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.
2261 Automatic Speech Recognition
2263 This filter uses PocketSphinx for speech recognition. To enable
2264 compilation of this filter, you need to configure FFmpeg with
2265 @code{--enable-pocketsphinx}.
2267 It accepts the following options:
2271 Set sampling rate of input audio. Defaults is @code{16000}.
2272 This need to match speech models, otherwise one will get poor results.
2275 Set dictionary containing acoustic model files.
2278 Set pronunciation dictionary.
2281 Set language model file.
2284 Set language model set.
2287 Set which language model to use.
2290 Set output for log messages.
2293 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2298 Display time domain statistical information about the audio channels.
2299 Statistics are calculated and displayed for each audio channel and,
2300 where applicable, an overall figure is also given.
2302 It accepts the following option:
2305 Short window length in seconds, used for peak and trough RMS measurement.
2306 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2310 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2311 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2314 Available keys for each channel are:
2356 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2357 this @code{lavfi.astats.Overall.Peak_count}.
2359 For description what each key means read below.
2362 Set number of frame after which stats are going to be recalculated.
2363 Default is disabled.
2365 @item measure_perchannel
2366 Select the entries which need to be measured per channel. The metadata keys can
2367 be used as flags, default is @option{all} which measures everything.
2368 @option{none} disables all per channel measurement.
2370 @item measure_overall
2371 Select the entries which need to be measured overall. The metadata keys can
2372 be used as flags, default is @option{all} which measures everything.
2373 @option{none} disables all overall measurement.
2377 A description of each shown parameter follows:
2381 Mean amplitude displacement from zero.
2384 Minimal sample level.
2387 Maximal sample level.
2389 @item Min difference
2390 Minimal difference between two consecutive samples.
2392 @item Max difference
2393 Maximal difference between two consecutive samples.
2395 @item Mean difference
2396 Mean difference between two consecutive samples.
2397 The average of each difference between two consecutive samples.
2399 @item RMS difference
2400 Root Mean Square difference between two consecutive samples.
2404 Standard peak and RMS level measured in dBFS.
2408 Peak and trough values for RMS level measured over a short window.
2411 Standard ratio of peak to RMS level (note: not in dB).
2414 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2415 (i.e. either @var{Min level} or @var{Max level}).
2418 Number of occasions (not the number of samples) that the signal attained either
2419 @var{Min level} or @var{Max level}.
2422 Overall bit depth of audio. Number of bits used for each sample.
2425 Measured dynamic range of audio in dB.
2427 @item Zero crossings
2428 Number of points where the waveform crosses the zero level axis.
2430 @item Zero crossings rate
2431 Rate of Zero crossings and number of audio samples.
2438 The filter accepts exactly one parameter, the audio tempo. If not
2439 specified then the filter will assume nominal 1.0 tempo. Tempo must
2440 be in the [0.5, 100.0] range.
2442 Note that tempo greater than 2 will skip some samples rather than
2443 blend them in. If for any reason this is a concern it is always
2444 possible to daisy-chain several instances of atempo to achieve the
2445 desired product tempo.
2447 @subsection Examples
2451 Slow down audio to 80% tempo:
2457 To speed up audio to 300% tempo:
2463 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2465 atempo=sqrt(3),atempo=sqrt(3)
2469 @subsection Commands
2471 This filter supports the following commands:
2474 Change filter tempo scale factor.
2475 Syntax for the command is : "@var{tempo}"
2480 Trim the input so that the output contains one continuous subpart of the input.
2482 It accepts the following parameters:
2485 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2486 sample with the timestamp @var{start} will be the first sample in the output.
2489 Specify time of the first audio sample that will be dropped, i.e. the
2490 audio sample immediately preceding the one with the timestamp @var{end} will be
2491 the last sample in the output.
2494 Same as @var{start}, except this option sets the start timestamp in samples
2498 Same as @var{end}, except this option sets the end timestamp in samples instead
2502 The maximum duration of the output in seconds.
2505 The number of the first sample that should be output.
2508 The number of the first sample that should be dropped.
2511 @option{start}, @option{end}, and @option{duration} are expressed as time
2512 duration specifications; see
2513 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2515 Note that the first two sets of the start/end options and the @option{duration}
2516 option look at the frame timestamp, while the _sample options simply count the
2517 samples that pass through the filter. So start/end_pts and start/end_sample will
2518 give different results when the timestamps are wrong, inexact or do not start at
2519 zero. Also note that this filter does not modify the timestamps. If you wish
2520 to have the output timestamps start at zero, insert the asetpts filter after the
2523 If multiple start or end options are set, this filter tries to be greedy and
2524 keep all samples that match at least one of the specified constraints. To keep
2525 only the part that matches all the constraints at once, chain multiple atrim
2528 The defaults are such that all the input is kept. So it is possible to set e.g.
2529 just the end values to keep everything before the specified time.
2534 Drop everything except the second minute of input:
2536 ffmpeg -i INPUT -af atrim=60:120
2540 Keep only the first 1000 samples:
2542 ffmpeg -i INPUT -af atrim=end_sample=1000
2547 @section axcorrelate
2548 Calculate normalized cross-correlation between two input audio streams.
2550 Resulted samples are always between -1 and 1 inclusive.
2551 If result is 1 it means two input samples are highly correlated in that selected segment.
2552 Result 0 means they are not correlated at all.
2553 If result is -1 it means two input samples are out of phase, which means they cancel each
2556 The filter accepts the following options:
2560 Set size of segment over which cross-correlation is calculated.
2561 Default is 256. Allowed range is from 2 to 131072.
2564 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2565 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2566 are always zero and thus need much less calculations to make.
2567 This is generally not true, but is valid for typical audio streams.
2570 @subsection Examples
2574 Calculate correlation between channels in stereo audio stream:
2576 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2582 Apply a two-pole Butterworth band-pass filter with central
2583 frequency @var{frequency}, and (3dB-point) band-width width.
2584 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2585 instead of the default: constant 0dB peak gain.
2586 The filter roll off at 6dB per octave (20dB per decade).
2588 The filter accepts the following options:
2592 Set the filter's central frequency. Default is @code{3000}.
2595 Constant skirt gain if set to 1. Defaults to 0.
2598 Set method to specify band-width of filter.
2613 Specify the band-width of a filter in width_type units.
2616 How much to use filtered signal in output. Default is 1.
2617 Range is between 0 and 1.
2620 Specify which channels to filter, by default all available are filtered.
2623 Normalize biquad coefficients, by default is disabled.
2624 Enabling it will normalize magnitude response at DC to 0dB.
2627 @subsection Commands
2629 This filter supports the following commands:
2632 Change bandpass frequency.
2633 Syntax for the command is : "@var{frequency}"
2636 Change bandpass width_type.
2637 Syntax for the command is : "@var{width_type}"
2640 Change bandpass width.
2641 Syntax for the command is : "@var{width}"
2644 Change bandpass mix.
2645 Syntax for the command is : "@var{mix}"
2650 Apply a two-pole Butterworth band-reject filter with central
2651 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2652 The filter roll off at 6dB per octave (20dB per decade).
2654 The filter accepts the following options:
2658 Set the filter's central frequency. Default is @code{3000}.
2661 Set method to specify band-width of filter.
2676 Specify the band-width of a filter in width_type units.
2679 How much to use filtered signal in output. Default is 1.
2680 Range is between 0 and 1.
2683 Specify which channels to filter, by default all available are filtered.
2686 Normalize biquad coefficients, by default is disabled.
2687 Enabling it will normalize magnitude response at DC to 0dB.
2690 @subsection Commands
2692 This filter supports the following commands:
2695 Change bandreject frequency.
2696 Syntax for the command is : "@var{frequency}"
2699 Change bandreject width_type.
2700 Syntax for the command is : "@var{width_type}"
2703 Change bandreject width.
2704 Syntax for the command is : "@var{width}"
2707 Change bandreject mix.
2708 Syntax for the command is : "@var{mix}"
2711 @section bass, lowshelf
2713 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2714 shelving filter with a response similar to that of a standard
2715 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2717 The filter accepts the following options:
2721 Give the gain at 0 Hz. Its useful range is about -20
2722 (for a large cut) to +20 (for a large boost).
2723 Beware of clipping when using a positive gain.
2726 Set the filter's central frequency and so can be used
2727 to extend or reduce the frequency range to be boosted or cut.
2728 The default value is @code{100} Hz.
2731 Set method to specify band-width of filter.
2746 Determine how steep is the filter's shelf transition.
2749 How much to use filtered signal in output. Default is 1.
2750 Range is between 0 and 1.
2753 Specify which channels to filter, by default all available are filtered.
2756 Normalize biquad coefficients, by default is disabled.
2757 Enabling it will normalize magnitude response at DC to 0dB.
2760 @subsection Commands
2762 This filter supports the following commands:
2765 Change bass frequency.
2766 Syntax for the command is : "@var{frequency}"
2769 Change bass width_type.
2770 Syntax for the command is : "@var{width_type}"
2774 Syntax for the command is : "@var{width}"
2778 Syntax for the command is : "@var{gain}"
2782 Syntax for the command is : "@var{mix}"
2787 Apply a biquad IIR filter with the given coefficients.
2788 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
2789 are the numerator and denominator coefficients respectively.
2790 and @var{channels}, @var{c} specify which channels to filter, by default all
2791 available are filtered.
2793 @subsection Commands
2795 This filter supports the following commands:
2803 Change biquad parameter.
2804 Syntax for the command is : "@var{value}"
2807 How much to use filtered signal in output. Default is 1.
2808 Range is between 0 and 1.
2811 Specify which channels to filter, by default all available are filtered.
2814 Normalize biquad coefficients, by default is disabled.
2815 Enabling it will normalize magnitude response at DC to 0dB.
2819 Bauer stereo to binaural transformation, which improves headphone listening of
2820 stereo audio records.
2822 To enable compilation of this filter you need to configure FFmpeg with
2823 @code{--enable-libbs2b}.
2825 It accepts the following parameters:
2829 Pre-defined crossfeed level.
2833 Default level (fcut=700, feed=50).
2836 Chu Moy circuit (fcut=700, feed=60).
2839 Jan Meier circuit (fcut=650, feed=95).
2844 Cut frequency (in Hz).
2853 Remap input channels to new locations.
2855 It accepts the following parameters:
2858 Map channels from input to output. The argument is a '|'-separated list of
2859 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
2860 @var{in_channel} form. @var{in_channel} can be either the name of the input
2861 channel (e.g. FL for front left) or its index in the input channel layout.
2862 @var{out_channel} is the name of the output channel or its index in the output
2863 channel layout. If @var{out_channel} is not given then it is implicitly an
2864 index, starting with zero and increasing by one for each mapping.
2866 @item channel_layout
2867 The channel layout of the output stream.
2870 If no mapping is present, the filter will implicitly map input channels to
2871 output channels, preserving indices.
2873 @subsection Examples
2877 For example, assuming a 5.1+downmix input MOV file,
2879 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
2881 will create an output WAV file tagged as stereo from the downmix channels of
2885 To fix a 5.1 WAV improperly encoded in AAC's native channel order
2887 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
2891 @section channelsplit
2893 Split each channel from an input audio stream into a separate output stream.
2895 It accepts the following parameters:
2897 @item channel_layout
2898 The channel layout of the input stream. The default is "stereo".
2900 A channel layout describing the channels to be extracted as separate output streams
2901 or "all" to extract each input channel as a separate stream. The default is "all".
2903 Choosing channels not present in channel layout in the input will result in an error.
2906 @subsection Examples
2910 For example, assuming a stereo input MP3 file,
2912 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
2914 will create an output Matroska file with two audio streams, one containing only
2915 the left channel and the other the right channel.
2918 Split a 5.1 WAV file into per-channel files:
2920 ffmpeg -i in.wav -filter_complex
2921 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
2922 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
2923 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
2928 Extract only LFE from a 5.1 WAV file:
2930 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
2931 -map '[LFE]' lfe.wav
2936 Add a chorus effect to the audio.
2938 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
2940 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
2941 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
2942 The modulation depth defines the range the modulated delay is played before or after
2943 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
2944 sound tuned around the original one, like in a chorus where some vocals are slightly
2947 It accepts the following parameters:
2950 Set input gain. Default is 0.4.
2953 Set output gain. Default is 0.4.
2956 Set delays. A typical delay is around 40ms to 60ms.
2968 @subsection Examples
2974 chorus=0.7:0.9:55:0.4:0.25:2
2980 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
2984 Fuller sounding chorus with three delays:
2986 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
2991 Compress or expand the audio's dynamic range.
2993 It accepts the following parameters:
2999 A list of times in seconds for each channel over which the instantaneous level
3000 of the input signal is averaged to determine its volume. @var{attacks} refers to
3001 increase of volume and @var{decays} refers to decrease of volume. For most
3002 situations, the attack time (response to the audio getting louder) should be
3003 shorter than the decay time, because the human ear is more sensitive to sudden
3004 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3005 a typical value for decay is 0.8 seconds.
3006 If specified number of attacks & decays is lower than number of channels, the last
3007 set attack/decay will be used for all remaining channels.
3010 A list of points for the transfer function, specified in dB relative to the
3011 maximum possible signal amplitude. Each key points list must be defined using
3012 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3013 @code{x0/y0 x1/y1 x2/y2 ....}
3015 The input values must be in strictly increasing order but the transfer function
3016 does not have to be monotonically rising. The point @code{0/0} is assumed but
3017 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3018 function are @code{-70/-70|-60/-20|1/0}.
3021 Set the curve radius in dB for all joints. It defaults to 0.01.
3024 Set the additional gain in dB to be applied at all points on the transfer
3025 function. This allows for easy adjustment of the overall gain.
3029 Set an initial volume, in dB, to be assumed for each channel when filtering
3030 starts. This permits the user to supply a nominal level initially, so that, for
3031 example, a very large gain is not applied to initial signal levels before the
3032 companding has begun to operate. A typical value for audio which is initially
3033 quiet is -90 dB. It defaults to 0.
3036 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3037 delayed before being fed to the volume adjuster. Specifying a delay
3038 approximately equal to the attack/decay times allows the filter to effectively
3039 operate in predictive rather than reactive mode. It defaults to 0.
3043 @subsection Examples
3047 Make music with both quiet and loud passages suitable for listening to in a
3050 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3053 Another example for audio with whisper and explosion parts:
3055 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3059 A noise gate for when the noise is at a lower level than the signal:
3061 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3065 Here is another noise gate, this time for when the noise is at a higher level
3066 than the signal (making it, in some ways, similar to squelch):
3068 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3072 2:1 compression starting at -6dB:
3074 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3078 2:1 compression starting at -9dB:
3080 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3084 2:1 compression starting at -12dB:
3086 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3090 2:1 compression starting at -18dB:
3092 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3096 3:1 compression starting at -15dB:
3098 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3104 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3110 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
3114 Hard limiter at -6dB:
3116 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3120 Hard limiter at -12dB:
3122 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3126 Hard noise gate at -35 dB:
3128 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3134 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3138 @section compensationdelay
3140 Compensation Delay Line is a metric based delay to compensate differing
3141 positions of microphones or speakers.
3143 For example, you have recorded guitar with two microphones placed in
3144 different locations. Because the front of sound wave has fixed speed in
3145 normal conditions, the phasing of microphones can vary and depends on
3146 their location and interposition. The best sound mix can be achieved when
3147 these microphones are in phase (synchronized). Note that a distance of
3148 ~30 cm between microphones makes one microphone capture the signal in
3149 antiphase to the other microphone. That makes the final mix sound moody.
3150 This filter helps to solve phasing problems by adding different delays
3151 to each microphone track and make them synchronized.
3153 The best result can be reached when you take one track as base and
3154 synchronize other tracks one by one with it.
3155 Remember that synchronization/delay tolerance depends on sample rate, too.
3156 Higher sample rates will give more tolerance.
3158 The filter accepts the following parameters:
3162 Set millimeters distance. This is compensation distance for fine tuning.
3166 Set cm distance. This is compensation distance for tightening distance setup.
3170 Set meters distance. This is compensation distance for hard distance setup.
3174 Set dry amount. Amount of unprocessed (dry) signal.
3178 Set wet amount. Amount of processed (wet) signal.
3182 Set temperature in degrees Celsius. This is the temperature of the environment.
3187 Apply headphone crossfeed filter.
3189 Crossfeed is the process of blending the left and right channels of stereo
3191 It is mainly used to reduce extreme stereo separation of low frequencies.
3193 The intent is to produce more speaker like sound to the listener.
3195 The filter accepts the following options:
3199 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3200 This sets gain of low shelf filter for side part of stereo image.
3201 Default is -6dB. Max allowed is -30db when strength is set to 1.
3204 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3205 This sets cut off frequency of low shelf filter. Default is cut off near
3206 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3209 Set input gain. Default is 0.9.
3212 Set output gain. Default is 1.
3215 @section crystalizer
3216 Simple algorithm to expand audio dynamic range.
3218 The filter accepts the following options:
3222 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3223 (unchanged sound) to 10.0 (maximum effect).
3226 Enable clipping. By default is enabled.
3229 @subsection Commands
3231 This filter supports the all above options as @ref{commands}.
3234 Apply a DC shift to the audio.
3236 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3237 in the recording chain) from the audio. The effect of a DC offset is reduced
3238 headroom and hence volume. The @ref{astats} filter can be used to determine if
3239 a signal has a DC offset.
3243 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3247 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3248 used to prevent clipping.
3253 Apply de-essing to the audio samples.
3257 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3261 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3265 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3269 Set the output mode.
3271 It accepts the following values:
3274 Pass input unchanged.
3277 Pass ess filtered out.
3282 Default value is @var{o}.
3288 Measure audio dynamic range.
3290 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3291 is found in transition material. And anything less that 8 have very poor dynamics
3292 and is very compressed.
3294 The filter accepts the following options:
3298 Set window length in seconds used to split audio into segments of equal length.
3299 Default is 3 seconds.
3303 Dynamic Audio Normalizer.
3305 This filter applies a certain amount of gain to the input audio in order
3306 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3307 contrast to more "simple" normalization algorithms, the Dynamic Audio
3308 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3309 This allows for applying extra gain to the "quiet" sections of the audio
3310 while avoiding distortions or clipping the "loud" sections. In other words:
3311 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3312 sections, in the sense that the volume of each section is brought to the
3313 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3314 this goal *without* applying "dynamic range compressing". It will retain 100%
3315 of the dynamic range *within* each section of the audio file.
3319 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3320 Default is 500 milliseconds.
3321 The Dynamic Audio Normalizer processes the input audio in small chunks,
3322 referred to as frames. This is required, because a peak magnitude has no
3323 meaning for just a single sample value. Instead, we need to determine the
3324 peak magnitude for a contiguous sequence of sample values. While a "standard"
3325 normalizer would simply use the peak magnitude of the complete file, the
3326 Dynamic Audio Normalizer determines the peak magnitude individually for each
3327 frame. The length of a frame is specified in milliseconds. By default, the
3328 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3329 been found to give good results with most files.
3330 Note that the exact frame length, in number of samples, will be determined
3331 automatically, based on the sampling rate of the individual input audio file.
3334 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3335 number. Default is 31.
3336 Probably the most important parameter of the Dynamic Audio Normalizer is the
3337 @code{window size} of the Gaussian smoothing filter. The filter's window size
3338 is specified in frames, centered around the current frame. For the sake of
3339 simplicity, this must be an odd number. Consequently, the default value of 31
3340 takes into account the current frame, as well as the 15 preceding frames and
3341 the 15 subsequent frames. Using a larger window results in a stronger
3342 smoothing effect and thus in less gain variation, i.e. slower gain
3343 adaptation. Conversely, using a smaller window results in a weaker smoothing
3344 effect and thus in more gain variation, i.e. faster gain adaptation.
3345 In other words, the more you increase this value, the more the Dynamic Audio
3346 Normalizer will behave like a "traditional" normalization filter. On the
3347 contrary, the more you decrease this value, the more the Dynamic Audio
3348 Normalizer will behave like a dynamic range compressor.
3351 Set the target peak value. This specifies the highest permissible magnitude
3352 level for the normalized audio input. This filter will try to approach the
3353 target peak magnitude as closely as possible, but at the same time it also
3354 makes sure that the normalized signal will never exceed the peak magnitude.
3355 A frame's maximum local gain factor is imposed directly by the target peak
3356 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3357 It is not recommended to go above this value.
3360 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3361 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3362 factor for each input frame, i.e. the maximum gain factor that does not
3363 result in clipping or distortion. The maximum gain factor is determined by
3364 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3365 additionally bounds the frame's maximum gain factor by a predetermined
3366 (global) maximum gain factor. This is done in order to avoid excessive gain
3367 factors in "silent" or almost silent frames. By default, the maximum gain
3368 factor is 10.0, For most inputs the default value should be sufficient and
3369 it usually is not recommended to increase this value. Though, for input
3370 with an extremely low overall volume level, it may be necessary to allow even
3371 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3372 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3373 Instead, a "sigmoid" threshold function will be applied. This way, the
3374 gain factors will smoothly approach the threshold value, but never exceed that
3378 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3379 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3380 This means that the maximum local gain factor for each frame is defined
3381 (only) by the frame's highest magnitude sample. This way, the samples can
3382 be amplified as much as possible without exceeding the maximum signal
3383 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3384 Normalizer can also take into account the frame's root mean square,
3385 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3386 determine the power of a time-varying signal. It is therefore considered
3387 that the RMS is a better approximation of the "perceived loudness" than
3388 just looking at the signal's peak magnitude. Consequently, by adjusting all
3389 frames to a constant RMS value, a uniform "perceived loudness" can be
3390 established. If a target RMS value has been specified, a frame's local gain
3391 factor is defined as the factor that would result in exactly that RMS value.
3392 Note, however, that the maximum local gain factor is still restricted by the
3393 frame's highest magnitude sample, in order to prevent clipping.
3396 Enable channels coupling. By default is enabled.
3397 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3398 amount. This means the same gain factor will be applied to all channels, i.e.
3399 the maximum possible gain factor is determined by the "loudest" channel.
3400 However, in some recordings, it may happen that the volume of the different
3401 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3402 In this case, this option can be used to disable the channel coupling. This way,
3403 the gain factor will be determined independently for each channel, depending
3404 only on the individual channel's highest magnitude sample. This allows for
3405 harmonizing the volume of the different channels.
3408 Enable DC bias correction. By default is disabled.
3409 An audio signal (in the time domain) is a sequence of sample values.
3410 In the Dynamic Audio Normalizer these sample values are represented in the
3411 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3412 audio signal, or "waveform", should be centered around the zero point.
3413 That means if we calculate the mean value of all samples in a file, or in a
3414 single frame, then the result should be 0.0 or at least very close to that
3415 value. If, however, there is a significant deviation of the mean value from
3416 0.0, in either positive or negative direction, this is referred to as a
3417 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3418 Audio Normalizer provides optional DC bias correction.
3419 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3420 the mean value, or "DC correction" offset, of each input frame and subtract
3421 that value from all of the frame's sample values which ensures those samples
3422 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3423 boundaries, the DC correction offset values will be interpolated smoothly
3424 between neighbouring frames.
3426 @item altboundary, b
3427 Enable alternative boundary mode. By default is disabled.
3428 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3429 around each frame. This includes the preceding frames as well as the
3430 subsequent frames. However, for the "boundary" frames, located at the very
3431 beginning and at the very end of the audio file, not all neighbouring
3432 frames are available. In particular, for the first few frames in the audio
3433 file, the preceding frames are not known. And, similarly, for the last few
3434 frames in the audio file, the subsequent frames are not known. Thus, the
3435 question arises which gain factors should be assumed for the missing frames
3436 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3437 to deal with this situation. The default boundary mode assumes a gain factor
3438 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3439 "fade out" at the beginning and at the end of the input, respectively.
3442 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3443 By default, the Dynamic Audio Normalizer does not apply "traditional"
3444 compression. This means that signal peaks will not be pruned and thus the
3445 full dynamic range will be retained within each local neighbourhood. However,
3446 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3447 normalization algorithm with a more "traditional" compression.
3448 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3449 (thresholding) function. If (and only if) the compression feature is enabled,
3450 all input frames will be processed by a soft knee thresholding function prior
3451 to the actual normalization process. Put simply, the thresholding function is
3452 going to prune all samples whose magnitude exceeds a certain threshold value.
3453 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3454 value. Instead, the threshold value will be adjusted for each individual
3456 In general, smaller parameters result in stronger compression, and vice versa.
3457 Values below 3.0 are not recommended, because audible distortion may appear.
3460 Set the target threshold value. This specifies the lowest permissible
3461 magnitude level for the audio input which will be normalized.
3462 If input frame volume is above this value frame will be normalized.
3463 Otherwise frame may not be normalized at all. The default value is set
3464 to 0, which means all input frames will be normalized.
3465 This option is mostly useful if digital noise is not wanted to be amplified.
3468 @subsection Commands
3470 This filter supports the all above options as @ref{commands}.
3474 Make audio easier to listen to on headphones.
3476 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3477 so that when listened to on headphones the stereo image is moved from
3478 inside your head (standard for headphones) to outside and in front of
3479 the listener (standard for speakers).
3485 Apply a two-pole peaking equalisation (EQ) filter. With this
3486 filter, the signal-level at and around a selected frequency can
3487 be increased or decreased, whilst (unlike bandpass and bandreject
3488 filters) that at all other frequencies is unchanged.
3490 In order to produce complex equalisation curves, this filter can
3491 be given several times, each with a different central frequency.
3493 The filter accepts the following options:
3497 Set the filter's central frequency in Hz.
3500 Set method to specify band-width of filter.
3515 Specify the band-width of a filter in width_type units.
3518 Set the required gain or attenuation in dB.
3519 Beware of clipping when using a positive gain.
3522 How much to use filtered signal in output. Default is 1.
3523 Range is between 0 and 1.
3526 Specify which channels to filter, by default all available are filtered.
3529 Normalize biquad coefficients, by default is disabled.
3530 Enabling it will normalize magnitude response at DC to 0dB.
3533 @subsection Examples
3536 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3538 equalizer=f=1000:t=h:width=200:g=-10
3542 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3544 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3548 @subsection Commands
3550 This filter supports the following commands:
3553 Change equalizer frequency.
3554 Syntax for the command is : "@var{frequency}"
3557 Change equalizer width_type.
3558 Syntax for the command is : "@var{width_type}"
3561 Change equalizer width.
3562 Syntax for the command is : "@var{width}"
3565 Change equalizer gain.
3566 Syntax for the command is : "@var{gain}"
3569 Change equalizer mix.
3570 Syntax for the command is : "@var{mix}"
3573 @section extrastereo
3575 Linearly increases the difference between left and right channels which
3576 adds some sort of "live" effect to playback.
3578 The filter accepts the following options:
3582 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3583 (average of both channels), with 1.0 sound will be unchanged, with
3584 -1.0 left and right channels will be swapped.
3587 Enable clipping. By default is enabled.
3590 @subsection Commands
3592 This filter supports the all above options as @ref{commands}.
3594 @section firequalizer
3595 Apply FIR Equalization using arbitrary frequency response.
3597 The filter accepts the following option:
3601 Set gain curve equation (in dB). The expression can contain variables:
3604 the evaluated frequency
3608 channel number, set to 0 when multichannels evaluation is disabled
3610 channel id, see libavutil/channel_layout.h, set to the first channel id when
3611 multichannels evaluation is disabled
3615 channel_layout, see libavutil/channel_layout.h
3620 @item gain_interpolate(f)
3621 interpolate gain on frequency f based on gain_entry
3622 @item cubic_interpolate(f)
3623 same as gain_interpolate, but smoother
3625 This option is also available as command. Default is @code{gain_interpolate(f)}.
3628 Set gain entry for gain_interpolate function. The expression can
3632 store gain entry at frequency f with value g
3634 This option is also available as command.
3637 Set filter delay in seconds. Higher value means more accurate.
3638 Default is @code{0.01}.
3641 Set filter accuracy in Hz. Lower value means more accurate.
3642 Default is @code{5}.
3645 Set window function. Acceptable values are:
3648 rectangular window, useful when gain curve is already smooth
3650 hann window (default)
3656 3-terms continuous 1st derivative nuttall window
3658 minimum 3-terms discontinuous nuttall window
3660 4-terms continuous 1st derivative nuttall window
3662 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3664 blackman-harris window
3670 If enabled, use fixed number of audio samples. This improves speed when
3671 filtering with large delay. Default is disabled.
3674 Enable multichannels evaluation on gain. Default is disabled.
3677 Enable zero phase mode by subtracting timestamp to compensate delay.
3678 Default is disabled.
3681 Set scale used by gain. Acceptable values are:
3684 linear frequency, linear gain
3686 linear frequency, logarithmic (in dB) gain (default)
3688 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3690 logarithmic frequency, logarithmic gain
3694 Set file for dumping, suitable for gnuplot.
3697 Set scale for dumpfile. Acceptable values are same with scale option.
3701 Enable 2-channel convolution using complex FFT. This improves speed significantly.
3702 Default is disabled.
3705 Enable minimum phase impulse response. Default is disabled.
3708 @subsection Examples
3713 firequalizer=gain='if(lt(f,1000), 0, -INF)'
3716 lowpass at 1000 Hz with gain_entry:
3718 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
3721 custom equalization:
3723 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
3726 higher delay with zero phase to compensate delay:
3728 firequalizer=delay=0.1:fixed=on:zero_phase=on
3731 lowpass on left channel, highpass on right channel:
3733 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
3734 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
3739 Apply a flanging effect to the audio.
3741 The filter accepts the following options:
3745 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
3748 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
3751 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
3755 Set percentage of delayed signal mixed with original. Range from 0 to 100.
3756 Default value is 71.
3759 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
3762 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
3763 Default value is @var{sinusoidal}.
3766 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
3767 Default value is 25.
3770 Set delay-line interpolation, @var{linear} or @var{quadratic}.
3771 Default is @var{linear}.
3775 Apply Haas effect to audio.
3777 Note that this makes most sense to apply on mono signals.
3778 With this filter applied to mono signals it give some directionality and
3779 stretches its stereo image.
3781 The filter accepts the following options:
3785 Set input level. By default is @var{1}, or 0dB
3788 Set output level. By default is @var{1}, or 0dB.
3791 Set gain applied to side part of signal. By default is @var{1}.
3794 Set kind of middle source. Can be one of the following:
3804 Pick middle part signal of stereo image.
3807 Pick side part signal of stereo image.
3811 Change middle phase. By default is disabled.
3814 Set left channel delay. By default is @var{2.05} milliseconds.
3817 Set left channel balance. By default is @var{-1}.
3820 Set left channel gain. By default is @var{1}.
3823 Change left phase. By default is disabled.
3826 Set right channel delay. By defaults is @var{2.12} milliseconds.
3829 Set right channel balance. By default is @var{1}.
3832 Set right channel gain. By default is @var{1}.
3835 Change right phase. By default is enabled.
3840 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
3841 embedded HDCD codes is expanded into a 20-bit PCM stream.
3843 The filter supports the Peak Extend and Low-level Gain Adjustment features
3844 of HDCD, and detects the Transient Filter flag.
3847 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
3850 When using the filter with wav, note the default encoding for wav is 16-bit,
3851 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
3852 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
3854 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
3855 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
3858 The filter accepts the following options:
3861 @item disable_autoconvert
3862 Disable any automatic format conversion or resampling in the filter graph.
3864 @item process_stereo
3865 Process the stereo channels together. If target_gain does not match between
3866 channels, consider it invalid and use the last valid target_gain.
3869 Set the code detect timer period in ms.
3872 Always extend peaks above -3dBFS even if PE isn't signaled.
3875 Replace audio with a solid tone and adjust the amplitude to signal some
3876 specific aspect of the decoding process. The output file can be loaded in
3877 an audio editor alongside the original to aid analysis.
3879 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
3886 Gain adjustment level at each sample
3888 Samples where peak extend occurs
3890 Samples where the code detect timer is active
3892 Samples where the target gain does not match between channels
3898 Apply head-related transfer functions (HRTFs) to create virtual
3899 loudspeakers around the user for binaural listening via headphones.
3900 The HRIRs are provided via additional streams, for each channel
3901 one stereo input stream is needed.
3903 The filter accepts the following options:
3907 Set mapping of input streams for convolution.
3908 The argument is a '|'-separated list of channel names in order as they
3909 are given as additional stream inputs for filter.
3910 This also specify number of input streams. Number of input streams
3911 must be not less than number of channels in first stream plus one.
3914 Set gain applied to audio. Value is in dB. Default is 0.
3917 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
3918 processing audio in time domain which is slow.
3919 @var{freq} is processing audio in frequency domain which is fast.
3920 Default is @var{freq}.
3923 Set custom gain for LFE channels. Value is in dB. Default is 0.
3926 Set size of frame in number of samples which will be processed at once.
3927 Default value is @var{1024}. Allowed range is from 1024 to 96000.
3930 Set format of hrir stream.
3931 Default value is @var{stereo}. Alternative value is @var{multich}.
3932 If value is set to @var{stereo}, number of additional streams should
3933 be greater or equal to number of input channels in first input stream.
3934 Also each additional stream should have stereo number of channels.
3935 If value is set to @var{multich}, number of additional streams should
3936 be exactly one. Also number of input channels of additional stream
3937 should be equal or greater than twice number of channels of first input
3941 @subsection Examples
3945 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3946 each amovie filter use stereo file with IR coefficients as input.
3947 The files give coefficients for each position of virtual loudspeaker:
3950 -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"
3955 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
3956 but now in @var{multich} @var{hrir} format.
3958 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"
3965 Apply a high-pass filter with 3dB point frequency.
3966 The filter can be either single-pole, or double-pole (the default).
3967 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
3969 The filter accepts the following options:
3973 Set frequency in Hz. Default is 3000.
3976 Set number of poles. Default is 2.
3979 Set method to specify band-width of filter.
3994 Specify the band-width of a filter in width_type units.
3995 Applies only to double-pole filter.
3996 The default is 0.707q and gives a Butterworth response.
3999 How much to use filtered signal in output. Default is 1.
4000 Range is between 0 and 1.
4003 Specify which channels to filter, by default all available are filtered.
4006 Normalize biquad coefficients, by default is disabled.
4007 Enabling it will normalize magnitude response at DC to 0dB.
4010 @subsection Commands
4012 This filter supports the following commands:
4015 Change highpass frequency.
4016 Syntax for the command is : "@var{frequency}"
4019 Change highpass width_type.
4020 Syntax for the command is : "@var{width_type}"
4023 Change highpass width.
4024 Syntax for the command is : "@var{width}"
4027 Change highpass mix.
4028 Syntax for the command is : "@var{mix}"
4033 Join multiple input streams into one multi-channel stream.
4035 It accepts the following parameters:
4039 The number of input streams. It defaults to 2.
4041 @item channel_layout
4042 The desired output channel layout. It defaults to stereo.
4045 Map channels from inputs to output. The argument is a '|'-separated list of
4046 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4047 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4048 can be either the name of the input channel (e.g. FL for front left) or its
4049 index in the specified input stream. @var{out_channel} is the name of the output
4053 The filter will attempt to guess the mappings when they are not specified
4054 explicitly. It does so by first trying to find an unused matching input channel
4055 and if that fails it picks the first unused input channel.
4057 Join 3 inputs (with properly set channel layouts):
4059 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4062 Build a 5.1 output from 6 single-channel streams:
4064 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4065 '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'
4071 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4073 To enable compilation of this filter you need to configure FFmpeg with
4074 @code{--enable-ladspa}.
4078 Specifies the name of LADSPA plugin library to load. If the environment
4079 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4080 each one of the directories specified by the colon separated list in
4081 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4082 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4083 @file{/usr/lib/ladspa/}.
4086 Specifies the plugin within the library. Some libraries contain only
4087 one plugin, but others contain many of them. If this is not set filter
4088 will list all available plugins within the specified library.
4091 Set the '|' separated list of controls which are zero or more floating point
4092 values that determine the behavior of the loaded plugin (for example delay,
4094 Controls need to be defined using the following syntax:
4095 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4096 @var{valuei} is the value set on the @var{i}-th control.
4097 Alternatively they can be also defined using the following syntax:
4098 @var{value0}|@var{value1}|@var{value2}|..., where
4099 @var{valuei} is the value set on the @var{i}-th control.
4100 If @option{controls} is set to @code{help}, all available controls and
4101 their valid ranges are printed.
4103 @item sample_rate, s
4104 Specify the sample rate, default to 44100. Only used if plugin have
4108 Set the number of samples per channel per each output frame, default
4109 is 1024. Only used if plugin have zero inputs.
4112 Set the minimum duration of the sourced audio. See
4113 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4114 for the accepted syntax.
4115 Note that the resulting duration may be greater than the specified duration,
4116 as the generated audio is always cut at the end of a complete frame.
4117 If not specified, or the expressed duration is negative, the audio is
4118 supposed to be generated forever.
4119 Only used if plugin have zero inputs.
4123 @subsection Examples
4127 List all available plugins within amp (LADSPA example plugin) library:
4133 List all available controls and their valid ranges for @code{vcf_notch}
4134 plugin from @code{VCF} library:
4136 ladspa=f=vcf:p=vcf_notch:c=help
4140 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4143 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4147 Add reverberation to the audio using TAP-plugins
4148 (Tom's Audio Processing plugins):
4150 ladspa=file=tap_reverb:tap_reverb
4154 Generate white noise, with 0.2 amplitude:
4156 ladspa=file=cmt:noise_source_white:c=c0=.2
4160 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4161 @code{C* Audio Plugin Suite} (CAPS) library:
4163 ladspa=file=caps:Click:c=c1=20'
4167 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4169 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4173 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4174 @code{SWH Plugins} collection:
4176 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4180 Attenuate low frequencies using Multiband EQ from Steve Harris
4181 @code{SWH Plugins} collection:
4183 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4187 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4190 ladspa=caps:Narrower
4194 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4196 ladspa=caps:White:.2
4200 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4202 ladspa=caps:Fractal:c=c1=1
4206 Dynamic volume normalization using @code{VLevel} plugin:
4208 ladspa=vlevel-ladspa:vlevel_mono
4212 @subsection Commands
4214 This filter supports the following commands:
4217 Modify the @var{N}-th control value.
4219 If the specified value is not valid, it is ignored and prior one is kept.
4224 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4225 Support for both single pass (livestreams, files) and double pass (files) modes.
4226 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4227 detect true peaks, the audio stream will be upsampled to 192 kHz.
4228 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4230 The filter accepts the following options:
4234 Set integrated loudness target.
4235 Range is -70.0 - -5.0. Default value is -24.0.
4238 Set loudness range target.
4239 Range is 1.0 - 20.0. Default value is 7.0.
4242 Set maximum true peak.
4243 Range is -9.0 - +0.0. Default value is -2.0.
4245 @item measured_I, measured_i
4246 Measured IL of input file.
4247 Range is -99.0 - +0.0.
4249 @item measured_LRA, measured_lra
4250 Measured LRA of input file.
4251 Range is 0.0 - 99.0.
4253 @item measured_TP, measured_tp
4254 Measured true peak of input file.
4255 Range is -99.0 - +99.0.
4257 @item measured_thresh
4258 Measured threshold of input file.
4259 Range is -99.0 - +0.0.
4262 Set offset gain. Gain is applied before the true-peak limiter.
4263 Range is -99.0 - +99.0. Default is +0.0.
4266 Normalize by linearly scaling the source audio.
4267 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4268 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4269 be lower than source LRA and the change in integrated loudness shouldn't
4270 result in a true peak which exceeds the target TP. If any of these
4271 conditions aren't met, normalization mode will revert to @var{dynamic}.
4272 Options are @code{true} or @code{false}. Default is @code{true}.
4275 Treat mono input files as "dual-mono". If a mono file is intended for playback
4276 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4277 If set to @code{true}, this option will compensate for this effect.
