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, can be @var{2nd}, @var{4th} or @var{8th}.
527 Default is @var{4th}.
532 Reduce audio bit resolution.
534 This filter is bit crusher with enhanced functionality. A bit crusher
535 is used to audibly reduce number of bits an audio signal is sampled
536 with. This doesn't change the bit depth at all, it just produces the
537 effect. Material reduced in bit depth sounds more harsh and "digital".
538 This filter is able to even round to continuous values instead of discrete
540 Additionally it has a D/C offset which results in different crushing of
541 the lower and the upper half of the signal.
542 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
544 Another feature of this filter is the logarithmic mode.
545 This setting switches from linear distances between bits to logarithmic ones.
546 The result is a much more "natural" sounding crusher which doesn't gate low
547 signals for example. The human ear has a logarithmic perception,
548 so this kind of crushing is much more pleasant.
549 Logarithmic crushing is also able to get anti-aliased.
551 The filter accepts the following options:
567 Can be linear: @code{lin} or logarithmic: @code{log}.
576 Set sample reduction.
579 Enable LFO. By default disabled.
590 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
594 Remove impulsive noise from input audio.
596 Samples detected as impulsive noise are replaced by interpolated samples using
597 autoregressive modelling.
601 Set window size, in milliseconds. Allowed range is from @code{10} to
602 @code{100}. Default value is @code{55} milliseconds.
603 This sets size of window which will be processed at once.
606 Set window overlap, in percentage of window size. Allowed range is from
607 @code{50} to @code{95}. Default value is @code{75} percent.
608 Setting this to a very high value increases impulsive noise removal but makes
609 whole process much slower.
612 Set autoregression order, in percentage of window size. Allowed range is from
613 @code{0} to @code{25}. Default value is @code{2} percent. This option also
614 controls quality of interpolated samples using neighbour good samples.
617 Set threshold value. Allowed range is from @code{1} to @code{100}.
618 Default value is @code{2}.
619 This controls the strength of impulsive noise which is going to be removed.
620 The lower value, the more samples will be detected as impulsive noise.
623 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
624 @code{10}. Default value is @code{2}.
625 If any two samples detected as noise are spaced less than this value then any
626 sample between those two samples will be also detected as noise.
631 It accepts the following values:
634 Select overlap-add method. Even not interpolated samples are slightly
635 changed with this method.
638 Select overlap-save method. Not interpolated samples remain unchanged.
641 Default value is @code{a}.
645 Remove clipped samples from input audio.
647 Samples detected as clipped are replaced by interpolated samples using
648 autoregressive modelling.
652 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
653 Default value is @code{55} milliseconds.
654 This sets size of window which will be processed at once.
657 Set window overlap, in percentage of window size. Allowed range is from @code{50}
658 to @code{95}. Default value is @code{75} percent.
661 Set autoregression order, in percentage of window size. Allowed range is from
662 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
663 quality of interpolated samples using neighbour good samples.
666 Set threshold value. Allowed range is from @code{1} to @code{100}.
667 Default value is @code{10}. Higher values make clip detection less aggressive.
670 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
671 Default value is @code{1000}. Higher values make clip detection less aggressive.
676 It accepts the following values:
679 Select overlap-add method. Even not interpolated samples are slightly changed
683 Select overlap-save method. Not interpolated samples remain unchanged.
686 Default value is @code{a}.
691 Delay one or more audio channels.
693 Samples in delayed channel are filled with silence.
695 The filter accepts the following option:
699 Set list of delays in milliseconds for each channel separated by '|'.
700 Unused delays will be silently ignored. If number of given delays is
701 smaller than number of channels all remaining channels will not be delayed.
702 If you want to delay exact number of samples, append 'S' to number.
703 If you want instead to delay in seconds, append 's' to number.
706 Use last set delay for all remaining channels. By default is disabled.
707 This option if enabled changes how option @code{delays} is interpreted.
714 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
715 the second channel (and any other channels that may be present) unchanged.
721 Delay second channel by 500 samples, the third channel by 700 samples and leave
722 the first channel (and any other channels that may be present) unchanged.
728 Delay all channels by same number of samples:
730 adelay=delays=64S:all=1
734 @section aderivative, aintegral
736 Compute derivative/integral of audio stream.
738 Applying both filters one after another produces original audio.
742 Apply echoing to the input audio.
744 Echoes are reflected sound and can occur naturally amongst mountains
745 (and sometimes large buildings) when talking or shouting; digital echo
746 effects emulate this behaviour and are often used to help fill out the
747 sound of a single instrument or vocal. The time difference between the
748 original signal and the reflection is the @code{delay}, and the
749 loudness of the reflected signal is the @code{decay}.
750 Multiple echoes can have different delays and decays.
752 A description of the accepted parameters follows.
756 Set input gain of reflected signal. Default is @code{0.6}.
759 Set output gain of reflected signal. Default is @code{0.3}.
762 Set list of time intervals in milliseconds between original signal and reflections
763 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
764 Default is @code{1000}.
767 Set list of loudness of reflected signals separated by '|'.
768 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
769 Default is @code{0.5}.
776 Make it sound as if there are twice as many instruments as are actually playing:
778 aecho=0.8:0.88:60:0.4
782 If delay is very short, then it sounds like a (metallic) robot playing music:
788 A longer delay will sound like an open air concert in the mountains:
790 aecho=0.8:0.9:1000:0.3
794 Same as above but with one more mountain:
796 aecho=0.8:0.9:1000|1800:0.3|0.25
801 Audio emphasis filter creates or restores material directly taken from LPs or
802 emphased CDs with different filter curves. E.g. to store music on vinyl the
803 signal has to be altered by a filter first to even out the disadvantages of
804 this recording medium.
805 Once the material is played back the inverse filter has to be applied to
806 restore the distortion of the frequency response.
808 The filter accepts the following options:
818 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
819 use @code{production} mode. Default is @code{reproduction} mode.
822 Set filter type. Selects medium. Can be one of the following:
834 select Compact Disc (CD).
840 select 50µs (FM-KF).
842 select 75µs (FM-KF).
848 Modify an audio signal according to the specified expressions.
850 This filter accepts one or more expressions (one for each channel),
851 which are evaluated and used to modify a corresponding audio signal.
853 It accepts the following parameters:
857 Set the '|'-separated expressions list for each separate channel. If
858 the number of input channels is greater than the number of
859 expressions, the last specified expression is used for the remaining
862 @item channel_layout, c
863 Set output channel layout. If not specified, the channel layout is
864 specified by the number of expressions. If set to @samp{same}, it will
865 use by default the same input channel layout.
868 Each expression in @var{exprs} can contain the following constants and functions:
872 channel number of the current expression
875 number of the evaluated sample, starting from 0
881 time of the evaluated sample expressed in seconds
884 @item nb_out_channels
885 input and output number of channels
888 the value of input channel with number @var{CH}
891 Note: this filter is slow. For faster processing you should use a
900 aeval=val(ch)/2:c=same
904 Invert phase of the second channel:
913 Apply fade-in/out effect to input audio.
915 A description of the accepted parameters follows.
919 Specify the effect type, can be either @code{in} for fade-in, or
920 @code{out} for a fade-out effect. Default is @code{in}.
922 @item start_sample, ss
923 Specify the number of the start sample for starting to apply the fade
924 effect. Default is 0.
927 Specify the number of samples for which the fade effect has to last. At
928 the end of the fade-in effect the output audio will have the same
929 volume as the input audio, at the end of the fade-out transition
930 the output audio will be silence. Default is 44100.
933 Specify the start time of the fade effect. Default is 0.
934 The value must be specified as a time duration; see
935 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
936 for the accepted syntax.
937 If set this option is used instead of @var{start_sample}.
940 Specify the duration of the fade effect. See
941 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
942 for the accepted syntax.
943 At the end of the fade-in effect the output audio will have the same
944 volume as the input audio, at the end of the fade-out transition
945 the output audio will be silence.
946 By default the duration is determined by @var{nb_samples}.
947 If set this option is used instead of @var{nb_samples}.
950 Set curve for fade transition.
952 It accepts the following values:
955 select triangular, linear slope (default)
957 select quarter of sine wave
959 select half of sine wave
961 select exponential sine wave
965 select inverted parabola
979 select inverted quarter of sine wave
981 select inverted half of sine wave
983 select double-exponential seat
985 select double-exponential sigmoid
987 select logistic sigmoid
997 Fade in first 15 seconds of audio:
1003 Fade out last 25 seconds of a 900 seconds audio:
1005 afade=t=out:st=875:d=25
1010 Denoise audio samples with FFT.
1012 A description of the accepted parameters follows.
1016 Set the noise reduction in dB, allowed range is 0.01 to 97.
1017 Default value is 12 dB.
1020 Set the noise floor in dB, allowed range is -80 to -20.
1021 Default value is -50 dB.
1026 It accepts the following values:
1035 Select shellac noise.
1038 Select custom noise, defined in @code{bn} option.
1040 Default value is white noise.
1044 Set custom band noise for every one of 15 bands.
1045 Bands are separated by ' ' or '|'.
1048 Set the residual floor in dB, allowed range is -80 to -20.
1049 Default value is -38 dB.
1052 Enable noise tracking. By default is disabled.
1053 With this enabled, noise floor is automatically adjusted.
1056 Enable residual tracking. By default is disabled.
1059 Set the output mode.
1061 It accepts the following values:
1064 Pass input unchanged.
1067 Pass noise filtered out.
1072 Default value is @var{o}.
1076 @subsection Commands
1078 This filter supports the following commands:
1080 @item sample_noise, sn
1081 Start or stop measuring noise profile.
1082 Syntax for the command is : "start" or "stop" string.
1083 After measuring noise profile is stopped it will be
1084 automatically applied in filtering.
1086 @item noise_reduction, nr
1087 Change noise reduction. Argument is single float number.
1088 Syntax for the command is : "@var{noise_reduction}"
1090 @item noise_floor, nf
1091 Change noise floor. Argument is single float number.
1092 Syntax for the command is : "@var{noise_floor}"
1094 @item output_mode, om
1095 Change output mode operation.
1096 Syntax for the command is : "i", "o" or "n" string.
1100 Apply arbitrary expressions to samples in frequency domain.
1104 Set frequency domain real expression for each separate channel separated
1105 by '|'. Default is "re".
1106 If the number of input channels is greater than the number of
1107 expressions, the last specified expression is used for the remaining
1111 Set frequency domain imaginary expression for each separate channel
1112 separated by '|'. Default is "im".
1114 Each expression in @var{real} and @var{imag} can contain the following
1115 constants and functions:
1122 current frequency bin number
1125 number of available bins
1128 channel number of the current expression
1137 current real part of frequency bin of current channel
1140 current imaginary part of frequency bin of current channel
1143 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1146 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1150 Set window size. Allowed range is from 16 to 131072.
1151 Default is @code{4096}
1154 Set window function. Default is @code{hann}.
1157 Set window overlap. If set to 1, the recommended overlap for selected
1158 window function will be picked. Default is @code{0.75}.
1161 @subsection Examples
1165 Leave almost only low frequencies in audio:
1167 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1171 Apply robotize effect:
1173 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1177 Apply whisper effect:
1179 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"
1186 Apply an arbitrary Finite Impulse Response filter.
1188 This filter is designed for applying long FIR filters,
1189 up to 60 seconds long.
1191 It can be used as component for digital crossover filters,
1192 room equalization, cross talk cancellation, wavefield synthesis,
1193 auralization, ambiophonics, ambisonics and spatialization.
1195 This filter uses the streams higher than first one as FIR coefficients.
1196 If the non-first stream holds a single channel, it will be used
1197 for all input channels in the first stream, otherwise
1198 the number of channels in the non-first stream must be same as
1199 the number of channels in the first stream.
1201 It accepts the following parameters:
1205 Set dry gain. This sets input gain.
1208 Set wet gain. This sets final output gain.
1211 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1214 Enable applying gain measured from power of IR.
1216 Set which approach to use for auto gain measurement.
1220 Do not apply any gain.
1223 select peak gain, very conservative approach. This is default value.
1226 select DC gain, limited application.
1229 select gain to noise approach, this is most popular one.
1233 Set gain to be applied to IR coefficients before filtering.
1234 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1237 Set format of IR stream. Can be @code{mono} or @code{input}.
1238 Default is @code{input}.
1241 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1242 Allowed range is 0.1 to 60 seconds.
1245 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1246 By default it is disabled.
1249 Set for which IR channel to display frequency response. By default is first channel
1250 displayed. This option is used only when @var{response} is enabled.
1253 Set video stream size. This option is used only when @var{response} is enabled.
1256 Set video stream frame rate. This option is used only when @var{response} is enabled.
1259 Set minimal partition size used for convolution. Default is @var{8192}.
1260 Allowed range is from @var{1} to @var{32768}.
1261 Lower values decreases latency at cost of higher CPU usage.
1264 Set maximal partition size used for convolution. Default is @var{8192}.
1265 Allowed range is from @var{8} to @var{32768}.
1266 Lower values may increase CPU usage.
1269 Set number of input impulse responses streams which will be switchable at runtime.
1270 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1273 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1274 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1275 This option can be changed at runtime via @ref{commands}.
1278 @subsection Examples
1282 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1284 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1291 Set output format constraints for the input audio. The framework will
1292 negotiate the most appropriate format to minimize conversions.
1294 It accepts the following parameters:
1297 @item sample_fmts, f
1298 A '|'-separated list of requested sample formats.
1300 @item sample_rates, r
1301 A '|'-separated list of requested sample rates.
1303 @item channel_layouts, cl
1304 A '|'-separated list of requested channel layouts.
1306 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1307 for the required syntax.
1310 If a parameter is omitted, all values are allowed.
1312 Force the output to either unsigned 8-bit or signed 16-bit stereo
1314 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1319 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1320 processing reduces disturbing noise between useful signals.
1322 Gating is done by detecting the volume below a chosen level @var{threshold}
1323 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1324 floor is set via @var{range}. Because an exact manipulation of the signal
1325 would cause distortion of the waveform the reduction can be levelled over
1326 time. This is done by setting @var{attack} and @var{release}.
1328 @var{attack} determines how long the signal has to fall below the threshold
1329 before any reduction will occur and @var{release} sets the time the signal
1330 has to rise above the threshold to reduce the reduction again.
1331 Shorter signals than the chosen attack time will be left untouched.
1335 Set input level before filtering.
1336 Default is 1. Allowed range is from 0.015625 to 64.
1339 Set the mode of operation. Can be @code{upward} or @code{downward}.
1340 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1341 will be amplified, expanding dynamic range in upward direction.
1342 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1345 Set the level of gain reduction when the signal is below the threshold.
1346 Default is 0.06125. Allowed range is from 0 to 1.
1347 Setting this to 0 disables reduction and then filter behaves like expander.
1350 If a signal rises above this level the gain reduction is released.
1351 Default is 0.125. Allowed range is from 0 to 1.
1354 Set a ratio by which the signal is reduced.
1355 Default is 2. Allowed range is from 1 to 9000.
1358 Amount of milliseconds the signal has to rise above the threshold before gain
1360 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1363 Amount of milliseconds the signal has to fall below the threshold before the
1364 reduction is increased again. Default is 250 milliseconds.
1365 Allowed range is from 0.01 to 9000.
1368 Set amount of amplification of signal after processing.
1369 Default is 1. Allowed range is from 1 to 64.
1372 Curve the sharp knee around the threshold to enter gain reduction more softly.
1373 Default is 2.828427125. Allowed range is from 1 to 8.
1376 Choose if exact signal should be taken for detection or an RMS like one.
1377 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1380 Choose if the average level between all channels or the louder channel affects
1382 Default is @code{average}. Can be @code{average} or @code{maximum}.
1387 Apply an arbitrary Infinite Impulse Response filter.
1389 It accepts the following parameters:
1393 Set numerator/zeros coefficients.
1396 Set denominator/poles coefficients.
1408 Set coefficients format.
1412 digital transfer function
1414 Z-plane zeros/poles, cartesian (default)
1416 Z-plane zeros/poles, polar radians
1418 Z-plane zeros/poles, polar degrees
1424 Set kind of processing.
1425 Can be @code{d} - direct or @code{s} - serial cascading. Default is @code{s}.
1428 Set filtering precision.
1432 double-precision floating-point (default)
1434 single-precision floating-point
1442 Normalize filter coefficients, by default is enabled.
1443 Enabling it will normalize magnitude response at DC to 0dB.
1446 How much to use filtered signal in output. Default is 1.
1447 Range is between 0 and 1.
1450 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1451 By default it is disabled.
1454 Set for which IR channel to display frequency response. By default is first channel
1455 displayed. This option is used only when @var{response} is enabled.
1458 Set video stream size. This option is used only when @var{response} is enabled.
1461 Coefficients in @code{tf} format are separated by spaces and are in ascending
1464 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1465 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1468 Different coefficients and gains can be provided for every channel, in such case
1469 use '|' to separate coefficients or gains. Last provided coefficients will be
1470 used for all remaining channels.
1472 @subsection Examples
1476 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1478 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
1482 Same as above but in @code{zp} format:
1484 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
1490 The limiter prevents an input signal from rising over a desired threshold.
1491 This limiter uses lookahead technology to prevent your signal from distorting.
1492 It means that there is a small delay after the signal is processed. Keep in mind
1493 that the delay it produces is the attack time you set.
1495 The filter accepts the following options:
1499 Set input gain. Default is 1.
1502 Set output gain. Default is 1.
1505 Don't let signals above this level pass the limiter. Default is 1.
1508 The limiter will reach its attenuation level in this amount of time in
1509 milliseconds. Default is 5 milliseconds.
1512 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1513 Default is 50 milliseconds.
1516 When gain reduction is always needed ASC takes care of releasing to an
1517 average reduction level rather than reaching a reduction of 0 in the release
1521 Select how much the release time is affected by ASC, 0 means nearly no changes
1522 in release time while 1 produces higher release times.
1525 Auto level output signal. Default is enabled.
1526 This normalizes audio back to 0dB if enabled.
1529 Depending on picked setting it is recommended to upsample input 2x or 4x times
1530 with @ref{aresample} before applying this filter.
1534 Apply a two-pole all-pass filter with central frequency (in Hz)
1535 @var{frequency}, and filter-width @var{width}.
1536 An all-pass filter changes the audio's frequency to phase relationship
1537 without changing its frequency to amplitude relationship.
1539 The filter accepts the following options:
1543 Set frequency in Hz.
1546 Set method to specify band-width of filter.
1561 Specify the band-width of a filter in width_type units.
1564 How much to use filtered signal in output. Default is 1.
1565 Range is between 0 and 1.
1568 Specify which channels to filter, by default all available are filtered.
1571 Normalize biquad coefficients, by default is disabled.
1572 Enabling it will normalize magnitude response at DC to 0dB.
1575 Set the filter order, can be 1 or 2. Default is 2.
1578 Set transform type of IIR filter.
1586 @subsection Commands
1588 This filter supports the following commands:
1591 Change allpass frequency.
1592 Syntax for the command is : "@var{frequency}"
1595 Change allpass width_type.
1596 Syntax for the command is : "@var{width_type}"
1599 Change allpass width.
1600 Syntax for the command is : "@var{width}"
1604 Syntax for the command is : "@var{mix}"
1611 The filter accepts the following options:
1615 Set the number of loops. Setting this value to -1 will result in infinite loops.
1619 Set maximal number of samples. Default is 0.
1622 Set first sample of loop. Default is 0.
1628 Merge two or more audio streams into a single multi-channel stream.
1630 The filter accepts the following options:
1635 Set the number of inputs. Default is 2.
1639 If the channel layouts of the inputs are disjoint, and therefore compatible,
1640 the channel layout of the output will be set accordingly and the channels
1641 will be reordered as necessary. If the channel layouts of the inputs are not
1642 disjoint, the output will have all the channels of the first input then all
1643 the channels of the second input, in that order, and the channel layout of
1644 the output will be the default value corresponding to the total number of
1647 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1648 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1649 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1650 first input, b1 is the first channel of the second input).
1652 On the other hand, if both input are in stereo, the output channels will be
1653 in the default order: a1, a2, b1, b2, and the channel layout will be
1654 arbitrarily set to 4.0, which may or may not be the expected value.
1656 All inputs must have the same sample rate, and format.
1658 If inputs do not have the same duration, the output will stop with the
1661 @subsection Examples
1665 Merge two mono files into a stereo stream:
1667 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1671 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1673 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
1679 Mixes multiple audio inputs into a single output.
1681 Note that this filter only supports float samples (the @var{amerge}
1682 and @var{pan} audio filters support many formats). If the @var{amix}
1683 input has integer samples then @ref{aresample} will be automatically
1684 inserted to perform the conversion to float samples.
1688 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1690 will mix 3 input audio streams to a single output with the same duration as the
1691 first input and a dropout transition time of 3 seconds.
1693 It accepts the following parameters:
1697 The number of inputs. If unspecified, it defaults to 2.
1700 How to determine the end-of-stream.
1704 The duration of the longest input. (default)
1707 The duration of the shortest input.
1710 The duration of the first input.
1714 @item dropout_transition
1715 The transition time, in seconds, for volume renormalization when an input
1716 stream ends. The default value is 2 seconds.
1719 Specify weight of each input audio stream as sequence.
1720 Each weight is separated by space. By default all inputs have same weight.
1723 @subsection Commands
1725 This filter supports the following commands:
1728 Syntax is same as option with same name.
1733 Multiply first audio stream with second audio stream and store result
1734 in output audio stream. Multiplication is done by multiplying each
1735 sample from first stream with sample at same position from second stream.
1737 With this element-wise multiplication one can create amplitude fades and
1738 amplitude modulations.
1740 @section anequalizer
1742 High-order parametric multiband equalizer for each channel.
1744 It accepts the following parameters:
1748 This option string is in format:
1749 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1750 Each equalizer band is separated by '|'.
1754 Set channel number to which equalization will be applied.
1755 If input doesn't have that channel the entry is ignored.
1758 Set central frequency for band.
1759 If input doesn't have that frequency the entry is ignored.
1762 Set band width in hertz.
1765 Set band gain in dB.
1768 Set filter type for band, optional, can be:
1772 Butterworth, this is default.
1783 With this option activated frequency response of anequalizer is displayed
1787 Set video stream size. Only useful if curves option is activated.
1790 Set max gain that will be displayed. Only useful if curves option is activated.
1791 Setting this to a reasonable value makes it possible to display gain which is derived from
1792 neighbour bands which are too close to each other and thus produce higher gain
1793 when both are activated.
1796 Set frequency scale used to draw frequency response in video output.
1797 Can be linear or logarithmic. Default is logarithmic.
1800 Set color for each channel curve which is going to be displayed in video stream.
1801 This is list of color names separated by space or by '|'.
1802 Unrecognised or missing colors will be replaced by white color.
1805 @subsection Examples
1809 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1810 for first 2 channels using Chebyshev type 1 filter:
1812 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1816 @subsection Commands
1818 This filter supports the following commands:
1821 Alter existing filter parameters.
1822 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1824 @var{fN} is existing filter number, starting from 0, if no such filter is available
1826 @var{freq} set new frequency parameter.
1827 @var{width} set new width parameter in herz.
1828 @var{gain} set new gain parameter in dB.
1830 Full filter invocation with asendcmd may look like this:
1831 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1836 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1838 Each sample is adjusted by looking for other samples with similar contexts. This
1839 context similarity is defined by comparing their surrounding patches of size
1840 @option{p}. Patches are searched in an area of @option{r} around the sample.
1842 The filter accepts the following options:
1846 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
1849 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
1850 Default value is 2 milliseconds.
1853 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
1854 Default value is 6 milliseconds.
1857 Set the output mode.
1859 It accepts the following values:
1862 Pass input unchanged.
1865 Pass noise filtered out.
1870 Default value is @var{o}.
1874 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
1877 @subsection Commands
1879 This filter supports the following commands:
1882 Change denoise strength. Argument is single float number.
1883 Syntax for the command is : "@var{s}"
1887 Syntax for the command is : "i", "o" or "n" string.
1891 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
1893 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
1894 relate to producing the least mean square of the error signal (difference between the desired,
1895 2nd input audio stream and the actual signal, the 1st input audio stream).
1897 A description of the accepted options follows.
1910 Set the filter leakage.
1913 It accepts the following values:
1922 Pass filtered samples.
1925 Pass difference between desired and filtered samples.
1927 Default value is @var{o}.
1931 @subsection Examples
1935 One of many usages of this filter is noise reduction, input audio is filtered
1936 with same samples that are delayed by fixed amount, one such example for stereo audio is:
1938 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
1942 @subsection Commands
1944 This filter supports the same commands as options, excluding option @code{order}.
1948 Pass the audio source unchanged to the output.
1952 Pad the end of an audio stream with silence.
1954 This can be used together with @command{ffmpeg} @option{-shortest} to
1955 extend audio streams to the same length as the video stream.
1957 A description of the accepted options follows.
1961 Set silence packet size. Default value is 4096.
1964 Set the number of samples of silence to add to the end. After the
1965 value is reached, the stream is terminated. This option is mutually
1966 exclusive with @option{whole_len}.
1969 Set the minimum total number of samples in the output audio stream. If
1970 the value is longer than the input audio length, silence is added to
1971 the end, until the value is reached. This option is mutually exclusive
1972 with @option{pad_len}.
1975 Specify the duration of samples of silence to add. See
1976 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1977 for the accepted syntax. Used only if set to non-zero value.
1980 Specify the minimum total duration in the output audio stream. See
1981 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1982 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
1983 the input audio length, silence is added to the end, until the value is reached.
1984 This option is mutually exclusive with @option{pad_dur}
1987 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
1988 nor @option{whole_dur} option is set, the filter will add silence to the end of
1989 the input stream indefinitely.
1991 @subsection Examples
1995 Add 1024 samples of silence to the end of the input:
2001 Make sure the audio output will contain at least 10000 samples, pad
2002 the input with silence if required:
2004 apad=whole_len=10000
2008 Use @command{ffmpeg} to pad the audio input with silence, so that the
2009 video stream will always result the shortest and will be converted
2010 until the end in the output file when using the @option{shortest}
2013 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2018 Add a phasing effect to the input audio.
2020 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2021 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2023 A description of the accepted parameters follows.
2027 Set input gain. Default is 0.4.
2030 Set output gain. Default is 0.74
2033 Set delay in milliseconds. Default is 3.0.
2036 Set decay. Default is 0.4.
2039 Set modulation speed in Hz. Default is 0.5.
2042 Set modulation type. Default is triangular.
2044 It accepts the following values:
2053 Audio pulsator is something between an autopanner and a tremolo.
2054 But it can produce funny stereo effects as well. Pulsator changes the volume
2055 of the left and right channel based on a LFO (low frequency oscillator) with
2056 different waveforms and shifted phases.
2057 This filter have the ability to define an offset between left and right
2058 channel. An offset of 0 means that both LFO shapes match each other.
2059 The left and right channel are altered equally - a conventional tremolo.
2060 An offset of 50% means that the shape of the right channel is exactly shifted
2061 in phase (or moved backwards about half of the frequency) - pulsator acts as
2062 an autopanner. At 1 both curves match again. Every setting in between moves the
2063 phase shift gapless between all stages and produces some "bypassing" sounds with
2064 sine and triangle waveforms. The more you set the offset near 1 (starting from
2065 the 0.5) the faster the signal passes from the left to the right speaker.
2067 The filter accepts the following options:
2071 Set input gain. By default it is 1. Range is [0.015625 - 64].
2074 Set output gain. By default it is 1. Range is [0.015625 - 64].
2077 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2078 sawup or sawdown. Default is sine.
2081 Set modulation. Define how much of original signal is affected by the LFO.
2084 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2087 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2090 Set pulse width. Default is 1. Allowed range is [0 - 2].
2093 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2096 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2100 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2104 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2105 if timing is set to hz.
2111 Resample the input audio to the specified parameters, using the
2112 libswresample library. If none are specified then the filter will
2113 automatically convert between its input and output.
2115 This filter is also able to stretch/squeeze the audio data to make it match
2116 the timestamps or to inject silence / cut out audio to make it match the
2117 timestamps, do a combination of both or do neither.
2119 The filter accepts the syntax
2120 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2121 expresses a sample rate and @var{resampler_options} is a list of
2122 @var{key}=@var{value} pairs, separated by ":". See the
2123 @ref{Resampler Options,,"Resampler Options" section in the
2124 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2125 for the complete list of supported options.
2127 @subsection Examples
2131 Resample the input audio to 44100Hz:
2137 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2138 samples per second compensation:
2140 aresample=async=1000
2146 Reverse an audio clip.
2148 Warning: This filter requires memory to buffer the entire clip, so trimming
2151 @subsection Examples
2155 Take the first 5 seconds of a clip, and reverse it.
2157 atrim=end=5,areverse
2163 Reduce noise from speech using Recurrent Neural Networks.
2165 This filter accepts the following options:
2169 Set train model file to load. This option is always required.
2172 @section asetnsamples
2174 Set the number of samples per each output audio frame.
2176 The last output packet may contain a different number of samples, as
2177 the filter will flush all the remaining samples when the input audio
2180 The filter accepts the following options:
2184 @item nb_out_samples, n
2185 Set the number of frames per each output audio frame. The number is
2186 intended as the number of samples @emph{per each channel}.
2187 Default value is 1024.
2190 If set to 1, the filter will pad the last audio frame with zeroes, so
2191 that the last frame will contain the same number of samples as the
2192 previous ones. Default value is 1.
2195 For example, to set the number of per-frame samples to 1234 and
2196 disable padding for the last frame, use:
2198 asetnsamples=n=1234:p=0
2203 Set the sample rate without altering the PCM data.
2204 This will result in a change of speed and pitch.
2206 The filter accepts the following options:
2209 @item sample_rate, r
2210 Set the output sample rate. Default is 44100 Hz.
2215 Show a line containing various information for each input audio frame.
2216 The input audio is not modified.
2218 The shown line contains a sequence of key/value pairs of the form
2219 @var{key}:@var{value}.
2221 The following values are shown in the output:
2225 The (sequential) number of the input frame, starting from 0.
2228 The presentation timestamp of the input frame, in time base units; the time base
2229 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2232 The presentation timestamp of the input frame in seconds.
2235 position of the frame in the input stream, -1 if this information in
2236 unavailable and/or meaningless (for example in case of synthetic audio)
2245 The sample rate for the audio frame.
2248 The number of samples (per channel) in the frame.
2251 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2252 audio, the data is treated as if all the planes were concatenated.
2254 @item plane_checksums
2255 A list of Adler-32 checksums for each data plane.
2259 Apply audio soft clipping.
2261 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2262 along a smooth curve, rather than the abrupt shape of hard-clipping.
2264 This filter accepts the following options:
2268 Set type of soft-clipping.
2270 It accepts the following values:
2282 Set additional parameter which controls sigmoid function.
2285 @subsection Commands
2287 This filter supports the all above options as @ref{commands}.
2290 Automatic Speech Recognition
2292 This filter uses PocketSphinx for speech recognition. To enable
2293 compilation of this filter, you need to configure FFmpeg with
2294 @code{--enable-pocketsphinx}.
2296 It accepts the following options:
2300 Set sampling rate of input audio. Defaults is @code{16000}.
2301 This need to match speech models, otherwise one will get poor results.
2304 Set dictionary containing acoustic model files.
2307 Set pronunciation dictionary.
2310 Set language model file.
2313 Set language model set.
2316 Set which language model to use.
2319 Set output for log messages.
2322 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2327 Display time domain statistical information about the audio channels.
2328 Statistics are calculated and displayed for each audio channel and,
2329 where applicable, an overall figure is also given.
2331 It accepts the following option:
2334 Short window length in seconds, used for peak and trough RMS measurement.
2335 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2339 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2340 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2343 Available keys for each channel are:
2389 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2390 this @code{lavfi.astats.Overall.Peak_count}.
2392 For description what each key means read below.
2395 Set number of frame after which stats are going to be recalculated.
2396 Default is disabled.
2398 @item measure_perchannel
2399 Select the entries which need to be measured per channel. The metadata keys can
2400 be used as flags, default is @option{all} which measures everything.
2401 @option{none} disables all per channel measurement.
2403 @item measure_overall
2404 Select the entries which need to be measured overall. The metadata keys can
2405 be used as flags, default is @option{all} which measures everything.
2406 @option{none} disables all overall measurement.
2410 A description of each shown parameter follows:
2414 Mean amplitude displacement from zero.
2417 Minimal sample level.
2420 Maximal sample level.
2422 @item Min difference
2423 Minimal difference between two consecutive samples.
2425 @item Max difference
2426 Maximal difference between two consecutive samples.
2428 @item Mean difference
2429 Mean difference between two consecutive samples.
2430 The average of each difference between two consecutive samples.
2432 @item RMS difference
2433 Root Mean Square difference between two consecutive samples.
2437 Standard peak and RMS level measured in dBFS.
2441 Peak and trough values for RMS level measured over a short window.
2444 Standard ratio of peak to RMS level (note: not in dB).
2447 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2448 (i.e. either @var{Min level} or @var{Max level}).
2451 Number of occasions (not the number of samples) that the signal attained either
2452 @var{Min level} or @var{Max level}.
2454 @item Noise floor dB
2455 Minimum local peak measured in dBFS over a short window.
2457 @item Noise floor count
2458 Number of occasions (not the number of samples) that the signal attained
2462 Overall bit depth of audio. Number of bits used for each sample.
2465 Measured dynamic range of audio in dB.
2467 @item Zero crossings
2468 Number of points where the waveform crosses the zero level axis.
2470 @item Zero crossings rate
2471 Rate of Zero crossings and number of audio samples.
2475 Boost subwoofer frequencies.
2477 The filter accepts the following options:
2481 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2482 Default value is 0.5.
2485 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2486 Default value is 0.8.
2489 Set delay line decay gain value. Allowed range is from 0 to 1.
2490 Default value is 0.7.
2493 Set delay line feedback gain value. Allowed range is from 0 to 1.
2494 Default value is 0.5.
2497 Set cutoff frequency in herz. Allowed range is 50 to 900.
2498 Default value is 100.
2501 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2502 Default value is 0.5.
2505 Set delay. Allowed range is from 1 to 100.
2506 Default value is 20.
2509 @subsection Commands
2511 This filter supports the all above options as @ref{commands}.
2517 The filter accepts exactly one parameter, the audio tempo. If not
2518 specified then the filter will assume nominal 1.0 tempo. Tempo must
2519 be in the [0.5, 100.0] range.
2521 Note that tempo greater than 2 will skip some samples rather than
2522 blend them in. If for any reason this is a concern it is always
2523 possible to daisy-chain several instances of atempo to achieve the
2524 desired product tempo.
2526 @subsection Examples
2530 Slow down audio to 80% tempo:
2536 To speed up audio to 300% tempo:
2542 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2544 atempo=sqrt(3),atempo=sqrt(3)
2548 @subsection Commands
2550 This filter supports the following commands:
2553 Change filter tempo scale factor.
2554 Syntax for the command is : "@var{tempo}"
2559 Trim the input so that the output contains one continuous subpart of the input.
2561 It accepts the following parameters:
2564 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2565 sample with the timestamp @var{start} will be the first sample in the output.
2568 Specify time of the first audio sample that will be dropped, i.e. the
2569 audio sample immediately preceding the one with the timestamp @var{end} will be
2570 the last sample in the output.
2573 Same as @var{start}, except this option sets the start timestamp in samples
2577 Same as @var{end}, except this option sets the end timestamp in samples instead
2581 The maximum duration of the output in seconds.
2584 The number of the first sample that should be output.
2587 The number of the first sample that should be dropped.
2590 @option{start}, @option{end}, and @option{duration} are expressed as time
2591 duration specifications; see
2592 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2594 Note that the first two sets of the start/end options and the @option{duration}
2595 option look at the frame timestamp, while the _sample options simply count the
2596 samples that pass through the filter. So start/end_pts and start/end_sample will
2597 give different results when the timestamps are wrong, inexact or do not start at
2598 zero. Also note that this filter does not modify the timestamps. If you wish
2599 to have the output timestamps start at zero, insert the asetpts filter after the
2602 If multiple start or end options are set, this filter tries to be greedy and
2603 keep all samples that match at least one of the specified constraints. To keep
2604 only the part that matches all the constraints at once, chain multiple atrim
2607 The defaults are such that all the input is kept. So it is possible to set e.g.
2608 just the end values to keep everything before the specified time.
2613 Drop everything except the second minute of input:
2615 ffmpeg -i INPUT -af atrim=60:120
2619 Keep only the first 1000 samples:
2621 ffmpeg -i INPUT -af atrim=end_sample=1000
2626 @section axcorrelate
2627 Calculate normalized cross-correlation between two input audio streams.
2629 Resulted samples are always between -1 and 1 inclusive.
2630 If result is 1 it means two input samples are highly correlated in that selected segment.
2631 Result 0 means they are not correlated at all.
2632 If result is -1 it means two input samples are out of phase, which means they cancel each
2635 The filter accepts the following options:
2639 Set size of segment over which cross-correlation is calculated.
2640 Default is 256. Allowed range is from 2 to 131072.
2643 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2644 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2645 are always zero and thus need much less calculations to make.
2646 This is generally not true, but is valid for typical audio streams.
2649 @subsection Examples
2653 Calculate correlation between channels in stereo audio stream:
2655 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2661 Apply a two-pole Butterworth band-pass filter with central
2662 frequency @var{frequency}, and (3dB-point) band-width width.
2663 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2664 instead of the default: constant 0dB peak gain.
2665 The filter roll off at 6dB per octave (20dB per decade).
2667 The filter accepts the following options:
2671 Set the filter's central frequency. Default is @code{3000}.
2674 Constant skirt gain if set to 1. Defaults to 0.
2677 Set method to specify band-width of filter.
2692 Specify the band-width of a filter in width_type units.
2695 How much to use filtered signal in output. Default is 1.
2696 Range is between 0 and 1.
2699 Specify which channels to filter, by default all available are filtered.
2702 Normalize biquad coefficients, by default is disabled.
2703 Enabling it will normalize magnitude response at DC to 0dB.
2706 Set transform type of IIR filter.
2714 @subsection Commands
2716 This filter supports the following commands:
2719 Change bandpass frequency.
2720 Syntax for the command is : "@var{frequency}"
2723 Change bandpass width_type.
2724 Syntax for the command is : "@var{width_type}"
2727 Change bandpass width.
2728 Syntax for the command is : "@var{width}"
2731 Change bandpass mix.
2732 Syntax for the command is : "@var{mix}"
2737 Apply a two-pole Butterworth band-reject filter with central
2738 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2739 The filter roll off at 6dB per octave (20dB per decade).
2741 The filter accepts the following options:
2745 Set the filter's central frequency. Default is @code{3000}.
2748 Set method to specify band-width of filter.
2763 Specify the band-width of a filter in width_type units.
2766 How much to use filtered signal in output. Default is 1.
2767 Range is between 0 and 1.
2770 Specify which channels to filter, by default all available are filtered.
2773 Normalize biquad coefficients, by default is disabled.
2774 Enabling it will normalize magnitude response at DC to 0dB.
2777 Set transform type of IIR filter.
2785 @subsection Commands
2787 This filter supports the following commands:
2790 Change bandreject frequency.
2791 Syntax for the command is : "@var{frequency}"
2794 Change bandreject width_type.
2795 Syntax for the command is : "@var{width_type}"
2798 Change bandreject width.
2799 Syntax for the command is : "@var{width}"
2802 Change bandreject mix.
2803 Syntax for the command is : "@var{mix}"
2806 @section bass, lowshelf
2808 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2809 shelving filter with a response similar to that of a standard
2810 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2812 The filter accepts the following options:
2816 Give the gain at 0 Hz. Its useful range is about -20
2817 (for a large cut) to +20 (for a large boost).
2818 Beware of clipping when using a positive gain.
2821 Set the filter's central frequency and so can be used
2822 to extend or reduce the frequency range to be boosted or cut.
2823 The default value is @code{100} Hz.
2826 Set method to specify band-width of filter.
2841 Determine how steep is the filter's shelf transition.
2844 How much to use filtered signal in output. Default is 1.
2845 Range is between 0 and 1.
2848 Specify which channels to filter, by default all available are filtered.
2851 Normalize biquad coefficients, by default is disabled.
2852 Enabling it will normalize magnitude response at DC to 0dB.
2855 Set transform type of IIR filter.
2863 @subsection Commands
2865 This filter supports the following commands:
2868 Change bass frequency.
2869 Syntax for the command is : "@var{frequency}"
2872 Change bass width_type.
2873 Syntax for the command is : "@var{width_type}"
2877 Syntax for the command is : "@var{width}"
2881 Syntax for the command is : "@var{gain}"
2885 Syntax for the command is : "@var{mix}"
2890 Apply a biquad IIR filter with the given coefficients.
2891 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2892 are the numerator and denominator coefficients respectively.
2893 and @var{channels}, @var{c} specify which channels to filter, by default all
2894 available are filtered.
2896 @subsection Commands
2898 This filter supports the following commands:
2906 Change biquad parameter.
2907 Syntax for the command is : "@var{value}"
2910 How much to use filtered signal in output. Default is 1.
2911 Range is between 0 and 1.
2914 Specify which channels to filter, by default all available are filtered.
2917 Normalize biquad coefficients, by default is disabled.
2918 Enabling it will normalize magnitude response at DC to 0dB.
2921 Set transform type of IIR filter.
2930 Bauer stereo to binaural transformation, which improves headphone listening of
2931 stereo audio records.
2933 To enable compilation of this filter you need to configure FFmpeg with
2934 @code{--enable-libbs2b}.
2936 It accepts the following parameters:
2940 Pre-defined crossfeed level.
2944 Default level (fcut=700, feed=50).
2947 Chu Moy circuit (fcut=700, feed=60).
2950 Jan Meier circuit (fcut=650, feed=95).
2955 Cut frequency (in Hz).
2964 Remap input channels to new locations.
2966 It accepts the following parameters:
2969 Map channels from input to output. The argument is a '|'-separated list of
2970 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2971 @var{in_channel} form. @var{in_channel} can be either the name of the input
2972 channel (e.g. FL for front left) or its index in the input channel layout.
2973 @var{out_channel} is the name of the output channel or its index in the output
2974 channel layout. If @var{out_channel} is not given then it is implicitly an
2975 index, starting with zero and increasing by one for each mapping.
2977 @item channel_layout
2978 The channel layout of the output stream.
2981 If no mapping is present, the filter will implicitly map input channels to
2982 output channels, preserving indices.
2984 @subsection Examples
2988 For example, assuming a 5.1+downmix input MOV file,
2990 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2992 will create an output WAV file tagged as stereo from the downmix channels of
2996 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2998 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3002 @section channelsplit
3004 Split each channel from an input audio stream into a separate output stream.
3006 It accepts the following parameters:
3008 @item channel_layout
3009 The channel layout of the input stream. The default is "stereo".
3011 A channel layout describing the channels to be extracted as separate output streams
3012 or "all" to extract each input channel as a separate stream. The default is "all".
3014 Choosing channels not present in channel layout in the input will result in an error.
3017 @subsection Examples
3021 For example, assuming a stereo input MP3 file,
3023 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3025 will create an output Matroska file with two audio streams, one containing only
3026 the left channel and the other the right channel.
3029 Split a 5.1 WAV file into per-channel files:
3031 ffmpeg -i in.wav -filter_complex
3032 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3033 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3034 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3039 Extract only LFE from a 5.1 WAV file:
3041 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3042 -map '[LFE]' lfe.wav
3047 Add a chorus effect to the audio.
3049 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3051 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3052 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3053 The modulation depth defines the range the modulated delay is played before or after
3054 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3055 sound tuned around the original one, like in a chorus where some vocals are slightly
3058 It accepts the following parameters:
3061 Set input gain. Default is 0.4.
3064 Set output gain. Default is 0.4.
3067 Set delays. A typical delay is around 40ms to 60ms.
3079 @subsection Examples
3085 chorus=0.7:0.9:55:0.4:0.25:2
3091 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3095 Fuller sounding chorus with three delays:
3097 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
3102 Compress or expand the audio's dynamic range.
3104 It accepts the following parameters:
3110 A list of times in seconds for each channel over which the instantaneous level
3111 of the input signal is averaged to determine its volume. @var{attacks} refers to
3112 increase of volume and @var{decays} refers to decrease of volume. For most
3113 situations, the attack time (response to the audio getting louder) should be
3114 shorter than the decay time, because the human ear is more sensitive to sudden
3115 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3116 a typical value for decay is 0.8 seconds.
3117 If specified number of attacks & decays is lower than number of channels, the last
3118 set attack/decay will be used for all remaining channels.
3121 A list of points for the transfer function, specified in dB relative to the
3122 maximum possible signal amplitude. Each key points list must be defined using
3123 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3124 @code{x0/y0 x1/y1 x2/y2 ....}
3126 The input values must be in strictly increasing order but the transfer function
3127 does not have to be monotonically rising. The point @code{0/0} is assumed but
3128 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3129 function are @code{-70/-70|-60/-20|1/0}.
3132 Set the curve radius in dB for all joints. It defaults to 0.01.
3135 Set the additional gain in dB to be applied at all points on the transfer
3136 function. This allows for easy adjustment of the overall gain.
3140 Set an initial volume, in dB, to be assumed for each channel when filtering
3141 starts. This permits the user to supply a nominal level initially, so that, for
3142 example, a very large gain is not applied to initial signal levels before the
3143 companding has begun to operate. A typical value for audio which is initially
3144 quiet is -90 dB. It defaults to 0.
3147 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3148 delayed before being fed to the volume adjuster. Specifying a delay
3149 approximately equal to the attack/decay times allows the filter to effectively
3150 operate in predictive rather than reactive mode. It defaults to 0.
3154 @subsection Examples
3158 Make music with both quiet and loud passages suitable for listening to in a
3161 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3164 Another example for audio with whisper and explosion parts:
3166 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3170 A noise gate for when the noise is at a lower level than the signal:
3172 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3176 Here is another noise gate, this time for when the noise is at a higher level
3177 than the signal (making it, in some ways, similar to squelch):
3179 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3183 2:1 compression starting at -6dB:
3185 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3189 2:1 compression starting at -9dB:
3191 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3195 2:1 compression starting at -12dB:
3197 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3201 2:1 compression starting at -18dB:
3203 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3207 3:1 compression starting at -15dB:
3209 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3215 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3221 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
3225 Hard limiter at -6dB:
3227 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3231 Hard limiter at -12dB:
3233 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3237 Hard noise gate at -35 dB:
3239 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3245 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3249 @section compensationdelay
3251 Compensation Delay Line is a metric based delay to compensate differing
3252 positions of microphones or speakers.
3254 For example, you have recorded guitar with two microphones placed in
3255 different locations. Because the front of sound wave has fixed speed in
3256 normal conditions, the phasing of microphones can vary and depends on
3257 their location and interposition. The best sound mix can be achieved when
3258 these microphones are in phase (synchronized). Note that a distance of
3259 ~30 cm between microphones makes one microphone capture the signal in
3260 antiphase to the other microphone. That makes the final mix sound moody.
3261 This filter helps to solve phasing problems by adding different delays
3262 to each microphone track and make them synchronized.
3264 The best result can be reached when you take one track as base and
3265 synchronize other tracks one by one with it.
3266 Remember that synchronization/delay tolerance depends on sample rate, too.
3267 Higher sample rates will give more tolerance.
3269 The filter accepts the following parameters:
3273 Set millimeters distance. This is compensation distance for fine tuning.
3277 Set cm distance. This is compensation distance for tightening distance setup.
3281 Set meters distance. This is compensation distance for hard distance setup.
3285 Set dry amount. Amount of unprocessed (dry) signal.
3289 Set wet amount. Amount of processed (wet) signal.
3293 Set temperature in degrees Celsius. This is the temperature of the environment.
3298 Apply headphone crossfeed filter.
3300 Crossfeed is the process of blending the left and right channels of stereo
3302 It is mainly used to reduce extreme stereo separation of low frequencies.
3304 The intent is to produce more speaker like sound to the listener.
3306 The filter accepts the following options:
3310 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3311 This sets gain of low shelf filter for side part of stereo image.
3312 Default is -6dB. Max allowed is -30db when strength is set to 1.
3315 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3316 This sets cut off frequency of low shelf filter. Default is cut off near
3317 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3320 Set curve slope of low shelf filter. Default is 0.5.
3321 Allowed range is from 0.01 to 1.
3324 Set input gain. Default is 0.9.
3327 Set output gain. Default is 1.
3330 @subsection Commands
3332 This filter supports the all above options as @ref{commands}.
3334 @section crystalizer
3335 Simple algorithm to expand audio dynamic range.
3337 The filter accepts the following options:
3341 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3342 (unchanged sound) to 10.0 (maximum effect).
3345 Enable clipping. By default is enabled.
3348 @subsection Commands
3350 This filter supports the all above options as @ref{commands}.
3353 Apply a DC shift to the audio.
3355 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3356 in the recording chain) from the audio. The effect of a DC offset is reduced
3357 headroom and hence volume. The @ref{astats} filter can be used to determine if
3358 a signal has a DC offset.
3362 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3366 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3367 used to prevent clipping.
3372 Apply de-essing to the audio samples.
3376 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3380 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3384 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3388 Set the output mode.
3390 It accepts the following values:
3393 Pass input unchanged.
3396 Pass ess filtered out.
3401 Default value is @var{o}.
3407 Measure audio dynamic range.
3409 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3410 is found in transition material. And anything less that 8 have very poor dynamics
3411 and is very compressed.
3413 The filter accepts the following options:
3417 Set window length in seconds used to split audio into segments of equal length.
3418 Default is 3 seconds.
3422 Dynamic Audio Normalizer.
3424 This filter applies a certain amount of gain to the input audio in order
3425 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3426 contrast to more "simple" normalization algorithms, the Dynamic Audio
3427 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3428 This allows for applying extra gain to the "quiet" sections of the audio
3429 while avoiding distortions or clipping the "loud" sections. In other words:
3430 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3431 sections, in the sense that the volume of each section is brought to the
3432 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3433 this goal *without* applying "dynamic range compressing". It will retain 100%
3434 of the dynamic range *within* each section of the audio file.
3438 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3439 Default is 500 milliseconds.
3440 The Dynamic Audio Normalizer processes the input audio in small chunks,
3441 referred to as frames. This is required, because a peak magnitude has no
3442 meaning for just a single sample value. Instead, we need to determine the
3443 peak magnitude for a contiguous sequence of sample values. While a "standard"
3444 normalizer would simply use the peak magnitude of the complete file, the
3445 Dynamic Audio Normalizer determines the peak magnitude individually for each
3446 frame. The length of a frame is specified in milliseconds. By default, the
3447 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3448 been found to give good results with most files.
3449 Note that the exact frame length, in number of samples, will be determined
3450 automatically, based on the sampling rate of the individual input audio file.
3453 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3454 number. Default is 31.
3455 Probably the most important parameter of the Dynamic Audio Normalizer is the
3456 @code{window size} of the Gaussian smoothing filter. The filter's window size
3457 is specified in frames, centered around the current frame. For the sake of
3458 simplicity, this must be an odd number. Consequently, the default value of 31
3459 takes into account the current frame, as well as the 15 preceding frames and
3460 the 15 subsequent frames. Using a larger window results in a stronger
3461 smoothing effect and thus in less gain variation, i.e. slower gain
3462 adaptation. Conversely, using a smaller window results in a weaker smoothing
3463 effect and thus in more gain variation, i.e. faster gain adaptation.
3464 In other words, the more you increase this value, the more the Dynamic Audio
3465 Normalizer will behave like a "traditional" normalization filter. On the
3466 contrary, the more you decrease this value, the more the Dynamic Audio
3467 Normalizer will behave like a dynamic range compressor.
3470 Set the target peak value. This specifies the highest permissible magnitude
3471 level for the normalized audio input. This filter will try to approach the
3472 target peak magnitude as closely as possible, but at the same time it also
3473 makes sure that the normalized signal will never exceed the peak magnitude.
3474 A frame's maximum local gain factor is imposed directly by the target peak
3475 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3476 It is not recommended to go above this value.
3479 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3480 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3481 factor for each input frame, i.e. the maximum gain factor that does not
3482 result in clipping or distortion. The maximum gain factor is determined by
3483 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3484 additionally bounds the frame's maximum gain factor by a predetermined
3485 (global) maximum gain factor. This is done in order to avoid excessive gain
3486 factors in "silent" or almost silent frames. By default, the maximum gain
3487 factor is 10.0, For most inputs the default value should be sufficient and
3488 it usually is not recommended to increase this value. Though, for input
3489 with an extremely low overall volume level, it may be necessary to allow even
3490 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3491 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3492 Instead, a "sigmoid" threshold function will be applied. This way, the
3493 gain factors will smoothly approach the threshold value, but never exceed that
3497 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3498 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3499 This means that the maximum local gain factor for each frame is defined
3500 (only) by the frame's highest magnitude sample. This way, the samples can
3501 be amplified as much as possible without exceeding the maximum signal
3502 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3503 Normalizer can also take into account the frame's root mean square,
3504 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3505 determine the power of a time-varying signal. It is therefore considered
3506 that the RMS is a better approximation of the "perceived loudness" than
3507 just looking at the signal's peak magnitude. Consequently, by adjusting all
3508 frames to a constant RMS value, a uniform "perceived loudness" can be
3509 established. If a target RMS value has been specified, a frame's local gain
3510 factor is defined as the factor that would result in exactly that RMS value.
3511 Note, however, that the maximum local gain factor is still restricted by the
3512 frame's highest magnitude sample, in order to prevent clipping.
3515 Enable channels coupling. By default is enabled.
3516 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3517 amount. This means the same gain factor will be applied to all channels, i.e.
3518 the maximum possible gain factor is determined by the "loudest" channel.
3519 However, in some recordings, it may happen that the volume of the different
3520 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3521 In this case, this option can be used to disable the channel coupling. This way,
3522 the gain factor will be determined independently for each channel, depending
3523 only on the individual channel's highest magnitude sample. This allows for
3524 harmonizing the volume of the different channels.
3527 Enable DC bias correction. By default is disabled.
3528 An audio signal (in the time domain) is a sequence of sample values.
3529 In the Dynamic Audio Normalizer these sample values are represented in the
3530 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3531 audio signal, or "waveform", should be centered around the zero point.
3532 That means if we calculate the mean value of all samples in a file, or in a
3533 single frame, then the result should be 0.0 or at least very close to that
3534 value. If, however, there is a significant deviation of the mean value from
3535 0.0, in either positive or negative direction, this is referred to as a
3536 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3537 Audio Normalizer provides optional DC bias correction.
3538 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3539 the mean value, or "DC correction" offset, of each input frame and subtract
3540 that value from all of the frame's sample values which ensures those samples
3541 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3542 boundaries, the DC correction offset values will be interpolated smoothly
3543 between neighbouring frames.
3545 @item altboundary, b
3546 Enable alternative boundary mode. By default is disabled.
3547 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3548 around each frame. This includes the preceding frames as well as the
3549 subsequent frames. However, for the "boundary" frames, located at the very
3550 beginning and at the very end of the audio file, not all neighbouring
3551 frames are available. In particular, for the first few frames in the audio
3552 file, the preceding frames are not known. And, similarly, for the last few
3553 frames in the audio file, the subsequent frames are not known. Thus, the
3554 question arises which gain factors should be assumed for the missing frames
3555 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3556 to deal with this situation. The default boundary mode assumes a gain factor
3557 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3558 "fade out" at the beginning and at the end of the input, respectively.
3561 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3562 By default, the Dynamic Audio Normalizer does not apply "traditional"
3563 compression. This means that signal peaks will not be pruned and thus the
3564 full dynamic range will be retained within each local neighbourhood. However,
3565 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3566 normalization algorithm with a more "traditional" compression.
3567 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3568 (thresholding) function. If (and only if) the compression feature is enabled,
3569 all input frames will be processed by a soft knee thresholding function prior
3570 to the actual normalization process. Put simply, the thresholding function is
3571 going to prune all samples whose magnitude exceeds a certain threshold value.
3572 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3573 value. Instead, the threshold value will be adjusted for each individual
3575 In general, smaller parameters result in stronger compression, and vice versa.
3576 Values below 3.0 are not recommended, because audible distortion may appear.
3579 Set the target threshold value. This specifies the lowest permissible
3580 magnitude level for the audio input which will be normalized.
3581 If input frame volume is above this value frame will be normalized.
3582 Otherwise frame may not be normalized at all. The default value is set
3583 to 0, which means all input frames will be normalized.
3584 This option is mostly useful if digital noise is not wanted to be amplified.
3587 @subsection Commands
3589 This filter supports the all above options as @ref{commands}.
3593 Make audio easier to listen to on headphones.
3595 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3596 so that when listened to on headphones the stereo image is moved from
3597 inside your head (standard for headphones) to outside and in front of
3598 the listener (standard for speakers).
3604 Apply a two-pole peaking equalisation (EQ) filter. With this
3605 filter, the signal-level at and around a selected frequency can
3606 be increased or decreased, whilst (unlike bandpass and bandreject
3607 filters) that at all other frequencies is unchanged.
3609 In order to produce complex equalisation curves, this filter can
3610 be given several times, each with a different central frequency.
