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.
1414 Z-plane zeros/poles, cartesian (default)
1416 Z-plane zeros/poles, polar radians
1418 Z-plane zeros/poles, polar degrees
1422 Set kind of processing.
1423 Can be @code{d} - direct or @code{s} - serial cascading. Default is @code{s}.
1426 Set filtering precision.
1430 double-precision floating-point (default)
1432 single-precision floating-point
1440 How much to use filtered signal in output. Default is 1.
1441 Range is between 0 and 1.
1444 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1445 By default it is disabled.
1448 Set for which IR channel to display frequency response. By default is first channel
1449 displayed. This option is used only when @var{response} is enabled.
1452 Set video stream size. This option is used only when @var{response} is enabled.
1455 Coefficients in @code{tf} format are separated by spaces and are in ascending
1458 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1459 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1462 Different coefficients and gains can be provided for every channel, in such case
1463 use '|' to separate coefficients or gains. Last provided coefficients will be
1464 used for all remaining channels.
1466 @subsection Examples
1470 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1472 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
1476 Same as above but in @code{zp} format:
1478 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
1484 The limiter prevents an input signal from rising over a desired threshold.
1485 This limiter uses lookahead technology to prevent your signal from distorting.
1486 It means that there is a small delay after the signal is processed. Keep in mind
1487 that the delay it produces is the attack time you set.
1489 The filter accepts the following options:
1493 Set input gain. Default is 1.
1496 Set output gain. Default is 1.
1499 Don't let signals above this level pass the limiter. Default is 1.
1502 The limiter will reach its attenuation level in this amount of time in
1503 milliseconds. Default is 5 milliseconds.
1506 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1507 Default is 50 milliseconds.
1510 When gain reduction is always needed ASC takes care of releasing to an
1511 average reduction level rather than reaching a reduction of 0 in the release
1515 Select how much the release time is affected by ASC, 0 means nearly no changes
1516 in release time while 1 produces higher release times.
1519 Auto level output signal. Default is enabled.
1520 This normalizes audio back to 0dB if enabled.
1523 Depending on picked setting it is recommended to upsample input 2x or 4x times
1524 with @ref{aresample} before applying this filter.
1528 Apply a two-pole all-pass filter with central frequency (in Hz)
1529 @var{frequency}, and filter-width @var{width}.
1530 An all-pass filter changes the audio's frequency to phase relationship
1531 without changing its frequency to amplitude relationship.
1533 The filter accepts the following options:
1537 Set frequency in Hz.
1540 Set method to specify band-width of filter.
1555 Specify the band-width of a filter in width_type units.
1558 How much to use filtered signal in output. Default is 1.
1559 Range is between 0 and 1.
1562 Specify which channels to filter, by default all available are filtered.
1565 Normalize biquad coefficients, by default is disabled.
1566 Enabling it will normalize magnitude response at DC to 0dB.
1569 @subsection Commands
1571 This filter supports the following commands:
1574 Change allpass frequency.
1575 Syntax for the command is : "@var{frequency}"
1578 Change allpass width_type.
1579 Syntax for the command is : "@var{width_type}"
1582 Change allpass width.
1583 Syntax for the command is : "@var{width}"
1587 Syntax for the command is : "@var{mix}"
1594 The filter accepts the following options:
1598 Set the number of loops. Setting this value to -1 will result in infinite loops.
1602 Set maximal number of samples. Default is 0.
1605 Set first sample of loop. Default is 0.
1611 Merge two or more audio streams into a single multi-channel stream.
1613 The filter accepts the following options:
1618 Set the number of inputs. Default is 2.
1622 If the channel layouts of the inputs are disjoint, and therefore compatible,
1623 the channel layout of the output will be set accordingly and the channels
1624 will be reordered as necessary. If the channel layouts of the inputs are not
1625 disjoint, the output will have all the channels of the first input then all
1626 the channels of the second input, in that order, and the channel layout of
1627 the output will be the default value corresponding to the total number of
1630 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1631 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1632 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1633 first input, b1 is the first channel of the second input).
1635 On the other hand, if both input are in stereo, the output channels will be
1636 in the default order: a1, a2, b1, b2, and the channel layout will be
1637 arbitrarily set to 4.0, which may or may not be the expected value.
1639 All inputs must have the same sample rate, and format.
1641 If inputs do not have the same duration, the output will stop with the
1644 @subsection Examples
1648 Merge two mono files into a stereo stream:
1650 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1654 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1656 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
1662 Mixes multiple audio inputs into a single output.
1664 Note that this filter only supports float samples (the @var{amerge}
1665 and @var{pan} audio filters support many formats). If the @var{amix}
1666 input has integer samples then @ref{aresample} will be automatically
1667 inserted to perform the conversion to float samples.
1671 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1673 will mix 3 input audio streams to a single output with the same duration as the
1674 first input and a dropout transition time of 3 seconds.
1676 It accepts the following parameters:
1680 The number of inputs. If unspecified, it defaults to 2.
1683 How to determine the end-of-stream.
1687 The duration of the longest input. (default)
1690 The duration of the shortest input.
1693 The duration of the first input.
1697 @item dropout_transition
1698 The transition time, in seconds, for volume renormalization when an input
1699 stream ends. The default value is 2 seconds.
1702 Specify weight of each input audio stream as sequence.
1703 Each weight is separated by space. By default all inputs have same weight.
1708 Multiply first audio stream with second audio stream and store result
1709 in output audio stream. Multiplication is done by multiplying each
1710 sample from first stream with sample at same position from second stream.
1712 With this element-wise multiplication one can create amplitude fades and
1713 amplitude modulations.
1715 @section anequalizer
1717 High-order parametric multiband equalizer for each channel.
1719 It accepts the following parameters:
1723 This option string is in format:
1724 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1725 Each equalizer band is separated by '|'.
1729 Set channel number to which equalization will be applied.
1730 If input doesn't have that channel the entry is ignored.
1733 Set central frequency for band.
1734 If input doesn't have that frequency the entry is ignored.
1737 Set band width in hertz.
1740 Set band gain in dB.
1743 Set filter type for band, optional, can be:
1747 Butterworth, this is default.
1758 With this option activated frequency response of anequalizer is displayed
1762 Set video stream size. Only useful if curves option is activated.
1765 Set max gain that will be displayed. Only useful if curves option is activated.
1766 Setting this to a reasonable value makes it possible to display gain which is derived from
1767 neighbour bands which are too close to each other and thus produce higher gain
1768 when both are activated.
1771 Set frequency scale used to draw frequency response in video output.
1772 Can be linear or logarithmic. Default is logarithmic.
1775 Set color for each channel curve which is going to be displayed in video stream.
1776 This is list of color names separated by space or by '|'.
1777 Unrecognised or missing colors will be replaced by white color.
1780 @subsection Examples
1784 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1785 for first 2 channels using Chebyshev type 1 filter:
1787 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1791 @subsection Commands
1793 This filter supports the following commands:
1796 Alter existing filter parameters.
1797 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1799 @var{fN} is existing filter number, starting from 0, if no such filter is available
1801 @var{freq} set new frequency parameter.
1802 @var{width} set new width parameter in herz.
1803 @var{gain} set new gain parameter in dB.
1805 Full filter invocation with asendcmd may look like this:
1806 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1811 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1813 Each sample is adjusted by looking for other samples with similar contexts. This
1814 context similarity is defined by comparing their surrounding patches of size
1815 @option{p}. Patches are searched in an area of @option{r} around the sample.
1817 The filter accepts the following options:
1821 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
1824 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
1825 Default value is 2 milliseconds.
1828 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
1829 Default value is 6 milliseconds.
1832 Set the output mode.
1834 It accepts the following values:
1837 Pass input unchanged.
1840 Pass noise filtered out.
1845 Default value is @var{o}.
1849 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
1852 @subsection Commands
1854 This filter supports the following commands:
1857 Change denoise strength. Argument is single float number.
1858 Syntax for the command is : "@var{s}"
1862 Syntax for the command is : "i", "o" or "n" string.
1866 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
1868 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
1869 relate to producing the least mean square of the error signal (difference between the desired,
1870 2nd input audio stream and the actual signal, the 1st input audio stream).
1872 A description of the accepted options follows.
1885 Set the filter leakage.
1888 It accepts the following values:
1897 Pass filtered samples.
1900 Pass difference between desired and filtered samples.
1902 Default value is @var{o}.
1906 @subsection Examples
1910 One of many usages of this filter is noise reduction, input audio is filtered
1911 with same samples that are delayed by fixed amount, one such example for stereo audio is:
1913 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
1917 @subsection Commands
1919 This filter supports the same commands as options, excluding option @code{order}.
1923 Pass the audio source unchanged to the output.
1927 Pad the end of an audio stream with silence.
1929 This can be used together with @command{ffmpeg} @option{-shortest} to
1930 extend audio streams to the same length as the video stream.
1932 A description of the accepted options follows.
1936 Set silence packet size. Default value is 4096.
1939 Set the number of samples of silence to add to the end. After the
1940 value is reached, the stream is terminated. This option is mutually
1941 exclusive with @option{whole_len}.
1944 Set the minimum total number of samples in the output audio stream. If
1945 the value is longer than the input audio length, silence is added to
1946 the end, until the value is reached. This option is mutually exclusive
1947 with @option{pad_len}.
1950 Specify the duration of samples of silence to add. See
1951 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1952 for the accepted syntax. Used only if set to non-zero value.
1955 Specify the minimum total duration in the output audio stream. See
1956 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1957 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
1958 the input audio length, silence is added to the end, until the value is reached.
1959 This option is mutually exclusive with @option{pad_dur}
1962 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
1963 nor @option{whole_dur} option is set, the filter will add silence to the end of
1964 the input stream indefinitely.
1966 @subsection Examples
1970 Add 1024 samples of silence to the end of the input:
1976 Make sure the audio output will contain at least 10000 samples, pad
1977 the input with silence if required:
1979 apad=whole_len=10000
1983 Use @command{ffmpeg} to pad the audio input with silence, so that the
1984 video stream will always result the shortest and will be converted
1985 until the end in the output file when using the @option{shortest}
1988 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
1993 Add a phasing effect to the input audio.
1995 A phaser filter creates series of peaks and troughs in the frequency spectrum.
1996 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
1998 A description of the accepted parameters follows.
2002 Set input gain. Default is 0.4.
2005 Set output gain. Default is 0.74
2008 Set delay in milliseconds. Default is 3.0.
2011 Set decay. Default is 0.4.
2014 Set modulation speed in Hz. Default is 0.5.
2017 Set modulation type. Default is triangular.
2019 It accepts the following values:
2028 Audio pulsator is something between an autopanner and a tremolo.
2029 But it can produce funny stereo effects as well. Pulsator changes the volume
2030 of the left and right channel based on a LFO (low frequency oscillator) with
2031 different waveforms and shifted phases.
2032 This filter have the ability to define an offset between left and right
2033 channel. An offset of 0 means that both LFO shapes match each other.
2034 The left and right channel are altered equally - a conventional tremolo.
2035 An offset of 50% means that the shape of the right channel is exactly shifted
2036 in phase (or moved backwards about half of the frequency) - pulsator acts as
2037 an autopanner. At 1 both curves match again. Every setting in between moves the
2038 phase shift gapless between all stages and produces some "bypassing" sounds with
2039 sine and triangle waveforms. The more you set the offset near 1 (starting from
2040 the 0.5) the faster the signal passes from the left to the right speaker.
2042 The filter accepts the following options:
2046 Set input gain. By default it is 1. Range is [0.015625 - 64].
2049 Set output gain. By default it is 1. Range is [0.015625 - 64].
2052 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2053 sawup or sawdown. Default is sine.
2056 Set modulation. Define how much of original signal is affected by the LFO.
2059 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2062 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2065 Set pulse width. Default is 1. Allowed range is [0 - 2].
2068 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2071 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2075 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2079 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2080 if timing is set to hz.
2086 Resample the input audio to the specified parameters, using the
2087 libswresample library. If none are specified then the filter will
2088 automatically convert between its input and output.
2090 This filter is also able to stretch/squeeze the audio data to make it match
2091 the timestamps or to inject silence / cut out audio to make it match the
2092 timestamps, do a combination of both or do neither.
2094 The filter accepts the syntax
2095 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2096 expresses a sample rate and @var{resampler_options} is a list of
2097 @var{key}=@var{value} pairs, separated by ":". See the
2098 @ref{Resampler Options,,"Resampler Options" section in the
2099 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2100 for the complete list of supported options.
2102 @subsection Examples
2106 Resample the input audio to 44100Hz:
2112 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2113 samples per second compensation:
2115 aresample=async=1000
2121 Reverse an audio clip.
2123 Warning: This filter requires memory to buffer the entire clip, so trimming
2126 @subsection Examples
2130 Take the first 5 seconds of a clip, and reverse it.
2132 atrim=end=5,areverse
2138 Reduce noise from speech using Recurrent Neural Networks.
2140 This filter accepts the following options:
2144 Set train model file to load. This option is always required.
2147 @section asetnsamples
2149 Set the number of samples per each output audio frame.
2151 The last output packet may contain a different number of samples, as
2152 the filter will flush all the remaining samples when the input audio
2155 The filter accepts the following options:
2159 @item nb_out_samples, n
2160 Set the number of frames per each output audio frame. The number is
2161 intended as the number of samples @emph{per each channel}.
2162 Default value is 1024.
2165 If set to 1, the filter will pad the last audio frame with zeroes, so
2166 that the last frame will contain the same number of samples as the
2167 previous ones. Default value is 1.
2170 For example, to set the number of per-frame samples to 1234 and
2171 disable padding for the last frame, use:
2173 asetnsamples=n=1234:p=0
2178 Set the sample rate without altering the PCM data.
2179 This will result in a change of speed and pitch.
2181 The filter accepts the following options:
2184 @item sample_rate, r
2185 Set the output sample rate. Default is 44100 Hz.
2190 Show a line containing various information for each input audio frame.
2191 The input audio is not modified.
2193 The shown line contains a sequence of key/value pairs of the form
2194 @var{key}:@var{value}.
2196 The following values are shown in the output:
2200 The (sequential) number of the input frame, starting from 0.
2203 The presentation timestamp of the input frame, in time base units; the time base
2204 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2207 The presentation timestamp of the input frame in seconds.
2210 position of the frame in the input stream, -1 if this information in
2211 unavailable and/or meaningless (for example in case of synthetic audio)
2220 The sample rate for the audio frame.
2223 The number of samples (per channel) in the frame.
2226 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2227 audio, the data is treated as if all the planes were concatenated.
2229 @item plane_checksums
2230 A list of Adler-32 checksums for each data plane.
2234 Apply audio soft clipping.
2236 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2237 along a smooth curve, rather than the abrupt shape of hard-clipping.
2239 This filter accepts the following options:
2243 Set type of soft-clipping.
2245 It accepts the following values:
2257 Set additional parameter which controls sigmoid function.
2260 @subsection Commands
2262 This filter supports the all above options as @ref{commands}.
2265 Automatic Speech Recognition
2267 This filter uses PocketSphinx for speech recognition. To enable
2268 compilation of this filter, you need to configure FFmpeg with
2269 @code{--enable-pocketsphinx}.
2271 It accepts the following options:
2275 Set sampling rate of input audio. Defaults is @code{16000}.
2276 This need to match speech models, otherwise one will get poor results.
2279 Set dictionary containing acoustic model files.
2282 Set pronunciation dictionary.
2285 Set language model file.
2288 Set language model set.
2291 Set which language model to use.
2294 Set output for log messages.
2297 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2302 Display time domain statistical information about the audio channels.
2303 Statistics are calculated and displayed for each audio channel and,
2304 where applicable, an overall figure is also given.
2306 It accepts the following option:
2309 Short window length in seconds, used for peak and trough RMS measurement.
2310 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2314 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2315 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2318 Available keys for each channel are:
2360 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2361 this @code{lavfi.astats.Overall.Peak_count}.
2363 For description what each key means read below.
2366 Set number of frame after which stats are going to be recalculated.
2367 Default is disabled.
2369 @item measure_perchannel
2370 Select the entries which need to be measured per channel. The metadata keys can
2371 be used as flags, default is @option{all} which measures everything.
2372 @option{none} disables all per channel measurement.
2374 @item measure_overall
2375 Select the entries which need to be measured overall. The metadata keys can
2376 be used as flags, default is @option{all} which measures everything.
2377 @option{none} disables all overall measurement.
2381 A description of each shown parameter follows:
2385 Mean amplitude displacement from zero.
2388 Minimal sample level.
2391 Maximal sample level.
2393 @item Min difference
2394 Minimal difference between two consecutive samples.
2396 @item Max difference
2397 Maximal difference between two consecutive samples.
2399 @item Mean difference
2400 Mean difference between two consecutive samples.
2401 The average of each difference between two consecutive samples.
2403 @item RMS difference
2404 Root Mean Square difference between two consecutive samples.
2408 Standard peak and RMS level measured in dBFS.
2412 Peak and trough values for RMS level measured over a short window.
2415 Standard ratio of peak to RMS level (note: not in dB).
2418 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2419 (i.e. either @var{Min level} or @var{Max level}).
2422 Number of occasions (not the number of samples) that the signal attained either
2423 @var{Min level} or @var{Max level}.
2426 Overall bit depth of audio. Number of bits used for each sample.
2429 Measured dynamic range of audio in dB.
2431 @item Zero crossings
2432 Number of points where the waveform crosses the zero level axis.
2434 @item Zero crossings rate
2435 Rate of Zero crossings and number of audio samples.
2442 The filter accepts exactly one parameter, the audio tempo. If not
2443 specified then the filter will assume nominal 1.0 tempo. Tempo must
2444 be in the [0.5, 100.0] range.
2446 Note that tempo greater than 2 will skip some samples rather than
2447 blend them in. If for any reason this is a concern it is always
2448 possible to daisy-chain several instances of atempo to achieve the
2449 desired product tempo.
2451 @subsection Examples
2455 Slow down audio to 80% tempo:
2461 To speed up audio to 300% tempo:
2467 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2469 atempo=sqrt(3),atempo=sqrt(3)
2473 @subsection Commands
2475 This filter supports the following commands:
2478 Change filter tempo scale factor.
2479 Syntax for the command is : "@var{tempo}"
2484 Trim the input so that the output contains one continuous subpart of the input.
2486 It accepts the following parameters:
2489 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2490 sample with the timestamp @var{start} will be the first sample in the output.
2493 Specify time of the first audio sample that will be dropped, i.e. the
2494 audio sample immediately preceding the one with the timestamp @var{end} will be
2495 the last sample in the output.
2498 Same as @var{start}, except this option sets the start timestamp in samples
2502 Same as @var{end}, except this option sets the end timestamp in samples instead
2506 The maximum duration of the output in seconds.
2509 The number of the first sample that should be output.
2512 The number of the first sample that should be dropped.
2515 @option{start}, @option{end}, and @option{duration} are expressed as time
2516 duration specifications; see
2517 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2519 Note that the first two sets of the start/end options and the @option{duration}
2520 option look at the frame timestamp, while the _sample options simply count the
2521 samples that pass through the filter. So start/end_pts and start/end_sample will
2522 give different results when the timestamps are wrong, inexact or do not start at
2523 zero. Also note that this filter does not modify the timestamps. If you wish
2524 to have the output timestamps start at zero, insert the asetpts filter after the
2527 If multiple start or end options are set, this filter tries to be greedy and
2528 keep all samples that match at least one of the specified constraints. To keep
2529 only the part that matches all the constraints at once, chain multiple atrim
2532 The defaults are such that all the input is kept. So it is possible to set e.g.
2533 just the end values to keep everything before the specified time.
2538 Drop everything except the second minute of input:
2540 ffmpeg -i INPUT -af atrim=60:120
2544 Keep only the first 1000 samples:
2546 ffmpeg -i INPUT -af atrim=end_sample=1000
2551 @section axcorrelate
2552 Calculate normalized cross-correlation between two input audio streams.
2554 Resulted samples are always between -1 and 1 inclusive.
2555 If result is 1 it means two input samples are highly correlated in that selected segment.
2556 Result 0 means they are not correlated at all.
2557 If result is -1 it means two input samples are out of phase, which means they cancel each
2560 The filter accepts the following options:
2564 Set size of segment over which cross-correlation is calculated.
2565 Default is 256. Allowed range is from 2 to 131072.
2568 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2569 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2570 are always zero and thus need much less calculations to make.
2571 This is generally not true, but is valid for typical audio streams.
2574 @subsection Examples
2578 Calculate correlation between channels in stereo audio stream:
2580 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2586 Apply a two-pole Butterworth band-pass filter with central
2587 frequency @var{frequency}, and (3dB-point) band-width width.
2588 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2589 instead of the default: constant 0dB peak gain.
2590 The filter roll off at 6dB per octave (20dB per decade).
2592 The filter accepts the following options:
2596 Set the filter's central frequency. Default is @code{3000}.
2599 Constant skirt gain if set to 1. Defaults to 0.
2602 Set method to specify band-width of filter.
2617 Specify the band-width of a filter in width_type units.
2620 How much to use filtered signal in output. Default is 1.
2621 Range is between 0 and 1.
2624 Specify which channels to filter, by default all available are filtered.
2627 Normalize biquad coefficients, by default is disabled.
2628 Enabling it will normalize magnitude response at DC to 0dB.
2631 @subsection Commands
2633 This filter supports the following commands:
2636 Change bandpass frequency.
2637 Syntax for the command is : "@var{frequency}"
2640 Change bandpass width_type.
2641 Syntax for the command is : "@var{width_type}"
2644 Change bandpass width.
2645 Syntax for the command is : "@var{width}"
2648 Change bandpass mix.
2649 Syntax for the command is : "@var{mix}"
2654 Apply a two-pole Butterworth band-reject filter with central
2655 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2656 The filter roll off at 6dB per octave (20dB per decade).
2658 The filter accepts the following options:
2662 Set the filter's central frequency. Default is @code{3000}.
2665 Set method to specify band-width of filter.
2680 Specify the band-width of a filter in width_type units.
2683 How much to use filtered signal in output. Default is 1.
2684 Range is between 0 and 1.
2687 Specify which channels to filter, by default all available are filtered.
2690 Normalize biquad coefficients, by default is disabled.
2691 Enabling it will normalize magnitude response at DC to 0dB.
2694 @subsection Commands
2696 This filter supports the following commands:
2699 Change bandreject frequency.
2700 Syntax for the command is : "@var{frequency}"
2703 Change bandreject width_type.
2704 Syntax for the command is : "@var{width_type}"
2707 Change bandreject width.
2708 Syntax for the command is : "@var{width}"
2711 Change bandreject mix.
2712 Syntax for the command is : "@var{mix}"
2715 @section bass, lowshelf
2717 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2718 shelving filter with a response similar to that of a standard
2719 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2721 The filter accepts the following options:
2725 Give the gain at 0 Hz. Its useful range is about -20
2726 (for a large cut) to +20 (for a large boost).
2727 Beware of clipping when using a positive gain.
2730 Set the filter's central frequency and so can be used
2731 to extend or reduce the frequency range to be boosted or cut.
2732 The default value is @code{100} Hz.
2735 Set method to specify band-width of filter.
2750 Determine how steep is the filter's shelf transition.
2753 How much to use filtered signal in output. Default is 1.
2754 Range is between 0 and 1.
2757 Specify which channels to filter, by default all available are filtered.
2760 Normalize biquad coefficients, by default is disabled.
2761 Enabling it will normalize magnitude response at DC to 0dB.
2764 @subsection Commands
2766 This filter supports the following commands:
2769 Change bass frequency.
2770 Syntax for the command is : "@var{frequency}"
2773 Change bass width_type.
2774 Syntax for the command is : "@var{width_type}"
2778 Syntax for the command is : "@var{width}"
2782 Syntax for the command is : "@var{gain}"
2786 Syntax for the command is : "@var{mix}"
2791 Apply a biquad IIR filter with the given coefficients.
2792 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2793 are the numerator and denominator coefficients respectively.
2794 and @var{channels}, @var{c} specify which channels to filter, by default all
2795 available are filtered.
2797 @subsection Commands
2799 This filter supports the following commands:
2807 Change biquad parameter.
2808 Syntax for the command is : "@var{value}"
2811 How much to use filtered signal in output. Default is 1.
2812 Range is between 0 and 1.
2815 Specify which channels to filter, by default all available are filtered.
2818 Normalize biquad coefficients, by default is disabled.
2819 Enabling it will normalize magnitude response at DC to 0dB.
2823 Bauer stereo to binaural transformation, which improves headphone listening of
2824 stereo audio records.
2826 To enable compilation of this filter you need to configure FFmpeg with
2827 @code{--enable-libbs2b}.
2829 It accepts the following parameters:
2833 Pre-defined crossfeed level.
2837 Default level (fcut=700, feed=50).
2840 Chu Moy circuit (fcut=700, feed=60).
2843 Jan Meier circuit (fcut=650, feed=95).
2848 Cut frequency (in Hz).
2857 Remap input channels to new locations.
2859 It accepts the following parameters:
2862 Map channels from input to output. The argument is a '|'-separated list of
2863 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2864 @var{in_channel} form. @var{in_channel} can be either the name of the input
2865 channel (e.g. FL for front left) or its index in the input channel layout.
2866 @var{out_channel} is the name of the output channel or its index in the output
2867 channel layout. If @var{out_channel} is not given then it is implicitly an
2868 index, starting with zero and increasing by one for each mapping.
2870 @item channel_layout
2871 The channel layout of the output stream.
2874 If no mapping is present, the filter will implicitly map input channels to
2875 output channels, preserving indices.
2877 @subsection Examples
2881 For example, assuming a 5.1+downmix input MOV file,
2883 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2885 will create an output WAV file tagged as stereo from the downmix channels of
2889 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2891 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
2895 @section channelsplit
2897 Split each channel from an input audio stream into a separate output stream.
2899 It accepts the following parameters:
2901 @item channel_layout
2902 The channel layout of the input stream. The default is "stereo".
2904 A channel layout describing the channels to be extracted as separate output streams
2905 or "all" to extract each input channel as a separate stream. The default is "all".
2907 Choosing channels not present in channel layout in the input will result in an error.
2910 @subsection Examples
2914 For example, assuming a stereo input MP3 file,
2916 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
2918 will create an output Matroska file with two audio streams, one containing only
2919 the left channel and the other the right channel.
2922 Split a 5.1 WAV file into per-channel files:
2924 ffmpeg -i in.wav -filter_complex
2925 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
2926 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
2927 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
2932 Extract only LFE from a 5.1 WAV file:
2934 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
2935 -map '[LFE]' lfe.wav
2940 Add a chorus effect to the audio.
2942 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
2944 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
2945 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
2946 The modulation depth defines the range the modulated delay is played before or after
2947 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
2948 sound tuned around the original one, like in a chorus where some vocals are slightly
2951 It accepts the following parameters:
2954 Set input gain. Default is 0.4.
2957 Set output gain. Default is 0.4.
2960 Set delays. A typical delay is around 40ms to 60ms.
2972 @subsection Examples
2978 chorus=0.7:0.9:55:0.4:0.25:2
2984 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
2988 Fuller sounding chorus with three delays:
2990 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
2995 Compress or expand the audio's dynamic range.
2997 It accepts the following parameters:
3003 A list of times in seconds for each channel over which the instantaneous level
3004 of the input signal is averaged to determine its volume. @var{attacks} refers to
3005 increase of volume and @var{decays} refers to decrease of volume. For most
3006 situations, the attack time (response to the audio getting louder) should be
3007 shorter than the decay time, because the human ear is more sensitive to sudden
3008 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3009 a typical value for decay is 0.8 seconds.
3010 If specified number of attacks & decays is lower than number of channels, the last
3011 set attack/decay will be used for all remaining channels.
3014 A list of points for the transfer function, specified in dB relative to the
3015 maximum possible signal amplitude. Each key points list must be defined using
3016 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3017 @code{x0/y0 x1/y1 x2/y2 ....}
3019 The input values must be in strictly increasing order but the transfer function
3020 does not have to be monotonically rising. The point @code{0/0} is assumed but
3021 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3022 function are @code{-70/-70|-60/-20|1/0}.
3025 Set the curve radius in dB for all joints. It defaults to 0.01.
3028 Set the additional gain in dB to be applied at all points on the transfer
3029 function. This allows for easy adjustment of the overall gain.
3033 Set an initial volume, in dB, to be assumed for each channel when filtering
3034 starts. This permits the user to supply a nominal level initially, so that, for
3035 example, a very large gain is not applied to initial signal levels before the
3036 companding has begun to operate. A typical value for audio which is initially
3037 quiet is -90 dB. It defaults to 0.
3040 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3041 delayed before being fed to the volume adjuster. Specifying a delay
3042 approximately equal to the attack/decay times allows the filter to effectively
3043 operate in predictive rather than reactive mode. It defaults to 0.
3047 @subsection Examples
3051 Make music with both quiet and loud passages suitable for listening to in a
3054 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3057 Another example for audio with whisper and explosion parts:
3059 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3063 A noise gate for when the noise is at a lower level than the signal:
3065 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3069 Here is another noise gate, this time for when the noise is at a higher level
3070 than the signal (making it, in some ways, similar to squelch):
3072 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3076 2:1 compression starting at -6dB:
3078 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3082 2:1 compression starting at -9dB:
3084 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3088 2:1 compression starting at -12dB:
3090 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3094 2:1 compression starting at -18dB:
3096 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3100 3:1 compression starting at -15dB:
3102 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3108 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3114 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
3118 Hard limiter at -6dB:
3120 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3124 Hard limiter at -12dB:
3126 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3130 Hard noise gate at -35 dB:
3132 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3138 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3142 @section compensationdelay
3144 Compensation Delay Line is a metric based delay to compensate differing
3145 positions of microphones or speakers.
3147 For example, you have recorded guitar with two microphones placed in
3148 different locations. Because the front of sound wave has fixed speed in
3149 normal conditions, the phasing of microphones can vary and depends on
3150 their location and interposition. The best sound mix can be achieved when
3151 these microphones are in phase (synchronized). Note that a distance of
3152 ~30 cm between microphones makes one microphone capture the signal in
3153 antiphase to the other microphone. That makes the final mix sound moody.
3154 This filter helps to solve phasing problems by adding different delays
3155 to each microphone track and make them synchronized.
3157 The best result can be reached when you take one track as base and
3158 synchronize other tracks one by one with it.
3159 Remember that synchronization/delay tolerance depends on sample rate, too.
3160 Higher sample rates will give more tolerance.
3162 The filter accepts the following parameters:
3166 Set millimeters distance. This is compensation distance for fine tuning.
3170 Set cm distance. This is compensation distance for tightening distance setup.
3174 Set meters distance. This is compensation distance for hard distance setup.
3178 Set dry amount. Amount of unprocessed (dry) signal.
3182 Set wet amount. Amount of processed (wet) signal.
3186 Set temperature in degrees Celsius. This is the temperature of the environment.
3191 Apply headphone crossfeed filter.
3193 Crossfeed is the process of blending the left and right channels of stereo
3195 It is mainly used to reduce extreme stereo separation of low frequencies.
3197 The intent is to produce more speaker like sound to the listener.
3199 The filter accepts the following options:
3203 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3204 This sets gain of low shelf filter for side part of stereo image.
3205 Default is -6dB. Max allowed is -30db when strength is set to 1.
3208 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3209 This sets cut off frequency of low shelf filter. Default is cut off near
3210 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3213 Set input gain. Default is 0.9.
3216 Set output gain. Default is 1.
3219 @section crystalizer
3220 Simple algorithm to expand audio dynamic range.
3222 The filter accepts the following options:
3226 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3227 (unchanged sound) to 10.0 (maximum effect).
3230 Enable clipping. By default is enabled.
3233 @subsection Commands
3235 This filter supports the all above options as @ref{commands}.
3238 Apply a DC shift to the audio.
3240 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3241 in the recording chain) from the audio. The effect of a DC offset is reduced
3242 headroom and hence volume. The @ref{astats} filter can be used to determine if
3243 a signal has a DC offset.
3247 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3251 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3252 used to prevent clipping.
3257 Apply de-essing to the audio samples.
3261 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3265 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3269 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3273 Set the output mode.
3275 It accepts the following values:
3278 Pass input unchanged.
3281 Pass ess filtered out.
3286 Default value is @var{o}.
3292 Measure audio dynamic range.
3294 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3295 is found in transition material. And anything less that 8 have very poor dynamics
3296 and is very compressed.
3298 The filter accepts the following options:
3302 Set window length in seconds used to split audio into segments of equal length.
3303 Default is 3 seconds.
3307 Dynamic Audio Normalizer.
3309 This filter applies a certain amount of gain to the input audio in order
3310 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3311 contrast to more "simple" normalization algorithms, the Dynamic Audio
3312 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3313 This allows for applying extra gain to the "quiet" sections of the audio
3314 while avoiding distortions or clipping the "loud" sections. In other words:
3315 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3316 sections, in the sense that the volume of each section is brought to the
3317 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3318 this goal *without* applying "dynamic range compressing". It will retain 100%
3319 of the dynamic range *within* each section of the audio file.
3323 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3324 Default is 500 milliseconds.
3325 The Dynamic Audio Normalizer processes the input audio in small chunks,
3326 referred to as frames. This is required, because a peak magnitude has no
3327 meaning for just a single sample value. Instead, we need to determine the
3328 peak magnitude for a contiguous sequence of sample values. While a "standard"
3329 normalizer would simply use the peak magnitude of the complete file, the
3330 Dynamic Audio Normalizer determines the peak magnitude individually for each
3331 frame. The length of a frame is specified in milliseconds. By default, the
3332 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3333 been found to give good results with most files.
3334 Note that the exact frame length, in number of samples, will be determined
3335 automatically, based on the sampling rate of the individual input audio file.
3338 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3339 number. Default is 31.
3340 Probably the most important parameter of the Dynamic Audio Normalizer is the
3341 @code{window size} of the Gaussian smoothing filter. The filter's window size
3342 is specified in frames, centered around the current frame. For the sake of
3343 simplicity, this must be an odd number. Consequently, the default value of 31
3344 takes into account the current frame, as well as the 15 preceding frames and
3345 the 15 subsequent frames. Using a larger window results in a stronger
3346 smoothing effect and thus in less gain variation, i.e. slower gain
3347 adaptation. Conversely, using a smaller window results in a weaker smoothing
3348 effect and thus in more gain variation, i.e. faster gain adaptation.
3349 In other words, the more you increase this value, the more the Dynamic Audio
3350 Normalizer will behave like a "traditional" normalization filter. On the
3351 contrary, the more you decrease this value, the more the Dynamic Audio
3352 Normalizer will behave like a dynamic range compressor.
3355 Set the target peak value. This specifies the highest permissible magnitude
3356 level for the normalized audio input. This filter will try to approach the
3357 target peak magnitude as closely as possible, but at the same time it also
3358 makes sure that the normalized signal will never exceed the peak magnitude.
3359 A frame's maximum local gain factor is imposed directly by the target peak
3360 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3361 It is not recommended to go above this value.
3364 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3365 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3366 factor for each input frame, i.e. the maximum gain factor that does not
3367 result in clipping or distortion. The maximum gain factor is determined by
3368 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3369 additionally bounds the frame's maximum gain factor by a predetermined
3370 (global) maximum gain factor. This is done in order to avoid excessive gain
3371 factors in "silent" or almost silent frames. By default, the maximum gain
3372 factor is 10.0, For most inputs the default value should be sufficient and
3373 it usually is not recommended to increase this value. Though, for input
3374 with an extremely low overall volume level, it may be necessary to allow even
3375 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3376 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3377 Instead, a "sigmoid" threshold function will be applied. This way, the
3378 gain factors will smoothly approach the threshold value, but never exceed that
3382 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3383 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3384 This means that the maximum local gain factor for each frame is defined
3385 (only) by the frame's highest magnitude sample. This way, the samples can
3386 be amplified as much as possible without exceeding the maximum signal
3387 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3388 Normalizer can also take into account the frame's root mean square,
3389 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3390 determine the power of a time-varying signal. It is therefore considered
3391 that the RMS is a better approximation of the "perceived loudness" than
3392 just looking at the signal's peak magnitude. Consequently, by adjusting all
3393 frames to a constant RMS value, a uniform "perceived loudness" can be
3394 established. If a target RMS value has been specified, a frame's local gain
3395 factor is defined as the factor that would result in exactly that RMS value.
3396 Note, however, that the maximum local gain factor is still restricted by the
3397 frame's highest magnitude sample, in order to prevent clipping.
3400 Enable channels coupling. By default is enabled.
3401 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3402 amount. This means the same gain factor will be applied to all channels, i.e.
3403 the maximum possible gain factor is determined by the "loudest" channel.
3404 However, in some recordings, it may happen that the volume of the different
3405 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3406 In this case, this option can be used to disable the channel coupling. This way,
3407 the gain factor will be determined independently for each channel, depending
3408 only on the individual channel's highest magnitude sample. This allows for
3409 harmonizing the volume of the different channels.
3412 Enable DC bias correction. By default is disabled.
3413 An audio signal (in the time domain) is a sequence of sample values.
3414 In the Dynamic Audio Normalizer these sample values are represented in the
3415 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3416 audio signal, or "waveform", should be centered around the zero point.
3417 That means if we calculate the mean value of all samples in a file, or in a
3418 single frame, then the result should be 0.0 or at least very close to that
3419 value. If, however, there is a significant deviation of the mean value from
3420 0.0, in either positive or negative direction, this is referred to as a
3421 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3422 Audio Normalizer provides optional DC bias correction.
3423 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3424 the mean value, or "DC correction" offset, of each input frame and subtract
3425 that value from all of the frame's sample values which ensures those samples
3426 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3427 boundaries, the DC correction offset values will be interpolated smoothly
3428 between neighbouring frames.
3430 @item altboundary, b
3431 Enable alternative boundary mode. By default is disabled.
3432 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3433 around each frame. This includes the preceding frames as well as the
3434 subsequent frames. However, for the "boundary" frames, located at the very
3435 beginning and at the very end of the audio file, not all neighbouring
3436 frames are available. In particular, for the first few frames in the audio
3437 file, the preceding frames are not known. And, similarly, for the last few
3438 frames in the audio file, the subsequent frames are not known. Thus, the
3439 question arises which gain factors should be assumed for the missing frames
3440 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3441 to deal with this situation. The default boundary mode assumes a gain factor
3442 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3443 "fade out" at the beginning and at the end of the input, respectively.
3446 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3447 By default, the Dynamic Audio Normalizer does not apply "traditional"
3448 compression. This means that signal peaks will not be pruned and thus the
3449 full dynamic range will be retained within each local neighbourhood. However,
3450 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3451 normalization algorithm with a more "traditional" compression.
3452 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3453 (thresholding) function. If (and only if) the compression feature is enabled,
3454 all input frames will be processed by a soft knee thresholding function prior
3455 to the actual normalization process. Put simply, the thresholding function is
3456 going to prune all samples whose magnitude exceeds a certain threshold value.
3457 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3458 value. Instead, the threshold value will be adjusted for each individual
3460 In general, smaller parameters result in stronger compression, and vice versa.
3461 Values below 3.0 are not recommended, because audible distortion may appear.
3464 Set the target threshold value. This specifies the lowest permissible
3465 magnitude level for the audio input which will be normalized.
3466 If input frame volume is above this value frame will be normalized.
3467 Otherwise frame may not be normalized at all. The default value is set
3468 to 0, which means all input frames will be normalized.
3469 This option is mostly useful if digital noise is not wanted to be amplified.
3472 @subsection Commands
3474 This filter supports the all above options as @ref{commands}.
3478 Make audio easier to listen to on headphones.
3480 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3481 so that when listened to on headphones the stereo image is moved from
3482 inside your head (standard for headphones) to outside and in front of
3483 the listener (standard for speakers).
3489 Apply a two-pole peaking equalisation (EQ) filter. With this
3490 filter, the signal-level at and around a selected frequency can
3491 be increased or decreased, whilst (unlike bandpass and bandreject
3492 filters) that at all other frequencies is unchanged.
3494 In order to produce complex equalisation curves, this filter can
3495 be given several times, each with a different central frequency.
3497 The filter accepts the following options:
3501 Set the filter's central frequency in Hz.
3504 Set method to specify band-width of filter.
3519 Specify the band-width of a filter in width_type units.
3522 Set the required gain or attenuation in dB.
3523 Beware of clipping when using a positive gain.
3526 How much to use filtered signal in output. Default is 1.
3527 Range is between 0 and 1.
3530 Specify which channels to filter, by default all available are filtered.
3533 Normalize biquad coefficients, by default is disabled.
3534 Enabling it will normalize magnitude response at DC to 0dB.
3537 @subsection Examples
3540 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3542 equalizer=f=1000:t=h:width=200:g=-10
3546 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3548 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3552 @subsection Commands
3554 This filter supports the following commands:
3557 Change equalizer frequency.
3558 Syntax for the command is : "@var{frequency}"
3561 Change equalizer width_type.
3562 Syntax for the command is : "@var{width_type}"
3565 Change equalizer width.
3566 Syntax for the command is : "@var{width}"
3569 Change equalizer gain.
3570 Syntax for the command is : "@var{gain}"
3573 Change equalizer mix.
3574 Syntax for the command is : "@var{mix}"
3577 @section extrastereo
3579 Linearly increases the difference between left and right channels which
3580 adds some sort of "live" effect to playback.
3582 The filter accepts the following options:
3586 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3587 (average of both channels), with 1.0 sound will be unchanged, with
3588 -1.0 left and right channels will be swapped.
3591 Enable clipping. By default is enabled.
3594 @subsection Commands
3596 This filter supports the all above options as @ref{commands}.
3598 @section firequalizer
3599 Apply FIR Equalization using arbitrary frequency response.
3601 The filter accepts the following option:
3605 Set gain curve equation (in dB). The expression can contain variables:
3608 the evaluated frequency
3612 channel number, set to 0 when multichannels evaluation is disabled
3614 channel id, see libavutil/channel_layout.h, set to the first channel id when
3615 multichannels evaluation is disabled
3619 channel_layout, see libavutil/channel_layout.h
3624 @item gain_interpolate(f)
3625 interpolate gain on frequency f based on gain_entry
3626 @item cubic_interpolate(f)
3627 same as gain_interpolate, but smoother
3629 This option is also available as command. Default is @code{gain_interpolate(f)}.
3632 Set gain entry for gain_interpolate function. The expression can
3636 store gain entry at frequency f with value g
3638 This option is also available as command.
3641 Set filter delay in seconds. Higher value means more accurate.
3642 Default is @code{0.01}.
3645 Set filter accuracy in Hz. Lower value means more accurate.
3646 Default is @code{5}.
3649 Set window function. Acceptable values are:
3652 rectangular window, useful when gain curve is already smooth
3654 hann window (default)
3660 3-terms continuous 1st derivative nuttall window
3662 minimum 3-terms discontinuous nuttall window
3664 4-terms continuous 1st derivative nuttall window
3666 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3668 blackman-harris window
3674 If enabled, use fixed number of audio samples. This improves speed when
3675 filtering with large delay. Default is disabled.
3678 Enable multichannels evaluation on gain. Default is disabled.
3681 Enable zero phase mode by subtracting timestamp to compensate delay.
3682 Default is disabled.
3685 Set scale used by gain. Acceptable values are:
3688 linear frequency, linear gain
3690 linear frequency, logarithmic (in dB) gain (default)
3692 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3694 logarithmic frequency, logarithmic gain
3698 Set file for dumping, suitable for gnuplot.
3701 Set scale for dumpfile. Acceptable values are same with scale option.
3705 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3706 Default is disabled.
3709 Enable minimum phase impulse response. Default is disabled.
3712 @subsection Examples
3717 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3720 lowpass at 1000 Hz with gain_entry:
3722 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3725 custom equalization:
3727 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3730 higher delay with zero phase to compensate delay:
3732 firequalizer=delay=0.1:fixed=on:zero_phase=on
3735 lowpass on left channel, highpass on right channel:
3737 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
3738 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
3743 Apply a flanging effect to the audio.
3745 The filter accepts the following options:
3749 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
3752 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
3755 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
3759 Set percentage of delayed signal mixed with original. Range from 0 to 100.
3760 Default value is 71.
3763 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
3766 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
3767 Default value is @var{sinusoidal}.
3770 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
3771 Default value is 25.
3774 Set delay-line interpolation, @var{linear} or @var{quadratic}.
3775 Default is @var{linear}.
3779 Apply Haas effect to audio.
3781 Note that this makes most sense to apply on mono signals.
3782 With this filter applied to mono signals it give some directionality and
3783 stretches its stereo image.
3785 The filter accepts the following options:
3789 Set input level. By default is @var{1}, or 0dB
3792 Set output level. By default is @var{1}, or 0dB.
3795 Set gain applied to side part of signal. By default is @var{1}.
3798 Set kind of middle source. Can be one of the following:
3808 Pick middle part signal of stereo image.
3811 Pick side part signal of stereo image.
3815 Change middle phase. By default is disabled.
3818 Set left channel delay. By default is @var{2.05} milliseconds.
3821 Set left channel balance. By default is @var{-1}.
3824 Set left channel gain. By default is @var{1}.
3827 Change left phase. By default is disabled.
3830 Set right channel delay. By defaults is @var{2.12} milliseconds.
3833 Set right channel balance. By default is @var{1}.
3836 Set right channel gain. By default is @var{1}.
3839 Change right phase. By default is enabled.
3844 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3845 embedded HDCD codes is expanded into a 20-bit PCM stream.
3847 The filter supports the Peak Extend and Low-level Gain Adjustment features
3848 of HDCD, and detects the Transient Filter flag.
3851 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3854 When using the filter with wav, note the default encoding for wav is 16-bit,
3855 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3856 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3858 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3859 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3862 The filter accepts the following options:
3865 @item disable_autoconvert
3866 Disable any automatic format conversion or resampling in the filter graph.
3868 @item process_stereo
3869 Process the stereo channels together. If target_gain does not match between
3870 channels, consider it invalid and use the last valid target_gain.
3873 Set the code detect timer period in ms.
3876 Always extend peaks above -3dBFS even if PE isn't signaled.
3879 Replace audio with a solid tone and adjust the amplitude to signal some
3880 specific aspect of the decoding process. The output file can be loaded in
3881 an audio editor alongside the original to aid analysis.
3883 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
3890 Gain adjustment level at each sample
3892 Samples where peak extend occurs
3894 Samples where the code detect timer is active
3896 Samples where the target gain does not match between channels
3902 Apply head-related transfer functions (HRTFs) to create virtual
3903 loudspeakers around the user for binaural listening via headphones.
3904 The HRIRs are provided via additional streams, for each channel
3905 one stereo input stream is needed.
3907 The filter accepts the following options:
3911 Set mapping of input streams for convolution.
3912 The argument is a '|'-separated list of channel names in order as they
3913 are given as additional stream inputs for filter.
3914 This also specify number of input streams. Number of input streams
3915 must be not less than number of channels in first stream plus one.
3918 Set gain applied to audio. Value is in dB. Default is 0.
3921 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
3922 processing audio in time domain which is slow.
3923 @var{freq} is processing audio in frequency domain which is fast.
3924 Default is @var{freq}.
3927 Set custom gain for LFE channels. Value is in dB. Default is 0.
3930 Set size of frame in number of samples which will be processed at once.
3931 Default value is @var{1024}. Allowed range is from 1024 to 96000.
3934 Set format of hrir stream.
3935 Default value is @var{stereo}. Alternative value is @var{multich}.
3936 If value is set to @var{stereo}, number of additional streams should
3937 be greater or equal to number of input channels in first input stream.
3938 Also each additional stream should have stereo number of channels.
3939 If value is set to @var{multich}, number of additional streams should
3940 be exactly one. Also number of input channels of additional stream
3941 should be equal or greater than twice number of channels of first input
3945 @subsection Examples
3949 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3950 each amovie filter use stereo file with IR coefficients as input.
3951 The files give coefficients for each position of virtual loudspeaker:
3954 -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"
3959 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3960 but now in @var{multich} @var{hrir} format.
3962 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"
3969 Apply a high-pass filter with 3dB point frequency.
3970 The filter can be either single-pole, or double-pole (the default).
3971 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3973 The filter accepts the following options:
3977 Set frequency in Hz. Default is 3000.
3980 Set number of poles. Default is 2.
3983 Set method to specify band-width of filter.
3998 Specify the band-width of a filter in width_type units.
3999 Applies only to double-pole filter.
4000 The default is 0.707q and gives a Butterworth response.
4003 How much to use filtered signal in output. Default is 1.
4004 Range is between 0 and 1.
4007 Specify which channels to filter, by default all available are filtered.
4010 Normalize biquad coefficients, by default is disabled.
4011 Enabling it will normalize magnitude response at DC to 0dB.
4014 @subsection Commands
4016 This filter supports the following commands:
4019 Change highpass frequency.
4020 Syntax for the command is : "@var{frequency}"
4023 Change highpass width_type.
4024 Syntax for the command is : "@var{width_type}"
4027 Change highpass width.
4028 Syntax for the command is : "@var{width}"
4031 Change highpass mix.
4032 Syntax for the command is : "@var{mix}"
4037 Join multiple input streams into one multi-channel stream.
4039 It accepts the following parameters:
4043 The number of input streams. It defaults to 2.
4045 @item channel_layout
4046 The desired output channel layout. It defaults to stereo.
4049 Map channels from inputs to output. The argument is a '|'-separated list of
4050 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4051 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4052 can be either the name of the input channel (e.g. FL for front left) or its
4053 index in the specified input stream. @var{out_channel} is the name of the output
4057 The filter will attempt to guess the mappings when they are not specified
4058 explicitly. It does so by first trying to find an unused matching input channel
4059 and if that fails it picks the first unused input channel.
4061 Join 3 inputs (with properly set channel layouts):
4063 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4066 Build a 5.1 output from 6 single-channel streams:
4068 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4069 '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'
4075 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4077 To enable compilation of this filter you need to configure FFmpeg with
4078 @code{--enable-ladspa}.
4082 Specifies the name of LADSPA plugin library to load. If the environment
4083 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4084 each one of the directories specified by the colon separated list in
4085 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4086 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4087 @file{/usr/lib/ladspa/}.
4090 Specifies the plugin within the library. Some libraries contain only
4091 one plugin, but others contain many of them. If this is not set filter
4092 will list all available plugins within the specified library.
4095 Set the '|' separated list of controls which are zero or more floating point
4096 values that determine the behavior of the loaded plugin (for example delay,
4098 Controls need to be defined using the following syntax:
4099 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4100 @var{valuei} is the value set on the @var{i}-th control.
4101 Alternatively they can be also defined using the following syntax:
4102 @var{value0}|@var{value1}|@var{value2}|..., where
4103 @var{valuei} is the value set on the @var{i}-th control.
4104 If @option{controls} is set to @code{help}, all available controls and
4105 their valid ranges are printed.
4107 @item sample_rate, s
4108 Specify the sample rate, default to 44100. Only used if plugin have
4112 Set the number of samples per channel per each output frame, default
4113 is 1024. Only used if plugin have zero inputs.
4116 Set the minimum duration of the sourced audio. See
4117 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4118 for the accepted syntax.
4119 Note that the resulting duration may be greater than the specified duration,
4120 as the generated audio is always cut at the end of a complete frame.
4121 If not specified, or the expressed duration is negative, the audio is
4122 supposed to be generated forever.
4123 Only used if plugin have zero inputs.
4127 @subsection Examples
4131 List all available plugins within amp (LADSPA example plugin) library:
4137 List all available controls and their valid ranges for @code{vcf_notch}
4138 plugin from @code{VCF} library:
4140 ladspa=f=vcf:p=vcf_notch:c=help
4144 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4147 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4151 Add reverberation to the audio using TAP-plugins
4152 (Tom's Audio Processing plugins):
4154 ladspa=file=tap_reverb:tap_reverb
4158 Generate white noise, with 0.2 amplitude:
4160 ladspa=file=cmt:noise_source_white:c=c0=.2
4164 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4165 @code{C* Audio Plugin Suite} (CAPS) library:
4167 ladspa=file=caps:Click:c=c1=20'
4171 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4173 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4177 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4178 @code{SWH Plugins} collection:
4180 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4184 Attenuate low frequencies using Multiband EQ from Steve Harris
4185 @code{SWH Plugins} collection:
4187 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4191 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4194 ladspa=caps:Narrower
4198 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4200 ladspa=caps:White:.2
4204 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4206 ladspa=caps:Fractal:c=c1=1
4210 Dynamic volume normalization using @code{VLevel} plugin:
4212 ladspa=vlevel-ladspa:vlevel_mono
4216 @subsection Commands
4218 This filter supports the following commands:
4221 Modify the @var{N}-th control value.
4223 If the specified value is not valid, it is ignored and prior one is kept.
4228 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4229 Support for both single pass (livestreams, files) and double pass (files) modes.
4230 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4231 detect true peaks, the audio stream will be upsampled to 192 kHz.
4232 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4234 The filter accepts the following options:
4238 Set integrated loudness target.
4239 Range is -70.0 - -5.0. Default value is -24.0.
4242 Set loudness range target.
4243 Range is 1.0 - 20.0. Default value is 7.0.
4246 Set maximum true peak.
4247 Range is -9.0 - +0.0. Default value is -2.0.
4249 @item measured_I, measured_i
4250 Measured IL of input file.
4251 Range is -99.0 - +0.0.
4253 @item measured_LRA, measured_lra
4254 Measured LRA of input file.
4255 Range is 0.0 - 99.0.
4257 @item measured_TP, measured_tp
4258 Measured true peak of input file.
4259 Range is -99.0 - +99.0.
4261 @item measured_thresh
4262 Measured threshold of input file.
4263 Range is -99.0 - +0.0.
4266 Set offset gain. Gain is applied before the true-peak limiter.
4267 Range is -99.0 - +99.0. Default is +0.0.
4270 Normalize by linearly scaling the source audio.
4271 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4272 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4273 be lower than source LRA and the change in integrated loudness shouldn't
4274 result in a true peak which exceeds the target TP. If any of these
4275 conditions aren't met, normalization mode will revert to @var{dynamic}.
4276 Options are @code{true} or @code{false}. Default is @code{true}.
4279 Treat mono input files as "dual-mono". If a mono file is intended for playback
4280 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4281 If set to @code{true}, this option will compensate for this effect.