4278 Multi-channel input files are not affected by this option.
4279 Options are true or false. Default is false.
4282 Set print format for stats. Options are summary, json, or none.
4283 Default value is none.
4288 Apply a low-pass filter with 3dB point frequency.
4289 The filter can be either single-pole or double-pole (the default).
4290 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4292 The filter accepts the following options:
4296 Set frequency in Hz. Default is 500.
4299 Set number of poles. Default is 2.
4302 Set method to specify band-width of filter.
4317 Specify the band-width of a filter in width_type units.
4318 Applies only to double-pole filter.
4319 The default is 0.707q and gives a Butterworth response.
4322 How much to use filtered signal in output. Default is 1.
4323 Range is between 0 and 1.
4326 Specify which channels to filter, by default all available are filtered.
4329 Normalize biquad coefficients, by default is disabled.
4330 Enabling it will normalize magnitude response at DC to 0dB.
4333 @subsection Examples
4336 Lowpass only LFE channel, it LFE is not present it does nothing:
4342 @subsection Commands
4344 This filter supports the following commands:
4347 Change lowpass frequency.
4348 Syntax for the command is : "@var{frequency}"
4351 Change lowpass width_type.
4352 Syntax for the command is : "@var{width_type}"
4355 Change lowpass width.
4356 Syntax for the command is : "@var{width}"
4360 Syntax for the command is : "@var{mix}"
4365 Load a LV2 (LADSPA Version 2) plugin.
4367 To enable compilation of this filter you need to configure FFmpeg with
4368 @code{--enable-lv2}.
4372 Specifies the plugin URI. You may need to escape ':'.
4375 Set the '|' separated list of controls which are zero or more floating point
4376 values that determine the behavior of the loaded plugin (for example delay,
4378 If @option{controls} is set to @code{help}, all available controls and
4379 their valid ranges are printed.
4381 @item sample_rate, s
4382 Specify the sample rate, default to 44100. Only used if plugin have
4386 Set the number of samples per channel per each output frame, default
4387 is 1024. Only used if plugin have zero inputs.
4390 Set the minimum duration of the sourced audio. See
4391 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4392 for the accepted syntax.
4393 Note that the resulting duration may be greater than the specified duration,
4394 as the generated audio is always cut at the end of a complete frame.
4395 If not specified, or the expressed duration is negative, the audio is
4396 supposed to be generated forever.
4397 Only used if plugin have zero inputs.
4400 @subsection Examples
4404 Apply bass enhancer plugin from Calf:
4406 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4410 Apply vinyl plugin from Calf:
4412 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4416 Apply bit crusher plugin from ArtyFX:
4418 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4423 Multiband Compress or expand the audio's dynamic range.
4425 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4426 This is akin to the crossover of a loudspeaker, and results in flat frequency
4427 response when absent compander action.
4429 It accepts the following parameters:
4433 This option syntax is:
4434 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4435 For explanation of each item refer to compand filter documentation.
4441 Mix channels with specific gain levels. The filter accepts the output
4442 channel layout followed by a set of channels definitions.
4444 This filter is also designed to efficiently remap the channels of an audio
4447 The filter accepts parameters of the form:
4448 "@var{l}|@var{outdef}|@var{outdef}|..."
4452 output channel layout or number of channels
4455 output channel specification, of the form:
4456 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4459 output channel to define, either a channel name (FL, FR, etc.) or a channel
4460 number (c0, c1, etc.)
4463 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4466 input channel to use, see out_name for details; it is not possible to mix
4467 named and numbered input channels
4470 If the `=' in a channel specification is replaced by `<', then the gains for
4471 that specification will be renormalized so that the total is 1, thus
4472 avoiding clipping noise.
4474 @subsection Mixing examples
4476 For example, if you want to down-mix from stereo to mono, but with a bigger
4477 factor for the left channel:
4479 pan=1c|c0=0.9*c0+0.1*c1
4482 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4483 7-channels surround:
4485 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4488 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4489 that should be preferred (see "-ac" option) unless you have very specific
4492 @subsection Remapping examples
4494 The channel remapping will be effective if, and only if:
4497 @item gain coefficients are zeroes or ones,
4498 @item only one input per channel output,
4501 If all these conditions are satisfied, the filter will notify the user ("Pure
4502 channel mapping detected"), and use an optimized and lossless method to do the
4505 For example, if you have a 5.1 source and want a stereo audio stream by
4506 dropping the extra channels:
4508 pan="stereo| c0=FL | c1=FR"
4511 Given the same source, you can also switch front left and front right channels
4512 and keep the input channel layout:
4514 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4517 If the input is a stereo audio stream, you can mute the front left channel (and
4518 still keep the stereo channel layout) with:
4523 Still with a stereo audio stream input, you can copy the right channel in both
4524 front left and right:
4526 pan="stereo| c0=FR | c1=FR"
4531 ReplayGain scanner filter. This filter takes an audio stream as an input and
4532 outputs it unchanged.
4533 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4537 Convert the audio sample format, sample rate and channel layout. It is
4538 not meant to be used directly.
4541 Apply time-stretching and pitch-shifting with librubberband.
4543 To enable compilation of this filter, you need to configure FFmpeg with
4544 @code{--enable-librubberband}.
4546 The filter accepts the following options:
4550 Set tempo scale factor.
4553 Set pitch scale factor.
4556 Set transients detector.
4557 Possible values are:
4566 Possible values are:
4575 Possible values are:
4582 Set processing window size.
4583 Possible values are:
4592 Possible values are:
4599 Enable formant preservation when shift pitching.
4600 Possible values are:
4608 Possible values are:
4617 Possible values are:
4624 @subsection Commands
4626 This filter supports the following commands:
4629 Change filter tempo scale factor.
4630 Syntax for the command is : "@var{tempo}"
4633 Change filter pitch scale factor.
4634 Syntax for the command is : "@var{pitch}"
4637 @section sidechaincompress
4639 This filter acts like normal compressor but has the ability to compress
4640 detected signal using second input signal.
4641 It needs two input streams and returns one output stream.
4642 First input stream will be processed depending on second stream signal.
4643 The filtered signal then can be filtered with other filters in later stages of
4644 processing. See @ref{pan} and @ref{amerge} filter.
4646 The filter accepts the following options:
4650 Set input gain. Default is 1. Range is between 0.015625 and 64.
4653 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4654 Default is @code{downward}.
4657 If a signal of second stream raises above this level it will affect the gain
4658 reduction of first stream.
4659 By default is 0.125. Range is between 0.00097563 and 1.
4662 Set a ratio about which the signal is reduced. 1:2 means that if the level
4663 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4664 Default is 2. Range is between 1 and 20.
4667 Amount of milliseconds the signal has to rise above the threshold before gain
4668 reduction starts. Default is 20. Range is between 0.01 and 2000.
4671 Amount of milliseconds the signal has to fall below the threshold before
4672 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4675 Set the amount by how much signal will be amplified after processing.
4676 Default is 1. Range is from 1 to 64.
4679 Curve the sharp knee around the threshold to enter gain reduction more softly.
4680 Default is 2.82843. Range is between 1 and 8.
4683 Choose if the @code{average} level between all channels of side-chain stream
4684 or the louder(@code{maximum}) channel of side-chain stream affects the
4685 reduction. Default is @code{average}.
4688 Should the exact signal be taken in case of @code{peak} or an RMS one in case
4689 of @code{rms}. Default is @code{rms} which is mainly smoother.
4692 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
4695 How much to use compressed signal in output. Default is 1.
4696 Range is between 0 and 1.
4699 @subsection Commands
4701 This filter supports the all above options as @ref{commands}.
4703 @subsection Examples
4707 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
4708 depending on the signal of 2nd input and later compressed signal to be
4709 merged with 2nd input:
4711 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
4715 @section sidechaingate
4717 A sidechain gate acts like a normal (wideband) gate but has the ability to
4718 filter the detected signal before sending it to the gain reduction stage.
4719 Normally a gate uses the full range signal to detect a level above the
4721 For example: If you cut all lower frequencies from your sidechain signal
4722 the gate will decrease the volume of your track only if not enough highs
4723 appear. With this technique you are able to reduce the resonation of a
4724 natural drum or remove "rumbling" of muted strokes from a heavily distorted
4726 It needs two input streams and returns one output stream.
4727 First input stream will be processed depending on second stream signal.
4729 The filter accepts the following options:
4733 Set input level before filtering.
4734 Default is 1. Allowed range is from 0.015625 to 64.
4737 Set the mode of operation. Can be @code{upward} or @code{downward}.
4738 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
4739 will be amplified, expanding dynamic range in upward direction.
4740 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
4743 Set the level of gain reduction when the signal is below the threshold.
4744 Default is 0.06125. Allowed range is from 0 to 1.
4745 Setting this to 0 disables reduction and then filter behaves like expander.
4748 If a signal rises above this level the gain reduction is released.
4749 Default is 0.125. Allowed range is from 0 to 1.
4752 Set a ratio about which the signal is reduced.
4753 Default is 2. Allowed range is from 1 to 9000.
4756 Amount of milliseconds the signal has to rise above the threshold before gain
4758 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
4761 Amount of milliseconds the signal has to fall below the threshold before the
4762 reduction is increased again. Default is 250 milliseconds.
4763 Allowed range is from 0.01 to 9000.
4766 Set amount of amplification of signal after processing.
4767 Default is 1. Allowed range is from 1 to 64.
4770 Curve the sharp knee around the threshold to enter gain reduction more softly.
4771 Default is 2.828427125. Allowed range is from 1 to 8.
4774 Choose if exact signal should be taken for detection or an RMS like one.
4775 Default is rms. Can be peak or rms.
4778 Choose if the average level between all channels or the louder channel affects
4780 Default is average. Can be average or maximum.
4783 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
4786 @section silencedetect
4788 Detect silence in an audio stream.
4790 This filter logs a message when it detects that the input audio volume is less
4791 or equal to a noise tolerance value for a duration greater or equal to the
4792 minimum detected noise duration.
4794 The printed times and duration are expressed in seconds. The
4795 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
4796 is set on the first frame whose timestamp equals or exceeds the detection
4797 duration and it contains the timestamp of the first frame of the silence.
4799 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
4800 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
4801 keys are set on the first frame after the silence. If @option{mono} is
4802 enabled, and each channel is evaluated separately, the @code{.X}
4803 suffixed keys are used, and @code{X} corresponds to the channel number.
4805 The filter accepts the following options:
4809 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
4810 specified value) or amplitude ratio. Default is -60dB, or 0.001.
4813 Set silence duration until notification (default is 2 seconds). See
4814 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4815 for the accepted syntax.
4818 Process each channel separately, instead of combined. By default is disabled.
4821 @subsection Examples
4825 Detect 5 seconds of silence with -50dB noise tolerance:
4827 silencedetect=n=-50dB:d=5
4831 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
4832 tolerance in @file{silence.mp3}:
4834 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
4838 @section silenceremove
4840 Remove silence from the beginning, middle or end of the audio.
4842 The filter accepts the following options:
4846 This value is used to indicate if audio should be trimmed at beginning of
4847 the audio. A value of zero indicates no silence should be trimmed from the
4848 beginning. When specifying a non-zero value, it trims audio up until it
4849 finds non-silence. Normally, when trimming silence from beginning of audio
4850 the @var{start_periods} will be @code{1} but it can be increased to higher
4851 values to trim all audio up to specific count of non-silence periods.
4852 Default value is @code{0}.
4854 @item start_duration
4855 Specify the amount of time that non-silence must be detected before it stops
4856 trimming audio. By increasing the duration, bursts of noises can be treated
4857 as silence and trimmed off. Default value is @code{0}.
4859 @item start_threshold
4860 This indicates what sample value should be treated as silence. For digital
4861 audio, a value of @code{0} may be fine but for audio recorded from analog,
4862 you may wish to increase the value to account for background noise.
4863 Can be specified in dB (in case "dB" is appended to the specified value)
4864 or amplitude ratio. Default value is @code{0}.
4867 Specify max duration of silence at beginning that will be kept after
4868 trimming. Default is 0, which is equal to trimming all samples detected
4872 Specify mode of detection of silence end in start of multi-channel audio.
4873 Can be @var{any} or @var{all}. Default is @var{any}.
4874 With @var{any}, any sample that is detected as non-silence will cause
4875 stopped trimming of silence.
4876 With @var{all}, only if all channels are detected as non-silence will cause
4877 stopped trimming of silence.
4880 Set the count for trimming silence from the end of audio.
4881 To remove silence from the middle of a file, specify a @var{stop_periods}
4882 that is negative. This value is then treated as a positive value and is
4883 used to indicate the effect should restart processing as specified by
4884 @var{start_periods}, making it suitable for removing periods of silence
4885 in the middle of the audio.
4886 Default value is @code{0}.
4889 Specify a duration of silence that must exist before audio is not copied any
4890 more. By specifying a higher duration, silence that is wanted can be left in
4892 Default value is @code{0}.
4894 @item stop_threshold
4895 This is the same as @option{start_threshold} but for trimming silence from
4897 Can be specified in dB (in case "dB" is appended to the specified value)
4898 or amplitude ratio. Default value is @code{0}.
4901 Specify max duration of silence at end that will be kept after
4902 trimming. Default is 0, which is equal to trimming all samples detected
4906 Specify mode of detection of silence start in end of multi-channel audio.
4907 Can be @var{any} or @var{all}. Default is @var{any}.
4908 With @var{any}, any sample that is detected as non-silence will cause
4909 stopped trimming of silence.
4910 With @var{all}, only if all channels are detected as non-silence will cause
4911 stopped trimming of silence.
4914 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
4915 and works better with digital silence which is exactly 0.
4916 Default value is @code{rms}.
4919 Set duration in number of seconds used to calculate size of window in number
4920 of samples for detecting silence.
4921 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
4924 @subsection Examples
4928 The following example shows how this filter can be used to start a recording
4929 that does not contain the delay at the start which usually occurs between
4930 pressing the record button and the start of the performance:
4932 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
4936 Trim all silence encountered from beginning to end where there is more than 1
4937 second of silence in audio:
4939 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
4943 Trim all digital silence samples, using peak detection, from beginning to end
4944 where there is more than 0 samples of digital silence in audio and digital
4945 silence is detected in all channels at same positions in stream:
4947 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
4953 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
4954 loudspeakers around the user for binaural listening via headphones (audio
4955 formats up to 9 channels supported).
4956 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
4957 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
4958 Austrian Academy of Sciences.
4960 To enable compilation of this filter you need to configure FFmpeg with
4961 @code{--enable-libmysofa}.
4963 The filter accepts the following options:
4967 Set the SOFA file used for rendering.
4970 Set gain applied to audio. Value is in dB. Default is 0.
4973 Set rotation of virtual loudspeakers in deg. Default is 0.
4976 Set elevation of virtual speakers in deg. Default is 0.
4979 Set distance in meters between loudspeakers and the listener with near-field
4980 HRTFs. Default is 1.
4983 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4984 processing audio in time domain which is slow.
4985 @var{freq} is processing audio in frequency domain which is fast.
4986 Default is @var{freq}.
4989 Set custom positions of virtual loudspeakers. Syntax for this option is:
4990 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
4991 Each virtual loudspeaker is described with short channel name following with
4992 azimuth and elevation in degrees.
4993 Each virtual loudspeaker description is separated by '|'.
4994 For example to override front left and front right channel positions use:
4995 'speakers=FL 45 15|FR 345 15'.
4996 Descriptions with unrecognised channel names are ignored.
4999 Set custom gain for LFE channels. Value is in dB. Default is 0.
5002 Set custom frame size in number of samples. Default is 1024.
5003 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5004 is set to @var{freq}.
5007 Should all IRs be normalized upon importing SOFA file.
5008 By default is enabled.
5011 Should nearest IRs be interpolated with neighbor IRs if exact position
5012 does not match. By default is disabled.
5015 Minphase all IRs upon loading of SOFA file. By default is disabled.
5018 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5021 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5024 @subsection Examples
5028 Using ClubFritz6 sofa file:
5030 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5034 Using ClubFritz12 sofa file and bigger radius with small rotation:
5036 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5040 Similar as above but with custom speaker positions for front left, front right, back left and back right
5041 and also with custom gain:
5043 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5047 @section stereotools
5049 This filter has some handy utilities to manage stereo signals, for converting
5050 M/S stereo recordings to L/R signal while having control over the parameters
5051 or spreading the stereo image of master track.
5053 The filter accepts the following options:
5057 Set input level before filtering for both channels. Defaults is 1.
5058 Allowed range is from 0.015625 to 64.
5061 Set output level after filtering for both channels. Defaults is 1.
5062 Allowed range is from 0.015625 to 64.
5065 Set input balance between both channels. Default is 0.
5066 Allowed range is from -1 to 1.
5069 Set output balance between both channels. Default is 0.
5070 Allowed range is from -1 to 1.
5073 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5074 clipping. Disabled by default.
5077 Mute the left channel. Disabled by default.
5080 Mute the right channel. Disabled by default.
5083 Change the phase of the left channel. Disabled by default.
5086 Change the phase of the right channel. Disabled by default.
5089 Set stereo mode. Available values are:
5093 Left/Right to Left/Right, this is default.
5096 Left/Right to Mid/Side.
5099 Mid/Side to Left/Right.
5102 Left/Right to Left/Left.
5105 Left/Right to Right/Right.
5108 Left/Right to Left + Right.
5111 Left/Right to Right/Left.
5114 Mid/Side to Left/Left.
5117 Mid/Side to Right/Right.
5121 Set level of side signal. Default is 1.
5122 Allowed range is from 0.015625 to 64.
5125 Set balance of side signal. Default is 0.
5126 Allowed range is from -1 to 1.
5129 Set level of the middle signal. Default is 1.
5130 Allowed range is from 0.015625 to 64.
5133 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5136 Set stereo base between mono and inversed channels. Default is 0.
5137 Allowed range is from -1 to 1.
5140 Set delay in milliseconds how much to delay left from right channel and
5141 vice versa. Default is 0. Allowed range is from -20 to 20.
5144 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5147 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5149 @item bmode_in, bmode_out
5150 Set balance mode for balance_in/balance_out option.
5152 Can be one of the following:
5156 Classic balance mode. Attenuate one channel at time.
5157 Gain is raised up to 1.
5160 Similar as classic mode above but gain is raised up to 2.
5163 Equal power distribution, from -6dB to +6dB range.
5167 @subsection Examples
5171 Apply karaoke like effect:
5173 stereotools=mlev=0.015625
5177 Convert M/S signal to L/R:
5179 "stereotools=mode=ms>lr"
5183 @section stereowiden
5185 This filter enhance the stereo effect by suppressing signal common to both
5186 channels and by delaying the signal of left into right and vice versa,
5187 thereby widening the stereo effect.
5189 The filter accepts the following options:
5193 Time in milliseconds of the delay of left signal into right and vice versa.
5194 Default is 20 milliseconds.
5197 Amount of gain in delayed signal into right and vice versa. Gives a delay
5198 effect of left signal in right output and vice versa which gives widening
5199 effect. Default is 0.3.
5202 Cross feed of left into right with inverted phase. This helps in suppressing
5203 the mono. If the value is 1 it will cancel all the signal common to both
5204 channels. Default is 0.3.
5207 Set level of input signal of original channel. Default is 0.8.
5210 @subsection Commands
5212 This filter supports the all above options except @code{delay} as @ref{commands}.
5214 @section superequalizer
5215 Apply 18 band equalizer.
5217 The filter accepts the following options:
5224 Set 131Hz band gain.
5226 Set 185Hz band gain.
5228 Set 262Hz band gain.
5230 Set 370Hz band gain.
5232 Set 523Hz band gain.
5234 Set 740Hz band gain.
5236 Set 1047Hz band gain.
5238 Set 1480Hz band gain.
5240 Set 2093Hz band gain.
5242 Set 2960Hz band gain.
5244 Set 4186Hz band gain.
5246 Set 5920Hz band gain.
5248 Set 8372Hz band gain.
5250 Set 11840Hz band gain.
5252 Set 16744Hz band gain.
5254 Set 20000Hz band gain.
5258 Apply audio surround upmix filter.
5260 This filter allows to produce multichannel output from audio stream.
5262 The filter accepts the following options:
5266 Set output channel layout. By default, this is @var{5.1}.
5268 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5269 for the required syntax.
5272 Set input channel layout. By default, this is @var{stereo}.
5274 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5275 for the required syntax.
5278 Set input volume level. By default, this is @var{1}.
5281 Set output volume level. By default, this is @var{1}.
5284 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5287 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5290 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5293 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5294 In @var{add} mode, LFE channel is created from input audio and added to output.
5295 In @var{sub} mode, LFE channel is created from input audio and added to output but
5296 also all non-LFE output channels are subtracted with output LFE channel.
5299 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5300 Default is @var{90}.
5303 Set front center input volume. By default, this is @var{1}.
5306 Set front center output volume. By default, this is @var{1}.
5309 Set front left input volume. By default, this is @var{1}.
5312 Set front left output volume. By default, this is @var{1}.
5315 Set front right input volume. By default, this is @var{1}.
5318 Set front right output volume. By default, this is @var{1}.
5321 Set side left input volume. By default, this is @var{1}.
5324 Set side left output volume. By default, this is @var{1}.
5327 Set side right input volume. By default, this is @var{1}.
5330 Set side right output volume. By default, this is @var{1}.
5333 Set back left input volume. By default, this is @var{1}.
5336 Set back left output volume. By default, this is @var{1}.
5339 Set back right input volume. By default, this is @var{1}.
5342 Set back right output volume. By default, this is @var{1}.
5345 Set back center input volume. By default, this is @var{1}.
5348 Set back center output volume. By default, this is @var{1}.
5351 Set LFE input volume. By default, this is @var{1}.
5354 Set LFE output volume. By default, this is @var{1}.
5357 Set spread usage of stereo image across X axis for all channels.
5360 Set spread usage of stereo image across Y axis for all channels.
5362 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5363 Set spread usage of stereo image across X axis for each channel.
5365 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5366 Set spread usage of stereo image across Y axis for each channel.
5369 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5372 Set window function.
5374 It accepts the following values:
5397 Default is @code{hann}.
5400 Set window overlap. If set to 1, the recommended overlap for selected
5401 window function will be picked. Default is @code{0.5}.
5404 @section treble, highshelf
5406 Boost or cut treble (upper) frequencies of the audio using a two-pole
5407 shelving filter with a response similar to that of a standard
5408 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5410 The filter accepts the following options:
5414 Give the gain at whichever is the lower of ~22 kHz and the
5415 Nyquist frequency. Its useful range is about -20 (for a large cut)
5416 to +20 (for a large boost). Beware of clipping when using a positive gain.
5419 Set the filter's central frequency and so can be used
5420 to extend or reduce the frequency range to be boosted or cut.
5421 The default value is @code{3000} Hz.
5424 Set method to specify band-width of filter.
5439 Determine how steep is the filter's shelf transition.
5442 How much to use filtered signal in output. Default is 1.
5443 Range is between 0 and 1.
5446 Specify which channels to filter, by default all available are filtered.
5449 Normalize biquad coefficients, by default is disabled.
5450 Enabling it will normalize magnitude response at DC to 0dB.
5453 @subsection Commands
5455 This filter supports the following commands:
5458 Change treble frequency.
5459 Syntax for the command is : "@var{frequency}"
5462 Change treble width_type.
5463 Syntax for the command is : "@var{width_type}"
5466 Change treble width.
5467 Syntax for the command is : "@var{width}"
5471 Syntax for the command is : "@var{gain}"
5475 Syntax for the command is : "@var{mix}"
5480 Sinusoidal amplitude modulation.
5482 The filter accepts the following options:
5486 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5487 (20 Hz or lower) will result in a tremolo effect.
5488 This filter may also be used as a ring modulator by specifying
5489 a modulation frequency higher than 20 Hz.
5490 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5493 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5494 Default value is 0.5.
5499 Sinusoidal phase modulation.
5501 The filter accepts the following options:
5505 Modulation frequency in Hertz.
5506 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5509 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5510 Default value is 0.5.
5515 Adjust the input audio volume.
5517 It accepts the following parameters:
5521 Set audio volume expression.
5523 Output values are clipped to the maximum value.
5525 The output audio volume is given by the relation:
5527 @var{output_volume} = @var{volume} * @var{input_volume}
5530 The default value for @var{volume} is "1.0".
5533 This parameter represents the mathematical precision.
5535 It determines which input sample formats will be allowed, which affects the
5536 precision of the volume scaling.
5540 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5542 32-bit floating-point; this limits input sample format to FLT. (default)
5544 64-bit floating-point; this limits input sample format to DBL.
5548 Choose the behaviour on encountering ReplayGain side data in input frames.
5552 Remove ReplayGain side data, ignoring its contents (the default).
5555 Ignore ReplayGain side data, but leave it in the frame.
5558 Prefer the track gain, if present.
5561 Prefer the album gain, if present.
5564 @item replaygain_preamp
5565 Pre-amplification gain in dB to apply to the selected replaygain gain.
5567 Default value for @var{replaygain_preamp} is 0.0.
5569 @item replaygain_noclip
5570 Prevent clipping by limiting the gain applied.
5572 Default value for @var{replaygain_noclip} is 1.
5575 Set when the volume expression is evaluated.
5577 It accepts the following values:
5580 only evaluate expression once during the filter initialization, or
5581 when the @samp{volume} command is sent
5584 evaluate expression for each incoming frame
5587 Default value is @samp{once}.
5590 The volume expression can contain the following parameters.
5594 frame number (starting at zero)
5597 @item nb_consumed_samples
5598 number of samples consumed by the filter
5600 number of samples in the current frame
5602 original frame position in the file
5608 PTS at start of stream
5610 time at start of stream
5616 last set volume value
5619 Note that when @option{eval} is set to @samp{once} only the
5620 @var{sample_rate} and @var{tb} variables are available, all other
5621 variables will evaluate to NAN.
5623 @subsection Commands
5625 This filter supports the following commands:
5628 Modify the volume expression.
5629 The command accepts the same syntax of the corresponding option.
5631 If the specified expression is not valid, it is kept at its current
5635 @subsection Examples
5639 Halve the input audio volume:
5643 volume=volume=-6.0206dB
5646 In all the above example the named key for @option{volume} can be
5647 omitted, for example like in:
5653 Increase input audio power by 6 decibels using fixed-point precision:
5655 volume=volume=6dB:precision=fixed
5659 Fade volume after time 10 with an annihilation period of 5 seconds:
5661 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
5665 @section volumedetect
5667 Detect the volume of the input video.
5669 The filter has no parameters. The input is not modified. Statistics about
5670 the volume will be printed in the log when the input stream end is reached.
5672 In particular it will show the mean volume (root mean square), maximum
5673 volume (on a per-sample basis), and the beginning of a histogram of the
5674 registered volume values (from the maximum value to a cumulated 1/1000 of
5677 All volumes are in decibels relative to the maximum PCM value.
5679 @subsection Examples
5681 Here is an excerpt of the output:
5683 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
5684 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
5685 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
5686 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
5687 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
5688 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
5689 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
5690 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
5691 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
5697 The mean square energy is approximately -27 dB, or 10^-2.7.
5699 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
5701 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
5704 In other words, raising the volume by +4 dB does not cause any clipping,
5705 raising it by +5 dB causes clipping for 6 samples, etc.
5707 @c man end AUDIO FILTERS
5709 @chapter Audio Sources
5710 @c man begin AUDIO SOURCES
5712 Below is a description of the currently available audio sources.
5716 Buffer audio frames, and make them available to the filter chain.
5718 This source is mainly intended for a programmatic use, in particular
5719 through the interface defined in @file{libavfilter/asrc_abuffer.h}.
5721 It accepts the following parameters:
5725 The timebase which will be used for timestamps of submitted frames. It must be
5726 either a floating-point number or in @var{numerator}/@var{denominator} form.
5729 The sample rate of the incoming audio buffers.
5732 The sample format of the incoming audio buffers.
5733 Either a sample format name or its corresponding integer representation from
5734 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
5736 @item channel_layout
5737 The channel layout of the incoming audio buffers.
5738 Either a channel layout name from channel_layout_map in
5739 @file{libavutil/channel_layout.c} or its corresponding integer representation
5740 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
5743 The number of channels of the incoming audio buffers.
5744 If both @var{channels} and @var{channel_layout} are specified, then they
5749 @subsection Examples
5752 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
5755 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
5756 Since the sample format with name "s16p" corresponds to the number
5757 6 and the "stereo" channel layout corresponds to the value 0x3, this is
5760 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
5765 Generate an audio signal specified by an expression.
5767 This source accepts in input one or more expressions (one for each
5768 channel), which are evaluated and used to generate a corresponding
5771 This source accepts the following options:
5775 Set the '|'-separated expressions list for each separate channel. In case the
5776 @option{channel_layout} option is not specified, the selected channel layout
5777 depends on the number of provided expressions. Otherwise the last
5778 specified expression is applied to the remaining output channels.
5780 @item channel_layout, c
5781 Set the channel layout. The number of channels in the specified layout
5782 must be equal to the number of specified expressions.
5785 Set the minimum duration of the sourced audio. See
5786 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5787 for the accepted syntax.
5788 Note that the resulting duration may be greater than the specified
5789 duration, as the generated audio is always cut at the end of a
5792 If not specified, or the expressed duration is negative, the audio is
5793 supposed to be generated forever.
5796 Set the number of samples per channel per each output frame,
5799 @item sample_rate, s
5800 Specify the sample rate, default to 44100.
5803 Each expression in @var{exprs} can contain the following constants:
5807 number of the evaluated sample, starting from 0
5810 time of the evaluated sample expressed in seconds, starting from 0
5817 @subsection Examples
5827 Generate a sin signal with frequency of 440 Hz, set sample rate to
5830 aevalsrc="sin(440*2*PI*t):s=8000"
5834 Generate a two channels signal, specify the channel layout (Front
5835 Center + Back Center) explicitly:
5837 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
5841 Generate white noise:
5843 aevalsrc="-2+random(0)"
5847 Generate an amplitude modulated signal:
5849 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
5853 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
5855 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
5862 The null audio source, return unprocessed audio frames. It is mainly useful
5863 as a template and to be employed in analysis / debugging tools, or as
5864 the source for filters which ignore the input data (for example the sox
5867 This source accepts the following options:
5871 @item channel_layout, cl
5873 Specifies the channel layout, and can be either an integer or a string
5874 representing a channel layout. The default value of @var{channel_layout}
5877 Check the channel_layout_map definition in
5878 @file{libavutil/channel_layout.c} for the mapping between strings and
5879 channel layout values.
5881 @item sample_rate, r
5882 Specifies the sample rate, and defaults to 44100.
5885 Set the number of samples per requested frames.
5889 @subsection Examples
5893 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
5895 anullsrc=r=48000:cl=4
5899 Do the same operation with a more obvious syntax:
5901 anullsrc=r=48000:cl=mono
5905 All the parameters need to be explicitly defined.
5909 Synthesize a voice utterance using the libflite library.
5911 To enable compilation of this filter you need to configure FFmpeg with
5912 @code{--enable-libflite}.
5914 Note that versions of the flite library prior to 2.0 are not thread-safe.
5916 The filter accepts the following options:
5921 If set to 1, list the names of the available voices and exit
5922 immediately. Default value is 0.
5925 Set the maximum number of samples per frame. Default value is 512.
5928 Set the filename containing the text to speak.
5931 Set the text to speak.
5934 Set the voice to use for the speech synthesis. Default value is
5935 @code{kal}. See also the @var{list_voices} option.
5938 @subsection Examples
5942 Read from file @file{speech.txt}, and synthesize the text using the
5943 standard flite voice:
5945 flite=textfile=speech.txt
5949 Read the specified text selecting the @code{slt} voice:
5951 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5955 Input text to ffmpeg:
5957 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
5961 Make @file{ffplay} speak the specified text, using @code{flite} and
5962 the @code{lavfi} device:
5964 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
5968 For more information about libflite, check:
5969 @url{http://www.festvox.org/flite/}
5973 Generate a noise audio signal.
5975 The filter accepts the following options:
5978 @item sample_rate, r
5979 Specify the sample rate. Default value is 48000 Hz.
5982 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
5986 Specify the duration of the generated audio stream. Not specifying this option
5987 results in noise with an infinite length.
5989 @item color, colour, c
5990 Specify the color of noise. Available noise colors are white, pink, brown,
5991 blue, violet and velvet. Default color is white.
5994 Specify a value used to seed the PRNG.
5997 Set the number of samples per each output frame, default is 1024.
6000 @subsection Examples
6005 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6007 anoisesrc=d=60:c=pink:r=44100:a=0.5
6013 Generate odd-tap Hilbert transform FIR coefficients.
6015 The resulting stream can be used with @ref{afir} filter for phase-shifting
6016 the signal by 90 degrees.
6018 This is used in many matrix coding schemes and for analytic signal generation.
6019 The process is often written as a multiplication by i (or j), the imaginary unit.
6021 The filter accepts the following options:
6025 @item sample_rate, s
6026 Set sample rate, default is 44100.
6029 Set length of FIR filter, default is 22051.
6032 Set number of samples per each frame.
6035 Set window function to be used when generating FIR coefficients.
6040 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6042 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6044 The filter accepts the following options:
6047 @item sample_rate, r
6048 Set sample rate, default is 44100.
6051 Set number of samples per each frame. Default is 1024.
6054 Set high-pass frequency. Default is 0.
6057 Set low-pass frequency. Default is 0.
6058 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6059 is higher than 0 then filter will create band-pass filter coefficients,
6060 otherwise band-reject filter coefficients.
6063 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6066 Set Kaiser window beta.
6069 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6072 Enable rounding, by default is disabled.
6075 Set number of taps for high-pass filter.
6078 Set number of taps for low-pass filter.
6083 Generate an audio signal made of a sine wave with amplitude 1/8.
6085 The audio signal is bit-exact.
6087 The filter accepts the following options:
6092 Set the carrier frequency. Default is 440 Hz.
6094 @item beep_factor, b
6095 Enable a periodic beep every second with frequency @var{beep_factor} times
6096 the carrier frequency. Default is 0, meaning the beep is disabled.
6098 @item sample_rate, r
6099 Specify the sample rate, default is 44100.
6102 Specify the duration of the generated audio stream.
6104 @item samples_per_frame
6105 Set the number of samples per output frame.
6107 The expression can contain the following constants:
6111 The (sequential) number of the output audio frame, starting from 0.
6114 The PTS (Presentation TimeStamp) of the output audio frame,
6115 expressed in @var{TB} units.
6118 The PTS of the output audio frame, expressed in seconds.
6121 The timebase of the output audio frames.
6124 Default is @code{1024}.
6127 @subsection Examples
6132 Generate a simple 440 Hz sine wave:
6138 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6142 sine=frequency=220:beep_factor=4:duration=5
6146 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6149 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6153 @c man end AUDIO SOURCES
6155 @chapter Audio Sinks
6156 @c man begin AUDIO SINKS
6158 Below is a description of the currently available audio sinks.
6160 @section abuffersink
6162 Buffer audio frames, and make them available to the end of filter chain.
6164 This sink is mainly intended for programmatic use, in particular
6165 through the interface defined in @file{libavfilter/buffersink.h}
6166 or the options system.
6168 It accepts a pointer to an AVABufferSinkContext structure, which
6169 defines the incoming buffers' formats, to be passed as the opaque
6170 parameter to @code{avfilter_init_filter} for initialization.
6173 Null audio sink; do absolutely nothing with the input audio. It is
6174 mainly useful as a template and for use in analysis / debugging
6177 @c man end AUDIO SINKS
6179 @chapter Video Filters
6180 @c man begin VIDEO FILTERS
6182 When you configure your FFmpeg build, you can disable any of the
6183 existing filters using @code{--disable-filters}.
6184 The configure output will show the video filters included in your
6187 Below is a description of the currently available video filters.
6191 Mark a region of interest in a video frame.
6193 The frame data is passed through unchanged, but metadata is attached
6194 to the frame indicating regions of interest which can affect the
6195 behaviour of later encoding. Multiple regions can be marked by
6196 applying the filter multiple times.
6200 Region distance in pixels from the left edge of the frame.
6202 Region distance in pixels from the top edge of the frame.
6204 Region width in pixels.
6206 Region height in pixels.
6208 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6209 and may contain the following variables:
6212 Width of the input frame.
6214 Height of the input frame.
6218 Quantisation offset to apply within the region.
6220 This must be a real value in the range -1 to +1. A value of zero
6221 indicates no quality change. A negative value asks for better quality
6222 (less quantisation), while a positive value asks for worse quality
6223 (greater quantisation).
6225 The range is calibrated so that the extreme values indicate the
6226 largest possible offset - if the rest of the frame is encoded with the
6227 worst possible quality, an offset of -1 indicates that this region
6228 should be encoded with the best possible quality anyway. Intermediate
6229 values are then interpolated in some codec-dependent way.
6231 For example, in 10-bit H.264 the quantisation parameter varies between
6232 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6233 this region should be encoded with a QP around one-tenth of the full
6234 range better than the rest of the frame. So, if most of the frame
6235 were to be encoded with a QP of around 30, this region would get a QP
6236 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6237 An extreme value of -1 would indicate that this region should be
6238 encoded with the best possible quality regardless of the treatment of
6239 the rest of the frame - that is, should be encoded at a QP of -12.
6241 If set to true, remove any existing regions of interest marked on the
6242 frame before adding the new one.
6245 @subsection Examples
6249 Mark the centre quarter of the frame as interesting.
6251 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6254 Mark the 100-pixel-wide region on the left edge of the frame as very
6255 uninteresting (to be encoded at much lower quality than the rest of
6258 addroi=0:0:100:ih:+1/5
6262 @section alphaextract
6264 Extract the alpha component from the input as a grayscale video. This
6265 is especially useful with the @var{alphamerge} filter.
6269 Add or replace the alpha component of the primary input with the
6270 grayscale value of a second input. This is intended for use with
6271 @var{alphaextract} to allow the transmission or storage of frame
6272 sequences that have alpha in a format that doesn't support an alpha
6275 For example, to reconstruct full frames from a normal YUV-encoded video
6276 and a separate video created with @var{alphaextract}, you might use:
6278 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6281 Since this filter is designed for reconstruction, it operates on frame
6282 sequences without considering timestamps, and terminates when either
6283 input reaches end of stream. This will cause problems if your encoding
6284 pipeline drops frames. If you're trying to apply an image as an
6285 overlay to a video stream, consider the @var{overlay} filter instead.
6289 Amplify differences between current pixel and pixels of adjacent frames in
6290 same pixel location.
6292 This filter accepts the following options:
6296 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6297 For example radius of 3 will instruct filter to calculate average of 7 frames.
6300 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6303 Set threshold for difference amplification. Any difference greater or equal to
6304 this value will not alter source pixel. Default is 10.
6305 Allowed range is from 0 to 65535.
6308 Set tolerance for difference amplification. Any difference lower to
6309 this value will not alter source pixel. Default is 0.
6310 Allowed range is from 0 to 65535.
6313 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6314 This option controls maximum possible value that will decrease source pixel value.
6317 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6318 This option controls maximum possible value that will increase source pixel value.
6321 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6324 @subsection Commands
6326 This filter supports the following @ref{commands} that corresponds to option of same name:
6338 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6339 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6340 Substation Alpha) subtitles files.
