite parity in the current frame
11784 @item @var{n} matches the field of the opposite parity in the next frame
11789 The @var{u} and @var{b} matching are a bit special in the sense that they match
11790 from the opposite parity flag. In the following examples, we assume that we are
11791 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11792 'x' is placed above and below each matched fields.
11794 With bottom matching (@option{field}=@var{bottom}):
11799 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11800 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11808 With top matching (@option{field}=@var{top}):
11813 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11814 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11822 @subsection Examples
11824 Simple IVTC of a top field first telecined stream:
11826 fieldmatch=order=tff:combmatch=none, decimate
11829 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11831 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11834 @section fieldorder
11836 Transform the field order of the input video.
11838 It accepts the following parameters:
11843 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11844 for bottom field first.
11847 The default value is @samp{tff}.
11849 The transformation is done by shifting the picture content up or down
11850 by one line, and filling the remaining line with appropriate picture content.
11851 This method is consistent with most broadcast field order converters.
11853 If the input video is not flagged as being interlaced, or it is already
11854 flagged as being of the required output field order, then this filter does
11855 not alter the incoming video.
11857 It is very useful when converting to or from PAL DV material,
11858 which is bottom field first.
11862 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11865 @section fifo, afifo
11867 Buffer input images and send them when they are requested.
11869 It is mainly useful when auto-inserted by the libavfilter
11872 It does not take parameters.
11874 @section fillborders
11876 Fill borders of the input video, without changing video stream dimensions.
11877 Sometimes video can have garbage at the four edges and you may not want to
11878 crop video input to keep size multiple of some number.
11880 This filter accepts the following options:
11884 Number of pixels to fill from left border.
11887 Number of pixels to fill from right border.
11890 Number of pixels to fill from top border.
11893 Number of pixels to fill from bottom border.
11898 It accepts the following values:
11901 fill pixels using outermost pixels
11904 fill pixels using mirroring (half sample symmetric)
11907 fill pixels with constant value
11910 fill pixels using reflecting (whole sample symmetric)
11913 fill pixels using wrapping
11916 fade pixels to constant value
11919 Default is @var{smear}.
11922 Set color for pixels in fixed or fade mode. Default is @var{black}.
11925 @subsection Commands
11926 This filter supports same @ref{commands} as options.
11927 The command accepts the same syntax of the corresponding option.
11929 If the specified expression is not valid, it is kept at its current
11934 Find a rectangular object
11936 It accepts the following options:
11940 Filepath of the object image, needs to be in gray8.
11943 Detection threshold, default is 0.5.
11946 Number of mipmaps, default is 3.
11948 @item xmin, ymin, xmax, ymax
11949 Specifies the rectangle in which to search.
11952 @subsection Examples
11956 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11958 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11964 Flood area with values of same pixel components with another values.
11966 It accepts the following options:
11969 Set pixel x coordinate.
11972 Set pixel y coordinate.
11975 Set source #0 component value.
11978 Set source #1 component value.
11981 Set source #2 component value.
11984 Set source #3 component value.
11987 Set destination #0 component value.
11990 Set destination #1 component value.
11993 Set destination #2 component value.
11996 Set destination #3 component value.
12002 Convert the input video to one of the specified pixel formats.
12003 Libavfilter will try to pick one that is suitable as input to
12006 It accepts the following parameters:
12010 A '|'-separated list of pixel format names, such as
12011 "pix_fmts=yuv420p|monow|rgb24".
12015 @subsection Examples
12019 Convert the input video to the @var{yuv420p} format
12021 format=pix_fmts=yuv420p
12024 Convert the input video to any of the formats in the list
12026 format=pix_fmts=yuv420p|yuv444p|yuv410p
12033 Convert the video to specified constant frame rate by duplicating or dropping
12034 frames as necessary.
12036 It accepts the following parameters:
12040 The desired output frame rate. The default is @code{25}.
12043 Assume the first PTS should be the given value, in seconds. This allows for
12044 padding/trimming at the start of stream. By default, no assumption is made
12045 about the first frame's expected PTS, so no padding or trimming is done.
12046 For example, this could be set to 0 to pad the beginning with duplicates of
12047 the first frame if a video stream starts after the audio stream or to trim any
12048 frames with a negative PTS.
12051 Timestamp (PTS) rounding method.
12053 Possible values are:
12060 round towards -infinity
12062 round towards +infinity
12066 The default is @code{near}.
12069 Action performed when reading the last frame.
12071 Possible values are:
12074 Use same timestamp rounding method as used for other frames.
12076 Pass through last frame if input duration has not been reached yet.
12078 The default is @code{round}.
12082 Alternatively, the options can be specified as a flat string:
12083 @var{fps}[:@var{start_time}[:@var{round}]].
12085 See also the @ref{setpts} filter.
12087 @subsection Examples
12091 A typical usage in order to set the fps to 25:
12097 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
12099 fps=fps=film:round=near
12105 Pack two different video streams into a stereoscopic video, setting proper
12106 metadata on supported codecs. The two views should have the same size and
12107 framerate and processing will stop when the shorter video ends. Please note
12108 that you may conveniently adjust view properties with the @ref{scale} and
12111 It accepts the following parameters:
12115 The desired packing format. Supported values are:
12120 The views are next to each other (default).
12123 The views are on top of each other.
12126 The views are packed by line.
12129 The views are packed by column.
12132 The views are temporally interleaved.
12141 # Convert left and right views into a frame-sequential video
12142 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
12144 # Convert views into a side-by-side video with the same output resolution as the input
12145 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
12150 Change the frame rate by interpolating new video output frames from the source
12153 This filter is not designed to function correctly with interlaced media. If
12154 you wish to change the frame rate of interlaced media then you are required
12155 to deinterlace before this filter and re-interlace after this filter.
12157 A description of the accepted options follows.
12161 Specify the output frames per second. This option can also be specified
12162 as a value alone. The default is @code{50}.
12165 Specify the start of a range where the output frame will be created as a
12166 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12167 the default is @code{15}.
12170 Specify the end of a range where the output frame will be created as a
12171 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12172 the default is @code{240}.
12175 Specify the level at which a scene change is detected as a value between
12176 0 and 100 to indicate a new scene; a low value reflects a low
12177 probability for the current frame to introduce a new scene, while a higher
12178 value means the current frame is more likely to be one.
12179 The default is @code{8.2}.
12182 Specify flags influencing the filter process.
12184 Available value for @var{flags} is:
12187 @item scene_change_detect, scd
12188 Enable scene change detection using the value of the option @var{scene}.
12189 This flag is enabled by default.
12195 Select one frame every N-th frame.
12197 This filter accepts the following option:
12200 Select frame after every @code{step} frames.
12201 Allowed values are positive integers higher than 0. Default value is @code{1}.
12204 @section freezedetect
12206 Detect frozen video.
12208 This filter logs a message and sets frame metadata when it detects that the
12209 input video has no significant change in content during a specified duration.
12210 Video freeze detection calculates the mean average absolute difference of all
12211 the components of video frames and compares it to a noise floor.
12213 The printed times and duration are expressed in seconds. The
12214 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12215 whose timestamp equals or exceeds the detection duration and it contains the
12216 timestamp of the first frame of the freeze. The
12217 @code{lavfi.freezedetect.freeze_duration} and
12218 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12221 The filter accepts the following options:
12225 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12226 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12230 Set freeze duration until notification (default is 2 seconds).
12233 @section freezeframes
12235 Freeze video frames.
12237 This filter freezes video frames using frame from 2nd input.
12239 The filter accepts the following options:
12243 Set number of first frame from which to start freeze.
12246 Set number of last frame from which to end freeze.
12249 Set number of frame from 2nd input which will be used instead of replaced frames.
12255 Apply a frei0r effect to the input video.
12257 To enable the compilation of this filter, you need to install the frei0r
12258 header and configure FFmpeg with @code{--enable-frei0r}.
12260 It accepts the following parameters:
12265 The name of the frei0r effect to load. If the environment variable
12266 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12267 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12268 Otherwise, the standard frei0r paths are searched, in this order:
12269 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12270 @file{/usr/lib/frei0r-1/}.
12272 @item filter_params
12273 A '|'-separated list of parameters to pass to the frei0r effect.
12277 A frei0r effect parameter can be a boolean (its value is either
12278 "y" or "n"), a double, a color (specified as
12279 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12280 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12281 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12282 a position (specified as @var{X}/@var{Y}, where
12283 @var{X} and @var{Y} are floating point numbers) and/or a string.
12285 The number and types of parameters depend on the loaded effect. If an
12286 effect parameter is not specified, the default value is set.
12288 @subsection Examples
12292 Apply the distort0r effect, setting the first two double parameters:
12294 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12298 Apply the colordistance effect, taking a color as the first parameter:
12300 frei0r=colordistance:0.2/0.3/0.4
12301 frei0r=colordistance:violet
12302 frei0r=colordistance:0x112233
12306 Apply the perspective effect, specifying the top left and top right image
12309 frei0r=perspective:0.2/0.2|0.8/0.2
12313 For more information, see
12314 @url{http://frei0r.dyne.org}
12316 @subsection Commands
12318 This filter supports the @option{filter_params} option as @ref{commands}.
12322 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12324 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12325 processing filter, one of them is performed once per block, not per pixel.
12326 This allows for much higher speed.
12328 The filter accepts the following options:
12332 Set quality. This option defines the number of levels for averaging. It accepts
12333 an integer in the range 4-5. Default value is @code{4}.
12336 Force a constant quantization parameter. It accepts an integer in range 0-63.
12337 If not set, the filter will use the QP from the video stream (if available).
12340 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12341 more details but also more artifacts, while higher values make the image smoother
12342 but also blurrier. Default value is @code{0} − PSNR optimal.
12344 @item use_bframe_qp
12345 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12346 option may cause flicker since the B-Frames have often larger QP. Default is
12347 @code{0} (not enabled).
12353 Apply Gaussian blur filter.
12355 The filter accepts the following options:
12359 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12362 Set number of steps for Gaussian approximation. Default is @code{1}.
12365 Set which planes to filter. By default all planes are filtered.
12368 Set vertical sigma, if negative it will be same as @code{sigma}.
12369 Default is @code{-1}.
12372 @subsection Commands
12373 This filter supports same commands as options.
12374 The command accepts the same syntax of the corresponding option.
12376 If the specified expression is not valid, it is kept at its current
12381 Apply generic equation to each pixel.
12383 The filter accepts the following options:
12386 @item lum_expr, lum
12387 Set the luminance expression.
12389 Set the chrominance blue expression.
12391 Set the chrominance red expression.
12392 @item alpha_expr, a
12393 Set the alpha expression.
12395 Set the red expression.
12396 @item green_expr, g
12397 Set the green expression.
12399 Set the blue expression.
12402 The colorspace is selected according to the specified options. If one
12403 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12404 options is specified, the filter will automatically select a YCbCr
12405 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12406 @option{blue_expr} options is specified, it will select an RGB
12409 If one of the chrominance expression is not defined, it falls back on the other
12410 one. If no alpha expression is specified it will evaluate to opaque value.
12411 If none of chrominance expressions are specified, they will evaluate
12412 to the luminance expression.
12414 The expressions can use the following variables and functions:
12418 The sequential number of the filtered frame, starting from @code{0}.
12422 The coordinates of the current sample.
12426 The width and height of the image.
12430 Width and height scale depending on the currently filtered plane. It is the
12431 ratio between the corresponding luma plane number of pixels and the current
12432 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12433 @code{0.5,0.5} for chroma planes.
12436 Time of the current frame, expressed in seconds.
12439 Return the value of the pixel at location (@var{x},@var{y}) of the current
12443 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12447 Return the value of the pixel at location (@var{x},@var{y}) of the
12448 blue-difference chroma plane. Return 0 if there is no such plane.
12451 Return the value of the pixel at location (@var{x},@var{y}) of the
12452 red-difference chroma plane. Return 0 if there is no such plane.
12457 Return the value of the pixel at location (@var{x},@var{y}) of the
12458 red/green/blue component. Return 0 if there is no such component.
12461 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12462 plane. Return 0 if there is no such plane.
12464 @item psum(x,y), lumsum(x, y), cbsum(x,y), crsum(x,y), rsum(x,y), gsum(x,y), bsum(x,y), alphasum(x,y)
12465 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12466 sums of samples within a rectangle. See the functions without the sum postfix.
12468 @item interpolation
12469 Set one of interpolation methods:
12474 Default is bilinear.
12477 For functions, if @var{x} and @var{y} are outside the area, the value will be
12478 automatically clipped to the closer edge.
12480 Please note that this filter can use multiple threads in which case each slice
12481 will have its own expression state. If you want to use only a single expression
12482 state because your expressions depend on previous state then you should limit
12483 the number of filter threads to 1.
12485 @subsection Examples
12489 Flip the image horizontally:
12495 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12496 wavelength of 100 pixels:
12498 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12502 Generate a fancy enigmatic moving light:
12504 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
12508 Generate a quick emboss effect:
12510 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12514 Modify RGB components depending on pixel position:
12516 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12520 Create a radial gradient that is the same size as the input (also see
12521 the @ref{vignette} filter):
12523 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12529 Fix the banding artifacts that are sometimes introduced into nearly flat
12530 regions by truncation to 8-bit color depth.
12531 Interpolate the gradients that should go where the bands are, and
12534 It is designed for playback only. Do not use it prior to
12535 lossy compression, because compression tends to lose the dither and
12536 bring back the bands.
12538 It accepts the following parameters:
12543 The maximum amount by which the filter will change any one pixel. This is also
12544 the threshold for detecting nearly flat regions. Acceptable values range from
12545 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12549 The neighborhood to fit the gradient to. A larger radius makes for smoother
12550 gradients, but also prevents the filter from modifying the pixels near detailed
12551 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12552 values will be clipped to the valid range.
12556 Alternatively, the options can be specified as a flat string:
12557 @var{strength}[:@var{radius}]
12559 @subsection Examples
12563 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12569 Specify radius, omitting the strength (which will fall-back to the default
12577 @anchor{graphmonitor}
12578 @section graphmonitor
12579 Show various filtergraph stats.
12581 With this filter one can debug complete filtergraph.
12582 Especially issues with links filling with queued frames.
12584 The filter accepts the following options:
12588 Set video output size. Default is @var{hd720}.
12591 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12594 Set output mode, can be @var{fulll} or @var{compact}.
12595 In @var{compact} mode only filters with some queued frames have displayed stats.
12598 Set flags which enable which stats are shown in video.
12600 Available values for flags are:
12603 Display number of queued frames in each link.
12605 @item frame_count_in
12606 Display number of frames taken from filter.
12608 @item frame_count_out
12609 Display number of frames given out from filter.
12612 Display current filtered frame pts.
12615 Display current filtered frame time.
12618 Display time base for filter link.
12621 Display used format for filter link.
12624 Display video size or number of audio channels in case of audio used by filter link.
12627 Display video frame rate or sample rate in case of audio used by filter link.
12630 Display link output status.
12634 Set upper limit for video rate of output stream, Default value is @var{25}.
12635 This guarantee that output video frame rate will not be higher than this value.
12639 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12640 and corrects the scene colors accordingly.
12642 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12644 The filter accepts the following options:
12648 The order of differentiation to be applied on the scene. Must be chosen in the range
12649 [0,2] and default value is 1.
12652 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12653 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12654 max value instead of calculating Minkowski distance.
12657 The standard deviation of Gaussian blur to be applied on the scene. Must be
12658 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12659 can't be equal to 0 if @var{difford} is greater than 0.
12662 @subsection Examples
12668 greyedge=difford=1:minknorm=5:sigma=2
12674 greyedge=difford=1:minknorm=0:sigma=2
12682 Apply a Hald CLUT to a video stream.
12684 First input is the video stream to process, and second one is the Hald CLUT.
12685 The Hald CLUT input can be a simple picture or a complete video stream.
12687 The filter accepts the following options:
12691 Force termination when the shortest input terminates. Default is @code{0}.
12693 Continue applying the last CLUT after the end of the stream. A value of
12694 @code{0} disable the filter after the last frame of the CLUT is reached.
12695 Default is @code{1}.
12698 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12699 filters share the same internals).
12701 This filter also supports the @ref{framesync} options.
12703 More information about the Hald CLUT can be found on Eskil Steenberg's website
12704 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12706 @subsection Workflow examples
12708 @subsubsection Hald CLUT video stream
12710 Generate an identity Hald CLUT stream altered with various effects:
12712 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
12715 Note: make sure you use a lossless codec.
12717 Then use it with @code{haldclut} to apply it on some random stream:
12719 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12722 The Hald CLUT will be applied to the 10 first seconds (duration of
12723 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12724 to the remaining frames of the @code{mandelbrot} stream.
12726 @subsubsection Hald CLUT with preview
12728 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12729 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12730 biggest possible square starting at the top left of the picture. The remaining
12731 padding pixels (bottom or right) will be ignored. This area can be used to add
12732 a preview of the Hald CLUT.
12734 Typically, the following generated Hald CLUT will be supported by the
12735 @code{haldclut} filter:
12738 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12739 pad=iw+320 [padded_clut];
12740 smptebars=s=320x256, split [a][b];
12741 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12742 [main][b] overlay=W-320" -frames:v 1 clut.png
12745 It contains the original and a preview of the effect of the CLUT: SMPTE color
12746 bars are displayed on the right-top, and below the same color bars processed by
12749 Then, the effect of this Hald CLUT can be visualized with:
12751 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12756 Flip the input video horizontally.
12758 For example, to horizontally flip the input video with @command{ffmpeg}:
12760 ffmpeg -i in.avi -vf "hflip" out.avi
12764 This filter applies a global color histogram equalization on a
12767 It can be used to correct video that has a compressed range of pixel
12768 intensities. The filter redistributes the pixel intensities to
12769 equalize their distribution across the intensity range. It may be
12770 viewed as an "automatically adjusting contrast filter". This filter is
12771 useful only for correcting degraded or poorly captured source
12774 The filter accepts the following options:
12778 Determine the amount of equalization to be applied. As the strength
12779 is reduced, the distribution of pixel intensities more-and-more
12780 approaches that of the input frame. The value must be a float number
12781 in the range [0,1] and defaults to 0.200.
12784 Set the maximum intensity that can generated and scale the output
12785 values appropriately. The strength should be set as desired and then
12786 the intensity can be limited if needed to avoid washing-out. The value
12787 must be a float number in the range [0,1] and defaults to 0.210.
12790 Set the antibanding level. If enabled the filter will randomly vary
12791 the luminance of output pixels by a small amount to avoid banding of
12792 the histogram. Possible values are @code{none}, @code{weak} or
12793 @code{strong}. It defaults to @code{none}.
12799 Compute and draw a color distribution histogram for the input video.
12801 The computed histogram is a representation of the color component
12802 distribution in an image.
12804 Standard histogram displays the color components distribution in an image.
12805 Displays color graph for each color component. Shows distribution of
12806 the Y, U, V, A or R, G, B components, depending on input format, in the
12807 current frame. Below each graph a color component scale meter is shown.
12809 The filter accepts the following options:
12813 Set height of level. Default value is @code{200}.
12814 Allowed range is [50, 2048].
12817 Set height of color scale. Default value is @code{12}.
12818 Allowed range is [0, 40].
12822 It accepts the following values:
12825 Per color component graphs are placed below each other.
12828 Per color component graphs are placed side by side.
12831 Presents information identical to that in the @code{parade}, except
12832 that the graphs representing color components are superimposed directly
12835 Default is @code{stack}.
12838 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12839 Default is @code{linear}.
12842 Set what color components to display.
12843 Default is @code{7}.
12846 Set foreground opacity. Default is @code{0.7}.
12849 Set background opacity. Default is @code{0.5}.
12852 @subsection Examples
12857 Calculate and draw histogram:
12859 ffplay -i input -vf histogram
12867 This is a high precision/quality 3d denoise filter. It aims to reduce
12868 image noise, producing smooth images and making still images really
12869 still. It should enhance compressibility.
12871 It accepts the following optional parameters:
12875 A non-negative floating point number which specifies spatial luma strength.
12876 It defaults to 4.0.
12878 @item chroma_spatial
12879 A non-negative floating point number which specifies spatial chroma strength.
12880 It defaults to 3.0*@var{luma_spatial}/4.0.
12883 A floating point number which specifies luma temporal strength. It defaults to
12884 6.0*@var{luma_spatial}/4.0.
12887 A floating point number which specifies chroma temporal strength. It defaults to
12888 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12891 @subsection Commands
12892 This filter supports same @ref{commands} as options.
12893 The command accepts the same syntax of the corresponding option.
12895 If the specified expression is not valid, it is kept at its current
12898 @anchor{hwdownload}
12899 @section hwdownload
12901 Download hardware frames to system memory.
12903 The input must be in hardware frames, and the output a non-hardware format.
12904 Not all formats will be supported on the output - it may be necessary to insert
12905 an additional @option{format} filter immediately following in the graph to get
12906 the output in a supported format.
12910 Map hardware frames to system memory or to another device.
12912 This filter has several different modes of operation; which one is used depends
12913 on the input and output formats:
12916 Hardware frame input, normal frame output
12918 Map the input frames to system memory and pass them to the output. If the
12919 original hardware frame is later required (for example, after overlaying
12920 something else on part of it), the @option{hwmap} filter can be used again
12921 in the next mode to retrieve it.
12923 Normal frame input, hardware frame output
12925 If the input is actually a software-mapped hardware frame, then unmap it -
12926 that is, return the original hardware frame.
12928 Otherwise, a device must be provided. Create new hardware surfaces on that
12929 device for the output, then map them back to the software format at the input
12930 and give those frames to the preceding filter. This will then act like the
12931 @option{hwupload} filter, but may be able to avoid an additional copy when
12932 the input is already in a compatible format.
12934 Hardware frame input and output
12936 A device must be supplied for the output, either directly or with the
12937 @option{derive_device} option. The input and output devices must be of
12938 different types and compatible - the exact meaning of this is
12939 system-dependent, but typically it means that they must refer to the same
12940 underlying hardware context (for example, refer to the same graphics card).
12942 If the input frames were originally created on the output device, then unmap
12943 to retrieve the original frames.
12945 Otherwise, map the frames to the output device - create new hardware frames
12946 on the output corresponding to the frames on the input.
12949 The following additional parameters are accepted:
12953 Set the frame mapping mode. Some combination of:
12956 The mapped frame should be readable.
12958 The mapped frame should be writeable.
12960 The mapping will always overwrite the entire frame.
12962 This may improve performance in some cases, as the original contents of the
12963 frame need not be loaded.
12965 The mapping must not involve any copying.
12967 Indirect mappings to copies of frames are created in some cases where either
12968 direct mapping is not possible or it would have unexpected properties.
12969 Setting this flag ensures that the mapping is direct and will fail if that is
12972 Defaults to @var{read+write} if not specified.
12974 @item derive_device @var{type}
12975 Rather than using the device supplied at initialisation, instead derive a new
12976 device of type @var{type} from the device the input frames exist on.
12979 In a hardware to hardware mapping, map in reverse - create frames in the sink
12980 and map them back to the source. This may be necessary in some cases where
12981 a mapping in one direction is required but only the opposite direction is
12982 supported by the devices being used.
12984 This option is dangerous - it may break the preceding filter in undefined
12985 ways if there are any additional constraints on that filter's output.
12986 Do not use it without fully understanding the implications of its use.
12992 Upload system memory frames to hardware surfaces.
12994 The device to upload to must be supplied when the filter is initialised. If
12995 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12996 option or with the @option{derive_device} option. The input and output devices
12997 must be of different types and compatible - the exact meaning of this is
12998 system-dependent, but typically it means that they must refer to the same
12999 underlying hardware context (for example, refer to the same graphics card).
13001 The following additional parameters are accepted:
13004 @item derive_device @var{type}
13005 Rather than using the device supplied at initialisation, instead derive a new
13006 device of type @var{type} from the device the input frames exist on.
13009 @anchor{hwupload_cuda}
13010 @section hwupload_cuda
13012 Upload system memory frames to a CUDA device.
13014 It accepts the following optional parameters:
13018 The number of the CUDA device to use
13023 Apply a high-quality magnification filter designed for pixel art. This filter
13024 was originally created by Maxim Stepin.
13026 It accepts the following option:
13030 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
13031 @code{hq3x} and @code{4} for @code{hq4x}.
13032 Default is @code{3}.
13036 Stack input videos horizontally.
13038 All streams must be of same pixel format and of same height.
13040 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
13041 to create same output.
13043 The filter accepts the following option:
13047 Set number of input streams. Default is 2.
13050 If set to 1, force the output to terminate when the shortest input
13051 terminates. Default value is 0.
13056 Modify the hue and/or the saturation of the input.
13058 It accepts the following parameters:
13062 Specify the hue angle as a number of degrees. It accepts an expression,
13063 and defaults to "0".
13066 Specify the saturation in the [-10,10] range. It accepts an expression and
13070 Specify the hue angle as a number of radians. It accepts an
13071 expression, and defaults to "0".
13074 Specify the brightness in the [-10,10] range. It accepts an expression and
13078 @option{h} and @option{H} are mutually exclusive, and can't be
13079 specified at the same time.
13081 The @option{b}, @option{h}, @option{H} and @option{s} option values are
13082 expressions containing the following constants:
13086 frame count of the input frame starting from 0
13089 presentation timestamp of the input frame expressed in time base units
13092 frame rate of the input video, NAN if the input frame rate is unknown
13095 timestamp expressed in seconds, NAN if the input timestamp is unknown
13098 time base of the input video
13101 @subsection Examples
13105 Set the hue to 90 degrees and the saturation to 1.0:
13111 Same command but expressing the hue in radians:
13117 Rotate hue and make the saturation swing between 0
13118 and 2 over a period of 1 second:
13120 hue="H=2*PI*t: s=sin(2*PI*t)+1"
13124 Apply a 3 seconds saturation fade-in effect starting at 0:
13126 hue="s=min(t/3\,1)"
13129 The general fade-in expression can be written as:
13131 hue="s=min(0\, max((t-START)/DURATION\, 1))"
13135 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
13137 hue="s=max(0\, min(1\, (8-t)/3))"
13140 The general fade-out expression can be written as:
13142 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
13147 @subsection Commands
13149 This filter supports the following commands:
13155 Modify the hue and/or the saturation and/or brightness of the input video.
13156 The command accepts the same syntax of the corresponding option.
13158 If the specified expression is not valid, it is kept at its current
13162 @section hysteresis
13164 Grow first stream into second stream by connecting components.
13165 This makes it possible to build more robust edge masks.
13167 This filter accepts the following options:
13171 Set which planes will be processed as bitmap, unprocessed planes will be
13172 copied from first stream.
13173 By default value 0xf, all planes will be processed.
13176 Set threshold which is used in filtering. If pixel component value is higher than
13177 this value filter algorithm for connecting components is activated.
13178 By default value is 0.
13181 The @code{hysteresis} filter also supports the @ref{framesync} options.
13185 Detect video interlacing type.
13187 This filter tries to detect if the input frames are interlaced, progressive,
13188 top or bottom field first. It will also try to detect fields that are
13189 repeated between adjacent frames (a sign of telecine).
13191 Single frame detection considers only immediately adjacent frames when classifying each frame.
13192 Multiple frame detection incorporates the classification history of previous frames.
13194 The filter will log these metadata values:
13197 @item single.current_frame
13198 Detected type of current frame using single-frame detection. One of:
13199 ``tff'' (top field first), ``bff'' (bottom field first),
13200 ``progressive'', or ``undetermined''
13203 Cumulative number of frames detected as top field first using single-frame detection.
13206 Cumulative number of frames detected as top field first using multiple-frame detection.
13209 Cumulative number of frames detected as bottom field first using single-frame detection.
13211 @item multiple.current_frame
13212 Detected type of current frame using multiple-frame detection. One of:
13213 ``tff'' (top field first), ``bff'' (bottom field first),
13214 ``progressive'', or ``undetermined''
13217 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13219 @item single.progressive
13220 Cumulative number of frames detected as progressive using single-frame detection.
13222 @item multiple.progressive
13223 Cumulative number of frames detected as progressive using multiple-frame detection.
13225 @item single.undetermined
13226 Cumulative number of frames that could not be classified using single-frame detection.
13228 @item multiple.undetermined
13229 Cumulative number of frames that could not be classified using multiple-frame detection.
13231 @item repeated.current_frame
13232 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13234 @item repeated.neither
13235 Cumulative number of frames with no repeated field.
13238 Cumulative number of frames with the top field repeated from the previous frame's top field.
13240 @item repeated.bottom
13241 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13244 The filter accepts the following options:
13248 Set interlacing threshold.
13250 Set progressive threshold.
13252 Threshold for repeated field detection.
13254 Number of frames after which a given frame's contribution to the
13255 statistics is halved (i.e., it contributes only 0.5 to its
13256 classification). The default of 0 means that all frames seen are given
13257 full weight of 1.0 forever.
13258 @item analyze_interlaced_flag
13259 When this is not 0 then idet will use the specified number of frames to determine
13260 if the interlaced flag is accurate, it will not count undetermined frames.
13261 If the flag is found to be accurate it will be used without any further
13262 computations, if it is found to be inaccurate it will be cleared without any
13263 further computations. This allows inserting the idet filter as a low computational
13264 method to clean up the interlaced flag
13269 Deinterleave or interleave fields.
13271 This filter allows one to process interlaced images fields without
13272 deinterlacing them. Deinterleaving splits the input frame into 2
13273 fields (so called half pictures). Odd lines are moved to the top
13274 half of the output image, even lines to the bottom half.
13275 You can process (filter) them independently and then re-interleave them.
13277 The filter accepts the following options:
13281 @item chroma_mode, c
13282 @item alpha_mode, a
13283 Available values for @var{luma_mode}, @var{chroma_mode} and
13284 @var{alpha_mode} are:
13290 @item deinterleave, d
13291 Deinterleave fields, placing one above the other.
13293 @item interleave, i
13294 Interleave fields. Reverse the effect of deinterleaving.
13296 Default value is @code{none}.
13298 @item luma_swap, ls
13299 @item chroma_swap, cs
13300 @item alpha_swap, as
13301 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13304 @subsection Commands
13306 This filter supports the all above options as @ref{commands}.
13310 Apply inflate effect to the video.
13312 This filter replaces the pixel by the local(3x3) average by taking into account
13313 only values higher than the pixel.
13315 It accepts the following options:
13322 Limit the maximum change for each plane, default is 65535.
13323 If 0, plane will remain unchanged.
13326 @subsection Commands
13328 This filter supports the all above options as @ref{commands}.
13332 Simple interlacing filter from progressive contents. This interleaves upper (or
13333 lower) lines from odd frames with lower (or upper) lines from even frames,
13334 halving the frame rate and preserving image height.
13337 Original Original New Frame
13338 Frame 'j' Frame 'j+1' (tff)
13339 ========== =========== ==================
13340 Line 0 --------------------> Frame 'j' Line 0
13341 Line 1 Line 1 ----> Frame 'j+1' Line 1
13342 Line 2 ---------------------> Frame 'j' Line 2
13343 Line 3 Line 3 ----> Frame 'j+1' Line 3
13345 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13348 It accepts the following optional parameters:
13352 This determines whether the interlaced frame is taken from the even
13353 (tff - default) or odd (bff) lines of the progressive frame.
13356 Vertical lowpass filter to avoid twitter interlacing and
13357 reduce moire patterns.
13361 Disable vertical lowpass filter
13364 Enable linear filter (default)
13367 Enable complex filter. This will slightly less reduce twitter and moire
13368 but better retain detail and subjective sharpness impression.
13375 Deinterlace input video by applying Donald Graft's adaptive kernel
13376 deinterling. Work on interlaced parts of a video to produce
13377 progressive frames.
13379 The description of the accepted parameters follows.
13383 Set the threshold which affects the filter's tolerance when
13384 determining if a pixel line must be processed. It must be an integer
13385 in the range [0,255] and defaults to 10. A value of 0 will result in
13386 applying the process on every pixels.
13389 Paint pixels exceeding the threshold value to white if set to 1.
13393 Set the fields order. Swap fields if set to 1, leave fields alone if
13397 Enable additional sharpening if set to 1. Default is 0.
13400 Enable twoway sharpening if set to 1. Default is 0.
13403 @subsection Examples
13407 Apply default values:
13409 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13413 Enable additional sharpening:
13419 Paint processed pixels in white:
13426 Apply kirsch operator to input video stream.
13428 The filter accepts the following option:
13432 Set which planes will be processed, unprocessed planes will be copied.
13433 By default value 0xf, all planes will be processed.
13436 Set value which will be multiplied with filtered result.
13439 Set value which will be added to filtered result.
13442 @subsection Commands
13444 This filter supports the all above options as @ref{commands}.
13448 Slowly update darker pixels.
13450 This filter makes short flashes of light appear longer.
13451 This filter accepts the following options:
13455 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13458 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13461 @section lenscorrection
13463 Correct radial lens distortion
13465 This filter can be used to correct for radial distortion as can result from the use
13466 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13467 one can use tools available for example as part of opencv or simply trial-and-error.
13468 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13469 and extract the k1 and k2 coefficients from the resulting matrix.
13471 Note that effectively the same filter is available in the open-source tools Krita and
13472 Digikam from the KDE project.
13474 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13475 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13476 brightness distribution, so you may want to use both filters together in certain
13477 cases, though you will have to take care of ordering, i.e. whether vignetting should
13478 be applied before or after lens correction.
13480 @subsection Options
13482 The filter accepts the following options:
13486 Relative x-coordinate of the focal point of the image, and thereby the center of the
13487 distortion. This value has a range [0,1] and is expressed as fractions of the image
13488 width. Default is 0.5.
13490 Relative y-coordinate of the focal point of the image, and thereby the center of the
13491 distortion. This value has a range [0,1] and is expressed as fractions of the image
13492 height. Default is 0.5.
13494 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13495 no correction. Default is 0.
13497 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13498 0 means no correction. Default is 0.