3612 The filter accepts the following options:
3616 Set the filter's central frequency in Hz.
3619 Set method to specify band-width of filter.
3634 Specify the band-width of a filter in width_type units.
3637 Set the required gain or attenuation in dB.
3638 Beware of clipping when using a positive gain.
3641 How much to use filtered signal in output. Default is 1.
3642 Range is between 0 and 1.
3645 Specify which channels to filter, by default all available are filtered.
3648 Normalize biquad coefficients, by default is disabled.
3649 Enabling it will normalize magnitude response at DC to 0dB.
3652 Set transform type of IIR filter.
3660 @subsection Examples
3663 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3665 equalizer=f=1000:t=h:width=200:g=-10
3669 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3671 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3675 @subsection Commands
3677 This filter supports the following commands:
3680 Change equalizer frequency.
3681 Syntax for the command is : "@var{frequency}"
3684 Change equalizer width_type.
3685 Syntax for the command is : "@var{width_type}"
3688 Change equalizer width.
3689 Syntax for the command is : "@var{width}"
3692 Change equalizer gain.
3693 Syntax for the command is : "@var{gain}"
3696 Change equalizer mix.
3697 Syntax for the command is : "@var{mix}"
3700 @section extrastereo
3702 Linearly increases the difference between left and right channels which
3703 adds some sort of "live" effect to playback.
3705 The filter accepts the following options:
3709 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3710 (average of both channels), with 1.0 sound will be unchanged, with
3711 -1.0 left and right channels will be swapped.
3714 Enable clipping. By default is enabled.
3717 @subsection Commands
3719 This filter supports the all above options as @ref{commands}.
3721 @section firequalizer
3722 Apply FIR Equalization using arbitrary frequency response.
3724 The filter accepts the following option:
3728 Set gain curve equation (in dB). The expression can contain variables:
3731 the evaluated frequency
3735 channel number, set to 0 when multichannels evaluation is disabled
3737 channel id, see libavutil/channel_layout.h, set to the first channel id when
3738 multichannels evaluation is disabled
3742 channel_layout, see libavutil/channel_layout.h
3747 @item gain_interpolate(f)
3748 interpolate gain on frequency f based on gain_entry
3749 @item cubic_interpolate(f)
3750 same as gain_interpolate, but smoother
3752 This option is also available as command. Default is @code{gain_interpolate(f)}.
3755 Set gain entry for gain_interpolate function. The expression can
3759 store gain entry at frequency f with value g
3761 This option is also available as command.
3764 Set filter delay in seconds. Higher value means more accurate.
3765 Default is @code{0.01}.
3768 Set filter accuracy in Hz. Lower value means more accurate.
3769 Default is @code{5}.
3772 Set window function. Acceptable values are:
3775 rectangular window, useful when gain curve is already smooth
3777 hann window (default)
3783 3-terms continuous 1st derivative nuttall window
3785 minimum 3-terms discontinuous nuttall window
3787 4-terms continuous 1st derivative nuttall window
3789 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3791 blackman-harris window
3797 If enabled, use fixed number of audio samples. This improves speed when
3798 filtering with large delay. Default is disabled.
3801 Enable multichannels evaluation on gain. Default is disabled.
3804 Enable zero phase mode by subtracting timestamp to compensate delay.
3805 Default is disabled.
3808 Set scale used by gain. Acceptable values are:
3811 linear frequency, linear gain
3813 linear frequency, logarithmic (in dB) gain (default)
3815 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3817 logarithmic frequency, logarithmic gain
3821 Set file for dumping, suitable for gnuplot.
3824 Set scale for dumpfile. Acceptable values are same with scale option.
3828 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3829 Default is disabled.
3832 Enable minimum phase impulse response. Default is disabled.
3835 @subsection Examples
3840 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3843 lowpass at 1000 Hz with gain_entry:
3845 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3848 custom equalization:
3850 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3853 higher delay with zero phase to compensate delay:
3855 firequalizer=delay=0.1:fixed=on:zero_phase=on
3858 lowpass on left channel, highpass on right channel:
3860 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
3861 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
3866 Apply a flanging effect to the audio.
3868 The filter accepts the following options:
3872 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
3875 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
3878 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
3882 Set percentage of delayed signal mixed with original. Range from 0 to 100.
3883 Default value is 71.
3886 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
3889 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
3890 Default value is @var{sinusoidal}.
3893 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
3894 Default value is 25.
3897 Set delay-line interpolation, @var{linear} or @var{quadratic}.
3898 Default is @var{linear}.
3902 Apply Haas effect to audio.
3904 Note that this makes most sense to apply on mono signals.
3905 With this filter applied to mono signals it give some directionality and
3906 stretches its stereo image.
3908 The filter accepts the following options:
3912 Set input level. By default is @var{1}, or 0dB
3915 Set output level. By default is @var{1}, or 0dB.
3918 Set gain applied to side part of signal. By default is @var{1}.
3921 Set kind of middle source. Can be one of the following:
3931 Pick middle part signal of stereo image.
3934 Pick side part signal of stereo image.
3938 Change middle phase. By default is disabled.
3941 Set left channel delay. By default is @var{2.05} milliseconds.
3944 Set left channel balance. By default is @var{-1}.
3947 Set left channel gain. By default is @var{1}.
3950 Change left phase. By default is disabled.
3953 Set right channel delay. By defaults is @var{2.12} milliseconds.
3956 Set right channel balance. By default is @var{1}.
3959 Set right channel gain. By default is @var{1}.
3962 Change right phase. By default is enabled.
3967 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3968 embedded HDCD codes is expanded into a 20-bit PCM stream.
3970 The filter supports the Peak Extend and Low-level Gain Adjustment features
3971 of HDCD, and detects the Transient Filter flag.
3974 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3977 When using the filter with wav, note the default encoding for wav is 16-bit,
3978 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3979 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3981 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3982 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3985 The filter accepts the following options:
3988 @item disable_autoconvert
3989 Disable any automatic format conversion or resampling in the filter graph.
3991 @item process_stereo
3992 Process the stereo channels together. If target_gain does not match between
3993 channels, consider it invalid and use the last valid target_gain.
3996 Set the code detect timer period in ms.
3999 Always extend peaks above -3dBFS even if PE isn't signaled.
4002 Replace audio with a solid tone and adjust the amplitude to signal some
4003 specific aspect of the decoding process. The output file can be loaded in
4004 an audio editor alongside the original to aid analysis.
4006 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4013 Gain adjustment level at each sample
4015 Samples where peak extend occurs
4017 Samples where the code detect timer is active
4019 Samples where the target gain does not match between channels
4025 Apply head-related transfer functions (HRTFs) to create virtual
4026 loudspeakers around the user for binaural listening via headphones.
4027 The HRIRs are provided via additional streams, for each channel
4028 one stereo input stream is needed.
4030 The filter accepts the following options:
4034 Set mapping of input streams for convolution.
4035 The argument is a '|'-separated list of channel names in order as they
4036 are given as additional stream inputs for filter.
4037 This also specify number of input streams. Number of input streams
4038 must be not less than number of channels in first stream plus one.
4041 Set gain applied to audio. Value is in dB. Default is 0.
4044 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4045 processing audio in time domain which is slow.
4046 @var{freq} is processing audio in frequency domain which is fast.
4047 Default is @var{freq}.
4050 Set custom gain for LFE channels. Value is in dB. Default is 0.
4053 Set size of frame in number of samples which will be processed at once.
4054 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4057 Set format of hrir stream.
4058 Default value is @var{stereo}. Alternative value is @var{multich}.
4059 If value is set to @var{stereo}, number of additional streams should
4060 be greater or equal to number of input channels in first input stream.
4061 Also each additional stream should have stereo number of channels.
4062 If value is set to @var{multich}, number of additional streams should
4063 be exactly one. Also number of input channels of additional stream
4064 should be equal or greater than twice number of channels of first input
4068 @subsection Examples
4072 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4073 each amovie filter use stereo file with IR coefficients as input.
4074 The files give coefficients for each position of virtual loudspeaker:
4077 -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"
4082 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4083 but now in @var{multich} @var{hrir} format.
4085 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"
4092 Apply a high-pass filter with 3dB point frequency.
4093 The filter can be either single-pole, or double-pole (the default).
4094 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4096 The filter accepts the following options:
4100 Set frequency in Hz. Default is 3000.
4103 Set number of poles. Default is 2.
4106 Set method to specify band-width of filter.
4121 Specify the band-width of a filter in width_type units.
4122 Applies only to double-pole filter.
4123 The default is 0.707q and gives a Butterworth response.
4126 How much to use filtered signal in output. Default is 1.
4127 Range is between 0 and 1.
4130 Specify which channels to filter, by default all available are filtered.
4133 Normalize biquad coefficients, by default is disabled.
4134 Enabling it will normalize magnitude response at DC to 0dB.
4137 Set transform type of IIR filter.
4145 @subsection Commands
4147 This filter supports the following commands:
4150 Change highpass frequency.
4151 Syntax for the command is : "@var{frequency}"
4154 Change highpass width_type.
4155 Syntax for the command is : "@var{width_type}"
4158 Change highpass width.
4159 Syntax for the command is : "@var{width}"
4162 Change highpass mix.
4163 Syntax for the command is : "@var{mix}"
4168 Join multiple input streams into one multi-channel stream.
4170 It accepts the following parameters:
4174 The number of input streams. It defaults to 2.
4176 @item channel_layout
4177 The desired output channel layout. It defaults to stereo.
4180 Map channels from inputs to output. The argument is a '|'-separated list of
4181 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4182 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4183 can be either the name of the input channel (e.g. FL for front left) or its
4184 index in the specified input stream. @var{out_channel} is the name of the output
4188 The filter will attempt to guess the mappings when they are not specified
4189 explicitly. It does so by first trying to find an unused matching input channel
4190 and if that fails it picks the first unused input channel.
4192 Join 3 inputs (with properly set channel layouts):
4194 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4197 Build a 5.1 output from 6 single-channel streams:
4199 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4200 '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'
4206 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4208 To enable compilation of this filter you need to configure FFmpeg with
4209 @code{--enable-ladspa}.
4213 Specifies the name of LADSPA plugin library to load. If the environment
4214 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4215 each one of the directories specified by the colon separated list in
4216 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4217 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4218 @file{/usr/lib/ladspa/}.
4221 Specifies the plugin within the library. Some libraries contain only
4222 one plugin, but others contain many of them. If this is not set filter
4223 will list all available plugins within the specified library.
4226 Set the '|' separated list of controls which are zero or more floating point
4227 values that determine the behavior of the loaded plugin (for example delay,
4229 Controls need to be defined using the following syntax:
4230 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4231 @var{valuei} is the value set on the @var{i}-th control.
4232 Alternatively they can be also defined using the following syntax:
4233 @var{value0}|@var{value1}|@var{value2}|..., where
4234 @var{valuei} is the value set on the @var{i}-th control.
4235 If @option{controls} is set to @code{help}, all available controls and
4236 their valid ranges are printed.
4238 @item sample_rate, s
4239 Specify the sample rate, default to 44100. Only used if plugin have
4243 Set the number of samples per channel per each output frame, default
4244 is 1024. Only used if plugin have zero inputs.
4247 Set the minimum duration of the sourced audio. See
4248 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4249 for the accepted syntax.
4250 Note that the resulting duration may be greater than the specified duration,
4251 as the generated audio is always cut at the end of a complete frame.
4252 If not specified, or the expressed duration is negative, the audio is
4253 supposed to be generated forever.
4254 Only used if plugin have zero inputs.
4257 Enable latency compensation, by default is disabled.
4258 Only used if plugin have inputs.
4261 @subsection Examples
4265 List all available plugins within amp (LADSPA example plugin) library:
4271 List all available controls and their valid ranges for @code{vcf_notch}
4272 plugin from @code{VCF} library:
4274 ladspa=f=vcf:p=vcf_notch:c=help
4278 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4281 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4285 Add reverberation to the audio using TAP-plugins
4286 (Tom's Audio Processing plugins):
4288 ladspa=file=tap_reverb:tap_reverb
4292 Generate white noise, with 0.2 amplitude:
4294 ladspa=file=cmt:noise_source_white:c=c0=.2
4298 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4299 @code{C* Audio Plugin Suite} (CAPS) library:
4301 ladspa=file=caps:Click:c=c1=20'
4305 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4307 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4311 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4312 @code{SWH Plugins} collection:
4314 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4318 Attenuate low frequencies using Multiband EQ from Steve Harris
4319 @code{SWH Plugins} collection:
4321 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4325 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4328 ladspa=caps:Narrower
4332 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4334 ladspa=caps:White:.2
4338 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4340 ladspa=caps:Fractal:c=c1=1
4344 Dynamic volume normalization using @code{VLevel} plugin:
4346 ladspa=vlevel-ladspa:vlevel_mono
4350 @subsection Commands
4352 This filter supports the following commands:
4355 Modify the @var{N}-th control value.
4357 If the specified value is not valid, it is ignored and prior one is kept.
4362 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4363 Support for both single pass (livestreams, files) and double pass (files) modes.
4364 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4365 detect true peaks, the audio stream will be upsampled to 192 kHz.
4366 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4368 The filter accepts the following options:
4372 Set integrated loudness target.
4373 Range is -70.0 - -5.0. Default value is -24.0.
4376 Set loudness range target.
4377 Range is 1.0 - 20.0. Default value is 7.0.
4380 Set maximum true peak.
4381 Range is -9.0 - +0.0. Default value is -2.0.
4383 @item measured_I, measured_i
4384 Measured IL of input file.
4385 Range is -99.0 - +0.0.
4387 @item measured_LRA, measured_lra
4388 Measured LRA of input file.
4389 Range is 0.0 - 99.0.
4391 @item measured_TP, measured_tp
4392 Measured true peak of input file.
4393 Range is -99.0 - +99.0.
4395 @item measured_thresh
4396 Measured threshold of input file.
4397 Range is -99.0 - +0.0.
4400 Set offset gain. Gain is applied before the true-peak limiter.
4401 Range is -99.0 - +99.0. Default is +0.0.
4404 Normalize by linearly scaling the source audio.
4405 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4406 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4407 be lower than source LRA and the change in integrated loudness shouldn't
4408 result in a true peak which exceeds the target TP. If any of these
4409 conditions aren't met, normalization mode will revert to @var{dynamic}.
4410 Options are @code{true} or @code{false}. Default is @code{true}.
4413 Treat mono input files as "dual-mono". If a mono file is intended for playback
4414 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4415 If set to @code{true}, this option will compensate for this effect.
4416 Multi-channel input files are not affected by this option.
4417 Options are true or false. Default is false.
4420 Set print format for stats. Options are summary, json, or none.
4421 Default value is none.
4426 Apply a low-pass filter with 3dB point frequency.
4427 The filter can be either single-pole or double-pole (the default).
4428 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4430 The filter accepts the following options:
4434 Set frequency in Hz. Default is 500.
4437 Set number of poles. Default is 2.
4440 Set method to specify band-width of filter.
4455 Specify the band-width of a filter in width_type units.
4456 Applies only to double-pole filter.
4457 The default is 0.707q and gives a Butterworth response.
4460 How much to use filtered signal in output. Default is 1.
4461 Range is between 0 and 1.
4464 Specify which channels to filter, by default all available are filtered.
4467 Normalize biquad coefficients, by default is disabled.
4468 Enabling it will normalize magnitude response at DC to 0dB.
4471 Set transform type of IIR filter.
4479 @subsection Examples
4482 Lowpass only LFE channel, it LFE is not present it does nothing:
4488 @subsection Commands
4490 This filter supports the following commands:
4493 Change lowpass frequency.
4494 Syntax for the command is : "@var{frequency}"
4497 Change lowpass width_type.
4498 Syntax for the command is : "@var{width_type}"
4501 Change lowpass width.
4502 Syntax for the command is : "@var{width}"
4506 Syntax for the command is : "@var{mix}"
4511 Load a LV2 (LADSPA Version 2) plugin.
4513 To enable compilation of this filter you need to configure FFmpeg with
4514 @code{--enable-lv2}.
4518 Specifies the plugin URI. You may need to escape ':'.
4521 Set the '|' separated list of controls which are zero or more floating point
4522 values that determine the behavior of the loaded plugin (for example delay,
4524 If @option{controls} is set to @code{help}, all available controls and
4525 their valid ranges are printed.
4527 @item sample_rate, s
4528 Specify the sample rate, default to 44100. Only used if plugin have
4532 Set the number of samples per channel per each output frame, default
4533 is 1024. Only used if plugin have zero inputs.
4536 Set the minimum duration of the sourced audio. See
4537 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4538 for the accepted syntax.
4539 Note that the resulting duration may be greater than the specified duration,
4540 as the generated audio is always cut at the end of a complete frame.
4541 If not specified, or the expressed duration is negative, the audio is
4542 supposed to be generated forever.
4543 Only used if plugin have zero inputs.
4546 @subsection Examples
4550 Apply bass enhancer plugin from Calf:
4552 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4556 Apply vinyl plugin from Calf:
4558 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4562 Apply bit crusher plugin from ArtyFX:
4564 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4569 Multiband Compress or expand the audio's dynamic range.
4571 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4572 This is akin to the crossover of a loudspeaker, and results in flat frequency
4573 response when absent compander action.
4575 It accepts the following parameters:
4579 This option syntax is:
4580 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4581 For explanation of each item refer to compand filter documentation.
4587 Mix channels with specific gain levels. The filter accepts the output
4588 channel layout followed by a set of channels definitions.
4590 This filter is also designed to efficiently remap the channels of an audio
4593 The filter accepts parameters of the form:
4594 "@var{l}|@var{outdef}|@var{outdef}|..."
4598 output channel layout or number of channels
4601 output channel specification, of the form:
4602 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4605 output channel to define, either a channel name (FL, FR, etc.) or a channel
4606 number (c0, c1, etc.)
4609 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4612 input channel to use, see out_name for details; it is not possible to mix
4613 named and numbered input channels
4616 If the `=' in a channel specification is replaced by `<', then the gains for
4617 that specification will be renormalized so that the total is 1, thus
4618 avoiding clipping noise.
4620 @subsection Mixing examples
4622 For example, if you want to down-mix from stereo to mono, but with a bigger
4623 factor for the left channel:
4625 pan=1c|c0=0.9*c0+0.1*c1
4628 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4629 7-channels surround:
4631 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4634 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4635 that should be preferred (see "-ac" option) unless you have very specific
4638 @subsection Remapping examples
4640 The channel remapping will be effective if, and only if:
4643 @item gain coefficients are zeroes or ones,
4644 @item only one input per channel output,
4647 If all these conditions are satisfied, the filter will notify the user ("Pure
4648 channel mapping detected"), and use an optimized and lossless method to do the
4651 For example, if you have a 5.1 source and want a stereo audio stream by
4652 dropping the extra channels:
4654 pan="stereo| c0=FL | c1=FR"
4657 Given the same source, you can also switch front left and front right channels
4658 and keep the input channel layout:
4660 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4663 If the input is a stereo audio stream, you can mute the front left channel (and
4664 still keep the stereo channel layout) with:
4669 Still with a stereo audio stream input, you can copy the right channel in both
4670 front left and right:
4672 pan="stereo| c0=FR | c1=FR"
4677 ReplayGain scanner filter. This filter takes an audio stream as an input and
4678 outputs it unchanged.
4679 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4683 Convert the audio sample format, sample rate and channel layout. It is
4684 not meant to be used directly.
4687 Apply time-stretching and pitch-shifting with librubberband.
4689 To enable compilation of this filter, you need to configure FFmpeg with
4690 @code{--enable-librubberband}.
4692 The filter accepts the following options:
4696 Set tempo scale factor.
4699 Set pitch scale factor.
4702 Set transients detector.
4703 Possible values are:
4712 Possible values are:
4721 Possible values are:
4728 Set processing window size.
4729 Possible values are:
4738 Possible values are:
4745 Enable formant preservation when shift pitching.
4746 Possible values are:
4754 Possible values are:
4763 Possible values are:
4770 @subsection Commands
4772 This filter supports the following commands:
4775 Change filter tempo scale factor.
4776 Syntax for the command is : "@var{tempo}"
4779 Change filter pitch scale factor.
4780 Syntax for the command is : "@var{pitch}"
4783 @section sidechaincompress
4785 This filter acts like normal compressor but has the ability to compress
4786 detected signal using second input signal.
4787 It needs two input streams and returns one output stream.
4788 First input stream will be processed depending on second stream signal.
4789 The filtered signal then can be filtered with other filters in later stages of
4790 processing. See @ref{pan} and @ref{amerge} filter.
4792 The filter accepts the following options:
4796 Set input gain. Default is 1. Range is between 0.015625 and 64.
4799 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4800 Default is @code{downward}.
4803 If a signal of second stream raises above this level it will affect the gain
4804 reduction of first stream.
4805 By default is 0.125. Range is between 0.00097563 and 1.
4808 Set a ratio about which the signal is reduced. 1:2 means that if the level
4809 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4810 Default is 2. Range is between 1 and 20.
4813 Amount of milliseconds the signal has to rise above the threshold before gain
4814 reduction starts. Default is 20. Range is between 0.01 and 2000.
4817 Amount of milliseconds the signal has to fall below the threshold before
4818 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4821 Set the amount by how much signal will be amplified after processing.
4822 Default is 1. Range is from 1 to 64.
4825 Curve the sharp knee around the threshold to enter gain reduction more softly.
4826 Default is 2.82843. Range is between 1 and 8.
4829 Choose if the @code{average} level between all channels of side-chain stream
4830 or the louder(@code{maximum}) channel of side-chain stream affects the
4831 reduction. Default is @code{average}.
4834 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4835 of @code{rms}. Default is @code{rms} which is mainly smoother.
4838 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4841 How much to use compressed signal in output. Default is 1.
4842 Range is between 0 and 1.
4845 @subsection Commands
4847 This filter supports the all above options as @ref{commands}.
4849 @subsection Examples
4853 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
4854 depending on the signal of 2nd input and later compressed signal to be
4855 merged with 2nd input:
4857 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
4861 @section sidechaingate
4863 A sidechain gate acts like a normal (wideband) gate but has the ability to
4864 filter the detected signal before sending it to the gain reduction stage.
4865 Normally a gate uses the full range signal to detect a level above the
4867 For example: If you cut all lower frequencies from your sidechain signal
4868 the gate will decrease the volume of your track only if not enough highs
4869 appear. With this technique you are able to reduce the resonation of a
4870 natural drum or remove "rumbling" of muted strokes from a heavily distorted
4872 It needs two input streams and returns one output stream.
4873 First input stream will be processed depending on second stream signal.
4875 The filter accepts the following options:
4879 Set input level before filtering.
4880 Default is 1. Allowed range is from 0.015625 to 64.
4883 Set the mode of operation. Can be @code{upward} or @code{downward}.
4884 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
4885 will be amplified, expanding dynamic range in upward direction.
4886 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
4889 Set the level of gain reduction when the signal is below the threshold.
4890 Default is 0.06125. Allowed range is from 0 to 1.
4891 Setting this to 0 disables reduction and then filter behaves like expander.
4894 If a signal rises above this level the gain reduction is released.
4895 Default is 0.125. Allowed range is from 0 to 1.
4898 Set a ratio about which the signal is reduced.
4899 Default is 2. Allowed range is from 1 to 9000.
4902 Amount of milliseconds the signal has to rise above the threshold before gain
4904 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
4907 Amount of milliseconds the signal has to fall below the threshold before the
4908 reduction is increased again. Default is 250 milliseconds.
4909 Allowed range is from 0.01 to 9000.
4912 Set amount of amplification of signal after processing.
4913 Default is 1. Allowed range is from 1 to 64.
4916 Curve the sharp knee around the threshold to enter gain reduction more softly.
4917 Default is 2.828427125. Allowed range is from 1 to 8.
4920 Choose if exact signal should be taken for detection or an RMS like one.
4921 Default is rms. Can be peak or rms.
4924 Choose if the average level between all channels or the louder channel affects
4926 Default is average. Can be average or maximum.
4929 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
4932 @section silencedetect
4934 Detect silence in an audio stream.
4936 This filter logs a message when it detects that the input audio volume is less
4937 or equal to a noise tolerance value for a duration greater or equal to the
4938 minimum detected noise duration.
4940 The printed times and duration are expressed in seconds. The
4941 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
4942 is set on the first frame whose timestamp equals or exceeds the detection
4943 duration and it contains the timestamp of the first frame of the silence.
4945 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
4946 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
4947 keys are set on the first frame after the silence. If @option{mono} is
4948 enabled, and each channel is evaluated separately, the @code{.X}
4949 suffixed keys are used, and @code{X} corresponds to the channel number.
4951 The filter accepts the following options:
4955 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4956 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4959 Set silence duration until notification (default is 2 seconds). See
4960 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4961 for the accepted syntax.
4964 Process each channel separately, instead of combined. By default is disabled.
4967 @subsection Examples
4971 Detect 5 seconds of silence with -50dB noise tolerance:
4973 silencedetect=n=-50dB:d=5
4977 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4978 tolerance in @file{silence.mp3}:
4980 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4984 @section silenceremove
4986 Remove silence from the beginning, middle or end of the audio.
4988 The filter accepts the following options:
4992 This value is used to indicate if audio should be trimmed at beginning of
4993 the audio. A value of zero indicates no silence should be trimmed from the
4994 beginning. When specifying a non-zero value, it trims audio up until it
4995 finds non-silence. Normally, when trimming silence from beginning of audio
4996 the @var{start_periods} will be @code{1} but it can be increased to higher
4997 values to trim all audio up to specific count of non-silence periods.
4998 Default value is @code{0}.
5000 @item start_duration
5001 Specify the amount of time that non-silence must be detected before it stops
5002 trimming audio. By increasing the duration, bursts of noises can be treated
5003 as silence and trimmed off. Default value is @code{0}.
5005 @item start_threshold
5006 This indicates what sample value should be treated as silence. For digital
5007 audio, a value of @code{0} may be fine but for audio recorded from analog,
5008 you may wish to increase the value to account for background noise.
5009 Can be specified in dB (in case "dB" is appended to the specified value)
5010 or amplitude ratio. Default value is @code{0}.
5013 Specify max duration of silence at beginning that will be kept after
5014 trimming. Default is 0, which is equal to trimming all samples detected
5018 Specify mode of detection of silence end in start of multi-channel audio.
5019 Can be @var{any} or @var{all}. Default is @var{any}.
5020 With @var{any}, any sample that is detected as non-silence will cause
5021 stopped trimming of silence.
5022 With @var{all}, only if all channels are detected as non-silence will cause
5023 stopped trimming of silence.
5026 Set the count for trimming silence from the end of audio.
5027 To remove silence from the middle of a file, specify a @var{stop_periods}
5028 that is negative. This value is then treated as a positive value and is
5029 used to indicate the effect should restart processing as specified by
5030 @var{start_periods}, making it suitable for removing periods of silence
5031 in the middle of the audio.
5032 Default value is @code{0}.
5035 Specify a duration of silence that must exist before audio is not copied any
5036 more. By specifying a higher duration, silence that is wanted can be left in
5038 Default value is @code{0}.
5040 @item stop_threshold
5041 This is the same as @option{start_threshold} but for trimming silence from
5043 Can be specified in dB (in case "dB" is appended to the specified value)
5044 or amplitude ratio. Default value is @code{0}.
5047 Specify max duration of silence at end that will be kept after
5048 trimming. Default is 0, which is equal to trimming all samples detected
5052 Specify mode of detection of silence start in end of multi-channel audio.
5053 Can be @var{any} or @var{all}. Default is @var{any}.
5054 With @var{any}, any sample that is detected as non-silence will cause
5055 stopped trimming of silence.
5056 With @var{all}, only if all channels are detected as non-silence will cause
5057 stopped trimming of silence.
5060 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5061 and works better with digital silence which is exactly 0.
5062 Default value is @code{rms}.
5065 Set duration in number of seconds used to calculate size of window in number
5066 of samples for detecting silence.
5067 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5070 @subsection Examples
5074 The following example shows how this filter can be used to start a recording
5075 that does not contain the delay at the start which usually occurs between
5076 pressing the record button and the start of the performance:
5078 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5082 Trim all silence encountered from beginning to end where there is more than 1
5083 second of silence in audio:
5085 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5089 Trim all digital silence samples, using peak detection, from beginning to end
5090 where there is more than 0 samples of digital silence in audio and digital
5091 silence is detected in all channels at same positions in stream:
5093 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5099 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5100 loudspeakers around the user for binaural listening via headphones (audio
5101 formats up to 9 channels supported).
5102 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5103 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5104 Austrian Academy of Sciences.
5106 To enable compilation of this filter you need to configure FFmpeg with
5107 @code{--enable-libmysofa}.
5109 The filter accepts the following options:
5113 Set the SOFA file used for rendering.
5116 Set gain applied to audio. Value is in dB. Default is 0.
5119 Set rotation of virtual loudspeakers in deg. Default is 0.
5122 Set elevation of virtual speakers in deg. Default is 0.
5125 Set distance in meters between loudspeakers and the listener with near-field
5126 HRTFs. Default is 1.
5129 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5130 processing audio in time domain which is slow.
5131 @var{freq} is processing audio in frequency domain which is fast.
5132 Default is @var{freq}.
5135 Set custom positions of virtual loudspeakers. Syntax for this option is:
5136 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5137 Each virtual loudspeaker is described with short channel name following with
5138 azimuth and elevation in degrees.
5139 Each virtual loudspeaker description is separated by '|'.
5140 For example to override front left and front right channel positions use:
5141 'speakers=FL 45 15|FR 345 15'.
5142 Descriptions with unrecognised channel names are ignored.
5145 Set custom gain for LFE channels. Value is in dB. Default is 0.
5148 Set custom frame size in number of samples. Default is 1024.
5149 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5150 is set to @var{freq}.
5153 Should all IRs be normalized upon importing SOFA file.
5154 By default is enabled.
5157 Should nearest IRs be interpolated with neighbor IRs if exact position
5158 does not match. By default is disabled.
5161 Minphase all IRs upon loading of SOFA file. By default is disabled.
5164 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5167 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5170 @subsection Examples
5174 Using ClubFritz6 sofa file:
5176 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5180 Using ClubFritz12 sofa file and bigger radius with small rotation:
5182 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5186 Similar as above but with custom speaker positions for front left, front right, back left and back right
5187 and also with custom gain:
5189 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5193 @section stereotools
5195 This filter has some handy utilities to manage stereo signals, for converting
5196 M/S stereo recordings to L/R signal while having control over the parameters
5197 or spreading the stereo image of master track.
5199 The filter accepts the following options:
5203 Set input level before filtering for both channels. Defaults is 1.
5204 Allowed range is from 0.015625 to 64.
5207 Set output level after filtering for both channels. Defaults is 1.
5208 Allowed range is from 0.015625 to 64.
5211 Set input balance between both channels. Default is 0.
5212 Allowed range is from -1 to 1.
5215 Set output balance between both channels. Default is 0.
5216 Allowed range is from -1 to 1.
5219 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5220 clipping. Disabled by default.
5223 Mute the left channel. Disabled by default.
5226 Mute the right channel. Disabled by default.
5229 Change the phase of the left channel. Disabled by default.
5232 Change the phase of the right channel. Disabled by default.
5235 Set stereo mode. Available values are:
5239 Left/Right to Left/Right, this is default.
5242 Left/Right to Mid/Side.
5245 Mid/Side to Left/Right.
5248 Left/Right to Left/Left.
5251 Left/Right to Right/Right.
5254 Left/Right to Left + Right.
5257 Left/Right to Right/Left.
5260 Mid/Side to Left/Left.
5263 Mid/Side to Right/Right.
5267 Set level of side signal. Default is 1.
5268 Allowed range is from 0.015625 to 64.
5271 Set balance of side signal. Default is 0.
5272 Allowed range is from -1 to 1.
5275 Set level of the middle signal. Default is 1.
5276 Allowed range is from 0.015625 to 64.
5279 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5282 Set stereo base between mono and inversed channels. Default is 0.
5283 Allowed range is from -1 to 1.
5286 Set delay in milliseconds how much to delay left from right channel and
5287 vice versa. Default is 0. Allowed range is from -20 to 20.
5290 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5293 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5295 @item bmode_in, bmode_out
5296 Set balance mode for balance_in/balance_out option.
5298 Can be one of the following:
5302 Classic balance mode. Attenuate one channel at time.
5303 Gain is raised up to 1.
5306 Similar as classic mode above but gain is raised up to 2.
5309 Equal power distribution, from -6dB to +6dB range.
5313 @subsection Examples
5317 Apply karaoke like effect:
5319 stereotools=mlev=0.015625
5323 Convert M/S signal to L/R:
5325 "stereotools=mode=ms>lr"
5329 @section stereowiden
5331 This filter enhance the stereo effect by suppressing signal common to both
5332 channels and by delaying the signal of left into right and vice versa,
5333 thereby widening the stereo effect.
5335 The filter accepts the following options:
5339 Time in milliseconds of the delay of left signal into right and vice versa.
5340 Default is 20 milliseconds.
5343 Amount of gain in delayed signal into right and vice versa. Gives a delay
5344 effect of left signal in right output and vice versa which gives widening
5345 effect. Default is 0.3.
5348 Cross feed of left into right with inverted phase. This helps in suppressing
5349 the mono. If the value is 1 it will cancel all the signal common to both
5350 channels. Default is 0.3.
5353 Set level of input signal of original channel. Default is 0.8.
5356 @subsection Commands
5358 This filter supports the all above options except @code{delay} as @ref{commands}.
5360 @section superequalizer
5361 Apply 18 band equalizer.
5363 The filter accepts the following options:
5370 Set 131Hz band gain.
5372 Set 185Hz band gain.
5374 Set 262Hz band gain.
5376 Set 370Hz band gain.
5378 Set 523Hz band gain.
5380 Set 740Hz band gain.
5382 Set 1047Hz band gain.
5384 Set 1480Hz band gain.
5386 Set 2093Hz band gain.
5388 Set 2960Hz band gain.
5390 Set 4186Hz band gain.
5392 Set 5920Hz band gain.
5394 Set 8372Hz band gain.
5396 Set 11840Hz band gain.
5398 Set 16744Hz band gain.
5400 Set 20000Hz band gain.
5404 Apply audio surround upmix filter.
5406 This filter allows to produce multichannel output from audio stream.
5408 The filter accepts the following options:
5412 Set output channel layout. By default, this is @var{5.1}.
5414 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5415 for the required syntax.
5418 Set input channel layout. By default, this is @var{stereo}.
5420 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5421 for the required syntax.
5424 Set input volume level. By default, this is @var{1}.
5427 Set output volume level. By default, this is @var{1}.
5430 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5433 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5436 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5439 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5440 In @var{add} mode, LFE channel is created from input audio and added to output.
5441 In @var{sub} mode, LFE channel is created from input audio and added to output but
5442 also all non-LFE output channels are subtracted with output LFE channel.
5445 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5446 Default is @var{90}.
5449 Set front center input volume. By default, this is @var{1}.
5452 Set front center output volume. By default, this is @var{1}.
5455 Set front left input volume. By default, this is @var{1}.
5458 Set front left output volume. By default, this is @var{1}.
5461 Set front right input volume. By default, this is @var{1}.
5464 Set front right output volume. By default, this is @var{1}.
5467 Set side left input volume. By default, this is @var{1}.
5470 Set side left output volume. By default, this is @var{1}.
5473 Set side right input volume. By default, this is @var{1}.
5476 Set side right output volume. By default, this is @var{1}.
5479 Set back left input volume. By default, this is @var{1}.
5482 Set back left output volume. By default, this is @var{1}.
5485 Set back right input volume. By default, this is @var{1}.
5488 Set back right output volume. By default, this is @var{1}.
5491 Set back center input volume. By default, this is @var{1}.
5494 Set back center output volume. By default, this is @var{1}.
5497 Set LFE input volume. By default, this is @var{1}.
5500 Set LFE output volume. By default, this is @var{1}.
5503 Set spread usage of stereo image across X axis for all channels.
5506 Set spread usage of stereo image across Y axis for all channels.
5508 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5509 Set spread usage of stereo image across X axis for each channel.
5511 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5512 Set spread usage of stereo image across Y axis for each channel.
5515 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5518 Set window function.
5520 It accepts the following values:
5543 Default is @code{hann}.
5546 Set window overlap. If set to 1, the recommended overlap for selected
5547 window function will be picked. Default is @code{0.5}.
5550 @section treble, highshelf
5552 Boost or cut treble (upper) frequencies of the audio using a two-pole
5553 shelving filter with a response similar to that of a standard
5554 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5556 The filter accepts the following options:
5560 Give the gain at whichever is the lower of ~22 kHz and the
5561 Nyquist frequency. Its useful range is about -20 (for a large cut)
5562 to +20 (for a large boost). Beware of clipping when using a positive gain.
5565 Set the filter's central frequency and so can be used
5566 to extend or reduce the frequency range to be boosted or cut.
5567 The default value is @code{3000} Hz.
5570 Set method to specify band-width of filter.
5585 Determine how steep is the filter's shelf transition.
5588 How much to use filtered signal in output. Default is 1.
5589 Range is between 0 and 1.
5592 Specify which channels to filter, by default all available are filtered.
5595 Normalize biquad coefficients, by default is disabled.
5596 Enabling it will normalize magnitude response at DC to 0dB.
5599 Set transform type of IIR filter.
5607 @subsection Commands
5609 This filter supports the following commands:
5612 Change treble frequency.
5613 Syntax for the command is : "@var{frequency}"
5616 Change treble width_type.
5617 Syntax for the command is : "@var{width_type}"
5620 Change treble width.
5621 Syntax for the command is : "@var{width}"
5625 Syntax for the command is : "@var{gain}"
5629 Syntax for the command is : "@var{mix}"
5634 Sinusoidal amplitude modulation.
5636 The filter accepts the following options:
5640 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5641 (20 Hz or lower) will result in a tremolo effect.
5642 This filter may also be used as a ring modulator by specifying
5643 a modulation frequency higher than 20 Hz.
5644 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5647 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5648 Default value is 0.5.
5653 Sinusoidal phase modulation.
5655 The filter accepts the following options:
5659 Modulation frequency in Hertz.
5660 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5663 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5664 Default value is 0.5.
5669 Adjust the input audio volume.
5671 It accepts the following parameters:
5675 Set audio volume expression.
5677 Output values are clipped to the maximum value.
5679 The output audio volume is given by the relation:
5681 @var{output_volume} = @var{volume} * @var{input_volume}
5684 The default value for @var{volume} is "1.0".
5687 This parameter represents the mathematical precision.
5689 It determines which input sample formats will be allowed, which affects the
5690 precision of the volume scaling.
5694 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5696 32-bit floating-point; this limits input sample format to FLT. (default)
5698 64-bit floating-point; this limits input sample format to DBL.
5702 Choose the behaviour on encountering ReplayGain side data in input frames.
5706 Remove ReplayGain side data, ignoring its contents (the default).
5709 Ignore ReplayGain side data, but leave it in the frame.
5712 Prefer the track gain, if present.
5715 Prefer the album gain, if present.
5718 @item replaygain_preamp
5719 Pre-amplification gain in dB to apply to the selected replaygain gain.
5721 Default value for @var{replaygain_preamp} is 0.0.
5723 @item replaygain_noclip
5724 Prevent clipping by limiting the gain applied.
5726 Default value for @var{replaygain_noclip} is 1.
5729 Set when the volume expression is evaluated.
5731 It accepts the following values:
5734 only evaluate expression once during the filter initialization, or
5735 when the @samp{volume} command is sent
5738 evaluate expression for each incoming frame
5741 Default value is @samp{once}.
5744 The volume expression can contain the following parameters.
5748 frame number (starting at zero)
5751 @item nb_consumed_samples
5752 number of samples consumed by the filter
5754 number of samples in the current frame
5756 original frame position in the file
5762 PTS at start of stream
5764 time at start of stream
5770 last set volume value
5773 Note that when @option{eval} is set to @samp{once} only the
5774 @var{sample_rate} and @var{tb} variables are available, all other
5775 variables will evaluate to NAN.
5777 @subsection Commands
5779 This filter supports the following commands:
5782 Modify the volume expression.
5783 The command accepts the same syntax of the corresponding option.
5785 If the specified expression is not valid, it is kept at its current
5789 @subsection Examples
5793 Halve the input audio volume:
5797 volume=volume=-6.0206dB
5800 In all the above example the named key for @option{volume} can be
5801 omitted, for example like in:
5807 Increase input audio power by 6 decibels using fixed-point precision:
5809 volume=volume=6dB:precision=fixed
5813 Fade volume after time 10 with an annihilation period of 5 seconds:
5815 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
5819 @section volumedetect
5821 Detect the volume of the input video.
5823 The filter has no parameters. The input is not modified. Statistics about
5824 the volume will be printed in the log when the input stream end is reached.
5826 In particular it will show the mean volume (root mean square), maximum
5827 volume (on a per-sample basis), and the beginning of a histogram of the
5828 registered volume values (from the maximum value to a cumulated 1/1000 of
5831 All volumes are in decibels relative to the maximum PCM value.
5833 @subsection Examples
5835 Here is an excerpt of the output:
5837 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
5838 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
5839 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
5840 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5841 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5842 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5843 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5844 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5845 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5851 The mean square energy is approximately -27 dB, or 10^-2.7.
5853 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5855 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5858 In other words, raising the volume by +4 dB does not cause any clipping,
5859 raising it by +5 dB causes clipping for 6 samples, etc.
5861 @c man end AUDIO FILTERS
5863 @chapter Audio Sources
5864 @c man begin AUDIO SOURCES
5866 Below is a description of the currently available audio sources.
5870 Buffer audio frames, and make them available to the filter chain.
5872 This source is mainly intended for a programmatic use, in particular
5873 through the interface defined in @file{libavfilter/buffersrc.h}.
5875 It accepts the following parameters:
5879 The timebase which will be used for timestamps of submitted frames. It must be
5880 either a floating-point number or in @var{numerator}/@var{denominator} form.
5883 The sample rate of the incoming audio buffers.
5886 The sample format of the incoming audio buffers.
5887 Either a sample format name or its corresponding integer representation from
5888 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5890 @item channel_layout
5891 The channel layout of the incoming audio buffers.
5892 Either a channel layout name from channel_layout_map in
5893 @file{libavutil/channel_layout.c} or its corresponding integer representation
5894 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5897 The number of channels of the incoming audio buffers.
5898 If both @var{channels} and @var{channel_layout} are specified, then they
5903 @subsection Examples
5906 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5909 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5910 Since the sample format with name "s16p" corresponds to the number
5911 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5914 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5919 Generate an audio signal specified by an expression.
5921 This source accepts in input one or more expressions (one for each
5922 channel), which are evaluated and used to generate a corresponding
5925 This source accepts the following options:
5929 Set the '|'-separated expressions list for each separate channel. In case the
5930 @option{channel_layout} option is not specified, the selected channel layout
5931 depends on the number of provided expressions. Otherwise the last
5932 specified expression is applied to the remaining output channels.
5934 @item channel_layout, c
5935 Set the channel layout. The number of channels in the specified layout
5936 must be equal to the number of specified expressions.
5939 Set the minimum duration of the sourced audio. See
5940 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5941 for the accepted syntax.
5942 Note that the resulting duration may be greater than the specified
5943 duration, as the generated audio is always cut at the end of a
5946 If not specified, or the expressed duration is negative, the audio is
5947 supposed to be generated forever.
5950 Set the number of samples per channel per each output frame,
5953 @item sample_rate, s
5954 Specify the sample rate, default to 44100.
5957 Each expression in @var{exprs} can contain the following constants:
5961 number of the evaluated sample, starting from 0
5964 time of the evaluated sample expressed in seconds, starting from 0
5971 @subsection Examples
5981 Generate a sin signal with frequency of 440 Hz, set sample rate to
5984 aevalsrc="sin(440*2*PI*t):s=8000"
5988 Generate a two channels signal, specify the channel layout (Front
5989 Center + Back Center) explicitly:
5991 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5995 Generate white noise:
5997 aevalsrc="-2+random(0)"
6001 Generate an amplitude modulated signal:
6003 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6007 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6009 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6016 Generate a FIR coefficients using frequency sampling method.
6018 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6020 The filter accepts the following options:
6024 Set number of filter coefficents in output audio stream.
6025 Default value is 1025.
6028 Set frequency points from where magnitude and phase are set.
6029 This must be in non decreasing order, and first element must be 0, while last element
6030 must be 1. Elements are separated by white spaces.
6033 Set magnitude value for every frequency point set by @option{frequency}.
6034 Number of values must be same as number of frequency points.
6035 Values are separated by white spaces.
6038 Set phase value for every frequency point set by @option{frequency}.
6039 Number of values must be same as number of frequency points.
6040 Values are separated by white spaces.
6042 @item sample_rate, r
6043 Set sample rate, default is 44100.
6046 Set number of samples per each frame. Default is 1024.
6049 Set window function. Default is blackman.
6054 The null audio source, return unprocessed audio frames. It is mainly useful
6055 as a template and to be employed in analysis / debugging tools, or as
6056 the source for filters which ignore the input data (for example the sox
6059 This source accepts the following options:
6063 @item channel_layout, cl
6065 Specifies the channel layout, and can be either an integer or a string
6066 representing a channel layout. The default value of @var{channel_layout}
6069 Check the channel_layout_map definition in
6070 @file{libavutil/channel_layout.c} for the mapping between strings and
6071 channel layout values.
6073 @item sample_rate, r
6074 Specifies the sample rate, and defaults to 44100.
6077 Set the number of samples per requested frames.
6080 Set the duration of the sourced audio. See
6081 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6082 for the accepted syntax.
6084 If not specified, or the expressed duration is negative, the audio is
6085 supposed to be generated forever.
6088 @subsection Examples
6092 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6094 anullsrc=r=48000:cl=4
6098 Do the same operation with a more obvious syntax:
6100 anullsrc=r=48000:cl=mono
6104 All the parameters need to be explicitly defined.
6108 Synthesize a voice utterance using the libflite library.
6110 To enable compilation of this filter you need to configure FFmpeg with
6111 @code{--enable-libflite}.
6113 Note that versions of the flite library prior to 2.0 are not thread-safe.
6115 The filter accepts the following options:
6120 If set to 1, list the names of the available voices and exit
6121 immediately. Default value is 0.
6124 Set the maximum number of samples per frame. Default value is 512.
6127 Set the filename containing the text to speak.
6130 Set the text to speak.
6133 Set the voice to use for the speech synthesis. Default value is
6134 @code{kal}. See also the @var{list_voices} option.
6137 @subsection Examples
6141 Read from file @file{speech.txt}, and synthesize the text using the
6142 standard flite voice:
6144 flite=textfile=speech.txt
6148 Read the specified text selecting the @code{slt} voice:
6150 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6154 Input text to ffmpeg:
6156 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6160 Make @file{ffplay} speak the specified text, using @code{flite} and
6161 the @code{lavfi} device:
6163 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6167 For more information about libflite, check:
6168 @url{http://www.festvox.org/flite/}
6172 Generate a noise audio signal.
6174 The filter accepts the following options:
6177 @item sample_rate, r
6178 Specify the sample rate. Default value is 48000 Hz.
6181 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6185 Specify the duration of the generated audio stream. Not specifying this option
6186 results in noise with an infinite length.
6188 @item color, colour, c
6189 Specify the color of noise. Available noise colors are white, pink, brown,
6190 blue, violet and velvet. Default color is white.
6193 Specify a value used to seed the PRNG.
6196 Set the number of samples per each output frame, default is 1024.
6199 @subsection Examples
6204 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6206 anoisesrc=d=60:c=pink:r=44100:a=0.5
6212 Generate odd-tap Hilbert transform FIR coefficients.
6214 The resulting stream can be used with @ref{afir} filter for phase-shifting
6215 the signal by 90 degrees.
6217 This is used in many matrix coding schemes and for analytic signal generation.
6218 The process is often written as a multiplication by i (or j), the imaginary unit.
6220 The filter accepts the following options:
6224 @item sample_rate, s
6225 Set sample rate, default is 44100.
6228 Set length of FIR filter, default is 22051.
6231 Set number of samples per each frame.
6234 Set window function to be used when generating FIR coefficients.
6239 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6241 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6243 The filter accepts the following options:
6246 @item sample_rate, r
6247 Set sample rate, default is 44100.
6250 Set number of samples per each frame. Default is 1024.
6253 Set high-pass frequency. Default is 0.
6256 Set low-pass frequency. Default is 0.
6257 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6258 is higher than 0 then filter will create band-pass filter coefficients,
6259 otherwise band-reject filter coefficients.
6262 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6265 Set Kaiser window beta.
6268 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6271 Enable rounding, by default is disabled.
6274 Set number of taps for high-pass filter.
6277 Set number of taps for low-pass filter.
6282 Generate an audio signal made of a sine wave with amplitude 1/8.
6284 The audio signal is bit-exact.
6286 The filter accepts the following options:
6291 Set the carrier frequency. Default is 440 Hz.
6293 @item beep_factor, b
6294 Enable a periodic beep every second with frequency @var{beep_factor} times
6295 the carrier frequency. Default is 0, meaning the beep is disabled.
6297 @item sample_rate, r
6298 Specify the sample rate, default is 44100.
6301 Specify the duration of the generated audio stream.
6303 @item samples_per_frame
6304 Set the number of samples per output frame.
6306 The expression can contain the following constants:
6310 The (sequential) number of the output audio frame, starting from 0.
6313 The PTS (Presentation TimeStamp) of the output audio frame,
6314 expressed in @var{TB} units.
6317 The PTS of the output audio frame, expressed in seconds.
6320 The timebase of the output audio frames.
6323 Default is @code{1024}.
6326 @subsection Examples
6331 Generate a simple 440 Hz sine wave:
6337 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6341 sine=frequency=220:beep_factor=4:duration=5
6345 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6348 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6352 @c man end AUDIO SOURCES
6354 @chapter Audio Sinks
6355 @c man begin AUDIO SINKS
6357 Below is a description of the currently available audio sinks.
6359 @section abuffersink
6361 Buffer audio frames, and make them available to the end of filter chain.
6363 This sink is mainly intended for programmatic use, in particular
6364 through the interface defined in @file{libavfilter/buffersink.h}
6365 or the options system.
6367 It accepts a pointer to an AVABufferSinkContext structure, which
6368 defines the incoming buffers' formats, to be passed as the opaque
6369 parameter to @code{avfilter_init_filter} for initialization.
6372 Null audio sink; do absolutely nothing with the input audio. It is
6373 mainly useful as a template and for use in analysis / debugging
6376 @c man end AUDIO SINKS
6378 @chapter Video Filters
6379 @c man begin VIDEO FILTERS
6381 When you configure your FFmpeg build, you can disable any of the
6382 existing filters using @code{--disable-filters}.
6383 The configure output will show the video filters included in your
6386 Below is a description of the currently available video filters.
6390 Mark a region of interest in a video frame.
6392 The frame data is passed through unchanged, but metadata is attached
6393 to the frame indicating regions of interest which can affect the
6394 behaviour of later encoding. Multiple regions can be marked by
6395 applying the filter multiple times.
6399 Region distance in pixels from the left edge of the frame.
6401 Region distance in pixels from the top edge of the frame.
6403 Region width in pixels.
6405 Region height in pixels.
6407 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6408 and may contain the following variables:
6411 Width of the input frame.
6413 Height of the input frame.
6417 Quantisation offset to apply within the region.
6419 This must be a real value in the range -1 to +1. A value of zero
6420 indicates no quality change. A negative value asks for better quality
6421 (less quantisation), while a positive value asks for worse quality
6422 (greater quantisation).
6424 The range is calibrated so that the extreme values indicate the
6425 largest possible offset - if the rest of the frame is encoded with the
6426 worst possible quality, an offset of -1 indicates that this region
6427 should be encoded with the best possible quality anyway. Intermediate
6428 values are then interpolated in some codec-dependent way.
6430 For example, in 10-bit H.264 the quantisation parameter varies between
6431 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6432 this region should be encoded with a QP around one-tenth of the full
6433 range better than the rest of the frame. So, if most of the frame
6434 were to be encoded with a QP of around 30, this region would get a QP
6435 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6436 An extreme value of -1 would indicate that this region should be
6437 encoded with the best possible quality regardless of the treatment of
6438 the rest of the frame - that is, should be encoded at a QP of -12.
6440 If set to true, remove any existing regions of interest marked on the
6441 frame before adding the new one.
6444 @subsection Examples
6448 Mark the centre quarter of the frame as interesting.
6450 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6453 Mark the 100-pixel-wide region on the left edge of the frame as very
6454 uninteresting (to be encoded at much lower quality than the rest of
6457 addroi=0:0:100:ih:+1/5
6461 @section alphaextract
6463 Extract the alpha component from the input as a grayscale video. This
6464 is especially useful with the @var{alphamerge} filter.
6468 Add or replace the alpha component of the primary input with the
6469 grayscale value of a second input. This is intended for use with
6470 @var{alphaextract} to allow the transmission or storage of frame
6471 sequences that have alpha in a format that doesn't support an alpha
6474 For example, to reconstruct full frames from a normal YUV-encoded video
6475 and a separate video created with @var{alphaextract}, you might use:
6477 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6482 Amplify differences between current pixel and pixels of adjacent frames in
6483 same pixel location.
6485 This filter accepts the following options:
6489 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6490 For example radius of 3 will instruct filter to calculate average of 7 frames.
6493 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6496 Set threshold for difference amplification. Any difference greater or equal to
6497 this value will not alter source pixel. Default is 10.
6498 Allowed range is from 0 to 65535.
6501 Set tolerance for difference amplification. Any difference lower to
6502 this value will not alter source pixel. Default is 0.
6503 Allowed range is from 0 to 65535.
6506 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6507 This option controls maximum possible value that will decrease source pixel value.
6510 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6511 This option controls maximum possible value that will increase source pixel value.
6514 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6517 @subsection Commands
6519 This filter supports the following @ref{commands} that corresponds to option of same name:
6531 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6532 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6533 Substation Alpha) subtitles files.
6535 This filter accepts the following option in addition to the common options from
6536 the @ref{subtitles} filter:
6540 Set the shaping engine
6542 Available values are:
6545 The default libass shaping engine, which is the best available.
6547 Fast, font-agnostic shaper that can do only substitutions
6549 Slower shaper using OpenType for substitutions and positioning
6552 The default is @code{auto}.
6556 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6558 The filter accepts the following options:
6562 Set threshold A for 1st plane. Default is 0.02.
6563 Valid range is 0 to 0.3.
6566 Set threshold B for 1st plane. Default is 0.04.
6567 Valid range is 0 to 5.
6570 Set threshold A for 2nd plane. Default is 0.02.
6571 Valid range is 0 to 0.3.
6574 Set threshold B for 2nd plane. Default is 0.04.
6575 Valid range is 0 to 5.
6578 Set threshold A for 3rd plane. Default is 0.02.
6579 Valid range is 0 to 0.3.
6582 Set threshold B for 3rd plane. Default is 0.04.
6583 Valid range is 0 to 5.
6585 Threshold A is designed to react on abrupt changes in the input signal and
6586 threshold B is designed to react on continuous changes in the input signal.
6589 Set number of frames filter will use for averaging. Default is 9. Must be odd
6590 number in range [5, 129].
6593 Set what planes of frame filter will use for averaging. Default is all.
6596 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
6597 Alternatively can be set to @code{s} serial.
6599 Parallel can be faster then serial, while other way around is never true.
6600 Parallel will abort early on first change being greater then thresholds, while serial
6601 will continue processing other side of frames if they are equal or bellow thresholds.
6604 @subsection Commands
6605 This filter supports same @ref{commands} as options except option @code{s}.
6606 The command accepts the same syntax of the corresponding option.
6610 Apply average blur filter.
6612 The filter accepts the following options:
6616 Set horizontal radius size.
6619 Set which planes to filter. By default all planes are filtered.
6622 Set vertical radius size, if zero it will be same as @code{sizeX}.
6623 Default is @code{0}.
6626 @subsection Commands
6627 This filter supports same commands as options.
6628 The command accepts the same syntax of the corresponding option.
6630 If the specified expression is not valid, it is kept at its current
6635 Compute the bounding box for the non-black pixels in the input frame
6638 This filter computes the bounding box containing all the pixels with a
6639 luminance value greater than the minimum allowed value.
6640 The parameters describing the bounding box are printed on the filter
6643 The filter accepts the following option:
6647 Set the minimal luminance value. Default is @code{16}.
6651 Apply bilateral filter, spatial smoothing while preserving edges.
6653 The filter accepts the following options:
6656 Set sigma of gaussian function to calculate spatial weight.
6657 Allowed range is 0 to 512. Default is 0.1.
6660 Set sigma of gaussian function to calculate range weight.
6661 Allowed range is 0 to 1. Default is 0.1.
6664 Set planes to filter. Default is first only.
6667 @section bitplanenoise
6669 Show and measure bit plane noise.
6671 The filter accepts the following options:
6675 Set which plane to analyze. Default is @code{1}.
6678 Filter out noisy pixels from @code{bitplane} set above.
6679 Default is disabled.
6682 @section blackdetect
6684 Detect video intervals that are (almost) completely black. Can be
6685 useful to detect chapter transitions, commercials, or invalid
6688 The filter outputs its detection analysis to both the log as well as
6689 frame metadata. If a black segment of at least the specified minimum
6690 duration is found, a line with the start and end timestamps as well
6691 as duration is printed to the log with level @code{info}. In addition,
6692 a log line with level @code{debug} is printed per frame showing the
6693 black amount detected for that frame.
6695 The filter also attaches metadata to the first frame of a black
6696 segment with key @code{lavfi.black_start} and to the first frame
6697 after the black segment ends with key @code{lavfi.black_end}. The
6698 value is the frame's timestamp. This metadata is added regardless
6699 of the minimum duration specified.