6342 This filter accepts the following option in addition to the common options from
6343 the @ref{subtitles} filter:
6347 Set the shaping engine
6349 Available values are:
6352 The default libass shaping engine, which is the best available.
6354 Fast, font-agnostic shaper that can do only substitutions
6356 Slower shaper using OpenType for substitutions and positioning
6359 The default is @code{auto}.
6363 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6365 The filter accepts the following options:
6369 Set threshold A for 1st plane. Default is 0.02.
6370 Valid range is 0 to 0.3.
6373 Set threshold B for 1st plane. Default is 0.04.
6374 Valid range is 0 to 5.
6377 Set threshold A for 2nd plane. Default is 0.02.
6378 Valid range is 0 to 0.3.
6381 Set threshold B for 2nd plane. Default is 0.04.
6382 Valid range is 0 to 5.
6385 Set threshold A for 3rd plane. Default is 0.02.
6386 Valid range is 0 to 0.3.
6389 Set threshold B for 3rd plane. Default is 0.04.
6390 Valid range is 0 to 5.
6392 Threshold A is designed to react on abrupt changes in the input signal and
6393 threshold B is designed to react on continuous changes in the input signal.
6396 Set number of frames filter will use for averaging. Default is 9. Must be odd
6397 number in range [5, 129].
6400 Set what planes of frame filter will use for averaging. Default is all.
6403 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
6404 Alternatively can be set to @code{s} serial.
6406 Parallel can be faster then serial, while other way around is never true.
6407 Parallel will abort early on first change being greater then thresholds, while serial
6408 will continue processing other side of frames if they are equal or bellow thresholds.
6411 @subsection Commands
6412 This filter supports same @ref{commands} as options except option @code{s}.
6413 The command accepts the same syntax of the corresponding option.
6417 Apply average blur filter.
6419 The filter accepts the following options:
6423 Set horizontal radius size.
6426 Set which planes to filter. By default all planes are filtered.
6429 Set vertical radius size, if zero it will be same as @code{sizeX}.
6430 Default is @code{0}.
6433 @subsection Commands
6434 This filter supports same commands as options.
6435 The command accepts the same syntax of the corresponding option.
6437 If the specified expression is not valid, it is kept at its current
6442 Compute the bounding box for the non-black pixels in the input frame
6445 This filter computes the bounding box containing all the pixels with a
6446 luminance value greater than the minimum allowed value.
6447 The parameters describing the bounding box are printed on the filter
6450 The filter accepts the following option:
6454 Set the minimal luminance value. Default is @code{16}.
6458 Apply bilateral filter, spatial smoothing while preserving edges.
6460 The filter accepts the following options:
6463 Set sigma of gaussian function to calculate spatial weight.
6464 Allowed range is 0 to 10. Default is 0.1.
6467 Set sigma of gaussian function to calculate range weight.
6468 Allowed range is 0 to 1. Default is 0.1.
6471 Set planes to filter. Default is first only.
6474 @section bitplanenoise
6476 Show and measure bit plane noise.
6478 The filter accepts the following options:
6482 Set which plane to analyze. Default is @code{1}.
6485 Filter out noisy pixels from @code{bitplane} set above.
6486 Default is disabled.
6489 @section blackdetect
6491 Detect video intervals that are (almost) completely black. Can be
6492 useful to detect chapter transitions, commercials, or invalid
6493 recordings. Output lines contains the time for the start, end and
6494 duration of the detected black interval expressed in seconds.
6496 In order to display the output lines, you need to set the loglevel at
6497 least to the AV_LOG_INFO value.
6499 The filter accepts the following options:
6502 @item black_min_duration, d
6503 Set the minimum detected black duration expressed in seconds. It must
6504 be a non-negative floating point number.
6506 Default value is 2.0.
6508 @item picture_black_ratio_th, pic_th
6509 Set the threshold for considering a picture "black".
6510 Express the minimum value for the ratio:
6512 @var{nb_black_pixels} / @var{nb_pixels}
6515 for which a picture is considered black.
6516 Default value is 0.98.
6518 @item pixel_black_th, pix_th
6519 Set the threshold for considering a pixel "black".
6521 The threshold expresses the maximum pixel luminance value for which a
6522 pixel is considered "black". The provided value is scaled according to
6523 the following equation:
6525 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6528 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6529 the input video format, the range is [0-255] for YUV full-range
6530 formats and [16-235] for YUV non full-range formats.
6532 Default value is 0.10.
6535 The following example sets the maximum pixel threshold to the minimum
6536 value, and detects only black intervals of 2 or more seconds:
6538 blackdetect=d=2:pix_th=0.00
6543 Detect frames that are (almost) completely black. Can be useful to
6544 detect chapter transitions or commercials. Output lines consist of
6545 the frame number of the detected frame, the percentage of blackness,
6546 the position in the file if known or -1 and the timestamp in seconds.
6548 In order to display the output lines, you need to set the loglevel at
6549 least to the AV_LOG_INFO value.
6551 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
6552 The value represents the percentage of pixels in the picture that
6553 are below the threshold value.
6555 It accepts the following parameters:
6560 The percentage of the pixels that have to be below the threshold; it defaults to
6563 @item threshold, thresh
6564 The threshold below which a pixel value is considered black; it defaults to
6569 @section blend, tblend
6571 Blend two video frames into each other.
6573 The @code{blend} filter takes two input streams and outputs one
6574 stream, the first input is the "top" layer and second input is
6575 "bottom" layer. By default, the output terminates when the longest input terminates.
6577 The @code{tblend} (time blend) filter takes two consecutive frames
6578 from one single stream, and outputs the result obtained by blending
6579 the new frame on top of the old frame.
6581 A description of the accepted options follows.
6589 Set blend mode for specific pixel component or all pixel components in case
6590 of @var{all_mode}. Default value is @code{normal}.
6592 Available values for component modes are:
6634 Set blend opacity for specific pixel component or all pixel components in case
6635 of @var{all_opacity}. Only used in combination with pixel component blend modes.
6642 Set blend expression for specific pixel component or all pixel components in case
6643 of @var{all_expr}. Note that related mode options will be ignored if those are set.
6645 The expressions can use the following variables:
6649 The sequential number of the filtered frame, starting from @code{0}.
6653 the coordinates of the current sample
6657 the width and height of currently filtered plane
6661 Width and height scale for the plane being filtered. It is the
6662 ratio between the dimensions of the current plane to the luma plane,
6663 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
6664 the luma plane and @code{0.5,0.5} for the chroma planes.
6667 Time of the current frame, expressed in seconds.
6670 Value of pixel component at current location for first video frame (top layer).
6673 Value of pixel component at current location for second video frame (bottom layer).
6677 The @code{blend} filter also supports the @ref{framesync} options.
6679 @subsection Examples
6683 Apply transition from bottom layer to top layer in first 10 seconds:
6685 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
6689 Apply linear horizontal transition from top layer to bottom layer:
6691 blend=all_expr='A*(X/W)+B*(1-X/W)'
6695 Apply 1x1 checkerboard effect:
6697 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
6701 Apply uncover left effect:
6703 blend=all_expr='if(gte(N*SW+X,W),A,B)'
6707 Apply uncover down effect:
6709 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
6713 Apply uncover up-left effect:
6715 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
6719 Split diagonally video and shows top and bottom layer on each side:
6721 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
6725 Display differences between the current and the previous frame:
6727 tblend=all_mode=grainextract
6733 Denoise frames using Block-Matching 3D algorithm.
6735 The filter accepts the following options.
6739 Set denoising strength. Default value is 1.
6740 Allowed range is from 0 to 999.9.
6741 The denoising algorithm is very sensitive to sigma, so adjust it
6742 according to the source.
6745 Set local patch size. This sets dimensions in 2D.
6748 Set sliding step for processing blocks. Default value is 4.
6749 Allowed range is from 1 to 64.
6750 Smaller values allows processing more reference blocks and is slower.
6753 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
6754 When set to 1, no block matching is done. Larger values allows more blocks
6756 Allowed range is from 1 to 256.
6759 Set radius for search block matching. Default is 9.
6760 Allowed range is from 1 to INT32_MAX.
6763 Set step between two search locations for block matching. Default is 1.
6764 Allowed range is from 1 to 64. Smaller is slower.
6767 Set threshold of mean square error for block matching. Valid range is 0 to
6771 Set thresholding parameter for hard thresholding in 3D transformed domain.
6772 Larger values results in stronger hard-thresholding filtering in frequency
6776 Set filtering estimation mode. Can be @code{basic} or @code{final}.
6777 Default is @code{basic}.
6780 If enabled, filter will use 2nd stream for block matching.
6781 Default is disabled for @code{basic} value of @var{estim} option,
6782 and always enabled if value of @var{estim} is @code{final}.
6785 Set planes to filter. Default is all available except alpha.
6788 @subsection Examples
6792 Basic filtering with bm3d:
6794 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
6798 Same as above, but filtering only luma:
6800 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
6804 Same as above, but with both estimation modes:
6806 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
6810 Same as above, but prefilter with @ref{nlmeans} filter instead:
6812 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
6818 Apply a boxblur algorithm to the input video.
6820 It accepts the following parameters:
6824 @item luma_radius, lr
6825 @item luma_power, lp
6826 @item chroma_radius, cr
6827 @item chroma_power, cp
6828 @item alpha_radius, ar
6829 @item alpha_power, ap
6833 A description of the accepted options follows.
6836 @item luma_radius, lr
6837 @item chroma_radius, cr
6838 @item alpha_radius, ar
6839 Set an expression for the box radius in pixels used for blurring the
6840 corresponding input plane.
6842 The radius value must be a non-negative number, and must not be
6843 greater than the value of the expression @code{min(w,h)/2} for the
6844 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
6847 Default value for @option{luma_radius} is "2". If not specified,
6848 @option{chroma_radius} and @option{alpha_radius} default to the
6849 corresponding value set for @option{luma_radius}.
6851 The expressions can contain the following constants:
6855 The input width and height in pixels.
6859 The input chroma image width and height in pixels.
6863 The horizontal and vertical chroma subsample values. For example, for the
6864 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
6867 @item luma_power, lp
6868 @item chroma_power, cp
6869 @item alpha_power, ap
6870 Specify how many times the boxblur filter is applied to the
6871 corresponding plane.
6873 Default value for @option{luma_power} is 2. If not specified,
6874 @option{chroma_power} and @option{alpha_power} default to the
6875 corresponding value set for @option{luma_power}.
6877 A value of 0 will disable the effect.
6880 @subsection Examples
6884 Apply a boxblur filter with the luma, chroma, and alpha radii
6887 boxblur=luma_radius=2:luma_power=1
6892 Set the luma radius to 2, and alpha and chroma radius to 0:
6894 boxblur=2:1:cr=0:ar=0
6898 Set the luma and chroma radii to a fraction of the video dimension:
6900 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
6906 Deinterlace the input video ("bwdif" stands for "Bob Weaver
6907 Deinterlacing Filter").
6909 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
6910 interpolation algorithms.
6911 It accepts the following parameters:
6915 The interlacing mode to adopt. It accepts one of the following values:
6919 Output one frame for each frame.
6921 Output one frame for each field.
6924 The default value is @code{send_field}.
6927 The picture field parity assumed for the input interlaced video. It accepts one
6928 of the following values:
6932 Assume the top field is first.
6934 Assume the bottom field is first.
6936 Enable automatic detection of field parity.
6939 The default value is @code{auto}.
6940 If the interlacing is unknown or the decoder does not export this information,
6941 top field first will be assumed.
6944 Specify which frames to deinterlace. Accepts one of the following
6949 Deinterlace all frames.
6951 Only deinterlace frames marked as interlaced.
6954 The default value is @code{all}.
6958 Remove all color information for all colors except for certain one.
6960 The filter accepts the following options:
6964 The color which will not be replaced with neutral chroma.
6967 Similarity percentage with the above color.
6968 0.01 matches only the exact key color, while 1.0 matches everything.
6972 0.0 makes pixels either fully gray, or not gray at all.
6973 Higher values result in more preserved color.
6976 Signals that the color passed is already in YUV instead of RGB.
6978 Literal colors like "green" or "red" don't make sense with this enabled anymore.
6979 This can be used to pass exact YUV values as hexadecimal numbers.
6982 @subsection Commands
6983 This filter supports same @ref{commands} as options.
6984 The command accepts the same syntax of the corresponding option.
6986 If the specified expression is not valid, it is kept at its current
6990 YUV colorspace color/chroma keying.
6992 The filter accepts the following options:
6996 The color which will be replaced with transparency.
6999 Similarity percentage with the key color.
7001 0.01 matches only the exact key color, while 1.0 matches everything.
7006 0.0 makes pixels either fully transparent, or not transparent at all.
7008 Higher values result in semi-transparent pixels, with a higher transparency
7009 the more similar the pixels color is to the key color.
7012 Signals that the color passed is already in YUV instead of RGB.
7014 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7015 This can be used to pass exact YUV values as hexadecimal numbers.
7018 @subsection Commands
7019 This filter supports same @ref{commands} as options.
7020 The command accepts the same syntax of the corresponding option.
7022 If the specified expression is not valid, it is kept at its current
7025 @subsection Examples
7029 Make every green pixel in the input image transparent:
7031 ffmpeg -i input.png -vf chromakey=green out.png
7035 Overlay a greenscreen-video on top of a static black background.
7037 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
7041 @section chromashift
7042 Shift chroma pixels horizontally and/or vertically.
7044 The filter accepts the following options:
7047 Set amount to shift chroma-blue horizontally.
7049 Set amount to shift chroma-blue vertically.
7051 Set amount to shift chroma-red horizontally.
7053 Set amount to shift chroma-red vertically.
7055 Set edge mode, can be @var{smear}, default, or @var{warp}.
7058 @subsection Commands
7060 This filter supports the all above options as @ref{commands}.
7064 Display CIE color diagram with pixels overlaid onto it.
7066 The filter accepts the following options:
7081 @item uhdtv, rec2020
7095 Set what gamuts to draw.
7097 See @code{system} option for available values.
7100 Set ciescope size, by default set to 512.
7103 Set intensity used to map input pixel values to CIE diagram.
7106 Set contrast used to draw tongue colors that are out of active color system gamut.
7109 Correct gamma displayed on scope, by default enabled.
7112 Show white point on CIE diagram, by default disabled.
7115 Set input gamma. Used only with XYZ input color space.
7120 Visualize information exported by some codecs.
7122 Some codecs can export information through frames using side-data or other
7123 means. For example, some MPEG based codecs export motion vectors through the
7124 @var{export_mvs} flag in the codec @option{flags2} option.
7126 The filter accepts the following option:
7130 Set motion vectors to visualize.
7132 Available flags for @var{mv} are:
7136 forward predicted MVs of P-frames
7138 forward predicted MVs of B-frames
7140 backward predicted MVs of B-frames
7144 Display quantization parameters using the chroma planes.
7147 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7149 Available flags for @var{mv_type} are:
7153 forward predicted MVs
7155 backward predicted MVs
7158 @item frame_type, ft
7159 Set frame type to visualize motion vectors of.
7161 Available flags for @var{frame_type} are:
7165 intra-coded frames (I-frames)
7167 predicted frames (P-frames)
7169 bi-directionally predicted frames (B-frames)
7173 @subsection Examples
7177 Visualize forward predicted MVs of all frames using @command{ffplay}:
7179 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7183 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7185 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7189 @section colorbalance
7190 Modify intensity of primary colors (red, green and blue) of input frames.
7192 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7193 regions for the red-cyan, green-magenta or blue-yellow balance.
7195 A positive adjustment value shifts the balance towards the primary color, a negative
7196 value towards the complementary color.
7198 The filter accepts the following options:
7204 Adjust red, green and blue shadows (darkest pixels).
7209 Adjust red, green and blue midtones (medium pixels).
7214 Adjust red, green and blue highlights (brightest pixels).
7216 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7219 Preserve lightness when changing color balance. Default is disabled.
7222 @subsection Examples
7226 Add red color cast to shadows:
7232 @subsection Commands
7234 This filter supports the all above options as @ref{commands}.
7236 @section colorchannelmixer
7238 Adjust video input frames by re-mixing color channels.
7240 This filter modifies a color channel by adding the values associated to
7241 the other channels of the same pixels. For example if the value to
7242 modify is red, the output value will be:
7244 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7247 The filter accepts the following options:
7254 Adjust contribution of input red, green, blue and alpha channels for output red channel.
7255 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
7261 Adjust contribution of input red, green, blue and alpha channels for output green channel.
7262 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
7268 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
7269 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
7275 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7276 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7278 Allowed ranges for options are @code{[-2.0, 2.0]}.
7281 @subsection Examples
7285 Convert source to grayscale:
7287 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7290 Simulate sepia tones:
7292 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7296 @subsection Commands
7298 This filter supports the all above options as @ref{commands}.
7301 RGB colorspace color keying.
7303 The filter accepts the following options:
7307 The color which will be replaced with transparency.
7310 Similarity percentage with the key color.
7312 0.01 matches only the exact key color, while 1.0 matches everything.
7317 0.0 makes pixels either fully transparent, or not transparent at all.
7319 Higher values result in semi-transparent pixels, with a higher transparency
7320 the more similar the pixels color is to the key color.
7323 @subsection Examples
7327 Make every green pixel in the input image transparent:
7329 ffmpeg -i input.png -vf colorkey=green out.png
7333 Overlay a greenscreen-video on top of a static background image.
7335 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
7340 Remove all color information for all RGB colors except for certain one.
7342 The filter accepts the following options:
7346 The color which will not be replaced with neutral gray.
7349 Similarity percentage with the above color.
7350 0.01 matches only the exact key color, while 1.0 matches everything.
7353 Blend percentage. 0.0 makes pixels fully gray.
7354 Higher values result in more preserved color.
7357 @section colorlevels
7359 Adjust video input frames using levels.
7361 The filter accepts the following options:
7368 Adjust red, green, blue and alpha input black point.
7369 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7375 Adjust red, green, blue and alpha input white point.
7376 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
7378 Input levels are used to lighten highlights (bright tones), darken shadows
7379 (dark tones), change the balance of bright and dark tones.
7385 Adjust red, green, blue and alpha output black point.
7386 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
7392 Adjust red, green, blue and alpha output white point.
7393 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
7395 Output levels allows manual selection of a constrained output level range.
7398 @subsection Examples
7402 Make video output darker:
7404 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
7410 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
7414 Make video output lighter:
7416 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
7420 Increase brightness:
7422 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
7426 @section colormatrix
7428 Convert color matrix.
7430 The filter accepts the following options:
7435 Specify the source and destination color matrix. Both values must be
7438 The accepted values are:
7466 For example to convert from BT.601 to SMPTE-240M, use the command:
7468 colormatrix=bt601:smpte240m
7473 Convert colorspace, transfer characteristics or color primaries.
7474 Input video needs to have an even size.
7476 The filter accepts the following options:
7481 Specify all color properties at once.
7483 The accepted values are:
7513 Specify output colorspace.
7515 The accepted values are:
7524 BT.470BG or BT.601-6 625
7527 SMPTE-170M or BT.601-6 525
7536 BT.2020 with non-constant luminance
7542 Specify output transfer characteristics.
7544 The accepted values are:
7556 Constant gamma of 2.2
7559 Constant gamma of 2.8
7562 SMPTE-170M, BT.601-6 625 or BT.601-6 525
7580 BT.2020 for 10-bits content
7583 BT.2020 for 12-bits content
7589 Specify output color primaries.
7591 The accepted values are:
7600 BT.470BG or BT.601-6 625
7603 SMPTE-170M or BT.601-6 525
7627 Specify output color range.
7629 The accepted values are:
7632 TV (restricted) range
7635 MPEG (restricted) range
7646 Specify output color format.
7648 The accepted values are:
7651 YUV 4:2:0 planar 8-bits
7654 YUV 4:2:0 planar 10-bits
7657 YUV 4:2:0 planar 12-bits
7660 YUV 4:2:2 planar 8-bits
7663 YUV 4:2:2 planar 10-bits
7666 YUV 4:2:2 planar 12-bits
7669 YUV 4:4:4 planar 8-bits
7672 YUV 4:4:4 planar 10-bits
7675 YUV 4:4:4 planar 12-bits
7680 Do a fast conversion, which skips gamma/primary correction. This will take
7681 significantly less CPU, but will be mathematically incorrect. To get output
7682 compatible with that produced by the colormatrix filter, use fast=1.
7685 Specify dithering mode.
7687 The accepted values are:
7693 Floyd-Steinberg dithering
7697 Whitepoint adaptation mode.
7699 The accepted values are:
7702 Bradford whitepoint adaptation
7705 von Kries whitepoint adaptation
7708 identity whitepoint adaptation (i.e. no whitepoint adaptation)
7712 Override all input properties at once. Same accepted values as @ref{all}.
7715 Override input colorspace. Same accepted values as @ref{space}.
7718 Override input color primaries. Same accepted values as @ref{primaries}.
7721 Override input transfer characteristics. Same accepted values as @ref{trc}.
7724 Override input color range. Same accepted values as @ref{range}.
7728 The filter converts the transfer characteristics, color space and color
7729 primaries to the specified user values. The output value, if not specified,
7730 is set to a default value based on the "all" property. If that property is
7731 also not specified, the filter will log an error. The output color range and
7732 format default to the same value as the input color range and format. The
7733 input transfer characteristics, color space, color primaries and color range
7734 should be set on the input data. If any of these are missing, the filter will
7735 log an error and no conversion will take place.
7737 For example to convert the input to SMPTE-240M, use the command:
7739 colorspace=smpte240m
7742 @section convolution
7744 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
7746 The filter accepts the following options:
7753 Set matrix for each plane.
7754 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
7755 and from 1 to 49 odd number of signed integers in @var{row} mode.
7761 Set multiplier for calculated value for each plane.
7762 If unset or 0, it will be sum of all matrix elements.
7768 Set bias for each plane. This value is added to the result of the multiplication.
7769 Useful for making the overall image brighter or darker. Default is 0.0.
7775 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
7776 Default is @var{square}.
7779 @subsection Examples
7785 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"
7791 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"
7797 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"
7803 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"
7807 Apply laplacian edge detector which includes diagonals:
7809 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"
7815 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"
7821 Apply 2D convolution of video stream in frequency domain using second stream
7824 The filter accepts the following options:
7828 Set which planes to process.
7831 Set which impulse video frames will be processed, can be @var{first}
7832 or @var{all}. Default is @var{all}.
7835 The @code{convolve} filter also supports the @ref{framesync} options.
7839 Copy the input video source unchanged to the output. This is mainly useful for
7844 Video filtering on GPU using Apple's CoreImage API on OSX.
7846 Hardware acceleration is based on an OpenGL context. Usually, this means it is
7847 processed by video hardware. However, software-based OpenGL implementations
7848 exist which means there is no guarantee for hardware processing. It depends on
7851 There are many filters and image generators provided by Apple that come with a
7852 large variety of options. The filter has to be referenced by its name along
7855 The coreimage filter accepts the following options:
7858 List all available filters and generators along with all their respective
7859 options as well as possible minimum and maximum values along with the default
7866 Specify all filters by their respective name and options.
7867 Use @var{list_filters} to determine all valid filter names and options.
7868 Numerical options are specified by a float value and are automatically clamped
7869 to their respective value range. Vector and color options have to be specified
7870 by a list of space separated float values. Character escaping has to be done.
7871 A special option name @code{default} is available to use default options for a
7874 It is required to specify either @code{default} or at least one of the filter options.
7875 All omitted options are used with their default values.
7876 The syntax of the filter string is as follows:
7878 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
7882 Specify a rectangle where the output of the filter chain is copied into the
7883 input image. It is given by a list of space separated float values:
7885 output_rect=x\ y\ width\ height
7887 If not given, the output rectangle equals the dimensions of the input image.
7888 The output rectangle is automatically cropped at the borders of the input
7889 image. Negative values are valid for each component.
7891 output_rect=25\ 25\ 100\ 100
7895 Several filters can be chained for successive processing without GPU-HOST
7896 transfers allowing for fast processing of complex filter chains.
7897 Currently, only filters with zero (generators) or exactly one (filters) input
7898 image and one output image are supported. Also, transition filters are not yet
7901 Some filters generate output images with additional padding depending on the
7902 respective filter kernel. The padding is automatically removed to ensure the
7903 filter output has the same size as the input image.
7905 For image generators, the size of the output image is determined by the
7906 previous output image of the filter chain or the input image of the whole
7907 filterchain, respectively. The generators do not use the pixel information of
7908 this image to generate their output. However, the generated output is
7909 blended onto this image, resulting in partial or complete coverage of the
7912 The @ref{coreimagesrc} video source can be used for generating input images
7913 which are directly fed into the filter chain. By using it, providing input
7914 images by another video source or an input video is not required.
7916 @subsection Examples
7921 List all filters available:
7923 coreimage=list_filters=true
7927 Use the CIBoxBlur filter with default options to blur an image:
7929 coreimage=filter=CIBoxBlur@@default
7933 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
7934 its center at 100x100 and a radius of 50 pixels:
7936 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
7940 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
7941 given as complete and escaped command-line for Apple's standard bash shell:
7943 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
7949 Cover a rectangular object
7951 It accepts the following options:
7955 Filepath of the optional cover image, needs to be in yuv420.
7960 It accepts the following values:
7963 cover it by the supplied image
7965 cover it by interpolating the surrounding pixels
7968 Default value is @var{blur}.
7971 @subsection Examples
7975 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
7977 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
7983 Crop the input video to given dimensions.
7985 It accepts the following parameters:
7989 The width of the output video. It defaults to @code{iw}.
7990 This expression is evaluated only once during the filter
7991 configuration, or when the @samp{w} or @samp{out_w} command is sent.
7994 The height of the output video. It defaults to @code{ih}.
7995 This expression is evaluated only once during the filter
7996 configuration, or when the @samp{h} or @samp{out_h} command is sent.
7999 The horizontal position, in the input video, of the left edge of the output
8000 video. It defaults to @code{(in_w-out_w)/2}.
8001 This expression is evaluated per-frame.
8004 The vertical position, in the input video, of the top edge of the output video.
8005 It defaults to @code{(in_h-out_h)/2}.
8006 This expression is evaluated per-frame.
8009 If set to 1 will force the output display aspect ratio
8010 to be the same of the input, by changing the output sample aspect
8011 ratio. It defaults to 0.
8014 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8015 width/height/x/y as specified and will not be rounded to nearest smaller value.
8019 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8020 expressions containing the following constants:
8025 The computed values for @var{x} and @var{y}. They are evaluated for
8030 The input width and height.
8034 These are the same as @var{in_w} and @var{in_h}.
8038 The output (cropped) width and height.
8042 These are the same as @var{out_w} and @var{out_h}.
8045 same as @var{iw} / @var{ih}
8048 input sample aspect ratio
8051 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8055 horizontal and vertical chroma subsample values. For example for the
8056 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8059 The number of the input frame, starting from 0.
8062 the position in the file of the input frame, NAN if unknown
8065 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8069 The expression for @var{out_w} may depend on the value of @var{out_h},
8070 and the expression for @var{out_h} may depend on @var{out_w}, but they
8071 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8072 evaluated after @var{out_w} and @var{out_h}.
8074 The @var{x} and @var{y} parameters specify the expressions for the
8075 position of the top-left corner of the output (non-cropped) area. They
8076 are evaluated for each frame. If the evaluated value is not valid, it
8077 is approximated to the nearest valid value.
8079 The expression for @var{x} may depend on @var{y}, and the expression
8080 for @var{y} may depend on @var{x}.
8082 @subsection Examples
8086 Crop area with size 100x100 at position (12,34).
8091 Using named options, the example above becomes:
8093 crop=w=100:h=100:x=12:y=34
8097 Crop the central input area with size 100x100:
8103 Crop the central input area with size 2/3 of the input video:
8105 crop=2/3*in_w:2/3*in_h
8109 Crop the input video central square:
8116 Delimit the rectangle with the top-left corner placed at position
8117 100:100 and the right-bottom corner corresponding to the right-bottom
8118 corner of the input image.
8120 crop=in_w-100:in_h-100:100:100
8124 Crop 10 pixels from the left and right borders, and 20 pixels from
8125 the top and bottom borders
8127 crop=in_w-2*10:in_h-2*20
8131 Keep only the bottom right quarter of the input image:
8133 crop=in_w/2:in_h/2:in_w/2:in_h/2
8137 Crop height for getting Greek harmony:
8139 crop=in_w:1/PHI*in_w
8143 Apply trembling effect:
8145 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)
8149 Apply erratic camera effect depending on timestamp:
8151 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)"
8155 Set x depending on the value of y:
8157 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8161 @subsection Commands
8163 This filter supports the following commands:
8169 Set width/height of the output video and the horizontal/vertical position
8171 The command accepts the same syntax of the corresponding option.
8173 If the specified expression is not valid, it is kept at its current
8179 Auto-detect the crop size.
8181 It calculates the necessary cropping parameters and prints the
8182 recommended parameters via the logging system. The detected dimensions
8183 correspond to the non-black area of the input video.
8185 It accepts the following parameters:
8190 Set higher black value threshold, which can be optionally specified
8191 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8192 value greater to the set value is considered non-black. It defaults to 24.
8193 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8194 on the bitdepth of the pixel format.
8197 The value which the width/height should be divisible by. It defaults to
8198 16. The offset is automatically adjusted to center the video. Use 2 to
8199 get only even dimensions (needed for 4:2:2 video). 16 is best when
8200 encoding to most video codecs.
8202 @item reset_count, reset
8203 Set the counter that determines after how many frames cropdetect will
8204 reset the previously detected largest video area and start over to
8205 detect the current optimal crop area. Default value is 0.
8207 This can be useful when channel logos distort the video area. 0
8208 indicates 'never reset', and returns the largest area encountered during
8215 Delay video filtering until a given wallclock timestamp. The filter first
8216 passes on @option{preroll} amount of frames, then it buffers at most
8217 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8218 it forwards the buffered frames and also any subsequent frames coming in its
8221 The filter can be used synchronize the output of multiple ffmpeg processes for
8222 realtime output devices like decklink. By putting the delay in the filtering
8223 chain and pre-buffering frames the process can pass on data to output almost
8224 immediately after the target wallclock timestamp is reached.
8226 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
8232 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
8235 The duration of content to pass on as preroll expressed in seconds. Default is 0.
8238 The maximum duration of content to buffer before waiting for the cue expressed
8239 in seconds. Default is 0.
8246 Apply color adjustments using curves.
8248 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
8249 component (red, green and blue) has its values defined by @var{N} key points
8250 tied from each other using a smooth curve. The x-axis represents the pixel
8251 values from the input frame, and the y-axis the new pixel values to be set for
8254 By default, a component curve is defined by the two points @var{(0;0)} and
8255 @var{(1;1)}. This creates a straight line where each original pixel value is
8256 "adjusted" to its own value, which means no change to the image.
8258 The filter allows you to redefine these two points and add some more. A new
8259 curve (using a natural cubic spline interpolation) will be define to pass
8260 smoothly through all these new coordinates. The new defined points needs to be
8261 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
8262 be in the @var{[0;1]} interval. If the computed curves happened to go outside
8263 the vector spaces, the values will be clipped accordingly.
8265 The filter accepts the following options:
8269 Select one of the available color presets. This option can be used in addition
8270 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
8271 options takes priority on the preset values.
8272 Available presets are:
8275 @item color_negative
8278 @item increase_contrast
8280 @item linear_contrast
8281 @item medium_contrast
8283 @item strong_contrast
8286 Default is @code{none}.
8288 Set the master key points. These points will define a second pass mapping. It
8289 is sometimes called a "luminance" or "value" mapping. It can be used with
8290 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
8291 post-processing LUT.
8293 Set the key points for the red component.
8295 Set the key points for the green component.
8297 Set the key points for the blue component.
8299 Set the key points for all components (not including master).
8300 Can be used in addition to the other key points component
8301 options. In this case, the unset component(s) will fallback on this
8302 @option{all} setting.
8304 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
8306 Save Gnuplot script of the curves in specified file.
8309 To avoid some filtergraph syntax conflicts, each key points list need to be
8310 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
8312 @subsection Examples
8316 Increase slightly the middle level of blue:
8318 curves=blue='0/0 0.5/0.58 1/1'
8324 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'
8326 Here we obtain the following coordinates for each components:
8329 @code{(0;0.11) (0.42;0.51) (1;0.95)}
8331 @code{(0;0) (0.50;0.48) (1;1)}
8333 @code{(0;0.22) (0.49;0.44) (1;0.80)}
8337 The previous example can also be achieved with the associated built-in preset:
8339 curves=preset=vintage
8349 Use a Photoshop preset and redefine the points of the green component:
8351 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
8355 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
8356 and @command{gnuplot}:
8358 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
8359 gnuplot -p /tmp/curves.plt
8365 Video data analysis filter.
8367 This filter shows hexadecimal pixel values of part of video.
8369 The filter accepts the following options:
8373 Set output video size.
8376 Set x offset from where to pick pixels.
8379 Set y offset from where to pick pixels.
8382 Set scope mode, can be one of the following:
8385 Draw hexadecimal pixel values with white color on black background.
8388 Draw hexadecimal pixel values with input video pixel color on black
8392 Draw hexadecimal pixel values on color background picked from input video,
8393 the text color is picked in such way so its always visible.
8397 Draw rows and columns numbers on left and top of video.
8400 Set background opacity.
8403 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
8408 Denoise frames using 2D DCT (frequency domain filtering).
8410 This filter is not designed for real time.
8412 The filter accepts the following options:
8416 Set the noise sigma constant.
8418 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
8419 coefficient (absolute value) below this threshold with be dropped.
8421 If you need a more advanced filtering, see @option{expr}.
8423 Default is @code{0}.
8426 Set number overlapping pixels for each block. Since the filter can be slow, you
8427 may want to reduce this value, at the cost of a less effective filter and the
8428 risk of various artefacts.
8430 If the overlapping value doesn't permit processing the whole input width or
8431 height, a warning will be displayed and according borders won't be denoised.
8433 Default value is @var{blocksize}-1, which is the best possible setting.
8436 Set the coefficient factor expression.
8438 For each coefficient of a DCT block, this expression will be evaluated as a
8439 multiplier value for the coefficient.
8441 If this is option is set, the @option{sigma} option will be ignored.
8443 The absolute value of the coefficient can be accessed through the @var{c}
8447 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8448 @var{blocksize}, which is the width and height of the processed blocks.
8450 The default value is @var{3} (8x8) and can be raised to @var{4} for a
8451 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
8452 on the speed processing. Also, a larger block size does not necessarily means a
8456 @subsection Examples
8458 Apply a denoise with a @option{sigma} of @code{4.5}:
8463 The same operation can be achieved using the expression system:
8465 dctdnoiz=e='gte(c, 4.5*3)'
8468 Violent denoise using a block size of @code{16x16}:
8475 Remove banding artifacts from input video.
8476 It works by replacing banded pixels with average value of referenced pixels.
8478 The filter accepts the following options:
8485 Set banding detection threshold for each plane. Default is 0.02.
8486 Valid range is 0.00003 to 0.5.
8487 If difference between current pixel and reference pixel is less than threshold,
8488 it will be considered as banded.
8491 Banding detection range in pixels. Default is 16. If positive, random number
8492 in range 0 to set value will be used. If negative, exact absolute value
8494 The range defines square of four pixels around current pixel.
8497 Set direction in radians from which four pixel will be compared. If positive,
8498 random direction from 0 to set direction will be picked. If negative, exact of
8499 absolute value will be picked. For example direction 0, -PI or -2*PI radians
8500 will pick only pixels on same row and -PI/2 will pick only pixels on same
8504 If enabled, current pixel is compared with average value of all four
8505 surrounding pixels. The default is enabled. If disabled current pixel is
8506 compared with all four surrounding pixels. The pixel is considered banded
8507 if only all four differences with surrounding pixels are less than threshold.
8510 If enabled, current pixel is changed if and only if all pixel components are banded,
8511 e.g. banding detection threshold is triggered for all color components.
8512 The default is disabled.
8517 Remove blocking artifacts from input video.
8519 The filter accepts the following options:
8523 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
8524 This controls what kind of deblocking is applied.
8527 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
8533 Set blocking detection thresholds. Allowed range is 0 to 1.
8534 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
8535 Using higher threshold gives more deblocking strength.
8536 Setting @var{alpha} controls threshold detection at exact edge of block.
8537 Remaining options controls threshold detection near the edge. Each one for
8538 below/above or left/right. Setting any of those to @var{0} disables
8542 Set planes to filter. Default is to filter all available planes.
8545 @subsection Examples
8549 Deblock using weak filter and block size of 4 pixels.
8551 deblock=filter=weak:block=4
8555 Deblock using strong filter, block size of 4 pixels and custom thresholds for
8556 deblocking more edges.
8558 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
8562 Similar as above, but filter only first plane.
8564 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
8568 Similar as above, but filter only second and third plane.
8570 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
8577 Drop duplicated frames at regular intervals.
8579 The filter accepts the following options:
8583 Set the number of frames from which one will be dropped. Setting this to
8584 @var{N} means one frame in every batch of @var{N} frames will be dropped.
8585 Default is @code{5}.
8588 Set the threshold for duplicate detection. If the difference metric for a frame
8589 is less than or equal to this value, then it is declared as duplicate. Default
8593 Set scene change threshold. Default is @code{15}.
8597 Set the size of the x and y-axis blocks used during metric calculations.
8598 Larger blocks give better noise suppression, but also give worse detection of
8599 small movements. Must be a power of two. Default is @code{32}.
8602 Mark main input as a pre-processed input and activate clean source input
8603 stream. This allows the input to be pre-processed with various filters to help
8604 the metrics calculation while keeping the frame selection lossless. When set to
8605 @code{1}, the first stream is for the pre-processed input, and the second
8606 stream is the clean source from where the kept frames are chosen. Default is
8610 Set whether or not chroma is considered in the metric calculations. Default is
8616 Apply 2D deconvolution of video stream in frequency domain using second stream
8619 The filter accepts the following options:
8623 Set which planes to process.
8626 Set which impulse video frames will be processed, can be @var{first}
8627 or @var{all}. Default is @var{all}.
8630 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
8631 and height are not same and not power of 2 or if stream prior to convolving
8635 The @code{deconvolve} filter also supports the @ref{framesync} options.
8639 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
8641 It accepts the following options:
8645 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
8646 @var{rainbows} for cross-color reduction.
8649 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
8652 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
8655 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
8658 Set temporal chroma threshold. Lower values increases reduction of cross-color.
8663 Apply deflate effect to the video.
8665 This filter replaces the pixel by the local(3x3) average by taking into account
8666 only values lower than the pixel.
8668 It accepts the following options:
8675 Limit the maximum change for each plane, default is 65535.
8676 If 0, plane will remain unchanged.
8679 @subsection Commands
8681 This filter supports the all above options as @ref{commands}.
8685 Remove temporal frame luminance variations.
8687 It accepts the following options:
8691 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
8694 Set averaging mode to smooth temporal luminance variations.
8696 Available values are:
8721 Do not actually modify frame. Useful when one only wants metadata.
8726 Remove judder produced by partially interlaced telecined content.
8728 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
8729 source was partially telecined content then the output of @code{pullup,dejudder}
8730 will have a variable frame rate. May change the recorded frame rate of the
8731 container. Aside from that change, this filter will not affect constant frame
8734 The option available in this filter is:
8738 Specify the length of the window over which the judder repeats.
8740 Accepts any integer greater than 1. Useful values are:
8744 If the original was telecined from 24 to 30 fps (Film to NTSC).
8747 If the original was telecined from 25 to 30 fps (PAL to NTSC).
8750 If a mixture of the two.
8753 The default is @samp{4}.
8758 Suppress a TV station logo by a simple interpolation of the surrounding
8759 pixels. Just set a rectangle covering the logo and watch it disappear
8760 (and sometimes something even uglier appear - your mileage may vary).
8762 It accepts the following parameters:
8767 Specify the top left corner coordinates of the logo. They must be
8772 Specify the width and height of the logo to clear. They must be
8776 Specify the thickness of the fuzzy edge of the rectangle (added to
8777 @var{w} and @var{h}). The default value is 1. This option is
8778 deprecated, setting higher values should no longer be necessary and
8782 When set to 1, a green rectangle is drawn on the screen to simplify
8783 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
8784 The default value is 0.
8786 The rectangle is drawn on the outermost pixels which will be (partly)
8787 replaced with interpolated values. The values of the next pixels
8788 immediately outside this rectangle in each direction will be used to
8789 compute the interpolated pixel values inside the rectangle.