13500 Set interpolation type. Can be @code{nearest} or @code{bilinear}.
13501 Default is @code{nearest}.
13503 Specify the color of the unmapped pixels. For the syntax of this option,
13504 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
13505 manual,ffmpeg-utils}. Default color is @code{black@@0}.
13508 The formula that generates the correction is:
13510 @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)
13512 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13513 distances from the focal point in the source and target images, respectively.
13515 @subsection Commands
13517 This filter supports the all above options as @ref{commands}.
13521 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13523 The @code{lensfun} filter requires the camera make, camera model, and lens model
13524 to apply the lens correction. The filter will load the lensfun database and
13525 query it to find the corresponding camera and lens entries in the database. As
13526 long as these entries can be found with the given options, the filter can
13527 perform corrections on frames. Note that incomplete strings will result in the
13528 filter choosing the best match with the given options, and the filter will
13529 output the chosen camera and lens models (logged with level "info"). You must
13530 provide the make, camera model, and lens model as they are required.
13532 The filter accepts the following options:
13536 The make of the camera (for example, "Canon"). This option is required.
13539 The model of the camera (for example, "Canon EOS 100D"). This option is
13543 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13544 option is required.
13547 The type of correction to apply. The following values are valid options:
13551 Enables fixing lens vignetting.
13554 Enables fixing lens geometry. This is the default.
13557 Enables fixing chromatic aberrations.
13560 Enables fixing lens vignetting and lens geometry.
13563 Enables fixing lens vignetting and chromatic aberrations.
13566 Enables fixing both lens geometry and chromatic aberrations.
13569 Enables all possible corrections.
13573 The focal length of the image/video (zoom; expected constant for video). For
13574 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13575 range should be chosen when using that lens. Default 18.
13578 The aperture of the image/video (expected constant for video). Note that
13579 aperture is only used for vignetting correction. Default 3.5.
13581 @item focus_distance
13582 The focus distance of the image/video (expected constant for video). Note that
13583 focus distance is only used for vignetting and only slightly affects the
13584 vignetting correction process. If unknown, leave it at the default value (which
13588 The scale factor which is applied after transformation. After correction the
13589 video is no longer necessarily rectangular. This parameter controls how much of
13590 the resulting image is visible. The value 0 means that a value will be chosen
13591 automatically such that there is little or no unmapped area in the output
13592 image. 1.0 means that no additional scaling is done. Lower values may result
13593 in more of the corrected image being visible, while higher values may avoid
13594 unmapped areas in the output.
13596 @item target_geometry
13597 The target geometry of the output image/video. The following values are valid
13601 @item rectilinear (default)
13604 @item equirectangular
13605 @item fisheye_orthographic
13606 @item fisheye_stereographic
13607 @item fisheye_equisolid
13608 @item fisheye_thoby
13611 Apply the reverse of image correction (instead of correcting distortion, apply
13614 @item interpolation
13615 The type of interpolation used when correcting distortion. The following values
13620 @item linear (default)
13625 @subsection Examples
13629 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13630 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13634 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
13638 Apply the same as before, but only for the first 5 seconds of video.
13641 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
13648 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13649 score between two input videos.
13651 The obtained VMAF score is printed through the logging system.
13653 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13654 After installing the library it can be enabled using:
13655 @code{./configure --enable-libvmaf}.
13656 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13658 The filter has following options:
13662 Set the model path which is to be used for SVM.
13663 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13666 Set the file path to be used to store logs.
13669 Set the format of the log file (csv, json or xml).
13671 @item enable_transform
13672 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13673 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13674 Default value: @code{false}
13677 Invokes the phone model which will generate VMAF scores higher than in the
13678 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13679 Default value: @code{false}
13682 Enables computing psnr along with vmaf.
13683 Default value: @code{false}
13686 Enables computing ssim along with vmaf.
13687 Default value: @code{false}
13690 Enables computing ms_ssim along with vmaf.
13691 Default value: @code{false}
13694 Set the pool method to be used for computing vmaf.
13695 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13698 Set number of threads to be used when computing vmaf.
13699 Default value: @code{0}, which makes use of all available logical processors.
13702 Set interval for frame subsampling used when computing vmaf.
13703 Default value: @code{1}
13705 @item enable_conf_interval
13706 Enables confidence interval.
13707 Default value: @code{false}
13710 This filter also supports the @ref{framesync} options.
13712 @subsection Examples
13715 On the below examples the input file @file{main.mpg} being processed is
13716 compared with the reference file @file{ref.mpg}.
13719 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13723 Example with options:
13725 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13729 Example with options and different containers:
13731 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]libvmaf=psnr=1:log_fmt=json" -f null -
13737 Limits the pixel components values to the specified range [min, max].
13739 The filter accepts the following options:
13743 Lower bound. Defaults to the lowest allowed value for the input.
13746 Upper bound. Defaults to the highest allowed value for the input.
13749 Specify which planes will be processed. Defaults to all available.
13752 @subsection Commands
13754 This filter supports the all above options as @ref{commands}.
13760 The filter accepts the following options:
13764 Set the number of loops. Setting this value to -1 will result in infinite loops.
13768 Set maximal size in number of frames. Default is 0.
13771 Set first frame of loop. Default is 0.
13774 @subsection Examples
13778 Loop single first frame infinitely:
13780 loop=loop=-1:size=1:start=0
13784 Loop single first frame 10 times:
13786 loop=loop=10:size=1:start=0
13790 Loop 10 first frames 5 times:
13792 loop=loop=5:size=10:start=0
13798 Apply a 1D LUT to an input video.
13800 The filter accepts the following options:
13804 Set the 1D LUT file name.
13806 Currently supported formats:
13815 Select interpolation mode.
13817 Available values are:
13821 Use values from the nearest defined point.
13823 Interpolate values using the linear interpolation.
13825 Interpolate values using the cosine interpolation.
13827 Interpolate values using the cubic interpolation.
13829 Interpolate values using the spline interpolation.
13836 Apply a 3D LUT to an input video.
13838 The filter accepts the following options:
13842 Set the 3D LUT file name.
13844 Currently supported formats:
13858 Select interpolation mode.
13860 Available values are:
13864 Use values from the nearest defined point.
13866 Interpolate values using the 8 points defining a cube.
13868 Interpolate values using a tetrahedron.
13874 Turn certain luma values into transparency.
13876 The filter accepts the following options:
13880 Set the luma which will be used as base for transparency.
13881 Default value is @code{0}.
13884 Set the range of luma values to be keyed out.
13885 Default value is @code{0.01}.
13888 Set the range of softness. Default value is @code{0}.
13889 Use this to control gradual transition from zero to full transparency.
13892 @subsection Commands
13893 This filter supports same @ref{commands} as options.
13894 The command accepts the same syntax of the corresponding option.
13896 If the specified expression is not valid, it is kept at its current
13899 @section lut, lutrgb, lutyuv
13901 Compute a look-up table for binding each pixel component input value
13902 to an output value, and apply it to the input video.
13904 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13905 to an RGB input video.
13907 These filters accept the following parameters:
13910 set first pixel component expression
13912 set second pixel component expression
13914 set third pixel component expression
13916 set fourth pixel component expression, corresponds to the alpha component
13919 set red component expression
13921 set green component expression
13923 set blue component expression
13925 alpha component expression
13928 set Y/luminance component expression
13930 set U/Cb component expression
13932 set V/Cr component expression
13935 Each of them specifies the expression to use for computing the lookup table for
13936 the corresponding pixel component values.
13938 The exact component associated to each of the @var{c*} options depends on the
13941 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13942 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13944 The expressions can contain the following constants and functions:
13949 The input width and height.
13952 The input value for the pixel component.
13955 The input value, clipped to the @var{minval}-@var{maxval} range.
13958 The maximum value for the pixel component.
13961 The minimum value for the pixel component.
13964 The negated value for the pixel component value, clipped to the
13965 @var{minval}-@var{maxval} range; it corresponds to the expression
13966 "maxval-clipval+minval".
13969 The computed value in @var{val}, clipped to the
13970 @var{minval}-@var{maxval} range.
13972 @item gammaval(gamma)
13973 The computed gamma correction value of the pixel component value,
13974 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13976 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13980 All expressions default to "val".
13982 @subsection Examples
13986 Negate input video:
13988 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13989 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13992 The above is the same as:
13994 lutrgb="r=negval:g=negval:b=negval"
13995 lutyuv="y=negval:u=negval:v=negval"
14005 Remove chroma components, turning the video into a graytone image:
14007 lutyuv="u=128:v=128"
14011 Apply a luma burning effect:
14017 Remove green and blue components:
14023 Set a constant alpha channel value on input:
14025 format=rgba,lutrgb=a="maxval-minval/2"
14029 Correct luminance gamma by a factor of 0.5:
14031 lutyuv=y=gammaval(0.5)
14035 Discard least significant bits of luma:
14037 lutyuv=y='bitand(val, 128+64+32)'
14041 Technicolor like effect:
14043 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
14047 @section lut2, tlut2
14049 The @code{lut2} filter takes two input streams and outputs one
14052 The @code{tlut2} (time lut2) filter takes two consecutive frames
14053 from one single stream.
14055 This filter accepts the following parameters:
14058 set first pixel component expression
14060 set second pixel component expression
14062 set third pixel component expression
14064 set fourth pixel component expression, corresponds to the alpha component
14067 set output bit depth, only available for @code{lut2} filter. By default is 0,
14068 which means bit depth is automatically picked from first input format.
14071 The @code{lut2} filter also supports the @ref{framesync} options.
14073 Each of them specifies the expression to use for computing the lookup table for
14074 the corresponding pixel component values.
14076 The exact component associated to each of the @var{c*} options depends on the
14079 The expressions can contain the following constants:
14084 The input width and height.
14087 The first input value for the pixel component.
14090 The second input value for the pixel component.
14093 The first input video bit depth.
14096 The second input video bit depth.
14099 All expressions default to "x".
14101 @subsection Examples
14105 Highlight differences between two RGB video streams:
14107 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)'
14111 Highlight differences between two YUV video streams:
14113 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)'
14117 Show max difference between two video streams:
14119 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)))'
14123 @section maskedclamp
14125 Clamp the first input stream with the second input and third input stream.
14127 Returns the value of first stream to be between second input
14128 stream - @code{undershoot} and third input stream + @code{overshoot}.
14130 This filter accepts the following options:
14133 Default value is @code{0}.
14136 Default value is @code{0}.
14139 Set which planes will be processed as bitmap, unprocessed planes will be
14140 copied from first stream.
14141 By default value 0xf, all planes will be processed.
14144 @subsection Commands
14146 This filter supports the all above options as @ref{commands}.
14150 Merge the second and third input stream into output stream using absolute differences
14151 between second input stream and first input stream and absolute difference between
14152 third input stream and first input stream. The picked value will be from second input
14153 stream if second absolute difference is greater than first one or from third input stream
14156 This filter accepts the following options:
14159 Set which planes will be processed as bitmap, unprocessed planes will be
14160 copied from first stream.
14161 By default value 0xf, all planes will be processed.
14164 @subsection Commands
14166 This filter supports the all above options as @ref{commands}.
14168 @section maskedmerge
14170 Merge the first input stream with the second input stream using per pixel
14171 weights in the third input stream.
14173 A value of 0 in the third stream pixel component means that pixel component
14174 from first stream is returned unchanged, while maximum value (eg. 255 for
14175 8-bit videos) means that pixel component from second stream is returned
14176 unchanged. Intermediate values define the amount of merging between both
14177 input stream's pixel components.
14179 This filter accepts the following options:
14182 Set which planes will be processed as bitmap, unprocessed planes will be
14183 copied from first stream.
14184 By default value 0xf, all planes will be processed.
14187 @subsection Commands
14189 This filter supports the all above options as @ref{commands}.
14193 Merge the second and third input stream into output stream using absolute differences
14194 between second input stream and first input stream and absolute difference between
14195 third input stream and first input stream. The picked value will be from second input
14196 stream if second absolute difference is less than first one or from third input stream
14199 This filter accepts the following options:
14202 Set which planes will be processed as bitmap, unprocessed planes will be
14203 copied from first stream.
14204 By default value 0xf, all planes will be processed.
14207 @subsection Commands
14209 This filter supports the all above options as @ref{commands}.
14211 @section maskedthreshold
14212 Pick pixels comparing absolute difference of two video streams with fixed
14215 If absolute difference between pixel component of first and second video
14216 stream is equal or lower than user supplied threshold than pixel component
14217 from first video stream is picked, otherwise pixel component from second
14218 video stream is picked.
14220 This filter accepts the following options:
14223 Set threshold used when picking pixels from absolute difference from two input
14227 Set which planes will be processed as bitmap, unprocessed planes will be
14228 copied from second stream.
14229 By default value 0xf, all planes will be processed.
14232 @subsection Commands
14234 This filter supports the all above options as @ref{commands}.
14237 Create mask from input video.
14239 For example it is useful to create motion masks after @code{tblend} filter.
14241 This filter accepts the following options:
14245 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14248 Set high threshold. Any pixel component higher than this value will be set to max value
14249 allowed for current pixel format.
14252 Set planes to filter, by default all available planes are filtered.
14255 Fill all frame pixels with this value.
14258 Set max average pixel value for frame. If sum of all pixel components is higher that this
14259 average, output frame will be completely filled with value set by @var{fill} option.
14260 Typically useful for scene changes when used in combination with @code{tblend} filter.
14265 Apply motion-compensation deinterlacing.
14267 It needs one field per frame as input and must thus be used together
14268 with yadif=1/3 or equivalent.
14270 This filter accepts the following options:
14273 Set the deinterlacing mode.
14275 It accepts one of the following values:
14280 use iterative motion estimation
14282 like @samp{slow}, but use multiple reference frames.
14284 Default value is @samp{fast}.
14287 Set the picture field parity assumed for the input video. It must be
14288 one of the following values:
14292 assume top field first
14294 assume bottom field first
14297 Default value is @samp{bff}.
14300 Set per-block quantization parameter (QP) used by the internal
14303 Higher values should result in a smoother motion vector field but less
14304 optimal individual vectors. Default value is 1.
14309 Pick median pixel from certain rectangle defined by radius.
14311 This filter accepts the following options:
14315 Set horizontal radius size. Default value is @code{1}.
14316 Allowed range is integer from 1 to 127.
14319 Set which planes to process. Default is @code{15}, which is all available planes.
14322 Set vertical radius size. Default value is @code{0}.
14323 Allowed range is integer from 0 to 127.
14324 If it is 0, value will be picked from horizontal @code{radius} option.
14327 Set median percentile. Default value is @code{0.5}.
14328 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14329 minimum values, and @code{1} maximum values.
14332 @subsection Commands
14333 This filter supports same @ref{commands} as options.
14334 The command accepts the same syntax of the corresponding option.
14336 If the specified expression is not valid, it is kept at its current
14339 @section mergeplanes
14341 Merge color channel components from several video streams.
14343 The filter accepts up to 4 input streams, and merge selected input
14344 planes to the output video.
14346 This filter accepts the following options:
14349 Set input to output plane mapping. Default is @code{0}.
14351 The mappings is specified as a bitmap. It should be specified as a
14352 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14353 mapping for the first plane of the output stream. 'A' sets the number of
14354 the input stream to use (from 0 to 3), and 'a' the plane number of the
14355 corresponding input to use (from 0 to 3). The rest of the mappings is
14356 similar, 'Bb' describes the mapping for the output stream second
14357 plane, 'Cc' describes the mapping for the output stream third plane and
14358 'Dd' describes the mapping for the output stream fourth plane.
14361 Set output pixel format. Default is @code{yuva444p}.
14364 @subsection Examples
14368 Merge three gray video streams of same width and height into single video stream:
14370 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14374 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14376 [a0][a1]mergeplanes=0x00010210:yuva444p
14380 Swap Y and A plane in yuva444p stream:
14382 format=yuva444p,mergeplanes=0x03010200:yuva444p
14386 Swap U and V plane in yuv420p stream:
14388 format=yuv420p,mergeplanes=0x000201:yuv420p
14392 Cast a rgb24 clip to yuv444p:
14394 format=rgb24,mergeplanes=0x000102:yuv444p
14400 Estimate and export motion vectors using block matching algorithms.
14401 Motion vectors are stored in frame side data to be used by other filters.
14403 This filter accepts the following options:
14406 Specify the motion estimation method. Accepts one of the following values:
14410 Exhaustive search algorithm.
14412 Three step search algorithm.
14414 Two dimensional logarithmic search algorithm.
14416 New three step search algorithm.
14418 Four step search algorithm.
14420 Diamond search algorithm.
14422 Hexagon-based search algorithm.
14424 Enhanced predictive zonal search algorithm.
14426 Uneven multi-hexagon search algorithm.
14428 Default value is @samp{esa}.
14431 Macroblock size. Default @code{16}.
14434 Search parameter. Default @code{7}.
14437 @section midequalizer
14439 Apply Midway Image Equalization effect using two video streams.
14441 Midway Image Equalization adjusts a pair of images to have the same
14442 histogram, while maintaining their dynamics as much as possible. It's
14443 useful for e.g. matching exposures from a pair of stereo cameras.
14445 This filter has two inputs and one output, which must be of same pixel format, but
14446 may be of different sizes. The output of filter is first input adjusted with
14447 midway histogram of both inputs.
14449 This filter accepts the following option:
14453 Set which planes to process. Default is @code{15}, which is all available planes.
14456 @section minterpolate
14458 Convert the video to specified frame rate using motion interpolation.
14460 This filter accepts the following options:
14463 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}.
14466 Motion interpolation mode. Following values are accepted:
14469 Duplicate previous or next frame for interpolating new ones.
14471 Blend source frames. Interpolated frame is mean of previous and next frames.
14473 Motion compensated interpolation. Following options are effective when this mode is selected:
14477 Motion compensation mode. Following values are accepted:
14480 Overlapped block motion compensation.
14482 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14484 Default mode is @samp{obmc}.
14487 Motion estimation mode. Following values are accepted:
14490 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14492 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14494 Default mode is @samp{bilat}.
14497 The algorithm to be used for motion estimation. Following values are accepted:
14500 Exhaustive search algorithm.
14502 Three step search algorithm.
14504 Two dimensional logarithmic search algorithm.
14506 New three step search algorithm.
14508 Four step search algorithm.
14510 Diamond search algorithm.
14512 Hexagon-based search algorithm.
14514 Enhanced predictive zonal search algorithm.
14516 Uneven multi-hexagon search algorithm.
14518 Default algorithm is @samp{epzs}.
14521 Macroblock size. Default @code{16}.
14524 Motion estimation search parameter. Default @code{32}.
14527 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).
14532 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:
14535 Disable scene change detection.
14537 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14539 Default method is @samp{fdiff}.
14541 @item scd_threshold
14542 Scene change detection threshold. Default is @code{10.}.
14547 Mix several video input streams into one video stream.
14549 A description of the accepted options follows.
14553 The number of inputs. If unspecified, it defaults to 2.
14556 Specify weight of each input video stream as sequence.
14557 Each weight is separated by space. If number of weights
14558 is smaller than number of @var{frames} last specified
14559 weight will be used for all remaining unset weights.
14562 Specify scale, if it is set it will be multiplied with sum
14563 of each weight multiplied with pixel values to give final destination
14564 pixel value. By default @var{scale} is auto scaled to sum of weights.
14567 Specify how end of stream is determined.
14570 The duration of the longest input. (default)
14573 The duration of the shortest input.
14576 The duration of the first input.
14580 @section mpdecimate
14582 Drop frames that do not differ greatly from the previous frame in
14583 order to reduce frame rate.
14585 The main use of this filter is for very-low-bitrate encoding
14586 (e.g. streaming over dialup modem), but it could in theory be used for
14587 fixing movies that were inverse-telecined incorrectly.
14589 A description of the accepted options follows.
14593 Set the maximum number of consecutive frames which can be dropped (if
14594 positive), or the minimum interval between dropped frames (if
14595 negative). If the value is 0, the frame is dropped disregarding the
14596 number of previous sequentially dropped frames.
14598 Default value is 0.
14603 Set the dropping threshold values.
14605 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14606 represent actual pixel value differences, so a threshold of 64
14607 corresponds to 1 unit of difference for each pixel, or the same spread
14608 out differently over the block.
14610 A frame is a candidate for dropping if no 8x8 blocks differ by more
14611 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14612 meaning the whole image) differ by more than a threshold of @option{lo}.
14614 Default value for @option{hi} is 64*12, default value for @option{lo} is
14615 64*5, and default value for @option{frac} is 0.33.
14621 Negate (invert) the input video.
14623 It accepts the following option:
14628 With value 1, it negates the alpha component, if present. Default value is 0.
14634 Denoise frames using Non-Local Means algorithm.
14636 Each pixel is adjusted by looking for other pixels with similar contexts. This
14637 context similarity is defined by comparing their surrounding patches of size
14638 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14641 Note that the research area defines centers for patches, which means some
14642 patches will be made of pixels outside that research area.
14644 The filter accepts the following options.
14648 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14651 Set patch size. Default is 7. Must be odd number in range [0, 99].
14654 Same as @option{p} but for chroma planes.
14656 The default value is @var{0} and means automatic.
14659 Set research size. Default is 15. Must be odd number in range [0, 99].
14662 Same as @option{r} but for chroma planes.
14664 The default value is @var{0} and means automatic.
14669 Deinterlace video using neural network edge directed interpolation.
14671 This filter accepts the following options:
14675 Mandatory option, without binary file filter can not work.
14676 Currently file can be found here:
14677 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14680 Set which frames to deinterlace, by default it is @code{all}.
14681 Can be @code{all} or @code{interlaced}.
14684 Set mode of operation.
14686 Can be one of the following:
14690 Use frame flags, both fields.
14692 Use frame flags, single field.
14694 Use top field only.
14696 Use bottom field only.
14698 Use both fields, top first.
14700 Use both fields, bottom first.
14704 Set which planes to process, by default filter process all frames.
14707 Set size of local neighborhood around each pixel, used by the predictor neural
14710 Can be one of the following:
14723 Set the number of neurons in predictor neural network.
14724 Can be one of the following:
14735 Controls the number of different neural network predictions that are blended
14736 together to compute the final output value. Can be @code{fast}, default or
14740 Set which set of weights to use in the predictor.
14741 Can be one of the following:
14745 weights trained to minimize absolute error
14747 weights trained to minimize squared error
14751 Controls whether or not the prescreener neural network is used to decide
14752 which pixels should be processed by the predictor neural network and which
14753 can be handled by simple cubic interpolation.
14754 The prescreener is trained to know whether cubic interpolation will be
14755 sufficient for a pixel or whether it should be predicted by the predictor nn.
14756 The computational complexity of the prescreener nn is much less than that of
14757 the predictor nn. Since most pixels can be handled by cubic interpolation,
14758 using the prescreener generally results in much faster processing.
14759 The prescreener is pretty accurate, so the difference between using it and not
14760 using it is almost always unnoticeable.
14762 Can be one of the following:
14772 Default is @code{new}.
14775 @subsection Commands
14776 This filter supports same @ref{commands} as options, excluding @var{weights} option.
14780 Force libavfilter not to use any of the specified pixel formats for the
14781 input to the next filter.
14783 It accepts the following parameters:
14787 A '|'-separated list of pixel format names, such as
14788 pix_fmts=yuv420p|monow|rgb24".
14792 @subsection Examples
14796 Force libavfilter to use a format different from @var{yuv420p} for the
14797 input to the vflip filter:
14799 noformat=pix_fmts=yuv420p,vflip
14803 Convert the input video to any of the formats not contained in the list:
14805 noformat=yuv420p|yuv444p|yuv410p
14811 Add noise on video input frame.
14813 The filter accepts the following options:
14821 Set noise seed for specific pixel component or all pixel components in case
14822 of @var{all_seed}. Default value is @code{123457}.
14824 @item all_strength, alls
14825 @item c0_strength, c0s
14826 @item c1_strength, c1s
14827 @item c2_strength, c2s
14828 @item c3_strength, c3s
14829 Set noise strength for specific pixel component or all pixel components in case
14830 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14832 @item all_flags, allf
14833 @item c0_flags, c0f
14834 @item c1_flags, c1f
14835 @item c2_flags, c2f
14836 @item c3_flags, c3f
14837 Set pixel component flags or set flags for all components if @var{all_flags}.
14838 Available values for component flags are:
14841 averaged temporal noise (smoother)
14843 mix random noise with a (semi)regular pattern
14845 temporal noise (noise pattern changes between frames)
14847 uniform noise (gaussian otherwise)
14851 @subsection Examples
14853 Add temporal and uniform noise to input video:
14855 noise=alls=20:allf=t+u
14860 Normalize RGB video (aka histogram stretching, contrast stretching).
14861 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14863 For each channel of each frame, the filter computes the input range and maps
14864 it linearly to the user-specified output range. The output range defaults
14865 to the full dynamic range from pure black to pure white.
14867 Temporal smoothing can be used on the input range to reduce flickering (rapid
14868 changes in brightness) caused when small dark or bright objects enter or leave
14869 the scene. This is similar to the auto-exposure (automatic gain control) on a
14870 video camera, and, like a video camera, it may cause a period of over- or
14871 under-exposure of the video.
14873 The R,G,B channels can be normalized independently, which may cause some
14874 color shifting, or linked together as a single channel, which prevents
14875 color shifting. Linked normalization preserves hue. Independent normalization
14876 does not, so it can be used to remove some color casts. Independent and linked
14877 normalization can be combined in any ratio.
14879 The normalize filter accepts the following options:
14884 Colors which define the output range. The minimum input value is mapped to
14885 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14886 The defaults are black and white respectively. Specifying white for
14887 @var{blackpt} and black for @var{whitept} will give color-inverted,
14888 normalized video. Shades of grey can be used to reduce the dynamic range
14889 (contrast). Specifying saturated colors here can create some interesting
14893 The number of previous frames to use for temporal smoothing. The input range
14894 of each channel is smoothed using a rolling average over the current frame
14895 and the @var{smoothing} previous frames. The default is 0 (no temporal
14899 Controls the ratio of independent (color shifting) channel normalization to
14900 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14901 independent. Defaults to 1.0 (fully independent).
14904 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14905 expensive no-op. Defaults to 1.0 (full strength).
14909 @subsection Commands
14910 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14911 The command accepts the same syntax of the corresponding option.
14913 If the specified expression is not valid, it is kept at its current
14916 @subsection Examples
14918 Stretch video contrast to use the full dynamic range, with no temporal
14919 smoothing; may flicker depending on the source content:
14921 normalize=blackpt=black:whitept=white:smoothing=0
14924 As above, but with 50 frames of temporal smoothing; flicker should be
14925 reduced, depending on the source content:
14927 normalize=blackpt=black:whitept=white:smoothing=50
14930 As above, but with hue-preserving linked channel normalization:
14932 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14935 As above, but with half strength:
14937 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14940 Map the darkest input color to red, the brightest input color to cyan:
14942 normalize=blackpt=red:whitept=cyan
14947 Pass the video source unchanged to the output.
14950 Optical Character Recognition
14952 This filter uses Tesseract for optical character recognition. To enable
14953 compilation of this filter, you need to configure FFmpeg with
14954 @code{--enable-libtesseract}.
14956 It accepts the following options:
14960 Set datapath to tesseract data. Default is to use whatever was
14961 set at installation.
14964 Set language, default is "eng".
14967 Set character whitelist.
14970 Set character blacklist.
14973 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14974 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14978 Apply a video transform using libopencv.
14980 To enable this filter, install the libopencv library and headers and
14981 configure FFmpeg with @code{--enable-libopencv}.
14983 It accepts the following parameters:
14988 The name of the libopencv filter to apply.
14990 @item filter_params
14991 The parameters to pass to the libopencv filter. If not specified, the default
14992 values are assumed.
14996 Refer to the official libopencv documentation for more precise
14998 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
15000 Several libopencv filters are supported; see the following subsections.
15005 Dilate an image by using a specific structuring element.
15006 It corresponds to the libopencv function @code{cvDilate}.
15008 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
15010 @var{struct_el} represents a structuring element, and has the syntax:
15011 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
15013 @var{cols} and @var{rows} represent the number of columns and rows of
15014 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
15015 point, and @var{shape} the shape for the structuring element. @var{shape}
15016 must be "rect", "cross", "ellipse", or "custom".
15018 If the value for @var{shape} is "custom", it must be followed by a
15019 string of the form "=@var{filename}". The file with name
15020 @var{filename} is assumed to represent a binary image, with each
15021 printable character corresponding to a bright pixel. When a custom
15022 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
15023 or columns and rows of the read file are assumed instead.
15025 The default value for @var{struct_el} is "3x3+0x0/rect".
15027 @var{nb_iterations} specifies the number of times the transform is
15028 applied to the image, and defaults to 1.
15032 # Use the default values
15035 # Dilate using a structuring element with a 5x5 cross, iterating two times
15036 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
15038 # Read the shape from the file diamond.shape, iterating two times.
15039 # The file diamond.shape may contain a pattern of characters like this
15045 # The specified columns and rows are ignored
15046 # but the anchor point coordinates are not
15047 ocv=dilate:0x0+2x2/custom=diamond.shape|2
15052 Erode an image by using a specific structuring element.
15053 It corresponds to the libopencv function @code{cvErode}.
15055 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
15056 with the same syntax and semantics as the @ref{dilate} filter.
15060 Smooth the input video.
15062 The filter takes the following parameters:
15063 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
15065 @var{type} is the type of smooth filter to apply, and must be one of
15066 the following values: "blur", "blur_no_scale", "median", "gaussian",
15067 or "bilateral". The default value is "gaussian".
15069 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
15070 depends on the smooth type. @var{param1} and
15071 @var{param2} accept integer positive values or 0. @var{param3} and
15072 @var{param4} accept floating point values.
15074 The default value for @var{param1} is 3. The default value for the
15075 other parameters is 0.
15077 These parameters correspond to the parameters assigned to the
15078 libopencv function @code{cvSmooth}.
15080 @section oscilloscope
15082 2D Video Oscilloscope.
15084 Useful to measure spatial impulse, step responses, chroma delays, etc.
15086 It accepts the following parameters:
15090 Set scope center x position.
15093 Set scope center y position.
15096 Set scope size, relative to frame diagonal.
15099 Set scope tilt/rotation.
15105 Set trace center x position.
15108 Set trace center y position.
15111 Set trace width, relative to width of frame.
15114 Set trace height, relative to height of frame.
15117 Set which components to trace. By default it traces first three components.
15120 Draw trace grid. By default is enabled.
15123 Draw some statistics. By default is enabled.
15126 Draw scope. By default is enabled.
15129 @subsection Commands
15130 This filter supports same @ref{commands} as options.
15131 The command accepts the same syntax of the corresponding option.
15133 If the specified expression is not valid, it is kept at its current
15136 @subsection Examples
15140 Inspect full first row of video frame.
15142 oscilloscope=x=0.5:y=0:s=1
15146 Inspect full last row of video frame.
15148 oscilloscope=x=0.5:y=1:s=1
15152 Inspect full 5th line of video frame of height 1080.
15154 oscilloscope=x=0.5:y=5/1080:s=1
15158 Inspect full last column of video frame.
15160 oscilloscope=x=1:y=0.5:s=1:t=1
15168 Overlay one video on top of another.
15170 It takes two inputs and has one output. The first input is the "main"
15171 video on which the second input is overlaid.
15173 It accepts the following parameters:
15175 A description of the accepted options follows.
15180 Set the expression for the x and y coordinates of the overlaid video
15181 on the main video. Default value is "0" for both expressions. In case
15182 the expression is invalid, it is set to a huge value (meaning that the
15183 overlay will not be displayed within the output visible area).
15186 See @ref{framesync}.
15189 Set when the expressions for @option{x}, and @option{y} are evaluated.
15191 It accepts the following values:
15194 only evaluate expressions once during the filter initialization or
15195 when a command is processed
15198 evaluate expressions for each incoming frame
15201 Default value is @samp{frame}.
15204 See @ref{framesync}.
15207 Set the format for the output video.
15209 It accepts the following values:
15212 force YUV420 output
15215 force YUV420p10 output
15218 force YUV422 output
15221 force YUV422p10 output
15224 force YUV444 output
15227 force packed RGB output
15230 force planar RGB output
15233 automatically pick format
15236 Default value is @samp{yuv420}.
15239 See @ref{framesync}.
15242 Set format of alpha of the overlaid video, it can be @var{straight} or
15243 @var{premultiplied}. Default is @var{straight}.
15246 The @option{x}, and @option{y} expressions can contain the following
15252 The main input width and height.
15256 The overlay input width and height.
15260 The computed values for @var{x} and @var{y}. They are evaluated for
15265 horizontal and vertical chroma subsample values of the output
15266 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15270 the number of input frame, starting from 0
15273 the position in the file of the input frame, NAN if unknown
15276 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15280 This filter also supports the @ref{framesync} options.
15282 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15283 when evaluation is done @emph{per frame}, and will evaluate to NAN
15284 when @option{eval} is set to @samp{init}.
15286 Be aware that frames are taken from each input video in timestamp
15287 order, hence, if their initial timestamps differ, it is a good idea
15288 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15289 have them begin in the same zero timestamp, as the example for
15290 the @var{movie} filter does.
15292 You can chain together more overlays but you should test the
15293 efficiency of such approach.