6701 The filter accepts the following options:
6704 @item black_min_duration, d
6705 Set the minimum detected black duration expressed in seconds. It must
6706 be a non-negative floating point number.
6708 Default value is 2.0.
6710 @item picture_black_ratio_th, pic_th
6711 Set the threshold for considering a picture "black".
6712 Express the minimum value for the ratio:
6714 @var{nb_black_pixels} / @var{nb_pixels}
6717 for which a picture is considered black.
6718 Default value is 0.98.
6720 @item pixel_black_th, pix_th
6721 Set the threshold for considering a pixel "black".
6723 The threshold expresses the maximum pixel luminance value for which a
6724 pixel is considered "black". The provided value is scaled according to
6725 the following equation:
6727 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6730 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6731 the input video format, the range is [0-255] for YUV full-range
6732 formats and [16-235] for YUV non full-range formats.
6734 Default value is 0.10.
6737 The following example sets the maximum pixel threshold to the minimum
6738 value, and detects only black intervals of 2 or more seconds:
6740 blackdetect=d=2:pix_th=0.00
6745 Detect frames that are (almost) completely black. Can be useful to
6746 detect chapter transitions or commercials. Output lines consist of
6747 the frame number of the detected frame, the percentage of blackness,
6748 the position in the file if known or -1 and the timestamp in seconds.
6750 In order to display the output lines, you need to set the loglevel at
6751 least to the AV_LOG_INFO value.
6753 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
6754 The value represents the percentage of pixels in the picture that
6755 are below the threshold value.
6757 It accepts the following parameters:
6762 The percentage of the pixels that have to be below the threshold; it defaults to
6765 @item threshold, thresh
6766 The threshold below which a pixel value is considered black; it defaults to
6774 Blend two video frames into each other.
6776 The @code{blend} filter takes two input streams and outputs one
6777 stream, the first input is the "top" layer and second input is
6778 "bottom" layer. By default, the output terminates when the longest input terminates.
6780 The @code{tblend} (time blend) filter takes two consecutive frames
6781 from one single stream, and outputs the result obtained by blending
6782 the new frame on top of the old frame.
6784 A description of the accepted options follows.
6792 Set blend mode for specific pixel component or all pixel components in case
6793 of @var{all_mode}. Default value is @code{normal}.
6795 Available values for component modes are:
6837 Set blend opacity for specific pixel component or all pixel components in case
6838 of @var{all_opacity}. Only used in combination with pixel component blend modes.
6845 Set blend expression for specific pixel component or all pixel components in case
6846 of @var{all_expr}. Note that related mode options will be ignored if those are set.
6848 The expressions can use the following variables:
6852 The sequential number of the filtered frame, starting from @code{0}.
6856 the coordinates of the current sample
6860 the width and height of currently filtered plane
6864 Width and height scale for the plane being filtered. It is the
6865 ratio between the dimensions of the current plane to the luma plane,
6866 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
6867 the luma plane and @code{0.5,0.5} for the chroma planes.
6870 Time of the current frame, expressed in seconds.
6873 Value of pixel component at current location for first video frame (top layer).
6876 Value of pixel component at current location for second video frame (bottom layer).
6880 The @code{blend} filter also supports the @ref{framesync} options.
6882 @subsection Examples
6886 Apply transition from bottom layer to top layer in first 10 seconds:
6888 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
6892 Apply linear horizontal transition from top layer to bottom layer:
6894 blend=all_expr='A*(X/W)+B*(1-X/W)'
6898 Apply 1x1 checkerboard effect:
6900 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
6904 Apply uncover left effect:
6906 blend=all_expr='if(gte(N*SW+X,W),A,B)'
6910 Apply uncover down effect:
6912 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
6916 Apply uncover up-left effect:
6918 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
6922 Split diagonally video and shows top and bottom layer on each side:
6924 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
6928 Display differences between the current and the previous frame:
6930 tblend=all_mode=grainextract
6936 Denoise frames using Block-Matching 3D algorithm.
6938 The filter accepts the following options.
6942 Set denoising strength. Default value is 1.
6943 Allowed range is from 0 to 999.9.
6944 The denoising algorithm is very sensitive to sigma, so adjust it
6945 according to the source.
6948 Set local patch size. This sets dimensions in 2D.
6951 Set sliding step for processing blocks. Default value is 4.
6952 Allowed range is from 1 to 64.
6953 Smaller values allows processing more reference blocks and is slower.
6956 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
6957 When set to 1, no block matching is done. Larger values allows more blocks
6959 Allowed range is from 1 to 256.
6962 Set radius for search block matching. Default is 9.
6963 Allowed range is from 1 to INT32_MAX.
6966 Set step between two search locations for block matching. Default is 1.
6967 Allowed range is from 1 to 64. Smaller is slower.
6970 Set threshold of mean square error for block matching. Valid range is 0 to
6974 Set thresholding parameter for hard thresholding in 3D transformed domain.
6975 Larger values results in stronger hard-thresholding filtering in frequency
6979 Set filtering estimation mode. Can be @code{basic} or @code{final}.
6980 Default is @code{basic}.
6983 If enabled, filter will use 2nd stream for block matching.
6984 Default is disabled for @code{basic} value of @var{estim} option,
6985 and always enabled if value of @var{estim} is @code{final}.
6988 Set planes to filter. Default is all available except alpha.
6991 @subsection Examples
6995 Basic filtering with bm3d:
6997 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7001 Same as above, but filtering only luma:
7003 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7007 Same as above, but with both estimation modes:
7009 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
7013 Same as above, but prefilter with @ref{nlmeans} filter instead:
7015 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
7021 Apply a boxblur algorithm to the input video.
7023 It accepts the following parameters:
7027 @item luma_radius, lr
7028 @item luma_power, lp
7029 @item chroma_radius, cr
7030 @item chroma_power, cp
7031 @item alpha_radius, ar
7032 @item alpha_power, ap
7036 A description of the accepted options follows.
7039 @item luma_radius, lr
7040 @item chroma_radius, cr
7041 @item alpha_radius, ar
7042 Set an expression for the box radius in pixels used for blurring the
7043 corresponding input plane.
7045 The radius value must be a non-negative number, and must not be
7046 greater than the value of the expression @code{min(w,h)/2} for the
7047 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7050 Default value for @option{luma_radius} is "2". If not specified,
7051 @option{chroma_radius} and @option{alpha_radius} default to the
7052 corresponding value set for @option{luma_radius}.
7054 The expressions can contain the following constants:
7058 The input width and height in pixels.
7062 The input chroma image width and height in pixels.
7066 The horizontal and vertical chroma subsample values. For example, for the
7067 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7070 @item luma_power, lp
7071 @item chroma_power, cp
7072 @item alpha_power, ap
7073 Specify how many times the boxblur filter is applied to the
7074 corresponding plane.
7076 Default value for @option{luma_power} is 2. If not specified,
7077 @option{chroma_power} and @option{alpha_power} default to the
7078 corresponding value set for @option{luma_power}.
7080 A value of 0 will disable the effect.
7083 @subsection Examples
7087 Apply a boxblur filter with the luma, chroma, and alpha radii
7090 boxblur=luma_radius=2:luma_power=1
7095 Set the luma radius to 2, and alpha and chroma radius to 0:
7097 boxblur=2:1:cr=0:ar=0
7101 Set the luma and chroma radii to a fraction of the video dimension:
7103 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7109 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7110 Deinterlacing Filter").
7112 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7113 interpolation algorithms.
7114 It accepts the following parameters:
7118 The interlacing mode to adopt. It accepts one of the following values:
7122 Output one frame for each frame.
7124 Output one frame for each field.
7127 The default value is @code{send_field}.
7130 The picture field parity assumed for the input interlaced video. It accepts one
7131 of the following values:
7135 Assume the top field is first.
7137 Assume the bottom field is first.
7139 Enable automatic detection of field parity.
7142 The default value is @code{auto}.
7143 If the interlacing is unknown or the decoder does not export this information,
7144 top field first will be assumed.
7147 Specify which frames to deinterlace. Accepts one of the following
7152 Deinterlace all frames.
7154 Only deinterlace frames marked as interlaced.
7157 The default value is @code{all}.
7162 Apply Contrast Adaptive Sharpen filter to video stream.
7164 The filter accepts the following options:
7168 Set the sharpening strength. Default value is 0.
7171 Set planes to filter. Default value is to filter all
7172 planes except alpha plane.
7176 Remove all color information for all colors except for certain one.
7178 The filter accepts the following options:
7182 The color which will not be replaced with neutral chroma.
7185 Similarity percentage with the above color.
7186 0.01 matches only the exact key color, while 1.0 matches everything.
7190 0.0 makes pixels either fully gray, or not gray at all.
7191 Higher values result in more preserved color.
7194 Signals that the color passed is already in YUV instead of RGB.
7196 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7197 This can be used to pass exact YUV values as hexadecimal numbers.
7200 @subsection Commands
7201 This filter supports same @ref{commands} as options.
7202 The command accepts the same syntax of the corresponding option.
7204 If the specified expression is not valid, it is kept at its current
7208 YUV colorspace color/chroma keying.
7210 The filter accepts the following options:
7214 The color which will be replaced with transparency.
7217 Similarity percentage with the key color.
7219 0.01 matches only the exact key color, while 1.0 matches everything.
7224 0.0 makes pixels either fully transparent, or not transparent at all.
7226 Higher values result in semi-transparent pixels, with a higher transparency
7227 the more similar the pixels color is to the key color.
7230 Signals that the color passed is already in YUV instead of RGB.
7232 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7233 This can be used to pass exact YUV values as hexadecimal numbers.
7236 @subsection Commands
7237 This filter supports same @ref{commands} as options.
7238 The command accepts the same syntax of the corresponding option.
7240 If the specified expression is not valid, it is kept at its current
7243 @subsection Examples
7247 Make every green pixel in the input image transparent:
7249 ffmpeg -i input.png -vf chromakey=green out.png
7253 Overlay a greenscreen-video on top of a static black background.
7255 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
7260 Reduce chrominance noise.
7262 The filter accepts the following options:
7266 Set threshold for averaging chrominance values.
7267 Sum of absolute difference of U and V pixel components or current
7268 pixel and neighbour pixels lower than this threshold will be used in
7269 averaging. Luma component is left unchanged and is copied to output.
7270 Default value is 30. Allowed range is from 1 to 200.
7273 Set horizontal radius of rectangle used for averaging.
7274 Allowed range is from 1 to 100. Default value is 5.
7277 Set vertical radius of rectangle used for averaging.
7278 Allowed range is from 1 to 100. Default value is 5.
7281 Set horizontal step when averaging. Default value is 1.
7282 Allowed range is from 1 to 50.
7283 Mostly useful to speed-up filtering.
7286 Set vertical step when averaging. Default value is 1.
7287 Allowed range is from 1 to 50.
7288 Mostly useful to speed-up filtering.
7291 @subsection Commands
7292 This filter supports same @ref{commands} as options.
7293 The command accepts the same syntax of the corresponding option.
7295 @section chromashift
7296 Shift chroma pixels horizontally and/or vertically.
7298 The filter accepts the following options:
7301 Set amount to shift chroma-blue horizontally.
7303 Set amount to shift chroma-blue vertically.
7305 Set amount to shift chroma-red horizontally.
7307 Set amount to shift chroma-red vertically.
7309 Set edge mode, can be @var{smear}, default, or @var{warp}.
7312 @subsection Commands
7314 This filter supports the all above options as @ref{commands}.
7318 Display CIE color diagram with pixels overlaid onto it.
7320 The filter accepts the following options:
7335 @item uhdtv, rec2020
7349 Set what gamuts to draw.
7351 See @code{system} option for available values.
7354 Set ciescope size, by default set to 512.
7357 Set intensity used to map input pixel values to CIE diagram.
7360 Set contrast used to draw tongue colors that are out of active color system gamut.
7363 Correct gamma displayed on scope, by default enabled.
7366 Show white point on CIE diagram, by default disabled.
7369 Set input gamma. Used only with XYZ input color space.
7374 Visualize information exported by some codecs.
7376 Some codecs can export information through frames using side-data or other
7377 means. For example, some MPEG based codecs export motion vectors through the
7378 @var{export_mvs} flag in the codec @option{flags2} option.
7380 The filter accepts the following option:
7384 Set motion vectors to visualize.
7386 Available flags for @var{mv} are:
7390 forward predicted MVs of P-frames
7392 forward predicted MVs of B-frames
7394 backward predicted MVs of B-frames
7398 Display quantization parameters using the chroma planes.
7401 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7403 Available flags for @var{mv_type} are:
7407 forward predicted MVs
7409 backward predicted MVs
7412 @item frame_type, ft
7413 Set frame type to visualize motion vectors of.
7415 Available flags for @var{frame_type} are:
7419 intra-coded frames (I-frames)
7421 predicted frames (P-frames)
7423 bi-directionally predicted frames (B-frames)
7427 @subsection Examples
7431 Visualize forward predicted MVs of all frames using @command{ffplay}:
7433 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7437 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7439 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7443 @section colorbalance
7444 Modify intensity of primary colors (red, green and blue) of input frames.
7446 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7447 regions for the red-cyan, green-magenta or blue-yellow balance.
7449 A positive adjustment value shifts the balance towards the primary color, a negative
7450 value towards the complementary color.
7452 The filter accepts the following options:
7458 Adjust red, green and blue shadows (darkest pixels).
7463 Adjust red, green and blue midtones (medium pixels).
7468 Adjust red, green and blue highlights (brightest pixels).
7470 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7473 Preserve lightness when changing color balance. Default is disabled.
7476 @subsection Examples
7480 Add red color cast to shadows:
7486 @subsection Commands
7488 This filter supports the all above options as @ref{commands}.
7490 @section colorchannelmixer
7492 Adjust video input frames by re-mixing color channels.
7494 This filter modifies a color channel by adding the values associated to
7495 the other channels of the same pixels. For example if the value to
7496 modify is red, the output value will be:
7498 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7501 The filter accepts the following options:
7508 Adjust contribution of input red, green, blue and alpha channels for output red channel.
7509 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
7515 Adjust contribution of input red, green, blue and alpha channels for output green channel.
7516 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
7522 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
7523 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
7529 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7530 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7532 Allowed ranges for options are @code{[-2.0, 2.0]}.
7535 @subsection Examples
7539 Convert source to grayscale:
7541 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7544 Simulate sepia tones:
7546 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7550 @subsection Commands
7552 This filter supports the all above options as @ref{commands}.
7555 RGB colorspace color keying.
7557 The filter accepts the following options:
7561 The color which will be replaced with transparency.
7564 Similarity percentage with the key color.
7566 0.01 matches only the exact key color, while 1.0 matches everything.
7571 0.0 makes pixels either fully transparent, or not transparent at all.
7573 Higher values result in semi-transparent pixels, with a higher transparency
7574 the more similar the pixels color is to the key color.
7577 @subsection Examples
7581 Make every green pixel in the input image transparent:
7583 ffmpeg -i input.png -vf colorkey=green out.png
7587 Overlay a greenscreen-video on top of a static background image.
7589 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
7593 @subsection Commands
7594 This filter supports same @ref{commands} as options.
7595 The command accepts the same syntax of the corresponding option.
7597 If the specified expression is not valid, it is kept at its current
7601 Remove all color information for all RGB colors except for certain one.
7603 The filter accepts the following options:
7607 The color which will not be replaced with neutral gray.
7610 Similarity percentage with the above color.
7611 0.01 matches only the exact key color, while 1.0 matches everything.
7614 Blend percentage. 0.0 makes pixels fully gray.
7615 Higher values result in more preserved color.
7618 @subsection Commands
7619 This filter supports same @ref{commands} as options.
7620 The command accepts the same syntax of the corresponding option.
7622 If the specified expression is not valid, it is kept at its current
7625 @section colorlevels
7627 Adjust video input frames using levels.
7629 The filter accepts the following options:
7636 Adjust red, green, blue and alpha input black point.
7637 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7643 Adjust red, green, blue and alpha input white point.
7644 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
7646 Input levels are used to lighten highlights (bright tones), darken shadows
7647 (dark tones), change the balance of bright and dark tones.
7653 Adjust red, green, blue and alpha output black point.
7654 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
7660 Adjust red, green, blue and alpha output white point.
7661 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
7663 Output levels allows manual selection of a constrained output level range.
7666 @subsection Examples
7670 Make video output darker:
7672 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
7678 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
7682 Make video output lighter:
7684 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
7688 Increase brightness:
7690 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
7694 @subsection Commands
7696 This filter supports the all above options as @ref{commands}.
7698 @section colormatrix
7700 Convert color matrix.
7702 The filter accepts the following options:
7707 Specify the source and destination color matrix. Both values must be
7710 The accepted values are:
7738 For example to convert from BT.601 to SMPTE-240M, use the command:
7740 colormatrix=bt601:smpte240m
7745 Convert colorspace, transfer characteristics or color primaries.
7746 Input video needs to have an even size.
7748 The filter accepts the following options:
7753 Specify all color properties at once.
7755 The accepted values are:
7785 Specify output colorspace.
7787 The accepted values are:
7796 BT.470BG or BT.601-6 625
7799 SMPTE-170M or BT.601-6 525
7808 BT.2020 with non-constant luminance
7814 Specify output transfer characteristics.
7816 The accepted values are:
7828 Constant gamma of 2.2
7831 Constant gamma of 2.8
7834 SMPTE-170M, BT.601-6 625 or BT.601-6 525
7852 BT.2020 for 10-bits content
7855 BT.2020 for 12-bits content
7861 Specify output color primaries.
7863 The accepted values are:
7872 BT.470BG or BT.601-6 625
7875 SMPTE-170M or BT.601-6 525
7899 Specify output color range.
7901 The accepted values are:
7904 TV (restricted) range
7907 MPEG (restricted) range
7918 Specify output color format.
7920 The accepted values are:
7923 YUV 4:2:0 planar 8-bits
7926 YUV 4:2:0 planar 10-bits
7929 YUV 4:2:0 planar 12-bits
7932 YUV 4:2:2 planar 8-bits
7935 YUV 4:2:2 planar 10-bits
7938 YUV 4:2:2 planar 12-bits
7941 YUV 4:4:4 planar 8-bits
7944 YUV 4:4:4 planar 10-bits
7947 YUV 4:4:4 planar 12-bits
7952 Do a fast conversion, which skips gamma/primary correction. This will take
7953 significantly less CPU, but will be mathematically incorrect. To get output
7954 compatible with that produced by the colormatrix filter, use fast=1.
7957 Specify dithering mode.
7959 The accepted values are:
7965 Floyd-Steinberg dithering
7969 Whitepoint adaptation mode.
7971 The accepted values are:
7974 Bradford whitepoint adaptation
7977 von Kries whitepoint adaptation
7980 identity whitepoint adaptation (i.e. no whitepoint adaptation)
7984 Override all input properties at once. Same accepted values as @ref{all}.
7987 Override input colorspace. Same accepted values as @ref{space}.
7990 Override input color primaries. Same accepted values as @ref{primaries}.
7993 Override input transfer characteristics. Same accepted values as @ref{trc}.
7996 Override input color range. Same accepted values as @ref{range}.
8000 The filter converts the transfer characteristics, color space and color
8001 primaries to the specified user values. The output value, if not specified,
8002 is set to a default value based on the "all" property. If that property is
8003 also not specified, the filter will log an error. The output color range and
8004 format default to the same value as the input color range and format. The
8005 input transfer characteristics, color space, color primaries and color range
8006 should be set on the input data. If any of these are missing, the filter will
8007 log an error and no conversion will take place.
8009 For example to convert the input to SMPTE-240M, use the command:
8011 colorspace=smpte240m
8014 @section convolution
8016 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8018 The filter accepts the following options:
8025 Set matrix for each plane.
8026 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8027 and from 1 to 49 odd number of signed integers in @var{row} mode.
8033 Set multiplier for calculated value for each plane.
8034 If unset or 0, it will be sum of all matrix elements.
8040 Set bias for each plane. This value is added to the result of the multiplication.
8041 Useful for making the overall image brighter or darker. Default is 0.0.
8047 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8048 Default is @var{square}.
8051 @subsection Examples
8057 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"
8063 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"
8069 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"
8075 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"
8079 Apply laplacian edge detector which includes diagonals:
8081 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"
8087 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"
8093 Apply 2D convolution of video stream in frequency domain using second stream
8096 The filter accepts the following options:
8100 Set which planes to process.
8103 Set which impulse video frames will be processed, can be @var{first}
8104 or @var{all}. Default is @var{all}.
8107 The @code{convolve} filter also supports the @ref{framesync} options.
8111 Copy the input video source unchanged to the output. This is mainly useful for
8116 Video filtering on GPU using Apple's CoreImage API on OSX.
8118 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8119 processed by video hardware. However, software-based OpenGL implementations
8120 exist which means there is no guarantee for hardware processing. It depends on
8123 There are many filters and image generators provided by Apple that come with a
8124 large variety of options. The filter has to be referenced by its name along
8127 The coreimage filter accepts the following options:
8130 List all available filters and generators along with all their respective
8131 options as well as possible minimum and maximum values along with the default
8138 Specify all filters by their respective name and options.
8139 Use @var{list_filters} to determine all valid filter names and options.
8140 Numerical options are specified by a float value and are automatically clamped
8141 to their respective value range. Vector and color options have to be specified
8142 by a list of space separated float values. Character escaping has to be done.
8143 A special option name @code{default} is available to use default options for a
8146 It is required to specify either @code{default} or at least one of the filter options.
8147 All omitted options are used with their default values.
8148 The syntax of the filter string is as follows:
8150 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8154 Specify a rectangle where the output of the filter chain is copied into the
8155 input image. It is given by a list of space separated float values:
8157 output_rect=x\ y\ width\ height
8159 If not given, the output rectangle equals the dimensions of the input image.
8160 The output rectangle is automatically cropped at the borders of the input
8161 image. Negative values are valid for each component.
8163 output_rect=25\ 25\ 100\ 100
8167 Several filters can be chained for successive processing without GPU-HOST
8168 transfers allowing for fast processing of complex filter chains.
8169 Currently, only filters with zero (generators) or exactly one (filters) input
8170 image and one output image are supported. Also, transition filters are not yet
8173 Some filters generate output images with additional padding depending on the
8174 respective filter kernel. The padding is automatically removed to ensure the
8175 filter output has the same size as the input image.
8177 For image generators, the size of the output image is determined by the
8178 previous output image of the filter chain or the input image of the whole
8179 filterchain, respectively. The generators do not use the pixel information of
8180 this image to generate their output. However, the generated output is
8181 blended onto this image, resulting in partial or complete coverage of the
8184 The @ref{coreimagesrc} video source can be used for generating input images
8185 which are directly fed into the filter chain. By using it, providing input
8186 images by another video source or an input video is not required.
8188 @subsection Examples
8193 List all filters available:
8195 coreimage=list_filters=true
8199 Use the CIBoxBlur filter with default options to blur an image:
8201 coreimage=filter=CIBoxBlur@@default
8205 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8206 its center at 100x100 and a radius of 50 pixels:
8208 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8212 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8213 given as complete and escaped command-line for Apple's standard bash shell:
8215 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8221 Cover a rectangular object
8223 It accepts the following options:
8227 Filepath of the optional cover image, needs to be in yuv420.
8232 It accepts the following values:
8235 cover it by the supplied image
8237 cover it by interpolating the surrounding pixels
8240 Default value is @var{blur}.
8243 @subsection Examples
8247 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8249 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8255 Crop the input video to given dimensions.
8257 It accepts the following parameters:
8261 The width of the output video. It defaults to @code{iw}.
8262 This expression is evaluated only once during the filter
8263 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8266 The height of the output video. It defaults to @code{ih}.
8267 This expression is evaluated only once during the filter
8268 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8271 The horizontal position, in the input video, of the left edge of the output
8272 video. It defaults to @code{(in_w-out_w)/2}.
8273 This expression is evaluated per-frame.
8276 The vertical position, in the input video, of the top edge of the output video.
8277 It defaults to @code{(in_h-out_h)/2}.
8278 This expression is evaluated per-frame.
8281 If set to 1 will force the output display aspect ratio
8282 to be the same of the input, by changing the output sample aspect
8283 ratio. It defaults to 0.
8286 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8287 width/height/x/y as specified and will not be rounded to nearest smaller value.
8291 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8292 expressions containing the following constants:
8297 The computed values for @var{x} and @var{y}. They are evaluated for
8302 The input width and height.
8306 These are the same as @var{in_w} and @var{in_h}.
8310 The output (cropped) width and height.
8314 These are the same as @var{out_w} and @var{out_h}.
8317 same as @var{iw} / @var{ih}
8320 input sample aspect ratio
8323 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8327 horizontal and vertical chroma subsample values. For example for the
8328 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8331 The number of the input frame, starting from 0.
8334 the position in the file of the input frame, NAN if unknown
8337 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8341 The expression for @var{out_w} may depend on the value of @var{out_h},
8342 and the expression for @var{out_h} may depend on @var{out_w}, but they
8343 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8344 evaluated after @var{out_w} and @var{out_h}.
8346 The @var{x} and @var{y} parameters specify the expressions for the
8347 position of the top-left corner of the output (non-cropped) area. They
8348 are evaluated for each frame. If the evaluated value is not valid, it
8349 is approximated to the nearest valid value.
8351 The expression for @var{x} may depend on @var{y}, and the expression
8352 for @var{y} may depend on @var{x}.
8354 @subsection Examples
8358 Crop area with size 100x100 at position (12,34).
8363 Using named options, the example above becomes:
8365 crop=w=100:h=100:x=12:y=34
8369 Crop the central input area with size 100x100:
8375 Crop the central input area with size 2/3 of the input video:
8377 crop=2/3*in_w:2/3*in_h
8381 Crop the input video central square:
8388 Delimit the rectangle with the top-left corner placed at position
8389 100:100 and the right-bottom corner corresponding to the right-bottom
8390 corner of the input image.
8392 crop=in_w-100:in_h-100:100:100
8396 Crop 10 pixels from the left and right borders, and 20 pixels from
8397 the top and bottom borders
8399 crop=in_w-2*10:in_h-2*20
8403 Keep only the bottom right quarter of the input image:
8405 crop=in_w/2:in_h/2:in_w/2:in_h/2
8409 Crop height for getting Greek harmony:
8411 crop=in_w:1/PHI*in_w
8415 Apply trembling effect:
8417 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)
8421 Apply erratic camera effect depending on timestamp:
8423 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)"
8427 Set x depending on the value of y:
8429 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8433 @subsection Commands
8435 This filter supports the following commands:
8441 Set width/height of the output video and the horizontal/vertical position
8443 The command accepts the same syntax of the corresponding option.
8445 If the specified expression is not valid, it is kept at its current
8451 Auto-detect the crop size.
8453 It calculates the necessary cropping parameters and prints the
8454 recommended parameters via the logging system. The detected dimensions
8455 correspond to the non-black area of the input video.
8457 It accepts the following parameters:
8462 Set higher black value threshold, which can be optionally specified
8463 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8464 value greater to the set value is considered non-black. It defaults to 24.
8465 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8466 on the bitdepth of the pixel format.
8469 The value which the width/height should be divisible by. It defaults to
8470 16. The offset is automatically adjusted to center the video. Use 2 to
8471 get only even dimensions (needed for 4:2:2 video). 16 is best when
8472 encoding to most video codecs.
8474 @item reset_count, reset
8475 Set the counter that determines after how many frames cropdetect will
8476 reset the previously detected largest video area and start over to
8477 detect the current optimal crop area. Default value is 0.
8479 This can be useful when channel logos distort the video area. 0
8480 indicates 'never reset', and returns the largest area encountered during
8487 Delay video filtering until a given wallclock timestamp. The filter first
8488 passes on @option{preroll} amount of frames, then it buffers at most
8489 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8490 it forwards the buffered frames and also any subsequent frames coming in its
8493 The filter can be used synchronize the output of multiple ffmpeg processes for
8494 realtime output devices like decklink. By putting the delay in the filtering
8495 chain and pre-buffering frames the process can pass on data to output almost
8496 immediately after the target wallclock timestamp is reached.
8498 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
8504 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
8507 The duration of content to pass on as preroll expressed in seconds. Default is 0.
8510 The maximum duration of content to buffer before waiting for the cue expressed
8511 in seconds. Default is 0.
8518 Apply color adjustments using curves.
8520 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
8521 component (red, green and blue) has its values defined by @var{N} key points
8522 tied from each other using a smooth curve. The x-axis represents the pixel
8523 values from the input frame, and the y-axis the new pixel values to be set for
8526 By default, a component curve is defined by the two points @var{(0;0)} and
8527 @var{(1;1)}. This creates a straight line where each original pixel value is
8528 "adjusted" to its own value, which means no change to the image.
8530 The filter allows you to redefine these two points and add some more. A new
8531 curve (using a natural cubic spline interpolation) will be define to pass
8532 smoothly through all these new coordinates. The new defined points needs to be
8533 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
8534 be in the @var{[0;1]} interval. If the computed curves happened to go outside
8535 the vector spaces, the values will be clipped accordingly.
8537 The filter accepts the following options:
8541 Select one of the available color presets. This option can be used in addition
8542 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
8543 options takes priority on the preset values.
8544 Available presets are:
8547 @item color_negative
8550 @item increase_contrast
8552 @item linear_contrast
8553 @item medium_contrast
8555 @item strong_contrast
8558 Default is @code{none}.
8560 Set the master key points. These points will define a second pass mapping. It
8561 is sometimes called a "luminance" or "value" mapping. It can be used with
8562 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
8563 post-processing LUT.
8565 Set the key points for the red component.
8567 Set the key points for the green component.
8569 Set the key points for the blue component.
8571 Set the key points for all components (not including master).
8572 Can be used in addition to the other key points component
8573 options. In this case, the unset component(s) will fallback on this
8574 @option{all} setting.
8576 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
8578 Save Gnuplot script of the curves in specified file.
8581 To avoid some filtergraph syntax conflicts, each key points list need to be
8582 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
8584 @subsection Examples
8588 Increase slightly the middle level of blue:
8590 curves=blue='0/0 0.5/0.58 1/1'
8596 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'
8598 Here we obtain the following coordinates for each components:
8601 @code{(0;0.11) (0.42;0.51) (1;0.95)}
8603 @code{(0;0) (0.50;0.48) (1;1)}
8605 @code{(0;0.22) (0.49;0.44) (1;0.80)}
8609 The previous example can also be achieved with the associated built-in preset:
8611 curves=preset=vintage
8621 Use a Photoshop preset and redefine the points of the green component:
8623 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
8627 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
8628 and @command{gnuplot}:
8630 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
8631 gnuplot -p /tmp/curves.plt
8637 Video data analysis filter.
8639 This filter shows hexadecimal pixel values of part of video.
8641 The filter accepts the following options:
8645 Set output video size.
8648 Set x offset from where to pick pixels.
8651 Set y offset from where to pick pixels.
8654 Set scope mode, can be one of the following:
8657 Draw hexadecimal pixel values with white color on black background.
8660 Draw hexadecimal pixel values with input video pixel color on black
8664 Draw hexadecimal pixel values on color background picked from input video,
8665 the text color is picked in such way so its always visible.
8669 Draw rows and columns numbers on left and top of video.
8672 Set background opacity.
8675 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
8679 Apply Directional blur filter.
8681 The filter accepts the following options:
8685 Set angle of directional blur. Default is @code{45}.
8688 Set radius of directional blur. Default is @code{5}.
8691 Set which planes to filter. By default all planes are filtered.
8694 @subsection Commands
8695 This filter supports same @ref{commands} as options.
8696 The command accepts the same syntax of the corresponding option.
8698 If the specified expression is not valid, it is kept at its current
8703 Denoise frames using 2D DCT (frequency domain filtering).
8705 This filter is not designed for real time.
8707 The filter accepts the following options:
8711 Set the noise sigma constant.
8713 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
8714 coefficient (absolute value) below this threshold with be dropped.
8716 If you need a more advanced filtering, see @option{expr}.
8718 Default is @code{0}.
8721 Set number overlapping pixels for each block. Since the filter can be slow, you
8722 may want to reduce this value, at the cost of a less effective filter and the
8723 risk of various artefacts.
8725 If the overlapping value doesn't permit processing the whole input width or
8726 height, a warning will be displayed and according borders won't be denoised.
8728 Default value is @var{blocksize}-1, which is the best possible setting.
8731 Set the coefficient factor expression.
8733 For each coefficient of a DCT block, this expression will be evaluated as a
8734 multiplier value for the coefficient.
8736 If this is option is set, the @option{sigma} option will be ignored.
8738 The absolute value of the coefficient can be accessed through the @var{c}
8742 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8743 @var{blocksize}, which is the width and height of the processed blocks.
8745 The default value is @var{3} (8x8) and can be raised to @var{4} for a
8746 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
8747 on the speed processing. Also, a larger block size does not necessarily means a
8751 @subsection Examples
8753 Apply a denoise with a @option{sigma} of @code{4.5}:
8758 The same operation can be achieved using the expression system:
8760 dctdnoiz=e='gte(c, 4.5*3)'
8763 Violent denoise using a block size of @code{16x16}:
8770 Remove banding artifacts from input video.
8771 It works by replacing banded pixels with average value of referenced pixels.
8773 The filter accepts the following options:
8780 Set banding detection threshold for each plane. Default is 0.02.
8781 Valid range is 0.00003 to 0.5.
8782 If difference between current pixel and reference pixel is less than threshold,
8783 it will be considered as banded.
8786 Banding detection range in pixels. Default is 16. If positive, random number
8787 in range 0 to set value will be used. If negative, exact absolute value
8789 The range defines square of four pixels around current pixel.
8792 Set direction in radians from which four pixel will be compared. If positive,
8793 random direction from 0 to set direction will be picked. If negative, exact of
8794 absolute value will be picked. For example direction 0, -PI or -2*PI radians
8795 will pick only pixels on same row and -PI/2 will pick only pixels on same
8799 If enabled, current pixel is compared with average value of all four
8800 surrounding pixels. The default is enabled. If disabled current pixel is
8801 compared with all four surrounding pixels. The pixel is considered banded
8802 if only all four differences with surrounding pixels are less than threshold.
8805 If enabled, current pixel is changed if and only if all pixel components are banded,
8806 e.g. banding detection threshold is triggered for all color components.
8807 The default is disabled.
8812 Remove blocking artifacts from input video.
8814 The filter accepts the following options:
8818 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
8819 This controls what kind of deblocking is applied.
8822 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
8828 Set blocking detection thresholds. Allowed range is 0 to 1.
8829 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
8830 Using higher threshold gives more deblocking strength.
8831 Setting @var{alpha} controls threshold detection at exact edge of block.
8832 Remaining options controls threshold detection near the edge. Each one for
8833 below/above or left/right. Setting any of those to @var{0} disables
8837 Set planes to filter. Default is to filter all available planes.
8840 @subsection Examples
8844 Deblock using weak filter and block size of 4 pixels.
8846 deblock=filter=weak:block=4
8850 Deblock using strong filter, block size of 4 pixels and custom thresholds for
8851 deblocking more edges.
8853 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
8857 Similar as above, but filter only first plane.
8859 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
8863 Similar as above, but filter only second and third plane.
8865 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
8872 Drop duplicated frames at regular intervals.
8874 The filter accepts the following options:
8878 Set the number of frames from which one will be dropped. Setting this to
8879 @var{N} means one frame in every batch of @var{N} frames will be dropped.
8880 Default is @code{5}.
8883 Set the threshold for duplicate detection. If the difference metric for a frame
8884 is less than or equal to this value, then it is declared as duplicate. Default
8888 Set scene change threshold. Default is @code{15}.
8892 Set the size of the x and y-axis blocks used during metric calculations.
8893 Larger blocks give better noise suppression, but also give worse detection of
8894 small movements. Must be a power of two. Default is @code{32}.
8897 Mark main input as a pre-processed input and activate clean source input
8898 stream. This allows the input to be pre-processed with various filters to help
8899 the metrics calculation while keeping the frame selection lossless. When set to
8900 @code{1}, the first stream is for the pre-processed input, and the second
8901 stream is the clean source from where the kept frames are chosen. Default is
8905 Set whether or not chroma is considered in the metric calculations. Default is
8911 Apply 2D deconvolution of video stream in frequency domain using second stream
8914 The filter accepts the following options:
8918 Set which planes to process.
8921 Set which impulse video frames will be processed, can be @var{first}
8922 or @var{all}. Default is @var{all}.
8925 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
8926 and height are not same and not power of 2 or if stream prior to convolving
8930 The @code{deconvolve} filter also supports the @ref{framesync} options.
8934 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
8936 It accepts the following options:
8940 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
8941 @var{rainbows} for cross-color reduction.
8944 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
8947 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
8950 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
8953 Set temporal chroma threshold. Lower values increases reduction of cross-color.
8958 Apply deflate effect to the video.
8960 This filter replaces the pixel by the local(3x3) average by taking into account
8961 only values lower than the pixel.
8963 It accepts the following options:
8970 Limit the maximum change for each plane, default is 65535.
8971 If 0, plane will remain unchanged.
8974 @subsection Commands
8976 This filter supports the all above options as @ref{commands}.
8980 Remove temporal frame luminance variations.
8982 It accepts the following options:
8986 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
8989 Set averaging mode to smooth temporal luminance variations.
8991 Available values are:
9016 Do not actually modify frame. Useful when one only wants metadata.
9021 Remove judder produced by partially interlaced telecined content.
9023 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9024 source was partially telecined content then the output of @code{pullup,dejudder}
9025 will have a variable frame rate. May change the recorded frame rate of the
9026 container. Aside from that change, this filter will not affect constant frame
9029 The option available in this filter is:
9033 Specify the length of the window over which the judder repeats.
9035 Accepts any integer greater than 1. Useful values are:
9039 If the original was telecined from 24 to 30 fps (Film to NTSC).
9042 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9045 If a mixture of the two.
9048 The default is @samp{4}.
9053 Suppress a TV station logo by a simple interpolation of the surrounding
9054 pixels. Just set a rectangle covering the logo and watch it disappear
9055 (and sometimes something even uglier appear - your mileage may vary).
9057 It accepts the following parameters:
9062 Specify the top left corner coordinates of the logo. They must be
9067 Specify the width and height of the logo to clear. They must be
9071 Specify the thickness of the fuzzy edge of the rectangle (added to
9072 @var{w} and @var{h}). The default value is 1. This option is
9073 deprecated, setting higher values should no longer be necessary and
9077 When set to 1, a green rectangle is drawn on the screen to simplify
9078 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9079 The default value is 0.
9081 The rectangle is drawn on the outermost pixels which will be (partly)
9082 replaced with interpolated values. The values of the next pixels
9083 immediately outside this rectangle in each direction will be used to
9084 compute the interpolated pixel values inside the rectangle.
9088 @subsection Examples
9092 Set a rectangle covering the area with top left corner coordinates 0,0
9093 and size 100x77, and a band of size 10:
9095 delogo=x=0:y=0:w=100:h=77:band=10
9103 Remove the rain in the input image/video by applying the derain methods based on
9104 convolutional neural networks. Supported models:
9108 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9109 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9112 Training as well as model generation scripts are provided in
9113 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9115 Native model files (.model) can be generated from TensorFlow model
9116 files (.pb) by using tools/python/convert.py
9118 The filter accepts the following options:
9122 Specify which filter to use. This option accepts the following values:
9126 Derain filter. To conduct derain filter, you need to use a derain model.
9129 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9131 Default value is @samp{derain}.
9134 Specify which DNN backend to use for model loading and execution. This option accepts
9135 the following values:
9139 Native implementation of DNN loading and execution.
9142 TensorFlow backend. To enable this backend you
9143 need to install the TensorFlow for C library (see
9144 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9145 @code{--enable-libtensorflow}
9147 Default value is @samp{native}.
9150 Set path to model file specifying network architecture and its parameters.
9151 Note that different backends use different file formats. TensorFlow and native
9152 backend can load files for only its format.
9155 It can also be finished with @ref{dnn_processing} filter.
9159 Attempt to fix small changes in horizontal and/or vertical shift. This
9160 filter helps remove camera shake from hand-holding a camera, bumping a
9161 tripod, moving on a vehicle, etc.
9163 The filter accepts the following options:
9171 Specify a rectangular area where to limit the search for motion
9173 If desired the search for motion vectors can be limited to a
9174 rectangular area of the frame defined by its top left corner, width
9175 and height. These parameters have the same meaning as the drawbox
9176 filter which can be used to visualise the position of the bounding
9179 This is useful when simultaneous movement of subjects within the frame
9180 might be confused for camera motion by the motion vector search.
9182 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9183 then the full frame is used. This allows later options to be set
9184 without specifying the bounding box for the motion vector search.
9186 Default - search the whole frame.
9190 Specify the maximum extent of movement in x and y directions in the
9191 range 0-64 pixels. Default 16.
9194 Specify how to generate pixels to fill blanks at the edge of the
9195 frame. Available values are:
9198 Fill zeroes at blank locations
9200 Original image at blank locations
9202 Extruded edge value at blank locations
9204 Mirrored edge at blank locations
9206 Default value is @samp{mirror}.
9209 Specify the blocksize to use for motion search. Range 4-128 pixels,
9213 Specify the contrast threshold for blocks. Only blocks with more than
9214 the specified contrast (difference between darkest and lightest
9215 pixels) will be considered. Range 1-255, default 125.
9218 Specify the search strategy. Available values are:
9221 Set exhaustive search
9223 Set less exhaustive search.
9225 Default value is @samp{exhaustive}.
9228 If set then a detailed log of the motion search is written to the
9235 Remove unwanted contamination of foreground colors, caused by reflected color of
9236 greenscreen or bluescreen.
9238 This filter accepts the following options:
9242 Set what type of despill to use.
9245 Set how spillmap will be generated.
9248 Set how much to get rid of still remaining spill.
9251 Controls amount of red in spill area.
9254 Controls amount of green in spill area.
9255 Should be -1 for greenscreen.
9258 Controls amount of blue in spill area.
9259 Should be -1 for bluescreen.
9262 Controls brightness of spill area, preserving colors.
9265 Modify alpha from generated spillmap.
9270 Apply an exact inverse of the telecine operation. It requires a predefined
9271 pattern specified using the pattern option which must be the same as that passed
9272 to the telecine filter.
9274 This filter accepts the following options:
9283 The default value is @code{top}.
9287 A string of numbers representing the pulldown pattern you wish to apply.
9288 The default value is @code{23}.
9291 A number representing position of the first frame with respect to the telecine
9292 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9297 Apply dilation effect to the video.
9299 This filter replaces the pixel by the local(3x3) maximum.
9301 It accepts the following options:
9308 Limit the maximum change for each plane, default is 65535.
9309 If 0, plane will remain unchanged.
9312 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9315 Flags to local 3x3 coordinates maps like this:
9322 @subsection Commands
9324 This filter supports the all above options as @ref{commands}.
9328 Displace pixels as indicated by second and third input stream.
9330 It takes three input streams and outputs one stream, the first input is the
9331 source, and second and third input are displacement maps.
9333 The second input specifies how much to displace pixels along the
9334 x-axis, while the third input specifies how much to displace pixels
9336 If one of displacement map streams terminates, last frame from that
9337 displacement map will be used.
9339 Note that once generated, displacements maps can be reused over and over again.
9341 A description of the accepted options follows.
9345 Set displace behavior for pixels that are out of range.
9347 Available values are:
9350 Missing pixels are replaced by black pixels.
9353 Adjacent pixels will spread out to replace missing pixels.
9356 Out of range pixels are wrapped so they point to pixels of other side.
9359 Out of range pixels will be replaced with mirrored pixels.
9361 Default is @samp{smear}.
9365 @subsection Examples
9369 Add ripple effect to rgb input of video size hd720:
9371 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
9375 Add wave effect to rgb input of video size hd720:
9377 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
9381 @anchor{dnn_processing}
9382 @section dnn_processing
9384 Do image processing with deep neural networks. It works together with another filter
9385 which converts the pixel format of the Frame to what the dnn network requires.
9387 The filter accepts the following options:
9391 Specify which DNN backend to use for model loading and execution. This option accepts
9392 the following values:
9396 Native implementation of DNN loading and execution.
9399 TensorFlow backend. To enable this backend you
9400 need to install the TensorFlow for C library (see
9401 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9402 @code{--enable-libtensorflow}
9405 OpenVINO backend. To enable this backend you
9406 need to build and install the OpenVINO for C library (see
9407 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9408 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9409 be needed if the header files and libraries are not installed into system path)
9413 Default value is @samp{native}.
9416 Set path to model file specifying network architecture and its parameters.
9417 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9418 backend can load files for only its format.
9420 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9423 Set the input name of the dnn network.
9426 Set the output name of the dnn network.
9430 @subsection Examples
9434 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9436 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9440 Halve the pixel value of the frame with format gray32f:
9442 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
9446 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9448 ./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
9452 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9454 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9461 Draw a colored box on the input image.
9463 It accepts the following parameters:
9468 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9472 The expressions which specify the width and height of the box; if 0 they are interpreted as
9473 the input width and height. It defaults to 0.
9476 Specify the color of the box to write. For the general syntax of this option,
9477 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9478 value @code{invert} is used, the box edge color is the same as the
9479 video with inverted luma.
9482 The expression which sets the thickness of the box edge.
9483 A value of @code{fill} will create a filled box. Default value is @code{3}.
9485 See below for the list of accepted constants.
9488 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
9489 will overwrite the video's color and alpha pixels.
9490 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
9493 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9494 following constants:
9498 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9502 horizontal and vertical chroma subsample values. For example for the
9503 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9507 The input width and height.
9510 The input sample aspect ratio.
9514 The x and y offset coordinates where the box is drawn.
9518 The width and height of the drawn box.
9521 The thickness of the drawn box.
9523 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9524 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9528 @subsection Examples
9532 Draw a black box around the edge of the input image:
9538 Draw a box with color red and an opacity of 50%:
9540 drawbox=10:20:200:60:red@@0.5
9543 The previous example can be specified as:
9545 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
9549 Fill the box with pink color:
9551 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
9555 Draw a 2-pixel red 2.40:1 mask:
9557 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
9561 @subsection Commands
9562 This filter supports same commands as options.
9563 The command accepts the same syntax of the corresponding option.
9565 If the specified expression is not valid, it is kept at its current
9570 Draw a graph using input video metadata.
9572 It accepts the following parameters:
9576 Set 1st frame metadata key from which metadata values will be used to draw a graph.
9579 Set 1st foreground color expression.
9582 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
9585 Set 2nd foreground color expression.
9588 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
9591 Set 3rd foreground color expression.
9594 Set 4th frame metadata key from which metadata values will be used to draw a graph.
9597 Set 4th foreground color expression.
9600 Set minimal value of metadata value.
9603 Set maximal value of metadata value.
9606 Set graph background color. Default is white.
9611 Available values for mode is:
9618 Default is @code{line}.
9623 Available values for slide is:
9626 Draw new frame when right border is reached.
9629 Replace old columns with new ones.
9632 Scroll from right to left.
9635 Scroll from left to right.
9638 Draw single picture.
9641 Default is @code{frame}.
9644 Set size of graph video. For the syntax of this option, check the
9645 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
9646 The default value is @code{900x256}.
9649 Set the output frame rate. Default value is @code{25}.
9651 The foreground color expressions can use the following variables:
9654 Minimal value of metadata value.
9657 Maximal value of metadata value.
9660 Current metadata key value.
9663 The color is defined as 0xAABBGGRR.
9666 Example using metadata from @ref{signalstats} filter:
9668 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
9671 Example using metadata from @ref{ebur128} filter:
9673 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
9678 Draw a grid on the input image.
9680 It accepts the following parameters:
9685 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
9689 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
9690 input width and height, respectively, minus @code{thickness}, so image gets
9691 framed. Default to 0.
9694 Specify the color of the grid. For the general syntax of this option,
9695 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9696 value @code{invert} is used, the grid color is the same as the
9697 video with inverted luma.
9700 The expression which sets the thickness of the grid line. Default value is @code{1}.
9702 See below for the list of accepted constants.
9705 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
9706 will overwrite the video's color and alpha pixels.
9707 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
9710 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9711 following constants:
9715 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9719 horizontal and vertical chroma subsample values. For example for the
9720 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9724 The input grid cell width and height.
9727 The input sample aspect ratio.
9731 The x and y coordinates of some point of grid intersection (meant to configure offset).
9735 The width and height of the drawn cell.
9738 The thickness of the drawn cell.
9740 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9741 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9745 @subsection Examples
9749 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
9751 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
9755 Draw a white 3x3 grid with an opacity of 50%:
9757 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
9761 @subsection Commands
9762 This filter supports same commands as options.
9763 The command accepts the same syntax of the corresponding option.
9765 If the specified expression is not valid, it is kept at its current
9771 Draw a text string or text from a specified file on top of a video, using the
9772 libfreetype library.
9774 To enable compilation of this filter, you need to configure FFmpeg with
9775 @code{--enable-libfreetype}.
9776 To enable default font fallback and the @var{font} option you need to
9777 configure FFmpeg with @code{--enable-libfontconfig}.
9778 To enable the @var{text_shaping} option, you need to configure FFmpeg with
9779 @code{--enable-libfribidi}.
9783 It accepts the following parameters:
9788 Used to draw a box around text using the background color.
9789 The value must be either 1 (enable) or 0 (disable).
9790 The default value of @var{box} is 0.
9793 Set the width of the border to be drawn around the box using @var{boxcolor}.
9794 The default value of @var{boxborderw} is 0.
9797 The color to be used for drawing box around text. For the syntax of this
9798 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9800 The default value of @var{boxcolor} is "white".
9803 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
9804 The default value of @var{line_spacing} is 0.
9807 Set the width of the border to be drawn around the text using @var{bordercolor}.
9808 The default value of @var{borderw} is 0.
9811 Set the color to be used for drawing border around text. For the syntax of this
9812 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9814 The default value of @var{bordercolor} is "black".
9817 Select how the @var{text} is expanded. Can be either @code{none},
9818 @code{strftime} (deprecated) or
9819 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
9823 Set a start time for the count. Value is in microseconds. Only applied
9824 in the deprecated strftime expansion mode. To emulate in normal expansion
9825 mode use the @code{pts} function, supplying the start time (in seconds)
9826 as the second argument.
9829 If true, check and fix text coords to avoid clipping.
9832 The color to be used for drawing fonts. For the syntax of this option, check
9833 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9835 The default value of @var{fontcolor} is "black".
9837 @item fontcolor_expr
9838 String which is expanded the same way as @var{text} to obtain dynamic
9839 @var{fontcolor} value. By default this option has empty value and is not
9840 processed. When this option is set, it overrides @var{fontcolor} option.
9843 The font family to be used for drawing text. By default Sans.
9846 The font file to be used for drawing text. The path must be included.
9847 This parameter is mandatory if the fontconfig support is disabled.
9850 Draw the text applying alpha blending. The value can
9851 be a number between 0.0 and 1.0.
9852 The expression accepts the same variables @var{x, y} as well.
9853 The default value is 1.
9854 Please see @var{fontcolor_expr}.
9857 The font size to be used for drawing text.
9858 The default value of @var{fontsize} is 16.
9861 If set to 1, attempt to shape the text (for example, reverse the order of
9862 right-to-left text and join Arabic characters) before drawing it.
9863 Otherwise, just draw the text exactly as given.
9864 By default 1 (if supported).
9867 The flags to be used for loading the fonts.
9869 The flags map the corresponding flags supported by libfreetype, and are
9870 a combination of the following values:
9877 @item vertical_layout
9878 @item force_autohint
9881 @item ignore_global_advance_width
9883 @item ignore_transform
9889 Default value is "default".
9891 For more information consult the documentation for the FT_LOAD_*
9895 The color to be used for drawing a shadow behind the drawn text. For the
9896 syntax of this option, check the @ref{color syntax,,"Color" section in the
9897 ffmpeg-utils manual,ffmpeg-utils}.
9899 The default value of @var{shadowcolor} is "black".
9903 The x and y offsets for the text shadow position with respect to the
9904 position of the text. They can be either positive or negative
9905 values. The default value for both is "0".
9908 The starting frame number for the n/frame_num variable. The default value
9912 The size in number of spaces to use for rendering the tab.
9916 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
9917 format. It can be used with or without text parameter. @var{timecode_rate}
9918 option must be specified.
9920 @item timecode_rate, rate, r
9921 Set the timecode frame rate (timecode only). Value will be rounded to nearest
9922 integer. Minimum value is "1".
9923 Drop-frame timecode is supported for frame rates 30 & 60.
9926 If set to 1, the output of the timecode option will wrap around at 24 hours.
9927 Default is 0 (disabled).
9930 The text string to be drawn. The text must be a sequence of UTF-8
9932 This parameter is mandatory if no file is specified with the parameter
9936 A text file containing text to be drawn. The text must be a sequence
9937 of UTF-8 encoded characters.
9939 This parameter is mandatory if no text string is specified with the
9940 parameter @var{text}.
9942 If both @var{text} and @var{textfile} are specified, an error is thrown.
9945 If set to 1, the @var{textfile} will be reloaded before each frame.
9946 Be sure to update it atomically, or it may be read partially, or even fail.
9950 The expressions which specify the offsets where text will be drawn
9951 within the video frame. They are relative to the top/left border of the
9954 The default value of @var{x} and @var{y} is "0".
9956 See below for the list of accepted constants and functions.
9959 The parameters for @var{x} and @var{y} are expressions containing the
9960 following constants and functions:
9964 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
9968 horizontal and vertical chroma subsample values. For example for the
9969 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9972 the height of each text line
9980 @item max_glyph_a, ascent
9981 the maximum distance from the baseline to the highest/upper grid
9982 coordinate used to place a glyph outline point, for all the rendered
9984 It is a positive value, due to the grid's orientation with the Y axis
9987 @item max_glyph_d, descent
9988 the maximum distance from the baseline to the lowest grid coordinate
9989 used to place a glyph outline point, for all the rendered glyphs.
9990 This is a negative value, due to the grid's orientation, with the Y axis
9994 maximum glyph height, that is the maximum height for all the glyphs
9995 contained in the rendered text, it is equivalent to @var{ascent} -
9999 maximum glyph width, that is the maximum width for all the glyphs
10000 contained in the rendered text
10003 the number of input frame, starting from 0
10005 @item rand(min, max)
10006 return a random number included between @var{min} and @var{max}
10009 The input sample aspect ratio.
10012 timestamp expressed in seconds, NAN if the input timestamp is unknown
10015 the height of the rendered text
10018 the width of the rendered text
10022 the x and y offset coordinates where the text is drawn.
10024 These parameters allow the @var{x} and @var{y} expressions to refer
10025 to each other, so you can for example specify @code{y=x/dar}.
10028 A one character description of the current frame's picture type.
10031 The current packet's position in the input file or stream
10032 (in bytes, from the start of the input). A value of -1 indicates
10033 this info is not available.
10036 The current packet's duration, in seconds.
10039 The current packet's size (in bytes).
10042 @anchor{drawtext_expansion}
10043 @subsection Text expansion
10045 If @option{expansion} is set to @code{strftime},
10046 the filter recognizes strftime() sequences in the provided text and
10047 expands them accordingly. Check the documentation of strftime(). This
10048 feature is deprecated.
10050 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10052 If @option{expansion} is set to @code{normal} (which is the default),
10053 the following expansion mechanism is used.
10055 The backslash character @samp{\}, followed by any character, always expands to
10056 the second character.
10058 Sequences of the form @code{%@{...@}} are expanded. The text between the
10059 braces is a function name, possibly followed by arguments separated by ':'.
10060 If the arguments contain special characters or delimiters (':' or '@}'),
10061 they should be escaped.
10063 Note that they probably must also be escaped as the value for the
10064 @option{text} option in the filter argument string and as the filter
10065 argument in the filtergraph description, and possibly also for the shell,
10066 that makes up to four levels of escaping; using a text file avoids these
10069 The following functions are available:
10074 The expression evaluation result.
10076 It must take one argument specifying the expression to be evaluated,
10077 which accepts the same constants and functions as the @var{x} and
10078 @var{y} values. Note that not all constants should be used, for
10079 example the text size is not known when evaluating the expression, so
10080 the constants @var{text_w} and @var{text_h} will have an undefined
10083 @item expr_int_format, eif
10084 Evaluate the expression's value and output as formatted integer.
10086 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10087 The second argument specifies the output format. Allowed values are @samp{x},
10088 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10089 @code{printf} function.
10090 The third parameter is optional and sets the number of positions taken by the output.
10091 It can be used to add padding with zeros from the left.
10094 The time at which the filter is running, expressed in UTC.
10095 It can accept an argument: a strftime() format string.
10098 The time at which the filter is running, expressed in the local time zone.
10099 It can accept an argument: a strftime() format string.
10102 Frame metadata. Takes one or two arguments.
10104 The first argument is mandatory and specifies the metadata key.
10106 The second argument is optional and specifies a default value, used when the
10107 metadata key is not found or empty.
10109 Available metadata can be identified by inspecting entries
10110 starting with TAG included within each frame section
10111 printed by running @code{ffprobe -show_frames}.
10113 String metadata generated in filters leading to
10114 the drawtext filter are also available.
10117 The frame number, starting from 0.
10120 A one character description of the current picture type.
10123 The timestamp of the current frame.
10124 It can take up to three arguments.
10126 The first argument is the format of the timestamp; it defaults to @code{flt}
10127 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10128 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10129 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10130 @code{localtime} stands for the timestamp of the frame formatted as
10131 local time zone time.