8793 @subsection Examples
8797 Set a rectangle covering the area with top left corner coordinates 0,0
8798 and size 100x77, and a band of size 10:
8800 delogo=x=0:y=0:w=100:h=77:band=10
8807 Remove the rain in the input image/video by applying the derain methods based on
8808 convolutional neural networks. Supported models:
8812 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
8813 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
8816 Training as well as model generation scripts are provided in
8817 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
8819 Native model files (.model) can be generated from TensorFlow model
8820 files (.pb) by using tools/python/convert.py
8822 The filter accepts the following options:
8826 Specify which filter to use. This option accepts the following values:
8830 Derain filter. To conduct derain filter, you need to use a derain model.
8833 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
8835 Default value is @samp{derain}.
8838 Specify which DNN backend to use for model loading and execution. This option accepts
8839 the following values:
8843 Native implementation of DNN loading and execution.
8846 TensorFlow backend. To enable this backend you
8847 need to install the TensorFlow for C library (see
8848 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
8849 @code{--enable-libtensorflow}
8851 Default value is @samp{native}.
8854 Set path to model file specifying network architecture and its parameters.
8855 Note that different backends use different file formats. TensorFlow and native
8856 backend can load files for only its format.
8861 Attempt to fix small changes in horizontal and/or vertical shift. This
8862 filter helps remove camera shake from hand-holding a camera, bumping a
8863 tripod, moving on a vehicle, etc.
8865 The filter accepts the following options:
8873 Specify a rectangular area where to limit the search for motion
8875 If desired the search for motion vectors can be limited to a
8876 rectangular area of the frame defined by its top left corner, width
8877 and height. These parameters have the same meaning as the drawbox
8878 filter which can be used to visualise the position of the bounding
8881 This is useful when simultaneous movement of subjects within the frame
8882 might be confused for camera motion by the motion vector search.
8884 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
8885 then the full frame is used. This allows later options to be set
8886 without specifying the bounding box for the motion vector search.
8888 Default - search the whole frame.
8892 Specify the maximum extent of movement in x and y directions in the
8893 range 0-64 pixels. Default 16.
8896 Specify how to generate pixels to fill blanks at the edge of the
8897 frame. Available values are:
8900 Fill zeroes at blank locations
8902 Original image at blank locations
8904 Extruded edge value at blank locations
8906 Mirrored edge at blank locations
8908 Default value is @samp{mirror}.
8911 Specify the blocksize to use for motion search. Range 4-128 pixels,
8915 Specify the contrast threshold for blocks. Only blocks with more than
8916 the specified contrast (difference between darkest and lightest
8917 pixels) will be considered. Range 1-255, default 125.
8920 Specify the search strategy. Available values are:
8923 Set exhaustive search
8925 Set less exhaustive search.
8927 Default value is @samp{exhaustive}.
8930 If set then a detailed log of the motion search is written to the
8937 Remove unwanted contamination of foreground colors, caused by reflected color of
8938 greenscreen or bluescreen.
8940 This filter accepts the following options:
8944 Set what type of despill to use.
8947 Set how spillmap will be generated.
8950 Set how much to get rid of still remaining spill.
8953 Controls amount of red in spill area.
8956 Controls amount of green in spill area.
8957 Should be -1 for greenscreen.
8960 Controls amount of blue in spill area.
8961 Should be -1 for bluescreen.
8964 Controls brightness of spill area, preserving colors.
8967 Modify alpha from generated spillmap.
8972 Apply an exact inverse of the telecine operation. It requires a predefined
8973 pattern specified using the pattern option which must be the same as that passed
8974 to the telecine filter.
8976 This filter accepts the following options:
8985 The default value is @code{top}.
8989 A string of numbers representing the pulldown pattern you wish to apply.
8990 The default value is @code{23}.
8993 A number representing position of the first frame with respect to the telecine
8994 pattern. This is to be used if the stream is cut. The default value is @code{0}.
8999 Apply dilation effect to the video.
9001 This filter replaces the pixel by the local(3x3) maximum.
9003 It accepts the following options:
9010 Limit the maximum change for each plane, default is 65535.
9011 If 0, plane will remain unchanged.
9014 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9017 Flags to local 3x3 coordinates maps like this:
9024 @subsection Commands
9026 This filter supports the all above options as @ref{commands}.
9030 Displace pixels as indicated by second and third input stream.
9032 It takes three input streams and outputs one stream, the first input is the
9033 source, and second and third input are displacement maps.
9035 The second input specifies how much to displace pixels along the
9036 x-axis, while the third input specifies how much to displace pixels
9038 If one of displacement map streams terminates, last frame from that
9039 displacement map will be used.
9041 Note that once generated, displacements maps can be reused over and over again.
9043 A description of the accepted options follows.
9047 Set displace behavior for pixels that are out of range.
9049 Available values are:
9052 Missing pixels are replaced by black pixels.
9055 Adjacent pixels will spread out to replace missing pixels.
9058 Out of range pixels are wrapped so they point to pixels of other side.
9061 Out of range pixels will be replaced with mirrored pixels.
9063 Default is @samp{smear}.
9067 @subsection Examples
9071 Add ripple effect to rgb input of video size hd720:
9073 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
9077 Add wave effect to rgb input of video size hd720:
9079 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
9083 @section dnn_processing
9085 Do image processing with deep neural networks. It works together with another filter
9086 which converts the pixel format of the Frame to what the dnn network requires.
9088 The filter accepts the following options:
9092 Specify which DNN backend to use for model loading and execution. This option accepts
9093 the following values:
9097 Native implementation of DNN loading and execution.
9100 TensorFlow backend. To enable this backend you
9101 need to install the TensorFlow for C library (see
9102 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9103 @code{--enable-libtensorflow}
9106 Default value is @samp{native}.
9109 Set path to model file specifying network architecture and its parameters.
9110 Note that different backends use different file formats. TensorFlow and native
9111 backend can load files for only its format.
9113 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9116 Set the input name of the dnn network.
9119 Set the output name of the dnn network.
9125 Halve the red channle of the frame with format rgb24:
9127 ffmpeg -i input.jpg -vf format=rgb24,dnn_processing=model=halve_first_channel.model:input=dnn_in:output=dnn_out:dnn_backend=native out.native.png
9131 Halve the pixel value of the frame with format gray32f:
9133 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
9140 Draw a colored box on the input image.
9142 It accepts the following parameters:
9147 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9151 The expressions which specify the width and height of the box; if 0 they are interpreted as
9152 the input width and height. It defaults to 0.
9155 Specify the color of the box to write. For the general syntax of this option,
9156 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9157 value @code{invert} is used, the box edge color is the same as the
9158 video with inverted luma.
9161 The expression which sets the thickness of the box edge.
9162 A value of @code{fill} will create a filled box. Default value is @code{3}.
9164 See below for the list of accepted constants.
9167 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
9168 will overwrite the video's color and alpha pixels.
9169 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
9172 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9173 following constants:
9177 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9181 horizontal and vertical chroma subsample values. For example for the
9182 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9186 The input width and height.
9189 The input sample aspect ratio.
9193 The x and y offset coordinates where the box is drawn.
9197 The width and height of the drawn box.
9200 The thickness of the drawn box.
9202 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9203 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9207 @subsection Examples
9211 Draw a black box around the edge of the input image:
9217 Draw a box with color red and an opacity of 50%:
9219 drawbox=10:20:200:60:red@@0.5
9222 The previous example can be specified as:
9224 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
9228 Fill the box with pink color:
9230 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
9234 Draw a 2-pixel red 2.40:1 mask:
9236 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
9240 @subsection Commands
9241 This filter supports same commands as options.
9242 The command accepts the same syntax of the corresponding option.
9244 If the specified expression is not valid, it is kept at its current
9249 Draw a graph using input video metadata.
9251 It accepts the following parameters:
9255 Set 1st frame metadata key from which metadata values will be used to draw a graph.
9258 Set 1st foreground color expression.
9261 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
9264 Set 2nd foreground color expression.
9267 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
9270 Set 3rd foreground color expression.
9273 Set 4th frame metadata key from which metadata values will be used to draw a graph.
9276 Set 4th foreground color expression.
9279 Set minimal value of metadata value.
9282 Set maximal value of metadata value.
9285 Set graph background color. Default is white.
9290 Available values for mode is:
9297 Default is @code{line}.
9302 Available values for slide is:
9305 Draw new frame when right border is reached.
9308 Replace old columns with new ones.
9311 Scroll from right to left.
9314 Scroll from left to right.
9317 Draw single picture.
9320 Default is @code{frame}.
9323 Set size of graph video. For the syntax of this option, check the
9324 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
9325 The default value is @code{900x256}.
9328 Set the output frame rate. Default value is @code{25}.
9330 The foreground color expressions can use the following variables:
9333 Minimal value of metadata value.
9336 Maximal value of metadata value.
9339 Current metadata key value.
9342 The color is defined as 0xAABBGGRR.
9345 Example using metadata from @ref{signalstats} filter:
9347 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
9350 Example using metadata from @ref{ebur128} filter:
9352 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
9357 Draw a grid on the input image.
9359 It accepts the following parameters:
9364 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
9368 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
9369 input width and height, respectively, minus @code{thickness}, so image gets
9370 framed. Default to 0.
9373 Specify the color of the grid. For the general syntax of this option,
9374 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9375 value @code{invert} is used, the grid color is the same as the
9376 video with inverted luma.
9379 The expression which sets the thickness of the grid line. Default value is @code{1}.
9381 See below for the list of accepted constants.
9384 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
9385 will overwrite the video's color and alpha pixels.
9386 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
9389 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9390 following constants:
9394 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9398 horizontal and vertical chroma subsample values. For example for the
9399 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9403 The input grid cell width and height.
9406 The input sample aspect ratio.
9410 The x and y coordinates of some point of grid intersection (meant to configure offset).
9414 The width and height of the drawn cell.
9417 The thickness of the drawn cell.
9419 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9420 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9424 @subsection Examples
9428 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
9430 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
9434 Draw a white 3x3 grid with an opacity of 50%:
9436 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
9440 @subsection Commands
9441 This filter supports same commands as options.
9442 The command accepts the same syntax of the corresponding option.
9444 If the specified expression is not valid, it is kept at its current
9450 Draw a text string or text from a specified file on top of a video, using the
9451 libfreetype library.
9453 To enable compilation of this filter, you need to configure FFmpeg with
9454 @code{--enable-libfreetype}.
9455 To enable default font fallback and the @var{font} option you need to
9456 configure FFmpeg with @code{--enable-libfontconfig}.
9457 To enable the @var{text_shaping} option, you need to configure FFmpeg with
9458 @code{--enable-libfribidi}.
9462 It accepts the following parameters:
9467 Used to draw a box around text using the background color.
9468 The value must be either 1 (enable) or 0 (disable).
9469 The default value of @var{box} is 0.
9472 Set the width of the border to be drawn around the box using @var{boxcolor}.
9473 The default value of @var{boxborderw} is 0.
9476 The color to be used for drawing box around text. For the syntax of this
9477 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9479 The default value of @var{boxcolor} is "white".
9482 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
9483 The default value of @var{line_spacing} is 0.
9486 Set the width of the border to be drawn around the text using @var{bordercolor}.
9487 The default value of @var{borderw} is 0.
9490 Set the color to be used for drawing border around text. For the syntax of this
9491 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9493 The default value of @var{bordercolor} is "black".
9496 Select how the @var{text} is expanded. Can be either @code{none},
9497 @code{strftime} (deprecated) or
9498 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
9502 Set a start time for the count. Value is in microseconds. Only applied
9503 in the deprecated strftime expansion mode. To emulate in normal expansion
9504 mode use the @code{pts} function, supplying the start time (in seconds)
9505 as the second argument.
9508 If true, check and fix text coords to avoid clipping.
9511 The color to be used for drawing fonts. For the syntax of this option, check
9512 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
9514 The default value of @var{fontcolor} is "black".
9516 @item fontcolor_expr
9517 String which is expanded the same way as @var{text} to obtain dynamic
9518 @var{fontcolor} value. By default this option has empty value and is not
9519 processed. When this option is set, it overrides @var{fontcolor} option.
9522 The font family to be used for drawing text. By default Sans.
9525 The font file to be used for drawing text. The path must be included.
9526 This parameter is mandatory if the fontconfig support is disabled.
9529 Draw the text applying alpha blending. The value can
9530 be a number between 0.0 and 1.0.
9531 The expression accepts the same variables @var{x, y} as well.
9532 The default value is 1.
9533 Please see @var{fontcolor_expr}.
9536 The font size to be used for drawing text.
9537 The default value of @var{fontsize} is 16.
9540 If set to 1, attempt to shape the text (for example, reverse the order of
9541 right-to-left text and join Arabic characters) before drawing it.
9542 Otherwise, just draw the text exactly as given.
9543 By default 1 (if supported).
9546 The flags to be used for loading the fonts.
9548 The flags map the corresponding flags supported by libfreetype, and are
9549 a combination of the following values:
9556 @item vertical_layout
9557 @item force_autohint
9560 @item ignore_global_advance_width
9562 @item ignore_transform
9568 Default value is "default".
9570 For more information consult the documentation for the FT_LOAD_*
9574 The color to be used for drawing a shadow behind the drawn text. For the
9575 syntax of this option, check the @ref{color syntax,,"Color" section in the
9576 ffmpeg-utils manual,ffmpeg-utils}.
9578 The default value of @var{shadowcolor} is "black".
9582 The x and y offsets for the text shadow position with respect to the
9583 position of the text. They can be either positive or negative
9584 values. The default value for both is "0".
9587 The starting frame number for the n/frame_num variable. The default value
9591 The size in number of spaces to use for rendering the tab.
9595 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
9596 format. It can be used with or without text parameter. @var{timecode_rate}
9597 option must be specified.
9599 @item timecode_rate, rate, r
9600 Set the timecode frame rate (timecode only). Value will be rounded to nearest
9601 integer. Minimum value is "1".
9602 Drop-frame timecode is supported for frame rates 30 & 60.
9605 If set to 1, the output of the timecode option will wrap around at 24 hours.
9606 Default is 0 (disabled).
9609 The text string to be drawn. The text must be a sequence of UTF-8
9611 This parameter is mandatory if no file is specified with the parameter
9615 A text file containing text to be drawn. The text must be a sequence
9616 of UTF-8 encoded characters.
9618 This parameter is mandatory if no text string is specified with the
9619 parameter @var{text}.
9621 If both @var{text} and @var{textfile} are specified, an error is thrown.
9624 If set to 1, the @var{textfile} will be reloaded before each frame.
9625 Be sure to update it atomically, or it may be read partially, or even fail.
9629 The expressions which specify the offsets where text will be drawn
9630 within the video frame. They are relative to the top/left border of the
9633 The default value of @var{x} and @var{y} is "0".
9635 See below for the list of accepted constants and functions.
9638 The parameters for @var{x} and @var{y} are expressions containing the
9639 following constants and functions:
9643 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
9647 horizontal and vertical chroma subsample values. For example for the
9648 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9651 the height of each text line
9659 @item max_glyph_a, ascent
9660 the maximum distance from the baseline to the highest/upper grid
9661 coordinate used to place a glyph outline point, for all the rendered
9663 It is a positive value, due to the grid's orientation with the Y axis
9666 @item max_glyph_d, descent
9667 the maximum distance from the baseline to the lowest grid coordinate
9668 used to place a glyph outline point, for all the rendered glyphs.
9669 This is a negative value, due to the grid's orientation, with the Y axis
9673 maximum glyph height, that is the maximum height for all the glyphs
9674 contained in the rendered text, it is equivalent to @var{ascent} -
9678 maximum glyph width, that is the maximum width for all the glyphs
9679 contained in the rendered text
9682 the number of input frame, starting from 0
9684 @item rand(min, max)
9685 return a random number included between @var{min} and @var{max}
9688 The input sample aspect ratio.
9691 timestamp expressed in seconds, NAN if the input timestamp is unknown
9694 the height of the rendered text
9697 the width of the rendered text
9701 the x and y offset coordinates where the text is drawn.
9703 These parameters allow the @var{x} and @var{y} expressions to refer
9704 to each other, so you can for example specify @code{y=x/dar}.
9707 A one character description of the current frame's picture type.
9710 The current packet's position in the input file or stream
9711 (in bytes, from the start of the input). A value of -1 indicates
9712 this info is not available.
9715 The current packet's duration, in seconds.
9718 The current packet's size (in bytes).
9721 @anchor{drawtext_expansion}
9722 @subsection Text expansion
9724 If @option{expansion} is set to @code{strftime},
9725 the filter recognizes strftime() sequences in the provided text and
9726 expands them accordingly. Check the documentation of strftime(). This
9727 feature is deprecated.
9729 If @option{expansion} is set to @code{none}, the text is printed verbatim.
9731 If @option{expansion} is set to @code{normal} (which is the default),
9732 the following expansion mechanism is used.
9734 The backslash character @samp{\}, followed by any character, always expands to
9735 the second character.
9737 Sequences of the form @code{%@{...@}} are expanded. The text between the
9738 braces is a function name, possibly followed by arguments separated by ':'.
9739 If the arguments contain special characters or delimiters (':' or '@}'),
9740 they should be escaped.
9742 Note that they probably must also be escaped as the value for the
9743 @option{text} option in the filter argument string and as the filter
9744 argument in the filtergraph description, and possibly also for the shell,
9745 that makes up to four levels of escaping; using a text file avoids these
9748 The following functions are available:
9753 The expression evaluation result.
9755 It must take one argument specifying the expression to be evaluated,
9756 which accepts the same constants and functions as the @var{x} and
9757 @var{y} values. Note that not all constants should be used, for
9758 example the text size is not known when evaluating the expression, so
9759 the constants @var{text_w} and @var{text_h} will have an undefined
9762 @item expr_int_format, eif
9763 Evaluate the expression's value and output as formatted integer.
9765 The first argument is the expression to be evaluated, just as for the @var{expr} function.
9766 The second argument specifies the output format. Allowed values are @samp{x},
9767 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
9768 @code{printf} function.
9769 The third parameter is optional and sets the number of positions taken by the output.
9770 It can be used to add padding with zeros from the left.
9773 The time at which the filter is running, expressed in UTC.
9774 It can accept an argument: a strftime() format string.
9777 The time at which the filter is running, expressed in the local time zone.
9778 It can accept an argument: a strftime() format string.
9781 Frame metadata. Takes one or two arguments.
9783 The first argument is mandatory and specifies the metadata key.
9785 The second argument is optional and specifies a default value, used when the
9786 metadata key is not found or empty.
9788 Available metadata can be identified by inspecting entries
9789 starting with TAG included within each frame section
9790 printed by running @code{ffprobe -show_frames}.
9792 String metadata generated in filters leading to
9793 the drawtext filter are also available.
9796 The frame number, starting from 0.
9799 A one character description of the current picture type.
9802 The timestamp of the current frame.
9803 It can take up to three arguments.
9805 The first argument is the format of the timestamp; it defaults to @code{flt}
9806 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
9807 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
9808 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
9809 @code{localtime} stands for the timestamp of the frame formatted as
9810 local time zone time.
9812 The second argument is an offset added to the timestamp.
9814 If the format is set to @code{hms}, a third argument @code{24HH} may be
9815 supplied to present the hour part of the formatted timestamp in 24h format
9818 If the format is set to @code{localtime} or @code{gmtime},
9819 a third argument may be supplied: a strftime() format string.
9820 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
9823 @subsection Commands
9825 This filter supports altering parameters via commands:
9828 Alter existing filter parameters.
9830 Syntax for the argument is the same as for filter invocation, e.g.
9833 fontsize=56:fontcolor=green:text='Hello World'
9836 Full filter invocation with sendcmd would look like this:
9839 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
9843 If the entire argument can't be parsed or applied as valid values then the filter will
9844 continue with its existing parameters.
9846 @subsection Examples
9850 Draw "Test Text" with font FreeSerif, using the default values for the
9851 optional parameters.
9854 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
9858 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
9859 and y=50 (counting from the top-left corner of the screen), text is
9860 yellow with a red box around it. Both the text and the box have an
9864 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
9865 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
9868 Note that the double quotes are not necessary if spaces are not used
9869 within the parameter list.
9872 Show the text at the center of the video frame:
9874 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
9878 Show the text at a random position, switching to a new position every 30 seconds:
9880 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)"
9884 Show a text line sliding from right to left in the last row of the video
9885 frame. The file @file{LONG_LINE} is assumed to contain a single line
9888 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
9892 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
9894 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
9898 Draw a single green letter "g", at the center of the input video.
9899 The glyph baseline is placed at half screen height.
9901 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
9905 Show text for 1 second every 3 seconds:
9907 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
9911 Use fontconfig to set the font. Note that the colons need to be escaped.
9913 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
9917 Print the date of a real-time encoding (see strftime(3)):
9919 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
9923 Show text fading in and out (appearing/disappearing):
9926 DS=1.0 # display start
9927 DE=10.0 # display end
9928 FID=1.5 # fade in duration
9929 FOD=5 # fade out duration
9930 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 @}"
9934 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
9935 and the @option{fontsize} value are included in the @option{y} offset.
9937 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
9938 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
9942 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
9943 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
9944 must have option @option{-export_path_metadata 1} for the special metadata fields
9945 to be available for filters.
9947 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
9952 For more information about libfreetype, check:
9953 @url{http://www.freetype.org/}.
9955 For more information about fontconfig, check:
9956 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
9958 For more information about libfribidi, check:
9959 @url{http://fribidi.org/}.
9963 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
9965 The filter accepts the following options:
9970 Set low and high threshold values used by the Canny thresholding
9973 The high threshold selects the "strong" edge pixels, which are then
9974 connected through 8-connectivity with the "weak" edge pixels selected
9975 by the low threshold.
9977 @var{low} and @var{high} threshold values must be chosen in the range
9978 [0,1], and @var{low} should be lesser or equal to @var{high}.
9980 Default value for @var{low} is @code{20/255}, and default value for @var{high}
9984 Define the drawing mode.
9988 Draw white/gray wires on black background.
9991 Mix the colors to create a paint/cartoon effect.
9994 Apply Canny edge detector on all selected planes.
9996 Default value is @var{wires}.
9999 Select planes for filtering. By default all available planes are filtered.
10002 @subsection Examples
10006 Standard edge detection with custom values for the hysteresis thresholding:
10008 edgedetect=low=0.1:high=0.4
10012 Painting effect without thresholding:
10014 edgedetect=mode=colormix:high=0
10020 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10022 For each input image, the filter will compute the optimal mapping from
10023 the input to the output given the codebook length, that is the number
10024 of distinct output colors.
10026 This filter accepts the following options.
10029 @item codebook_length, l
10030 Set codebook length. The value must be a positive integer, and
10031 represents the number of distinct output colors. Default value is 256.
10034 Set the maximum number of iterations to apply for computing the optimal
10035 mapping. The higher the value the better the result and the higher the
10036 computation time. Default value is 1.
10039 Set a random seed, must be an integer included between 0 and
10040 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10041 will try to use a good random seed on a best effort basis.
10044 Set pal8 output pixel format. This option does not work with codebook
10045 length greater than 256.
10050 Measure graylevel entropy in histogram of color channels of video frames.
10052 It accepts the following parameters:
10056 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10058 @var{diff} mode measures entropy of histogram delta values, absolute differences
10059 between neighbour histogram values.
10063 Set brightness, contrast, saturation and approximate gamma adjustment.
10065 The filter accepts the following options:
10069 Set the contrast expression. The value must be a float value in range
10070 @code{-1000.0} to @code{1000.0}. The default value is "1".
10073 Set the brightness expression. The value must be a float value in
10074 range @code{-1.0} to @code{1.0}. The default value is "0".
10077 Set the saturation expression. The value must be a float in
10078 range @code{0.0} to @code{3.0}. The default value is "1".
10081 Set the gamma expression. The value must be a float in range
10082 @code{0.1} to @code{10.0}. The default value is "1".
10085 Set the gamma expression for red. The value must be a float in
10086 range @code{0.1} to @code{10.0}. The default value is "1".
10089 Set the gamma expression for green. The value must be a float in range
10090 @code{0.1} to @code{10.0}. The default value is "1".
10093 Set the gamma expression for blue. The value must be a float in range
10094 @code{0.1} to @code{10.0}. The default value is "1".
10097 Set the gamma weight expression. It can be used to reduce the effect
10098 of a high gamma value on bright image areas, e.g. keep them from
10099 getting overamplified and just plain white. The value must be a float
10100 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10101 gamma correction all the way down while @code{1.0} leaves it at its
10102 full strength. Default is "1".
10105 Set when the expressions for brightness, contrast, saturation and
10106 gamma expressions are evaluated.
10108 It accepts the following values:
10111 only evaluate expressions once during the filter initialization or
10112 when a command is processed
10115 evaluate expressions for each incoming frame
10118 Default value is @samp{init}.
10121 The expressions accept the following parameters:
10124 frame count of the input frame starting from 0
10127 byte position of the corresponding packet in the input file, NAN if
10131 frame rate of the input video, NAN if the input frame rate is unknown
10134 timestamp expressed in seconds, NAN if the input timestamp is unknown
10137 @subsection Commands
10138 The filter supports the following commands:
10142 Set the contrast expression.
10145 Set the brightness expression.
10148 Set the saturation expression.
10151 Set the gamma expression.
10154 Set the gamma_r expression.
10157 Set gamma_g expression.
10160 Set gamma_b expression.
10163 Set gamma_weight expression.
10165 The command accepts the same syntax of the corresponding option.
10167 If the specified expression is not valid, it is kept at its current
10174 Apply erosion effect to the video.
10176 This filter replaces the pixel by the local(3x3) minimum.
10178 It accepts the following options:
10185 Limit the maximum change for each plane, default is 65535.
10186 If 0, plane will remain unchanged.
10189 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
10192 Flags to local 3x3 coordinates maps like this:
10199 @subsection Commands
10201 This filter supports the all above options as @ref{commands}.
10203 @section extractplanes
10205 Extract color channel components from input video stream into
10206 separate grayscale video streams.
10208 The filter accepts the following option:
10212 Set plane(s) to extract.
10214 Available values for planes are:
10225 Choosing planes not available in the input will result in an error.
10226 That means you cannot select @code{r}, @code{g}, @code{b} planes
10227 with @code{y}, @code{u}, @code{v} planes at same time.
10230 @subsection Examples
10234 Extract luma, u and v color channel component from input video frame
10235 into 3 grayscale outputs:
10237 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
10243 Apply a fade-in/out effect to the input video.
10245 It accepts the following parameters:
10249 The effect type can be either "in" for a fade-in, or "out" for a fade-out
10251 Default is @code{in}.
10253 @item start_frame, s
10254 Specify the number of the frame to start applying the fade
10255 effect at. Default is 0.
10258 The number of frames that the fade effect lasts. At the end of the
10259 fade-in effect, the output video will have the same intensity as the input video.
10260 At the end of the fade-out transition, the output video will be filled with the
10261 selected @option{color}.
10265 If set to 1, fade only alpha channel, if one exists on the input.
10266 Default value is 0.
10268 @item start_time, st
10269 Specify the timestamp (in seconds) of the frame to start to apply the fade
10270 effect. If both start_frame and start_time are specified, the fade will start at
10271 whichever comes last. Default is 0.
10274 The number of seconds for which the fade effect has to last. At the end of the
10275 fade-in effect the output video will have the same intensity as the input video,
10276 at the end of the fade-out transition the output video will be filled with the
10277 selected @option{color}.
10278 If both duration and nb_frames are specified, duration is used. Default is 0
10279 (nb_frames is used by default).
10282 Specify the color of the fade. Default is "black".
10285 @subsection Examples
10289 Fade in the first 30 frames of video:
10294 The command above is equivalent to:
10300 Fade out the last 45 frames of a 200-frame video:
10303 fade=type=out:start_frame=155:nb_frames=45
10307 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
10309 fade=in:0:25, fade=out:975:25
10313 Make the first 5 frames yellow, then fade in from frame 5-24:
10315 fade=in:5:20:color=yellow
10319 Fade in alpha over first 25 frames of video:
10321 fade=in:0:25:alpha=1
10325 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
10327 fade=t=in:st=5.5:d=0.5
10333 Denoise frames using 3D FFT (frequency domain filtering).
10335 The filter accepts the following options:
10339 Set the noise sigma constant. This sets denoising strength.
10340 Default value is 1. Allowed range is from 0 to 30.
10341 Using very high sigma with low overlap may give blocking artifacts.
10344 Set amount of denoising. By default all detected noise is reduced.
10345 Default value is 1. Allowed range is from 0 to 1.
10348 Set size of block, Default is 4, can be 3, 4, 5 or 6.
10349 Actual size of block in pixels is 2 to power of @var{block}, so by default
10350 block size in pixels is 2^4 which is 16.
10353 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
10356 Set number of previous frames to use for denoising. By default is set to 0.
10359 Set number of next frames to to use for denoising. By default is set to 0.
10362 Set planes which will be filtered, by default are all available filtered
10367 Apply arbitrary expressions to samples in frequency domain
10371 Adjust the dc value (gain) of the luma plane of the image. The filter
10372 accepts an integer value in range @code{0} to @code{1000}. The default
10373 value is set to @code{0}.
10376 Adjust the dc value (gain) of the 1st chroma plane of the image. The
10377 filter accepts an integer value in range @code{0} to @code{1000}. The
10378 default value is set to @code{0}.
10381 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
10382 filter accepts an integer value in range @code{0} to @code{1000}. The
10383 default value is set to @code{0}.
10386 Set the frequency domain weight expression for the luma plane.
10389 Set the frequency domain weight expression for the 1st chroma plane.
10392 Set the frequency domain weight expression for the 2nd chroma plane.
10395 Set when the expressions are evaluated.
10397 It accepts the following values:
10400 Only evaluate expressions once during the filter initialization.
10403 Evaluate expressions for each incoming frame.
10406 Default value is @samp{init}.
10408 The filter accepts the following variables:
10411 The coordinates of the current sample.
10415 The width and height of the image.
10418 The number of input frame, starting from 0.
10421 @subsection Examples
10427 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
10433 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
10439 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
10445 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
10452 Extract a single field from an interlaced image using stride
10453 arithmetic to avoid wasting CPU time. The output frames are marked as
10456 The filter accepts the following options:
10460 Specify whether to extract the top (if the value is @code{0} or
10461 @code{top}) or the bottom field (if the value is @code{1} or
10467 Create new frames by copying the top and bottom fields from surrounding frames
10468 supplied as numbers by the hint file.
10472 Set file containing hints: absolute/relative frame numbers.
10474 There must be one line for each frame in a clip. Each line must contain two
10475 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
10476 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
10477 is current frame number for @code{absolute} mode or out of [-1, 1] range
10478 for @code{relative} mode. First number tells from which frame to pick up top
10479 field and second number tells from which frame to pick up bottom field.
10481 If optionally followed by @code{+} output frame will be marked as interlaced,
10482 else if followed by @code{-} output frame will be marked as progressive, else
10483 it will be marked same as input frame.
10484 If optionally followed by @code{t} output frame will use only top field, or in
10485 case of @code{b} it will use only bottom field.
10486 If line starts with @code{#} or @code{;} that line is skipped.
10489 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
10492 Example of first several lines of @code{hint} file for @code{relative} mode:
10494 0,0 - # first frame
10495 1,0 - # second frame, use third's frame top field and second's frame bottom field
10496 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
10511 @section fieldmatch
10513 Field matching filter for inverse telecine. It is meant to reconstruct the
10514 progressive frames from a telecined stream. The filter does not drop duplicated
10515 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
10516 followed by a decimation filter such as @ref{decimate} in the filtergraph.
10518 The separation of the field matching and the decimation is notably motivated by
10519 the possibility of inserting a de-interlacing filter fallback between the two.
10520 If the source has mixed telecined and real interlaced content,
10521 @code{fieldmatch} will not be able to match fields for the interlaced parts.
10522 But these remaining combed frames will be marked as interlaced, and thus can be
10523 de-interlaced by a later filter such as @ref{yadif} before decimation.
10525 In addition to the various configuration options, @code{fieldmatch} can take an
10526 optional second stream, activated through the @option{ppsrc} option. If
10527 enabled, the frames reconstruction will be based on the fields and frames from
10528 this second stream. This allows the first input to be pre-processed in order to
10529 help the various algorithms of the filter, while keeping the output lossless
10530 (assuming the fields are matched properly). Typically, a field-aware denoiser,
10531 or brightness/contrast adjustments can help.
10533 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
10534 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
10535 which @code{fieldmatch} is based on. While the semantic and usage are very
10536 close, some behaviour and options names can differ.
10538 The @ref{decimate} filter currently only works for constant frame rate input.
10539 If your input has mixed telecined (30fps) and progressive content with a lower
10540 framerate like 24fps use the following filterchain to produce the necessary cfr
10541 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
10543 The filter accepts the following options:
10547 Specify the assumed field order of the input stream. Available values are:
10551 Auto detect parity (use FFmpeg's internal parity value).
10553 Assume bottom field first.
10555 Assume top field first.
10558 Note that it is sometimes recommended not to trust the parity announced by the
10561 Default value is @var{auto}.
10564 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
10565 sense that it won't risk creating jerkiness due to duplicate frames when
10566 possible, but if there are bad edits or blended fields it will end up
10567 outputting combed frames when a good match might actually exist. On the other
10568 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
10569 but will almost always find a good frame if there is one. The other values are
10570 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
10571 jerkiness and creating duplicate frames versus finding good matches in sections
10572 with bad edits, orphaned fields, blended fields, etc.
10574 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
10576 Available values are:
10580 2-way matching (p/c)
10582 2-way matching, and trying 3rd match if still combed (p/c + n)
10584 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
10586 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
10587 still combed (p/c + n + u/b)
10589 3-way matching (p/c/n)
10591 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
10592 detected as combed (p/c/n + u/b)
10595 The parenthesis at the end indicate the matches that would be used for that
10596 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
10599 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
10602 Default value is @var{pc_n}.
10605 Mark the main input stream as a pre-processed input, and enable the secondary
10606 input stream as the clean source to pick the fields from. See the filter
10607 introduction for more details. It is similar to the @option{clip2} feature from
10610 Default value is @code{0} (disabled).
10613 Set the field to match from. It is recommended to set this to the same value as
10614 @option{order} unless you experience matching failures with that setting. In
10615 certain circumstances changing the field that is used to match from can have a
10616 large impact on matching performance. Available values are:
10620 Automatic (same value as @option{order}).
10622 Match from the bottom field.
10624 Match from the top field.
10627 Default value is @var{auto}.
10630 Set whether or not chroma is included during the match comparisons. In most
10631 cases it is recommended to leave this enabled. You should set this to @code{0}
10632 only if your clip has bad chroma problems such as heavy rainbowing or other
10633 artifacts. Setting this to @code{0} could also be used to speed things up at
10634 the cost of some accuracy.
10636 Default value is @code{1}.
10640 These define an exclusion band which excludes the lines between @option{y0} and
10641 @option{y1} from being included in the field matching decision. An exclusion
10642 band can be used to ignore subtitles, a logo, or other things that may
10643 interfere with the matching. @option{y0} sets the starting scan line and
10644 @option{y1} sets the ending line; all lines in between @option{y0} and
10645 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
10646 @option{y0} and @option{y1} to the same value will disable the feature.
10647 @option{y0} and @option{y1} defaults to @code{0}.
10650 Set the scene change detection threshold as a percentage of maximum change on
10651 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
10652 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
10653 @option{scthresh} is @code{[0.0, 100.0]}.
10655 Default value is @code{12.0}.
10658 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
10659 account the combed scores of matches when deciding what match to use as the
10660 final match. Available values are:
10664 No final matching based on combed scores.
10666 Combed scores are only used when a scene change is detected.
10668 Use combed scores all the time.
10671 Default is @var{sc}.
10674 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
10675 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
10676 Available values are:
10680 No forced calculation.
10682 Force p/c/n calculations.
10684 Force p/c/n/u/b calculations.
10687 Default value is @var{none}.
10690 This is the area combing threshold used for combed frame detection. This
10691 essentially controls how "strong" or "visible" combing must be to be detected.
10692 Larger values mean combing must be more visible and smaller values mean combing
10693 can be less visible or strong and still be detected. Valid settings are from
10694 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
10695 be detected as combed). This is basically a pixel difference value. A good
10696 range is @code{[8, 12]}.
10698 Default value is @code{9}.
10701 Sets whether or not chroma is considered in the combed frame decision. Only
10702 disable this if your source has chroma problems (rainbowing, etc.) that are
10703 causing problems for the combed frame detection with chroma enabled. Actually,
10704 using @option{chroma}=@var{0} is usually more reliable, except for the case
10705 where there is chroma only combing in the source.
10707 Default value is @code{0}.
10711 Respectively set the x-axis and y-axis size of the window used during combed
10712 frame detection. This has to do with the size of the area in which
10713 @option{combpel} pixels are required to be detected as combed for a frame to be
10714 declared combed. See the @option{combpel} parameter description for more info.
10715 Possible values are any number that is a power of 2 starting at 4 and going up
10718 Default value is @code{16}.
10721 The number of combed pixels inside any of the @option{blocky} by
10722 @option{blockx} size blocks on the frame for the frame to be detected as
10723 combed. While @option{cthresh} controls how "visible" the combing must be, this
10724 setting controls "how much" combing there must be in any localized area (a
10725 window defined by the @option{blockx} and @option{blocky} settings) on the
10726 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
10727 which point no frames will ever be detected as combed). This setting is known
10728 as @option{MI} in TFM/VFM vocabulary.
10730 Default value is @code{80}.
10733 @anchor{p/c/n/u/b meaning}
10734 @subsection p/c/n/u/b meaning
10736 @subsubsection p/c/n
10738 We assume the following telecined stream:
10741 Top fields: 1 2 2 3 4
10742 Bottom fields: 1 2 3 4 4
10745 The numbers correspond to the progressive frame the fields relate to. Here, the
10746 first two frames are progressive, the 3rd and 4th are combed, and so on.
10748 When @code{fieldmatch} is configured to run a matching from bottom
10749 (@option{field}=@var{bottom}) this is how this input stream get transformed:
10754 B 1 2 3 4 4 <-- matching reference
10763 As a result of the field matching, we can see that some frames get duplicated.
10764 To perform a complete inverse telecine, you need to rely on a decimation filter
10765 after this operation. See for instance the @ref{decimate} filter.
10767 The same operation now matching from top fields (@option{field}=@var{top})
10772 T 1 2 2 3 4 <-- matching reference
10782 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
10783 basically, they refer to the frame and field of the opposite parity:
10786 @item @var{p} matches the field of the opposite parity in the previous frame
10787 @item @var{c} matches the field of the opposite parity in the current frame
10788 @item @var{n} matches the field of the opposite parity in the next frame
10793 The @var{u} and @var{b} matching are a bit special in the sense that they match
10794 from the opposite parity flag. In the following examples, we assume that we are
10795 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
10796 'x' is placed above and below each matched fields.
10798 With bottom matching (@option{field}=@var{bottom}):
10803 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10804 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10812 With top matching (@option{field}=@var{top}):
10817 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
10818 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
10826 @subsection Examples
10828 Simple IVTC of a top field first telecined stream:
10830 fieldmatch=order=tff:combmatch=none, decimate
10833 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
10835 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
10838 @section fieldorder
10840 Transform the field order of the input video.
10842 It accepts the following parameters:
10847 The output field order. Valid values are @var{tff} for top field first or @var{bff}
10848 for bottom field first.