4282 Multi-channel input files are not affected by this option.
4283 Options are true or false. Default is false.
4286 Set print format for stats. Options are summary, json, or none.
4287 Default value is none.
4292 Apply a low-pass filter with 3dB point frequency.
4293 The filter can be either single-pole or double-pole (the default).
4294 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4296 The filter accepts the following options:
4300 Set frequency in Hz. Default is 500.
4303 Set number of poles. Default is 2.
4306 Set method to specify band-width of filter.
4321 Specify the band-width of a filter in width_type units.
4322 Applies only to double-pole filter.
4323 The default is 0.707q and gives a Butterworth response.
4326 How much to use filtered signal in output. Default is 1.
4327 Range is between 0 and 1.
4330 Specify which channels to filter, by default all available are filtered.
4333 Normalize biquad coefficients, by default is disabled.
4334 Enabling it will normalize magnitude response at DC to 0dB.
4337 @subsection Examples
4340 Lowpass only LFE channel, it LFE is not present it does nothing:
4346 @subsection Commands
4348 This filter supports the following commands:
4351 Change lowpass frequency.
4352 Syntax for the command is : "@var{frequency}"
4355 Change lowpass width_type.
4356 Syntax for the command is : "@var{width_type}"
4359 Change lowpass width.
4360 Syntax for the command is : "@var{width}"
4364 Syntax for the command is : "@var{mix}"
4369 Load a LV2 (LADSPA Version 2) plugin.
4371 To enable compilation of this filter you need to configure FFmpeg with
4372 @code{--enable-lv2}.
4376 Specifies the plugin URI. You may need to escape ':'.
4379 Set the '|' separated list of controls which are zero or more floating point
4380 values that determine the behavior of the loaded plugin (for example delay,
4382 If @option{controls} is set to @code{help}, all available controls and
4383 their valid ranges are printed.
4385 @item sample_rate, s
4386 Specify the sample rate, default to 44100. Only used if plugin have
4390 Set the number of samples per channel per each output frame, default
4391 is 1024. Only used if plugin have zero inputs.
4394 Set the minimum duration of the sourced audio. See
4395 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4396 for the accepted syntax.
4397 Note that the resulting duration may be greater than the specified duration,
4398 as the generated audio is always cut at the end of a complete frame.
4399 If not specified, or the expressed duration is negative, the audio is
4400 supposed to be generated forever.
4401 Only used if plugin have zero inputs.
4404 @subsection Examples
4408 Apply bass enhancer plugin from Calf:
4410 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4414 Apply vinyl plugin from Calf:
4416 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4420 Apply bit crusher plugin from ArtyFX:
4422 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4427 Multiband Compress or expand the audio's dynamic range.
4429 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4430 This is akin to the crossover of a loudspeaker, and results in flat frequency
4431 response when absent compander action.
4433 It accepts the following parameters:
4437 This option syntax is:
4438 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4439 For explanation of each item refer to compand filter documentation.
4445 Mix channels with specific gain levels. The filter accepts the output
4446 channel layout followed by a set of channels definitions.
4448 This filter is also designed to efficiently remap the channels of an audio
4451 The filter accepts parameters of the form:
4452 "@var{l}|@var{outdef}|@var{outdef}|..."
4456 output channel layout or number of channels
4459 output channel specification, of the form:
4460 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4463 output channel to define, either a channel name (FL, FR, etc.) or a channel
4464 number (c0, c1, etc.)
4467 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4470 input channel to use, see out_name for details; it is not possible to mix
4471 named and numbered input channels
4474 If the `=' in a channel specification is replaced by `<', then the gains for
4475 that specification will be renormalized so that the total is 1, thus
4476 avoiding clipping noise.
4478 @subsection Mixing examples
4480 For example, if you want to down-mix from stereo to mono, but with a bigger
4481 factor for the left channel:
4483 pan=1c|c0=0.9*c0+0.1*c1
4486 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4487 7-channels surround:
4489 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4492 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4493 that should be preferred (see "-ac" option) unless you have very specific
4496 @subsection Remapping examples
4498 The channel remapping will be effective if, and only if:
4501 @item gain coefficients are zeroes or ones,
4502 @item only one input per channel output,
4505 If all these conditions are satisfied, the filter will notify the user ("Pure
4506 channel mapping detected"), and use an optimized and lossless method to do the
4509 For example, if you have a 5.1 source and want a stereo audio stream by
4510 dropping the extra channels:
4512 pan="stereo| c0=FL | c1=FR"
4515 Given the same source, you can also switch front left and front right channels
4516 and keep the input channel layout:
4518 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4521 If the input is a stereo audio stream, you can mute the front left channel (and
4522 still keep the stereo channel layout) with:
4527 Still with a stereo audio stream input, you can copy the right channel in both
4528 front left and right:
4530 pan="stereo| c0=FR | c1=FR"
4535 ReplayGain scanner filter. This filter takes an audio stream as an input and
4536 outputs it unchanged.
4537 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4541 Convert the audio sample format, sample rate and channel layout. It is
4542 not meant to be used directly.
4545 Apply time-stretching and pitch-shifting with librubberband.
4547 To enable compilation of this filter, you need to configure FFmpeg with
4548 @code{--enable-librubberband}.
4550 The filter accepts the following options:
4554 Set tempo scale factor.
4557 Set pitch scale factor.
4560 Set transients detector.
4561 Possible values are:
4570 Possible values are:
4579 Possible values are:
4586 Set processing window size.
4587 Possible values are:
4596 Possible values are:
4603 Enable formant preservation when shift pitching.
4604 Possible values are:
4612 Possible values are:
4621 Possible values are:
4628 @subsection Commands
4630 This filter supports the following commands:
4633 Change filter tempo scale factor.
4634 Syntax for the command is : "@var{tempo}"
4637 Change filter pitch scale factor.
4638 Syntax for the command is : "@var{pitch}"
4641 @section sidechaincompress
4643 This filter acts like normal compressor but has the ability to compress
4644 detected signal using second input signal.
4645 It needs two input streams and returns one output stream.
4646 First input stream will be processed depending on second stream signal.
4647 The filtered signal then can be filtered with other filters in later stages of
4648 processing. See @ref{pan} and @ref{amerge} filter.
4650 The filter accepts the following options:
4654 Set input gain. Default is 1. Range is between 0.015625 and 64.
4657 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4658 Default is @code{downward}.
4661 If a signal of second stream raises above this level it will affect the gain
4662 reduction of first stream.
4663 By default is 0.125. Range is between 0.00097563 and 1.
4666 Set a ratio about which the signal is reduced. 1:2 means that if the level
4667 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4668 Default is 2. Range is between 1 and 20.
4671 Amount of milliseconds the signal has to rise above the threshold before gain
4672 reduction starts. Default is 20. Range is between 0.01 and 2000.
4675 Amount of milliseconds the signal has to fall below the threshold before
4676 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4679 Set the amount by how much signal will be amplified after processing.
4680 Default is 1. Range is from 1 to 64.
4683 Curve the sharp knee around the threshold to enter gain reduction more softly.
4684 Default is 2.82843. Range is between 1 and 8.
4687 Choose if the @code{average} level between all channels of side-chain stream
4688 or the louder(@code{maximum}) channel of side-chain stream affects the
4689 reduction. Default is @code{average}.
4692 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4693 of @code{rms}. Default is @code{rms} which is mainly smoother.
4696 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4699 How much to use compressed signal in output. Default is 1.
4700 Range is between 0 and 1.
4703 @subsection Commands
4705 This filter supports the all above options as @ref{commands}.
4707 @subsection Examples
4711 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
4712 depending on the signal of 2nd input and later compressed signal to be
4713 merged with 2nd input:
4715 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
4719 @section sidechaingate
4721 A sidechain gate acts like a normal (wideband) gate but has the ability to
4722 filter the detected signal before sending it to the gain reduction stage.
4723 Normally a gate uses the full range signal to detect a level above the
4725 For example: If you cut all lower frequencies from your sidechain signal
4726 the gate will decrease the volume of your track only if not enough highs
4727 appear. With this technique you are able to reduce the resonation of a
4728 natural drum or remove "rumbling" of muted strokes from a heavily distorted
4730 It needs two input streams and returns one output stream.
4731 First input stream will be processed depending on second stream signal.
4733 The filter accepts the following options:
4737 Set input level before filtering.
4738 Default is 1. Allowed range is from 0.015625 to 64.
4741 Set the mode of operation. Can be @code{upward} or @code{downward}.
4742 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
4743 will be amplified, expanding dynamic range in upward direction.
4744 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
4747 Set the level of gain reduction when the signal is below the threshold.
4748 Default is 0.06125. Allowed range is from 0 to 1.
4749 Setting this to 0 disables reduction and then filter behaves like expander.
4752 If a signal rises above this level the gain reduction is released.
4753 Default is 0.125. Allowed range is from 0 to 1.
4756 Set a ratio about which the signal is reduced.
4757 Default is 2. Allowed range is from 1 to 9000.
4760 Amount of milliseconds the signal has to rise above the threshold before gain
4762 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
4765 Amount of milliseconds the signal has to fall below the threshold before the
4766 reduction is increased again. Default is 250 milliseconds.
4767 Allowed range is from 0.01 to 9000.
4770 Set amount of amplification of signal after processing.
4771 Default is 1. Allowed range is from 1 to 64.
4774 Curve the sharp knee around the threshold to enter gain reduction more softly.
4775 Default is 2.828427125. Allowed range is from 1 to 8.
4778 Choose if exact signal should be taken for detection or an RMS like one.
4779 Default is rms. Can be peak or rms.
4782 Choose if the average level between all channels or the louder channel affects
4784 Default is average. Can be average or maximum.
4787 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
4790 @section silencedetect
4792 Detect silence in an audio stream.
4794 This filter logs a message when it detects that the input audio volume is less
4795 or equal to a noise tolerance value for a duration greater or equal to the
4796 minimum detected noise duration.
4798 The printed times and duration are expressed in seconds. The
4799 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
4800 is set on the first frame whose timestamp equals or exceeds the detection
4801 duration and it contains the timestamp of the first frame of the silence.
4803 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
4804 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
4805 keys are set on the first frame after the silence. If @option{mono} is
4806 enabled, and each channel is evaluated separately, the @code{.X}
4807 suffixed keys are used, and @code{X} corresponds to the channel number.
4809 The filter accepts the following options:
4813 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4814 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4817 Set silence duration until notification (default is 2 seconds). See
4818 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4819 for the accepted syntax.
4822 Process each channel separately, instead of combined. By default is disabled.
4825 @subsection Examples
4829 Detect 5 seconds of silence with -50dB noise tolerance:
4831 silencedetect=n=-50dB:d=5
4835 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4836 tolerance in @file{silence.mp3}:
4838 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4842 @section silenceremove
4844 Remove silence from the beginning, middle or end of the audio.
4846 The filter accepts the following options:
4850 This value is used to indicate if audio should be trimmed at beginning of
4851 the audio. A value of zero indicates no silence should be trimmed from the
4852 beginning. When specifying a non-zero value, it trims audio up until it
4853 finds non-silence. Normally, when trimming silence from beginning of audio
4854 the @var{start_periods} will be @code{1} but it can be increased to higher
4855 values to trim all audio up to specific count of non-silence periods.
4856 Default value is @code{0}.
4858 @item start_duration
4859 Specify the amount of time that non-silence must be detected before it stops
4860 trimming audio. By increasing the duration, bursts of noises can be treated
4861 as silence and trimmed off. Default value is @code{0}.
4863 @item start_threshold
4864 This indicates what sample value should be treated as silence. For digital
4865 audio, a value of @code{0} may be fine but for audio recorded from analog,
4866 you may wish to increase the value to account for background noise.
4867 Can be specified in dB (in case "dB" is appended to the specified value)
4868 or amplitude ratio. Default value is @code{0}.
4871 Specify max duration of silence at beginning that will be kept after
4872 trimming. Default is 0, which is equal to trimming all samples detected
4876 Specify mode of detection of silence end in start of multi-channel audio.
4877 Can be @var{any} or @var{all}. Default is @var{any}.
4878 With @var{any}, any sample that is detected as non-silence will cause
4879 stopped trimming of silence.
4880 With @var{all}, only if all channels are detected as non-silence will cause
4881 stopped trimming of silence.
4884 Set the count for trimming silence from the end of audio.
4885 To remove silence from the middle of a file, specify a @var{stop_periods}
4886 that is negative. This value is then treated as a positive value and is
4887 used to indicate the effect should restart processing as specified by
4888 @var{start_periods}, making it suitable for removing periods of silence
4889 in the middle of the audio.
4890 Default value is @code{0}.
4893 Specify a duration of silence that must exist before audio is not copied any
4894 more. By specifying a higher duration, silence that is wanted can be left in
4896 Default value is @code{0}.
4898 @item stop_threshold
4899 This is the same as @option{start_threshold} but for trimming silence from
4901 Can be specified in dB (in case "dB" is appended to the specified value)
4902 or amplitude ratio. Default value is @code{0}.
4905 Specify max duration of silence at end that will be kept after
4906 trimming. Default is 0, which is equal to trimming all samples detected
4910 Specify mode of detection of silence start in end of multi-channel audio.
4911 Can be @var{any} or @var{all}. Default is @var{any}.
4912 With @var{any}, any sample that is detected as non-silence will cause
4913 stopped trimming of silence.
4914 With @var{all}, only if all channels are detected as non-silence will cause
4915 stopped trimming of silence.
4918 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4919 and works better with digital silence which is exactly 0.
4920 Default value is @code{rms}.
4923 Set duration in number of seconds used to calculate size of window in number
4924 of samples for detecting silence.
4925 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4928 @subsection Examples
4932 The following example shows how this filter can be used to start a recording
4933 that does not contain the delay at the start which usually occurs between
4934 pressing the record button and the start of the performance:
4936 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
4940 Trim all silence encountered from beginning to end where there is more than 1
4941 second of silence in audio:
4943 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
4947 Trim all digital silence samples, using peak detection, from beginning to end
4948 where there is more than 0 samples of digital silence in audio and digital
4949 silence is detected in all channels at same positions in stream:
4951 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
4957 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4958 loudspeakers around the user for binaural listening via headphones (audio
4959 formats up to 9 channels supported).
4960 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4961 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4962 Austrian Academy of Sciences.
4964 To enable compilation of this filter you need to configure FFmpeg with
4965 @code{--enable-libmysofa}.
4967 The filter accepts the following options:
4971 Set the SOFA file used for rendering.
4974 Set gain applied to audio. Value is in dB. Default is 0.
4977 Set rotation of virtual loudspeakers in deg. Default is 0.
4980 Set elevation of virtual speakers in deg. Default is 0.
4983 Set distance in meters between loudspeakers and the listener with near-field
4984 HRTFs. Default is 1.
4987 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4988 processing audio in time domain which is slow.
4989 @var{freq} is processing audio in frequency domain which is fast.
4990 Default is @var{freq}.
4993 Set custom positions of virtual loudspeakers. Syntax for this option is:
4994 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
4995 Each virtual loudspeaker is described with short channel name following with
4996 azimuth and elevation in degrees.
4997 Each virtual loudspeaker description is separated by '|'.
4998 For example to override front left and front right channel positions use:
4999 'speakers=FL 45 15|FR 345 15'.
5000 Descriptions with unrecognised channel names are ignored.
5003 Set custom gain for LFE channels. Value is in dB. Default is 0.
5006 Set custom frame size in number of samples. Default is 1024.
5007 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5008 is set to @var{freq}.
5011 Should all IRs be normalized upon importing SOFA file.
5012 By default is enabled.
5015 Should nearest IRs be interpolated with neighbor IRs if exact position
5016 does not match. By default is disabled.
5019 Minphase all IRs upon loading of SOFA file. By default is disabled.
5022 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5025 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5028 @subsection Examples
5032 Using ClubFritz6 sofa file:
5034 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5038 Using ClubFritz12 sofa file and bigger radius with small rotation:
5040 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5044 Similar as above but with custom speaker positions for front left, front right, back left and back right
5045 and also with custom gain:
5047 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5051 @section stereotools
5053 This filter has some handy utilities to manage stereo signals, for converting
5054 M/S stereo recordings to L/R signal while having control over the parameters
5055 or spreading the stereo image of master track.
5057 The filter accepts the following options:
5061 Set input level before filtering for both channels. Defaults is 1.
5062 Allowed range is from 0.015625 to 64.
5065 Set output level after filtering for both channels. Defaults is 1.
5066 Allowed range is from 0.015625 to 64.
5069 Set input balance between both channels. Default is 0.
5070 Allowed range is from -1 to 1.
5073 Set output balance between both channels. Default is 0.
5074 Allowed range is from -1 to 1.
5077 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5078 clipping. Disabled by default.
5081 Mute the left channel. Disabled by default.
5084 Mute the right channel. Disabled by default.
5087 Change the phase of the left channel. Disabled by default.
5090 Change the phase of the right channel. Disabled by default.
5093 Set stereo mode. Available values are:
5097 Left/Right to Left/Right, this is default.
5100 Left/Right to Mid/Side.
5103 Mid/Side to Left/Right.
5106 Left/Right to Left/Left.
5109 Left/Right to Right/Right.
5112 Left/Right to Left + Right.
5115 Left/Right to Right/Left.
5118 Mid/Side to Left/Left.
5121 Mid/Side to Right/Right.
5125 Set level of side signal. Default is 1.
5126 Allowed range is from 0.015625 to 64.
5129 Set balance of side signal. Default is 0.
5130 Allowed range is from -1 to 1.
5133 Set level of the middle signal. Default is 1.
5134 Allowed range is from 0.015625 to 64.
5137 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5140 Set stereo base between mono and inversed channels. Default is 0.
5141 Allowed range is from -1 to 1.
5144 Set delay in milliseconds how much to delay left from right channel and
5145 vice versa. Default is 0. Allowed range is from -20 to 20.
5148 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5151 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5153 @item bmode_in, bmode_out
5154 Set balance mode for balance_in/balance_out option.
5156 Can be one of the following:
5160 Classic balance mode. Attenuate one channel at time.
5161 Gain is raised up to 1.
5164 Similar as classic mode above but gain is raised up to 2.
5167 Equal power distribution, from -6dB to +6dB range.
5171 @subsection Examples
5175 Apply karaoke like effect:
5177 stereotools=mlev=0.015625
5181 Convert M/S signal to L/R:
5183 "stereotools=mode=ms>lr"
5187 @section stereowiden
5189 This filter enhance the stereo effect by suppressing signal common to both
5190 channels and by delaying the signal of left into right and vice versa,
5191 thereby widening the stereo effect.
5193 The filter accepts the following options:
5197 Time in milliseconds of the delay of left signal into right and vice versa.
5198 Default is 20 milliseconds.
5201 Amount of gain in delayed signal into right and vice versa. Gives a delay
5202 effect of left signal in right output and vice versa which gives widening
5203 effect. Default is 0.3.
5206 Cross feed of left into right with inverted phase. This helps in suppressing
5207 the mono. If the value is 1 it will cancel all the signal common to both
5208 channels. Default is 0.3.
5211 Set level of input signal of original channel. Default is 0.8.
5214 @subsection Commands
5216 This filter supports the all above options except @code{delay} as @ref{commands}.
5218 @section superequalizer
5219 Apply 18 band equalizer.
5221 The filter accepts the following options:
5228 Set 131Hz band gain.
5230 Set 185Hz band gain.
5232 Set 262Hz band gain.
5234 Set 370Hz band gain.
5236 Set 523Hz band gain.
5238 Set 740Hz band gain.
5240 Set 1047Hz band gain.
5242 Set 1480Hz band gain.
5244 Set 2093Hz band gain.
5246 Set 2960Hz band gain.
5248 Set 4186Hz band gain.
5250 Set 5920Hz band gain.
5252 Set 8372Hz band gain.
5254 Set 11840Hz band gain.
5256 Set 16744Hz band gain.
5258 Set 20000Hz band gain.
5262 Apply audio surround upmix filter.
5264 This filter allows to produce multichannel output from audio stream.
5266 The filter accepts the following options:
5270 Set output channel layout. By default, this is @var{5.1}.
5272 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5273 for the required syntax.
5276 Set input channel layout. By default, this is @var{stereo}.
5278 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5279 for the required syntax.
5282 Set input volume level. By default, this is @var{1}.
5285 Set output volume level. By default, this is @var{1}.
5288 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5291 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5294 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5297 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5298 In @var{add} mode, LFE channel is created from input audio and added to output.
5299 In @var{sub} mode, LFE channel is created from input audio and added to output but
5300 also all non-LFE output channels are subtracted with output LFE channel.
5303 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5304 Default is @var{90}.
5307 Set front center input volume. By default, this is @var{1}.
5310 Set front center output volume. By default, this is @var{1}.
5313 Set front left input volume. By default, this is @var{1}.
5316 Set front left output volume. By default, this is @var{1}.
5319 Set front right input volume. By default, this is @var{1}.
5322 Set front right output volume. By default, this is @var{1}.
5325 Set side left input volume. By default, this is @var{1}.
5328 Set side left output volume. By default, this is @var{1}.
5331 Set side right input volume. By default, this is @var{1}.
5334 Set side right output volume. By default, this is @var{1}.
5337 Set back left input volume. By default, this is @var{1}.
5340 Set back left output volume. By default, this is @var{1}.
5343 Set back right input volume. By default, this is @var{1}.
5346 Set back right output volume. By default, this is @var{1}.
5349 Set back center input volume. By default, this is @var{1}.
5352 Set back center output volume. By default, this is @var{1}.
5355 Set LFE input volume. By default, this is @var{1}.
5358 Set LFE output volume. By default, this is @var{1}.
5361 Set spread usage of stereo image across X axis for all channels.
5364 Set spread usage of stereo image across Y axis for all channels.
5366 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5367 Set spread usage of stereo image across X axis for each channel.
5369 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5370 Set spread usage of stereo image across Y axis for each channel.
5373 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5376 Set window function.
5378 It accepts the following values:
5401 Default is @code{hann}.
5404 Set window overlap. If set to 1, the recommended overlap for selected
5405 window function will be picked. Default is @code{0.5}.
5408 @section treble, highshelf
5410 Boost or cut treble (upper) frequencies of the audio using a two-pole
5411 shelving filter with a response similar to that of a standard
5412 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5414 The filter accepts the following options:
5418 Give the gain at whichever is the lower of ~22 kHz and the
5419 Nyquist frequency. Its useful range is about -20 (for a large cut)
5420 to +20 (for a large boost). Beware of clipping when using a positive gain.
5423 Set the filter's central frequency and so can be used
5424 to extend or reduce the frequency range to be boosted or cut.
5425 The default value is @code{3000} Hz.
5428 Set method to specify band-width of filter.
5443 Determine how steep is the filter's shelf transition.
5446 How much to use filtered signal in output. Default is 1.
5447 Range is between 0 and 1.
5450 Specify which channels to filter, by default all available are filtered.
5453 Normalize biquad coefficients, by default is disabled.
5454 Enabling it will normalize magnitude response at DC to 0dB.
5457 @subsection Commands
5459 This filter supports the following commands:
5462 Change treble frequency.
5463 Syntax for the command is : "@var{frequency}"
5466 Change treble width_type.
5467 Syntax for the command is : "@var{width_type}"
5470 Change treble width.
5471 Syntax for the command is : "@var{width}"
5475 Syntax for the command is : "@var{gain}"
5479 Syntax for the command is : "@var{mix}"
5484 Sinusoidal amplitude modulation.
5486 The filter accepts the following options:
5490 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5491 (20 Hz or lower) will result in a tremolo effect.
5492 This filter may also be used as a ring modulator by specifying
5493 a modulation frequency higher than 20 Hz.
5494 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5497 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5498 Default value is 0.5.
5503 Sinusoidal phase modulation.
5505 The filter accepts the following options:
5509 Modulation frequency in Hertz.
5510 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5513 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5514 Default value is 0.5.
5519 Adjust the input audio volume.
5521 It accepts the following parameters:
5525 Set audio volume expression.
5527 Output values are clipped to the maximum value.
5529 The output audio volume is given by the relation:
5531 @var{output_volume} = @var{volume} * @var{input_volume}
5534 The default value for @var{volume} is "1.0".
5537 This parameter represents the mathematical precision.
5539 It determines which input sample formats will be allowed, which affects the
5540 precision of the volume scaling.
5544 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5546 32-bit floating-point; this limits input sample format to FLT. (default)
5548 64-bit floating-point; this limits input sample format to DBL.
5552 Choose the behaviour on encountering ReplayGain side data in input frames.
5556 Remove ReplayGain side data, ignoring its contents (the default).
5559 Ignore ReplayGain side data, but leave it in the frame.
5562 Prefer the track gain, if present.
5565 Prefer the album gain, if present.
5568 @item replaygain_preamp
5569 Pre-amplification gain in dB to apply to the selected replaygain gain.
5571 Default value for @var{replaygain_preamp} is 0.0.
5573 @item replaygain_noclip
5574 Prevent clipping by limiting the gain applied.
5576 Default value for @var{replaygain_noclip} is 1.
5579 Set when the volume expression is evaluated.
5581 It accepts the following values:
5584 only evaluate expression once during the filter initialization, or
5585 when the @samp{volume} command is sent
5588 evaluate expression for each incoming frame
5591 Default value is @samp{once}.
5594 The volume expression can contain the following parameters.
5598 frame number (starting at zero)
5601 @item nb_consumed_samples
5602 number of samples consumed by the filter
5604 number of samples in the current frame
5606 original frame position in the file
5612 PTS at start of stream
5614 time at start of stream
5620 last set volume value
5623 Note that when @option{eval} is set to @samp{once} only the
5624 @var{sample_rate} and @var{tb} variables are available, all other
5625 variables will evaluate to NAN.
5627 @subsection Commands
5629 This filter supports the following commands:
5632 Modify the volume expression.
5633 The command accepts the same syntax of the corresponding option.
5635 If the specified expression is not valid, it is kept at its current
5639 @subsection Examples
5643 Halve the input audio volume:
5647 volume=volume=-6.0206dB
5650 In all the above example the named key for @option{volume} can be
5651 omitted, for example like in:
5657 Increase input audio power by 6 decibels using fixed-point precision:
5659 volume=volume=6dB:precision=fixed
5663 Fade volume after time 10 with an annihilation period of 5 seconds:
5665 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
5669 @section volumedetect
5671 Detect the volume of the input video.
5673 The filter has no parameters. The input is not modified. Statistics about
5674 the volume will be printed in the log when the input stream end is reached.
5676 In particular it will show the mean volume (root mean square), maximum
5677 volume (on a per-sample basis), and the beginning of a histogram of the
5678 registered volume values (from the maximum value to a cumulated 1/1000 of
5681 All volumes are in decibels relative to the maximum PCM value.
5683 @subsection Examples
5685 Here is an excerpt of the output:
5687 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
5688 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
5689 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
5690 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5691 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5692 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5693 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5694 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5695 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5701 The mean square energy is approximately -27 dB, or 10^-2.7.
5703 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5705 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5708 In other words, raising the volume by +4 dB does not cause any clipping,
5709 raising it by +5 dB causes clipping for 6 samples, etc.
5711 @c man end AUDIO FILTERS
5713 @chapter Audio Sources
5714 @c man begin AUDIO SOURCES
5716 Below is a description of the currently available audio sources.
5720 Buffer audio frames, and make them available to the filter chain.
5722 This source is mainly intended for a programmatic use, in particular
5723 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
5725 It accepts the following parameters:
5729 The timebase which will be used for timestamps of submitted frames. It must be
5730 either a floating-point number or in @var{numerator}/@var{denominator} form.
5733 The sample rate of the incoming audio buffers.
5736 The sample format of the incoming audio buffers.
5737 Either a sample format name or its corresponding integer representation from
5738 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5740 @item channel_layout
5741 The channel layout of the incoming audio buffers.
5742 Either a channel layout name from channel_layout_map in
5743 @file{libavutil/channel_layout.c} or its corresponding integer representation
5744 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5747 The number of channels of the incoming audio buffers.
5748 If both @var{channels} and @var{channel_layout} are specified, then they
5753 @subsection Examples
5756 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5759 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5760 Since the sample format with name "s16p" corresponds to the number
5761 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5764 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5769 Generate an audio signal specified by an expression.
5771 This source accepts in input one or more expressions (one for each
5772 channel), which are evaluated and used to generate a corresponding
5775 This source accepts the following options:
5779 Set the '|'-separated expressions list for each separate channel. In case the
5780 @option{channel_layout} option is not specified, the selected channel layout
5781 depends on the number of provided expressions. Otherwise the last
5782 specified expression is applied to the remaining output channels.
5784 @item channel_layout, c
5785 Set the channel layout. The number of channels in the specified layout
5786 must be equal to the number of specified expressions.
5789 Set the minimum duration of the sourced audio. See
5790 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5791 for the accepted syntax.
5792 Note that the resulting duration may be greater than the specified
5793 duration, as the generated audio is always cut at the end of a
5796 If not specified, or the expressed duration is negative, the audio is
5797 supposed to be generated forever.
5800 Set the number of samples per channel per each output frame,
5803 @item sample_rate, s
5804 Specify the sample rate, default to 44100.
5807 Each expression in @var{exprs} can contain the following constants:
5811 number of the evaluated sample, starting from 0
5814 time of the evaluated sample expressed in seconds, starting from 0
5821 @subsection Examples
5831 Generate a sin signal with frequency of 440 Hz, set sample rate to
5834 aevalsrc="sin(440*2*PI*t):s=8000"
5838 Generate a two channels signal, specify the channel layout (Front
5839 Center + Back Center) explicitly:
5841 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5845 Generate white noise:
5847 aevalsrc="-2+random(0)"
5851 Generate an amplitude modulated signal:
5853 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
5857 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
5859 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
5866 Generate a FIR coefficients using frequency sampling method.
5868 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
5870 The filter accepts the following options:
5874 Set number of filter coefficents in output audio stream.
5875 Default value is 1025.
5878 Set frequency points from where magnitude and phase are set.
5879 This must be in non decreasing order, and first element must be 0, while last element
5880 must be 1. Elements are separated by white spaces.
5883 Set magnitude value for every frequency point set by @option{frequency}.
5884 Number of values must be same as number of frequency points.
5885 Values are separated by white spaces.
5888 Set phase value for every frequency point set by @option{frequency}.
5889 Number of values must be same as number of frequency points.
5890 Values are separated by white spaces.
5892 @item sample_rate, r
5893 Set sample rate, default is 44100.
5896 Set number of samples per each frame. Default is 1024.
5899 Set window function. Default is blackman.
5904 The null audio source, return unprocessed audio frames. It is mainly useful
5905 as a template and to be employed in analysis / debugging tools, or as
5906 the source for filters which ignore the input data (for example the sox
5909 This source accepts the following options:
5913 @item channel_layout, cl
5915 Specifies the channel layout, and can be either an integer or a string
5916 representing a channel layout. The default value of @var{channel_layout}
5919 Check the channel_layout_map definition in
5920 @file{libavutil/channel_layout.c} for the mapping between strings and
5921 channel layout values.
5923 @item sample_rate, r
5924 Specifies the sample rate, and defaults to 44100.
5927 Set the number of samples per requested frames.
5931 @subsection Examples
5935 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
5937 anullsrc=r=48000:cl=4
5941 Do the same operation with a more obvious syntax:
5943 anullsrc=r=48000:cl=mono
5947 All the parameters need to be explicitly defined.
5951 Synthesize a voice utterance using the libflite library.
5953 To enable compilation of this filter you need to configure FFmpeg with
5954 @code{--enable-libflite}.
5956 Note that versions of the flite library prior to 2.0 are not thread-safe.
5958 The filter accepts the following options:
5963 If set to 1, list the names of the available voices and exit
5964 immediately. Default value is 0.
5967 Set the maximum number of samples per frame. Default value is 512.
5970 Set the filename containing the text to speak.
5973 Set the text to speak.
5976 Set the voice to use for the speech synthesis. Default value is
5977 @code{kal}. See also the @var{list_voices} option.
5980 @subsection Examples
5984 Read from file @file{speech.txt}, and synthesize the text using the
5985 standard flite voice:
5987 flite=textfile=speech.txt
5991 Read the specified text selecting the @code{slt} voice:
5993 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5997 Input text to ffmpeg:
5999 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6003 Make @file{ffplay} speak the specified text, using @code{flite} and
6004 the @code{lavfi} device:
6006 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6010 For more information about libflite, check:
6011 @url{http://www.festvox.org/flite/}
6015 Generate a noise audio signal.
6017 The filter accepts the following options:
6020 @item sample_rate, r
6021 Specify the sample rate. Default value is 48000 Hz.
6024 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6028 Specify the duration of the generated audio stream. Not specifying this option
6029 results in noise with an infinite length.
6031 @item color, colour, c
6032 Specify the color of noise. Available noise colors are white, pink, brown,
6033 blue, violet and velvet. Default color is white.
6036 Specify a value used to seed the PRNG.
6039 Set the number of samples per each output frame, default is 1024.
6042 @subsection Examples
6047 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6049 anoisesrc=d=60:c=pink:r=44100:a=0.5
6055 Generate odd-tap Hilbert transform FIR coefficients.
6057 The resulting stream can be used with @ref{afir} filter for phase-shifting
6058 the signal by 90 degrees.
6060 This is used in many matrix coding schemes and for analytic signal generation.
6061 The process is often written as a multiplication by i (or j), the imaginary unit.
6063 The filter accepts the following options:
6067 @item sample_rate, s
6068 Set sample rate, default is 44100.
6071 Set length of FIR filter, default is 22051.
6074 Set number of samples per each frame.
6077 Set window function to be used when generating FIR coefficients.
6082 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6084 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6086 The filter accepts the following options:
6089 @item sample_rate, r
6090 Set sample rate, default is 44100.
6093 Set number of samples per each frame. Default is 1024.
6096 Set high-pass frequency. Default is 0.
6099 Set low-pass frequency. Default is 0.
6100 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6101 is higher than 0 then filter will create band-pass filter coefficients,
6102 otherwise band-reject filter coefficients.
6105 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6108 Set Kaiser window beta.
6111 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6114 Enable rounding, by default is disabled.
6117 Set number of taps for high-pass filter.
6120 Set number of taps for low-pass filter.
6125 Generate an audio signal made of a sine wave with amplitude 1/8.
6127 The audio signal is bit-exact.
6129 The filter accepts the following options:
6134 Set the carrier frequency. Default is 440 Hz.
6136 @item beep_factor, b
6137 Enable a periodic beep every second with frequency @var{beep_factor} times
6138 the carrier frequency. Default is 0, meaning the beep is disabled.
6140 @item sample_rate, r
6141 Specify the sample rate, default is 44100.
6144 Specify the duration of the generated audio stream.
6146 @item samples_per_frame
6147 Set the number of samples per output frame.
6149 The expression can contain the following constants:
6153 The (sequential) number of the output audio frame, starting from 0.
6156 The PTS (Presentation TimeStamp) of the output audio frame,
6157 expressed in @var{TB} units.
6160 The PTS of the output audio frame, expressed in seconds.
6163 The timebase of the output audio frames.
6166 Default is @code{1024}.
6169 @subsection Examples
6174 Generate a simple 440 Hz sine wave:
6180 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6184 sine=frequency=220:beep_factor=4:duration=5
6188 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6191 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6195 @c man end AUDIO SOURCES
6197 @chapter Audio Sinks
6198 @c man begin AUDIO SINKS
6200 Below is a description of the currently available audio sinks.
6202 @section abuffersink
6204 Buffer audio frames, and make them available to the end of filter chain.
6206 This sink is mainly intended for programmatic use, in particular
6207 through the interface defined in @file{libavfilter/buffersink.h}
6208 or the options system.
6210 It accepts a pointer to an AVABufferSinkContext structure, which
6211 defines the incoming buffers' formats, to be passed as the opaque
6212 parameter to @code{avfilter_init_filter} for initialization.
6215 Null audio sink; do absolutely nothing with the input audio. It is
6216 mainly useful as a template and for use in analysis / debugging
6219 @c man end AUDIO SINKS
6221 @chapter Video Filters
6222 @c man begin VIDEO FILTERS
6224 When you configure your FFmpeg build, you can disable any of the
6225 existing filters using @code{--disable-filters}.
6226 The configure output will show the video filters included in your
6229 Below is a description of the currently available video filters.
6233 Mark a region of interest in a video frame.
6235 The frame data is passed through unchanged, but metadata is attached
6236 to the frame indicating regions of interest which can affect the
6237 behaviour of later encoding. Multiple regions can be marked by
6238 applying the filter multiple times.
6242 Region distance in pixels from the left edge of the frame.
6244 Region distance in pixels from the top edge of the frame.
6246 Region width in pixels.
6248 Region height in pixels.
6250 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6251 and may contain the following variables:
6254 Width of the input frame.
6256 Height of the input frame.
6260 Quantisation offset to apply within the region.
6262 This must be a real value in the range -1 to +1. A value of zero
6263 indicates no quality change. A negative value asks for better quality
6264 (less quantisation), while a positive value asks for worse quality
6265 (greater quantisation).
6267 The range is calibrated so that the extreme values indicate the
6268 largest possible offset - if the rest of the frame is encoded with the
6269 worst possible quality, an offset of -1 indicates that this region
6270 should be encoded with the best possible quality anyway. Intermediate
6271 values are then interpolated in some codec-dependent way.
6273 For example, in 10-bit H.264 the quantisation parameter varies between
6274 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6275 this region should be encoded with a QP around one-tenth of the full
6276 range better than the rest of the frame. So, if most of the frame
6277 were to be encoded with a QP of around 30, this region would get a QP
6278 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6279 An extreme value of -1 would indicate that this region should be
6280 encoded with the best possible quality regardless of the treatment of
6281 the rest of the frame - that is, should be encoded at a QP of -12.
6283 If set to true, remove any existing regions of interest marked on the
6284 frame before adding the new one.
6287 @subsection Examples
6291 Mark the centre quarter of the frame as interesting.
6293 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6296 Mark the 100-pixel-wide region on the left edge of the frame as very
6297 uninteresting (to be encoded at much lower quality than the rest of
6300 addroi=0:0:100:ih:+1/5
6304 @section alphaextract
6306 Extract the alpha component from the input as a grayscale video. This
6307 is especially useful with the @var{alphamerge} filter.
6311 Add or replace the alpha component of the primary input with the
6312 grayscale value of a second input. This is intended for use with
6313 @var{alphaextract} to allow the transmission or storage of frame
6314 sequences that have alpha in a format that doesn't support an alpha
6317 For example, to reconstruct full frames from a normal YUV-encoded video
6318 and a separate video created with @var{alphaextract}, you might use:
6320 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6323 Since this filter is designed for reconstruction, it operates on frame
6324 sequences without considering timestamps, and terminates when either
6325 input reaches end of stream. This will cause problems if your encoding
6326 pipeline drops frames. If you're trying to apply an image as an
6327 overlay to a video stream, consider the @var{overlay} filter instead.
6331 Amplify differences between current pixel and pixels of adjacent frames in
6332 same pixel location.
6334 This filter accepts the following options:
6338 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6339 For example radius of 3 will instruct filter to calculate average of 7 frames.
6342 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6345 Set threshold for difference amplification. Any difference greater or equal to
6346 this value will not alter source pixel. Default is 10.
6347 Allowed range is from 0 to 65535.
6350 Set tolerance for difference amplification. Any difference lower to
6351 this value will not alter source pixel. Default is 0.
6352 Allowed range is from 0 to 65535.
6355 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6356 This option controls maximum possible value that will decrease source pixel value.
6359 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6360 This option controls maximum possible value that will increase source pixel value.
6363 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6366 @subsection Commands
6368 This filter supports the following @ref{commands} that corresponds to option of same name:
6380 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6381 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6382 Substation Alpha) subtitles files.
6384 This filter accepts the following option in addition to the common options from
6385 the @ref{subtitles} filter:
6389 Set the shaping engine
6391 Available values are:
6394 The default libass shaping engine, which is the best available.
6396 Fast, font-agnostic shaper that can do only substitutions
6398 Slower shaper using OpenType for substitutions and positioning
6401 The default is @code{auto}.
6405 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6407 The filter accepts the following options:
6411 Set threshold A for 1st plane. Default is 0.02.
6412 Valid range is 0 to 0.3.
6415 Set threshold B for 1st plane. Default is 0.04.
6416 Valid range is 0 to 5.
6419 Set threshold A for 2nd plane. Default is 0.02.
6420 Valid range is 0 to 0.3.
6423 Set threshold B for 2nd plane. Default is 0.04.
6424 Valid range is 0 to 5.
6427 Set threshold A for 3rd plane. Default is 0.02.
6428 Valid range is 0 to 0.3.
6431 Set threshold B for 3rd plane. Default is 0.04.
6432 Valid range is 0 to 5.
6434 Threshold A is designed to react on abrupt changes in the input signal and
6435 threshold B is designed to react on continuous changes in the input signal.
6438 Set number of frames filter will use for averaging. Default is 9. Must be odd
6439 number in range [5, 129].
6442 Set what planes of frame filter will use for averaging. Default is all.
6445 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
6446 Alternatively can be set to @code{s} serial.
6448 Parallel can be faster then serial, while other way around is never true.
6449 Parallel will abort early on first change being greater then thresholds, while serial
6450 will continue processing other side of frames if they are equal or bellow thresholds.
6453 @subsection Commands
6454 This filter supports same @ref{commands} as options except option @code{s}.
6455 The command accepts the same syntax of the corresponding option.
6459 Apply average blur filter.
6461 The filter accepts the following options:
6465 Set horizontal radius size.
6468 Set which planes to filter. By default all planes are filtered.
6471 Set vertical radius size, if zero it will be same as @code{sizeX}.
6472 Default is @code{0}.
6475 @subsection Commands
6476 This filter supports same commands as options.
6477 The command accepts the same syntax of the corresponding option.
6479 If the specified expression is not valid, it is kept at its current
6484 Compute the bounding box for the non-black pixels in the input frame
6487 This filter computes the bounding box containing all the pixels with a
6488 luminance value greater than the minimum allowed value.
6489 The parameters describing the bounding box are printed on the filter
6492 The filter accepts the following option:
6496 Set the minimal luminance value. Default is @code{16}.
6500 Apply bilateral filter, spatial smoothing while preserving edges.
6502 The filter accepts the following options:
6505 Set sigma of gaussian function to calculate spatial weight.
6506 Allowed range is 0 to 10. Default is 0.1.
6509 Set sigma of gaussian function to calculate range weight.
6510 Allowed range is 0 to 1. Default is 0.1.
6513 Set planes to filter. Default is first only.
6516 @section bitplanenoise
6518 Show and measure bit plane noise.
6520 The filter accepts the following options:
6524 Set which plane to analyze. Default is @code{1}.
6527 Filter out noisy pixels from @code{bitplane} set above.
6528 Default is disabled.
6531 @section blackdetect
6533 Detect video intervals that are (almost) completely black. Can be
6534 useful to detect chapter transitions, commercials, or invalid
6537 The filter outputs its detection analysis to both the log as well as
6538 frame metadata. If a black segment of at least the specified minimum
6539 duration is found, a line with the start and end timestamps as well
6540 as duration is printed to the log with level @code{info}. In addition,
6541 a log line with level @code{debug} is printed per frame showing the
6542 black amount detected for that frame.
6544 The filter also attaches metadata to the first frame of a black
6545 segment with key @code{lavfi.black_start} and to the first frame
6546 after the black segment ends with key @code{lavfi.black_end}. The
6547 value is the frame's timestamp. This metadata is added regardless
6548 of the minimum duration specified.
6550 The filter accepts the following options:
6553 @item black_min_duration, d
6554 Set the minimum detected black duration expressed in seconds. It must
6555 be a non-negative floating point number.
6557 Default value is 2.0.
6559 @item picture_black_ratio_th, pic_th
6560 Set the threshold for considering a picture "black".
6561 Express the minimum value for the ratio:
6563 @var{nb_black_pixels} / @var{nb_pixels}
6566 for which a picture is considered black.
6567 Default value is 0.98.
6569 @item pixel_black_th, pix_th
6570 Set the threshold for considering a pixel "black".
6572 The threshold expresses the maximum pixel luminance value for which a
6573 pixel is considered "black". The provided value is scaled according to
6574 the following equation:
6576 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6579 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6580 the input video format, the range is [0-255] for YUV full-range
6581 formats and [16-235] for YUV non full-range formats.
6583 Default value is 0.10.
6586 The following example sets the maximum pixel threshold to the minimum
6587 value, and detects only black intervals of 2 or more seconds:
6589 blackdetect=d=2:pix_th=0.00
6594 Detect frames that are (almost) completely black. Can be useful to
6595 detect chapter transitions or commercials. Output lines consist of
6596 the frame number of the detected frame, the percentage of blackness,
6597 the position in the file if known or -1 and the timestamp in seconds.
6599 In order to display the output lines, you need to set the loglevel at
6600 least to the AV_LOG_INFO value.
6602 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
6603 The value represents the percentage of pixels in the picture that
6604 are below the threshold value.
6606 It accepts the following parameters:
6611 The percentage of the pixels that have to be below the threshold; it defaults to
6614 @item threshold, thresh
6615 The threshold below which a pixel value is considered black; it defaults to
6623 Blend two video frames into each other.
6625 The @code{blend} filter takes two input streams and outputs one
6626 stream, the first input is the "top" layer and second input is
6627 "bottom" layer. By default, the output terminates when the longest input terminates.
6629 The @code{tblend} (time blend) filter takes two consecutive frames
6630 from one single stream, and outputs the result obtained by blending
6631 the new frame on top of the old frame.
6633 A description of the accepted options follows.
6641 Set blend mode for specific pixel component or all pixel components in case
6642 of @var{all_mode}. Default value is @code{normal}.
6644 Available values for component modes are:
6686 Set blend opacity for specific pixel component or all pixel components in case
6687 of @var{all_opacity}. Only used in combination with pixel component blend modes.
6694 Set blend expression for specific pixel component or all pixel components in case
6695 of @var{all_expr}. Note that related mode options will be ignored if those are set.
6697 The expressions can use the following variables:
6701 The sequential number of the filtered frame, starting from @code{0}.
6705 the coordinates of the current sample
6709 the width and height of currently filtered plane
6713 Width and height scale for the plane being filtered. It is the
6714 ratio between the dimensions of the current plane to the luma plane,
6715 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
6716 the luma plane and @code{0.5,0.5} for the chroma planes.
6719 Time of the current frame, expressed in seconds.
6722 Value of pixel component at current location for first video frame (top layer).
6725 Value of pixel component at current location for second video frame (bottom layer).
6729 The @code{blend} filter also supports the @ref{framesync} options.
6731 @subsection Examples
6735 Apply transition from bottom layer to top layer in first 10 seconds:
6737 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
6741 Apply linear horizontal transition from top layer to bottom layer:
6743 blend=all_expr='A*(X/W)+B*(1-X/W)'
6747 Apply 1x1 checkerboard effect:
6749 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
6753 Apply uncover left effect:
6755 blend=all_expr='if(gte(N*SW+X,W),A,B)'
6759 Apply uncover down effect:
6761 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
6765 Apply uncover up-left effect:
6767 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
6771 Split diagonally video and shows top and bottom layer on each side:
6773 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
6777 Display differences between the current and the previous frame:
6779 tblend=all_mode=grainextract
6785 Denoise frames using Block-Matching 3D algorithm.
6787 The filter accepts the following options.
6791 Set denoising strength. Default value is 1.
6792 Allowed range is from 0 to 999.9.
6793 The denoising algorithm is very sensitive to sigma, so adjust it
6794 according to the source.
6797 Set local patch size. This sets dimensions in 2D.
6800 Set sliding step for processing blocks. Default value is 4.
6801 Allowed range is from 1 to 64.
6802 Smaller values allows processing more reference blocks and is slower.
6805 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
6806 When set to 1, no block matching is done. Larger values allows more blocks
6808 Allowed range is from 1 to 256.
6811 Set radius for search block matching. Default is 9.
6812 Allowed range is from 1 to INT32_MAX.
6815 Set step between two search locations for block matching. Default is 1.
6816 Allowed range is from 1 to 64. Smaller is slower.
6819 Set threshold of mean square error for block matching. Valid range is 0 to
6823 Set thresholding parameter for hard thresholding in 3D transformed domain.
6824 Larger values results in stronger hard-thresholding filtering in frequency
6828 Set filtering estimation mode. Can be @code{basic} or @code{final}.
6829 Default is @code{basic}.
6832 If enabled, filter will use 2nd stream for block matching.
6833 Default is disabled for @code{basic} value of @var{estim} option,
6834 and always enabled if value of @var{estim} is @code{final}.
6837 Set planes to filter. Default is all available except alpha.
6840 @subsection Examples
6844 Basic filtering with bm3d:
6846 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
6850 Same as above, but filtering only luma:
6852 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
6856 Same as above, but with both estimation modes:
6858 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
6862 Same as above, but prefilter with @ref{nlmeans} filter instead:
6864 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
6870 Apply a boxblur algorithm to the input video.
6872 It accepts the following parameters:
6876 @item luma_radius, lr
6877 @item luma_power, lp
6878 @item chroma_radius, cr
6879 @item chroma_power, cp
6880 @item alpha_radius, ar
6881 @item alpha_power, ap
6885 A description of the accepted options follows.
6888 @item luma_radius, lr
6889 @item chroma_radius, cr
6890 @item alpha_radius, ar
6891 Set an expression for the box radius in pixels used for blurring the
6892 corresponding input plane.
6894 The radius value must be a non-negative number, and must not be
6895 greater than the value of the expression @code{min(w,h)/2} for the
6896 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
6899 Default value for @option{luma_radius} is "2". If not specified,
6900 @option{chroma_radius} and @option{alpha_radius} default to the
6901 corresponding value set for @option{luma_radius}.
6903 The expressions can contain the following constants:
6907 The input width and height in pixels.
6911 The input chroma image width and height in pixels.
6915 The horizontal and vertical chroma subsample values. For example, for the
6916 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
6919 @item luma_power, lp
6920 @item chroma_power, cp
6921 @item alpha_power, ap
6922 Specify how many times the boxblur filter is applied to the
6923 corresponding plane.
6925 Default value for @option{luma_power} is 2. If not specified,
6926 @option{chroma_power} and @option{alpha_power} default to the
6927 corresponding value set for @option{luma_power}.
6929 A value of 0 will disable the effect.
6932 @subsection Examples
6936 Apply a boxblur filter with the luma, chroma, and alpha radii
6939 boxblur=luma_radius=2:luma_power=1
6944 Set the luma radius to 2, and alpha and chroma radius to 0:
6946 boxblur=2:1:cr=0:ar=0
6950 Set the luma and chroma radii to a fraction of the video dimension:
6952 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
6958 Deinterlace the input video ("bwdif" stands for "Bob Weaver
6959 Deinterlacing Filter").
6961 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
6962 interpolation algorithms.
6963 It accepts the following parameters:
6967 The interlacing mode to adopt. It accepts one of the following values:
6971 Output one frame for each frame.
6973 Output one frame for each field.
6976 The default value is @code{send_field}.
6979 The picture field parity assumed for the input interlaced video. It accepts one
6980 of the following values:
6984 Assume the top field is first.
6986 Assume the bottom field is first.
6988 Enable automatic detection of field parity.
6991 The default value is @code{auto}.
6992 If the interlacing is unknown or the decoder does not export this information,
6993 top field first will be assumed.
6996 Specify which frames to deinterlace. Accepts one of the following
7001 Deinterlace all frames.
7003 Only deinterlace frames marked as interlaced.
7006 The default value is @code{all}.
7011 Apply Contrast Adaptive Sharpen filter to video stream.
7013 The filter accepts the following options:
7017 Set the sharpening strength. Default value is 0.
7020 Set planes to filter. Default value is to filter all
7021 planes except alpha plane.
7025 Remove all color information for all colors except for certain one.
7027 The filter accepts the following options:
7031 The color which will not be replaced with neutral chroma.
7034 Similarity percentage with the above color.
7035 0.01 matches only the exact key color, while 1.0 matches everything.
7039 0.0 makes pixels either fully gray, or not gray at all.
7040 Higher values result in more preserved color.
7043 Signals that the color passed is already in YUV instead of RGB.
7045 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7046 This can be used to pass exact YUV values as hexadecimal numbers.
7049 @subsection Commands
7050 This filter supports same @ref{commands} as options.
7051 The command accepts the same syntax of the corresponding option.
7053 If the specified expression is not valid, it is kept at its current
7057 YUV colorspace color/chroma keying.
7059 The filter accepts the following options:
7063 The color which will be replaced with transparency.
7066 Similarity percentage with the key color.
7068 0.01 matches only the exact key color, while 1.0 matches everything.
7073 0.0 makes pixels either fully transparent, or not transparent at all.
7075 Higher values result in semi-transparent pixels, with a higher transparency
7076 the more similar the pixels color is to the key color.
7079 Signals that the color passed is already in YUV instead of RGB.
7081 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7082 This can be used to pass exact YUV values as hexadecimal numbers.
7085 @subsection Commands
7086 This filter supports same @ref{commands} as options.
7087 The command accepts the same syntax of the corresponding option.
7089 If the specified expression is not valid, it is kept at its current
7092 @subsection Examples
7096 Make every green pixel in the input image transparent:
7098 ffmpeg -i input.png -vf chromakey=green out.png
7102 Overlay a greenscreen-video on top of a static black background.
7104 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
7108 @section chromashift
7109 Shift chroma pixels horizontally and/or vertically.
7111 The filter accepts the following options:
7114 Set amount to shift chroma-blue horizontally.
7116 Set amount to shift chroma-blue vertically.
7118 Set amount to shift chroma-red horizontally.
7120 Set amount to shift chroma-red vertically.
7122 Set edge mode, can be @var{smear}, default, or @var{warp}.
7125 @subsection Commands
7127 This filter supports the all above options as @ref{commands}.
7131 Display CIE color diagram with pixels overlaid onto it.
7133 The filter accepts the following options:
7148 @item uhdtv, rec2020
7162 Set what gamuts to draw.
7164 See @code{system} option for available values.
7167 Set ciescope size, by default set to 512.
7170 Set intensity used to map input pixel values to CIE diagram.
7173 Set contrast used to draw tongue colors that are out of active color system gamut.
7176 Correct gamma displayed on scope, by default enabled.
7179 Show white point on CIE diagram, by default disabled.
7182 Set input gamma. Used only with XYZ input color space.
7187 Visualize information exported by some codecs.
7189 Some codecs can export information through frames using side-data or other
7190 means. For example, some MPEG based codecs export motion vectors through the
7191 @var{export_mvs} flag in the codec @option{flags2} option.
7193 The filter accepts the following option:
7197 Set motion vectors to visualize.
7199 Available flags for @var{mv} are:
7203 forward predicted MVs of P-frames
7205 forward predicted MVs of B-frames
7207 backward predicted MVs of B-frames
7211 Display quantization parameters using the chroma planes.
7214 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7216 Available flags for @var{mv_type} are:
7220 forward predicted MVs
7222 backward predicted MVs
7225 @item frame_type, ft
7226 Set frame type to visualize motion vectors of.
7228 Available flags for @var{frame_type} are:
7232 intra-coded frames (I-frames)
7234 predicted frames (P-frames)
7236 bi-directionally predicted frames (B-frames)
7240 @subsection Examples
7244 Visualize forward predicted MVs of all frames using @command{ffplay}:
7246 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7250 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7252 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7256 @section colorbalance
7257 Modify intensity of primary colors (red, green and blue) of input frames.
7259 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7260 regions for the red-cyan, green-magenta or blue-yellow balance.