15295 @subsection Commands
15297 This filter supports the following commands:
15301 Modify the x and y of the overlay input.
15302 The command accepts the same syntax of the corresponding option.
15304 If the specified expression is not valid, it is kept at its current
15308 @subsection Examples
15312 Draw the overlay at 10 pixels from the bottom right corner of the main
15315 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15318 Using named options the example above becomes:
15320 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15324 Insert a transparent PNG logo in the bottom left corner of the input,
15325 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15327 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15331 Insert 2 different transparent PNG logos (second logo on bottom
15332 right corner) using the @command{ffmpeg} tool:
15334 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
15338 Add a transparent color layer on top of the main video; @code{WxH}
15339 must specify the size of the main input to the overlay filter:
15341 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15345 Play an original video and a filtered version (here with the deshake
15346 filter) side by side using the @command{ffplay} tool:
15348 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15351 The above command is the same as:
15353 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15357 Make a sliding overlay appearing from the left to the right top part of the
15358 screen starting since time 2:
15360 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15364 Compose output by putting two input videos side to side:
15366 ffmpeg -i left.avi -i right.avi -filter_complex "
15367 nullsrc=size=200x100 [background];
15368 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15369 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15370 [background][left] overlay=shortest=1 [background+left];
15371 [background+left][right] overlay=shortest=1:x=100 [left+right]
15376 Mask 10-20 seconds of a video by applying the delogo filter to a section
15378 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15379 -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]'
15384 Chain several overlays in cascade:
15386 nullsrc=s=200x200 [bg];
15387 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15388 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15389 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15390 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15391 [in3] null, [mid2] overlay=100:100 [out0]
15396 @anchor{overlay_cuda}
15397 @section overlay_cuda
15399 Overlay one video on top of another.
15401 This is the CUDA variant of the @ref{overlay} filter.
15402 It only accepts CUDA frames. The underlying input pixel formats have to match.
15404 It takes two inputs and has one output. The first input is the "main"
15405 video on which the second input is overlaid.
15407 It accepts the following parameters:
15412 Set the x and y coordinates of the overlaid video on the main video.
15413 Default value is "0" for both expressions.
15416 See @ref{framesync}.
15419 See @ref{framesync}.
15422 See @ref{framesync}.
15426 This filter also supports the @ref{framesync} options.
15430 Apply Overcomplete Wavelet denoiser.
15432 The filter accepts the following options:
15438 Larger depth values will denoise lower frequency components more, but
15439 slow down filtering.
15441 Must be an int in the range 8-16, default is @code{8}.
15443 @item luma_strength, ls
15446 Must be a double value in the range 0-1000, default is @code{1.0}.
15448 @item chroma_strength, cs
15449 Set chroma strength.
15451 Must be a double value in the range 0-1000, default is @code{1.0}.
15457 Add paddings to the input image, and place the original input at the
15458 provided @var{x}, @var{y} coordinates.
15460 It accepts the following parameters:
15465 Specify an expression for the size of the output image with the
15466 paddings added. If the value for @var{width} or @var{height} is 0, the
15467 corresponding input size is used for the output.
15469 The @var{width} expression can reference the value set by the
15470 @var{height} expression, and vice versa.
15472 The default value of @var{width} and @var{height} is 0.
15476 Specify the offsets to place the input image at within the padded area,
15477 with respect to the top/left border of the output image.
15479 The @var{x} expression can reference the value set by the @var{y}
15480 expression, and vice versa.
15482 The default value of @var{x} and @var{y} is 0.
15484 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15485 so the input image is centered on the padded area.
15488 Specify the color of the padded area. For the syntax of this option,
15489 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15490 manual,ffmpeg-utils}.
15492 The default value of @var{color} is "black".
15495 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15497 It accepts the following values:
15501 Only evaluate expressions once during the filter initialization or when
15502 a command is processed.
15505 Evaluate expressions for each incoming frame.
15509 Default value is @samp{init}.
15512 Pad to aspect instead to a resolution.
15516 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15517 options are expressions containing the following constants:
15522 The input video width and height.
15526 These are the same as @var{in_w} and @var{in_h}.
15530 The output width and height (the size of the padded area), as
15531 specified by the @var{width} and @var{height} expressions.
15535 These are the same as @var{out_w} and @var{out_h}.
15539 The x and y offsets as specified by the @var{x} and @var{y}
15540 expressions, or NAN if not yet specified.
15543 same as @var{iw} / @var{ih}
15546 input sample aspect ratio
15549 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15553 The horizontal and vertical chroma subsample values. For example for the
15554 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15557 @subsection Examples
15561 Add paddings with the color "violet" to the input video. The output video
15562 size is 640x480, and the top-left corner of the input video is placed at
15565 pad=640:480:0:40:violet
15568 The example above is equivalent to the following command:
15570 pad=width=640:height=480:x=0:y=40:color=violet
15574 Pad the input to get an output with dimensions increased by 3/2,
15575 and put the input video at the center of the padded area:
15577 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15581 Pad the input to get a squared output with size equal to the maximum
15582 value between the input width and height, and put the input video at
15583 the center of the padded area:
15585 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15589 Pad the input to get a final w/h ratio of 16:9:
15591 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15595 In case of anamorphic video, in order to set the output display aspect
15596 correctly, it is necessary to use @var{sar} in the expression,
15597 according to the relation:
15599 (ih * X / ih) * sar = output_dar
15600 X = output_dar / sar
15603 Thus the previous example needs to be modified to:
15605 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15609 Double the output size and put the input video in the bottom-right
15610 corner of the output padded area:
15612 pad="2*iw:2*ih:ow-iw:oh-ih"
15616 @anchor{palettegen}
15617 @section palettegen
15619 Generate one palette for a whole video stream.
15621 It accepts the following options:
15625 Set the maximum number of colors to quantize in the palette.
15626 Note: the palette will still contain 256 colors; the unused palette entries
15629 @item reserve_transparent
15630 Create a palette of 255 colors maximum and reserve the last one for
15631 transparency. Reserving the transparency color is useful for GIF optimization.
15632 If not set, the maximum of colors in the palette will be 256. You probably want
15633 to disable this option for a standalone image.
15636 @item transparency_color
15637 Set the color that will be used as background for transparency.
15640 Set statistics mode.
15642 It accepts the following values:
15645 Compute full frame histograms.
15647 Compute histograms only for the part that differs from previous frame. This
15648 might be relevant to give more importance to the moving part of your input if
15649 the background is static.
15651 Compute new histogram for each frame.
15654 Default value is @var{full}.
15657 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15658 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15659 color quantization of the palette. This information is also visible at
15660 @var{info} logging level.
15662 @subsection Examples
15666 Generate a representative palette of a given video using @command{ffmpeg}:
15668 ffmpeg -i input.mkv -vf palettegen palette.png
15672 @section paletteuse
15674 Use a palette to downsample an input video stream.
15676 The filter takes two inputs: one video stream and a palette. The palette must
15677 be a 256 pixels image.
15679 It accepts the following options:
15683 Select dithering mode. Available algorithms are:
15686 Ordered 8x8 bayer dithering (deterministic)
15688 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15689 Note: this dithering is sometimes considered "wrong" and is included as a
15691 @item floyd_steinberg
15692 Floyd and Steingberg dithering (error diffusion)
15694 Frankie Sierra dithering v2 (error diffusion)
15696 Frankie Sierra dithering v2 "Lite" (error diffusion)
15699 Default is @var{sierra2_4a}.
15702 When @var{bayer} dithering is selected, this option defines the scale of the
15703 pattern (how much the crosshatch pattern is visible). A low value means more
15704 visible pattern for less banding, and higher value means less visible pattern
15705 at the cost of more banding.
15707 The option must be an integer value in the range [0,5]. Default is @var{2}.
15710 If set, define the zone to process
15714 Only the changing rectangle will be reprocessed. This is similar to GIF
15715 cropping/offsetting compression mechanism. This option can be useful for speed
15716 if only a part of the image is changing, and has use cases such as limiting the
15717 scope of the error diffusal @option{dither} to the rectangle that bounds the
15718 moving scene (it leads to more deterministic output if the scene doesn't change
15719 much, and as a result less moving noise and better GIF compression).
15722 Default is @var{none}.
15725 Take new palette for each output frame.
15727 @item alpha_threshold
15728 Sets the alpha threshold for transparency. Alpha values above this threshold
15729 will be treated as completely opaque, and values below this threshold will be
15730 treated as completely transparent.
15732 The option must be an integer value in the range [0,255]. Default is @var{128}.
15735 @subsection Examples
15739 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15740 using @command{ffmpeg}:
15742 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15746 @section perspective
15748 Correct perspective of video not recorded perpendicular to the screen.
15750 A description of the accepted parameters follows.
15761 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15762 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15763 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15764 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15765 then the corners of the source will be sent to the specified coordinates.
15767 The expressions can use the following variables:
15772 the width and height of video frame.
15776 Output frame count.
15779 @item interpolation
15780 Set interpolation for perspective correction.
15782 It accepts the following values:
15788 Default value is @samp{linear}.
15791 Set interpretation of coordinate options.
15793 It accepts the following values:
15797 Send point in the source specified by the given coordinates to
15798 the corners of the destination.
15800 @item 1, destination
15802 Send the corners of the source to the point in the destination specified
15803 by the given coordinates.
15805 Default value is @samp{source}.
15809 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15811 It accepts the following values:
15814 only evaluate expressions once during the filter initialization or
15815 when a command is processed
15818 evaluate expressions for each incoming frame
15821 Default value is @samp{init}.
15826 Delay interlaced video by one field time so that the field order changes.
15828 The intended use is to fix PAL movies that have been captured with the
15829 opposite field order to the film-to-video transfer.
15831 A description of the accepted parameters follows.
15837 It accepts the following values:
15840 Capture field order top-first, transfer bottom-first.
15841 Filter will delay the bottom field.
15844 Capture field order bottom-first, transfer top-first.
15845 Filter will delay the top field.
15848 Capture and transfer with the same field order. This mode only exists
15849 for the documentation of the other options to refer to, but if you
15850 actually select it, the filter will faithfully do nothing.
15853 Capture field order determined automatically by field flags, transfer
15855 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15856 basis using field flags. If no field information is available,
15857 then this works just like @samp{u}.
15860 Capture unknown or varying, transfer opposite.
15861 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15862 analyzing the images and selecting the alternative that produces best
15863 match between the fields.
15866 Capture top-first, transfer unknown or varying.
15867 Filter selects among @samp{t} and @samp{p} using image analysis.
15870 Capture bottom-first, transfer unknown or varying.
15871 Filter selects among @samp{b} and @samp{p} using image analysis.
15874 Capture determined by field flags, transfer unknown or varying.
15875 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15876 image analysis. If no field information is available, then this works just
15877 like @samp{U}. This is the default mode.
15880 Both capture and transfer unknown or varying.
15881 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15885 @subsection Commands
15887 This filter supports the all above options as @ref{commands}.
15889 @section photosensitivity
15890 Reduce various flashes in video, so to help users with epilepsy.
15892 It accepts the following options:
15895 Set how many frames to use when filtering. Default is 30.
15898 Set detection threshold factor. Default is 1.
15902 Set how many pixels to skip when sampling frames. Default is 1.
15903 Allowed range is from 1 to 1024.
15906 Leave frames unchanged. Default is disabled.
15909 @section pixdesctest
15911 Pixel format descriptor test filter, mainly useful for internal
15912 testing. The output video should be equal to the input video.
15916 format=monow, pixdesctest
15919 can be used to test the monowhite pixel format descriptor definition.
15923 Display sample values of color channels. Mainly useful for checking color
15924 and levels. Minimum supported resolution is 640x480.
15926 The filters accept the following options:
15930 Set scope X position, relative offset on X axis.
15933 Set scope Y position, relative offset on Y axis.
15942 Set window opacity. This window also holds statistics about pixel area.
15945 Set window X position, relative offset on X axis.
15948 Set window Y position, relative offset on Y axis.
15953 Enable the specified chain of postprocessing subfilters using libpostproc. This
15954 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15955 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15956 Each subfilter and some options have a short and a long name that can be used
15957 interchangeably, i.e. dr/dering are the same.
15959 The filters accept the following options:
15963 Set postprocessing subfilters string.
15966 All subfilters share common options to determine their scope:
15970 Honor the quality commands for this subfilter.
15973 Do chrominance filtering, too (default).
15976 Do luminance filtering only (no chrominance).
15979 Do chrominance filtering only (no luminance).
15982 These options can be appended after the subfilter name, separated by a '|'.
15984 Available subfilters are:
15987 @item hb/hdeblock[|difference[|flatness]]
15988 Horizontal deblocking filter
15991 Difference factor where higher values mean more deblocking (default: @code{32}).
15993 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15996 @item vb/vdeblock[|difference[|flatness]]
15997 Vertical deblocking filter
16000 Difference factor where higher values mean more deblocking (default: @code{32}).
16002 Flatness threshold where lower values mean more deblocking (default: @code{39}).
16005 @item ha/hadeblock[|difference[|flatness]]
16006 Accurate horizontal deblocking filter
16009 Difference factor where higher values mean more deblocking (default: @code{32}).
16011 Flatness threshold where lower values mean more deblocking (default: @code{39}).
16014 @item va/vadeblock[|difference[|flatness]]
16015 Accurate vertical deblocking filter
16018 Difference factor where higher values mean more deblocking (default: @code{32}).
16020 Flatness threshold where lower values mean more deblocking (default: @code{39}).
16024 The horizontal and vertical deblocking filters share the difference and
16025 flatness values so you cannot set different horizontal and vertical
16029 @item h1/x1hdeblock
16030 Experimental horizontal deblocking filter
16032 @item v1/x1vdeblock
16033 Experimental vertical deblocking filter
16038 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
16041 larger -> stronger filtering
16043 larger -> stronger filtering
16045 larger -> stronger filtering
16048 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
16051 Stretch luminance to @code{0-255}.
16054 @item lb/linblenddeint
16055 Linear blend deinterlacing filter that deinterlaces the given block by
16056 filtering all lines with a @code{(1 2 1)} filter.
16058 @item li/linipoldeint
16059 Linear interpolating deinterlacing filter that deinterlaces the given block by
16060 linearly interpolating every second line.
16062 @item ci/cubicipoldeint
16063 Cubic interpolating deinterlacing filter deinterlaces the given block by
16064 cubically interpolating every second line.
16066 @item md/mediandeint
16067 Median deinterlacing filter that deinterlaces the given block by applying a
16068 median filter to every second line.
16070 @item fd/ffmpegdeint
16071 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
16072 second line with a @code{(-1 4 2 4 -1)} filter.
16075 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
16076 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
16078 @item fq/forceQuant[|quantizer]
16079 Overrides the quantizer table from the input with the constant quantizer you
16087 Default pp filter combination (@code{hb|a,vb|a,dr|a})
16090 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
16093 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
16096 @subsection Examples
16100 Apply horizontal and vertical deblocking, deringing and automatic
16101 brightness/contrast:
16107 Apply default filters without brightness/contrast correction:
16113 Apply default filters and temporal denoiser:
16115 pp=default/tmpnoise|1|2|3
16119 Apply deblocking on luminance only, and switch vertical deblocking on or off
16120 automatically depending on available CPU time:
16127 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
16128 similar to spp = 6 with 7 point DCT, where only the center sample is
16131 The filter accepts the following options:
16135 Force a constant quantization parameter. It accepts an integer in range
16136 0 to 63. If not set, the filter will use the QP from the video stream
16140 Set thresholding mode. Available modes are:
16144 Set hard thresholding.
16146 Set soft thresholding (better de-ringing effect, but likely blurrier).
16148 Set medium thresholding (good results, default).
16152 @section premultiply
16153 Apply alpha premultiply effect to input video stream using first plane
16154 of second stream as alpha.
16156 Both streams must have same dimensions and same pixel format.
16158 The filter accepts the following option:
16162 Set which planes will be processed, unprocessed planes will be copied.
16163 By default value 0xf, all planes will be processed.
16166 Do not require 2nd input for processing, instead use alpha plane from input stream.
16170 Apply prewitt 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.
16186 @subsection Commands
16188 This filter supports the all above options as @ref{commands}.
16190 @section pseudocolor
16192 Alter frame colors in video with pseudocolors.
16194 This filter accepts the following options:
16198 set pixel first component expression
16201 set pixel second component expression
16204 set pixel third component expression
16207 set pixel fourth component expression, corresponds to the alpha component
16210 set component to use as base for altering colors
16213 Each of them specifies the expression to use for computing the lookup table for
16214 the corresponding pixel component values.
16216 The expressions can contain the following constants and functions:
16221 The input width and height.
16224 The input value for the pixel component.
16226 @item ymin, umin, vmin, amin
16227 The minimum allowed component value.
16229 @item ymax, umax, vmax, amax
16230 The maximum allowed component value.
16233 All expressions default to "val".
16235 @subsection Examples
16239 Change too high luma values to gradient:
16241 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'"
16247 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16248 Ratio) between two input videos.
16250 This filter takes in input two input videos, the first input is
16251 considered the "main" source and is passed unchanged to the
16252 output. The second input is used as a "reference" video for computing
16255 Both video inputs must have the same resolution and pixel format for
16256 this filter to work correctly. Also it assumes that both inputs
16257 have the same number of frames, which are compared one by one.
16259 The obtained average PSNR is printed through the logging system.
16261 The filter stores the accumulated MSE (mean squared error) of each
16262 frame, and at the end of the processing it is averaged across all frames
16263 equally, and the following formula is applied to obtain the PSNR:
16266 PSNR = 10*log10(MAX^2/MSE)
16269 Where MAX is the average of the maximum values of each component of the
16272 The description of the accepted parameters follows.
16275 @item stats_file, f
16276 If specified the filter will use the named file to save the PSNR of
16277 each individual frame. When filename equals "-" the data is sent to
16280 @item stats_version
16281 Specifies which version of the stats file format to use. Details of
16282 each format are written below.
16283 Default value is 1.
16285 @item stats_add_max
16286 Determines whether the max value is output to the stats log.
16287 Default value is 0.
16288 Requires stats_version >= 2. If this is set and stats_version < 2,
16289 the filter will return an error.
16292 This filter also supports the @ref{framesync} options.
16294 The file printed if @var{stats_file} is selected, contains a sequence of
16295 key/value pairs of the form @var{key}:@var{value} for each compared
16298 If a @var{stats_version} greater than 1 is specified, a header line precedes
16299 the list of per-frame-pair stats, with key value pairs following the frame
16300 format with the following parameters:
16303 @item psnr_log_version
16304 The version of the log file format. Will match @var{stats_version}.
16307 A comma separated list of the per-frame-pair parameters included in
16311 A description of each shown per-frame-pair parameter follows:
16315 sequential number of the input frame, starting from 1
16318 Mean Square Error pixel-by-pixel average difference of the compared
16319 frames, averaged over all the image components.
16321 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16322 Mean Square Error pixel-by-pixel average difference of the compared
16323 frames for the component specified by the suffix.
16325 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16326 Peak Signal to Noise ratio of the compared frames for the component
16327 specified by the suffix.
16329 @item max_avg, max_y, max_u, max_v
16330 Maximum allowed value for each channel, and average over all
16334 @subsection Examples
16339 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16340 [main][ref] psnr="stats_file=stats.log" [out]
16343 On this example the input file being processed is compared with the
16344 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16345 is stored in @file{stats.log}.
16348 Another example with different containers:
16350 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]psnr" -f null -
16357 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16358 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16361 The pullup filter is designed to take advantage of future context in making
16362 its decisions. This filter is stateless in the sense that it does not lock
16363 onto a pattern to follow, but it instead looks forward to the following
16364 fields in order to identify matches and rebuild progressive frames.
16366 To produce content with an even framerate, insert the fps filter after
16367 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16368 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16370 The filter accepts the following options:
16377 These options set the amount of "junk" to ignore at the left, right, top, and
16378 bottom of the image, respectively. Left and right are in units of 8 pixels,
16379 while top and bottom are in units of 2 lines.
16380 The default is 8 pixels on each side.
16383 Set the strict breaks. Setting this option to 1 will reduce the chances of
16384 filter generating an occasional mismatched frame, but it may also cause an
16385 excessive number of frames to be dropped during high motion sequences.
16386 Conversely, setting it to -1 will make filter match fields more easily.
16387 This may help processing of video where there is slight blurring between
16388 the fields, but may also cause there to be interlaced frames in the output.
16389 Default value is @code{0}.
16392 Set the metric plane to use. It accepts the following values:
16398 Use chroma blue plane.
16401 Use chroma red plane.
16404 This option may be set to use chroma plane instead of the default luma plane
16405 for doing filter's computations. This may improve accuracy on very clean
16406 source material, but more likely will decrease accuracy, especially if there
16407 is chroma noise (rainbow effect) or any grayscale video.
16408 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16409 load and make pullup usable in realtime on slow machines.
16412 For best results (without duplicated frames in the output file) it is
16413 necessary to change the output frame rate. For example, to inverse
16414 telecine NTSC input:
16416 ffmpeg -i input -vf pullup -r 24000/1001 ...
16421 Change video quantization parameters (QP).
16423 The filter accepts the following option:
16427 Set expression for quantization parameter.
16430 The expression is evaluated through the eval API and can contain, among others,
16431 the following constants:
16435 1 if index is not 129, 0 otherwise.
16438 Sequential index starting from -129 to 128.
16441 @subsection Examples
16445 Some equation like:
16453 Flush video frames from internal cache of frames into a random order.
16454 No frame is discarded.
16455 Inspired by @ref{frei0r} nervous filter.
16459 Set size in number of frames of internal cache, in range from @code{2} to
16460 @code{512}. Default is @code{30}.
16463 Set seed for random number generator, must be an integer included between
16464 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16465 less than @code{0}, the filter will try to use a good random seed on a
16469 @section readeia608
16471 Read closed captioning (EIA-608) information from the top lines of a video frame.
16473 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16474 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16475 with EIA-608 data (starting from 0). A description of each metadata value follows:
16478 @item lavfi.readeia608.X.cc
16479 The two bytes stored as EIA-608 data (printed in hexadecimal).
16481 @item lavfi.readeia608.X.line
16482 The number of the line on which the EIA-608 data was identified and read.
16485 This filter accepts the following options:
16489 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16492 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16495 Set the ratio of width reserved for sync code detection.
16496 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16499 Enable checking the parity bit. In the event of a parity error, the filter will output
16500 @code{0x00} for that character. Default is false.
16503 Lowpass lines prior to further processing. Default is enabled.
16506 @subsection Commands
16508 This filter supports the all above options as @ref{commands}.
16510 @subsection Examples
16514 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16516 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
16522 Read vertical interval timecode (VITC) information from the top lines of a
16525 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16526 timecode value, if a valid timecode has been detected. Further metadata key
16527 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16528 timecode data has been found or not.
16530 This filter accepts the following options:
16534 Set the maximum number of lines to scan for VITC data. If the value is set to
16535 @code{-1} the full video frame is scanned. Default is @code{45}.
16538 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16539 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16542 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16543 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16546 @subsection Examples
16550 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16551 draw @code{--:--:--:--} as a placeholder:
16553 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16559 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16561 Destination pixel at position (X, Y) will be picked from source (x, y) position
16562 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16563 value for pixel will be used for destination pixel.
16565 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16566 will have Xmap/Ymap video stream dimensions.
16567 Xmap and Ymap input video streams are 16bit depth, single channel.
16571 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16572 Default is @code{color}.
16575 Specify the color of the unmapped pixels. For the syntax of this option,
16576 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16577 manual,ffmpeg-utils}. Default color is @code{black}.
16580 @section removegrain
16582 The removegrain filter is a spatial denoiser for progressive video.
16586 Set mode for the first plane.
16589 Set mode for the second plane.
16592 Set mode for the third plane.
16595 Set mode for the fourth plane.
16598 Range of mode is from 0 to 24. Description of each mode follows:
16602 Leave input plane unchanged. Default.
16605 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16608 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16611 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16614 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16615 This is equivalent to a median filter.
16618 Line-sensitive clipping giving the minimal change.
16621 Line-sensitive clipping, intermediate.
16624 Line-sensitive clipping, intermediate.
16627 Line-sensitive clipping, intermediate.
16630 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16633 Replaces the target pixel with the closest neighbour.
16636 [1 2 1] horizontal and vertical kernel blur.
16642 Bob mode, interpolates top field from the line where the neighbours
16643 pixels are the closest.
16646 Bob mode, interpolates bottom field from the line where the neighbours
16647 pixels are the closest.
16650 Bob mode, interpolates top field. Same as 13 but with a more complicated
16651 interpolation formula.
16654 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16655 interpolation formula.
16658 Clips the pixel with the minimum and maximum of respectively the maximum and
16659 minimum of each pair of opposite neighbour pixels.
16662 Line-sensitive clipping using opposite neighbours whose greatest distance from
16663 the current pixel is minimal.
16666 Replaces the pixel with the average of its 8 neighbours.
16669 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16672 Clips pixels using the averages of opposite neighbour.
16675 Same as mode 21 but simpler and faster.
16678 Small edge and halo removal, but reputed useless.
16684 @section removelogo
16686 Suppress a TV station logo, using an image file to determine which
16687 pixels comprise the logo. It works by filling in the pixels that
16688 comprise the logo with neighboring pixels.
16690 The filter accepts the following options:
16694 Set the filter bitmap file, which can be any image format supported by
16695 libavformat. The width and height of the image file must match those of the
16696 video stream being processed.
16699 Pixels in the provided bitmap image with a value of zero are not
16700 considered part of the logo, non-zero pixels are considered part of
16701 the logo. If you use white (255) for the logo and black (0) for the
16702 rest, you will be safe. For making the filter bitmap, it is
16703 recommended to take a screen capture of a black frame with the logo
16704 visible, and then using a threshold filter followed by the erode
16705 filter once or twice.
16707 If needed, little splotches can be fixed manually. Remember that if
16708 logo pixels are not covered, the filter quality will be much
16709 reduced. Marking too many pixels as part of the logo does not hurt as
16710 much, but it will increase the amount of blurring needed to cover over
16711 the image and will destroy more information than necessary, and extra
16712 pixels will slow things down on a large logo.
16714 @section repeatfields
16716 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16717 fields based on its value.
16721 Reverse a video clip.
16723 Warning: This filter requires memory to buffer the entire clip, so trimming
16726 @subsection Examples
16730 Take the first 5 seconds of a clip, and reverse it.
16737 Shift R/G/B/A pixels horizontally and/or vertically.
16739 The filter accepts the following options:
16742 Set amount to shift red horizontally.
16744 Set amount to shift red vertically.
16746 Set amount to shift green horizontally.
16748 Set amount to shift green vertically.
16750 Set amount to shift blue horizontally.
16752 Set amount to shift blue vertically.
16754 Set amount to shift alpha horizontally.
16756 Set amount to shift alpha vertically.
16758 Set edge mode, can be @var{smear}, default, or @var{warp}.
16761 @subsection Commands
16763 This filter supports the all above options as @ref{commands}.
16766 Apply roberts cross operator to input video stream.
16768 The filter accepts the following option:
16772 Set which planes will be processed, unprocessed planes will be copied.
16773 By default value 0xf, all planes will be processed.
16776 Set value which will be multiplied with filtered result.
16779 Set value which will be added to filtered result.
16782 @subsection Commands
16784 This filter supports the all above options as @ref{commands}.
16788 Rotate video by an arbitrary angle expressed in radians.
16790 The filter accepts the following options:
16792 A description of the optional parameters follows.
16795 Set an expression for the angle by which to rotate the input video
16796 clockwise, expressed as a number of radians. A negative value will
16797 result in a counter-clockwise rotation. By default it is set to "0".
16799 This expression is evaluated for each frame.
16802 Set the output width expression, default value is "iw".
16803 This expression is evaluated just once during configuration.
16806 Set the output height expression, default value is "ih".
16807 This expression is evaluated just once during configuration.
16810 Enable bilinear interpolation if set to 1, a value of 0 disables
16811 it. Default value is 1.
16814 Set the color used to fill the output area not covered by the rotated
16815 image. For the general syntax of this option, check the
16816 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16817 If the special value "none" is selected then no
16818 background is printed (useful for example if the background is never shown).
16820 Default value is "black".
16823 The expressions for the angle and the output size can contain the
16824 following constants and functions:
16828 sequential number of the input frame, starting from 0. It is always NAN
16829 before the first frame is filtered.
16832 time in seconds of the input frame, it is set to 0 when the filter is
16833 configured. It is always NAN before the first frame is filtered.
16837 horizontal and vertical chroma subsample values. For example for the
16838 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16842 the input video width and height
16846 the output width and height, that is the size of the padded area as
16847 specified by the @var{width} and @var{height} expressions
16851 the minimal width/height required for completely containing the input
16852 video rotated by @var{a} radians.
16854 These are only available when computing the @option{out_w} and
16855 @option{out_h} expressions.
16858 @subsection Examples
16862 Rotate the input by PI/6 radians clockwise:
16868 Rotate the input by PI/6 radians counter-clockwise:
16874 Rotate the input by 45 degrees clockwise:
16880 Apply a constant rotation with period T, starting from an angle of PI/3:
16882 rotate=PI/3+2*PI*t/T
16886 Make the input video rotation oscillating with a period of T
16887 seconds and an amplitude of A radians:
16889 rotate=A*sin(2*PI/T*t)
16893 Rotate the video, output size is chosen so that the whole rotating
16894 input video is always completely contained in the output:
16896 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16900 Rotate the video, reduce the output size so that no background is ever
16903 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16907 @subsection Commands
16909 The filter supports the following commands:
16913 Set the angle expression.
16914 The command accepts the same syntax of the corresponding option.
16916 If the specified expression is not valid, it is kept at its current
16922 Apply Shape Adaptive Blur.
16924 The filter accepts the following options:
16927 @item luma_radius, lr
16928 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16929 value is 1.0. A greater value will result in a more blurred image, and
16930 in slower processing.
16932 @item luma_pre_filter_radius, lpfr
16933 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16936 @item luma_strength, ls
16937 Set luma maximum difference between pixels to still be considered, must
16938 be a value in the 0.1-100.0 range, default value is 1.0.
16940 @item chroma_radius, cr
16941 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16942 greater value will result in a more blurred image, and in slower
16945 @item chroma_pre_filter_radius, cpfr
16946 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16948 @item chroma_strength, cs
16949 Set chroma maximum difference between pixels to still be considered,
16950 must be a value in the -0.9-100.0 range.
16953 Each chroma option value, if not explicitly specified, is set to the
16954 corresponding luma option value.
16959 Scale (resize) the input video, using the libswscale library.
16961 The scale filter forces the output display aspect ratio to be the same
16962 of the input, by changing the output sample aspect ratio.
16964 If the input image format is different from the format requested by
16965 the next filter, the scale filter will convert the input to the
16968 @subsection Options
16969 The filter accepts the following options, or any of the options
16970 supported by the libswscale scaler.
16972 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16973 the complete list of scaler options.
16978 Set the output video dimension expression. Default value is the input
16981 If the @var{width} or @var{w} value is 0, the input width is used for
16982 the output. If the @var{height} or @var{h} value is 0, the input height
16983 is used for the output.
16985 If one and only one of the values is -n with n >= 1, the scale filter
16986 will use a value that maintains the aspect ratio of the input image,
16987 calculated from the other specified dimension. After that it will,
16988 however, make sure that the calculated dimension is divisible by n and
16989 adjust the value if necessary.
16991 If both values are -n with n >= 1, the behavior will be identical to
16992 both values being set to 0 as previously detailed.
16994 See below for the list of accepted constants for use in the dimension
16998 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
17002 Only evaluate expressions once during the filter initialization or when a command is processed.
17005 Evaluate expressions for each incoming frame.
17009 Default value is @samp{init}.
17013 Set the interlacing mode. It accepts the following values:
17017 Force interlaced aware scaling.
17020 Do not apply interlaced scaling.
17023 Select interlaced aware scaling depending on whether the source frames
17024 are flagged as interlaced or not.
17027 Default value is @samp{0}.
17030 Set libswscale scaling flags. See
17031 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
17032 complete list of values. If not explicitly specified the filter applies
17036 @item param0, param1
17037 Set libswscale input parameters for scaling algorithms that need them. See
17038 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
17039 complete documentation. If not explicitly specified the filter applies
17045 Set the video size. For the syntax of this option, check the
17046 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17048 @item in_color_matrix
17049 @item out_color_matrix
17050 Set in/output YCbCr color space type.
17052 This allows the autodetected value to be overridden as well as allows forcing
17053 a specific value used for the output and encoder.
17055 If not specified, the color space type depends on the pixel format.
17061 Choose automatically.
17064 Format conforming to International Telecommunication Union (ITU)
17065 Recommendation BT.709.
17068 Set color space conforming to the United States Federal Communications
17069 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
17074 Set color space conforming to:
17078 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
17081 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
17084 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
17089 Set color space conforming to SMPTE ST 240:1999.
17092 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
17097 Set in/output YCbCr sample range.
17099 This allows the autodetected value to be overridden as well as allows forcing
17100 a specific value used for the output and encoder. If not specified, the
17101 range depends on the pixel format. Possible values:
17105 Choose automatically.
17108 Set full range (0-255 in case of 8-bit luma).
17110 @item mpeg/limited/tv
17111 Set "MPEG" range (16-235 in case of 8-bit luma).
17114 @item force_original_aspect_ratio
17115 Enable decreasing or increasing output video width or height if necessary to
17116 keep the original aspect ratio. Possible values:
17120 Scale the video as specified and disable this feature.
17123 The output video dimensions will automatically be decreased if needed.
17126 The output video dimensions will automatically be increased if needed.
17130 One useful instance of this option is that when you know a specific device's
17131 maximum allowed resolution, you can use this to limit the output video to
17132 that, while retaining the aspect ratio. For example, device A allows
17133 1280x720 playback, and your video is 1920x800. Using this option (set it to
17134 decrease) and specifying 1280x720 to the command line makes the output
17137 Please note that this is a different thing than specifying -1 for @option{w}
17138 or @option{h}, you still need to specify the output resolution for this option
17141 @item force_divisible_by
17142 Ensures that both the output dimensions, width and height, are divisible by the
17143 given integer when used together with @option{force_original_aspect_ratio}. This
17144 works similar to using @code{-n} in the @option{w} and @option{h} options.
17146 This option respects the value set for @option{force_original_aspect_ratio},
17147 increasing or decreasing the resolution accordingly. The video's aspect ratio
17148 may be slightly modified.