10851 The default value is @samp{tff}.
10853 The transformation is done by shifting the picture content up or down
10854 by one line, and filling the remaining line with appropriate picture content.
10855 This method is consistent with most broadcast field order converters.
10857 If the input video is not flagged as being interlaced, or it is already
10858 flagged as being of the required output field order, then this filter does
10859 not alter the incoming video.
10861 It is very useful when converting to or from PAL DV material,
10862 which is bottom field first.
10866 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
10869 @section fifo, afifo
10871 Buffer input images and send them when they are requested.
10873 It is mainly useful when auto-inserted by the libavfilter
10876 It does not take parameters.
10878 @section fillborders
10880 Fill borders of the input video, without changing video stream dimensions.
10881 Sometimes video can have garbage at the four edges and you may not want to
10882 crop video input to keep size multiple of some number.
10884 This filter accepts the following options:
10888 Number of pixels to fill from left border.
10891 Number of pixels to fill from right border.
10894 Number of pixels to fill from top border.
10897 Number of pixels to fill from bottom border.
10902 It accepts the following values:
10905 fill pixels using outermost pixels
10908 fill pixels using mirroring
10911 fill pixels with constant value
10914 Default is @var{smear}.
10917 Set color for pixels in fixed mode. Default is @var{black}.
10920 @subsection Commands
10921 This filter supports same @ref{commands} as options.
10922 The command accepts the same syntax of the corresponding option.
10924 If the specified expression is not valid, it is kept at its current
10929 Find a rectangular object
10931 It accepts the following options:
10935 Filepath of the object image, needs to be in gray8.
10938 Detection threshold, default is 0.5.
10941 Number of mipmaps, default is 3.
10943 @item xmin, ymin, xmax, ymax
10944 Specifies the rectangle in which to search.
10947 @subsection Examples
10951 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
10953 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
10959 Flood area with values of same pixel components with another values.
10961 It accepts the following options:
10964 Set pixel x coordinate.
10967 Set pixel y coordinate.
10970 Set source #0 component value.
10973 Set source #1 component value.
10976 Set source #2 component value.
10979 Set source #3 component value.
10982 Set destination #0 component value.
10985 Set destination #1 component value.
10988 Set destination #2 component value.
10991 Set destination #3 component value.
10997 Convert the input video to one of the specified pixel formats.
10998 Libavfilter will try to pick one that is suitable as input to
11001 It accepts the following parameters:
11005 A '|'-separated list of pixel format names, such as
11006 "pix_fmts=yuv420p|monow|rgb24".
11010 @subsection Examples
11014 Convert the input video to the @var{yuv420p} format
11016 format=pix_fmts=yuv420p
11019 Convert the input video to any of the formats in the list
11021 format=pix_fmts=yuv420p|yuv444p|yuv410p
11028 Convert the video to specified constant frame rate by duplicating or dropping
11029 frames as necessary.
11031 It accepts the following parameters:
11035 The desired output frame rate. The default is @code{25}.
11038 Assume the first PTS should be the given value, in seconds. This allows for
11039 padding/trimming at the start of stream. By default, no assumption is made
11040 about the first frame's expected PTS, so no padding or trimming is done.
11041 For example, this could be set to 0 to pad the beginning with duplicates of
11042 the first frame if a video stream starts after the audio stream or to trim any
11043 frames with a negative PTS.
11046 Timestamp (PTS) rounding method.
11048 Possible values are:
11055 round towards -infinity
11057 round towards +infinity
11061 The default is @code{near}.
11064 Action performed when reading the last frame.
11066 Possible values are:
11069 Use same timestamp rounding method as used for other frames.
11071 Pass through last frame if input duration has not been reached yet.
11073 The default is @code{round}.
11077 Alternatively, the options can be specified as a flat string:
11078 @var{fps}[:@var{start_time}[:@var{round}]].
11080 See also the @ref{setpts} filter.
11082 @subsection Examples
11086 A typical usage in order to set the fps to 25:
11092 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11094 fps=fps=film:round=near
11100 Pack two different video streams into a stereoscopic video, setting proper
11101 metadata on supported codecs. The two views should have the same size and
11102 framerate and processing will stop when the shorter video ends. Please note
11103 that you may conveniently adjust view properties with the @ref{scale} and
11106 It accepts the following parameters:
11110 The desired packing format. Supported values are:
11115 The views are next to each other (default).
11118 The views are on top of each other.
11121 The views are packed by line.
11124 The views are packed by column.
11127 The views are temporally interleaved.
11136 # Convert left and right views into a frame-sequential video
11137 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11139 # Convert views into a side-by-side video with the same output resolution as the input
11140 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
11145 Change the frame rate by interpolating new video output frames from the source
11148 This filter is not designed to function correctly with interlaced media. If
11149 you wish to change the frame rate of interlaced media then you are required
11150 to deinterlace before this filter and re-interlace after this filter.
11152 A description of the accepted options follows.
11156 Specify the output frames per second. This option can also be specified
11157 as a value alone. The default is @code{50}.
11160 Specify the start of a range where the output frame will be created as a
11161 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11162 the default is @code{15}.
11165 Specify the end of a range where the output frame will be created as a
11166 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11167 the default is @code{240}.
11170 Specify the level at which a scene change is detected as a value between
11171 0 and 100 to indicate a new scene; a low value reflects a low
11172 probability for the current frame to introduce a new scene, while a higher
11173 value means the current frame is more likely to be one.
11174 The default is @code{8.2}.
11177 Specify flags influencing the filter process.
11179 Available value for @var{flags} is:
11182 @item scene_change_detect, scd
11183 Enable scene change detection using the value of the option @var{scene}.
11184 This flag is enabled by default.
11190 Select one frame every N-th frame.
11192 This filter accepts the following option:
11195 Select frame after every @code{step} frames.
11196 Allowed values are positive integers higher than 0. Default value is @code{1}.
11199 @section freezedetect
11201 Detect frozen video.
11203 This filter logs a message and sets frame metadata when it detects that the
11204 input video has no significant change in content during a specified duration.
11205 Video freeze detection calculates the mean average absolute difference of all
11206 the components of video frames and compares it to a noise floor.
11208 The printed times and duration are expressed in seconds. The
11209 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
11210 whose timestamp equals or exceeds the detection duration and it contains the
11211 timestamp of the first frame of the freeze. The
11212 @code{lavfi.freezedetect.freeze_duration} and
11213 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
11216 The filter accepts the following options:
11220 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
11221 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
11225 Set freeze duration until notification (default is 2 seconds).
11228 @section freezeframes
11230 Freeze video frames.
11232 This filter freezes video frames using frame from 2nd input.
11234 The filter accepts the following options:
11238 Set number of first frame from which to start freeze.
11241 Set number of last frame from which to end freeze.
11244 Set number of frame from 2nd input which will be used instead of replaced frames.
11250 Apply a frei0r effect to the input video.
11252 To enable the compilation of this filter, you need to install the frei0r
11253 header and configure FFmpeg with @code{--enable-frei0r}.
11255 It accepts the following parameters:
11260 The name of the frei0r effect to load. If the environment variable
11261 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
11262 directories specified by the colon-separated list in @env{FREI0R_PATH}.
11263 Otherwise, the standard frei0r paths are searched, in this order:
11264 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
11265 @file{/usr/lib/frei0r-1/}.
11267 @item filter_params
11268 A '|'-separated list of parameters to pass to the frei0r effect.
11272 A frei0r effect parameter can be a boolean (its value is either
11273 "y" or "n"), a double, a color (specified as
11274 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
11275 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
11276 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
11277 a position (specified as @var{X}/@var{Y}, where
11278 @var{X} and @var{Y} are floating point numbers) and/or a string.
11280 The number and types of parameters depend on the loaded effect. If an
11281 effect parameter is not specified, the default value is set.
11283 @subsection Examples
11287 Apply the distort0r effect, setting the first two double parameters:
11289 frei0r=filter_name=distort0r:filter_params=0.5|0.01
11293 Apply the colordistance effect, taking a color as the first parameter:
11295 frei0r=colordistance:0.2/0.3/0.4
11296 frei0r=colordistance:violet
11297 frei0r=colordistance:0x112233
11301 Apply the perspective effect, specifying the top left and top right image
11304 frei0r=perspective:0.2/0.2|0.8/0.2
11308 For more information, see
11309 @url{http://frei0r.dyne.org}
11313 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
11315 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
11316 processing filter, one of them is performed once per block, not per pixel.
11317 This allows for much higher speed.
11319 The filter accepts the following options:
11323 Set quality. This option defines the number of levels for averaging. It accepts
11324 an integer in the range 4-5. Default value is @code{4}.
11327 Force a constant quantization parameter. It accepts an integer in range 0-63.
11328 If not set, the filter will use the QP from the video stream (if available).
11331 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
11332 more details but also more artifacts, while higher values make the image smoother
11333 but also blurrier. Default value is @code{0} − PSNR optimal.
11335 @item use_bframe_qp
11336 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
11337 option may cause flicker since the B-Frames have often larger QP. Default is
11338 @code{0} (not enabled).
11344 Apply Gaussian blur filter.
11346 The filter accepts the following options:
11350 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
11353 Set number of steps for Gaussian approximation. Default is @code{1}.
11356 Set which planes to filter. By default all planes are filtered.
11359 Set vertical sigma, if negative it will be same as @code{sigma}.
11360 Default is @code{-1}.
11363 @subsection Commands
11364 This filter supports same commands as options.
11365 The command accepts the same syntax of the corresponding option.
11367 If the specified expression is not valid, it is kept at its current
11372 Apply generic equation to each pixel.
11374 The filter accepts the following options:
11377 @item lum_expr, lum
11378 Set the luminance expression.
11380 Set the chrominance blue expression.
11382 Set the chrominance red expression.
11383 @item alpha_expr, a
11384 Set the alpha expression.
11386 Set the red expression.
11387 @item green_expr, g
11388 Set the green expression.
11390 Set the blue expression.
11393 The colorspace is selected according to the specified options. If one
11394 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
11395 options is specified, the filter will automatically select a YCbCr
11396 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
11397 @option{blue_expr} options is specified, it will select an RGB
11400 If one of the chrominance expression is not defined, it falls back on the other
11401 one. If no alpha expression is specified it will evaluate to opaque value.
11402 If none of chrominance expressions are specified, they will evaluate
11403 to the luminance expression.
11405 The expressions can use the following variables and functions:
11409 The sequential number of the filtered frame, starting from @code{0}.
11413 The coordinates of the current sample.
11417 The width and height of the image.
11421 Width and height scale depending on the currently filtered plane. It is the
11422 ratio between the corresponding luma plane number of pixels and the current
11423 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
11424 @code{0.5,0.5} for chroma planes.
11427 Time of the current frame, expressed in seconds.
11430 Return the value of the pixel at location (@var{x},@var{y}) of the current
11434 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
11438 Return the value of the pixel at location (@var{x},@var{y}) of the
11439 blue-difference chroma plane. Return 0 if there is no such plane.
11442 Return the value of the pixel at location (@var{x},@var{y}) of the
11443 red-difference chroma plane. Return 0 if there is no such plane.
11448 Return the value of the pixel at location (@var{x},@var{y}) of the
11449 red/green/blue component. Return 0 if there is no such component.
11452 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
11453 plane. Return 0 if there is no such plane.
11455 @item interpolation
11456 Set one of interpolation methods:
11461 Default is bilinear.
11464 For functions, if @var{x} and @var{y} are outside the area, the value will be
11465 automatically clipped to the closer edge.
11467 @subsection Examples
11471 Flip the image horizontally:
11477 Generate a bidimensional sine wave, with angle @code{PI/3} and a
11478 wavelength of 100 pixels:
11480 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
11484 Generate a fancy enigmatic moving light:
11486 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
11490 Generate a quick emboss effect:
11492 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
11496 Modify RGB components depending on pixel position:
11498 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
11502 Create a radial gradient that is the same size as the input (also see
11503 the @ref{vignette} filter):
11505 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
11511 Fix the banding artifacts that are sometimes introduced into nearly flat
11512 regions by truncation to 8-bit color depth.
11513 Interpolate the gradients that should go where the bands are, and
11516 It is designed for playback only. Do not use it prior to
11517 lossy compression, because compression tends to lose the dither and
11518 bring back the bands.
11520 It accepts the following parameters:
11525 The maximum amount by which the filter will change any one pixel. This is also
11526 the threshold for detecting nearly flat regions. Acceptable values range from
11527 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
11531 The neighborhood to fit the gradient to. A larger radius makes for smoother
11532 gradients, but also prevents the filter from modifying the pixels near detailed
11533 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
11534 values will be clipped to the valid range.
11538 Alternatively, the options can be specified as a flat string:
11539 @var{strength}[:@var{radius}]
11541 @subsection Examples
11545 Apply the filter with a @code{3.5} strength and radius of @code{8}:
11551 Specify radius, omitting the strength (which will fall-back to the default
11559 @anchor{graphmonitor}
11560 @section graphmonitor
11561 Show various filtergraph stats.
11563 With this filter one can debug complete filtergraph.
11564 Especially issues with links filling with queued frames.
11566 The filter accepts the following options:
11570 Set video output size. Default is @var{hd720}.
11573 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
11576 Set output mode, can be @var{fulll} or @var{compact}.
11577 In @var{compact} mode only filters with some queued frames have displayed stats.
11580 Set flags which enable which stats are shown in video.
11582 Available values for flags are:
11585 Display number of queued frames in each link.
11587 @item frame_count_in
11588 Display number of frames taken from filter.
11590 @item frame_count_out
11591 Display number of frames given out from filter.
11594 Display current filtered frame pts.
11597 Display current filtered frame time.
11600 Display time base for filter link.
11603 Display used format for filter link.
11606 Display video size or number of audio channels in case of audio used by filter link.
11609 Display video frame rate or sample rate in case of audio used by filter link.
11613 Set upper limit for video rate of output stream, Default value is @var{25}.
11614 This guarantee that output video frame rate will not be higher than this value.
11618 A color constancy variation filter which estimates scene illumination via grey edge algorithm
11619 and corrects the scene colors accordingly.
11621 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
11623 The filter accepts the following options:
11627 The order of differentiation to be applied on the scene. Must be chosen in the range
11628 [0,2] and default value is 1.
11631 The Minkowski parameter to be used for calculating the Minkowski distance. Must
11632 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
11633 max value instead of calculating Minkowski distance.
11636 The standard deviation of Gaussian blur to be applied on the scene. Must be
11637 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
11638 can't be equal to 0 if @var{difford} is greater than 0.
11641 @subsection Examples
11647 greyedge=difford=1:minknorm=5:sigma=2
11653 greyedge=difford=1:minknorm=0:sigma=2
11661 Apply a Hald CLUT to a video stream.
11663 First input is the video stream to process, and second one is the Hald CLUT.
11664 The Hald CLUT input can be a simple picture or a complete video stream.
11666 The filter accepts the following options:
11670 Force termination when the shortest input terminates. Default is @code{0}.
11672 Continue applying the last CLUT after the end of the stream. A value of
11673 @code{0} disable the filter after the last frame of the CLUT is reached.
11674 Default is @code{1}.
11677 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
11678 filters share the same internals).
11680 This filter also supports the @ref{framesync} options.
11682 More information about the Hald CLUT can be found on Eskil Steenberg's website
11683 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
11685 @subsection Workflow examples
11687 @subsubsection Hald CLUT video stream
11689 Generate an identity Hald CLUT stream altered with various effects:
11691 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
11694 Note: make sure you use a lossless codec.
11696 Then use it with @code{haldclut} to apply it on some random stream:
11698 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
11701 The Hald CLUT will be applied to the 10 first seconds (duration of
11702 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
11703 to the remaining frames of the @code{mandelbrot} stream.
11705 @subsubsection Hald CLUT with preview
11707 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
11708 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
11709 biggest possible square starting at the top left of the picture. The remaining
11710 padding pixels (bottom or right) will be ignored. This area can be used to add
11711 a preview of the Hald CLUT.
11713 Typically, the following generated Hald CLUT will be supported by the
11714 @code{haldclut} filter:
11717 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
11718 pad=iw+320 [padded_clut];
11719 smptebars=s=320x256, split [a][b];
11720 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
11721 [main][b] overlay=W-320" -frames:v 1 clut.png
11724 It contains the original and a preview of the effect of the CLUT: SMPTE color
11725 bars are displayed on the right-top, and below the same color bars processed by
11728 Then, the effect of this Hald CLUT can be visualized with:
11730 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
11735 Flip the input video horizontally.
11737 For example, to horizontally flip the input video with @command{ffmpeg}:
11739 ffmpeg -i in.avi -vf "hflip" out.avi
11743 This filter applies a global color histogram equalization on a
11746 It can be used to correct video that has a compressed range of pixel
11747 intensities. The filter redistributes the pixel intensities to
11748 equalize their distribution across the intensity range. It may be
11749 viewed as an "automatically adjusting contrast filter". This filter is
11750 useful only for correcting degraded or poorly captured source
11753 The filter accepts the following options:
11757 Determine the amount of equalization to be applied. As the strength
11758 is reduced, the distribution of pixel intensities more-and-more
11759 approaches that of the input frame. The value must be a float number
11760 in the range [0,1] and defaults to 0.200.
11763 Set the maximum intensity that can generated and scale the output
11764 values appropriately. The strength should be set as desired and then
11765 the intensity can be limited if needed to avoid washing-out. The value
11766 must be a float number in the range [0,1] and defaults to 0.210.
11769 Set the antibanding level. If enabled the filter will randomly vary
11770 the luminance of output pixels by a small amount to avoid banding of
11771 the histogram. Possible values are @code{none}, @code{weak} or
11772 @code{strong}. It defaults to @code{none}.
11778 Compute and draw a color distribution histogram for the input video.
11780 The computed histogram is a representation of the color component
11781 distribution in an image.
11783 Standard histogram displays the color components distribution in an image.
11784 Displays color graph for each color component. Shows distribution of
11785 the Y, U, V, A or R, G, B components, depending on input format, in the
11786 current frame. Below each graph a color component scale meter is shown.
11788 The filter accepts the following options:
11792 Set height of level. Default value is @code{200}.
11793 Allowed range is [50, 2048].
11796 Set height of color scale. Default value is @code{12}.
11797 Allowed range is [0, 40].
11801 It accepts the following values:
11804 Per color component graphs are placed below each other.
11807 Per color component graphs are placed side by side.
11810 Presents information identical to that in the @code{parade}, except
11811 that the graphs representing color components are superimposed directly
11814 Default is @code{stack}.
11817 Set mode. Can be either @code{linear}, or @code{logarithmic}.
11818 Default is @code{linear}.
11821 Set what color components to display.
11822 Default is @code{7}.
11825 Set foreground opacity. Default is @code{0.7}.
11828 Set background opacity. Default is @code{0.5}.
11831 @subsection Examples
11836 Calculate and draw histogram:
11838 ffplay -i input -vf histogram
11846 This is a high precision/quality 3d denoise filter. It aims to reduce
11847 image noise, producing smooth images and making still images really
11848 still. It should enhance compressibility.
11850 It accepts the following optional parameters:
11854 A non-negative floating point number which specifies spatial luma strength.
11855 It defaults to 4.0.
11857 @item chroma_spatial
11858 A non-negative floating point number which specifies spatial chroma strength.
11859 It defaults to 3.0*@var{luma_spatial}/4.0.
11862 A floating point number which specifies luma temporal strength. It defaults to
11863 6.0*@var{luma_spatial}/4.0.
11866 A floating point number which specifies chroma temporal strength. It defaults to
11867 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
11870 @subsection Commands
11871 This filter supports same @ref{commands} as options.
11872 The command accepts the same syntax of the corresponding option.
11874 If the specified expression is not valid, it is kept at its current
11877 @anchor{hwdownload}
11878 @section hwdownload
11880 Download hardware frames to system memory.
11882 The input must be in hardware frames, and the output a non-hardware format.
11883 Not all formats will be supported on the output - it may be necessary to insert
11884 an additional @option{format} filter immediately following in the graph to get
11885 the output in a supported format.
11889 Map hardware frames to system memory or to another device.
11891 This filter has several different modes of operation; which one is used depends
11892 on the input and output formats:
11895 Hardware frame input, normal frame output
11897 Map the input frames to system memory and pass them to the output. If the
11898 original hardware frame is later required (for example, after overlaying
11899 something else on part of it), the @option{hwmap} filter can be used again
11900 in the next mode to retrieve it.
11902 Normal frame input, hardware frame output
11904 If the input is actually a software-mapped hardware frame, then unmap it -
11905 that is, return the original hardware frame.
11907 Otherwise, a device must be provided. Create new hardware surfaces on that
11908 device for the output, then map them back to the software format at the input
11909 and give those frames to the preceding filter. This will then act like the
11910 @option{hwupload} filter, but may be able to avoid an additional copy when
11911 the input is already in a compatible format.
11913 Hardware frame input and output
11915 A device must be supplied for the output, either directly or with the
11916 @option{derive_device} option. The input and output devices must be of
11917 different types and compatible - the exact meaning of this is
11918 system-dependent, but typically it means that they must refer to the same
11919 underlying hardware context (for example, refer to the same graphics card).
11921 If the input frames were originally created on the output device, then unmap
11922 to retrieve the original frames.
11924 Otherwise, map the frames to the output device - create new hardware frames
11925 on the output corresponding to the frames on the input.
11928 The following additional parameters are accepted:
11932 Set the frame mapping mode. Some combination of:
11935 The mapped frame should be readable.
11937 The mapped frame should be writeable.
11939 The mapping will always overwrite the entire frame.
11941 This may improve performance in some cases, as the original contents of the
11942 frame need not be loaded.
11944 The mapping must not involve any copying.
11946 Indirect mappings to copies of frames are created in some cases where either
11947 direct mapping is not possible or it would have unexpected properties.
11948 Setting this flag ensures that the mapping is direct and will fail if that is
11951 Defaults to @var{read+write} if not specified.
11953 @item derive_device @var{type}
11954 Rather than using the device supplied at initialisation, instead derive a new
11955 device of type @var{type} from the device the input frames exist on.
11958 In a hardware to hardware mapping, map in reverse - create frames in the sink
11959 and map them back to the source. This may be necessary in some cases where
11960 a mapping in one direction is required but only the opposite direction is
11961 supported by the devices being used.
11963 This option is dangerous - it may break the preceding filter in undefined
11964 ways if there are any additional constraints on that filter's output.
11965 Do not use it without fully understanding the implications of its use.
11971 Upload system memory frames to hardware surfaces.
11973 The device to upload to must be supplied when the filter is initialised. If
11974 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
11977 @anchor{hwupload_cuda}
11978 @section hwupload_cuda
11980 Upload system memory frames to a CUDA device.
11982 It accepts the following optional parameters:
11986 The number of the CUDA device to use
11991 Apply a high-quality magnification filter designed for pixel art. This filter
11992 was originally created by Maxim Stepin.
11994 It accepts the following option:
11998 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
11999 @code{hq3x} and @code{4} for @code{hq4x}.
12000 Default is @code{3}.
12004 Stack input videos horizontally.
12006 All streams must be of same pixel format and of same height.
12008 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12009 to create same output.
12011 The filter accepts the following option:
12015 Set number of input streams. Default is 2.
12018 If set to 1, force the output to terminate when the shortest input
12019 terminates. Default value is 0.
12024 Modify the hue and/or the saturation of the input.
12026 It accepts the following parameters:
12030 Specify the hue angle as a number of degrees. It accepts an expression,
12031 and defaults to "0".
12034 Specify the saturation in the [-10,10] range. It accepts an expression and
12038 Specify the hue angle as a number of radians. It accepts an
12039 expression, and defaults to "0".
12042 Specify the brightness in the [-10,10] range. It accepts an expression and
12046 @option{h} and @option{H} are mutually exclusive, and can't be
12047 specified at the same time.
12049 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12050 expressions containing the following constants:
12054 frame count of the input frame starting from 0
12057 presentation timestamp of the input frame expressed in time base units
12060 frame rate of the input video, NAN if the input frame rate is unknown
12063 timestamp expressed in seconds, NAN if the input timestamp is unknown
12066 time base of the input video
12069 @subsection Examples
12073 Set the hue to 90 degrees and the saturation to 1.0:
12079 Same command but expressing the hue in radians:
12085 Rotate hue and make the saturation swing between 0
12086 and 2 over a period of 1 second:
12088 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12092 Apply a 3 seconds saturation fade-in effect starting at 0:
12094 hue="s=min(t/3\,1)"
12097 The general fade-in expression can be written as:
12099 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12103 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12105 hue="s=max(0\, min(1\, (8-t)/3))"
12108 The general fade-out expression can be written as:
12110 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12115 @subsection Commands
12117 This filter supports the following commands:
12123 Modify the hue and/or the saturation and/or brightness of the input video.
12124 The command accepts the same syntax of the corresponding option.
12126 If the specified expression is not valid, it is kept at its current
12130 @section hysteresis
12132 Grow first stream into second stream by connecting components.
12133 This makes it possible to build more robust edge masks.
12135 This filter accepts the following options:
12139 Set which planes will be processed as bitmap, unprocessed planes will be
12140 copied from first stream.
12141 By default value 0xf, all planes will be processed.
12144 Set threshold which is used in filtering. If pixel component value is higher than
12145 this value filter algorithm for connecting components is activated.
12146 By default value is 0.
12151 Detect video interlacing type.
12153 This filter tries to detect if the input frames are interlaced, progressive,
12154 top or bottom field first. It will also try to detect fields that are
12155 repeated between adjacent frames (a sign of telecine).
12157 Single frame detection considers only immediately adjacent frames when classifying each frame.
12158 Multiple frame detection incorporates the classification history of previous frames.
12160 The filter will log these metadata values:
12163 @item single.current_frame
12164 Detected type of current frame using single-frame detection. One of:
12165 ``tff'' (top field first), ``bff'' (bottom field first),
12166 ``progressive'', or ``undetermined''
12169 Cumulative number of frames detected as top field first using single-frame detection.
12172 Cumulative number of frames detected as top field first using multiple-frame detection.
12175 Cumulative number of frames detected as bottom field first using single-frame detection.
12177 @item multiple.current_frame
12178 Detected type of current frame using multiple-frame detection. One of:
12179 ``tff'' (top field first), ``bff'' (bottom field first),
12180 ``progressive'', or ``undetermined''
12183 Cumulative number of frames detected as bottom field first using multiple-frame detection.
12185 @item single.progressive
12186 Cumulative number of frames detected as progressive using single-frame detection.
12188 @item multiple.progressive
12189 Cumulative number of frames detected as progressive using multiple-frame detection.
12191 @item single.undetermined
12192 Cumulative number of frames that could not be classified using single-frame detection.
12194 @item multiple.undetermined
12195 Cumulative number of frames that could not be classified using multiple-frame detection.
12197 @item repeated.current_frame
12198 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
12200 @item repeated.neither
12201 Cumulative number of frames with no repeated field.
12204 Cumulative number of frames with the top field repeated from the previous frame's top field.
12206 @item repeated.bottom
12207 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
12210 The filter accepts the following options:
12214 Set interlacing threshold.
12216 Set progressive threshold.
12218 Threshold for repeated field detection.
12220 Number of frames after which a given frame's contribution to the
12221 statistics is halved (i.e., it contributes only 0.5 to its
12222 classification). The default of 0 means that all frames seen are given
12223 full weight of 1.0 forever.
12224 @item analyze_interlaced_flag
12225 When this is not 0 then idet will use the specified number of frames to determine
12226 if the interlaced flag is accurate, it will not count undetermined frames.
12227 If the flag is found to be accurate it will be used without any further
12228 computations, if it is found to be inaccurate it will be cleared without any
12229 further computations. This allows inserting the idet filter as a low computational
12230 method to clean up the interlaced flag
12235 Deinterleave or interleave fields.
12237 This filter allows one to process interlaced images fields without
12238 deinterlacing them. Deinterleaving splits the input frame into 2
12239 fields (so called half pictures). Odd lines are moved to the top
12240 half of the output image, even lines to the bottom half.
12241 You can process (filter) them independently and then re-interleave them.
12243 The filter accepts the following options:
12247 @item chroma_mode, c
12248 @item alpha_mode, a
12249 Available values for @var{luma_mode}, @var{chroma_mode} and
12250 @var{alpha_mode} are:
12256 @item deinterleave, d
12257 Deinterleave fields, placing one above the other.
12259 @item interleave, i
12260 Interleave fields. Reverse the effect of deinterleaving.
12262 Default value is @code{none}.
12264 @item luma_swap, ls
12265 @item chroma_swap, cs
12266 @item alpha_swap, as
12267 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
12270 @subsection Commands
12272 This filter supports the all above options as @ref{commands}.
12276 Apply inflate effect to the video.
12278 This filter replaces the pixel by the local(3x3) average by taking into account
12279 only values higher than the pixel.
12281 It accepts the following options:
12288 Limit the maximum change for each plane, default is 65535.
12289 If 0, plane will remain unchanged.
12292 @subsection Commands
12294 This filter supports the all above options as @ref{commands}.
12298 Simple interlacing filter from progressive contents. This interleaves upper (or
12299 lower) lines from odd frames with lower (or upper) lines from even frames,
12300 halving the frame rate and preserving image height.
12303 Original Original New Frame
12304 Frame 'j' Frame 'j+1' (tff)
12305 ========== =========== ==================
12306 Line 0 --------------------> Frame 'j' Line 0
12307 Line 1 Line 1 ----> Frame 'j+1' Line 1
12308 Line 2 ---------------------> Frame 'j' Line 2
12309 Line 3 Line 3 ----> Frame 'j+1' Line 3
12311 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
12314 It accepts the following optional parameters:
12318 This determines whether the interlaced frame is taken from the even
12319 (tff - default) or odd (bff) lines of the progressive frame.
12322 Vertical lowpass filter to avoid twitter interlacing and
12323 reduce moire patterns.
12327 Disable vertical lowpass filter
12330 Enable linear filter (default)
12333 Enable complex filter. This will slightly less reduce twitter and moire
12334 but better retain detail and subjective sharpness impression.
12341 Deinterlace input video by applying Donald Graft's adaptive kernel
12342 deinterling. Work on interlaced parts of a video to produce
12343 progressive frames.
12345 The description of the accepted parameters follows.
12349 Set the threshold which affects the filter's tolerance when
12350 determining if a pixel line must be processed. It must be an integer
12351 in the range [0,255] and defaults to 10. A value of 0 will result in
12352 applying the process on every pixels.
12355 Paint pixels exceeding the threshold value to white if set to 1.
12359 Set the fields order. Swap fields if set to 1, leave fields alone if
12363 Enable additional sharpening if set to 1. Default is 0.
12366 Enable twoway sharpening if set to 1. Default is 0.
12369 @subsection Examples
12373 Apply default values:
12375 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
12379 Enable additional sharpening:
12385 Paint processed pixels in white:
12393 Slowly update darker pixels.
12395 This filter makes short flashes of light appear longer.
12396 This filter accepts the following options:
12400 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
12403 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
12406 @section lenscorrection
12408 Correct radial lens distortion
12410 This filter can be used to correct for radial distortion as can result from the use
12411 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
12412 one can use tools available for example as part of opencv or simply trial-and-error.
12413 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
12414 and extract the k1 and k2 coefficients from the resulting matrix.
12416 Note that effectively the same filter is available in the open-source tools Krita and
12417 Digikam from the KDE project.
12419 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
12420 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
12421 brightness distribution, so you may want to use both filters together in certain
12422 cases, though you will have to take care of ordering, i.e. whether vignetting should
12423 be applied before or after lens correction.
12425 @subsection Options
12427 The filter accepts the following options:
12431 Relative x-coordinate of the focal point of the image, and thereby the center of the
12432 distortion. This value has a range [0,1] and is expressed as fractions of the image
12433 width. Default is 0.5.
12435 Relative y-coordinate of the focal point of the image, and thereby the center of the
12436 distortion. This value has a range [0,1] and is expressed as fractions of the image
12437 height. Default is 0.5.
12439 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
12440 no correction. Default is 0.
12442 Coefficient of the double quadratic correction term. This value has a range [-1,1].
12443 0 means no correction. Default is 0.
12446 The formula that generates the correction is:
12448 @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)
12450 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
12451 distances from the focal point in the source and target images, respectively.
12455 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
12457 The @code{lensfun} filter requires the camera make, camera model, and lens model
12458 to apply the lens correction. The filter will load the lensfun database and
12459 query it to find the corresponding camera and lens entries in the database. As
12460 long as these entries can be found with the given options, the filter can
12461 perform corrections on frames. Note that incomplete strings will result in the
12462 filter choosing the best match with the given options, and the filter will
12463 output the chosen camera and lens models (logged with level "info"). You must
12464 provide the make, camera model, and lens model as they are required.
12466 The filter accepts the following options:
12470 The make of the camera (for example, "Canon"). This option is required.
12473 The model of the camera (for example, "Canon EOS 100D"). This option is
12477 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
12478 option is required.
12481 The type of correction to apply. The following values are valid options:
12485 Enables fixing lens vignetting.
12488 Enables fixing lens geometry. This is the default.
12491 Enables fixing chromatic aberrations.
12494 Enables fixing lens vignetting and lens geometry.
12497 Enables fixing lens vignetting and chromatic aberrations.
12500 Enables fixing both lens geometry and chromatic aberrations.
12503 Enables all possible corrections.
12507 The focal length of the image/video (zoom; expected constant for video). For
12508 example, a 18--55mm lens has focal length range of [18--55], so a value in that
12509 range should be chosen when using that lens. Default 18.
12512 The aperture of the image/video (expected constant for video). Note that
12513 aperture is only used for vignetting correction. Default 3.5.
12515 @item focus_distance
12516 The focus distance of the image/video (expected constant for video). Note that
12517 focus distance is only used for vignetting and only slightly affects the
12518 vignetting correction process. If unknown, leave it at the default value (which
12522 The scale factor which is applied after transformation. After correction the
12523 video is no longer necessarily rectangular. This parameter controls how much of
12524 the resulting image is visible. The value 0 means that a value will be chosen
12525 automatically such that there is little or no unmapped area in the output
12526 image. 1.0 means that no additional scaling is done. Lower values may result
12527 in more of the corrected image being visible, while higher values may avoid
12528 unmapped areas in the output.
12530 @item target_geometry
12531 The target geometry of the output image/video. The following values are valid
12535 @item rectilinear (default)
12538 @item equirectangular
12539 @item fisheye_orthographic
12540 @item fisheye_stereographic
12541 @item fisheye_equisolid
12542 @item fisheye_thoby
12545 Apply the reverse of image correction (instead of correcting distortion, apply
12548 @item interpolation
12549 The type of interpolation used when correcting distortion. The following values
12554 @item linear (default)
12559 @subsection Examples
12563 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
12564 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
12568 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
12572 Apply the same as before, but only for the first 5 seconds of video.
12575 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
12582 Obtain the VMAF (Video Multi-Method Assessment Fusion)
12583 score between two input videos.
12585 The obtained VMAF score is printed through the logging system.
12587 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
12588 After installing the library it can be enabled using:
12589 @code{./configure --enable-libvmaf --enable-version3}.
12590 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
12592 The filter has following options:
12596 Set the model path which is to be used for SVM.
12597 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
12600 Set the file path to be used to store logs.
12603 Set the format of the log file (xml or json).
12605 @item enable_transform
12606 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
12607 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
12608 Default value: @code{false}
12611 Invokes the phone model which will generate VMAF scores higher than in the
12612 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
12613 Default value: @code{false}
12616 Enables computing psnr along with vmaf.
12617 Default value: @code{false}
12620 Enables computing ssim along with vmaf.
12621 Default value: @code{false}
12624 Enables computing ms_ssim along with vmaf.
12625 Default value: @code{false}
12628 Set the pool method to be used for computing vmaf.
12629 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
12632 Set number of threads to be used when computing vmaf.
12633 Default value: @code{0}, which makes use of all available logical processors.
12636 Set interval for frame subsampling used when computing vmaf.
12637 Default value: @code{1}
12639 @item enable_conf_interval
12640 Enables confidence interval.
12641 Default value: @code{false}
12644 This filter also supports the @ref{framesync} options.
12646 @subsection Examples
12649 On the below examples the input file @file{main.mpg} being processed is
12650 compared with the reference file @file{ref.mpg}.
12653 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
12657 Example with options:
12659 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
12663 Example with options and different containers:
12665 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 -
12671 Limits the pixel components values to the specified range [min, max].
12673 The filter accepts the following options:
12677 Lower bound. Defaults to the lowest allowed value for the input.
12680 Upper bound. Defaults to the highest allowed value for the input.
12683 Specify which planes will be processed. Defaults to all available.
12690 The filter accepts the following options:
12694 Set the number of loops. Setting this value to -1 will result in infinite loops.
12698 Set maximal size in number of frames. Default is 0.
12701 Set first frame of loop. Default is 0.
12704 @subsection Examples
12708 Loop single first frame infinitely:
12710 loop=loop=-1:size=1:start=0
12714 Loop single first frame 10 times:
12716 loop=loop=10:size=1:start=0
12720 Loop 10 first frames 5 times:
12722 loop=loop=5:size=10:start=0
12728 Apply a 1D LUT to an input video.
12730 The filter accepts the following options:
12734 Set the 1D LUT file name.
12736 Currently supported formats:
12745 Select interpolation mode.
12747 Available values are:
12751 Use values from the nearest defined point.
12753 Interpolate values using the linear interpolation.
12755 Interpolate values using the cosine interpolation.
12757 Interpolate values using the cubic interpolation.
12759 Interpolate values using the spline interpolation.
12766 Apply a 3D LUT to an input video.
12768 The filter accepts the following options:
12772 Set the 3D LUT file name.
12774 Currently supported formats:
12788 Select interpolation mode.
12790 Available values are:
12794 Use values from the nearest defined point.
12796 Interpolate values using the 8 points defining a cube.
12798 Interpolate values using a tetrahedron.
12804 Turn certain luma values into transparency.
12806 The filter accepts the following options:
12810 Set the luma which will be used as base for transparency.
12811 Default value is @code{0}.
12814 Set the range of luma values to be keyed out.
12815 Default value is @code{0.01}.
12818 Set the range of softness. Default value is @code{0}.
12819 Use this to control gradual transition from zero to full transparency.
12822 @subsection Commands
12823 This filter supports same @ref{commands} as options.
12824 The command accepts the same syntax of the corresponding option.
12826 If the specified expression is not valid, it is kept at its current
12829 @section lut, lutrgb, lutyuv
12831 Compute a look-up table for binding each pixel component input value
12832 to an output value, and apply it to the input video.
12834 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
12835 to an RGB input video.
12837 These filters accept the following parameters:
12840 set first pixel component expression
12842 set second pixel component expression
12844 set third pixel component expression
12846 set fourth pixel component expression, corresponds to the alpha component
12849 set red component expression
12851 set green component expression
12853 set blue component expression
12855 alpha component expression
12858 set Y/luminance component expression
12860 set U/Cb component expression
12862 set V/Cr component expression
12865 Each of them specifies the expression to use for computing the lookup table for
12866 the corresponding pixel component values.
12868 The exact component associated to each of the @var{c*} options depends on the
12871 The @var{lut} filter requires either YUV or RGB pixel formats in input,
12872 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
12874 The expressions can contain the following constants and functions:
12879 The input width and height.
12882 The input value for the pixel component.
12885 The input value, clipped to the @var{minval}-@var{maxval} range.
12888 The maximum value for the pixel component.
12891 The minimum value for the pixel component.
12894 The negated value for the pixel component value, clipped to the
12895 @var{minval}-@var{maxval} range; it corresponds to the expression
12896 "maxval-clipval+minval".
12899 The computed value in @var{val}, clipped to the
12900 @var{minval}-@var{maxval} range.