7262 A positive adjustment value shifts the balance towards the primary color, a negative
7263 value towards the complementary color.
7265 The filter accepts the following options:
7271 Adjust red, green and blue shadows (darkest pixels).
7276 Adjust red, green and blue midtones (medium pixels).
7281 Adjust red, green and blue highlights (brightest pixels).
7283 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7286 Preserve lightness when changing color balance. Default is disabled.
7289 @subsection Examples
7293 Add red color cast to shadows:
7299 @subsection Commands
7301 This filter supports the all above options as @ref{commands}.
7303 @section colorchannelmixer
7305 Adjust video input frames by re-mixing color channels.
7307 This filter modifies a color channel by adding the values associated to
7308 the other channels of the same pixels. For example if the value to
7309 modify is red, the output value will be:
7311 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7314 The filter accepts the following options:
7321 Adjust contribution of input red, green, blue and alpha channels for output red channel.
7322 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
7328 Adjust contribution of input red, green, blue and alpha channels for output green channel.
7329 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
7335 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
7336 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
7342 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7343 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7345 Allowed ranges for options are @code{[-2.0, 2.0]}.
7348 @subsection Examples
7352 Convert source to grayscale:
7354 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7357 Simulate sepia tones:
7359 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7363 @subsection Commands
7365 This filter supports the all above options as @ref{commands}.
7368 RGB colorspace color keying.
7370 The filter accepts the following options:
7374 The color which will be replaced with transparency.
7377 Similarity percentage with the key color.
7379 0.01 matches only the exact key color, while 1.0 matches everything.
7384 0.0 makes pixels either fully transparent, or not transparent at all.
7386 Higher values result in semi-transparent pixels, with a higher transparency
7387 the more similar the pixels color is to the key color.
7390 @subsection Examples
7394 Make every green pixel in the input image transparent:
7396 ffmpeg -i input.png -vf colorkey=green out.png
7400 Overlay a greenscreen-video on top of a static background image.
7402 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
7406 @subsection Commands
7407 This filter supports same @ref{commands} as options.
7408 The command accepts the same syntax of the corresponding option.
7410 If the specified expression is not valid, it is kept at its current
7414 Remove all color information for all RGB colors except for certain one.
7416 The filter accepts the following options:
7420 The color which will not be replaced with neutral gray.
7423 Similarity percentage with the above color.
7424 0.01 matches only the exact key color, while 1.0 matches everything.
7427 Blend percentage. 0.0 makes pixels fully gray.
7428 Higher values result in more preserved color.
7431 @subsection Commands
7432 This filter supports same @ref{commands} as options.
7433 The command accepts the same syntax of the corresponding option.
7435 If the specified expression is not valid, it is kept at its current
7438 @section colorlevels
7440 Adjust video input frames using levels.
7442 The filter accepts the following options:
7449 Adjust red, green, blue and alpha input black point.
7450 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7456 Adjust red, green, blue and alpha input white point.
7457 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
7459 Input levels are used to lighten highlights (bright tones), darken shadows
7460 (dark tones), change the balance of bright and dark tones.
7466 Adjust red, green, blue and alpha output black point.
7467 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
7473 Adjust red, green, blue and alpha output white point.
7474 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
7476 Output levels allows manual selection of a constrained output level range.
7479 @subsection Examples
7483 Make video output darker:
7485 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
7491 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
7495 Make video output lighter:
7497 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
7501 Increase brightness:
7503 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
7507 @subsection Commands
7509 This filter supports the all above options as @ref{commands}.
7511 @section colormatrix
7513 Convert color matrix.
7515 The filter accepts the following options:
7520 Specify the source and destination color matrix. Both values must be
7523 The accepted values are:
7551 For example to convert from BT.601 to SMPTE-240M, use the command:
7553 colormatrix=bt601:smpte240m
7558 Convert colorspace, transfer characteristics or color primaries.
7559 Input video needs to have an even size.
7561 The filter accepts the following options:
7566 Specify all color properties at once.
7568 The accepted values are:
7598 Specify output colorspace.
7600 The accepted values are:
7609 BT.470BG or BT.601-6 625
7612 SMPTE-170M or BT.601-6 525
7621 BT.2020 with non-constant luminance
7627 Specify output transfer characteristics.
7629 The accepted values are:
7641 Constant gamma of 2.2
7644 Constant gamma of 2.8
7647 SMPTE-170M, BT.601-6 625 or BT.601-6 525
7665 BT.2020 for 10-bits content
7668 BT.2020 for 12-bits content
7674 Specify output color primaries.
7676 The accepted values are:
7685 BT.470BG or BT.601-6 625
7688 SMPTE-170M or BT.601-6 525
7712 Specify output color range.
7714 The accepted values are:
7717 TV (restricted) range
7720 MPEG (restricted) range
7731 Specify output color format.
7733 The accepted values are:
7736 YUV 4:2:0 planar 8-bits
7739 YUV 4:2:0 planar 10-bits
7742 YUV 4:2:0 planar 12-bits
7745 YUV 4:2:2 planar 8-bits
7748 YUV 4:2:2 planar 10-bits
7751 YUV 4:2:2 planar 12-bits
7754 YUV 4:4:4 planar 8-bits
7757 YUV 4:4:4 planar 10-bits
7760 YUV 4:4:4 planar 12-bits
7765 Do a fast conversion, which skips gamma/primary correction. This will take
7766 significantly less CPU, but will be mathematically incorrect. To get output
7767 compatible with that produced by the colormatrix filter, use fast=1.
7770 Specify dithering mode.
7772 The accepted values are:
7778 Floyd-Steinberg dithering
7782 Whitepoint adaptation mode.
7784 The accepted values are:
7787 Bradford whitepoint adaptation
7790 von Kries whitepoint adaptation
7793 identity whitepoint adaptation (i.e. no whitepoint adaptation)
7797 Override all input properties at once. Same accepted values as @ref{all}.
7800 Override input colorspace. Same accepted values as @ref{space}.
7803 Override input color primaries. Same accepted values as @ref{primaries}.
7806 Override input transfer characteristics. Same accepted values as @ref{trc}.
7809 Override input color range. Same accepted values as @ref{range}.
7813 The filter converts the transfer characteristics, color space and color
7814 primaries to the specified user values. The output value, if not specified,
7815 is set to a default value based on the "all" property. If that property is
7816 also not specified, the filter will log an error. The output color range and
7817 format default to the same value as the input color range and format. The
7818 input transfer characteristics, color space, color primaries and color range
7819 should be set on the input data. If any of these are missing, the filter will
7820 log an error and no conversion will take place.
7822 For example to convert the input to SMPTE-240M, use the command:
7824 colorspace=smpte240m
7827 @section convolution
7829 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
7831 The filter accepts the following options:
7838 Set matrix for each plane.
7839 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
7840 and from 1 to 49 odd number of signed integers in @var{row} mode.
7846 Set multiplier for calculated value for each plane.
7847 If unset or 0, it will be sum of all matrix elements.
7853 Set bias for each plane. This value is added to the result of the multiplication.
7854 Useful for making the overall image brighter or darker. Default is 0.0.
7860 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
7861 Default is @var{square}.
7864 @subsection Examples
7870 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"
7876 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"
7882 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"
7888 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"
7892 Apply laplacian edge detector which includes diagonals:
7894 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"
7900 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"
7906 Apply 2D convolution of video stream in frequency domain using second stream
7909 The filter accepts the following options:
7913 Set which planes to process.
7916 Set which impulse video frames will be processed, can be @var{first}
7917 or @var{all}. Default is @var{all}.
7920 The @code{convolve} filter also supports the @ref{framesync} options.
7924 Copy the input video source unchanged to the output. This is mainly useful for
7929 Video filtering on GPU using Apple's CoreImage API on OSX.
7931 Hardware acceleration is based on an OpenGL context. Usually, this means it is
7932 processed by video hardware. However, software-based OpenGL implementations
7933 exist which means there is no guarantee for hardware processing. It depends on
7936 There are many filters and image generators provided by Apple that come with a
7937 large variety of options. The filter has to be referenced by its name along
7940 The coreimage filter accepts the following options:
7943 List all available filters and generators along with all their respective
7944 options as well as possible minimum and maximum values along with the default
7951 Specify all filters by their respective name and options.
7952 Use @var{list_filters} to determine all valid filter names and options.
7953 Numerical options are specified by a float value and are automatically clamped
7954 to their respective value range. Vector and color options have to be specified
7955 by a list of space separated float values. Character escaping has to be done.
7956 A special option name @code{default} is available to use default options for a
7959 It is required to specify either @code{default} or at least one of the filter options.
7960 All omitted options are used with their default values.
7961 The syntax of the filter string is as follows:
7963 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
7967 Specify a rectangle where the output of the filter chain is copied into the
7968 input image. It is given by a list of space separated float values:
7970 output_rect=x\ y\ width\ height
7972 If not given, the output rectangle equals the dimensions of the input image.
7973 The output rectangle is automatically cropped at the borders of the input
7974 image. Negative values are valid for each component.
7976 output_rect=25\ 25\ 100\ 100
7980 Several filters can be chained for successive processing without GPU-HOST
7981 transfers allowing for fast processing of complex filter chains.
7982 Currently, only filters with zero (generators) or exactly one (filters) input
7983 image and one output image are supported. Also, transition filters are not yet
7986 Some filters generate output images with additional padding depending on the
7987 respective filter kernel. The padding is automatically removed to ensure the
7988 filter output has the same size as the input image.
7990 For image generators, the size of the output image is determined by the
7991 previous output image of the filter chain or the input image of the whole
7992 filterchain, respectively. The generators do not use the pixel information of
7993 this image to generate their output. However, the generated output is
7994 blended onto this image, resulting in partial or complete coverage of the
7997 The @ref{coreimagesrc} video source can be used for generating input images
7998 which are directly fed into the filter chain. By using it, providing input
7999 images by another video source or an input video is not required.
8001 @subsection Examples
8006 List all filters available:
8008 coreimage=list_filters=true
8012 Use the CIBoxBlur filter with default options to blur an image:
8014 coreimage=filter=CIBoxBlur@@default
8018 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8019 its center at 100x100 and a radius of 50 pixels:
8021 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8025 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8026 given as complete and escaped command-line for Apple's standard bash shell:
8028 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8034 Cover a rectangular object
8036 It accepts the following options:
8040 Filepath of the optional cover image, needs to be in yuv420.
8045 It accepts the following values:
8048 cover it by the supplied image
8050 cover it by interpolating the surrounding pixels
8053 Default value is @var{blur}.
8056 @subsection Examples
8060 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8062 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8068 Crop the input video to given dimensions.
8070 It accepts the following parameters:
8074 The width of the output video. It defaults to @code{iw}.
8075 This expression is evaluated only once during the filter
8076 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8079 The height of the output video. It defaults to @code{ih}.
8080 This expression is evaluated only once during the filter
8081 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8084 The horizontal position, in the input video, of the left edge of the output
8085 video. It defaults to @code{(in_w-out_w)/2}.
8086 This expression is evaluated per-frame.
8089 The vertical position, in the input video, of the top edge of the output video.
8090 It defaults to @code{(in_h-out_h)/2}.
8091 This expression is evaluated per-frame.
8094 If set to 1 will force the output display aspect ratio
8095 to be the same of the input, by changing the output sample aspect
8096 ratio. It defaults to 0.
8099 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8100 width/height/x/y as specified and will not be rounded to nearest smaller value.
8104 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8105 expressions containing the following constants:
8110 The computed values for @var{x} and @var{y}. They are evaluated for
8115 The input width and height.
8119 These are the same as @var{in_w} and @var{in_h}.
8123 The output (cropped) width and height.
8127 These are the same as @var{out_w} and @var{out_h}.
8130 same as @var{iw} / @var{ih}
8133 input sample aspect ratio
8136 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8140 horizontal and vertical chroma subsample values. For example for the
8141 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8144 The number of the input frame, starting from 0.
8147 the position in the file of the input frame, NAN if unknown
8150 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8154 The expression for @var{out_w} may depend on the value of @var{out_h},
8155 and the expression for @var{out_h} may depend on @var{out_w}, but they
8156 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8157 evaluated after @var{out_w} and @var{out_h}.
8159 The @var{x} and @var{y} parameters specify the expressions for the
8160 position of the top-left corner of the output (non-cropped) area. They
8161 are evaluated for each frame. If the evaluated value is not valid, it
8162 is approximated to the nearest valid value.
8164 The expression for @var{x} may depend on @var{y}, and the expression
8165 for @var{y} may depend on @var{x}.
8167 @subsection Examples
8171 Crop area with size 100x100 at position (12,34).
8176 Using named options, the example above becomes:
8178 crop=w=100:h=100:x=12:y=34
8182 Crop the central input area with size 100x100:
8188 Crop the central input area with size 2/3 of the input video:
8190 crop=2/3*in_w:2/3*in_h
8194 Crop the input video central square:
8201 Delimit the rectangle with the top-left corner placed at position
8202 100:100 and the right-bottom corner corresponding to the right-bottom
8203 corner of the input image.
8205 crop=in_w-100:in_h-100:100:100
8209 Crop 10 pixels from the left and right borders, and 20 pixels from
8210 the top and bottom borders
8212 crop=in_w-2*10:in_h-2*20
8216 Keep only the bottom right quarter of the input image:
8218 crop=in_w/2:in_h/2:in_w/2:in_h/2
8222 Crop height for getting Greek harmony:
8224 crop=in_w:1/PHI*in_w
8228 Apply trembling effect:
8230 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)
8234 Apply erratic camera effect depending on timestamp:
8236 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)"
8240 Set x depending on the value of y:
8242 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8246 @subsection Commands
8248 This filter supports the following commands:
8254 Set width/height of the output video and the horizontal/vertical position
8256 The command accepts the same syntax of the corresponding option.
8258 If the specified expression is not valid, it is kept at its current
8264 Auto-detect the crop size.
8266 It calculates the necessary cropping parameters and prints the
8267 recommended parameters via the logging system. The detected dimensions
8268 correspond to the non-black area of the input video.
8270 It accepts the following parameters:
8275 Set higher black value threshold, which can be optionally specified
8276 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8277 value greater to the set value is considered non-black. It defaults to 24.
8278 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8279 on the bitdepth of the pixel format.
8282 The value which the width/height should be divisible by. It defaults to
8283 16. The offset is automatically adjusted to center the video. Use 2 to
8284 get only even dimensions (needed for 4:2:2 video). 16 is best when
8285 encoding to most video codecs.
8287 @item reset_count, reset
8288 Set the counter that determines after how many frames cropdetect will
8289 reset the previously detected largest video area and start over to
8290 detect the current optimal crop area. Default value is 0.
8292 This can be useful when channel logos distort the video area. 0
8293 indicates 'never reset', and returns the largest area encountered during
8300 Delay video filtering until a given wallclock timestamp. The filter first
8301 passes on @option{preroll} amount of frames, then it buffers at most
8302 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8303 it forwards the buffered frames and also any subsequent frames coming in its
8306 The filter can be used synchronize the output of multiple ffmpeg processes for
8307 realtime output devices like decklink. By putting the delay in the filtering
8308 chain and pre-buffering frames the process can pass on data to output almost
8309 immediately after the target wallclock timestamp is reached.
8311 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
8317 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
8320 The duration of content to pass on as preroll expressed in seconds. Default is 0.
8323 The maximum duration of content to buffer before waiting for the cue expressed
8324 in seconds. Default is 0.
8331 Apply color adjustments using curves.
8333 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
8334 component (red, green and blue) has its values defined by @var{N} key points
8335 tied from each other using a smooth curve. The x-axis represents the pixel
8336 values from the input frame, and the y-axis the new pixel values to be set for
8339 By default, a component curve is defined by the two points @var{(0;0)} and
8340 @var{(1;1)}. This creates a straight line where each original pixel value is
8341 "adjusted" to its own value, which means no change to the image.
8343 The filter allows you to redefine these two points and add some more. A new
8344 curve (using a natural cubic spline interpolation) will be define to pass
8345 smoothly through all these new coordinates. The new defined points needs to be
8346 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
8347 be in the @var{[0;1]} interval. If the computed curves happened to go outside
8348 the vector spaces, the values will be clipped accordingly.
8350 The filter accepts the following options:
8354 Select one of the available color presets. This option can be used in addition
8355 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
8356 options takes priority on the preset values.
8357 Available presets are:
8360 @item color_negative
8363 @item increase_contrast
8365 @item linear_contrast
8366 @item medium_contrast
8368 @item strong_contrast
8371 Default is @code{none}.
8373 Set the master key points. These points will define a second pass mapping. It
8374 is sometimes called a "luminance" or "value" mapping. It can be used with
8375 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
8376 post-processing LUT.
8378 Set the key points for the red component.
8380 Set the key points for the green component.
8382 Set the key points for the blue component.
8384 Set the key points for all components (not including master).
8385 Can be used in addition to the other key points component
8386 options. In this case, the unset component(s) will fallback on this
8387 @option{all} setting.
8389 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
8391 Save Gnuplot script of the curves in specified file.
8394 To avoid some filtergraph syntax conflicts, each key points list need to be
8395 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
8397 @subsection Examples
8401 Increase slightly the middle level of blue:
8403 curves=blue='0/0 0.5/0.58 1/1'
8409 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'
8411 Here we obtain the following coordinates for each components:
8414 @code{(0;0.11) (0.42;0.51) (1;0.95)}
8416 @code{(0;0) (0.50;0.48) (1;1)}
8418 @code{(0;0.22) (0.49;0.44) (1;0.80)}
8422 The previous example can also be achieved with the associated built-in preset:
8424 curves=preset=vintage
8434 Use a Photoshop preset and redefine the points of the green component:
8436 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
8440 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
8441 and @command{gnuplot}:
8443 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
8444 gnuplot -p /tmp/curves.plt
8450 Video data analysis filter.
8452 This filter shows hexadecimal pixel values of part of video.
8454 The filter accepts the following options:
8458 Set output video size.
8461 Set x offset from where to pick pixels.
8464 Set y offset from where to pick pixels.
8467 Set scope mode, can be one of the following:
8470 Draw hexadecimal pixel values with white color on black background.
8473 Draw hexadecimal pixel values with input video pixel color on black
8477 Draw hexadecimal pixel values on color background picked from input video,
8478 the text color is picked in such way so its always visible.
8482 Draw rows and columns numbers on left and top of video.
8485 Set background opacity.
8488 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
8493 Denoise frames using 2D DCT (frequency domain filtering).
8495 This filter is not designed for real time.
8497 The filter accepts the following options:
8501 Set the noise sigma constant.
8503 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
8504 coefficient (absolute value) below this threshold with be dropped.
8506 If you need a more advanced filtering, see @option{expr}.
8508 Default is @code{0}.
8511 Set number overlapping pixels for each block. Since the filter can be slow, you
8512 may want to reduce this value, at the cost of a less effective filter and the
8513 risk of various artefacts.
8515 If the overlapping value doesn't permit processing the whole input width or
8516 height, a warning will be displayed and according borders won't be denoised.
8518 Default value is @var{blocksize}-1, which is the best possible setting.
8521 Set the coefficient factor expression.
8523 For each coefficient of a DCT block, this expression will be evaluated as a
8524 multiplier value for the coefficient.
8526 If this is option is set, the @option{sigma} option will be ignored.
8528 The absolute value of the coefficient can be accessed through the @var{c}
8532 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8533 @var{blocksize}, which is the width and height of the processed blocks.
8535 The default value is @var{3} (8x8) and can be raised to @var{4} for a
8536 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
8537 on the speed processing. Also, a larger block size does not necessarily means a
8541 @subsection Examples
8543 Apply a denoise with a @option{sigma} of @code{4.5}:
8548 The same operation can be achieved using the expression system:
8550 dctdnoiz=e='gte(c, 4.5*3)'
8553 Violent denoise using a block size of @code{16x16}:
8560 Remove banding artifacts from input video.
8561 It works by replacing banded pixels with average value of referenced pixels.
8563 The filter accepts the following options:
8570 Set banding detection threshold for each plane. Default is 0.02.
8571 Valid range is 0.00003 to 0.5.
8572 If difference between current pixel and reference pixel is less than threshold,
8573 it will be considered as banded.
8576 Banding detection range in pixels. Default is 16. If positive, random number
8577 in range 0 to set value will be used. If negative, exact absolute value
8579 The range defines square of four pixels around current pixel.
8582 Set direction in radians from which four pixel will be compared. If positive,
8583 random direction from 0 to set direction will be picked. If negative, exact of
8584 absolute value will be picked. For example direction 0, -PI or -2*PI radians
8585 will pick only pixels on same row and -PI/2 will pick only pixels on same
8589 If enabled, current pixel is compared with average value of all four
8590 surrounding pixels. The default is enabled. If disabled current pixel is
8591 compared with all four surrounding pixels. The pixel is considered banded
8592 if only all four differences with surrounding pixels are less than threshold.
8595 If enabled, current pixel is changed if and only if all pixel components are banded,
8596 e.g. banding detection threshold is triggered for all color components.
8597 The default is disabled.
8602 Remove blocking artifacts from input video.
8604 The filter accepts the following options:
8608 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
8609 This controls what kind of deblocking is applied.
8612 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
8618 Set blocking detection thresholds. Allowed range is 0 to 1.
8619 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
8620 Using higher threshold gives more deblocking strength.
8621 Setting @var{alpha} controls threshold detection at exact edge of block.
8622 Remaining options controls threshold detection near the edge. Each one for
8623 below/above or left/right. Setting any of those to @var{0} disables
8627 Set planes to filter. Default is to filter all available planes.
8630 @subsection Examples
8634 Deblock using weak filter and block size of 4 pixels.
8636 deblock=filter=weak:block=4
8640 Deblock using strong filter, block size of 4 pixels and custom thresholds for
8641 deblocking more edges.
8643 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
8647 Similar as above, but filter only first plane.
8649 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
8653 Similar as above, but filter only second and third plane.
8655 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
8662 Drop duplicated frames at regular intervals.
8664 The filter accepts the following options:
8668 Set the number of frames from which one will be dropped. Setting this to
8669 @var{N} means one frame in every batch of @var{N} frames will be dropped.
8670 Default is @code{5}.
8673 Set the threshold for duplicate detection. If the difference metric for a frame
8674 is less than or equal to this value, then it is declared as duplicate. Default
8678 Set scene change threshold. Default is @code{15}.
8682 Set the size of the x and y-axis blocks used during metric calculations.
8683 Larger blocks give better noise suppression, but also give worse detection of
8684 small movements. Must be a power of two. Default is @code{32}.
8687 Mark main input as a pre-processed input and activate clean source input
8688 stream. This allows the input to be pre-processed with various filters to help
8689 the metrics calculation while keeping the frame selection lossless. When set to
8690 @code{1}, the first stream is for the pre-processed input, and the second
8691 stream is the clean source from where the kept frames are chosen. Default is
8695 Set whether or not chroma is considered in the metric calculations. Default is
8701 Apply 2D deconvolution of video stream in frequency domain using second stream
8704 The filter accepts the following options:
8708 Set which planes to process.
8711 Set which impulse video frames will be processed, can be @var{first}
8712 or @var{all}. Default is @var{all}.
8715 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
8716 and height are not same and not power of 2 or if stream prior to convolving
8720 The @code{deconvolve} filter also supports the @ref{framesync} options.
8724 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
8726 It accepts the following options:
8730 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
8731 @var{rainbows} for cross-color reduction.
8734 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
8737 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
8740 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
8743 Set temporal chroma threshold. Lower values increases reduction of cross-color.
8748 Apply deflate effect to the video.
8750 This filter replaces the pixel by the local(3x3) average by taking into account
8751 only values lower than the pixel.
8753 It accepts the following options:
8760 Limit the maximum change for each plane, default is 65535.
8761 If 0, plane will remain unchanged.
8764 @subsection Commands
8766 This filter supports the all above options as @ref{commands}.
8770 Remove temporal frame luminance variations.
8772 It accepts the following options:
8776 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
8779 Set averaging mode to smooth temporal luminance variations.
8781 Available values are:
8806 Do not actually modify frame. Useful when one only wants metadata.
8811 Remove judder produced by partially interlaced telecined content.
8813 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
8814 source was partially telecined content then the output of @code{pullup,dejudder}
8815 will have a variable frame rate. May change the recorded frame rate of the
8816 container. Aside from that change, this filter will not affect constant frame
8819 The option available in this filter is:
8823 Specify the length of the window over which the judder repeats.
8825 Accepts any integer greater than 1. Useful values are:
8829 If the original was telecined from 24 to 30 fps (Film to NTSC).
8832 If the original was telecined from 25 to 30 fps (PAL to NTSC).
8835 If a mixture of the two.
8838 The default is @samp{4}.
8843 Suppress a TV station logo by a simple interpolation of the surrounding
8844 pixels. Just set a rectangle covering the logo and watch it disappear
8845 (and sometimes something even uglier appear - your mileage may vary).
8847 It accepts the following parameters:
8852 Specify the top left corner coordinates of the logo. They must be
8857 Specify the width and height of the logo to clear. They must be
8861 Specify the thickness of the fuzzy edge of the rectangle (added to
8862 @var{w} and @var{h}). The default value is 1. This option is
8863 deprecated, setting higher values should no longer be necessary and
8867 When set to 1, a green rectangle is drawn on the screen to simplify
8868 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
8869 The default value is 0.
8871 The rectangle is drawn on the outermost pixels which will be (partly)
8872 replaced with interpolated values. The values of the next pixels
8873 immediately outside this rectangle in each direction will be used to
8874 compute the interpolated pixel values inside the rectangle.
8878 @subsection Examples
8882 Set a rectangle covering the area with top left corner coordinates 0,0
8883 and size 100x77, and a band of size 10:
8885 delogo=x=0:y=0:w=100:h=77:band=10
8893 Remove the rain in the input image/video by applying the derain methods based on
8894 convolutional neural networks. Supported models:
8898 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
8899 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
8902 Training as well as model generation scripts are provided in
8903 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
8905 Native model files (.model) can be generated from TensorFlow model
8906 files (.pb) by using tools/python/convert.py
8908 The filter accepts the following options:
8912 Specify which filter to use. This option accepts the following values:
8916 Derain filter. To conduct derain filter, you need to use a derain model.
8919 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
8921 Default value is @samp{derain}.
8924 Specify which DNN backend to use for model loading and execution. This option accepts
8925 the following values:
8929 Native implementation of DNN loading and execution.
8932 TensorFlow backend. To enable this backend you
8933 need to install the TensorFlow for C library (see
8934 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
8935 @code{--enable-libtensorflow}
8937 Default value is @samp{native}.
8940 Set path to model file specifying network architecture and its parameters.
8941 Note that different backends use different file formats. TensorFlow and native
8942 backend can load files for only its format.
8945 It can also be finished with @ref{dnn_processing} filter.
8949 Attempt to fix small changes in horizontal and/or vertical shift. This
8950 filter helps remove camera shake from hand-holding a camera, bumping a
8951 tripod, moving on a vehicle, etc.
8953 The filter accepts the following options:
8961 Specify a rectangular area where to limit the search for motion
8963 If desired the search for motion vectors can be limited to a
8964 rectangular area of the frame defined by its top left corner, width
8965 and height. These parameters have the same meaning as the drawbox
8966 filter which can be used to visualise the position of the bounding
8969 This is useful when simultaneous movement of subjects within the frame
8970 might be confused for camera motion by the motion vector search.
8972 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
8973 then the full frame is used. This allows later options to be set
8974 without specifying the bounding box for the motion vector search.
8976 Default - search the whole frame.
8980 Specify the maximum extent of movement in x and y directions in the
8981 range 0-64 pixels. Default 16.
8984 Specify how to generate pixels to fill blanks at the edge of the
8985 frame. Available values are:
8988 Fill zeroes at blank locations
8990 Original image at blank locations
8992 Extruded edge value at blank locations
8994 Mirrored edge at blank locations
8996 Default value is @samp{mirror}.
8999 Specify the blocksize to use for motion search. Range 4-128 pixels,
9003 Specify the contrast threshold for blocks. Only blocks with more than
9004 the specified contrast (difference between darkest and lightest
9005 pixels) will be considered. Range 1-255, default 125.
9008 Specify the search strategy. Available values are:
9011 Set exhaustive search
9013 Set less exhaustive search.
9015 Default value is @samp{exhaustive}.
9018 If set then a detailed log of the motion search is written to the
9025 Remove unwanted contamination of foreground colors, caused by reflected color of
9026 greenscreen or bluescreen.
9028 This filter accepts the following options:
9032 Set what type of despill to use.
9035 Set how spillmap will be generated.
9038 Set how much to get rid of still remaining spill.
9041 Controls amount of red in spill area.
9044 Controls amount of green in spill area.
9045 Should be -1 for greenscreen.
9048 Controls amount of blue in spill area.
9049 Should be -1 for bluescreen.
9052 Controls brightness of spill area, preserving colors.
9055 Modify alpha from generated spillmap.
9060 Apply an exact inverse of the telecine operation. It requires a predefined
9061 pattern specified using the pattern option which must be the same as that passed
9062 to the telecine filter.
9064 This filter accepts the following options:
9073 The default value is @code{top}.
9077 A string of numbers representing the pulldown pattern you wish to apply.
9078 The default value is @code{23}.
9081 A number representing position of the first frame with respect to the telecine
9082 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9087 Apply dilation effect to the video.
9089 This filter replaces the pixel by the local(3x3) maximum.
9091 It accepts the following options:
9098 Limit the maximum change for each plane, default is 65535.
9099 If 0, plane will remain unchanged.
9102 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9105 Flags to local 3x3 coordinates maps like this:
9112 @subsection Commands
9114 This filter supports the all above options as @ref{commands}.
9118 Displace pixels as indicated by second and third input stream.
9120 It takes three input streams and outputs one stream, the first input is the
9121 source, and second and third input are displacement maps.
9123 The second input specifies how much to displace pixels along the
9124 x-axis, while the third input specifies how much to displace pixels
9126 If one of displacement map streams terminates, last frame from that
9127 displacement map will be used.
9129 Note that once generated, displacements maps can be reused over and over again.
9131 A description of the accepted options follows.
9135 Set displace behavior for pixels that are out of range.
9137 Available values are:
9140 Missing pixels are replaced by black pixels.
9143 Adjacent pixels will spread out to replace missing pixels.
9146 Out of range pixels are wrapped so they point to pixels of other side.
9149 Out of range pixels will be replaced with mirrored pixels.
9151 Default is @samp{smear}.
9155 @subsection Examples
9159 Add ripple effect to rgb input of video size hd720:
9161 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
9165 Add wave effect to rgb input of video size hd720:
9167 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
9171 @anchor{dnn_processing}
9172 @section dnn_processing
9174 Do image processing with deep neural networks. It works together with another filter
9175 which converts the pixel format of the Frame to what the dnn network requires.
9177 The filter accepts the following options:
9181 Specify which DNN backend to use for model loading and execution. This option accepts
9182 the following values:
9186 Native implementation of DNN loading and execution.
9189 TensorFlow backend. To enable this backend you
9190 need to install the TensorFlow for C library (see
9191 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9192 @code{--enable-libtensorflow}
9195 Default value is @samp{native}.
9198 Set path to model file specifying network architecture and its parameters.
9199 Note that different backends use different file formats. TensorFlow and native
9200 backend can load files for only its format.
9202 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9205 Set the input name of the dnn network.
9208 Set the output name of the dnn network.
9212 @subsection Examples
9216 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9218 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9222 Halve the pixel value of the frame with format gray32f:
9224 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
9228 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9230 ./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
9234 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9236 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9243 Draw a colored box on the input image.
9245 It accepts the following parameters:
9250 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9254 The expressions which specify the width and height of the box; if 0 they are interpreted as
9255 the input width and height. It defaults to 0.
9258 Specify the color of the box to write. For the general syntax of this option,
9259 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9260 value @code{invert} is used, the box edge color is the same as the
9261 video with inverted luma.
9264 The expression which sets the thickness of the box edge.
9265 A value of @code{fill} will create a filled box. Default value is @code{3}.
9267 See below for the list of accepted constants.
9270 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
9271 will overwrite the video's color and alpha pixels.
9272 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
9275 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9276 following constants:
9280 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9284 horizontal and vertical chroma subsample values. For example for the
9285 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9289 The input width and height.
9292 The input sample aspect ratio.
9296 The x and y offset coordinates where the box is drawn.
9300 The width and height of the drawn box.
9303 The thickness of the drawn box.
9305 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9306 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9310 @subsection Examples
9314 Draw a black box around the edge of the input image:
9320 Draw a box with color red and an opacity of 50%:
9322 drawbox=10:20:200:60:red@@0.5
9325 The previous example can be specified as:
9327 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
9331 Fill the box with pink color:
9333 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
9337 Draw a 2-pixel red 2.40:1 mask:
9339 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
9343 @subsection Commands
9344 This filter supports same commands as options.
9345 The command accepts the same syntax of the corresponding option.
9347 If the specified expression is not valid, it is kept at its current
9352 Draw a graph using input video metadata.
9354 It accepts the following parameters:
9358 Set 1st frame metadata key from which metadata values will be used to draw a graph.
9361 Set 1st foreground color expression.
9364 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
9367 Set 2nd foreground color expression.
9370 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
9373 Set 3rd foreground color expression.
9376 Set 4th frame metadata key from which metadata values will be used to draw a graph.
9379 Set 4th foreground color expression.
9382 Set minimal value of metadata value.
9385 Set maximal value of metadata value.
9388 Set graph background color. Default is white.
9393 Available values for mode is:
9400 Default is @code{line}.
9405 Available values for slide is:
9408 Draw new frame when right border is reached.
9411 Replace old columns with new ones.
9414 Scroll from right to left.
9417 Scroll from left to right.
9420 Draw single picture.
9423 Default is @code{frame}.
9426 Set size of graph video. For the syntax of this option, check the
9427 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
9428 The default value is @code{900x256}.
9431 Set the output frame rate. Default value is @code{25}.
9433 The foreground color expressions can use the following variables:
9436 Minimal value of metadata value.
9439 Maximal value of metadata value.
9442 Current metadata key value.
9445 The color is defined as 0xAABBGGRR.
9448 Example using metadata from @ref{signalstats} filter:
9450 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
9453 Example using metadata from @ref{ebur128} filter:
9455 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
9460 Draw a grid on the input image.
9462 It accepts the following parameters:
9467 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
9471 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
9472 input width and height, respectively, minus @code{thickness}, so image gets
9473 framed. Default to 0.
9476 Specify the color of the grid. 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 grid color is the same as the
9479 video with inverted luma.
9482 The expression which sets the thickness of the grid line. Default value is @code{1}.
9484 See below for the list of accepted constants.
9487 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
9488 will overwrite the video's color and alpha pixels.
9489 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
9492 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9493 following constants:
9497 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9501 horizontal and vertical chroma subsample values. For example for the
9502 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9506 The input grid cell width and height.
9509 The input sample aspect ratio.
9513 The x and y coordinates of some point of grid intersection (meant to configure offset).
9517 The width and height of the drawn cell.
9520 The thickness of the drawn cell.
9522 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9523 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9527 @subsection Examples
9531 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
9533 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
9537 Draw a white 3x3 grid with an opacity of 50%:
9539 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
9543 @subsection Commands
9544 This filter supports same commands as options.
9545 The command accepts the same syntax of the corresponding option.
9547 If the specified expression is not valid, it is kept at its current
9553 Draw a text string or text from a specified file on top of a video, using the
9554 libfreetype library.
9556 To enable compilation of this filter, you need to configure FFmpeg with
9557 @code{--enable-libfreetype}.
9558 To enable default font fallback and the @var{font} option you need to
9559 configure FFmpeg with @code{--enable-libfontconfig}.
9560 To enable the @var{text_shaping} option, you need to configure FFmpeg with
9561 @code{--enable-libfribidi}.
9565 It accepts the following parameters:
9570 Used to draw a box around text using the background color.
9571 The value must be either 1 (enable) or 0 (disable).
9572 The default value of @var{box} is 0.
9575 Set the width of the border to be drawn around the box using @var{boxcolor}.
9576 The default value of @var{boxborderw} is 0.
9579 The color to be used for drawing box around text. For the syntax of this
9580 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9582 The default value of @var{boxcolor} is "white".
9585 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
9586 The default value of @var{line_spacing} is 0.
9589 Set the width of the border to be drawn around the text using @var{bordercolor}.
9590 The default value of @var{borderw} is 0.
9593 Set the color to be used for drawing border around text. For the syntax of this
9594 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9596 The default value of @var{bordercolor} is "black".
9599 Select how the @var{text} is expanded. Can be either @code{none},
9600 @code{strftime} (deprecated) or
9601 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
9605 Set a start time for the count. Value is in microseconds. Only applied
9606 in the deprecated strftime expansion mode. To emulate in normal expansion
9607 mode use the @code{pts} function, supplying the start time (in seconds)
9608 as the second argument.
9611 If true, check and fix text coords to avoid clipping.
9614 The color to be used for drawing fonts. For the syntax of this option, check
9615 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9617 The default value of @var{fontcolor} is "black".
9619 @item fontcolor_expr
9620 String which is expanded the same way as @var{text} to obtain dynamic
9621 @var{fontcolor} value. By default this option has empty value and is not
9622 processed. When this option is set, it overrides @var{fontcolor} option.
9625 The font family to be used for drawing text. By default Sans.
9628 The font file to be used for drawing text. The path must be included.
9629 This parameter is mandatory if the fontconfig support is disabled.
9632 Draw the text applying alpha blending. The value can
9633 be a number between 0.0 and 1.0.
9634 The expression accepts the same variables @var{x, y} as well.
9635 The default value is 1.
9636 Please see @var{fontcolor_expr}.
9639 The font size to be used for drawing text.
9640 The default value of @var{fontsize} is 16.
9643 If set to 1, attempt to shape the text (for example, reverse the order of
9644 right-to-left text and join Arabic characters) before drawing it.
9645 Otherwise, just draw the text exactly as given.
9646 By default 1 (if supported).
9649 The flags to be used for loading the fonts.
9651 The flags map the corresponding flags supported by libfreetype, and are
9652 a combination of the following values:
9659 @item vertical_layout
9660 @item force_autohint
9663 @item ignore_global_advance_width
9665 @item ignore_transform
9671 Default value is "default".
9673 For more information consult the documentation for the FT_LOAD_*
9677 The color to be used for drawing a shadow behind the drawn text. For the
9678 syntax of this option, check the @ref{color syntax,,"Color" section in the
9679 ffmpeg-utils manual,ffmpeg-utils}.
9681 The default value of @var{shadowcolor} is "black".
9685 The x and y offsets for the text shadow position with respect to the
9686 position of the text. They can be either positive or negative
9687 values. The default value for both is "0".
9690 The starting frame number for the n/frame_num variable. The default value
9694 The size in number of spaces to use for rendering the tab.
9698 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
9699 format. It can be used with or without text parameter. @var{timecode_rate}
9700 option must be specified.
9702 @item timecode_rate, rate, r
9703 Set the timecode frame rate (timecode only). Value will be rounded to nearest
9704 integer. Minimum value is "1".
9705 Drop-frame timecode is supported for frame rates 30 & 60.
9708 If set to 1, the output of the timecode option will wrap around at 24 hours.
9709 Default is 0 (disabled).
9712 The text string to be drawn. The text must be a sequence of UTF-8
9714 This parameter is mandatory if no file is specified with the parameter
9718 A text file containing text to be drawn. The text must be a sequence
9719 of UTF-8 encoded characters.
9721 This parameter is mandatory if no text string is specified with the
9722 parameter @var{text}.
9724 If both @var{text} and @var{textfile} are specified, an error is thrown.
9727 If set to 1, the @var{textfile} will be reloaded before each frame.
9728 Be sure to update it atomically, or it may be read partially, or even fail.
9732 The expressions which specify the offsets where text will be drawn
9733 within the video frame. They are relative to the top/left border of the
9736 The default value of @var{x} and @var{y} is "0".
9738 See below for the list of accepted constants and functions.
9741 The parameters for @var{x} and @var{y} are expressions containing the
9742 following constants and functions:
9746 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
9750 horizontal and vertical chroma subsample values. For example for the
9751 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9754 the height of each text line
9762 @item max_glyph_a, ascent
9763 the maximum distance from the baseline to the highest/upper grid
9764 coordinate used to place a glyph outline point, for all the rendered
9766 It is a positive value, due to the grid's orientation with the Y axis
9769 @item max_glyph_d, descent
9770 the maximum distance from the baseline to the lowest grid coordinate
9771 used to place a glyph outline point, for all the rendered glyphs.
9772 This is a negative value, due to the grid's orientation, with the Y axis
9776 maximum glyph height, that is the maximum height for all the glyphs
9777 contained in the rendered text, it is equivalent to @var{ascent} -
9781 maximum glyph width, that is the maximum width for all the glyphs
9782 contained in the rendered text
9785 the number of input frame, starting from 0
9787 @item rand(min, max)
9788 return a random number included between @var{min} and @var{max}
9791 The input sample aspect ratio.
9794 timestamp expressed in seconds, NAN if the input timestamp is unknown
9797 the height of the rendered text
9800 the width of the rendered text
9804 the x and y offset coordinates where the text is drawn.
9806 These parameters allow the @var{x} and @var{y} expressions to refer
9807 to each other, so you can for example specify @code{y=x/dar}.
9810 A one character description of the current frame's picture type.
9813 The current packet's position in the input file or stream
9814 (in bytes, from the start of the input). A value of -1 indicates
9815 this info is not available.
9818 The current packet's duration, in seconds.
9821 The current packet's size (in bytes).
9824 @anchor{drawtext_expansion}
9825 @subsection Text expansion
9827 If @option{expansion} is set to @code{strftime},
9828 the filter recognizes strftime() sequences in the provided text and
9829 expands them accordingly. Check the documentation of strftime(). This
9830 feature is deprecated.
9832 If @option{expansion} is set to @code{none}, the text is printed verbatim.
9834 If @option{expansion} is set to @code{normal} (which is the default),
9835 the following expansion mechanism is used.
9837 The backslash character @samp{\}, followed by any character, always expands to
9838 the second character.
9840 Sequences of the form @code{%@{...@}} are expanded. The text between the
9841 braces is a function name, possibly followed by arguments separated by ':'.
9842 If the arguments contain special characters or delimiters (':' or '@}'),
9843 they should be escaped.
9845 Note that they probably must also be escaped as the value for the
9846 @option{text} option in the filter argument string and as the filter
9847 argument in the filtergraph description, and possibly also for the shell,
9848 that makes up to four levels of escaping; using a text file avoids these
9851 The following functions are available:
9856 The expression evaluation result.
9858 It must take one argument specifying the expression to be evaluated,
9859 which accepts the same constants and functions as the @var{x} and
9860 @var{y} values. Note that not all constants should be used, for
9861 example the text size is not known when evaluating the expression, so
9862 the constants @var{text_w} and @var{text_h} will have an undefined
9865 @item expr_int_format, eif
9866 Evaluate the expression's value and output as formatted integer.
9868 The first argument is the expression to be evaluated, just as for the @var{expr} function.
9869 The second argument specifies the output format. Allowed values are @samp{x},
9870 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
9871 @code{printf} function.
9872 The third parameter is optional and sets the number of positions taken by the output.
9873 It can be used to add padding with zeros from the left.
9876 The time at which the filter is running, expressed in UTC.
9877 It can accept an argument: a strftime() format string.
9880 The time at which the filter is running, expressed in the local time zone.
9881 It can accept an argument: a strftime() format string.
9884 Frame metadata. Takes one or two arguments.
9886 The first argument is mandatory and specifies the metadata key.
9888 The second argument is optional and specifies a default value, used when the
9889 metadata key is not found or empty.
9891 Available metadata can be identified by inspecting entries
9892 starting with TAG included within each frame section
9893 printed by running @code{ffprobe -show_frames}.
9895 String metadata generated in filters leading to
9896 the drawtext filter are also available.
9899 The frame number, starting from 0.
9902 A one character description of the current picture type.
9905 The timestamp of the current frame.
9906 It can take up to three arguments.
9908 The first argument is the format of the timestamp; it defaults to @code{flt}
9909 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
9910 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
9911 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
9912 @code{localtime} stands for the timestamp of the frame formatted as
9913 local time zone time.
9915 The second argument is an offset added to the timestamp.
9917 If the format is set to @code{hms}, a third argument @code{24HH} may be
9918 supplied to present the hour part of the formatted timestamp in 24h format
9921 If the format is set to @code{localtime} or @code{gmtime},
9922 a third argument may be supplied: a strftime() format string.
9923 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
9926 @subsection Commands
9928 This filter supports altering parameters via commands:
9931 Alter existing filter parameters.
9933 Syntax for the argument is the same as for filter invocation, e.g.
9936 fontsize=56:fontcolor=green:text='Hello World'
9939 Full filter invocation with sendcmd would look like this:
9942 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
9946 If the entire argument can't be parsed or applied as valid values then the filter will
9947 continue with its existing parameters.
9949 @subsection Examples
9953 Draw "Test Text" with font FreeSerif, using the default values for the
9954 optional parameters.
9957 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
9961 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
9962 and y=50 (counting from the top-left corner of the screen), text is
9963 yellow with a red box around it. Both the text and the box have an
9967 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
9968 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
9971 Note that the double quotes are not necessary if spaces are not used
9972 within the parameter list.
9975 Show the text at the center of the video frame:
9977 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
9981 Show the text at a random position, switching to a new position every 30 seconds:
9983 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)"
9987 Show a text line sliding from right to left in the last row of the video
9988 frame. The file @file{LONG_LINE} is assumed to contain a single line
9991 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
9995 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
9997 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10001 Draw a single green letter "g", at the center of the input video.
10002 The glyph baseline is placed at half screen height.
10004 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10008 Show text for 1 second every 3 seconds:
10010 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10014 Use fontconfig to set the font. Note that the colons need to be escaped.
10016 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10020 Print the date of a real-time encoding (see strftime(3)):
10022 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10026 Show text fading in and out (appearing/disappearing):
10029 DS=1.0 # display start
10030 DE=10.0 # display end
10031 FID=1.5 # fade in duration
10032 FOD=5 # fade out duration
10033 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 @}"
10037 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10038 and the @option{fontsize} value are included in the @option{y} offset.
10040 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10041 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10045 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10046 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10047 must have option @option{-export_path_metadata 1} for the special metadata fields
10048 to be available for filters.
10050 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10055 For more information about libfreetype, check:
10056 @url{http://www.freetype.org/}.
10058 For more information about fontconfig, check:
10059 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10061 For more information about libfribidi, check:
10062 @url{http://fribidi.org/}.
10064 @section edgedetect
10066 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10068 The filter accepts the following options:
10073 Set low and high threshold values used by the Canny thresholding
10076 The high threshold selects the "strong" edge pixels, which are then
10077 connected through 8-connectivity with the "weak" edge pixels selected
10078 by the low threshold.
10080 @var{low} and @var{high} threshold values must be chosen in the range
10081 [0,1], and @var{low} should be lesser or equal to @var{high}.
10083 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10087 Define the drawing mode.
10091 Draw white/gray wires on black background.
10094 Mix the colors to create a paint/cartoon effect.
10097 Apply Canny edge detector on all selected planes.
10099 Default value is @var{wires}.
10102 Select planes for filtering. By default all available planes are filtered.
10105 @subsection Examples
10109 Standard edge detection with custom values for the hysteresis thresholding:
10111 edgedetect=low=0.1:high=0.4
10115 Painting effect without thresholding:
10117 edgedetect=mode=colormix:high=0
10123 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10125 For each input image, the filter will compute the optimal mapping from
10126 the input to the output given the codebook length, that is the number
10127 of distinct output colors.
10129 This filter accepts the following options.
10132 @item codebook_length, l
10133 Set codebook length. The value must be a positive integer, and
10134 represents the number of distinct output colors. Default value is 256.
10137 Set the maximum number of iterations to apply for computing the optimal
10138 mapping. The higher the value the better the result and the higher the
10139 computation time. Default value is 1.
10142 Set a random seed, must be an integer included between 0 and
10143 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10144 will try to use a good random seed on a best effort basis.
10147 Set pal8 output pixel format. This option does not work with codebook
10148 length greater than 256.
10153 Measure graylevel entropy in histogram of color channels of video frames.
10155 It accepts the following parameters:
10159 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10161 @var{diff} mode measures entropy of histogram delta values, absolute differences
10162 between neighbour histogram values.
10166 Set brightness, contrast, saturation and approximate gamma adjustment.
10168 The filter accepts the following options:
10172 Set the contrast expression. The value must be a float value in range
10173 @code{-1000.0} to @code{1000.0}. The default value is "1".
10176 Set the brightness expression. The value must be a float value in
10177 range @code{-1.0} to @code{1.0}. The default value is "0".
10180 Set the saturation expression. The value must be a float in
10181 range @code{0.0} to @code{3.0}. The default value is "1".
10184 Set the gamma expression. The value must be a float in range
10185 @code{0.1} to @code{10.0}. The default value is "1".
10188 Set the gamma expression for red. The value must be a float in
10189 range @code{0.1} to @code{10.0}. The default value is "1".
10192 Set the gamma expression for green. The value must be a float in range
10193 @code{0.1} to @code{10.0}. The default value is "1".
10196 Set the gamma expression for blue. The value must be a float in range
10197 @code{0.1} to @code{10.0}. The default value is "1".
10200 Set the gamma weight expression. It can be used to reduce the effect
10201 of a high gamma value on bright image areas, e.g. keep them from
10202 getting overamplified and just plain white. The value must be a float
10203 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10204 gamma correction all the way down while @code{1.0} leaves it at its
10205 full strength. Default is "1".
10208 Set when the expressions for brightness, contrast, saturation and
10209 gamma expressions are evaluated.
10211 It accepts the following values:
10214 only evaluate expressions once during the filter initialization or
10215 when a command is processed
10218 evaluate expressions for each incoming frame
10221 Default value is @samp{init}.
10224 The expressions accept the following parameters:
10227 frame count of the input frame starting from 0
10230 byte position of the corresponding packet in the input file, NAN if
10234 frame rate of the input video, NAN if the input frame rate is unknown
10237 timestamp expressed in seconds, NAN if the input timestamp is unknown
10240 @subsection Commands
10241 The filter supports the following commands:
10245 Set the contrast expression.
10248 Set the brightness expression.
10251 Set the saturation expression.
10254 Set the gamma expression.
10257 Set the gamma_r expression.
10260 Set gamma_g expression.
10263 Set gamma_b expression.
10266 Set gamma_weight expression.
10268 The command accepts the same syntax of the corresponding option.
10270 If the specified expression is not valid, it is kept at its current
10277 Apply erosion effect to the video.
10279 This filter replaces the pixel by the local(3x3) minimum.
10281 It accepts the following options:
10288 Limit the maximum change for each plane, default is 65535.
10289 If 0, plane will remain unchanged.
10292 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
10295 Flags to local 3x3 coordinates maps like this:
10302 @subsection Commands
10304 This filter supports the all above options as @ref{commands}.
10306 @section extractplanes
10308 Extract color channel components from input video stream into
10309 separate grayscale video streams.
10311 The filter accepts the following option:
10315 Set plane(s) to extract.
10317 Available values for planes are:
10328 Choosing planes not available in the input will result in an error.
10329 That means you cannot select @code{r}, @code{g}, @code{b} planes
10330 with @code{y}, @code{u}, @code{v} planes at same time.
10333 @subsection Examples
10337 Extract luma, u and v color channel component from input video frame
10338 into 3 grayscale outputs:
10340 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
10346 Apply a fade-in/out effect to the input video.
10348 It accepts the following parameters:
10352 The effect type can be either "in" for a fade-in, or "out" for a fade-out
10354 Default is @code{in}.
10356 @item start_frame, s
10357 Specify the number of the frame to start applying the fade
10358 effect at. Default is 0.
10361 The number of frames that the fade effect lasts. At the end of the
10362 fade-in effect, the output video will have the same intensity as the input video.
10363 At the end of the fade-out transition, the output video will be filled with the
10364 selected @option{color}.
10368 If set to 1, fade only alpha channel, if one exists on the input.
10369 Default value is 0.
10371 @item start_time, st
10372 Specify the timestamp (in seconds) of the frame to start to apply the fade
10373 effect. If both start_frame and start_time are specified, the fade will start at
10374 whichever comes last. Default is 0.
10377 The number of seconds for which the fade effect has to last. At the end of the
10378 fade-in effect the output video will have the same intensity as the input video,
10379 at the end of the fade-out transition the output video will be filled with the
10380 selected @option{color}.
10381 If both duration and nb_frames are specified, duration is used. Default is 0
10382 (nb_frames is used by default).
10385 Specify the color of the fade. Default is "black".
10388 @subsection Examples
10392 Fade in the first 30 frames of video:
10397 The command above is equivalent to:
10403 Fade out the last 45 frames of a 200-frame video:
10406 fade=type=out:start_frame=155:nb_frames=45
10410 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
10412 fade=in:0:25, fade=out:975:25
10416 Make the first 5 frames yellow, then fade in from frame 5-24:
10418 fade=in:5:20:color=yellow
10422 Fade in alpha over first 25 frames of video:
10424 fade=in:0:25:alpha=1
10428 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
10430 fade=t=in:st=5.5:d=0.5
10436 Denoise frames using 3D FFT (frequency domain filtering).
10438 The filter accepts the following options:
10442 Set the noise sigma constant. This sets denoising strength.
10443 Default value is 1. Allowed range is from 0 to 30.
10444 Using very high sigma with low overlap may give blocking artifacts.
10447 Set amount of denoising. By default all detected noise is reduced.
10448 Default value is 1. Allowed range is from 0 to 1.
10451 Set size of block, Default is 4, can be 3, 4, 5 or 6.
10452 Actual size of block in pixels is 2 to power of @var{block}, so by default
10453 block size in pixels is 2^4 which is 16.
10456 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
10459 Set number of previous frames to use for denoising. By default is set to 0.
10462 Set number of next frames to to use for denoising. By default is set to 0.
10465 Set planes which will be filtered, by default are all available filtered
10470 Apply arbitrary expressions to samples in frequency domain
10474 Adjust the dc value (gain) of the luma plane of the image. The filter
10475 accepts an integer value in range @code{0} to @code{1000}. The default
10476 value is set to @code{0}.
10479 Adjust the dc value (gain) of the 1st chroma plane of the image. The
10480 filter accepts an integer value in range @code{0} to @code{1000}. The
10481 default value is set to @code{0}.
10484 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
10485 filter accepts an integer value in range @code{0} to @code{1000}. The
10486 default value is set to @code{0}.