17150 This option can be handy if you need to have a video fit within or exceed
17151 a defined resolution using @option{force_original_aspect_ratio} but also have
17152 encoder restrictions on width or height divisibility.
17156 The values of the @option{w} and @option{h} options are expressions
17157 containing the following constants:
17162 The input width and height
17166 These are the same as @var{in_w} and @var{in_h}.
17170 The output (scaled) width and height
17174 These are the same as @var{out_w} and @var{out_h}
17177 The same as @var{iw} / @var{ih}
17180 input sample aspect ratio
17183 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17187 horizontal and vertical input chroma subsample values. For example for the
17188 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17192 horizontal and vertical output chroma subsample values. For example for the
17193 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17196 The (sequential) number of the input frame, starting from 0.
17197 Only available with @code{eval=frame}.
17200 The presentation timestamp of the input frame, expressed as a number of
17201 seconds. Only available with @code{eval=frame}.
17204 The position (byte offset) of the frame in the input stream, or NaN if
17205 this information is unavailable and/or meaningless (for example in case of synthetic video).
17206 Only available with @code{eval=frame}.
17209 @subsection Examples
17213 Scale the input video to a size of 200x100
17218 This is equivalent to:
17229 Specify a size abbreviation for the output size:
17234 which can also be written as:
17240 Scale the input to 2x:
17242 scale=w=2*iw:h=2*ih
17246 The above is the same as:
17248 scale=2*in_w:2*in_h
17252 Scale the input to 2x with forced interlaced scaling:
17254 scale=2*iw:2*ih:interl=1
17258 Scale the input to half size:
17260 scale=w=iw/2:h=ih/2
17264 Increase the width, and set the height to the same size:
17270 Seek Greek harmony:
17277 Increase the height, and set the width to 3/2 of the height:
17279 scale=w=3/2*oh:h=3/5*ih
17283 Increase the size, making the size a multiple of the chroma
17286 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17290 Increase the width to a maximum of 500 pixels,
17291 keeping the same aspect ratio as the input:
17293 scale=w='min(500\, iw*3/2):h=-1'
17297 Make pixels square by combining scale and setsar:
17299 scale='trunc(ih*dar):ih',setsar=1/1
17303 Make pixels square by combining scale and setsar,
17304 making sure the resulting resolution is even (required by some codecs):
17306 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17310 @subsection Commands
17312 This filter supports the following commands:
17316 Set the output video dimension expression.
17317 The command accepts the same syntax of the corresponding option.
17319 If the specified expression is not valid, it is kept at its current
17325 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17326 format conversion on CUDA video frames. Setting the output width and height
17327 works in the same way as for the @var{scale} filter.
17329 The following additional options are accepted:
17332 The pixel format of the output CUDA frames. If set to the string "same" (the
17333 default), the input format will be kept. Note that automatic format negotiation
17334 and conversion is not yet supported for hardware frames
17337 The interpolation algorithm used for resizing. One of the following:
17344 @item cubic2p_bspline
17345 2-parameter cubic (B=1, C=0)
17347 @item cubic2p_catmullrom
17348 2-parameter cubic (B=0, C=1/2)
17350 @item cubic2p_b05c03
17351 2-parameter cubic (B=1/2, C=3/10)
17359 @item force_original_aspect_ratio
17360 Enable decreasing or increasing output video width or height if necessary to
17361 keep the original aspect ratio. Possible values:
17365 Scale the video as specified and disable this feature.
17368 The output video dimensions will automatically be decreased if needed.
17371 The output video dimensions will automatically be increased if needed.
17375 One useful instance of this option is that when you know a specific device's
17376 maximum allowed resolution, you can use this to limit the output video to
17377 that, while retaining the aspect ratio. For example, device A allows
17378 1280x720 playback, and your video is 1920x800. Using this option (set it to
17379 decrease) and specifying 1280x720 to the command line makes the output
17382 Please note that this is a different thing than specifying -1 for @option{w}
17383 or @option{h}, you still need to specify the output resolution for this option
17386 @item force_divisible_by
17387 Ensures that both the output dimensions, width and height, are divisible by the
17388 given integer when used together with @option{force_original_aspect_ratio}. This
17389 works similar to using @code{-n} in the @option{w} and @option{h} options.
17391 This option respects the value set for @option{force_original_aspect_ratio},
17392 increasing or decreasing the resolution accordingly. The video's aspect ratio
17393 may be slightly modified.
17395 This option can be handy if you need to have a video fit within or exceed
17396 a defined resolution using @option{force_original_aspect_ratio} but also have
17397 encoder restrictions on width or height divisibility.
17403 Scale (resize) the input video, based on a reference video.
17405 See the scale filter for available options, scale2ref supports the same but
17406 uses the reference video instead of the main input as basis. scale2ref also
17407 supports the following additional constants for the @option{w} and
17408 @option{h} options:
17413 The main input video's width and height
17416 The same as @var{main_w} / @var{main_h}
17419 The main input video's sample aspect ratio
17421 @item main_dar, mdar
17422 The main input video's display aspect ratio. Calculated from
17423 @code{(main_w / main_h) * main_sar}.
17427 The main input video's horizontal and vertical chroma subsample values.
17428 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17432 The (sequential) number of the main input frame, starting from 0.
17433 Only available with @code{eval=frame}.
17436 The presentation timestamp of the main input frame, expressed as a number of
17437 seconds. Only available with @code{eval=frame}.
17440 The position (byte offset) of the frame in the main input stream, or NaN if
17441 this information is unavailable and/or meaningless (for example in case of synthetic video).
17442 Only available with @code{eval=frame}.
17445 @subsection Examples
17449 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17451 'scale2ref[b][a];[a][b]overlay'
17455 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17457 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17461 @subsection Commands
17463 This filter supports the following commands:
17467 Set the output video dimension expression.
17468 The command accepts the same syntax of the corresponding option.
17470 If the specified expression is not valid, it is kept at its current
17475 Scroll input video horizontally and/or vertically by constant speed.
17477 The filter accepts the following options:
17479 @item horizontal, h
17480 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17481 Negative values changes scrolling direction.
17484 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17485 Negative values changes scrolling direction.
17488 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17491 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17494 @subsection Commands
17496 This filter supports the following @ref{commands}:
17498 @item horizontal, h
17499 Set the horizontal scrolling speed.
17501 Set the vertical scrolling speed.
17507 Detect video scene change.
17509 This filter sets frame metadata with mafd between frame, the scene score, and
17510 forward the frame to the next filter, so they can use these metadata to detect
17511 scene change or others.
17513 In addition, this filter logs a message and sets frame metadata when it detects
17514 a scene change by @option{threshold}.
17516 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17518 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17519 to detect scene change.
17521 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17522 detect scene change with @option{threshold}.
17524 The filter accepts the following options:
17528 Set the scene change detection threshold as a percentage of maximum change. Good
17529 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17532 Default value is @code{10.}.
17535 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17536 You can enable it if you want to get snapshot of scene change frames only.
17539 @anchor{selectivecolor}
17540 @section selectivecolor
17542 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17543 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17544 by the "purity" of the color (that is, how saturated it already is).
17546 This filter is similar to the Adobe Photoshop Selective Color tool.
17548 The filter accepts the following options:
17551 @item correction_method
17552 Select color correction method.
17554 Available values are:
17557 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17560 Specified adjustments are relative to the original component value.
17562 Default is @code{absolute}.
17564 Adjustments for red pixels (pixels where the red component is the maximum)
17566 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17568 Adjustments for green pixels (pixels where the green component is the maximum)
17570 Adjustments for cyan pixels (pixels where the red component is the minimum)
17572 Adjustments for blue pixels (pixels where the blue component is the maximum)
17574 Adjustments for magenta pixels (pixels where the green component is the minimum)
17576 Adjustments for white pixels (pixels where all components are greater than 128)
17578 Adjustments for all pixels except pure black and pure white
17580 Adjustments for black pixels (pixels where all components are lesser than 128)
17582 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17585 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17586 4 space separated floating point adjustment values in the [-1,1] range,
17587 respectively to adjust the amount of cyan, magenta, yellow and black for the
17588 pixels of its range.
17590 @subsection Examples
17594 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17595 increase magenta by 27% in blue areas:
17597 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17601 Use a Photoshop selective color preset:
17603 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17607 @anchor{separatefields}
17608 @section separatefields
17610 The @code{separatefields} takes a frame-based video input and splits
17611 each frame into its components fields, producing a new half height clip
17612 with twice the frame rate and twice the frame count.
17614 This filter use field-dominance information in frame to decide which
17615 of each pair of fields to place first in the output.
17616 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17618 @section setdar, setsar
17620 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17623 This is done by changing the specified Sample (aka Pixel) Aspect
17624 Ratio, according to the following equation:
17626 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17629 Keep in mind that the @code{setdar} filter does not modify the pixel
17630 dimensions of the video frame. Also, the display aspect ratio set by
17631 this filter may be changed by later filters in the filterchain,
17632 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17635 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17636 the filter output video.
17638 Note that as a consequence of the application of this filter, the
17639 output display aspect ratio will change according to the equation
17642 Keep in mind that the sample aspect ratio set by the @code{setsar}
17643 filter may be changed by later filters in the filterchain, e.g. if
17644 another "setsar" or a "setdar" filter is applied.
17646 It accepts the following parameters:
17649 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17650 Set the aspect ratio used by the filter.
17652 The parameter can be a floating point number string, an expression, or
17653 a string of the form @var{num}:@var{den}, where @var{num} and
17654 @var{den} are the numerator and denominator of the aspect ratio. If
17655 the parameter is not specified, it is assumed the value "0".
17656 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17660 Set the maximum integer value to use for expressing numerator and
17661 denominator when reducing the expressed aspect ratio to a rational.
17662 Default value is @code{100}.
17666 The parameter @var{sar} is an expression containing
17667 the following constants:
17671 These are approximated values for the mathematical constants e
17672 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17675 The input width and height.
17678 These are the same as @var{w} / @var{h}.
17681 The input sample aspect ratio.
17684 The input display aspect ratio. It is the same as
17685 (@var{w} / @var{h}) * @var{sar}.
17688 Horizontal and vertical chroma subsample values. For example, for the
17689 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17692 @subsection Examples
17697 To change the display aspect ratio to 16:9, specify one of the following:
17704 To change the sample aspect ratio to 10:11, specify:
17710 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17711 1000 in the aspect ratio reduction, use the command:
17713 setdar=ratio=16/9:max=1000
17721 Force field for the output video frame.
17723 The @code{setfield} filter marks the interlace type field for the
17724 output frames. It does not change the input frame, but only sets the
17725 corresponding property, which affects how the frame is treated by
17726 following filters (e.g. @code{fieldorder} or @code{yadif}).
17728 The filter accepts the following options:
17733 Available values are:
17737 Keep the same field property.
17740 Mark the frame as bottom-field-first.
17743 Mark the frame as top-field-first.
17746 Mark the frame as progressive.
17753 Force frame parameter for the output video frame.
17755 The @code{setparams} filter marks interlace and color range for the
17756 output frames. It does not change the input frame, but only sets the
17757 corresponding property, which affects how the frame is treated by
17762 Available values are:
17766 Keep the same field property (default).
17769 Mark the frame as bottom-field-first.
17772 Mark the frame as top-field-first.
17775 Mark the frame as progressive.
17779 Available values are:
17783 Keep the same color range property (default).
17785 @item unspecified, unknown
17786 Mark the frame as unspecified color range.
17788 @item limited, tv, mpeg
17789 Mark the frame as limited range.
17791 @item full, pc, jpeg
17792 Mark the frame as full range.
17795 @item color_primaries
17796 Set the color primaries.
17797 Available values are:
17801 Keep the same color primaries property (default).
17818 Set the color transfer.
17819 Available values are:
17823 Keep the same color trc property (default).
17845 Set the colorspace.
17846 Available values are:
17850 Keep the same colorspace property (default).
17863 @item chroma-derived-nc
17864 @item chroma-derived-c
17870 Apply shear transform to input video.
17872 This filter supports the following options:
17876 Shear factor in X-direction. Default value is 0.
17877 Allowed range is from -2 to 2.
17880 Shear factor in Y-direction. Default value is 0.
17881 Allowed range is from -2 to 2.
17884 Set the color used to fill the output area not covered by the transformed
17885 video. For the general syntax of this option, check the
17886 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
17887 If the special value "none" is selected then no
17888 background is printed (useful for example if the background is never shown).
17890 Default value is "black".
17893 Set interpolation type. Can be @code{bilinear} or @code{nearest}. Default is @code{bilinear}.
17896 @subsection Commands
17898 This filter supports the all above options as @ref{commands}.
17902 Show a line containing various information for each input video frame.
17903 The input video is not modified.
17905 This filter supports the following options:
17909 Calculate checksums of each plane. By default enabled.
17912 The shown line contains a sequence of key/value pairs of the form
17913 @var{key}:@var{value}.
17915 The following values are shown in the output:
17919 The (sequential) number of the input frame, starting from 0.
17922 The Presentation TimeStamp of the input frame, expressed as a number of
17923 time base units. The time base unit depends on the filter input pad.
17926 The Presentation TimeStamp of the input frame, expressed as a number of
17930 The position of the frame in the input stream, or -1 if this information is
17931 unavailable and/or meaningless (for example in case of synthetic video).
17934 The pixel format name.
17937 The sample aspect ratio of the input frame, expressed in the form
17938 @var{num}/@var{den}.
17941 The size of the input frame. For the syntax of this option, check the
17942 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17945 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17946 for bottom field first).
17949 This is 1 if the frame is a key frame, 0 otherwise.
17952 The picture type of the input frame ("I" for an I-frame, "P" for a
17953 P-frame, "B" for a B-frame, or "?" for an unknown type).
17954 Also refer to the documentation of the @code{AVPictureType} enum and of
17955 the @code{av_get_picture_type_char} function defined in
17956 @file{libavutil/avutil.h}.
17959 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17961 @item plane_checksum
17962 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17963 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17966 The mean value of pixels in each plane of the input frame, expressed in the form
17967 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17970 The standard deviation of pixel values in each plane of the input frame, expressed
17971 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17975 @section showpalette
17977 Displays the 256 colors palette of each frame. This filter is only relevant for
17978 @var{pal8} pixel format frames.
17980 It accepts the following option:
17984 Set the size of the box used to represent one palette color entry. Default is
17985 @code{30} (for a @code{30x30} pixel box).
17988 @section shuffleframes
17990 Reorder and/or duplicate and/or drop video frames.
17992 It accepts the following parameters:
17996 Set the destination indexes of input frames.
17997 This is space or '|' separated list of indexes that maps input frames to output
17998 frames. Number of indexes also sets maximal value that each index may have.
17999 '-1' index have special meaning and that is to drop frame.
18002 The first frame has the index 0. The default is to keep the input unchanged.
18004 @subsection Examples
18008 Swap second and third frame of every three frames of the input:
18010 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
18014 Swap 10th and 1st frame of every ten frames of the input:
18016 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
18020 @section shufflepixels
18022 Reorder pixels in video frames.
18024 This filter accepts the following options:
18028 Set shuffle direction. Can be forward or inverse direction.
18029 Default direction is forward.
18032 Set shuffle mode. Can be horizontal, vertical or block mode.
18036 Set shuffle block_size. In case of horizontal shuffle mode only width
18037 part of size is used, and in case of vertical shuffle mode only height
18038 part of size is used.
18041 Set random seed used with shuffling pixels. Mainly useful to set to be able
18042 to reverse filtering process to get original input.
18043 For example, to reverse forward shuffle you need to use same parameters
18044 and exact same seed and to set direction to inverse.
18047 @section shuffleplanes
18049 Reorder and/or duplicate video planes.
18051 It accepts the following parameters:
18056 The index of the input plane to be used as the first output plane.
18059 The index of the input plane to be used as the second output plane.
18062 The index of the input plane to be used as the third output plane.
18065 The index of the input plane to be used as the fourth output plane.
18069 The first plane has the index 0. The default is to keep the input unchanged.
18071 @subsection Examples
18075 Swap the second and third planes of the input:
18077 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
18081 @anchor{signalstats}
18082 @section signalstats
18083 Evaluate various visual metrics that assist in determining issues associated
18084 with the digitization of analog video media.
18086 By default the filter will log these metadata values:
18090 Display the minimal Y value contained within the input frame. Expressed in
18094 Display the Y value at the 10% percentile within the input frame. Expressed in
18098 Display the average Y value within the input frame. Expressed in range of
18102 Display the Y value at the 90% percentile within the input frame. Expressed in
18106 Display the maximum Y value contained within the input frame. Expressed in
18110 Display the minimal U value contained within the input frame. Expressed in
18114 Display the U value at the 10% percentile within the input frame. Expressed in
18118 Display the average U value within the input frame. Expressed in range of
18122 Display the U value at the 90% percentile within the input frame. Expressed in
18126 Display the maximum U value contained within the input frame. Expressed in
18130 Display the minimal V value contained within the input frame. Expressed in
18134 Display the V value at the 10% percentile within the input frame. Expressed in
18138 Display the average V value within the input frame. Expressed in range of
18142 Display the V value at the 90% percentile within the input frame. Expressed in
18146 Display the maximum V value contained within the input frame. Expressed in
18150 Display the minimal saturation value contained within the input frame.
18151 Expressed in range of [0-~181.02].
18154 Display the saturation value at the 10% percentile within the input frame.
18155 Expressed in range of [0-~181.02].
18158 Display the average saturation value within the input frame. Expressed in range
18162 Display the saturation value at the 90% percentile within the input frame.
18163 Expressed in range of [0-~181.02].
18166 Display the maximum saturation value contained within the input frame.
18167 Expressed in range of [0-~181.02].
18170 Display the median value for hue within the input frame. Expressed in range of
18174 Display the average value for hue within the input frame. Expressed in range of
18178 Display the average of sample value difference between all values of the Y
18179 plane in the current frame and corresponding values of the previous input frame.
18180 Expressed in range of [0-255].
18183 Display the average of sample value difference between all values of the U
18184 plane in the current frame and corresponding values of the previous input frame.
18185 Expressed in range of [0-255].
18188 Display the average of sample value difference between all values of the V
18189 plane in the current frame and corresponding values of the previous input frame.
18190 Expressed in range of [0-255].
18193 Display bit depth of Y plane in current frame.
18194 Expressed in range of [0-16].
18197 Display bit depth of U plane in current frame.
18198 Expressed in range of [0-16].
18201 Display bit depth of V plane in current frame.
18202 Expressed in range of [0-16].
18205 The filter accepts the following options:
18211 @option{stat} specify an additional form of image analysis.
18212 @option{out} output video with the specified type of pixel highlighted.
18214 Both options accept the following values:
18218 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
18219 unlike the neighboring pixels of the same field. Examples of temporal outliers
18220 include the results of video dropouts, head clogs, or tape tracking issues.
18223 Identify @var{vertical line repetition}. Vertical line repetition includes
18224 similar rows of pixels within a frame. In born-digital video vertical line
18225 repetition is common, but this pattern is uncommon in video digitized from an
18226 analog source. When it occurs in video that results from the digitization of an
18227 analog source it can indicate concealment from a dropout compensator.
18230 Identify pixels that fall outside of legal broadcast range.
18234 Set the highlight color for the @option{out} option. The default color is
18238 @subsection Examples
18242 Output data of various video metrics:
18244 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
18248 Output specific data about the minimum and maximum values of the Y plane per frame:
18250 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
18254 Playback video while highlighting pixels that are outside of broadcast range in red.
18256 ffplay example.mov -vf signalstats="out=brng:color=red"
18260 Playback video with signalstats metadata drawn over the frame.
18262 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
18265 The contents of signalstat_drawtext.txt used in the command are:
18268 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
18269 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18270 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18271 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18279 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18280 input. In this case the matching between the inputs can be calculated additionally.
18281 The filter always passes through the first input. The signature of each stream can
18282 be written into a file.
18284 It accepts the following options:
18288 Enable or disable the matching process.
18290 Available values are:
18294 Disable the calculation of a matching (default).
18296 Calculate the matching for the whole video and output whether the whole video
18297 matches or only parts.
18299 Calculate only until a matching is found or the video ends. Should be faster in
18304 Set the number of inputs. The option value must be a non negative integer.
18305 Default value is 1.
18308 Set the path to which the output is written. If there is more than one input,
18309 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18310 integer), that will be replaced with the input number. If no filename is
18311 specified, no output will be written. This is the default.
18314 Choose the output format.
18316 Available values are:
18320 Use the specified binary representation (default).
18322 Use the specified xml representation.
18326 Set threshold to detect one word as similar. The option value must be an integer
18327 greater than zero. The default value is 9000.
18330 Set threshold to detect all words as similar. The option value must be an integer
18331 greater than zero. The default value is 60000.
18334 Set threshold to detect frames as similar. The option value must be an integer
18335 greater than zero. The default value is 116.
18338 Set the minimum length of a sequence in frames to recognize it as matching
18339 sequence. The option value must be a non negative integer value.
18340 The default value is 0.
18343 Set the minimum relation, that matching frames to all frames must have.
18344 The option value must be a double value between 0 and 1. The default value is 0.5.
18347 @subsection Examples
18351 To calculate the signature of an input video and store it in signature.bin:
18353 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18357 To detect whether two videos match and store the signatures in XML format in
18358 signature0.xml and signature1.xml:
18360 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 -
18368 Blur the input video without impacting the outlines.
18370 It accepts the following options:
18373 @item luma_radius, lr
18374 Set the luma radius. The option value must be a float number in
18375 the range [0.1,5.0] that specifies the variance of the gaussian filter
18376 used to blur the image (slower if larger). Default value is 1.0.
18378 @item luma_strength, ls
18379 Set the luma strength. The option value must be a float number
18380 in the range [-1.0,1.0] that configures the blurring. A value included
18381 in [0.0,1.0] will blur the image whereas a value included in
18382 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18384 @item luma_threshold, lt
18385 Set the luma threshold used as a coefficient to determine
18386 whether a pixel should be blurred or not. The option value must be an
18387 integer in the range [-30,30]. A value of 0 will filter all the image,
18388 a value included in [0,30] will filter flat areas and a value included
18389 in [-30,0] will filter edges. Default value is 0.
18391 @item chroma_radius, cr
18392 Set the chroma radius. The option value must be a float number in
18393 the range [0.1,5.0] that specifies the variance of the gaussian filter
18394 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18396 @item chroma_strength, cs
18397 Set the chroma strength. The option value must be a float number
18398 in the range [-1.0,1.0] that configures the blurring. A value included
18399 in [0.0,1.0] will blur the image whereas a value included in
18400 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18402 @item chroma_threshold, ct
18403 Set the chroma threshold used as a coefficient to determine
18404 whether a pixel should be blurred or not. The option value must be an
18405 integer in the range [-30,30]. A value of 0 will filter all the image,
18406 a value included in [0,30] will filter flat areas and a value included
18407 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18410 If a chroma option is not explicitly set, the corresponding luma value
18414 Apply sobel operator to input video stream.
18416 The filter accepts the following option:
18420 Set which planes will be processed, unprocessed planes will be copied.
18421 By default value 0xf, all planes will be processed.
18424 Set value which will be multiplied with filtered result.
18427 Set value which will be added to filtered result.
18430 @subsection Commands
18432 This filter supports the all above options as @ref{commands}.
18437 Apply a simple postprocessing filter that compresses and decompresses the image
18438 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18439 and average the results.
18441 The filter accepts the following options:
18445 Set quality. This option defines the number of levels for averaging. It accepts
18446 an integer in the range 0-6. If set to @code{0}, the filter will have no
18447 effect. A value of @code{6} means the higher quality. For each increment of
18448 that value the speed drops by a factor of approximately 2. Default value is
18452 Force a constant quantization parameter. If not set, the filter will use the QP
18453 from the video stream (if available).
18456 Set thresholding mode. Available modes are:
18460 Set hard thresholding (default).
18462 Set soft thresholding (better de-ringing effect, but likely blurrier).
18465 @item use_bframe_qp
18466 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18467 option may cause flicker since the B-Frames have often larger QP. Default is
18468 @code{0} (not enabled).
18471 @subsection Commands
18473 This filter supports the following commands:
18475 @item quality, level
18476 Set quality level. The value @code{max} can be used to set the maximum level,
18477 currently @code{6}.
18483 Scale the input by applying one of the super-resolution methods based on
18484 convolutional neural networks. Supported models:
18488 Super-Resolution Convolutional Neural Network model (SRCNN).
18489 See @url{https://arxiv.org/abs/1501.00092}.
18492 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18493 See @url{https://arxiv.org/abs/1609.05158}.
18496 Training scripts as well as scripts for model file (.pb) saving can be found at
18497 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18498 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18500 Native model files (.model) can be generated from TensorFlow model
18501 files (.pb) by using tools/python/convert.py
18503 The filter accepts the following options:
18507 Specify which DNN backend to use for model loading and execution. This option accepts
18508 the following values:
18512 Native implementation of DNN loading and execution.
18515 TensorFlow backend. To enable this backend you
18516 need to install the TensorFlow for C library (see
18517 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18518 @code{--enable-libtensorflow}
18521 Default value is @samp{native}.
18524 Set path to model file specifying network architecture and its parameters.
18525 Note that different backends use different file formats. TensorFlow backend
18526 can load files for both formats, while native backend can load files for only
18530 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18531 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18532 input upscaled using bicubic upscaling with proper scale factor.
18535 This feature can also be finished with @ref{dnn_processing} filter.
18539 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18541 This filter takes in input two input videos, the first input is
18542 considered the "main" source and is passed unchanged to the
18543 output. The second input is used as a "reference" video for computing
18546 Both video inputs must have the same resolution and pixel format for
18547 this filter to work correctly. Also it assumes that both inputs
18548 have the same number of frames, which are compared one by one.
18550 The filter stores the calculated SSIM of each frame.
18552 The description of the accepted parameters follows.
18555 @item stats_file, f
18556 If specified the filter will use the named file to save the SSIM of
18557 each individual frame. When filename equals "-" the data is sent to
18561 The file printed if @var{stats_file} is selected, contains a sequence of
18562 key/value pairs of the form @var{key}:@var{value} for each compared
18565 A description of each shown parameter follows:
18569 sequential number of the input frame, starting from 1
18571 @item Y, U, V, R, G, B
18572 SSIM of the compared frames for the component specified by the suffix.
18575 SSIM of the compared frames for the whole frame.
18578 Same as above but in dB representation.
18581 This filter also supports the @ref{framesync} options.
18583 @subsection Examples
18588 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18589 [main][ref] ssim="stats_file=stats.log" [out]
18592 On this example the input file being processed is compared with the
18593 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18594 is stored in @file{stats.log}.
18597 Another example with both psnr and ssim at same time:
18599 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18603 Another example with different containers:
18605 ffmpeg -i main.mpg -i ref.mkv -lavfi "[0:v]settb=AVTB,setpts=PTS-STARTPTS[main];[1:v]settb=AVTB,setpts=PTS-STARTPTS[ref];[main][ref]ssim" -f null -
18611 Convert between different stereoscopic image formats.
18613 The filters accept the following options:
18617 Set stereoscopic image format of input.
18619 Available values for input image formats are:
18622 side by side parallel (left eye left, right eye right)
18625 side by side crosseye (right eye left, left eye right)
18628 side by side parallel with half width resolution
18629 (left eye left, right eye right)
18632 side by side crosseye with half width resolution
18633 (right eye left, left eye right)
18637 above-below (left eye above, right eye below)
18641 above-below (right eye above, left eye below)
18645 above-below with half height resolution
18646 (left eye above, right eye below)
18650 above-below with half height resolution
18651 (right eye above, left eye below)
18654 alternating frames (left eye first, right eye second)
18657 alternating frames (right eye first, left eye second)
18660 interleaved rows (left eye has top row, right eye starts on next row)
18663 interleaved rows (right eye has top row, left eye starts on next row)
18666 interleaved columns, left eye first
18669 interleaved columns, right eye first
18671 Default value is @samp{sbsl}.
18675 Set stereoscopic image format of output.
18679 side by side parallel (left eye left, right eye right)
18682 side by side crosseye (right eye left, left eye right)
18685 side by side parallel with half width resolution
18686 (left eye left, right eye right)
18689 side by side crosseye with half width resolution
18690 (right eye left, left eye right)
18694 above-below (left eye above, right eye below)
18698 above-below (right eye above, left eye below)
18702 above-below with half height resolution
18703 (left eye above, right eye below)
18707 above-below with half height resolution
18708 (right eye above, left eye below)
18711 alternating frames (left eye first, right eye second)
18714 alternating frames (right eye first, left eye second)
18717 interleaved rows (left eye has top row, right eye starts on next row)
18720 interleaved rows (right eye has top row, left eye starts on next row)
18723 anaglyph red/blue gray
18724 (red filter on left eye, blue filter on right eye)
18727 anaglyph red/green gray
18728 (red filter on left eye, green filter on right eye)
18731 anaglyph red/cyan gray
18732 (red filter on left eye, cyan filter on right eye)
18735 anaglyph red/cyan half colored
18736 (red filter on left eye, cyan filter on right eye)
18739 anaglyph red/cyan color
18740 (red filter on left eye, cyan filter on right eye)
18743 anaglyph red/cyan color optimized with the least squares projection of dubois
18744 (red filter on left eye, cyan filter on right eye)
18747 anaglyph green/magenta gray
18748 (green filter on left eye, magenta filter on right eye)
18751 anaglyph green/magenta half colored
18752 (green filter on left eye, magenta filter on right eye)
18755 anaglyph green/magenta colored
18756 (green filter on left eye, magenta filter on right eye)
18759 anaglyph green/magenta color optimized with the least squares projection of dubois
18760 (green filter on left eye, magenta filter on right eye)
18763 anaglyph yellow/blue gray
18764 (yellow filter on left eye, blue filter on right eye)
18767 anaglyph yellow/blue half colored
18768 (yellow filter on left eye, blue filter on right eye)
18771 anaglyph yellow/blue colored
18772 (yellow filter on left eye, blue filter on right eye)
18775 anaglyph yellow/blue color optimized with the least squares projection of dubois
18776 (yellow filter on left eye, blue filter on right eye)
18779 mono output (left eye only)
18782 mono output (right eye only)
18785 checkerboard, left eye first
18788 checkerboard, right eye first
18791 interleaved columns, left eye first
18794 interleaved columns, right eye first
18800 Default value is @samp{arcd}.
18803 @subsection Examples
18807 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18813 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18819 @section streamselect, astreamselect
18820 Select video or audio streams.
18822 The filter accepts the following options:
18826 Set number of inputs. Default is 2.
18829 Set input indexes to remap to outputs.
18832 @subsection Commands
18834 The @code{streamselect} and @code{astreamselect} filter supports the following
18839 Set input indexes to remap to outputs.
18842 @subsection Examples
18846 Select first 5 seconds 1st stream and rest of time 2nd stream:
18848 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18852 Same as above, but for audio:
18854 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18861 Draw subtitles on top of input video using the libass library.
18863 To enable compilation of this filter you need to configure FFmpeg with
18864 @code{--enable-libass}. This filter also requires a build with libavcodec and
18865 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18866 Alpha) subtitles format.
18868 The filter accepts the following options:
18872 Set the filename of the subtitle file to read. It must be specified.
18874 @item original_size
18875 Specify the size of the original video, the video for which the ASS file
18876 was composed. For the syntax of this option, check the
18877 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18878 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18879 correctly scale the fonts if the aspect ratio has been changed.
18882 Set a directory path containing fonts that can be used by the filter.
18883 These fonts will be used in addition to whatever the font provider uses.
18886 Process alpha channel, by default alpha channel is untouched.
18889 Set subtitles input character encoding. @code{subtitles} filter only. Only
18890 useful if not UTF-8.
18892 @item stream_index, si
18893 Set subtitles stream index. @code{subtitles} filter only.
18896 Override default style or script info parameters of the subtitles. It accepts a
18897 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18900 If the first key is not specified, it is assumed that the first value
18901 specifies the @option{filename}.
18903 For example, to render the file @file{sub.srt} on top of the input
18904 video, use the command:
18909 which is equivalent to:
18911 subtitles=filename=sub.srt
18914 To render the default subtitles stream from file @file{video.mkv}, use:
18916 subtitles=video.mkv
18919 To render the second subtitles stream from that file, use:
18921 subtitles=video.mkv:si=1
18924 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18925 @code{DejaVu Serif}, use:
18927 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18930 @section super2xsai
18932 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18933 Interpolate) pixel art scaling algorithm.
18935 Useful for enlarging pixel art images without reducing sharpness.
18939 Swap two rectangular objects in video.
18941 This filter accepts the following options:
18951 Set 1st rect x coordinate.
18954 Set 1st rect y coordinate.
18957 Set 2nd rect x coordinate.
18960 Set 2nd rect y coordinate.
18962 All expressions are evaluated once for each frame.
18965 The all options are expressions containing the following constants:
18970 The input width and height.
18973 same as @var{w} / @var{h}
18976 input sample aspect ratio
18979 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18982 The number of the input frame, starting from 0.
18985 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18988 the position in the file of the input frame, NAN if unknown
18995 Blend successive video frames.
19001 Apply telecine process to the video.
19003 This filter accepts the following options:
19012 The default value is @code{top}.
19016 A string of numbers representing the pulldown pattern you wish to apply.
19017 The default value is @code{23}.
19021 Some typical patterns:
19026 24p: 2332 (preferred)
19033 24p: 222222222223 ("Euro pulldown")
19038 @section thistogram
19040 Compute and draw a color distribution histogram for the input video across time.
19042 Unlike @ref{histogram} video filter which only shows histogram of single input frame
19043 at certain time, this filter shows also past histograms of number of frames defined
19044 by @code{width} option.
19046 The computed histogram is a representation of the color component
19047 distribution in an image.