12902 @item gammaval(gamma)
12903 The computed gamma correction value of the pixel component value,
12904 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
12906 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
12910 All expressions default to "val".
12912 @subsection Examples
12916 Negate input video:
12918 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
12919 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
12922 The above is the same as:
12924 lutrgb="r=negval:g=negval:b=negval"
12925 lutyuv="y=negval:u=negval:v=negval"
12935 Remove chroma components, turning the video into a graytone image:
12937 lutyuv="u=128:v=128"
12941 Apply a luma burning effect:
12947 Remove green and blue components:
12953 Set a constant alpha channel value on input:
12955 format=rgba,lutrgb=a="maxval-minval/2"
12959 Correct luminance gamma by a factor of 0.5:
12961 lutyuv=y=gammaval(0.5)
12965 Discard least significant bits of luma:
12967 lutyuv=y='bitand(val, 128+64+32)'
12971 Technicolor like effect:
12973 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
12977 @section lut2, tlut2
12979 The @code{lut2} filter takes two input streams and outputs one
12982 The @code{tlut2} (time lut2) filter takes two consecutive frames
12983 from one single stream.
12985 This filter accepts the following parameters:
12988 set first pixel component expression
12990 set second pixel component expression
12992 set third pixel component expression
12994 set fourth pixel component expression, corresponds to the alpha component
12997 set output bit depth, only available for @code{lut2} filter. By default is 0,
12998 which means bit depth is automatically picked from first input format.
13001 Each of them specifies the expression to use for computing the lookup table for
13002 the corresponding pixel component values.
13004 The exact component associated to each of the @var{c*} options depends on the
13007 The expressions can contain the following constants:
13012 The input width and height.
13015 The first input value for the pixel component.
13018 The second input value for the pixel component.
13021 The first input video bit depth.
13024 The second input video bit depth.
13027 All expressions default to "x".
13029 @subsection Examples
13033 Highlight differences between two RGB video streams:
13035 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)'
13039 Highlight differences between two YUV video streams:
13041 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)'
13045 Show max difference between two video streams:
13047 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)))'
13051 @section maskedclamp
13053 Clamp the first input stream with the second input and third input stream.
13055 Returns the value of first stream to be between second input
13056 stream - @code{undershoot} and third input stream + @code{overshoot}.
13058 This filter accepts the following options:
13061 Default value is @code{0}.
13064 Default value is @code{0}.
13067 Set which planes will be processed as bitmap, unprocessed planes will be
13068 copied from first stream.
13069 By default value 0xf, all planes will be processed.
13074 Merge the second and third input stream into output stream using absolute differences
13075 between second input stream and first input stream and absolute difference between
13076 third input stream and first input stream. The picked value will be from second input
13077 stream if second absolute difference is greater than first one or from third input stream
13080 This filter accepts the following options:
13083 Set which planes will be processed as bitmap, unprocessed planes will be
13084 copied from first stream.
13085 By default value 0xf, all planes will be processed.
13088 @section maskedmerge
13090 Merge the first input stream with the second input stream using per pixel
13091 weights in the third input stream.
13093 A value of 0 in the third stream pixel component means that pixel component
13094 from first stream is returned unchanged, while maximum value (eg. 255 for
13095 8-bit videos) means that pixel component from second stream is returned
13096 unchanged. Intermediate values define the amount of merging between both
13097 input stream's pixel components.
13099 This filter accepts the following options:
13102 Set which planes will be processed as bitmap, unprocessed planes will be
13103 copied from first stream.
13104 By default value 0xf, all planes will be processed.
13109 Merge the second and third input stream into output stream using absolute differences
13110 between second input stream and first input stream and absolute difference between
13111 third input stream and first input stream. The picked value will be from second input
13112 stream if second absolute difference is less than first one or from third input stream
13115 This filter accepts the following options:
13118 Set which planes will be processed as bitmap, unprocessed planes will be
13119 copied from first stream.
13120 By default value 0xf, all planes will be processed.
13124 Create mask from input video.
13126 For example it is useful to create motion masks after @code{tblend} filter.
13128 This filter accepts the following options:
13132 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
13135 Set high threshold. Any pixel component higher than this value will be set to max value
13136 allowed for current pixel format.
13139 Set planes to filter, by default all available planes are filtered.
13142 Fill all frame pixels with this value.
13145 Set max average pixel value for frame. If sum of all pixel components is higher that this
13146 average, output frame will be completely filled with value set by @var{fill} option.
13147 Typically useful for scene changes when used in combination with @code{tblend} filter.
13152 Apply motion-compensation deinterlacing.
13154 It needs one field per frame as input and must thus be used together
13155 with yadif=1/3 or equivalent.
13157 This filter accepts the following options:
13160 Set the deinterlacing mode.
13162 It accepts one of the following values:
13167 use iterative motion estimation
13169 like @samp{slow}, but use multiple reference frames.
13171 Default value is @samp{fast}.
13174 Set the picture field parity assumed for the input video. It must be
13175 one of the following values:
13179 assume top field first
13181 assume bottom field first
13184 Default value is @samp{bff}.
13187 Set per-block quantization parameter (QP) used by the internal
13190 Higher values should result in a smoother motion vector field but less
13191 optimal individual vectors. Default value is 1.
13196 Pick median pixel from certain rectangle defined by radius.
13198 This filter accepts the following options:
13202 Set horizontal radius size. Default value is @code{1}.
13203 Allowed range is integer from 1 to 127.
13206 Set which planes to process. Default is @code{15}, which is all available planes.
13209 Set vertical radius size. Default value is @code{0}.
13210 Allowed range is integer from 0 to 127.
13211 If it is 0, value will be picked from horizontal @code{radius} option.
13214 @subsection Commands
13215 This filter supports same @ref{commands} as options.
13216 The command accepts the same syntax of the corresponding option.
13218 If the specified expression is not valid, it is kept at its current
13221 @section mergeplanes
13223 Merge color channel components from several video streams.
13225 The filter accepts up to 4 input streams, and merge selected input
13226 planes to the output video.
13228 This filter accepts the following options:
13231 Set input to output plane mapping. Default is @code{0}.
13233 The mappings is specified as a bitmap. It should be specified as a
13234 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
13235 mapping for the first plane of the output stream. 'A' sets the number of
13236 the input stream to use (from 0 to 3), and 'a' the plane number of the
13237 corresponding input to use (from 0 to 3). The rest of the mappings is
13238 similar, 'Bb' describes the mapping for the output stream second
13239 plane, 'Cc' describes the mapping for the output stream third plane and
13240 'Dd' describes the mapping for the output stream fourth plane.
13243 Set output pixel format. Default is @code{yuva444p}.
13246 @subsection Examples
13250 Merge three gray video streams of same width and height into single video stream:
13252 [a0][a1][a2]mergeplanes=0x001020:yuv444p
13256 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
13258 [a0][a1]mergeplanes=0x00010210:yuva444p
13262 Swap Y and A plane in yuva444p stream:
13264 format=yuva444p,mergeplanes=0x03010200:yuva444p
13268 Swap U and V plane in yuv420p stream:
13270 format=yuv420p,mergeplanes=0x000201:yuv420p
13274 Cast a rgb24 clip to yuv444p:
13276 format=rgb24,mergeplanes=0x000102:yuv444p
13282 Estimate and export motion vectors using block matching algorithms.
13283 Motion vectors are stored in frame side data to be used by other filters.
13285 This filter accepts the following options:
13288 Specify the motion estimation method. Accepts one of the following values:
13292 Exhaustive search algorithm.
13294 Three step search algorithm.
13296 Two dimensional logarithmic search algorithm.
13298 New three step search algorithm.
13300 Four step search algorithm.
13302 Diamond search algorithm.
13304 Hexagon-based search algorithm.
13306 Enhanced predictive zonal search algorithm.
13308 Uneven multi-hexagon search algorithm.
13310 Default value is @samp{esa}.
13313 Macroblock size. Default @code{16}.
13316 Search parameter. Default @code{7}.
13319 @section midequalizer
13321 Apply Midway Image Equalization effect using two video streams.
13323 Midway Image Equalization adjusts a pair of images to have the same
13324 histogram, while maintaining their dynamics as much as possible. It's
13325 useful for e.g. matching exposures from a pair of stereo cameras.
13327 This filter has two inputs and one output, which must be of same pixel format, but
13328 may be of different sizes. The output of filter is first input adjusted with
13329 midway histogram of both inputs.
13331 This filter accepts the following option:
13335 Set which planes to process. Default is @code{15}, which is all available planes.
13338 @section minterpolate
13340 Convert the video to specified frame rate using motion interpolation.
13342 This filter accepts the following options:
13345 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}.
13348 Motion interpolation mode. Following values are accepted:
13351 Duplicate previous or next frame for interpolating new ones.
13353 Blend source frames. Interpolated frame is mean of previous and next frames.
13355 Motion compensated interpolation. Following options are effective when this mode is selected:
13359 Motion compensation mode. Following values are accepted:
13362 Overlapped block motion compensation.
13364 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
13366 Default mode is @samp{obmc}.
13369 Motion estimation mode. Following values are accepted:
13372 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
13374 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
13376 Default mode is @samp{bilat}.
13379 The algorithm to be used for motion estimation. Following values are accepted:
13382 Exhaustive search algorithm.
13384 Three step search algorithm.
13386 Two dimensional logarithmic search algorithm.
13388 New three step search algorithm.
13390 Four step search algorithm.
13392 Diamond search algorithm.
13394 Hexagon-based search algorithm.
13396 Enhanced predictive zonal search algorithm.
13398 Uneven multi-hexagon search algorithm.
13400 Default algorithm is @samp{epzs}.
13403 Macroblock size. Default @code{16}.
13406 Motion estimation search parameter. Default @code{32}.
13409 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).
13414 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:
13417 Disable scene change detection.
13419 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
13421 Default method is @samp{fdiff}.
13423 @item scd_threshold
13424 Scene change detection threshold. Default is @code{5.0}.
13429 Mix several video input streams into one video stream.
13431 A description of the accepted options follows.
13435 The number of inputs. If unspecified, it defaults to 2.
13438 Specify weight of each input video stream as sequence.
13439 Each weight is separated by space. If number of weights
13440 is smaller than number of @var{frames} last specified
13441 weight will be used for all remaining unset weights.
13444 Specify scale, if it is set it will be multiplied with sum
13445 of each weight multiplied with pixel values to give final destination
13446 pixel value. By default @var{scale} is auto scaled to sum of weights.
13449 Specify how end of stream is determined.
13452 The duration of the longest input. (default)
13455 The duration of the shortest input.
13458 The duration of the first input.
13462 @section mpdecimate
13464 Drop frames that do not differ greatly from the previous frame in
13465 order to reduce frame rate.
13467 The main use of this filter is for very-low-bitrate encoding
13468 (e.g. streaming over dialup modem), but it could in theory be used for
13469 fixing movies that were inverse-telecined incorrectly.
13471 A description of the accepted options follows.
13475 Set the maximum number of consecutive frames which can be dropped (if
13476 positive), or the minimum interval between dropped frames (if
13477 negative). If the value is 0, the frame is dropped disregarding the
13478 number of previous sequentially dropped frames.
13480 Default value is 0.
13485 Set the dropping threshold values.
13487 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
13488 represent actual pixel value differences, so a threshold of 64
13489 corresponds to 1 unit of difference for each pixel, or the same spread
13490 out differently over the block.
13492 A frame is a candidate for dropping if no 8x8 blocks differ by more
13493 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
13494 meaning the whole image) differ by more than a threshold of @option{lo}.
13496 Default value for @option{hi} is 64*12, default value for @option{lo} is
13497 64*5, and default value for @option{frac} is 0.33.
13503 Negate (invert) the input video.
13505 It accepts the following option:
13510 With value 1, it negates the alpha component, if present. Default value is 0.
13516 Denoise frames using Non-Local Means algorithm.
13518 Each pixel is adjusted by looking for other pixels with similar contexts. This
13519 context similarity is defined by comparing their surrounding patches of size
13520 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
13523 Note that the research area defines centers for patches, which means some
13524 patches will be made of pixels outside that research area.
13526 The filter accepts the following options.
13530 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
13533 Set patch size. Default is 7. Must be odd number in range [0, 99].
13536 Same as @option{p} but for chroma planes.
13538 The default value is @var{0} and means automatic.
13541 Set research size. Default is 15. Must be odd number in range [0, 99].
13544 Same as @option{r} but for chroma planes.
13546 The default value is @var{0} and means automatic.
13551 Deinterlace video using neural network edge directed interpolation.
13553 This filter accepts the following options:
13557 Mandatory option, without binary file filter can not work.
13558 Currently file can be found here:
13559 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
13562 Set which frames to deinterlace, by default it is @code{all}.
13563 Can be @code{all} or @code{interlaced}.
13566 Set mode of operation.
13568 Can be one of the following:
13572 Use frame flags, both fields.
13574 Use frame flags, single field.
13576 Use top field only.
13578 Use bottom field only.
13580 Use both fields, top first.
13582 Use both fields, bottom first.
13586 Set which planes to process, by default filter process all frames.
13589 Set size of local neighborhood around each pixel, used by the predictor neural
13592 Can be one of the following:
13605 Set the number of neurons in predictor neural network.
13606 Can be one of the following:
13617 Controls the number of different neural network predictions that are blended
13618 together to compute the final output value. Can be @code{fast}, default or
13622 Set which set of weights to use in the predictor.
13623 Can be one of the following:
13627 weights trained to minimize absolute error
13629 weights trained to minimize squared error
13633 Controls whether or not the prescreener neural network is used to decide
13634 which pixels should be processed by the predictor neural network and which
13635 can be handled by simple cubic interpolation.
13636 The prescreener is trained to know whether cubic interpolation will be
13637 sufficient for a pixel or whether it should be predicted by the predictor nn.
13638 The computational complexity of the prescreener nn is much less than that of
13639 the predictor nn. Since most pixels can be handled by cubic interpolation,
13640 using the prescreener generally results in much faster processing.
13641 The prescreener is pretty accurate, so the difference between using it and not
13642 using it is almost always unnoticeable.
13644 Can be one of the following:
13652 Default is @code{new}.
13655 Set various debugging flags.
13660 Force libavfilter not to use any of the specified pixel formats for the
13661 input to the next filter.
13663 It accepts the following parameters:
13667 A '|'-separated list of pixel format names, such as
13668 pix_fmts=yuv420p|monow|rgb24".
13672 @subsection Examples
13676 Force libavfilter to use a format different from @var{yuv420p} for the
13677 input to the vflip filter:
13679 noformat=pix_fmts=yuv420p,vflip
13683 Convert the input video to any of the formats not contained in the list:
13685 noformat=yuv420p|yuv444p|yuv410p
13691 Add noise on video input frame.
13693 The filter accepts the following options:
13701 Set noise seed for specific pixel component or all pixel components in case
13702 of @var{all_seed}. Default value is @code{123457}.
13704 @item all_strength, alls
13705 @item c0_strength, c0s
13706 @item c1_strength, c1s
13707 @item c2_strength, c2s
13708 @item c3_strength, c3s
13709 Set noise strength for specific pixel component or all pixel components in case
13710 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
13712 @item all_flags, allf
13713 @item c0_flags, c0f
13714 @item c1_flags, c1f
13715 @item c2_flags, c2f
13716 @item c3_flags, c3f
13717 Set pixel component flags or set flags for all components if @var{all_flags}.
13718 Available values for component flags are:
13721 averaged temporal noise (smoother)
13723 mix random noise with a (semi)regular pattern
13725 temporal noise (noise pattern changes between frames)
13727 uniform noise (gaussian otherwise)
13731 @subsection Examples
13733 Add temporal and uniform noise to input video:
13735 noise=alls=20:allf=t+u
13740 Normalize RGB video (aka histogram stretching, contrast stretching).
13741 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
13743 For each channel of each frame, the filter computes the input range and maps
13744 it linearly to the user-specified output range. The output range defaults
13745 to the full dynamic range from pure black to pure white.
13747 Temporal smoothing can be used on the input range to reduce flickering (rapid
13748 changes in brightness) caused when small dark or bright objects enter or leave
13749 the scene. This is similar to the auto-exposure (automatic gain control) on a
13750 video camera, and, like a video camera, it may cause a period of over- or
13751 under-exposure of the video.
13753 The R,G,B channels can be normalized independently, which may cause some
13754 color shifting, or linked together as a single channel, which prevents
13755 color shifting. Linked normalization preserves hue. Independent normalization
13756 does not, so it can be used to remove some color casts. Independent and linked
13757 normalization can be combined in any ratio.
13759 The normalize filter accepts the following options:
13764 Colors which define the output range. The minimum input value is mapped to
13765 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
13766 The defaults are black and white respectively. Specifying white for
13767 @var{blackpt} and black for @var{whitept} will give color-inverted,
13768 normalized video. Shades of grey can be used to reduce the dynamic range
13769 (contrast). Specifying saturated colors here can create some interesting
13773 The number of previous frames to use for temporal smoothing. The input range
13774 of each channel is smoothed using a rolling average over the current frame
13775 and the @var{smoothing} previous frames. The default is 0 (no temporal
13779 Controls the ratio of independent (color shifting) channel normalization to
13780 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
13781 independent. Defaults to 1.0 (fully independent).
13784 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
13785 expensive no-op. Defaults to 1.0 (full strength).
13789 @subsection Commands
13790 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
13791 The command accepts the same syntax of the corresponding option.
13793 If the specified expression is not valid, it is kept at its current
13796 @subsection Examples
13798 Stretch video contrast to use the full dynamic range, with no temporal
13799 smoothing; may flicker depending on the source content:
13801 normalize=blackpt=black:whitept=white:smoothing=0
13804 As above, but with 50 frames of temporal smoothing; flicker should be
13805 reduced, depending on the source content:
13807 normalize=blackpt=black:whitept=white:smoothing=50
13810 As above, but with hue-preserving linked channel normalization:
13812 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
13815 As above, but with half strength:
13817 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
13820 Map the darkest input color to red, the brightest input color to cyan:
13822 normalize=blackpt=red:whitept=cyan
13827 Pass the video source unchanged to the output.
13830 Optical Character Recognition
13832 This filter uses Tesseract for optical character recognition. To enable
13833 compilation of this filter, you need to configure FFmpeg with
13834 @code{--enable-libtesseract}.
13836 It accepts the following options:
13840 Set datapath to tesseract data. Default is to use whatever was
13841 set at installation.
13844 Set language, default is "eng".
13847 Set character whitelist.
13850 Set character blacklist.
13853 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
13854 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
13858 Apply a video transform using libopencv.
13860 To enable this filter, install the libopencv library and headers and
13861 configure FFmpeg with @code{--enable-libopencv}.
13863 It accepts the following parameters:
13868 The name of the libopencv filter to apply.
13870 @item filter_params
13871 The parameters to pass to the libopencv filter. If not specified, the default
13872 values are assumed.
13876 Refer to the official libopencv documentation for more precise
13878 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
13880 Several libopencv filters are supported; see the following subsections.
13885 Dilate an image by using a specific structuring element.
13886 It corresponds to the libopencv function @code{cvDilate}.
13888 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
13890 @var{struct_el} represents a structuring element, and has the syntax:
13891 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
13893 @var{cols} and @var{rows} represent the number of columns and rows of
13894 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
13895 point, and @var{shape} the shape for the structuring element. @var{shape}
13896 must be "rect", "cross", "ellipse", or "custom".
13898 If the value for @var{shape} is "custom", it must be followed by a
13899 string of the form "=@var{filename}". The file with name
13900 @var{filename} is assumed to represent a binary image, with each
13901 printable character corresponding to a bright pixel. When a custom
13902 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
13903 or columns and rows of the read file are assumed instead.
13905 The default value for @var{struct_el} is "3x3+0x0/rect".
13907 @var{nb_iterations} specifies the number of times the transform is
13908 applied to the image, and defaults to 1.
13912 # Use the default values
13915 # Dilate using a structuring element with a 5x5 cross, iterating two times
13916 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
13918 # Read the shape from the file diamond.shape, iterating two times.
13919 # The file diamond.shape may contain a pattern of characters like this
13925 # The specified columns and rows are ignored
13926 # but the anchor point coordinates are not
13927 ocv=dilate:0x0+2x2/custom=diamond.shape|2
13932 Erode an image by using a specific structuring element.
13933 It corresponds to the libopencv function @code{cvErode}.
13935 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
13936 with the same syntax and semantics as the @ref{dilate} filter.
13940 Smooth the input video.
13942 The filter takes the following parameters:
13943 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
13945 @var{type} is the type of smooth filter to apply, and must be one of
13946 the following values: "blur", "blur_no_scale", "median", "gaussian",
13947 or "bilateral". The default value is "gaussian".
13949 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
13950 depends on the smooth type. @var{param1} and
13951 @var{param2} accept integer positive values or 0. @var{param3} and
13952 @var{param4} accept floating point values.
13954 The default value for @var{param1} is 3. The default value for the
13955 other parameters is 0.
13957 These parameters correspond to the parameters assigned to the
13958 libopencv function @code{cvSmooth}.
13960 @section oscilloscope
13962 2D Video Oscilloscope.
13964 Useful to measure spatial impulse, step responses, chroma delays, etc.
13966 It accepts the following parameters:
13970 Set scope center x position.
13973 Set scope center y position.
13976 Set scope size, relative to frame diagonal.
13979 Set scope tilt/rotation.
13985 Set trace center x position.
13988 Set trace center y position.
13991 Set trace width, relative to width of frame.
13994 Set trace height, relative to height of frame.
13997 Set which components to trace. By default it traces first three components.
14000 Draw trace grid. By default is enabled.
14003 Draw some statistics. By default is enabled.
14006 Draw scope. By default is enabled.
14009 @subsection Commands
14010 This filter supports same @ref{commands} as options.
14011 The command accepts the same syntax of the corresponding option.
14013 If the specified expression is not valid, it is kept at its current
14016 @subsection Examples
14020 Inspect full first row of video frame.
14022 oscilloscope=x=0.5:y=0:s=1
14026 Inspect full last row of video frame.
14028 oscilloscope=x=0.5:y=1:s=1
14032 Inspect full 5th line of video frame of height 1080.
14034 oscilloscope=x=0.5:y=5/1080:s=1
14038 Inspect full last column of video frame.
14040 oscilloscope=x=1:y=0.5:s=1:t=1
14048 Overlay one video on top of another.
14050 It takes two inputs and has one output. The first input is the "main"
14051 video on which the second input is overlaid.
14053 It accepts the following parameters:
14055 A description of the accepted options follows.
14060 Set the expression for the x and y coordinates of the overlaid video
14061 on the main video. Default value is "0" for both expressions. In case
14062 the expression is invalid, it is set to a huge value (meaning that the
14063 overlay will not be displayed within the output visible area).
14066 See @ref{framesync}.
14069 Set when the expressions for @option{x}, and @option{y} are evaluated.
14071 It accepts the following values:
14074 only evaluate expressions once during the filter initialization or
14075 when a command is processed
14078 evaluate expressions for each incoming frame
14081 Default value is @samp{frame}.
14084 See @ref{framesync}.
14087 Set the format for the output video.
14089 It accepts the following values:
14092 force YUV420 output
14095 force YUV422 output
14098 force YUV444 output
14101 force packed RGB output
14104 force planar RGB output
14107 automatically pick format
14110 Default value is @samp{yuv420}.
14113 See @ref{framesync}.
14116 Set format of alpha of the overlaid video, it can be @var{straight} or
14117 @var{premultiplied}. Default is @var{straight}.
14120 The @option{x}, and @option{y} expressions can contain the following
14126 The main input width and height.
14130 The overlay input width and height.
14134 The computed values for @var{x} and @var{y}. They are evaluated for
14139 horizontal and vertical chroma subsample values of the output
14140 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
14144 the number of input frame, starting from 0
14147 the position in the file of the input frame, NAN if unknown
14150 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
14154 This filter also supports the @ref{framesync} options.
14156 Note that the @var{n}, @var{pos}, @var{t} variables are available only
14157 when evaluation is done @emph{per frame}, and will evaluate to NAN
14158 when @option{eval} is set to @samp{init}.
14160 Be aware that frames are taken from each input video in timestamp
14161 order, hence, if their initial timestamps differ, it is a good idea
14162 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
14163 have them begin in the same zero timestamp, as the example for
14164 the @var{movie} filter does.
14166 You can chain together more overlays but you should test the
14167 efficiency of such approach.
14169 @subsection Commands
14171 This filter supports the following commands:
14175 Modify the x and y of the overlay input.
14176 The command accepts the same syntax of the corresponding option.
14178 If the specified expression is not valid, it is kept at its current
14182 @subsection Examples
14186 Draw the overlay at 10 pixels from the bottom right corner of the main
14189 overlay=main_w-overlay_w-10:main_h-overlay_h-10
14192 Using named options the example above becomes:
14194 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
14198 Insert a transparent PNG logo in the bottom left corner of the input,
14199 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
14201 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
14205 Insert 2 different transparent PNG logos (second logo on bottom
14206 right corner) using the @command{ffmpeg} tool:
14208 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
14212 Add a transparent color layer on top of the main video; @code{WxH}
14213 must specify the size of the main input to the overlay filter:
14215 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
14219 Play an original video and a filtered version (here with the deshake
14220 filter) side by side using the @command{ffplay} tool:
14222 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
14225 The above command is the same as:
14227 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
14231 Make a sliding overlay appearing from the left to the right top part of the
14232 screen starting since time 2:
14234 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
14238 Compose output by putting two input videos side to side:
14240 ffmpeg -i left.avi -i right.avi -filter_complex "
14241 nullsrc=size=200x100 [background];
14242 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
14243 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
14244 [background][left] overlay=shortest=1 [background+left];
14245 [background+left][right] overlay=shortest=1:x=100 [left+right]
14250 Mask 10-20 seconds of a video by applying the delogo filter to a section
14252 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
14253 -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]'
14258 Chain several overlays in cascade:
14260 nullsrc=s=200x200 [bg];
14261 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
14262 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
14263 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
14264 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
14265 [in3] null, [mid2] overlay=100:100 [out0]
14272 Apply Overcomplete Wavelet denoiser.
14274 The filter accepts the following options:
14280 Larger depth values will denoise lower frequency components more, but
14281 slow down filtering.
14283 Must be an int in the range 8-16, default is @code{8}.
14285 @item luma_strength, ls
14288 Must be a double value in the range 0-1000, default is @code{1.0}.
14290 @item chroma_strength, cs
14291 Set chroma strength.
14293 Must be a double value in the range 0-1000, default is @code{1.0}.
14299 Add paddings to the input image, and place the original input at the
14300 provided @var{x}, @var{y} coordinates.
14302 It accepts the following parameters:
14307 Specify an expression for the size of the output image with the
14308 paddings added. If the value for @var{width} or @var{height} is 0, the
14309 corresponding input size is used for the output.
14311 The @var{width} expression can reference the value set by the
14312 @var{height} expression, and vice versa.
14314 The default value of @var{width} and @var{height} is 0.
14318 Specify the offsets to place the input image at within the padded area,
14319 with respect to the top/left border of the output image.
14321 The @var{x} expression can reference the value set by the @var{y}
14322 expression, and vice versa.
14324 The default value of @var{x} and @var{y} is 0.
14326 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
14327 so the input image is centered on the padded area.
14330 Specify the color of the padded area. For the syntax of this option,
14331 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
14332 manual,ffmpeg-utils}.
14334 The default value of @var{color} is "black".
14337 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
14339 It accepts the following values:
14343 Only evaluate expressions once during the filter initialization or when
14344 a command is processed.
14347 Evaluate expressions for each incoming frame.
14351 Default value is @samp{init}.
14354 Pad to aspect instead to a resolution.
14358 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
14359 options are expressions containing the following constants:
14364 The input video width and height.
14368 These are the same as @var{in_w} and @var{in_h}.
14372 The output width and height (the size of the padded area), as
14373 specified by the @var{width} and @var{height} expressions.
14377 These are the same as @var{out_w} and @var{out_h}.
14381 The x and y offsets as specified by the @var{x} and @var{y}
14382 expressions, or NAN if not yet specified.
14385 same as @var{iw} / @var{ih}
14388 input sample aspect ratio
14391 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
14395 The horizontal and vertical chroma subsample values. For example for the
14396 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
14399 @subsection Examples
14403 Add paddings with the color "violet" to the input video. The output video
14404 size is 640x480, and the top-left corner of the input video is placed at
14407 pad=640:480:0:40:violet
14410 The example above is equivalent to the following command:
14412 pad=width=640:height=480:x=0:y=40:color=violet
14416 Pad the input to get an output with dimensions increased by 3/2,
14417 and put the input video at the center of the padded area:
14419 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
14423 Pad the input to get a squared output with size equal to the maximum
14424 value between the input width and height, and put the input video at
14425 the center of the padded area:
14427 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
14431 Pad the input to get a final w/h ratio of 16:9:
14433 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
14437 In case of anamorphic video, in order to set the output display aspect
14438 correctly, it is necessary to use @var{sar} in the expression,
14439 according to the relation:
14441 (ih * X / ih) * sar = output_dar
14442 X = output_dar / sar
14445 Thus the previous example needs to be modified to:
14447 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
14451 Double the output size and put the input video in the bottom-right
14452 corner of the output padded area:
14454 pad="2*iw:2*ih:ow-iw:oh-ih"
14458 @anchor{palettegen}
14459 @section palettegen
14461 Generate one palette for a whole video stream.
14463 It accepts the following options:
14467 Set the maximum number of colors to quantize in the palette.
14468 Note: the palette will still contain 256 colors; the unused palette entries
14471 @item reserve_transparent
14472 Create a palette of 255 colors maximum and reserve the last one for
14473 transparency. Reserving the transparency color is useful for GIF optimization.
14474 If not set, the maximum of colors in the palette will be 256. You probably want
14475 to disable this option for a standalone image.
14478 @item transparency_color
14479 Set the color that will be used as background for transparency.
14482 Set statistics mode.
14484 It accepts the following values:
14487 Compute full frame histograms.
14489 Compute histograms only for the part that differs from previous frame. This
14490 might be relevant to give more importance to the moving part of your input if
14491 the background is static.
14493 Compute new histogram for each frame.
14496 Default value is @var{full}.
14499 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
14500 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
14501 color quantization of the palette. This information is also visible at
14502 @var{info} logging level.
14504 @subsection Examples
14508 Generate a representative palette of a given video using @command{ffmpeg}:
14510 ffmpeg -i input.mkv -vf palettegen palette.png
14514 @section paletteuse
14516 Use a palette to downsample an input video stream.
14518 The filter takes two inputs: one video stream and a palette. The palette must
14519 be a 256 pixels image.
14521 It accepts the following options:
14525 Select dithering mode. Available algorithms are:
14528 Ordered 8x8 bayer dithering (deterministic)
14530 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
14531 Note: this dithering is sometimes considered "wrong" and is included as a
14533 @item floyd_steinberg
14534 Floyd and Steingberg dithering (error diffusion)
14536 Frankie Sierra dithering v2 (error diffusion)
14538 Frankie Sierra dithering v2 "Lite" (error diffusion)
14541 Default is @var{sierra2_4a}.
14544 When @var{bayer} dithering is selected, this option defines the scale of the
14545 pattern (how much the crosshatch pattern is visible). A low value means more
14546 visible pattern for less banding, and higher value means less visible pattern
14547 at the cost of more banding.
14549 The option must be an integer value in the range [0,5]. Default is @var{2}.
14552 If set, define the zone to process
14556 Only the changing rectangle will be reprocessed. This is similar to GIF
14557 cropping/offsetting compression mechanism. This option can be useful for speed
14558 if only a part of the image is changing, and has use cases such as limiting the
14559 scope of the error diffusal @option{dither} to the rectangle that bounds the
14560 moving scene (it leads to more deterministic output if the scene doesn't change
14561 much, and as a result less moving noise and better GIF compression).
14564 Default is @var{none}.
14567 Take new palette for each output frame.
14569 @item alpha_threshold
14570 Sets the alpha threshold for transparency. Alpha values above this threshold
14571 will be treated as completely opaque, and values below this threshold will be
14572 treated as completely transparent.
14574 The option must be an integer value in the range [0,255]. Default is @var{128}.
14577 @subsection Examples
14581 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
14582 using @command{ffmpeg}:
14584 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
14588 @section perspective
14590 Correct perspective of video not recorded perpendicular to the screen.
14592 A description of the accepted parameters follows.
14603 Set coordinates expression for top left, top right, bottom left and bottom right corners.
14604 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
14605 If the @code{sense} option is set to @code{source}, then the specified points will be sent
14606 to the corners of the destination. If the @code{sense} option is set to @code{destination},
14607 then the corners of the source will be sent to the specified coordinates.
14609 The expressions can use the following variables:
14614 the width and height of video frame.
14618 Output frame count.
14621 @item interpolation
14622 Set interpolation for perspective correction.
14624 It accepts the following values:
14630 Default value is @samp{linear}.
14633 Set interpretation of coordinate options.
14635 It accepts the following values:
14639 Send point in the source specified by the given coordinates to
14640 the corners of the destination.
14642 @item 1, destination
14644 Send the corners of the source to the point in the destination specified
14645 by the given coordinates.
14647 Default value is @samp{source}.
14651 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
14653 It accepts the following values:
14656 only evaluate expressions once during the filter initialization or
14657 when a command is processed
14660 evaluate expressions for each incoming frame
14663 Default value is @samp{init}.
14668 Delay interlaced video by one field time so that the field order changes.
14670 The intended use is to fix PAL movies that have been captured with the
14671 opposite field order to the film-to-video transfer.
14673 A description of the accepted parameters follows.
14679 It accepts the following values:
14682 Capture field order top-first, transfer bottom-first.
14683 Filter will delay the bottom field.
14686 Capture field order bottom-first, transfer top-first.
14687 Filter will delay the top field.
14690 Capture and transfer with the same field order. This mode only exists
14691 for the documentation of the other options to refer to, but if you
14692 actually select it, the filter will faithfully do nothing.
14695 Capture field order determined automatically by field flags, transfer
14697 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
14698 basis using field flags. If no field information is available,
14699 then this works just like @samp{u}.
14702 Capture unknown or varying, transfer opposite.
14703 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
14704 analyzing the images and selecting the alternative that produces best
14705 match between the fields.
14708 Capture top-first, transfer unknown or varying.
14709 Filter selects among @samp{t} and @samp{p} using image analysis.
14712 Capture bottom-first, transfer unknown or varying.
14713 Filter selects among @samp{b} and @samp{p} using image analysis.
14716 Capture determined by field flags, transfer unknown or varying.
14717 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
14718 image analysis. If no field information is available, then this works just
14719 like @samp{U}. This is the default mode.
14722 Both capture and transfer unknown or varying.
14723 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
14727 @section photosensitivity
14728 Reduce various flashes in video, so to help users with epilepsy.
14730 It accepts the following options:
14733 Set how many frames to use when filtering. Default is 30.
14736 Set detection threshold factor. Default is 1.
14740 Set how many pixels to skip when sampling frames. Default is 1.
14741 Allowed range is from 1 to 1024.
14744 Leave frames unchanged. Default is disabled.
14747 @section pixdesctest
14749 Pixel format descriptor test filter, mainly useful for internal
14750 testing. The output video should be equal to the input video.
14754 format=monow, pixdesctest
14757 can be used to test the monowhite pixel format descriptor definition.
14761 Display sample values of color channels. Mainly useful for checking color
14762 and levels. Minimum supported resolution is 640x480.
14764 The filters accept the following options:
14768 Set scope X position, relative offset on X axis.
14771 Set scope Y position, relative offset on Y axis.
14780 Set window opacity. This window also holds statistics about pixel area.
14783 Set window X position, relative offset on X axis.
14786 Set window Y position, relative offset on Y axis.
14791 Enable the specified chain of postprocessing subfilters using libpostproc. This
14792 library should be automatically selected with a GPL build (@code{--enable-gpl}).
14793 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
14794 Each subfilter and some options have a short and a long name that can be used
14795 interchangeably, i.e. dr/dering are the same.
14797 The filters accept the following options:
14801 Set postprocessing subfilters string.
14804 All subfilters share common options to determine their scope:
14808 Honor the quality commands for this subfilter.
14811 Do chrominance filtering, too (default).
14814 Do luminance filtering only (no chrominance).
14817 Do chrominance filtering only (no luminance).
14820 These options can be appended after the subfilter name, separated by a '|'.
14822 Available subfilters are:
14825 @item hb/hdeblock[|difference[|flatness]]
14826 Horizontal deblocking filter
14829 Difference factor where higher values mean more deblocking (default: @code{32}).
14831 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14834 @item vb/vdeblock[|difference[|flatness]]
14835 Vertical deblocking filter
14838 Difference factor where higher values mean more deblocking (default: @code{32}).
14840 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14843 @item ha/hadeblock[|difference[|flatness]]
14844 Accurate horizontal deblocking filter
14847 Difference factor where higher values mean more deblocking (default: @code{32}).
14849 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14852 @item va/vadeblock[|difference[|flatness]]
14853 Accurate vertical deblocking filter
14856 Difference factor where higher values mean more deblocking (default: @code{32}).
14858 Flatness threshold where lower values mean more deblocking (default: @code{39}).
14862 The horizontal and vertical deblocking filters share the difference and
14863 flatness values so you cannot set different horizontal and vertical
14867 @item h1/x1hdeblock
14868 Experimental horizontal deblocking filter
14870 @item v1/x1vdeblock
14871 Experimental vertical deblocking filter
14876 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
14879 larger -> stronger filtering
14881 larger -> stronger filtering
14883 larger -> stronger filtering
14886 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
14889 Stretch luminance to @code{0-255}.
14892 @item lb/linblenddeint
14893 Linear blend deinterlacing filter that deinterlaces the given block by
14894 filtering all lines with a @code{(1 2 1)} filter.
14896 @item li/linipoldeint
14897 Linear interpolating deinterlacing filter that deinterlaces the given block by
14898 linearly interpolating every second line.
14900 @item ci/cubicipoldeint
14901 Cubic interpolating deinterlacing filter deinterlaces the given block by
14902 cubically interpolating every second line.
14904 @item md/mediandeint
14905 Median deinterlacing filter that deinterlaces the given block by applying a
14906 median filter to every second line.
14908 @item fd/ffmpegdeint
14909 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
14910 second line with a @code{(-1 4 2 4 -1)} filter.
14913 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
14914 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
14916 @item fq/forceQuant[|quantizer]
14917 Overrides the quantizer table from the input with the constant quantizer you
14925 Default pp filter combination (@code{hb|a,vb|a,dr|a})
14928 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
14931 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
14934 @subsection Examples
14938 Apply horizontal and vertical deblocking, deringing and automatic
14939 brightness/contrast:
14945 Apply default filters without brightness/contrast correction:
14951 Apply default filters and temporal denoiser:
14953 pp=default/tmpnoise|1|2|3
14957 Apply deblocking on luminance only, and switch vertical deblocking on or off
14958 automatically depending on available CPU time:
14965 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
14966 similar to spp = 6 with 7 point DCT, where only the center sample is
14969 The filter accepts the following options:
14973 Force a constant quantization parameter. It accepts an integer in range
14974 0 to 63. If not set, the filter will use the QP from the video stream
14978 Set thresholding mode. Available modes are:
14982 Set hard thresholding.
14984 Set soft thresholding (better de-ringing effect, but likely blurrier).
14986 Set medium thresholding (good results, default).
14990 @section premultiply
14991 Apply alpha premultiply effect to input video stream using first plane
14992 of second stream as alpha.