10133 The second argument is an offset added to the timestamp.
10135 If the format is set to @code{hms}, a third argument @code{24HH} may be
10136 supplied to present the hour part of the formatted timestamp in 24h format
10139 If the format is set to @code{localtime} or @code{gmtime},
10140 a third argument may be supplied: a strftime() format string.
10141 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10144 @subsection Commands
10146 This filter supports altering parameters via commands:
10149 Alter existing filter parameters.
10151 Syntax for the argument is the same as for filter invocation, e.g.
10154 fontsize=56:fontcolor=green:text='Hello World'
10157 Full filter invocation with sendcmd would look like this:
10160 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10164 If the entire argument can't be parsed or applied as valid values then the filter will
10165 continue with its existing parameters.
10167 @subsection Examples
10171 Draw "Test Text" with font FreeSerif, using the default values for the
10172 optional parameters.
10175 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10179 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10180 and y=50 (counting from the top-left corner of the screen), text is
10181 yellow with a red box around it. Both the text and the box have an
10185 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10186 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10189 Note that the double quotes are not necessary if spaces are not used
10190 within the parameter list.
10193 Show the text at the center of the video frame:
10195 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10199 Show the text at a random position, switching to a new position every 30 seconds:
10201 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)"
10205 Show a text line sliding from right to left in the last row of the video
10206 frame. The file @file{LONG_LINE} is assumed to contain a single line
10209 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10213 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10215 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10219 Draw a single green letter "g", at the center of the input video.
10220 The glyph baseline is placed at half screen height.
10222 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10226 Show text for 1 second every 3 seconds:
10228 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10232 Use fontconfig to set the font. Note that the colons need to be escaped.
10234 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10238 Draw "Test Text" with font size dependent on height of the video.
10240 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10244 Print the date of a real-time encoding (see strftime(3)):
10246 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10250 Show text fading in and out (appearing/disappearing):
10253 DS=1.0 # display start
10254 DE=10.0 # display end
10255 FID=1.5 # fade in duration
10256 FOD=5 # fade out duration
10257 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 @}"
10261 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10262 and the @option{fontsize} value are included in the @option{y} offset.
10264 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10265 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10269 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10270 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10271 must have option @option{-export_path_metadata 1} for the special metadata fields
10272 to be available for filters.
10274 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10279 For more information about libfreetype, check:
10280 @url{http://www.freetype.org/}.
10282 For more information about fontconfig, check:
10283 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10285 For more information about libfribidi, check:
10286 @url{http://fribidi.org/}.
10288 @section edgedetect
10290 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10292 The filter accepts the following options:
10297 Set low and high threshold values used by the Canny thresholding
10300 The high threshold selects the "strong" edge pixels, which are then
10301 connected through 8-connectivity with the "weak" edge pixels selected
10302 by the low threshold.
10304 @var{low} and @var{high} threshold values must be chosen in the range
10305 [0,1], and @var{low} should be lesser or equal to @var{high}.
10307 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10311 Define the drawing mode.
10315 Draw white/gray wires on black background.
10318 Mix the colors to create a paint/cartoon effect.
10321 Apply Canny edge detector on all selected planes.
10323 Default value is @var{wires}.
10326 Select planes for filtering. By default all available planes are filtered.
10329 @subsection Examples
10333 Standard edge detection with custom values for the hysteresis thresholding:
10335 edgedetect=low=0.1:high=0.4
10339 Painting effect without thresholding:
10341 edgedetect=mode=colormix:high=0
10347 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10349 For each input image, the filter will compute the optimal mapping from
10350 the input to the output given the codebook length, that is the number
10351 of distinct output colors.
10353 This filter accepts the following options.
10356 @item codebook_length, l
10357 Set codebook length. The value must be a positive integer, and
10358 represents the number of distinct output colors. Default value is 256.
10361 Set the maximum number of iterations to apply for computing the optimal
10362 mapping. The higher the value the better the result and the higher the
10363 computation time. Default value is 1.
10366 Set a random seed, must be an integer included between 0 and
10367 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10368 will try to use a good random seed on a best effort basis.
10371 Set pal8 output pixel format. This option does not work with codebook
10372 length greater than 256.
10377 Measure graylevel entropy in histogram of color channels of video frames.
10379 It accepts the following parameters:
10383 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10385 @var{diff} mode measures entropy of histogram delta values, absolute differences
10386 between neighbour histogram values.
10390 Set brightness, contrast, saturation and approximate gamma adjustment.
10392 The filter accepts the following options:
10396 Set the contrast expression. The value must be a float value in range
10397 @code{-1000.0} to @code{1000.0}. The default value is "1".
10400 Set the brightness expression. The value must be a float value in
10401 range @code{-1.0} to @code{1.0}. The default value is "0".
10404 Set the saturation expression. The value must be a float in
10405 range @code{0.0} to @code{3.0}. The default value is "1".
10408 Set the gamma expression. The value must be a float in range
10409 @code{0.1} to @code{10.0}. The default value is "1".
10412 Set the gamma expression for red. The value must be a float in
10413 range @code{0.1} to @code{10.0}. The default value is "1".
10416 Set the gamma expression for green. The value must be a float in range
10417 @code{0.1} to @code{10.0}. The default value is "1".
10420 Set the gamma expression for blue. The value must be a float in range
10421 @code{0.1} to @code{10.0}. The default value is "1".
10424 Set the gamma weight expression. It can be used to reduce the effect
10425 of a high gamma value on bright image areas, e.g. keep them from
10426 getting overamplified and just plain white. The value must be a float
10427 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10428 gamma correction all the way down while @code{1.0} leaves it at its
10429 full strength. Default is "1".
10432 Set when the expressions for brightness, contrast, saturation and
10433 gamma expressions are evaluated.
10435 It accepts the following values:
10438 only evaluate expressions once during the filter initialization or
10439 when a command is processed
10442 evaluate expressions for each incoming frame
10445 Default value is @samp{init}.
10448 The expressions accept the following parameters:
10451 frame count of the input frame starting from 0
10454 byte position of the corresponding packet in the input file, NAN if
10458 frame rate of the input video, NAN if the input frame rate is unknown
10461 timestamp expressed in seconds, NAN if the input timestamp is unknown
10464 @subsection Commands
10465 The filter supports the following commands:
10469 Set the contrast expression.
10472 Set the brightness expression.
10475 Set the saturation expression.
10478 Set the gamma expression.
10481 Set the gamma_r expression.
10484 Set gamma_g expression.
10487 Set gamma_b expression.
10490 Set gamma_weight expression.
10492 The command accepts the same syntax of the corresponding option.
10494 If the specified expression is not valid, it is kept at its current
10501 Apply erosion effect to the video.
10503 This filter replaces the pixel by the local(3x3) minimum.
10505 It accepts the following options:
10512 Limit the maximum change for each plane, default is 65535.
10513 If 0, plane will remain unchanged.
10516 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
10519 Flags to local 3x3 coordinates maps like this:
10526 @subsection Commands
10528 This filter supports the all above options as @ref{commands}.
10530 @section extractplanes
10532 Extract color channel components from input video stream into
10533 separate grayscale video streams.
10535 The filter accepts the following option:
10539 Set plane(s) to extract.
10541 Available values for planes are:
10552 Choosing planes not available in the input will result in an error.
10553 That means you cannot select @code{r}, @code{g}, @code{b} planes
10554 with @code{y}, @code{u}, @code{v} planes at same time.
10557 @subsection Examples
10561 Extract luma, u and v color channel component from input video frame
10562 into 3 grayscale outputs:
10564 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
10570 Apply a fade-in/out effect to the input video.
10572 It accepts the following parameters:
10576 The effect type can be either "in" for a fade-in, or "out" for a fade-out
10578 Default is @code{in}.
10580 @item start_frame, s
10581 Specify the number of the frame to start applying the fade
10582 effect at. Default is 0.
10585 The number of frames that the fade effect lasts. At the end of the
10586 fade-in effect, the output video will have the same intensity as the input video.
10587 At the end of the fade-out transition, the output video will be filled with the
10588 selected @option{color}.
10592 If set to 1, fade only alpha channel, if one exists on the input.
10593 Default value is 0.
10595 @item start_time, st
10596 Specify the timestamp (in seconds) of the frame to start to apply the fade
10597 effect. If both start_frame and start_time are specified, the fade will start at
10598 whichever comes last. Default is 0.
10601 The number of seconds for which the fade effect has to last. At the end of the
10602 fade-in effect the output video will have the same intensity as the input video,
10603 at the end of the fade-out transition the output video will be filled with the
10604 selected @option{color}.
10605 If both duration and nb_frames are specified, duration is used. Default is 0
10606 (nb_frames is used by default).
10609 Specify the color of the fade. Default is "black".
10612 @subsection Examples
10616 Fade in the first 30 frames of video:
10621 The command above is equivalent to:
10627 Fade out the last 45 frames of a 200-frame video:
10630 fade=type=out:start_frame=155:nb_frames=45
10634 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
10636 fade=in:0:25, fade=out:975:25
10640 Make the first 5 frames yellow, then fade in from frame 5-24:
10642 fade=in:5:20:color=yellow
10646 Fade in alpha over first 25 frames of video:
10648 fade=in:0:25:alpha=1
10652 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
10654 fade=t=in:st=5.5:d=0.5
10660 Denoise frames using 3D FFT (frequency domain filtering).
10662 The filter accepts the following options:
10666 Set the noise sigma constant. This sets denoising strength.
10667 Default value is 1. Allowed range is from 0 to 30.
10668 Using very high sigma with low overlap may give blocking artifacts.
10671 Set amount of denoising. By default all detected noise is reduced.
10672 Default value is 1. Allowed range is from 0 to 1.
10675 Set size of block, Default is 4, can be 3, 4, 5 or 6.
10676 Actual size of block in pixels is 2 to power of @var{block}, so by default
10677 block size in pixels is 2^4 which is 16.
10680 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
10683 Set number of previous frames to use for denoising. By default is set to 0.
10686 Set number of next frames to to use for denoising. By default is set to 0.
10689 Set planes which will be filtered, by default are all available filtered
10694 Apply arbitrary expressions to samples in frequency domain
10698 Adjust the dc value (gain) of the luma plane of the image. The filter
10699 accepts an integer value in range @code{0} to @code{1000}. The default
10700 value is set to @code{0}.
10703 Adjust the dc value (gain) of the 1st chroma plane of the image. The
10704 filter accepts an integer value in range @code{0} to @code{1000}. The
10705 default value is set to @code{0}.
10708 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
10709 filter accepts an integer value in range @code{0} to @code{1000}. The
10710 default value is set to @code{0}.
10713 Set the frequency domain weight expression for the luma plane.
10716 Set the frequency domain weight expression for the 1st chroma plane.
10719 Set the frequency domain weight expression for the 2nd chroma plane.
10722 Set when the expressions are evaluated.
10724 It accepts the following values:
10727 Only evaluate expressions once during the filter initialization.
10730 Evaluate expressions for each incoming frame.
10733 Default value is @samp{init}.
10735 The filter accepts the following variables:
10738 The coordinates of the current sample.
10742 The width and height of the image.
10745 The number of input frame, starting from 0.
10748 @subsection Examples
10754 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
10760 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
10766 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
10772 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
10779 Extract a single field from an interlaced image using stride
10780 arithmetic to avoid wasting CPU time. The output frames are marked as
10783 The filter accepts the following options:
10787 Specify whether to extract the top (if the value is @code{0} or
10788 @code{top}) or the bottom field (if the value is @code{1} or
10794 Create new frames by copying the top and bottom fields from surrounding frames
10795 supplied as numbers by the hint file.
10799 Set file containing hints: absolute/relative frame numbers.
10801 There must be one line for each frame in a clip. Each line must contain two
10802 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
10803 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
10804 is current frame number for @code{absolute} mode or out of [-1, 1] range
10805 for @code{relative} mode. First number tells from which frame to pick up top
10806 field and second number tells from which frame to pick up bottom field.
10808 If optionally followed by @code{+} output frame will be marked as interlaced,
10809 else if followed by @code{-} output frame will be marked as progressive, else
10810 it will be marked same as input frame.
10811 If optionally followed by @code{t} output frame will use only top field, or in
10812 case of @code{b} it will use only bottom field.
10813 If line starts with @code{#} or @code{;} that line is skipped.
10816 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
10819 Example of first several lines of @code{hint} file for @code{relative} mode:
10821 0,0 - # first frame
10822 1,0 - # second frame, use third's frame top field and second's frame bottom field
10823 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
10838 @section fieldmatch
10840 Field matching filter for inverse telecine. It is meant to reconstruct the
10841 progressive frames from a telecined stream. The filter does not drop duplicated
10842 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
10843 followed by a decimation filter such as @ref{decimate} in the filtergraph.
10845 The separation of the field matching and the decimation is notably motivated by
10846 the possibility of inserting a de-interlacing filter fallback between the two.
10847 If the source has mixed telecined and real interlaced content,
10848 @code{fieldmatch} will not be able to match fields for the interlaced parts.
10849 But these remaining combed frames will be marked as interlaced, and thus can be
10850 de-interlaced by a later filter such as @ref{yadif} before decimation.
10852 In addition to the various configuration options, @code{fieldmatch} can take an
10853 optional second stream, activated through the @option{ppsrc} option. If
10854 enabled, the frames reconstruction will be based on the fields and frames from
10855 this second stream. This allows the first input to be pre-processed in order to
10856 help the various algorithms of the filter, while keeping the output lossless
10857 (assuming the fields are matched properly). Typically, a field-aware denoiser,
10858 or brightness/contrast adjustments can help.
10860 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
10861 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
10862 which @code{fieldmatch} is based on. While the semantic and usage are very
10863 close, some behaviour and options names can differ.
10865 The @ref{decimate} filter currently only works for constant frame rate input.
10866 If your input has mixed telecined (30fps) and progressive content with a lower
10867 framerate like 24fps use the following filterchain to produce the necessary cfr
10868 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
10870 The filter accepts the following options:
10874 Specify the assumed field order of the input stream. Available values are:
10878 Auto detect parity (use FFmpeg's internal parity value).
10880 Assume bottom field first.
10882 Assume top field first.
10885 Note that it is sometimes recommended not to trust the parity announced by the
10888 Default value is @var{auto}.
10891 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
10892 sense that it won't risk creating jerkiness due to duplicate frames when
10893 possible, but if there are bad edits or blended fields it will end up
10894 outputting combed frames when a good match might actually exist. On the other
10895 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
10896 but will almost always find a good frame if there is one. The other values are
10897 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
10898 jerkiness and creating duplicate frames versus finding good matches in sections
10899 with bad edits, orphaned fields, blended fields, etc.
10901 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
10903 Available values are:
10907 2-way matching (p/c)
10909 2-way matching, and trying 3rd match if still combed (p/c + n)
10911 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
10913 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
10914 still combed (p/c + n + u/b)
10916 3-way matching (p/c/n)
10918 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
10919 detected as combed (p/c/n + u/b)
10922 The parenthesis at the end indicate the matches that would be used for that
10923 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
10926 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
10929 Default value is @var{pc_n}.
10932 Mark the main input stream as a pre-processed input, and enable the secondary
10933 input stream as the clean source to pick the fields from. See the filter
10934 introduction for more details. It is similar to the @option{clip2} feature from
10937 Default value is @code{0} (disabled).
10940 Set the field to match from. It is recommended to set this to the same value as
10941 @option{order} unless you experience matching failures with that setting. In
10942 certain circumstances changing the field that is used to match from can have a
10943 large impact on matching performance. Available values are:
10947 Automatic (same value as @option{order}).
10949 Match from the bottom field.
10951 Match from the top field.
10954 Default value is @var{auto}.
10957 Set whether or not chroma is included during the match comparisons. In most
10958 cases it is recommended to leave this enabled. You should set this to @code{0}
10959 only if your clip has bad chroma problems such as heavy rainbowing or other
10960 artifacts. Setting this to @code{0} could also be used to speed things up at
10961 the cost of some accuracy.
10963 Default value is @code{1}.
10967 These define an exclusion band which excludes the lines between @option{y0} and
10968 @option{y1} from being included in the field matching decision. An exclusion
10969 band can be used to ignore subtitles, a logo, or other things that may
10970 interfere with the matching. @option{y0} sets the starting scan line and
10971 @option{y1} sets the ending line; all lines in between @option{y0} and
10972 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
10973 @option{y0} and @option{y1} to the same value will disable the feature.
10974 @option{y0} and @option{y1} defaults to @code{0}.
10977 Set the scene change detection threshold as a percentage of maximum change on
10978 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
10979 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
10980 @option{scthresh} is @code{[0.0, 100.0]}.
10982 Default value is @code{12.0}.
10985 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
10986 account the combed scores of matches when deciding what match to use as the
10987 final match. Available values are:
10991 No final matching based on combed scores.
10993 Combed scores are only used when a scene change is detected.
10995 Use combed scores all the time.
10998 Default is @var{sc}.
11001 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11002 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11003 Available values are:
11007 No forced calculation.
11009 Force p/c/n calculations.
11011 Force p/c/n/u/b calculations.
11014 Default value is @var{none}.
11017 This is the area combing threshold used for combed frame detection. This
11018 essentially controls how "strong" or "visible" combing must be to be detected.
11019 Larger values mean combing must be more visible and smaller values mean combing
11020 can be less visible or strong and still be detected. Valid settings are from
11021 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11022 be detected as combed). This is basically a pixel difference value. A good
11023 range is @code{[8, 12]}.
11025 Default value is @code{9}.
11028 Sets whether or not chroma is considered in the combed frame decision. Only
11029 disable this if your source has chroma problems (rainbowing, etc.) that are
11030 causing problems for the combed frame detection with chroma enabled. Actually,
11031 using @option{chroma}=@var{0} is usually more reliable, except for the case
11032 where there is chroma only combing in the source.
11034 Default value is @code{0}.
11038 Respectively set the x-axis and y-axis size of the window used during combed
11039 frame detection. This has to do with the size of the area in which
11040 @option{combpel} pixels are required to be detected as combed for a frame to be
11041 declared combed. See the @option{combpel} parameter description for more info.
11042 Possible values are any number that is a power of 2 starting at 4 and going up
11045 Default value is @code{16}.
11048 The number of combed pixels inside any of the @option{blocky} by
11049 @option{blockx} size blocks on the frame for the frame to be detected as
11050 combed. While @option{cthresh} controls how "visible" the combing must be, this
11051 setting controls "how much" combing there must be in any localized area (a
11052 window defined by the @option{blockx} and @option{blocky} settings) on the
11053 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11054 which point no frames will ever be detected as combed). This setting is known
11055 as @option{MI} in TFM/VFM vocabulary.
11057 Default value is @code{80}.
11060 @anchor{p/c/n/u/b meaning}
11061 @subsection p/c/n/u/b meaning
11063 @subsubsection p/c/n
11065 We assume the following telecined stream:
11068 Top fields: 1 2 2 3 4
11069 Bottom fields: 1 2 3 4 4
11072 The numbers correspond to the progressive frame the fields relate to. Here, the
11073 first two frames are progressive, the 3rd and 4th are combed, and so on.
11075 When @code{fieldmatch} is configured to run a matching from bottom
11076 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11081 B 1 2 3 4 4 <-- matching reference
11090 As a result of the field matching, we can see that some frames get duplicated.
11091 To perform a complete inverse telecine, you need to rely on a decimation filter
11092 after this operation. See for instance the @ref{decimate} filter.
11094 The same operation now matching from top fields (@option{field}=@var{top})
11099 T 1 2 2 3 4 <-- matching reference
11109 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11110 basically, they refer to the frame and field of the opposite parity:
11113 @item @var{p} matches the field of the opposite parity in the previous frame
11114 @item @var{c} matches the field of the opposite parity in the current frame
11115 @item @var{n} matches the field of the opposite parity in the next frame
11120 The @var{u} and @var{b} matching are a bit special in the sense that they match
11121 from the opposite parity flag. In the following examples, we assume that we are
11122 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11123 'x' is placed above and below each matched fields.
11125 With bottom matching (@option{field}=@var{bottom}):
11130 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11131 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11139 With top matching (@option{field}=@var{top}):
11144 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11145 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11153 @subsection Examples
11155 Simple IVTC of a top field first telecined stream:
11157 fieldmatch=order=tff:combmatch=none, decimate
11160 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11162 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11165 @section fieldorder
11167 Transform the field order of the input video.
11169 It accepts the following parameters:
11174 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11175 for bottom field first.
11178 The default value is @samp{tff}.
11180 The transformation is done by shifting the picture content up or down
11181 by one line, and filling the remaining line with appropriate picture content.
11182 This method is consistent with most broadcast field order converters.
11184 If the input video is not flagged as being interlaced, or it is already
11185 flagged as being of the required output field order, then this filter does
11186 not alter the incoming video.
11188 It is very useful when converting to or from PAL DV material,
11189 which is bottom field first.
11193 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11196 @section fifo, afifo
11198 Buffer input images and send them when they are requested.
11200 It is mainly useful when auto-inserted by the libavfilter
11203 It does not take parameters.
11205 @section fillborders
11207 Fill borders of the input video, without changing video stream dimensions.
11208 Sometimes video can have garbage at the four edges and you may not want to
11209 crop video input to keep size multiple of some number.
11211 This filter accepts the following options:
11215 Number of pixels to fill from left border.
11218 Number of pixels to fill from right border.
11221 Number of pixels to fill from top border.
11224 Number of pixels to fill from bottom border.
11229 It accepts the following values:
11232 fill pixels using outermost pixels
11235 fill pixels using mirroring
11238 fill pixels with constant value
11241 Default is @var{smear}.
11244 Set color for pixels in fixed mode. Default is @var{black}.
11247 @subsection Commands
11248 This filter supports same @ref{commands} as options.
11249 The command accepts the same syntax of the corresponding option.
11251 If the specified expression is not valid, it is kept at its current
11256 Find a rectangular object
11258 It accepts the following options:
11262 Filepath of the object image, needs to be in gray8.
11265 Detection threshold, default is 0.5.
11268 Number of mipmaps, default is 3.
11270 @item xmin, ymin, xmax, ymax
11271 Specifies the rectangle in which to search.
11274 @subsection Examples
11278 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11280 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11286 Flood area with values of same pixel components with another values.
11288 It accepts the following options:
11291 Set pixel x coordinate.
11294 Set pixel y coordinate.
11297 Set source #0 component value.
11300 Set source #1 component value.
11303 Set source #2 component value.
11306 Set source #3 component value.
11309 Set destination #0 component value.
11312 Set destination #1 component value.
11315 Set destination #2 component value.
11318 Set destination #3 component value.
11324 Convert the input video to one of the specified pixel formats.
11325 Libavfilter will try to pick one that is suitable as input to
11328 It accepts the following parameters:
11332 A '|'-separated list of pixel format names, such as
11333 "pix_fmts=yuv420p|monow|rgb24".
11337 @subsection Examples
11341 Convert the input video to the @var{yuv420p} format
11343 format=pix_fmts=yuv420p
11346 Convert the input video to any of the formats in the list
11348 format=pix_fmts=yuv420p|yuv444p|yuv410p
11355 Convert the video to specified constant frame rate by duplicating or dropping
11356 frames as necessary.
11358 It accepts the following parameters:
11362 The desired output frame rate. The default is @code{25}.
11365 Assume the first PTS should be the given value, in seconds. This allows for
11366 padding/trimming at the start of stream. By default, no assumption is made
11367 about the first frame's expected PTS, so no padding or trimming is done.
11368 For example, this could be set to 0 to pad the beginning with duplicates of
11369 the first frame if a video stream starts after the audio stream or to trim any
11370 frames with a negative PTS.
11373 Timestamp (PTS) rounding method.
11375 Possible values are:
11382 round towards -infinity
11384 round towards +infinity
11388 The default is @code{near}.
11391 Action performed when reading the last frame.
11393 Possible values are:
11396 Use same timestamp rounding method as used for other frames.
11398 Pass through last frame if input duration has not been reached yet.
11400 The default is @code{round}.
11404 Alternatively, the options can be specified as a flat string:
11405 @var{fps}[:@var{start_time}[:@var{round}]].
11407 See also the @ref{setpts} filter.
11409 @subsection Examples
11413 A typical usage in order to set the fps to 25:
11419 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11421 fps=fps=film:round=near
11427 Pack two different video streams into a stereoscopic video, setting proper
11428 metadata on supported codecs. The two views should have the same size and
11429 framerate and processing will stop when the shorter video ends. Please note
11430 that you may conveniently adjust view properties with the @ref{scale} and
11433 It accepts the following parameters:
11437 The desired packing format. Supported values are:
11442 The views are next to each other (default).
11445 The views are on top of each other.
11448 The views are packed by line.
11451 The views are packed by column.
11454 The views are temporally interleaved.
11463 # Convert left and right views into a frame-sequential video
11464 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11466 # Convert views into a side-by-side video with the same output resolution as the input
11467 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
11472 Change the frame rate by interpolating new video output frames from the source
11475 This filter is not designed to function correctly with interlaced media. If
11476 you wish to change the frame rate of interlaced media then you are required
11477 to deinterlace before this filter and re-interlace after this filter.
11479 A description of the accepted options follows.
11483 Specify the output frames per second. This option can also be specified
11484 as a value alone. The default is @code{50}.
11487 Specify the start of a range where the output frame will be created as a
11488 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11489 the default is @code{15}.
11492 Specify the end of a range where the output frame will be created as a
11493 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11494 the default is @code{240}.
11497 Specify the level at which a scene change is detected as a value between
11498 0 and 100 to indicate a new scene; a low value reflects a low
11499 probability for the current frame to introduce a new scene, while a higher
11500 value means the current frame is more likely to be one.
11501 The default is @code{8.2}.
11504 Specify flags influencing the filter process.
11506 Available value for @var{flags} is:
11509 @item scene_change_detect, scd
11510 Enable scene change detection using the value of the option @var{scene}.
11511 This flag is enabled by default.
11517 Select one frame every N-th frame.
11519 This filter accepts the following option:
11522 Select frame after every @code{step} frames.
11523 Allowed values are positive integers higher than 0. Default value is @code{1}.
11526 @section freezedetect
11528 Detect frozen video.
11530 This filter logs a message and sets frame metadata when it detects that the
11531 input video has no significant change in content during a specified duration.
11532 Video freeze detection calculates the mean average absolute difference of all
11533 the components of video frames and compares it to a noise floor.
11535 The printed times and duration are expressed in seconds. The
11536 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
11537 whose timestamp equals or exceeds the detection duration and it contains the
11538 timestamp of the first frame of the freeze. The
11539 @code{lavfi.freezedetect.freeze_duration} and
11540 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
11543 The filter accepts the following options:
11547 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
11548 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
11552 Set freeze duration until notification (default is 2 seconds).
11555 @section freezeframes
11557 Freeze video frames.
11559 This filter freezes video frames using frame from 2nd input.
11561 The filter accepts the following options:
11565 Set number of first frame from which to start freeze.
11568 Set number of last frame from which to end freeze.
11571 Set number of frame from 2nd input which will be used instead of replaced frames.
11577 Apply a frei0r effect to the input video.
11579 To enable the compilation of this filter, you need to install the frei0r
11580 header and configure FFmpeg with @code{--enable-frei0r}.
11582 It accepts the following parameters:
11587 The name of the frei0r effect to load. If the environment variable
11588 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
11589 directories specified by the colon-separated list in @env{FREI0R_PATH}.
11590 Otherwise, the standard frei0r paths are searched, in this order:
11591 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
11592 @file{/usr/lib/frei0r-1/}.
11594 @item filter_params
11595 A '|'-separated list of parameters to pass to the frei0r effect.
11599 A frei0r effect parameter can be a boolean (its value is either
11600 "y" or "n"), a double, a color (specified as
11601 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
11602 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
11603 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
11604 a position (specified as @var{X}/@var{Y}, where
11605 @var{X} and @var{Y} are floating point numbers) and/or a string.
11607 The number and types of parameters depend on the loaded effect. If an
11608 effect parameter is not specified, the default value is set.
11610 @subsection Examples
11614 Apply the distort0r effect, setting the first two double parameters:
11616 frei0r=filter_name=distort0r:filter_params=0.5|0.01
11620 Apply the colordistance effect, taking a color as the first parameter:
11622 frei0r=colordistance:0.2/0.3/0.4
11623 frei0r=colordistance:violet
11624 frei0r=colordistance:0x112233
11628 Apply the perspective effect, specifying the top left and top right image
11631 frei0r=perspective:0.2/0.2|0.8/0.2
11635 For more information, see
11636 @url{http://frei0r.dyne.org}
11640 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
11642 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
11643 processing filter, one of them is performed once per block, not per pixel.
11644 This allows for much higher speed.
11646 The filter accepts the following options:
11650 Set quality. This option defines the number of levels for averaging. It accepts
11651 an integer in the range 4-5. Default value is @code{4}.
11654 Force a constant quantization parameter. It accepts an integer in range 0-63.
11655 If not set, the filter will use the QP from the video stream (if available).
11658 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
11659 more details but also more artifacts, while higher values make the image smoother
11660 but also blurrier. Default value is @code{0} − PSNR optimal.
11662 @item use_bframe_qp
11663 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
11664 option may cause flicker since the B-Frames have often larger QP. Default is
11665 @code{0} (not enabled).
11671 Apply Gaussian blur filter.
11673 The filter accepts the following options:
11677 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
11680 Set number of steps for Gaussian approximation. Default is @code{1}.
11683 Set which planes to filter. By default all planes are filtered.
11686 Set vertical sigma, if negative it will be same as @code{sigma}.
11687 Default is @code{-1}.
11690 @subsection Commands
11691 This filter supports same commands as options.
11692 The command accepts the same syntax of the corresponding option.
11694 If the specified expression is not valid, it is kept at its current
11699 Apply generic equation to each pixel.
11701 The filter accepts the following options:
11704 @item lum_expr, lum
11705 Set the luminance expression.
11707 Set the chrominance blue expression.
11709 Set the chrominance red expression.
11710 @item alpha_expr, a
11711 Set the alpha expression.
11713 Set the red expression.
11714 @item green_expr, g
11715 Set the green expression.
11717 Set the blue expression.
11720 The colorspace is selected according to the specified options. If one
11721 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
11722 options is specified, the filter will automatically select a YCbCr
11723 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
11724 @option{blue_expr} options is specified, it will select an RGB
11727 If one of the chrominance expression is not defined, it falls back on the other
11728 one. If no alpha expression is specified it will evaluate to opaque value.
11729 If none of chrominance expressions are specified, they will evaluate
11730 to the luminance expression.
11732 The expressions can use the following variables and functions:
11736 The sequential number of the filtered frame, starting from @code{0}.
11740 The coordinates of the current sample.
11744 The width and height of the image.
11748 Width and height scale depending on the currently filtered plane. It is the
11749 ratio between the corresponding luma plane number of pixels and the current
11750 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
11751 @code{0.5,0.5} for chroma planes.
11754 Time of the current frame, expressed in seconds.
11757 Return the value of the pixel at location (@var{x},@var{y}) of the current
11761 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
11765 Return the value of the pixel at location (@var{x},@var{y}) of the
11766 blue-difference chroma plane. Return 0 if there is no such plane.
11769 Return the value of the pixel at location (@var{x},@var{y}) of the
11770 red-difference chroma plane. Return 0 if there is no such plane.
11775 Return the value of the pixel at location (@var{x},@var{y}) of the
11776 red/green/blue component. Return 0 if there is no such component.
11779 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
11780 plane. Return 0 if there is no such plane.
11782 @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)
11783 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
11784 sums of samples within a rectangle. See the functions without the sum postfix.
11786 @item interpolation
11787 Set one of interpolation methods:
11792 Default is bilinear.
11795 For functions, if @var{x} and @var{y} are outside the area, the value will be
11796 automatically clipped to the closer edge.
11798 Please note that this filter can use multiple threads in which case each slice
11799 will have its own expression state. If you want to use only a single expression
11800 state because your expressions depend on previous state then you should limit
11801 the number of filter threads to 1.
11803 @subsection Examples
11807 Flip the image horizontally:
11813 Generate a bidimensional sine wave, with angle @code{PI/3} and a
11814 wavelength of 100 pixels:
11816 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
11820 Generate a fancy enigmatic moving light:
11822 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
11826 Generate a quick emboss effect:
11828 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
11832 Modify RGB components depending on pixel position:
11834 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
11838 Create a radial gradient that is the same size as the input (also see
11839 the @ref{vignette} filter):
11841 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
11847 Fix the banding artifacts that are sometimes introduced into nearly flat
11848 regions by truncation to 8-bit color depth.
11849 Interpolate the gradients that should go where the bands are, and
11852 It is designed for playback only. Do not use it prior to
11853 lossy compression, because compression tends to lose the dither and
11854 bring back the bands.
11856 It accepts the following parameters:
11861 The maximum amount by which the filter will change any one pixel. This is also
11862 the threshold for detecting nearly flat regions. Acceptable values range from
11863 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
11867 The neighborhood to fit the gradient to. A larger radius makes for smoother
11868 gradients, but also prevents the filter from modifying the pixels near detailed
11869 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
11870 values will be clipped to the valid range.
11874 Alternatively, the options can be specified as a flat string:
11875 @var{strength}[:@var{radius}]
11877 @subsection Examples
11881 Apply the filter with a @code{3.5} strength and radius of @code{8}:
11887 Specify radius, omitting the strength (which will fall-back to the default
11895 @anchor{graphmonitor}
11896 @section graphmonitor
11897 Show various filtergraph stats.
11899 With this filter one can debug complete filtergraph.
11900 Especially issues with links filling with queued frames.
11902 The filter accepts the following options:
11906 Set video output size. Default is @var{hd720}.
11909 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
11912 Set output mode, can be @var{fulll} or @var{compact}.
11913 In @var{compact} mode only filters with some queued frames have displayed stats.
11916 Set flags which enable which stats are shown in video.
11918 Available values for flags are:
11921 Display number of queued frames in each link.
11923 @item frame_count_in
11924 Display number of frames taken from filter.
11926 @item frame_count_out
11927 Display number of frames given out from filter.
11930 Display current filtered frame pts.
11933 Display current filtered frame time.
11936 Display time base for filter link.
11939 Display used format for filter link.
11942 Display video size or number of audio channels in case of audio used by filter link.
11945 Display video frame rate or sample rate in case of audio used by filter link.
11948 Display link output status.
11952 Set upper limit for video rate of output stream, Default value is @var{25}.
11953 This guarantee that output video frame rate will not be higher than this value.
11957 A color constancy variation filter which estimates scene illumination via grey edge algorithm
11958 and corrects the scene colors accordingly.
11960 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
11962 The filter accepts the following options:
11966 The order of differentiation to be applied on the scene. Must be chosen in the range
11967 [0,2] and default value is 1.
11970 The Minkowski parameter to be used for calculating the Minkowski distance. Must
11971 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
11972 max value instead of calculating Minkowski distance.
11975 The standard deviation of Gaussian blur to be applied on the scene. Must be
11976 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
11977 can't be equal to 0 if @var{difford} is greater than 0.
11980 @subsection Examples
11986 greyedge=difford=1:minknorm=5:sigma=2
11992 greyedge=difford=1:minknorm=0:sigma=2
12000 Apply a Hald CLUT to a video stream.
12002 First input is the video stream to process, and second one is the Hald CLUT.
12003 The Hald CLUT input can be a simple picture or a complete video stream.
12005 The filter accepts the following options:
12009 Force termination when the shortest input terminates. Default is @code{0}.
12011 Continue applying the last CLUT after the end of the stream. A value of
12012 @code{0} disable the filter after the last frame of the CLUT is reached.
12013 Default is @code{1}.
12016 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12017 filters share the same internals).
12019 This filter also supports the @ref{framesync} options.
12021 More information about the Hald CLUT can be found on Eskil Steenberg's website
12022 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12024 @subsection Workflow examples
12026 @subsubsection Hald CLUT video stream
12028 Generate an identity Hald CLUT stream altered with various effects:
12030 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
12033 Note: make sure you use a lossless codec.
12035 Then use it with @code{haldclut} to apply it on some random stream:
12037 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12040 The Hald CLUT will be applied to the 10 first seconds (duration of
12041 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12042 to the remaining frames of the @code{mandelbrot} stream.
12044 @subsubsection Hald CLUT with preview
12046 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12047 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12048 biggest possible square starting at the top left of the picture. The remaining
12049 padding pixels (bottom or right) will be ignored. This area can be used to add
12050 a preview of the Hald CLUT.
12052 Typically, the following generated Hald CLUT will be supported by the
12053 @code{haldclut} filter:
12056 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12057 pad=iw+320 [padded_clut];
12058 smptebars=s=320x256, split [a][b];
12059 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12060 [main][b] overlay=W-320" -frames:v 1 clut.png
12063 It contains the original and a preview of the effect of the CLUT: SMPTE color
12064 bars are displayed on the right-top, and below the same color bars processed by
12067 Then, the effect of this Hald CLUT can be visualized with:
12069 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12074 Flip the input video horizontally.
12076 For example, to horizontally flip the input video with @command{ffmpeg}:
12078 ffmpeg -i in.avi -vf "hflip" out.avi
12082 This filter applies a global color histogram equalization on a
12085 It can be used to correct video that has a compressed range of pixel
12086 intensities. The filter redistributes the pixel intensities to
12087 equalize their distribution across the intensity range. It may be
12088 viewed as an "automatically adjusting contrast filter". This filter is
12089 useful only for correcting degraded or poorly captured source
12092 The filter accepts the following options:
12096 Determine the amount of equalization to be applied. As the strength
12097 is reduced, the distribution of pixel intensities more-and-more
12098 approaches that of the input frame. The value must be a float number
12099 in the range [0,1] and defaults to 0.200.
12102 Set the maximum intensity that can generated and scale the output
12103 values appropriately. The strength should be set as desired and then
12104 the intensity can be limited if needed to avoid washing-out. The value
12105 must be a float number in the range [0,1] and defaults to 0.210.
12108 Set the antibanding level. If enabled the filter will randomly vary
12109 the luminance of output pixels by a small amount to avoid banding of
12110 the histogram. Possible values are @code{none}, @code{weak} or
12111 @code{strong}. It defaults to @code{none}.
12117 Compute and draw a color distribution histogram for the input video.
12119 The computed histogram is a representation of the color component
12120 distribution in an image.
12122 Standard histogram displays the color components distribution in an image.
12123 Displays color graph for each color component. Shows distribution of
12124 the Y, U, V, A or R, G, B components, depending on input format, in the
12125 current frame. Below each graph a color component scale meter is shown.
12127 The filter accepts the following options:
12131 Set height of level. Default value is @code{200}.
12132 Allowed range is [50, 2048].
12135 Set height of color scale. Default value is @code{12}.
12136 Allowed range is [0, 40].
12140 It accepts the following values:
12143 Per color component graphs are placed below each other.
12146 Per color component graphs are placed side by side.
12149 Presents information identical to that in the @code{parade}, except
12150 that the graphs representing color components are superimposed directly
12153 Default is @code{stack}.
12156 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12157 Default is @code{linear}.
12160 Set what color components to display.
12161 Default is @code{7}.
12164 Set foreground opacity. Default is @code{0.7}.
12167 Set background opacity. Default is @code{0.5}.
12170 @subsection Examples
12175 Calculate and draw histogram:
12177 ffplay -i input -vf histogram
12185 This is a high precision/quality 3d denoise filter. It aims to reduce
12186 image noise, producing smooth images and making still images really
12187 still. It should enhance compressibility.
12189 It accepts the following optional parameters:
12193 A non-negative floating point number which specifies spatial luma strength.
12194 It defaults to 4.0.
12196 @item chroma_spatial
12197 A non-negative floating point number which specifies spatial chroma strength.
12198 It defaults to 3.0*@var{luma_spatial}/4.0.
12201 A floating point number which specifies luma temporal strength. It defaults to
12202 6.0*@var{luma_spatial}/4.0.
12205 A floating point number which specifies chroma temporal strength. It defaults to
12206 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12209 @subsection Commands
12210 This filter supports same @ref{commands} as options.
12211 The command accepts the same syntax of the corresponding option.
12213 If the specified expression is not valid, it is kept at its current
12216 @anchor{hwdownload}
12217 @section hwdownload
12219 Download hardware frames to system memory.
12221 The input must be in hardware frames, and the output a non-hardware format.
12222 Not all formats will be supported on the output - it may be necessary to insert
12223 an additional @option{format} filter immediately following in the graph to get
12224 the output in a supported format.
12228 Map hardware frames to system memory or to another device.
12230 This filter has several different modes of operation; which one is used depends
12231 on the input and output formats:
12234 Hardware frame input, normal frame output
12236 Map the input frames to system memory and pass them to the output. If the
12237 original hardware frame is later required (for example, after overlaying
12238 something else on part of it), the @option{hwmap} filter can be used again
12239 in the next mode to retrieve it.
12241 Normal frame input, hardware frame output
12243 If the input is actually a software-mapped hardware frame, then unmap it -
12244 that is, return the original hardware frame.
12246 Otherwise, a device must be provided. Create new hardware surfaces on that
12247 device for the output, then map them back to the software format at the input
12248 and give those frames to the preceding filter. This will then act like the
12249 @option{hwupload} filter, but may be able to avoid an additional copy when
12250 the input is already in a compatible format.
12252 Hardware frame input and output
12254 A device must be supplied for the output, either directly or with the
12255 @option{derive_device} option. The input and output devices must be of
12256 different types and compatible - the exact meaning of this is
12257 system-dependent, but typically it means that they must refer to the same
12258 underlying hardware context (for example, refer to the same graphics card).
12260 If the input frames were originally created on the output device, then unmap
12261 to retrieve the original frames.
12263 Otherwise, map the frames to the output device - create new hardware frames
12264 on the output corresponding to the frames on the input.
12267 The following additional parameters are accepted:
12271 Set the frame mapping mode. Some combination of:
12274 The mapped frame should be readable.
12276 The mapped frame should be writeable.
12278 The mapping will always overwrite the entire frame.
12280 This may improve performance in some cases, as the original contents of the
12281 frame need not be loaded.
12283 The mapping must not involve any copying.
12285 Indirect mappings to copies of frames are created in some cases where either
12286 direct mapping is not possible or it would have unexpected properties.
12287 Setting this flag ensures that the mapping is direct and will fail if that is
12290 Defaults to @var{read+write} if not specified.
12292 @item derive_device @var{type}
12293 Rather than using the device supplied at initialisation, instead derive a new
12294 device of type @var{type} from the device the input frames exist on.
12297 In a hardware to hardware mapping, map in reverse - create frames in the sink
12298 and map them back to the source. This may be necessary in some cases where
12299 a mapping in one direction is required but only the opposite direction is
12300 supported by the devices being used.
12302 This option is dangerous - it may break the preceding filter in undefined
12303 ways if there are any additional constraints on that filter's output.
12304 Do not use it without fully understanding the implications of its use.
12310 Upload system memory frames to hardware surfaces.
12312 The device to upload to must be supplied when the filter is initialised. If
12313 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12314 option or with the @option{derive_device} option. The input and output devices
12315 must be of different types and compatible - the exact meaning of this is
12316 system-dependent, but typically it means that they must refer to the same
12317 underlying hardware context (for example, refer to the same graphics card).
12319 The following additional parameters are accepted:
12322 @item derive_device @var{type}
12323 Rather than using the device supplied at initialisation, instead derive a new
12324 device of type @var{type} from the device the input frames exist on.
12327 @anchor{hwupload_cuda}
12328 @section hwupload_cuda
12330 Upload system memory frames to a CUDA device.
12332 It accepts the following optional parameters:
12336 The number of the CUDA device to use
12341 Apply a high-quality magnification filter designed for pixel art. This filter
12342 was originally created by Maxim Stepin.
12344 It accepts the following option:
12348 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12349 @code{hq3x} and @code{4} for @code{hq4x}.
12350 Default is @code{3}.
12354 Stack input videos horizontally.
12356 All streams must be of same pixel format and of same height.
12358 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12359 to create same output.
12361 The filter accepts the following option:
12365 Set number of input streams. Default is 2.
12368 If set to 1, force the output to terminate when the shortest input
12369 terminates. Default value is 0.
12374 Modify the hue and/or the saturation of the input.
12376 It accepts the following parameters:
12380 Specify the hue angle as a number of degrees. It accepts an expression,
12381 and defaults to "0".
12384 Specify the saturation in the [-10,10] range. It accepts an expression and
12388 Specify the hue angle as a number of radians. It accepts an
12389 expression, and defaults to "0".
12392 Specify the brightness in the [-10,10] range. It accepts an expression and
12396 @option{h} and @option{H} are mutually exclusive, and can't be
12397 specified at the same time.
12399 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12400 expressions containing the following constants:
12404 frame count of the input frame starting from 0
12407 presentation timestamp of the input frame expressed in time base units
12410 frame rate of the input video, NAN if the input frame rate is unknown
12413 timestamp expressed in seconds, NAN if the input timestamp is unknown
12416 time base of the input video
12419 @subsection Examples
12423 Set the hue to 90 degrees and the saturation to 1.0:
12429 Same command but expressing the hue in radians:
12435 Rotate hue and make the saturation swing between 0
12436 and 2 over a period of 1 second:
12438 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12442 Apply a 3 seconds saturation fade-in effect starting at 0:
12444 hue="s=min(t/3\,1)"
12447 The general fade-in expression can be written as:
12449 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12453 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12455 hue="s=max(0\, min(1\, (8-t)/3))"
12458 The general fade-out expression can be written as:
12460 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12465 @subsection Commands
12467 This filter supports the following commands:
12473 Modify the hue and/or the saturation and/or brightness of the input video.
12474 The command accepts the same syntax of the corresponding option.
12476 If the specified expression is not valid, it is kept at its current
12480 @section hysteresis
12482 Grow first stream into second stream by connecting components.
12483 This makes it possible to build more robust edge masks.
12485 This filter accepts the following options:
12489 Set which planes will be processed as bitmap, unprocessed planes will be
12490 copied from first stream.
12491 By default value 0xf, all planes will be processed.
12494 Set threshold which is used in filtering. If pixel component value is higher than
12495 this value filter algorithm for connecting components is activated.
12496 By default value is 0.
12499 The @code{hysteresis} filter also supports the @ref{framesync} options.
12503 Detect video interlacing type.
12505 This filter tries to detect if the input frames are interlaced, progressive,
12506 top or bottom field first. It will also try to detect fields that are
12507 repeated between adjacent frames (a sign of telecine).
12509 Single frame detection considers only immediately adjacent frames when classifying each frame.
12510 Multiple frame detection incorporates the classification history of previous frames.
12512 The filter will log these metadata values:
12515 @item single.current_frame
12516 Detected type of current frame using single-frame detection. One of:
12517 ``tff'' (top field first), ``bff'' (bottom field first),
12518 ``progressive'', or ``undetermined''
12521 Cumulative number of frames detected as top field first using single-frame detection.
12524 Cumulative number of frames detected as top field first using multiple-frame detection.
12527 Cumulative number of frames detected as bottom field first using single-frame detection.
12529 @item multiple.current_frame
12530 Detected type of current frame using multiple-frame detection. One of:
12531 ``tff'' (top field first), ``bff'' (bottom field first),
12532 ``progressive'', or ``undetermined''
12535 Cumulative number of frames detected as bottom field first using multiple-frame detection.
12537 @item single.progressive
12538 Cumulative number of frames detected as progressive using single-frame detection.
12540 @item multiple.progressive
12541 Cumulative number of frames detected as progressive using multiple-frame detection.
12543 @item single.undetermined
12544 Cumulative number of frames that could not be classified using single-frame detection.
12546 @item multiple.undetermined
12547 Cumulative number of frames that could not be classified using multiple-frame detection.
12549 @item repeated.current_frame
12550 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
12552 @item repeated.neither
12553 Cumulative number of frames with no repeated field.
12556 Cumulative number of frames with the top field repeated from the previous frame's top field.
12558 @item repeated.bottom
12559 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
12562 The filter accepts the following options:
12566 Set interlacing threshold.
12568 Set progressive threshold.
12570 Threshold for repeated field detection.
12572 Number of frames after which a given frame's contribution to the
12573 statistics is halved (i.e., it contributes only 0.5 to its
12574 classification). The default of 0 means that all frames seen are given
12575 full weight of 1.0 forever.
12576 @item analyze_interlaced_flag
12577 When this is not 0 then idet will use the specified number of frames to determine
12578 if the interlaced flag is accurate, it will not count undetermined frames.
12579 If the flag is found to be accurate it will be used without any further
12580 computations, if it is found to be inaccurate it will be cleared without any
12581 further computations. This allows inserting the idet filter as a low computational
12582 method to clean up the interlaced flag
12587 Deinterleave or interleave fields.
12589 This filter allows one to process interlaced images fields without
12590 deinterlacing them. Deinterleaving splits the input frame into 2
12591 fields (so called half pictures). Odd lines are moved to the top
12592 half of the output image, even lines to the bottom half.
12593 You can process (filter) them independently and then re-interleave them.
12595 The filter accepts the following options:
12599 @item chroma_mode, c
12600 @item alpha_mode, a
12601 Available values for @var{luma_mode}, @var{chroma_mode} and
12602 @var{alpha_mode} are:
12608 @item deinterleave, d
12609 Deinterleave fields, placing one above the other.
12611 @item interleave, i
12612 Interleave fields. Reverse the effect of deinterleaving.
12614 Default value is @code{none}.
12616 @item luma_swap, ls
12617 @item chroma_swap, cs
12618 @item alpha_swap, as
12619 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
12622 @subsection Commands
12624 This filter supports the all above options as @ref{commands}.
12628 Apply inflate effect to the video.
12630 This filter replaces the pixel by the local(3x3) average by taking into account
12631 only values higher than the pixel.
12633 It accepts the following options:
12640 Limit the maximum change for each plane, default is 65535.
12641 If 0, plane will remain unchanged.
12644 @subsection Commands
12646 This filter supports the all above options as @ref{commands}.
12650 Simple interlacing filter from progressive contents. This interleaves upper (or
12651 lower) lines from odd frames with lower (or upper) lines from even frames,
12652 halving the frame rate and preserving image height.
12655 Original Original New Frame
12656 Frame 'j' Frame 'j+1' (tff)
12657 ========== =========== ==================
12658 Line 0 --------------------> Frame 'j' Line 0
12659 Line 1 Line 1 ----> Frame 'j+1' Line 1
12660 Line 2 ---------------------> Frame 'j' Line 2
12661 Line 3 Line 3 ----> Frame 'j+1' Line 3
12663 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
12666 It accepts the following optional parameters:
12670 This determines whether the interlaced frame is taken from the even
12671 (tff - default) or odd (bff) lines of the progressive frame.
12674 Vertical lowpass filter to avoid twitter interlacing and
12675 reduce moire patterns.
12679 Disable vertical lowpass filter
12682 Enable linear filter (default)
12685 Enable complex filter. This will slightly less reduce twitter and moire
12686 but better retain detail and subjective sharpness impression.
12693 Deinterlace input video by applying Donald Graft's adaptive kernel
12694 deinterling. Work on interlaced parts of a video to produce
12695 progressive frames.
12697 The description of the accepted parameters follows.
12701 Set the threshold which affects the filter's tolerance when
12702 determining if a pixel line must be processed. It must be an integer
12703 in the range [0,255] and defaults to 10. A value of 0 will result in
12704 applying the process on every pixels.
12707 Paint pixels exceeding the threshold value to white if set to 1.
12711 Set the fields order. Swap fields if set to 1, leave fields alone if
12715 Enable additional sharpening if set to 1. Default is 0.
12718 Enable twoway sharpening if set to 1. Default is 0.
12721 @subsection Examples
12725 Apply default values:
12727 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
12731 Enable additional sharpening:
12737 Paint processed pixels in white:
12745 Slowly update darker pixels.
12747 This filter makes short flashes of light appear longer.
12748 This filter accepts the following options:
12752 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
12755 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
12758 @section lenscorrection
12760 Correct radial lens distortion
12762 This filter can be used to correct for radial distortion as can result from the use
12763 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
12764 one can use tools available for example as part of opencv or simply trial-and-error.
12765 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
12766 and extract the k1 and k2 coefficients from the resulting matrix.
12768 Note that effectively the same filter is available in the open-source tools Krita and
12769 Digikam from the KDE project.
12771 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
12772 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
12773 brightness distribution, so you may want to use both filters together in certain
12774 cases, though you will have to take care of ordering, i.e. whether vignetting should
12775 be applied before or after lens correction.
12777 @subsection Options
12779 The filter accepts the following options:
12783 Relative x-coordinate of the focal point of the image, and thereby the center of the
12784 distortion. This value has a range [0,1] and is expressed as fractions of the image
12785 width. Default is 0.5.
12787 Relative y-coordinate of the focal point of the image, and thereby the center of the
12788 distortion. This value has a range [0,1] and is expressed as fractions of the image
12789 height. Default is 0.5.
12791 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
12792 no correction. Default is 0.
12794 Coefficient of the double quadratic correction term. This value has a range [-1,1].
12795 0 means no correction. Default is 0.
12798 The formula that generates the correction is:
12800 @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)
12802 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
12803 distances from the focal point in the source and target images, respectively.
12807 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
12809 The @code{lensfun} filter requires the camera make, camera model, and lens model
12810 to apply the lens correction. The filter will load the lensfun database and
12811 query it to find the corresponding camera and lens entries in the database. As
12812 long as these entries can be found with the given options, the filter can
12813 perform corrections on frames. Note that incomplete strings will result in the
12814 filter choosing the best match with the given options, and the filter will
12815 output the chosen camera and lens models (logged with level "info"). You must
12816 provide the make, camera model, and lens model as they are required.
12818 The filter accepts the following options:
12822 The make of the camera (for example, "Canon"). This option is required.
12825 The model of the camera (for example, "Canon EOS 100D"). This option is
12829 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
12830 option is required.
12833 The type of correction to apply. The following values are valid options:
12837 Enables fixing lens vignetting.
12840 Enables fixing lens geometry. This is the default.
12843 Enables fixing chromatic aberrations.
12846 Enables fixing lens vignetting and lens geometry.
12849 Enables fixing lens vignetting and chromatic aberrations.
12852 Enables fixing both lens geometry and chromatic aberrations.
12855 Enables all possible corrections.
12859 The focal length of the image/video (zoom; expected constant for video). For
12860 example, a 18--55mm lens has focal length range of [18--55], so a value in that
12861 range should be chosen when using that lens. Default 18.
12864 The aperture of the image/video (expected constant for video). Note that
12865 aperture is only used for vignetting correction. Default 3.5.
12867 @item focus_distance
12868 The focus distance of the image/video (expected constant for video). Note that
12869 focus distance is only used for vignetting and only slightly affects the
12870 vignetting correction process. If unknown, leave it at the default value (which
12874 The scale factor which is applied after transformation. After correction the
12875 video is no longer necessarily rectangular. This parameter controls how much of
12876 the resulting image is visible. The value 0 means that a value will be chosen
12877 automatically such that there is little or no unmapped area in the output
12878 image. 1.0 means that no additional scaling is done. Lower values may result
12879 in more of the corrected image being visible, while higher values may avoid
12880 unmapped areas in the output.
12882 @item target_geometry
12883 The target geometry of the output image/video. The following values are valid
12887 @item rectilinear (default)
12890 @item equirectangular
12891 @item fisheye_orthographic
12892 @item fisheye_stereographic
12893 @item fisheye_equisolid
12894 @item fisheye_thoby
12897 Apply the reverse of image correction (instead of correcting distortion, apply
12900 @item interpolation
12901 The type of interpolation used when correcting distortion. The following values
12906 @item linear (default)
12911 @subsection Examples
12915 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
12916 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
12920 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
12924 Apply the same as before, but only for the first 5 seconds of video.
12927 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
12934 Obtain the VMAF (Video Multi-Method Assessment Fusion)
12935 score between two input videos.
12937 The obtained VMAF score is printed through the logging system.
12939 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
12940 After installing the library it can be enabled using:
12941 @code{./configure --enable-libvmaf}.
12942 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
12944 The filter has following options:
12948 Set the model path which is to be used for SVM.
12949 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
12952 Set the file path to be used to store logs.
12955 Set the format of the log file (csv, json or xml).
12957 @item enable_transform
12958 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
12959 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
12960 Default value: @code{false}
12963 Invokes the phone model which will generate VMAF scores higher than in the
12964 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
12965 Default value: @code{false}
12968 Enables computing psnr along with vmaf.
12969 Default value: @code{false}
12972 Enables computing ssim along with vmaf.
12973 Default value: @code{false}
12976 Enables computing ms_ssim along with vmaf.
12977 Default value: @code{false}
12980 Set the pool method to be used for computing vmaf.
12981 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
12984 Set number of threads to be used when computing vmaf.
12985 Default value: @code{0}, which makes use of all available logical processors.
12988 Set interval for frame subsampling used when computing vmaf.
12989 Default value: @code{1}
12991 @item enable_conf_interval
12992 Enables confidence interval.
12993 Default value: @code{false}
12996 This filter also supports the @ref{framesync} options.
12998 @subsection Examples
13001 On the below examples the input file @file{main.mpg} being processed is
13002 compared with the reference file @file{ref.mpg}.