19049 The filter accepts the following options:
19053 Set width of single color component output. Default value is @code{0}.
19054 Value of @code{0} means width will be picked from input video.
19055 This also set number of passed histograms to keep.
19056 Allowed range is [0, 8192].
19058 @item display_mode, d
19060 It accepts the following values:
19063 Per color component graphs are placed below each other.
19066 Per color component graphs are placed side by side.
19069 Presents information identical to that in the @code{parade}, except
19070 that the graphs representing color components are superimposed directly
19073 Default is @code{stack}.
19075 @item levels_mode, m
19076 Set mode. Can be either @code{linear}, or @code{logarithmic}.
19077 Default is @code{linear}.
19079 @item components, c
19080 Set what color components to display.
19081 Default is @code{7}.
19084 Set background opacity. Default is @code{0.9}.
19087 Show envelope. Default is disabled.
19090 Set envelope color. Default is @code{gold}.
19095 Available values for slide is:
19098 Draw new frame when right border is reached.
19101 Replace old columns with new ones.
19104 Scroll from right to left.
19107 Scroll from left to right.
19110 Draw single picture.
19113 Default is @code{replace}.
19118 Apply threshold effect to video stream.
19120 This filter needs four video streams to perform thresholding.
19121 First stream is stream we are filtering.
19122 Second stream is holding threshold values, third stream is holding min values,
19123 and last, fourth stream is holding max values.
19125 The filter accepts the following option:
19129 Set which planes will be processed, unprocessed planes will be copied.
19130 By default value 0xf, all planes will be processed.
19133 For example if first stream pixel's component value is less then threshold value
19134 of pixel component from 2nd threshold stream, third stream value will picked,
19135 otherwise fourth stream pixel component value will be picked.
19137 Using color source filter one can perform various types of thresholding:
19139 @subsection Examples
19143 Binary threshold, using gray color as threshold:
19145 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
19149 Inverted binary threshold, using gray color as threshold:
19151 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
19155 Truncate binary threshold, using gray color as threshold:
19157 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
19161 Threshold to zero, using gray color as threshold:
19163 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
19167 Inverted threshold to zero, using gray color as threshold:
19169 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
19174 Select the most representative frame in a given sequence of consecutive frames.
19176 The filter accepts the following options:
19180 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
19181 will pick one of them, and then handle the next batch of @var{n} frames until
19182 the end. Default is @code{100}.
19185 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
19186 value will result in a higher memory usage, so a high value is not recommended.
19188 @subsection Examples
19192 Extract one picture each 50 frames:
19198 Complete example of a thumbnail creation with @command{ffmpeg}:
19200 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
19207 Tile several successive frames together.
19209 The @ref{untile} filter can do the reverse.
19211 The filter accepts the following options:
19216 Set the grid size (i.e. the number of lines and columns). For the syntax of
19217 this option, check the
19218 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19221 Set the maximum number of frames to render in the given area. It must be less
19222 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
19223 the area will be used.
19226 Set the outer border margin in pixels.
19229 Set the inner border thickness (i.e. the number of pixels between frames). For
19230 more advanced padding options (such as having different values for the edges),
19231 refer to the pad video filter.
19234 Specify the color of the unused area. For the syntax of this option, check the
19235 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19236 The default value of @var{color} is "black".
19239 Set the number of frames to overlap when tiling several successive frames together.
19240 The value must be between @code{0} and @var{nb_frames - 1}.
19243 Set the number of frames to initially be empty before displaying first output frame.
19244 This controls how soon will one get first output frame.
19245 The value must be between @code{0} and @var{nb_frames - 1}.
19248 @subsection Examples
19252 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
19254 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
19256 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
19257 duplicating each output frame to accommodate the originally detected frame
19261 Display @code{5} pictures in an area of @code{3x2} frames,
19262 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
19263 mixed flat and named options:
19265 tile=3x2:nb_frames=5:padding=7:margin=2
19269 @section tinterlace
19271 Perform various types of temporal field interlacing.
19273 Frames are counted starting from 1, so the first input frame is
19276 The filter accepts the following options:
19281 Specify the mode of the interlacing. This option can also be specified
19282 as a value alone. See below for a list of values for this option.
19284 Available values are:
19288 Move odd frames into the upper field, even into the lower field,
19289 generating a double height frame at half frame rate.
19293 Frame 1 Frame 2 Frame 3 Frame 4
19295 11111 22222 33333 44444
19296 11111 22222 33333 44444
19297 11111 22222 33333 44444
19298 11111 22222 33333 44444
19312 Only output odd frames, even frames are dropped, generating a frame with
19313 unchanged height at half frame rate.
19318 Frame 1 Frame 2 Frame 3 Frame 4
19320 11111 22222 33333 44444
19321 11111 22222 33333 44444
19322 11111 22222 33333 44444
19323 11111 22222 33333 44444
19333 Only output even frames, odd frames are dropped, generating a frame with
19334 unchanged height at half frame rate.
19339 Frame 1 Frame 2 Frame 3 Frame 4
19341 11111 22222 33333 44444
19342 11111 22222 33333 44444
19343 11111 22222 33333 44444
19344 11111 22222 33333 44444
19354 Expand each frame to full height, but pad alternate lines with black,
19355 generating a frame with double height at the same input frame rate.
19360 Frame 1 Frame 2 Frame 3 Frame 4
19362 11111 22222 33333 44444
19363 11111 22222 33333 44444
19364 11111 22222 33333 44444
19365 11111 22222 33333 44444
19368 11111 ..... 33333 .....
19369 ..... 22222 ..... 44444
19370 11111 ..... 33333 .....
19371 ..... 22222 ..... 44444
19372 11111 ..... 33333 .....
19373 ..... 22222 ..... 44444
19374 11111 ..... 33333 .....
19375 ..... 22222 ..... 44444
19379 @item interleave_top, 4
19380 Interleave the upper field from odd frames with the lower field from
19381 even frames, generating a frame with unchanged height at half frame rate.
19386 Frame 1 Frame 2 Frame 3 Frame 4
19388 11111<- 22222 33333<- 44444
19389 11111 22222<- 33333 44444<-
19390 11111<- 22222 33333<- 44444
19391 11111 22222<- 33333 44444<-
19401 @item interleave_bottom, 5
19402 Interleave the lower field from odd frames with the upper field from
19403 even frames, generating a frame with unchanged height at half frame rate.
19408 Frame 1 Frame 2 Frame 3 Frame 4
19410 11111 22222<- 33333 44444<-
19411 11111<- 22222 33333<- 44444
19412 11111 22222<- 33333 44444<-
19413 11111<- 22222 33333<- 44444
19423 @item interlacex2, 6
19424 Double frame rate with unchanged height. Frames are inserted each
19425 containing the second temporal field from the previous input frame and
19426 the first temporal field from the next input frame. This mode relies on
19427 the top_field_first flag. Useful for interlaced video displays with no
19428 field synchronisation.
19433 Frame 1 Frame 2 Frame 3 Frame 4
19435 11111 22222 33333 44444
19436 11111 22222 33333 44444
19437 11111 22222 33333 44444
19438 11111 22222 33333 44444
19441 11111 22222 22222 33333 33333 44444 44444
19442 11111 11111 22222 22222 33333 33333 44444
19443 11111 22222 22222 33333 33333 44444 44444
19444 11111 11111 22222 22222 33333 33333 44444
19449 Move odd frames into the upper field, even into the lower field,
19450 generating a double height frame at same frame rate.
19455 Frame 1 Frame 2 Frame 3 Frame 4
19457 11111 22222 33333 44444
19458 11111 22222 33333 44444
19459 11111 22222 33333 44444
19460 11111 22222 33333 44444
19463 11111 33333 33333 55555
19464 22222 22222 44444 44444
19465 11111 33333 33333 55555
19466 22222 22222 44444 44444
19467 11111 33333 33333 55555
19468 22222 22222 44444 44444
19469 11111 33333 33333 55555
19470 22222 22222 44444 44444
19475 Numeric values are deprecated but are accepted for backward
19476 compatibility reasons.
19478 Default mode is @code{merge}.
19481 Specify flags influencing the filter process.
19483 Available value for @var{flags} is:
19486 @item low_pass_filter, vlpf
19487 Enable linear vertical low-pass filtering in the filter.
19488 Vertical low-pass filtering is required when creating an interlaced
19489 destination from a progressive source which contains high-frequency
19490 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19493 @item complex_filter, cvlpf
19494 Enable complex vertical low-pass filtering.
19495 This will slightly less reduce interlace 'twitter' and Moire
19496 patterning but better retain detail and subjective sharpness impression.
19499 Bypass already interlaced frames, only adjust the frame rate.
19502 Vertical low-pass filtering and bypassing already interlaced frames can only be
19503 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19508 Pick median pixels from several successive input video frames.
19510 The filter accepts the following options:
19514 Set radius of median filter.
19515 Default is 1. Allowed range is from 1 to 127.
19518 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19521 Set median percentile. Default value is @code{0.5}.
19522 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19523 minimum values, and @code{1} maximum values.
19526 @subsection Commands
19528 This filter supports all above options as @ref{commands}, excluding option @code{radius}.
19530 @section tmidequalizer
19532 Apply Temporal Midway Video Equalization effect.
19534 Midway Video Equalization adjusts a sequence of video frames to have the same
19535 histograms, while maintaining their dynamics as much as possible. It's
19536 useful for e.g. matching exposures from a video frames sequence.
19538 This filter accepts the following option:
19542 Set filtering radius. Default is @code{5}. Allowed range is from 1 to 127.
19545 Set filtering sigma. Default is @code{0.5}. This controls strength of filtering.
19546 Setting this option to 0 effectively does nothing.
19549 Set which planes to process. Default is @code{15}, which is all available planes.
19554 Mix successive video frames.
19556 A description of the accepted options follows.
19560 The number of successive frames to mix. If unspecified, it defaults to 3.
19563 Specify weight of each input video frame.
19564 Each weight is separated by space. If number of weights is smaller than
19565 number of @var{frames} last specified weight will be used for all remaining
19569 Specify scale, if it is set it will be multiplied with sum
19570 of each weight multiplied with pixel values to give final destination
19571 pixel value. By default @var{scale} is auto scaled to sum of weights.
19574 @subsection Examples
19578 Average 7 successive frames:
19580 tmix=frames=7:weights="1 1 1 1 1 1 1"
19584 Apply simple temporal convolution:
19586 tmix=frames=3:weights="-1 3 -1"
19590 Similar as above but only showing temporal differences:
19592 tmix=frames=3:weights="-1 2 -1":scale=1
19598 Tone map colors from different dynamic ranges.
19600 This filter expects data in single precision floating point, as it needs to
19601 operate on (and can output) out-of-range values. Another filter, such as
19602 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19604 The tonemapping algorithms implemented only work on linear light, so input
19605 data should be linearized beforehand (and possibly correctly tagged).
19608 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19611 @subsection Options
19612 The filter accepts the following options.
19616 Set the tone map algorithm to use.
19618 Possible values are:
19621 Do not apply any tone map, only desaturate overbright pixels.
19624 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19625 in-range values, while distorting out-of-range values.
19628 Stretch the entire reference gamut to a linear multiple of the display.
19631 Fit a logarithmic transfer between the tone curves.
19634 Preserve overall image brightness with a simple curve, using nonlinear
19635 contrast, which results in flattening details and degrading color accuracy.
19638 Preserve both dark and bright details better than @var{reinhard}, at the cost
19639 of slightly darkening everything. Use it when detail preservation is more
19640 important than color and brightness accuracy.
19643 Smoothly map out-of-range values, while retaining contrast and colors for
19644 in-range material as much as possible. Use it when color accuracy is more
19645 important than detail preservation.
19651 Tune the tone mapping algorithm.
19653 This affects the following algorithms:
19659 Specifies the scale factor to use while stretching.
19663 Specifies the exponent of the function.
19667 Specify an extra linear coefficient to multiply into the signal before clipping.
19671 Specify the local contrast coefficient at the display peak.
19672 Default to 0.5, which means that in-gamut values will be about half as bright
19679 Specify the transition point from linear to mobius transform. Every value
19680 below this point is guaranteed to be mapped 1:1. The higher the value, the
19681 more accurate the result will be, at the cost of losing bright details.
19682 Default to 0.3, which due to the steep initial slope still preserves in-range
19683 colors fairly accurately.
19687 Apply desaturation for highlights that exceed this level of brightness. The
19688 higher the parameter, the more color information will be preserved. This
19689 setting helps prevent unnaturally blown-out colors for super-highlights, by
19690 (smoothly) turning into white instead. This makes images feel more natural,
19691 at the cost of reducing information about out-of-range colors.
19693 The default of 2.0 is somewhat conservative and will mostly just apply to
19694 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19696 This option works only if the input frame has a supported color tag.
19699 Override signal/nominal/reference peak with this value. Useful when the
19700 embedded peak information in display metadata is not reliable or when tone
19701 mapping from a lower range to a higher range.
19706 Temporarily pad video frames.
19708 The filter accepts the following options:
19712 Specify number of delay frames before input video stream. Default is 0.
19715 Specify number of padding frames after input video stream.
19716 Set to -1 to pad indefinitely. Default is 0.
19719 Set kind of frames added to beginning of stream.
19720 Can be either @var{add} or @var{clone}.
19721 With @var{add} frames of solid-color are added.
19722 With @var{clone} frames are clones of first frame.
19723 Default is @var{add}.
19726 Set kind of frames added to end of stream.
19727 Can be either @var{add} or @var{clone}.
19728 With @var{add} frames of solid-color are added.
19729 With @var{clone} frames are clones of last frame.
19730 Default is @var{add}.
19732 @item start_duration, stop_duration
19733 Specify the duration of the start/stop delay. See
19734 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19735 for the accepted syntax.
19736 These options override @var{start} and @var{stop}. Default is 0.
19739 Specify the color of the padded area. For the syntax of this option,
19740 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19741 manual,ffmpeg-utils}.
19743 The default value of @var{color} is "black".
19749 Transpose rows with columns in the input video and optionally flip it.
19751 It accepts the following parameters:
19756 Specify the transposition direction.
19758 Can assume the following values:
19760 @item 0, 4, cclock_flip
19761 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19769 Rotate by 90 degrees clockwise, that is:
19777 Rotate by 90 degrees counterclockwise, that is:
19784 @item 3, 7, clock_flip
19785 Rotate by 90 degrees clockwise and vertically flip, that is:
19793 For values between 4-7, the transposition is only done if the input
19794 video geometry is portrait and not landscape. These values are
19795 deprecated, the @code{passthrough} option should be used instead.
19797 Numerical values are deprecated, and should be dropped in favor of
19798 symbolic constants.
19801 Do not apply the transposition if the input geometry matches the one
19802 specified by the specified value. It accepts the following values:
19805 Always apply transposition.
19807 Preserve portrait geometry (when @var{height} >= @var{width}).
19809 Preserve landscape geometry (when @var{width} >= @var{height}).
19812 Default value is @code{none}.
19815 For example to rotate by 90 degrees clockwise and preserve portrait
19818 transpose=dir=1:passthrough=portrait
19821 The command above can also be specified as:
19823 transpose=1:portrait
19826 @section transpose_npp
19828 Transpose rows with columns in the input video and optionally flip it.
19829 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19831 It accepts the following parameters:
19836 Specify the transposition direction.
19838 Can assume the following values:
19841 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19844 Rotate by 90 degrees clockwise.
19847 Rotate by 90 degrees counterclockwise.
19850 Rotate by 90 degrees clockwise and vertically flip.
19854 Do not apply the transposition if the input geometry matches the one
19855 specified by the specified value. It accepts the following values:
19858 Always apply transposition. (default)
19860 Preserve portrait geometry (when @var{height} >= @var{width}).
19862 Preserve landscape geometry (when @var{width} >= @var{height}).
19868 Trim the input so that the output contains one continuous subpart of the input.
19870 It accepts the following parameters:
19873 Specify the time of the start of the kept section, i.e. the frame with the
19874 timestamp @var{start} will be the first frame in the output.
19877 Specify the time of the first frame that will be dropped, i.e. the frame
19878 immediately preceding the one with the timestamp @var{end} will be the last
19879 frame in the output.
19882 This is the same as @var{start}, except this option sets the start timestamp
19883 in timebase units instead of seconds.
19886 This is the same as @var{end}, except this option sets the end timestamp
19887 in timebase units instead of seconds.
19890 The maximum duration of the output in seconds.
19893 The number of the first frame that should be passed to the output.
19896 The number of the first frame that should be dropped.
19899 @option{start}, @option{end}, and @option{duration} are expressed as time
19900 duration specifications; see
19901 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19902 for the accepted syntax.
19904 Note that the first two sets of the start/end options and the @option{duration}
19905 option look at the frame timestamp, while the _frame variants simply count the
19906 frames that pass through the filter. Also note that this filter does not modify
19907 the timestamps. If you wish for the output timestamps to start at zero, insert a
19908 setpts filter after the trim filter.
19910 If multiple start or end options are set, this filter tries to be greedy and
19911 keep all the frames that match at least one of the specified constraints. To keep
19912 only the part that matches all the constraints at once, chain multiple trim
19915 The defaults are such that all the input is kept. So it is possible to set e.g.
19916 just the end values to keep everything before the specified time.
19921 Drop everything except the second minute of input:
19923 ffmpeg -i INPUT -vf trim=60:120
19927 Keep only the first second:
19929 ffmpeg -i INPUT -vf trim=duration=1
19934 @section unpremultiply
19935 Apply alpha unpremultiply effect to input video stream using first plane
19936 of second stream as alpha.
19938 Both streams must have same dimensions and same pixel format.
19940 The filter accepts the following option:
19944 Set which planes will be processed, unprocessed planes will be copied.
19945 By default value 0xf, all planes will be processed.
19947 If the format has 1 or 2 components, then luma is bit 0.
19948 If the format has 3 or 4 components:
19949 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19950 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19951 If present, the alpha channel is always the last bit.
19954 Do not require 2nd input for processing, instead use alpha plane from input stream.
19960 Sharpen or blur the input video.
19962 It accepts the following parameters:
19965 @item luma_msize_x, lx
19966 Set the luma matrix horizontal size. It must be an odd integer between
19967 3 and 23. The default value is 5.
19969 @item luma_msize_y, ly
19970 Set the luma matrix vertical size. It must be an odd integer between 3
19971 and 23. The default value is 5.
19973 @item luma_amount, la
19974 Set the luma effect strength. It must be a floating point number, reasonable
19975 values lay between -1.5 and 1.5.
19977 Negative values will blur the input video, while positive values will
19978 sharpen it, a value of zero will disable the effect.
19980 Default value is 1.0.
19982 @item chroma_msize_x, cx
19983 Set the chroma matrix horizontal size. It must be an odd integer
19984 between 3 and 23. The default value is 5.
19986 @item chroma_msize_y, cy
19987 Set the chroma matrix vertical size. It must be an odd integer
19988 between 3 and 23. The default value is 5.
19990 @item chroma_amount, ca
19991 Set the chroma effect strength. It must be a floating point number, reasonable
19992 values lay between -1.5 and 1.5.
19994 Negative values will blur the input video, while positive values will
19995 sharpen it, a value of zero will disable the effect.
19997 Default value is 0.0.
20001 All parameters are optional and default to the equivalent of the
20002 string '5:5:1.0:5:5:0.0'.
20004 @subsection Examples
20008 Apply strong luma sharpen effect:
20010 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
20014 Apply a strong blur of both luma and chroma parameters:
20016 unsharp=7:7:-2:7:7:-2
20023 Decompose a video made of tiled images into the individual images.
20025 The frame rate of the output video is the frame rate of the input video
20026 multiplied by the number of tiles.
20028 This filter does the reverse of @ref{tile}.
20030 The filter accepts the following options:
20035 Set the grid size (i.e. the number of lines and columns). For the syntax of
20036 this option, check the
20037 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20040 @subsection Examples
20044 Produce a 1-second video from a still image file made of 25 frames stacked
20045 vertically, like an analogic film reel:
20047 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
20053 Apply ultra slow/simple postprocessing filter that compresses and decompresses
20054 the image at several (or - in the case of @option{quality} level @code{8} - all)
20055 shifts and average the results.
20057 The way this differs from the behavior of spp is that uspp actually encodes &
20058 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
20059 DCT similar to MJPEG.
20061 The filter accepts the following options:
20065 Set quality. This option defines the number of levels for averaging. It accepts
20066 an integer in the range 0-8. If set to @code{0}, the filter will have no
20067 effect. A value of @code{8} means the higher quality. For each increment of
20068 that value the speed drops by a factor of approximately 2. Default value is
20072 Force a constant quantization parameter. If not set, the filter will use the QP
20073 from the video stream (if available).
20078 Convert 360 videos between various formats.
20080 The filter accepts the following options:
20086 Set format of the input/output video.
20094 Equirectangular projection.
20099 Cubemap with 3x2/6x1/1x6 layout.
20101 Format specific options:
20106 Set padding proportion for the input/output cubemap. Values in decimals.
20113 1% of face is padding. For example, with 1920x1280 resolution face size would be 640x640 and padding would be 3 pixels from each side. (640 * 0.01 = 6 pixels)
20116 Default value is @b{@samp{0}}.
20117 Maximum value is @b{@samp{0.1}}.
20121 Set fixed padding for the input/output cubemap. Values in pixels.
20123 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
20127 Set order of faces for the input/output cubemap. Choose one direction for each position.
20129 Designation of directions:
20145 Default value is @b{@samp{rludfb}}.
20149 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
20151 Designation of angles:
20154 0 degrees clockwise
20156 90 degrees clockwise
20158 180 degrees clockwise
20160 270 degrees clockwise
20163 Default value is @b{@samp{000000}}.
20167 Equi-Angular Cubemap.
20174 Format specific options:
20179 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20181 If diagonal field of view is set it overrides horizontal and vertical field of view.
20186 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20188 If diagonal field of view is set it overrides horizontal and vertical field of view.
20194 Format specific options:
20199 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20201 If diagonal field of view is set it overrides horizontal and vertical field of view.
20206 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20208 If diagonal field of view is set it overrides horizontal and vertical field of view.
20214 Facebook's 360 formats.
20217 Stereographic format.
20219 Format specific options:
20224 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20226 If diagonal field of view is set it overrides horizontal and vertical field of view.
20231 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20233 If diagonal field of view is set it overrides horizontal and vertical field of view.
20240 Ball format, gives significant distortion toward the back.
20243 Hammer-Aitoff map projection format.
20246 Sinusoidal map projection format.
20249 Fisheye projection.
20251 Format specific options:
20256 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20258 If diagonal field of view is set it overrides horizontal and vertical field of view.
20263 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20265 If diagonal field of view is set it overrides horizontal and vertical field of view.
20269 Pannini projection.
20271 Format specific options:
20274 Set output pannini parameter.
20277 Set input pannini parameter.
20281 Cylindrical projection.
20283 Format specific options:
20288 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20290 If diagonal field of view is set it overrides horizontal and vertical field of view.
20295 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20297 If diagonal field of view is set it overrides horizontal and vertical field of view.
20301 Perspective projection. @i{(output only)}
20303 Format specific options:
20306 Set perspective parameter.
20310 Tetrahedron projection.
20313 Truncated square pyramid projection.
20317 Half equirectangular projection.
20322 Format specific options:
20327 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20329 If diagonal field of view is set it overrides horizontal and vertical field of view.
20334 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20336 If diagonal field of view is set it overrides horizontal and vertical field of view.
20340 Orthographic format.
20342 Format specific options:
20347 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20349 If diagonal field of view is set it overrides horizontal and vertical field of view.
20354 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20356 If diagonal field of view is set it overrides horizontal and vertical field of view.
20360 Octahedron projection.
20364 Set interpolation method.@*
20365 @i{Note: more complex interpolation methods require much more memory to run.}
20375 Bilinear interpolation.
20377 Lagrange9 interpolation.
20380 Bicubic interpolation.
20383 Lanczos interpolation.
20386 Spline16 interpolation.
20389 Gaussian interpolation.
20391 Mitchell interpolation.
20394 Default value is @b{@samp{line}}.
20398 Set the output video resolution.
20400 Default resolution depends on formats.
20404 Set the input/output stereo format.
20415 Default value is @b{@samp{2d}} for input and output format.
20420 Set rotation for the output video. Values in degrees.
20423 Set rotation order for the output video. Choose one item for each position.
20434 Default value is @b{@samp{ypr}}.
20439 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20443 Set if input video is flipped horizontally/vertically. Boolean values.
20446 Set if input video is transposed. Boolean value, by default disabled.
20449 Set if output video needs to be transposed. Boolean value, by default disabled.
20452 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20455 @subsection Examples
20459 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20461 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20464 Extract back view of Equi-Angular Cubemap:
20466 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20469 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20471 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20475 @subsection Commands
20477 This filter supports subset of above options as @ref{commands}.
20479 @section vaguedenoiser
20481 Apply a wavelet based denoiser.
20483 It transforms each frame from the video input into the wavelet domain,
20484 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20485 the obtained coefficients. It does an inverse wavelet transform after.
20486 Due to wavelet properties, it should give a nice smoothed result, and
20487 reduced noise, without blurring picture features.
20489 This filter accepts the following options:
20493 The filtering strength. The higher, the more filtered the video will be.
20494 Hard thresholding can use a higher threshold than soft thresholding
20495 before the video looks overfiltered. Default value is 2.
20498 The filtering method the filter will use.
20500 It accepts the following values:
20503 All values under the threshold will be zeroed.
20506 All values under the threshold will be zeroed. All values above will be
20507 reduced by the threshold.
20510 Scales or nullifies coefficients - intermediary between (more) soft and
20511 (less) hard thresholding.
20514 Default is garrote.
20517 Number of times, the wavelet will decompose the picture. Picture can't
20518 be decomposed beyond a particular point (typically, 8 for a 640x480
20519 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20522 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20525 A list of the planes to process. By default all planes are processed.
20528 The threshold type the filter will use.
20530 It accepts the following values:
20533 Threshold used is same for all decompositions.
20536 Threshold used depends also on each decomposition coefficients.
20539 Default is universal.
20542 @section vectorscope
20544 Display 2 color component values in the two dimensional graph (which is called
20547 This filter accepts the following options:
20551 Set vectorscope mode.
20553 It accepts the following values:
20557 Gray values are displayed on graph, higher brightness means more pixels have
20558 same component color value on location in graph. This is the default mode.
20561 Gray values are displayed on graph. Surrounding pixels values which are not
20562 present in video frame are drawn in gradient of 2 color components which are
20563 set by option @code{x} and @code{y}. The 3rd color component is static.
20566 Actual color components values present in video frame are displayed on graph.
20569 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20570 on graph increases value of another color component, which is luminance by
20571 default values of @code{x} and @code{y}.
20574 Actual colors present in video frame are displayed on graph. If two different
20575 colors map to same position on graph then color with higher value of component
20576 not present in graph is picked.
20579 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20580 component picked from radial gradient.
20584 Set which color component will be represented on X-axis. Default is @code{1}.
20587 Set which color component will be represented on Y-axis. Default is @code{2}.
20590 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20591 of color component which represents frequency of (X, Y) location in graph.
20596 No envelope, this is default.
20599 Instant envelope, even darkest single pixel will be clearly highlighted.
20602 Hold maximum and minimum values presented in graph over time. This way you
20603 can still spot out of range values without constantly looking at vectorscope.
20606 Peak and instant envelope combined together.
20610 Set what kind of graticule to draw.
20619 Set graticule opacity.
20622 Set graticule flags.
20626 Draw graticule for white point.
20629 Draw graticule for black point.
20632 Draw color points short names.
20636 Set background opacity.
20638 @item lthreshold, l
20639 Set low threshold for color component not represented on X or Y axis.
20640 Values lower than this value will be ignored. Default is 0.
20641 Note this value is multiplied with actual max possible value one pixel component
20642 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20645 @item hthreshold, h
20646 Set high threshold for color component not represented on X or Y axis.
20647 Values higher than this value will be ignored. Default is 1.
20648 Note this value is multiplied with actual max possible value one pixel component
20649 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20650 is 0.9 * 255 = 230.
20652 @item colorspace, c
20653 Set what kind of colorspace to use when drawing graticule.
20663 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20664 This means no tint, and output will remain gray.
20667 @anchor{vidstabdetect}
20668 @section vidstabdetect
20670 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20671 @ref{vidstabtransform} for pass 2.
20673 This filter generates a file with relative translation and rotation
20674 transform information about subsequent frames, which is then used by
20675 the @ref{vidstabtransform} filter.
20677 To enable compilation of this filter you need to configure FFmpeg with
20678 @code{--enable-libvidstab}.
20680 This filter accepts the following options:
20684 Set the path to the file used to write the transforms information.
20685 Default value is @file{transforms.trf}.
20688 Set how shaky the video is and how quick the camera is. It accepts an
20689 integer in the range 1-10, a value of 1 means little shakiness, a
20690 value of 10 means strong shakiness. Default value is 5.
20693 Set the accuracy of the detection process. It must be a value in the
20694 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20695 accuracy. Default value is 15.
20698 Set stepsize of the search process. The region around minimum is
20699 scanned with 1 pixel resolution. Default value is 6.
20702 Set minimum contrast. Below this value a local measurement field is
20703 discarded. Must be a floating point value in the range 0-1. Default
20707 Set reference frame number for tripod mode.
20709 If enabled, the motion of the frames is compared to a reference frame
20710 in the filtered stream, identified by the specified number. The idea
20711 is to compensate all movements in a more-or-less static scene and keep
20712 the camera view absolutely still.
20714 If set to 0, it is disabled. The frames are counted starting from 1.
20717 Show fields and transforms in the resulting frames. It accepts an
20718 integer in the range 0-2. Default value is 0, which disables any
20722 @subsection Examples
20726 Use default values:
20732 Analyze strongly shaky movie and put the results in file
20733 @file{mytransforms.trf}:
20735 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20739 Visualize the result of internal transformations in the resulting
20742 vidstabdetect=show=1
20746 Analyze a video with medium shakiness using @command{ffmpeg}:
20748 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20752 @anchor{vidstabtransform}
20753 @section vidstabtransform
20755 Video stabilization/deshaking: pass 2 of 2,
20756 see @ref{vidstabdetect} for pass 1.
20758 Read a file with transform information for each frame and
20759 apply/compensate them. Together with the @ref{vidstabdetect}
20760 filter this can be used to deshake videos. See also
20761 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20762 the @ref{unsharp} filter, see below.
20764 To enable compilation of this filter you need to configure FFmpeg with
20765 @code{--enable-libvidstab}.
20767 @subsection Options
20771 Set path to the file used to read the transforms. Default value is
20772 @file{transforms.trf}.
20775 Set the number of frames (value*2 + 1) used for lowpass filtering the
20776 camera movements. Default value is 10.
20778 For example a number of 10 means that 21 frames are used (10 in the
20779 past and 10 in the future) to smoothen the motion in the video. A
20780 larger value leads to a smoother video, but limits the acceleration of
20781 the camera (pan/tilt movements). 0 is a special case where a static
20782 camera is simulated.
20785 Set the camera path optimization algorithm.
20787 Accepted values are:
20790 gaussian kernel low-pass filter on camera motion (default)
20792 averaging on transformations
20796 Set maximal number of pixels to translate frames. Default value is -1,
20800 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20801 value is -1, meaning no limit.
20804 Specify how to deal with borders that may be visible due to movement
20807 Available values are:
20810 keep image information from previous frame (default)
20812 fill the border black
20816 Invert transforms if set to 1. Default value is 0.
20819 Consider transforms as relative to previous frame if set to 1,
20820 absolute if set to 0. Default value is 0.
20823 Set percentage to zoom. A positive value will result in a zoom-in
20824 effect, a negative value in a zoom-out effect. Default value is 0 (no
20828 Set optimal zooming to avoid borders.
20830 Accepted values are:
20835 optimal static zoom value is determined (only very strong movements
20836 will lead to visible borders) (default)
20838 optimal adaptive zoom value is determined (no borders will be
20839 visible), see @option{zoomspeed}
20842 Note that the value given at zoom is added to the one calculated here.
20845 Set percent to zoom maximally each frame (enabled when
20846 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20850 Specify type of interpolation.
20852 Available values are:
20857 linear only horizontal
20859 linear in both directions (default)
20861 cubic in both directions (slow)
20865 Enable virtual tripod mode if set to 1, which is equivalent to
20866 @code{relative=0:smoothing=0}. Default value is 0.
20868 Use also @code{tripod} option of @ref{vidstabdetect}.
20871 Increase log verbosity if set to 1. Also the detected global motions
20872 are written to the temporary file @file{global_motions.trf}. Default
20876 @subsection Examples
20880 Use @command{ffmpeg} for a typical stabilization with default values:
20882 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20885 Note the use of the @ref{unsharp} filter which is always recommended.
20888 Zoom in a bit more and load transform data from a given file:
20890 vidstabtransform=zoom=5:input="mytransforms.trf"
20894 Smoothen the video even more:
20896 vidstabtransform=smoothing=30
20902 Flip the input video vertically.
20904 For example, to vertically flip a video with @command{ffmpeg}:
20906 ffmpeg -i in.avi -vf "vflip" out.avi
20911 Detect variable frame rate video.
20913 This filter tries to detect if the input is variable or constant frame rate.
20915 At end it will output number of frames detected as having variable delta pts,
20916 and ones with constant delta pts.
20917 If there was frames with variable delta, than it will also show min, max and
20918 average delta encountered.
20922 Boost or alter saturation.
20924 The filter accepts the following options:
20927 Set strength of boost if positive value or strength of alter if negative value.
20928 Default is 0. Allowed range is from -2 to 2.
20931 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20934 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20937 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20940 Set the red luma coefficient.
20943 Set the green luma coefficient.
20946 Set the blue luma coefficient.
20949 If @code{intensity} is negative and this is set to 1, colors will change,
20950 otherwise colors will be less saturated, more towards gray.
20953 @subsection Commands
20955 This filter supports the all above options as @ref{commands}.
20960 Make or reverse a natural vignetting effect.
20962 The filter accepts the following options:
20966 Set lens angle expression as a number of radians.
20968 The value is clipped in the @code{[0,PI/2]} range.
20970 Default value: @code{"PI/5"}
20974 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20978 Set forward/backward mode.