10489 Set the frequency domain weight expression for the luma plane.
10492 Set the frequency domain weight expression for the 1st chroma plane.
10495 Set the frequency domain weight expression for the 2nd chroma plane.
10498 Set when the expressions are evaluated.
10500 It accepts the following values:
10503 Only evaluate expressions once during the filter initialization.
10506 Evaluate expressions for each incoming frame.
10509 Default value is @samp{init}.
10511 The filter accepts the following variables:
10514 The coordinates of the current sample.
10518 The width and height of the image.
10521 The number of input frame, starting from 0.
10524 @subsection Examples
10530 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
10536 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
10542 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
10548 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
10555 Extract a single field from an interlaced image using stride
10556 arithmetic to avoid wasting CPU time. The output frames are marked as
10559 The filter accepts the following options:
10563 Specify whether to extract the top (if the value is @code{0} or
10564 @code{top}) or the bottom field (if the value is @code{1} or
10570 Create new frames by copying the top and bottom fields from surrounding frames
10571 supplied as numbers by the hint file.
10575 Set file containing hints: absolute/relative frame numbers.
10577 There must be one line for each frame in a clip. Each line must contain two
10578 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
10579 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
10580 is current frame number for @code{absolute} mode or out of [-1, 1] range
10581 for @code{relative} mode. First number tells from which frame to pick up top
10582 field and second number tells from which frame to pick up bottom field.
10584 If optionally followed by @code{+} output frame will be marked as interlaced,
10585 else if followed by @code{-} output frame will be marked as progressive, else
10586 it will be marked same as input frame.
10587 If optionally followed by @code{t} output frame will use only top field, or in
10588 case of @code{b} it will use only bottom field.
10589 If line starts with @code{#} or @code{;} that line is skipped.
10592 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
10595 Example of first several lines of @code{hint} file for @code{relative} mode:
10597 0,0 - # first frame
10598 1,0 - # second frame, use third's frame top field and second's frame bottom field
10599 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
10614 @section fieldmatch
10616 Field matching filter for inverse telecine. It is meant to reconstruct the
10617 progressive frames from a telecined stream. The filter does not drop duplicated
10618 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
10619 followed by a decimation filter such as @ref{decimate} in the filtergraph.
10621 The separation of the field matching and the decimation is notably motivated by
10622 the possibility of inserting a de-interlacing filter fallback between the two.
10623 If the source has mixed telecined and real interlaced content,
10624 @code{fieldmatch} will not be able to match fields for the interlaced parts.
10625 But these remaining combed frames will be marked as interlaced, and thus can be
10626 de-interlaced by a later filter such as @ref{yadif} before decimation.
10628 In addition to the various configuration options, @code{fieldmatch} can take an
10629 optional second stream, activated through the @option{ppsrc} option. If
10630 enabled, the frames reconstruction will be based on the fields and frames from
10631 this second stream. This allows the first input to be pre-processed in order to
10632 help the various algorithms of the filter, while keeping the output lossless
10633 (assuming the fields are matched properly). Typically, a field-aware denoiser,
10634 or brightness/contrast adjustments can help.
10636 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
10637 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
10638 which @code{fieldmatch} is based on. While the semantic and usage are very
10639 close, some behaviour and options names can differ.
10641 The @ref{decimate} filter currently only works for constant frame rate input.
10642 If your input has mixed telecined (30fps) and progressive content with a lower
10643 framerate like 24fps use the following filterchain to produce the necessary cfr
10644 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
10646 The filter accepts the following options:
10650 Specify the assumed field order of the input stream. Available values are:
10654 Auto detect parity (use FFmpeg's internal parity value).
10656 Assume bottom field first.
10658 Assume top field first.
10661 Note that it is sometimes recommended not to trust the parity announced by the
10664 Default value is @var{auto}.
10667 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
10668 sense that it won't risk creating jerkiness due to duplicate frames when
10669 possible, but if there are bad edits or blended fields it will end up
10670 outputting combed frames when a good match might actually exist. On the other
10671 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
10672 but will almost always find a good frame if there is one. The other values are
10673 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
10674 jerkiness and creating duplicate frames versus finding good matches in sections
10675 with bad edits, orphaned fields, blended fields, etc.
10677 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
10679 Available values are:
10683 2-way matching (p/c)
10685 2-way matching, and trying 3rd match if still combed (p/c + n)
10687 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
10689 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
10690 still combed (p/c + n + u/b)
10692 3-way matching (p/c/n)
10694 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
10695 detected as combed (p/c/n + u/b)
10698 The parenthesis at the end indicate the matches that would be used for that
10699 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
10702 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
10705 Default value is @var{pc_n}.
10708 Mark the main input stream as a pre-processed input, and enable the secondary
10709 input stream as the clean source to pick the fields from. See the filter
10710 introduction for more details. It is similar to the @option{clip2} feature from
10713 Default value is @code{0} (disabled).
10716 Set the field to match from. It is recommended to set this to the same value as
10717 @option{order} unless you experience matching failures with that setting. In
10718 certain circumstances changing the field that is used to match from can have a
10719 large impact on matching performance. Available values are:
10723 Automatic (same value as @option{order}).
10725 Match from the bottom field.
10727 Match from the top field.
10730 Default value is @var{auto}.
10733 Set whether or not chroma is included during the match comparisons. In most
10734 cases it is recommended to leave this enabled. You should set this to @code{0}
10735 only if your clip has bad chroma problems such as heavy rainbowing or other
10736 artifacts. Setting this to @code{0} could also be used to speed things up at
10737 the cost of some accuracy.
10739 Default value is @code{1}.
10743 These define an exclusion band which excludes the lines between @option{y0} and
10744 @option{y1} from being included in the field matching decision. An exclusion
10745 band can be used to ignore subtitles, a logo, or other things that may
10746 interfere with the matching. @option{y0} sets the starting scan line and
10747 @option{y1} sets the ending line; all lines in between @option{y0} and
10748 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
10749 @option{y0} and @option{y1} to the same value will disable the feature.
10750 @option{y0} and @option{y1} defaults to @code{0}.
10753 Set the scene change detection threshold as a percentage of maximum change on
10754 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
10755 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
10756 @option{scthresh} is @code{[0.0, 100.0]}.
10758 Default value is @code{12.0}.
10761 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
10762 account the combed scores of matches when deciding what match to use as the
10763 final match. Available values are:
10767 No final matching based on combed scores.
10769 Combed scores are only used when a scene change is detected.
10771 Use combed scores all the time.
10774 Default is @var{sc}.
10777 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
10778 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
10779 Available values are:
10783 No forced calculation.
10785 Force p/c/n calculations.
10787 Force p/c/n/u/b calculations.
10790 Default value is @var{none}.
10793 This is the area combing threshold used for combed frame detection. This
10794 essentially controls how "strong" or "visible" combing must be to be detected.
10795 Larger values mean combing must be more visible and smaller values mean combing
10796 can be less visible or strong and still be detected. Valid settings are from
10797 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
10798 be detected as combed). This is basically a pixel difference value. A good
10799 range is @code{[8, 12]}.
10801 Default value is @code{9}.
10804 Sets whether or not chroma is considered in the combed frame decision. Only
10805 disable this if your source has chroma problems (rainbowing, etc.) that are
10806 causing problems for the combed frame detection with chroma enabled. Actually,
10807 using @option{chroma}=@var{0} is usually more reliable, except for the case
10808 where there is chroma only combing in the source.
10810 Default value is @code{0}.
10814 Respectively set the x-axis and y-axis size of the window used during combed
10815 frame detection. This has to do with the size of the area in which
10816 @option{combpel} pixels are required to be detected as combed for a frame to be
10817 declared combed. See the @option{combpel} parameter description for more info.
10818 Possible values are any number that is a power of 2 starting at 4 and going up
10821 Default value is @code{16}.
10824 The number of combed pixels inside any of the @option{blocky} by
10825 @option{blockx} size blocks on the frame for the frame to be detected as
10826 combed. While @option{cthresh} controls how "visible" the combing must be, this
10827 setting controls "how much" combing there must be in any localized area (a
10828 window defined by the @option{blockx} and @option{blocky} settings) on the
10829 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
10830 which point no frames will ever be detected as combed). This setting is known
10831 as @option{MI} in TFM/VFM vocabulary.
10833 Default value is @code{80}.
10836 @anchor{p/c/n/u/b meaning}
10837 @subsection p/c/n/u/b meaning
10839 @subsubsection p/c/n
10841 We assume the following telecined stream:
10844 Top fields: 1 2 2 3 4
10845 Bottom fields: 1 2 3 4 4
10848 The numbers correspond to the progressive frame the fields relate to. Here, the
10849 first two frames are progressive, the 3rd and 4th are combed, and so on.
10851 When @code{fieldmatch} is configured to run a matching from bottom
10852 (@option{field}=@var{bottom}) this is how this input stream get transformed:
10857 B 1 2 3 4 4 <-- matching reference
10866 As a result of the field matching, we can see that some frames get duplicated.
10867 To perform a complete inverse telecine, you need to rely on a decimation filter
10868 after this operation. See for instance the @ref{decimate} filter.
10870 The same operation now matching from top fields (@option{field}=@var{top})
10875 T 1 2 2 3 4 <-- matching reference
10885 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
10886 basically, they refer to the frame and field of the opposite parity:
10889 @item @var{p} matches the field of the opposite parity in the previous frame
10890 @item @var{c} matches the field of the opposite parity in the current frame
10891 @item @var{n} matches the field of the opposite parity in the next frame
10896 The @var{u} and @var{b} matching are a bit special in the sense that they match
10897 from the opposite parity flag. In the following examples, we assume that we are
10898 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
10899 'x' is placed above and below each matched fields.
10901 With bottom matching (@option{field}=@var{bottom}):
10906 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10907 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10915 With top matching (@option{field}=@var{top}):
10920 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10921 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10929 @subsection Examples
10931 Simple IVTC of a top field first telecined stream:
10933 fieldmatch=order=tff:combmatch=none, decimate
10936 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
10938 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
10941 @section fieldorder
10943 Transform the field order of the input video.
10945 It accepts the following parameters:
10950 The output field order. Valid values are @var{tff} for top field first or @var{bff}
10951 for bottom field first.
10954 The default value is @samp{tff}.
10956 The transformation is done by shifting the picture content up or down
10957 by one line, and filling the remaining line with appropriate picture content.
10958 This method is consistent with most broadcast field order converters.
10960 If the input video is not flagged as being interlaced, or it is already
10961 flagged as being of the required output field order, then this filter does
10962 not alter the incoming video.
10964 It is very useful when converting to or from PAL DV material,
10965 which is bottom field first.
10969 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
10972 @section fifo, afifo
10974 Buffer input images and send them when they are requested.
10976 It is mainly useful when auto-inserted by the libavfilter
10979 It does not take parameters.
10981 @section fillborders
10983 Fill borders of the input video, without changing video stream dimensions.
10984 Sometimes video can have garbage at the four edges and you may not want to
10985 crop video input to keep size multiple of some number.
10987 This filter accepts the following options:
10991 Number of pixels to fill from left border.
10994 Number of pixels to fill from right border.
10997 Number of pixels to fill from top border.
11000 Number of pixels to fill from bottom border.
11005 It accepts the following values:
11008 fill pixels using outermost pixels
11011 fill pixels using mirroring
11014 fill pixels with constant value
11017 Default is @var{smear}.
11020 Set color for pixels in fixed mode. Default is @var{black}.
11023 @subsection Commands
11024 This filter supports same @ref{commands} as options.
11025 The command accepts the same syntax of the corresponding option.
11027 If the specified expression is not valid, it is kept at its current
11032 Find a rectangular object
11034 It accepts the following options:
11038 Filepath of the object image, needs to be in gray8.
11041 Detection threshold, default is 0.5.
11044 Number of mipmaps, default is 3.
11046 @item xmin, ymin, xmax, ymax
11047 Specifies the rectangle in which to search.
11050 @subsection Examples
11054 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11056 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11062 Flood area with values of same pixel components with another values.
11064 It accepts the following options:
11067 Set pixel x coordinate.
11070 Set pixel y coordinate.
11073 Set source #0 component value.
11076 Set source #1 component value.
11079 Set source #2 component value.
11082 Set source #3 component value.
11085 Set destination #0 component value.
11088 Set destination #1 component value.
11091 Set destination #2 component value.
11094 Set destination #3 component value.
11100 Convert the input video to one of the specified pixel formats.
11101 Libavfilter will try to pick one that is suitable as input to
11104 It accepts the following parameters:
11108 A '|'-separated list of pixel format names, such as
11109 "pix_fmts=yuv420p|monow|rgb24".
11113 @subsection Examples
11117 Convert the input video to the @var{yuv420p} format
11119 format=pix_fmts=yuv420p
11122 Convert the input video to any of the formats in the list
11124 format=pix_fmts=yuv420p|yuv444p|yuv410p
11131 Convert the video to specified constant frame rate by duplicating or dropping
11132 frames as necessary.
11134 It accepts the following parameters:
11138 The desired output frame rate. The default is @code{25}.
11141 Assume the first PTS should be the given value, in seconds. This allows for
11142 padding/trimming at the start of stream. By default, no assumption is made
11143 about the first frame's expected PTS, so no padding or trimming is done.
11144 For example, this could be set to 0 to pad the beginning with duplicates of
11145 the first frame if a video stream starts after the audio stream or to trim any
11146 frames with a negative PTS.
11149 Timestamp (PTS) rounding method.
11151 Possible values are:
11158 round towards -infinity
11160 round towards +infinity
11164 The default is @code{near}.
11167 Action performed when reading the last frame.
11169 Possible values are:
11172 Use same timestamp rounding method as used for other frames.
11174 Pass through last frame if input duration has not been reached yet.
11176 The default is @code{round}.
11180 Alternatively, the options can be specified as a flat string:
11181 @var{fps}[:@var{start_time}[:@var{round}]].
11183 See also the @ref{setpts} filter.
11185 @subsection Examples
11189 A typical usage in order to set the fps to 25:
11195 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11197 fps=fps=film:round=near
11203 Pack two different video streams into a stereoscopic video, setting proper
11204 metadata on supported codecs. The two views should have the same size and
11205 framerate and processing will stop when the shorter video ends. Please note
11206 that you may conveniently adjust view properties with the @ref{scale} and
11209 It accepts the following parameters:
11213 The desired packing format. Supported values are:
11218 The views are next to each other (default).
11221 The views are on top of each other.
11224 The views are packed by line.
11227 The views are packed by column.
11230 The views are temporally interleaved.
11239 # Convert left and right views into a frame-sequential video
11240 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11242 # Convert views into a side-by-side video with the same output resolution as the input
11243 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
11248 Change the frame rate by interpolating new video output frames from the source
11251 This filter is not designed to function correctly with interlaced media. If
11252 you wish to change the frame rate of interlaced media then you are required
11253 to deinterlace before this filter and re-interlace after this filter.
11255 A description of the accepted options follows.
11259 Specify the output frames per second. This option can also be specified
11260 as a value alone. The default is @code{50}.
11263 Specify the start of a range where the output frame will be created as a
11264 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11265 the default is @code{15}.
11268 Specify the end of a range where the output frame will be created as a
11269 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11270 the default is @code{240}.
11273 Specify the level at which a scene change is detected as a value between
11274 0 and 100 to indicate a new scene; a low value reflects a low
11275 probability for the current frame to introduce a new scene, while a higher
11276 value means the current frame is more likely to be one.
11277 The default is @code{8.2}.
11280 Specify flags influencing the filter process.
11282 Available value for @var{flags} is:
11285 @item scene_change_detect, scd
11286 Enable scene change detection using the value of the option @var{scene}.
11287 This flag is enabled by default.
11293 Select one frame every N-th frame.
11295 This filter accepts the following option:
11298 Select frame after every @code{step} frames.
11299 Allowed values are positive integers higher than 0. Default value is @code{1}.
11302 @section freezedetect
11304 Detect frozen video.
11306 This filter logs a message and sets frame metadata when it detects that the
11307 input video has no significant change in content during a specified duration.
11308 Video freeze detection calculates the mean average absolute difference of all
11309 the components of video frames and compares it to a noise floor.
11311 The printed times and duration are expressed in seconds. The
11312 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
11313 whose timestamp equals or exceeds the detection duration and it contains the
11314 timestamp of the first frame of the freeze. The
11315 @code{lavfi.freezedetect.freeze_duration} and
11316 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
11319 The filter accepts the following options:
11323 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
11324 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
11328 Set freeze duration until notification (default is 2 seconds).
11331 @section freezeframes
11333 Freeze video frames.
11335 This filter freezes video frames using frame from 2nd input.
11337 The filter accepts the following options:
11341 Set number of first frame from which to start freeze.
11344 Set number of last frame from which to end freeze.
11347 Set number of frame from 2nd input which will be used instead of replaced frames.
11353 Apply a frei0r effect to the input video.
11355 To enable the compilation of this filter, you need to install the frei0r
11356 header and configure FFmpeg with @code{--enable-frei0r}.
11358 It accepts the following parameters:
11363 The name of the frei0r effect to load. If the environment variable
11364 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
11365 directories specified by the colon-separated list in @env{FREI0R_PATH}.
11366 Otherwise, the standard frei0r paths are searched, in this order:
11367 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
11368 @file{/usr/lib/frei0r-1/}.
11370 @item filter_params
11371 A '|'-separated list of parameters to pass to the frei0r effect.
11375 A frei0r effect parameter can be a boolean (its value is either
11376 "y" or "n"), a double, a color (specified as
11377 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
11378 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
11379 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
11380 a position (specified as @var{X}/@var{Y}, where
11381 @var{X} and @var{Y} are floating point numbers) and/or a string.
11383 The number and types of parameters depend on the loaded effect. If an
11384 effect parameter is not specified, the default value is set.
11386 @subsection Examples
11390 Apply the distort0r effect, setting the first two double parameters:
11392 frei0r=filter_name=distort0r:filter_params=0.5|0.01
11396 Apply the colordistance effect, taking a color as the first parameter:
11398 frei0r=colordistance:0.2/0.3/0.4
11399 frei0r=colordistance:violet
11400 frei0r=colordistance:0x112233
11404 Apply the perspective effect, specifying the top left and top right image
11407 frei0r=perspective:0.2/0.2|0.8/0.2
11411 For more information, see
11412 @url{http://frei0r.dyne.org}
11416 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
11418 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
11419 processing filter, one of them is performed once per block, not per pixel.
11420 This allows for much higher speed.
11422 The filter accepts the following options:
11426 Set quality. This option defines the number of levels for averaging. It accepts
11427 an integer in the range 4-5. Default value is @code{4}.
11430 Force a constant quantization parameter. It accepts an integer in range 0-63.
11431 If not set, the filter will use the QP from the video stream (if available).
11434 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
11435 more details but also more artifacts, while higher values make the image smoother
11436 but also blurrier. Default value is @code{0} − PSNR optimal.
11438 @item use_bframe_qp
11439 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
11440 option may cause flicker since the B-Frames have often larger QP. Default is
11441 @code{0} (not enabled).
11447 Apply Gaussian blur filter.
11449 The filter accepts the following options:
11453 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
11456 Set number of steps for Gaussian approximation. Default is @code{1}.
11459 Set which planes to filter. By default all planes are filtered.
11462 Set vertical sigma, if negative it will be same as @code{sigma}.
11463 Default is @code{-1}.
11466 @subsection Commands
11467 This filter supports same commands as options.
11468 The command accepts the same syntax of the corresponding option.
11470 If the specified expression is not valid, it is kept at its current
11475 Apply generic equation to each pixel.
11477 The filter accepts the following options:
11480 @item lum_expr, lum
11481 Set the luminance expression.
11483 Set the chrominance blue expression.
11485 Set the chrominance red expression.
11486 @item alpha_expr, a
11487 Set the alpha expression.
11489 Set the red expression.
11490 @item green_expr, g
11491 Set the green expression.
11493 Set the blue expression.
11496 The colorspace is selected according to the specified options. If one
11497 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
11498 options is specified, the filter will automatically select a YCbCr
11499 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
11500 @option{blue_expr} options is specified, it will select an RGB
11503 If one of the chrominance expression is not defined, it falls back on the other
11504 one. If no alpha expression is specified it will evaluate to opaque value.
11505 If none of chrominance expressions are specified, they will evaluate
11506 to the luminance expression.
11508 The expressions can use the following variables and functions:
11512 The sequential number of the filtered frame, starting from @code{0}.
11516 The coordinates of the current sample.
11520 The width and height of the image.
11524 Width and height scale depending on the currently filtered plane. It is the
11525 ratio between the corresponding luma plane number of pixels and the current
11526 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
11527 @code{0.5,0.5} for chroma planes.
11530 Time of the current frame, expressed in seconds.
11533 Return the value of the pixel at location (@var{x},@var{y}) of the current
11537 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
11541 Return the value of the pixel at location (@var{x},@var{y}) of the
11542 blue-difference chroma plane. Return 0 if there is no such plane.
11545 Return the value of the pixel at location (@var{x},@var{y}) of the
11546 red-difference chroma plane. Return 0 if there is no such plane.
11551 Return the value of the pixel at location (@var{x},@var{y}) of the
11552 red/green/blue component. Return 0 if there is no such component.
11555 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
11556 plane. Return 0 if there is no such plane.
11558 @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)
11559 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
11560 sums of samples within a rectangle. See the functions without the sum postfix.
11562 @item interpolation
11563 Set one of interpolation methods:
11568 Default is bilinear.
11571 For functions, if @var{x} and @var{y} are outside the area, the value will be
11572 automatically clipped to the closer edge.
11574 Please note that this filter can use multiple threads in which case each slice
11575 will have its own expression state. If you want to use only a single expression
11576 state because your expressions depend on previous state then you should limit
11577 the number of filter threads to 1.
11579 @subsection Examples
11583 Flip the image horizontally:
11589 Generate a bidimensional sine wave, with angle @code{PI/3} and a
11590 wavelength of 100 pixels:
11592 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
11596 Generate a fancy enigmatic moving light:
11598 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
11602 Generate a quick emboss effect:
11604 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
11608 Modify RGB components depending on pixel position:
11610 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
11614 Create a radial gradient that is the same size as the input (also see
11615 the @ref{vignette} filter):
11617 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
11623 Fix the banding artifacts that are sometimes introduced into nearly flat
11624 regions by truncation to 8-bit color depth.
11625 Interpolate the gradients that should go where the bands are, and
11628 It is designed for playback only. Do not use it prior to
11629 lossy compression, because compression tends to lose the dither and
11630 bring back the bands.
11632 It accepts the following parameters:
11637 The maximum amount by which the filter will change any one pixel. This is also
11638 the threshold for detecting nearly flat regions. Acceptable values range from
11639 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
11643 The neighborhood to fit the gradient to. A larger radius makes for smoother
11644 gradients, but also prevents the filter from modifying the pixels near detailed
11645 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
11646 values will be clipped to the valid range.
11650 Alternatively, the options can be specified as a flat string:
11651 @var{strength}[:@var{radius}]
11653 @subsection Examples
11657 Apply the filter with a @code{3.5} strength and radius of @code{8}:
11663 Specify radius, omitting the strength (which will fall-back to the default
11671 @anchor{graphmonitor}
11672 @section graphmonitor
11673 Show various filtergraph stats.
11675 With this filter one can debug complete filtergraph.
11676 Especially issues with links filling with queued frames.
11678 The filter accepts the following options:
11682 Set video output size. Default is @var{hd720}.
11685 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
11688 Set output mode, can be @var{fulll} or @var{compact}.
11689 In @var{compact} mode only filters with some queued frames have displayed stats.
11692 Set flags which enable which stats are shown in video.
11694 Available values for flags are:
11697 Display number of queued frames in each link.
11699 @item frame_count_in
11700 Display number of frames taken from filter.
11702 @item frame_count_out
11703 Display number of frames given out from filter.
11706 Display current filtered frame pts.
11709 Display current filtered frame time.
11712 Display time base for filter link.
11715 Display used format for filter link.
11718 Display video size or number of audio channels in case of audio used by filter link.
11721 Display video frame rate or sample rate in case of audio used by filter link.
11725 Set upper limit for video rate of output stream, Default value is @var{25}.
11726 This guarantee that output video frame rate will not be higher than this value.
11730 A color constancy variation filter which estimates scene illumination via grey edge algorithm
11731 and corrects the scene colors accordingly.
11733 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
11735 The filter accepts the following options:
11739 The order of differentiation to be applied on the scene. Must be chosen in the range
11740 [0,2] and default value is 1.
11743 The Minkowski parameter to be used for calculating the Minkowski distance. Must
11744 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
11745 max value instead of calculating Minkowski distance.
11748 The standard deviation of Gaussian blur to be applied on the scene. Must be
11749 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
11750 can't be equal to 0 if @var{difford} is greater than 0.
11753 @subsection Examples
11759 greyedge=difford=1:minknorm=5:sigma=2
11765 greyedge=difford=1:minknorm=0:sigma=2
11773 Apply a Hald CLUT to a video stream.
11775 First input is the video stream to process, and second one is the Hald CLUT.
11776 The Hald CLUT input can be a simple picture or a complete video stream.
11778 The filter accepts the following options:
11782 Force termination when the shortest input terminates. Default is @code{0}.
11784 Continue applying the last CLUT after the end of the stream. A value of
11785 @code{0} disable the filter after the last frame of the CLUT is reached.
11786 Default is @code{1}.
11789 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
11790 filters share the same internals).
11792 This filter also supports the @ref{framesync} options.
11794 More information about the Hald CLUT can be found on Eskil Steenberg's website
11795 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
11797 @subsection Workflow examples
11799 @subsubsection Hald CLUT video stream
11801 Generate an identity Hald CLUT stream altered with various effects:
11803 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
11806 Note: make sure you use a lossless codec.
11808 Then use it with @code{haldclut} to apply it on some random stream:
11810 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
11813 The Hald CLUT will be applied to the 10 first seconds (duration of
11814 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
11815 to the remaining frames of the @code{mandelbrot} stream.
11817 @subsubsection Hald CLUT with preview
11819 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
11820 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
11821 biggest possible square starting at the top left of the picture. The remaining
11822 padding pixels (bottom or right) will be ignored. This area can be used to add
11823 a preview of the Hald CLUT.
11825 Typically, the following generated Hald CLUT will be supported by the
11826 @code{haldclut} filter:
11829 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
11830 pad=iw+320 [padded_clut];
11831 smptebars=s=320x256, split [a][b];
11832 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
11833 [main][b] overlay=W-320" -frames:v 1 clut.png
11836 It contains the original and a preview of the effect of the CLUT: SMPTE color
11837 bars are displayed on the right-top, and below the same color bars processed by
11840 Then, the effect of this Hald CLUT can be visualized with:
11842 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
11847 Flip the input video horizontally.
11849 For example, to horizontally flip the input video with @command{ffmpeg}:
11851 ffmpeg -i in.avi -vf "hflip" out.avi
11855 This filter applies a global color histogram equalization on a
11858 It can be used to correct video that has a compressed range of pixel
11859 intensities. The filter redistributes the pixel intensities to
11860 equalize their distribution across the intensity range. It may be
11861 viewed as an "automatically adjusting contrast filter". This filter is
11862 useful only for correcting degraded or poorly captured source
11865 The filter accepts the following options:
11869 Determine the amount of equalization to be applied. As the strength
11870 is reduced, the distribution of pixel intensities more-and-more
11871 approaches that of the input frame. The value must be a float number
11872 in the range [0,1] and defaults to 0.200.
11875 Set the maximum intensity that can generated and scale the output
11876 values appropriately. The strength should be set as desired and then
11877 the intensity can be limited if needed to avoid washing-out. The value
11878 must be a float number in the range [0,1] and defaults to 0.210.
11881 Set the antibanding level. If enabled the filter will randomly vary
11882 the luminance of output pixels by a small amount to avoid banding of
11883 the histogram. Possible values are @code{none}, @code{weak} or
11884 @code{strong}. It defaults to @code{none}.
11890 Compute and draw a color distribution histogram for the input video.
11892 The computed histogram is a representation of the color component
11893 distribution in an image.
11895 Standard histogram displays the color components distribution in an image.
11896 Displays color graph for each color component. Shows distribution of
11897 the Y, U, V, A or R, G, B components, depending on input format, in the
11898 current frame. Below each graph a color component scale meter is shown.
11900 The filter accepts the following options:
11904 Set height of level. Default value is @code{200}.
11905 Allowed range is [50, 2048].
11908 Set height of color scale. Default value is @code{12}.
11909 Allowed range is [0, 40].
11913 It accepts the following values:
11916 Per color component graphs are placed below each other.
11919 Per color component graphs are placed side by side.
11922 Presents information identical to that in the @code{parade}, except
11923 that the graphs representing color components are superimposed directly
11926 Default is @code{stack}.
11929 Set mode. Can be either @code{linear}, or @code{logarithmic}.
11930 Default is @code{linear}.
11933 Set what color components to display.
11934 Default is @code{7}.
11937 Set foreground opacity. Default is @code{0.7}.
11940 Set background opacity. Default is @code{0.5}.
11943 @subsection Examples
11948 Calculate and draw histogram:
11950 ffplay -i input -vf histogram
11958 This is a high precision/quality 3d denoise filter. It aims to reduce
11959 image noise, producing smooth images and making still images really
11960 still. It should enhance compressibility.
11962 It accepts the following optional parameters:
11966 A non-negative floating point number which specifies spatial luma strength.
11967 It defaults to 4.0.
11969 @item chroma_spatial
11970 A non-negative floating point number which specifies spatial chroma strength.
11971 It defaults to 3.0*@var{luma_spatial}/4.0.
11974 A floating point number which specifies luma temporal strength. It defaults to
11975 6.0*@var{luma_spatial}/4.0.
11978 A floating point number which specifies chroma temporal strength. It defaults to
11979 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
11982 @subsection Commands
11983 This filter supports same @ref{commands} as options.
11984 The command accepts the same syntax of the corresponding option.
11986 If the specified expression is not valid, it is kept at its current
11989 @anchor{hwdownload}
11990 @section hwdownload
11992 Download hardware frames to system memory.
11994 The input must be in hardware frames, and the output a non-hardware format.
11995 Not all formats will be supported on the output - it may be necessary to insert
11996 an additional @option{format} filter immediately following in the graph to get
11997 the output in a supported format.
12001 Map hardware frames to system memory or to another device.
12003 This filter has several different modes of operation; which one is used depends
12004 on the input and output formats:
12007 Hardware frame input, normal frame output
12009 Map the input frames to system memory and pass them to the output. If the
12010 original hardware frame is later required (for example, after overlaying
12011 something else on part of it), the @option{hwmap} filter can be used again
12012 in the next mode to retrieve it.
12014 Normal frame input, hardware frame output
12016 If the input is actually a software-mapped hardware frame, then unmap it -
12017 that is, return the original hardware frame.
12019 Otherwise, a device must be provided. Create new hardware surfaces on that
12020 device for the output, then map them back to the software format at the input
12021 and give those frames to the preceding filter. This will then act like the
12022 @option{hwupload} filter, but may be able to avoid an additional copy when
12023 the input is already in a compatible format.
12025 Hardware frame input and output
12027 A device must be supplied for the output, either directly or with the
12028 @option{derive_device} option. The input and output devices must be of
12029 different types and compatible - the exact meaning of this is
12030 system-dependent, but typically it means that they must refer to the same
12031 underlying hardware context (for example, refer to the same graphics card).
12033 If the input frames were originally created on the output device, then unmap
12034 to retrieve the original frames.
12036 Otherwise, map the frames to the output device - create new hardware frames
12037 on the output corresponding to the frames on the input.
12040 The following additional parameters are accepted:
12044 Set the frame mapping mode. Some combination of:
12047 The mapped frame should be readable.
12049 The mapped frame should be writeable.
12051 The mapping will always overwrite the entire frame.
12053 This may improve performance in some cases, as the original contents of the
12054 frame need not be loaded.
12056 The mapping must not involve any copying.
12058 Indirect mappings to copies of frames are created in some cases where either
12059 direct mapping is not possible or it would have unexpected properties.
12060 Setting this flag ensures that the mapping is direct and will fail if that is
12063 Defaults to @var{read+write} if not specified.
12065 @item derive_device @var{type}
12066 Rather than using the device supplied at initialisation, instead derive a new
12067 device of type @var{type} from the device the input frames exist on.
12070 In a hardware to hardware mapping, map in reverse - create frames in the sink
12071 and map them back to the source. This may be necessary in some cases where
12072 a mapping in one direction is required but only the opposite direction is
12073 supported by the devices being used.
12075 This option is dangerous - it may break the preceding filter in undefined
12076 ways if there are any additional constraints on that filter's output.
12077 Do not use it without fully understanding the implications of its use.
12083 Upload system memory frames to hardware surfaces.
12085 The device to upload to must be supplied when the filter is initialised. If
12086 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12087 option or with the @option{derive_device} option. The input and output devices
12088 must be of different types and compatible - the exact meaning of this is
12089 system-dependent, but typically it means that they must refer to the same
12090 underlying hardware context (for example, refer to the same graphics card).
12092 The following additional parameters are accepted:
12095 @item derive_device @var{type}
12096 Rather than using the device supplied at initialisation, instead derive a new
12097 device of type @var{type} from the device the input frames exist on.
12100 @anchor{hwupload_cuda}
12101 @section hwupload_cuda
12103 Upload system memory frames to a CUDA device.
12105 It accepts the following optional parameters:
12109 The number of the CUDA device to use
12114 Apply a high-quality magnification filter designed for pixel art. This filter
12115 was originally created by Maxim Stepin.
12117 It accepts the following option:
12121 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12122 @code{hq3x} and @code{4} for @code{hq4x}.
12123 Default is @code{3}.
12127 Stack input videos horizontally.
12129 All streams must be of same pixel format and of same height.
12131 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12132 to create same output.
12134 The filter accepts the following option:
12138 Set number of input streams. Default is 2.
12141 If set to 1, force the output to terminate when the shortest input
12142 terminates. Default value is 0.
12147 Modify the hue and/or the saturation of the input.
12149 It accepts the following parameters:
12153 Specify the hue angle as a number of degrees. It accepts an expression,
12154 and defaults to "0".
12157 Specify the saturation in the [-10,10] range. It accepts an expression and
12161 Specify the hue angle as a number of radians. It accepts an
12162 expression, and defaults to "0".
12165 Specify the brightness in the [-10,10] range. It accepts an expression and
12169 @option{h} and @option{H} are mutually exclusive, and can't be
12170 specified at the same time.
12172 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12173 expressions containing the following constants:
12177 frame count of the input frame starting from 0
12180 presentation timestamp of the input frame expressed in time base units
12183 frame rate of the input video, NAN if the input frame rate is unknown
12186 timestamp expressed in seconds, NAN if the input timestamp is unknown
12189 time base of the input video
12192 @subsection Examples
12196 Set the hue to 90 degrees and the saturation to 1.0:
12202 Same command but expressing the hue in radians:
12208 Rotate hue and make the saturation swing between 0
12209 and 2 over a period of 1 second:
12211 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12215 Apply a 3 seconds saturation fade-in effect starting at 0:
12217 hue="s=min(t/3\,1)"
12220 The general fade-in expression can be written as:
12222 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12226 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12228 hue="s=max(0\, min(1\, (8-t)/3))"
12231 The general fade-out expression can be written as:
12233 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12238 @subsection Commands
12240 This filter supports the following commands:
12246 Modify the hue and/or the saturation and/or brightness of the input video.
12247 The command accepts the same syntax of the corresponding option.
12249 If the specified expression is not valid, it is kept at its current
12253 @section hysteresis
12255 Grow first stream into second stream by connecting components.
12256 This makes it possible to build more robust edge masks.
12258 This filter accepts the following options:
12262 Set which planes will be processed as bitmap, unprocessed planes will be
12263 copied from first stream.
12264 By default value 0xf, all planes will be processed.
12267 Set threshold which is used in filtering. If pixel component value is higher than
12268 this value filter algorithm for connecting components is activated.
12269 By default value is 0.
12272 The @code{hysteresis} filter also supports the @ref{framesync} options.
12276 Detect video interlacing type.
12278 This filter tries to detect if the input frames are interlaced, progressive,
12279 top or bottom field first. It will also try to detect fields that are
12280 repeated between adjacent frames (a sign of telecine).
12282 Single frame detection considers only immediately adjacent frames when classifying each frame.
12283 Multiple frame detection incorporates the classification history of previous frames.
12285 The filter will log these metadata values:
12288 @item single.current_frame
12289 Detected type of current frame using single-frame detection. One of:
12290 ``tff'' (top field first), ``bff'' (bottom field first),
12291 ``progressive'', or ``undetermined''
12294 Cumulative number of frames detected as top field first using single-frame detection.
12297 Cumulative number of frames detected as top field first using multiple-frame detection.
12300 Cumulative number of frames detected as bottom field first using single-frame detection.
12302 @item multiple.current_frame
12303 Detected type of current frame using multiple-frame detection. One of:
12304 ``tff'' (top field first), ``bff'' (bottom field first),
12305 ``progressive'', or ``undetermined''
12308 Cumulative number of frames detected as bottom field first using multiple-frame detection.
12310 @item single.progressive
12311 Cumulative number of frames detected as progressive using single-frame detection.
12313 @item multiple.progressive
12314 Cumulative number of frames detected as progressive using multiple-frame detection.
12316 @item single.undetermined
12317 Cumulative number of frames that could not be classified using single-frame detection.
12319 @item multiple.undetermined
12320 Cumulative number of frames that could not be classified using multiple-frame detection.
12322 @item repeated.current_frame
12323 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
12325 @item repeated.neither
12326 Cumulative number of frames with no repeated field.
12329 Cumulative number of frames with the top field repeated from the previous frame's top field.
12331 @item repeated.bottom
12332 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
12335 The filter accepts the following options:
12339 Set interlacing threshold.
12341 Set progressive threshold.
12343 Threshold for repeated field detection.
12345 Number of frames after which a given frame's contribution to the
12346 statistics is halved (i.e., it contributes only 0.5 to its
12347 classification). The default of 0 means that all frames seen are given
12348 full weight of 1.0 forever.
12349 @item analyze_interlaced_flag
12350 When this is not 0 then idet will use the specified number of frames to determine
12351 if the interlaced flag is accurate, it will not count undetermined frames.
12352 If the flag is found to be accurate it will be used without any further
12353 computations, if it is found to be inaccurate it will be cleared without any
12354 further computations. This allows inserting the idet filter as a low computational
12355 method to clean up the interlaced flag
12360 Deinterleave or interleave fields.
12362 This filter allows one to process interlaced images fields without
12363 deinterlacing them. Deinterleaving splits the input frame into 2
12364 fields (so called half pictures). Odd lines are moved to the top
12365 half of the output image, even lines to the bottom half.
12366 You can process (filter) them independently and then re-interleave them.
12368 The filter accepts the following options:
12372 @item chroma_mode, c
12373 @item alpha_mode, a
12374 Available values for @var{luma_mode}, @var{chroma_mode} and
12375 @var{alpha_mode} are:
12381 @item deinterleave, d
12382 Deinterleave fields, placing one above the other.
12384 @item interleave, i
12385 Interleave fields. Reverse the effect of deinterleaving.
12387 Default value is @code{none}.
12389 @item luma_swap, ls
12390 @item chroma_swap, cs
12391 @item alpha_swap, as
12392 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
12395 @subsection Commands
12397 This filter supports the all above options as @ref{commands}.
12401 Apply inflate effect to the video.
12403 This filter replaces the pixel by the local(3x3) average by taking into account
12404 only values higher than the pixel.
12406 It accepts the following options:
12413 Limit the maximum change for each plane, default is 65535.
12414 If 0, plane will remain unchanged.
12417 @subsection Commands
12419 This filter supports the all above options as @ref{commands}.
12423 Simple interlacing filter from progressive contents. This interleaves upper (or
12424 lower) lines from odd frames with lower (or upper) lines from even frames,
12425 halving the frame rate and preserving image height.
12428 Original Original New Frame
12429 Frame 'j' Frame 'j+1' (tff)
12430 ========== =========== ==================
12431 Line 0 --------------------> Frame 'j' Line 0
12432 Line 1 Line 1 ----> Frame 'j+1' Line 1
12433 Line 2 ---------------------> Frame 'j' Line 2
12434 Line 3 Line 3 ----> Frame 'j+1' Line 3
12436 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
12439 It accepts the following optional parameters:
12443 This determines whether the interlaced frame is taken from the even
12444 (tff - default) or odd (bff) lines of the progressive frame.
12447 Vertical lowpass filter to avoid twitter interlacing and
12448 reduce moire patterns.
12452 Disable vertical lowpass filter
12455 Enable linear filter (default)
12458 Enable complex filter. This will slightly less reduce twitter and moire
12459 but better retain detail and subjective sharpness impression.
12466 Deinterlace input video by applying Donald Graft's adaptive kernel
12467 deinterling. Work on interlaced parts of a video to produce
12468 progressive frames.
12470 The description of the accepted parameters follows.
12474 Set the threshold which affects the filter's tolerance when
12475 determining if a pixel line must be processed. It must be an integer
12476 in the range [0,255] and defaults to 10. A value of 0 will result in
12477 applying the process on every pixels.
12480 Paint pixels exceeding the threshold value to white if set to 1.
12484 Set the fields order. Swap fields if set to 1, leave fields alone if
12488 Enable additional sharpening if set to 1. Default is 0.
12491 Enable twoway sharpening if set to 1. Default is 0.
12494 @subsection Examples
12498 Apply default values:
12500 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
12504 Enable additional sharpening:
12510 Paint processed pixels in white:
12518 Slowly update darker pixels.
12520 This filter makes short flashes of light appear longer.
12521 This filter accepts the following options:
12525 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
12528 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
12531 @section lenscorrection
12533 Correct radial lens distortion
12535 This filter can be used to correct for radial distortion as can result from the use
12536 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
12537 one can use tools available for example as part of opencv or simply trial-and-error.
12538 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
12539 and extract the k1 and k2 coefficients from the resulting matrix.
12541 Note that effectively the same filter is available in the open-source tools Krita and
12542 Digikam from the KDE project.
12544 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
12545 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
12546 brightness distribution, so you may want to use both filters together in certain
12547 cases, though you will have to take care of ordering, i.e. whether vignetting should
12548 be applied before or after lens correction.
12550 @subsection Options
12552 The filter accepts the following options:
12556 Relative x-coordinate of the focal point of the image, and thereby the center of the
12557 distortion. This value has a range [0,1] and is expressed as fractions of the image
12558 width. Default is 0.5.
12560 Relative y-coordinate of the focal point of the image, and thereby the center of the
12561 distortion. This value has a range [0,1] and is expressed as fractions of the image
12562 height. Default is 0.5.
12564 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
12565 no correction. Default is 0.
12567 Coefficient of the double quadratic correction term. This value has a range [-1,1].
12568 0 means no correction. Default is 0.
12571 The formula that generates the correction is:
12573 @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)
12575 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
12576 distances from the focal point in the source and target images, respectively.
12580 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
12582 The @code{lensfun} filter requires the camera make, camera model, and lens model
12583 to apply the lens correction. The filter will load the lensfun database and
12584 query it to find the corresponding camera and lens entries in the database. As
12585 long as these entries can be found with the given options, the filter can
12586 perform corrections on frames. Note that incomplete strings will result in the
12587 filter choosing the best match with the given options, and the filter will
12588 output the chosen camera and lens models (logged with level "info"). You must
12589 provide the make, camera model, and lens model as they are required.
12591 The filter accepts the following options:
12595 The make of the camera (for example, "Canon"). This option is required.
12598 The model of the camera (for example, "Canon EOS 100D"). This option is
12602 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
12603 option is required.
12606 The type of correction to apply. The following values are valid options:
12610 Enables fixing lens vignetting.
12613 Enables fixing lens geometry. This is the default.
12616 Enables fixing chromatic aberrations.
12619 Enables fixing lens vignetting and lens geometry.
12622 Enables fixing lens vignetting and chromatic aberrations.
12625 Enables fixing both lens geometry and chromatic aberrations.
12628 Enables all possible corrections.
12632 The focal length of the image/video (zoom; expected constant for video). For
12633 example, a 18--55mm lens has focal length range of [18--55], so a value in that
12634 range should be chosen when using that lens. Default 18.
12637 The aperture of the image/video (expected constant for video). Note that
12638 aperture is only used for vignetting correction. Default 3.5.
12640 @item focus_distance
12641 The focus distance of the image/video (expected constant for video). Note that
12642 focus distance is only used for vignetting and only slightly affects the
12643 vignetting correction process. If unknown, leave it at the default value (which
12647 The scale factor which is applied after transformation. After correction the
12648 video is no longer necessarily rectangular. This parameter controls how much of
12649 the resulting image is visible. The value 0 means that a value will be chosen
12650 automatically such that there is little or no unmapped area in the output
12651 image. 1.0 means that no additional scaling is done. Lower values may result
12652 in more of the corrected image being visible, while higher values may avoid
12653 unmapped areas in the output.
12655 @item target_geometry
12656 The target geometry of the output image/video. The following values are valid
12660 @item rectilinear (default)
12663 @item equirectangular
12664 @item fisheye_orthographic
12665 @item fisheye_stereographic
12666 @item fisheye_equisolid
12667 @item fisheye_thoby
12670 Apply the reverse of image correction (instead of correcting distortion, apply
12673 @item interpolation
12674 The type of interpolation used when correcting distortion. The following values
12679 @item linear (default)
12684 @subsection Examples
12688 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
12689 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
12693 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
12697 Apply the same as before, but only for the first 5 seconds of video.
12700 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
12707 Obtain the VMAF (Video Multi-Method Assessment Fusion)
12708 score between two input videos.
12710 The obtained VMAF score is printed through the logging system.
12712 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
12713 After installing the library it can be enabled using:
12714 @code{./configure --enable-libvmaf --enable-version3}.
12715 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
12717 The filter has following options:
12721 Set the model path which is to be used for SVM.
12722 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
12725 Set the file path to be used to store logs.
12728 Set the format of the log file (xml or json).
12730 @item enable_transform
12731 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
12732 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
12733 Default value: @code{false}
12736 Invokes the phone model which will generate VMAF scores higher than in the
12737 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
12738 Default value: @code{false}
12741 Enables computing psnr along with vmaf.
12742 Default value: @code{false}
12745 Enables computing ssim along with vmaf.
12746 Default value: @code{false}
12749 Enables computing ms_ssim along with vmaf.
12750 Default value: @code{false}
12753 Set the pool method to be used for computing vmaf.
12754 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
12757 Set number of threads to be used when computing vmaf.
12758 Default value: @code{0}, which makes use of all available logical processors.
12761 Set interval for frame subsampling used when computing vmaf.
12762 Default value: @code{1}
12764 @item enable_conf_interval
12765 Enables confidence interval.
12766 Default value: @code{false}
12769 This filter also supports the @ref{framesync} options.
12771 @subsection Examples
12774 On the below examples the input file @file{main.mpg} being processed is
12775 compared with the reference file @file{ref.mpg}.
12778 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
12782 Example with options:
12784 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
12788 Example with options and different containers:
12790 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 -
12796 Limits the pixel components values to the specified range [min, max].
12798 The filter accepts the following options:
12802 Lower bound. Defaults to the lowest allowed value for the input.
12805 Upper bound. Defaults to the highest allowed value for the input.
12808 Specify which planes will be processed. Defaults to all available.
12815 The filter accepts the following options:
12819 Set the number of loops. Setting this value to -1 will result in infinite loops.
12823 Set maximal size in number of frames. Default is 0.
12826 Set first frame of loop. Default is 0.
12829 @subsection Examples
12833 Loop single first frame infinitely:
12835 loop=loop=-1:size=1:start=0
12839 Loop single first frame 10 times:
12841 loop=loop=10:size=1:start=0
12845 Loop 10 first frames 5 times:
12847 loop=loop=5:size=10:start=0
12853 Apply a 1D LUT to an input video.
12855 The filter accepts the following options:
12859 Set the 1D LUT file name.
12861 Currently supported formats:
12870 Select interpolation mode.
12872 Available values are:
12876 Use values from the nearest defined point.
12878 Interpolate values using the linear interpolation.
12880 Interpolate values using the cosine interpolation.
12882 Interpolate values using the cubic interpolation.
12884 Interpolate values using the spline interpolation.
12891 Apply a 3D LUT to an input video.
12893 The filter accepts the following options:
12897 Set the 3D LUT file name.
12899 Currently supported formats:
12913 Select interpolation mode.
12915 Available values are:
12919 Use values from the nearest defined point.
12921 Interpolate values using the 8 points defining a cube.
12923 Interpolate values using a tetrahedron.
12929 Turn certain luma values into transparency.
12931 The filter accepts the following options:
12935 Set the luma which will be used as base for transparency.
12936 Default value is @code{0}.
12939 Set the range of luma values to be keyed out.
12940 Default value is @code{0.01}.
12943 Set the range of softness. Default value is @code{0}.
12944 Use this to control gradual transition from zero to full transparency.
12947 @subsection Commands
12948 This filter supports same @ref{commands} as options.
12949 The command accepts the same syntax of the corresponding option.
12951 If the specified expression is not valid, it is kept at its current
12954 @section lut, lutrgb, lutyuv
12956 Compute a look-up table for binding each pixel component input value
12957 to an output value, and apply it to the input video.
12959 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
12960 to an RGB input video.
12962 These filters accept the following parameters:
12965 set first pixel component expression
12967 set second pixel component expression
12969 set third pixel component expression
12971 set fourth pixel component expression, corresponds to the alpha component
12974 set red component expression
12976 set green component expression
12978 set blue component expression
12980 alpha component expression
12983 set Y/luminance component expression
12985 set U/Cb component expression
12987 set V/Cr component expression
12990 Each of them specifies the expression to use for computing the lookup table for
12991 the corresponding pixel component values.
12993 The exact component associated to each of the @var{c*} options depends on the
12996 The @var{lut} filter requires either YUV or RGB pixel formats in input,
12997 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
12999 The expressions can contain the following constants and functions:
13004 The input width and height.
13007 The input value for the pixel component.
13010 The input value, clipped to the @var{minval}-@var{maxval} range.
13013 The maximum value for the pixel component.
13016 The minimum value for the pixel component.
13019 The negated value for the pixel component value, clipped to the
13020 @var{minval}-@var{maxval} range; it corresponds to the expression
13021 "maxval-clipval+minval".
13024 The computed value in @var{val}, clipped to the
13025 @var{minval}-@var{maxval} range.
13027 @item gammaval(gamma)
13028 The computed gamma correction value of the pixel component value,
13029 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13031 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13035 All expressions default to "val".
13037 @subsection Examples
13041 Negate input video:
13043 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13044 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13047 The above is the same as:
13049 lutrgb="r=negval:g=negval:b=negval"
13050 lutyuv="y=negval:u=negval:v=negval"
13060 Remove chroma components, turning the video into a graytone image:
13062 lutyuv="u=128:v=128"
13066 Apply a luma burning effect:
13072 Remove green and blue components:
13078 Set a constant alpha channel value on input:
13080 format=rgba,lutrgb=a="maxval-minval/2"
13084 Correct luminance gamma by a factor of 0.5:
13086 lutyuv=y=gammaval(0.5)
13090 Discard least significant bits of luma:
13092 lutyuv=y='bitand(val, 128+64+32)'
13096 Technicolor like effect:
13098 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13102 @section lut2, tlut2
13104 The @code{lut2} filter takes two input streams and outputs one
13107 The @code{tlut2} (time lut2) filter takes two consecutive frames
13108 from one single stream.
13110 This filter accepts the following parameters:
13113 set first pixel component expression
13115 set second pixel component expression
13117 set third pixel component expression
13119 set fourth pixel component expression, corresponds to the alpha component
13122 set output bit depth, only available for @code{lut2} filter. By default is 0,
13123 which means bit depth is automatically picked from first input format.
13126 The @code{lut2} filter also supports the @ref{framesync} options.
13128 Each of them specifies the expression to use for computing the lookup table for
13129 the corresponding pixel component values.
13131 The exact component associated to each of the @var{c*} options depends on the
13134 The expressions can contain the following constants:
13139 The input width and height.
13142 The first input value for the pixel component.
13145 The second input value for the pixel component.
13148 The first input video bit depth.
13151 The second input video bit depth.
13154 All expressions default to "x".
13156 @subsection Examples
13160 Highlight differences between two RGB video streams:
13162 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)'
13166 Highlight differences between two YUV video streams:
13168 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)'
13172 Show max difference between two video streams:
13174 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)))'
13178 @section maskedclamp
13180 Clamp the first input stream with the second input and third input stream.
13182 Returns the value of first stream to be between second input
13183 stream - @code{undershoot} and third input stream + @code{overshoot}.
13185 This filter accepts the following options:
13188 Default value is @code{0}.
13191 Default value is @code{0}.
13194 Set which planes will be processed as bitmap, unprocessed planes will be
13195 copied from first stream.
13196 By default value 0xf, all planes will be processed.
13201 Merge the second and third input stream into output stream using absolute differences
13202 between second input stream and first input stream and absolute difference between
13203 third input stream and first input stream. The picked value will be from second input
13204 stream if second absolute difference is greater than first one or from third input stream
13207 This filter accepts the following options:
13210 Set which planes will be processed as bitmap, unprocessed planes will be
13211 copied from first stream.
13212 By default value 0xf, all planes will be processed.
13215 @section maskedmerge
13217 Merge the first input stream with the second input stream using per pixel
13218 weights in the third input stream.
13220 A value of 0 in the third stream pixel component means that pixel component
13221 from first stream is returned unchanged, while maximum value (eg. 255 for
13222 8-bit videos) means that pixel component from second stream is returned
13223 unchanged. Intermediate values define the amount of merging between both
13224 input stream's pixel components.
13226 This filter accepts the following options:
13229 Set which planes will be processed as bitmap, unprocessed planes will be
13230 copied from first stream.
13231 By default value 0xf, all planes will be processed.
13236 Merge the second and third input stream into output stream using absolute differences
13237 between second input stream and first input stream and absolute difference between
13238 third input stream and first input stream. The picked value will be from second input
13239 stream if second absolute difference is less than first one or from third input stream
13242 This filter accepts the following options:
13245 Set which planes will be processed as bitmap, unprocessed planes will be
13246 copied from first stream.
13247 By default value 0xf, all planes will be processed.
13251 Create mask from input video.
13253 For example it is useful to create motion masks after @code{tblend} filter.
13255 This filter accepts the following options:
13259 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
13262 Set high threshold. Any pixel component higher than this value will be set to max value
13263 allowed for current pixel format.
13266 Set planes to filter, by default all available planes are filtered.
13269 Fill all frame pixels with this value.
13272 Set max average pixel value for frame. If sum of all pixel components is higher that this
13273 average, output frame will be completely filled with value set by @var{fill} option.
13274 Typically useful for scene changes when used in combination with @code{tblend} filter.
13279 Apply motion-compensation deinterlacing.
13281 It needs one field per frame as input and must thus be used together
13282 with yadif=1/3 or equivalent.
13284 This filter accepts the following options:
13287 Set the deinterlacing mode.
13289 It accepts one of the following values:
13294 use iterative motion estimation
13296 like @samp{slow}, but use multiple reference frames.