13005 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13009 Example with options:
13011 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13015 Example with options and different containers:
13017 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 -
13023 Limits the pixel components values to the specified range [min, max].
13025 The filter accepts the following options:
13029 Lower bound. Defaults to the lowest allowed value for the input.
13032 Upper bound. Defaults to the highest allowed value for the input.
13035 Specify which planes will be processed. Defaults to all available.
13042 The filter accepts the following options:
13046 Set the number of loops. Setting this value to -1 will result in infinite loops.
13050 Set maximal size in number of frames. Default is 0.
13053 Set first frame of loop. Default is 0.
13056 @subsection Examples
13060 Loop single first frame infinitely:
13062 loop=loop=-1:size=1:start=0
13066 Loop single first frame 10 times:
13068 loop=loop=10:size=1:start=0
13072 Loop 10 first frames 5 times:
13074 loop=loop=5:size=10:start=0
13080 Apply a 1D LUT to an input video.
13082 The filter accepts the following options:
13086 Set the 1D LUT file name.
13088 Currently supported formats:
13097 Select interpolation mode.
13099 Available values are:
13103 Use values from the nearest defined point.
13105 Interpolate values using the linear interpolation.
13107 Interpolate values using the cosine interpolation.
13109 Interpolate values using the cubic interpolation.
13111 Interpolate values using the spline interpolation.
13118 Apply a 3D LUT to an input video.
13120 The filter accepts the following options:
13124 Set the 3D LUT file name.
13126 Currently supported formats:
13140 Select interpolation mode.
13142 Available values are:
13146 Use values from the nearest defined point.
13148 Interpolate values using the 8 points defining a cube.
13150 Interpolate values using a tetrahedron.
13156 Turn certain luma values into transparency.
13158 The filter accepts the following options:
13162 Set the luma which will be used as base for transparency.
13163 Default value is @code{0}.
13166 Set the range of luma values to be keyed out.
13167 Default value is @code{0.01}.
13170 Set the range of softness. Default value is @code{0}.
13171 Use this to control gradual transition from zero to full transparency.
13174 @subsection Commands
13175 This filter supports same @ref{commands} as options.
13176 The command accepts the same syntax of the corresponding option.
13178 If the specified expression is not valid, it is kept at its current
13181 @section lut, lutrgb, lutyuv
13183 Compute a look-up table for binding each pixel component input value
13184 to an output value, and apply it to the input video.
13186 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13187 to an RGB input video.
13189 These filters accept the following parameters:
13192 set first pixel component expression
13194 set second pixel component expression
13196 set third pixel component expression
13198 set fourth pixel component expression, corresponds to the alpha component
13201 set red component expression
13203 set green component expression
13205 set blue component expression
13207 alpha component expression
13210 set Y/luminance component expression
13212 set U/Cb component expression
13214 set V/Cr component expression
13217 Each of them specifies the expression to use for computing the lookup table for
13218 the corresponding pixel component values.
13220 The exact component associated to each of the @var{c*} options depends on the
13223 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13224 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13226 The expressions can contain the following constants and functions:
13231 The input width and height.
13234 The input value for the pixel component.
13237 The input value, clipped to the @var{minval}-@var{maxval} range.
13240 The maximum value for the pixel component.
13243 The minimum value for the pixel component.
13246 The negated value for the pixel component value, clipped to the
13247 @var{minval}-@var{maxval} range; it corresponds to the expression
13248 "maxval-clipval+minval".
13251 The computed value in @var{val}, clipped to the
13252 @var{minval}-@var{maxval} range.
13254 @item gammaval(gamma)
13255 The computed gamma correction value of the pixel component value,
13256 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13258 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13262 All expressions default to "val".
13264 @subsection Examples
13268 Negate input video:
13270 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13271 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13274 The above is the same as:
13276 lutrgb="r=negval:g=negval:b=negval"
13277 lutyuv="y=negval:u=negval:v=negval"
13287 Remove chroma components, turning the video into a graytone image:
13289 lutyuv="u=128:v=128"
13293 Apply a luma burning effect:
13299 Remove green and blue components:
13305 Set a constant alpha channel value on input:
13307 format=rgba,lutrgb=a="maxval-minval/2"
13311 Correct luminance gamma by a factor of 0.5:
13313 lutyuv=y=gammaval(0.5)
13317 Discard least significant bits of luma:
13319 lutyuv=y='bitand(val, 128+64+32)'
13323 Technicolor like effect:
13325 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13329 @section lut2, tlut2
13331 The @code{lut2} filter takes two input streams and outputs one
13334 The @code{tlut2} (time lut2) filter takes two consecutive frames
13335 from one single stream.
13337 This filter accepts the following parameters:
13340 set first pixel component expression
13342 set second pixel component expression
13344 set third pixel component expression
13346 set fourth pixel component expression, corresponds to the alpha component
13349 set output bit depth, only available for @code{lut2} filter. By default is 0,
13350 which means bit depth is automatically picked from first input format.
13353 The @code{lut2} filter also supports the @ref{framesync} options.
13355 Each of them specifies the expression to use for computing the lookup table for
13356 the corresponding pixel component values.
13358 The exact component associated to each of the @var{c*} options depends on the
13361 The expressions can contain the following constants:
13366 The input width and height.
13369 The first input value for the pixel component.
13372 The second input value for the pixel component.
13375 The first input video bit depth.
13378 The second input video bit depth.
13381 All expressions default to "x".
13383 @subsection Examples
13387 Highlight differences between two RGB video streams:
13389 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)'
13393 Highlight differences between two YUV video streams:
13395 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)'
13399 Show max difference between two video streams:
13401 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)))'
13405 @section maskedclamp
13407 Clamp the first input stream with the second input and third input stream.
13409 Returns the value of first stream to be between second input
13410 stream - @code{undershoot} and third input stream + @code{overshoot}.
13412 This filter accepts the following options:
13415 Default value is @code{0}.
13418 Default value is @code{0}.
13421 Set which planes will be processed as bitmap, unprocessed planes will be
13422 copied from first stream.
13423 By default value 0xf, all planes will be processed.
13428 Merge the second and third input stream into output stream using absolute differences
13429 between second input stream and first input stream and absolute difference between
13430 third input stream and first input stream. The picked value will be from second input
13431 stream if second absolute difference is greater than first one or from third input stream
13434 This filter accepts the following options:
13437 Set which planes will be processed as bitmap, unprocessed planes will be
13438 copied from first stream.
13439 By default value 0xf, all planes will be processed.
13442 @section maskedmerge
13444 Merge the first input stream with the second input stream using per pixel
13445 weights in the third input stream.
13447 A value of 0 in the third stream pixel component means that pixel component
13448 from first stream is returned unchanged, while maximum value (eg. 255 for
13449 8-bit videos) means that pixel component from second stream is returned
13450 unchanged. Intermediate values define the amount of merging between both
13451 input stream's pixel components.
13453 This filter accepts the following options:
13456 Set which planes will be processed as bitmap, unprocessed planes will be
13457 copied from first stream.
13458 By default value 0xf, all planes will be processed.
13463 Merge the second and third input stream into output stream using absolute differences
13464 between second input stream and first input stream and absolute difference between
13465 third input stream and first input stream. The picked value will be from second input
13466 stream if second absolute difference is less than first one or from third input stream
13469 This filter accepts the following options:
13472 Set which planes will be processed as bitmap, unprocessed planes will be
13473 copied from first stream.
13474 By default value 0xf, all planes will be processed.
13477 @section maskedthreshold
13478 Pick pixels comparing absolute difference of two video streams with fixed
13481 If absolute difference between pixel component of first and second video
13482 stream is equal or lower than user supplied threshold than pixel component
13483 from first video stream is picked, otherwise pixel component from second
13484 video stream is picked.
13486 This filter accepts the following options:
13489 Set threshold used when picking pixels from absolute difference from two input
13493 Set which planes will be processed as bitmap, unprocessed planes will be
13494 copied from second stream.
13495 By default value 0xf, all planes will be processed.
13499 Create mask from input video.
13501 For example it is useful to create motion masks after @code{tblend} filter.
13503 This filter accepts the following options:
13507 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
13510 Set high threshold. Any pixel component higher than this value will be set to max value
13511 allowed for current pixel format.
13514 Set planes to filter, by default all available planes are filtered.
13517 Fill all frame pixels with this value.
13520 Set max average pixel value for frame. If sum of all pixel components is higher that this
13521 average, output frame will be completely filled with value set by @var{fill} option.
13522 Typically useful for scene changes when used in combination with @code{tblend} filter.
13527 Apply motion-compensation deinterlacing.
13529 It needs one field per frame as input and must thus be used together
13530 with yadif=1/3 or equivalent.
13532 This filter accepts the following options:
13535 Set the deinterlacing mode.
13537 It accepts one of the following values:
13542 use iterative motion estimation
13544 like @samp{slow}, but use multiple reference frames.
13546 Default value is @samp{fast}.
13549 Set the picture field parity assumed for the input video. It must be
13550 one of the following values:
13554 assume top field first
13556 assume bottom field first
13559 Default value is @samp{bff}.
13562 Set per-block quantization parameter (QP) used by the internal
13565 Higher values should result in a smoother motion vector field but less
13566 optimal individual vectors. Default value is 1.
13571 Pick median pixel from certain rectangle defined by radius.
13573 This filter accepts the following options:
13577 Set horizontal radius size. Default value is @code{1}.
13578 Allowed range is integer from 1 to 127.
13581 Set which planes to process. Default is @code{15}, which is all available planes.
13584 Set vertical radius size. Default value is @code{0}.
13585 Allowed range is integer from 0 to 127.
13586 If it is 0, value will be picked from horizontal @code{radius} option.
13589 Set median percentile. Default value is @code{0.5}.
13590 Default value of @code{0.5} will pick always median values, while @code{0} will pick
13591 minimum values, and @code{1} maximum values.
13594 @subsection Commands
13595 This filter supports same @ref{commands} as options.
13596 The command accepts the same syntax of the corresponding option.
13598 If the specified expression is not valid, it is kept at its current
13601 @section mergeplanes
13603 Merge color channel components from several video streams.
13605 The filter accepts up to 4 input streams, and merge selected input
13606 planes to the output video.
13608 This filter accepts the following options:
13611 Set input to output plane mapping. Default is @code{0}.
13613 The mappings is specified as a bitmap. It should be specified as a
13614 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
13615 mapping for the first plane of the output stream. 'A' sets the number of
13616 the input stream to use (from 0 to 3), and 'a' the plane number of the
13617 corresponding input to use (from 0 to 3). The rest of the mappings is
13618 similar, 'Bb' describes the mapping for the output stream second
13619 plane, 'Cc' describes the mapping for the output stream third plane and
13620 'Dd' describes the mapping for the output stream fourth plane.
13623 Set output pixel format. Default is @code{yuva444p}.
13626 @subsection Examples
13630 Merge three gray video streams of same width and height into single video stream:
13632 [a0][a1][a2]mergeplanes=0x001020:yuv444p
13636 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
13638 [a0][a1]mergeplanes=0x00010210:yuva444p
13642 Swap Y and A plane in yuva444p stream:
13644 format=yuva444p,mergeplanes=0x03010200:yuva444p
13648 Swap U and V plane in yuv420p stream:
13650 format=yuv420p,mergeplanes=0x000201:yuv420p
13654 Cast a rgb24 clip to yuv444p:
13656 format=rgb24,mergeplanes=0x000102:yuv444p
13662 Estimate and export motion vectors using block matching algorithms.
13663 Motion vectors are stored in frame side data to be used by other filters.
13665 This filter accepts the following options:
13668 Specify the motion estimation method. Accepts one of the following values:
13672 Exhaustive search algorithm.
13674 Three step search algorithm.
13676 Two dimensional logarithmic search algorithm.
13678 New three step search algorithm.
13680 Four step search algorithm.
13682 Diamond search algorithm.
13684 Hexagon-based search algorithm.
13686 Enhanced predictive zonal search algorithm.
13688 Uneven multi-hexagon search algorithm.
13690 Default value is @samp{esa}.
13693 Macroblock size. Default @code{16}.
13696 Search parameter. Default @code{7}.
13699 @section midequalizer
13701 Apply Midway Image Equalization effect using two video streams.
13703 Midway Image Equalization adjusts a pair of images to have the same
13704 histogram, while maintaining their dynamics as much as possible. It's
13705 useful for e.g. matching exposures from a pair of stereo cameras.
13707 This filter has two inputs and one output, which must be of same pixel format, but
13708 may be of different sizes. The output of filter is first input adjusted with
13709 midway histogram of both inputs.
13711 This filter accepts the following option:
13715 Set which planes to process. Default is @code{15}, which is all available planes.
13718 @section minterpolate
13720 Convert the video to specified frame rate using motion interpolation.
13722 This filter accepts the following options:
13725 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}.
13728 Motion interpolation mode. Following values are accepted:
13731 Duplicate previous or next frame for interpolating new ones.
13733 Blend source frames. Interpolated frame is mean of previous and next frames.
13735 Motion compensated interpolation. Following options are effective when this mode is selected:
13739 Motion compensation mode. Following values are accepted:
13742 Overlapped block motion compensation.
13744 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
13746 Default mode is @samp{obmc}.
13749 Motion estimation mode. Following values are accepted:
13752 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
13754 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
13756 Default mode is @samp{bilat}.
13759 The algorithm to be used for motion estimation. Following values are accepted:
13762 Exhaustive search algorithm.
13764 Three step search algorithm.
13766 Two dimensional logarithmic search algorithm.
13768 New three step search algorithm.
13770 Four step search algorithm.
13772 Diamond search algorithm.
13774 Hexagon-based search algorithm.
13776 Enhanced predictive zonal search algorithm.
13778 Uneven multi-hexagon search algorithm.
13780 Default algorithm is @samp{epzs}.
13783 Macroblock size. Default @code{16}.
13786 Motion estimation search parameter. Default @code{32}.
13789 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).
13794 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:
13797 Disable scene change detection.
13799 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
13801 Default method is @samp{fdiff}.
13803 @item scd_threshold
13804 Scene change detection threshold. Default is @code{10.}.
13809 Mix several video input streams into one video stream.
13811 A description of the accepted options follows.
13815 The number of inputs. If unspecified, it defaults to 2.
13818 Specify weight of each input video stream as sequence.
13819 Each weight is separated by space. If number of weights
13820 is smaller than number of @var{frames} last specified
13821 weight will be used for all remaining unset weights.
13824 Specify scale, if it is set it will be multiplied with sum
13825 of each weight multiplied with pixel values to give final destination
13826 pixel value. By default @var{scale} is auto scaled to sum of weights.
13829 Specify how end of stream is determined.
13832 The duration of the longest input. (default)
13835 The duration of the shortest input.
13838 The duration of the first input.
13842 @section mpdecimate
13844 Drop frames that do not differ greatly from the previous frame in
13845 order to reduce frame rate.
13847 The main use of this filter is for very-low-bitrate encoding
13848 (e.g. streaming over dialup modem), but it could in theory be used for
13849 fixing movies that were inverse-telecined incorrectly.
13851 A description of the accepted options follows.
13855 Set the maximum number of consecutive frames which can be dropped (if
13856 positive), or the minimum interval between dropped frames (if
13857 negative). If the value is 0, the frame is dropped disregarding the
13858 number of previous sequentially dropped frames.
13860 Default value is 0.
13865 Set the dropping threshold values.
13867 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
13868 represent actual pixel value differences, so a threshold of 64
13869 corresponds to 1 unit of difference for each pixel, or the same spread
13870 out differently over the block.
13872 A frame is a candidate for dropping if no 8x8 blocks differ by more
13873 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
13874 meaning the whole image) differ by more than a threshold of @option{lo}.
13876 Default value for @option{hi} is 64*12, default value for @option{lo} is
13877 64*5, and default value for @option{frac} is 0.33.
13883 Negate (invert) the input video.
13885 It accepts the following option:
13890 With value 1, it negates the alpha component, if present. Default value is 0.
13896 Denoise frames using Non-Local Means algorithm.
13898 Each pixel is adjusted by looking for other pixels with similar contexts. This
13899 context similarity is defined by comparing their surrounding patches of size
13900 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
13903 Note that the research area defines centers for patches, which means some
13904 patches will be made of pixels outside that research area.
13906 The filter accepts the following options.
13910 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
13913 Set patch size. Default is 7. Must be odd number in range [0, 99].
13916 Same as @option{p} but for chroma planes.
13918 The default value is @var{0} and means automatic.
13921 Set research size. Default is 15. Must be odd number in range [0, 99].
13924 Same as @option{r} but for chroma planes.
13926 The default value is @var{0} and means automatic.
13931 Deinterlace video using neural network edge directed interpolation.
13933 This filter accepts the following options:
13937 Mandatory option, without binary file filter can not work.
13938 Currently file can be found here:
13939 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
13942 Set which frames to deinterlace, by default it is @code{all}.
13943 Can be @code{all} or @code{interlaced}.
13946 Set mode of operation.
13948 Can be one of the following:
13952 Use frame flags, both fields.
13954 Use frame flags, single field.
13956 Use top field only.
13958 Use bottom field only.
13960 Use both fields, top first.
13962 Use both fields, bottom first.
13966 Set which planes to process, by default filter process all frames.
13969 Set size of local neighborhood around each pixel, used by the predictor neural
13972 Can be one of the following:
13985 Set the number of neurons in predictor neural network.
13986 Can be one of the following:
13997 Controls the number of different neural network predictions that are blended
13998 together to compute the final output value. Can be @code{fast}, default or
14002 Set which set of weights to use in the predictor.
14003 Can be one of the following:
14007 weights trained to minimize absolute error
14009 weights trained to minimize squared error
14013 Controls whether or not the prescreener neural network is used to decide
14014 which pixels should be processed by the predictor neural network and which
14015 can be handled by simple cubic interpolation.
14016 The prescreener is trained to know whether cubic interpolation will be
14017 sufficient for a pixel or whether it should be predicted by the predictor nn.
14018 The computational complexity of the prescreener nn is much less than that of
14019 the predictor nn. Since most pixels can be handled by cubic interpolation,
14020 using the prescreener generally results in much faster processing.
14021 The prescreener is pretty accurate, so the difference between using it and not
14022 using it is almost always unnoticeable.
14024 Can be one of the following:
14032 Default is @code{new}.
14035 Set various debugging flags.
14040 Force libavfilter not to use any of the specified pixel formats for the
14041 input to the next filter.
14043 It accepts the following parameters:
14047 A '|'-separated list of pixel format names, such as
14048 pix_fmts=yuv420p|monow|rgb24".
14052 @subsection Examples
14056 Force libavfilter to use a format different from @var{yuv420p} for the
14057 input to the vflip filter:
14059 noformat=pix_fmts=yuv420p,vflip
14063 Convert the input video to any of the formats not contained in the list:
14065 noformat=yuv420p|yuv444p|yuv410p
14071 Add noise on video input frame.
14073 The filter accepts the following options:
14081 Set noise seed for specific pixel component or all pixel components in case
14082 of @var{all_seed}. Default value is @code{123457}.
14084 @item all_strength, alls
14085 @item c0_strength, c0s
14086 @item c1_strength, c1s
14087 @item c2_strength, c2s
14088 @item c3_strength, c3s
14089 Set noise strength for specific pixel component or all pixel components in case
14090 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14092 @item all_flags, allf
14093 @item c0_flags, c0f
14094 @item c1_flags, c1f
14095 @item c2_flags, c2f
14096 @item c3_flags, c3f
14097 Set pixel component flags or set flags for all components if @var{all_flags}.
14098 Available values for component flags are:
14101 averaged temporal noise (smoother)
14103 mix random noise with a (semi)regular pattern
14105 temporal noise (noise pattern changes between frames)
14107 uniform noise (gaussian otherwise)
14111 @subsection Examples
14113 Add temporal and uniform noise to input video:
14115 noise=alls=20:allf=t+u
14120 Normalize RGB video (aka histogram stretching, contrast stretching).
14121 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14123 For each channel of each frame, the filter computes the input range and maps
14124 it linearly to the user-specified output range. The output range defaults
14125 to the full dynamic range from pure black to pure white.
14127 Temporal smoothing can be used on the input range to reduce flickering (rapid
14128 changes in brightness) caused when small dark or bright objects enter or leave
14129 the scene. This is similar to the auto-exposure (automatic gain control) on a
14130 video camera, and, like a video camera, it may cause a period of over- or
14131 under-exposure of the video.
14133 The R,G,B channels can be normalized independently, which may cause some
14134 color shifting, or linked together as a single channel, which prevents
14135 color shifting. Linked normalization preserves hue. Independent normalization
14136 does not, so it can be used to remove some color casts. Independent and linked
14137 normalization can be combined in any ratio.
14139 The normalize filter accepts the following options:
14144 Colors which define the output range. The minimum input value is mapped to
14145 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14146 The defaults are black and white respectively. Specifying white for
14147 @var{blackpt} and black for @var{whitept} will give color-inverted,
14148 normalized video. Shades of grey can be used to reduce the dynamic range
14149 (contrast). Specifying saturated colors here can create some interesting
14153 The number of previous frames to use for temporal smoothing. The input range
14154 of each channel is smoothed using a rolling average over the current frame
14155 and the @var{smoothing} previous frames. The default is 0 (no temporal
14159 Controls the ratio of independent (color shifting) channel normalization to
14160 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14161 independent. Defaults to 1.0 (fully independent).
14164 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14165 expensive no-op. Defaults to 1.0 (full strength).
14169 @subsection Commands
14170 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14171 The command accepts the same syntax of the corresponding option.
14173 If the specified expression is not valid, it is kept at its current
14176 @subsection Examples
14178 Stretch video contrast to use the full dynamic range, with no temporal
14179 smoothing; may flicker depending on the source content:
14181 normalize=blackpt=black:whitept=white:smoothing=0
14184 As above, but with 50 frames of temporal smoothing; flicker should be
14185 reduced, depending on the source content:
14187 normalize=blackpt=black:whitept=white:smoothing=50
14190 As above, but with hue-preserving linked channel normalization:
14192 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14195 As above, but with half strength:
14197 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14200 Map the darkest input color to red, the brightest input color to cyan:
14202 normalize=blackpt=red:whitept=cyan
14207 Pass the video source unchanged to the output.
14210 Optical Character Recognition
14212 This filter uses Tesseract for optical character recognition. To enable
14213 compilation of this filter, you need to configure FFmpeg with
14214 @code{--enable-libtesseract}.
14216 It accepts the following options:
14220 Set datapath to tesseract data. Default is to use whatever was
14221 set at installation.
14224 Set language, default is "eng".
14227 Set character whitelist.
14230 Set character blacklist.
14233 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14234 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14238 Apply a video transform using libopencv.
14240 To enable this filter, install the libopencv library and headers and
14241 configure FFmpeg with @code{--enable-libopencv}.
14243 It accepts the following parameters:
14248 The name of the libopencv filter to apply.
14250 @item filter_params
14251 The parameters to pass to the libopencv filter. If not specified, the default
14252 values are assumed.
14256 Refer to the official libopencv documentation for more precise
14258 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14260 Several libopencv filters are supported; see the following subsections.
14265 Dilate an image by using a specific structuring element.
14266 It corresponds to the libopencv function @code{cvDilate}.
14268 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14270 @var{struct_el} represents a structuring element, and has the syntax:
14271 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14273 @var{cols} and @var{rows} represent the number of columns and rows of
14274 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14275 point, and @var{shape} the shape for the structuring element. @var{shape}
14276 must be "rect", "cross", "ellipse", or "custom".
14278 If the value for @var{shape} is "custom", it must be followed by a
14279 string of the form "=@var{filename}". The file with name
14280 @var{filename} is assumed to represent a binary image, with each
14281 printable character corresponding to a bright pixel. When a custom
14282 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14283 or columns and rows of the read file are assumed instead.
14285 The default value for @var{struct_el} is "3x3+0x0/rect".
14287 @var{nb_iterations} specifies the number of times the transform is
14288 applied to the image, and defaults to 1.
14292 # Use the default values
14295 # Dilate using a structuring element with a 5x5 cross, iterating two times
14296 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14298 # Read the shape from the file diamond.shape, iterating two times.
14299 # The file diamond.shape may contain a pattern of characters like this
14305 # The specified columns and rows are ignored
14306 # but the anchor point coordinates are not
14307 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14312 Erode an image by using a specific structuring element.
14313 It corresponds to the libopencv function @code{cvErode}.
14315 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14316 with the same syntax and semantics as the @ref{dilate} filter.
14320 Smooth the input video.
14322 The filter takes the following parameters:
14323 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14325 @var{type} is the type of smooth filter to apply, and must be one of
14326 the following values: "blur", "blur_no_scale", "median", "gaussian",
14327 or "bilateral". The default value is "gaussian".
14329 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14330 depends on the smooth type. @var{param1} and
14331 @var{param2} accept integer positive values or 0. @var{param3} and
14332 @var{param4} accept floating point values.
14334 The default value for @var{param1} is 3. The default value for the
14335 other parameters is 0.
14337 These parameters correspond to the parameters assigned to the
14338 libopencv function @code{cvSmooth}.
14340 @section oscilloscope
14342 2D Video Oscilloscope.
14344 Useful to measure spatial impulse, step responses, chroma delays, etc.
14346 It accepts the following parameters:
14350 Set scope center x position.
14353 Set scope center y position.
14356 Set scope size, relative to frame diagonal.
14359 Set scope tilt/rotation.
14365 Set trace center x position.
14368 Set trace center y position.
14371 Set trace width, relative to width of frame.
14374 Set trace height, relative to height of frame.
14377 Set which components to trace. By default it traces first three components.
14380 Draw trace grid. By default is enabled.
14383 Draw some statistics. By default is enabled.
14386 Draw scope. By default is enabled.
14389 @subsection Commands
14390 This filter supports same @ref{commands} as options.
14391 The command accepts the same syntax of the corresponding option.
14393 If the specified expression is not valid, it is kept at its current
14396 @subsection Examples
14400 Inspect full first row of video frame.
14402 oscilloscope=x=0.5:y=0:s=1
14406 Inspect full last row of video frame.
14408 oscilloscope=x=0.5:y=1:s=1
14412 Inspect full 5th line of video frame of height 1080.
14414 oscilloscope=x=0.5:y=5/1080:s=1
14418 Inspect full last column of video frame.
14420 oscilloscope=x=1:y=0.5:s=1:t=1
14428 Overlay one video on top of another.
14430 It takes two inputs and has one output. The first input is the "main"
14431 video on which the second input is overlaid.
14433 It accepts the following parameters:
14435 A description of the accepted options follows.
14440 Set the expression for the x and y coordinates of the overlaid video
14441 on the main video. Default value is "0" for both expressions. In case
14442 the expression is invalid, it is set to a huge value (meaning that the
14443 overlay will not be displayed within the output visible area).
14446 See @ref{framesync}.
14449 Set when the expressions for @option{x}, and @option{y} are evaluated.
14451 It accepts the following values:
14454 only evaluate expressions once during the filter initialization or
14455 when a command is processed
14458 evaluate expressions for each incoming frame
14461 Default value is @samp{frame}.
14464 See @ref{framesync}.
14467 Set the format for the output video.
14469 It accepts the following values:
14472 force YUV420 output
14475 force YUV420p10 output
14478 force YUV422 output
14481 force YUV422p10 output
14484 force YUV444 output
14487 force packed RGB output
14490 force planar RGB output
14493 automatically pick format
14496 Default value is @samp{yuv420}.
14499 See @ref{framesync}.
14502 Set format of alpha of the overlaid video, it can be @var{straight} or
14503 @var{premultiplied}. Default is @var{straight}.
14506 The @option{x}, and @option{y} expressions can contain the following
14512 The main input width and height.
14516 The overlay input width and height.
14520 The computed values for @var{x} and @var{y}. They are evaluated for
14525 horizontal and vertical chroma subsample values of the output
14526 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
14530 the number of input frame, starting from 0
14533 the position in the file of the input frame, NAN if unknown
14536 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
14540 This filter also supports the @ref{framesync} options.
14542 Note that the @var{n}, @var{pos}, @var{t} variables are available only
14543 when evaluation is done @emph{per frame}, and will evaluate to NAN
14544 when @option{eval} is set to @samp{init}.
14546 Be aware that frames are taken from each input video in timestamp
14547 order, hence, if their initial timestamps differ, it is a good idea
14548 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
14549 have them begin in the same zero timestamp, as the example for
14550 the @var{movie} filter does.
14552 You can chain together more overlays but you should test the
14553 efficiency of such approach.
14555 @subsection Commands
14557 This filter supports the following commands:
14561 Modify the x and y of the overlay input.
14562 The command accepts the same syntax of the corresponding option.
14564 If the specified expression is not valid, it is kept at its current
14568 @subsection Examples
14572 Draw the overlay at 10 pixels from the bottom right corner of the main
14575 overlay=main_w-overlay_w-10:main_h-overlay_h-10
14578 Using named options the example above becomes:
14580 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
14584 Insert a transparent PNG logo in the bottom left corner of the input,
14585 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
14587 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
14591 Insert 2 different transparent PNG logos (second logo on bottom
14592 right corner) using the @command{ffmpeg} tool:
14594 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
14598 Add a transparent color layer on top of the main video; @code{WxH}
14599 must specify the size of the main input to the overlay filter:
14601 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
14605 Play an original video and a filtered version (here with the deshake
14606 filter) side by side using the @command{ffplay} tool:
14608 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
14611 The above command is the same as:
14613 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
14617 Make a sliding overlay appearing from the left to the right top part of the
14618 screen starting since time 2:
14620 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
14624 Compose output by putting two input videos side to side:
14626 ffmpeg -i left.avi -i right.avi -filter_complex "
14627 nullsrc=size=200x100 [background];
14628 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
14629 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
14630 [background][left] overlay=shortest=1 [background+left];
14631 [background+left][right] overlay=shortest=1:x=100 [left+right]
14636 Mask 10-20 seconds of a video by applying the delogo filter to a section
14638 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
14639 -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]'
14644 Chain several overlays in cascade:
14646 nullsrc=s=200x200 [bg];
14647 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
14648 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
14649 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
14650 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
14651 [in3] null, [mid2] overlay=100:100 [out0]
14656 @anchor{overlay_cuda}
14657 @section overlay_cuda
14659 Overlay one video on top of another.
14661 This is the CUDA cariant of the @ref{overlay} filter.
14662 It only accepts CUDA frames. The underlying input pixel formats have to match.
14664 It takes two inputs and has one output. The first input is the "main"
14665 video on which the second input is overlaid.
14667 It accepts the following parameters:
14672 Set the x and y coordinates of the overlaid video on the main video.
14673 Default value is "0" for both expressions.
14676 See @ref{framesync}.
14679 See @ref{framesync}.
14682 See @ref{framesync}.
14686 This filter also supports the @ref{framesync} options.
14690 Apply Overcomplete Wavelet denoiser.
14692 The filter accepts the following options:
14698 Larger depth values will denoise lower frequency components more, but
14699 slow down filtering.
14701 Must be an int in the range 8-16, default is @code{8}.
14703 @item luma_strength, ls
14706 Must be a double value in the range 0-1000, default is @code{1.0}.
14708 @item chroma_strength, cs
14709 Set chroma strength.
14711 Must be a double value in the range 0-1000, default is @code{1.0}.
14717 Add paddings to the input image, and place the original input at the
14718 provided @var{x}, @var{y} coordinates.
14720 It accepts the following parameters:
14725 Specify an expression for the size of the output image with the
14726 paddings added. If the value for @var{width} or @var{height} is 0, the
14727 corresponding input size is used for the output.
14729 The @var{width} expression can reference the value set by the
14730 @var{height} expression, and vice versa.
14732 The default value of @var{width} and @var{height} is 0.
14736 Specify the offsets to place the input image at within the padded area,
14737 with respect to the top/left border of the output image.
14739 The @var{x} expression can reference the value set by the @var{y}
14740 expression, and vice versa.
14742 The default value of @var{x} and @var{y} is 0.
14744 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
14745 so the input image is centered on the padded area.
14748 Specify the color of the padded area. For the syntax of this option,
14749 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
14750 manual,ffmpeg-utils}.
14752 The default value of @var{color} is "black".
14755 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
14757 It accepts the following values:
14761 Only evaluate expressions once during the filter initialization or when
14762 a command is processed.
14765 Evaluate expressions for each incoming frame.
14769 Default value is @samp{init}.
14772 Pad to aspect instead to a resolution.
14776 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
14777 options are expressions containing the following constants:
14782 The input video width and height.
14786 These are the same as @var{in_w} and @var{in_h}.
14790 The output width and height (the size of the padded area), as
14791 specified by the @var{width} and @var{height} expressions.
14795 These are the same as @var{out_w} and @var{out_h}.
14799 The x and y offsets as specified by the @var{x} and @var{y}
14800 expressions, or NAN if not yet specified.
14803 same as @var{iw} / @var{ih}
14806 input sample aspect ratio
14809 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
14813 The horizontal and vertical chroma subsample values. For example for the
14814 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14817 @subsection Examples
14821 Add paddings with the color "violet" to the input video. The output video
14822 size is 640x480, and the top-left corner of the input video is placed at
14825 pad=640:480:0:40:violet
14828 The example above is equivalent to the following command:
14830 pad=width=640:height=480:x=0:y=40:color=violet
14834 Pad the input to get an output with dimensions increased by 3/2,
14835 and put the input video at the center of the padded area:
14837 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
14841 Pad the input to get a squared output with size equal to the maximum
14842 value between the input width and height, and put the input video at
14843 the center of the padded area:
14845 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
14849 Pad the input to get a final w/h ratio of 16:9:
14851 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
14855 In case of anamorphic video, in order to set the output display aspect
14856 correctly, it is necessary to use @var{sar} in the expression,
14857 according to the relation:
14859 (ih * X / ih) * sar = output_dar
14860 X = output_dar / sar
14863 Thus the previous example needs to be modified to:
14865 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
14869 Double the output size and put the input video in the bottom-right
14870 corner of the output padded area:
14872 pad="2*iw:2*ih:ow-iw:oh-ih"
14876 @anchor{palettegen}
14877 @section palettegen
14879 Generate one palette for a whole video stream.
14881 It accepts the following options:
14885 Set the maximum number of colors to quantize in the palette.
14886 Note: the palette will still contain 256 colors; the unused palette entries
14889 @item reserve_transparent
14890 Create a palette of 255 colors maximum and reserve the last one for
14891 transparency. Reserving the transparency color is useful for GIF optimization.
14892 If not set, the maximum of colors in the palette will be 256. You probably want
14893 to disable this option for a standalone image.
14896 @item transparency_color
14897 Set the color that will be used as background for transparency.
14900 Set statistics mode.
14902 It accepts the following values:
14905 Compute full frame histograms.
14907 Compute histograms only for the part that differs from previous frame. This
14908 might be relevant to give more importance to the moving part of your input if
14909 the background is static.
14911 Compute new histogram for each frame.
14914 Default value is @var{full}.
14917 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
14918 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
14919 color quantization of the palette. This information is also visible at
14920 @var{info} logging level.
14922 @subsection Examples
14926 Generate a representative palette of a given video using @command{ffmpeg}:
14928 ffmpeg -i input.mkv -vf palettegen palette.png
14932 @section paletteuse
14934 Use a palette to downsample an input video stream.
14936 The filter takes two inputs: one video stream and a palette. The palette must
14937 be a 256 pixels image.
14939 It accepts the following options:
14943 Select dithering mode. Available algorithms are:
14946 Ordered 8x8 bayer dithering (deterministic)
14948 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
14949 Note: this dithering is sometimes considered "wrong" and is included as a
14951 @item floyd_steinberg
14952 Floyd and Steingberg dithering (error diffusion)
14954 Frankie Sierra dithering v2 (error diffusion)
14956 Frankie Sierra dithering v2 "Lite" (error diffusion)
14959 Default is @var{sierra2_4a}.
14962 When @var{bayer} dithering is selected, this option defines the scale of the
14963 pattern (how much the crosshatch pattern is visible). A low value means more
14964 visible pattern for less banding, and higher value means less visible pattern
14965 at the cost of more banding.
14967 The option must be an integer value in the range [0,5]. Default is @var{2}.
14970 If set, define the zone to process
14974 Only the changing rectangle will be reprocessed. This is similar to GIF
14975 cropping/offsetting compression mechanism. This option can be useful for speed
14976 if only a part of the image is changing, and has use cases such as limiting the
14977 scope of the error diffusal @option{dither} to the rectangle that bounds the
14978 moving scene (it leads to more deterministic output if the scene doesn't change
14979 much, and as a result less moving noise and better GIF compression).
14982 Default is @var{none}.
14985 Take new palette for each output frame.
14987 @item alpha_threshold
14988 Sets the alpha threshold for transparency. Alpha values above this threshold
14989 will be treated as completely opaque, and values below this threshold will be
14990 treated as completely transparent.
14992 The option must be an integer value in the range [0,255]. Default is @var{128}.
14995 @subsection Examples
14999 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15000 using @command{ffmpeg}:
15002 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15006 @section perspective
15008 Correct perspective of video not recorded perpendicular to the screen.
15010 A description of the accepted parameters follows.
15021 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15022 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15023 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15024 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15025 then the corners of the source will be sent to the specified coordinates.
15027 The expressions can use the following variables:
15032 the width and height of video frame.
15036 Output frame count.
15039 @item interpolation
15040 Set interpolation for perspective correction.
15042 It accepts the following values:
15048 Default value is @samp{linear}.
15051 Set interpretation of coordinate options.
15053 It accepts the following values:
15057 Send point in the source specified by the given coordinates to
15058 the corners of the destination.
15060 @item 1, destination
15062 Send the corners of the source to the point in the destination specified
15063 by the given coordinates.
15065 Default value is @samp{source}.
15069 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15071 It accepts the following values:
15074 only evaluate expressions once during the filter initialization or
15075 when a command is processed
15078 evaluate expressions for each incoming frame
15081 Default value is @samp{init}.
15086 Delay interlaced video by one field time so that the field order changes.
15088 The intended use is to fix PAL movies that have been captured with the
15089 opposite field order to the film-to-video transfer.
15091 A description of the accepted parameters follows.
15097 It accepts the following values:
15100 Capture field order top-first, transfer bottom-first.
15101 Filter will delay the bottom field.
15104 Capture field order bottom-first, transfer top-first.
15105 Filter will delay the top field.
15108 Capture and transfer with the same field order. This mode only exists
15109 for the documentation of the other options to refer to, but if you
15110 actually select it, the filter will faithfully do nothing.
15113 Capture field order determined automatically by field flags, transfer
15115 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15116 basis using field flags. If no field information is available,
15117 then this works just like @samp{u}.
15120 Capture unknown or varying, transfer opposite.
15121 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15122 analyzing the images and selecting the alternative that produces best
15123 match between the fields.
15126 Capture top-first, transfer unknown or varying.
15127 Filter selects among @samp{t} and @samp{p} using image analysis.
15130 Capture bottom-first, transfer unknown or varying.
15131 Filter selects among @samp{b} and @samp{p} using image analysis.
15134 Capture determined by field flags, transfer unknown or varying.
15135 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15136 image analysis. If no field information is available, then this works just
15137 like @samp{U}. This is the default mode.
15140 Both capture and transfer unknown or varying.
15141 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15145 @section photosensitivity
15146 Reduce various flashes in video, so to help users with epilepsy.
15148 It accepts the following options:
15151 Set how many frames to use when filtering. Default is 30.
15154 Set detection threshold factor. Default is 1.
15158 Set how many pixels to skip when sampling frames. Default is 1.
15159 Allowed range is from 1 to 1024.
15162 Leave frames unchanged. Default is disabled.
15165 @section pixdesctest
15167 Pixel format descriptor test filter, mainly useful for internal
15168 testing. The output video should be equal to the input video.
15172 format=monow, pixdesctest
15175 can be used to test the monowhite pixel format descriptor definition.
15179 Display sample values of color channels. Mainly useful for checking color
15180 and levels. Minimum supported resolution is 640x480.
15182 The filters accept the following options:
15186 Set scope X position, relative offset on X axis.
15189 Set scope Y position, relative offset on Y axis.
15198 Set window opacity. This window also holds statistics about pixel area.
15201 Set window X position, relative offset on X axis.
15204 Set window Y position, relative offset on Y axis.
15209 Enable the specified chain of postprocessing subfilters using libpostproc. This
15210 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15211 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15212 Each subfilter and some options have a short and a long name that can be used
15213 interchangeably, i.e. dr/dering are the same.
15215 The filters accept the following options:
15219 Set postprocessing subfilters string.
15222 All subfilters share common options to determine their scope:
15226 Honor the quality commands for this subfilter.
15229 Do chrominance filtering, too (default).
15232 Do luminance filtering only (no chrominance).
15235 Do chrominance filtering only (no luminance).
15238 These options can be appended after the subfilter name, separated by a '|'.
15240 Available subfilters are:
15243 @item hb/hdeblock[|difference[|flatness]]
15244 Horizontal deblocking filter
15247 Difference factor where higher values mean more deblocking (default: @code{32}).
15249 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15252 @item vb/vdeblock[|difference[|flatness]]
15253 Vertical deblocking filter
15256 Difference factor where higher values mean more deblocking (default: @code{32}).
15258 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15261 @item ha/hadeblock[|difference[|flatness]]
15262 Accurate horizontal deblocking filter
15265 Difference factor where higher values mean more deblocking (default: @code{32}).
15267 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15270 @item va/vadeblock[|difference[|flatness]]
15271 Accurate vertical deblocking filter
15274 Difference factor where higher values mean more deblocking (default: @code{32}).
15276 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15280 The horizontal and vertical deblocking filters share the difference and
15281 flatness values so you cannot set different horizontal and vertical
15285 @item h1/x1hdeblock
15286 Experimental horizontal deblocking filter
15288 @item v1/x1vdeblock
15289 Experimental vertical deblocking filter
15294 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15297 larger -> stronger filtering
15299 larger -> stronger filtering
15301 larger -> stronger filtering
15304 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15307 Stretch luminance to @code{0-255}.
15310 @item lb/linblenddeint
15311 Linear blend deinterlacing filter that deinterlaces the given block by
15312 filtering all lines with a @code{(1 2 1)} filter.
15314 @item li/linipoldeint
15315 Linear interpolating deinterlacing filter that deinterlaces the given block by
15316 linearly interpolating every second line.
15318 @item ci/cubicipoldeint
15319 Cubic interpolating deinterlacing filter deinterlaces the given block by
15320 cubically interpolating every second line.
15322 @item md/mediandeint
15323 Median deinterlacing filter that deinterlaces the given block by applying a
15324 median filter to every second line.
15326 @item fd/ffmpegdeint
15327 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15328 second line with a @code{(-1 4 2 4 -1)} filter.
15331 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15332 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15334 @item fq/forceQuant[|quantizer]
15335 Overrides the quantizer table from the input with the constant quantizer you
15343 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15346 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15349 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15352 @subsection Examples
15356 Apply horizontal and vertical deblocking, deringing and automatic
15357 brightness/contrast:
15363 Apply default filters without brightness/contrast correction:
15369 Apply default filters and temporal denoiser:
15371 pp=default/tmpnoise|1|2|3
15375 Apply deblocking on luminance only, and switch vertical deblocking on or off
15376 automatically depending on available CPU time:
15383 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15384 similar to spp = 6 with 7 point DCT, where only the center sample is
15387 The filter accepts the following options:
15391 Force a constant quantization parameter. It accepts an integer in range
15392 0 to 63. If not set, the filter will use the QP from the video stream
15396 Set thresholding mode. Available modes are:
15400 Set hard thresholding.
15402 Set soft thresholding (better de-ringing effect, but likely blurrier).
15404 Set medium thresholding (good results, default).
15408 @section premultiply
15409 Apply alpha premultiply effect to input video stream using first plane
15410 of second stream as alpha.
15412 Both streams must have same dimensions and same pixel format.
15414 The filter accepts the following option:
15418 Set which planes will be processed, unprocessed planes will be copied.
15419 By default value 0xf, all planes will be processed.
15422 Do not require 2nd input for processing, instead use alpha plane from input stream.
15426 Apply prewitt operator to input video stream.
15428 The filter accepts the following option:
15432 Set which planes will be processed, unprocessed planes will be copied.
15433 By default value 0xf, all planes will be processed.
15436 Set value which will be multiplied with filtered result.
15439 Set value which will be added to filtered result.
15442 @section pseudocolor
15444 Alter frame colors in video with pseudocolors.
15446 This filter accepts the following options:
15450 set pixel first component expression
15453 set pixel second component expression
15456 set pixel third component expression
15459 set pixel fourth component expression, corresponds to the alpha component
15462 set component to use as base for altering colors
15465 Each of them specifies the expression to use for computing the lookup table for
15466 the corresponding pixel component values.
15468 The expressions can contain the following constants and functions:
15473 The input width and height.
15476 The input value for the pixel component.
15478 @item ymin, umin, vmin, amin
15479 The minimum allowed component value.
15481 @item ymax, umax, vmax, amax
15482 The maximum allowed component value.
15485 All expressions default to "val".
15487 @subsection Examples
15491 Change too high luma values to gradient:
15493 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'"
15499 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
15500 Ratio) between two input videos.
15502 This filter takes in input two input videos, the first input is
15503 considered the "main" source and is passed unchanged to the
15504 output. The second input is used as a "reference" video for computing
15507 Both video inputs must have the same resolution and pixel format for
15508 this filter to work correctly. Also it assumes that both inputs
15509 have the same number of frames, which are compared one by one.
15511 The obtained average PSNR is printed through the logging system.
15513 The filter stores the accumulated MSE (mean squared error) of each
15514 frame, and at the end of the processing it is averaged across all frames
15515 equally, and the following formula is applied to obtain the PSNR:
15518 PSNR = 10*log10(MAX^2/MSE)
15521 Where MAX is the average of the maximum values of each component of the
15524 The description of the accepted parameters follows.
15527 @item stats_file, f
15528 If specified the filter will use the named file to save the PSNR of
15529 each individual frame. When filename equals "-" the data is sent to
15532 @item stats_version
15533 Specifies which version of the stats file format to use. Details of
15534 each format are written below.
15535 Default value is 1.
15537 @item stats_add_max
15538 Determines whether the max value is output to the stats log.
15539 Default value is 0.
15540 Requires stats_version >= 2. If this is set and stats_version < 2,
15541 the filter will return an error.
15544 This filter also supports the @ref{framesync} options.
15546 The file printed if @var{stats_file} is selected, contains a sequence of
15547 key/value pairs of the form @var{key}:@var{value} for each compared
15550 If a @var{stats_version} greater than 1 is specified, a header line precedes
15551 the list of per-frame-pair stats, with key value pairs following the frame
15552 format with the following parameters:
15555 @item psnr_log_version
15556 The version of the log file format. Will match @var{stats_version}.
15559 A comma separated list of the per-frame-pair parameters included in
15563 A description of each shown per-frame-pair parameter follows:
15567 sequential number of the input frame, starting from 1
15570 Mean Square Error pixel-by-pixel average difference of the compared
15571 frames, averaged over all the image components.
15573 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
15574 Mean Square Error pixel-by-pixel average difference of the compared
15575 frames for the component specified by the suffix.
15577 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
15578 Peak Signal to Noise ratio of the compared frames for the component
15579 specified by the suffix.
15581 @item max_avg, max_y, max_u, max_v
15582 Maximum allowed value for each channel, and average over all
15586 @subsection Examples
15591 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
15592 [main][ref] psnr="stats_file=stats.log" [out]
15595 On this example the input file being processed is compared with the
15596 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
15597 is stored in @file{stats.log}.
15600 Another example with different containers:
15602 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 -
15609 Pulldown reversal (inverse telecine) filter, capable of handling mixed
15610 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
15613 The pullup filter is designed to take advantage of future context in making
15614 its decisions. This filter is stateless in the sense that it does not lock
15615 onto a pattern to follow, but it instead looks forward to the following
15616 fields in order to identify matches and rebuild progressive frames.
15618 To produce content with an even framerate, insert the fps filter after
15619 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
15620 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
15622 The filter accepts the following options:
15629 These options set the amount of "junk" to ignore at the left, right, top, and
15630 bottom of the image, respectively. Left and right are in units of 8 pixels,
15631 while top and bottom are in units of 2 lines.
15632 The default is 8 pixels on each side.
15635 Set the strict breaks. Setting this option to 1 will reduce the chances of
15636 filter generating an occasional mismatched frame, but it may also cause an
15637 excessive number of frames to be dropped during high motion sequences.
15638 Conversely, setting it to -1 will make filter match fields more easily.
15639 This may help processing of video where there is slight blurring between
15640 the fields, but may also cause there to be interlaced frames in the output.
15641 Default value is @code{0}.
15644 Set the metric plane to use. It accepts the following values:
15650 Use chroma blue plane.
15653 Use chroma red plane.
15656 This option may be set to use chroma plane instead of the default luma plane
15657 for doing filter's computations. This may improve accuracy on very clean
15658 source material, but more likely will decrease accuracy, especially if there
15659 is chroma noise (rainbow effect) or any grayscale video.
15660 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
15661 load and make pullup usable in realtime on slow machines.
15664 For best results (without duplicated frames in the output file) it is
15665 necessary to change the output frame rate. For example, to inverse
15666 telecine NTSC input:
15668 ffmpeg -i input -vf pullup -r 24000/1001 ...
15673 Change video quantization parameters (QP).
15675 The filter accepts the following option:
15679 Set expression for quantization parameter.
15682 The expression is evaluated through the eval API and can contain, among others,
15683 the following constants:
15687 1 if index is not 129, 0 otherwise.
15690 Sequential index starting from -129 to 128.
15693 @subsection Examples
15697 Some equation like:
15705 Flush video frames from internal cache of frames into a random order.
15706 No frame is discarded.
15707 Inspired by @ref{frei0r} nervous filter.
15711 Set size in number of frames of internal cache, in range from @code{2} to
15712 @code{512}. Default is @code{30}.
15715 Set seed for random number generator, must be an integer included between
15716 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
15717 less than @code{0}, the filter will try to use a good random seed on a
15721 @section readeia608
15723 Read closed captioning (EIA-608) information from the top lines of a video frame.
15725 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
15726 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
15727 with EIA-608 data (starting from 0). A description of each metadata value follows:
15730 @item lavfi.readeia608.X.cc
15731 The two bytes stored as EIA-608 data (printed in hexadecimal).
15733 @item lavfi.readeia608.X.line
15734 The number of the line on which the EIA-608 data was identified and read.
15737 This filter accepts the following options:
15741 Set the line to start scanning for EIA-608 data. Default is @code{0}.
15744 Set the line to end scanning for EIA-608 data. Default is @code{29}.
15747 Set the ratio of width reserved for sync code detection.
15748 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
15751 Enable checking the parity bit. In the event of a parity error, the filter will output
15752 @code{0x00} for that character. Default is false.
15755 Lowpass lines prior to further processing. Default is enabled.
15758 @subsection Examples
15762 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
15764 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
15770 Read vertical interval timecode (VITC) information from the top lines of a
15773 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
15774 timecode value, if a valid timecode has been detected. Further metadata key
15775 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
15776 timecode data has been found or not.
15778 This filter accepts the following options:
15782 Set the maximum number of lines to scan for VITC data. If the value is set to
15783 @code{-1} the full video frame is scanned. Default is @code{45}.
15786 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
15787 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
15790 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
15791 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
15794 @subsection Examples
15798 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
15799 draw @code{--:--:--:--} as a placeholder:
15801 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
15807 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
15809 Destination pixel at position (X, Y) will be picked from source (x, y) position
15810 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
15811 value for pixel will be used for destination pixel.
15813 Xmap and Ymap input video streams must be of same dimensions. Output video stream
15814 will have Xmap/Ymap video stream dimensions.
15815 Xmap and Ymap input video streams are 16bit depth, single channel.
15819 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
15820 Default is @code{color}.
15823 Specify the color of the unmapped pixels. For the syntax of this option,
15824 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15825 manual,ffmpeg-utils}. Default color is @code{black}.
15828 @section removegrain
15830 The removegrain filter is a spatial denoiser for progressive video.
15834 Set mode for the first plane.
15837 Set mode for the second plane.
15840 Set mode for the third plane.
15843 Set mode for the fourth plane.
15846 Range of mode is from 0 to 24. Description of each mode follows:
15850 Leave input plane unchanged. Default.
15853 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
15856 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
15859 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
15862 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
15863 This is equivalent to a median filter.
15866 Line-sensitive clipping giving the minimal change.
15869 Line-sensitive clipping, intermediate.
15872 Line-sensitive clipping, intermediate.
15875 Line-sensitive clipping, intermediate.
15878 Line-sensitive clipping on a line where the neighbours pixels are the closest.
15881 Replaces the target pixel with the closest neighbour.
15884 [1 2 1] horizontal and vertical kernel blur.
15890 Bob mode, interpolates top field from the line where the neighbours
15891 pixels are the closest.
15894 Bob mode, interpolates bottom field from the line where the neighbours
15895 pixels are the closest.
15898 Bob mode, interpolates top field. Same as 13 but with a more complicated
15899 interpolation formula.
15902 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
15903 interpolation formula.
15906 Clips the pixel with the minimum and maximum of respectively the maximum and
15907 minimum of each pair of opposite neighbour pixels.
15910 Line-sensitive clipping using opposite neighbours whose greatest distance from
15911 the current pixel is minimal.
15914 Replaces the pixel with the average of its 8 neighbours.
15917 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
15920 Clips pixels using the averages of opposite neighbour.
15923 Same as mode 21 but simpler and faster.
15926 Small edge and halo removal, but reputed useless.
15932 @section removelogo
15934 Suppress a TV station logo, using an image file to determine which
15935 pixels comprise the logo. It works by filling in the pixels that
15936 comprise the logo with neighboring pixels.
15938 The filter accepts the following options:
15942 Set the filter bitmap file, which can be any image format supported by
15943 libavformat. The width and height of the image file must match those of the
15944 video stream being processed.
15947 Pixels in the provided bitmap image with a value of zero are not
15948 considered part of the logo, non-zero pixels are considered part of
15949 the logo. If you use white (255) for the logo and black (0) for the
15950 rest, you will be safe. For making the filter bitmap, it is
15951 recommended to take a screen capture of a black frame with the logo
15952 visible, and then using a threshold filter followed by the erode
15953 filter once or twice.
15955 If needed, little splotches can be fixed manually. Remember that if
15956 logo pixels are not covered, the filter quality will be much
15957 reduced. Marking too many pixels as part of the logo does not hurt as
15958 much, but it will increase the amount of blurring needed to cover over
15959 the image and will destroy more information than necessary, and extra
15960 pixels will slow things down on a large logo.
15962 @section repeatfields
15964 This filter uses the repeat_field flag from the Video ES headers and hard repeats
15965 fields based on its value.
15969 Reverse a video clip.
15971 Warning: This filter requires memory to buffer the entire clip, so trimming
15974 @subsection Examples
15978 Take the first 5 seconds of a clip, and reverse it.
15985 Shift R/G/B/A pixels horizontally and/or vertically.
15987 The filter accepts the following options:
15990 Set amount to shift red horizontally.
15992 Set amount to shift red vertically.
15994 Set amount to shift green horizontally.
15996 Set amount to shift green vertically.
15998 Set amount to shift blue horizontally.
16000 Set amount to shift blue vertically.
16002 Set amount to shift alpha horizontally.
16004 Set amount to shift alpha vertically.
16006 Set edge mode, can be @var{smear}, default, or @var{warp}.
16009 @subsection Commands
16011 This filter supports the all above options as @ref{commands}.
16014 Apply roberts cross operator to input video stream.
16016 The filter accepts the following option:
16020 Set which planes will be processed, unprocessed planes will be copied.
16021 By default value 0xf, all planes will be processed.
16024 Set value which will be multiplied with filtered result.
16027 Set value which will be added to filtered result.
16032 Rotate video by an arbitrary angle expressed in radians.
16034 The filter accepts the following options:
16036 A description of the optional parameters follows.
16039 Set an expression for the angle by which to rotate the input video
16040 clockwise, expressed as a number of radians. A negative value will
16041 result in a counter-clockwise rotation. By default it is set to "0".
16043 This expression is evaluated for each frame.
16046 Set the output width expression, default value is "iw".
16047 This expression is evaluated just once during configuration.
16050 Set the output height expression, default value is "ih".
16051 This expression is evaluated just once during configuration.
16054 Enable bilinear interpolation if set to 1, a value of 0 disables
16055 it. Default value is 1.
16058 Set the color used to fill the output area not covered by the rotated
16059 image. For the general syntax of this option, check the
16060 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16061 If the special value "none" is selected then no
16062 background is printed (useful for example if the background is never shown).
16064 Default value is "black".
16067 The expressions for the angle and the output size can contain the
16068 following constants and functions:
16072 sequential number of the input frame, starting from 0. It is always NAN
16073 before the first frame is filtered.
16076 time in seconds of the input frame, it is set to 0 when the filter is
16077 configured. It is always NAN before the first frame is filtered.
16081 horizontal and vertical chroma subsample values. For example for the
16082 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16086 the input video width and height
16090 the output width and height, that is the size of the padded area as
16091 specified by the @var{width} and @var{height} expressions
16095 the minimal width/height required for completely containing the input
16096 video rotated by @var{a} radians.
16098 These are only available when computing the @option{out_w} and
16099 @option{out_h} expressions.