14994 Both streams must have same dimensions and same pixel format.
14996 The filter accepts the following option:
15000 Set which planes will be processed, unprocessed planes will be copied.
15001 By default value 0xf, all planes will be processed.
15004 Do not require 2nd input for processing, instead use alpha plane from input stream.
15008 Apply prewitt operator to input video stream.
15010 The filter accepts the following option:
15014 Set which planes will be processed, unprocessed planes will be copied.
15015 By default value 0xf, all planes will be processed.
15018 Set value which will be multiplied with filtered result.
15021 Set value which will be added to filtered result.
15024 @anchor{program_opencl}
15025 @section program_opencl
15027 Filter video using an OpenCL program.
15032 OpenCL program source file.
15035 Kernel name in program.
15038 Number of inputs to the filter. Defaults to 1.
15041 Size of output frames. Defaults to the same as the first input.
15045 The program source file must contain a kernel function with the given name,
15046 which will be run once for each plane of the output. Each run on a plane
15047 gets enqueued as a separate 2D global NDRange with one work-item for each
15048 pixel to be generated. The global ID offset for each work-item is therefore
15049 the coordinates of a pixel in the destination image.
15051 The kernel function needs to take the following arguments:
15054 Destination image, @var{__write_only image2d_t}.
15056 This image will become the output; the kernel should write all of it.
15058 Frame index, @var{unsigned int}.
15060 This is a counter starting from zero and increasing by one for each frame.
15062 Source images, @var{__read_only image2d_t}.
15064 These are the most recent images on each input. The kernel may read from
15065 them to generate the output, but they can't be written to.
15072 Copy the input to the output (output must be the same size as the input).
15074 __kernel void copy(__write_only image2d_t destination,
15075 unsigned int index,
15076 __read_only image2d_t source)
15078 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
15080 int2 location = (int2)(get_global_id(0), get_global_id(1));
15082 float4 value = read_imagef(source, sampler, location);
15084 write_imagef(destination, location, value);
15089 Apply a simple transformation, rotating the input by an amount increasing
15090 with the index counter. Pixel values are linearly interpolated by the
15091 sampler, and the output need not have the same dimensions as the input.
15093 __kernel void rotate_image(__write_only image2d_t dst,
15094 unsigned int index,
15095 __read_only image2d_t src)
15097 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
15098 CLK_FILTER_LINEAR);
15100 float angle = (float)index / 100.0f;
15102 float2 dst_dim = convert_float2(get_image_dim(dst));
15103 float2 src_dim = convert_float2(get_image_dim(src));
15105 float2 dst_cen = dst_dim / 2.0f;
15106 float2 src_cen = src_dim / 2.0f;
15108 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
15110 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
15112 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
15113 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
15115 src_pos = src_pos * src_dim / dst_dim;
15117 float2 src_loc = src_pos + src_cen;
15119 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
15120 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
15121 write_imagef(dst, dst_loc, 0.5f);
15123 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
15128 Blend two inputs together, with the amount of each input used varying
15129 with the index counter.
15131 __kernel void blend_images(__write_only image2d_t dst,
15132 unsigned int index,
15133 __read_only image2d_t src1,
15134 __read_only image2d_t src2)
15136 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
15137 CLK_FILTER_LINEAR);
15139 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
15141 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
15142 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
15143 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
15145 float4 val1 = read_imagef(src1, sampler, src1_loc);
15146 float4 val2 = read_imagef(src2, sampler, src2_loc);
15148 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
15154 @section pseudocolor
15156 Alter frame colors in video with pseudocolors.
15158 This filter accepts the following options:
15162 set pixel first component expression
15165 set pixel second component expression
15168 set pixel third component expression
15171 set pixel fourth component expression, corresponds to the alpha component
15174 set component to use as base for altering colors
15177 Each of them specifies the expression to use for computing the lookup table for
15178 the corresponding pixel component values.
15180 The expressions can contain the following constants and functions:
15185 The input width and height.
15188 The input value for the pixel component.
15190 @item ymin, umin, vmin, amin
15191 The minimum allowed component value.
15193 @item ymax, umax, vmax, amax
15194 The maximum allowed component value.
15197 All expressions default to "val".
15199 @subsection Examples
15203 Change too high luma values to gradient:
15205 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'"
15211 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
15212 Ratio) between two input videos.
15214 This filter takes in input two input videos, the first input is
15215 considered the "main" source and is passed unchanged to the
15216 output. The second input is used as a "reference" video for computing
15219 Both video inputs must have the same resolution and pixel format for
15220 this filter to work correctly. Also it assumes that both inputs
15221 have the same number of frames, which are compared one by one.
15223 The obtained average PSNR is printed through the logging system.
15225 The filter stores the accumulated MSE (mean squared error) of each
15226 frame, and at the end of the processing it is averaged across all frames
15227 equally, and the following formula is applied to obtain the PSNR:
15230 PSNR = 10*log10(MAX^2/MSE)
15233 Where MAX is the average of the maximum values of each component of the
15236 The description of the accepted parameters follows.
15239 @item stats_file, f
15240 If specified the filter will use the named file to save the PSNR of
15241 each individual frame. When filename equals "-" the data is sent to
15244 @item stats_version
15245 Specifies which version of the stats file format to use. Details of
15246 each format are written below.
15247 Default value is 1.
15249 @item stats_add_max
15250 Determines whether the max value is output to the stats log.
15251 Default value is 0.
15252 Requires stats_version >= 2. If this is set and stats_version < 2,
15253 the filter will return an error.
15256 This filter also supports the @ref{framesync} options.
15258 The file printed if @var{stats_file} is selected, contains a sequence of
15259 key/value pairs of the form @var{key}:@var{value} for each compared
15262 If a @var{stats_version} greater than 1 is specified, a header line precedes
15263 the list of per-frame-pair stats, with key value pairs following the frame
15264 format with the following parameters:
15267 @item psnr_log_version
15268 The version of the log file format. Will match @var{stats_version}.
15271 A comma separated list of the per-frame-pair parameters included in
15275 A description of each shown per-frame-pair parameter follows:
15279 sequential number of the input frame, starting from 1
15282 Mean Square Error pixel-by-pixel average difference of the compared
15283 frames, averaged over all the image components.
15285 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
15286 Mean Square Error pixel-by-pixel average difference of the compared
15287 frames for the component specified by the suffix.
15289 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
15290 Peak Signal to Noise ratio of the compared frames for the component
15291 specified by the suffix.
15293 @item max_avg, max_y, max_u, max_v
15294 Maximum allowed value for each channel, and average over all
15298 @subsection Examples
15303 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
15304 [main][ref] psnr="stats_file=stats.log" [out]
15307 On this example the input file being processed is compared with the
15308 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
15309 is stored in @file{stats.log}.
15312 Another example with different containers:
15314 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 -
15321 Pulldown reversal (inverse telecine) filter, capable of handling mixed
15322 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
15325 The pullup filter is designed to take advantage of future context in making
15326 its decisions. This filter is stateless in the sense that it does not lock
15327 onto a pattern to follow, but it instead looks forward to the following
15328 fields in order to identify matches and rebuild progressive frames.
15330 To produce content with an even framerate, insert the fps filter after
15331 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
15332 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
15334 The filter accepts the following options:
15341 These options set the amount of "junk" to ignore at the left, right, top, and
15342 bottom of the image, respectively. Left and right are in units of 8 pixels,
15343 while top and bottom are in units of 2 lines.
15344 The default is 8 pixels on each side.
15347 Set the strict breaks. Setting this option to 1 will reduce the chances of
15348 filter generating an occasional mismatched frame, but it may also cause an
15349 excessive number of frames to be dropped during high motion sequences.
15350 Conversely, setting it to -1 will make filter match fields more easily.
15351 This may help processing of video where there is slight blurring between
15352 the fields, but may also cause there to be interlaced frames in the output.
15353 Default value is @code{0}.
15356 Set the metric plane to use. It accepts the following values:
15362 Use chroma blue plane.
15365 Use chroma red plane.
15368 This option may be set to use chroma plane instead of the default luma plane
15369 for doing filter's computations. This may improve accuracy on very clean
15370 source material, but more likely will decrease accuracy, especially if there
15371 is chroma noise (rainbow effect) or any grayscale video.
15372 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
15373 load and make pullup usable in realtime on slow machines.
15376 For best results (without duplicated frames in the output file) it is
15377 necessary to change the output frame rate. For example, to inverse
15378 telecine NTSC input:
15380 ffmpeg -i input -vf pullup -r 24000/1001 ...
15385 Change video quantization parameters (QP).
15387 The filter accepts the following option:
15391 Set expression for quantization parameter.
15394 The expression is evaluated through the eval API and can contain, among others,
15395 the following constants:
15399 1 if index is not 129, 0 otherwise.
15402 Sequential index starting from -129 to 128.
15405 @subsection Examples
15409 Some equation like:
15417 Flush video frames from internal cache of frames into a random order.
15418 No frame is discarded.
15419 Inspired by @ref{frei0r} nervous filter.
15423 Set size in number of frames of internal cache, in range from @code{2} to
15424 @code{512}. Default is @code{30}.
15427 Set seed for random number generator, must be an integer included between
15428 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
15429 less than @code{0}, the filter will try to use a good random seed on a
15433 @section readeia608
15435 Read closed captioning (EIA-608) information from the top lines of a video frame.
15437 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
15438 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
15439 with EIA-608 data (starting from 0). A description of each metadata value follows:
15442 @item lavfi.readeia608.X.cc
15443 The two bytes stored as EIA-608 data (printed in hexadecimal).
15445 @item lavfi.readeia608.X.line
15446 The number of the line on which the EIA-608 data was identified and read.
15449 This filter accepts the following options:
15453 Set the line to start scanning for EIA-608 data. Default is @code{0}.
15456 Set the line to end scanning for EIA-608 data. Default is @code{29}.
15459 Set the ratio of width reserved for sync code detection.
15460 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
15463 Enable checking the parity bit. In the event of a parity error, the filter will output
15464 @code{0x00} for that character. Default is false.
15467 Lowpass lines prior to further processing. Default is enabled.
15470 @subsection Examples
15474 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
15476 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
15482 Read vertical interval timecode (VITC) information from the top lines of a
15485 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
15486 timecode value, if a valid timecode has been detected. Further metadata key
15487 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
15488 timecode data has been found or not.
15490 This filter accepts the following options:
15494 Set the maximum number of lines to scan for VITC data. If the value is set to
15495 @code{-1} the full video frame is scanned. Default is @code{45}.
15498 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
15499 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
15502 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
15503 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
15506 @subsection Examples
15510 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
15511 draw @code{--:--:--:--} as a placeholder:
15513 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
15519 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
15521 Destination pixel at position (X, Y) will be picked from source (x, y) position
15522 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
15523 value for pixel will be used for destination pixel.
15525 Xmap and Ymap input video streams must be of same dimensions. Output video stream
15526 will have Xmap/Ymap video stream dimensions.
15527 Xmap and Ymap input video streams are 16bit depth, single channel.
15531 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
15532 Default is @code{color}.
15535 @section removegrain
15537 The removegrain filter is a spatial denoiser for progressive video.
15541 Set mode for the first plane.
15544 Set mode for the second plane.
15547 Set mode for the third plane.
15550 Set mode for the fourth plane.
15553 Range of mode is from 0 to 24. Description of each mode follows:
15557 Leave input plane unchanged. Default.
15560 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
15563 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
15566 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
15569 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
15570 This is equivalent to a median filter.
15573 Line-sensitive clipping giving the minimal change.
15576 Line-sensitive clipping, intermediate.
15579 Line-sensitive clipping, intermediate.
15582 Line-sensitive clipping, intermediate.
15585 Line-sensitive clipping on a line where the neighbours pixels are the closest.
15588 Replaces the target pixel with the closest neighbour.
15591 [1 2 1] horizontal and vertical kernel blur.
15597 Bob mode, interpolates top field from the line where the neighbours
15598 pixels are the closest.
15601 Bob mode, interpolates bottom field from the line where the neighbours
15602 pixels are the closest.
15605 Bob mode, interpolates top field. Same as 13 but with a more complicated
15606 interpolation formula.
15609 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
15610 interpolation formula.
15613 Clips the pixel with the minimum and maximum of respectively the maximum and
15614 minimum of each pair of opposite neighbour pixels.
15617 Line-sensitive clipping using opposite neighbours whose greatest distance from
15618 the current pixel is minimal.
15621 Replaces the pixel with the average of its 8 neighbours.
15624 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
15627 Clips pixels using the averages of opposite neighbour.
15630 Same as mode 21 but simpler and faster.
15633 Small edge and halo removal, but reputed useless.
15639 @section removelogo
15641 Suppress a TV station logo, using an image file to determine which
15642 pixels comprise the logo. It works by filling in the pixels that
15643 comprise the logo with neighboring pixels.
15645 The filter accepts the following options:
15649 Set the filter bitmap file, which can be any image format supported by
15650 libavformat. The width and height of the image file must match those of the
15651 video stream being processed.
15654 Pixels in the provided bitmap image with a value of zero are not
15655 considered part of the logo, non-zero pixels are considered part of
15656 the logo. If you use white (255) for the logo and black (0) for the
15657 rest, you will be safe. For making the filter bitmap, it is
15658 recommended to take a screen capture of a black frame with the logo
15659 visible, and then using a threshold filter followed by the erode
15660 filter once or twice.
15662 If needed, little splotches can be fixed manually. Remember that if
15663 logo pixels are not covered, the filter quality will be much
15664 reduced. Marking too many pixels as part of the logo does not hurt as
15665 much, but it will increase the amount of blurring needed to cover over
15666 the image and will destroy more information than necessary, and extra
15667 pixels will slow things down on a large logo.
15669 @section repeatfields
15671 This filter uses the repeat_field flag from the Video ES headers and hard repeats
15672 fields based on its value.
15676 Reverse a video clip.
15678 Warning: This filter requires memory to buffer the entire clip, so trimming
15681 @subsection Examples
15685 Take the first 5 seconds of a clip, and reverse it.
15692 Shift R/G/B/A pixels horizontally and/or vertically.
15694 The filter accepts the following options:
15697 Set amount to shift red horizontally.
15699 Set amount to shift red vertically.
15701 Set amount to shift green horizontally.
15703 Set amount to shift green vertically.
15705 Set amount to shift blue horizontally.
15707 Set amount to shift blue vertically.
15709 Set amount to shift alpha horizontally.
15711 Set amount to shift alpha vertically.
15713 Set edge mode, can be @var{smear}, default, or @var{warp}.
15716 @subsection Commands
15718 This filter supports the all above options as @ref{commands}.
15721 Apply roberts cross operator to input video stream.
15723 The filter accepts the following option:
15727 Set which planes will be processed, unprocessed planes will be copied.
15728 By default value 0xf, all planes will be processed.
15731 Set value which will be multiplied with filtered result.
15734 Set value which will be added to filtered result.
15739 Rotate video by an arbitrary angle expressed in radians.
15741 The filter accepts the following options:
15743 A description of the optional parameters follows.
15746 Set an expression for the angle by which to rotate the input video
15747 clockwise, expressed as a number of radians. A negative value will
15748 result in a counter-clockwise rotation. By default it is set to "0".
15750 This expression is evaluated for each frame.
15753 Set the output width expression, default value is "iw".
15754 This expression is evaluated just once during configuration.
15757 Set the output height expression, default value is "ih".
15758 This expression is evaluated just once during configuration.
15761 Enable bilinear interpolation if set to 1, a value of 0 disables
15762 it. Default value is 1.
15765 Set the color used to fill the output area not covered by the rotated
15766 image. For the general syntax of this option, check the
15767 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
15768 If the special value "none" is selected then no
15769 background is printed (useful for example if the background is never shown).
15771 Default value is "black".
15774 The expressions for the angle and the output size can contain the
15775 following constants and functions:
15779 sequential number of the input frame, starting from 0. It is always NAN
15780 before the first frame is filtered.
15783 time in seconds of the input frame, it is set to 0 when the filter is
15784 configured. It is always NAN before the first frame is filtered.
15788 horizontal and vertical chroma subsample values. For example for the
15789 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15793 the input video width and height
15797 the output width and height, that is the size of the padded area as
15798 specified by the @var{width} and @var{height} expressions
15802 the minimal width/height required for completely containing the input
15803 video rotated by @var{a} radians.
15805 These are only available when computing the @option{out_w} and
15806 @option{out_h} expressions.
15809 @subsection Examples
15813 Rotate the input by PI/6 radians clockwise:
15819 Rotate the input by PI/6 radians counter-clockwise:
15825 Rotate the input by 45 degrees clockwise:
15831 Apply a constant rotation with period T, starting from an angle of PI/3:
15833 rotate=PI/3+2*PI*t/T
15837 Make the input video rotation oscillating with a period of T
15838 seconds and an amplitude of A radians:
15840 rotate=A*sin(2*PI/T*t)
15844 Rotate the video, output size is chosen so that the whole rotating
15845 input video is always completely contained in the output:
15847 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
15851 Rotate the video, reduce the output size so that no background is ever
15854 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
15858 @subsection Commands
15860 The filter supports the following commands:
15864 Set the angle expression.
15865 The command accepts the same syntax of the corresponding option.
15867 If the specified expression is not valid, it is kept at its current
15873 Apply Shape Adaptive Blur.
15875 The filter accepts the following options:
15878 @item luma_radius, lr
15879 Set luma blur filter strength, must be a value in range 0.1-4.0, default
15880 value is 1.0. A greater value will result in a more blurred image, and
15881 in slower processing.
15883 @item luma_pre_filter_radius, lpfr
15884 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
15887 @item luma_strength, ls
15888 Set luma maximum difference between pixels to still be considered, must
15889 be a value in the 0.1-100.0 range, default value is 1.0.
15891 @item chroma_radius, cr
15892 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
15893 greater value will result in a more blurred image, and in slower
15896 @item chroma_pre_filter_radius, cpfr
15897 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
15899 @item chroma_strength, cs
15900 Set chroma maximum difference between pixels to still be considered,
15901 must be a value in the -0.9-100.0 range.
15904 Each chroma option value, if not explicitly specified, is set to the
15905 corresponding luma option value.
15910 Scale (resize) the input video, using the libswscale library.
15912 The scale filter forces the output display aspect ratio to be the same
15913 of the input, by changing the output sample aspect ratio.
15915 If the input image format is different from the format requested by
15916 the next filter, the scale filter will convert the input to the
15919 @subsection Options
15920 The filter accepts the following options, or any of the options
15921 supported by the libswscale scaler.
15923 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
15924 the complete list of scaler options.
15929 Set the output video dimension expression. Default value is the input
15932 If the @var{width} or @var{w} value is 0, the input width is used for
15933 the output. If the @var{height} or @var{h} value is 0, the input height
15934 is used for the output.
15936 If one and only one of the values is -n with n >= 1, the scale filter
15937 will use a value that maintains the aspect ratio of the input image,
15938 calculated from the other specified dimension. After that it will,
15939 however, make sure that the calculated dimension is divisible by n and
15940 adjust the value if necessary.
15942 If both values are -n with n >= 1, the behavior will be identical to
15943 both values being set to 0 as previously detailed.
15945 See below for the list of accepted constants for use in the dimension
15949 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
15953 Only evaluate expressions once during the filter initialization or when a command is processed.
15956 Evaluate expressions for each incoming frame.
15960 Default value is @samp{init}.
15964 Set the interlacing mode. It accepts the following values:
15968 Force interlaced aware scaling.
15971 Do not apply interlaced scaling.
15974 Select interlaced aware scaling depending on whether the source frames
15975 are flagged as interlaced or not.
15978 Default value is @samp{0}.
15981 Set libswscale scaling flags. See
15982 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15983 complete list of values. If not explicitly specified the filter applies
15987 @item param0, param1
15988 Set libswscale input parameters for scaling algorithms that need them. See
15989 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
15990 complete documentation. If not explicitly specified the filter applies
15996 Set the video size. For the syntax of this option, check the
15997 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
15999 @item in_color_matrix
16000 @item out_color_matrix
16001 Set in/output YCbCr color space type.
16003 This allows the autodetected value to be overridden as well as allows forcing
16004 a specific value used for the output and encoder.
16006 If not specified, the color space type depends on the pixel format.
16012 Choose automatically.
16015 Format conforming to International Telecommunication Union (ITU)
16016 Recommendation BT.709.
16019 Set color space conforming to the United States Federal Communications
16020 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16025 Set color space conforming to:
16029 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16032 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16035 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16040 Set color space conforming to SMPTE ST 240:1999.
16043 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16048 Set in/output YCbCr sample range.
16050 This allows the autodetected value to be overridden as well as allows forcing
16051 a specific value used for the output and encoder. If not specified, the
16052 range depends on the pixel format. Possible values:
16056 Choose automatically.
16059 Set full range (0-255 in case of 8-bit luma).
16061 @item mpeg/limited/tv
16062 Set "MPEG" range (16-235 in case of 8-bit luma).
16065 @item force_original_aspect_ratio
16066 Enable decreasing or increasing output video width or height if necessary to
16067 keep the original aspect ratio. Possible values:
16071 Scale the video as specified and disable this feature.
16074 The output video dimensions will automatically be decreased if needed.
16077 The output video dimensions will automatically be increased if needed.
16081 One useful instance of this option is that when you know a specific device's
16082 maximum allowed resolution, you can use this to limit the output video to
16083 that, while retaining the aspect ratio. For example, device A allows
16084 1280x720 playback, and your video is 1920x800. Using this option (set it to
16085 decrease) and specifying 1280x720 to the command line makes the output
16088 Please note that this is a different thing than specifying -1 for @option{w}
16089 or @option{h}, you still need to specify the output resolution for this option
16092 @item force_divisible_by
16093 Ensures that both the output dimensions, width and height, are divisible by the
16094 given integer when used together with @option{force_original_aspect_ratio}. This
16095 works similar to using @code{-n} in the @option{w} and @option{h} options.
16097 This option respects the value set for @option{force_original_aspect_ratio},
16098 increasing or decreasing the resolution accordingly. The video's aspect ratio
16099 may be slightly modified.
16101 This option can be handy if you need to have a video fit within or exceed
16102 a defined resolution using @option{force_original_aspect_ratio} but also have
16103 encoder restrictions on width or height divisibility.
16107 The values of the @option{w} and @option{h} options are expressions
16108 containing the following constants:
16113 The input width and height
16117 These are the same as @var{in_w} and @var{in_h}.
16121 The output (scaled) width and height
16125 These are the same as @var{out_w} and @var{out_h}
16128 The same as @var{iw} / @var{ih}
16131 input sample aspect ratio
16134 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16138 horizontal and vertical input chroma subsample values. For example for the
16139 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16143 horizontal and vertical output chroma subsample values. For example for the
16144 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16147 The (sequential) number of the input frame, starting from 0.
16148 Only available with @code{eval=frame}.
16151 The presentation timestamp of the input frame, expressed as a number of
16152 seconds. Only available with @code{eval=frame}.
16155 The position (byte offset) of the frame in the input stream, or NaN if
16156 this information is unavailable and/or meaningless (for example in case of synthetic video).
16157 Only available with @code{eval=frame}.
16160 @subsection Examples
16164 Scale the input video to a size of 200x100
16169 This is equivalent to:
16180 Specify a size abbreviation for the output size:
16185 which can also be written as:
16191 Scale the input to 2x:
16193 scale=w=2*iw:h=2*ih
16197 The above is the same as:
16199 scale=2*in_w:2*in_h
16203 Scale the input to 2x with forced interlaced scaling:
16205 scale=2*iw:2*ih:interl=1
16209 Scale the input to half size:
16211 scale=w=iw/2:h=ih/2
16215 Increase the width, and set the height to the same size:
16221 Seek Greek harmony:
16228 Increase the height, and set the width to 3/2 of the height:
16230 scale=w=3/2*oh:h=3/5*ih
16234 Increase the size, making the size a multiple of the chroma
16237 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
16241 Increase the width to a maximum of 500 pixels,
16242 keeping the same aspect ratio as the input:
16244 scale=w='min(500\, iw*3/2):h=-1'
16248 Make pixels square by combining scale and setsar:
16250 scale='trunc(ih*dar):ih',setsar=1/1
16254 Make pixels square by combining scale and setsar,
16255 making sure the resulting resolution is even (required by some codecs):
16257 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
16261 @subsection Commands
16263 This filter supports the following commands:
16267 Set the output video dimension expression.
16268 The command accepts the same syntax of the corresponding option.
16270 If the specified expression is not valid, it is kept at its current
16276 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
16277 format conversion on CUDA video frames. Setting the output width and height
16278 works in the same way as for the @var{scale} filter.
16280 The following additional options are accepted:
16283 The pixel format of the output CUDA frames. If set to the string "same" (the
16284 default), the input format will be kept. Note that automatic format negotiation
16285 and conversion is not yet supported for hardware frames
16288 The interpolation algorithm used for resizing. One of the following:
16295 @item cubic2p_bspline
16296 2-parameter cubic (B=1, C=0)
16298 @item cubic2p_catmullrom
16299 2-parameter cubic (B=0, C=1/2)
16301 @item cubic2p_b05c03
16302 2-parameter cubic (B=1/2, C=3/10)
16310 @item force_original_aspect_ratio
16311 Enable decreasing or increasing output video width or height if necessary to
16312 keep the original aspect ratio. Possible values:
16316 Scale the video as specified and disable this feature.
16319 The output video dimensions will automatically be decreased if needed.
16322 The output video dimensions will automatically be increased if needed.
16326 One useful instance of this option is that when you know a specific device's
16327 maximum allowed resolution, you can use this to limit the output video to
16328 that, while retaining the aspect ratio. For example, device A allows
16329 1280x720 playback, and your video is 1920x800. Using this option (set it to
16330 decrease) and specifying 1280x720 to the command line makes the output
16333 Please note that this is a different thing than specifying -1 for @option{w}
16334 or @option{h}, you still need to specify the output resolution for this option
16337 @item force_divisible_by
16338 Ensures that both the output dimensions, width and height, are divisible by the
16339 given integer when used together with @option{force_original_aspect_ratio}. This
16340 works similar to using @code{-n} in the @option{w} and @option{h} options.
16342 This option respects the value set for @option{force_original_aspect_ratio},
16343 increasing or decreasing the resolution accordingly. The video's aspect ratio
16344 may be slightly modified.
16346 This option can be handy if you need to have a video fit within or exceed
16347 a defined resolution using @option{force_original_aspect_ratio} but also have
16348 encoder restrictions on width or height divisibility.
16354 Scale (resize) the input video, based on a reference video.
16356 See the scale filter for available options, scale2ref supports the same but
16357 uses the reference video instead of the main input as basis. scale2ref also
16358 supports the following additional constants for the @option{w} and
16359 @option{h} options:
16364 The main input video's width and height
16367 The same as @var{main_w} / @var{main_h}
16370 The main input video's sample aspect ratio
16372 @item main_dar, mdar
16373 The main input video's display aspect ratio. Calculated from
16374 @code{(main_w / main_h) * main_sar}.
16378 The main input video's horizontal and vertical chroma subsample values.
16379 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
16383 The (sequential) number of the main input frame, starting from 0.
16384 Only available with @code{eval=frame}.
16387 The presentation timestamp of the main input frame, expressed as a number of
16388 seconds. Only available with @code{eval=frame}.
16391 The position (byte offset) of the frame in the main input stream, or NaN if
16392 this information is unavailable and/or meaningless (for example in case of synthetic video).
16393 Only available with @code{eval=frame}.
16396 @subsection Examples
16400 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
16402 'scale2ref[b][a];[a][b]overlay'
16406 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
16408 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
16412 @subsection Commands
16414 This filter supports the following commands:
16418 Set the output video dimension expression.
16419 The command accepts the same syntax of the corresponding option.
16421 If the specified expression is not valid, it is kept at its current
16426 Scroll input video horizontally and/or vertically by constant speed.
16428 The filter accepts the following options:
16430 @item horizontal, h
16431 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
16432 Negative values changes scrolling direction.
16435 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
16436 Negative values changes scrolling direction.
16439 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
16442 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
16445 @subsection Commands
16447 This filter supports the following @ref{commands}:
16449 @item horizontal, h
16450 Set the horizontal scrolling speed.
16452 Set the vertical scrolling speed.
16455 @anchor{selectivecolor}
16456 @section selectivecolor
16458 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
16459 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
16460 by the "purity" of the color (that is, how saturated it already is).
16462 This filter is similar to the Adobe Photoshop Selective Color tool.
16464 The filter accepts the following options:
16467 @item correction_method
16468 Select color correction method.
16470 Available values are:
16473 Specified adjustments are applied "as-is" (added/subtracted to original pixel
16476 Specified adjustments are relative to the original component value.
16478 Default is @code{absolute}.
16480 Adjustments for red pixels (pixels where the red component is the maximum)
16482 Adjustments for yellow pixels (pixels where the blue component is the minimum)
16484 Adjustments for green pixels (pixels where the green component is the maximum)
16486 Adjustments for cyan pixels (pixels where the red component is the minimum)
16488 Adjustments for blue pixels (pixels where the blue component is the maximum)
16490 Adjustments for magenta pixels (pixels where the green component is the minimum)
16492 Adjustments for white pixels (pixels where all components are greater than 128)
16494 Adjustments for all pixels except pure black and pure white
16496 Adjustments for black pixels (pixels where all components are lesser than 128)
16498 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
16501 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
16502 4 space separated floating point adjustment values in the [-1,1] range,
16503 respectively to adjust the amount of cyan, magenta, yellow and black for the
16504 pixels of its range.
16506 @subsection Examples
16510 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
16511 increase magenta by 27% in blue areas:
16513 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
16517 Use a Photoshop selective color preset:
16519 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
16523 @anchor{separatefields}
16524 @section separatefields
16526 The @code{separatefields} takes a frame-based video input and splits
16527 each frame into its components fields, producing a new half height clip
16528 with twice the frame rate and twice the frame count.
16530 This filter use field-dominance information in frame to decide which
16531 of each pair of fields to place first in the output.
16532 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
16534 @section setdar, setsar
16536 The @code{setdar} filter sets the Display Aspect Ratio for the filter
16539 This is done by changing the specified Sample (aka Pixel) Aspect
16540 Ratio, according to the following equation:
16542 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
16545 Keep in mind that the @code{setdar} filter does not modify the pixel
16546 dimensions of the video frame. Also, the display aspect ratio set by
16547 this filter may be changed by later filters in the filterchain,
16548 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
16551 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
16552 the filter output video.
16554 Note that as a consequence of the application of this filter, the
16555 output display aspect ratio will change according to the equation
16558 Keep in mind that the sample aspect ratio set by the @code{setsar}
16559 filter may be changed by later filters in the filterchain, e.g. if
16560 another "setsar" or a "setdar" filter is applied.
16562 It accepts the following parameters:
16565 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
16566 Set the aspect ratio used by the filter.
16568 The parameter can be a floating point number string, an expression, or
16569 a string of the form @var{num}:@var{den}, where @var{num} and
16570 @var{den} are the numerator and denominator of the aspect ratio. If
16571 the parameter is not specified, it is assumed the value "0".
16572 In case the form "@var{num}:@var{den}" is used, the @code{:} character
16576 Set the maximum integer value to use for expressing numerator and
16577 denominator when reducing the expressed aspect ratio to a rational.
16578 Default value is @code{100}.
16582 The parameter @var{sar} is an expression containing
16583 the following constants:
16587 These are approximated values for the mathematical constants e
16588 (Euler's number), pi (Greek pi), and phi (the golden ratio).
16591 The input width and height.
16594 These are the same as @var{w} / @var{h}.
16597 The input sample aspect ratio.
16600 The input display aspect ratio. It is the same as
16601 (@var{w} / @var{h}) * @var{sar}.
16604 Horizontal and vertical chroma subsample values. For example, for the
16605 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16608 @subsection Examples
16613 To change the display aspect ratio to 16:9, specify one of the following:
16620 To change the sample aspect ratio to 10:11, specify:
16626 To set a display aspect ratio of 16:9, and specify a maximum integer value of
16627 1000 in the aspect ratio reduction, use the command:
16629 setdar=ratio=16/9:max=1000
16637 Force field for the output video frame.
16639 The @code{setfield} filter marks the interlace type field for the
16640 output frames. It does not change the input frame, but only sets the
16641 corresponding property, which affects how the frame is treated by
16642 following filters (e.g. @code{fieldorder} or @code{yadif}).
16644 The filter accepts the following options:
16649 Available values are:
16653 Keep the same field property.
16656 Mark the frame as bottom-field-first.
16659 Mark the frame as top-field-first.
16662 Mark the frame as progressive.
16669 Force frame parameter for the output video frame.
16671 The @code{setparams} filter marks interlace and color range for the
16672 output frames. It does not change the input frame, but only sets the
16673 corresponding property, which affects how the frame is treated by
16678 Available values are:
16682 Keep the same field property (default).
16685 Mark the frame as bottom-field-first.
16688 Mark the frame as top-field-first.
16691 Mark the frame as progressive.
16695 Available values are:
16699 Keep the same color range property (default).
16701 @item unspecified, unknown
16702 Mark the frame as unspecified color range.
16704 @item limited, tv, mpeg
16705 Mark the frame as limited range.
16707 @item full, pc, jpeg
16708 Mark the frame as full range.
16711 @item color_primaries
16712 Set the color primaries.
16713 Available values are:
16717 Keep the same color primaries property (default).
16734 Set the color transfer.
16735 Available values are:
16739 Keep the same color trc property (default).
16761 Set the colorspace.
16762 Available values are:
16766 Keep the same colorspace property (default).
16779 @item chroma-derived-nc
16780 @item chroma-derived-c
16787 Show a line containing various information for each input video frame.
16788 The input video is not modified.
16790 This filter supports the following options:
16794 Calculate checksums of each plane. By default enabled.
16797 The shown line contains a sequence of key/value pairs of the form
16798 @var{key}:@var{value}.
16800 The following values are shown in the output:
16804 The (sequential) number of the input frame, starting from 0.
16807 The Presentation TimeStamp of the input frame, expressed as a number of
16808 time base units. The time base unit depends on the filter input pad.
16811 The Presentation TimeStamp of the input frame, expressed as a number of
16815 The position of the frame in the input stream, or -1 if this information is
16816 unavailable and/or meaningless (for example in case of synthetic video).
16819 The pixel format name.
16822 The sample aspect ratio of the input frame, expressed in the form
16823 @var{num}/@var{den}.
16826 The size of the input frame. For the syntax of this option, check the
16827 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16830 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
16831 for bottom field first).
16834 This is 1 if the frame is a key frame, 0 otherwise.
16837 The picture type of the input frame ("I" for an I-frame, "P" for a
16838 P-frame, "B" for a B-frame, or "?" for an unknown type).
16839 Also refer to the documentation of the @code{AVPictureType} enum and of
16840 the @code{av_get_picture_type_char} function defined in
16841 @file{libavutil/avutil.h}.
16844 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
16846 @item plane_checksum
16847 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
16848 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
16851 The mean value of pixels in each plane of the input frame, expressed in the form
16852 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
16855 The standard deviation of pixel values in each plane of the input frame, expressed
16856 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
16860 @section showpalette
16862 Displays the 256 colors palette of each frame. This filter is only relevant for
16863 @var{pal8} pixel format frames.
16865 It accepts the following option:
16869 Set the size of the box used to represent one palette color entry. Default is
16870 @code{30} (for a @code{30x30} pixel box).
16873 @section shuffleframes
16875 Reorder and/or duplicate and/or drop video frames.
16877 It accepts the following parameters:
16881 Set the destination indexes of input frames.
16882 This is space or '|' separated list of indexes that maps input frames to output
16883 frames. Number of indexes also sets maximal value that each index may have.
16884 '-1' index have special meaning and that is to drop frame.
16887 The first frame has the index 0. The default is to keep the input unchanged.
16889 @subsection Examples
16893 Swap second and third frame of every three frames of the input:
16895 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
16899 Swap 10th and 1st frame of every ten frames of the input:
16901 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
16905 @section shuffleplanes
16907 Reorder and/or duplicate video planes.
16909 It accepts the following parameters:
16914 The index of the input plane to be used as the first output plane.
16917 The index of the input plane to be used as the second output plane.
16920 The index of the input plane to be used as the third output plane.
16923 The index of the input plane to be used as the fourth output plane.
16927 The first plane has the index 0. The default is to keep the input unchanged.
16929 @subsection Examples
16933 Swap the second and third planes of the input:
16935 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
16939 @anchor{signalstats}
16940 @section signalstats
16941 Evaluate various visual metrics that assist in determining issues associated
16942 with the digitization of analog video media.
16944 By default the filter will log these metadata values:
16948 Display the minimal Y value contained within the input frame. Expressed in
16952 Display the Y value at the 10% percentile within the input frame. Expressed in
16956 Display the average Y value within the input frame. Expressed in range of
16960 Display the Y value at the 90% percentile within the input frame. Expressed in
16964 Display the maximum Y value contained within the input frame. Expressed in
16968 Display the minimal U value contained within the input frame. Expressed in
16972 Display the U value at the 10% percentile within the input frame. Expressed in
16976 Display the average U value within the input frame. Expressed in range of
16980 Display the U value at the 90% percentile within the input frame. Expressed in
16984 Display the maximum U value contained within the input frame. Expressed in
16988 Display the minimal V value contained within the input frame. Expressed in
16992 Display the V value at the 10% percentile within the input frame. Expressed in
16996 Display the average V value within the input frame. Expressed in range of
17000 Display the V value at the 90% percentile within the input frame. Expressed in
17004 Display the maximum V value contained within the input frame. Expressed in
17008 Display the minimal saturation value contained within the input frame.
17009 Expressed in range of [0-~181.02].
17012 Display the saturation value at the 10% percentile within the input frame.
17013 Expressed in range of [0-~181.02].
17016 Display the average saturation value within the input frame. Expressed in range
17020 Display the saturation value at the 90% percentile within the input frame.
17021 Expressed in range of [0-~181.02].
17024 Display the maximum saturation value contained within the input frame.
17025 Expressed in range of [0-~181.02].
17028 Display the median value for hue within the input frame. Expressed in range of
17032 Display the average value for hue within the input frame. Expressed in range of
17036 Display the average of sample value difference between all values of the Y
17037 plane in the current frame and corresponding values of the previous input frame.
17038 Expressed in range of [0-255].
17041 Display the average of sample value difference between all values of the U
17042 plane in the current frame and corresponding values of the previous input frame.
17043 Expressed in range of [0-255].
17046 Display the average of sample value difference between all values of the V
17047 plane in the current frame and corresponding values of the previous input frame.
17048 Expressed in range of [0-255].
17051 Display bit depth of Y plane in current frame.
17052 Expressed in range of [0-16].
17055 Display bit depth of U plane in current frame.
17056 Expressed in range of [0-16].
17059 Display bit depth of V plane in current frame.
17060 Expressed in range of [0-16].
17063 The filter accepts the following options:
17069 @option{stat} specify an additional form of image analysis.
17070 @option{out} output video with the specified type of pixel highlighted.
17072 Both options accept the following values:
17076 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17077 unlike the neighboring pixels of the same field. Examples of temporal outliers
17078 include the results of video dropouts, head clogs, or tape tracking issues.
17081 Identify @var{vertical line repetition}. Vertical line repetition includes
17082 similar rows of pixels within a frame. In born-digital video vertical line
17083 repetition is common, but this pattern is uncommon in video digitized from an
17084 analog source. When it occurs in video that results from the digitization of an
17085 analog source it can indicate concealment from a dropout compensator.
17088 Identify pixels that fall outside of legal broadcast range.
17092 Set the highlight color for the @option{out} option. The default color is
17096 @subsection Examples
17100 Output data of various video metrics:
17102 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17106 Output specific data about the minimum and maximum values of the Y plane per frame:
17108 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17112 Playback video while highlighting pixels that are outside of broadcast range in red.
17114 ffplay example.mov -vf signalstats="out=brng:color=red"
17118 Playback video with signalstats metadata drawn over the frame.
17120 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17123 The contents of signalstat_drawtext.txt used in the command are:
17126 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17127 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
17128 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
17129 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
17137 Calculates the MPEG-7 Video Signature. The filter can handle more than one
17138 input. In this case the matching between the inputs can be calculated additionally.