13298 Default value is @samp{fast}.
13301 Set the picture field parity assumed for the input video. It must be
13302 one of the following values:
13306 assume top field first
13308 assume bottom field first
13311 Default value is @samp{bff}.
13314 Set per-block quantization parameter (QP) used by the internal
13317 Higher values should result in a smoother motion vector field but less
13318 optimal individual vectors. Default value is 1.
13323 Pick median pixel from certain rectangle defined by radius.
13325 This filter accepts the following options:
13329 Set horizontal radius size. Default value is @code{1}.
13330 Allowed range is integer from 1 to 127.
13333 Set which planes to process. Default is @code{15}, which is all available planes.
13336 Set vertical radius size. Default value is @code{0}.
13337 Allowed range is integer from 0 to 127.
13338 If it is 0, value will be picked from horizontal @code{radius} option.
13341 Set median percentile. Default value is @code{0.5}.
13342 Default value of @code{0.5} will pick always median values, while @code{0} will pick
13343 minimum values, and @code{1} maximum values.
13346 @subsection Commands
13347 This filter supports same @ref{commands} as options.
13348 The command accepts the same syntax of the corresponding option.
13350 If the specified expression is not valid, it is kept at its current
13353 @section mergeplanes
13355 Merge color channel components from several video streams.
13357 The filter accepts up to 4 input streams, and merge selected input
13358 planes to the output video.
13360 This filter accepts the following options:
13363 Set input to output plane mapping. Default is @code{0}.
13365 The mappings is specified as a bitmap. It should be specified as a
13366 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
13367 mapping for the first plane of the output stream. 'A' sets the number of
13368 the input stream to use (from 0 to 3), and 'a' the plane number of the
13369 corresponding input to use (from 0 to 3). The rest of the mappings is
13370 similar, 'Bb' describes the mapping for the output stream second
13371 plane, 'Cc' describes the mapping for the output stream third plane and
13372 'Dd' describes the mapping for the output stream fourth plane.
13375 Set output pixel format. Default is @code{yuva444p}.
13378 @subsection Examples
13382 Merge three gray video streams of same width and height into single video stream:
13384 [a0][a1][a2]mergeplanes=0x001020:yuv444p
13388 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
13390 [a0][a1]mergeplanes=0x00010210:yuva444p
13394 Swap Y and A plane in yuva444p stream:
13396 format=yuva444p,mergeplanes=0x03010200:yuva444p
13400 Swap U and V plane in yuv420p stream:
13402 format=yuv420p,mergeplanes=0x000201:yuv420p
13406 Cast a rgb24 clip to yuv444p:
13408 format=rgb24,mergeplanes=0x000102:yuv444p
13414 Estimate and export motion vectors using block matching algorithms.
13415 Motion vectors are stored in frame side data to be used by other filters.
13417 This filter accepts the following options:
13420 Specify the motion estimation method. Accepts one of the following values:
13424 Exhaustive search algorithm.
13426 Three step search algorithm.
13428 Two dimensional logarithmic search algorithm.
13430 New three step search algorithm.
13432 Four step search algorithm.
13434 Diamond search algorithm.
13436 Hexagon-based search algorithm.
13438 Enhanced predictive zonal search algorithm.
13440 Uneven multi-hexagon search algorithm.
13442 Default value is @samp{esa}.
13445 Macroblock size. Default @code{16}.
13448 Search parameter. Default @code{7}.
13451 @section midequalizer
13453 Apply Midway Image Equalization effect using two video streams.
13455 Midway Image Equalization adjusts a pair of images to have the same
13456 histogram, while maintaining their dynamics as much as possible. It's
13457 useful for e.g. matching exposures from a pair of stereo cameras.
13459 This filter has two inputs and one output, which must be of same pixel format, but
13460 may be of different sizes. The output of filter is first input adjusted with
13461 midway histogram of both inputs.
13463 This filter accepts the following option:
13467 Set which planes to process. Default is @code{15}, which is all available planes.
13470 @section minterpolate
13472 Convert the video to specified frame rate using motion interpolation.
13474 This filter accepts the following options:
13477 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}.
13480 Motion interpolation mode. Following values are accepted:
13483 Duplicate previous or next frame for interpolating new ones.
13485 Blend source frames. Interpolated frame is mean of previous and next frames.
13487 Motion compensated interpolation. Following options are effective when this mode is selected:
13491 Motion compensation mode. Following values are accepted:
13494 Overlapped block motion compensation.
13496 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
13498 Default mode is @samp{obmc}.
13501 Motion estimation mode. Following values are accepted:
13504 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
13506 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
13508 Default mode is @samp{bilat}.
13511 The algorithm to be used for motion estimation. Following values are accepted:
13514 Exhaustive search algorithm.
13516 Three step search algorithm.
13518 Two dimensional logarithmic search algorithm.
13520 New three step search algorithm.
13522 Four step search algorithm.
13524 Diamond search algorithm.
13526 Hexagon-based search algorithm.
13528 Enhanced predictive zonal search algorithm.
13530 Uneven multi-hexagon search algorithm.
13532 Default algorithm is @samp{epzs}.
13535 Macroblock size. Default @code{16}.
13538 Motion estimation search parameter. Default @code{32}.
13541 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).
13546 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:
13549 Disable scene change detection.
13551 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
13553 Default method is @samp{fdiff}.
13555 @item scd_threshold
13556 Scene change detection threshold. Default is @code{5.0}.
13561 Mix several video input streams into one video stream.
13563 A description of the accepted options follows.
13567 The number of inputs. If unspecified, it defaults to 2.
13570 Specify weight of each input video stream as sequence.
13571 Each weight is separated by space. If number of weights
13572 is smaller than number of @var{frames} last specified
13573 weight will be used for all remaining unset weights.
13576 Specify scale, if it is set it will be multiplied with sum
13577 of each weight multiplied with pixel values to give final destination
13578 pixel value. By default @var{scale} is auto scaled to sum of weights.
13581 Specify how end of stream is determined.
13584 The duration of the longest input. (default)
13587 The duration of the shortest input.
13590 The duration of the first input.
13594 @section mpdecimate
13596 Drop frames that do not differ greatly from the previous frame in
13597 order to reduce frame rate.
13599 The main use of this filter is for very-low-bitrate encoding
13600 (e.g. streaming over dialup modem), but it could in theory be used for
13601 fixing movies that were inverse-telecined incorrectly.
13603 A description of the accepted options follows.
13607 Set the maximum number of consecutive frames which can be dropped (if
13608 positive), or the minimum interval between dropped frames (if
13609 negative). If the value is 0, the frame is dropped disregarding the
13610 number of previous sequentially dropped frames.
13612 Default value is 0.
13617 Set the dropping threshold values.
13619 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
13620 represent actual pixel value differences, so a threshold of 64
13621 corresponds to 1 unit of difference for each pixel, or the same spread
13622 out differently over the block.
13624 A frame is a candidate for dropping if no 8x8 blocks differ by more
13625 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
13626 meaning the whole image) differ by more than a threshold of @option{lo}.
13628 Default value for @option{hi} is 64*12, default value for @option{lo} is
13629 64*5, and default value for @option{frac} is 0.33.
13635 Negate (invert) the input video.
13637 It accepts the following option:
13642 With value 1, it negates the alpha component, if present. Default value is 0.
13648 Denoise frames using Non-Local Means algorithm.
13650 Each pixel is adjusted by looking for other pixels with similar contexts. This
13651 context similarity is defined by comparing their surrounding patches of size
13652 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
13655 Note that the research area defines centers for patches, which means some
13656 patches will be made of pixels outside that research area.
13658 The filter accepts the following options.
13662 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
13665 Set patch size. Default is 7. Must be odd number in range [0, 99].
13668 Same as @option{p} but for chroma planes.
13670 The default value is @var{0} and means automatic.
13673 Set research size. Default is 15. Must be odd number in range [0, 99].
13676 Same as @option{r} but for chroma planes.
13678 The default value is @var{0} and means automatic.
13683 Deinterlace video using neural network edge directed interpolation.
13685 This filter accepts the following options:
13689 Mandatory option, without binary file filter can not work.
13690 Currently file can be found here:
13691 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
13694 Set which frames to deinterlace, by default it is @code{all}.
13695 Can be @code{all} or @code{interlaced}.
13698 Set mode of operation.
13700 Can be one of the following:
13704 Use frame flags, both fields.
13706 Use frame flags, single field.
13708 Use top field only.
13710 Use bottom field only.
13712 Use both fields, top first.
13714 Use both fields, bottom first.
13718 Set which planes to process, by default filter process all frames.
13721 Set size of local neighborhood around each pixel, used by the predictor neural
13724 Can be one of the following:
13737 Set the number of neurons in predictor neural network.
13738 Can be one of the following:
13749 Controls the number of different neural network predictions that are blended
13750 together to compute the final output value. Can be @code{fast}, default or
13754 Set which set of weights to use in the predictor.
13755 Can be one of the following:
13759 weights trained to minimize absolute error
13761 weights trained to minimize squared error
13765 Controls whether or not the prescreener neural network is used to decide
13766 which pixels should be processed by the predictor neural network and which
13767 can be handled by simple cubic interpolation.
13768 The prescreener is trained to know whether cubic interpolation will be
13769 sufficient for a pixel or whether it should be predicted by the predictor nn.
13770 The computational complexity of the prescreener nn is much less than that of
13771 the predictor nn. Since most pixels can be handled by cubic interpolation,
13772 using the prescreener generally results in much faster processing.
13773 The prescreener is pretty accurate, so the difference between using it and not
13774 using it is almost always unnoticeable.
13776 Can be one of the following:
13784 Default is @code{new}.
13787 Set various debugging flags.
13792 Force libavfilter not to use any of the specified pixel formats for the
13793 input to the next filter.
13795 It accepts the following parameters:
13799 A '|'-separated list of pixel format names, such as
13800 pix_fmts=yuv420p|monow|rgb24".
13804 @subsection Examples
13808 Force libavfilter to use a format different from @var{yuv420p} for the
13809 input to the vflip filter:
13811 noformat=pix_fmts=yuv420p,vflip
13815 Convert the input video to any of the formats not contained in the list:
13817 noformat=yuv420p|yuv444p|yuv410p
13823 Add noise on video input frame.
13825 The filter accepts the following options:
13833 Set noise seed for specific pixel component or all pixel components in case
13834 of @var{all_seed}. Default value is @code{123457}.
13836 @item all_strength, alls
13837 @item c0_strength, c0s
13838 @item c1_strength, c1s
13839 @item c2_strength, c2s
13840 @item c3_strength, c3s
13841 Set noise strength for specific pixel component or all pixel components in case
13842 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
13844 @item all_flags, allf
13845 @item c0_flags, c0f
13846 @item c1_flags, c1f
13847 @item c2_flags, c2f
13848 @item c3_flags, c3f
13849 Set pixel component flags or set flags for all components if @var{all_flags}.
13850 Available values for component flags are:
13853 averaged temporal noise (smoother)
13855 mix random noise with a (semi)regular pattern
13857 temporal noise (noise pattern changes between frames)
13859 uniform noise (gaussian otherwise)
13863 @subsection Examples
13865 Add temporal and uniform noise to input video:
13867 noise=alls=20:allf=t+u
13872 Normalize RGB video (aka histogram stretching, contrast stretching).
13873 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
13875 For each channel of each frame, the filter computes the input range and maps
13876 it linearly to the user-specified output range. The output range defaults
13877 to the full dynamic range from pure black to pure white.
13879 Temporal smoothing can be used on the input range to reduce flickering (rapid
13880 changes in brightness) caused when small dark or bright objects enter or leave
13881 the scene. This is similar to the auto-exposure (automatic gain control) on a
13882 video camera, and, like a video camera, it may cause a period of over- or
13883 under-exposure of the video.
13885 The R,G,B channels can be normalized independently, which may cause some
13886 color shifting, or linked together as a single channel, which prevents
13887 color shifting. Linked normalization preserves hue. Independent normalization
13888 does not, so it can be used to remove some color casts. Independent and linked
13889 normalization can be combined in any ratio.
13891 The normalize filter accepts the following options:
13896 Colors which define the output range. The minimum input value is mapped to
13897 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
13898 The defaults are black and white respectively. Specifying white for
13899 @var{blackpt} and black for @var{whitept} will give color-inverted,
13900 normalized video. Shades of grey can be used to reduce the dynamic range
13901 (contrast). Specifying saturated colors here can create some interesting
13905 The number of previous frames to use for temporal smoothing. The input range
13906 of each channel is smoothed using a rolling average over the current frame
13907 and the @var{smoothing} previous frames. The default is 0 (no temporal
13911 Controls the ratio of independent (color shifting) channel normalization to
13912 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
13913 independent. Defaults to 1.0 (fully independent).
13916 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
13917 expensive no-op. Defaults to 1.0 (full strength).
13921 @subsection Commands
13922 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
13923 The command accepts the same syntax of the corresponding option.
13925 If the specified expression is not valid, it is kept at its current
13928 @subsection Examples
13930 Stretch video contrast to use the full dynamic range, with no temporal
13931 smoothing; may flicker depending on the source content:
13933 normalize=blackpt=black:whitept=white:smoothing=0
13936 As above, but with 50 frames of temporal smoothing; flicker should be
13937 reduced, depending on the source content:
13939 normalize=blackpt=black:whitept=white:smoothing=50
13942 As above, but with hue-preserving linked channel normalization:
13944 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
13947 As above, but with half strength:
13949 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
13952 Map the darkest input color to red, the brightest input color to cyan:
13954 normalize=blackpt=red:whitept=cyan
13959 Pass the video source unchanged to the output.
13962 Optical Character Recognition
13964 This filter uses Tesseract for optical character recognition. To enable
13965 compilation of this filter, you need to configure FFmpeg with
13966 @code{--enable-libtesseract}.
13968 It accepts the following options:
13972 Set datapath to tesseract data. Default is to use whatever was
13973 set at installation.
13976 Set language, default is "eng".
13979 Set character whitelist.
13982 Set character blacklist.
13985 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
13986 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
13990 Apply a video transform using libopencv.
13992 To enable this filter, install the libopencv library and headers and
13993 configure FFmpeg with @code{--enable-libopencv}.
13995 It accepts the following parameters:
14000 The name of the libopencv filter to apply.
14002 @item filter_params
14003 The parameters to pass to the libopencv filter. If not specified, the default
14004 values are assumed.
14008 Refer to the official libopencv documentation for more precise
14010 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14012 Several libopencv filters are supported; see the following subsections.
14017 Dilate an image by using a specific structuring element.
14018 It corresponds to the libopencv function @code{cvDilate}.
14020 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14022 @var{struct_el} represents a structuring element, and has the syntax:
14023 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14025 @var{cols} and @var{rows} represent the number of columns and rows of
14026 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14027 point, and @var{shape} the shape for the structuring element. @var{shape}
14028 must be "rect", "cross", "ellipse", or "custom".
14030 If the value for @var{shape} is "custom", it must be followed by a
14031 string of the form "=@var{filename}". The file with name
14032 @var{filename} is assumed to represent a binary image, with each
14033 printable character corresponding to a bright pixel. When a custom
14034 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14035 or columns and rows of the read file are assumed instead.
14037 The default value for @var{struct_el} is "3x3+0x0/rect".
14039 @var{nb_iterations} specifies the number of times the transform is
14040 applied to the image, and defaults to 1.
14044 # Use the default values
14047 # Dilate using a structuring element with a 5x5 cross, iterating two times
14048 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14050 # Read the shape from the file diamond.shape, iterating two times.
14051 # The file diamond.shape may contain a pattern of characters like this
14057 # The specified columns and rows are ignored
14058 # but the anchor point coordinates are not
14059 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14064 Erode an image by using a specific structuring element.
14065 It corresponds to the libopencv function @code{cvErode}.
14067 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14068 with the same syntax and semantics as the @ref{dilate} filter.
14072 Smooth the input video.
14074 The filter takes the following parameters:
14075 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14077 @var{type} is the type of smooth filter to apply, and must be one of
14078 the following values: "blur", "blur_no_scale", "median", "gaussian",
14079 or "bilateral". The default value is "gaussian".
14081 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14082 depends on the smooth type. @var{param1} and
14083 @var{param2} accept integer positive values or 0. @var{param3} and
14084 @var{param4} accept floating point values.
14086 The default value for @var{param1} is 3. The default value for the
14087 other parameters is 0.
14089 These parameters correspond to the parameters assigned to the
14090 libopencv function @code{cvSmooth}.
14092 @section oscilloscope
14094 2D Video Oscilloscope.
14096 Useful to measure spatial impulse, step responses, chroma delays, etc.
14098 It accepts the following parameters:
14102 Set scope center x position.
14105 Set scope center y position.
14108 Set scope size, relative to frame diagonal.
14111 Set scope tilt/rotation.
14117 Set trace center x position.
14120 Set trace center y position.
14123 Set trace width, relative to width of frame.
14126 Set trace height, relative to height of frame.
14129 Set which components to trace. By default it traces first three components.
14132 Draw trace grid. By default is enabled.
14135 Draw some statistics. By default is enabled.
14138 Draw scope. By default is enabled.
14141 @subsection Commands
14142 This filter supports same @ref{commands} as options.
14143 The command accepts the same syntax of the corresponding option.
14145 If the specified expression is not valid, it is kept at its current
14148 @subsection Examples
14152 Inspect full first row of video frame.
14154 oscilloscope=x=0.5:y=0:s=1
14158 Inspect full last row of video frame.
14160 oscilloscope=x=0.5:y=1:s=1
14164 Inspect full 5th line of video frame of height 1080.
14166 oscilloscope=x=0.5:y=5/1080:s=1
14170 Inspect full last column of video frame.
14172 oscilloscope=x=1:y=0.5:s=1:t=1
14180 Overlay one video on top of another.
14182 It takes two inputs and has one output. The first input is the "main"
14183 video on which the second input is overlaid.
14185 It accepts the following parameters:
14187 A description of the accepted options follows.
14192 Set the expression for the x and y coordinates of the overlaid video
14193 on the main video. Default value is "0" for both expressions. In case
14194 the expression is invalid, it is set to a huge value (meaning that the
14195 overlay will not be displayed within the output visible area).
14198 See @ref{framesync}.
14201 Set when the expressions for @option{x}, and @option{y} are evaluated.
14203 It accepts the following values:
14206 only evaluate expressions once during the filter initialization or
14207 when a command is processed
14210 evaluate expressions for each incoming frame
14213 Default value is @samp{frame}.
14216 See @ref{framesync}.
14219 Set the format for the output video.
14221 It accepts the following values:
14224 force YUV420 output
14227 force YUV422 output
14230 force YUV444 output
14233 force packed RGB output
14236 force planar RGB output
14239 automatically pick format
14242 Default value is @samp{yuv420}.
14245 See @ref{framesync}.
14248 Set format of alpha of the overlaid video, it can be @var{straight} or
14249 @var{premultiplied}. Default is @var{straight}.
14252 The @option{x}, and @option{y} expressions can contain the following
14258 The main input width and height.
14262 The overlay input width and height.
14266 The computed values for @var{x} and @var{y}. They are evaluated for
14271 horizontal and vertical chroma subsample values of the output
14272 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
14276 the number of input frame, starting from 0
14279 the position in the file of the input frame, NAN if unknown
14282 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
14286 This filter also supports the @ref{framesync} options.
14288 Note that the @var{n}, @var{pos}, @var{t} variables are available only
14289 when evaluation is done @emph{per frame}, and will evaluate to NAN
14290 when @option{eval} is set to @samp{init}.
14292 Be aware that frames are taken from each input video in timestamp
14293 order, hence, if their initial timestamps differ, it is a good idea
14294 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
14295 have them begin in the same zero timestamp, as the example for
14296 the @var{movie} filter does.
14298 You can chain together more overlays but you should test the
14299 efficiency of such approach.
14301 @subsection Commands
14303 This filter supports the following commands:
14307 Modify the x and y of the overlay input.
14308 The command accepts the same syntax of the corresponding option.
14310 If the specified expression is not valid, it is kept at its current
14314 @subsection Examples
14318 Draw the overlay at 10 pixels from the bottom right corner of the main
14321 overlay=main_w-overlay_w-10:main_h-overlay_h-10
14324 Using named options the example above becomes:
14326 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
14330 Insert a transparent PNG logo in the bottom left corner of the input,
14331 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
14333 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
14337 Insert 2 different transparent PNG logos (second logo on bottom
14338 right corner) using the @command{ffmpeg} tool:
14340 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
14344 Add a transparent color layer on top of the main video; @code{WxH}
14345 must specify the size of the main input to the overlay filter:
14347 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
14351 Play an original video and a filtered version (here with the deshake
14352 filter) side by side using the @command{ffplay} tool:
14354 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
14357 The above command is the same as:
14359 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
14363 Make a sliding overlay appearing from the left to the right top part of the
14364 screen starting since time 2:
14366 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
14370 Compose output by putting two input videos side to side:
14372 ffmpeg -i left.avi -i right.avi -filter_complex "
14373 nullsrc=size=200x100 [background];
14374 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
14375 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
14376 [background][left] overlay=shortest=1 [background+left];
14377 [background+left][right] overlay=shortest=1:x=100 [left+right]
14382 Mask 10-20 seconds of a video by applying the delogo filter to a section
14384 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
14385 -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]'
14390 Chain several overlays in cascade:
14392 nullsrc=s=200x200 [bg];
14393 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
14394 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
14395 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
14396 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
14397 [in3] null, [mid2] overlay=100:100 [out0]
14402 @anchor{overlay_cuda}
14403 @section overlay_cuda
14405 Overlay one video on top of another.
14407 This is the CUDA cariant of the @ref{overlay} filter.
14408 It only accepts CUDA frames. The underlying input pixel formats have to match.
14410 It takes two inputs and has one output. The first input is the "main"
14411 video on which the second input is overlaid.
14413 It accepts the following parameters:
14418 Set the x and y coordinates of the overlaid video on the main video.
14419 Default value is "0" for both expressions.
14422 See @ref{framesync}.
14425 See @ref{framesync}.
14428 See @ref{framesync}.
14432 This filter also supports the @ref{framesync} options.
14436 Apply Overcomplete Wavelet denoiser.
14438 The filter accepts the following options:
14444 Larger depth values will denoise lower frequency components more, but
14445 slow down filtering.
14447 Must be an int in the range 8-16, default is @code{8}.
14449 @item luma_strength, ls
14452 Must be a double value in the range 0-1000, default is @code{1.0}.
14454 @item chroma_strength, cs
14455 Set chroma strength.
14457 Must be a double value in the range 0-1000, default is @code{1.0}.
14463 Add paddings to the input image, and place the original input at the
14464 provided @var{x}, @var{y} coordinates.
14466 It accepts the following parameters:
14471 Specify an expression for the size of the output image with the
14472 paddings added. If the value for @var{width} or @var{height} is 0, the
14473 corresponding input size is used for the output.
14475 The @var{width} expression can reference the value set by the
14476 @var{height} expression, and vice versa.
14478 The default value of @var{width} and @var{height} is 0.
14482 Specify the offsets to place the input image at within the padded area,
14483 with respect to the top/left border of the output image.
14485 The @var{x} expression can reference the value set by the @var{y}
14486 expression, and vice versa.
14488 The default value of @var{x} and @var{y} is 0.
14490 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
14491 so the input image is centered on the padded area.
14494 Specify the color of the padded area. For the syntax of this option,
14495 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
14496 manual,ffmpeg-utils}.
14498 The default value of @var{color} is "black".
14501 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
14503 It accepts the following values:
14507 Only evaluate expressions once during the filter initialization or when
14508 a command is processed.
14511 Evaluate expressions for each incoming frame.
14515 Default value is @samp{init}.
14518 Pad to aspect instead to a resolution.
14522 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
14523 options are expressions containing the following constants:
14528 The input video width and height.
14532 These are the same as @var{in_w} and @var{in_h}.
14536 The output width and height (the size of the padded area), as
14537 specified by the @var{width} and @var{height} expressions.
14541 These are the same as @var{out_w} and @var{out_h}.
14545 The x and y offsets as specified by the @var{x} and @var{y}
14546 expressions, or NAN if not yet specified.
14549 same as @var{iw} / @var{ih}
14552 input sample aspect ratio
14555 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
14559 The horizontal and vertical chroma subsample values. For example for the
14560 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14563 @subsection Examples
14567 Add paddings with the color "violet" to the input video. The output video
14568 size is 640x480, and the top-left corner of the input video is placed at
14571 pad=640:480:0:40:violet
14574 The example above is equivalent to the following command:
14576 pad=width=640:height=480:x=0:y=40:color=violet
14580 Pad the input to get an output with dimensions increased by 3/2,
14581 and put the input video at the center of the padded area:
14583 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
14587 Pad the input to get a squared output with size equal to the maximum
14588 value between the input width and height, and put the input video at
14589 the center of the padded area:
14591 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
14595 Pad the input to get a final w/h ratio of 16:9:
14597 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
14601 In case of anamorphic video, in order to set the output display aspect
14602 correctly, it is necessary to use @var{sar} in the expression,
14603 according to the relation:
14605 (ih * X / ih) * sar = output_dar
14606 X = output_dar / sar
14609 Thus the previous example needs to be modified to:
14611 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
14615 Double the output size and put the input video in the bottom-right
14616 corner of the output padded area:
14618 pad="2*iw:2*ih:ow-iw:oh-ih"
14622 @anchor{palettegen}
14623 @section palettegen
14625 Generate one palette for a whole video stream.
14627 It accepts the following options:
14631 Set the maximum number of colors to quantize in the palette.
14632 Note: the palette will still contain 256 colors; the unused palette entries
14635 @item reserve_transparent
14636 Create a palette of 255 colors maximum and reserve the last one for
14637 transparency. Reserving the transparency color is useful for GIF optimization.
14638 If not set, the maximum of colors in the palette will be 256. You probably want
14639 to disable this option for a standalone image.
14642 @item transparency_color
14643 Set the color that will be used as background for transparency.
14646 Set statistics mode.
14648 It accepts the following values:
14651 Compute full frame histograms.
14653 Compute histograms only for the part that differs from previous frame. This
14654 might be relevant to give more importance to the moving part of your input if
14655 the background is static.
14657 Compute new histogram for each frame.
14660 Default value is @var{full}.
14663 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
14664 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
14665 color quantization of the palette. This information is also visible at
14666 @var{info} logging level.
14668 @subsection Examples
14672 Generate a representative palette of a given video using @command{ffmpeg}:
14674 ffmpeg -i input.mkv -vf palettegen palette.png
14678 @section paletteuse
14680 Use a palette to downsample an input video stream.
14682 The filter takes two inputs: one video stream and a palette. The palette must
14683 be a 256 pixels image.
14685 It accepts the following options:
14689 Select dithering mode. Available algorithms are:
14692 Ordered 8x8 bayer dithering (deterministic)
14694 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
14695 Note: this dithering is sometimes considered "wrong" and is included as a
14697 @item floyd_steinberg
14698 Floyd and Steingberg dithering (error diffusion)
14700 Frankie Sierra dithering v2 (error diffusion)
14702 Frankie Sierra dithering v2 "Lite" (error diffusion)
14705 Default is @var{sierra2_4a}.
14708 When @var{bayer} dithering is selected, this option defines the scale of the
14709 pattern (how much the crosshatch pattern is visible). A low value means more
14710 visible pattern for less banding, and higher value means less visible pattern
14711 at the cost of more banding.
14713 The option must be an integer value in the range [0,5]. Default is @var{2}.
14716 If set, define the zone to process
14720 Only the changing rectangle will be reprocessed. This is similar to GIF
14721 cropping/offsetting compression mechanism. This option can be useful for speed
14722 if only a part of the image is changing, and has use cases such as limiting the
14723 scope of the error diffusal @option{dither} to the rectangle that bounds the
14724 moving scene (it leads to more deterministic output if the scene doesn't change
14725 much, and as a result less moving noise and better GIF compression).
14728 Default is @var{none}.
14731 Take new palette for each output frame.
14733 @item alpha_threshold
14734 Sets the alpha threshold for transparency. Alpha values above this threshold
14735 will be treated as completely opaque, and values below this threshold will be
14736 treated as completely transparent.
14738 The option must be an integer value in the range [0,255]. Default is @var{128}.
14741 @subsection Examples
14745 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
14746 using @command{ffmpeg}:
14748 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
14752 @section perspective
14754 Correct perspective of video not recorded perpendicular to the screen.
14756 A description of the accepted parameters follows.
14767 Set coordinates expression for top left, top right, bottom left and bottom right corners.
14768 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
14769 If the @code{sense} option is set to @code{source}, then the specified points will be sent
14770 to the corners of the destination. If the @code{sense} option is set to @code{destination},
14771 then the corners of the source will be sent to the specified coordinates.
14773 The expressions can use the following variables:
14778 the width and height of video frame.
14782 Output frame count.
14785 @item interpolation
14786 Set interpolation for perspective correction.
14788 It accepts the following values:
14794 Default value is @samp{linear}.
14797 Set interpretation of coordinate options.
14799 It accepts the following values:
14803 Send point in the source specified by the given coordinates to
14804 the corners of the destination.
14806 @item 1, destination
14808 Send the corners of the source to the point in the destination specified
14809 by the given coordinates.
14811 Default value is @samp{source}.
14815 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
14817 It accepts the following values:
14820 only evaluate expressions once during the filter initialization or
14821 when a command is processed
14824 evaluate expressions for each incoming frame
14827 Default value is @samp{init}.
14832 Delay interlaced video by one field time so that the field order changes.
14834 The intended use is to fix PAL movies that have been captured with the
14835 opposite field order to the film-to-video transfer.
14837 A description of the accepted parameters follows.
14843 It accepts the following values:
14846 Capture field order top-first, transfer bottom-first.
14847 Filter will delay the bottom field.
14850 Capture field order bottom-first, transfer top-first.
14851 Filter will delay the top field.
14854 Capture and transfer with the same field order. This mode only exists
14855 for the documentation of the other options to refer to, but if you
14856 actually select it, the filter will faithfully do nothing.
14859 Capture field order determined automatically by field flags, transfer
14861 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
14862 basis using field flags. If no field information is available,
14863 then this works just like @samp{u}.
14866 Capture unknown or varying, transfer opposite.
14867 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
14868 analyzing the images and selecting the alternative that produces best
14869 match between the fields.
14872 Capture top-first, transfer unknown or varying.
14873 Filter selects among @samp{t} and @samp{p} using image analysis.
14876 Capture bottom-first, transfer unknown or varying.
14877 Filter selects among @samp{b} and @samp{p} using image analysis.
14880 Capture determined by field flags, transfer unknown or varying.
14881 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
14882 image analysis. If no field information is available, then this works just
14883 like @samp{U}. This is the default mode.
14886 Both capture and transfer unknown or varying.
14887 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
14891 @section photosensitivity
14892 Reduce various flashes in video, so to help users with epilepsy.
14894 It accepts the following options:
14897 Set how many frames to use when filtering. Default is 30.
14900 Set detection threshold factor. Default is 1.
14904 Set how many pixels to skip when sampling frames. Default is 1.
14905 Allowed range is from 1 to 1024.
14908 Leave frames unchanged. Default is disabled.
14911 @section pixdesctest
14913 Pixel format descriptor test filter, mainly useful for internal
14914 testing. The output video should be equal to the input video.
14918 format=monow, pixdesctest
14921 can be used to test the monowhite pixel format descriptor definition.
14925 Display sample values of color channels. Mainly useful for checking color
14926 and levels. Minimum supported resolution is 640x480.
14928 The filters accept the following options:
14932 Set scope X position, relative offset on X axis.
14935 Set scope Y position, relative offset on Y axis.
14944 Set window opacity. This window also holds statistics about pixel area.
14947 Set window X position, relative offset on X axis.
14950 Set window Y position, relative offset on Y axis.
14955 Enable the specified chain of postprocessing subfilters using libpostproc. This
14956 library should be automatically selected with a GPL build (@code{--enable-gpl}).
14957 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
14958 Each subfilter and some options have a short and a long name that can be used
14959 interchangeably, i.e. dr/dering are the same.
14961 The filters accept the following options:
14965 Set postprocessing subfilters string.
14968 All subfilters share common options to determine their scope:
14972 Honor the quality commands for this subfilter.
14975 Do chrominance filtering, too (default).
14978 Do luminance filtering only (no chrominance).
14981 Do chrominance filtering only (no luminance).
14984 These options can be appended after the subfilter name, separated by a '|'.
14986 Available subfilters are:
14989 @item hb/hdeblock[|difference[|flatness]]
14990 Horizontal deblocking filter
14993 Difference factor where higher values mean more deblocking (default: @code{32}).
14995 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14998 @item vb/vdeblock[|difference[|flatness]]
14999 Vertical deblocking filter
15002 Difference factor where higher values mean more deblocking (default: @code{32}).
15004 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15007 @item ha/hadeblock[|difference[|flatness]]
15008 Accurate horizontal deblocking filter
15011 Difference factor where higher values mean more deblocking (default: @code{32}).
15013 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15016 @item va/vadeblock[|difference[|flatness]]
15017 Accurate vertical deblocking filter
15020 Difference factor where higher values mean more deblocking (default: @code{32}).
15022 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15026 The horizontal and vertical deblocking filters share the difference and
15027 flatness values so you cannot set different horizontal and vertical
15031 @item h1/x1hdeblock
15032 Experimental horizontal deblocking filter
15034 @item v1/x1vdeblock
15035 Experimental vertical deblocking filter
15040 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15043 larger -> stronger filtering
15045 larger -> stronger filtering
15047 larger -> stronger filtering
15050 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15053 Stretch luminance to @code{0-255}.
15056 @item lb/linblenddeint
15057 Linear blend deinterlacing filter that deinterlaces the given block by
15058 filtering all lines with a @code{(1 2 1)} filter.
15060 @item li/linipoldeint
15061 Linear interpolating deinterlacing filter that deinterlaces the given block by
15062 linearly interpolating every second line.
15064 @item ci/cubicipoldeint
15065 Cubic interpolating deinterlacing filter deinterlaces the given block by
15066 cubically interpolating every second line.
15068 @item md/mediandeint
15069 Median deinterlacing filter that deinterlaces the given block by applying a
15070 median filter to every second line.
15072 @item fd/ffmpegdeint
15073 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15074 second line with a @code{(-1 4 2 4 -1)} filter.
15077 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15078 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15080 @item fq/forceQuant[|quantizer]
15081 Overrides the quantizer table from the input with the constant quantizer you
15089 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15092 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15095 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15098 @subsection Examples
15102 Apply horizontal and vertical deblocking, deringing and automatic
15103 brightness/contrast:
15109 Apply default filters without brightness/contrast correction:
15115 Apply default filters and temporal denoiser:
15117 pp=default/tmpnoise|1|2|3
15121 Apply deblocking on luminance only, and switch vertical deblocking on or off
15122 automatically depending on available CPU time:
15129 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15130 similar to spp = 6 with 7 point DCT, where only the center sample is
15133 The filter accepts the following options:
15137 Force a constant quantization parameter. It accepts an integer in range
15138 0 to 63. If not set, the filter will use the QP from the video stream
15142 Set thresholding mode. Available modes are:
15146 Set hard thresholding.
15148 Set soft thresholding (better de-ringing effect, but likely blurrier).
15150 Set medium thresholding (good results, default).
15154 @section premultiply
15155 Apply alpha premultiply effect to input video stream using first plane
15156 of second stream as alpha.
15158 Both streams must have same dimensions and same pixel format.
15160 The filter accepts the following option:
15164 Set which planes will be processed, unprocessed planes will be copied.
15165 By default value 0xf, all planes will be processed.
15168 Do not require 2nd input for processing, instead use alpha plane from input stream.
15172 Apply prewitt operator to input video stream.
15174 The filter accepts the following option:
15178 Set which planes will be processed, unprocessed planes will be copied.
15179 By default value 0xf, all planes will be processed.
15182 Set value which will be multiplied with filtered result.
15185 Set value which will be added to filtered result.
15188 @section pseudocolor
15190 Alter frame colors in video with pseudocolors.
15192 This filter accepts the following options:
15196 set pixel first component expression
15199 set pixel second component expression
15202 set pixel third component expression
15205 set pixel fourth component expression, corresponds to the alpha component
15208 set component to use as base for altering colors
15211 Each of them specifies the expression to use for computing the lookup table for
15212 the corresponding pixel component values.
15214 The expressions can contain the following constants and functions:
15219 The input width and height.
15222 The input value for the pixel component.
15224 @item ymin, umin, vmin, amin
15225 The minimum allowed component value.
15227 @item ymax, umax, vmax, amax
15228 The maximum allowed component value.
15231 All expressions default to "val".
15233 @subsection Examples
15237 Change too high luma values to gradient:
15239 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'"
15245 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
15246 Ratio) between two input videos.
15248 This filter takes in input two input videos, the first input is
15249 considered the "main" source and is passed unchanged to the
15250 output. The second input is used as a "reference" video for computing
15253 Both video inputs must have the same resolution and pixel format for
15254 this filter to work correctly. Also it assumes that both inputs
15255 have the same number of frames, which are compared one by one.
15257 The obtained average PSNR is printed through the logging system.
15259 The filter stores the accumulated MSE (mean squared error) of each
15260 frame, and at the end of the processing it is averaged across all frames
15261 equally, and the following formula is applied to obtain the PSNR:
15264 PSNR = 10*log10(MAX^2/MSE)
15267 Where MAX is the average of the maximum values of each component of the
15270 The description of the accepted parameters follows.
15273 @item stats_file, f
15274 If specified the filter will use the named file to save the PSNR of
15275 each individual frame. When filename equals "-" the data is sent to
15278 @item stats_version
15279 Specifies which version of the stats file format to use. Details of
15280 each format are written below.
15281 Default value is 1.
15283 @item stats_add_max
15284 Determines whether the max value is output to the stats log.
15285 Default value is 0.
15286 Requires stats_version >= 2. If this is set and stats_version < 2,
15287 the filter will return an error.
15290 This filter also supports the @ref{framesync} options.
15292 The file printed if @var{stats_file} is selected, contains a sequence of
15293 key/value pairs of the form @var{key}:@var{value} for each compared
15296 If a @var{stats_version} greater than 1 is specified, a header line precedes
15297 the list of per-frame-pair stats, with key value pairs following the frame
15298 format with the following parameters:
15301 @item psnr_log_version
15302 The version of the log file format. Will match @var{stats_version}.
15305 A comma separated list of the per-frame-pair parameters included in
15309 A description of each shown per-frame-pair parameter follows:
15313 sequential number of the input frame, starting from 1
15316 Mean Square Error pixel-by-pixel average difference of the compared
15317 frames, averaged over all the image components.
15319 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
15320 Mean Square Error pixel-by-pixel average difference of the compared
15321 frames for the component specified by the suffix.
15323 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
15324 Peak Signal to Noise ratio of the compared frames for the component
15325 specified by the suffix.
15327 @item max_avg, max_y, max_u, max_v
15328 Maximum allowed value for each channel, and average over all
15332 @subsection Examples
15337 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
15338 [main][ref] psnr="stats_file=stats.log" [out]
15341 On this example the input file being processed is compared with the
15342 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
15343 is stored in @file{stats.log}.
15346 Another example with different containers:
15348 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 -
15355 Pulldown reversal (inverse telecine) filter, capable of handling mixed
15356 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
15359 The pullup filter is designed to take advantage of future context in making
15360 its decisions. This filter is stateless in the sense that it does not lock
15361 onto a pattern to follow, but it instead looks forward to the following
15362 fields in order to identify matches and rebuild progressive frames.
15364 To produce content with an even framerate, insert the fps filter after
15365 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
15366 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
15368 The filter accepts the following options:
15375 These options set the amount of "junk" to ignore at the left, right, top, and
15376 bottom of the image, respectively. Left and right are in units of 8 pixels,
15377 while top and bottom are in units of 2 lines.
15378 The default is 8 pixels on each side.
15381 Set the strict breaks. Setting this option to 1 will reduce the chances of
15382 filter generating an occasional mismatched frame, but it may also cause an
15383 excessive number of frames to be dropped during high motion sequences.
15384 Conversely, setting it to -1 will make filter match fields more easily.
15385 This may help processing of video where there is slight blurring between
15386 the fields, but may also cause there to be interlaced frames in the output.
15387 Default value is @code{0}.
15390 Set the metric plane to use. It accepts the following values:
15396 Use chroma blue plane.
15399 Use chroma red plane.
15402 This option may be set to use chroma plane instead of the default luma plane
15403 for doing filter's computations. This may improve accuracy on very clean
15404 source material, but more likely will decrease accuracy, especially if there
15405 is chroma noise (rainbow effect) or any grayscale video.
15406 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
15407 load and make pullup usable in realtime on slow machines.
15410 For best results (without duplicated frames in the output file) it is
15411 necessary to change the output frame rate. For example, to inverse
15412 telecine NTSC input:
15414 ffmpeg -i input -vf pullup -r 24000/1001 ...
15419 Change video quantization parameters (QP).
15421 The filter accepts the following option:
15425 Set expression for quantization parameter.
15428 The expression is evaluated through the eval API and can contain, among others,
15429 the following constants:
15433 1 if index is not 129, 0 otherwise.
15436 Sequential index starting from -129 to 128.
15439 @subsection Examples
15443 Some equation like:
15451 Flush video frames from internal cache of frames into a random order.
15452 No frame is discarded.
15453 Inspired by @ref{frei0r} nervous filter.
15457 Set size in number of frames of internal cache, in range from @code{2} to
15458 @code{512}. Default is @code{30}.
15461 Set seed for random number generator, must be an integer included between
15462 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
15463 less than @code{0}, the filter will try to use a good random seed on a
15467 @section readeia608
15469 Read closed captioning (EIA-608) information from the top lines of a video frame.
15471 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
15472 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
15473 with EIA-608 data (starting from 0). A description of each metadata value follows:
15476 @item lavfi.readeia608.X.cc
15477 The two bytes stored as EIA-608 data (printed in hexadecimal).
15479 @item lavfi.readeia608.X.line
15480 The number of the line on which the EIA-608 data was identified and read.
15483 This filter accepts the following options:
15487 Set the line to start scanning for EIA-608 data. Default is @code{0}.
15490 Set the line to end scanning for EIA-608 data. Default is @code{29}.
15493 Set the ratio of width reserved for sync code detection.
15494 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
15497 Enable checking the parity bit. In the event of a parity error, the filter will output
15498 @code{0x00} for that character. Default is false.
15501 Lowpass lines prior to further processing. Default is enabled.
15504 @subsection Examples
15508 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
15510 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
15516 Read vertical interval timecode (VITC) information from the top lines of a
15519 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
15520 timecode value, if a valid timecode has been detected. Further metadata key
15521 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
15522 timecode data has been found or not.
15524 This filter accepts the following options:
15528 Set the maximum number of lines to scan for VITC data. If the value is set to
15529 @code{-1} the full video frame is scanned. Default is @code{45}.
15532 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
15533 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
15536 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
15537 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
15540 @subsection Examples
15544 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
15545 draw @code{--:--:--:--} as a placeholder:
15547 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
15553 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
15555 Destination pixel at position (X, Y) will be picked from source (x, y) position
15556 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
15557 value for pixel will be used for destination pixel.
15559 Xmap and Ymap input video streams must be of same dimensions. Output video stream
15560 will have Xmap/Ymap video stream dimensions.
15561 Xmap and Ymap input video streams are 16bit depth, single channel.
15565 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
15566 Default is @code{color}.
15569 Specify the color of the unmapped pixels. For the syntax of this option,
15570 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15571 manual,ffmpeg-utils}. Default color is @code{black}.
15574 @section removegrain
15576 The removegrain filter is a spatial denoiser for progressive video.
15580 Set mode for the first plane.
15583 Set mode for the second plane.
15586 Set mode for the third plane.
15589 Set mode for the fourth plane.
15592 Range of mode is from 0 to 24. Description of each mode follows:
15596 Leave input plane unchanged. Default.
15599 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
15602 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
15605 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
15608 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
15609 This is equivalent to a median filter.
15612 Line-sensitive clipping giving the minimal change.
15615 Line-sensitive clipping, intermediate.
15618 Line-sensitive clipping, intermediate.
15621 Line-sensitive clipping, intermediate.
15624 Line-sensitive clipping on a line where the neighbours pixels are the closest.
15627 Replaces the target pixel with the closest neighbour.
15630 [1 2 1] horizontal and vertical kernel blur.
15636 Bob mode, interpolates top field from the line where the neighbours
15637 pixels are the closest.
15640 Bob mode, interpolates bottom field from the line where the neighbours
15641 pixels are the closest.
15644 Bob mode, interpolates top field. Same as 13 but with a more complicated
15645 interpolation formula.
15648 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
15649 interpolation formula.
15652 Clips the pixel with the minimum and maximum of respectively the maximum and
15653 minimum of each pair of opposite neighbour pixels.
15656 Line-sensitive clipping using opposite neighbours whose greatest distance from
15657 the current pixel is minimal.
15660 Replaces the pixel with the average of its 8 neighbours.
15663 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
15666 Clips pixels using the averages of opposite neighbour.
15669 Same as mode 21 but simpler and faster.
15672 Small edge and halo removal, but reputed useless.
15678 @section removelogo
15680 Suppress a TV station logo, using an image file to determine which
15681 pixels comprise the logo. It works by filling in the pixels that
15682 comprise the logo with neighboring pixels.
15684 The filter accepts the following options:
15688 Set the filter bitmap file, which can be any image format supported by
15689 libavformat. The width and height of the image file must match those of the
15690 video stream being processed.
15693 Pixels in the provided bitmap image with a value of zero are not
15694 considered part of the logo, non-zero pixels are considered part of
15695 the logo. If you use white (255) for the logo and black (0) for the
15696 rest, you will be safe. For making the filter bitmap, it is
15697 recommended to take a screen capture of a black frame with the logo
15698 visible, and then using a threshold filter followed by the erode
15699 filter once or twice.
15701 If needed, little splotches can be fixed manually. Remember that if
15702 logo pixels are not covered, the filter quality will be much
15703 reduced. Marking too many pixels as part of the logo does not hurt as
15704 much, but it will increase the amount of blurring needed to cover over
15705 the image and will destroy more information than necessary, and extra
15706 pixels will slow things down on a large logo.
15708 @section repeatfields
15710 This filter uses the repeat_field flag from the Video ES headers and hard repeats
15711 fields based on its value.
15715 Reverse a video clip.
15717 Warning: This filter requires memory to buffer the entire clip, so trimming
15720 @subsection Examples
15724 Take the first 5 seconds of a clip, and reverse it.
15731 Shift R/G/B/A pixels horizontally and/or vertically.
15733 The filter accepts the following options:
15736 Set amount to shift red horizontally.
15738 Set amount to shift red vertically.
15740 Set amount to shift green horizontally.
15742 Set amount to shift green vertically.
15744 Set amount to shift blue horizontally.
15746 Set amount to shift blue vertically.
15748 Set amount to shift alpha horizontally.
15750 Set amount to shift alpha vertically.
15752 Set edge mode, can be @var{smear}, default, or @var{warp}.
15755 @subsection Commands
15757 This filter supports the all above options as @ref{commands}.
15760 Apply roberts cross operator to input video stream.
15762 The filter accepts the following option:
15766 Set which planes will be processed, unprocessed planes will be copied.
15767 By default value 0xf, all planes will be processed.
15770 Set value which will be multiplied with filtered result.
15773 Set value which will be added to filtered result.
15778 Rotate video by an arbitrary angle expressed in radians.
15780 The filter accepts the following options:
15782 A description of the optional parameters follows.
15785 Set an expression for the angle by which to rotate the input video
15786 clockwise, expressed as a number of radians. A negative value will
15787 result in a counter-clockwise rotation. By default it is set to "0".
15789 This expression is evaluated for each frame.
15792 Set the output width expression, default value is "iw".
15793 This expression is evaluated just once during configuration.
15796 Set the output height expression, default value is "ih".
15797 This expression is evaluated just once during configuration.
15800 Enable bilinear interpolation if set to 1, a value of 0 disables
15801 it. Default value is 1.
15804 Set the color used to fill the output area not covered by the rotated
15805 image. For the general syntax of this option, check the
15806 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
15807 If the special value "none" is selected then no
15808 background is printed (useful for example if the background is never shown).
15810 Default value is "black".
15813 The expressions for the angle and the output size can contain the
15814 following constants and functions:
15818 sequential number of the input frame, starting from 0. It is always NAN
15819 before the first frame is filtered.
15822 time in seconds of the input frame, it is set to 0 when the filter is
15823 configured. It is always NAN before the first frame is filtered.
15827 horizontal and vertical chroma subsample values. For example for the
15828 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15832 the input video width and height
15836 the output width and height, that is the size of the padded area as
15837 specified by the @var{width} and @var{height} expressions
15841 the minimal width/height required for completely containing the input
15842 video rotated by @var{a} radians.
15844 These are only available when computing the @option{out_w} and
15845 @option{out_h} expressions.
15848 @subsection Examples
15852 Rotate the input by PI/6 radians clockwise:
15858 Rotate the input by PI/6 radians counter-clockwise:
15864 Rotate the input by 45 degrees clockwise:
15870 Apply a constant rotation with period T, starting from an angle of PI/3:
15872 rotate=PI/3+2*PI*t/T
15876 Make the input video rotation oscillating with a period of T
15877 seconds and an amplitude of A radians:
15879 rotate=A*sin(2*PI/T*t)
15883 Rotate the video, output size is chosen so that the whole rotating
15884 input video is always completely contained in the output:
15886 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
15890 Rotate the video, reduce the output size so that no background is ever
15893 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
15897 @subsection Commands
15899 The filter supports the following commands:
15903 Set the angle expression.
15904 The command accepts the same syntax of the corresponding option.
15906 If the specified expression is not valid, it is kept at its current
15912 Apply Shape Adaptive Blur.
15914 The filter accepts the following options:
15917 @item luma_radius, lr
15918 Set luma blur filter strength, must be a value in range 0.1-4.0, default
15919 value is 1.0. A greater value will result in a more blurred image, and
15920 in slower processing.
15922 @item luma_pre_filter_radius, lpfr
15923 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
15926 @item luma_strength, ls
15927 Set luma maximum difference between pixels to still be considered, must
15928 be a value in the 0.1-100.0 range, default value is 1.0.
15930 @item chroma_radius, cr
15931 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
15932 greater value will result in a more blurred image, and in slower
15935 @item chroma_pre_filter_radius, cpfr
15936 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
15938 @item chroma_strength, cs
15939 Set chroma maximum difference between pixels to still be considered,
15940 must be a value in the -0.9-100.0 range.
15943 Each chroma option value, if not explicitly specified, is set to the
15944 corresponding luma option value.
15949 Scale (resize) the input video, using the libswscale library.
15951 The scale filter forces the output display aspect ratio to be the same
15952 of the input, by changing the output sample aspect ratio.
15954 If the input image format is different from the format requested by
15955 the next filter, the scale filter will convert the input to the
15958 @subsection Options
15959 The filter accepts the following options, or any of the options
15960 supported by the libswscale scaler.
15962 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
15963 the complete list of scaler options.
15968 Set the output video dimension expression. Default value is the input
15971 If the @var{width} or @var{w} value is 0, the input width is used for
15972 the output. If the @var{height} or @var{h} value is 0, the input height
15973 is used for the output.
15975 If one and only one of the values is -n with n >= 1, the scale filter
15976 will use a value that maintains the aspect ratio of the input image,
15977 calculated from the other specified dimension. After that it will,
15978 however, make sure that the calculated dimension is divisible by n and
15979 adjust the value if necessary.
15981 If both values are -n with n >= 1, the behavior will be identical to
15982 both values being set to 0 as previously detailed.
15984 See below for the list of accepted constants for use in the dimension
15988 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
15992 Only evaluate expressions once during the filter initialization or when a command is processed.
15995 Evaluate expressions for each incoming frame.
15999 Default value is @samp{init}.
16003 Set the interlacing mode. It accepts the following values:
16007 Force interlaced aware scaling.
16010 Do not apply interlaced scaling.
16013 Select interlaced aware scaling depending on whether the source frames
16014 are flagged as interlaced or not.
16017 Default value is @samp{0}.
16020 Set libswscale scaling flags. See
16021 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16022 complete list of values. If not explicitly specified the filter applies
16026 @item param0, param1
16027 Set libswscale input parameters for scaling algorithms that need them. See
16028 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16029 complete documentation. If not explicitly specified the filter applies
16035 Set the video size. For the syntax of this option, check the
16036 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16038 @item in_color_matrix
16039 @item out_color_matrix
16040 Set in/output YCbCr color space type.
16042 This allows the autodetected value to be overridden as well as allows forcing
16043 a specific value used for the output and encoder.
16045 If not specified, the color space type depends on the pixel format.
16051 Choose automatically.
16054 Format conforming to International Telecommunication Union (ITU)
16055 Recommendation BT.709.
16058 Set color space conforming to the United States Federal Communications
16059 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16064 Set color space conforming to:
16068 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16071 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16074 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16079 Set color space conforming to SMPTE ST 240:1999.
16082 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16087 Set in/output YCbCr sample range.
16089 This allows the autodetected value to be overridden as well as allows forcing
16090 a specific value used for the output and encoder. If not specified, the
16091 range depends on the pixel format. Possible values:
16095 Choose automatically.
16098 Set full range (0-255 in case of 8-bit luma).
16100 @item mpeg/limited/tv
16101 Set "MPEG" range (16-235 in case of 8-bit luma).
16104 @item force_original_aspect_ratio
16105 Enable decreasing or increasing output video width or height if necessary to
16106 keep the original aspect ratio. Possible values:
16110 Scale the video as specified and disable this feature.
16113 The output video dimensions will automatically be decreased if needed.
16116 The output video dimensions will automatically be increased if needed.
16120 One useful instance of this option is that when you know a specific device's
16121 maximum allowed resolution, you can use this to limit the output video to
16122 that, while retaining the aspect ratio. For example, device A allows
16123 1280x720 playback, and your video is 1920x800. Using this option (set it to
16124 decrease) and specifying 1280x720 to the command line makes the output
16127 Please note that this is a different thing than specifying -1 for @option{w}
16128 or @option{h}, you still need to specify the output resolution for this option
16131 @item force_divisible_by
16132 Ensures that both the output dimensions, width and height, are divisible by the
16133 given integer when used together with @option{force_original_aspect_ratio}. This
16134 works similar to using @code{-n} in the @option{w} and @option{h} options.
16136 This option respects the value set for @option{force_original_aspect_ratio},
16137 increasing or decreasing the resolution accordingly. The video's aspect ratio
16138 may be slightly modified.
16140 This option can be handy if you need to have a video fit within or exceed
16141 a defined resolution using @option{force_original_aspect_ratio} but also have
16142 encoder restrictions on width or height divisibility.