16102 @subsection Examples
16106 Rotate the input by PI/6 radians clockwise:
16112 Rotate the input by PI/6 radians counter-clockwise:
16118 Rotate the input by 45 degrees clockwise:
16124 Apply a constant rotation with period T, starting from an angle of PI/3:
16126 rotate=PI/3+2*PI*t/T
16130 Make the input video rotation oscillating with a period of T
16131 seconds and an amplitude of A radians:
16133 rotate=A*sin(2*PI/T*t)
16137 Rotate the video, output size is chosen so that the whole rotating
16138 input video is always completely contained in the output:
16140 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16144 Rotate the video, reduce the output size so that no background is ever
16147 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16151 @subsection Commands
16153 The filter supports the following commands:
16157 Set the angle expression.
16158 The command accepts the same syntax of the corresponding option.
16160 If the specified expression is not valid, it is kept at its current
16166 Apply Shape Adaptive Blur.
16168 The filter accepts the following options:
16171 @item luma_radius, lr
16172 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16173 value is 1.0. A greater value will result in a more blurred image, and
16174 in slower processing.
16176 @item luma_pre_filter_radius, lpfr
16177 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16180 @item luma_strength, ls
16181 Set luma maximum difference between pixels to still be considered, must
16182 be a value in the 0.1-100.0 range, default value is 1.0.
16184 @item chroma_radius, cr
16185 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16186 greater value will result in a more blurred image, and in slower
16189 @item chroma_pre_filter_radius, cpfr
16190 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16192 @item chroma_strength, cs
16193 Set chroma maximum difference between pixels to still be considered,
16194 must be a value in the -0.9-100.0 range.
16197 Each chroma option value, if not explicitly specified, is set to the
16198 corresponding luma option value.
16203 Scale (resize) the input video, using the libswscale library.
16205 The scale filter forces the output display aspect ratio to be the same
16206 of the input, by changing the output sample aspect ratio.
16208 If the input image format is different from the format requested by
16209 the next filter, the scale filter will convert the input to the
16212 @subsection Options
16213 The filter accepts the following options, or any of the options
16214 supported by the libswscale scaler.
16216 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16217 the complete list of scaler options.
16222 Set the output video dimension expression. Default value is the input
16225 If the @var{width} or @var{w} value is 0, the input width is used for
16226 the output. If the @var{height} or @var{h} value is 0, the input height
16227 is used for the output.
16229 If one and only one of the values is -n with n >= 1, the scale filter
16230 will use a value that maintains the aspect ratio of the input image,
16231 calculated from the other specified dimension. After that it will,
16232 however, make sure that the calculated dimension is divisible by n and
16233 adjust the value if necessary.
16235 If both values are -n with n >= 1, the behavior will be identical to
16236 both values being set to 0 as previously detailed.
16238 See below for the list of accepted constants for use in the dimension
16242 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16246 Only evaluate expressions once during the filter initialization or when a command is processed.
16249 Evaluate expressions for each incoming frame.
16253 Default value is @samp{init}.
16257 Set the interlacing mode. It accepts the following values:
16261 Force interlaced aware scaling.
16264 Do not apply interlaced scaling.
16267 Select interlaced aware scaling depending on whether the source frames
16268 are flagged as interlaced or not.
16271 Default value is @samp{0}.
16274 Set libswscale scaling flags. See
16275 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16276 complete list of values. If not explicitly specified the filter applies
16280 @item param0, param1
16281 Set libswscale input parameters for scaling algorithms that need them. See
16282 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16283 complete documentation. If not explicitly specified the filter applies
16289 Set the video size. For the syntax of this option, check the
16290 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16292 @item in_color_matrix
16293 @item out_color_matrix
16294 Set in/output YCbCr color space type.
16296 This allows the autodetected value to be overridden as well as allows forcing
16297 a specific value used for the output and encoder.
16299 If not specified, the color space type depends on the pixel format.
16305 Choose automatically.
16308 Format conforming to International Telecommunication Union (ITU)
16309 Recommendation BT.709.
16312 Set color space conforming to the United States Federal Communications
16313 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16318 Set color space conforming to:
16322 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16325 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16328 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16333 Set color space conforming to SMPTE ST 240:1999.
16336 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16341 Set in/output YCbCr sample range.
16343 This allows the autodetected value to be overridden as well as allows forcing
16344 a specific value used for the output and encoder. If not specified, the
16345 range depends on the pixel format. Possible values:
16349 Choose automatically.
16352 Set full range (0-255 in case of 8-bit luma).
16354 @item mpeg/limited/tv
16355 Set "MPEG" range (16-235 in case of 8-bit luma).
16358 @item force_original_aspect_ratio
16359 Enable decreasing or increasing output video width or height if necessary to
16360 keep the original aspect ratio. Possible values:
16364 Scale the video as specified and disable this feature.
16367 The output video dimensions will automatically be decreased if needed.
16370 The output video dimensions will automatically be increased if needed.
16374 One useful instance of this option is that when you know a specific device's
16375 maximum allowed resolution, you can use this to limit the output video to
16376 that, while retaining the aspect ratio. For example, device A allows
16377 1280x720 playback, and your video is 1920x800. Using this option (set it to
16378 decrease) and specifying 1280x720 to the command line makes the output
16381 Please note that this is a different thing than specifying -1 for @option{w}
16382 or @option{h}, you still need to specify the output resolution for this option
16385 @item force_divisible_by
16386 Ensures that both the output dimensions, width and height, are divisible by the
16387 given integer when used together with @option{force_original_aspect_ratio}. This
16388 works similar to using @code{-n} in the @option{w} and @option{h} options.
16390 This option respects the value set for @option{force_original_aspect_ratio},
16391 increasing or decreasing the resolution accordingly. The video's aspect ratio
16392 may be slightly modified.
16394 This option can be handy if you need to have a video fit within or exceed
16395 a defined resolution using @option{force_original_aspect_ratio} but also have
16396 encoder restrictions on width or height divisibility.
16400 The values of the @option{w} and @option{h} options are expressions
16401 containing the following constants:
16406 The input width and height
16410 These are the same as @var{in_w} and @var{in_h}.
16414 The output (scaled) width and height
16418 These are the same as @var{out_w} and @var{out_h}
16421 The same as @var{iw} / @var{ih}
16424 input sample aspect ratio
16427 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16431 horizontal and vertical input chroma subsample values. For example for the
16432 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16436 horizontal and vertical output chroma subsample values. For example for the
16437 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16440 The (sequential) number of the input frame, starting from 0.
16441 Only available with @code{eval=frame}.
16444 The presentation timestamp of the input frame, expressed as a number of
16445 seconds. Only available with @code{eval=frame}.
16448 The position (byte offset) of the frame in the input stream, or NaN if
16449 this information is unavailable and/or meaningless (for example in case of synthetic video).
16450 Only available with @code{eval=frame}.
16453 @subsection Examples
16457 Scale the input video to a size of 200x100
16462 This is equivalent to:
16473 Specify a size abbreviation for the output size:
16478 which can also be written as:
16484 Scale the input to 2x:
16486 scale=w=2*iw:h=2*ih
16490 The above is the same as:
16492 scale=2*in_w:2*in_h
16496 Scale the input to 2x with forced interlaced scaling:
16498 scale=2*iw:2*ih:interl=1
16502 Scale the input to half size:
16504 scale=w=iw/2:h=ih/2
16508 Increase the width, and set the height to the same size:
16514 Seek Greek harmony:
16521 Increase the height, and set the width to 3/2 of the height:
16523 scale=w=3/2*oh:h=3/5*ih
16527 Increase the size, making the size a multiple of the chroma
16530 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
16534 Increase the width to a maximum of 500 pixels,
16535 keeping the same aspect ratio as the input:
16537 scale=w='min(500\, iw*3/2):h=-1'
16541 Make pixels square by combining scale and setsar:
16543 scale='trunc(ih*dar):ih',setsar=1/1
16547 Make pixels square by combining scale and setsar,
16548 making sure the resulting resolution is even (required by some codecs):
16550 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
16554 @subsection Commands
16556 This filter supports the following commands:
16560 Set the output video dimension expression.
16561 The command accepts the same syntax of the corresponding option.
16563 If the specified expression is not valid, it is kept at its current
16569 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
16570 format conversion on CUDA video frames. Setting the output width and height
16571 works in the same way as for the @var{scale} filter.
16573 The following additional options are accepted:
16576 The pixel format of the output CUDA frames. If set to the string "same" (the
16577 default), the input format will be kept. Note that automatic format negotiation
16578 and conversion is not yet supported for hardware frames
16581 The interpolation algorithm used for resizing. One of the following:
16588 @item cubic2p_bspline
16589 2-parameter cubic (B=1, C=0)
16591 @item cubic2p_catmullrom
16592 2-parameter cubic (B=0, C=1/2)
16594 @item cubic2p_b05c03
16595 2-parameter cubic (B=1/2, C=3/10)
16603 @item force_original_aspect_ratio
16604 Enable decreasing or increasing output video width or height if necessary to
16605 keep the original aspect ratio. Possible values:
16609 Scale the video as specified and disable this feature.
16612 The output video dimensions will automatically be decreased if needed.
16615 The output video dimensions will automatically be increased if needed.
16619 One useful instance of this option is that when you know a specific device's
16620 maximum allowed resolution, you can use this to limit the output video to
16621 that, while retaining the aspect ratio. For example, device A allows
16622 1280x720 playback, and your video is 1920x800. Using this option (set it to
16623 decrease) and specifying 1280x720 to the command line makes the output
16626 Please note that this is a different thing than specifying -1 for @option{w}
16627 or @option{h}, you still need to specify the output resolution for this option
16630 @item force_divisible_by
16631 Ensures that both the output dimensions, width and height, are divisible by the
16632 given integer when used together with @option{force_original_aspect_ratio}. This
16633 works similar to using @code{-n} in the @option{w} and @option{h} options.
16635 This option respects the value set for @option{force_original_aspect_ratio},
16636 increasing or decreasing the resolution accordingly. The video's aspect ratio
16637 may be slightly modified.
16639 This option can be handy if you need to have a video fit within or exceed
16640 a defined resolution using @option{force_original_aspect_ratio} but also have
16641 encoder restrictions on width or height divisibility.
16647 Scale (resize) the input video, based on a reference video.
16649 See the scale filter for available options, scale2ref supports the same but
16650 uses the reference video instead of the main input as basis. scale2ref also
16651 supports the following additional constants for the @option{w} and
16652 @option{h} options:
16657 The main input video's width and height
16660 The same as @var{main_w} / @var{main_h}
16663 The main input video's sample aspect ratio
16665 @item main_dar, mdar
16666 The main input video's display aspect ratio. Calculated from
16667 @code{(main_w / main_h) * main_sar}.
16671 The main input video's horizontal and vertical chroma subsample values.
16672 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
16676 The (sequential) number of the main input frame, starting from 0.
16677 Only available with @code{eval=frame}.
16680 The presentation timestamp of the main input frame, expressed as a number of
16681 seconds. Only available with @code{eval=frame}.
16684 The position (byte offset) of the frame in the main input stream, or NaN if
16685 this information is unavailable and/or meaningless (for example in case of synthetic video).
16686 Only available with @code{eval=frame}.
16689 @subsection Examples
16693 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
16695 'scale2ref[b][a];[a][b]overlay'
16699 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
16701 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
16705 @subsection Commands
16707 This filter supports the following commands:
16711 Set the output video dimension expression.
16712 The command accepts the same syntax of the corresponding option.
16714 If the specified expression is not valid, it is kept at its current
16719 Scroll input video horizontally and/or vertically by constant speed.
16721 The filter accepts the following options:
16723 @item horizontal, h
16724 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
16725 Negative values changes scrolling direction.
16728 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
16729 Negative values changes scrolling direction.
16732 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
16735 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
16738 @subsection Commands
16740 This filter supports the following @ref{commands}:
16742 @item horizontal, h
16743 Set the horizontal scrolling speed.
16745 Set the vertical scrolling speed.
16751 Detect video scene change.
16753 This filter sets frame metadata with mafd between frame, the scene score, and
16754 forward the frame to the next filter, so they can use these metadata to detect
16755 scene change or others.
16757 In addition, this filter logs a message and sets frame metadata when it detects
16758 a scene change by @option{threshold}.
16760 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
16762 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
16763 to detect scene change.
16765 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
16766 detect scene change with @option{threshold}.
16768 The filter accepts the following options:
16772 Set the scene change detection threshold as a percentage of maximum change. Good
16773 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
16776 Default value is @code{10.}.
16779 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
16780 You can enable it if you want to get snapshot of scene change frames only.
16783 @anchor{selectivecolor}
16784 @section selectivecolor
16786 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
16787 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
16788 by the "purity" of the color (that is, how saturated it already is).
16790 This filter is similar to the Adobe Photoshop Selective Color tool.
16792 The filter accepts the following options:
16795 @item correction_method
16796 Select color correction method.
16798 Available values are:
16801 Specified adjustments are applied "as-is" (added/subtracted to original pixel
16804 Specified adjustments are relative to the original component value.
16806 Default is @code{absolute}.
16808 Adjustments for red pixels (pixels where the red component is the maximum)
16810 Adjustments for yellow pixels (pixels where the blue component is the minimum)
16812 Adjustments for green pixels (pixels where the green component is the maximum)
16814 Adjustments for cyan pixels (pixels where the red component is the minimum)
16816 Adjustments for blue pixels (pixels where the blue component is the maximum)
16818 Adjustments for magenta pixels (pixels where the green component is the minimum)
16820 Adjustments for white pixels (pixels where all components are greater than 128)
16822 Adjustments for all pixels except pure black and pure white
16824 Adjustments for black pixels (pixels where all components are lesser than 128)
16826 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
16829 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
16830 4 space separated floating point adjustment values in the [-1,1] range,
16831 respectively to adjust the amount of cyan, magenta, yellow and black for the
16832 pixels of its range.
16834 @subsection Examples
16838 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
16839 increase magenta by 27% in blue areas:
16841 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
16845 Use a Photoshop selective color preset:
16847 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
16851 @anchor{separatefields}
16852 @section separatefields
16854 The @code{separatefields} takes a frame-based video input and splits
16855 each frame into its components fields, producing a new half height clip
16856 with twice the frame rate and twice the frame count.
16858 This filter use field-dominance information in frame to decide which
16859 of each pair of fields to place first in the output.
16860 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
16862 @section setdar, setsar
16864 The @code{setdar} filter sets the Display Aspect Ratio for the filter
16867 This is done by changing the specified Sample (aka Pixel) Aspect
16868 Ratio, according to the following equation:
16870 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
16873 Keep in mind that the @code{setdar} filter does not modify the pixel
16874 dimensions of the video frame. Also, the display aspect ratio set by
16875 this filter may be changed by later filters in the filterchain,
16876 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
16879 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
16880 the filter output video.
16882 Note that as a consequence of the application of this filter, the
16883 output display aspect ratio will change according to the equation
16886 Keep in mind that the sample aspect ratio set by the @code{setsar}
16887 filter may be changed by later filters in the filterchain, e.g. if
16888 another "setsar" or a "setdar" filter is applied.
16890 It accepts the following parameters:
16893 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
16894 Set the aspect ratio used by the filter.
16896 The parameter can be a floating point number string, an expression, or
16897 a string of the form @var{num}:@var{den}, where @var{num} and
16898 @var{den} are the numerator and denominator of the aspect ratio. If
16899 the parameter is not specified, it is assumed the value "0".
16900 In case the form "@var{num}:@var{den}" is used, the @code{:} character
16904 Set the maximum integer value to use for expressing numerator and
16905 denominator when reducing the expressed aspect ratio to a rational.
16906 Default value is @code{100}.
16910 The parameter @var{sar} is an expression containing
16911 the following constants:
16915 These are approximated values for the mathematical constants e
16916 (Euler's number), pi (Greek pi), and phi (the golden ratio).
16919 The input width and height.
16922 These are the same as @var{w} / @var{h}.
16925 The input sample aspect ratio.
16928 The input display aspect ratio. It is the same as
16929 (@var{w} / @var{h}) * @var{sar}.
16932 Horizontal and vertical chroma subsample values. For example, for the
16933 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16936 @subsection Examples
16941 To change the display aspect ratio to 16:9, specify one of the following:
16948 To change the sample aspect ratio to 10:11, specify:
16954 To set a display aspect ratio of 16:9, and specify a maximum integer value of
16955 1000 in the aspect ratio reduction, use the command:
16957 setdar=ratio=16/9:max=1000
16965 Force field for the output video frame.
16967 The @code{setfield} filter marks the interlace type field for the
16968 output frames. It does not change the input frame, but only sets the
16969 corresponding property, which affects how the frame is treated by
16970 following filters (e.g. @code{fieldorder} or @code{yadif}).
16972 The filter accepts the following options:
16977 Available values are:
16981 Keep the same field property.
16984 Mark the frame as bottom-field-first.
16987 Mark the frame as top-field-first.
16990 Mark the frame as progressive.
16997 Force frame parameter for the output video frame.
16999 The @code{setparams} filter marks interlace and color range for the
17000 output frames. It does not change the input frame, but only sets the
17001 corresponding property, which affects how the frame is treated by
17006 Available values are:
17010 Keep the same field property (default).
17013 Mark the frame as bottom-field-first.
17016 Mark the frame as top-field-first.
17019 Mark the frame as progressive.
17023 Available values are:
17027 Keep the same color range property (default).
17029 @item unspecified, unknown
17030 Mark the frame as unspecified color range.
17032 @item limited, tv, mpeg
17033 Mark the frame as limited range.
17035 @item full, pc, jpeg
17036 Mark the frame as full range.
17039 @item color_primaries
17040 Set the color primaries.
17041 Available values are:
17045 Keep the same color primaries property (default).
17062 Set the color transfer.
17063 Available values are:
17067 Keep the same color trc property (default).
17089 Set the colorspace.
17090 Available values are:
17094 Keep the same colorspace property (default).
17107 @item chroma-derived-nc
17108 @item chroma-derived-c
17115 Show a line containing various information for each input video frame.
17116 The input video is not modified.
17118 This filter supports the following options:
17122 Calculate checksums of each plane. By default enabled.
17125 The shown line contains a sequence of key/value pairs of the form
17126 @var{key}:@var{value}.
17128 The following values are shown in the output:
17132 The (sequential) number of the input frame, starting from 0.
17135 The Presentation TimeStamp of the input frame, expressed as a number of
17136 time base units. The time base unit depends on the filter input pad.
17139 The Presentation TimeStamp of the input frame, expressed as a number of
17143 The position of the frame in the input stream, or -1 if this information is
17144 unavailable and/or meaningless (for example in case of synthetic video).
17147 The pixel format name.
17150 The sample aspect ratio of the input frame, expressed in the form
17151 @var{num}/@var{den}.
17154 The size of the input frame. For the syntax of this option, check the
17155 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17158 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17159 for bottom field first).
17162 This is 1 if the frame is a key frame, 0 otherwise.
17165 The picture type of the input frame ("I" for an I-frame, "P" for a
17166 P-frame, "B" for a B-frame, or "?" for an unknown type).
17167 Also refer to the documentation of the @code{AVPictureType} enum and of
17168 the @code{av_get_picture_type_char} function defined in
17169 @file{libavutil/avutil.h}.
17172 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17174 @item plane_checksum
17175 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17176 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17179 The mean value of pixels in each plane of the input frame, expressed in the form
17180 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17183 The standard deviation of pixel values in each plane of the input frame, expressed
17184 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17188 @section showpalette
17190 Displays the 256 colors palette of each frame. This filter is only relevant for
17191 @var{pal8} pixel format frames.
17193 It accepts the following option:
17197 Set the size of the box used to represent one palette color entry. Default is
17198 @code{30} (for a @code{30x30} pixel box).
17201 @section shuffleframes
17203 Reorder and/or duplicate and/or drop video frames.
17205 It accepts the following parameters:
17209 Set the destination indexes of input frames.
17210 This is space or '|' separated list of indexes that maps input frames to output
17211 frames. Number of indexes also sets maximal value that each index may have.
17212 '-1' index have special meaning and that is to drop frame.
17215 The first frame has the index 0. The default is to keep the input unchanged.
17217 @subsection Examples
17221 Swap second and third frame of every three frames of the input:
17223 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17227 Swap 10th and 1st frame of every ten frames of the input:
17229 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17233 @section shuffleplanes
17235 Reorder and/or duplicate video planes.
17237 It accepts the following parameters:
17242 The index of the input plane to be used as the first output plane.
17245 The index of the input plane to be used as the second output plane.
17248 The index of the input plane to be used as the third output plane.
17251 The index of the input plane to be used as the fourth output plane.
17255 The first plane has the index 0. The default is to keep the input unchanged.
17257 @subsection Examples
17261 Swap the second and third planes of the input:
17263 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17267 @anchor{signalstats}
17268 @section signalstats
17269 Evaluate various visual metrics that assist in determining issues associated
17270 with the digitization of analog video media.
17272 By default the filter will log these metadata values:
17276 Display the minimal Y value contained within the input frame. Expressed in
17280 Display the Y value at the 10% percentile within the input frame. Expressed in
17284 Display the average Y value within the input frame. Expressed in range of
17288 Display the Y value at the 90% percentile within the input frame. Expressed in
17292 Display the maximum Y value contained within the input frame. Expressed in
17296 Display the minimal U value contained within the input frame. Expressed in
17300 Display the U value at the 10% percentile within the input frame. Expressed in
17304 Display the average U value within the input frame. Expressed in range of
17308 Display the U value at the 90% percentile within the input frame. Expressed in
17312 Display the maximum U value contained within the input frame. Expressed in
17316 Display the minimal V value contained within the input frame. Expressed in
17320 Display the V value at the 10% percentile within the input frame. Expressed in
17324 Display the average V value within the input frame. Expressed in range of
17328 Display the V value at the 90% percentile within the input frame. Expressed in
17332 Display the maximum V value contained within the input frame. Expressed in
17336 Display the minimal saturation value contained within the input frame.
17337 Expressed in range of [0-~181.02].
17340 Display the saturation value at the 10% percentile within the input frame.
17341 Expressed in range of [0-~181.02].
17344 Display the average saturation value within the input frame. Expressed in range
17348 Display the saturation value at the 90% percentile within the input frame.
17349 Expressed in range of [0-~181.02].
17352 Display the maximum saturation value contained within the input frame.
17353 Expressed in range of [0-~181.02].
17356 Display the median value for hue within the input frame. Expressed in range of
17360 Display the average value for hue within the input frame. Expressed in range of
17364 Display the average of sample value difference between all values of the Y
17365 plane in the current frame and corresponding values of the previous input frame.
17366 Expressed in range of [0-255].
17369 Display the average of sample value difference between all values of the U
17370 plane in the current frame and corresponding values of the previous input frame.
17371 Expressed in range of [0-255].
17374 Display the average of sample value difference between all values of the V
17375 plane in the current frame and corresponding values of the previous input frame.
17376 Expressed in range of [0-255].
17379 Display bit depth of Y plane in current frame.
17380 Expressed in range of [0-16].
17383 Display bit depth of U plane in current frame.
17384 Expressed in range of [0-16].
17387 Display bit depth of V plane in current frame.
17388 Expressed in range of [0-16].
17391 The filter accepts the following options:
17397 @option{stat} specify an additional form of image analysis.
17398 @option{out} output video with the specified type of pixel highlighted.
17400 Both options accept the following values:
17404 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17405 unlike the neighboring pixels of the same field. Examples of temporal outliers
17406 include the results of video dropouts, head clogs, or tape tracking issues.
17409 Identify @var{vertical line repetition}. Vertical line repetition includes
17410 similar rows of pixels within a frame. In born-digital video vertical line
17411 repetition is common, but this pattern is uncommon in video digitized from an
17412 analog source. When it occurs in video that results from the digitization of an
17413 analog source it can indicate concealment from a dropout compensator.
17416 Identify pixels that fall outside of legal broadcast range.
17420 Set the highlight color for the @option{out} option. The default color is
17424 @subsection Examples
17428 Output data of various video metrics:
17430 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17434 Output specific data about the minimum and maximum values of the Y plane per frame:
17436 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17440 Playback video while highlighting pixels that are outside of broadcast range in red.
17442 ffplay example.mov -vf signalstats="out=brng:color=red"
17446 Playback video with signalstats metadata drawn over the frame.
17448 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17451 The contents of signalstat_drawtext.txt used in the command are:
17454 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17455 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
17456 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
17457 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
17465 Calculates the MPEG-7 Video Signature. The filter can handle more than one
17466 input. In this case the matching between the inputs can be calculated additionally.
17467 The filter always passes through the first input. The signature of each stream can
17468 be written into a file.
17470 It accepts the following options:
17474 Enable or disable the matching process.
17476 Available values are:
17480 Disable the calculation of a matching (default).
17482 Calculate the matching for the whole video and output whether the whole video
17483 matches or only parts.
17485 Calculate only until a matching is found or the video ends. Should be faster in
17490 Set the number of inputs. The option value must be a non negative integer.
17491 Default value is 1.
17494 Set the path to which the output is written. If there is more than one input,
17495 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
17496 integer), that will be replaced with the input number. If no filename is
17497 specified, no output will be written. This is the default.
17500 Choose the output format.
17502 Available values are:
17506 Use the specified binary representation (default).
17508 Use the specified xml representation.
17512 Set threshold to detect one word as similar. The option value must be an integer
17513 greater than zero. The default value is 9000.
17516 Set threshold to detect all words as similar. The option value must be an integer
17517 greater than zero. The default value is 60000.
17520 Set threshold to detect frames as similar. The option value must be an integer
17521 greater than zero. The default value is 116.
17524 Set the minimum length of a sequence in frames to recognize it as matching
17525 sequence. The option value must be a non negative integer value.
17526 The default value is 0.
17529 Set the minimum relation, that matching frames to all frames must have.
17530 The option value must be a double value between 0 and 1. The default value is 0.5.
17533 @subsection Examples
17537 To calculate the signature of an input video and store it in signature.bin:
17539 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
17543 To detect whether two videos match and store the signatures in XML format in
17544 signature0.xml and signature1.xml:
17546 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 -
17554 Blur the input video without impacting the outlines.
17556 It accepts the following options:
17559 @item luma_radius, lr
17560 Set the luma radius. The option value must be a float number in
17561 the range [0.1,5.0] that specifies the variance of the gaussian filter
17562 used to blur the image (slower if larger). Default value is 1.0.
17564 @item luma_strength, ls
17565 Set the luma strength. The option value must be a float number
17566 in the range [-1.0,1.0] that configures the blurring. A value included
17567 in [0.0,1.0] will blur the image whereas a value included in
17568 [-1.0,0.0] will sharpen the image. Default value is 1.0.
17570 @item luma_threshold, lt
17571 Set the luma threshold used as a coefficient to determine
17572 whether a pixel should be blurred or not. The option value must be an
17573 integer in the range [-30,30]. A value of 0 will filter all the image,
17574 a value included in [0,30] will filter flat areas and a value included
17575 in [-30,0] will filter edges. Default value is 0.
17577 @item chroma_radius, cr
17578 Set the chroma radius. The option value must be a float number in
17579 the range [0.1,5.0] that specifies the variance of the gaussian filter
17580 used to blur the image (slower if larger). Default value is @option{luma_radius}.
17582 @item chroma_strength, cs
17583 Set the chroma strength. The option value must be a float number
17584 in the range [-1.0,1.0] that configures the blurring. A value included
17585 in [0.0,1.0] will blur the image whereas a value included in
17586 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
17588 @item chroma_threshold, ct
17589 Set the chroma threshold used as a coefficient to determine
17590 whether a pixel should be blurred or not. The option value must be an
17591 integer in the range [-30,30]. A value of 0 will filter all the image,
17592 a value included in [0,30] will filter flat areas and a value included
17593 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
17596 If a chroma option is not explicitly set, the corresponding luma value
17600 Apply sobel operator to input video stream.
17602 The filter accepts the following option:
17606 Set which planes will be processed, unprocessed planes will be copied.
17607 By default value 0xf, all planes will be processed.
17610 Set value which will be multiplied with filtered result.
17613 Set value which will be added to filtered result.
17619 Apply a simple postprocessing filter that compresses and decompresses the image
17620 at several (or - in the case of @option{quality} level @code{6} - all) shifts
17621 and average the results.
17623 The filter accepts the following options:
17627 Set quality. This option defines the number of levels for averaging. It accepts
17628 an integer in the range 0-6. If set to @code{0}, the filter will have no
17629 effect. A value of @code{6} means the higher quality. For each increment of
17630 that value the speed drops by a factor of approximately 2. Default value is
17634 Force a constant quantization parameter. If not set, the filter will use the QP
17635 from the video stream (if available).
17638 Set thresholding mode. Available modes are:
17642 Set hard thresholding (default).
17644 Set soft thresholding (better de-ringing effect, but likely blurrier).
17647 @item use_bframe_qp
17648 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
17649 option may cause flicker since the B-Frames have often larger QP. Default is
17650 @code{0} (not enabled).
17653 @subsection Commands
17655 This filter supports the following commands:
17657 @item quality, level
17658 Set quality level. The value @code{max} can be used to set the maximum level,
17659 currently @code{6}.
17665 Scale the input by applying one of the super-resolution methods based on
17666 convolutional neural networks. Supported models:
17670 Super-Resolution Convolutional Neural Network model (SRCNN).
17671 See @url{https://arxiv.org/abs/1501.00092}.
17674 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
17675 See @url{https://arxiv.org/abs/1609.05158}.
17678 Training scripts as well as scripts for model file (.pb) saving can be found at
17679 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
17680 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
17682 Native model files (.model) can be generated from TensorFlow model
17683 files (.pb) by using tools/python/convert.py
17685 The filter accepts the following options:
17689 Specify which DNN backend to use for model loading and execution. This option accepts
17690 the following values:
17694 Native implementation of DNN loading and execution.
17697 TensorFlow backend. To enable this backend you
17698 need to install the TensorFlow for C library (see
17699 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
17700 @code{--enable-libtensorflow}
17703 Default value is @samp{native}.
17706 Set path to model file specifying network architecture and its parameters.
17707 Note that different backends use different file formats. TensorFlow backend
17708 can load files for both formats, while native backend can load files for only
17712 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
17713 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
17714 input upscaled using bicubic upscaling with proper scale factor.
17717 This feature can also be finished with @ref{dnn_processing} filter.
17721 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
17723 This filter takes in input two input videos, the first input is
17724 considered the "main" source and is passed unchanged to the
17725 output. The second input is used as a "reference" video for computing
17728 Both video inputs must have the same resolution and pixel format for
17729 this filter to work correctly. Also it assumes that both inputs
17730 have the same number of frames, which are compared one by one.
17732 The filter stores the calculated SSIM of each frame.
17734 The description of the accepted parameters follows.
17737 @item stats_file, f
17738 If specified the filter will use the named file to save the SSIM of
17739 each individual frame. When filename equals "-" the data is sent to
17743 The file printed if @var{stats_file} is selected, contains a sequence of
17744 key/value pairs of the form @var{key}:@var{value} for each compared
17747 A description of each shown parameter follows:
17751 sequential number of the input frame, starting from 1
17753 @item Y, U, V, R, G, B
17754 SSIM of the compared frames for the component specified by the suffix.
17757 SSIM of the compared frames for the whole frame.
17760 Same as above but in dB representation.
17763 This filter also supports the @ref{framesync} options.
17765 @subsection Examples
17770 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
17771 [main][ref] ssim="stats_file=stats.log" [out]
17774 On this example the input file being processed is compared with the
17775 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
17776 is stored in @file{stats.log}.
17779 Another example with both psnr and ssim at same time:
17781 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
17785 Another example with different containers:
17787 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 -
17793 Convert between different stereoscopic image formats.
17795 The filters accept the following options:
17799 Set stereoscopic image format of input.
17801 Available values for input image formats are:
17804 side by side parallel (left eye left, right eye right)
17807 side by side crosseye (right eye left, left eye right)
17810 side by side parallel with half width resolution
17811 (left eye left, right eye right)
17814 side by side crosseye with half width resolution
17815 (right eye left, left eye right)
17819 above-below (left eye above, right eye below)
17823 above-below (right eye above, left eye below)
17827 above-below with half height resolution
17828 (left eye above, right eye below)
17832 above-below with half height resolution
17833 (right eye above, left eye below)
17836 alternating frames (left eye first, right eye second)
17839 alternating frames (right eye first, left eye second)
17842 interleaved rows (left eye has top row, right eye starts on next row)
17845 interleaved rows (right eye has top row, left eye starts on next row)
17848 interleaved columns, left eye first
17851 interleaved columns, right eye first
17853 Default value is @samp{sbsl}.
17857 Set stereoscopic image format of output.
17861 side by side parallel (left eye left, right eye right)
17864 side by side crosseye (right eye left, left eye right)
17867 side by side parallel with half width resolution
17868 (left eye left, right eye right)
17871 side by side crosseye with half width resolution
17872 (right eye left, left eye right)
17876 above-below (left eye above, right eye below)
17880 above-below (right eye above, left eye below)
17884 above-below with half height resolution
17885 (left eye above, right eye below)
17889 above-below with half height resolution
17890 (right eye above, left eye below)
17893 alternating frames (left eye first, right eye second)
17896 alternating frames (right eye first, left eye second)
17899 interleaved rows (left eye has top row, right eye starts on next row)
17902 interleaved rows (right eye has top row, left eye starts on next row)
17905 anaglyph red/blue gray
17906 (red filter on left eye, blue filter on right eye)
17909 anaglyph red/green gray
17910 (red filter on left eye, green filter on right eye)
17913 anaglyph red/cyan gray
17914 (red filter on left eye, cyan filter on right eye)
17917 anaglyph red/cyan half colored
17918 (red filter on left eye, cyan filter on right eye)
17921 anaglyph red/cyan color
17922 (red filter on left eye, cyan filter on right eye)
17925 anaglyph red/cyan color optimized with the least squares projection of dubois
17926 (red filter on left eye, cyan filter on right eye)
17929 anaglyph green/magenta gray
17930 (green filter on left eye, magenta filter on right eye)
17933 anaglyph green/magenta half colored
17934 (green filter on left eye, magenta filter on right eye)
17937 anaglyph green/magenta colored
17938 (green filter on left eye, magenta filter on right eye)
17941 anaglyph green/magenta color optimized with the least squares projection of dubois
17942 (green filter on left eye, magenta filter on right eye)
17945 anaglyph yellow/blue gray
17946 (yellow filter on left eye, blue filter on right eye)
17949 anaglyph yellow/blue half colored
17950 (yellow filter on left eye, blue filter on right eye)
17953 anaglyph yellow/blue colored
17954 (yellow filter on left eye, blue filter on right eye)
17957 anaglyph yellow/blue color optimized with the least squares projection of dubois
17958 (yellow filter on left eye, blue filter on right eye)
17961 mono output (left eye only)
17964 mono output (right eye only)
17967 checkerboard, left eye first
17970 checkerboard, right eye first
17973 interleaved columns, left eye first
17976 interleaved columns, right eye first
17982 Default value is @samp{arcd}.
17985 @subsection Examples
17989 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
17995 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18001 @section streamselect, astreamselect
18002 Select video or audio streams.
18004 The filter accepts the following options:
18008 Set number of inputs. Default is 2.
18011 Set input indexes to remap to outputs.
18014 @subsection Commands
18016 The @code{streamselect} and @code{astreamselect} filter supports the following
18021 Set input indexes to remap to outputs.
18024 @subsection Examples
18028 Select first 5 seconds 1st stream and rest of time 2nd stream:
18030 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18034 Same as above, but for audio:
18036 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18043 Draw subtitles on top of input video using the libass library.
18045 To enable compilation of this filter you need to configure FFmpeg with
18046 @code{--enable-libass}. This filter also requires a build with libavcodec and
18047 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18048 Alpha) subtitles format.
18050 The filter accepts the following options:
18054 Set the filename of the subtitle file to read. It must be specified.
18056 @item original_size
18057 Specify the size of the original video, the video for which the ASS file
18058 was composed. For the syntax of this option, check the
18059 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18060 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18061 correctly scale the fonts if the aspect ratio has been changed.
18064 Set a directory path containing fonts that can be used by the filter.
18065 These fonts will be used in addition to whatever the font provider uses.
18068 Process alpha channel, by default alpha channel is untouched.
18071 Set subtitles input character encoding. @code{subtitles} filter only. Only
18072 useful if not UTF-8.
18074 @item stream_index, si
18075 Set subtitles stream index. @code{subtitles} filter only.
18078 Override default style or script info parameters of the subtitles. It accepts a
18079 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18082 If the first key is not specified, it is assumed that the first value
18083 specifies the @option{filename}.
18085 For example, to render the file @file{sub.srt} on top of the input
18086 video, use the command:
18091 which is equivalent to:
18093 subtitles=filename=sub.srt
18096 To render the default subtitles stream from file @file{video.mkv}, use:
18098 subtitles=video.mkv
18101 To render the second subtitles stream from that file, use:
18103 subtitles=video.mkv:si=1
18106 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18107 @code{DejaVu Serif}, use:
18109 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18112 @section super2xsai
18114 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18115 Interpolate) pixel art scaling algorithm.
18117 Useful for enlarging pixel art images without reducing sharpness.
18121 Swap two rectangular objects in video.
18123 This filter accepts the following options:
18133 Set 1st rect x coordinate.
18136 Set 1st rect y coordinate.
18139 Set 2nd rect x coordinate.
18142 Set 2nd rect y coordinate.
18144 All expressions are evaluated once for each frame.
18147 The all options are expressions containing the following constants:
18152 The input width and height.
18155 same as @var{w} / @var{h}
18158 input sample aspect ratio
18161 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18164 The number of the input frame, starting from 0.
18167 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18170 the position in the file of the input frame, NAN if unknown
18177 Blend successive video frames.
18183 Apply telecine process to the video.
18185 This filter accepts the following options:
18194 The default value is @code{top}.
18198 A string of numbers representing the pulldown pattern you wish to apply.
18199 The default value is @code{23}.
18203 Some typical patterns:
18208 24p: 2332 (preferred)
18215 24p: 222222222223 ("Euro pulldown")
18220 @section thistogram
18222 Compute and draw a color distribution histogram for the input video across time.
18224 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18225 at certain time, this filter shows also past histograms of number of frames defined
18226 by @code{width} option.
18228 The computed histogram is a representation of the color component
18229 distribution in an image.
18231 The filter accepts the following options:
18235 Set width of single color component output. Default value is @code{0}.
18236 Value of @code{0} means width will be picked from input video.
18237 This also set number of passed histograms to keep.
18238 Allowed range is [0, 8192].
18240 @item display_mode, d
18242 It accepts the following values:
18245 Per color component graphs are placed below each other.
18248 Per color component graphs are placed side by side.
18251 Presents information identical to that in the @code{parade}, except
18252 that the graphs representing color components are superimposed directly
18255 Default is @code{stack}.
18257 @item levels_mode, m
18258 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18259 Default is @code{linear}.
18261 @item components, c
18262 Set what color components to display.
18263 Default is @code{7}.
18266 Set background opacity. Default is @code{0.9}.
18269 Show envelope. Default is disabled.
18272 Set envelope color. Default is @code{gold}.
18277 Available values for slide is:
18280 Draw new frame when right border is reached.
18283 Replace old columns with new ones.
18286 Scroll from right to left.
18289 Scroll from left to right.
18292 Draw single picture.
18295 Default is @code{replace}.
18300 Apply threshold effect to video stream.
18302 This filter needs four video streams to perform thresholding.
18303 First stream is stream we are filtering.
18304 Second stream is holding threshold values, third stream is holding min values,
18305 and last, fourth stream is holding max values.
18307 The filter accepts the following option:
18311 Set which planes will be processed, unprocessed planes will be copied.
18312 By default value 0xf, all planes will be processed.
18315 For example if first stream pixel's component value is less then threshold value
18316 of pixel component from 2nd threshold stream, third stream value will picked,
18317 otherwise fourth stream pixel component value will be picked.
18319 Using color source filter one can perform various types of thresholding:
18321 @subsection Examples
18325 Binary threshold, using gray color as threshold:
18327 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18331 Inverted binary threshold, using gray color as threshold:
18333 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18337 Truncate binary threshold, using gray color as threshold:
18339 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18343 Threshold to zero, using gray color as threshold:
18345 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18349 Inverted threshold to zero, using gray color as threshold:
18351 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18356 Select the most representative frame in a given sequence of consecutive frames.
18358 The filter accepts the following options:
18362 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18363 will pick one of them, and then handle the next batch of @var{n} frames until
18364 the end. Default is @code{100}.
18367 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18368 value will result in a higher memory usage, so a high value is not recommended.
18370 @subsection Examples
18374 Extract one picture each 50 frames:
18380 Complete example of a thumbnail creation with @command{ffmpeg}:
18382 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18389 Tile several successive frames together.
18391 The @ref{untile} filter can do the reverse.
18393 The filter accepts the following options:
18398 Set the grid size (i.e. the number of lines and columns). For the syntax of
18399 this option, check the
18400 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18403 Set the maximum number of frames to render in the given area. It must be less
18404 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18405 the area will be used.
18408 Set the outer border margin in pixels.
18411 Set the inner border thickness (i.e. the number of pixels between frames). For
18412 more advanced padding options (such as having different values for the edges),
18413 refer to the pad video filter.
18416 Specify the color of the unused area. For the syntax of this option, check the
18417 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18418 The default value of @var{color} is "black".
18421 Set the number of frames to overlap when tiling several successive frames together.
18422 The value must be between @code{0} and @var{nb_frames - 1}.
18425 Set the number of frames to initially be empty before displaying first output frame.
18426 This controls how soon will one get first output frame.
18427 The value must be between @code{0} and @var{nb_frames - 1}.
18430 @subsection Examples
18434 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18436 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18438 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18439 duplicating each output frame to accommodate the originally detected frame
18443 Display @code{5} pictures in an area of @code{3x2} frames,
18444 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18445 mixed flat and named options:
18447 tile=3x2:nb_frames=5:padding=7:margin=2
18451 @section tinterlace
18453 Perform various types of temporal field interlacing.
18455 Frames are counted starting from 1, so the first input frame is
18458 The filter accepts the following options:
18463 Specify the mode of the interlacing. This option can also be specified
18464 as a value alone. See below for a list of values for this option.
18466 Available values are:
18470 Move odd frames into the upper field, even into the lower field,
18471 generating a double height frame at half frame rate.
18475 Frame 1 Frame 2 Frame 3 Frame 4
18477 11111 22222 33333 44444
18478 11111 22222 33333 44444
18479 11111 22222 33333 44444
18480 11111 22222 33333 44444
18494 Only output odd frames, even frames are dropped, generating a frame with
18495 unchanged height at half frame rate.
18500 Frame 1 Frame 2 Frame 3 Frame 4
18502 11111 22222 33333 44444
18503 11111 22222 33333 44444
18504 11111 22222 33333 44444
18505 11111 22222 33333 44444
18515 Only output even frames, odd frames are dropped, generating a frame with
18516 unchanged height at half frame rate.
18521 Frame 1 Frame 2 Frame 3 Frame 4
18523 11111 22222 33333 44444
18524 11111 22222 33333 44444
18525 11111 22222 33333 44444
18526 11111 22222 33333 44444
18536 Expand each frame to full height, but pad alternate lines with black,
18537 generating a frame with double height at the same input frame rate.
18542 Frame 1 Frame 2 Frame 3 Frame 4
18544 11111 22222 33333 44444
18545 11111 22222 33333 44444
18546 11111 22222 33333 44444
18547 11111 22222 33333 44444
18550 11111 ..... 33333 .....
18551 ..... 22222 ..... 44444
18552 11111 ..... 33333 .....
18553 ..... 22222 ..... 44444
18554 11111 ..... 33333 .....
18555 ..... 22222 ..... 44444
18556 11111 ..... 33333 .....
18557 ..... 22222 ..... 44444
18561 @item interleave_top, 4
18562 Interleave the upper field from odd frames with the lower field from
18563 even frames, generating a frame with unchanged height at half frame rate.
18568 Frame 1 Frame 2 Frame 3 Frame 4
18570 11111<- 22222 33333<- 44444
18571 11111 22222<- 33333 44444<-
18572 11111<- 22222 33333<- 44444
18573 11111 22222<- 33333 44444<-
18583 @item interleave_bottom, 5
18584 Interleave the lower field from odd frames with the upper field from
18585 even frames, generating a frame with unchanged height at half frame rate.
18590 Frame 1 Frame 2 Frame 3 Frame 4
18592 11111 22222<- 33333 44444<-
18593 11111<- 22222 33333<- 44444
18594 11111 22222<- 33333 44444<-
18595 11111<- 22222 33333<- 44444
18605 @item interlacex2, 6
18606 Double frame rate with unchanged height. Frames are inserted each
18607 containing the second temporal field from the previous input frame and
18608 the first temporal field from the next input frame. This mode relies on
18609 the top_field_first flag. Useful for interlaced video displays with no
18610 field synchronisation.
18615 Frame 1 Frame 2 Frame 3 Frame 4
18617 11111 22222 33333 44444
18618 11111 22222 33333 44444
18619 11111 22222 33333 44444
18620 11111 22222 33333 44444
18623 11111 22222 22222 33333 33333 44444 44444
18624 11111 11111 22222 22222 33333 33333 44444
18625 11111 22222 22222 33333 33333 44444 44444
18626 11111 11111 22222 22222 33333 33333 44444
18631 Move odd frames into the upper field, even into the lower field,
18632 generating a double height frame at same frame rate.
18637 Frame 1 Frame 2 Frame 3 Frame 4
18639 11111 22222 33333 44444
18640 11111 22222 33333 44444
18641 11111 22222 33333 44444
18642 11111 22222 33333 44444
18645 11111 33333 33333 55555
18646 22222 22222 44444 44444
18647 11111 33333 33333 55555
18648 22222 22222 44444 44444
18649 11111 33333 33333 55555
18650 22222 22222 44444 44444
18651 11111 33333 33333 55555
18652 22222 22222 44444 44444
18657 Numeric values are deprecated but are accepted for backward
18658 compatibility reasons.
18660 Default mode is @code{merge}.
18663 Specify flags influencing the filter process.
18665 Available value for @var{flags} is:
18668 @item low_pass_filter, vlpf
18669 Enable linear vertical low-pass filtering in the filter.
18670 Vertical low-pass filtering is required when creating an interlaced
18671 destination from a progressive source which contains high-frequency
18672 vertical detail. Filtering will reduce interlace 'twitter' and Moire
18675 @item complex_filter, cvlpf
18676 Enable complex vertical low-pass filtering.
18677 This will slightly less reduce interlace 'twitter' and Moire
18678 patterning but better retain detail and subjective sharpness impression.
18681 Bypass already interlaced frames, only adjust the frame rate.
18684 Vertical low-pass filtering and bypassing already interlaced frames can only be
18685 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
18690 Pick median pixels from several successive input video frames.
18692 The filter accepts the following options:
18696 Set radius of median filter.
18697 Default is 1. Allowed range is from 1 to 127.
18700 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
18703 Set median percentile. Default value is @code{0.5}.
18704 Default value of @code{0.5} will pick always median values, while @code{0} will pick
18705 minimum values, and @code{1} maximum values.
18710 Mix successive video frames.
18712 A description of the accepted options follows.
18716 The number of successive frames to mix. If unspecified, it defaults to 3.
18719 Specify weight of each input video frame.
18720 Each weight is separated by space. If number of weights is smaller than
18721 number of @var{frames} last specified weight will be used for all remaining
18725 Specify scale, if it is set it will be multiplied with sum
18726 of each weight multiplied with pixel values to give final destination
18727 pixel value. By default @var{scale} is auto scaled to sum of weights.
18730 @subsection Examples
18734 Average 7 successive frames:
18736 tmix=frames=7:weights="1 1 1 1 1 1 1"
18740 Apply simple temporal convolution:
18742 tmix=frames=3:weights="-1 3 -1"
18746 Similar as above but only showing temporal differences:
18748 tmix=frames=3:weights="-1 2 -1":scale=1
18754 Tone map colors from different dynamic ranges.
18756 This filter expects data in single precision floating point, as it needs to
18757 operate on (and can output) out-of-range values. Another filter, such as
18758 @ref{zscale}, is needed to convert the resulting frame to a usable format.
18760 The tonemapping algorithms implemented only work on linear light, so input
18761 data should be linearized beforehand (and possibly correctly tagged).
18764 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
18767 @subsection Options
18768 The filter accepts the following options.
18772 Set the tone map algorithm to use.
18774 Possible values are:
18777 Do not apply any tone map, only desaturate overbright pixels.
18780 Hard-clip any out-of-range values. Use it for perfect color accuracy for
18781 in-range values, while distorting out-of-range values.
18784 Stretch the entire reference gamut to a linear multiple of the display.
18787 Fit a logarithmic transfer between the tone curves.
18790 Preserve overall image brightness with a simple curve, using nonlinear
18791 contrast, which results in flattening details and degrading color accuracy.
18794 Preserve both dark and bright details better than @var{reinhard}, at the cost
18795 of slightly darkening everything. Use it when detail preservation is more
18796 important than color and brightness accuracy.
18799 Smoothly map out-of-range values, while retaining contrast and colors for
18800 in-range material as much as possible. Use it when color accuracy is more
18801 important than detail preservation.
18807 Tune the tone mapping algorithm.
18809 This affects the following algorithms:
18815 Specifies the scale factor to use while stretching.
18819 Specifies the exponent of the function.
18823 Specify an extra linear coefficient to multiply into the signal before clipping.
18827 Specify the local contrast coefficient at the display peak.
18828 Default to 0.5, which means that in-gamut values will be about half as bright
18835 Specify the transition point from linear to mobius transform. Every value
18836 below this point is guaranteed to be mapped 1:1. The higher the value, the
18837 more accurate the result will be, at the cost of losing bright details.
18838 Default to 0.3, which due to the steep initial slope still preserves in-range
18839 colors fairly accurately.
18843 Apply desaturation for highlights that exceed this level of brightness. The
18844 higher the parameter, the more color information will be preserved. This
18845 setting helps prevent unnaturally blown-out colors for super-highlights, by
18846 (smoothly) turning into white instead. This makes images feel more natural,
18847 at the cost of reducing information about out-of-range colors.
18849 The default of 2.0 is somewhat conservative and will mostly just apply to
18850 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
18852 This option works only if the input frame has a supported color tag.
18855 Override signal/nominal/reference peak with this value. Useful when the
18856 embedded peak information in display metadata is not reliable or when tone
18857 mapping from a lower range to a higher range.
18862 Temporarily pad video frames.
18864 The filter accepts the following options:
18868 Specify number of delay frames before input video stream. Default is 0.
18871 Specify number of padding frames after input video stream.
18872 Set to -1 to pad indefinitely. Default is 0.
18875 Set kind of frames added to beginning of stream.
18876 Can be either @var{add} or @var{clone}.
18877 With @var{add} frames of solid-color are added.
18878 With @var{clone} frames are clones of first frame.
18879 Default is @var{add}.
18882 Set kind of frames added to end of stream.
18883 Can be either @var{add} or @var{clone}.
18884 With @var{add} frames of solid-color are added.
18885 With @var{clone} frames are clones of last frame.
18886 Default is @var{add}.
18888 @item start_duration, stop_duration
18889 Specify the duration of the start/stop delay. See
18890 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18891 for the accepted syntax.
18892 These options override @var{start} and @var{stop}. Default is 0.
18895 Specify the color of the padded area. For the syntax of this option,
18896 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
18897 manual,ffmpeg-utils}.
18899 The default value of @var{color} is "black".
18905 Transpose rows with columns in the input video and optionally flip it.
18907 It accepts the following parameters:
18912 Specify the transposition direction.
18914 Can assume the following values:
18916 @item 0, 4, cclock_flip
18917 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
18925 Rotate by 90 degrees clockwise, that is:
18933 Rotate by 90 degrees counterclockwise, that is:
18940 @item 3, 7, clock_flip
18941 Rotate by 90 degrees clockwise and vertically flip, that is:
18949 For values between 4-7, the transposition is only done if the input
18950 video geometry is portrait and not landscape. These values are
18951 deprecated, the @code{passthrough} option should be used instead.
18953 Numerical values are deprecated, and should be dropped in favor of
18954 symbolic constants.
18957 Do not apply the transposition if the input geometry matches the one
18958 specified by the specified value. It accepts the following values:
18961 Always apply transposition.
18963 Preserve portrait geometry (when @var{height} >= @var{width}).
18965 Preserve landscape geometry (when @var{width} >= @var{height}).
18968 Default value is @code{none}.
18971 For example to rotate by 90 degrees clockwise and preserve portrait
18974 transpose=dir=1:passthrough=portrait
18977 The command above can also be specified as:
18979 transpose=1:portrait
18982 @section transpose_npp
18984 Transpose rows with columns in the input video and optionally flip it.
18985 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
18987 It accepts the following parameters:
18992 Specify the transposition direction.
18994 Can assume the following values:
18997 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19000 Rotate by 90 degrees clockwise.
19003 Rotate by 90 degrees counterclockwise.
19006 Rotate by 90 degrees clockwise and vertically flip.
19010 Do not apply the transposition if the input geometry matches the one
19011 specified by the specified value. It accepts the following values:
19014 Always apply transposition. (default)
19016 Preserve portrait geometry (when @var{height} >= @var{width}).
19018 Preserve landscape geometry (when @var{width} >= @var{height}).
19024 Trim the input so that the output contains one continuous subpart of the input.
19026 It accepts the following parameters:
19029 Specify the time of the start of the kept section, i.e. the frame with the
19030 timestamp @var{start} will be the first frame in the output.
19033 Specify the time of the first frame that will be dropped, i.e. the frame
19034 immediately preceding the one with the timestamp @var{end} will be the last
19035 frame in the output.
19038 This is the same as @var{start}, except this option sets the start timestamp
19039 in timebase units instead of seconds.
19042 This is the same as @var{end}, except this option sets the end timestamp
19043 in timebase units instead of seconds.
19046 The maximum duration of the output in seconds.
19049 The number of the first frame that should be passed to the output.
19052 The number of the first frame that should be dropped.
19055 @option{start}, @option{end}, and @option{duration} are expressed as time
19056 duration specifications; see
19057 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19058 for the accepted syntax.
19060 Note that the first two sets of the start/end options and the @option{duration}
19061 option look at the frame timestamp, while the _frame variants simply count the
19062 frames that pass through the filter. Also note that this filter does not modify
19063 the timestamps. If you wish for the output timestamps to start at zero, insert a
19064 setpts filter after the trim filter.
19066 If multiple start or end options are set, this filter tries to be greedy and
19067 keep all the frames that match at least one of the specified constraints. To keep
19068 only the part that matches all the constraints at once, chain multiple trim
19071 The defaults are such that all the input is kept. So it is possible to set e.g.
19072 just the end values to keep everything before the specified time.
19077 Drop everything except the second minute of input:
19079 ffmpeg -i INPUT -vf trim=60:120
19083 Keep only the first second:
19085 ffmpeg -i INPUT -vf trim=duration=1
19090 @section unpremultiply
19091 Apply alpha unpremultiply effect to input video stream using first plane
19092 of second stream as alpha.
19094 Both streams must have same dimensions and same pixel format.
19096 The filter accepts the following option:
19100 Set which planes will be processed, unprocessed planes will be copied.
19101 By default value 0xf, all planes will be processed.
19103 If the format has 1 or 2 components, then luma is bit 0.
19104 If the format has 3 or 4 components:
19105 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19106 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19107 If present, the alpha channel is always the last bit.
19110 Do not require 2nd input for processing, instead use alpha plane from input stream.
19116 Sharpen or blur the input video.
19118 It accepts the following parameters:
19121 @item luma_msize_x, lx
19122 Set the luma matrix horizontal size. It must be an odd integer between
19123 3 and 23. The default value is 5.
19125 @item luma_msize_y, ly
19126 Set the luma matrix vertical size. It must be an odd integer between 3
19127 and 23. The default value is 5.
19129 @item luma_amount, la
19130 Set the luma effect strength. It must be a floating point number, reasonable
19131 values lay between -1.5 and 1.5.
19133 Negative values will blur the input video, while positive values will
19134 sharpen it, a value of zero will disable the effect.
19136 Default value is 1.0.
19138 @item chroma_msize_x, cx
19139 Set the chroma matrix horizontal size. It must be an odd integer
19140 between 3 and 23. The default value is 5.
19142 @item chroma_msize_y, cy
19143 Set the chroma matrix vertical size. It must be an odd integer
19144 between 3 and 23. The default value is 5.
19146 @item chroma_amount, ca
19147 Set the chroma effect strength. It must be a floating point number, reasonable
19148 values lay between -1.5 and 1.5.
19150 Negative values will blur the input video, while positive values will
19151 sharpen it, a value of zero will disable the effect.
19153 Default value is 0.0.
19157 All parameters are optional and default to the equivalent of the
19158 string '5:5:1.0:5:5:0.0'.
19160 @subsection Examples
19164 Apply strong luma sharpen effect:
19166 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19170 Apply a strong blur of both luma and chroma parameters:
19172 unsharp=7:7:-2:7:7:-2
19179 Decompose a video made of tiled images into the individual images.
19181 The frame rate of the output video is the frame rate of the input video
19182 multiplied by the number of tiles.
19184 This filter does the reverse of @ref{tile}.
19186 The filter accepts the following options:
19191 Set the grid size (i.e. the number of lines and columns). For the syntax of
19192 this option, check the
19193 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19196 @subsection Examples
19200 Produce a 1-second video from a still image file made of 25 frames stacked
19201 vertically, like an analogic film reel:
19203 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19209 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19210 the image at several (or - in the case of @option{quality} level @code{8} - all)
19211 shifts and average the results.
19213 The way this differs from the behavior of spp is that uspp actually encodes &
19214 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19215 DCT similar to MJPEG.