20980 Available modes are:
20983 The larger the distance from the central point, the darker the image becomes.
20986 The larger the distance from the central point, the brighter the image becomes.
20987 This can be used to reverse a vignette effect, though there is no automatic
20988 detection to extract the lens @option{angle} and other settings (yet). It can
20989 also be used to create a burning effect.
20992 Default value is @samp{forward}.
20995 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20997 It accepts the following values:
21000 Evaluate expressions only once during the filter initialization.
21003 Evaluate expressions for each incoming frame. This is way slower than the
21004 @samp{init} mode since it requires all the scalers to be re-computed, but it
21005 allows advanced dynamic expressions.
21008 Default value is @samp{init}.
21011 Set dithering to reduce the circular banding effects. Default is @code{1}
21015 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
21016 Setting this value to the SAR of the input will make a rectangular vignetting
21017 following the dimensions of the video.
21019 Default is @code{1/1}.
21022 @subsection Expressions
21024 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
21025 following parameters.
21030 input width and height
21033 the number of input frame, starting from 0
21036 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
21037 @var{TB} units, NAN if undefined
21040 frame rate of the input video, NAN if the input frame rate is unknown
21043 the PTS (Presentation TimeStamp) of the filtered video frame,
21044 expressed in seconds, NAN if undefined
21047 time base of the input video
21051 @subsection Examples
21055 Apply simple strong vignetting effect:
21061 Make a flickering vignetting:
21063 vignette='PI/4+random(1)*PI/50':eval=frame
21068 @section vmafmotion
21070 Obtain the average VMAF motion score of a video.
21071 It is one of the component metrics of VMAF.
21073 The obtained average motion score is printed through the logging system.
21075 The filter accepts the following options:
21079 If specified, the filter will use the named file to save the motion score of
21080 each frame with respect to the previous frame.
21081 When filename equals "-" the data is sent to standard output.
21086 ffmpeg -i ref.mpg -vf vmafmotion -f null -
21090 Stack input videos vertically.
21092 All streams must be of same pixel format and of same width.
21094 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
21095 to create same output.
21097 The filter accepts the following options:
21101 Set number of input streams. Default is 2.
21104 If set to 1, force the output to terminate when the shortest input
21105 terminates. Default value is 0.
21110 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
21111 Deinterlacing Filter").
21113 Based on the process described by Martin Weston for BBC R&D, and
21114 implemented based on the de-interlace algorithm written by Jim
21115 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
21116 uses filter coefficients calculated by BBC R&D.
21118 This filter uses field-dominance information in frame to decide which
21119 of each pair of fields to place first in the output.
21120 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
21122 There are two sets of filter coefficients, so called "simple"
21123 and "complex". Which set of filter coefficients is used can
21124 be set by passing an optional parameter:
21128 Set the interlacing filter coefficients. Accepts one of the following values:
21132 Simple filter coefficient set.
21134 More-complex filter coefficient set.
21136 Default value is @samp{complex}.
21139 The interlacing mode to adopt. It accepts one of the following values:
21143 Output one frame for each frame.
21145 Output one frame for each field.
21148 The default value is @code{field}.
21151 The picture field parity assumed for the input interlaced video. It accepts one
21152 of the following values:
21156 Assume the top field is first.
21158 Assume the bottom field is first.
21160 Enable automatic detection of field parity.
21163 The default value is @code{auto}.
21164 If the interlacing is unknown or the decoder does not export this information,
21165 top field first will be assumed.
21168 Specify which frames to deinterlace. Accepts one of the following values:
21172 Deinterlace all frames,
21174 Only deinterlace frames marked as interlaced.
21177 Default value is @samp{all}.
21180 @subsection Commands
21181 This filter supports same @ref{commands} as options.
21184 Video waveform monitor.
21186 The waveform monitor plots color component intensity. By default luminance
21187 only. Each column of the waveform corresponds to a column of pixels in the
21190 It accepts the following options:
21194 Can be either @code{row}, or @code{column}. Default is @code{column}.
21195 In row mode, the graph on the left side represents color component value 0 and
21196 the right side represents value = 255. In column mode, the top side represents
21197 color component value = 0 and bottom side represents value = 255.
21200 Set intensity. Smaller values are useful to find out how many values of the same
21201 luminance are distributed across input rows/columns.
21202 Default value is @code{0.04}. Allowed range is [0, 1].
21205 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
21206 In mirrored mode, higher values will be represented on the left
21207 side for @code{row} mode and at the top for @code{column} mode. Default is
21208 @code{1} (mirrored).
21212 It accepts the following values:
21215 Presents information identical to that in the @code{parade}, except
21216 that the graphs representing color components are superimposed directly
21219 This display mode makes it easier to spot relative differences or similarities
21220 in overlapping areas of the color components that are supposed to be identical,
21221 such as neutral whites, grays, or blacks.
21224 Display separate graph for the color components side by side in
21225 @code{row} mode or one below the other in @code{column} mode.
21228 Display separate graph for the color components side by side in
21229 @code{column} mode or one below the other in @code{row} mode.
21231 Using this display mode makes it easy to spot color casts in the highlights
21232 and shadows of an image, by comparing the contours of the top and the bottom
21233 graphs of each waveform. Since whites, grays, and blacks are characterized
21234 by exactly equal amounts of red, green, and blue, neutral areas of the picture
21235 should display three waveforms of roughly equal width/height. If not, the
21236 correction is easy to perform by making level adjustments the three waveforms.
21238 Default is @code{stack}.
21240 @item components, c
21241 Set which color components to display. Default is 1, which means only luminance
21242 or red color component if input is in RGB colorspace. If is set for example to
21243 7 it will display all 3 (if) available color components.
21248 No envelope, this is default.
21251 Instant envelope, minimum and maximum values presented in graph will be easily
21252 visible even with small @code{step} value.
21255 Hold minimum and maximum values presented in graph across time. This way you
21256 can still spot out of range values without constantly looking at waveforms.
21259 Peak and instant envelope combined together.
21265 No filtering, this is default.
21268 Luma and chroma combined together.
21271 Similar as above, but shows difference between blue and red chroma.
21274 Similar as above, but use different colors.
21277 Similar as above, but again with different colors.
21280 Displays only chroma.
21283 Displays actual color value on waveform.
21286 Similar as above, but with luma showing frequency of chroma values.
21290 Set which graticule to display.
21294 Do not display graticule.
21297 Display green graticule showing legal broadcast ranges.
21300 Display orange graticule showing legal broadcast ranges.
21303 Display invert graticule showing legal broadcast ranges.
21307 Set graticule opacity.
21310 Set graticule flags.
21314 Draw numbers above lines. By default enabled.
21317 Draw dots instead of lines.
21321 Set scale used for displaying graticule.
21328 Default is digital.
21331 Set background opacity.
21335 Set tint for output.
21336 Only used with lowpass filter and when display is not overlay and input
21337 pixel formats are not RGB.
21340 @section weave, doubleweave
21342 The @code{weave} takes a field-based video input and join
21343 each two sequential fields into single frame, producing a new double
21344 height clip with half the frame rate and half the frame count.
21346 The @code{doubleweave} works same as @code{weave} but without
21347 halving frame rate and frame count.
21349 It accepts the following option:
21353 Set first field. Available values are:
21357 Set the frame as top-field-first.
21360 Set the frame as bottom-field-first.
21364 @subsection Examples
21368 Interlace video using @ref{select} and @ref{separatefields} filter:
21370 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21375 Apply the xBR high-quality magnification filter which is designed for pixel
21376 art. It follows a set of edge-detection rules, see
21377 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21379 It accepts the following option:
21383 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21384 @code{3xBR} and @code{4} for @code{4xBR}.
21385 Default is @code{3}.
21390 Apply cross fade from one input video stream to another input video stream.
21391 The cross fade is applied for specified duration.
21393 The filter accepts the following options:
21397 Set one of available transition effects:
21445 Default transition effect is fade.
21448 Set cross fade duration in seconds.
21449 Default duration is 1 second.
21452 Set cross fade start relative to first input stream in seconds.
21453 Default offset is 0.
21456 Set expression for custom transition effect.
21458 The expressions can use the following variables and functions:
21463 The coordinates of the current sample.
21467 The width and height of the image.
21470 Progress of transition effect.
21473 Currently processed plane.
21476 Return value of first input at current location and plane.
21479 Return value of second input at current location and plane.
21485 Return the value of the pixel at location (@var{x},@var{y}) of the
21486 first/second/third/fourth component of first input.
21492 Return the value of the pixel at location (@var{x},@var{y}) of the
21493 first/second/third/fourth component of second input.
21497 @subsection Examples
21501 Cross fade from one input video to another input video, with fade transition and duration of transition
21502 of 2 seconds starting at offset of 5 seconds:
21504 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21509 Pick median pixels from several input videos.
21511 The filter accepts the following options:
21515 Set number of inputs.
21516 Default is 3. Allowed range is from 3 to 255.
21517 If number of inputs is even number, than result will be mean value between two median values.
21520 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21523 Set median percentile. Default value is @code{0.5}.
21524 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21525 minimum values, and @code{1} maximum values.
21528 @subsection Commands
21530 This filter supports all above options as @ref{commands}, excluding option @code{inputs}.
21533 Stack video inputs into custom layout.
21535 All streams must be of same pixel format.
21537 The filter accepts the following options:
21541 Set number of input streams. Default is 2.
21544 Specify layout of inputs.
21545 This option requires the desired layout configuration to be explicitly set by the user.
21546 This sets position of each video input in output. Each input
21547 is separated by '|'.
21548 The first number represents the column, and the second number represents the row.
21549 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21550 where X is video input from which to take width or height.
21551 Multiple values can be used when separated by '+'. In such
21552 case values are summed together.
21554 Note that if inputs are of different sizes gaps may appear, as not all of
21555 the output video frame will be filled. Similarly, videos can overlap each
21556 other if their position doesn't leave enough space for the full frame of
21559 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21560 a layout must be set by the user.
21563 If set to 1, force the output to terminate when the shortest input
21564 terminates. Default value is 0.
21567 If set to valid color, all unused pixels will be filled with that color.
21568 By default fill is set to none, so it is disabled.
21571 @subsection Examples
21575 Display 4 inputs into 2x2 grid.
21579 input1(0, 0) | input3(w0, 0)
21580 input2(0, h0) | input4(w0, h0)
21584 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21587 Note that if inputs are of different sizes, gaps or overlaps may occur.
21590 Display 4 inputs into 1x4 grid.
21597 input4(0, h0+h1+h2)
21601 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21604 Note that if inputs are of different widths, unused space will appear.
21607 Display 9 inputs into 3x3 grid.
21611 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21612 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21613 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21617 xstack=inputs=9:layout=0_0|0_h0|0_h0+h1|w0_0|w0_h0|w0_h0+h1|w0+w3_0|w0+w3_h0|w0+w3_h0+h1
21620 Note that if inputs are of different sizes, gaps or overlaps may occur.
21623 Display 16 inputs into 4x4 grid.
21627 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21628 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21629 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21630 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21634 xstack=inputs=16:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2|w0_0|w0_h0|w0_h0+h1|w0_h0+h1+h2|w0+w4_0|
21635 w0+w4_h0|w0+w4_h0+h1|w0+w4_h0+h1+h2|w0+w4+w8_0|w0+w4+w8_h0|w0+w4+w8_h0+h1|w0+w4+w8_h0+h1+h2
21638 Note that if inputs are of different sizes, gaps or overlaps may occur.
21645 Deinterlace the input video ("yadif" means "yet another deinterlacing
21648 It accepts the following parameters:
21654 The interlacing mode to adopt. It accepts one of the following values:
21657 @item 0, send_frame
21658 Output one frame for each frame.
21659 @item 1, send_field
21660 Output one frame for each field.
21661 @item 2, send_frame_nospatial
21662 Like @code{send_frame}, but it skips the spatial interlacing check.
21663 @item 3, send_field_nospatial
21664 Like @code{send_field}, but it skips the spatial interlacing check.
21667 The default value is @code{send_frame}.
21670 The picture field parity assumed for the input interlaced video. It accepts one
21671 of the following values:
21675 Assume the top field is first.
21677 Assume the bottom field is first.
21679 Enable automatic detection of field parity.
21682 The default value is @code{auto}.
21683 If the interlacing is unknown or the decoder does not export this information,
21684 top field first will be assumed.
21687 Specify which frames to deinterlace. Accepts one of the following
21692 Deinterlace all frames.
21693 @item 1, interlaced
21694 Only deinterlace frames marked as interlaced.
21697 The default value is @code{all}.
21700 @section yadif_cuda
21702 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21703 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21706 It accepts the following parameters:
21712 The interlacing mode to adopt. It accepts one of the following values:
21715 @item 0, send_frame
21716 Output one frame for each frame.
21717 @item 1, send_field
21718 Output one frame for each field.
21719 @item 2, send_frame_nospatial
21720 Like @code{send_frame}, but it skips the spatial interlacing check.
21721 @item 3, send_field_nospatial
21722 Like @code{send_field}, but it skips the spatial interlacing check.
21725 The default value is @code{send_frame}.
21728 The picture field parity assumed for the input interlaced video. It accepts one
21729 of the following values:
21733 Assume the top field is first.
21735 Assume the bottom field is first.
21737 Enable automatic detection of field parity.
21740 The default value is @code{auto}.
21741 If the interlacing is unknown or the decoder does not export this information,
21742 top field first will be assumed.
21745 Specify which frames to deinterlace. Accepts one of the following
21750 Deinterlace all frames.
21751 @item 1, interlaced
21752 Only deinterlace frames marked as interlaced.
21755 The default value is @code{all}.
21760 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21761 The algorithm is described in
21762 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21764 It accepts the following parameters:
21768 Set the window radius. Default value is 3.
21771 Set which planes to filter. Default is only the first plane.
21774 Set blur strength. Default value is 128.
21777 @subsection Commands
21778 This filter supports same @ref{commands} as options.
21782 Apply Zoom & Pan effect.
21784 This filter accepts the following options:
21788 Set the zoom expression. Range is 1-10. Default is 1.
21792 Set the x and y expression. Default is 0.
21795 Set the duration expression in number of frames.
21796 This sets for how many number of frames effect will last for
21797 single input image.
21800 Set the output image size, default is 'hd720'.
21803 Set the output frame rate, default is '25'.
21806 Each expression can contain the following constants:
21825 Output frame count.
21828 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21830 @item out_time, time, ot
21831 The output timestamp expressed in seconds.
21835 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21836 for current input frame.
21840 'x' and 'y' of last output frame of previous input frame or 0 when there was
21841 not yet such frame (first input frame).
21844 Last calculated zoom from 'z' expression for current input frame.
21847 Last calculated zoom of last output frame of previous input frame.
21850 Number of output frames for current input frame. Calculated from 'd' expression
21851 for each input frame.
21854 number of output frames created for previous input frame
21857 Rational number: input width / input height
21860 sample aspect ratio
21863 display aspect ratio
21867 @subsection Examples
21871 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21873 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
21877 Zoom in up to 1.5x and pan always at center of picture:
21879 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21883 Same as above but without pausing:
21885 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21889 Zoom in 2x into center of picture only for the first second of the input video:
21891 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21898 Scale (resize) the input video, using the z.lib library:
21899 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21900 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21902 The zscale filter forces the output display aspect ratio to be the same
21903 as the input, by changing the output sample aspect ratio.
21905 If the input image format is different from the format requested by
21906 the next filter, the zscale filter will convert the input to the
21909 @subsection Options
21910 The filter accepts the following options.
21915 Set the output video dimension expression. Default value is the input
21918 If the @var{width} or @var{w} value is 0, the input width is used for
21919 the output. If the @var{height} or @var{h} value is 0, the input height
21920 is used for the output.
21922 If one and only one of the values is -n with n >= 1, the zscale filter
21923 will use a value that maintains the aspect ratio of the input image,
21924 calculated from the other specified dimension. After that it will,
21925 however, make sure that the calculated dimension is divisible by n and
21926 adjust the value if necessary.
21928 If both values are -n with n >= 1, the behavior will be identical to
21929 both values being set to 0 as previously detailed.
21931 See below for the list of accepted constants for use in the dimension
21935 Set the video size. For the syntax of this option, check the
21936 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21939 Set the dither type.
21941 Possible values are:
21946 @item error_diffusion
21952 Set the resize filter type.
21954 Possible values are:
21964 Default is bilinear.
21967 Set the color range.
21969 Possible values are:
21976 Default is same as input.
21979 Set the color primaries.
21981 Possible values are:
21991 Default is same as input.
21994 Set the transfer characteristics.
21996 Possible values are:
22010 Default is same as input.
22013 Set the colorspace matrix.
22015 Possible value are:
22026 Default is same as input.
22029 Set the input color range.
22031 Possible values are:
22038 Default is same as input.
22040 @item primariesin, pin
22041 Set the input color primaries.
22043 Possible values are:
22053 Default is same as input.
22055 @item transferin, tin
22056 Set the input transfer characteristics.
22058 Possible values are:
22069 Default is same as input.
22071 @item matrixin, min
22072 Set the input colorspace matrix.
22074 Possible value are:
22086 Set the output chroma location.
22088 Possible values are:
22099 @item chromalin, cin
22100 Set the input chroma location.
22102 Possible values are:
22114 Set the nominal peak luminance.
22117 The values of the @option{w} and @option{h} options are expressions
22118 containing the following constants:
22123 The input width and height
22127 These are the same as @var{in_w} and @var{in_h}.
22131 The output (scaled) width and height
22135 These are the same as @var{out_w} and @var{out_h}
22138 The same as @var{iw} / @var{ih}
22141 input sample aspect ratio
22144 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
22148 horizontal and vertical input chroma subsample values. For example for the
22149 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22153 horizontal and vertical output chroma subsample values. For example for the
22154 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22157 @subsection Commands
22159 This filter supports the following commands:
22163 Set the output video dimension expression.
22164 The command accepts the same syntax of the corresponding option.
22166 If the specified expression is not valid, it is kept at its current
22170 @c man end VIDEO FILTERS
22172 @chapter OpenCL Video Filters
22173 @c man begin OPENCL VIDEO FILTERS
22175 Below is a description of the currently available OpenCL video filters.
22177 To enable compilation of these filters you need to configure FFmpeg with
22178 @code{--enable-opencl}.
22180 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
22183 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
22184 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
22185 given device parameters.
22187 @item -filter_hw_device @var{name}
22188 Pass the hardware device called @var{name} to all filters in any filter graph.
22192 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
22196 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
22198 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
22202 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.
22204 @section avgblur_opencl
22206 Apply average blur filter.
22208 The filter accepts the following options:
22212 Set horizontal radius size.
22213 Range is @code{[1, 1024]} and default value is @code{1}.
22216 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22219 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
22222 @subsection Example
22226 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.
22228 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
22232 @section boxblur_opencl
22234 Apply a boxblur algorithm to the input video.
22236 It accepts the following parameters:
22240 @item luma_radius, lr
22241 @item luma_power, lp
22242 @item chroma_radius, cr
22243 @item chroma_power, cp
22244 @item alpha_radius, ar
22245 @item alpha_power, ap
22249 A description of the accepted options follows.
22252 @item luma_radius, lr
22253 @item chroma_radius, cr
22254 @item alpha_radius, ar
22255 Set an expression for the box radius in pixels used for blurring the
22256 corresponding input plane.
22258 The radius value must be a non-negative number, and must not be
22259 greater than the value of the expression @code{min(w,h)/2} for the
22260 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
22263 Default value for @option{luma_radius} is "2". If not specified,
22264 @option{chroma_radius} and @option{alpha_radius} default to the
22265 corresponding value set for @option{luma_radius}.
22267 The expressions can contain the following constants:
22271 The input width and height in pixels.
22275 The input chroma image width and height in pixels.
22279 The horizontal and vertical chroma subsample values. For example, for the
22280 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
22283 @item luma_power, lp
22284 @item chroma_power, cp
22285 @item alpha_power, ap
22286 Specify how many times the boxblur filter is applied to the
22287 corresponding plane.
22289 Default value for @option{luma_power} is 2. If not specified,
22290 @option{chroma_power} and @option{alpha_power} default to the
22291 corresponding value set for @option{luma_power}.
22293 A value of 0 will disable the effect.
22296 @subsection Examples
22298 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.
22302 Apply a boxblur filter with the luma, chroma, and alpha radius
22303 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.
22305 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
22306 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
22310 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.
22312 For the luma plane, a 2x2 box radius will be run once.
22314 For the chroma plane, a 4x4 box radius will be run 5 times.
22316 For the alpha plane, a 3x3 box radius will be run 7 times.
22318 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
22322 @section colorkey_opencl
22323 RGB colorspace color keying.
22325 The filter accepts the following options:
22329 The color which will be replaced with transparency.
22332 Similarity percentage with the key color.
22334 0.01 matches only the exact key color, while 1.0 matches everything.
22339 0.0 makes pixels either fully transparent, or not transparent at all.
22341 Higher values result in semi-transparent pixels, with a higher transparency
22342 the more similar the pixels color is to the key color.
22345 @subsection Examples
22349 Make every semi-green pixel in the input transparent with some slight blending:
22351 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22355 @section convolution_opencl
22357 Apply convolution of 3x3, 5x5, 7x7 matrix.
22359 The filter accepts the following options:
22366 Set matrix for each plane.
22367 Matrix is sequence of 9, 25 or 49 signed numbers.
22368 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22374 Set multiplier for calculated value for each plane.
22375 If unset or 0, it will be sum of all matrix elements.
22376 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22382 Set bias for each plane. This value is added to the result of the multiplication.
22383 Useful for making the overall image brighter or darker.
22384 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22388 @subsection Examples
22394 -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
22400 -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
22404 Apply edge enhance:
22406 -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
22412 -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
22416 Apply laplacian edge detector which includes diagonals:
22418 -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
22424 -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
22428 @section erosion_opencl
22430 Apply erosion effect to the video.
22432 This filter replaces the pixel by the local(3x3) minimum.
22434 It accepts the following options:
22441 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22442 If @code{0}, plane will remain unchanged.
22445 Flag which specifies the pixel to refer to.
22446 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22448 Flags to local 3x3 coordinates region centered on @code{x}:
22457 @subsection Example
22461 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.
22463 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22467 @section deshake_opencl
22468 Feature-point based video stabilization filter.
22470 The filter accepts the following options:
22474 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22477 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22479 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22481 Viewing point matches in the output video is only supported for RGB input.
22483 Defaults to @code{0}.
22485 @item adaptive_crop
22486 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22488 Defaults to @code{1}.
22490 @item refine_features
22491 Whether or not feature points should be refined at a sub-pixel level.
22493 This can be turned off for a slight performance gain at the cost of precision.
22495 Defaults to @code{1}.
22497 @item smooth_strength
22498 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22500 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22502 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22504 Defaults to @code{0.0}.
22506 @item smooth_window_multiplier
22507 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22509 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22511 Acceptable values range from @code{0.1} to @code{10.0}.
22513 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22514 potentially improving smoothness, but also increase latency and memory usage.
22516 Defaults to @code{2.0}.
22520 @subsection Examples
22524 Stabilize a video with a fixed, medium smoothing strength:
22526 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22530 Stabilize a video with debugging (both in console and in rendered video):
22532 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22536 @section dilation_opencl
22538 Apply dilation effect to the video.
22540 This filter replaces the pixel by the local(3x3) maximum.
22542 It accepts the following options:
22549 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22550 If @code{0}, plane will remain unchanged.
22553 Flag which specifies the pixel to refer to.
22554 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22556 Flags to local 3x3 coordinates region centered on @code{x}:
22565 @subsection Example
22569 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.
22571 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22575 @section nlmeans_opencl
22577 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22579 @section overlay_opencl
22581 Overlay one video on top of another.
22583 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22584 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22586 The filter accepts the following options:
22591 Set the x coordinate of the overlaid video on the main video.
22592 Default value is @code{0}.
22595 Set the y coordinate of the overlaid video on the main video.
22596 Default value is @code{0}.
22600 @subsection Examples
22604 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22606 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22609 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22611 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22616 @section pad_opencl
22618 Add paddings to the input image, and place the original input at the
22619 provided @var{x}, @var{y} coordinates.
22621 It accepts the following options:
22626 Specify an expression for the size of the output image with the
22627 paddings added. If the value for @var{width} or @var{height} is 0, the
22628 corresponding input size is used for the output.
22630 The @var{width} expression can reference the value set by the
22631 @var{height} expression, and vice versa.
22633 The default value of @var{width} and @var{height} is 0.
22637 Specify the offsets to place the input image at within the padded area,
22638 with respect to the top/left border of the output image.
22640 The @var{x} expression can reference the value set by the @var{y}
22641 expression, and vice versa.
22643 The default value of @var{x} and @var{y} is 0.
22645 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22646 so the input image is centered on the padded area.
22649 Specify the color of the padded area. For the syntax of this option,
22650 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22651 manual,ffmpeg-utils}.
22654 Pad to an aspect instead to a resolution.
22657 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22658 options are expressions containing the following constants:
22663 The input video width and height.
22667 These are the same as @var{in_w} and @var{in_h}.
22671 The output width and height (the size of the padded area), as
22672 specified by the @var{width} and @var{height} expressions.
22676 These are the same as @var{out_w} and @var{out_h}.
22680 The x and y offsets as specified by the @var{x} and @var{y}
22681 expressions, or NAN if not yet specified.
22684 same as @var{iw} / @var{ih}
22687 input sample aspect ratio
22690 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22693 @section prewitt_opencl
22695 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22697 The filter accepts the following option:
22701 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22704 Set value which will be multiplied with filtered result.
22705 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22708 Set value which will be added to filtered result.
22709 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22712 @subsection Example
22716 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22718 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22722 @anchor{program_opencl}
22723 @section program_opencl
22725 Filter video using an OpenCL program.
22730 OpenCL program source file.
22733 Kernel name in program.
22736 Number of inputs to the filter. Defaults to 1.
22739 Size of output frames. Defaults to the same as the first input.
22743 The @code{program_opencl} filter also supports the @ref{framesync} options.
22745 The program source file must contain a kernel function with the given name,
22746 which will be run once for each plane of the output. Each run on a plane
22747 gets enqueued as a separate 2D global NDRange with one work-item for each
22748 pixel to be generated. The global ID offset for each work-item is therefore
22749 the coordinates of a pixel in the destination image.
22751 The kernel function needs to take the following arguments:
22754 Destination image, @var{__write_only image2d_t}.
22756 This image will become the output; the kernel should write all of it.
22758 Frame index, @var{unsigned int}.
22760 This is a counter starting from zero and increasing by one for each frame.
22762 Source images, @var{__read_only image2d_t}.
22764 These are the most recent images on each input. The kernel may read from
22765 them to generate the output, but they can't be written to.
22772 Copy the input to the output (output must be the same size as the input).
22774 __kernel void copy(__write_only image2d_t destination,
22775 unsigned int index,
22776 __read_only image2d_t source)
22778 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22780 int2 location = (int2)(get_global_id(0), get_global_id(1));
22782 float4 value = read_imagef(source, sampler, location);
22784 write_imagef(destination, location, value);
22789 Apply a simple transformation, rotating the input by an amount increasing
22790 with the index counter. Pixel values are linearly interpolated by the
22791 sampler, and the output need not have the same dimensions as the input.
22793 __kernel void rotate_image(__write_only image2d_t dst,
22794 unsigned int index,
22795 __read_only image2d_t src)
22797 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22798 CLK_FILTER_LINEAR);
22800 float angle = (float)index / 100.0f;
22802 float2 dst_dim = convert_float2(get_image_dim(dst));
22803 float2 src_dim = convert_float2(get_image_dim(src));
22805 float2 dst_cen = dst_dim / 2.0f;
22806 float2 src_cen = src_dim / 2.0f;
22808 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22810 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22812 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22813 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22815 src_pos = src_pos * src_dim / dst_dim;
22817 float2 src_loc = src_pos + src_cen;
22819 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22820 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22821 write_imagef(dst, dst_loc, 0.5f);
22823 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22828 Blend two inputs together, with the amount of each input used varying
22829 with the index counter.
22831 __kernel void blend_images(__write_only image2d_t dst,
22832 unsigned int index,
22833 __read_only image2d_t src1,
22834 __read_only image2d_t src2)
22836 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22837 CLK_FILTER_LINEAR);
22839 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22841 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22842 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22843 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22845 float4 val1 = read_imagef(src1, sampler, src1_loc);
22846 float4 val2 = read_imagef(src2, sampler, src2_loc);
22848 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22854 @section roberts_opencl
22855 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22857 The filter accepts the following option:
22861 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22864 Set value which will be multiplied with filtered result.
22865 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22868 Set value which will be added to filtered result.
22869 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22872 @subsection Example
22876 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22878 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22882 @section sobel_opencl
22884 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22886 The filter accepts the following option:
22890 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22893 Set value which will be multiplied with filtered result.
22894 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22897 Set value which will be added to filtered result.
22898 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22901 @subsection Example
22905 Apply sobel operator with scale set to 2 and delta set to 10
22907 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22911 @section tonemap_opencl
22913 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22915 It accepts the following parameters:
22919 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22922 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22925 Apply desaturation for highlights that exceed this level of brightness. The
22926 higher the parameter, the more color information will be preserved. This
22927 setting helps prevent unnaturally blown-out colors for super-highlights, by
22928 (smoothly) turning into white instead. This makes images feel more natural,
22929 at the cost of reducing information about out-of-range colors.
22931 The default value is 0.5, and the algorithm here is a little different from
22932 the cpu version tonemap currently. A setting of 0.0 disables this option.
22935 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22936 is used to detect whether the scene has changed or not. If the distance between
22937 the current frame average brightness and the current running average exceeds
22938 a threshold value, we would re-calculate scene average and peak brightness.
22939 The default value is 0.2.
22942 Specify the output pixel format.
22944 Currently supported formats are:
22951 Set the output color range.
22953 Possible values are:
22959 Default is same as input.
22962 Set the output color primaries.
22964 Possible values are:
22970 Default is same as input.
22973 Set the output transfer characteristics.
22975 Possible values are:
22984 Set the output colorspace matrix.
22986 Possible value are:
22992 Default is same as input.
22996 @subsection Example
23000 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
23002 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
23006 @section unsharp_opencl
23008 Sharpen or blur the input video.
23010 It accepts the following parameters:
23013 @item luma_msize_x, lx
23014 Set the luma matrix horizontal size.
23015 Range is @code{[1, 23]} and default value is @code{5}.
23017 @item luma_msize_y, ly
23018 Set the luma matrix vertical size.
23019 Range is @code{[1, 23]} and default value is @code{5}.
23021 @item luma_amount, la
23022 Set the luma effect strength.
23023 Range is @code{[-10, 10]} and default value is @code{1.0}.
23025 Negative values will blur the input video, while positive values will
23026 sharpen it, a value of zero will disable the effect.
23028 @item chroma_msize_x, cx
23029 Set the chroma matrix horizontal size.
23030 Range is @code{[1, 23]} and default value is @code{5}.
23032 @item chroma_msize_y, cy
23033 Set the chroma matrix vertical size.
23034 Range is @code{[1, 23]} and default value is @code{5}.
23036 @item chroma_amount, ca
23037 Set the chroma effect strength.
23038 Range is @code{[-10, 10]} and default value is @code{0.0}.
23040 Negative values will blur the input video, while positive values will
23041 sharpen it, a value of zero will disable the effect.
23045 All parameters are optional and default to the equivalent of the
23046 string '5:5:1.0:5:5:0.0'.
23048 @subsection Examples
23052 Apply strong luma sharpen effect:
23054 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
23058 Apply a strong blur of both luma and chroma parameters:
23060 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
23064 @section xfade_opencl
23066 Cross fade two videos with custom transition effect by using OpenCL.
23068 It accepts the following options:
23072 Set one of possible transition effects.
23076 Select custom transition effect, the actual transition description
23077 will be picked from source and kernel options.
23089 Default transition is fade.
23093 OpenCL program source file for custom transition.
23096 Set name of kernel to use for custom transition from program source file.
23099 Set duration of video transition.
23102 Set time of start of transition relative to first video.
23105 The program source file must contain a kernel function with the given name,
23106 which will be run once for each plane of the output. Each run on a plane
23107 gets enqueued as a separate 2D global NDRange with one work-item for each
23108 pixel to be generated. The global ID offset for each work-item is therefore
23109 the coordinates of a pixel in the destination image.
23111 The kernel function needs to take the following arguments:
23114 Destination image, @var{__write_only image2d_t}.
23116 This image will become the output; the kernel should write all of it.
23119 First Source image, @var{__read_only image2d_t}.
23120 Second Source image, @var{__read_only image2d_t}.
23122 These are the most recent images on each input. The kernel may read from
23123 them to generate the output, but they can't be written to.
23126 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
23133 Apply dots curtain transition effect:
23135 __kernel void blend_images(__write_only image2d_t dst,
23136 __read_only image2d_t src1,
23137 __read_only image2d_t src2,
23140 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23141 CLK_FILTER_LINEAR);
23142 int2 p = (int2)(get_global_id(0), get_global_id(1));
23143 float2 rp = (float2)(get_global_id(0), get_global_id(1));
23144 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
23147 float2 dots = (float2)(20.0, 20.0);
23148 float2 center = (float2)(0,0);
23151 float4 val1 = read_imagef(src1, sampler, p);
23152 float4 val2 = read_imagef(src2, sampler, p);
23153 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
23155 write_imagef(dst, p, next ? val1 : val2);
23161 @c man end OPENCL VIDEO FILTERS
23163 @chapter VAAPI Video Filters
23164 @c man begin VAAPI VIDEO FILTERS
23166 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
23168 To enable compilation of these filters you need to configure FFmpeg with
23169 @code{--enable-vaapi}.