16146 The values of the @option{w} and @option{h} options are expressions
16147 containing the following constants:
16152 The input width and height
16156 These are the same as @var{in_w} and @var{in_h}.
16160 The output (scaled) width and height
16164 These are the same as @var{out_w} and @var{out_h}
16167 The same as @var{iw} / @var{ih}
16170 input sample aspect ratio
16173 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16177 horizontal and vertical input chroma subsample values. For example for the
16178 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16182 horizontal and vertical output chroma subsample values. For example for the
16183 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16186 The (sequential) number of the input frame, starting from 0.
16187 Only available with @code{eval=frame}.
16190 The presentation timestamp of the input frame, expressed as a number of
16191 seconds. Only available with @code{eval=frame}.
16194 The position (byte offset) of the frame in the input stream, or NaN if
16195 this information is unavailable and/or meaningless (for example in case of synthetic video).
16196 Only available with @code{eval=frame}.
16199 @subsection Examples
16203 Scale the input video to a size of 200x100
16208 This is equivalent to:
16219 Specify a size abbreviation for the output size:
16224 which can also be written as:
16230 Scale the input to 2x:
16232 scale=w=2*iw:h=2*ih
16236 The above is the same as:
16238 scale=2*in_w:2*in_h
16242 Scale the input to 2x with forced interlaced scaling:
16244 scale=2*iw:2*ih:interl=1
16248 Scale the input to half size:
16250 scale=w=iw/2:h=ih/2
16254 Increase the width, and set the height to the same size:
16260 Seek Greek harmony:
16267 Increase the height, and set the width to 3/2 of the height:
16269 scale=w=3/2*oh:h=3/5*ih
16273 Increase the size, making the size a multiple of the chroma
16276 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
16280 Increase the width to a maximum of 500 pixels,
16281 keeping the same aspect ratio as the input:
16283 scale=w='min(500\, iw*3/2):h=-1'
16287 Make pixels square by combining scale and setsar:
16289 scale='trunc(ih*dar):ih',setsar=1/1
16293 Make pixels square by combining scale and setsar,
16294 making sure the resulting resolution is even (required by some codecs):
16296 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
16300 @subsection Commands
16302 This filter supports the following commands:
16306 Set the output video dimension expression.
16307 The command accepts the same syntax of the corresponding option.
16309 If the specified expression is not valid, it is kept at its current
16315 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
16316 format conversion on CUDA video frames. Setting the output width and height
16317 works in the same way as for the @var{scale} filter.
16319 The following additional options are accepted:
16322 The pixel format of the output CUDA frames. If set to the string "same" (the
16323 default), the input format will be kept. Note that automatic format negotiation
16324 and conversion is not yet supported for hardware frames
16327 The interpolation algorithm used for resizing. One of the following:
16334 @item cubic2p_bspline
16335 2-parameter cubic (B=1, C=0)
16337 @item cubic2p_catmullrom
16338 2-parameter cubic (B=0, C=1/2)
16340 @item cubic2p_b05c03
16341 2-parameter cubic (B=1/2, C=3/10)
16349 @item force_original_aspect_ratio
16350 Enable decreasing or increasing output video width or height if necessary to
16351 keep the original aspect ratio. Possible values:
16355 Scale the video as specified and disable this feature.
16358 The output video dimensions will automatically be decreased if needed.
16361 The output video dimensions will automatically be increased if needed.
16365 One useful instance of this option is that when you know a specific device's
16366 maximum allowed resolution, you can use this to limit the output video to
16367 that, while retaining the aspect ratio. For example, device A allows
16368 1280x720 playback, and your video is 1920x800. Using this option (set it to
16369 decrease) and specifying 1280x720 to the command line makes the output
16372 Please note that this is a different thing than specifying -1 for @option{w}
16373 or @option{h}, you still need to specify the output resolution for this option
16376 @item force_divisible_by
16377 Ensures that both the output dimensions, width and height, are divisible by the
16378 given integer when used together with @option{force_original_aspect_ratio}. This
16379 works similar to using @code{-n} in the @option{w} and @option{h} options.
16381 This option respects the value set for @option{force_original_aspect_ratio},
16382 increasing or decreasing the resolution accordingly. The video's aspect ratio
16383 may be slightly modified.
16385 This option can be handy if you need to have a video fit within or exceed
16386 a defined resolution using @option{force_original_aspect_ratio} but also have
16387 encoder restrictions on width or height divisibility.
16393 Scale (resize) the input video, based on a reference video.
16395 See the scale filter for available options, scale2ref supports the same but
16396 uses the reference video instead of the main input as basis. scale2ref also
16397 supports the following additional constants for the @option{w} and
16398 @option{h} options:
16403 The main input video's width and height
16406 The same as @var{main_w} / @var{main_h}
16409 The main input video's sample aspect ratio
16411 @item main_dar, mdar
16412 The main input video's display aspect ratio. Calculated from
16413 @code{(main_w / main_h) * main_sar}.
16417 The main input video's horizontal and vertical chroma subsample values.
16418 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
16422 The (sequential) number of the main input frame, starting from 0.
16423 Only available with @code{eval=frame}.
16426 The presentation timestamp of the main input frame, expressed as a number of
16427 seconds. Only available with @code{eval=frame}.
16430 The position (byte offset) of the frame in the main input stream, or NaN if
16431 this information is unavailable and/or meaningless (for example in case of synthetic video).
16432 Only available with @code{eval=frame}.
16435 @subsection Examples
16439 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
16441 'scale2ref[b][a];[a][b]overlay'
16445 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
16447 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
16451 @subsection Commands
16453 This filter supports the following commands:
16457 Set the output video dimension expression.
16458 The command accepts the same syntax of the corresponding option.
16460 If the specified expression is not valid, it is kept at its current
16465 Scroll input video horizontally and/or vertically by constant speed.
16467 The filter accepts the following options:
16469 @item horizontal, h
16470 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
16471 Negative values changes scrolling direction.
16474 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
16475 Negative values changes scrolling direction.
16478 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
16481 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
16484 @subsection Commands
16486 This filter supports the following @ref{commands}:
16488 @item horizontal, h
16489 Set the horizontal scrolling speed.
16491 Set the vertical scrolling speed.
16494 @anchor{selectivecolor}
16495 @section selectivecolor
16497 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
16498 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
16499 by the "purity" of the color (that is, how saturated it already is).
16501 This filter is similar to the Adobe Photoshop Selective Color tool.
16503 The filter accepts the following options:
16506 @item correction_method
16507 Select color correction method.
16509 Available values are:
16512 Specified adjustments are applied "as-is" (added/subtracted to original pixel
16515 Specified adjustments are relative to the original component value.
16517 Default is @code{absolute}.
16519 Adjustments for red pixels (pixels where the red component is the maximum)
16521 Adjustments for yellow pixels (pixels where the blue component is the minimum)
16523 Adjustments for green pixels (pixels where the green component is the maximum)
16525 Adjustments for cyan pixels (pixels where the red component is the minimum)
16527 Adjustments for blue pixels (pixels where the blue component is the maximum)
16529 Adjustments for magenta pixels (pixels where the green component is the minimum)
16531 Adjustments for white pixels (pixels where all components are greater than 128)
16533 Adjustments for all pixels except pure black and pure white
16535 Adjustments for black pixels (pixels where all components are lesser than 128)
16537 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
16540 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
16541 4 space separated floating point adjustment values in the [-1,1] range,
16542 respectively to adjust the amount of cyan, magenta, yellow and black for the
16543 pixels of its range.
16545 @subsection Examples
16549 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
16550 increase magenta by 27% in blue areas:
16552 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
16556 Use a Photoshop selective color preset:
16558 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
16562 @anchor{separatefields}
16563 @section separatefields
16565 The @code{separatefields} takes a frame-based video input and splits
16566 each frame into its components fields, producing a new half height clip
16567 with twice the frame rate and twice the frame count.
16569 This filter use field-dominance information in frame to decide which
16570 of each pair of fields to place first in the output.
16571 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
16573 @section setdar, setsar
16575 The @code{setdar} filter sets the Display Aspect Ratio for the filter
16578 This is done by changing the specified Sample (aka Pixel) Aspect
16579 Ratio, according to the following equation:
16581 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
16584 Keep in mind that the @code{setdar} filter does not modify the pixel
16585 dimensions of the video frame. Also, the display aspect ratio set by
16586 this filter may be changed by later filters in the filterchain,
16587 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
16590 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
16591 the filter output video.
16593 Note that as a consequence of the application of this filter, the
16594 output display aspect ratio will change according to the equation
16597 Keep in mind that the sample aspect ratio set by the @code{setsar}
16598 filter may be changed by later filters in the filterchain, e.g. if
16599 another "setsar" or a "setdar" filter is applied.
16601 It accepts the following parameters:
16604 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
16605 Set the aspect ratio used by the filter.
16607 The parameter can be a floating point number string, an expression, or
16608 a string of the form @var{num}:@var{den}, where @var{num} and
16609 @var{den} are the numerator and denominator of the aspect ratio. If
16610 the parameter is not specified, it is assumed the value "0".
16611 In case the form "@var{num}:@var{den}" is used, the @code{:} character
16615 Set the maximum integer value to use for expressing numerator and
16616 denominator when reducing the expressed aspect ratio to a rational.
16617 Default value is @code{100}.
16621 The parameter @var{sar} is an expression containing
16622 the following constants:
16626 These are approximated values for the mathematical constants e
16627 (Euler's number), pi (Greek pi), and phi (the golden ratio).
16630 The input width and height.
16633 These are the same as @var{w} / @var{h}.
16636 The input sample aspect ratio.
16639 The input display aspect ratio. It is the same as
16640 (@var{w} / @var{h}) * @var{sar}.
16643 Horizontal and vertical chroma subsample values. For example, for the
16644 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16647 @subsection Examples
16652 To change the display aspect ratio to 16:9, specify one of the following:
16659 To change the sample aspect ratio to 10:11, specify:
16665 To set a display aspect ratio of 16:9, and specify a maximum integer value of
16666 1000 in the aspect ratio reduction, use the command:
16668 setdar=ratio=16/9:max=1000
16676 Force field for the output video frame.
16678 The @code{setfield} filter marks the interlace type field for the
16679 output frames. It does not change the input frame, but only sets the
16680 corresponding property, which affects how the frame is treated by
16681 following filters (e.g. @code{fieldorder} or @code{yadif}).
16683 The filter accepts the following options:
16688 Available values are:
16692 Keep the same field property.
16695 Mark the frame as bottom-field-first.
16698 Mark the frame as top-field-first.
16701 Mark the frame as progressive.
16708 Force frame parameter for the output video frame.
16710 The @code{setparams} filter marks interlace and color range for the
16711 output frames. It does not change the input frame, but only sets the
16712 corresponding property, which affects how the frame is treated by
16717 Available values are:
16721 Keep the same field property (default).
16724 Mark the frame as bottom-field-first.
16727 Mark the frame as top-field-first.
16730 Mark the frame as progressive.
16734 Available values are:
16738 Keep the same color range property (default).
16740 @item unspecified, unknown
16741 Mark the frame as unspecified color range.
16743 @item limited, tv, mpeg
16744 Mark the frame as limited range.
16746 @item full, pc, jpeg
16747 Mark the frame as full range.
16750 @item color_primaries
16751 Set the color primaries.
16752 Available values are:
16756 Keep the same color primaries property (default).
16773 Set the color transfer.
16774 Available values are:
16778 Keep the same color trc property (default).
16800 Set the colorspace.
16801 Available values are:
16805 Keep the same colorspace property (default).
16818 @item chroma-derived-nc
16819 @item chroma-derived-c
16826 Show a line containing various information for each input video frame.
16827 The input video is not modified.
16829 This filter supports the following options:
16833 Calculate checksums of each plane. By default enabled.
16836 The shown line contains a sequence of key/value pairs of the form
16837 @var{key}:@var{value}.
16839 The following values are shown in the output:
16843 The (sequential) number of the input frame, starting from 0.
16846 The Presentation TimeStamp of the input frame, expressed as a number of
16847 time base units. The time base unit depends on the filter input pad.
16850 The Presentation TimeStamp of the input frame, expressed as a number of
16854 The position of the frame in the input stream, or -1 if this information is
16855 unavailable and/or meaningless (for example in case of synthetic video).
16858 The pixel format name.
16861 The sample aspect ratio of the input frame, expressed in the form
16862 @var{num}/@var{den}.
16865 The size of the input frame. For the syntax of this option, check the
16866 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16869 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
16870 for bottom field first).
16873 This is 1 if the frame is a key frame, 0 otherwise.
16876 The picture type of the input frame ("I" for an I-frame, "P" for a
16877 P-frame, "B" for a B-frame, or "?" for an unknown type).
16878 Also refer to the documentation of the @code{AVPictureType} enum and of
16879 the @code{av_get_picture_type_char} function defined in
16880 @file{libavutil/avutil.h}.
16883 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
16885 @item plane_checksum
16886 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
16887 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
16890 The mean value of pixels in each plane of the input frame, expressed in the form
16891 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
16894 The standard deviation of pixel values in each plane of the input frame, expressed
16895 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
16899 @section showpalette
16901 Displays the 256 colors palette of each frame. This filter is only relevant for
16902 @var{pal8} pixel format frames.
16904 It accepts the following option:
16908 Set the size of the box used to represent one palette color entry. Default is
16909 @code{30} (for a @code{30x30} pixel box).
16912 @section shuffleframes
16914 Reorder and/or duplicate and/or drop video frames.
16916 It accepts the following parameters:
16920 Set the destination indexes of input frames.
16921 This is space or '|' separated list of indexes that maps input frames to output
16922 frames. Number of indexes also sets maximal value that each index may have.
16923 '-1' index have special meaning and that is to drop frame.
16926 The first frame has the index 0. The default is to keep the input unchanged.
16928 @subsection Examples
16932 Swap second and third frame of every three frames of the input:
16934 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
16938 Swap 10th and 1st frame of every ten frames of the input:
16940 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
16944 @section shuffleplanes
16946 Reorder and/or duplicate video planes.
16948 It accepts the following parameters:
16953 The index of the input plane to be used as the first output plane.
16956 The index of the input plane to be used as the second output plane.
16959 The index of the input plane to be used as the third output plane.
16962 The index of the input plane to be used as the fourth output plane.
16966 The first plane has the index 0. The default is to keep the input unchanged.
16968 @subsection Examples
16972 Swap the second and third planes of the input:
16974 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
16978 @anchor{signalstats}
16979 @section signalstats
16980 Evaluate various visual metrics that assist in determining issues associated
16981 with the digitization of analog video media.
16983 By default the filter will log these metadata values:
16987 Display the minimal Y value contained within the input frame. Expressed in
16991 Display the Y value at the 10% percentile within the input frame. Expressed in
16995 Display the average Y value within the input frame. Expressed in range of
16999 Display the Y value at the 90% percentile within the input frame. Expressed in
17003 Display the maximum Y value contained within the input frame. Expressed in
17007 Display the minimal U value contained within the input frame. Expressed in
17011 Display the U value at the 10% percentile within the input frame. Expressed in
17015 Display the average U value within the input frame. Expressed in range of
17019 Display the U value at the 90% percentile within the input frame. Expressed in
17023 Display the maximum U value contained within the input frame. Expressed in
17027 Display the minimal V value contained within the input frame. Expressed in
17031 Display the V value at the 10% percentile within the input frame. Expressed in
17035 Display the average V value within the input frame. Expressed in range of
17039 Display the V value at the 90% percentile within the input frame. Expressed in
17043 Display the maximum V value contained within the input frame. Expressed in
17047 Display the minimal saturation value contained within the input frame.
17048 Expressed in range of [0-~181.02].
17051 Display the saturation value at the 10% percentile within the input frame.
17052 Expressed in range of [0-~181.02].
17055 Display the average saturation value within the input frame. Expressed in range
17059 Display the saturation value at the 90% percentile within the input frame.
17060 Expressed in range of [0-~181.02].
17063 Display the maximum saturation value contained within the input frame.
17064 Expressed in range of [0-~181.02].
17067 Display the median value for hue within the input frame. Expressed in range of
17071 Display the average value for hue within the input frame. Expressed in range of
17075 Display the average of sample value difference between all values of the Y
17076 plane in the current frame and corresponding values of the previous input frame.
17077 Expressed in range of [0-255].
17080 Display the average of sample value difference between all values of the U
17081 plane in the current frame and corresponding values of the previous input frame.
17082 Expressed in range of [0-255].
17085 Display the average of sample value difference between all values of the V
17086 plane in the current frame and corresponding values of the previous input frame.
17087 Expressed in range of [0-255].
17090 Display bit depth of Y plane in current frame.
17091 Expressed in range of [0-16].
17094 Display bit depth of U plane in current frame.
17095 Expressed in range of [0-16].
17098 Display bit depth of V plane in current frame.
17099 Expressed in range of [0-16].
17102 The filter accepts the following options:
17108 @option{stat} specify an additional form of image analysis.
17109 @option{out} output video with the specified type of pixel highlighted.
17111 Both options accept the following values:
17115 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17116 unlike the neighboring pixels of the same field. Examples of temporal outliers
17117 include the results of video dropouts, head clogs, or tape tracking issues.
17120 Identify @var{vertical line repetition}. Vertical line repetition includes
17121 similar rows of pixels within a frame. In born-digital video vertical line
17122 repetition is common, but this pattern is uncommon in video digitized from an
17123 analog source. When it occurs in video that results from the digitization of an
17124 analog source it can indicate concealment from a dropout compensator.
17127 Identify pixels that fall outside of legal broadcast range.
17131 Set the highlight color for the @option{out} option. The default color is
17135 @subsection Examples
17139 Output data of various video metrics:
17141 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17145 Output specific data about the minimum and maximum values of the Y plane per frame:
17147 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17151 Playback video while highlighting pixels that are outside of broadcast range in red.
17153 ffplay example.mov -vf signalstats="out=brng:color=red"
17157 Playback video with signalstats metadata drawn over the frame.
17159 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17162 The contents of signalstat_drawtext.txt used in the command are:
17165 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17166 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
17167 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
17168 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
17176 Calculates the MPEG-7 Video Signature. The filter can handle more than one
17177 input. In this case the matching between the inputs can be calculated additionally.
17178 The filter always passes through the first input. The signature of each stream can
17179 be written into a file.
17181 It accepts the following options:
17185 Enable or disable the matching process.
17187 Available values are:
17191 Disable the calculation of a matching (default).
17193 Calculate the matching for the whole video and output whether the whole video
17194 matches or only parts.
17196 Calculate only until a matching is found or the video ends. Should be faster in
17201 Set the number of inputs. The option value must be a non negative integer.
17202 Default value is 1.
17205 Set the path to which the output is written. If there is more than one input,
17206 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
17207 integer), that will be replaced with the input number. If no filename is
17208 specified, no output will be written. This is the default.
17211 Choose the output format.
17213 Available values are:
17217 Use the specified binary representation (default).
17219 Use the specified xml representation.
17223 Set threshold to detect one word as similar. The option value must be an integer
17224 greater than zero. The default value is 9000.
17227 Set threshold to detect all words as similar. The option value must be an integer
17228 greater than zero. The default value is 60000.
17231 Set threshold to detect frames as similar. The option value must be an integer
17232 greater than zero. The default value is 116.
17235 Set the minimum length of a sequence in frames to recognize it as matching
17236 sequence. The option value must be a non negative integer value.
17237 The default value is 0.
17240 Set the minimum relation, that matching frames to all frames must have.
17241 The option value must be a double value between 0 and 1. The default value is 0.5.
17244 @subsection Examples
17248 To calculate the signature of an input video and store it in signature.bin:
17250 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
17254 To detect whether two videos match and store the signatures in XML format in
17255 signature0.xml and signature1.xml:
17257 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 -
17265 Blur the input video without impacting the outlines.
17267 It accepts the following options:
17270 @item luma_radius, lr
17271 Set the luma radius. The option value must be a float number in
17272 the range [0.1,5.0] that specifies the variance of the gaussian filter
17273 used to blur the image (slower if larger). Default value is 1.0.
17275 @item luma_strength, ls
17276 Set the luma strength. The option value must be a float number
17277 in the range [-1.0,1.0] that configures the blurring. A value included
17278 in [0.0,1.0] will blur the image whereas a value included in
17279 [-1.0,0.0] will sharpen the image. Default value is 1.0.
17281 @item luma_threshold, lt
17282 Set the luma threshold used as a coefficient to determine
17283 whether a pixel should be blurred or not. The option value must be an
17284 integer in the range [-30,30]. A value of 0 will filter all the image,
17285 a value included in [0,30] will filter flat areas and a value included
17286 in [-30,0] will filter edges. Default value is 0.
17288 @item chroma_radius, cr
17289 Set the chroma radius. The option value must be a float number in
17290 the range [0.1,5.0] that specifies the variance of the gaussian filter
17291 used to blur the image (slower if larger). Default value is @option{luma_radius}.
17293 @item chroma_strength, cs
17294 Set the chroma strength. The option value must be a float number
17295 in the range [-1.0,1.0] that configures the blurring. A value included
17296 in [0.0,1.0] will blur the image whereas a value included in
17297 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
17299 @item chroma_threshold, ct
17300 Set the chroma threshold used as a coefficient to determine
17301 whether a pixel should be blurred or not. The option value must be an
17302 integer in the range [-30,30]. A value of 0 will filter all the image,
17303 a value included in [0,30] will filter flat areas and a value included
17304 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
17307 If a chroma option is not explicitly set, the corresponding luma value
17311 Apply sobel operator to input video stream.
17313 The filter accepts the following option:
17317 Set which planes will be processed, unprocessed planes will be copied.
17318 By default value 0xf, all planes will be processed.
17321 Set value which will be multiplied with filtered result.
17324 Set value which will be added to filtered result.
17330 Apply a simple postprocessing filter that compresses and decompresses the image
17331 at several (or - in the case of @option{quality} level @code{6} - all) shifts
17332 and average the results.
17334 The filter accepts the following options:
17338 Set quality. This option defines the number of levels for averaging. It accepts
17339 an integer in the range 0-6. If set to @code{0}, the filter will have no
17340 effect. A value of @code{6} means the higher quality. For each increment of
17341 that value the speed drops by a factor of approximately 2. Default value is
17345 Force a constant quantization parameter. If not set, the filter will use the QP
17346 from the video stream (if available).
17349 Set thresholding mode. Available modes are:
17353 Set hard thresholding (default).
17355 Set soft thresholding (better de-ringing effect, but likely blurrier).
17358 @item use_bframe_qp
17359 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
17360 option may cause flicker since the B-Frames have often larger QP. Default is
17361 @code{0} (not enabled).
17364 @subsection Commands
17366 This filter supports the following commands:
17368 @item quality, level
17369 Set quality level. The value @code{max} can be used to set the maximum level,
17370 currently @code{6}.
17376 Scale the input by applying one of the super-resolution methods based on
17377 convolutional neural networks. Supported models:
17381 Super-Resolution Convolutional Neural Network model (SRCNN).
17382 See @url{https://arxiv.org/abs/1501.00092}.
17385 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
17386 See @url{https://arxiv.org/abs/1609.05158}.
17389 Training scripts as well as scripts for model file (.pb) saving can be found at
17390 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
17391 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
17393 Native model files (.model) can be generated from TensorFlow model
17394 files (.pb) by using tools/python/convert.py
17396 The filter accepts the following options:
17400 Specify which DNN backend to use for model loading and execution. This option accepts
17401 the following values:
17405 Native implementation of DNN loading and execution.
17408 TensorFlow backend. To enable this backend you
17409 need to install the TensorFlow for C library (see
17410 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
17411 @code{--enable-libtensorflow}
17414 Default value is @samp{native}.
17417 Set path to model file specifying network architecture and its parameters.
17418 Note that different backends use different file formats. TensorFlow backend
17419 can load files for both formats, while native backend can load files for only
17423 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
17424 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
17425 input upscaled using bicubic upscaling with proper scale factor.
17428 This feature can also be finished with @ref{dnn_processing} filter.
17432 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
17434 This filter takes in input two input videos, the first input is
17435 considered the "main" source and is passed unchanged to the
17436 output. The second input is used as a "reference" video for computing
17439 Both video inputs must have the same resolution and pixel format for
17440 this filter to work correctly. Also it assumes that both inputs
17441 have the same number of frames, which are compared one by one.
17443 The filter stores the calculated SSIM of each frame.
17445 The description of the accepted parameters follows.
17448 @item stats_file, f
17449 If specified the filter will use the named file to save the SSIM of
17450 each individual frame. When filename equals "-" the data is sent to
17454 The file printed if @var{stats_file} is selected, contains a sequence of
17455 key/value pairs of the form @var{key}:@var{value} for each compared
17458 A description of each shown parameter follows:
17462 sequential number of the input frame, starting from 1
17464 @item Y, U, V, R, G, B
17465 SSIM of the compared frames for the component specified by the suffix.
17468 SSIM of the compared frames for the whole frame.
17471 Same as above but in dB representation.
17474 This filter also supports the @ref{framesync} options.
17476 @subsection Examples
17481 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
17482 [main][ref] ssim="stats_file=stats.log" [out]
17485 On this example the input file being processed is compared with the
17486 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
17487 is stored in @file{stats.log}.
17490 Another example with both psnr and ssim at same time:
17492 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
17496 Another example with different containers:
17498 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 -
17504 Convert between different stereoscopic image formats.
17506 The filters accept the following options:
17510 Set stereoscopic image format of input.
17512 Available values for input image formats are:
17515 side by side parallel (left eye left, right eye right)
17518 side by side crosseye (right eye left, left eye right)
17521 side by side parallel with half width resolution
17522 (left eye left, right eye right)
17525 side by side crosseye with half width resolution
17526 (right eye left, left eye right)
17530 above-below (left eye above, right eye below)
17534 above-below (right eye above, left eye below)
17538 above-below with half height resolution
17539 (left eye above, right eye below)
17543 above-below with half height resolution
17544 (right eye above, left eye below)
17547 alternating frames (left eye first, right eye second)
17550 alternating frames (right eye first, left eye second)
17553 interleaved rows (left eye has top row, right eye starts on next row)
17556 interleaved rows (right eye has top row, left eye starts on next row)
17559 interleaved columns, left eye first
17562 interleaved columns, right eye first
17564 Default value is @samp{sbsl}.
17568 Set stereoscopic image format of output.
17572 side by side parallel (left eye left, right eye right)
17575 side by side crosseye (right eye left, left eye right)
17578 side by side parallel with half width resolution
17579 (left eye left, right eye right)
17582 side by side crosseye with half width resolution
17583 (right eye left, left eye right)
17587 above-below (left eye above, right eye below)
17591 above-below (right eye above, left eye below)
17595 above-below with half height resolution
17596 (left eye above, right eye below)
17600 above-below with half height resolution
17601 (right eye above, left eye below)
17604 alternating frames (left eye first, right eye second)
17607 alternating frames (right eye first, left eye second)
17610 interleaved rows (left eye has top row, right eye starts on next row)
17613 interleaved rows (right eye has top row, left eye starts on next row)
17616 anaglyph red/blue gray
17617 (red filter on left eye, blue filter on right eye)
17620 anaglyph red/green gray
17621 (red filter on left eye, green filter on right eye)
17624 anaglyph red/cyan gray
17625 (red filter on left eye, cyan filter on right eye)
17628 anaglyph red/cyan half colored
17629 (red filter on left eye, cyan filter on right eye)
17632 anaglyph red/cyan color
17633 (red filter on left eye, cyan filter on right eye)
17636 anaglyph red/cyan color optimized with the least squares projection of dubois
17637 (red filter on left eye, cyan filter on right eye)
17640 anaglyph green/magenta gray
17641 (green filter on left eye, magenta filter on right eye)
17644 anaglyph green/magenta half colored
17645 (green filter on left eye, magenta filter on right eye)
17648 anaglyph green/magenta colored
17649 (green filter on left eye, magenta filter on right eye)
17652 anaglyph green/magenta color optimized with the least squares projection of dubois
17653 (green filter on left eye, magenta filter on right eye)
17656 anaglyph yellow/blue gray
17657 (yellow filter on left eye, blue filter on right eye)
17660 anaglyph yellow/blue half colored
17661 (yellow filter on left eye, blue filter on right eye)
17664 anaglyph yellow/blue colored
17665 (yellow filter on left eye, blue filter on right eye)
17668 anaglyph yellow/blue color optimized with the least squares projection of dubois
17669 (yellow filter on left eye, blue filter on right eye)
17672 mono output (left eye only)
17675 mono output (right eye only)
17678 checkerboard, left eye first
17681 checkerboard, right eye first
17684 interleaved columns, left eye first
17687 interleaved columns, right eye first
17693 Default value is @samp{arcd}.
17696 @subsection Examples
17700 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
17706 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
17712 @section streamselect, astreamselect
17713 Select video or audio streams.
17715 The filter accepts the following options:
17719 Set number of inputs. Default is 2.
17722 Set input indexes to remap to outputs.
17725 @subsection Commands
17727 The @code{streamselect} and @code{astreamselect} filter supports the following
17732 Set input indexes to remap to outputs.
17735 @subsection Examples
17739 Select first 5 seconds 1st stream and rest of time 2nd stream:
17741 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
17745 Same as above, but for audio:
17747 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
17754 Draw subtitles on top of input video using the libass library.
17756 To enable compilation of this filter you need to configure FFmpeg with
17757 @code{--enable-libass}. This filter also requires a build with libavcodec and
17758 libavformat to convert the passed subtitles file to ASS (Advanced Substation
17759 Alpha) subtitles format.
17761 The filter accepts the following options:
17765 Set the filename of the subtitle file to read. It must be specified.
17767 @item original_size
17768 Specify the size of the original video, the video for which the ASS file
17769 was composed. For the syntax of this option, check the
17770 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17771 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
17772 correctly scale the fonts if the aspect ratio has been changed.
17775 Set a directory path containing fonts that can be used by the filter.
17776 These fonts will be used in addition to whatever the font provider uses.
17779 Process alpha channel, by default alpha channel is untouched.
17782 Set subtitles input character encoding. @code{subtitles} filter only. Only
17783 useful if not UTF-8.
17785 @item stream_index, si
17786 Set subtitles stream index. @code{subtitles} filter only.
17789 Override default style or script info parameters of the subtitles. It accepts a
17790 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
17793 If the first key is not specified, it is assumed that the first value
17794 specifies the @option{filename}.
17796 For example, to render the file @file{sub.srt} on top of the input
17797 video, use the command:
17802 which is equivalent to:
17804 subtitles=filename=sub.srt
17807 To render the default subtitles stream from file @file{video.mkv}, use:
17809 subtitles=video.mkv
17812 To render the second subtitles stream from that file, use:
17814 subtitles=video.mkv:si=1
17817 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
17818 @code{DejaVu Serif}, use:
17820 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HCCFF0000'
17823 @section super2xsai
17825 Scale the input by 2x and smooth using the Super2xSaI (Scale and
17826 Interpolate) pixel art scaling algorithm.
17828 Useful for enlarging pixel art images without reducing sharpness.
17832 Swap two rectangular objects in video.
17834 This filter accepts the following options:
17844 Set 1st rect x coordinate.
17847 Set 1st rect y coordinate.
17850 Set 2nd rect x coordinate.
17853 Set 2nd rect y coordinate.
17855 All expressions are evaluated once for each frame.
17858 The all options are expressions containing the following constants:
17863 The input width and height.
17866 same as @var{w} / @var{h}
17869 input sample aspect ratio
17872 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
17875 The number of the input frame, starting from 0.
17878 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
17881 the position in the file of the input frame, NAN if unknown
17888 Blend successive video frames.
17894 Apply telecine process to the video.
17896 This filter accepts the following options:
17905 The default value is @code{top}.
17909 A string of numbers representing the pulldown pattern you wish to apply.
17910 The default value is @code{23}.
17914 Some typical patterns:
17919 24p: 2332 (preferred)
17926 24p: 222222222223 ("Euro pulldown")
17931 @section thistogram
17933 Compute and draw a color distribution histogram for the input video across time.
17935 Unlike @ref{histogram} video filter which only shows histogram of single input frame
17936 at certain time, this filter shows also past histograms of number of frames defined
17937 by @code{width} option.
17939 The computed histogram is a representation of the color component
17940 distribution in an image.
17942 The filter accepts the following options:
17946 Set width of single color component output. Default value is @code{0}.
17947 Value of @code{0} means width will be picked from input video.
17948 This also set number of passed histograms to keep.
17949 Allowed range is [0, 8192].
17951 @item display_mode, d
17953 It accepts the following values:
17956 Per color component graphs are placed below each other.
17959 Per color component graphs are placed side by side.
17962 Presents information identical to that in the @code{parade}, except
17963 that the graphs representing color components are superimposed directly
17966 Default is @code{stack}.
17968 @item levels_mode, m
17969 Set mode. Can be either @code{linear}, or @code{logarithmic}.
17970 Default is @code{linear}.
17972 @item components, c
17973 Set what color components to display.
17974 Default is @code{7}.
17977 Set background opacity. Default is @code{0.9}.
17980 Show envelope. Default is disabled.
17983 Set envelope color. Default is @code{gold}.
17988 Apply threshold effect to video stream.
17990 This filter needs four video streams to perform thresholding.
17991 First stream is stream we are filtering.
17992 Second stream is holding threshold values, third stream is holding min values,
17993 and last, fourth stream is holding max values.
17995 The filter accepts the following option:
17999 Set which planes will be processed, unprocessed planes will be copied.
18000 By default value 0xf, all planes will be processed.
18003 For example if first stream pixel's component value is less then threshold value
18004 of pixel component from 2nd threshold stream, third stream value will picked,
18005 otherwise fourth stream pixel component value will be picked.
18007 Using color source filter one can perform various types of thresholding:
18009 @subsection Examples
18013 Binary threshold, using gray color as threshold:
18015 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18019 Inverted binary threshold, using gray color as threshold:
18021 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18025 Truncate binary threshold, using gray color as threshold:
18027 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18031 Threshold to zero, using gray color as threshold:
18033 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18037 Inverted threshold to zero, using gray color as threshold:
18039 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18044 Select the most representative frame in a given sequence of consecutive frames.
18046 The filter accepts the following options:
18050 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18051 will pick one of them, and then handle the next batch of @var{n} frames until
18052 the end. Default is @code{100}.
18055 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18056 value will result in a higher memory usage, so a high value is not recommended.
18058 @subsection Examples
18062 Extract one picture each 50 frames:
18068 Complete example of a thumbnail creation with @command{ffmpeg}:
18070 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18076 Tile several successive frames together.
18078 The filter accepts the following options:
18083 Set the grid size (i.e. the number of lines and columns). For the syntax of
18084 this option, check the
18085 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18088 Set the maximum number of frames to render in the given area. It must be less
18089 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18090 the area will be used.
18093 Set the outer border margin in pixels.
18096 Set the inner border thickness (i.e. the number of pixels between frames). For
18097 more advanced padding options (such as having different values for the edges),
18098 refer to the pad video filter.
18101 Specify the color of the unused area. For the syntax of this option, check the
18102 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18103 The default value of @var{color} is "black".
18106 Set the number of frames to overlap when tiling several successive frames together.
18107 The value must be between @code{0} and @var{nb_frames - 1}.
18110 Set the number of frames to initially be empty before displaying first output frame.
18111 This controls how soon will one get first output frame.
18112 The value must be between @code{0} and @var{nb_frames - 1}.
18115 @subsection Examples
18119 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18121 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18123 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18124 duplicating each output frame to accommodate the originally detected frame
18128 Display @code{5} pictures in an area of @code{3x2} frames,
18129 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18130 mixed flat and named options:
18132 tile=3x2:nb_frames=5:padding=7:margin=2
18136 @section tinterlace
18138 Perform various types of temporal field interlacing.
18140 Frames are counted starting from 1, so the first input frame is
18143 The filter accepts the following options:
18148 Specify the mode of the interlacing. This option can also be specified
18149 as a value alone. See below for a list of values for this option.
18151 Available values are:
18155 Move odd frames into the upper field, even into the lower field,
18156 generating a double height frame at half frame rate.
18160 Frame 1 Frame 2 Frame 3 Frame 4
18162 11111 22222 33333 44444
18163 11111 22222 33333 44444
18164 11111 22222 33333 44444
18165 11111 22222 33333 44444
18179 Only output odd frames, even frames are dropped, generating a frame with
18180 unchanged height at half frame rate.
18185 Frame 1 Frame 2 Frame 3 Frame 4
18187 11111 22222 33333 44444
18188 11111 22222 33333 44444
18189 11111 22222 33333 44444
18190 11111 22222 33333 44444
18200 Only output even frames, odd frames are dropped, generating a frame with
18201 unchanged height at half frame rate.
18206 Frame 1 Frame 2 Frame 3 Frame 4
18208 11111 22222 33333 44444
18209 11111 22222 33333 44444
18210 11111 22222 33333 44444
18211 11111 22222 33333 44444
18221 Expand each frame to full height, but pad alternate lines with black,
18222 generating a frame with double height at the same input frame rate.
18227 Frame 1 Frame 2 Frame 3 Frame 4
18229 11111 22222 33333 44444
18230 11111 22222 33333 44444
18231 11111 22222 33333 44444
18232 11111 22222 33333 44444
18235 11111 ..... 33333 .....
18236 ..... 22222 ..... 44444
18237 11111 ..... 33333 .....
18238 ..... 22222 ..... 44444
18239 11111 ..... 33333 .....
18240 ..... 22222 ..... 44444
18241 11111 ..... 33333 .....
18242 ..... 22222 ..... 44444
18246 @item interleave_top, 4
18247 Interleave the upper field from odd frames with the lower field from
18248 even frames, generating a frame with unchanged height at half frame rate.
18253 Frame 1 Frame 2 Frame 3 Frame 4
18255 11111<- 22222 33333<- 44444
18256 11111 22222<- 33333 44444<-
18257 11111<- 22222 33333<- 44444
18258 11111 22222<- 33333 44444<-
18268 @item interleave_bottom, 5
18269 Interleave the lower field from odd frames with the upper field from
18270 even frames, generating a frame with unchanged height at half frame rate.
18275 Frame 1 Frame 2 Frame 3 Frame 4
18277 11111 22222<- 33333 44444<-
18278 11111<- 22222 33333<- 44444
18279 11111 22222<- 33333 44444<-
18280 11111<- 22222 33333<- 44444
18290 @item interlacex2, 6
18291 Double frame rate with unchanged height. Frames are inserted each
18292 containing the second temporal field from the previous input frame and
18293 the first temporal field from the next input frame. This mode relies on
18294 the top_field_first flag. Useful for interlaced video displays with no
18295 field synchronisation.
18300 Frame 1 Frame 2 Frame 3 Frame 4
18302 11111 22222 33333 44444
18303 11111 22222 33333 44444
18304 11111 22222 33333 44444
18305 11111 22222 33333 44444
18308 11111 22222 22222 33333 33333 44444 44444
18309 11111 11111 22222 22222 33333 33333 44444
18310 11111 22222 22222 33333 33333 44444 44444
18311 11111 11111 22222 22222 33333 33333 44444
18316 Move odd frames into the upper field, even into the lower field,
18317 generating a double height frame at same frame rate.
18322 Frame 1 Frame 2 Frame 3 Frame 4
18324 11111 22222 33333 44444
18325 11111 22222 33333 44444
18326 11111 22222 33333 44444
18327 11111 22222 33333 44444
18330 11111 33333 33333 55555
18331 22222 22222 44444 44444
18332 11111 33333 33333 55555
18333 22222 22222 44444 44444
18334 11111 33333 33333 55555
18335 22222 22222 44444 44444
18336 11111 33333 33333 55555
18337 22222 22222 44444 44444
18342 Numeric values are deprecated but are accepted for backward
18343 compatibility reasons.
18345 Default mode is @code{merge}.
18348 Specify flags influencing the filter process.
18350 Available value for @var{flags} is:
18353 @item low_pass_filter, vlpf
18354 Enable linear vertical low-pass filtering in the filter.
18355 Vertical low-pass filtering is required when creating an interlaced
18356 destination from a progressive source which contains high-frequency
18357 vertical detail. Filtering will reduce interlace 'twitter' and Moire
18360 @item complex_filter, cvlpf
18361 Enable complex vertical low-pass filtering.
18362 This will slightly less reduce interlace 'twitter' and Moire
18363 patterning but better retain detail and subjective sharpness impression.
18366 Bypass already interlaced frames, only adjust the frame rate.
18369 Vertical low-pass filtering and bypassing already interlaced frames can only be
18370 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
18376 Mix successive video frames.
18378 A description of the accepted options follows.
18382 The number of successive frames to mix. If unspecified, it defaults to 3.
18385 Specify weight of each input video frame.
18386 Each weight is separated by space. If number of weights is smaller than
18387 number of @var{frames} last specified weight will be used for all remaining
18391 Specify scale, if it is set it will be multiplied with sum
18392 of each weight multiplied with pixel values to give final destination
18393 pixel value. By default @var{scale} is auto scaled to sum of weights.
18396 @subsection Examples
18400 Average 7 successive frames:
18402 tmix=frames=7:weights="1 1 1 1 1 1 1"
18406 Apply simple temporal convolution:
18408 tmix=frames=3:weights="-1 3 -1"
18412 Similar as above but only showing temporal differences:
18414 tmix=frames=3:weights="-1 2 -1":scale=1
18420 Tone map colors from different dynamic ranges.
18422 This filter expects data in single precision floating point, as it needs to
18423 operate on (and can output) out-of-range values. Another filter, such as
18424 @ref{zscale}, is needed to convert the resulting frame to a usable format.
18426 The tonemapping algorithms implemented only work on linear light, so input
18427 data should be linearized beforehand (and possibly correctly tagged).
18430 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
18433 @subsection Options
18434 The filter accepts the following options.
18438 Set the tone map algorithm to use.
18440 Possible values are:
18443 Do not apply any tone map, only desaturate overbright pixels.
18446 Hard-clip any out-of-range values. Use it for perfect color accuracy for
18447 in-range values, while distorting out-of-range values.
18450 Stretch the entire reference gamut to a linear multiple of the display.
18453 Fit a logarithmic transfer between the tone curves.
18456 Preserve overall image brightness with a simple curve, using nonlinear
18457 contrast, which results in flattening details and degrading color accuracy.
18460 Preserve both dark and bright details better than @var{reinhard}, at the cost
18461 of slightly darkening everything. Use it when detail preservation is more
18462 important than color and brightness accuracy.
18465 Smoothly map out-of-range values, while retaining contrast and colors for
18466 in-range material as much as possible. Use it when color accuracy is more
18467 important than detail preservation.
18473 Tune the tone mapping algorithm.
18475 This affects the following algorithms:
18481 Specifies the scale factor to use while stretching.
18485 Specifies the exponent of the function.
18489 Specify an extra linear coefficient to multiply into the signal before clipping.
18493 Specify the local contrast coefficient at the display peak.
18494 Default to 0.5, which means that in-gamut values will be about half as bright
18501 Specify the transition point from linear to mobius transform. Every value
18502 below this point is guaranteed to be mapped 1:1. The higher the value, the
18503 more accurate the result will be, at the cost of losing bright details.
18504 Default to 0.3, which due to the steep initial slope still preserves in-range
18505 colors fairly accurately.
18509 Apply desaturation for highlights that exceed this level of brightness. The
18510 higher the parameter, the more color information will be preserved. This
18511 setting helps prevent unnaturally blown-out colors for super-highlights, by
18512 (smoothly) turning into white instead. This makes images feel more natural,
18513 at the cost of reducing information about out-of-range colors.
18515 The default of 2.0 is somewhat conservative and will mostly just apply to
18516 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
18518 This option works only if the input frame has a supported color tag.
18521 Override signal/nominal/reference peak with this value. Useful when the
18522 embedded peak information in display metadata is not reliable or when tone
18523 mapping from a lower range to a higher range.
18528 Temporarily pad video frames.
18530 The filter accepts the following options:
18534 Specify number of delay frames before input video stream. Default is 0.
18537 Specify number of padding frames after input video stream.
18538 Set to -1 to pad indefinitely. Default is 0.
18541 Set kind of frames added to beginning of stream.
18542 Can be either @var{add} or @var{clone}.
18543 With @var{add} frames of solid-color are added.
18544 With @var{clone} frames are clones of first frame.
18545 Default is @var{add}.
18548 Set kind of frames added to end of stream.
18549 Can be either @var{add} or @var{clone}.
18550 With @var{add} frames of solid-color are added.
18551 With @var{clone} frames are clones of last frame.
18552 Default is @var{add}.
18554 @item start_duration, stop_duration
18555 Specify the duration of the start/stop delay. See
18556 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18557 for the accepted syntax.
18558 These options override @var{start} and @var{stop}. Default is 0.
18561 Specify the color of the padded area. For the syntax of this option,
18562 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
18563 manual,ffmpeg-utils}.
18565 The default value of @var{color} is "black".
18571 Transpose rows with columns in the input video and optionally flip it.
18573 It accepts the following parameters:
18578 Specify the transposition direction.
18580 Can assume the following values:
18582 @item 0, 4, cclock_flip
18583 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
18591 Rotate by 90 degrees clockwise, that is:
18599 Rotate by 90 degrees counterclockwise, that is:
18606 @item 3, 7, clock_flip
18607 Rotate by 90 degrees clockwise and vertically flip, that is:
18615 For values between 4-7, the transposition is only done if the input
18616 video geometry is portrait and not landscape. These values are
18617 deprecated, the @code{passthrough} option should be used instead.
18619 Numerical values are deprecated, and should be dropped in favor of
18620 symbolic constants.
18623 Do not apply the transposition if the input geometry matches the one
18624 specified by the specified value. It accepts the following values:
18627 Always apply transposition.
18629 Preserve portrait geometry (when @var{height} >= @var{width}).
18631 Preserve landscape geometry (when @var{width} >= @var{height}).
18634 Default value is @code{none}.
18637 For example to rotate by 90 degrees clockwise and preserve portrait
18640 transpose=dir=1:passthrough=portrait
18643 The command above can also be specified as:
18645 transpose=1:portrait
18648 @section transpose_npp
18650 Transpose rows with columns in the input video and optionally flip it.
18651 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
18653 It accepts the following parameters:
18658 Specify the transposition direction.
18660 Can assume the following values:
18663 Rotate by 90 degrees counterclockwise and vertically flip. (default)
18666 Rotate by 90 degrees clockwise.
18669 Rotate by 90 degrees counterclockwise.
18672 Rotate by 90 degrees clockwise and vertically flip.
18676 Do not apply the transposition if the input geometry matches the one
18677 specified by the specified value. It accepts the following values:
18680 Always apply transposition. (default)
18682 Preserve portrait geometry (when @var{height} >= @var{width}).
18684 Preserve landscape geometry (when @var{width} >= @var{height}).
18690 Trim the input so that the output contains one continuous subpart of the input.
18692 It accepts the following parameters:
18695 Specify the time of the start of the kept section, i.e. the frame with the
18696 timestamp @var{start} will be the first frame in the output.
18699 Specify the time of the first frame that will be dropped, i.e. the frame
18700 immediately preceding the one with the timestamp @var{end} will be the last
18701 frame in the output.
18704 This is the same as @var{start}, except this option sets the start timestamp
18705 in timebase units instead of seconds.
18708 This is the same as @var{end}, except this option sets the end timestamp
18709 in timebase units instead of seconds.
18712 The maximum duration of the output in seconds.
18715 The number of the first frame that should be passed to the output.
18718 The number of the first frame that should be dropped.
18721 @option{start}, @option{end}, and @option{duration} are expressed as time
18722 duration specifications; see
18723 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18724 for the accepted syntax.
18726 Note that the first two sets of the start/end options and the @option{duration}
18727 option look at the frame timestamp, while the _frame variants simply count the
18728 frames that pass through the filter. Also note that this filter does not modify
18729 the timestamps. If you wish for the output timestamps to start at zero, insert a
18730 setpts filter after the trim filter.
18732 If multiple start or end options are set, this filter tries to be greedy and
18733 keep all the frames that match at least one of the specified constraints. To keep
18734 only the part that matches all the constraints at once, chain multiple trim
18737 The defaults are such that all the input is kept. So it is possible to set e.g.
18738 just the end values to keep everything before the specified time.
18743 Drop everything except the second minute of input:
18745 ffmpeg -i INPUT -vf trim=60:120
18749 Keep only the first second:
18751 ffmpeg -i INPUT -vf trim=duration=1
18756 @section unpremultiply
18757 Apply alpha unpremultiply effect to input video stream using first plane
18758 of second stream as alpha.
18760 Both streams must have same dimensions and same pixel format.
18762 The filter accepts the following option:
18766 Set which planes will be processed, unprocessed planes will be copied.
18767 By default value 0xf, all planes will be processed.
18769 If the format has 1 or 2 components, then luma is bit 0.
18770 If the format has 3 or 4 components:
18771 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
18772 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
18773 If present, the alpha channel is always the last bit.
18776 Do not require 2nd input for processing, instead use alpha plane from input stream.
18782 Sharpen or blur the input video.
18784 It accepts the following parameters:
18787 @item luma_msize_x, lx
18788 Set the luma matrix horizontal size. It must be an odd integer between
18789 3 and 23. The default value is 5.
18791 @item luma_msize_y, ly
18792 Set the luma matrix vertical size. It must be an odd integer between 3
18793 and 23. The default value is 5.
18795 @item luma_amount, la
18796 Set the luma effect strength. It must be a floating point number, reasonable
18797 values lay between -1.5 and 1.5.
18799 Negative values will blur the input video, while positive values will
18800 sharpen it, a value of zero will disable the effect.
18802 Default value is 1.0.
18804 @item chroma_msize_x, cx
18805 Set the chroma matrix horizontal size. It must be an odd integer
18806 between 3 and 23. The default value is 5.
18808 @item chroma_msize_y, cy
18809 Set the chroma matrix vertical size. It must be an odd integer
18810 between 3 and 23. The default value is 5.
18812 @item chroma_amount, ca
18813 Set the chroma effect strength. It must be a floating point number, reasonable
18814 values lay between -1.5 and 1.5.
18816 Negative values will blur the input video, while positive values will
18817 sharpen it, a value of zero will disable the effect.
18819 Default value is 0.0.
18823 All parameters are optional and default to the equivalent of the
18824 string '5:5:1.0:5:5:0.0'.
18826 @subsection Examples
18830 Apply strong luma sharpen effect:
18832 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
18836 Apply a strong blur of both luma and chroma parameters:
18838 unsharp=7:7:-2:7:7:-2
18844 Apply ultra slow/simple postprocessing filter that compresses and decompresses
18845 the image at several (or - in the case of @option{quality} level @code{8} - all)
18846 shifts and average the results.
18848 The way this differs from the behavior of spp is that uspp actually encodes &
18849 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
18850 DCT similar to MJPEG.
18852 The filter accepts the following options:
18856 Set quality. This option defines the number of levels for averaging. It accepts
18857 an integer in the range 0-8. If set to @code{0}, the filter will have no
18858 effect. A value of @code{8} means the higher quality. For each increment of
18859 that value the speed drops by a factor of approximately 2. Default value is
18863 Force a constant quantization parameter. If not set, the filter will use the QP
18864 from the video stream (if available).
18869 Convert 360 videos between various formats.
18871 The filter accepts the following options:
18877 Set format of the input/output video.
18885 Equirectangular projection.
18890 Cubemap with 3x2/6x1/1x6 layout.
18892 Format specific options:
18897 Set padding proportion for the input/output cubemap. Values in decimals.
18904 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)
18907 Default value is @b{@samp{0}}.
18911 Set fixed padding for the input/output cubemap. Values in pixels.
18913 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
18917 Set order of faces for the input/output cubemap. Choose one direction for each position.
18919 Designation of directions:
18935 Default value is @b{@samp{rludfb}}.
18939 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
18941 Designation of angles:
18944 0 degrees clockwise
18946 90 degrees clockwise
18948 180 degrees clockwise
18950 270 degrees clockwise
18953 Default value is @b{@samp{000000}}.
18957 Equi-Angular Cubemap.
18964 Format specific options:
18969 Set output horizontal/vertical/diagonal field of view. Values in degrees.
18971 If diagonal field of view is set it overrides horizontal and vertical field of view.
18976 Set input horizontal/vertical/diagonal field of view. Values in degrees.
18978 If diagonal field of view is set it overrides horizontal and vertical field of view.
18984 Format specific options:
18988 Set padding proportion. Values in decimals.
18998 Default value is @b{@samp{0}}.
19004 Facebook's 360 formats.
19007 Stereographic format.
19009 Format specific options:
19014 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19016 If diagonal field of view is set it overrides horizontal and vertical field of view.
19021 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19023 If diagonal field of view is set it overrides horizontal and vertical field of view.
19030 Ball format, gives significant distortion toward the back.
19033 Hammer-Aitoff map projection format.
19036 Sinusoidal map projection format.
19039 Fisheye projection.
19041 Format specific options:
19046 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19048 If diagonal field of view is set it overrides horizontal and vertical field of view.
19053 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19055 If diagonal field of view is set it overrides horizontal and vertical field of view.
19059 Pannini projection.
19061 Format specific options:
19064 Set output pannini parameter.
19067 Set input pannini parameter.
19071 Cylindrical projection.
19073 Format specific options:
19078 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19080 If diagonal field of view is set it overrides horizontal and vertical field of view.
19085 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19087 If diagonal field of view is set it overrides horizontal and vertical field of view.
19091 Perspective projection. @i{(output only)}
19093 Format specific options:
19096 Set perspective parameter.
19100 Tetrahedron projection.
19103 Truncated square pyramid projection.
19107 Half equirectangular projection.
19111 Set interpolation method.@*
19112 @i{Note: more complex interpolation methods require much more memory to run.}
19122 Bilinear interpolation.
19124 Lagrange9 interpolation.
19127 Bicubic interpolation.
19130 Lanczos interpolation.
19133 Spline16 interpolation.
19136 Gaussian interpolation.
19139 Default value is @b{@samp{line}}.
19143 Set the output video resolution.
19145 Default resolution depends on formats.
19149 Set the input/output stereo format.
19160 Default value is @b{@samp{2d}} for input and output format.
19165 Set rotation for the output video. Values in degrees.
19168 Set rotation order for the output video. Choose one item for each position.
19179 Default value is @b{@samp{ypr}}.
19184 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
19188 Set if input video is flipped horizontally/vertically. Boolean values.
19191 Set if input video is transposed. Boolean value, by default disabled.
19194 Set if output video needs to be transposed. Boolean value, by default disabled.
19197 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
19200 @subsection Examples
19204 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
19206 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
19209 Extract back view of Equi-Angular Cubemap:
19211 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
19214 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
19216 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
19220 @subsection Commands
19222 This filter supports subset of above options as @ref{commands}.
19224 @section vaguedenoiser
19226 Apply a wavelet based denoiser.
19228 It transforms each frame from the video input into the wavelet domain,
19229 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
19230 the obtained coefficients. It does an inverse wavelet transform after.
19231 Due to wavelet properties, it should give a nice smoothed result, and
19232 reduced noise, without blurring picture features.