19217 The filter accepts the following options:
19221 Set quality. This option defines the number of levels for averaging. It accepts
19222 an integer in the range 0-8. If set to @code{0}, the filter will have no
19223 effect. A value of @code{8} means the higher quality. For each increment of
19224 that value the speed drops by a factor of approximately 2. Default value is
19228 Force a constant quantization parameter. If not set, the filter will use the QP
19229 from the video stream (if available).
19234 Convert 360 videos between various formats.
19236 The filter accepts the following options:
19242 Set format of the input/output video.
19250 Equirectangular projection.
19255 Cubemap with 3x2/6x1/1x6 layout.
19257 Format specific options:
19262 Set padding proportion for the input/output cubemap. Values in decimals.
19269 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)
19272 Default value is @b{@samp{0}}.
19273 Maximum value is @b{@samp{0.1}}.
19277 Set fixed padding for the input/output cubemap. Values in pixels.
19279 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19283 Set order of faces for the input/output cubemap. Choose one direction for each position.
19285 Designation of directions:
19301 Default value is @b{@samp{rludfb}}.
19305 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
19307 Designation of angles:
19310 0 degrees clockwise
19312 90 degrees clockwise
19314 180 degrees clockwise
19316 270 degrees clockwise
19319 Default value is @b{@samp{000000}}.
19323 Equi-Angular Cubemap.
19330 Format specific options:
19335 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19337 If diagonal field of view is set it overrides horizontal and vertical field of view.
19342 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19344 If diagonal field of view is set it overrides horizontal and vertical field of view.
19350 Format specific options:
19355 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19357 If diagonal field of view is set it overrides horizontal and vertical field of view.
19362 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19364 If diagonal field of view is set it overrides horizontal and vertical field of view.
19370 Facebook's 360 formats.
19373 Stereographic format.
19375 Format specific options:
19380 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19382 If diagonal field of view is set it overrides horizontal and vertical field of view.
19387 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19389 If diagonal field of view is set it overrides horizontal and vertical field of view.
19396 Ball format, gives significant distortion toward the back.
19399 Hammer-Aitoff map projection format.
19402 Sinusoidal map projection format.
19405 Fisheye projection.
19407 Format specific options:
19412 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19414 If diagonal field of view is set it overrides horizontal and vertical field of view.
19419 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19421 If diagonal field of view is set it overrides horizontal and vertical field of view.
19425 Pannini projection.
19427 Format specific options:
19430 Set output pannini parameter.
19433 Set input pannini parameter.
19437 Cylindrical projection.
19439 Format specific options:
19444 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19446 If diagonal field of view is set it overrides horizontal and vertical field of view.
19451 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19453 If diagonal field of view is set it overrides horizontal and vertical field of view.
19457 Perspective projection. @i{(output only)}
19459 Format specific options:
19462 Set perspective parameter.
19466 Tetrahedron projection.
19469 Truncated square pyramid projection.
19473 Half equirectangular projection.
19478 Format specific options:
19483 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19485 If diagonal field of view is set it overrides horizontal and vertical field of view.
19490 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19492 If diagonal field of view is set it overrides horizontal and vertical field of view.
19496 Orthographic format.
19498 Format specific options:
19503 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19505 If diagonal field of view is set it overrides horizontal and vertical field of view.
19510 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19512 If diagonal field of view is set it overrides horizontal and vertical field of view.
19516 Octahedron projection.
19520 Set interpolation method.@*
19521 @i{Note: more complex interpolation methods require much more memory to run.}
19531 Bilinear interpolation.
19533 Lagrange9 interpolation.
19536 Bicubic interpolation.
19539 Lanczos interpolation.
19542 Spline16 interpolation.
19545 Gaussian interpolation.
19548 Default value is @b{@samp{line}}.
19552 Set the output video resolution.
19554 Default resolution depends on formats.
19558 Set the input/output stereo format.
19569 Default value is @b{@samp{2d}} for input and output format.
19574 Set rotation for the output video. Values in degrees.
19577 Set rotation order for the output video. Choose one item for each position.
19588 Default value is @b{@samp{ypr}}.
19593 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
19597 Set if input video is flipped horizontally/vertically. Boolean values.
19600 Set if input video is transposed. Boolean value, by default disabled.
19603 Set if output video needs to be transposed. Boolean value, by default disabled.
19606 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
19609 @subsection Examples
19613 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
19615 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
19618 Extract back view of Equi-Angular Cubemap:
19620 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
19623 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
19625 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
19629 @subsection Commands
19631 This filter supports subset of above options as @ref{commands}.
19633 @section vaguedenoiser
19635 Apply a wavelet based denoiser.
19637 It transforms each frame from the video input into the wavelet domain,
19638 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
19639 the obtained coefficients. It does an inverse wavelet transform after.
19640 Due to wavelet properties, it should give a nice smoothed result, and
19641 reduced noise, without blurring picture features.
19643 This filter accepts the following options:
19647 The filtering strength. The higher, the more filtered the video will be.
19648 Hard thresholding can use a higher threshold than soft thresholding
19649 before the video looks overfiltered. Default value is 2.
19652 The filtering method the filter will use.
19654 It accepts the following values:
19657 All values under the threshold will be zeroed.
19660 All values under the threshold will be zeroed. All values above will be
19661 reduced by the threshold.
19664 Scales or nullifies coefficients - intermediary between (more) soft and
19665 (less) hard thresholding.
19668 Default is garrote.
19671 Number of times, the wavelet will decompose the picture. Picture can't
19672 be decomposed beyond a particular point (typically, 8 for a 640x480
19673 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
19676 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
19679 A list of the planes to process. By default all planes are processed.
19682 The threshold type the filter will use.
19684 It accepts the following values:
19687 Threshold used is same for all decompositions.
19690 Threshold used depends also on each decomposition coefficients.
19693 Default is universal.
19696 @section vectorscope
19698 Display 2 color component values in the two dimensional graph (which is called
19701 This filter accepts the following options:
19705 Set vectorscope mode.
19707 It accepts the following values:
19711 Gray values are displayed on graph, higher brightness means more pixels have
19712 same component color value on location in graph. This is the default mode.
19715 Gray values are displayed on graph. Surrounding pixels values which are not
19716 present in video frame are drawn in gradient of 2 color components which are
19717 set by option @code{x} and @code{y}. The 3rd color component is static.
19720 Actual color components values present in video frame are displayed on graph.
19723 Similar as color2 but higher frequency of same values @code{x} and @code{y}
19724 on graph increases value of another color component, which is luminance by
19725 default values of @code{x} and @code{y}.
19728 Actual colors present in video frame are displayed on graph. If two different
19729 colors map to same position on graph then color with higher value of component
19730 not present in graph is picked.
19733 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
19734 component picked from radial gradient.
19738 Set which color component will be represented on X-axis. Default is @code{1}.
19741 Set which color component will be represented on Y-axis. Default is @code{2}.
19744 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
19745 of color component which represents frequency of (X, Y) location in graph.
19750 No envelope, this is default.
19753 Instant envelope, even darkest single pixel will be clearly highlighted.
19756 Hold maximum and minimum values presented in graph over time. This way you
19757 can still spot out of range values without constantly looking at vectorscope.
19760 Peak and instant envelope combined together.
19764 Set what kind of graticule to draw.
19773 Set graticule opacity.
19776 Set graticule flags.
19780 Draw graticule for white point.
19783 Draw graticule for black point.
19786 Draw color points short names.
19790 Set background opacity.
19792 @item lthreshold, l
19793 Set low threshold for color component not represented on X or Y axis.
19794 Values lower than this value will be ignored. Default is 0.
19795 Note this value is multiplied with actual max possible value one pixel component
19796 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
19799 @item hthreshold, h
19800 Set high threshold for color component not represented on X or Y axis.
19801 Values higher than this value will be ignored. Default is 1.
19802 Note this value is multiplied with actual max possible value one pixel component
19803 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
19804 is 0.9 * 255 = 230.
19806 @item colorspace, c
19807 Set what kind of colorspace to use when drawing graticule.
19817 Set color tint for gray/tint vectorscope mode. By default both options are zero.
19818 This means no tint, and output will remain gray.
19821 @anchor{vidstabdetect}
19822 @section vidstabdetect
19824 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
19825 @ref{vidstabtransform} for pass 2.
19827 This filter generates a file with relative translation and rotation
19828 transform information about subsequent frames, which is then used by
19829 the @ref{vidstabtransform} filter.
19831 To enable compilation of this filter you need to configure FFmpeg with
19832 @code{--enable-libvidstab}.
19834 This filter accepts the following options:
19838 Set the path to the file used to write the transforms information.
19839 Default value is @file{transforms.trf}.
19842 Set how shaky the video is and how quick the camera is. It accepts an
19843 integer in the range 1-10, a value of 1 means little shakiness, a
19844 value of 10 means strong shakiness. Default value is 5.
19847 Set the accuracy of the detection process. It must be a value in the
19848 range 1-15. A value of 1 means low accuracy, a value of 15 means high
19849 accuracy. Default value is 15.
19852 Set stepsize of the search process. The region around minimum is
19853 scanned with 1 pixel resolution. Default value is 6.
19856 Set minimum contrast. Below this value a local measurement field is
19857 discarded. Must be a floating point value in the range 0-1. Default
19861 Set reference frame number for tripod mode.
19863 If enabled, the motion of the frames is compared to a reference frame
19864 in the filtered stream, identified by the specified number. The idea
19865 is to compensate all movements in a more-or-less static scene and keep
19866 the camera view absolutely still.
19868 If set to 0, it is disabled. The frames are counted starting from 1.
19871 Show fields and transforms in the resulting frames. It accepts an
19872 integer in the range 0-2. Default value is 0, which disables any
19876 @subsection Examples
19880 Use default values:
19886 Analyze strongly shaky movie and put the results in file
19887 @file{mytransforms.trf}:
19889 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
19893 Visualize the result of internal transformations in the resulting
19896 vidstabdetect=show=1
19900 Analyze a video with medium shakiness using @command{ffmpeg}:
19902 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
19906 @anchor{vidstabtransform}
19907 @section vidstabtransform
19909 Video stabilization/deshaking: pass 2 of 2,
19910 see @ref{vidstabdetect} for pass 1.
19912 Read a file with transform information for each frame and
19913 apply/compensate them. Together with the @ref{vidstabdetect}
19914 filter this can be used to deshake videos. See also
19915 @url{http://public.hronopik.de/vid.stab}. It is important to also use
19916 the @ref{unsharp} filter, see below.
19918 To enable compilation of this filter you need to configure FFmpeg with
19919 @code{--enable-libvidstab}.
19921 @subsection Options
19925 Set path to the file used to read the transforms. Default value is
19926 @file{transforms.trf}.
19929 Set the number of frames (value*2 + 1) used for lowpass filtering the
19930 camera movements. Default value is 10.
19932 For example a number of 10 means that 21 frames are used (10 in the
19933 past and 10 in the future) to smoothen the motion in the video. A
19934 larger value leads to a smoother video, but limits the acceleration of
19935 the camera (pan/tilt movements). 0 is a special case where a static
19936 camera is simulated.
19939 Set the camera path optimization algorithm.
19941 Accepted values are:
19944 gaussian kernel low-pass filter on camera motion (default)
19946 averaging on transformations
19950 Set maximal number of pixels to translate frames. Default value is -1,
19954 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
19955 value is -1, meaning no limit.
19958 Specify how to deal with borders that may be visible due to movement
19961 Available values are:
19964 keep image information from previous frame (default)
19966 fill the border black
19970 Invert transforms if set to 1. Default value is 0.
19973 Consider transforms as relative to previous frame if set to 1,
19974 absolute if set to 0. Default value is 0.
19977 Set percentage to zoom. A positive value will result in a zoom-in
19978 effect, a negative value in a zoom-out effect. Default value is 0 (no
19982 Set optimal zooming to avoid borders.
19984 Accepted values are:
19989 optimal static zoom value is determined (only very strong movements
19990 will lead to visible borders) (default)
19992 optimal adaptive zoom value is determined (no borders will be
19993 visible), see @option{zoomspeed}
19996 Note that the value given at zoom is added to the one calculated here.
19999 Set percent to zoom maximally each frame (enabled when
20000 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20004 Specify type of interpolation.
20006 Available values are:
20011 linear only horizontal
20013 linear in both directions (default)
20015 cubic in both directions (slow)
20019 Enable virtual tripod mode if set to 1, which is equivalent to
20020 @code{relative=0:smoothing=0}. Default value is 0.
20022 Use also @code{tripod} option of @ref{vidstabdetect}.
20025 Increase log verbosity if set to 1. Also the detected global motions
20026 are written to the temporary file @file{global_motions.trf}. Default
20030 @subsection Examples
20034 Use @command{ffmpeg} for a typical stabilization with default values:
20036 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20039 Note the use of the @ref{unsharp} filter which is always recommended.
20042 Zoom in a bit more and load transform data from a given file:
20044 vidstabtransform=zoom=5:input="mytransforms.trf"
20048 Smoothen the video even more:
20050 vidstabtransform=smoothing=30
20056 Flip the input video vertically.
20058 For example, to vertically flip a video with @command{ffmpeg}:
20060 ffmpeg -i in.avi -vf "vflip" out.avi
20065 Detect variable frame rate video.
20067 This filter tries to detect if the input is variable or constant frame rate.
20069 At end it will output number of frames detected as having variable delta pts,
20070 and ones with constant delta pts.
20071 If there was frames with variable delta, than it will also show min, max and
20072 average delta encountered.
20076 Boost or alter saturation.
20078 The filter accepts the following options:
20081 Set strength of boost if positive value or strength of alter if negative value.
20082 Default is 0. Allowed range is from -2 to 2.
20085 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20088 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20091 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20094 Set the red luma coefficient.
20097 Set the green luma coefficient.
20100 Set the blue luma coefficient.
20103 If @code{intensity} is negative and this is set to 1, colors will change,
20104 otherwise colors will be less saturated, more towards gray.
20107 @subsection Commands
20109 This filter supports the all above options as @ref{commands}.
20114 Make or reverse a natural vignetting effect.
20116 The filter accepts the following options:
20120 Set lens angle expression as a number of radians.
20122 The value is clipped in the @code{[0,PI/2]} range.
20124 Default value: @code{"PI/5"}
20128 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20132 Set forward/backward mode.
20134 Available modes are:
20137 The larger the distance from the central point, the darker the image becomes.
20140 The larger the distance from the central point, the brighter the image becomes.
20141 This can be used to reverse a vignette effect, though there is no automatic
20142 detection to extract the lens @option{angle} and other settings (yet). It can
20143 also be used to create a burning effect.
20146 Default value is @samp{forward}.
20149 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20151 It accepts the following values:
20154 Evaluate expressions only once during the filter initialization.
20157 Evaluate expressions for each incoming frame. This is way slower than the
20158 @samp{init} mode since it requires all the scalers to be re-computed, but it
20159 allows advanced dynamic expressions.
20162 Default value is @samp{init}.
20165 Set dithering to reduce the circular banding effects. Default is @code{1}
20169 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20170 Setting this value to the SAR of the input will make a rectangular vignetting
20171 following the dimensions of the video.
20173 Default is @code{1/1}.
20176 @subsection Expressions
20178 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20179 following parameters.
20184 input width and height
20187 the number of input frame, starting from 0
20190 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20191 @var{TB} units, NAN if undefined
20194 frame rate of the input video, NAN if the input frame rate is unknown
20197 the PTS (Presentation TimeStamp) of the filtered video frame,
20198 expressed in seconds, NAN if undefined
20201 time base of the input video
20205 @subsection Examples
20209 Apply simple strong vignetting effect:
20215 Make a flickering vignetting:
20217 vignette='PI/4+random(1)*PI/50':eval=frame
20222 @section vmafmotion
20224 Obtain the average VMAF motion score of a video.
20225 It is one of the component metrics of VMAF.
20227 The obtained average motion score is printed through the logging system.
20229 The filter accepts the following options:
20233 If specified, the filter will use the named file to save the motion score of
20234 each frame with respect to the previous frame.
20235 When filename equals "-" the data is sent to standard output.
20240 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20244 Stack input videos vertically.
20246 All streams must be of same pixel format and of same width.
20248 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20249 to create same output.
20251 The filter accepts the following options:
20255 Set number of input streams. Default is 2.
20258 If set to 1, force the output to terminate when the shortest input
20259 terminates. Default value is 0.
20264 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20265 Deinterlacing Filter").
20267 Based on the process described by Martin Weston for BBC R&D, and
20268 implemented based on the de-interlace algorithm written by Jim
20269 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20270 uses filter coefficients calculated by BBC R&D.
20272 This filter uses field-dominance information in frame to decide which
20273 of each pair of fields to place first in the output.
20274 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20276 There are two sets of filter coefficients, so called "simple"
20277 and "complex". Which set of filter coefficients is used can
20278 be set by passing an optional parameter:
20282 Set the interlacing filter coefficients. Accepts one of the following values:
20286 Simple filter coefficient set.
20288 More-complex filter coefficient set.
20290 Default value is @samp{complex}.
20293 Specify which frames to deinterlace. Accepts one of the following values:
20297 Deinterlace all frames,
20299 Only deinterlace frames marked as interlaced.
20302 Default value is @samp{all}.
20306 Video waveform monitor.
20308 The waveform monitor plots color component intensity. By default luminance
20309 only. Each column of the waveform corresponds to a column of pixels in the
20312 It accepts the following options:
20316 Can be either @code{row}, or @code{column}. Default is @code{column}.
20317 In row mode, the graph on the left side represents color component value 0 and
20318 the right side represents value = 255. In column mode, the top side represents
20319 color component value = 0 and bottom side represents value = 255.
20322 Set intensity. Smaller values are useful to find out how many values of the same
20323 luminance are distributed across input rows/columns.
20324 Default value is @code{0.04}. Allowed range is [0, 1].
20327 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
20328 In mirrored mode, higher values will be represented on the left
20329 side for @code{row} mode and at the top for @code{column} mode. Default is
20330 @code{1} (mirrored).
20334 It accepts the following values:
20337 Presents information identical to that in the @code{parade}, except
20338 that the graphs representing color components are superimposed directly
20341 This display mode makes it easier to spot relative differences or similarities
20342 in overlapping areas of the color components that are supposed to be identical,
20343 such as neutral whites, grays, or blacks.
20346 Display separate graph for the color components side by side in
20347 @code{row} mode or one below the other in @code{column} mode.
20350 Display separate graph for the color components side by side in
20351 @code{column} mode or one below the other in @code{row} mode.
20353 Using this display mode makes it easy to spot color casts in the highlights
20354 and shadows of an image, by comparing the contours of the top and the bottom
20355 graphs of each waveform. Since whites, grays, and blacks are characterized
20356 by exactly equal amounts of red, green, and blue, neutral areas of the picture
20357 should display three waveforms of roughly equal width/height. If not, the
20358 correction is easy to perform by making level adjustments the three waveforms.
20360 Default is @code{stack}.
20362 @item components, c
20363 Set which color components to display. Default is 1, which means only luminance
20364 or red color component if input is in RGB colorspace. If is set for example to
20365 7 it will display all 3 (if) available color components.
20370 No envelope, this is default.
20373 Instant envelope, minimum and maximum values presented in graph will be easily
20374 visible even with small @code{step} value.
20377 Hold minimum and maximum values presented in graph across time. This way you
20378 can still spot out of range values without constantly looking at waveforms.
20381 Peak and instant envelope combined together.
20387 No filtering, this is default.
20390 Luma and chroma combined together.
20393 Similar as above, but shows difference between blue and red chroma.
20396 Similar as above, but use different colors.
20399 Similar as above, but again with different colors.
20402 Displays only chroma.
20405 Displays actual color value on waveform.
20408 Similar as above, but with luma showing frequency of chroma values.
20412 Set which graticule to display.
20416 Do not display graticule.
20419 Display green graticule showing legal broadcast ranges.
20422 Display orange graticule showing legal broadcast ranges.
20425 Display invert graticule showing legal broadcast ranges.
20429 Set graticule opacity.
20432 Set graticule flags.
20436 Draw numbers above lines. By default enabled.
20439 Draw dots instead of lines.
20443 Set scale used for displaying graticule.
20450 Default is digital.
20453 Set background opacity.
20457 Set tint for output.
20458 Only used with lowpass filter and when display is not overlay and input
20459 pixel formats are not RGB.
20462 @section weave, doubleweave
20464 The @code{weave} takes a field-based video input and join
20465 each two sequential fields into single frame, producing a new double
20466 height clip with half the frame rate and half the frame count.
20468 The @code{doubleweave} works same as @code{weave} but without
20469 halving frame rate and frame count.
20471 It accepts the following option:
20475 Set first field. Available values are:
20479 Set the frame as top-field-first.
20482 Set the frame as bottom-field-first.
20486 @subsection Examples
20490 Interlace video using @ref{select} and @ref{separatefields} filter:
20492 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
20497 Apply the xBR high-quality magnification filter which is designed for pixel
20498 art. It follows a set of edge-detection rules, see
20499 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
20501 It accepts the following option:
20505 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
20506 @code{3xBR} and @code{4} for @code{4xBR}.
20507 Default is @code{3}.
20512 Apply cross fade from one input video stream to another input video stream.
20513 The cross fade is applied for specified duration.
20515 The filter accepts the following options:
20519 Set one of available transition effects:
20565 Default transition effect is fade.
20568 Set cross fade duration in seconds.
20569 Default duration is 1 second.
20572 Set cross fade start relative to first input stream in seconds.
20573 Default offset is 0.
20576 Set expression for custom transition effect.
20578 The expressions can use the following variables and functions:
20583 The coordinates of the current sample.
20587 The width and height of the image.
20590 Progress of transition effect.
20593 Currently processed plane.
20596 Return value of first input at current location and plane.
20599 Return value of second input at current location and plane.
20605 Return the value of the pixel at location (@var{x},@var{y}) of the
20606 first/second/third/fourth component of first input.
20612 Return the value of the pixel at location (@var{x},@var{y}) of the
20613 first/second/third/fourth component of second input.
20617 @subsection Examples
20621 Cross fade from one input video to another input video, with fade transition and duration of transition
20622 of 2 seconds starting at offset of 5 seconds:
20624 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
20629 Pick median pixels from several input videos.
20631 The filter accepts the following options:
20635 Set number of inputs.
20636 Default is 3. Allowed range is from 3 to 255.
20637 If number of inputs is even number, than result will be mean value between two median values.
20640 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
20643 Set median percentile. Default value is @code{0.5}.
20644 Default value of @code{0.5} will pick always median values, while @code{0} will pick
20645 minimum values, and @code{1} maximum values.
20649 Stack video inputs into custom layout.
20651 All streams must be of same pixel format.
20653 The filter accepts the following options:
20657 Set number of input streams. Default is 2.
20660 Specify layout of inputs.
20661 This option requires the desired layout configuration to be explicitly set by the user.
20662 This sets position of each video input in output. Each input
20663 is separated by '|'.
20664 The first number represents the column, and the second number represents the row.
20665 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
20666 where X is video input from which to take width or height.
20667 Multiple values can be used when separated by '+'. In such
20668 case values are summed together.
20670 Note that if inputs are of different sizes gaps may appear, as not all of
20671 the output video frame will be filled. Similarly, videos can overlap each
20672 other if their position doesn't leave enough space for the full frame of
20675 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
20676 a layout must be set by the user.
20679 If set to 1, force the output to terminate when the shortest input
20680 terminates. Default value is 0.
20683 If set to valid color, all unused pixels will be filled with that color.
20684 By default fill is set to none, so it is disabled.
20687 @subsection Examples
20691 Display 4 inputs into 2x2 grid.
20695 input1(0, 0) | input3(w0, 0)
20696 input2(0, h0) | input4(w0, h0)
20700 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
20703 Note that if inputs are of different sizes, gaps or overlaps may occur.
20706 Display 4 inputs into 1x4 grid.
20713 input4(0, h0+h1+h2)
20717 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
20720 Note that if inputs are of different widths, unused space will appear.
20723 Display 9 inputs into 3x3 grid.
20727 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
20728 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
20729 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
20733 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
20736 Note that if inputs are of different sizes, gaps or overlaps may occur.
20739 Display 16 inputs into 4x4 grid.
20743 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
20744 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
20745 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
20746 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
20750 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|
20751 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
20754 Note that if inputs are of different sizes, gaps or overlaps may occur.
20761 Deinterlace the input video ("yadif" means "yet another deinterlacing
20764 It accepts the following parameters:
20770 The interlacing mode to adopt. It accepts one of the following values:
20773 @item 0, send_frame
20774 Output one frame for each frame.
20775 @item 1, send_field
20776 Output one frame for each field.
20777 @item 2, send_frame_nospatial
20778 Like @code{send_frame}, but it skips the spatial interlacing check.
20779 @item 3, send_field_nospatial
20780 Like @code{send_field}, but it skips the spatial interlacing check.
20783 The default value is @code{send_frame}.
20786 The picture field parity assumed for the input interlaced video. It accepts one
20787 of the following values:
20791 Assume the top field is first.
20793 Assume the bottom field is first.
20795 Enable automatic detection of field parity.
20798 The default value is @code{auto}.
20799 If the interlacing is unknown or the decoder does not export this information,
20800 top field first will be assumed.
20803 Specify which frames to deinterlace. Accepts one of the following
20808 Deinterlace all frames.
20809 @item 1, interlaced
20810 Only deinterlace frames marked as interlaced.
20813 The default value is @code{all}.
20816 @section yadif_cuda
20818 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
20819 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
20822 It accepts the following parameters:
20828 The interlacing mode to adopt. It accepts one of the following values:
20831 @item 0, send_frame
20832 Output one frame for each frame.
20833 @item 1, send_field
20834 Output one frame for each field.
20835 @item 2, send_frame_nospatial
20836 Like @code{send_frame}, but it skips the spatial interlacing check.
20837 @item 3, send_field_nospatial
20838 Like @code{send_field}, but it skips the spatial interlacing check.
20841 The default value is @code{send_frame}.
20844 The picture field parity assumed for the input interlaced video. It accepts one
20845 of the following values:
20849 Assume the top field is first.
20851 Assume the bottom field is first.
20853 Enable automatic detection of field parity.
20856 The default value is @code{auto}.
20857 If the interlacing is unknown or the decoder does not export this information,
20858 top field first will be assumed.
20861 Specify which frames to deinterlace. Accepts one of the following
20866 Deinterlace all frames.
20867 @item 1, interlaced
20868 Only deinterlace frames marked as interlaced.
20871 The default value is @code{all}.
20876 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
20877 The algorithm is described in
20878 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
20880 It accepts the following parameters:
20884 Set the window radius. Default value is 3.
20887 Set which planes to filter. Default is only the first plane.
20890 Set blur strength. Default value is 128.
20893 @subsection Commands
20894 This filter supports same @ref{commands} as options.
20898 Apply Zoom & Pan effect.
20900 This filter accepts the following options:
20904 Set the zoom expression. Range is 1-10. Default is 1.
20908 Set the x and y expression. Default is 0.
20911 Set the duration expression in number of frames.
20912 This sets for how many number of frames effect will last for
20913 single input image.
20916 Set the output image size, default is 'hd720'.
20919 Set the output frame rate, default is '25'.
20922 Each expression can contain the following constants:
20941 Output frame count.
20944 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
20946 @item out_time, time, ot
20947 The output timestamp expressed in seconds.
20951 Last calculated 'x' and 'y' position from 'x' and 'y' expression
20952 for current input frame.
20956 'x' and 'y' of last output frame of previous input frame or 0 when there was
20957 not yet such frame (first input frame).
20960 Last calculated zoom from 'z' expression for current input frame.
20963 Last calculated zoom of last output frame of previous input frame.
20966 Number of output frames for current input frame. Calculated from 'd' expression
20967 for each input frame.
20970 number of output frames created for previous input frame
20973 Rational number: input width / input height
20976 sample aspect ratio
20979 display aspect ratio
20983 @subsection Examples
20987 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
20989 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
20993 Zoom in up to 1.5x and pan always at center of picture:
20995 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
20999 Same as above but without pausing:
21001 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21005 Zoom in 2x into center of picture only for the first second of the input video:
21007 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21014 Scale (resize) the input video, using the z.lib library:
21015 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21016 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21018 The zscale filter forces the output display aspect ratio to be the same
21019 as the input, by changing the output sample aspect ratio.
21021 If the input image format is different from the format requested by
21022 the next filter, the zscale filter will convert the input to the
21025 @subsection Options
21026 The filter accepts the following options.
21031 Set the output video dimension expression. Default value is the input
21034 If the @var{width} or @var{w} value is 0, the input width is used for
21035 the output. If the @var{height} or @var{h} value is 0, the input height
21036 is used for the output.
21038 If one and only one of the values is -n with n >= 1, the zscale filter
21039 will use a value that maintains the aspect ratio of the input image,
21040 calculated from the other specified dimension. After that it will,
21041 however, make sure that the calculated dimension is divisible by n and
21042 adjust the value if necessary.
21044 If both values are -n with n >= 1, the behavior will be identical to
21045 both values being set to 0 as previously detailed.
21047 See below for the list of accepted constants for use in the dimension
21051 Set the video size. For the syntax of this option, check the
21052 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21055 Set the dither type.
21057 Possible values are:
21062 @item error_diffusion
21068 Set the resize filter type.
21070 Possible values are:
21080 Default is bilinear.
21083 Set the color range.
21085 Possible values are:
21092 Default is same as input.
21095 Set the color primaries.
21097 Possible values are:
21107 Default is same as input.
21110 Set the transfer characteristics.
21112 Possible values are:
21126 Default is same as input.
21129 Set the colorspace matrix.
21131 Possible value are:
21142 Default is same as input.
21145 Set the input color range.
21147 Possible values are:
21154 Default is same as input.
21156 @item primariesin, pin
21157 Set the input color primaries.
21159 Possible values are:
21169 Default is same as input.
21171 @item transferin, tin
21172 Set the input transfer characteristics.
21174 Possible values are:
21185 Default is same as input.
21187 @item matrixin, min
21188 Set the input colorspace matrix.
21190 Possible value are:
21202 Set the output chroma location.
21204 Possible values are:
21215 @item chromalin, cin
21216 Set the input chroma location.
21218 Possible values are:
21230 Set the nominal peak luminance.
21233 The values of the @option{w} and @option{h} options are expressions
21234 containing the following constants:
21239 The input width and height
21243 These are the same as @var{in_w} and @var{in_h}.
21247 The output (scaled) width and height
21251 These are the same as @var{out_w} and @var{out_h}
21254 The same as @var{iw} / @var{ih}
21257 input sample aspect ratio
21260 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
21264 horizontal and vertical input chroma subsample values. For example for the
21265 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21269 horizontal and vertical output chroma subsample values. For example for the
21270 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21273 @subsection Commands
21275 This filter supports the following commands:
21279 Set the output video dimension expression.
21280 The command accepts the same syntax of the corresponding option.
21282 If the specified expression is not valid, it is kept at its current
21286 @c man end VIDEO FILTERS
21288 @chapter OpenCL Video Filters
21289 @c man begin OPENCL VIDEO FILTERS
21291 Below is a description of the currently available OpenCL video filters.
21293 To enable compilation of these filters you need to configure FFmpeg with
21294 @code{--enable-opencl}.
21296 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
21299 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
21300 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
21301 given device parameters.
21303 @item -filter_hw_device @var{name}
21304 Pass the hardware device called @var{name} to all filters in any filter graph.
21308 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
21312 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
21314 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
21318 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.
21320 @section avgblur_opencl
21322 Apply average blur filter.
21324 The filter accepts the following options:
21328 Set horizontal radius size.
21329 Range is @code{[1, 1024]} and default value is @code{1}.
21332 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21335 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
21338 @subsection Example
21342 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.
21344 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
21348 @section boxblur_opencl
21350 Apply a boxblur algorithm to the input video.
21352 It accepts the following parameters:
21356 @item luma_radius, lr
21357 @item luma_power, lp
21358 @item chroma_radius, cr
21359 @item chroma_power, cp
21360 @item alpha_radius, ar
21361 @item alpha_power, ap
21365 A description of the accepted options follows.
21368 @item luma_radius, lr
21369 @item chroma_radius, cr
21370 @item alpha_radius, ar
21371 Set an expression for the box radius in pixels used for blurring the
21372 corresponding input plane.
21374 The radius value must be a non-negative number, and must not be
21375 greater than the value of the expression @code{min(w,h)/2} for the
21376 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
21379 Default value for @option{luma_radius} is "2". If not specified,
21380 @option{chroma_radius} and @option{alpha_radius} default to the
21381 corresponding value set for @option{luma_radius}.
21383 The expressions can contain the following constants:
21387 The input width and height in pixels.
21391 The input chroma image width and height in pixels.
21395 The horizontal and vertical chroma subsample values. For example, for the
21396 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
21399 @item luma_power, lp
21400 @item chroma_power, cp
21401 @item alpha_power, ap
21402 Specify how many times the boxblur filter is applied to the
21403 corresponding plane.
21405 Default value for @option{luma_power} is 2. If not specified,
21406 @option{chroma_power} and @option{alpha_power} default to the
21407 corresponding value set for @option{luma_power}.
21409 A value of 0 will disable the effect.
21412 @subsection Examples
21414 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.
21418 Apply a boxblur filter with the luma, chroma, and alpha radius
21419 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.
21421 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
21422 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
21426 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.
21428 For the luma plane, a 2x2 box radius will be run once.
21430 For the chroma plane, a 4x4 box radius will be run 5 times.
21432 For the alpha plane, a 3x3 box radius will be run 7 times.
21434 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
21438 @section colorkey_opencl
21439 RGB colorspace color keying.
21441 The filter accepts the following options:
21445 The color which will be replaced with transparency.
21448 Similarity percentage with the key color.
21450 0.01 matches only the exact key color, while 1.0 matches everything.
21455 0.0 makes pixels either fully transparent, or not transparent at all.
21457 Higher values result in semi-transparent pixels, with a higher transparency
21458 the more similar the pixels color is to the key color.
21461 @subsection Examples
21465 Make every semi-green pixel in the input transparent with some slight blending:
21467 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
21471 @section convolution_opencl
21473 Apply convolution of 3x3, 5x5, 7x7 matrix.
21475 The filter accepts the following options:
21482 Set matrix for each plane.
21483 Matrix is sequence of 9, 25 or 49 signed numbers.
21484 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
21490 Set multiplier for calculated value for each plane.
21491 If unset or 0, it will be sum of all matrix elements.
21492 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
21498 Set bias for each plane. This value is added to the result of the multiplication.
21499 Useful for making the overall image brighter or darker.
21500 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
21504 @subsection Examples
21510 -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
21516 -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
21520 Apply edge enhance:
21522 -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
21528 -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
21532 Apply laplacian edge detector which includes diagonals:
21534 -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
21540 -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
21544 @section erosion_opencl
21546 Apply erosion effect to the video.
21548 This filter replaces the pixel by the local(3x3) minimum.
21550 It accepts the following options:
21557 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21558 If @code{0}, plane will remain unchanged.
21561 Flag which specifies the pixel to refer to.
21562 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21564 Flags to local 3x3 coordinates region centered on @code{x}:
21573 @subsection Example
21577 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.
21579 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21583 @section deshake_opencl
21584 Feature-point based video stabilization filter.
21586 The filter accepts the following options:
21590 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
21593 Whether or not additional debug info should be displayed, both in the processed output and in the console.
21595 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
21597 Viewing point matches in the output video is only supported for RGB input.
21599 Defaults to @code{0}.
21601 @item adaptive_crop
21602 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
21604 Defaults to @code{1}.
21606 @item refine_features
21607 Whether or not feature points should be refined at a sub-pixel level.
21609 This can be turned off for a slight performance gain at the cost of precision.
21611 Defaults to @code{1}.
21613 @item smooth_strength
21614 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
21616 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
21618 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
21620 Defaults to @code{0.0}.
21622 @item smooth_window_multiplier
21623 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
21625 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
21627 Acceptable values range from @code{0.1} to @code{10.0}.
21629 Larger values increase the amount of motion data available for determining how to smooth the camera path,
21630 potentially improving smoothness, but also increase latency and memory usage.
21632 Defaults to @code{2.0}.
21636 @subsection Examples
21640 Stabilize a video with a fixed, medium smoothing strength:
21642 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
21646 Stabilize a video with debugging (both in console and in rendered video):
21648 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
21652 @section dilation_opencl
21654 Apply dilation effect to the video.
21656 This filter replaces the pixel by the local(3x3) maximum.
21658 It accepts the following options:
21665 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21666 If @code{0}, plane will remain unchanged.
21669 Flag which specifies the pixel to refer to.
21670 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21672 Flags to local 3x3 coordinates region centered on @code{x}:
21681 @subsection Example
21685 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.
21687 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21691 @section nlmeans_opencl
21693 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
21695 @section overlay_opencl
21697 Overlay one video on top of another.
21699 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
21700 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
21702 The filter accepts the following options:
21707 Set the x coordinate of the overlaid video on the main video.
21708 Default value is @code{0}.
21711 Set the y coordinate of the overlaid video on the main video.
21712 Default value is @code{0}.
21716 @subsection Examples
21720 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
21722 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21725 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
21727 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21732 @section pad_opencl
21734 Add paddings to the input image, and place the original input at the
21735 provided @var{x}, @var{y} coordinates.
21737 It accepts the following options:
21742 Specify an expression for the size of the output image with the
21743 paddings added. If the value for @var{width} or @var{height} is 0, the
21744 corresponding input size is used for the output.
21746 The @var{width} expression can reference the value set by the
21747 @var{height} expression, and vice versa.
21749 The default value of @var{width} and @var{height} is 0.
21753 Specify the offsets to place the input image at within the padded area,
21754 with respect to the top/left border of the output image.
21756 The @var{x} expression can reference the value set by the @var{y}
21757 expression, and vice versa.
21759 The default value of @var{x} and @var{y} is 0.
21761 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
21762 so the input image is centered on the padded area.
21765 Specify the color of the padded area. For the syntax of this option,
21766 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
21767 manual,ffmpeg-utils}.
21770 Pad to an aspect instead to a resolution.
21773 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
21774 options are expressions containing the following constants:
21779 The input video width and height.
21783 These are the same as @var{in_w} and @var{in_h}.
21787 The output width and height (the size of the padded area), as
21788 specified by the @var{width} and @var{height} expressions.
21792 These are the same as @var{out_w} and @var{out_h}.
21796 The x and y offsets as specified by the @var{x} and @var{y}
21797 expressions, or NAN if not yet specified.
21800 same as @var{iw} / @var{ih}
21803 input sample aspect ratio
21806 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
21809 @section prewitt_opencl
21811 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
21813 The filter accepts the following option:
21817 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21820 Set value which will be multiplied with filtered result.
21821 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21824 Set value which will be added to filtered result.
21825 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21828 @subsection Example
21832 Apply the Prewitt operator with scale set to 2 and delta set to 10.
21834 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
21838 @anchor{program_opencl}
21839 @section program_opencl
21841 Filter video using an OpenCL program.
21846 OpenCL program source file.
21849 Kernel name in program.
21852 Number of inputs to the filter. Defaults to 1.
21855 Size of output frames. Defaults to the same as the first input.
21859 The @code{program_opencl} filter also supports the @ref{framesync} options.
21861 The program source file must contain a kernel function with the given name,
21862 which will be run once for each plane of the output. Each run on a plane
21863 gets enqueued as a separate 2D global NDRange with one work-item for each
21864 pixel to be generated. The global ID offset for each work-item is therefore
21865 the coordinates of a pixel in the destination image.
21867 The kernel function needs to take the following arguments:
21870 Destination image, @var{__write_only image2d_t}.
21872 This image will become the output; the kernel should write all of it.
21874 Frame index, @var{unsigned int}.
21876 This is a counter starting from zero and increasing by one for each frame.
21878 Source images, @var{__read_only image2d_t}.
21880 These are the most recent images on each input. The kernel may read from
21881 them to generate the output, but they can't be written to.
21888 Copy the input to the output (output must be the same size as the input).
21890 __kernel void copy(__write_only image2d_t destination,
21891 unsigned int index,
21892 __read_only image2d_t source)
21894 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
21896 int2 location = (int2)(get_global_id(0), get_global_id(1));
21898 float4 value = read_imagef(source, sampler, location);
21900 write_imagef(destination, location, value);
21905 Apply a simple transformation, rotating the input by an amount increasing
21906 with the index counter. Pixel values are linearly interpolated by the
21907 sampler, and the output need not have the same dimensions as the input.
21909 __kernel void rotate_image(__write_only image2d_t dst,
21910 unsigned int index,
21911 __read_only image2d_t src)
21913 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
21914 CLK_FILTER_LINEAR);
21916 float angle = (float)index / 100.0f;
21918 float2 dst_dim = convert_float2(get_image_dim(dst));
21919 float2 src_dim = convert_float2(get_image_dim(src));
21921 float2 dst_cen = dst_dim / 2.0f;
21922 float2 src_cen = src_dim / 2.0f;
21924 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
21926 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
21928 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
21929 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
21931 src_pos = src_pos * src_dim / dst_dim;
21933 float2 src_loc = src_pos + src_cen;
21935 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
21936 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
21937 write_imagef(dst, dst_loc, 0.5f);
21939 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
21944 Blend two inputs together, with the amount of each input used varying
21945 with the index counter.
21947 __kernel void blend_images(__write_only image2d_t dst,
21948 unsigned int index,
21949 __read_only image2d_t src1,
21950 __read_only image2d_t src2)
21952 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
21953 CLK_FILTER_LINEAR);
21955 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
21957 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
21958 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
21959 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
21961 float4 val1 = read_imagef(src1, sampler, src1_loc);
21962 float4 val2 = read_imagef(src2, sampler, src2_loc);
21964 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
21970 @section roberts_opencl
21971 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
21973 The filter accepts the following option:
21977 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21980 Set value which will be multiplied with filtered result.
21981 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21984 Set value which will be added to filtered result.
21985 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21988 @subsection Example
21992 Apply the Roberts cross operator with scale set to 2 and delta set to 10
21994 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
21998 @section sobel_opencl
22000 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22002 The filter accepts the following option:
22006 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22009 Set value which will be multiplied with filtered result.
22010 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22013 Set value which will be added to filtered result.
22014 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22017 @subsection Example
22021 Apply sobel operator with scale set to 2 and delta set to 10
22023 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22027 @section tonemap_opencl
22029 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22031 It accepts the following parameters:
22035 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22038 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22041 Apply desaturation for highlights that exceed this level of brightness. The
22042 higher the parameter, the more color information will be preserved. This
22043 setting helps prevent unnaturally blown-out colors for super-highlights, by
22044 (smoothly) turning into white instead. This makes images feel more natural,
22045 at the cost of reducing information about out-of-range colors.
22047 The default value is 0.5, and the algorithm here is a little different from
22048 the cpu version tonemap currently. A setting of 0.0 disables this option.
22051 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22052 is used to detect whether the scene has changed or not. If the distance between
22053 the current frame average brightness and the current running average exceeds
22054 a threshold value, we would re-calculate scene average and peak brightness.
22055 The default value is 0.2.
22058 Specify the output pixel format.
22060 Currently supported formats are:
22067 Set the output color range.
22069 Possible values are:
22075 Default is same as input.
22078 Set the output color primaries.
22080 Possible values are:
22086 Default is same as input.
22089 Set the output transfer characteristics.
22091 Possible values are:
22100 Set the output colorspace matrix.
22102 Possible value are:
22108 Default is same as input.
22112 @subsection Example
22116 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22118 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22122 @section unsharp_opencl
22124 Sharpen or blur the input video.
22126 It accepts the following parameters:
22129 @item luma_msize_x, lx
22130 Set the luma matrix horizontal size.
22131 Range is @code{[1, 23]} and default value is @code{5}.
22133 @item luma_msize_y, ly
22134 Set the luma matrix vertical size.
22135 Range is @code{[1, 23]} and default value is @code{5}.
22137 @item luma_amount, la
22138 Set the luma effect strength.
22139 Range is @code{[-10, 10]} and default value is @code{1.0}.
22141 Negative values will blur the input video, while positive values will
22142 sharpen it, a value of zero will disable the effect.
22144 @item chroma_msize_x, cx
22145 Set the chroma matrix horizontal size.
22146 Range is @code{[1, 23]} and default value is @code{5}.
22148 @item chroma_msize_y, cy
22149 Set the chroma matrix vertical size.
22150 Range is @code{[1, 23]} and default value is @code{5}.
22152 @item chroma_amount, ca
22153 Set the chroma effect strength.
22154 Range is @code{[-10, 10]} and default value is @code{0.0}.
22156 Negative values will blur the input video, while positive values will
22157 sharpen it, a value of zero will disable the effect.
22161 All parameters are optional and default to the equivalent of the
22162 string '5:5:1.0:5:5:0.0'.
22164 @subsection Examples
22168 Apply strong luma sharpen effect:
22170 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22174 Apply a strong blur of both luma and chroma parameters:
22176 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22180 @section xfade_opencl
22182 Cross fade two videos with custom transition effect by using OpenCL.
22184 It accepts the following options:
22188 Set one of possible transition effects.
22192 Select custom transition effect, the actual transition description
22193 will be picked from source and kernel options.
22205 Default transition is fade.
22209 OpenCL program source file for custom transition.
22212 Set name of kernel to use for custom transition from program source file.
22215 Set duration of video transition.
22218 Set time of start of transition relative to first video.
22221 The program source file must contain a kernel function with the given name,
22222 which will be run once for each plane of the output. Each run on a plane
22223 gets enqueued as a separate 2D global NDRange with one work-item for each
22224 pixel to be generated. The global ID offset for each work-item is therefore
22225 the coordinates of a pixel in the destination image.
22227 The kernel function needs to take the following arguments:
22230 Destination image, @var{__write_only image2d_t}.
22232 This image will become the output; the kernel should write all of it.
22235 First Source image, @var{__read_only image2d_t}.
22236 Second Source image, @var{__read_only image2d_t}.
22238 These are the most recent images on each input. The kernel may read from
22239 them to generate the output, but they can't be written to.
22242 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22249 Apply dots curtain transition effect:
22251 __kernel void blend_images(__write_only image2d_t dst,
22252 __read_only image2d_t src1,
22253 __read_only image2d_t src2,
22256 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22257 CLK_FILTER_LINEAR);
22258 int2 p = (int2)(get_global_id(0), get_global_id(1));
22259 float2 rp = (float2)(get_global_id(0), get_global_id(1));
22260 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
22263 float2 dots = (float2)(20.0, 20.0);
22264 float2 center = (float2)(0,0);
22267 float4 val1 = read_imagef(src1, sampler, p);
22268 float4 val2 = read_imagef(src2, sampler, p);
22269 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
22271 write_imagef(dst, p, next ? val1 : val2);
22277 @c man end OPENCL VIDEO FILTERS
22279 @chapter VAAPI Video Filters
22280 @c man begin VAAPI VIDEO FILTERS
22282 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
22284 To enable compilation of these filters you need to configure FFmpeg with
22285 @code{--enable-vaapi}.
22287 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}
22289 @section tonemap_vaapi
22291 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
22292 It maps the dynamic range of HDR10 content to the SDR content.
22293 It currently only accepts HDR10 as input.
22295 It accepts the following parameters:
22299 Specify the output pixel format.
22301 Currently supported formats are:
22310 Set the output color primaries.
22312 Default is same as input.
22315 Set the output transfer characteristics.
22320 Set the output colorspace matrix.
22322 Default is same as input.
22326 @subsection Example
22330 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
22332 tonemap_vaapi=format=p010:t=bt2020-10
22336 @c man end VAAPI VIDEO FILTERS
22338 @chapter Video Sources
22339 @c man begin VIDEO SOURCES
22341 Below is a description of the currently available video sources.
22345 Buffer video frames, and make them available to the filter chain.
22347 This source is mainly intended for a programmatic use, in particular
22348 through the interface defined in @file{libavfilter/buffersrc.h}.
22350 It accepts the following parameters:
22355 Specify the size (width and height) of the buffered video frames. For the
22356 syntax of this option, check the
22357 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22360 The input video width.
22363 The input video height.
22366 A string representing the pixel format of the buffered video frames.
22367 It may be a number corresponding to a pixel format, or a pixel format
22371 Specify the timebase assumed by the timestamps of the buffered frames.
22374 Specify the frame rate expected for the video stream.
22376 @item pixel_aspect, sar
22377 The sample (pixel) aspect ratio of the input video.
22380 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
22381 to the filtergraph description to specify swscale flags for automatically
22382 inserted scalers. See @ref{Filtergraph syntax}.
22384 @item hw_frames_ctx
22385 When using a hardware pixel format, this should be a reference to an
22386 AVHWFramesContext describing input frames.
22391 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
22394 will instruct the source to accept video frames with size 320x240 and
22395 with format "yuv410p", assuming 1/24 as the timestamps timebase and
22396 square pixels (1:1 sample aspect ratio).
22397 Since the pixel format with name "yuv410p" corresponds to the number 6
22398 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
22399 this example corresponds to:
22401 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
22404 Alternatively, the options can be specified as a flat string, but this
22405 syntax is deprecated:
22407 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
22411 Create a pattern generated by an elementary cellular automaton.
22413 The initial state of the cellular automaton can be defined through the
22414 @option{filename} and @option{pattern} options. If such options are
22415 not specified an initial state is created randomly.
22417 At each new frame a new row in the video is filled with the result of
22418 the cellular automaton next generation. The behavior when the whole
22419 frame is filled is defined by the @option{scroll} option.
22421 This source accepts the following options:
22425 Read the initial cellular automaton state, i.e. the starting row, from
22426 the specified file.
22427 In the file, each non-whitespace character is considered an alive
22428 cell, a newline will terminate the row, and further characters in the
22429 file will be ignored.
22432 Read the initial cellular automaton state, i.e. the starting row, from
22433 the specified string.
22435 Each non-whitespace character in the string is considered an alive
22436 cell, a newline will terminate the row, and further characters in the
22437 string will be ignored.
22440 Set the video rate, that is the number of frames generated per second.
22443 @item random_fill_ratio, ratio
22444 Set the random fill ratio for the initial cellular automaton row. It
22445 is a floating point number value ranging from 0 to 1, defaults to
22448 This option is ignored when a file or a pattern is specified.
22450 @item random_seed, seed
22451 Set the seed for filling randomly the initial row, must be an integer
22452 included between 0 and UINT32_MAX. If not specified, or if explicitly
22453 set to -1, the filter will try to use a good random seed on a best
22457 Set the cellular automaton rule, it is a number ranging from 0 to 255.
22458 Default value is 110.
22461 Set the size of the output video. For the syntax of this option, check the
22462 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22464 If @option{filename} or @option{pattern} is specified, the size is set
22465 by default to the width of the specified initial state row, and the
22466 height is set to @var{width} * PHI.
22468 If @option{size} is set, it must contain the width of the specified
22469 pattern string, and the specified pattern will be centered in the
22472 If a filename or a pattern string is not specified, the size value
22473 defaults to "320x518" (used for a randomly generated initial state).
22476 If set to 1, scroll the output upward when all the rows in the output
22477 have been already filled. If set to 0, the new generated row will be
22478 written over the top row just after the bottom row is filled.
22481 @item start_full, full
22482 If set to 1, completely fill the output with generated rows before
22483 outputting the first frame.
22484 This is the default behavior, for disabling set the value to 0.
22487 If set to 1, stitch the left and right row edges together.
22488 This is the default behavior, for disabling set the value to 0.
22491 @subsection Examples
22495 Read the initial state from @file{pattern}, and specify an output of
22498 cellauto=f=pattern:s=200x400
22502 Generate a random initial row with a width of 200 cells, with a fill
22505 cellauto=ratio=2/3:s=200x200
22509 Create a pattern generated by rule 18 starting by a single alive cell
22510 centered on an initial row with width 100:
22512 cellauto=p=@@:s=100x400:full=0:rule=18
22516 Specify a more elaborated initial pattern:
22518 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
22523 @anchor{coreimagesrc}
22524 @section coreimagesrc
22525 Video source generated on GPU using Apple's CoreImage API on OSX.
22527 This video source is a specialized version of the @ref{coreimage} video filter.
22528 Use a core image generator at the beginning of the applied filterchain to
22529 generate the content.
22531 The coreimagesrc video source accepts the following options:
22533 @item list_generators
22534 List all available generators along with all their respective options as well as
22535 possible minimum and maximum values along with the default values.
22537 list_generators=true
22541 Specify the size of the sourced video. For the syntax of this option, check the
22542 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22543 The default value is @code{320x240}.
22546 Specify the frame rate of the sourced video, as the number of frames
22547 generated per second. It has to be a string in the format
22548 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22549 number or a valid video frame rate abbreviation. The default value is
22553 Set the sample aspect ratio of the sourced video.
22556 Set the duration of the sourced video. See
22557 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22558 for the accepted syntax.
22560 If not specified, or the expressed duration is negative, the video is
22561 supposed to be generated forever.
22564 Additionally, all options of the @ref{coreimage} video filter are accepted.
22565 A complete filterchain can be used for further processing of the
22566 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
22567 and examples for details.
22569 @subsection Examples
22574 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
22575 given as complete and escaped command-line for Apple's standard bash shell:
22577 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
22579 This example is equivalent to the QRCode example of @ref{coreimage} without the
22580 need for a nullsrc video source.
22585 Generate several gradients.
22589 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22590 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22593 Set frame rate, expressed as number of frames per second. Default
22596 @item c0, c1, c2, c3, c4, c5, c6, c7
22597 Set 8 colors. Default values for colors is to pick random one.
22599 @item x0, y0, y0, y1
22600 Set gradient line source and destination points. If negative or out of range, random ones
22604 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
22607 Set seed for picking gradient line points.
22610 Set the duration of the sourced video. See
22611 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22612 for the accepted syntax.
22614 If not specified, or the expressed duration is negative, the video is
22615 supposed to be generated forever.
22618 Set speed of gradients rotation.
22622 @section mandelbrot
22624 Generate a Mandelbrot set fractal, and progressively zoom towards the
22625 point specified with @var{start_x} and @var{start_y}.
22627 This source accepts the following options:
22632 Set the terminal pts value. Default value is 400.
22635 Set the terminal scale value.
22636 Must be a floating point value. Default value is 0.3.
22639 Set the inner coloring mode, that is the algorithm used to draw the
22640 Mandelbrot fractal internal region.
22642 It shall assume one of the following values:
22647 Show time until convergence.
22649 Set color based on point closest to the origin of the iterations.
22654 Default value is @var{mincol}.
22657 Set the bailout value. Default value is 10.0.
22660 Set the maximum of iterations performed by the rendering
22661 algorithm. Default value is 7189.
22664 Set outer coloring mode.
22665 It shall assume one of following values:
22667 @item iteration_count
22668 Set iteration count mode.
22669 @item normalized_iteration_count
22670 set normalized iteration count mode.
22672 Default value is @var{normalized_iteration_count}.
22675 Set frame rate, expressed as number of frames per second. Default
22679 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22680 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22683 Set the initial scale value. Default value is 3.0.
22686 Set the initial x position. Must be a floating point value between
22687 -100 and 100. Default value is -0.743643887037158704752191506114774.
22690 Set the initial y position. Must be a floating point value between
22691 -100 and 100. Default value is -0.131825904205311970493132056385139.
22696 Generate various test patterns, as generated by the MPlayer test filter.
22698 The size of the generated video is fixed, and is 256x256.
22699 This source is useful in particular for testing encoding features.
22701 This source accepts the following options:
22706 Specify the frame rate of the sourced video, as the number of frames
22707 generated per second. It has to be a string in the format
22708 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22709 number or a valid video frame rate abbreviation. The default value is
22713 Set the duration of the sourced video. See
22714 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22715 for the accepted syntax.
22717 If not specified, or the expressed duration is negative, the video is
22718 supposed to be generated forever.
22722 Set the number or the name of the test to perform. Supported tests are:
22736 @item max_frames, m
22737 Set the maximum number of frames generated for each test, default value is 30.
22741 Default value is "all", which will cycle through the list of all tests.
22746 mptestsrc=t=dc_luma
22749 will generate a "dc_luma" test pattern.
22751 @section frei0r_src
22753 Provide a frei0r source.
22755 To enable compilation of this filter you need to install the frei0r
22756 header and configure FFmpeg with @code{--enable-frei0r}.
22758 This source accepts the following parameters:
22763 The size of the video to generate. For the syntax of this option, check the
22764 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22767 The framerate of the generated video. It may be a string of the form
22768 @var{num}/@var{den} or a frame rate abbreviation.
22771 The name to the frei0r source to load. For more information regarding frei0r and
22772 how to set the parameters, read the @ref{frei0r} section in the video filters
22775 @item filter_params
22776 A '|'-separated list of parameters to pass to the frei0r source.
22780 For example, to generate a frei0r partik0l source with size 200x200
22781 and frame rate 10 which is overlaid on the overlay filter main input:
22783 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
22788 Generate a life pattern.
22790 This source is based on a generalization of John Conway's life game.
22792 The sourced input represents a life grid, each pixel represents a cell
22793 which can be in one of two possible states, alive or dead. Every cell
22794 interacts with its eight neighbours, which are the cells that are
22795 horizontally, vertically, or diagonally adjacent.
22797 At each interaction the grid evolves according to the adopted rule,
22798 which specifies the number of neighbor alive cells which will make a
22799 cell stay alive or born. The @option{rule} option allows one to specify
22802 This source accepts the following options:
22806 Set the file from which to read the initial grid state. In the file,
22807 each non-whitespace character is considered an alive cell, and newline
22808 is used to delimit the end of each row.