23171 To use vaapi filters, you need to setup the vaapi device correctly. For more information, please read @url{https://trac.ffmpeg.org/wiki/Hardware/VAAPI}
23173 @section tonemap_vaapi
23175 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
23176 It maps the dynamic range of HDR10 content to the SDR content.
23177 It currently only accepts HDR10 as input.
23179 It accepts the following parameters:
23183 Specify the output pixel format.
23185 Currently supported formats are:
23194 Set the output color primaries.
23196 Default is same as input.
23199 Set the output transfer characteristics.
23204 Set the output colorspace matrix.
23206 Default is same as input.
23210 @subsection Example
23214 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
23216 tonemap_vaapi=format=p010:t=bt2020-10
23220 @c man end VAAPI VIDEO FILTERS
23222 @chapter Video Sources
23223 @c man begin VIDEO SOURCES
23225 Below is a description of the currently available video sources.
23229 Buffer video frames, and make them available to the filter chain.
23231 This source is mainly intended for a programmatic use, in particular
23232 through the interface defined in @file{libavfilter/buffersrc.h}.
23234 It accepts the following parameters:
23239 Specify the size (width and height) of the buffered video frames. For the
23240 syntax of this option, check the
23241 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23244 The input video width.
23247 The input video height.
23250 A string representing the pixel format of the buffered video frames.
23251 It may be a number corresponding to a pixel format, or a pixel format
23255 Specify the timebase assumed by the timestamps of the buffered frames.
23258 Specify the frame rate expected for the video stream.
23260 @item pixel_aspect, sar
23261 The sample (pixel) aspect ratio of the input video.
23264 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
23265 to the filtergraph description to specify swscale flags for automatically
23266 inserted scalers. See @ref{Filtergraph syntax}.
23268 @item hw_frames_ctx
23269 When using a hardware pixel format, this should be a reference to an
23270 AVHWFramesContext describing input frames.
23275 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
23278 will instruct the source to accept video frames with size 320x240 and
23279 with format "yuv410p", assuming 1/24 as the timestamps timebase and
23280 square pixels (1:1 sample aspect ratio).
23281 Since the pixel format with name "yuv410p" corresponds to the number 6
23282 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
23283 this example corresponds to:
23285 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
23288 Alternatively, the options can be specified as a flat string, but this
23289 syntax is deprecated:
23291 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
23295 Create a pattern generated by an elementary cellular automaton.
23297 The initial state of the cellular automaton can be defined through the
23298 @option{filename} and @option{pattern} options. If such options are
23299 not specified an initial state is created randomly.
23301 At each new frame a new row in the video is filled with the result of
23302 the cellular automaton next generation. The behavior when the whole
23303 frame is filled is defined by the @option{scroll} option.
23305 This source accepts the following options:
23309 Read the initial cellular automaton state, i.e. the starting row, from
23310 the specified file.
23311 In the file, each non-whitespace character is considered an alive
23312 cell, a newline will terminate the row, and further characters in the
23313 file will be ignored.
23316 Read the initial cellular automaton state, i.e. the starting row, from
23317 the specified string.
23319 Each non-whitespace character in the string is considered an alive
23320 cell, a newline will terminate the row, and further characters in the
23321 string will be ignored.
23324 Set the video rate, that is the number of frames generated per second.
23327 @item random_fill_ratio, ratio
23328 Set the random fill ratio for the initial cellular automaton row. It
23329 is a floating point number value ranging from 0 to 1, defaults to
23332 This option is ignored when a file or a pattern is specified.
23334 @item random_seed, seed
23335 Set the seed for filling randomly the initial row, must be an integer
23336 included between 0 and UINT32_MAX. If not specified, or if explicitly
23337 set to -1, the filter will try to use a good random seed on a best
23341 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23342 Default value is 110.
23345 Set the size of the output video. For the syntax of this option, check the
23346 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23348 If @option{filename} or @option{pattern} is specified, the size is set
23349 by default to the width of the specified initial state row, and the
23350 height is set to @var{width} * PHI.
23352 If @option{size} is set, it must contain the width of the specified
23353 pattern string, and the specified pattern will be centered in the
23356 If a filename or a pattern string is not specified, the size value
23357 defaults to "320x518" (used for a randomly generated initial state).
23360 If set to 1, scroll the output upward when all the rows in the output
23361 have been already filled. If set to 0, the new generated row will be
23362 written over the top row just after the bottom row is filled.
23365 @item start_full, full
23366 If set to 1, completely fill the output with generated rows before
23367 outputting the first frame.
23368 This is the default behavior, for disabling set the value to 0.
23371 If set to 1, stitch the left and right row edges together.
23372 This is the default behavior, for disabling set the value to 0.
23375 @subsection Examples
23379 Read the initial state from @file{pattern}, and specify an output of
23382 cellauto=f=pattern:s=200x400
23386 Generate a random initial row with a width of 200 cells, with a fill
23389 cellauto=ratio=2/3:s=200x200
23393 Create a pattern generated by rule 18 starting by a single alive cell
23394 centered on an initial row with width 100:
23396 cellauto=p=@@:s=100x400:full=0:rule=18
23400 Specify a more elaborated initial pattern:
23402 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23407 @anchor{coreimagesrc}
23408 @section coreimagesrc
23409 Video source generated on GPU using Apple's CoreImage API on OSX.
23411 This video source is a specialized version of the @ref{coreimage} video filter.
23412 Use a core image generator at the beginning of the applied filterchain to
23413 generate the content.
23415 The coreimagesrc video source accepts the following options:
23417 @item list_generators
23418 List all available generators along with all their respective options as well as
23419 possible minimum and maximum values along with the default values.
23421 list_generators=true
23425 Specify the size of the sourced video. For the syntax of this option, check the
23426 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23427 The default value is @code{320x240}.
23430 Specify the frame rate of the sourced video, as the number of frames
23431 generated per second. It has to be a string in the format
23432 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23433 number or a valid video frame rate abbreviation. The default value is
23437 Set the sample aspect ratio of the sourced video.
23440 Set the duration of the sourced video. See
23441 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23442 for the accepted syntax.
23444 If not specified, or the expressed duration is negative, the video is
23445 supposed to be generated forever.
23448 Additionally, all options of the @ref{coreimage} video filter are accepted.
23449 A complete filterchain can be used for further processing of the
23450 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23451 and examples for details.
23453 @subsection Examples
23458 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23459 given as complete and escaped command-line for Apple's standard bash shell:
23461 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23463 This example is equivalent to the QRCode example of @ref{coreimage} without the
23464 need for a nullsrc video source.
23469 Generate several gradients.
23473 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23474 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23477 Set frame rate, expressed as number of frames per second. Default
23480 @item c0, c1, c2, c3, c4, c5, c6, c7
23481 Set 8 colors. Default values for colors is to pick random one.
23483 @item x0, y0, y0, y1
23484 Set gradient line source and destination points. If negative or out of range, random ones
23488 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23491 Set seed for picking gradient line points.
23494 Set the duration of the sourced video. See
23495 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23496 for the accepted syntax.
23498 If not specified, or the expressed duration is negative, the video is
23499 supposed to be generated forever.
23502 Set speed of gradients rotation.
23506 @section mandelbrot
23508 Generate a Mandelbrot set fractal, and progressively zoom towards the
23509 point specified with @var{start_x} and @var{start_y}.
23511 This source accepts the following options:
23516 Set the terminal pts value. Default value is 400.
23519 Set the terminal scale value.
23520 Must be a floating point value. Default value is 0.3.
23523 Set the inner coloring mode, that is the algorithm used to draw the
23524 Mandelbrot fractal internal region.
23526 It shall assume one of the following values:
23531 Show time until convergence.
23533 Set color based on point closest to the origin of the iterations.
23538 Default value is @var{mincol}.
23541 Set the bailout value. Default value is 10.0.
23544 Set the maximum of iterations performed by the rendering
23545 algorithm. Default value is 7189.
23548 Set outer coloring mode.
23549 It shall assume one of following values:
23551 @item iteration_count
23552 Set iteration count mode.
23553 @item normalized_iteration_count
23554 set normalized iteration count mode.
23556 Default value is @var{normalized_iteration_count}.
23559 Set frame rate, expressed as number of frames per second. Default
23563 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23564 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23567 Set the initial scale value. Default value is 3.0.
23570 Set the initial x position. Must be a floating point value between
23571 -100 and 100. Default value is -0.743643887037158704752191506114774.
23574 Set the initial y position. Must be a floating point value between
23575 -100 and 100. Default value is -0.131825904205311970493132056385139.
23580 Generate various test patterns, as generated by the MPlayer test filter.
23582 The size of the generated video is fixed, and is 256x256.
23583 This source is useful in particular for testing encoding features.
23585 This source accepts the following options:
23590 Specify the frame rate of the sourced video, as the number of frames
23591 generated per second. It has to be a string in the format
23592 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23593 number or a valid video frame rate abbreviation. The default value is
23597 Set the duration of the sourced video. See
23598 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23599 for the accepted syntax.
23601 If not specified, or the expressed duration is negative, the video is
23602 supposed to be generated forever.
23606 Set the number or the name of the test to perform. Supported tests are:
23620 @item max_frames, m
23621 Set the maximum number of frames generated for each test, default value is 30.
23625 Default value is "all", which will cycle through the list of all tests.
23630 mptestsrc=t=dc_luma
23633 will generate a "dc_luma" test pattern.
23635 @section frei0r_src
23637 Provide a frei0r source.
23639 To enable compilation of this filter you need to install the frei0r
23640 header and configure FFmpeg with @code{--enable-frei0r}.
23642 This source accepts the following parameters:
23647 The size of the video to generate. For the syntax of this option, check the
23648 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23651 The framerate of the generated video. It may be a string of the form
23652 @var{num}/@var{den} or a frame rate abbreviation.
23655 The name to the frei0r source to load. For more information regarding frei0r and
23656 how to set the parameters, read the @ref{frei0r} section in the video filters
23659 @item filter_params
23660 A '|'-separated list of parameters to pass to the frei0r source.
23664 For example, to generate a frei0r partik0l source with size 200x200
23665 and frame rate 10 which is overlaid on the overlay filter main input:
23667 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23672 Generate a life pattern.
23674 This source is based on a generalization of John Conway's life game.
23676 The sourced input represents a life grid, each pixel represents a cell
23677 which can be in one of two possible states, alive or dead. Every cell
23678 interacts with its eight neighbours, which are the cells that are
23679 horizontally, vertically, or diagonally adjacent.
23681 At each interaction the grid evolves according to the adopted rule,
23682 which specifies the number of neighbor alive cells which will make a
23683 cell stay alive or born. The @option{rule} option allows one to specify
23686 This source accepts the following options:
23690 Set the file from which to read the initial grid state. In the file,
23691 each non-whitespace character is considered an alive cell, and newline
23692 is used to delimit the end of each row.
23694 If this option is not specified, the initial grid is generated
23698 Set the video rate, that is the number of frames generated per second.
23701 @item random_fill_ratio, ratio
23702 Set the random fill ratio for the initial random grid. It is a
23703 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23704 It is ignored when a file is specified.
23706 @item random_seed, seed
23707 Set the seed for filling the initial random grid, must be an integer
23708 included between 0 and UINT32_MAX. If not specified, or if explicitly
23709 set to -1, the filter will try to use a good random seed on a best
23715 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23716 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23717 @var{NS} specifies the number of alive neighbor cells which make a
23718 live cell stay alive, and @var{NB} the number of alive neighbor cells
23719 which make a dead cell to become alive (i.e. to "born").
23720 "s" and "b" can be used in place of "S" and "B", respectively.
23722 Alternatively a rule can be specified by an 18-bits integer. The 9
23723 high order bits are used to encode the next cell state if it is alive
23724 for each number of neighbor alive cells, the low order bits specify
23725 the rule for "borning" new cells. Higher order bits encode for an
23726 higher number of neighbor cells.
23727 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23728 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23730 Default value is "S23/B3", which is the original Conway's game of life
23731 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23732 cells, and will born a new cell if there are three alive cells around
23736 Set the size of the output video. For the syntax of this option, check the
23737 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23739 If @option{filename} is specified, the size is set by default to the
23740 same size of the input file. If @option{size} is set, it must contain
23741 the size specified in the input file, and the initial grid defined in
23742 that file is centered in the larger resulting area.
23744 If a filename is not specified, the size value defaults to "320x240"
23745 (used for a randomly generated initial grid).
23748 If set to 1, stitch the left and right grid edges together, and the
23749 top and bottom edges also. Defaults to 1.
23752 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23753 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23754 value from 0 to 255.
23757 Set the color of living (or new born) cells.
23760 Set the color of dead cells. If @option{mold} is set, this is the first color
23761 used to represent a dead cell.
23764 Set mold color, for definitely dead and moldy cells.
23766 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23767 ffmpeg-utils manual,ffmpeg-utils}.
23770 @subsection Examples
23774 Read a grid from @file{pattern}, and center it on a grid of size
23777 life=f=pattern:s=300x300
23781 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23783 life=ratio=2/3:s=200x200
23787 Specify a custom rule for evolving a randomly generated grid:
23793 Full example with slow death effect (mold) using @command{ffplay}:
23795 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23802 @anchor{haldclutsrc}
23805 @anchor{pal100bars}
23806 @anchor{rgbtestsrc}
23808 @anchor{smptehdbars}
23811 @anchor{yuvtestsrc}
23812 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23814 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23816 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23818 The @code{color} source provides an uniformly colored input.
23820 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23821 @ref{haldclut} filter.
23823 The @code{nullsrc} source returns unprocessed video frames. It is
23824 mainly useful to be employed in analysis / debugging tools, or as the
23825 source for filters which ignore the input data.
23827 The @code{pal75bars} source generates a color bars pattern, based on
23828 EBU PAL recommendations with 75% color levels.
23830 The @code{pal100bars} source generates a color bars pattern, based on
23831 EBU PAL recommendations with 100% color levels.
23833 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23834 detecting RGB vs BGR issues. You should see a red, green and blue
23835 stripe from top to bottom.
23837 The @code{smptebars} source generates a color bars pattern, based on
23838 the SMPTE Engineering Guideline EG 1-1990.
23840 The @code{smptehdbars} source generates a color bars pattern, based on
23841 the SMPTE RP 219-2002.
23843 The @code{testsrc} source generates a test video pattern, showing a
23844 color pattern, a scrolling gradient and a timestamp. This is mainly
23845 intended for testing purposes.
23847 The @code{testsrc2} source is similar to testsrc, but supports more
23848 pixel formats instead of just @code{rgb24}. This allows using it as an
23849 input for other tests without requiring a format conversion.
23851 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23852 see a y, cb and cr stripe from top to bottom.
23854 The sources accept the following parameters:
23859 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23860 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23861 pixels to be used as identity matrix for 3D lookup tables. Each component is
23862 coded on a @code{1/(N*N)} scale.
23865 Specify the color of the source, only available in the @code{color}
23866 source. For the syntax of this option, check the
23867 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23870 Specify the size of the sourced video. For the syntax of this option, check the
23871 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23872 The default value is @code{320x240}.
23874 This option is not available with the @code{allrgb}, @code{allyuv}, and
23875 @code{haldclutsrc} filters.
23878 Specify the frame rate of the sourced video, as the number of frames
23879 generated per second. It has to be a string in the format
23880 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23881 number or a valid video frame rate abbreviation. The default value is
23885 Set the duration of the sourced video. See
23886 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23887 for the accepted syntax.
23889 If not specified, or the expressed duration is negative, the video is
23890 supposed to be generated forever.
23892 Since the frame rate is used as time base, all frames including the last one
23893 will have their full duration. If the specified duration is not a multiple
23894 of the frame duration, it will be rounded up.
23897 Set the sample aspect ratio of the sourced video.
23900 Specify the alpha (opacity) of the background, only available in the
23901 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23902 255 (fully opaque, the default).
23905 Set the number of decimals to show in the timestamp, only available in the
23906 @code{testsrc} source.
23908 The displayed timestamp value will correspond to the original
23909 timestamp value multiplied by the power of 10 of the specified
23910 value. Default value is 0.
23913 @subsection Examples
23917 Generate a video with a duration of 5.3 seconds, with size
23918 176x144 and a frame rate of 10 frames per second:
23920 testsrc=duration=5.3:size=qcif:rate=10
23924 The following graph description will generate a red source
23925 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23928 color=c=red@@0.2:s=qcif:r=10
23932 If the input content is to be ignored, @code{nullsrc} can be used. The
23933 following command generates noise in the luminance plane by employing
23934 the @code{geq} filter:
23936 nullsrc=s=256x256, geq=random(1)*255:128:128
23940 @subsection Commands
23942 The @code{color} source supports the following commands:
23946 Set the color of the created image. Accepts the same syntax of the
23947 corresponding @option{color} option.
23952 Generate video using an OpenCL program.
23957 OpenCL program source file.
23960 Kernel name in program.
23963 Size of frames to generate. This must be set.
23966 Pixel format to use for the generated frames. This must be set.
23969 Number of frames generated every second. Default value is '25'.
23973 For details of how the program loading works, see the @ref{program_opencl}
23980 Generate a colour ramp by setting pixel values from the position of the pixel
23981 in the output image. (Note that this will work with all pixel formats, but
23982 the generated output will not be the same.)
23984 __kernel void ramp(__write_only image2d_t dst,
23985 unsigned int index)
23987 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23990 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23992 write_imagef(dst, loc, val);
23997 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23999 __kernel void sierpinski_carpet(__write_only image2d_t dst,
24000 unsigned int index)
24002 int2 loc = (int2)(get_global_id(0), get_global_id(1));
24004 float4 value = 0.0f;
24005 int x = loc.x + index;
24006 int y = loc.y + index;
24007 while (x > 0 || y > 0) {
24008 if (x % 3 == 1 && y % 3 == 1) {
24016 write_imagef(dst, loc, value);
24022 @section sierpinski
24024 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
24026 This source accepts the following options:
24030 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
24031 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
24034 Set frame rate, expressed as number of frames per second. Default
24038 Set seed which is used for random panning.
24041 Set max jump for single pan destination. Allowed range is from 1 to 10000.
24044 Set fractal type, can be default @code{carpet} or @code{triangle}.
24047 @c man end VIDEO SOURCES
24049 @chapter Video Sinks
24050 @c man begin VIDEO SINKS
24052 Below is a description of the currently available video sinks.
24054 @section buffersink
24056 Buffer video frames, and make them available to the end of the filter
24059 This sink is mainly intended for programmatic use, in particular
24060 through the interface defined in @file{libavfilter/buffersink.h}
24061 or the options system.
24063 It accepts a pointer to an AVBufferSinkContext structure, which
24064 defines the incoming buffers' formats, to be passed as the opaque
24065 parameter to @code{avfilter_init_filter} for initialization.
24069 Null video sink: do absolutely nothing with the input video. It is
24070 mainly useful as a template and for use in analysis / debugging
24073 @c man end VIDEO SINKS
24075 @chapter Multimedia Filters
24076 @c man begin MULTIMEDIA FILTERS
24078 Below is a description of the currently available multimedia filters.
24082 Convert input audio to a video output, displaying the audio bit scope.
24084 The filter accepts the following options:
24088 Set frame rate, expressed as number of frames per second. Default
24092 Specify the video size for the output. For the syntax of this option, check the
24093 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24094 Default value is @code{1024x256}.
24097 Specify list of colors separated by space or by '|' which will be used to
24098 draw channels. Unrecognized or missing colors will be replaced
24102 @section adrawgraph
24103 Draw a graph using input audio metadata.
24105 See @ref{drawgraph}
24107 @section agraphmonitor
24109 See @ref{graphmonitor}.
24111 @section ahistogram
24113 Convert input audio to a video output, displaying the volume histogram.
24115 The filter accepts the following options:
24119 Specify how histogram is calculated.
24121 It accepts the following values:
24124 Use single histogram for all channels.
24126 Use separate histogram for each channel.
24128 Default is @code{single}.
24131 Set frame rate, expressed as number of frames per second. Default
24135 Specify the video size for the output. For the syntax of this option, check the
24136 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24137 Default value is @code{hd720}.
24142 It accepts the following values:
24153 reverse logarithmic
24155 Default is @code{log}.
24158 Set amplitude scale.
24160 It accepts the following values:
24167 Default is @code{log}.
24170 Set how much frames to accumulate in histogram.
24171 Default is 1. Setting this to -1 accumulates all frames.
24174 Set histogram ratio of window height.
24177 Set sonogram sliding.
24179 It accepts the following values:
24182 replace old rows with new ones.
24184 scroll from top to bottom.
24186 Default is @code{replace}.
24189 @section aphasemeter
24191 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
24192 representing mean phase of current audio frame. A video output can also be produced and is
24193 enabled by default. The audio is passed through as first output.
24195 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
24196 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
24197 and @code{1} means channels are in phase.
24199 The filter accepts the following options, all related to its video output:
24203 Set the output frame rate. Default value is @code{25}.
24206 Set the video size for the output. For the syntax of this option, check the
24207 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24208 Default value is @code{800x400}.
24213 Specify the red, green, blue contrast. Default values are @code{2},
24214 @code{7} and @code{1}.
24215 Allowed range is @code{[0, 255]}.
24218 Set color which will be used for drawing median phase. If color is
24219 @code{none} which is default, no median phase value will be drawn.
24222 Enable video output. Default is enabled.
24225 @subsection phasing detection
24227 The filter also detects out of phase and mono sequences in stereo streams.
24228 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
24230 The filter accepts the following options for this detection:
24234 Enable mono and out of phase detection. Default is disabled.
24237 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
24238 Allowed range is @code{[0, 1]}.
24241 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
24242 Allowed range is @code{[90, 180]}.
24245 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
24248 @subsection Examples
24252 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
24254 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
24258 @section avectorscope
24260 Convert input audio to a video output, representing the audio vector
24263 The filter is used to measure the difference between channels of stereo
24264 audio stream. A monaural signal, consisting of identical left and right
24265 signal, results in straight vertical line. Any stereo separation is visible
24266 as a deviation from this line, creating a Lissajous figure.
24267 If the straight (or deviation from it) but horizontal line appears this
24268 indicates that the left and right channels are out of phase.
24270 The filter accepts the following options:
24274 Set the vectorscope mode.
24276 Available values are:
24279 Lissajous rotated by 45 degrees.
24282 Same as above but not rotated.
24285 Shape resembling half of circle.
24288 Default value is @samp{lissajous}.
24291 Set the video size for the output. For the syntax of this option, check the
24292 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24293 Default value is @code{400x400}.
24296 Set the output frame rate. Default value is @code{25}.
24302 Specify the red, green, blue and alpha contrast. Default values are @code{40},
24303 @code{160}, @code{80} and @code{255}.
24304 Allowed range is @code{[0, 255]}.
24310 Specify the red, green, blue and alpha fade. Default values are @code{15},
24311 @code{10}, @code{5} and @code{5}.
24312 Allowed range is @code{[0, 255]}.
24315 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
24316 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
24319 Set the vectorscope drawing mode.
24321 Available values are:
24324 Draw dot for each sample.
24327 Draw line between previous and current sample.
24330 Default value is @samp{dot}.
24333 Specify amplitude scale of audio samples.
24335 Available values are:
24351 Swap left channel axis with right channel axis.
24361 Mirror only x axis.
24364 Mirror only y axis.
24372 @subsection Examples
24376 Complete example using @command{ffplay}:
24378 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24379 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24383 @section bench, abench
24385 Benchmark part of a filtergraph.
24387 The filter accepts the following options:
24391 Start or stop a timer.
24393 Available values are:
24396 Get the current time, set it as frame metadata (using the key
24397 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24400 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24401 the input frame metadata to get the time difference. Time difference, average,
24402 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24403 @code{min}) are then printed. The timestamps are expressed in seconds.
24407 @subsection Examples
24411 Benchmark @ref{selectivecolor} filter:
24413 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24419 Concatenate audio and video streams, joining them together one after the
24422 The filter works on segments of synchronized video and audio streams. All
24423 segments must have the same number of streams of each type, and that will
24424 also be the number of streams at output.
24426 The filter accepts the following options:
24431 Set the number of segments. Default is 2.
24434 Set the number of output video streams, that is also the number of video
24435 streams in each segment. Default is 1.
24438 Set the number of output audio streams, that is also the number of audio
24439 streams in each segment. Default is 0.
24442 Activate unsafe mode: do not fail if segments have a different format.
24446 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24447 @var{a} audio outputs.
24449 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24450 segment, in the same order as the outputs, then the inputs for the second
24453 Related streams do not always have exactly the same duration, for various
24454 reasons including codec frame size or sloppy authoring. For that reason,
24455 related synchronized streams (e.g. a video and its audio track) should be
24456 concatenated at once. The concat filter will use the duration of the longest
24457 stream in each segment (except the last one), and if necessary pad shorter
24458 audio streams with silence.
24460 For this filter to work correctly, all segments must start at timestamp 0.
24462 All corresponding streams must have the same parameters in all segments; the
24463 filtering system will automatically select a common pixel format for video
24464 streams, and a common sample format, sample rate and channel layout for
24465 audio streams, but other settings, such as resolution, must be converted
24466 explicitly by the user.
24468 Different frame rates are acceptable but will result in variable frame rate
24469 at output; be sure to configure the output file to handle it.
24471 @subsection Examples
24475 Concatenate an opening, an episode and an ending, all in bilingual version
24476 (video in stream 0, audio in streams 1 and 2):
24478 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24479 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24480 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24481 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24485 Concatenate two parts, handling audio and video separately, using the
24486 (a)movie sources, and adjusting the resolution:
24488 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24489 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24490 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24492 Note that a desync will happen at the stitch if the audio and video streams
24493 do not have exactly the same duration in the first file.
24497 @subsection Commands
24499 This filter supports the following commands:
24502 Close the current segment and step to the next one
24508 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24509 level. By default, it logs a message at a frequency of 10Hz with the
24510 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24511 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24513 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24514 sample format is double-precision floating point. The input stream will be converted to
24515 this specification, if needed. Users may need to insert aformat and/or aresample filters
24516 after this filter to obtain the original parameters.
24518 The filter also has a video output (see the @var{video} option) with a real
24519 time graph to observe the loudness evolution. The graphic contains the logged
24520 message mentioned above, so it is not printed anymore when this option is set,
24521 unless the verbose logging is set. The main graphing area contains the
24522 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24523 the momentary loudness (400 milliseconds), but can optionally be configured
24524 to instead display short-term loudness (see @var{gauge}).
24526 The green area marks a +/- 1LU target range around the target loudness
24527 (-23LUFS by default, unless modified through @var{target}).
24529 More information about the Loudness Recommendation EBU R128 on
24530 @url{http://tech.ebu.ch/loudness}.
24532 The filter accepts the following options:
24537 Activate the video output. The audio stream is passed unchanged whether this
24538 option is set or no. The video stream will be the first output stream if
24539 activated. Default is @code{0}.
24542 Set the video size. This option is for video only. For the syntax of this
24544 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24545 Default and minimum resolution is @code{640x480}.
24548 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24549 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24550 other integer value between this range is allowed.
24553 Set metadata injection. If set to @code{1}, the audio input will be segmented
24554 into 100ms output frames, each of them containing various loudness information
24555 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24557 Default is @code{0}.
24560 Force the frame logging level.
24562 Available values are:
24565 information logging level
24567 verbose logging level
24570 By default, the logging level is set to @var{info}. If the @option{video} or
24571 the @option{metadata} options are set, it switches to @var{verbose}.
24576 Available modes can be cumulated (the option is a @code{flag} type). Possible
24580 Disable any peak mode (default).
24582 Enable sample-peak mode.
24584 Simple peak mode looking for the higher sample value. It logs a message
24585 for sample-peak (identified by @code{SPK}).
24587 Enable true-peak mode.
24589 If enabled, the peak lookup is done on an over-sampled version of the input
24590 stream for better peak accuracy. It logs a message for true-peak.
24591 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24592 This mode requires a build with @code{libswresample}.
24596 Treat mono input files as "dual mono". If a mono file is intended for playback
24597 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24598 If set to @code{true}, this option will compensate for this effect.
24599 Multi-channel input files are not affected by this option.
24602 Set a specific pan law to be used for the measurement of dual mono files.
24603 This parameter is optional, and has a default value of -3.01dB.
24606 Set a specific target level (in LUFS) used as relative zero in the visualization.
24607 This parameter is optional and has a default value of -23LUFS as specified
24608 by EBU R128. However, material published online may prefer a level of -16LUFS
24609 (e.g. for use with podcasts or video platforms).
24612 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24613 @code{shortterm}. By default the momentary value will be used, but in certain
24614 scenarios it may be more useful to observe the short term value instead (e.g.
24618 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24619 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24620 video output, not the summary or continuous log output.
24623 @subsection Examples
24627 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24629 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24633 Run an analysis with @command{ffmpeg}:
24635 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24639 @section interleave, ainterleave
24641 Temporally interleave frames from several inputs.
24643 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24645 These filters read frames from several inputs and send the oldest
24646 queued frame to the output.
24648 Input streams must have well defined, monotonically increasing frame
24651 In order to submit one frame to output, these filters need to enqueue
24652 at least one frame for each input, so they cannot work in case one
24653 input is not yet terminated and will not receive incoming frames.
24655 For example consider the case when one input is a @code{select} filter
24656 which always drops input frames. The @code{interleave} filter will keep
24657 reading from that input, but it will never be able to send new frames
24658 to output until the input sends an end-of-stream signal.
24660 Also, depending on inputs synchronization, the filters will drop
24661 frames in case one input receives more frames than the other ones, and
24662 the queue is already filled.
24664 These filters accept the following options:
24668 Set the number of different inputs, it is 2 by default.
24671 How to determine the end-of-stream.
24675 The duration of the longest input. (default)
24678 The duration of the shortest input.
24681 The duration of the first input.
24686 @subsection Examples
24690 Interleave frames belonging to different streams using @command{ffmpeg}:
24692 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24696 Add flickering blur effect:
24698 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24702 @section metadata, ametadata
24704 Manipulate frame metadata.
24706 This filter accepts the following options:
24710 Set mode of operation of the filter.
24712 Can be one of the following:
24716 If both @code{value} and @code{key} is set, select frames
24717 which have such metadata. If only @code{key} is set, select
24718 every frame that has such key in metadata.
24721 Add new metadata @code{key} and @code{value}. If key is already available
24725 Modify value of already present key.
24728 If @code{value} is set, delete only keys that have such value.
24729 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24733 Print key and its value if metadata was found. If @code{key} is not set print all
24734 metadata values available in frame.
24738 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24741 Set metadata value which will be used. This option is mandatory for
24742 @code{modify} and @code{add} mode.
24745 Which function to use when comparing metadata value and @code{value}.
24747 Can be one of following:
24751 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24754 Values are interpreted as strings, returns true if metadata value starts with
24755 the @code{value} option string.
24758 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24761 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24764 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24767 Values are interpreted as floats, returns true if expression from option @code{expr}
24771 Values are interpreted as strings, returns true if metadata value ends with
24772 the @code{value} option string.
24776 Set expression which is used when @code{function} is set to @code{expr}.
24777 The expression is evaluated through the eval API and can contain the following
24782 Float representation of @code{value} from metadata key.
24785 Float representation of @code{value} as supplied by user in @code{value} option.
24789 If specified in @code{print} mode, output is written to the named file. Instead of
24790 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24791 for standard output. If @code{file} option is not set, output is written to the log
24792 with AV_LOG_INFO loglevel.
24795 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24799 @subsection Examples
24803 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24806 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24809 Print silencedetect output to file @file{metadata.txt}.
24811 silencedetect,ametadata=mode=print:file=metadata.txt
24814 Direct all metadata to a pipe with file descriptor 4.
24816 metadata=mode=print:file='pipe\:4'
24820 @section perms, aperms
24822 Set read/write permissions for the output frames.
24824 These filters are mainly aimed at developers to test direct path in the
24825 following filter in the filtergraph.
24827 The filters accept the following options:
24831 Select the permissions mode.
24833 It accepts the following values:
24836 Do nothing. This is the default.
24838 Set all the output frames read-only.
24840 Set all the output frames directly writable.
24842 Make the frame read-only if writable, and writable if read-only.
24844 Set each output frame read-only or writable randomly.
24848 Set the seed for the @var{random} mode, must be an integer included between
24849 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24850 @code{-1}, the filter will try to use a good random seed on a best effort
24854 Note: in case of auto-inserted filter between the permission filter and the
24855 following one, the permission might not be received as expected in that
24856 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24857 perms/aperms filter can avoid this problem.
24859 @section realtime, arealtime
24861 Slow down filtering to match real time approximately.
24863 These filters will pause the filtering for a variable amount of time to
24864 match the output rate with the input timestamps.
24865 They are similar to the @option{re} option to @code{ffmpeg}.
24867 They accept the following options:
24871 Time limit for the pauses. Any pause longer than that will be considered
24872 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24874 Speed factor for processing. The value must be a float larger than zero.
24875 Values larger than 1.0 will result in faster than realtime processing,
24876 smaller will slow processing down. The @var{limit} is automatically adapted
24877 accordingly. Default is 1.0.
24879 A processing speed faster than what is possible without these filters cannot
24884 @section select, aselect
24886 Select frames to pass in output.
24888 This filter accepts the following options:
24893 Set expression, which is evaluated for each input frame.
24895 If the expression is evaluated to zero, the frame is discarded.
24897 If the evaluation result is negative or NaN, the frame is sent to the
24898 first output; otherwise it is sent to the output with index
24899 @code{ceil(val)-1}, assuming that the input index starts from 0.
24901 For example a value of @code{1.2} corresponds to the output with index
24902 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24905 Set the number of outputs. The output to which to send the selected
24906 frame is based on the result of the evaluation. Default value is 1.
24909 The expression can contain the following constants:
24913 The (sequential) number of the filtered frame, starting from 0.
24916 The (sequential) number of the selected frame, starting from 0.
24918 @item prev_selected_n
24919 The sequential number of the last selected frame. It's NAN if undefined.
24922 The timebase of the input timestamps.
24925 The PTS (Presentation TimeStamp) of the filtered video frame,
24926 expressed in @var{TB} units. It's NAN if undefined.