19234 This filter accepts the following options:
19238 The filtering strength. The higher, the more filtered the video will be.
19239 Hard thresholding can use a higher threshold than soft thresholding
19240 before the video looks overfiltered. Default value is 2.
19243 The filtering method the filter will use.
19245 It accepts the following values:
19248 All values under the threshold will be zeroed.
19251 All values under the threshold will be zeroed. All values above will be
19252 reduced by the threshold.
19255 Scales or nullifies coefficients - intermediary between (more) soft and
19256 (less) hard thresholding.
19259 Default is garrote.
19262 Number of times, the wavelet will decompose the picture. Picture can't
19263 be decomposed beyond a particular point (typically, 8 for a 640x480
19264 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
19267 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
19270 A list of the planes to process. By default all planes are processed.
19273 @section vectorscope
19275 Display 2 color component values in the two dimensional graph (which is called
19278 This filter accepts the following options:
19282 Set vectorscope mode.
19284 It accepts the following values:
19288 Gray values are displayed on graph, higher brightness means more pixels have
19289 same component color value on location in graph. This is the default mode.
19292 Gray values are displayed on graph. Surrounding pixels values which are not
19293 present in video frame are drawn in gradient of 2 color components which are
19294 set by option @code{x} and @code{y}. The 3rd color component is static.
19297 Actual color components values present in video frame are displayed on graph.
19300 Similar as color2 but higher frequency of same values @code{x} and @code{y}
19301 on graph increases value of another color component, which is luminance by
19302 default values of @code{x} and @code{y}.
19305 Actual colors present in video frame are displayed on graph. If two different
19306 colors map to same position on graph then color with higher value of component
19307 not present in graph is picked.
19310 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
19311 component picked from radial gradient.
19315 Set which color component will be represented on X-axis. Default is @code{1}.
19318 Set which color component will be represented on Y-axis. Default is @code{2}.
19321 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
19322 of color component which represents frequency of (X, Y) location in graph.
19327 No envelope, this is default.
19330 Instant envelope, even darkest single pixel will be clearly highlighted.
19333 Hold maximum and minimum values presented in graph over time. This way you
19334 can still spot out of range values without constantly looking at vectorscope.
19337 Peak and instant envelope combined together.
19341 Set what kind of graticule to draw.
19350 Set graticule opacity.
19353 Set graticule flags.
19357 Draw graticule for white point.
19360 Draw graticule for black point.
19363 Draw color points short names.
19367 Set background opacity.
19369 @item lthreshold, l
19370 Set low threshold for color component not represented on X or Y axis.
19371 Values lower than this value will be ignored. Default is 0.
19372 Note this value is multiplied with actual max possible value one pixel component
19373 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
19376 @item hthreshold, h
19377 Set high threshold for color component not represented on X or Y axis.
19378 Values higher than this value will be ignored. Default is 1.
19379 Note this value is multiplied with actual max possible value one pixel component
19380 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
19381 is 0.9 * 255 = 230.
19383 @item colorspace, c
19384 Set what kind of colorspace to use when drawing graticule.
19394 Set color tint for gray/tint vectorscope mode. By default both options are zero.
19395 This means no tint, and output will remain gray.
19398 @anchor{vidstabdetect}
19399 @section vidstabdetect
19401 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
19402 @ref{vidstabtransform} for pass 2.
19404 This filter generates a file with relative translation and rotation
19405 transform information about subsequent frames, which is then used by
19406 the @ref{vidstabtransform} filter.
19408 To enable compilation of this filter you need to configure FFmpeg with
19409 @code{--enable-libvidstab}.
19411 This filter accepts the following options:
19415 Set the path to the file used to write the transforms information.
19416 Default value is @file{transforms.trf}.
19419 Set how shaky the video is and how quick the camera is. It accepts an
19420 integer in the range 1-10, a value of 1 means little shakiness, a
19421 value of 10 means strong shakiness. Default value is 5.
19424 Set the accuracy of the detection process. It must be a value in the
19425 range 1-15. A value of 1 means low accuracy, a value of 15 means high
19426 accuracy. Default value is 15.
19429 Set stepsize of the search process. The region around minimum is
19430 scanned with 1 pixel resolution. Default value is 6.
19433 Set minimum contrast. Below this value a local measurement field is
19434 discarded. Must be a floating point value in the range 0-1. Default
19438 Set reference frame number for tripod mode.
19440 If enabled, the motion of the frames is compared to a reference frame
19441 in the filtered stream, identified by the specified number. The idea
19442 is to compensate all movements in a more-or-less static scene and keep
19443 the camera view absolutely still.
19445 If set to 0, it is disabled. The frames are counted starting from 1.
19448 Show fields and transforms in the resulting frames. It accepts an
19449 integer in the range 0-2. Default value is 0, which disables any
19453 @subsection Examples
19457 Use default values:
19463 Analyze strongly shaky movie and put the results in file
19464 @file{mytransforms.trf}:
19466 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
19470 Visualize the result of internal transformations in the resulting
19473 vidstabdetect=show=1
19477 Analyze a video with medium shakiness using @command{ffmpeg}:
19479 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
19483 @anchor{vidstabtransform}
19484 @section vidstabtransform
19486 Video stabilization/deshaking: pass 2 of 2,
19487 see @ref{vidstabdetect} for pass 1.
19489 Read a file with transform information for each frame and
19490 apply/compensate them. Together with the @ref{vidstabdetect}
19491 filter this can be used to deshake videos. See also
19492 @url{http://public.hronopik.de/vid.stab}. It is important to also use
19493 the @ref{unsharp} filter, see below.
19495 To enable compilation of this filter you need to configure FFmpeg with
19496 @code{--enable-libvidstab}.
19498 @subsection Options
19502 Set path to the file used to read the transforms. Default value is
19503 @file{transforms.trf}.
19506 Set the number of frames (value*2 + 1) used for lowpass filtering the
19507 camera movements. Default value is 10.
19509 For example a number of 10 means that 21 frames are used (10 in the
19510 past and 10 in the future) to smoothen the motion in the video. A
19511 larger value leads to a smoother video, but limits the acceleration of
19512 the camera (pan/tilt movements). 0 is a special case where a static
19513 camera is simulated.
19516 Set the camera path optimization algorithm.
19518 Accepted values are:
19521 gaussian kernel low-pass filter on camera motion (default)
19523 averaging on transformations
19527 Set maximal number of pixels to translate frames. Default value is -1,
19531 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
19532 value is -1, meaning no limit.
19535 Specify how to deal with borders that may be visible due to movement
19538 Available values are:
19541 keep image information from previous frame (default)
19543 fill the border black
19547 Invert transforms if set to 1. Default value is 0.
19550 Consider transforms as relative to previous frame if set to 1,
19551 absolute if set to 0. Default value is 0.
19554 Set percentage to zoom. A positive value will result in a zoom-in
19555 effect, a negative value in a zoom-out effect. Default value is 0 (no
19559 Set optimal zooming to avoid borders.
19561 Accepted values are:
19566 optimal static zoom value is determined (only very strong movements
19567 will lead to visible borders) (default)
19569 optimal adaptive zoom value is determined (no borders will be
19570 visible), see @option{zoomspeed}
19573 Note that the value given at zoom is added to the one calculated here.
19576 Set percent to zoom maximally each frame (enabled when
19577 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
19581 Specify type of interpolation.
19583 Available values are:
19588 linear only horizontal
19590 linear in both directions (default)
19592 cubic in both directions (slow)
19596 Enable virtual tripod mode if set to 1, which is equivalent to
19597 @code{relative=0:smoothing=0}. Default value is 0.
19599 Use also @code{tripod} option of @ref{vidstabdetect}.
19602 Increase log verbosity if set to 1. Also the detected global motions
19603 are written to the temporary file @file{global_motions.trf}. Default
19607 @subsection Examples
19611 Use @command{ffmpeg} for a typical stabilization with default values:
19613 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
19616 Note the use of the @ref{unsharp} filter which is always recommended.
19619 Zoom in a bit more and load transform data from a given file:
19621 vidstabtransform=zoom=5:input="mytransforms.trf"
19625 Smoothen the video even more:
19627 vidstabtransform=smoothing=30
19633 Flip the input video vertically.
19635 For example, to vertically flip a video with @command{ffmpeg}:
19637 ffmpeg -i in.avi -vf "vflip" out.avi
19642 Detect variable frame rate video.
19644 This filter tries to detect if the input is variable or constant frame rate.
19646 At end it will output number of frames detected as having variable delta pts,
19647 and ones with constant delta pts.
19648 If there was frames with variable delta, than it will also show min, max and
19649 average delta encountered.
19653 Boost or alter saturation.
19655 The filter accepts the following options:
19658 Set strength of boost if positive value or strength of alter if negative value.
19659 Default is 0. Allowed range is from -2 to 2.
19662 Set the red balance. Default is 1. Allowed range is from -10 to 10.
19665 Set the green balance. Default is 1. Allowed range is from -10 to 10.
19668 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
19671 Set the red luma coefficient.
19674 Set the green luma coefficient.
19677 Set the blue luma coefficient.
19680 If @code{intensity} is negative and this is set to 1, colors will change,
19681 otherwise colors will be less saturated, more towards gray.
19684 @subsection Commands
19686 This filter supports the all above options as @ref{commands}.
19691 Make or reverse a natural vignetting effect.
19693 The filter accepts the following options:
19697 Set lens angle expression as a number of radians.
19699 The value is clipped in the @code{[0,PI/2]} range.
19701 Default value: @code{"PI/5"}
19705 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
19709 Set forward/backward mode.
19711 Available modes are:
19714 The larger the distance from the central point, the darker the image becomes.
19717 The larger the distance from the central point, the brighter the image becomes.
19718 This can be used to reverse a vignette effect, though there is no automatic
19719 detection to extract the lens @option{angle} and other settings (yet). It can
19720 also be used to create a burning effect.
19723 Default value is @samp{forward}.
19726 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
19728 It accepts the following values:
19731 Evaluate expressions only once during the filter initialization.
19734 Evaluate expressions for each incoming frame. This is way slower than the
19735 @samp{init} mode since it requires all the scalers to be re-computed, but it
19736 allows advanced dynamic expressions.
19739 Default value is @samp{init}.
19742 Set dithering to reduce the circular banding effects. Default is @code{1}
19746 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
19747 Setting this value to the SAR of the input will make a rectangular vignetting
19748 following the dimensions of the video.
19750 Default is @code{1/1}.
19753 @subsection Expressions
19755 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
19756 following parameters.
19761 input width and height
19764 the number of input frame, starting from 0
19767 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
19768 @var{TB} units, NAN if undefined
19771 frame rate of the input video, NAN if the input frame rate is unknown
19774 the PTS (Presentation TimeStamp) of the filtered video frame,
19775 expressed in seconds, NAN if undefined
19778 time base of the input video
19782 @subsection Examples
19786 Apply simple strong vignetting effect:
19792 Make a flickering vignetting:
19794 vignette='PI/4+random(1)*PI/50':eval=frame
19799 @section vmafmotion
19801 Obtain the average VMAF motion score of a video.
19802 It is one of the component metrics of VMAF.
19804 The obtained average motion score is printed through the logging system.
19806 The filter accepts the following options:
19810 If specified, the filter will use the named file to save the motion score of
19811 each frame with respect to the previous frame.
19812 When filename equals "-" the data is sent to standard output.
19817 ffmpeg -i ref.mpg -vf vmafmotion -f null -
19821 Stack input videos vertically.
19823 All streams must be of same pixel format and of same width.
19825 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
19826 to create same output.
19828 The filter accepts the following options:
19832 Set number of input streams. Default is 2.
19835 If set to 1, force the output to terminate when the shortest input
19836 terminates. Default value is 0.
19841 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
19842 Deinterlacing Filter").
19844 Based on the process described by Martin Weston for BBC R&D, and
19845 implemented based on the de-interlace algorithm written by Jim
19846 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
19847 uses filter coefficients calculated by BBC R&D.
19849 This filter uses field-dominance information in frame to decide which
19850 of each pair of fields to place first in the output.
19851 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
19853 There are two sets of filter coefficients, so called "simple"
19854 and "complex". Which set of filter coefficients is used can
19855 be set by passing an optional parameter:
19859 Set the interlacing filter coefficients. Accepts one of the following values:
19863 Simple filter coefficient set.
19865 More-complex filter coefficient set.
19867 Default value is @samp{complex}.
19870 Specify which frames to deinterlace. Accepts one of the following values:
19874 Deinterlace all frames,
19876 Only deinterlace frames marked as interlaced.
19879 Default value is @samp{all}.
19883 Video waveform monitor.
19885 The waveform monitor plots color component intensity. By default luminance
19886 only. Each column of the waveform corresponds to a column of pixels in the
19889 It accepts the following options:
19893 Can be either @code{row}, or @code{column}. Default is @code{column}.
19894 In row mode, the graph on the left side represents color component value 0 and
19895 the right side represents value = 255. In column mode, the top side represents
19896 color component value = 0 and bottom side represents value = 255.
19899 Set intensity. Smaller values are useful to find out how many values of the same
19900 luminance are distributed across input rows/columns.
19901 Default value is @code{0.04}. Allowed range is [0, 1].
19904 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
19905 In mirrored mode, higher values will be represented on the left
19906 side for @code{row} mode and at the top for @code{column} mode. Default is
19907 @code{1} (mirrored).
19911 It accepts the following values:
19914 Presents information identical to that in the @code{parade}, except
19915 that the graphs representing color components are superimposed directly
19918 This display mode makes it easier to spot relative differences or similarities
19919 in overlapping areas of the color components that are supposed to be identical,
19920 such as neutral whites, grays, or blacks.
19923 Display separate graph for the color components side by side in
19924 @code{row} mode or one below the other in @code{column} mode.
19927 Display separate graph for the color components side by side in
19928 @code{column} mode or one below the other in @code{row} mode.
19930 Using this display mode makes it easy to spot color casts in the highlights
19931 and shadows of an image, by comparing the contours of the top and the bottom
19932 graphs of each waveform. Since whites, grays, and blacks are characterized
19933 by exactly equal amounts of red, green, and blue, neutral areas of the picture
19934 should display three waveforms of roughly equal width/height. If not, the
19935 correction is easy to perform by making level adjustments the three waveforms.
19937 Default is @code{stack}.
19939 @item components, c
19940 Set which color components to display. Default is 1, which means only luminance
19941 or red color component if input is in RGB colorspace. If is set for example to
19942 7 it will display all 3 (if) available color components.
19947 No envelope, this is default.
19950 Instant envelope, minimum and maximum values presented in graph will be easily
19951 visible even with small @code{step} value.
19954 Hold minimum and maximum values presented in graph across time. This way you
19955 can still spot out of range values without constantly looking at waveforms.
19958 Peak and instant envelope combined together.
19964 No filtering, this is default.
19967 Luma and chroma combined together.
19970 Similar as above, but shows difference between blue and red chroma.
19973 Similar as above, but use different colors.
19976 Similar as above, but again with different colors.
19979 Displays only chroma.
19982 Displays actual color value on waveform.
19985 Similar as above, but with luma showing frequency of chroma values.
19989 Set which graticule to display.
19993 Do not display graticule.
19996 Display green graticule showing legal broadcast ranges.
19999 Display orange graticule showing legal broadcast ranges.
20002 Display invert graticule showing legal broadcast ranges.
20006 Set graticule opacity.
20009 Set graticule flags.
20013 Draw numbers above lines. By default enabled.
20016 Draw dots instead of lines.
20020 Set scale used for displaying graticule.
20027 Default is digital.
20030 Set background opacity.
20034 Set tint for output.
20035 Only used with lowpass filter and when display is not overlay and input
20036 pixel formats are not RGB.
20039 @section weave, doubleweave
20041 The @code{weave} takes a field-based video input and join
20042 each two sequential fields into single frame, producing a new double
20043 height clip with half the frame rate and half the frame count.
20045 The @code{doubleweave} works same as @code{weave} but without
20046 halving frame rate and frame count.
20048 It accepts the following option:
20052 Set first field. Available values are:
20056 Set the frame as top-field-first.
20059 Set the frame as bottom-field-first.
20063 @subsection Examples
20067 Interlace video using @ref{select} and @ref{separatefields} filter:
20069 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
20074 Apply the xBR high-quality magnification filter which is designed for pixel
20075 art. It follows a set of edge-detection rules, see
20076 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
20078 It accepts the following option:
20082 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
20083 @code{3xBR} and @code{4} for @code{4xBR}.
20084 Default is @code{3}.
20089 Apply cross fade from one input video stream to another input video stream.
20090 The cross fade is applied for specified duration.
20092 The filter accepts the following options:
20096 Set one of available transition effects:
20136 Default transition effect is fade.
20139 Set cross fade duration in seconds.
20140 Default duration is 1 second.
20143 Set cross fade start relative to first input stream in seconds.
20144 Default offset is 0.
20147 Set expression for custom transition effect.
20149 The expressions can use the following variables and functions:
20154 The coordinates of the current sample.
20158 The width and height of the image.
20161 Progress of transition effect.
20164 Currently processed plane.
20167 Return value of first input at current location and plane.
20170 Return value of second input at current location and plane.
20176 Return the value of the pixel at location (@var{x},@var{y}) of the
20177 first/second/third/fourth component of first input.
20183 Return the value of the pixel at location (@var{x},@var{y}) of the
20184 first/second/third/fourth component of second input.
20188 @subsection Examples
20192 Cross fade from one input video to another input video, with fade transition and duration of transition
20193 of 2 seconds starting at offset of 5 seconds:
20195 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
20200 Pick median pixels from several input videos.
20202 The filter accepts the following options:
20206 Set number of inputs.
20207 Default is 3. Allowed range is from 3 to 255.
20208 If number of inputs is even number, than result will be mean value between two median values.
20211 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
20214 Set median percentile. Default value is @code{0.5}.
20215 Default value of @code{0.5} will pick always median values, while @code{0} will pick
20216 minimum values, and @code{1} maximum values.
20220 Stack video inputs into custom layout.
20222 All streams must be of same pixel format.
20224 The filter accepts the following options:
20228 Set number of input streams. Default is 2.
20231 Specify layout of inputs.
20232 This option requires the desired layout configuration to be explicitly set by the user.
20233 This sets position of each video input in output. Each input
20234 is separated by '|'.
20235 The first number represents the column, and the second number represents the row.
20236 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
20237 where X is video input from which to take width or height.
20238 Multiple values can be used when separated by '+'. In such
20239 case values are summed together.
20241 Note that if inputs are of different sizes gaps may appear, as not all of
20242 the output video frame will be filled. Similarly, videos can overlap each
20243 other if their position doesn't leave enough space for the full frame of
20246 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
20247 a layout must be set by the user.
20250 If set to 1, force the output to terminate when the shortest input
20251 terminates. Default value is 0.
20254 If set to valid color, all unused pixels will be filled with that color.
20255 By default fill is set to none, so it is disabled.
20258 @subsection Examples
20262 Display 4 inputs into 2x2 grid.
20266 input1(0, 0) | input3(w0, 0)
20267 input2(0, h0) | input4(w0, h0)
20271 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
20274 Note that if inputs are of different sizes, gaps or overlaps may occur.
20277 Display 4 inputs into 1x4 grid.
20284 input4(0, h0+h1+h2)
20288 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
20291 Note that if inputs are of different widths, unused space will appear.
20294 Display 9 inputs into 3x3 grid.
20298 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
20299 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
20300 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
20304 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
20307 Note that if inputs are of different sizes, gaps or overlaps may occur.
20310 Display 16 inputs into 4x4 grid.
20314 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
20315 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
20316 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
20317 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
20321 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|
20322 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
20325 Note that if inputs are of different sizes, gaps or overlaps may occur.
20332 Deinterlace the input video ("yadif" means "yet another deinterlacing
20335 It accepts the following parameters:
20341 The interlacing mode to adopt. It accepts one of the following values:
20344 @item 0, send_frame
20345 Output one frame for each frame.
20346 @item 1, send_field
20347 Output one frame for each field.
20348 @item 2, send_frame_nospatial
20349 Like @code{send_frame}, but it skips the spatial interlacing check.
20350 @item 3, send_field_nospatial
20351 Like @code{send_field}, but it skips the spatial interlacing check.
20354 The default value is @code{send_frame}.
20357 The picture field parity assumed for the input interlaced video. It accepts one
20358 of the following values:
20362 Assume the top field is first.
20364 Assume the bottom field is first.
20366 Enable automatic detection of field parity.
20369 The default value is @code{auto}.
20370 If the interlacing is unknown or the decoder does not export this information,
20371 top field first will be assumed.
20374 Specify which frames to deinterlace. Accepts one of the following
20379 Deinterlace all frames.
20380 @item 1, interlaced
20381 Only deinterlace frames marked as interlaced.
20384 The default value is @code{all}.
20387 @section yadif_cuda
20389 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
20390 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
20393 It accepts the following parameters:
20399 The interlacing mode to adopt. It accepts one of the following values:
20402 @item 0, send_frame
20403 Output one frame for each frame.
20404 @item 1, send_field
20405 Output one frame for each field.
20406 @item 2, send_frame_nospatial
20407 Like @code{send_frame}, but it skips the spatial interlacing check.
20408 @item 3, send_field_nospatial
20409 Like @code{send_field}, but it skips the spatial interlacing check.
20412 The default value is @code{send_frame}.
20415 The picture field parity assumed for the input interlaced video. It accepts one
20416 of the following values:
20420 Assume the top field is first.
20422 Assume the bottom field is first.
20424 Enable automatic detection of field parity.
20427 The default value is @code{auto}.
20428 If the interlacing is unknown or the decoder does not export this information,
20429 top field first will be assumed.
20432 Specify which frames to deinterlace. Accepts one of the following
20437 Deinterlace all frames.
20438 @item 1, interlaced
20439 Only deinterlace frames marked as interlaced.
20442 The default value is @code{all}.
20447 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
20448 The algorithm is described in
20449 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
20451 It accepts the following parameters:
20455 Set the window radius. Default value is 3.
20458 Set which planes to filter. Default is only the first plane.
20461 Set blur strength. Default value is 128.
20464 @subsection Commands
20465 This filter supports same @ref{commands} as options.
20469 Apply Zoom & Pan effect.
20471 This filter accepts the following options:
20475 Set the zoom expression. Range is 1-10. Default is 1.
20479 Set the x and y expression. Default is 0.
20482 Set the duration expression in number of frames.
20483 This sets for how many number of frames effect will last for
20484 single input image.
20487 Set the output image size, default is 'hd720'.
20490 Set the output frame rate, default is '25'.
20493 Each expression can contain the following constants:
20512 Output frame count.
20516 Last calculated 'x' and 'y' position from 'x' and 'y' expression
20517 for current input frame.
20521 'x' and 'y' of last output frame of previous input frame or 0 when there was
20522 not yet such frame (first input frame).
20525 Last calculated zoom from 'z' expression for current input frame.
20528 Last calculated zoom of last output frame of previous input frame.
20531 Number of output frames for current input frame. Calculated from 'd' expression
20532 for each input frame.
20535 number of output frames created for previous input frame
20538 Rational number: input width / input height
20541 sample aspect ratio
20544 display aspect ratio
20548 @subsection Examples
20552 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
20554 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
20558 Zoom-in up to 1.5 and pan always at center of picture:
20560 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
20564 Same as above but without pausing:
20566 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
20572 Scale (resize) the input video, using the z.lib library:
20573 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
20574 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
20576 The zscale filter forces the output display aspect ratio to be the same
20577 as the input, by changing the output sample aspect ratio.
20579 If the input image format is different from the format requested by
20580 the next filter, the zscale filter will convert the input to the
20583 @subsection Options
20584 The filter accepts the following options.
20589 Set the output video dimension expression. Default value is the input
20592 If the @var{width} or @var{w} value is 0, the input width is used for
20593 the output. If the @var{height} or @var{h} value is 0, the input height
20594 is used for the output.
20596 If one and only one of the values is -n with n >= 1, the zscale filter
20597 will use a value that maintains the aspect ratio of the input image,
20598 calculated from the other specified dimension. After that it will,
20599 however, make sure that the calculated dimension is divisible by n and
20600 adjust the value if necessary.
20602 If both values are -n with n >= 1, the behavior will be identical to
20603 both values being set to 0 as previously detailed.
20605 See below for the list of accepted constants for use in the dimension
20609 Set the video size. For the syntax of this option, check the
20610 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20613 Set the dither type.
20615 Possible values are:
20620 @item error_diffusion
20626 Set the resize filter type.
20628 Possible values are:
20638 Default is bilinear.
20641 Set the color range.
20643 Possible values are:
20650 Default is same as input.
20653 Set the color primaries.
20655 Possible values are:
20665 Default is same as input.
20668 Set the transfer characteristics.
20670 Possible values are:
20684 Default is same as input.
20687 Set the colorspace matrix.
20689 Possible value are:
20700 Default is same as input.
20703 Set the input color range.
20705 Possible values are:
20712 Default is same as input.
20714 @item primariesin, pin
20715 Set the input color primaries.
20717 Possible values are:
20727 Default is same as input.
20729 @item transferin, tin
20730 Set the input transfer characteristics.
20732 Possible values are:
20743 Default is same as input.
20745 @item matrixin, min
20746 Set the input colorspace matrix.
20748 Possible value are:
20760 Set the output chroma location.
20762 Possible values are:
20773 @item chromalin, cin
20774 Set the input chroma location.
20776 Possible values are:
20788 Set the nominal peak luminance.
20791 The values of the @option{w} and @option{h} options are expressions
20792 containing the following constants:
20797 The input width and height
20801 These are the same as @var{in_w} and @var{in_h}.
20805 The output (scaled) width and height
20809 These are the same as @var{out_w} and @var{out_h}
20812 The same as @var{iw} / @var{ih}
20815 input sample aspect ratio
20818 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
20822 horizontal and vertical input chroma subsample values. For example for the
20823 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
20827 horizontal and vertical output chroma subsample values. For example for the
20828 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
20831 @subsection Commands
20833 This filter supports the following commands:
20837 Set the output video dimension expression.
20838 The command accepts the same syntax of the corresponding option.
20840 If the specified expression is not valid, it is kept at its current
20844 @c man end VIDEO FILTERS
20846 @chapter OpenCL Video Filters
20847 @c man begin OPENCL VIDEO FILTERS
20849 Below is a description of the currently available OpenCL video filters.
20851 To enable compilation of these filters you need to configure FFmpeg with
20852 @code{--enable-opencl}.
20854 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
20857 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
20858 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
20859 given device parameters.
20861 @item -filter_hw_device @var{name}
20862 Pass the hardware device called @var{name} to all filters in any filter graph.
20866 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
20870 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
20872 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
20876 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.
20878 @section avgblur_opencl
20880 Apply average blur filter.
20882 The filter accepts the following options:
20886 Set horizontal radius size.
20887 Range is @code{[1, 1024]} and default value is @code{1}.
20890 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20893 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
20896 @subsection Example
20900 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.
20902 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
20906 @section boxblur_opencl
20908 Apply a boxblur algorithm to the input video.
20910 It accepts the following parameters:
20914 @item luma_radius, lr
20915 @item luma_power, lp
20916 @item chroma_radius, cr
20917 @item chroma_power, cp
20918 @item alpha_radius, ar
20919 @item alpha_power, ap
20923 A description of the accepted options follows.
20926 @item luma_radius, lr
20927 @item chroma_radius, cr
20928 @item alpha_radius, ar
20929 Set an expression for the box radius in pixels used for blurring the
20930 corresponding input plane.
20932 The radius value must be a non-negative number, and must not be
20933 greater than the value of the expression @code{min(w,h)/2} for the
20934 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
20937 Default value for @option{luma_radius} is "2". If not specified,
20938 @option{chroma_radius} and @option{alpha_radius} default to the
20939 corresponding value set for @option{luma_radius}.
20941 The expressions can contain the following constants:
20945 The input width and height in pixels.
20949 The input chroma image width and height in pixels.
20953 The horizontal and vertical chroma subsample values. For example, for the
20954 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
20957 @item luma_power, lp
20958 @item chroma_power, cp
20959 @item alpha_power, ap
20960 Specify how many times the boxblur filter is applied to the
20961 corresponding plane.
20963 Default value for @option{luma_power} is 2. If not specified,
20964 @option{chroma_power} and @option{alpha_power} default to the
20965 corresponding value set for @option{luma_power}.
20967 A value of 0 will disable the effect.
20970 @subsection Examples
20972 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.
20976 Apply a boxblur filter with the luma, chroma, and alpha radius
20977 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.
20979 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
20980 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
20984 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.
20986 For the luma plane, a 2x2 box radius will be run once.
20988 For the chroma plane, a 4x4 box radius will be run 5 times.
20990 For the alpha plane, a 3x3 box radius will be run 7 times.
20992 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
20996 @section colorkey_opencl
20997 RGB colorspace color keying.
20999 The filter accepts the following options:
21003 The color which will be replaced with transparency.
21006 Similarity percentage with the key color.
21008 0.01 matches only the exact key color, while 1.0 matches everything.
21013 0.0 makes pixels either fully transparent, or not transparent at all.
21015 Higher values result in semi-transparent pixels, with a higher transparency
21016 the more similar the pixels color is to the key color.
21019 @subsection Examples
21023 Make every semi-green pixel in the input transparent with some slight blending:
21025 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
21029 @section convolution_opencl
21031 Apply convolution of 3x3, 5x5, 7x7 matrix.
21033 The filter accepts the following options:
21040 Set matrix for each plane.
21041 Matrix is sequence of 9, 25 or 49 signed numbers.
21042 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
21048 Set multiplier for calculated value for each plane.
21049 If unset or 0, it will be sum of all matrix elements.
21050 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
21056 Set bias for each plane. This value is added to the result of the multiplication.
21057 Useful for making the overall image brighter or darker.
21058 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
21062 @subsection Examples
21068 -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
21074 -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
21078 Apply edge enhance:
21080 -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
21086 -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
21090 Apply laplacian edge detector which includes diagonals:
21092 -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
21098 -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
21102 @section erosion_opencl
21104 Apply erosion effect to the video.
21106 This filter replaces the pixel by the local(3x3) minimum.
21108 It accepts the following options:
21115 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21116 If @code{0}, plane will remain unchanged.
21119 Flag which specifies the pixel to refer to.
21120 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21122 Flags to local 3x3 coordinates region centered on @code{x}:
21131 @subsection Example
21135 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.
21137 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21141 @section deshake_opencl
21142 Feature-point based video stabilization filter.
21144 The filter accepts the following options:
21148 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
21151 Whether or not additional debug info should be displayed, both in the processed output and in the console.
21153 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
21155 Viewing point matches in the output video is only supported for RGB input.
21157 Defaults to @code{0}.
21159 @item adaptive_crop
21160 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
21162 Defaults to @code{1}.
21164 @item refine_features
21165 Whether or not feature points should be refined at a sub-pixel level.
21167 This can be turned off for a slight performance gain at the cost of precision.
21169 Defaults to @code{1}.
21171 @item smooth_strength
21172 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
21174 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
21176 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
21178 Defaults to @code{0.0}.
21180 @item smooth_window_multiplier
21181 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
21183 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
21185 Acceptable values range from @code{0.1} to @code{10.0}.
21187 Larger values increase the amount of motion data available for determining how to smooth the camera path,
21188 potentially improving smoothness, but also increase latency and memory usage.
21190 Defaults to @code{2.0}.
21194 @subsection Examples
21198 Stabilize a video with a fixed, medium smoothing strength:
21200 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
21204 Stabilize a video with debugging (both in console and in rendered video):
21206 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
21210 @section dilation_opencl
21212 Apply dilation effect to the video.
21214 This filter replaces the pixel by the local(3x3) maximum.
21216 It accepts the following options:
21223 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21224 If @code{0}, plane will remain unchanged.
21227 Flag which specifies the pixel to refer to.
21228 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21230 Flags to local 3x3 coordinates region centered on @code{x}:
21239 @subsection Example
21243 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.
21245 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21249 @section nlmeans_opencl
21251 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
21253 @section overlay_opencl
21255 Overlay one video on top of another.
21257 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
21258 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
21260 The filter accepts the following options:
21265 Set the x coordinate of the overlaid video on the main video.
21266 Default value is @code{0}.
21269 Set the y coordinate of the overlaid video on the main video.
21270 Default value is @code{0}.
21274 @subsection Examples
21278 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
21280 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21283 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
21285 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21290 @section pad_opencl
21292 Add paddings to the input image, and place the original input at the
21293 provided @var{x}, @var{y} coordinates.
21295 It accepts the following options:
21300 Specify an expression for the size of the output image with the
21301 paddings added. If the value for @var{width} or @var{height} is 0, the
21302 corresponding input size is used for the output.
21304 The @var{width} expression can reference the value set by the
21305 @var{height} expression, and vice versa.
21307 The default value of @var{width} and @var{height} is 0.
21311 Specify the offsets to place the input image at within the padded area,
21312 with respect to the top/left border of the output image.
21314 The @var{x} expression can reference the value set by the @var{y}
21315 expression, and vice versa.
21317 The default value of @var{x} and @var{y} is 0.
21319 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
21320 so the input image is centered on the padded area.
21323 Specify the color of the padded area. For the syntax of this option,
21324 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
21325 manual,ffmpeg-utils}.
21328 Pad to an aspect instead to a resolution.
21331 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
21332 options are expressions containing the following constants:
21337 The input video width and height.
21341 These are the same as @var{in_w} and @var{in_h}.
21345 The output width and height (the size of the padded area), as
21346 specified by the @var{width} and @var{height} expressions.
21350 These are the same as @var{out_w} and @var{out_h}.
21354 The x and y offsets as specified by the @var{x} and @var{y}
21355 expressions, or NAN if not yet specified.
21358 same as @var{iw} / @var{ih}
21361 input sample aspect ratio
21364 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
21367 @section prewitt_opencl
21369 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
21371 The filter accepts the following option:
21375 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21378 Set value which will be multiplied with filtered result.
21379 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21382 Set value which will be added to filtered result.
21383 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21386 @subsection Example
21390 Apply the Prewitt operator with scale set to 2 and delta set to 10.
21392 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
21396 @anchor{program_opencl}
21397 @section program_opencl
21399 Filter video using an OpenCL program.
21404 OpenCL program source file.
21407 Kernel name in program.
21410 Number of inputs to the filter. Defaults to 1.
21413 Size of output frames. Defaults to the same as the first input.
21417 The @code{program_opencl} filter also supports the @ref{framesync} options.
21419 The program source file must contain a kernel function with the given name,
21420 which will be run once for each plane of the output. Each run on a plane
21421 gets enqueued as a separate 2D global NDRange with one work-item for each
21422 pixel to be generated. The global ID offset for each work-item is therefore
21423 the coordinates of a pixel in the destination image.
21425 The kernel function needs to take the following arguments:
21428 Destination image, @var{__write_only image2d_t}.
21430 This image will become the output; the kernel should write all of it.
21432 Frame index, @var{unsigned int}.
21434 This is a counter starting from zero and increasing by one for each frame.
21436 Source images, @var{__read_only image2d_t}.
21438 These are the most recent images on each input. The kernel may read from
21439 them to generate the output, but they can't be written to.
21446 Copy the input to the output (output must be the same size as the input).
21448 __kernel void copy(__write_only image2d_t destination,
21449 unsigned int index,
21450 __read_only image2d_t source)
21452 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
21454 int2 location = (int2)(get_global_id(0), get_global_id(1));
21456 float4 value = read_imagef(source, sampler, location);
21458 write_imagef(destination, location, value);
21463 Apply a simple transformation, rotating the input by an amount increasing
21464 with the index counter. Pixel values are linearly interpolated by the
21465 sampler, and the output need not have the same dimensions as the input.
21467 __kernel void rotate_image(__write_only image2d_t dst,
21468 unsigned int index,
21469 __read_only image2d_t src)
21471 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
21472 CLK_FILTER_LINEAR);
21474 float angle = (float)index / 100.0f;
21476 float2 dst_dim = convert_float2(get_image_dim(dst));
21477 float2 src_dim = convert_float2(get_image_dim(src));
21479 float2 dst_cen = dst_dim / 2.0f;
21480 float2 src_cen = src_dim / 2.0f;
21482 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
21484 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
21486 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
21487 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
21489 src_pos = src_pos * src_dim / dst_dim;
21491 float2 src_loc = src_pos + src_cen;
21493 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
21494 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
21495 write_imagef(dst, dst_loc, 0.5f);
21497 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
21502 Blend two inputs together, with the amount of each input used varying
21503 with the index counter.
21505 __kernel void blend_images(__write_only image2d_t dst,
21506 unsigned int index,
21507 __read_only image2d_t src1,
21508 __read_only image2d_t src2)
21510 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
21511 CLK_FILTER_LINEAR);
21513 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
21515 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
21516 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
21517 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
21519 float4 val1 = read_imagef(src1, sampler, src1_loc);
21520 float4 val2 = read_imagef(src2, sampler, src2_loc);
21522 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
21528 @section roberts_opencl
21529 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
21531 The filter accepts the following option:
21535 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21538 Set value which will be multiplied with filtered result.
21539 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21542 Set value which will be added to filtered result.
21543 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21546 @subsection Example
21550 Apply the Roberts cross operator with scale set to 2 and delta set to 10
21552 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
21556 @section sobel_opencl
21558 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
21560 The filter accepts the following option:
21564 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21567 Set value which will be multiplied with filtered result.
21568 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21571 Set value which will be added to filtered result.
21572 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21575 @subsection Example
21579 Apply sobel operator with scale set to 2 and delta set to 10
21581 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
21585 @section tonemap_opencl
21587 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
21589 It accepts the following parameters:
21593 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
21596 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
21599 Apply desaturation for highlights that exceed this level of brightness. The
21600 higher the parameter, the more color information will be preserved. This
21601 setting helps prevent unnaturally blown-out colors for super-highlights, by
21602 (smoothly) turning into white instead. This makes images feel more natural,
21603 at the cost of reducing information about out-of-range colors.
21605 The default value is 0.5, and the algorithm here is a little different from
21606 the cpu version tonemap currently. A setting of 0.0 disables this option.
21609 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
21610 is used to detect whether the scene has changed or not. If the distance between
21611 the current frame average brightness and the current running average exceeds
21612 a threshold value, we would re-calculate scene average and peak brightness.
21613 The default value is 0.2.
21616 Specify the output pixel format.
21618 Currently supported formats are:
21625 Set the output color range.
21627 Possible values are:
21633 Default is same as input.
21636 Set the output color primaries.
21638 Possible values are:
21644 Default is same as input.
21647 Set the output transfer characteristics.
21649 Possible values are:
21658 Set the output colorspace matrix.
21660 Possible value are:
21666 Default is same as input.
21670 @subsection Example
21674 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
21676 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
21680 @section unsharp_opencl
21682 Sharpen or blur the input video.
21684 It accepts the following parameters:
21687 @item luma_msize_x, lx
21688 Set the luma matrix horizontal size.
21689 Range is @code{[1, 23]} and default value is @code{5}.
21691 @item luma_msize_y, ly
21692 Set the luma matrix vertical size.
21693 Range is @code{[1, 23]} and default value is @code{5}.
21695 @item luma_amount, la
21696 Set the luma effect strength.
21697 Range is @code{[-10, 10]} and default value is @code{1.0}.
21699 Negative values will blur the input video, while positive values will
21700 sharpen it, a value of zero will disable the effect.
21702 @item chroma_msize_x, cx
21703 Set the chroma matrix horizontal size.
21704 Range is @code{[1, 23]} and default value is @code{5}.
21706 @item chroma_msize_y, cy
21707 Set the chroma matrix vertical size.
21708 Range is @code{[1, 23]} and default value is @code{5}.
21710 @item chroma_amount, ca
21711 Set the chroma effect strength.
21712 Range is @code{[-10, 10]} and default value is @code{0.0}.
21714 Negative values will blur the input video, while positive values will
21715 sharpen it, a value of zero will disable the effect.
21719 All parameters are optional and default to the equivalent of the
21720 string '5:5:1.0:5:5:0.0'.
21722 @subsection Examples
21726 Apply strong luma sharpen effect:
21728 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
21732 Apply a strong blur of both luma and chroma parameters:
21734 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
21738 @section xfade_opencl
21740 Cross fade two videos with custom transition effect by using OpenCL.
21742 It accepts the following options:
21746 Set one of possible transition effects.
21750 Select custom transition effect, the actual transition description
21751 will be picked from source and kernel options.
21763 Default transition is fade.
21767 OpenCL program source file for custom transition.
21770 Set name of kernel to use for custom transition from program source file.
21773 Set duration of video transition.
21776 Set time of start of transition relative to first video.
21779 The program source file must contain a kernel function with the given name,
21780 which will be run once for each plane of the output. Each run on a plane
21781 gets enqueued as a separate 2D global NDRange with one work-item for each
21782 pixel to be generated. The global ID offset for each work-item is therefore
21783 the coordinates of a pixel in the destination image.
21785 The kernel function needs to take the following arguments:
21788 Destination image, @var{__write_only image2d_t}.
21790 This image will become the output; the kernel should write all of it.
21793 First Source image, @var{__read_only image2d_t}.
21794 Second Source image, @var{__read_only image2d_t}.
21796 These are the most recent images on each input. The kernel may read from
21797 them to generate the output, but they can't be written to.
21800 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
21807 Apply dots curtain transition effect:
21809 __kernel void blend_images(__write_only image2d_t dst,
21810 __read_only image2d_t src1,
21811 __read_only image2d_t src2,
21814 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
21815 CLK_FILTER_LINEAR);
21816 int2 p = (int2)(get_global_id(0), get_global_id(1));
21817 float2 rp = (float2)(get_global_id(0), get_global_id(1));
21818 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
21821 float2 dots = (float2)(20.0, 20.0);
21822 float2 center = (float2)(0,0);
21825 float4 val1 = read_imagef(src1, sampler, p);
21826 float4 val2 = read_imagef(src2, sampler, p);
21827 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
21829 write_imagef(dst, p, next ? val1 : val2);
21835 @c man end OPENCL VIDEO FILTERS
21837 @chapter VAAPI Video Filters
21838 @c man begin VAAPI VIDEO FILTERS
21840 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
21842 To enable compilation of these filters you need to configure FFmpeg with
21843 @code{--enable-vaapi}.
21845 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}
21847 @section tonemap_vaapi
21849 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
21850 It maps the dynamic range of HDR10 content to the SDR content.
21851 It currently only accepts HDR10 as input.
21853 It accepts the following parameters:
21857 Specify the output pixel format.
21859 Currently supported formats are:
21868 Set the output color primaries.
21870 Default is same as input.
21873 Set the output transfer characteristics.
21878 Set the output colorspace matrix.
21880 Default is same as input.
21884 @subsection Example
21888 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
21890 tonemap_vaapi=format=p010:t=bt2020-10
21894 @c man end VAAPI VIDEO FILTERS
21896 @chapter Video Sources
21897 @c man begin VIDEO SOURCES
21899 Below is a description of the currently available video sources.
21903 Buffer video frames, and make them available to the filter chain.
21905 This source is mainly intended for a programmatic use, in particular
21906 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
21908 It accepts the following parameters:
21913 Specify the size (width and height) of the buffered video frames. For the
21914 syntax of this option, check the
21915 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21918 The input video width.
21921 The input video height.
21924 A string representing the pixel format of the buffered video frames.
21925 It may be a number corresponding to a pixel format, or a pixel format
21929 Specify the timebase assumed by the timestamps of the buffered frames.
21932 Specify the frame rate expected for the video stream.
21934 @item pixel_aspect, sar
21935 The sample (pixel) aspect ratio of the input video.
21938 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
21939 to the filtergraph description to specify swscale flags for automatically
21940 inserted scalers. See @ref{Filtergraph syntax}.
21942 @item hw_frames_ctx
21943 When using a hardware pixel format, this should be a reference to an
21944 AVHWFramesContext describing input frames.
21949 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
21952 will instruct the source to accept video frames with size 320x240 and
21953 with format "yuv410p", assuming 1/24 as the timestamps timebase and
21954 square pixels (1:1 sample aspect ratio).
21955 Since the pixel format with name "yuv410p" corresponds to the number 6
21956 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
21957 this example corresponds to:
21959 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
21962 Alternatively, the options can be specified as a flat string, but this
21963 syntax is deprecated:
21965 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
21969 Create a pattern generated by an elementary cellular automaton.
21971 The initial state of the cellular automaton can be defined through the
21972 @option{filename} and @option{pattern} options. If such options are
21973 not specified an initial state is created randomly.
21975 At each new frame a new row in the video is filled with the result of
21976 the cellular automaton next generation. The behavior when the whole
21977 frame is filled is defined by the @option{scroll} option.
21979 This source accepts the following options:
21983 Read the initial cellular automaton state, i.e. the starting row, from
21984 the specified file.
21985 In the file, each non-whitespace character is considered an alive
21986 cell, a newline will terminate the row, and further characters in the
21987 file will be ignored.
21990 Read the initial cellular automaton state, i.e. the starting row, from
21991 the specified string.
21993 Each non-whitespace character in the string is considered an alive
21994 cell, a newline will terminate the row, and further characters in the
21995 string will be ignored.
21998 Set the video rate, that is the number of frames generated per second.
22001 @item random_fill_ratio, ratio
22002 Set the random fill ratio for the initial cellular automaton row. It
22003 is a floating point number value ranging from 0 to 1, defaults to
22006 This option is ignored when a file or a pattern is specified.
22008 @item random_seed, seed
22009 Set the seed for filling randomly the initial row, must be an integer
22010 included between 0 and UINT32_MAX. If not specified, or if explicitly
22011 set to -1, the filter will try to use a good random seed on a best
22015 Set the cellular automaton rule, it is a number ranging from 0 to 255.
22016 Default value is 110.
22019 Set the size of the output video. For the syntax of this option, check the
22020 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22022 If @option{filename} or @option{pattern} is specified, the size is set
22023 by default to the width of the specified initial state row, and the
22024 height is set to @var{width} * PHI.
22026 If @option{size} is set, it must contain the width of the specified
22027 pattern string, and the specified pattern will be centered in the
22030 If a filename or a pattern string is not specified, the size value
22031 defaults to "320x518" (used for a randomly generated initial state).
22034 If set to 1, scroll the output upward when all the rows in the output
22035 have been already filled. If set to 0, the new generated row will be
22036 written over the top row just after the bottom row is filled.
22039 @item start_full, full
22040 If set to 1, completely fill the output with generated rows before
22041 outputting the first frame.
22042 This is the default behavior, for disabling set the value to 0.
22045 If set to 1, stitch the left and right row edges together.
22046 This is the default behavior, for disabling set the value to 0.
22049 @subsection Examples
22053 Read the initial state from @file{pattern}, and specify an output of
22056 cellauto=f=pattern:s=200x400
22060 Generate a random initial row with a width of 200 cells, with a fill
22063 cellauto=ratio=2/3:s=200x200
22067 Create a pattern generated by rule 18 starting by a single alive cell
22068 centered on an initial row with width 100:
22070 cellauto=p=@@:s=100x400:full=0:rule=18
22074 Specify a more elaborated initial pattern:
22076 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
22081 @anchor{coreimagesrc}
22082 @section coreimagesrc
22083 Video source generated on GPU using Apple's CoreImage API on OSX.
22085 This video source is a specialized version of the @ref{coreimage} video filter.
22086 Use a core image generator at the beginning of the applied filterchain to
22087 generate the content.
22089 The coreimagesrc video source accepts the following options:
22091 @item list_generators
22092 List all available generators along with all their respective options as well as
22093 possible minimum and maximum values along with the default values.
22095 list_generators=true
22099 Specify the size of the sourced video. For the syntax of this option, check the
22100 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22101 The default value is @code{320x240}.
22104 Specify the frame rate of the sourced video, as the number of frames
22105 generated per second. It has to be a string in the format
22106 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22107 number or a valid video frame rate abbreviation. The default value is
22111 Set the sample aspect ratio of the sourced video.
22114 Set the duration of the sourced video. See
22115 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22116 for the accepted syntax.
22118 If not specified, or the expressed duration is negative, the video is
22119 supposed to be generated forever.
22122 Additionally, all options of the @ref{coreimage} video filter are accepted.
22123 A complete filterchain can be used for further processing of the
22124 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
22125 and examples for details.
22127 @subsection Examples
22132 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
22133 given as complete and escaped command-line for Apple's standard bash shell:
22135 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
22137 This example is equivalent to the QRCode example of @ref{coreimage} without the
22138 need for a nullsrc video source.
22142 @section mandelbrot
22144 Generate a Mandelbrot set fractal, and progressively zoom towards the
22145 point specified with @var{start_x} and @var{start_y}.
22147 This source accepts the following options:
22152 Set the terminal pts value. Default value is 400.
22155 Set the terminal scale value.
22156 Must be a floating point value. Default value is 0.3.
22159 Set the inner coloring mode, that is the algorithm used to draw the
22160 Mandelbrot fractal internal region.
22162 It shall assume one of the following values:
22167 Show time until convergence.
22169 Set color based on point closest to the origin of the iterations.
22174 Default value is @var{mincol}.
22177 Set the bailout value. Default value is 10.0.
22180 Set the maximum of iterations performed by the rendering
22181 algorithm. Default value is 7189.
22184 Set outer coloring mode.
22185 It shall assume one of following values:
22187 @item iteration_count
22188 Set iteration count mode.
22189 @item normalized_iteration_count
22190 set normalized iteration count mode.
22192 Default value is @var{normalized_iteration_count}.
22195 Set frame rate, expressed as number of frames per second. Default
22199 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22200 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22203 Set the initial scale value. Default value is 3.0.
22206 Set the initial x position. Must be a floating point value between
22207 -100 and 100. Default value is -0.743643887037158704752191506114774.
22210 Set the initial y position. Must be a floating point value between
22211 -100 and 100. Default value is -0.131825904205311970493132056385139.
22216 Generate various test patterns, as generated by the MPlayer test filter.
22218 The size of the generated video is fixed, and is 256x256.
22219 This source is useful in particular for testing encoding features.
22221 This source accepts the following options:
22226 Specify the frame rate of the sourced video, as the number of frames
22227 generated per second. It has to be a string in the format
22228 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22229 number or a valid video frame rate abbreviation. The default value is
22233 Set the duration of the sourced video. See
22234 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22235 for the accepted syntax.
22237 If not specified, or the expressed duration is negative, the video is
22238 supposed to be generated forever.
22242 Set the number or the name of the test to perform. Supported tests are:
22256 @item max_frames, m
22257 Set the maximum number of frames generated for each test, default value is 30.
22261 Default value is "all", which will cycle through the list of all tests.
22266 mptestsrc=t=dc_luma
22269 will generate a "dc_luma" test pattern.
22271 @section frei0r_src
22273 Provide a frei0r source.
22275 To enable compilation of this filter you need to install the frei0r
22276 header and configure FFmpeg with @code{--enable-frei0r}.
22278 This source accepts the following parameters:
22283 The size of the video to generate. For the syntax of this option, check the
22284 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22287 The framerate of the generated video. It may be a string of the form
22288 @var{num}/@var{den} or a frame rate abbreviation.
22291 The name to the frei0r source to load. For more information regarding frei0r and
22292 how to set the parameters, read the @ref{frei0r} section in the video filters
22295 @item filter_params
22296 A '|'-separated list of parameters to pass to the frei0r source.
22300 For example, to generate a frei0r partik0l source with size 200x200
22301 and frame rate 10 which is overlaid on the overlay filter main input:
22303 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
22308 Generate a life pattern.
22310 This source is based on a generalization of John Conway's life game.
22312 The sourced input represents a life grid, each pixel represents a cell
22313 which can be in one of two possible states, alive or dead. Every cell
22314 interacts with its eight neighbours, which are the cells that are
22315 horizontally, vertically, or diagonally adjacent.
22317 At each interaction the grid evolves according to the adopted rule,
22318 which specifies the number of neighbor alive cells which will make a
22319 cell stay alive or born. The @option{rule} option allows one to specify
22322 This source accepts the following options:
22326 Set the file from which to read the initial grid state. In the file,
22327 each non-whitespace character is considered an alive cell, and newline
22328 is used to delimit the end of each row.
22330 If this option is not specified, the initial grid is generated
22334 Set the video rate, that is the number of frames generated per second.
22337 @item random_fill_ratio, ratio
22338 Set the random fill ratio for the initial random grid. It is a
22339 floating point number value ranging from 0 to 1, defaults to 1/PHI.
22340 It is ignored when a file is specified.
22342 @item random_seed, seed
22343 Set the seed for filling the initial random grid, must be an integer
22344 included between 0 and UINT32_MAX. If not specified, or if explicitly
22345 set to -1, the filter will try to use a good random seed on a best
22351 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
22352 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
22353 @var{NS} specifies the number of alive neighbor cells which make a
22354 live cell stay alive, and @var{NB} the number of alive neighbor cells
22355 which make a dead cell to become alive (i.e. to "born").
22356 "s" and "b" can be used in place of "S" and "B", respectively.
22358 Alternatively a rule can be specified by an 18-bits integer. The 9
22359 high order bits are used to encode the next cell state if it is alive
22360 for each number of neighbor alive cells, the low order bits specify
22361 the rule for "borning" new cells. Higher order bits encode for an
22362 higher number of neighbor cells.
22363 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
22364 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
22366 Default value is "S23/B3", which is the original Conway's game of life
22367 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
22368 cells, and will born a new cell if there are three alive cells around
22372 Set the size of the output video. For the syntax of this option, check the
22373 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22375 If @option{filename} is specified, the size is set by default to the
22376 same size of the input file. If @option{size} is set, it must contain
22377 the size specified in the input file, and the initial grid defined in
22378 that file is centered in the larger resulting area.
22380 If a filename is not specified, the size value defaults to "320x240"
22381 (used for a randomly generated initial grid).
22384 If set to 1, stitch the left and right grid edges together, and the
22385 top and bottom edges also. Defaults to 1.
22388 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
22389 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
22390 value from 0 to 255.
22393 Set the color of living (or new born) cells.
22396 Set the color of dead cells. If @option{mold} is set, this is the first color
22397 used to represent a dead cell.
22400 Set mold color, for definitely dead and moldy cells.
22402 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
22403 ffmpeg-utils manual,ffmpeg-utils}.
22406 @subsection Examples
22410 Read a grid from @file{pattern}, and center it on a grid of size
22413 life=f=pattern:s=300x300
22417 Generate a random grid of size 200x200, with a fill ratio of 2/3:
22419 life=ratio=2/3:s=200x200
22423 Specify a custom rule for evolving a randomly generated grid:
22429 Full example with slow death effect (mold) using @command{ffplay}:
22431 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
22438 @anchor{haldclutsrc}
22441 @anchor{pal100bars}
22442 @anchor{rgbtestsrc}
22444 @anchor{smptehdbars}
22447 @anchor{yuvtestsrc}
22448 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
22450 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
22452 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
22454 The @code{color} source provides an uniformly colored input.
22456 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
22457 @ref{haldclut} filter.