22810 If this option is not specified, the initial grid is generated
22814 Set the video rate, that is the number of frames generated per second.
22817 @item random_fill_ratio, ratio
22818 Set the random fill ratio for the initial random grid. It is a
22819 floating point number value ranging from 0 to 1, defaults to 1/PHI.
22820 It is ignored when a file is specified.
22822 @item random_seed, seed
22823 Set the seed for filling the initial random grid, must be an integer
22824 included between 0 and UINT32_MAX. If not specified, or if explicitly
22825 set to -1, the filter will try to use a good random seed on a best
22831 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
22832 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
22833 @var{NS} specifies the number of alive neighbor cells which make a
22834 live cell stay alive, and @var{NB} the number of alive neighbor cells
22835 which make a dead cell to become alive (i.e. to "born").
22836 "s" and "b" can be used in place of "S" and "B", respectively.
22838 Alternatively a rule can be specified by an 18-bits integer. The 9
22839 high order bits are used to encode the next cell state if it is alive
22840 for each number of neighbor alive cells, the low order bits specify
22841 the rule for "borning" new cells. Higher order bits encode for an
22842 higher number of neighbor cells.
22843 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
22844 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
22846 Default value is "S23/B3", which is the original Conway's game of life
22847 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
22848 cells, and will born a new cell if there are three alive cells around
22852 Set the size of the output video. For the syntax of this option, check the
22853 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22855 If @option{filename} is specified, the size is set by default to the
22856 same size of the input file. If @option{size} is set, it must contain
22857 the size specified in the input file, and the initial grid defined in
22858 that file is centered in the larger resulting area.
22860 If a filename is not specified, the size value defaults to "320x240"
22861 (used for a randomly generated initial grid).
22864 If set to 1, stitch the left and right grid edges together, and the
22865 top and bottom edges also. Defaults to 1.
22868 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
22869 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
22870 value from 0 to 255.
22873 Set the color of living (or new born) cells.
22876 Set the color of dead cells. If @option{mold} is set, this is the first color
22877 used to represent a dead cell.
22880 Set mold color, for definitely dead and moldy cells.
22882 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
22883 ffmpeg-utils manual,ffmpeg-utils}.
22886 @subsection Examples
22890 Read a grid from @file{pattern}, and center it on a grid of size
22893 life=f=pattern:s=300x300
22897 Generate a random grid of size 200x200, with a fill ratio of 2/3:
22899 life=ratio=2/3:s=200x200
22903 Specify a custom rule for evolving a randomly generated grid:
22909 Full example with slow death effect (mold) using @command{ffplay}:
22911 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
22918 @anchor{haldclutsrc}
22921 @anchor{pal100bars}
22922 @anchor{rgbtestsrc}
22924 @anchor{smptehdbars}
22927 @anchor{yuvtestsrc}
22928 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
22930 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
22932 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
22934 The @code{color} source provides an uniformly colored input.
22936 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
22937 @ref{haldclut} filter.
22939 The @code{nullsrc} source returns unprocessed video frames. It is
22940 mainly useful to be employed in analysis / debugging tools, or as the
22941 source for filters which ignore the input data.
22943 The @code{pal75bars} source generates a color bars pattern, based on
22944 EBU PAL recommendations with 75% color levels.
22946 The @code{pal100bars} source generates a color bars pattern, based on
22947 EBU PAL recommendations with 100% color levels.
22949 The @code{rgbtestsrc} source generates an RGB test pattern useful for
22950 detecting RGB vs BGR issues. You should see a red, green and blue
22951 stripe from top to bottom.
22953 The @code{smptebars} source generates a color bars pattern, based on
22954 the SMPTE Engineering Guideline EG 1-1990.
22956 The @code{smptehdbars} source generates a color bars pattern, based on
22957 the SMPTE RP 219-2002.
22959 The @code{testsrc} source generates a test video pattern, showing a
22960 color pattern, a scrolling gradient and a timestamp. This is mainly
22961 intended for testing purposes.
22963 The @code{testsrc2} source is similar to testsrc, but supports more
22964 pixel formats instead of just @code{rgb24}. This allows using it as an
22965 input for other tests without requiring a format conversion.
22967 The @code{yuvtestsrc} source generates an YUV test pattern. You should
22968 see a y, cb and cr stripe from top to bottom.
22970 The sources accept the following parameters:
22975 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
22976 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
22977 pixels to be used as identity matrix for 3D lookup tables. Each component is
22978 coded on a @code{1/(N*N)} scale.
22981 Specify the color of the source, only available in the @code{color}
22982 source. For the syntax of this option, check the
22983 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
22986 Specify the size of the sourced video. For the syntax of this option, check the
22987 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22988 The default value is @code{320x240}.
22990 This option is not available with the @code{allrgb}, @code{allyuv}, and
22991 @code{haldclutsrc} filters.
22994 Specify the frame rate of the sourced video, as the number of frames
22995 generated per second. It has to be a string in the format
22996 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22997 number or a valid video frame rate abbreviation. The default value is
23001 Set the duration of the sourced video. See
23002 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23003 for the accepted syntax.
23005 If not specified, or the expressed duration is negative, the video is
23006 supposed to be generated forever.
23008 Since the frame rate is used as time base, all frames including the last one
23009 will have their full duration. If the specified duration is not a multiple
23010 of the frame duration, it will be rounded up.
23013 Set the sample aspect ratio of the sourced video.
23016 Specify the alpha (opacity) of the background, only available in the
23017 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23018 255 (fully opaque, the default).
23021 Set the number of decimals to show in the timestamp, only available in the
23022 @code{testsrc} source.
23024 The displayed timestamp value will correspond to the original
23025 timestamp value multiplied by the power of 10 of the specified
23026 value. Default value is 0.
23029 @subsection Examples
23033 Generate a video with a duration of 5.3 seconds, with size
23034 176x144 and a frame rate of 10 frames per second:
23036 testsrc=duration=5.3:size=qcif:rate=10
23040 The following graph description will generate a red source
23041 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23044 color=c=red@@0.2:s=qcif:r=10
23048 If the input content is to be ignored, @code{nullsrc} can be used. The
23049 following command generates noise in the luminance plane by employing
23050 the @code{geq} filter:
23052 nullsrc=s=256x256, geq=random(1)*255:128:128
23056 @subsection Commands
23058 The @code{color} source supports the following commands:
23062 Set the color of the created image. Accepts the same syntax of the
23063 corresponding @option{color} option.
23068 Generate video using an OpenCL program.
23073 OpenCL program source file.
23076 Kernel name in program.
23079 Size of frames to generate. This must be set.
23082 Pixel format to use for the generated frames. This must be set.
23085 Number of frames generated every second. Default value is '25'.
23089 For details of how the program loading works, see the @ref{program_opencl}
23096 Generate a colour ramp by setting pixel values from the position of the pixel
23097 in the output image. (Note that this will work with all pixel formats, but
23098 the generated output will not be the same.)
23100 __kernel void ramp(__write_only image2d_t dst,
23101 unsigned int index)
23103 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23106 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23108 write_imagef(dst, loc, val);
23113 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23115 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23116 unsigned int index)
23118 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23120 float4 value = 0.0f;
23121 int x = loc.x + index;
23122 int y = loc.y + index;
23123 while (x > 0 || y > 0) {
23124 if (x % 3 == 1 && y % 3 == 1) {
23132 write_imagef(dst, loc, value);
23138 @section sierpinski
23140 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23142 This source accepts the following options:
23146 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23147 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23150 Set frame rate, expressed as number of frames per second. Default
23154 Set seed which is used for random panning.
23157 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23160 Set fractal type, can be default @code{carpet} or @code{triangle}.
23163 @c man end VIDEO SOURCES
23165 @chapter Video Sinks
23166 @c man begin VIDEO SINKS
23168 Below is a description of the currently available video sinks.
23170 @section buffersink
23172 Buffer video frames, and make them available to the end of the filter
23175 This sink is mainly intended for programmatic use, in particular
23176 through the interface defined in @file{libavfilter/buffersink.h}
23177 or the options system.
23179 It accepts a pointer to an AVBufferSinkContext structure, which
23180 defines the incoming buffers' formats, to be passed as the opaque
23181 parameter to @code{avfilter_init_filter} for initialization.
23185 Null video sink: do absolutely nothing with the input video. It is
23186 mainly useful as a template and for use in analysis / debugging
23189 @c man end VIDEO SINKS
23191 @chapter Multimedia Filters
23192 @c man begin MULTIMEDIA FILTERS
23194 Below is a description of the currently available multimedia filters.
23198 Convert input audio to a video output, displaying the audio bit scope.
23200 The filter accepts the following options:
23204 Set frame rate, expressed as number of frames per second. Default
23208 Specify the video size for the output. For the syntax of this option, check the
23209 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23210 Default value is @code{1024x256}.
23213 Specify list of colors separated by space or by '|' which will be used to
23214 draw channels. Unrecognized or missing colors will be replaced
23218 @section adrawgraph
23219 Draw a graph using input audio metadata.
23221 See @ref{drawgraph}
23223 @section agraphmonitor
23225 See @ref{graphmonitor}.
23227 @section ahistogram
23229 Convert input audio to a video output, displaying the volume histogram.
23231 The filter accepts the following options:
23235 Specify how histogram is calculated.
23237 It accepts the following values:
23240 Use single histogram for all channels.
23242 Use separate histogram for each channel.
23244 Default is @code{single}.
23247 Set frame rate, expressed as number of frames per second. Default
23251 Specify the video size for the output. For the syntax of this option, check the
23252 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23253 Default value is @code{hd720}.
23258 It accepts the following values:
23269 reverse logarithmic
23271 Default is @code{log}.
23274 Set amplitude scale.
23276 It accepts the following values:
23283 Default is @code{log}.
23286 Set how much frames to accumulate in histogram.
23287 Default is 1. Setting this to -1 accumulates all frames.
23290 Set histogram ratio of window height.
23293 Set sonogram sliding.
23295 It accepts the following values:
23298 replace old rows with new ones.
23300 scroll from top to bottom.
23302 Default is @code{replace}.
23305 @section aphasemeter
23307 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
23308 representing mean phase of current audio frame. A video output can also be produced and is
23309 enabled by default. The audio is passed through as first output.
23311 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
23312 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
23313 and @code{1} means channels are in phase.
23315 The filter accepts the following options, all related to its video output:
23319 Set the output frame rate. Default value is @code{25}.
23322 Set the video size for the output. For the syntax of this option, check the
23323 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23324 Default value is @code{800x400}.
23329 Specify the red, green, blue contrast. Default values are @code{2},
23330 @code{7} and @code{1}.
23331 Allowed range is @code{[0, 255]}.
23334 Set color which will be used for drawing median phase. If color is
23335 @code{none} which is default, no median phase value will be drawn.
23338 Enable video output. Default is enabled.
23341 @section avectorscope
23343 Convert input audio to a video output, representing the audio vector
23346 The filter is used to measure the difference between channels of stereo
23347 audio stream. A monaural signal, consisting of identical left and right
23348 signal, results in straight vertical line. Any stereo separation is visible
23349 as a deviation from this line, creating a Lissajous figure.
23350 If the straight (or deviation from it) but horizontal line appears this
23351 indicates that the left and right channels are out of phase.
23353 The filter accepts the following options:
23357 Set the vectorscope mode.
23359 Available values are:
23362 Lissajous rotated by 45 degrees.
23365 Same as above but not rotated.
23368 Shape resembling half of circle.
23371 Default value is @samp{lissajous}.
23374 Set the video size for the output. For the syntax of this option, check the
23375 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23376 Default value is @code{400x400}.
23379 Set the output frame rate. Default value is @code{25}.
23385 Specify the red, green, blue and alpha contrast. Default values are @code{40},
23386 @code{160}, @code{80} and @code{255}.
23387 Allowed range is @code{[0, 255]}.
23393 Specify the red, green, blue and alpha fade. Default values are @code{15},
23394 @code{10}, @code{5} and @code{5}.
23395 Allowed range is @code{[0, 255]}.
23398 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
23399 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
23402 Set the vectorscope drawing mode.
23404 Available values are:
23407 Draw dot for each sample.
23410 Draw line between previous and current sample.
23413 Default value is @samp{dot}.
23416 Specify amplitude scale of audio samples.
23418 Available values are:
23434 Swap left channel axis with right channel axis.
23444 Mirror only x axis.
23447 Mirror only y axis.
23455 @subsection Examples
23459 Complete example using @command{ffplay}:
23461 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
23462 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
23466 @section bench, abench
23468 Benchmark part of a filtergraph.
23470 The filter accepts the following options:
23474 Start or stop a timer.
23476 Available values are:
23479 Get the current time, set it as frame metadata (using the key
23480 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
23483 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
23484 the input frame metadata to get the time difference. Time difference, average,
23485 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
23486 @code{min}) are then printed. The timestamps are expressed in seconds.
23490 @subsection Examples
23494 Benchmark @ref{selectivecolor} filter:
23496 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
23502 Concatenate audio and video streams, joining them together one after the
23505 The filter works on segments of synchronized video and audio streams. All
23506 segments must have the same number of streams of each type, and that will
23507 also be the number of streams at output.
23509 The filter accepts the following options:
23514 Set the number of segments. Default is 2.
23517 Set the number of output video streams, that is also the number of video
23518 streams in each segment. Default is 1.
23521 Set the number of output audio streams, that is also the number of audio
23522 streams in each segment. Default is 0.
23525 Activate unsafe mode: do not fail if segments have a different format.
23529 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
23530 @var{a} audio outputs.
23532 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
23533 segment, in the same order as the outputs, then the inputs for the second
23536 Related streams do not always have exactly the same duration, for various
23537 reasons including codec frame size or sloppy authoring. For that reason,
23538 related synchronized streams (e.g. a video and its audio track) should be
23539 concatenated at once. The concat filter will use the duration of the longest
23540 stream in each segment (except the last one), and if necessary pad shorter
23541 audio streams with silence.
23543 For this filter to work correctly, all segments must start at timestamp 0.
23545 All corresponding streams must have the same parameters in all segments; the
23546 filtering system will automatically select a common pixel format for video
23547 streams, and a common sample format, sample rate and channel layout for
23548 audio streams, but other settings, such as resolution, must be converted
23549 explicitly by the user.
23551 Different frame rates are acceptable but will result in variable frame rate
23552 at output; be sure to configure the output file to handle it.
23554 @subsection Examples
23558 Concatenate an opening, an episode and an ending, all in bilingual version
23559 (video in stream 0, audio in streams 1 and 2):
23561 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
23562 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
23563 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
23564 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
23568 Concatenate two parts, handling audio and video separately, using the
23569 (a)movie sources, and adjusting the resolution:
23571 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
23572 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
23573 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
23575 Note that a desync will happen at the stitch if the audio and video streams
23576 do not have exactly the same duration in the first file.
23580 @subsection Commands
23582 This filter supports the following commands:
23585 Close the current segment and step to the next one
23591 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
23592 level. By default, it logs a message at a frequency of 10Hz with the
23593 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
23594 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
23596 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
23597 sample format is double-precision floating point. The input stream will be converted to
23598 this specification, if needed. Users may need to insert aformat and/or aresample filters
23599 after this filter to obtain the original parameters.
23601 The filter also has a video output (see the @var{video} option) with a real
23602 time graph to observe the loudness evolution. The graphic contains the logged
23603 message mentioned above, so it is not printed anymore when this option is set,
23604 unless the verbose logging is set. The main graphing area contains the
23605 short-term loudness (3 seconds of analysis), and the gauge on the right is for
23606 the momentary loudness (400 milliseconds), but can optionally be configured
23607 to instead display short-term loudness (see @var{gauge}).
23609 The green area marks a +/- 1LU target range around the target loudness
23610 (-23LUFS by default, unless modified through @var{target}).
23612 More information about the Loudness Recommendation EBU R128 on
23613 @url{http://tech.ebu.ch/loudness}.
23615 The filter accepts the following options:
23620 Activate the video output. The audio stream is passed unchanged whether this
23621 option is set or no. The video stream will be the first output stream if
23622 activated. Default is @code{0}.
23625 Set the video size. This option is for video only. For the syntax of this
23627 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23628 Default and minimum resolution is @code{640x480}.
23631 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
23632 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
23633 other integer value between this range is allowed.
23636 Set metadata injection. If set to @code{1}, the audio input will be segmented
23637 into 100ms output frames, each of them containing various loudness information
23638 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
23640 Default is @code{0}.
23643 Force the frame logging level.
23645 Available values are:
23648 information logging level
23650 verbose logging level
23653 By default, the logging level is set to @var{info}. If the @option{video} or
23654 the @option{metadata} options are set, it switches to @var{verbose}.
23659 Available modes can be cumulated (the option is a @code{flag} type). Possible
23663 Disable any peak mode (default).
23665 Enable sample-peak mode.
23667 Simple peak mode looking for the higher sample value. It logs a message
23668 for sample-peak (identified by @code{SPK}).
23670 Enable true-peak mode.
23672 If enabled, the peak lookup is done on an over-sampled version of the input
23673 stream for better peak accuracy. It logs a message for true-peak.
23674 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
23675 This mode requires a build with @code{libswresample}.
23679 Treat mono input files as "dual mono". If a mono file is intended for playback
23680 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
23681 If set to @code{true}, this option will compensate for this effect.
23682 Multi-channel input files are not affected by this option.
23685 Set a specific pan law to be used for the measurement of dual mono files.
23686 This parameter is optional, and has a default value of -3.01dB.
23689 Set a specific target level (in LUFS) used as relative zero in the visualization.
23690 This parameter is optional and has a default value of -23LUFS as specified
23691 by EBU R128. However, material published online may prefer a level of -16LUFS
23692 (e.g. for use with podcasts or video platforms).
23695 Set the value displayed by the gauge. Valid values are @code{momentary} and s
23696 @code{shortterm}. By default the momentary value will be used, but in certain
23697 scenarios it may be more useful to observe the short term value instead (e.g.
23701 Sets the display scale for the loudness. Valid parameters are @code{absolute}
23702 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
23703 video output, not the summary or continuous log output.
23706 @subsection Examples
23710 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
23712 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
23716 Run an analysis with @command{ffmpeg}:
23718 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
23722 @section interleave, ainterleave
23724 Temporally interleave frames from several inputs.
23726 @code{interleave} works with video inputs, @code{ainterleave} with audio.
23728 These filters read frames from several inputs and send the oldest
23729 queued frame to the output.
23731 Input streams must have well defined, monotonically increasing frame
23734 In order to submit one frame to output, these filters need to enqueue
23735 at least one frame for each input, so they cannot work in case one
23736 input is not yet terminated and will not receive incoming frames.
23738 For example consider the case when one input is a @code{select} filter
23739 which always drops input frames. The @code{interleave} filter will keep
23740 reading from that input, but it will never be able to send new frames
23741 to output until the input sends an end-of-stream signal.
23743 Also, depending on inputs synchronization, the filters will drop
23744 frames in case one input receives more frames than the other ones, and
23745 the queue is already filled.
23747 These filters accept the following options:
23751 Set the number of different inputs, it is 2 by default.
23754 How to determine the end-of-stream.
23758 The duration of the longest input. (default)
23761 The duration of the shortest input.
23764 The duration of the first input.
23769 @subsection Examples
23773 Interleave frames belonging to different streams using @command{ffmpeg}:
23775 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
23779 Add flickering blur effect:
23781 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
23785 @section metadata, ametadata
23787 Manipulate frame metadata.
23789 This filter accepts the following options:
23793 Set mode of operation of the filter.
23795 Can be one of the following:
23799 If both @code{value} and @code{key} is set, select frames
23800 which have such metadata. If only @code{key} is set, select
23801 every frame that has such key in metadata.
23804 Add new metadata @code{key} and @code{value}. If key is already available
23808 Modify value of already present key.
23811 If @code{value} is set, delete only keys that have such value.
23812 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
23816 Print key and its value if metadata was found. If @code{key} is not set print all
23817 metadata values available in frame.
23821 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
23824 Set metadata value which will be used. This option is mandatory for
23825 @code{modify} and @code{add} mode.
23828 Which function to use when comparing metadata value and @code{value}.
23830 Can be one of following:
23834 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
23837 Values are interpreted as strings, returns true if metadata value starts with
23838 the @code{value} option string.
23841 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
23844 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
23847 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
23850 Values are interpreted as floats, returns true if expression from option @code{expr}
23854 Values are interpreted as strings, returns true if metadata value ends with
23855 the @code{value} option string.
23859 Set expression which is used when @code{function} is set to @code{expr}.
23860 The expression is evaluated through the eval API and can contain the following
23865 Float representation of @code{value} from metadata key.
23868 Float representation of @code{value} as supplied by user in @code{value} option.
23872 If specified in @code{print} mode, output is written to the named file. Instead of
23873 plain filename any writable url can be specified. Filename ``-'' is a shorthand
23874 for standard output. If @code{file} option is not set, output is written to the log
23875 with AV_LOG_INFO loglevel.
23878 Reduces buffering in print mode when output is written to a URL set using @var{file}.
23882 @subsection Examples
23886 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
23889 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
23892 Print silencedetect output to file @file{metadata.txt}.
23894 silencedetect,ametadata=mode=print:file=metadata.txt
23897 Direct all metadata to a pipe with file descriptor 4.
23899 metadata=mode=print:file='pipe\:4'
23903 @section perms, aperms
23905 Set read/write permissions for the output frames.
23907 These filters are mainly aimed at developers to test direct path in the
23908 following filter in the filtergraph.
23910 The filters accept the following options:
23914 Select the permissions mode.
23916 It accepts the following values:
23919 Do nothing. This is the default.
23921 Set all the output frames read-only.
23923 Set all the output frames directly writable.
23925 Make the frame read-only if writable, and writable if read-only.
23927 Set each output frame read-only or writable randomly.
23931 Set the seed for the @var{random} mode, must be an integer included between
23932 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
23933 @code{-1}, the filter will try to use a good random seed on a best effort
23937 Note: in case of auto-inserted filter between the permission filter and the
23938 following one, the permission might not be received as expected in that
23939 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
23940 perms/aperms filter can avoid this problem.
23942 @section realtime, arealtime
23944 Slow down filtering to match real time approximately.
23946 These filters will pause the filtering for a variable amount of time to
23947 match the output rate with the input timestamps.
23948 They are similar to the @option{re} option to @code{ffmpeg}.
23950 They accept the following options:
23954 Time limit for the pauses. Any pause longer than that will be considered
23955 a timestamp discontinuity and reset the timer. Default is 2 seconds.
23957 Speed factor for processing. The value must be a float larger than zero.
23958 Values larger than 1.0 will result in faster than realtime processing,
23959 smaller will slow processing down. The @var{limit} is automatically adapted
23960 accordingly. Default is 1.0.
23962 A processing speed faster than what is possible without these filters cannot
23967 @section select, aselect
23969 Select frames to pass in output.
23971 This filter accepts the following options:
23976 Set expression, which is evaluated for each input frame.
23978 If the expression is evaluated to zero, the frame is discarded.
23980 If the evaluation result is negative or NaN, the frame is sent to the
23981 first output; otherwise it is sent to the output with index
23982 @code{ceil(val)-1}, assuming that the input index starts from 0.
23984 For example a value of @code{1.2} corresponds to the output with index
23985 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
23988 Set the number of outputs. The output to which to send the selected
23989 frame is based on the result of the evaluation. Default value is 1.
23992 The expression can contain the following constants:
23996 The (sequential) number of the filtered frame, starting from 0.
23999 The (sequential) number of the selected frame, starting from 0.
24001 @item prev_selected_n
24002 The sequential number of the last selected frame. It's NAN if undefined.
24005 The timebase of the input timestamps.
24008 The PTS (Presentation TimeStamp) of the filtered video frame,
24009 expressed in @var{TB} units. It's NAN if undefined.
24012 The PTS of the filtered video frame,
24013 expressed in seconds. It's NAN if undefined.
24016 The PTS of the previously filtered video frame. It's NAN if undefined.
24018 @item prev_selected_pts
24019 The PTS of the last previously filtered video frame. It's NAN if undefined.
24021 @item prev_selected_t
24022 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24025 The PTS of the first video frame in the video. It's NAN if undefined.
24028 The time of the first video frame in the video. It's NAN if undefined.
24030 @item pict_type @emph{(video only)}
24031 The type of the filtered frame. It can assume one of the following
24043 @item interlace_type @emph{(video only)}
24044 The frame interlace type. It can assume one of the following values:
24047 The frame is progressive (not interlaced).
24049 The frame is top-field-first.
24051 The frame is bottom-field-first.
24054 @item consumed_sample_n @emph{(audio only)}
24055 the number of selected samples before the current frame
24057 @item samples_n @emph{(audio only)}
24058 the number of samples in the current frame
24060 @item sample_rate @emph{(audio only)}
24061 the input sample rate
24064 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24067 the position in the file of the filtered frame, -1 if the information
24068 is not available (e.g. for synthetic video)
24070 @item scene @emph{(video only)}
24071 value between 0 and 1 to indicate a new scene; a low value reflects a low
24072 probability for the current frame to introduce a new scene, while a higher
24073 value means the current frame is more likely to be one (see the example below)
24075 @item concatdec_select
24076 The concat demuxer can select only part of a concat input file by setting an
24077 inpoint and an outpoint, but the output packets may not be entirely contained
24078 in the selected interval. By using this variable, it is possible to skip frames
24079 generated by the concat demuxer which are not exactly contained in the selected
24082 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24083 and the @var{lavf.concat.duration} packet metadata values which are also
24084 present in the decoded frames.
24086 The @var{concatdec_select} variable is -1 if the frame pts is at least
24087 start_time and either the duration metadata is missing or the frame pts is less
24088 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24091 That basically means that an input frame is selected if its pts is within the
24092 interval set by the concat demuxer.
24096 The default value of the select expression is "1".
24098 @subsection Examples
24102 Select all frames in input:
24107 The example above is the same as:
24119 Select only I-frames:
24121 select='eq(pict_type\,I)'
24125 Select one frame every 100:
24127 select='not(mod(n\,100))'
24131 Select only frames contained in the 10-20 time interval:
24133 select=between(t\,10\,20)
24137 Select only I-frames contained in the 10-20 time interval:
24139 select=between(t\,10\,20)*eq(pict_type\,I)
24143 Select frames with a minimum distance of 10 seconds:
24145 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24149 Use aselect to select only audio frames with samples number > 100:
24151 aselect='gt(samples_n\,100)'
24155 Create a mosaic of the first scenes:
24157 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24160 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24164 Send even and odd frames to separate outputs, and compose them:
24166 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24170 Select useful frames from an ffconcat file which is using inpoints and
24171 outpoints but where the source files are not intra frame only.
24173 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24177 @section sendcmd, asendcmd
24179 Send commands to filters in the filtergraph.
24181 These filters read commands to be sent to other filters in the
24184 @code{sendcmd} must be inserted between two video filters,
24185 @code{asendcmd} must be inserted between two audio filters, but apart
24186 from that they act the same way.
24188 The specification of commands can be provided in the filter arguments
24189 with the @var{commands} option, or in a file specified by the
24190 @var{filename} option.
24192 These filters accept the following options:
24195 Set the commands to be read and sent to the other filters.
24197 Set the filename of the commands to be read and sent to the other
24201 @subsection Commands syntax
24203 A commands description consists of a sequence of interval
24204 specifications, comprising a list of commands to be executed when a
24205 particular event related to that interval occurs. The occurring event
24206 is typically the current frame time entering or leaving a given time
24209 An interval is specified by the following syntax:
24211 @var{START}[-@var{END}] @var{COMMANDS};
24214 The time interval is specified by the @var{START} and @var{END} times.
24215 @var{END} is optional and defaults to the maximum time.
24217 The current frame time is considered within the specified interval if
24218 it is included in the interval [@var{START}, @var{END}), that is when
24219 the time is greater or equal to @var{START} and is lesser than
24222 @var{COMMANDS} consists of a sequence of one or more command
24223 specifications, separated by ",", relating to that interval. The
24224 syntax of a command specification is given by:
24226 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
24229 @var{FLAGS} is optional and specifies the type of events relating to
24230 the time interval which enable sending the specified command, and must
24231 be a non-null sequence of identifier flags separated by "+" or "|" and
24232 enclosed between "[" and "]".
24234 The following flags are recognized:
24237 The command is sent when the current frame timestamp enters the
24238 specified interval. In other words, the command is sent when the
24239 previous frame timestamp was not in the given interval, and the
24243 The command is sent when the current frame timestamp leaves the
24244 specified interval. In other words, the command is sent when the
24245 previous frame timestamp was in the given interval, and the
24249 The command @var{ARG} is interpreted as expression and result of
24250 expression is passed as @var{ARG}.
24252 The expression is evaluated through the eval API and can contain the following
24257 Original position in the file of the frame, or undefined if undefined
24258 for the current frame.
24261 The presentation timestamp in input.
24264 The count of the input frame for video or audio, starting from 0.
24267 The time in seconds of the current frame.
24270 The start time in seconds of the current command interval.
24273 The end time in seconds of the current command interval.
24276 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
24281 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
24284 @var{TARGET} specifies the target of the command, usually the name of
24285 the filter class or a specific filter instance name.
24287 @var{COMMAND} specifies the name of the command for the target filter.
24289 @var{ARG} is optional and specifies the optional list of argument for
24290 the given @var{COMMAND}.
24292 Between one interval specification and another, whitespaces, or
24293 sequences of characters starting with @code{#} until the end of line,
24294 are ignored and can be used to annotate comments.
24296 A simplified BNF description of the commands specification syntax
24299 @var{COMMAND_FLAG} ::= "enter" | "leave"
24300 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
24301 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
24302 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
24303 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
24304 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
24307 @subsection Examples
24311 Specify audio tempo change at second 4:
24313 asendcmd=c='4.0 atempo tempo 1.5',atempo
24317 Target a specific filter instance:
24319 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
24323 Specify a list of drawtext and hue commands in a file.
24325 # show text in the interval 5-10
24326 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
24327 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
24329 # desaturate the image in the interval 15-20
24330 15.0-20.0 [enter] hue s 0,
24331 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
24333 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
24335 # apply an exponential saturation fade-out effect, starting from time 25
24336 25 [enter] hue s exp(25-t)
24339 A filtergraph allowing to read and process the above command list
24340 stored in a file @file{test.cmd}, can be specified with:
24342 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
24347 @section setpts, asetpts
24349 Change the PTS (presentation timestamp) of the input frames.
24351 @code{setpts} works on video frames, @code{asetpts} on audio frames.
24353 This filter accepts the following options:
24358 The expression which is evaluated for each frame to construct its timestamp.
24362 The expression is evaluated through the eval API and can contain the following
24366 @item FRAME_RATE, FR
24367 frame rate, only defined for constant frame-rate video
24370 The presentation timestamp in input
24373 The count of the input frame for video or the number of consumed samples,
24374 not including the current frame for audio, starting from 0.
24376 @item NB_CONSUMED_SAMPLES
24377 The number of consumed samples, not including the current frame (only
24380 @item NB_SAMPLES, S
24381 The number of samples in the current frame (only audio)
24383 @item SAMPLE_RATE, SR
24384 The audio sample rate.
24387 The PTS of the first frame.
24390 the time in seconds of the first frame
24393 State whether the current frame is interlaced.
24396 the time in seconds of the current frame
24399 original position in the file of the frame, or undefined if undefined
24400 for the current frame
24403 The previous input PTS.
24406 previous input time in seconds
24409 The previous output PTS.
24412 previous output time in seconds
24415 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
24419 The wallclock (RTC) time at the start of the movie in microseconds.
24422 The timebase of the input timestamps.
24426 @subsection Examples
24430 Start counting PTS from zero
24432 setpts=PTS-STARTPTS
24436 Apply fast motion effect:
24442 Apply slow motion effect:
24448 Set fixed rate of 25 frames per second:
24454 Set fixed rate 25 fps with some jitter:
24456 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
24460 Apply an offset of 10 seconds to the input PTS:
24466 Generate timestamps from a "live source" and rebase onto the current timebase:
24468 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
24472 Generate timestamps by counting samples:
24481 Force color range for the output video frame.
24483 The @code{setrange} filter marks the color range property for the
24484 output frames. It does not change the input frame, but only sets the
24485 corresponding property, which affects how the frame is treated by
24488 The filter accepts the following options:
24493 Available values are:
24497 Keep the same color range property.
24499 @item unspecified, unknown
24500 Set the color range as unspecified.
24502 @item limited, tv, mpeg
24503 Set the color range as limited.
24505 @item full, pc, jpeg
24506 Set the color range as full.
24510 @section settb, asettb
24512 Set the timebase to use for the output frames timestamps.
24513 It is mainly useful for testing timebase configuration.
24515 It accepts the following parameters:
24520 The expression which is evaluated into the output timebase.
24524 The value for @option{tb} is an arithmetic expression representing a
24525 rational. The expression can contain the constants "AVTB" (the default
24526 timebase), "intb" (the input timebase) and "sr" (the sample rate,
24527 audio only). Default value is "intb".
24529 @subsection Examples
24533 Set the timebase to 1/25:
24539 Set the timebase to 1/10:
24545 Set the timebase to 1001/1000:
24551 Set the timebase to 2*intb:
24557 Set the default timebase value:
24564 Convert input audio to a video output representing frequency spectrum
24565 logarithmically using Brown-Puckette constant Q transform algorithm with
24566 direct frequency domain coefficient calculation (but the transform itself
24567 is not really constant Q, instead the Q factor is actually variable/clamped),
24568 with musical tone scale, from E0 to D#10.
24570 The filter accepts the following options:
24574 Specify the video size for the output. It must be even. For the syntax of this option,
24575 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24576 Default value is @code{1920x1080}.
24579 Set the output frame rate. Default value is @code{25}.
24582 Set the bargraph height. It must be even. Default value is @code{-1} which
24583 computes the bargraph height automatically.
24586 Set the axis height. It must be even. Default value is @code{-1} which computes
24587 the axis height automatically.
24590 Set the sonogram height. It must be even. Default value is @code{-1} which
24591 computes the sonogram height automatically.
24594 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
24595 instead. Default value is @code{1}.
24597 @item sono_v, volume
24598 Specify the sonogram volume expression. It can contain variables:
24601 the @var{bar_v} evaluated expression
24602 @item frequency, freq, f
24603 the frequency where it is evaluated
24604 @item timeclamp, tc
24605 the value of @var{timeclamp} option
24609 @item a_weighting(f)
24610 A-weighting of equal loudness
24611 @item b_weighting(f)
24612 B-weighting of equal loudness
24613 @item c_weighting(f)
24614 C-weighting of equal loudness.
24616 Default value is @code{16}.
24618 @item bar_v, volume2
24619 Specify the bargraph volume expression. It can contain variables:
24622 the @var{sono_v} evaluated expression
24623 @item frequency, freq, f
24624 the frequency where it is evaluated
24625 @item timeclamp, tc
24626 the value of @var{timeclamp} option
24630 @item a_weighting(f)
24631 A-weighting of equal loudness
24632 @item b_weighting(f)
24633 B-weighting of equal loudness
24634 @item c_weighting(f)
24635 C-weighting of equal loudness.
24637 Default value is @code{sono_v}.
24639 @item sono_g, gamma
24640 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
24641 higher gamma makes the spectrum having more range. Default value is @code{3}.
24642 Acceptable range is @code{[1, 7]}.
24644 @item bar_g, gamma2
24645 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
24649 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
24650 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
24652 @item timeclamp, tc
24653 Specify the transform timeclamp. At low frequency, there is trade-off between
24654 accuracy in time domain and frequency domain. If timeclamp is lower,
24655 event in time domain is represented more accurately (such as fast bass drum),
24656 otherwise event in frequency domain is represented more accurately
24657 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
24660 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
24661 limits future samples by applying asymmetric windowing in time domain, useful
24662 when low latency is required. Accepted range is @code{[0, 1]}.
24665 Specify the transform base frequency. Default value is @code{20.01523126408007475},
24666 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
24669 Specify the transform end frequency. Default value is @code{20495.59681441799654},
24670 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
24673 This option is deprecated and ignored.
24676 Specify the transform length in time domain. Use this option to control accuracy
24677 trade-off between time domain and frequency domain at every frequency sample.
24678 It can contain variables:
24680 @item frequency, freq, f
24681 the frequency where it is evaluated
24682 @item timeclamp, tc
24683 the value of @var{timeclamp} option.
24685 Default value is @code{384*tc/(384+tc*f)}.
24688 Specify the transform count for every video frame. Default value is @code{6}.
24689 Acceptable range is @code{[1, 30]}.
24692 Specify the transform count for every single pixel. Default value is @code{0},
24693 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
24696 Specify font file for use with freetype to draw the axis. If not specified,
24697 use embedded font. Note that drawing with font file or embedded font is not
24698 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
24702 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
24703 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
24707 Specify font color expression. This is arithmetic expression that should return
24708 integer value 0xRRGGBB. It can contain variables:
24710 @item frequency, freq, f
24711 the frequency where it is evaluated
24712 @item timeclamp, tc
24713 the value of @var{timeclamp} option
24718 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
24719 @item r(x), g(x), b(x)
24720 red, green, and blue value of intensity x.
24722 Default value is @code{st(0, (midi(f)-59.5)/12);
24723 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
24724 r(1-ld(1)) + b(ld(1))}.
24727 Specify image file to draw the axis. This option override @var{fontfile} and
24728 @var{fontcolor} option.
24731 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
24732 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
24733 Default value is @code{1}.
24736 Set colorspace. The accepted values are:
24739 Unspecified (default)
24748 BT.470BG or BT.601-6 625
24751 SMPTE-170M or BT.601-6 525
24757 BT.2020 with non-constant luminance
24762 Set spectrogram color scheme. This is list of floating point values with format
24763 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
24764 The default is @code{1|0.5|0|0|0.5|1}.
24768 @subsection Examples
24772 Playing audio while showing the spectrum:
24774 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
24778 Same as above, but with frame rate 30 fps:
24780 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
24784 Playing at 1280x720:
24786 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
24790 Disable sonogram display:
24796 A1 and its harmonics: A1, A2, (near)E3, A3:
24798 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),
24799 asplit[a][out1]; [a] showcqt [out0]'
24803 Same as above, but with more accuracy in frequency domain:
24805 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),
24806 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
24812 bar_v=10:sono_v=bar_v*a_weighting(f)
24816 Custom gamma, now spectrum is linear to the amplitude.
24822 Custom tlength equation:
24824 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)))'
24828 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
24830 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
24834 Custom font using fontconfig:
24836 font='Courier New,Monospace,mono|bold'
24840 Custom frequency range with custom axis using image file:
24842 axisfile=myaxis.png:basefreq=40:endfreq=10000
24848 Convert input audio to video output representing the audio power spectrum.
24849 Audio amplitude is on Y-axis while frequency is on X-axis.
24851 The filter accepts the following options:
24855 Specify size of video. For the syntax of this option, check the
24856 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24857 Default is @code{1024x512}.
24861 This set how each frequency bin will be represented.
24863 It accepts the following values:
24869 Default is @code{bar}.
24872 Set amplitude scale.
24874 It accepts the following values:
24888 Default is @code{log}.
24891 Set frequency scale.
24893 It accepts the following values:
24902 Reverse logarithmic scale.
24904 Default is @code{lin}.
24907 Set window size. Allowed range is from 16 to 65536.
24909 Default is @code{2048}
24912 Set windowing function.
24914 It accepts the following values:
24937 Default is @code{hanning}.
24940 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
24941 which means optimal overlap for selected window function will be picked.
24944 Set time averaging. Setting this to 0 will display current maximal peaks.
24945 Default is @code{1}, which means time averaging is disabled.
24948 Specify list of colors separated by space or by '|' which will be used to
24949 draw channel frequencies. Unrecognized or missing colors will be replaced
24953 Set channel display mode.
24955 It accepts the following values:
24960 Default is @code{combined}.
24963 Set minimum amplitude used in @code{log} amplitude scaler.
24967 @section showspatial
24969 Convert stereo input audio to a video output, representing the spatial relationship
24970 between two channels.
24972 The filter accepts the following options:
24976 Specify the video size for the output. For the syntax of this option, check the
24977 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24978 Default value is @code{512x512}.
24981 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
24984 Set window function.
24986 It accepts the following values:
25011 Default value is @code{hann}.
25014 Set ratio of overlap window. Default value is @code{0.5}.
25015 When value is @code{1} overlap is set to recommended size for specific
25016 window function currently used.
25019 @anchor{showspectrum}
25020 @section showspectrum
25022 Convert input audio to a video output, representing the audio frequency
25025 The filter accepts the following options:
25029 Specify the video size for the output. For the syntax of this option, check the
25030 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25031 Default value is @code{640x512}.
25034 Specify how the spectrum should slide along the window.
25036 It accepts the following values:
25039 the samples start again on the left when they reach the right
25041 the samples scroll from right to left
25043 frames are only produced when the samples reach the right
25045 the samples scroll from left to right
25048 Default value is @code{replace}.
25051 Specify display mode.
25053 It accepts the following values:
25056 all channels are displayed in the same row
25058 all channels are displayed in separate rows
25061 Default value is @samp{combined}.
25064 Specify display color mode.
25066 It accepts the following values:
25069 each channel is displayed in a separate color
25071 each channel is displayed using the same color scheme
25073 each channel is displayed using the rainbow color scheme
25075 each channel is displayed using the moreland color scheme
25077 each channel is displayed using the nebulae color scheme
25079 each channel is displayed using the fire color scheme
25081 each channel is displayed using the fiery color scheme
25083 each channel is displayed using the fruit color scheme
25085 each channel is displayed using the cool color scheme
25087 each channel is displayed using the magma color scheme
25089 each channel is displayed using the green color scheme
25091 each channel is displayed using the viridis color scheme
25093 each channel is displayed using the plasma color scheme
25095 each channel is displayed using the cividis color scheme
25097 each channel is displayed using the terrain color scheme
25100 Default value is @samp{channel}.
25103 Specify scale used for calculating intensity color values.
25105 It accepts the following values:
25110 square root, default
25121 Default value is @samp{sqrt}.
25124 Specify frequency scale.
25126 It accepts the following values:
25134 Default value is @samp{lin}.
25137 Set saturation modifier for displayed colors. Negative values provide
25138 alternative color scheme. @code{0} is no saturation at all.
25139 Saturation must be in [-10.0, 10.0] range.
25140 Default value is @code{1}.
25143 Set window function.
25145 It accepts the following values:
25170 Default value is @code{hann}.
25173 Set orientation of time vs frequency axis. Can be @code{vertical} or
25174 @code{horizontal}. Default is @code{vertical}.
25177 Set ratio of overlap window. Default value is @code{0}.
25178 When value is @code{1} overlap is set to recommended size for specific
25179 window function currently used.
25182 Set scale gain for calculating intensity color values.
25183 Default value is @code{1}.
25186 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25189 Set color rotation, must be in [-1.0, 1.0] range.
25190 Default value is @code{0}.
25193 Set start frequency from which to display spectrogram. Default is @code{0}.
25196 Set stop frequency to which to display spectrogram. Default is @code{0}.
25199 Set upper frame rate limit. Default is @code{auto}, unlimited.
25202 Draw time and frequency axes and legends. Default is disabled.
25205 The usage is very similar to the showwaves filter; see the examples in that
25208 @subsection Examples
25212 Large window with logarithmic color scaling:
25214 showspectrum=s=1280x480:scale=log
25218 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
25220 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
25221 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
25225 @section showspectrumpic
25227 Convert input audio to a single video frame, representing the audio frequency
25230 The filter accepts the following options:
25234 Specify the video size for the output. For the syntax of this option, check the
25235 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25236 Default value is @code{4096x2048}.
25239 Specify display mode.
25241 It accepts the following values:
25244 all channels are displayed in the same row
25246 all channels are displayed in separate rows
25248 Default value is @samp{combined}.
25251 Specify display color mode.
25253 It accepts the following values:
25256 each channel is displayed in a separate color
25258 each channel is displayed using the same color scheme
25260 each channel is displayed using the rainbow color scheme
25262 each channel is displayed using the moreland color scheme
25264 each channel is displayed using the nebulae color scheme
25266 each channel is displayed using the fire color scheme
25268 each channel is displayed using the fiery color scheme
25270 each channel is displayed using the fruit color scheme
25272 each channel is displayed using the cool color scheme
25274 each channel is displayed using the magma color scheme
25276 each channel is displayed using the green color scheme
25278 each channel is displayed using the viridis color scheme
25280 each channel is displayed using the plasma color scheme
25282 each channel is displayed using the cividis color scheme
25284 each channel is displayed using the terrain color scheme
25286 Default value is @samp{intensity}.
25289 Specify scale used for calculating intensity color values.
25291 It accepts the following values:
25296 square root, default
25306 Default value is @samp{log}.
25309 Specify frequency scale.
25311 It accepts the following values:
25319 Default value is @samp{lin}.
25322 Set saturation modifier for displayed colors. Negative values provide
25323 alternative color scheme. @code{0} is no saturation at all.
25324 Saturation must be in [-10.0, 10.0] range.
25325 Default value is @code{1}.
25328 Set window function.
25330 It accepts the following values:
25354 Default value is @code{hann}.
25357 Set orientation of time vs frequency axis. Can be @code{vertical} or
25358 @code{horizontal}. Default is @code{vertical}.
25361 Set scale gain for calculating intensity color values.
25362 Default value is @code{1}.
25365 Draw time and frequency axes and legends. Default is enabled.
25368 Set color rotation, must be in [-1.0, 1.0] range.
25369 Default value is @code{0}.
25372 Set start frequency from which to display spectrogram. Default is @code{0}.
25375 Set stop frequency to which to display spectrogram. Default is @code{0}.
25378 @subsection Examples
25382 Extract an audio spectrogram of a whole audio track
25383 in a 1024x1024 picture using @command{ffmpeg}:
25385 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
25389 @section showvolume
25391 Convert input audio volume to a video output.
25393 The filter accepts the following options:
25400 Set border width, allowed range is [0, 5]. Default is 1.
25403 Set channel width, allowed range is [80, 8192]. Default is 400.
25406 Set channel height, allowed range is [1, 900]. Default is 20.
25409 Set fade, allowed range is [0, 1]. Default is 0.95.
25412 Set volume color expression.
25414 The expression can use the following variables:
25418 Current max volume of channel in dB.
25424 Current channel number, starting from 0.
25428 If set, displays channel names. Default is enabled.
25431 If set, displays volume values. Default is enabled.
25434 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
25435 default is @code{h}.
25438 Set step size, allowed range is [0, 5]. Default is 0, which means
25442 Set background opacity, allowed range is [0, 1]. Default is 0.
25445 Set metering mode, can be peak: @code{p} or rms: @code{r},
25446 default is @code{p}.
25449 Set display scale, can be linear: @code{lin} or log: @code{log},
25450 default is @code{lin}.
25454 If set to > 0., display a line for the max level
25455 in the previous seconds.
25456 default is disabled: @code{0.}
25459 The color of the max line. Use when @code{dm} option is set to > 0.
25460 default is: @code{orange}
25465 Convert input audio to a video output, representing the samples waves.
25467 The filter accepts the following options:
25471 Specify the video size for the output. For the syntax of this option, check the
25472 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25473 Default value is @code{600x240}.
25478 Available values are:
25481 Draw a point for each sample.
25484 Draw a vertical line for each sample.
25487 Draw a point for each sample and a line between them.
25490 Draw a centered vertical line for each sample.
25493 Default value is @code{point}.
25496 Set the number of samples which are printed on the same column. A
25497 larger value will decrease the frame rate. Must be a positive
25498 integer. This option can be set only if the value for @var{rate}
25499 is not explicitly specified.
25502 Set the (approximate) output frame rate. This is done by setting the
25503 option @var{n}. Default value is "25".
25505 @item split_channels
25506 Set if channels should be drawn separately or overlap. Default value is 0.
25509 Set colors separated by '|' which are going to be used for drawing of each channel.
25512 Set amplitude scale.
25514 Available values are:
25532 Set the draw mode. This is mostly useful to set for high @var{n}.
25534 Available values are:
25537 Scale pixel values for each drawn sample.
25540 Draw every sample directly.
25543 Default value is @code{scale}.
25546 @subsection Examples
25550 Output the input file audio and the corresponding video representation
25553 amovie=a.mp3,asplit[out0],showwaves[out1]
25557 Create a synthetic signal and show it with showwaves, forcing a
25558 frame rate of 30 frames per second:
25560 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
25564 @section showwavespic
25566 Convert input audio to a single video frame, representing the samples waves.
25568 The filter accepts the following options:
25572 Specify the video size for the output. For the syntax of this option, check the
25573 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25574 Default value is @code{600x240}.
25576 @item split_channels
25577 Set if channels should be drawn separately or overlap. Default value is 0.
25580 Set colors separated by '|' which are going to be used for drawing of each channel.
25583 Set amplitude scale.
25585 Available values are:
25605 Available values are:
25608 Scale pixel values for each drawn sample.
25611 Draw every sample directly.
25614 Default value is @code{scale}.
25617 Set the filter mode.
25619 Available values are:
25622 Use average samples values for each drawn sample.
25625 Use peak samples values for each drawn sample.
25628 Default value is @code{average}.
25631 @subsection Examples
25635 Extract a channel split representation of the wave form of a whole audio track
25636 in a 1024x800 picture using @command{ffmpeg}:
25638 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
25642 @section sidedata, asidedata
25644 Delete frame side data, or select frames based on it.
25646 This filter accepts the following options:
25650 Set mode of operation of the filter.
25652 Can be one of the following:
25656 Select every frame with side data of @code{type}.
25659 Delete side data of @code{type}. If @code{type} is not set, delete all side
25665 Set side data type used with all modes. Must be set for @code{select} mode. For
25666 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
25667 in @file{libavutil/frame.h}. For example, to choose
25668 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
25672 @section spectrumsynth
25674 Synthesize audio from 2 input video spectrums, first input stream represents
25675 magnitude across time and second represents phase across time.
25676 The filter will transform from frequency domain as displayed in videos back
25677 to time domain as presented in audio output.
25679 This filter is primarily created for reversing processed @ref{showspectrum}
25680 filter outputs, but can synthesize sound from other spectrograms too.
25681 But in such case results are going to be poor if the phase data is not
25682 available, because in such cases phase data need to be recreated, usually
25683 it's just recreated from random noise.
25684 For best results use gray only output (@code{channel} color mode in
25685 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
25686 @code{lin} scale for phase video. To produce phase, for 2nd video, use
25687 @code{data} option. Inputs videos should generally use @code{fullframe}
25688 slide mode as that saves resources needed for decoding video.
25690 The filter accepts the following options:
25694 Specify sample rate of output audio, the sample rate of audio from which
25695 spectrum was generated may differ.
25698 Set number of channels represented in input video spectrums.
25701 Set scale which was used when generating magnitude input spectrum.
25702 Can be @code{lin} or @code{log}. Default is @code{log}.
25705 Set slide which was used when generating inputs spectrums.
25706 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
25707 Default is @code{fullframe}.
25710 Set window function used for resynthesis.
25713 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25714 which means optimal overlap for selected window function will be picked.
25717 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
25718 Default is @code{vertical}.
25721 @subsection Examples
25725 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
25726 then resynthesize videos back to audio with spectrumsynth:
25728 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
25729 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
25730 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
25734 @section split, asplit
25736 Split input into several identical outputs.
25738 @code{asplit} works with audio input, @code{split} with video.
25740 The filter accepts a single parameter which specifies the number of outputs. If
25741 unspecified, it defaults to 2.
25743 @subsection Examples
25747 Create two separate outputs from the same input:
25749 [in] split [out0][out1]
25753 To create 3 or more outputs, you need to specify the number of
25756 [in] asplit=3 [out0][out1][out2]
25760 Create two separate outputs from the same input, one cropped and
25763 [in] split [splitout1][splitout2];
25764 [splitout1] crop=100:100:0:0 [cropout];
25765 [splitout2] pad=200:200:100:100 [padout];
25769 Create 5 copies of the input audio with @command{ffmpeg}:
25771 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
25777 Receive commands sent through a libzmq client, and forward them to
25778 filters in the filtergraph.
25780 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
25781 must be inserted between two video filters, @code{azmq} between two
25782 audio filters. Both are capable to send messages to any filter type.
25784 To enable these filters you need to install the libzmq library and
25785 headers and configure FFmpeg with @code{--enable-libzmq}.
25787 For more information about libzmq see:
25788 @url{http://www.zeromq.org/}
25790 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
25791 receives messages sent through a network interface defined by the
25792 @option{bind_address} (or the abbreviation "@option{b}") option.
25793 Default value of this option is @file{tcp://localhost:5555}. You may
25794 want to alter this value to your needs, but do not forget to escape any
25795 ':' signs (see @ref{filtergraph escaping}).
25797 The received message must be in the form:
25799 @var{TARGET} @var{COMMAND} [@var{ARG}]
25802 @var{TARGET} specifies the target of the command, usually the name of
25803 the filter class or a specific filter instance name. The default
25804 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
25805 but you can override this by using the @samp{filter_name@@id} syntax
25806 (see @ref{Filtergraph syntax}).
25808 @var{COMMAND} specifies the name of the command for the target filter.
25810 @var{ARG} is optional and specifies the optional argument list for the
25811 given @var{COMMAND}.
25813 Upon reception, the message is processed and the corresponding command
25814 is injected into the filtergraph. Depending on the result, the filter
25815 will send a reply to the client, adopting the format:
25817 @var{ERROR_CODE} @var{ERROR_REASON}
25821 @var{MESSAGE} is optional.
25823 @subsection Examples
25825 Look at @file{tools/zmqsend} for an example of a zmq client which can
25826 be used to send commands processed by these filters.
25828 Consider the following filtergraph generated by @command{ffplay}.
25829 In this example the last overlay filter has an instance name. All other
25830 filters will have default instance names.
25833 ffplay -dumpgraph 1 -f lavfi "
25834 color=s=100x100:c=red [l];
25835 color=s=100x100:c=blue [r];
25836 nullsrc=s=200x100, zmq [bg];
25837 [bg][l] overlay [bg+l];
25838 [bg+l][r] overlay@@my=x=100 "
25841 To change the color of the left side of the video, the following
25842 command can be used:
25844 echo Parsed_color_0 c yellow | tools/zmqsend
25847 To change the right side:
25849 echo Parsed_color_1 c pink | tools/zmqsend
25852 To change the position of the right side:
25854 echo overlay@@my x 150 | tools/zmqsend
25858 @c man end MULTIMEDIA FILTERS
25860 @chapter Multimedia Sources
25861 @c man begin MULTIMEDIA SOURCES
25863 Below is a description of the currently available multimedia sources.
25867 This is the same as @ref{movie} source, except it selects an audio
25873 Read audio and/or video stream(s) from a movie container.
25875 It accepts the following parameters:
25879 The name of the resource to read (not necessarily a file; it can also be a
25880 device or a stream accessed through some protocol).
25882 @item format_name, f
25883 Specifies the format assumed for the movie to read, and can be either
25884 the name of a container or an input device. If not specified, the
25885 format is guessed from @var{movie_name} or by probing.
25887 @item seek_point, sp
25888 Specifies the seek point in seconds. The frames will be output
25889 starting from this seek point. The parameter is evaluated with
25890 @code{av_strtod}, so the numerical value may be suffixed by an IS
25891 postfix. The default value is "0".
25894 Specifies the streams to read. Several streams can be specified,
25895 separated by "+". The source will then have as many outputs, in the
25896 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
25897 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
25898 respectively the default (best suited) video and audio stream. Default
25899 is "dv", or "da" if the filter is called as "amovie".
25901 @item stream_index, si
25902 Specifies the index of the video stream to read. If the value is -1,
25903 the most suitable video stream will be automatically selected. The default
25904 value is "-1". Deprecated. If the filter is called "amovie", it will select
25905 audio instead of video.
25908 Specifies how many times to read the stream in sequence.
25909 If the value is 0, the stream will be looped infinitely.
25910 Default value is "1".
25912 Note that when the movie is looped the source timestamps are not
25913 changed, so it will generate non monotonically increasing timestamps.
25915 @item discontinuity
25916 Specifies the time difference between frames above which the point is
25917 considered a timestamp discontinuity which is removed by adjusting the later
25921 It allows overlaying a second video on top of the main input of
25922 a filtergraph, as shown in this graph:
25924 input -----------> deltapts0 --> overlay --> output
25927 movie --> scale--> deltapts1 -------+
25929 @subsection Examples
25933 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
25934 on top of the input labelled "in":
25936 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
25937 [in] setpts=PTS-STARTPTS [main];
25938 [main][over] overlay=16:16 [out]
25942 Read from a video4linux2 device, and overlay it on top of the input
25945 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
25946 [in] setpts=PTS-STARTPTS [main];
25947 [main][over] overlay=16:16 [out]
25951 Read the first video stream and the audio stream with id 0x81 from
25952 dvd.vob; the video is connected to the pad named "video" and the audio is
25953 connected to the pad named "audio":
25955 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
25959 @subsection Commands
25961 Both movie and amovie support the following commands:
25964 Perform seek using "av_seek_frame".
25965 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
25968 @var{stream_index}: If stream_index is -1, a default
25969 stream is selected, and @var{timestamp} is automatically converted
25970 from AV_TIME_BASE units to the stream specific time_base.
25972 @var{timestamp}: Timestamp in AVStream.time_base units
25973 or, if no stream is specified, in AV_TIME_BASE units.
25975 @var{flags}: Flags which select direction and seeking mode.
25979 Get movie duration in AV_TIME_BASE units.
25983 @c man end MULTIMEDIA SOURCES