24929 The PTS of the filtered video frame,
24930 expressed in seconds. It's NAN if undefined.
24933 The PTS of the previously filtered video frame. It's NAN if undefined.
24935 @item prev_selected_pts
24936 The PTS of the last previously filtered video frame. It's NAN if undefined.
24938 @item prev_selected_t
24939 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24942 The PTS of the first video frame in the video. It's NAN if undefined.
24945 The time of the first video frame in the video. It's NAN if undefined.
24947 @item pict_type @emph{(video only)}
24948 The type of the filtered frame. It can assume one of the following
24960 @item interlace_type @emph{(video only)}
24961 The frame interlace type. It can assume one of the following values:
24964 The frame is progressive (not interlaced).
24966 The frame is top-field-first.
24968 The frame is bottom-field-first.
24971 @item consumed_sample_n @emph{(audio only)}
24972 the number of selected samples before the current frame
24974 @item samples_n @emph{(audio only)}
24975 the number of samples in the current frame
24977 @item sample_rate @emph{(audio only)}
24978 the input sample rate
24981 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24984 the position in the file of the filtered frame, -1 if the information
24985 is not available (e.g. for synthetic video)
24987 @item scene @emph{(video only)}
24988 value between 0 and 1 to indicate a new scene; a low value reflects a low
24989 probability for the current frame to introduce a new scene, while a higher
24990 value means the current frame is more likely to be one (see the example below)
24992 @item concatdec_select
24993 The concat demuxer can select only part of a concat input file by setting an
24994 inpoint and an outpoint, but the output packets may not be entirely contained
24995 in the selected interval. By using this variable, it is possible to skip frames
24996 generated by the concat demuxer which are not exactly contained in the selected
24999 This works by comparing the frame pts against the @var{lavf.concat.start_time}
25000 and the @var{lavf.concat.duration} packet metadata values which are also
25001 present in the decoded frames.
25003 The @var{concatdec_select} variable is -1 if the frame pts is at least
25004 start_time and either the duration metadata is missing or the frame pts is less
25005 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
25008 That basically means that an input frame is selected if its pts is within the
25009 interval set by the concat demuxer.
25013 The default value of the select expression is "1".
25015 @subsection Examples
25019 Select all frames in input:
25024 The example above is the same as:
25036 Select only I-frames:
25038 select='eq(pict_type\,I)'
25042 Select one frame every 100:
25044 select='not(mod(n\,100))'
25048 Select only frames contained in the 10-20 time interval:
25050 select=between(t\,10\,20)
25054 Select only I-frames contained in the 10-20 time interval:
25056 select=between(t\,10\,20)*eq(pict_type\,I)
25060 Select frames with a minimum distance of 10 seconds:
25062 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
25066 Use aselect to select only audio frames with samples number > 100:
25068 aselect='gt(samples_n\,100)'
25072 Create a mosaic of the first scenes:
25074 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
25077 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
25081 Send even and odd frames to separate outputs, and compose them:
25083 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
25087 Select useful frames from an ffconcat file which is using inpoints and
25088 outpoints but where the source files are not intra frame only.
25090 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
25094 @section sendcmd, asendcmd
25096 Send commands to filters in the filtergraph.
25098 These filters read commands to be sent to other filters in the
25101 @code{sendcmd} must be inserted between two video filters,
25102 @code{asendcmd} must be inserted between two audio filters, but apart
25103 from that they act the same way.
25105 The specification of commands can be provided in the filter arguments
25106 with the @var{commands} option, or in a file specified by the
25107 @var{filename} option.
25109 These filters accept the following options:
25112 Set the commands to be read and sent to the other filters.
25114 Set the filename of the commands to be read and sent to the other
25118 @subsection Commands syntax
25120 A commands description consists of a sequence of interval
25121 specifications, comprising a list of commands to be executed when a
25122 particular event related to that interval occurs. The occurring event
25123 is typically the current frame time entering or leaving a given time
25126 An interval is specified by the following syntax:
25128 @var{START}[-@var{END}] @var{COMMANDS};
25131 The time interval is specified by the @var{START} and @var{END} times.
25132 @var{END} is optional and defaults to the maximum time.
25134 The current frame time is considered within the specified interval if
25135 it is included in the interval [@var{START}, @var{END}), that is when
25136 the time is greater or equal to @var{START} and is lesser than
25139 @var{COMMANDS} consists of a sequence of one or more command
25140 specifications, separated by ",", relating to that interval. The
25141 syntax of a command specification is given by:
25143 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
25146 @var{FLAGS} is optional and specifies the type of events relating to
25147 the time interval which enable sending the specified command, and must
25148 be a non-null sequence of identifier flags separated by "+" or "|" and
25149 enclosed between "[" and "]".
25151 The following flags are recognized:
25154 The command is sent when the current frame timestamp enters the
25155 specified interval. In other words, the command is sent when the
25156 previous frame timestamp was not in the given interval, and the
25160 The command is sent when the current frame timestamp leaves the
25161 specified interval. In other words, the command is sent when the
25162 previous frame timestamp was in the given interval, and the
25166 The command @var{ARG} is interpreted as expression and result of
25167 expression is passed as @var{ARG}.
25169 The expression is evaluated through the eval API and can contain the following
25174 Original position in the file of the frame, or undefined if undefined
25175 for the current frame.
25178 The presentation timestamp in input.
25181 The count of the input frame for video or audio, starting from 0.
25184 The time in seconds of the current frame.
25187 The start time in seconds of the current command interval.
25190 The end time in seconds of the current command interval.
25193 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
25198 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
25201 @var{TARGET} specifies the target of the command, usually the name of
25202 the filter class or a specific filter instance name.
25204 @var{COMMAND} specifies the name of the command for the target filter.
25206 @var{ARG} is optional and specifies the optional list of argument for
25207 the given @var{COMMAND}.
25209 Between one interval specification and another, whitespaces, or
25210 sequences of characters starting with @code{#} until the end of line,
25211 are ignored and can be used to annotate comments.
25213 A simplified BNF description of the commands specification syntax
25216 @var{COMMAND_FLAG} ::= "enter" | "leave"
25217 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
25218 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
25219 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
25220 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
25221 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
25224 @subsection Examples
25228 Specify audio tempo change at second 4:
25230 asendcmd=c='4.0 atempo tempo 1.5',atempo
25234 Target a specific filter instance:
25236 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
25240 Specify a list of drawtext and hue commands in a file.
25242 # show text in the interval 5-10
25243 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
25244 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
25246 # desaturate the image in the interval 15-20
25247 15.0-20.0 [enter] hue s 0,
25248 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
25250 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
25252 # apply an exponential saturation fade-out effect, starting from time 25
25253 25 [enter] hue s exp(25-t)
25256 A filtergraph allowing to read and process the above command list
25257 stored in a file @file{test.cmd}, can be specified with:
25259 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
25264 @section setpts, asetpts
25266 Change the PTS (presentation timestamp) of the input frames.
25268 @code{setpts} works on video frames, @code{asetpts} on audio frames.
25270 This filter accepts the following options:
25275 The expression which is evaluated for each frame to construct its timestamp.
25279 The expression is evaluated through the eval API and can contain the following
25283 @item FRAME_RATE, FR
25284 frame rate, only defined for constant frame-rate video
25287 The presentation timestamp in input
25290 The count of the input frame for video or the number of consumed samples,
25291 not including the current frame for audio, starting from 0.
25293 @item NB_CONSUMED_SAMPLES
25294 The number of consumed samples, not including the current frame (only
25297 @item NB_SAMPLES, S
25298 The number of samples in the current frame (only audio)
25300 @item SAMPLE_RATE, SR
25301 The audio sample rate.
25304 The PTS of the first frame.
25307 the time in seconds of the first frame
25310 State whether the current frame is interlaced.
25313 the time in seconds of the current frame
25316 original position in the file of the frame, or undefined if undefined
25317 for the current frame
25320 The previous input PTS.
25323 previous input time in seconds
25326 The previous output PTS.
25329 previous output time in seconds
25332 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25336 The wallclock (RTC) time at the start of the movie in microseconds.
25339 The timebase of the input timestamps.
25343 @subsection Examples
25347 Start counting PTS from zero
25349 setpts=PTS-STARTPTS
25353 Apply fast motion effect:
25359 Apply slow motion effect:
25365 Set fixed rate of 25 frames per second:
25371 Set fixed rate 25 fps with some jitter:
25373 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25377 Apply an offset of 10 seconds to the input PTS:
25383 Generate timestamps from a "live source" and rebase onto the current timebase:
25385 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25389 Generate timestamps by counting samples:
25398 Force color range for the output video frame.
25400 The @code{setrange} filter marks the color range property for the
25401 output frames. It does not change the input frame, but only sets the
25402 corresponding property, which affects how the frame is treated by
25405 The filter accepts the following options:
25410 Available values are:
25414 Keep the same color range property.
25416 @item unspecified, unknown
25417 Set the color range as unspecified.
25419 @item limited, tv, mpeg
25420 Set the color range as limited.
25422 @item full, pc, jpeg
25423 Set the color range as full.
25427 @section settb, asettb
25429 Set the timebase to use for the output frames timestamps.
25430 It is mainly useful for testing timebase configuration.
25432 It accepts the following parameters:
25437 The expression which is evaluated into the output timebase.
25441 The value for @option{tb} is an arithmetic expression representing a
25442 rational. The expression can contain the constants "AVTB" (the default
25443 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25444 audio only). Default value is "intb".
25446 @subsection Examples
25450 Set the timebase to 1/25:
25456 Set the timebase to 1/10:
25462 Set the timebase to 1001/1000:
25468 Set the timebase to 2*intb:
25474 Set the default timebase value:
25481 Convert input audio to a video output representing frequency spectrum
25482 logarithmically using Brown-Puckette constant Q transform algorithm with
25483 direct frequency domain coefficient calculation (but the transform itself
25484 is not really constant Q, instead the Q factor is actually variable/clamped),
25485 with musical tone scale, from E0 to D#10.
25487 The filter accepts the following options:
25491 Specify the video size for the output. It must be even. For the syntax of this option,
25492 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25493 Default value is @code{1920x1080}.
25496 Set the output frame rate. Default value is @code{25}.
25499 Set the bargraph height. It must be even. Default value is @code{-1} which
25500 computes the bargraph height automatically.
25503 Set the axis height. It must be even. Default value is @code{-1} which computes
25504 the axis height automatically.
25507 Set the sonogram height. It must be even. Default value is @code{-1} which
25508 computes the sonogram height automatically.
25511 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25512 instead. Default value is @code{1}.
25514 @item sono_v, volume
25515 Specify the sonogram volume expression. It can contain variables:
25518 the @var{bar_v} evaluated expression
25519 @item frequency, freq, f
25520 the frequency where it is evaluated
25521 @item timeclamp, tc
25522 the value of @var{timeclamp} option
25526 @item a_weighting(f)
25527 A-weighting of equal loudness
25528 @item b_weighting(f)
25529 B-weighting of equal loudness
25530 @item c_weighting(f)
25531 C-weighting of equal loudness.
25533 Default value is @code{16}.
25535 @item bar_v, volume2
25536 Specify the bargraph volume expression. It can contain variables:
25539 the @var{sono_v} evaluated expression
25540 @item frequency, freq, f
25541 the frequency where it is evaluated
25542 @item timeclamp, tc
25543 the value of @var{timeclamp} option
25547 @item a_weighting(f)
25548 A-weighting of equal loudness
25549 @item b_weighting(f)
25550 B-weighting of equal loudness
25551 @item c_weighting(f)
25552 C-weighting of equal loudness.
25554 Default value is @code{sono_v}.
25556 @item sono_g, gamma
25557 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25558 higher gamma makes the spectrum having more range. Default value is @code{3}.
25559 Acceptable range is @code{[1, 7]}.
25561 @item bar_g, gamma2
25562 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25566 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25567 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25569 @item timeclamp, tc
25570 Specify the transform timeclamp. At low frequency, there is trade-off between
25571 accuracy in time domain and frequency domain. If timeclamp is lower,
25572 event in time domain is represented more accurately (such as fast bass drum),
25573 otherwise event in frequency domain is represented more accurately
25574 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25577 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25578 limits future samples by applying asymmetric windowing in time domain, useful
25579 when low latency is required. Accepted range is @code{[0, 1]}.
25582 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25583 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25586 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25587 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25590 This option is deprecated and ignored.
25593 Specify the transform length in time domain. Use this option to control accuracy
25594 trade-off between time domain and frequency domain at every frequency sample.
25595 It can contain variables:
25597 @item frequency, freq, f
25598 the frequency where it is evaluated
25599 @item timeclamp, tc
25600 the value of @var{timeclamp} option.
25602 Default value is @code{384*tc/(384+tc*f)}.
25605 Specify the transform count for every video frame. Default value is @code{6}.
25606 Acceptable range is @code{[1, 30]}.
25609 Specify the transform count for every single pixel. Default value is @code{0},
25610 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25613 Specify font file for use with freetype to draw the axis. If not specified,
25614 use embedded font. Note that drawing with font file or embedded font is not
25615 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25619 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25620 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25624 Specify font color expression. This is arithmetic expression that should return
25625 integer value 0xRRGGBB. It can contain variables:
25627 @item frequency, freq, f
25628 the frequency where it is evaluated
25629 @item timeclamp, tc
25630 the value of @var{timeclamp} option
25635 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25636 @item r(x), g(x), b(x)
25637 red, green, and blue value of intensity x.
25639 Default value is @code{st(0, (midi(f)-59.5)/12);
25640 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25641 r(1-ld(1)) + b(ld(1))}.
25644 Specify image file to draw the axis. This option override @var{fontfile} and
25645 @var{fontcolor} option.
25648 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25649 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25650 Default value is @code{1}.
25653 Set colorspace. The accepted values are:
25656 Unspecified (default)
25665 BT.470BG or BT.601-6 625
25668 SMPTE-170M or BT.601-6 525
25674 BT.2020 with non-constant luminance
25679 Set spectrogram color scheme. This is list of floating point values with format
25680 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25681 The default is @code{1|0.5|0|0|0.5|1}.
25685 @subsection Examples
25689 Playing audio while showing the spectrum:
25691 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25695 Same as above, but with frame rate 30 fps:
25697 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25701 Playing at 1280x720:
25703 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25707 Disable sonogram display:
25713 A1 and its harmonics: A1, A2, (near)E3, A3:
25715 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),
25716 asplit[a][out1]; [a] showcqt [out0]'
25720 Same as above, but with more accuracy in frequency domain:
25722 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),
25723 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25729 bar_v=10:sono_v=bar_v*a_weighting(f)
25733 Custom gamma, now spectrum is linear to the amplitude.
25739 Custom tlength equation:
25741 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)))'
25745 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25747 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25751 Custom font using fontconfig:
25753 font='Courier New,Monospace,mono|bold'
25757 Custom frequency range with custom axis using image file:
25759 axisfile=myaxis.png:basefreq=40:endfreq=10000
25765 Convert input audio to video output representing the audio power spectrum.
25766 Audio amplitude is on Y-axis while frequency is on X-axis.
25768 The filter accepts the following options:
25772 Specify size of video. For the syntax of this option, check the
25773 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25774 Default is @code{1024x512}.
25778 This set how each frequency bin will be represented.
25780 It accepts the following values:
25786 Default is @code{bar}.
25789 Set amplitude scale.
25791 It accepts the following values:
25805 Default is @code{log}.
25808 Set frequency scale.
25810 It accepts the following values:
25819 Reverse logarithmic scale.
25821 Default is @code{lin}.
25824 Set window size. Allowed range is from 16 to 65536.
25826 Default is @code{2048}
25829 Set windowing function.
25831 It accepts the following values:
25854 Default is @code{hanning}.
25857 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25858 which means optimal overlap for selected window function will be picked.
25861 Set time averaging. Setting this to 0 will display current maximal peaks.
25862 Default is @code{1}, which means time averaging is disabled.
25865 Specify list of colors separated by space or by '|' which will be used to
25866 draw channel frequencies. Unrecognized or missing colors will be replaced
25870 Set channel display mode.
25872 It accepts the following values:
25877 Default is @code{combined}.
25880 Set minimum amplitude used in @code{log} amplitude scaler.
25883 Set data display mode.
25885 It accepts the following values:
25891 Default is @code{magnitude}.
25894 @section showspatial
25896 Convert stereo input audio to a video output, representing the spatial relationship
25897 between two channels.
25899 The filter accepts the following options:
25903 Specify the video size for the output. For the syntax of this option, check the
25904 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25905 Default value is @code{512x512}.
25908 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25911 Set window function.
25913 It accepts the following values:
25938 Default value is @code{hann}.
25941 Set ratio of overlap window. Default value is @code{0.5}.
25942 When value is @code{1} overlap is set to recommended size for specific
25943 window function currently used.
25946 @anchor{showspectrum}
25947 @section showspectrum
25949 Convert input audio to a video output, representing the audio frequency
25952 The filter accepts the following options:
25956 Specify the video size for the output. For the syntax of this option, check the
25957 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25958 Default value is @code{640x512}.
25961 Specify how the spectrum should slide along the window.
25963 It accepts the following values:
25966 the samples start again on the left when they reach the right
25968 the samples scroll from right to left
25970 frames are only produced when the samples reach the right
25972 the samples scroll from left to right
25975 Default value is @code{replace}.
25978 Specify display mode.
25980 It accepts the following values:
25983 all channels are displayed in the same row
25985 all channels are displayed in separate rows
25988 Default value is @samp{combined}.
25991 Specify display color mode.
25993 It accepts the following values:
25996 each channel is displayed in a separate color
25998 each channel is displayed using the same color scheme
26000 each channel is displayed using the rainbow color scheme
26002 each channel is displayed using the moreland color scheme
26004 each channel is displayed using the nebulae color scheme
26006 each channel is displayed using the fire color scheme
26008 each channel is displayed using the fiery color scheme
26010 each channel is displayed using the fruit color scheme
26012 each channel is displayed using the cool color scheme
26014 each channel is displayed using the magma color scheme
26016 each channel is displayed using the green color scheme
26018 each channel is displayed using the viridis color scheme
26020 each channel is displayed using the plasma color scheme
26022 each channel is displayed using the cividis color scheme
26024 each channel is displayed using the terrain color scheme
26027 Default value is @samp{channel}.
26030 Specify scale used for calculating intensity color values.
26032 It accepts the following values:
26037 square root, default
26048 Default value is @samp{sqrt}.
26051 Specify frequency scale.
26053 It accepts the following values:
26061 Default value is @samp{lin}.
26064 Set saturation modifier for displayed colors. Negative values provide
26065 alternative color scheme. @code{0} is no saturation at all.
26066 Saturation must be in [-10.0, 10.0] range.
26067 Default value is @code{1}.
26070 Set window function.
26072 It accepts the following values:
26097 Default value is @code{hann}.
26100 Set orientation of time vs frequency axis. Can be @code{vertical} or
26101 @code{horizontal}. Default is @code{vertical}.
26104 Set ratio of overlap window. Default value is @code{0}.
26105 When value is @code{1} overlap is set to recommended size for specific
26106 window function currently used.
26109 Set scale gain for calculating intensity color values.
26110 Default value is @code{1}.
26113 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
26116 Set color rotation, must be in [-1.0, 1.0] range.
26117 Default value is @code{0}.
26120 Set start frequency from which to display spectrogram. Default is @code{0}.
26123 Set stop frequency to which to display spectrogram. Default is @code{0}.
26126 Set upper frame rate limit. Default is @code{auto}, unlimited.
26129 Draw time and frequency axes and legends. Default is disabled.
26132 The usage is very similar to the showwaves filter; see the examples in that
26135 @subsection Examples
26139 Large window with logarithmic color scaling:
26141 showspectrum=s=1280x480:scale=log
26145 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
26147 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
26148 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
26152 @section showspectrumpic
26154 Convert input audio to a single video frame, representing the audio frequency
26157 The filter accepts the following options:
26161 Specify the video size for the output. For the syntax of this option, check the
26162 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26163 Default value is @code{4096x2048}.
26166 Specify display mode.
26168 It accepts the following values:
26171 all channels are displayed in the same row
26173 all channels are displayed in separate rows
26175 Default value is @samp{combined}.
26178 Specify display color mode.
26180 It accepts the following values:
26183 each channel is displayed in a separate color
26185 each channel is displayed using the same color scheme
26187 each channel is displayed using the rainbow color scheme
26189 each channel is displayed using the moreland color scheme
26191 each channel is displayed using the nebulae color scheme
26193 each channel is displayed using the fire color scheme
26195 each channel is displayed using the fiery color scheme
26197 each channel is displayed using the fruit color scheme
26199 each channel is displayed using the cool color scheme
26201 each channel is displayed using the magma color scheme
26203 each channel is displayed using the green color scheme
26205 each channel is displayed using the viridis color scheme
26207 each channel is displayed using the plasma color scheme
26209 each channel is displayed using the cividis color scheme
26211 each channel is displayed using the terrain color scheme
26213 Default value is @samp{intensity}.
26216 Specify scale used for calculating intensity color values.
26218 It accepts the following values:
26223 square root, default
26233 Default value is @samp{log}.
26236 Specify frequency scale.
26238 It accepts the following values:
26246 Default value is @samp{lin}.
26249 Set saturation modifier for displayed colors. Negative values provide
26250 alternative color scheme. @code{0} is no saturation at all.
26251 Saturation must be in [-10.0, 10.0] range.
26252 Default value is @code{1}.
26255 Set window function.
26257 It accepts the following values:
26281 Default value is @code{hann}.
26284 Set orientation of time vs frequency axis. Can be @code{vertical} or
26285 @code{horizontal}. Default is @code{vertical}.
26288 Set scale gain for calculating intensity color values.
26289 Default value is @code{1}.
26292 Draw time and frequency axes and legends. Default is enabled.
26295 Set color rotation, must be in [-1.0, 1.0] range.
26296 Default value is @code{0}.
26299 Set start frequency from which to display spectrogram. Default is @code{0}.
26302 Set stop frequency to which to display spectrogram. Default is @code{0}.
26305 @subsection Examples
26309 Extract an audio spectrogram of a whole audio track
26310 in a 1024x1024 picture using @command{ffmpeg}:
26312 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
26316 @section showvolume
26318 Convert input audio volume to a video output.
26320 The filter accepts the following options:
26327 Set border width, allowed range is [0, 5]. Default is 1.
26330 Set channel width, allowed range is [80, 8192]. Default is 400.
26333 Set channel height, allowed range is [1, 900]. Default is 20.
26336 Set fade, allowed range is [0, 1]. Default is 0.95.
26339 Set volume color expression.
26341 The expression can use the following variables:
26345 Current max volume of channel in dB.
26351 Current channel number, starting from 0.
26355 If set, displays channel names. Default is enabled.
26358 If set, displays volume values. Default is enabled.
26361 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26362 default is @code{h}.
26365 Set step size, allowed range is [0, 5]. Default is 0, which means
26369 Set background opacity, allowed range is [0, 1]. Default is 0.
26372 Set metering mode, can be peak: @code{p} or rms: @code{r},
26373 default is @code{p}.
26376 Set display scale, can be linear: @code{lin} or log: @code{log},
26377 default is @code{lin}.
26381 If set to > 0., display a line for the max level
26382 in the previous seconds.
26383 default is disabled: @code{0.}
26386 The color of the max line. Use when @code{dm} option is set to > 0.
26387 default is: @code{orange}
26392 Convert input audio to a video output, representing the samples waves.
26394 The filter accepts the following options:
26398 Specify the video size for the output. For the syntax of this option, check the
26399 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26400 Default value is @code{600x240}.
26405 Available values are:
26408 Draw a point for each sample.
26411 Draw a vertical line for each sample.
26414 Draw a point for each sample and a line between them.
26417 Draw a centered vertical line for each sample.
26420 Default value is @code{point}.
26423 Set the number of samples which are printed on the same column. A
26424 larger value will decrease the frame rate. Must be a positive
26425 integer. This option can be set only if the value for @var{rate}
26426 is not explicitly specified.
26429 Set the (approximate) output frame rate. This is done by setting the
26430 option @var{n}. Default value is "25".
26432 @item split_channels
26433 Set if channels should be drawn separately or overlap. Default value is 0.
26436 Set colors separated by '|' which are going to be used for drawing of each channel.
26439 Set amplitude scale.
26441 Available values are:
26459 Set the draw mode. This is mostly useful to set for high @var{n}.
26461 Available values are:
26464 Scale pixel values for each drawn sample.
26467 Draw every sample directly.
26470 Default value is @code{scale}.
26473 @subsection Examples
26477 Output the input file audio and the corresponding video representation
26480 amovie=a.mp3,asplit[out0],showwaves[out1]
26484 Create a synthetic signal and show it with showwaves, forcing a
26485 frame rate of 30 frames per second:
26487 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26491 @section showwavespic
26493 Convert input audio to a single video frame, representing the samples waves.
26495 The filter accepts the following options:
26499 Specify the video size for the output. For the syntax of this option, check the
26500 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26501 Default value is @code{600x240}.
26503 @item split_channels
26504 Set if channels should be drawn separately or overlap. Default value is 0.
26507 Set colors separated by '|' which are going to be used for drawing of each channel.
26510 Set amplitude scale.
26512 Available values are:
26532 Available values are:
26535 Scale pixel values for each drawn sample.
26538 Draw every sample directly.
26541 Default value is @code{scale}.
26544 Set the filter mode.
26546 Available values are:
26549 Use average samples values for each drawn sample.
26552 Use peak samples values for each drawn sample.
26555 Default value is @code{average}.
26558 @subsection Examples
26562 Extract a channel split representation of the wave form of a whole audio track
26563 in a 1024x800 picture using @command{ffmpeg}:
26565 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26569 @section sidedata, asidedata
26571 Delete frame side data, or select frames based on it.
26573 This filter accepts the following options:
26577 Set mode of operation of the filter.
26579 Can be one of the following:
26583 Select every frame with side data of @code{type}.
26586 Delete side data of @code{type}. If @code{type} is not set, delete all side
26592 Set side data type used with all modes. Must be set for @code{select} mode. For
26593 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26594 in @file{libavutil/frame.h}. For example, to choose
26595 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26599 @section spectrumsynth
26601 Synthesize audio from 2 input video spectrums, first input stream represents
26602 magnitude across time and second represents phase across time.
26603 The filter will transform from frequency domain as displayed in videos back
26604 to time domain as presented in audio output.
26606 This filter is primarily created for reversing processed @ref{showspectrum}
26607 filter outputs, but can synthesize sound from other spectrograms too.
26608 But in such case results are going to be poor if the phase data is not
26609 available, because in such cases phase data need to be recreated, usually
26610 it's just recreated from random noise.
26611 For best results use gray only output (@code{channel} color mode in
26612 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26613 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26614 @code{data} option. Inputs videos should generally use @code{fullframe}
26615 slide mode as that saves resources needed for decoding video.
26617 The filter accepts the following options:
26621 Specify sample rate of output audio, the sample rate of audio from which
26622 spectrum was generated may differ.
26625 Set number of channels represented in input video spectrums.
26628 Set scale which was used when generating magnitude input spectrum.
26629 Can be @code{lin} or @code{log}. Default is @code{log}.
26632 Set slide which was used when generating inputs spectrums.
26633 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26634 Default is @code{fullframe}.
26637 Set window function used for resynthesis.
26640 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26641 which means optimal overlap for selected window function will be picked.
26644 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26645 Default is @code{vertical}.
26648 @subsection Examples
26652 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26653 then resynthesize videos back to audio with spectrumsynth:
26655 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
26656 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
26657 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26661 @section split, asplit
26663 Split input into several identical outputs.
26665 @code{asplit} works with audio input, @code{split} with video.
26667 The filter accepts a single parameter which specifies the number of outputs. If
26668 unspecified, it defaults to 2.
26670 @subsection Examples
26674 Create two separate outputs from the same input:
26676 [in] split [out0][out1]
26680 To create 3 or more outputs, you need to specify the number of
26683 [in] asplit=3 [out0][out1][out2]
26687 Create two separate outputs from the same input, one cropped and
26690 [in] split [splitout1][splitout2];
26691 [splitout1] crop=100:100:0:0 [cropout];
26692 [splitout2] pad=200:200:100:100 [padout];
26696 Create 5 copies of the input audio with @command{ffmpeg}:
26698 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26704 Receive commands sent through a libzmq client, and forward them to
26705 filters in the filtergraph.
26707 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26708 must be inserted between two video filters, @code{azmq} between two
26709 audio filters. Both are capable to send messages to any filter type.
26711 To enable these filters you need to install the libzmq library and
26712 headers and configure FFmpeg with @code{--enable-libzmq}.
26714 For more information about libzmq see:
26715 @url{http://www.zeromq.org/}
26717 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26718 receives messages sent through a network interface defined by the
26719 @option{bind_address} (or the abbreviation "@option{b}") option.
26720 Default value of this option is @file{tcp://localhost:5555}. You may
26721 want to alter this value to your needs, but do not forget to escape any
26722 ':' signs (see @ref{filtergraph escaping}).
26724 The received message must be in the form:
26726 @var{TARGET} @var{COMMAND} [@var{ARG}]
26729 @var{TARGET} specifies the target of the command, usually the name of
26730 the filter class or a specific filter instance name. The default
26731 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26732 but you can override this by using the @samp{filter_name@@id} syntax
26733 (see @ref{Filtergraph syntax}).
26735 @var{COMMAND} specifies the name of the command for the target filter.
26737 @var{ARG} is optional and specifies the optional argument list for the
26738 given @var{COMMAND}.
26740 Upon reception, the message is processed and the corresponding command
26741 is injected into the filtergraph. Depending on the result, the filter
26742 will send a reply to the client, adopting the format:
26744 @var{ERROR_CODE} @var{ERROR_REASON}
26748 @var{MESSAGE} is optional.
26750 @subsection Examples
26752 Look at @file{tools/zmqsend} for an example of a zmq client which can
26753 be used to send commands processed by these filters.
26755 Consider the following filtergraph generated by @command{ffplay}.
26756 In this example the last overlay filter has an instance name. All other
26757 filters will have default instance names.
26760 ffplay -dumpgraph 1 -f lavfi "
26761 color=s=100x100:c=red [l];
26762 color=s=100x100:c=blue [r];
26763 nullsrc=s=200x100, zmq [bg];
26764 [bg][l] overlay [bg+l];
26765 [bg+l][r] overlay@@my=x=100 "
26768 To change the color of the left side of the video, the following
26769 command can be used:
26771 echo Parsed_color_0 c yellow | tools/zmqsend
26774 To change the right side:
26776 echo Parsed_color_1 c pink | tools/zmqsend
26779 To change the position of the right side:
26781 echo overlay@@my x 150 | tools/zmqsend
26785 @c man end MULTIMEDIA FILTERS
26787 @chapter Multimedia Sources
26788 @c man begin MULTIMEDIA SOURCES
26790 Below is a description of the currently available multimedia sources.
26794 This is the same as @ref{movie} source, except it selects an audio
26800 Read audio and/or video stream(s) from a movie container.
26802 It accepts the following parameters:
26806 The name of the resource to read (not necessarily a file; it can also be a
26807 device or a stream accessed through some protocol).
26809 @item format_name, f
26810 Specifies the format assumed for the movie to read, and can be either
26811 the name of a container or an input device. If not specified, the
26812 format is guessed from @var{movie_name} or by probing.
26814 @item seek_point, sp
26815 Specifies the seek point in seconds. The frames will be output
26816 starting from this seek point. The parameter is evaluated with
26817 @code{av_strtod}, so the numerical value may be suffixed by an IS
26818 postfix. The default value is "0".
26821 Specifies the streams to read. Several streams can be specified,
26822 separated by "+". The source will then have as many outputs, in the
26823 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26824 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26825 respectively the default (best suited) video and audio stream. Default
26826 is "dv", or "da" if the filter is called as "amovie".
26828 @item stream_index, si
26829 Specifies the index of the video stream to read. If the value is -1,
26830 the most suitable video stream will be automatically selected. The default
26831 value is "-1". Deprecated. If the filter is called "amovie", it will select
26832 audio instead of video.
26835 Specifies how many times to read the stream in sequence.
26836 If the value is 0, the stream will be looped infinitely.
26837 Default value is "1".
26839 Note that when the movie is looped the source timestamps are not
26840 changed, so it will generate non monotonically increasing timestamps.
26842 @item discontinuity
26843 Specifies the time difference between frames above which the point is
26844 considered a timestamp discontinuity which is removed by adjusting the later
26848 It allows overlaying a second video on top of the main input of
26849 a filtergraph, as shown in this graph:
26851 input -----------> deltapts0 --> overlay --> output
26854 movie --> scale--> deltapts1 -------+
26856 @subsection Examples
26860 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26861 on top of the input labelled "in":
26863 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26864 [in] setpts=PTS-STARTPTS [main];
26865 [main][over] overlay=16:16 [out]
26869 Read from a video4linux2 device, and overlay it on top of the input
26872 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26873 [in] setpts=PTS-STARTPTS [main];
26874 [main][over] overlay=16:16 [out]
26878 Read the first video stream and the audio stream with id 0x81 from
26879 dvd.vob; the video is connected to the pad named "video" and the audio is
26880 connected to the pad named "audio":
26882 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26886 @subsection Commands
26888 Both movie and amovie support the following commands:
26891 Perform seek using "av_seek_frame".
26892 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26895 @var{stream_index}: If stream_index is -1, a default
26896 stream is selected, and @var{timestamp} is automatically converted
26897 from AV_TIME_BASE units to the stream specific time_base.
26899 @var{timestamp}: Timestamp in AVStream.time_base units
26900 or, if no stream is specified, in AV_TIME_BASE units.
26902 @var{flags}: Flags which select direction and seeking mode.
26906 Get movie duration in AV_TIME_BASE units.
26910 @c man end MULTIMEDIA SOURCES