22459 The @code{nullsrc} source returns unprocessed video frames. It is
22460 mainly useful to be employed in analysis / debugging tools, or as the
22461 source for filters which ignore the input data.
22463 The @code{pal75bars} source generates a color bars pattern, based on
22464 EBU PAL recommendations with 75% color levels.
22466 The @code{pal100bars} source generates a color bars pattern, based on
22467 EBU PAL recommendations with 100% color levels.
22469 The @code{rgbtestsrc} source generates an RGB test pattern useful for
22470 detecting RGB vs BGR issues. You should see a red, green and blue
22471 stripe from top to bottom.
22473 The @code{smptebars} source generates a color bars pattern, based on
22474 the SMPTE Engineering Guideline EG 1-1990.
22476 The @code{smptehdbars} source generates a color bars pattern, based on
22477 the SMPTE RP 219-2002.
22479 The @code{testsrc} source generates a test video pattern, showing a
22480 color pattern, a scrolling gradient and a timestamp. This is mainly
22481 intended for testing purposes.
22483 The @code{testsrc2} source is similar to testsrc, but supports more
22484 pixel formats instead of just @code{rgb24}. This allows using it as an
22485 input for other tests without requiring a format conversion.
22487 The @code{yuvtestsrc} source generates an YUV test pattern. You should
22488 see a y, cb and cr stripe from top to bottom.
22490 The sources accept the following parameters:
22495 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
22496 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
22497 pixels to be used as identity matrix for 3D lookup tables. Each component is
22498 coded on a @code{1/(N*N)} scale.
22501 Specify the color of the source, only available in the @code{color}
22502 source. For the syntax of this option, check the
22503 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
22506 Specify the size of the sourced video. For the syntax of this option, check the
22507 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22508 The default value is @code{320x240}.
22510 This option is not available with the @code{allrgb}, @code{allyuv}, and
22511 @code{haldclutsrc} filters.
22514 Specify the frame rate of the sourced video, as the number of frames
22515 generated per second. It has to be a string in the format
22516 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22517 number or a valid video frame rate abbreviation. The default value is
22521 Set the duration of the sourced video. See
22522 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22523 for the accepted syntax.
22525 If not specified, or the expressed duration is negative, the video is
22526 supposed to be generated forever.
22529 Set the sample aspect ratio of the sourced video.
22532 Specify the alpha (opacity) of the background, only available in the
22533 @code{testsrc2} source. The value must be between 0 (fully transparent) and
22534 255 (fully opaque, the default).
22537 Set the number of decimals to show in the timestamp, only available in the
22538 @code{testsrc} source.
22540 The displayed timestamp value will correspond to the original
22541 timestamp value multiplied by the power of 10 of the specified
22542 value. Default value is 0.
22545 @subsection Examples
22549 Generate a video with a duration of 5.3 seconds, with size
22550 176x144 and a frame rate of 10 frames per second:
22552 testsrc=duration=5.3:size=qcif:rate=10
22556 The following graph description will generate a red source
22557 with an opacity of 0.2, with size "qcif" and a frame rate of 10
22560 color=c=red@@0.2:s=qcif:r=10
22564 If the input content is to be ignored, @code{nullsrc} can be used. The
22565 following command generates noise in the luminance plane by employing
22566 the @code{geq} filter:
22568 nullsrc=s=256x256, geq=random(1)*255:128:128
22572 @subsection Commands
22574 The @code{color} source supports the following commands:
22578 Set the color of the created image. Accepts the same syntax of the
22579 corresponding @option{color} option.
22584 Generate video using an OpenCL program.
22589 OpenCL program source file.
22592 Kernel name in program.
22595 Size of frames to generate. This must be set.
22598 Pixel format to use for the generated frames. This must be set.
22601 Number of frames generated every second. Default value is '25'.
22605 For details of how the program loading works, see the @ref{program_opencl}
22612 Generate a colour ramp by setting pixel values from the position of the pixel
22613 in the output image. (Note that this will work with all pixel formats, but
22614 the generated output will not be the same.)
22616 __kernel void ramp(__write_only image2d_t dst,
22617 unsigned int index)
22619 int2 loc = (int2)(get_global_id(0), get_global_id(1));
22622 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
22624 write_imagef(dst, loc, val);
22629 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
22631 __kernel void sierpinski_carpet(__write_only image2d_t dst,
22632 unsigned int index)
22634 int2 loc = (int2)(get_global_id(0), get_global_id(1));
22636 float4 value = 0.0f;
22637 int x = loc.x + index;
22638 int y = loc.y + index;
22639 while (x > 0 || y > 0) {
22640 if (x % 3 == 1 && y % 3 == 1) {
22648 write_imagef(dst, loc, value);
22654 @section sierpinski
22656 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
22658 This source accepts the following options:
22662 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22663 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22666 Set frame rate, expressed as number of frames per second. Default
22670 Set seed which is used for random panning.
22673 Set max jump for single pan destination. Allowed range is from 1 to 10000.
22676 Set fractal type, can be default @code{carpet} or @code{triangle}.
22679 @c man end VIDEO SOURCES
22681 @chapter Video Sinks
22682 @c man begin VIDEO SINKS
22684 Below is a description of the currently available video sinks.
22686 @section buffersink
22688 Buffer video frames, and make them available to the end of the filter
22691 This sink is mainly intended for programmatic use, in particular
22692 through the interface defined in @file{libavfilter/buffersink.h}
22693 or the options system.
22695 It accepts a pointer to an AVBufferSinkContext structure, which
22696 defines the incoming buffers' formats, to be passed as the opaque
22697 parameter to @code{avfilter_init_filter} for initialization.
22701 Null video sink: do absolutely nothing with the input video. It is
22702 mainly useful as a template and for use in analysis / debugging
22705 @c man end VIDEO SINKS
22707 @chapter Multimedia Filters
22708 @c man begin MULTIMEDIA FILTERS
22710 Below is a description of the currently available multimedia filters.
22714 Convert input audio to a video output, displaying the audio bit scope.
22716 The filter accepts the following options:
22720 Set frame rate, expressed as number of frames per second. Default
22724 Specify the video size for the output. For the syntax of this option, check the
22725 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22726 Default value is @code{1024x256}.
22729 Specify list of colors separated by space or by '|' which will be used to
22730 draw channels. Unrecognized or missing colors will be replaced
22734 @section adrawgraph
22735 Draw a graph using input audio metadata.
22737 See @ref{drawgraph}
22739 @section agraphmonitor
22741 See @ref{graphmonitor}.
22743 @section ahistogram
22745 Convert input audio to a video output, displaying the volume histogram.
22747 The filter accepts the following options:
22751 Specify how histogram is calculated.
22753 It accepts the following values:
22756 Use single histogram for all channels.
22758 Use separate histogram for each channel.
22760 Default is @code{single}.
22763 Set frame rate, expressed as number of frames per second. Default
22767 Specify the video size for the output. For the syntax of this option, check the
22768 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22769 Default value is @code{hd720}.
22774 It accepts the following values:
22785 reverse logarithmic
22787 Default is @code{log}.
22790 Set amplitude scale.
22792 It accepts the following values:
22799 Default is @code{log}.
22802 Set how much frames to accumulate in histogram.
22803 Default is 1. Setting this to -1 accumulates all frames.
22806 Set histogram ratio of window height.
22809 Set sonogram sliding.
22811 It accepts the following values:
22814 replace old rows with new ones.
22816 scroll from top to bottom.
22818 Default is @code{replace}.
22821 @section aphasemeter
22823 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
22824 representing mean phase of current audio frame. A video output can also be produced and is
22825 enabled by default. The audio is passed through as first output.
22827 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
22828 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
22829 and @code{1} means channels are in phase.
22831 The filter accepts the following options, all related to its video output:
22835 Set the output frame rate. Default value is @code{25}.
22838 Set the video size for the output. For the syntax of this option, check the
22839 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22840 Default value is @code{800x400}.
22845 Specify the red, green, blue contrast. Default values are @code{2},
22846 @code{7} and @code{1}.
22847 Allowed range is @code{[0, 255]}.
22850 Set color which will be used for drawing median phase. If color is
22851 @code{none} which is default, no median phase value will be drawn.
22854 Enable video output. Default is enabled.
22857 @section avectorscope
22859 Convert input audio to a video output, representing the audio vector
22862 The filter is used to measure the difference between channels of stereo
22863 audio stream. A monaural signal, consisting of identical left and right
22864 signal, results in straight vertical line. Any stereo separation is visible
22865 as a deviation from this line, creating a Lissajous figure.
22866 If the straight (or deviation from it) but horizontal line appears this
22867 indicates that the left and right channels are out of phase.
22869 The filter accepts the following options:
22873 Set the vectorscope mode.
22875 Available values are:
22878 Lissajous rotated by 45 degrees.
22881 Same as above but not rotated.
22884 Shape resembling half of circle.
22887 Default value is @samp{lissajous}.
22890 Set the video size for the output. For the syntax of this option, check the
22891 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22892 Default value is @code{400x400}.
22895 Set the output frame rate. Default value is @code{25}.
22901 Specify the red, green, blue and alpha contrast. Default values are @code{40},
22902 @code{160}, @code{80} and @code{255}.
22903 Allowed range is @code{[0, 255]}.
22909 Specify the red, green, blue and alpha fade. Default values are @code{15},
22910 @code{10}, @code{5} and @code{5}.
22911 Allowed range is @code{[0, 255]}.
22914 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
22915 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
22918 Set the vectorscope drawing mode.
22920 Available values are:
22923 Draw dot for each sample.
22926 Draw line between previous and current sample.
22929 Default value is @samp{dot}.
22932 Specify amplitude scale of audio samples.
22934 Available values are:
22950 Swap left channel axis with right channel axis.
22960 Mirror only x axis.
22963 Mirror only y axis.
22971 @subsection Examples
22975 Complete example using @command{ffplay}:
22977 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
22978 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
22982 @section bench, abench
22984 Benchmark part of a filtergraph.
22986 The filter accepts the following options:
22990 Start or stop a timer.
22992 Available values are:
22995 Get the current time, set it as frame metadata (using the key
22996 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
22999 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
23000 the input frame metadata to get the time difference. Time difference, average,
23001 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
23002 @code{min}) are then printed. The timestamps are expressed in seconds.
23006 @subsection Examples
23010 Benchmark @ref{selectivecolor} filter:
23012 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
23018 Concatenate audio and video streams, joining them together one after the
23021 The filter works on segments of synchronized video and audio streams. All
23022 segments must have the same number of streams of each type, and that will
23023 also be the number of streams at output.
23025 The filter accepts the following options:
23030 Set the number of segments. Default is 2.
23033 Set the number of output video streams, that is also the number of video
23034 streams in each segment. Default is 1.
23037 Set the number of output audio streams, that is also the number of audio
23038 streams in each segment. Default is 0.
23041 Activate unsafe mode: do not fail if segments have a different format.
23045 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
23046 @var{a} audio outputs.
23048 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
23049 segment, in the same order as the outputs, then the inputs for the second
23052 Related streams do not always have exactly the same duration, for various
23053 reasons including codec frame size or sloppy authoring. For that reason,
23054 related synchronized streams (e.g. a video and its audio track) should be
23055 concatenated at once. The concat filter will use the duration of the longest
23056 stream in each segment (except the last one), and if necessary pad shorter
23057 audio streams with silence.
23059 For this filter to work correctly, all segments must start at timestamp 0.
23061 All corresponding streams must have the same parameters in all segments; the
23062 filtering system will automatically select a common pixel format for video
23063 streams, and a common sample format, sample rate and channel layout for
23064 audio streams, but other settings, such as resolution, must be converted
23065 explicitly by the user.
23067 Different frame rates are acceptable but will result in variable frame rate
23068 at output; be sure to configure the output file to handle it.
23070 @subsection Examples
23074 Concatenate an opening, an episode and an ending, all in bilingual version
23075 (video in stream 0, audio in streams 1 and 2):
23077 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
23078 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
23079 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
23080 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
23084 Concatenate two parts, handling audio and video separately, using the
23085 (a)movie sources, and adjusting the resolution:
23087 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
23088 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
23089 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
23091 Note that a desync will happen at the stitch if the audio and video streams
23092 do not have exactly the same duration in the first file.
23096 @subsection Commands
23098 This filter supports the following commands:
23101 Close the current segment and step to the next one
23107 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
23108 level. By default, it logs a message at a frequency of 10Hz with the
23109 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
23110 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
23112 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
23113 sample format is double-precision floating point. The input stream will be converted to
23114 this specification, if needed. Users may need to insert aformat and/or aresample filters
23115 after this filter to obtain the original parameters.
23117 The filter also has a video output (see the @var{video} option) with a real
23118 time graph to observe the loudness evolution. The graphic contains the logged
23119 message mentioned above, so it is not printed anymore when this option is set,
23120 unless the verbose logging is set. The main graphing area contains the
23121 short-term loudness (3 seconds of analysis), and the gauge on the right is for
23122 the momentary loudness (400 milliseconds), but can optionally be configured
23123 to instead display short-term loudness (see @var{gauge}).
23125 The green area marks a +/- 1LU target range around the target loudness
23126 (-23LUFS by default, unless modified through @var{target}).
23128 More information about the Loudness Recommendation EBU R128 on
23129 @url{http://tech.ebu.ch/loudness}.
23131 The filter accepts the following options:
23136 Activate the video output. The audio stream is passed unchanged whether this
23137 option is set or no. The video stream will be the first output stream if
23138 activated. Default is @code{0}.
23141 Set the video size. This option is for video only. For the syntax of this
23143 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23144 Default and minimum resolution is @code{640x480}.
23147 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
23148 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
23149 other integer value between this range is allowed.
23152 Set metadata injection. If set to @code{1}, the audio input will be segmented
23153 into 100ms output frames, each of them containing various loudness information
23154 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
23156 Default is @code{0}.
23159 Force the frame logging level.
23161 Available values are:
23164 information logging level
23166 verbose logging level
23169 By default, the logging level is set to @var{info}. If the @option{video} or
23170 the @option{metadata} options are set, it switches to @var{verbose}.
23175 Available modes can be cumulated (the option is a @code{flag} type). Possible
23179 Disable any peak mode (default).
23181 Enable sample-peak mode.
23183 Simple peak mode looking for the higher sample value. It logs a message
23184 for sample-peak (identified by @code{SPK}).
23186 Enable true-peak mode.
23188 If enabled, the peak lookup is done on an over-sampled version of the input
23189 stream for better peak accuracy. It logs a message for true-peak.
23190 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
23191 This mode requires a build with @code{libswresample}.
23195 Treat mono input files as "dual mono". If a mono file is intended for playback
23196 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
23197 If set to @code{true}, this option will compensate for this effect.
23198 Multi-channel input files are not affected by this option.
23201 Set a specific pan law to be used for the measurement of dual mono files.
23202 This parameter is optional, and has a default value of -3.01dB.
23205 Set a specific target level (in LUFS) used as relative zero in the visualization.
23206 This parameter is optional and has a default value of -23LUFS as specified
23207 by EBU R128. However, material published online may prefer a level of -16LUFS
23208 (e.g. for use with podcasts or video platforms).
23211 Set the value displayed by the gauge. Valid values are @code{momentary} and s
23212 @code{shortterm}. By default the momentary value will be used, but in certain
23213 scenarios it may be more useful to observe the short term value instead (e.g.
23217 Sets the display scale for the loudness. Valid parameters are @code{absolute}
23218 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
23219 video output, not the summary or continuous log output.
23222 @subsection Examples
23226 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
23228 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
23232 Run an analysis with @command{ffmpeg}:
23234 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
23238 @section interleave, ainterleave
23240 Temporally interleave frames from several inputs.
23242 @code{interleave} works with video inputs, @code{ainterleave} with audio.
23244 These filters read frames from several inputs and send the oldest
23245 queued frame to the output.
23247 Input streams must have well defined, monotonically increasing frame
23250 In order to submit one frame to output, these filters need to enqueue
23251 at least one frame for each input, so they cannot work in case one
23252 input is not yet terminated and will not receive incoming frames.
23254 For example consider the case when one input is a @code{select} filter
23255 which always drops input frames. The @code{interleave} filter will keep
23256 reading from that input, but it will never be able to send new frames
23257 to output until the input sends an end-of-stream signal.
23259 Also, depending on inputs synchronization, the filters will drop
23260 frames in case one input receives more frames than the other ones, and
23261 the queue is already filled.
23263 These filters accept the following options:
23267 Set the number of different inputs, it is 2 by default.
23270 @subsection Examples
23274 Interleave frames belonging to different streams using @command{ffmpeg}:
23276 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
23280 Add flickering blur effect:
23282 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
23286 @section metadata, ametadata
23288 Manipulate frame metadata.
23290 This filter accepts the following options:
23294 Set mode of operation of the filter.
23296 Can be one of the following:
23300 If both @code{value} and @code{key} is set, select frames
23301 which have such metadata. If only @code{key} is set, select
23302 every frame that has such key in metadata.
23305 Add new metadata @code{key} and @code{value}. If key is already available
23309 Modify value of already present key.
23312 If @code{value} is set, delete only keys that have such value.
23313 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
23317 Print key and its value if metadata was found. If @code{key} is not set print all
23318 metadata values available in frame.
23322 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
23325 Set metadata value which will be used. This option is mandatory for
23326 @code{modify} and @code{add} mode.
23329 Which function to use when comparing metadata value and @code{value}.
23331 Can be one of following:
23335 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
23338 Values are interpreted as strings, returns true if metadata value starts with
23339 the @code{value} option string.
23342 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
23345 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
23348 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
23351 Values are interpreted as floats, returns true if expression from option @code{expr}
23355 Values are interpreted as strings, returns true if metadata value ends with
23356 the @code{value} option string.
23360 Set expression which is used when @code{function} is set to @code{expr}.
23361 The expression is evaluated through the eval API and can contain the following
23366 Float representation of @code{value} from metadata key.
23369 Float representation of @code{value} as supplied by user in @code{value} option.
23373 If specified in @code{print} mode, output is written to the named file. Instead of
23374 plain filename any writable url can be specified. Filename ``-'' is a shorthand
23375 for standard output. If @code{file} option is not set, output is written to the log
23376 with AV_LOG_INFO loglevel.
23379 Reduces buffering in print mode when output is written to a URL set using @var{file}.
23383 @subsection Examples
23387 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
23390 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
23393 Print silencedetect output to file @file{metadata.txt}.
23395 silencedetect,ametadata=mode=print:file=metadata.txt
23398 Direct all metadata to a pipe with file descriptor 4.
23400 metadata=mode=print:file='pipe\:4'
23404 @section perms, aperms
23406 Set read/write permissions for the output frames.
23408 These filters are mainly aimed at developers to test direct path in the
23409 following filter in the filtergraph.
23411 The filters accept the following options:
23415 Select the permissions mode.
23417 It accepts the following values:
23420 Do nothing. This is the default.
23422 Set all the output frames read-only.
23424 Set all the output frames directly writable.
23426 Make the frame read-only if writable, and writable if read-only.
23428 Set each output frame read-only or writable randomly.
23432 Set the seed for the @var{random} mode, must be an integer included between
23433 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
23434 @code{-1}, the filter will try to use a good random seed on a best effort
23438 Note: in case of auto-inserted filter between the permission filter and the
23439 following one, the permission might not be received as expected in that
23440 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
23441 perms/aperms filter can avoid this problem.
23443 @section realtime, arealtime
23445 Slow down filtering to match real time approximately.
23447 These filters will pause the filtering for a variable amount of time to
23448 match the output rate with the input timestamps.
23449 They are similar to the @option{re} option to @code{ffmpeg}.
23451 They accept the following options:
23455 Time limit for the pauses. Any pause longer than that will be considered
23456 a timestamp discontinuity and reset the timer. Default is 2 seconds.
23458 Speed factor for processing. The value must be a float larger than zero.
23459 Values larger than 1.0 will result in faster than realtime processing,
23460 smaller will slow processing down. The @var{limit} is automatically adapted
23461 accordingly. Default is 1.0.
23463 A processing speed faster than what is possible without these filters cannot
23468 @section select, aselect
23470 Select frames to pass in output.
23472 This filter accepts the following options:
23477 Set expression, which is evaluated for each input frame.
23479 If the expression is evaluated to zero, the frame is discarded.
23481 If the evaluation result is negative or NaN, the frame is sent to the
23482 first output; otherwise it is sent to the output with index
23483 @code{ceil(val)-1}, assuming that the input index starts from 0.
23485 For example a value of @code{1.2} corresponds to the output with index
23486 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
23489 Set the number of outputs. The output to which to send the selected
23490 frame is based on the result of the evaluation. Default value is 1.
23493 The expression can contain the following constants:
23497 The (sequential) number of the filtered frame, starting from 0.
23500 The (sequential) number of the selected frame, starting from 0.
23502 @item prev_selected_n
23503 The sequential number of the last selected frame. It's NAN if undefined.
23506 The timebase of the input timestamps.
23509 The PTS (Presentation TimeStamp) of the filtered video frame,
23510 expressed in @var{TB} units. It's NAN if undefined.
23513 The PTS of the filtered video frame,
23514 expressed in seconds. It's NAN if undefined.
23517 The PTS of the previously filtered video frame. It's NAN if undefined.
23519 @item prev_selected_pts
23520 The PTS of the last previously filtered video frame. It's NAN if undefined.
23522 @item prev_selected_t
23523 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
23526 The PTS of the first video frame in the video. It's NAN if undefined.
23529 The time of the first video frame in the video. It's NAN if undefined.
23531 @item pict_type @emph{(video only)}
23532 The type of the filtered frame. It can assume one of the following
23544 @item interlace_type @emph{(video only)}
23545 The frame interlace type. It can assume one of the following values:
23548 The frame is progressive (not interlaced).
23550 The frame is top-field-first.
23552 The frame is bottom-field-first.
23555 @item consumed_sample_n @emph{(audio only)}
23556 the number of selected samples before the current frame
23558 @item samples_n @emph{(audio only)}
23559 the number of samples in the current frame
23561 @item sample_rate @emph{(audio only)}
23562 the input sample rate
23565 This is 1 if the filtered frame is a key-frame, 0 otherwise.
23568 the position in the file of the filtered frame, -1 if the information
23569 is not available (e.g. for synthetic video)
23571 @item scene @emph{(video only)}
23572 value between 0 and 1 to indicate a new scene; a low value reflects a low
23573 probability for the current frame to introduce a new scene, while a higher
23574 value means the current frame is more likely to be one (see the example below)
23576 @item concatdec_select
23577 The concat demuxer can select only part of a concat input file by setting an
23578 inpoint and an outpoint, but the output packets may not be entirely contained
23579 in the selected interval. By using this variable, it is possible to skip frames
23580 generated by the concat demuxer which are not exactly contained in the selected
23583 This works by comparing the frame pts against the @var{lavf.concat.start_time}
23584 and the @var{lavf.concat.duration} packet metadata values which are also
23585 present in the decoded frames.
23587 The @var{concatdec_select} variable is -1 if the frame pts is at least
23588 start_time and either the duration metadata is missing or the frame pts is less
23589 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
23592 That basically means that an input frame is selected if its pts is within the
23593 interval set by the concat demuxer.
23597 The default value of the select expression is "1".
23599 @subsection Examples
23603 Select all frames in input:
23608 The example above is the same as:
23620 Select only I-frames:
23622 select='eq(pict_type\,I)'
23626 Select one frame every 100:
23628 select='not(mod(n\,100))'
23632 Select only frames contained in the 10-20 time interval:
23634 select=between(t\,10\,20)
23638 Select only I-frames contained in the 10-20 time interval:
23640 select=between(t\,10\,20)*eq(pict_type\,I)
23644 Select frames with a minimum distance of 10 seconds:
23646 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
23650 Use aselect to select only audio frames with samples number > 100:
23652 aselect='gt(samples_n\,100)'
23656 Create a mosaic of the first scenes:
23658 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
23661 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
23665 Send even and odd frames to separate outputs, and compose them:
23667 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
23671 Select useful frames from an ffconcat file which is using inpoints and
23672 outpoints but where the source files are not intra frame only.
23674 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
23678 @section sendcmd, asendcmd
23680 Send commands to filters in the filtergraph.
23682 These filters read commands to be sent to other filters in the
23685 @code{sendcmd} must be inserted between two video filters,
23686 @code{asendcmd} must be inserted between two audio filters, but apart
23687 from that they act the same way.
23689 The specification of commands can be provided in the filter arguments
23690 with the @var{commands} option, or in a file specified by the
23691 @var{filename} option.
23693 These filters accept the following options:
23696 Set the commands to be read and sent to the other filters.
23698 Set the filename of the commands to be read and sent to the other
23702 @subsection Commands syntax
23704 A commands description consists of a sequence of interval
23705 specifications, comprising a list of commands to be executed when a
23706 particular event related to that interval occurs. The occurring event
23707 is typically the current frame time entering or leaving a given time
23710 An interval is specified by the following syntax:
23712 @var{START}[-@var{END}] @var{COMMANDS};
23715 The time interval is specified by the @var{START} and @var{END} times.
23716 @var{END} is optional and defaults to the maximum time.
23718 The current frame time is considered within the specified interval if
23719 it is included in the interval [@var{START}, @var{END}), that is when
23720 the time is greater or equal to @var{START} and is lesser than
23723 @var{COMMANDS} consists of a sequence of one or more command
23724 specifications, separated by ",", relating to that interval. The
23725 syntax of a command specification is given by:
23727 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
23730 @var{FLAGS} is optional and specifies the type of events relating to
23731 the time interval which enable sending the specified command, and must
23732 be a non-null sequence of identifier flags separated by "+" or "|" and
23733 enclosed between "[" and "]".
23735 The following flags are recognized:
23738 The command is sent when the current frame timestamp enters the
23739 specified interval. In other words, the command is sent when the
23740 previous frame timestamp was not in the given interval, and the
23744 The command is sent when the current frame timestamp leaves the
23745 specified interval. In other words, the command is sent when the
23746 previous frame timestamp was in the given interval, and the
23750 The command @var{ARG} is interpreted as expression and result of
23751 expression is passed as @var{ARG}.
23753 The expression is evaluated through the eval API and can contain the following
23758 Original position in the file of the frame, or undefined if undefined
23759 for the current frame.
23762 The presentation timestamp in input.
23765 The count of the input frame for video or audio, starting from 0.
23768 The time in seconds of the current frame.
23771 The start time in seconds of the current command interval.
23774 The end time in seconds of the current command interval.
23777 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
23782 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
23785 @var{TARGET} specifies the target of the command, usually the name of
23786 the filter class or a specific filter instance name.
23788 @var{COMMAND} specifies the name of the command for the target filter.
23790 @var{ARG} is optional and specifies the optional list of argument for
23791 the given @var{COMMAND}.
23793 Between one interval specification and another, whitespaces, or
23794 sequences of characters starting with @code{#} until the end of line,
23795 are ignored and can be used to annotate comments.
23797 A simplified BNF description of the commands specification syntax
23800 @var{COMMAND_FLAG} ::= "enter" | "leave"
23801 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
23802 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
23803 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
23804 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
23805 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
23808 @subsection Examples
23812 Specify audio tempo change at second 4:
23814 asendcmd=c='4.0 atempo tempo 1.5',atempo
23818 Target a specific filter instance:
23820 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
23824 Specify a list of drawtext and hue commands in a file.
23826 # show text in the interval 5-10
23827 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
23828 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
23830 # desaturate the image in the interval 15-20
23831 15.0-20.0 [enter] hue s 0,
23832 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
23834 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
23836 # apply an exponential saturation fade-out effect, starting from time 25
23837 25 [enter] hue s exp(25-t)
23840 A filtergraph allowing to read and process the above command list
23841 stored in a file @file{test.cmd}, can be specified with:
23843 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
23848 @section setpts, asetpts
23850 Change the PTS (presentation timestamp) of the input frames.
23852 @code{setpts} works on video frames, @code{asetpts} on audio frames.
23854 This filter accepts the following options:
23859 The expression which is evaluated for each frame to construct its timestamp.
23863 The expression is evaluated through the eval API and can contain the following
23867 @item FRAME_RATE, FR
23868 frame rate, only defined for constant frame-rate video
23871 The presentation timestamp in input
23874 The count of the input frame for video or the number of consumed samples,
23875 not including the current frame for audio, starting from 0.
23877 @item NB_CONSUMED_SAMPLES
23878 The number of consumed samples, not including the current frame (only
23881 @item NB_SAMPLES, S
23882 The number of samples in the current frame (only audio)
23884 @item SAMPLE_RATE, SR
23885 The audio sample rate.
23888 The PTS of the first frame.
23891 the time in seconds of the first frame
23894 State whether the current frame is interlaced.
23897 the time in seconds of the current frame
23900 original position in the file of the frame, or undefined if undefined
23901 for the current frame
23904 The previous input PTS.
23907 previous input time in seconds
23910 The previous output PTS.
23913 previous output time in seconds
23916 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
23920 The wallclock (RTC) time at the start of the movie in microseconds.
23923 The timebase of the input timestamps.
23927 @subsection Examples
23931 Start counting PTS from zero
23933 setpts=PTS-STARTPTS
23937 Apply fast motion effect:
23943 Apply slow motion effect:
23949 Set fixed rate of 25 frames per second:
23955 Set fixed rate 25 fps with some jitter:
23957 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
23961 Apply an offset of 10 seconds to the input PTS:
23967 Generate timestamps from a "live source" and rebase onto the current timebase:
23969 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
23973 Generate timestamps by counting samples:
23982 Force color range for the output video frame.
23984 The @code{setrange} filter marks the color range property for the
23985 output frames. It does not change the input frame, but only sets the
23986 corresponding property, which affects how the frame is treated by
23989 The filter accepts the following options:
23994 Available values are:
23998 Keep the same color range property.
24000 @item unspecified, unknown
24001 Set the color range as unspecified.
24003 @item limited, tv, mpeg
24004 Set the color range as limited.
24006 @item full, pc, jpeg
24007 Set the color range as full.
24011 @section settb, asettb
24013 Set the timebase to use for the output frames timestamps.
24014 It is mainly useful for testing timebase configuration.
24016 It accepts the following parameters:
24021 The expression which is evaluated into the output timebase.
24025 The value for @option{tb} is an arithmetic expression representing a
24026 rational. The expression can contain the constants "AVTB" (the default
24027 timebase), "intb" (the input timebase) and "sr" (the sample rate,
24028 audio only). Default value is "intb".
24030 @subsection Examples
24034 Set the timebase to 1/25:
24040 Set the timebase to 1/10:
24046 Set the timebase to 1001/1000:
24052 Set the timebase to 2*intb:
24058 Set the default timebase value:
24065 Convert input audio to a video output representing frequency spectrum
24066 logarithmically using Brown-Puckette constant Q transform algorithm with
24067 direct frequency domain coefficient calculation (but the transform itself
24068 is not really constant Q, instead the Q factor is actually variable/clamped),
24069 with musical tone scale, from E0 to D#10.
24071 The filter accepts the following options:
24075 Specify the video size for the output. It must be even. For the syntax of this option,
24076 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24077 Default value is @code{1920x1080}.
24080 Set the output frame rate. Default value is @code{25}.
24083 Set the bargraph height. It must be even. Default value is @code{-1} which
24084 computes the bargraph height automatically.
24087 Set the axis height. It must be even. Default value is @code{-1} which computes
24088 the axis height automatically.
24091 Set the sonogram height. It must be even. Default value is @code{-1} which
24092 computes the sonogram height automatically.
24095 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
24096 instead. Default value is @code{1}.
24098 @item sono_v, volume
24099 Specify the sonogram volume expression. It can contain variables:
24102 the @var{bar_v} evaluated expression
24103 @item frequency, freq, f
24104 the frequency where it is evaluated
24105 @item timeclamp, tc
24106 the value of @var{timeclamp} option
24110 @item a_weighting(f)
24111 A-weighting of equal loudness
24112 @item b_weighting(f)
24113 B-weighting of equal loudness
24114 @item c_weighting(f)
24115 C-weighting of equal loudness.
24117 Default value is @code{16}.
24119 @item bar_v, volume2
24120 Specify the bargraph volume expression. It can contain variables:
24123 the @var{sono_v} evaluated expression
24124 @item frequency, freq, f
24125 the frequency where it is evaluated
24126 @item timeclamp, tc
24127 the value of @var{timeclamp} option
24131 @item a_weighting(f)
24132 A-weighting of equal loudness
24133 @item b_weighting(f)
24134 B-weighting of equal loudness
24135 @item c_weighting(f)
24136 C-weighting of equal loudness.
24138 Default value is @code{sono_v}.
24140 @item sono_g, gamma
24141 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
24142 higher gamma makes the spectrum having more range. Default value is @code{3}.
24143 Acceptable range is @code{[1, 7]}.
24145 @item bar_g, gamma2
24146 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
24150 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
24151 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
24153 @item timeclamp, tc
24154 Specify the transform timeclamp. At low frequency, there is trade-off between
24155 accuracy in time domain and frequency domain. If timeclamp is lower,
24156 event in time domain is represented more accurately (such as fast bass drum),
24157 otherwise event in frequency domain is represented more accurately
24158 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
24161 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
24162 limits future samples by applying asymmetric windowing in time domain, useful
24163 when low latency is required. Accepted range is @code{[0, 1]}.
24166 Specify the transform base frequency. Default value is @code{20.01523126408007475},
24167 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
24170 Specify the transform end frequency. Default value is @code{20495.59681441799654},
24171 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
24174 This option is deprecated and ignored.
24177 Specify the transform length in time domain. Use this option to control accuracy
24178 trade-off between time domain and frequency domain at every frequency sample.
24179 It can contain variables:
24181 @item frequency, freq, f
24182 the frequency where it is evaluated
24183 @item timeclamp, tc
24184 the value of @var{timeclamp} option.
24186 Default value is @code{384*tc/(384+tc*f)}.
24189 Specify the transform count for every video frame. Default value is @code{6}.
24190 Acceptable range is @code{[1, 30]}.
24193 Specify the transform count for every single pixel. Default value is @code{0},
24194 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
24197 Specify font file for use with freetype to draw the axis. If not specified,
24198 use embedded font. Note that drawing with font file or embedded font is not
24199 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
24203 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
24204 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
24208 Specify font color expression. This is arithmetic expression that should return
24209 integer value 0xRRGGBB. It can contain variables:
24211 @item frequency, freq, f
24212 the frequency where it is evaluated
24213 @item timeclamp, tc
24214 the value of @var{timeclamp} option
24219 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
24220 @item r(x), g(x), b(x)
24221 red, green, and blue value of intensity x.
24223 Default value is @code{st(0, (midi(f)-59.5)/12);
24224 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
24225 r(1-ld(1)) + b(ld(1))}.
24228 Specify image file to draw the axis. This option override @var{fontfile} and
24229 @var{fontcolor} option.
24232 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
24233 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
24234 Default value is @code{1}.
24237 Set colorspace. The accepted values are:
24240 Unspecified (default)
24249 BT.470BG or BT.601-6 625
24252 SMPTE-170M or BT.601-6 525
24258 BT.2020 with non-constant luminance
24263 Set spectrogram color scheme. This is list of floating point values with format
24264 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
24265 The default is @code{1|0.5|0|0|0.5|1}.
24269 @subsection Examples
24273 Playing audio while showing the spectrum:
24275 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
24279 Same as above, but with frame rate 30 fps:
24281 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
24285 Playing at 1280x720:
24287 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
24291 Disable sonogram display:
24297 A1 and its harmonics: A1, A2, (near)E3, A3:
24299 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),
24300 asplit[a][out1]; [a] showcqt [out0]'
24304 Same as above, but with more accuracy in frequency domain:
24306 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),
24307 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
24313 bar_v=10:sono_v=bar_v*a_weighting(f)
24317 Custom gamma, now spectrum is linear to the amplitude.
24323 Custom tlength equation:
24325 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)))'
24329 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
24331 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
24335 Custom font using fontconfig:
24337 font='Courier New,Monospace,mono|bold'
24341 Custom frequency range with custom axis using image file:
24343 axisfile=myaxis.png:basefreq=40:endfreq=10000
24349 Convert input audio to video output representing the audio power spectrum.
24350 Audio amplitude is on Y-axis while frequency is on X-axis.
24352 The filter accepts the following options:
24356 Specify size of video. For the syntax of this option, check the
24357 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24358 Default is @code{1024x512}.
24362 This set how each frequency bin will be represented.
24364 It accepts the following values:
24370 Default is @code{bar}.
24373 Set amplitude scale.
24375 It accepts the following values:
24389 Default is @code{log}.
24392 Set frequency scale.
24394 It accepts the following values:
24403 Reverse logarithmic scale.
24405 Default is @code{lin}.
24408 Set window size. Allowed range is from 16 to 65536.
24410 Default is @code{2048}
24413 Set windowing function.
24415 It accepts the following values:
24438 Default is @code{hanning}.
24441 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
24442 which means optimal overlap for selected window function will be picked.
24445 Set time averaging. Setting this to 0 will display current maximal peaks.
24446 Default is @code{1}, which means time averaging is disabled.
24449 Specify list of colors separated by space or by '|' which will be used to
24450 draw channel frequencies. Unrecognized or missing colors will be replaced
24454 Set channel display mode.
24456 It accepts the following values:
24461 Default is @code{combined}.
24464 Set minimum amplitude used in @code{log} amplitude scaler.
24468 @section showspatial
24470 Convert stereo input audio to a video output, representing the spatial relationship
24471 between two channels.
24473 The filter accepts the following options:
24477 Specify the video size for the output. For the syntax of this option, check the
24478 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24479 Default value is @code{512x512}.
24482 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
24485 Set window function.
24487 It accepts the following values:
24512 Default value is @code{hann}.
24515 Set ratio of overlap window. Default value is @code{0.5}.
24516 When value is @code{1} overlap is set to recommended size for specific
24517 window function currently used.
24520 @anchor{showspectrum}
24521 @section showspectrum
24523 Convert input audio to a video output, representing the audio frequency
24526 The filter accepts the following options:
24530 Specify the video size for the output. For the syntax of this option, check the
24531 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24532 Default value is @code{640x512}.
24535 Specify how the spectrum should slide along the window.
24537 It accepts the following values:
24540 the samples start again on the left when they reach the right
24542 the samples scroll from right to left
24544 frames are only produced when the samples reach the right
24546 the samples scroll from left to right
24549 Default value is @code{replace}.
24552 Specify display mode.
24554 It accepts the following values:
24557 all channels are displayed in the same row
24559 all channels are displayed in separate rows
24562 Default value is @samp{combined}.
24565 Specify display color mode.
24567 It accepts the following values:
24570 each channel is displayed in a separate color
24572 each channel is displayed using the same color scheme
24574 each channel is displayed using the rainbow color scheme
24576 each channel is displayed using the moreland color scheme
24578 each channel is displayed using the nebulae color scheme
24580 each channel is displayed using the fire color scheme
24582 each channel is displayed using the fiery color scheme
24584 each channel is displayed using the fruit color scheme
24586 each channel is displayed using the cool color scheme
24588 each channel is displayed using the magma color scheme
24590 each channel is displayed using the green color scheme
24592 each channel is displayed using the viridis color scheme
24594 each channel is displayed using the plasma color scheme
24596 each channel is displayed using the cividis color scheme
24598 each channel is displayed using the terrain color scheme
24601 Default value is @samp{channel}.
24604 Specify scale used for calculating intensity color values.
24606 It accepts the following values:
24611 square root, default
24622 Default value is @samp{sqrt}.
24625 Specify frequency scale.
24627 It accepts the following values:
24635 Default value is @samp{lin}.
24638 Set saturation modifier for displayed colors. Negative values provide
24639 alternative color scheme. @code{0} is no saturation at all.
24640 Saturation must be in [-10.0, 10.0] range.
24641 Default value is @code{1}.
24644 Set window function.
24646 It accepts the following values:
24671 Default value is @code{hann}.
24674 Set orientation of time vs frequency axis. Can be @code{vertical} or
24675 @code{horizontal}. Default is @code{vertical}.
24678 Set ratio of overlap window. Default value is @code{0}.
24679 When value is @code{1} overlap is set to recommended size for specific
24680 window function currently used.
24683 Set scale gain for calculating intensity color values.
24684 Default value is @code{1}.
24687 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
24690 Set color rotation, must be in [-1.0, 1.0] range.
24691 Default value is @code{0}.
24694 Set start frequency from which to display spectrogram. Default is @code{0}.
24697 Set stop frequency to which to display spectrogram. Default is @code{0}.
24700 Set upper frame rate limit. Default is @code{auto}, unlimited.
24703 Draw time and frequency axes and legends. Default is disabled.
24706 The usage is very similar to the showwaves filter; see the examples in that
24709 @subsection Examples
24713 Large window with logarithmic color scaling:
24715 showspectrum=s=1280x480:scale=log
24719 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
24721 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24722 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
24726 @section showspectrumpic
24728 Convert input audio to a single video frame, representing the audio frequency
24731 The filter accepts the following options:
24735 Specify the video size for the output. For the syntax of this option, check the
24736 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24737 Default value is @code{4096x2048}.
24740 Specify display mode.
24742 It accepts the following values:
24745 all channels are displayed in the same row
24747 all channels are displayed in separate rows
24749 Default value is @samp{combined}.
24752 Specify display color mode.
24754 It accepts the following values:
24757 each channel is displayed in a separate color
24759 each channel is displayed using the same color scheme
24761 each channel is displayed using the rainbow color scheme
24763 each channel is displayed using the moreland color scheme
24765 each channel is displayed using the nebulae color scheme
24767 each channel is displayed using the fire color scheme
24769 each channel is displayed using the fiery color scheme
24771 each channel is displayed using the fruit color scheme
24773 each channel is displayed using the cool color scheme
24775 each channel is displayed using the magma color scheme
24777 each channel is displayed using the green color scheme
24779 each channel is displayed using the viridis color scheme
24781 each channel is displayed using the plasma color scheme
24783 each channel is displayed using the cividis color scheme
24785 each channel is displayed using the terrain color scheme
24787 Default value is @samp{intensity}.
24790 Specify scale used for calculating intensity color values.
24792 It accepts the following values:
24797 square root, default
24807 Default value is @samp{log}.
24810 Specify frequency scale.
24812 It accepts the following values:
24820 Default value is @samp{lin}.
24823 Set saturation modifier for displayed colors. Negative values provide
24824 alternative color scheme. @code{0} is no saturation at all.
24825 Saturation must be in [-10.0, 10.0] range.
24826 Default value is @code{1}.
24829 Set window function.
24831 It accepts the following values:
24855 Default value is @code{hann}.
24858 Set orientation of time vs frequency axis. Can be @code{vertical} or
24859 @code{horizontal}. Default is @code{vertical}.
24862 Set scale gain for calculating intensity color values.
24863 Default value is @code{1}.
24866 Draw time and frequency axes and legends. Default is enabled.
24869 Set color rotation, must be in [-1.0, 1.0] range.
24870 Default value is @code{0}.
24873 Set start frequency from which to display spectrogram. Default is @code{0}.
24876 Set stop frequency to which to display spectrogram. Default is @code{0}.
24879 @subsection Examples
24883 Extract an audio spectrogram of a whole audio track
24884 in a 1024x1024 picture using @command{ffmpeg}:
24886 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
24890 @section showvolume
24892 Convert input audio volume to a video output.
24894 The filter accepts the following options:
24901 Set border width, allowed range is [0, 5]. Default is 1.
24904 Set channel width, allowed range is [80, 8192]. Default is 400.
24907 Set channel height, allowed range is [1, 900]. Default is 20.
24910 Set fade, allowed range is [0, 1]. Default is 0.95.
24913 Set volume color expression.
24915 The expression can use the following variables:
24919 Current max volume of channel in dB.
24925 Current channel number, starting from 0.
24929 If set, displays channel names. Default is enabled.
24932 If set, displays volume values. Default is enabled.
24935 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
24936 default is @code{h}.
24939 Set step size, allowed range is [0, 5]. Default is 0, which means
24943 Set background opacity, allowed range is [0, 1]. Default is 0.
24946 Set metering mode, can be peak: @code{p} or rms: @code{r},
24947 default is @code{p}.
24950 Set display scale, can be linear: @code{lin} or log: @code{log},
24951 default is @code{lin}.
24955 If set to > 0., display a line for the max level
24956 in the previous seconds.
24957 default is disabled: @code{0.}
24960 The color of the max line. Use when @code{dm} option is set to > 0.
24961 default is: @code{orange}
24966 Convert input audio to a video output, representing the samples waves.
24968 The filter accepts the following options:
24972 Specify the video size for the output. For the syntax of this option, check the
24973 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24974 Default value is @code{600x240}.
24979 Available values are:
24982 Draw a point for each sample.
24985 Draw a vertical line for each sample.
24988 Draw a point for each sample and a line between them.
24991 Draw a centered vertical line for each sample.
24994 Default value is @code{point}.
24997 Set the number of samples which are printed on the same column. A
24998 larger value will decrease the frame rate. Must be a positive
24999 integer. This option can be set only if the value for @var{rate}
25000 is not explicitly specified.
25003 Set the (approximate) output frame rate. This is done by setting the
25004 option @var{n}. Default value is "25".
25006 @item split_channels
25007 Set if channels should be drawn separately or overlap. Default value is 0.
25010 Set colors separated by '|' which are going to be used for drawing of each channel.
25013 Set amplitude scale.
25015 Available values are:
25033 Set the draw mode. This is mostly useful to set for high @var{n}.
25035 Available values are:
25038 Scale pixel values for each drawn sample.
25041 Draw every sample directly.
25044 Default value is @code{scale}.
25047 @subsection Examples
25051 Output the input file audio and the corresponding video representation
25054 amovie=a.mp3,asplit[out0],showwaves[out1]
25058 Create a synthetic signal and show it with showwaves, forcing a
25059 frame rate of 30 frames per second:
25061 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
25065 @section showwavespic
25067 Convert input audio to a single video frame, representing the samples waves.
25069 The filter accepts the following options:
25073 Specify the video size for the output. For the syntax of this option, check the
25074 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25075 Default value is @code{600x240}.
25077 @item split_channels
25078 Set if channels should be drawn separately or overlap. Default value is 0.
25081 Set colors separated by '|' which are going to be used for drawing of each channel.
25084 Set amplitude scale.
25086 Available values are:
25106 Available values are:
25109 Scale pixel values for each drawn sample.
25112 Draw every sample directly.
25115 Default value is @code{scale}.
25118 @subsection Examples
25122 Extract a channel split representation of the wave form of a whole audio track
25123 in a 1024x800 picture using @command{ffmpeg}:
25125 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
25129 @section sidedata, asidedata
25131 Delete frame side data, or select frames based on it.
25133 This filter accepts the following options:
25137 Set mode of operation of the filter.
25139 Can be one of the following:
25143 Select every frame with side data of @code{type}.
25146 Delete side data of @code{type}. If @code{type} is not set, delete all side
25152 Set side data type used with all modes. Must be set for @code{select} mode. For
25153 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
25154 in @file{libavutil/frame.h}. For example, to choose
25155 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
25159 @section spectrumsynth
25161 Synthesize audio from 2 input video spectrums, first input stream represents
25162 magnitude across time and second represents phase across time.
25163 The filter will transform from frequency domain as displayed in videos back
25164 to time domain as presented in audio output.
25166 This filter is primarily created for reversing processed @ref{showspectrum}
25167 filter outputs, but can synthesize sound from other spectrograms too.
25168 But in such case results are going to be poor if the phase data is not
25169 available, because in such cases phase data need to be recreated, usually
25170 it's just recreated from random noise.
25171 For best results use gray only output (@code{channel} color mode in
25172 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
25173 @code{lin} scale for phase video. To produce phase, for 2nd video, use
25174 @code{data} option. Inputs videos should generally use @code{fullframe}
25175 slide mode as that saves resources needed for decoding video.
25177 The filter accepts the following options:
25181 Specify sample rate of output audio, the sample rate of audio from which
25182 spectrum was generated may differ.
25185 Set number of channels represented in input video spectrums.
25188 Set scale which was used when generating magnitude input spectrum.
25189 Can be @code{lin} or @code{log}. Default is @code{log}.
25192 Set slide which was used when generating inputs spectrums.
25193 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
25194 Default is @code{fullframe}.
25197 Set window function used for resynthesis.
25200 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25201 which means optimal overlap for selected window function will be picked.
25204 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
25205 Default is @code{vertical}.
25208 @subsection Examples
25212 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
25213 then resynthesize videos back to audio with spectrumsynth:
25215 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
25216 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
25217 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
25221 @section split, asplit
25223 Split input into several identical outputs.
25225 @code{asplit} works with audio input, @code{split} with video.
25227 The filter accepts a single parameter which specifies the number of outputs. If
25228 unspecified, it defaults to 2.
25230 @subsection Examples
25234 Create two separate outputs from the same input:
25236 [in] split [out0][out1]
25240 To create 3 or more outputs, you need to specify the number of
25243 [in] asplit=3 [out0][out1][out2]
25247 Create two separate outputs from the same input, one cropped and
25250 [in] split [splitout1][splitout2];
25251 [splitout1] crop=100:100:0:0 [cropout];
25252 [splitout2] pad=200:200:100:100 [padout];
25256 Create 5 copies of the input audio with @command{ffmpeg}:
25258 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
25264 Receive commands sent through a libzmq client, and forward them to
25265 filters in the filtergraph.
25267 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
25268 must be inserted between two video filters, @code{azmq} between two
25269 audio filters. Both are capable to send messages to any filter type.
25271 To enable these filters you need to install the libzmq library and
25272 headers and configure FFmpeg with @code{--enable-libzmq}.
25274 For more information about libzmq see:
25275 @url{http://www.zeromq.org/}
25277 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
25278 receives messages sent through a network interface defined by the
25279 @option{bind_address} (or the abbreviation "@option{b}") option.
25280 Default value of this option is @file{tcp://localhost:5555}. You may
25281 want to alter this value to your needs, but do not forget to escape any
25282 ':' signs (see @ref{filtergraph escaping}).
25284 The received message must be in the form:
25286 @var{TARGET} @var{COMMAND} [@var{ARG}]
25289 @var{TARGET} specifies the target of the command, usually the name of
25290 the filter class or a specific filter instance name. The default
25291 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
25292 but you can override this by using the @samp{filter_name@@id} syntax
25293 (see @ref{Filtergraph syntax}).
25295 @var{COMMAND} specifies the name of the command for the target filter.
25297 @var{ARG} is optional and specifies the optional argument list for the
25298 given @var{COMMAND}.
25300 Upon reception, the message is processed and the corresponding command
25301 is injected into the filtergraph. Depending on the result, the filter
25302 will send a reply to the client, adopting the format:
25304 @var{ERROR_CODE} @var{ERROR_REASON}
25308 @var{MESSAGE} is optional.
25310 @subsection Examples
25312 Look at @file{tools/zmqsend} for an example of a zmq client which can
25313 be used to send commands processed by these filters.
25315 Consider the following filtergraph generated by @command{ffplay}.
25316 In this example the last overlay filter has an instance name. All other
25317 filters will have default instance names.
25320 ffplay -dumpgraph 1 -f lavfi "
25321 color=s=100x100:c=red [l];
25322 color=s=100x100:c=blue [r];
25323 nullsrc=s=200x100, zmq [bg];
25324 [bg][l] overlay [bg+l];
25325 [bg+l][r] overlay@@my=x=100 "
25328 To change the color of the left side of the video, the following
25329 command can be used:
25331 echo Parsed_color_0 c yellow | tools/zmqsend
25334 To change the right side:
25336 echo Parsed_color_1 c pink | tools/zmqsend
25339 To change the position of the right side:
25341 echo overlay@@my x 150 | tools/zmqsend
25345 @c man end MULTIMEDIA FILTERS
25347 @chapter Multimedia Sources
25348 @c man begin MULTIMEDIA SOURCES
25350 Below is a description of the currently available multimedia sources.
25354 This is the same as @ref{movie} source, except it selects an audio
25360 Read audio and/or video stream(s) from a movie container.
25362 It accepts the following parameters:
25366 The name of the resource to read (not necessarily a file; it can also be a
25367 device or a stream accessed through some protocol).
25369 @item format_name, f
25370 Specifies the format assumed for the movie to read, and can be either
25371 the name of a container or an input device. If not specified, the
25372 format is guessed from @var{movie_name} or by probing.
25374 @item seek_point, sp
25375 Specifies the seek point in seconds. The frames will be output
25376 starting from this seek point. The parameter is evaluated with
25377 @code{av_strtod}, so the numerical value may be suffixed by an IS
25378 postfix. The default value is "0".
25381 Specifies the streams to read. Several streams can be specified,
25382 separated by "+". The source will then have as many outputs, in the
25383 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
25384 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
25385 respectively the default (best suited) video and audio stream. Default
25386 is "dv", or "da" if the filter is called as "amovie".
25388 @item stream_index, si
25389 Specifies the index of the video stream to read. If the value is -1,
25390 the most suitable video stream will be automatically selected. The default
25391 value is "-1". Deprecated. If the filter is called "amovie", it will select
25392 audio instead of video.
25395 Specifies how many times to read the stream in sequence.
25396 If the value is 0, the stream will be looped infinitely.
25397 Default value is "1".
25399 Note that when the movie is looped the source timestamps are not
25400 changed, so it will generate non monotonically increasing timestamps.
25402 @item discontinuity
25403 Specifies the time difference between frames above which the point is
25404 considered a timestamp discontinuity which is removed by adjusting the later
25408 It allows overlaying a second video on top of the main input of
25409 a filtergraph, as shown in this graph:
25411 input -----------> deltapts0 --> overlay --> output
25414 movie --> scale--> deltapts1 -------+
25416 @subsection Examples
25420 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
25421 on top of the input labelled "in":
25423 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
25424 [in] setpts=PTS-STARTPTS [main];
25425 [main][over] overlay=16:16 [out]
25429 Read from a video4linux2 device, and overlay it on top of the input
25432 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
25433 [in] setpts=PTS-STARTPTS [main];
25434 [main][over] overlay=16:16 [out]
25438 Read the first video stream and the audio stream with id 0x81 from
25439 dvd.vob; the video is connected to the pad named "video" and the audio is
25440 connected to the pad named "audio":
25442 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
25446 @subsection Commands
25448 Both movie and amovie support the following commands:
25451 Perform seek using "av_seek_frame".
25452 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
25455 @var{stream_index}: If stream_index is -1, a default
25456 stream is selected, and @var{timestamp} is automatically converted
25457 from AV_TIME_BASE units to the stream specific time_base.
25459 @var{timestamp}: Timestamp in AVStream.time_base units
25460 or, if no stream is specified, in AV_TIME_BASE units.
25462 @var{flags}: Flags which select direction and seeking mode.
25466 Get movie duration in AV_TIME_BASE units.
25470 @c man end MULTIMEDIA SOURCES