17139 The filter always passes through the first input. The signature of each stream can
17140 be written into a file.
17142 It accepts the following options:
17146 Enable or disable the matching process.
17148 Available values are:
17152 Disable the calculation of a matching (default).
17154 Calculate the matching for the whole video and output whether the whole video
17155 matches or only parts.
17157 Calculate only until a matching is found or the video ends. Should be faster in
17162 Set the number of inputs. The option value must be a non negative integer.
17163 Default value is 1.
17166 Set the path to which the output is written. If there is more than one input,
17167 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
17168 integer), that will be replaced with the input number. If no filename is
17169 specified, no output will be written. This is the default.
17172 Choose the output format.
17174 Available values are:
17178 Use the specified binary representation (default).
17180 Use the specified xml representation.
17184 Set threshold to detect one word as similar. The option value must be an integer
17185 greater than zero. The default value is 9000.
17188 Set threshold to detect all words as similar. The option value must be an integer
17189 greater than zero. The default value is 60000.
17192 Set threshold to detect frames as similar. The option value must be an integer
17193 greater than zero. The default value is 116.
17196 Set the minimum length of a sequence in frames to recognize it as matching
17197 sequence. The option value must be a non negative integer value.
17198 The default value is 0.
17201 Set the minimum relation, that matching frames to all frames must have.
17202 The option value must be a double value between 0 and 1. The default value is 0.5.
17205 @subsection Examples
17209 To calculate the signature of an input video and store it in signature.bin:
17211 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
17215 To detect whether two videos match and store the signatures in XML format in
17216 signature0.xml and signature1.xml:
17218 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 -
17226 Blur the input video without impacting the outlines.
17228 It accepts the following options:
17231 @item luma_radius, lr
17232 Set the luma radius. The option value must be a float number in
17233 the range [0.1,5.0] that specifies the variance of the gaussian filter
17234 used to blur the image (slower if larger). Default value is 1.0.
17236 @item luma_strength, ls
17237 Set the luma strength. The option value must be a float number
17238 in the range [-1.0,1.0] that configures the blurring. A value included
17239 in [0.0,1.0] will blur the image whereas a value included in
17240 [-1.0,0.0] will sharpen the image. Default value is 1.0.
17242 @item luma_threshold, lt
17243 Set the luma threshold used as a coefficient to determine
17244 whether a pixel should be blurred or not. The option value must be an
17245 integer in the range [-30,30]. A value of 0 will filter all the image,
17246 a value included in [0,30] will filter flat areas and a value included
17247 in [-30,0] will filter edges. Default value is 0.
17249 @item chroma_radius, cr
17250 Set the chroma radius. The option value must be a float number in
17251 the range [0.1,5.0] that specifies the variance of the gaussian filter
17252 used to blur the image (slower if larger). Default value is @option{luma_radius}.
17254 @item chroma_strength, cs
17255 Set the chroma strength. The option value must be a float number
17256 in the range [-1.0,1.0] that configures the blurring. A value included
17257 in [0.0,1.0] will blur the image whereas a value included in
17258 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
17260 @item chroma_threshold, ct
17261 Set the chroma threshold used as a coefficient to determine
17262 whether a pixel should be blurred or not. The option value must be an
17263 integer in the range [-30,30]. A value of 0 will filter all the image,
17264 a value included in [0,30] will filter flat areas and a value included
17265 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
17268 If a chroma option is not explicitly set, the corresponding luma value
17272 Apply sobel operator to input video stream.
17274 The filter accepts the following option:
17278 Set which planes will be processed, unprocessed planes will be copied.
17279 By default value 0xf, all planes will be processed.
17282 Set value which will be multiplied with filtered result.
17285 Set value which will be added to filtered result.
17291 Apply a simple postprocessing filter that compresses and decompresses the image
17292 at several (or - in the case of @option{quality} level @code{6} - all) shifts
17293 and average the results.
17295 The filter accepts the following options:
17299 Set quality. This option defines the number of levels for averaging. It accepts
17300 an integer in the range 0-6. If set to @code{0}, the filter will have no
17301 effect. A value of @code{6} means the higher quality. For each increment of
17302 that value the speed drops by a factor of approximately 2. Default value is
17306 Force a constant quantization parameter. If not set, the filter will use the QP
17307 from the video stream (if available).
17310 Set thresholding mode. Available modes are:
17314 Set hard thresholding (default).
17316 Set soft thresholding (better de-ringing effect, but likely blurrier).
17319 @item use_bframe_qp
17320 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
17321 option may cause flicker since the B-Frames have often larger QP. Default is
17322 @code{0} (not enabled).
17325 @subsection Commands
17327 This filter supports the following commands:
17329 @item quality, level
17330 Set quality level. The value @code{max} can be used to set the maximum level,
17331 currently @code{6}.
17336 Scale the input by applying one of the super-resolution methods based on
17337 convolutional neural networks. Supported models:
17341 Super-Resolution Convolutional Neural Network model (SRCNN).
17342 See @url{https://arxiv.org/abs/1501.00092}.
17345 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
17346 See @url{https://arxiv.org/abs/1609.05158}.
17349 Training scripts as well as scripts for model file (.pb) saving can be found at
17350 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
17351 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
17353 Native model files (.model) can be generated from TensorFlow model
17354 files (.pb) by using tools/python/convert.py
17356 The filter accepts the following options:
17360 Specify which DNN backend to use for model loading and execution. This option accepts
17361 the following values:
17365 Native implementation of DNN loading and execution.
17368 TensorFlow backend. To enable this backend you
17369 need to install the TensorFlow for C library (see
17370 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
17371 @code{--enable-libtensorflow}
17374 Default value is @samp{native}.
17377 Set path to model file specifying network architecture and its parameters.
17378 Note that different backends use different file formats. TensorFlow backend
17379 can load files for both formats, while native backend can load files for only
17383 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
17384 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
17385 input upscaled using bicubic upscaling with proper scale factor.
17390 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
17392 This filter takes in input two input videos, the first input is
17393 considered the "main" source and is passed unchanged to the
17394 output. The second input is used as a "reference" video for computing
17397 Both video inputs must have the same resolution and pixel format for
17398 this filter to work correctly. Also it assumes that both inputs
17399 have the same number of frames, which are compared one by one.
17401 The filter stores the calculated SSIM of each frame.
17403 The description of the accepted parameters follows.
17406 @item stats_file, f
17407 If specified the filter will use the named file to save the SSIM of
17408 each individual frame. When filename equals "-" the data is sent to
17412 The file printed if @var{stats_file} is selected, contains a sequence of
17413 key/value pairs of the form @var{key}:@var{value} for each compared
17416 A description of each shown parameter follows:
17420 sequential number of the input frame, starting from 1
17422 @item Y, U, V, R, G, B
17423 SSIM of the compared frames for the component specified by the suffix.
17426 SSIM of the compared frames for the whole frame.
17429 Same as above but in dB representation.
17432 This filter also supports the @ref{framesync} options.
17434 @subsection Examples
17439 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
17440 [main][ref] ssim="stats_file=stats.log" [out]
17443 On this example the input file being processed is compared with the
17444 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
17445 is stored in @file{stats.log}.
17448 Another example with both psnr and ssim at same time:
17450 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
17454 Another example with different containers:
17456 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 -
17462 Convert between different stereoscopic image formats.
17464 The filters accept the following options:
17468 Set stereoscopic image format of input.
17470 Available values for input image formats are:
17473 side by side parallel (left eye left, right eye right)
17476 side by side crosseye (right eye left, left eye right)
17479 side by side parallel with half width resolution
17480 (left eye left, right eye right)
17483 side by side crosseye with half width resolution
17484 (right eye left, left eye right)
17488 above-below (left eye above, right eye below)
17492 above-below (right eye above, left eye below)
17496 above-below with half height resolution
17497 (left eye above, right eye below)
17501 above-below with half height resolution
17502 (right eye above, left eye below)
17505 alternating frames (left eye first, right eye second)
17508 alternating frames (right eye first, left eye second)
17511 interleaved rows (left eye has top row, right eye starts on next row)
17514 interleaved rows (right eye has top row, left eye starts on next row)
17517 interleaved columns, left eye first
17520 interleaved columns, right eye first
17522 Default value is @samp{sbsl}.
17526 Set stereoscopic image format of output.
17530 side by side parallel (left eye left, right eye right)
17533 side by side crosseye (right eye left, left eye right)
17536 side by side parallel with half width resolution
17537 (left eye left, right eye right)
17540 side by side crosseye with half width resolution
17541 (right eye left, left eye right)
17545 above-below (left eye above, right eye below)
17549 above-below (right eye above, left eye below)
17553 above-below with half height resolution
17554 (left eye above, right eye below)
17558 above-below with half height resolution
17559 (right eye above, left eye below)
17562 alternating frames (left eye first, right eye second)
17565 alternating frames (right eye first, left eye second)
17568 interleaved rows (left eye has top row, right eye starts on next row)
17571 interleaved rows (right eye has top row, left eye starts on next row)
17574 anaglyph red/blue gray
17575 (red filter on left eye, blue filter on right eye)
17578 anaglyph red/green gray
17579 (red filter on left eye, green filter on right eye)
17582 anaglyph red/cyan gray
17583 (red filter on left eye, cyan filter on right eye)
17586 anaglyph red/cyan half colored
17587 (red filter on left eye, cyan filter on right eye)
17590 anaglyph red/cyan color
17591 (red filter on left eye, cyan filter on right eye)
17594 anaglyph red/cyan color optimized with the least squares projection of dubois
17595 (red filter on left eye, cyan filter on right eye)
17598 anaglyph green/magenta gray
17599 (green filter on left eye, magenta filter on right eye)
17602 anaglyph green/magenta half colored
17603 (green filter on left eye, magenta filter on right eye)
17606 anaglyph green/magenta colored
17607 (green filter on left eye, magenta filter on right eye)
17610 anaglyph green/magenta color optimized with the least squares projection of dubois
17611 (green filter on left eye, magenta filter on right eye)
17614 anaglyph yellow/blue gray
17615 (yellow filter on left eye, blue filter on right eye)
17618 anaglyph yellow/blue half colored
17619 (yellow filter on left eye, blue filter on right eye)
17622 anaglyph yellow/blue colored
17623 (yellow filter on left eye, blue filter on right eye)
17626 anaglyph yellow/blue color optimized with the least squares projection of dubois
17627 (yellow filter on left eye, blue filter on right eye)
17630 mono output (left eye only)
17633 mono output (right eye only)
17636 checkerboard, left eye first
17639 checkerboard, right eye first
17642 interleaved columns, left eye first
17645 interleaved columns, right eye first
17651 Default value is @samp{arcd}.
17654 @subsection Examples
17658 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
17664 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
17670 @section streamselect, astreamselect
17671 Select video or audio streams.
17673 The filter accepts the following options:
17677 Set number of inputs. Default is 2.
17680 Set input indexes to remap to outputs.
17683 @subsection Commands
17685 The @code{streamselect} and @code{astreamselect} filter supports the following
17690 Set input indexes to remap to outputs.
17693 @subsection Examples
17697 Select first 5 seconds 1st stream and rest of time 2nd stream:
17699 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
17703 Same as above, but for audio:
17705 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
17712 Draw subtitles on top of input video using the libass library.
17714 To enable compilation of this filter you need to configure FFmpeg with
17715 @code{--enable-libass}. This filter also requires a build with libavcodec and
17716 libavformat to convert the passed subtitles file to ASS (Advanced Substation
17717 Alpha) subtitles format.
17719 The filter accepts the following options:
17723 Set the filename of the subtitle file to read. It must be specified.
17725 @item original_size
17726 Specify the size of the original video, the video for which the ASS file
17727 was composed. For the syntax of this option, check the
17728 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17729 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
17730 correctly scale the fonts if the aspect ratio has been changed.
17733 Set a directory path containing fonts that can be used by the filter.
17734 These fonts will be used in addition to whatever the font provider uses.
17737 Process alpha channel, by default alpha channel is untouched.
17740 Set subtitles input character encoding. @code{subtitles} filter only. Only
17741 useful if not UTF-8.
17743 @item stream_index, si
17744 Set subtitles stream index. @code{subtitles} filter only.
17747 Override default style or script info parameters of the subtitles. It accepts a
17748 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
17751 If the first key is not specified, it is assumed that the first value
17752 specifies the @option{filename}.
17754 For example, to render the file @file{sub.srt} on top of the input
17755 video, use the command:
17760 which is equivalent to:
17762 subtitles=filename=sub.srt
17765 To render the default subtitles stream from file @file{video.mkv}, use:
17767 subtitles=video.mkv
17770 To render the second subtitles stream from that file, use:
17772 subtitles=video.mkv:si=1
17775 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
17776 @code{DejaVu Serif}, use:
17778 subtitles=sub.srt:force_style='FontName=DejaVu Serif,PrimaryColour=&HCCFF0000'
17781 @section super2xsai
17783 Scale the input by 2x and smooth using the Super2xSaI (Scale and
17784 Interpolate) pixel art scaling algorithm.
17786 Useful for enlarging pixel art images without reducing sharpness.
17790 Swap two rectangular objects in video.
17792 This filter accepts the following options:
17802 Set 1st rect x coordinate.
17805 Set 1st rect y coordinate.
17808 Set 2nd rect x coordinate.
17811 Set 2nd rect y coordinate.
17813 All expressions are evaluated once for each frame.
17816 The all options are expressions containing the following constants:
17821 The input width and height.
17824 same as @var{w} / @var{h}
17827 input sample aspect ratio
17830 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
17833 The number of the input frame, starting from 0.
17836 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
17839 the position in the file of the input frame, NAN if unknown
17847 Apply telecine process to the video.
17849 This filter accepts the following options:
17858 The default value is @code{top}.
17862 A string of numbers representing the pulldown pattern you wish to apply.
17863 The default value is @code{23}.
17867 Some typical patterns:
17872 24p: 2332 (preferred)
17879 24p: 222222222223 ("Euro pulldown")
17884 @section thistogram
17886 Compute and draw a color distribution histogram for the input video across time.
17888 Unlike @ref{histogram} video filter which only shows histogram of single input frame
17889 at certain time, this filter shows also past histograms of number of frames defined
17890 by @code{width} option.
17892 The computed histogram is a representation of the color component
17893 distribution in an image.
17895 The filter accepts the following options:
17899 Set width of single color component output. Default value is @code{0}.
17900 Value of @code{0} means width will be picked from input video.
17901 This also set number of passed histograms to keep.
17902 Allowed range is [0, 8192].
17904 @item display_mode, d
17906 It accepts the following values:
17909 Per color component graphs are placed below each other.
17912 Per color component graphs are placed side by side.
17915 Presents information identical to that in the @code{parade}, except
17916 that the graphs representing color components are superimposed directly
17919 Default is @code{stack}.
17921 @item levels_mode, m
17922 Set mode. Can be either @code{linear}, or @code{logarithmic}.
17923 Default is @code{linear}.
17925 @item components, c
17926 Set what color components to display.
17927 Default is @code{7}.
17930 Set background opacity. Default is @code{0.9}.
17933 Show envelope. Default is disabled.
17936 Set envelope color. Default is @code{gold}.
17941 Apply threshold effect to video stream.
17943 This filter needs four video streams to perform thresholding.
17944 First stream is stream we are filtering.
17945 Second stream is holding threshold values, third stream is holding min values,
17946 and last, fourth stream is holding max values.
17948 The filter accepts the following option:
17952 Set which planes will be processed, unprocessed planes will be copied.
17953 By default value 0xf, all planes will be processed.
17956 For example if first stream pixel's component value is less then threshold value
17957 of pixel component from 2nd threshold stream, third stream value will picked,
17958 otherwise fourth stream pixel component value will be picked.
17960 Using color source filter one can perform various types of thresholding:
17962 @subsection Examples
17966 Binary threshold, using gray color as threshold:
17968 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
17972 Inverted binary threshold, using gray color as threshold:
17974 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
17978 Truncate binary threshold, using gray color as threshold:
17980 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
17984 Threshold to zero, using gray color as threshold:
17986 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
17990 Inverted threshold to zero, using gray color as threshold:
17992 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
17997 Select the most representative frame in a given sequence of consecutive frames.
17999 The filter accepts the following options:
18003 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18004 will pick one of them, and then handle the next batch of @var{n} frames until
18005 the end. Default is @code{100}.
18008 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18009 value will result in a higher memory usage, so a high value is not recommended.
18011 @subsection Examples
18015 Extract one picture each 50 frames:
18021 Complete example of a thumbnail creation with @command{ffmpeg}:
18023 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18029 Tile several successive frames together.
18031 The filter accepts the following options:
18036 Set the grid size (i.e. the number of lines and columns). For the syntax of
18037 this option, check the
18038 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18041 Set the maximum number of frames to render in the given area. It must be less
18042 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18043 the area will be used.
18046 Set the outer border margin in pixels.
18049 Set the inner border thickness (i.e. the number of pixels between frames). For
18050 more advanced padding options (such as having different values for the edges),
18051 refer to the pad video filter.
18054 Specify the color of the unused area. For the syntax of this option, check the
18055 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18056 The default value of @var{color} is "black".
18059 Set the number of frames to overlap when tiling several successive frames together.
18060 The value must be between @code{0} and @var{nb_frames - 1}.
18063 Set the number of frames to initially be empty before displaying first output frame.
18064 This controls how soon will one get first output frame.
18065 The value must be between @code{0} and @var{nb_frames - 1}.
18068 @subsection Examples
18072 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18074 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18076 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18077 duplicating each output frame to accommodate the originally detected frame
18081 Display @code{5} pictures in an area of @code{3x2} frames,
18082 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18083 mixed flat and named options:
18085 tile=3x2:nb_frames=5:padding=7:margin=2
18089 @section tinterlace
18091 Perform various types of temporal field interlacing.
18093 Frames are counted starting from 1, so the first input frame is
18096 The filter accepts the following options:
18101 Specify the mode of the interlacing. This option can also be specified
18102 as a value alone. See below for a list of values for this option.
18104 Available values are:
18108 Move odd frames into the upper field, even into the lower field,
18109 generating a double height frame at half frame rate.
18113 Frame 1 Frame 2 Frame 3 Frame 4
18115 11111 22222 33333 44444
18116 11111 22222 33333 44444
18117 11111 22222 33333 44444
18118 11111 22222 33333 44444
18132 Only output odd frames, even frames are dropped, generating a frame with
18133 unchanged height at half frame rate.
18138 Frame 1 Frame 2 Frame 3 Frame 4
18140 11111 22222 33333 44444
18141 11111 22222 33333 44444
18142 11111 22222 33333 44444
18143 11111 22222 33333 44444
18153 Only output even frames, odd frames are dropped, generating a frame with
18154 unchanged height at half frame rate.
18159 Frame 1 Frame 2 Frame 3 Frame 4
18161 11111 22222 33333 44444
18162 11111 22222 33333 44444
18163 11111 22222 33333 44444
18164 11111 22222 33333 44444
18174 Expand each frame to full height, but pad alternate lines with black,
18175 generating a frame with double height at the same input frame rate.
18180 Frame 1 Frame 2 Frame 3 Frame 4
18182 11111 22222 33333 44444
18183 11111 22222 33333 44444
18184 11111 22222 33333 44444
18185 11111 22222 33333 44444
18188 11111 ..... 33333 .....
18189 ..... 22222 ..... 44444
18190 11111 ..... 33333 .....
18191 ..... 22222 ..... 44444
18192 11111 ..... 33333 .....
18193 ..... 22222 ..... 44444
18194 11111 ..... 33333 .....
18195 ..... 22222 ..... 44444
18199 @item interleave_top, 4
18200 Interleave the upper field from odd frames with the lower field from
18201 even frames, generating a frame with 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 @item interleave_bottom, 5
18222 Interleave the lower field from odd frames with the upper field from
18223 even frames, generating a frame with unchanged height at half frame rate.
18228 Frame 1 Frame 2 Frame 3 Frame 4
18230 11111 22222<- 33333 44444<-
18231 11111<- 22222 33333<- 44444
18232 11111 22222<- 33333 44444<-
18233 11111<- 22222 33333<- 44444
18243 @item interlacex2, 6
18244 Double frame rate with unchanged height. Frames are inserted each
18245 containing the second temporal field from the previous input frame and
18246 the first temporal field from the next input frame. This mode relies on
18247 the top_field_first flag. Useful for interlaced video displays with no
18248 field synchronisation.
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
18261 11111 22222 22222 33333 33333 44444 44444
18262 11111 11111 22222 22222 33333 33333 44444
18263 11111 22222 22222 33333 33333 44444 44444
18264 11111 11111 22222 22222 33333 33333 44444
18269 Move odd frames into the upper field, even into the lower field,
18270 generating a double height frame at same 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
18283 11111 33333 33333 55555
18284 22222 22222 44444 44444
18285 11111 33333 33333 55555
18286 22222 22222 44444 44444
18287 11111 33333 33333 55555
18288 22222 22222 44444 44444
18289 11111 33333 33333 55555
18290 22222 22222 44444 44444
18295 Numeric values are deprecated but are accepted for backward
18296 compatibility reasons.
18298 Default mode is @code{merge}.
18301 Specify flags influencing the filter process.
18303 Available value for @var{flags} is:
18306 @item low_pass_filter, vlpf
18307 Enable linear vertical low-pass filtering in the filter.
18308 Vertical low-pass filtering is required when creating an interlaced
18309 destination from a progressive source which contains high-frequency
18310 vertical detail. Filtering will reduce interlace 'twitter' and Moire
18313 @item complex_filter, cvlpf
18314 Enable complex vertical low-pass filtering.
18315 This will slightly less reduce interlace 'twitter' and Moire
18316 patterning but better retain detail and subjective sharpness impression.
18319 Bypass already interlaced frames, only adjust the frame rate.
18322 Vertical low-pass filtering and bypassing already interlaced frames can only be
18323 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
18329 Mix successive video frames.
18331 A description of the accepted options follows.
18335 The number of successive frames to mix. If unspecified, it defaults to 3.
18338 Specify weight of each input video frame.
18339 Each weight is separated by space. If number of weights is smaller than
18340 number of @var{frames} last specified weight will be used for all remaining
18344 Specify scale, if it is set it will be multiplied with sum
18345 of each weight multiplied with pixel values to give final destination
18346 pixel value. By default @var{scale} is auto scaled to sum of weights.
18349 @subsection Examples
18353 Average 7 successive frames:
18355 tmix=frames=7:weights="1 1 1 1 1 1 1"
18359 Apply simple temporal convolution:
18361 tmix=frames=3:weights="-1 3 -1"
18365 Similar as above but only showing temporal differences:
18367 tmix=frames=3:weights="-1 2 -1":scale=1
18373 Tone map colors from different dynamic ranges.
18375 This filter expects data in single precision floating point, as it needs to
18376 operate on (and can output) out-of-range values. Another filter, such as
18377 @ref{zscale}, is needed to convert the resulting frame to a usable format.
18379 The tonemapping algorithms implemented only work on linear light, so input
18380 data should be linearized beforehand (and possibly correctly tagged).
18383 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
18386 @subsection Options
18387 The filter accepts the following options.
18391 Set the tone map algorithm to use.
18393 Possible values are:
18396 Do not apply any tone map, only desaturate overbright pixels.
18399 Hard-clip any out-of-range values. Use it for perfect color accuracy for
18400 in-range values, while distorting out-of-range values.
18403 Stretch the entire reference gamut to a linear multiple of the display.
18406 Fit a logarithmic transfer between the tone curves.
18409 Preserve overall image brightness with a simple curve, using nonlinear
18410 contrast, which results in flattening details and degrading color accuracy.
18413 Preserve both dark and bright details better than @var{reinhard}, at the cost
18414 of slightly darkening everything. Use it when detail preservation is more
18415 important than color and brightness accuracy.
18418 Smoothly map out-of-range values, while retaining contrast and colors for
18419 in-range material as much as possible. Use it when color accuracy is more
18420 important than detail preservation.
18426 Tune the tone mapping algorithm.
18428 This affects the following algorithms:
18434 Specifies the scale factor to use while stretching.
18438 Specifies the exponent of the function.
18442 Specify an extra linear coefficient to multiply into the signal before clipping.
18446 Specify the local contrast coefficient at the display peak.
18447 Default to 0.5, which means that in-gamut values will be about half as bright
18454 Specify the transition point from linear to mobius transform. Every value
18455 below this point is guaranteed to be mapped 1:1. The higher the value, the
18456 more accurate the result will be, at the cost of losing bright details.
18457 Default to 0.3, which due to the steep initial slope still preserves in-range
18458 colors fairly accurately.
18462 Apply desaturation for highlights that exceed this level of brightness. The
18463 higher the parameter, the more color information will be preserved. This
18464 setting helps prevent unnaturally blown-out colors for super-highlights, by
18465 (smoothly) turning into white instead. This makes images feel more natural,
18466 at the cost of reducing information about out-of-range colors.
18468 The default of 2.0 is somewhat conservative and will mostly just apply to
18469 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
18471 This option works only if the input frame has a supported color tag.
18474 Override signal/nominal/reference peak with this value. Useful when the
18475 embedded peak information in display metadata is not reliable or when tone
18476 mapping from a lower range to a higher range.
18481 Temporarily pad video frames.
18483 The filter accepts the following options:
18487 Specify number of delay frames before input video stream.
18490 Specify number of padding frames after input video stream.
18491 Set to -1 to pad indefinitely.
18494 Set kind of frames added to beginning of stream.
18495 Can be either @var{add} or @var{clone}.
18496 With @var{add} frames of solid-color are added.
18497 With @var{clone} frames are clones of first frame.
18500 Set kind of frames added to end of stream.
18501 Can be either @var{add} or @var{clone}.
18502 With @var{add} frames of solid-color are added.
18503 With @var{clone} frames are clones of last frame.
18505 @item start_duration, stop_duration
18506 Specify the duration of the start/stop delay. See
18507 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18508 for the accepted syntax.
18509 These options override @var{start} and @var{stop}.
18512 Specify the color of the padded area. For the syntax of this option,
18513 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
18514 manual,ffmpeg-utils}.
18516 The default value of @var{color} is "black".
18522 Transpose rows with columns in the input video and optionally flip it.
18524 It accepts the following parameters:
18529 Specify the transposition direction.
18531 Can assume the following values:
18533 @item 0, 4, cclock_flip
18534 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
18542 Rotate by 90 degrees clockwise, that is:
18550 Rotate by 90 degrees counterclockwise, that is:
18557 @item 3, 7, clock_flip
18558 Rotate by 90 degrees clockwise and vertically flip, that is:
18566 For values between 4-7, the transposition is only done if the input
18567 video geometry is portrait and not landscape. These values are
18568 deprecated, the @code{passthrough} option should be used instead.
18570 Numerical values are deprecated, and should be dropped in favor of
18571 symbolic constants.
18574 Do not apply the transposition if the input geometry matches the one
18575 specified by the specified value. It accepts the following values:
18578 Always apply transposition.
18580 Preserve portrait geometry (when @var{height} >= @var{width}).
18582 Preserve landscape geometry (when @var{width} >= @var{height}).
18585 Default value is @code{none}.
18588 For example to rotate by 90 degrees clockwise and preserve portrait
18591 transpose=dir=1:passthrough=portrait
18594 The command above can also be specified as:
18596 transpose=1:portrait
18599 @section transpose_npp
18601 Transpose rows with columns in the input video and optionally flip it.
18602 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
18604 It accepts the following parameters:
18609 Specify the transposition direction.
18611 Can assume the following values:
18614 Rotate by 90 degrees counterclockwise and vertically flip. (default)
18617 Rotate by 90 degrees clockwise.
18620 Rotate by 90 degrees counterclockwise.
18623 Rotate by 90 degrees clockwise and vertically flip.
18627 Do not apply the transposition if the input geometry matches the one
18628 specified by the specified value. It accepts the following values:
18631 Always apply transposition. (default)
18633 Preserve portrait geometry (when @var{height} >= @var{width}).
18635 Preserve landscape geometry (when @var{width} >= @var{height}).
18641 Trim the input so that the output contains one continuous subpart of the input.
18643 It accepts the following parameters:
18646 Specify the time of the start of the kept section, i.e. the frame with the
18647 timestamp @var{start} will be the first frame in the output.
18650 Specify the time of the first frame that will be dropped, i.e. the frame
18651 immediately preceding the one with the timestamp @var{end} will be the last
18652 frame in the output.
18655 This is the same as @var{start}, except this option sets the start timestamp
18656 in timebase units instead of seconds.
18659 This is the same as @var{end}, except this option sets the end timestamp
18660 in timebase units instead of seconds.
18663 The maximum duration of the output in seconds.
18666 The number of the first frame that should be passed to the output.
18669 The number of the first frame that should be dropped.
18672 @option{start}, @option{end}, and @option{duration} are expressed as time
18673 duration specifications; see
18674 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
18675 for the accepted syntax.
18677 Note that the first two sets of the start/end options and the @option{duration}
18678 option look at the frame timestamp, while the _frame variants simply count the
18679 frames that pass through the filter. Also note that this filter does not modify
18680 the timestamps. If you wish for the output timestamps to start at zero, insert a
18681 setpts filter after the trim filter.
18683 If multiple start or end options are set, this filter tries to be greedy and
18684 keep all the frames that match at least one of the specified constraints. To keep
18685 only the part that matches all the constraints at once, chain multiple trim
18688 The defaults are such that all the input is kept. So it is possible to set e.g.
18689 just the end values to keep everything before the specified time.
18694 Drop everything except the second minute of input:
18696 ffmpeg -i INPUT -vf trim=60:120
18700 Keep only the first second:
18702 ffmpeg -i INPUT -vf trim=duration=1
18707 @section unpremultiply
18708 Apply alpha unpremultiply effect to input video stream using first plane
18709 of second stream as alpha.
18711 Both streams must have same dimensions and same pixel format.
18713 The filter accepts the following option:
18717 Set which planes will be processed, unprocessed planes will be copied.
18718 By default value 0xf, all planes will be processed.
18720 If the format has 1 or 2 components, then luma is bit 0.
18721 If the format has 3 or 4 components:
18722 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
18723 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
18724 If present, the alpha channel is always the last bit.
18727 Do not require 2nd input for processing, instead use alpha plane from input stream.
18733 Sharpen or blur the input video.
18735 It accepts the following parameters:
18738 @item luma_msize_x, lx
18739 Set the luma matrix horizontal size. It must be an odd integer between
18740 3 and 23. The default value is 5.
18742 @item luma_msize_y, ly
18743 Set the luma matrix vertical size. It must be an odd integer between 3
18744 and 23. The default value is 5.
18746 @item luma_amount, la
18747 Set the luma effect strength. It must be a floating point number, reasonable
18748 values lay between -1.5 and 1.5.
18750 Negative values will blur the input video, while positive values will
18751 sharpen it, a value of zero will disable the effect.
18753 Default value is 1.0.
18755 @item chroma_msize_x, cx
18756 Set the chroma matrix horizontal size. It must be an odd integer
18757 between 3 and 23. The default value is 5.
18759 @item chroma_msize_y, cy
18760 Set the chroma matrix vertical size. It must be an odd integer
18761 between 3 and 23. The default value is 5.
18763 @item chroma_amount, ca
18764 Set the chroma effect strength. It must be a floating point number, reasonable
18765 values lay between -1.5 and 1.5.
18767 Negative values will blur the input video, while positive values will
18768 sharpen it, a value of zero will disable the effect.
18770 Default value is 0.0.
18774 All parameters are optional and default to the equivalent of the
18775 string '5:5:1.0:5:5:0.0'.
18777 @subsection Examples
18781 Apply strong luma sharpen effect:
18783 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
18787 Apply a strong blur of both luma and chroma parameters:
18789 unsharp=7:7:-2:7:7:-2
18795 Apply ultra slow/simple postprocessing filter that compresses and decompresses
18796 the image at several (or - in the case of @option{quality} level @code{8} - all)
18797 shifts and average the results.
18799 The way this differs from the behavior of spp is that uspp actually encodes &
18800 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
18801 DCT similar to MJPEG.
18803 The filter accepts the following options:
18807 Set quality. This option defines the number of levels for averaging. It accepts
18808 an integer in the range 0-8. If set to @code{0}, the filter will have no
18809 effect. A value of @code{8} means the higher quality. For each increment of
18810 that value the speed drops by a factor of approximately 2. Default value is
18814 Force a constant quantization parameter. If not set, the filter will use the QP
18815 from the video stream (if available).
18820 Convert 360 videos between various formats.
18822 The filter accepts the following options:
18828 Set format of the input/output video.
18836 Equirectangular projection.
18841 Cubemap with 3x2/6x1/1x6 layout.
18843 Format specific options:
18848 Set padding proportion for the input/output cubemap. Values in decimals.
18855 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)
18858 Default value is @b{@samp{0}}.
18862 Set fixed padding for the input/output cubemap. Values in pixels.
18864 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
18868 Set order of faces for the input/output cubemap. Choose one direction for each position.
18870 Designation of directions:
18886 Default value is @b{@samp{rludfb}}.
18890 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
18892 Designation of angles:
18895 0 degrees clockwise
18897 90 degrees clockwise
18899 180 degrees clockwise
18901 270 degrees clockwise
18904 Default value is @b{@samp{000000}}.
18908 Equi-Angular Cubemap.
18915 Format specific options:
18920 Set output horizontal/vertical/diagonal field of view. Values in degrees.
18922 If diagonal field of view is set it overrides horizontal and vertical field of view.
18927 Set input horizontal/vertical/diagonal field of view. Values in degrees.
18929 If diagonal field of view is set it overrides horizontal and vertical field of view.
18935 Format specific options:
18939 Set padding proportion. Values in decimals.
18949 Default value is @b{@samp{0}}.
18954 Facebook's 360 format.
18957 Stereographic format.
18959 Format specific options:
18964 Set output horizontal/vertical/diagonal field of view. Values in degrees.
18966 If diagonal field of view is set it overrides horizontal and vertical field of view.
18971 Set input horizontal/vertical/diagonal field of view. Values in degrees.
18973 If diagonal field of view is set it overrides horizontal and vertical field of view.
18980 Ball format, gives significant distortion toward the back.
18983 Hammer-Aitoff map projection format.
18986 Sinusoidal map projection format.
18989 Fisheye projection.
18991 Format specific options:
18996 Set output horizontal/vertical/diagonal field of view. Values in degrees.
18998 If diagonal field of view is set it overrides horizontal and vertical field of view.
19003 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19005 If diagonal field of view is set it overrides horizontal and vertical field of view.
19009 Pannini projection. @i{(output only)}
19011 Format specific options:
19014 Set pannini parameter.
19018 Cylindrical projection.
19020 Format specific options:
19025 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19027 If diagonal field of view is set it overrides horizontal and vertical field of view.
19032 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19034 If diagonal field of view is set it overrides horizontal and vertical field of view.
19038 Perspective projection. @i{(output only)}
19040 Format specific options:
19043 Set perspective parameter.
19047 Tetrahedron projection.
19051 Set interpolation method.@*
19052 @i{Note: more complex interpolation methods require much more memory to run.}
19062 Bilinear interpolation.
19065 Bicubic interpolation.
19068 Lanczos interpolation.
19071 Spline16 interpolation.
19074 Gaussian interpolation.
19077 Default value is @b{@samp{line}}.
19081 Set the output video resolution.
19083 Default resolution depends on formats.
19087 Set the input/output stereo format.
19098 Default value is @b{@samp{2d}} for input and output format.
19103 Set rotation for the output video. Values in degrees.
19106 Set rotation order for the output video. Choose one item for each position.
19117 Default value is @b{@samp{ypr}}.
19122 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
19126 Set if input video is flipped horizontally/vertically. Boolean values.
19129 Set if input video is transposed. Boolean value, by default disabled.
19132 Set if output video needs to be transposed. Boolean value, by default disabled.
19135 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
19138 @subsection Examples
19142 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
19144 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
19147 Extract back view of Equi-Angular Cubemap:
19149 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
19152 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
19154 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
19158 @section vaguedenoiser
19160 Apply a wavelet based denoiser.
19162 It transforms each frame from the video input into the wavelet domain,
19163 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
19164 the obtained coefficients. It does an inverse wavelet transform after.
19165 Due to wavelet properties, it should give a nice smoothed result, and
19166 reduced noise, without blurring picture features.
19168 This filter accepts the following options:
19172 The filtering strength. The higher, the more filtered the video will be.
19173 Hard thresholding can use a higher threshold than soft thresholding
19174 before the video looks overfiltered. Default value is 2.
19177 The filtering method the filter will use.
19179 It accepts the following values:
19182 All values under the threshold will be zeroed.
19185 All values under the threshold will be zeroed. All values above will be
19186 reduced by the threshold.
19189 Scales or nullifies coefficients - intermediary between (more) soft and
19190 (less) hard thresholding.
19193 Default is garrote.
19196 Number of times, the wavelet will decompose the picture. Picture can't
19197 be decomposed beyond a particular point (typically, 8 for a 640x480
19198 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
19201 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
19204 A list of the planes to process. By default all planes are processed.
19207 @section vectorscope
19209 Display 2 color component values in the two dimensional graph (which is called
19212 This filter accepts the following options:
19216 Set vectorscope mode.
19218 It accepts the following values:
19222 Gray values are displayed on graph, higher brightness means more pixels have
19223 same component color value on location in graph. This is the default mode.
19226 Gray values are displayed on graph. Surrounding pixels values which are not
19227 present in video frame are drawn in gradient of 2 color components which are
19228 set by option @code{x} and @code{y}. The 3rd color component is static.
19231 Actual color components values present in video frame are displayed on graph.
19234 Similar as color2 but higher frequency of same values @code{x} and @code{y}
19235 on graph increases value of another color component, which is luminance by
19236 default values of @code{x} and @code{y}.
19239 Actual colors present in video frame are displayed on graph. If two different
19240 colors map to same position on graph then color with higher value of component
19241 not present in graph is picked.
19244 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
19245 component picked from radial gradient.
19249 Set which color component will be represented on X-axis. Default is @code{1}.
19252 Set which color component will be represented on Y-axis. Default is @code{2}.
19255 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
19256 of color component which represents frequency of (X, Y) location in graph.
19261 No envelope, this is default.
19264 Instant envelope, even darkest single pixel will be clearly highlighted.
19267 Hold maximum and minimum values presented in graph over time. This way you
19268 can still spot out of range values without constantly looking at vectorscope.
19271 Peak and instant envelope combined together.
19275 Set what kind of graticule to draw.
19284 Set graticule opacity.
19287 Set graticule flags.
19291 Draw graticule for white point.
19294 Draw graticule for black point.
19297 Draw color points short names.
19301 Set background opacity.
19303 @item lthreshold, l
19304 Set low threshold for color component not represented on X or Y axis.
19305 Values lower than this value will be ignored. Default is 0.
19306 Note this value is multiplied with actual max possible value one pixel component
19307 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
19310 @item hthreshold, h
19311 Set high threshold for color component not represented on X or Y axis.
19312 Values higher than this value will be ignored. Default is 1.
19313 Note this value is multiplied with actual max possible value one pixel component
19314 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
19315 is 0.9 * 255 = 230.
19317 @item colorspace, c
19318 Set what kind of colorspace to use when drawing graticule.
19328 Set color tint for gray/tint vectorscope mode. By default both options are zero.
19329 This means no tint, and output will remain gray.
19332 @anchor{vidstabdetect}
19333 @section vidstabdetect
19335 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
19336 @ref{vidstabtransform} for pass 2.
19338 This filter generates a file with relative translation and rotation
19339 transform information about subsequent frames, which is then used by
19340 the @ref{vidstabtransform} filter.
19342 To enable compilation of this filter you need to configure FFmpeg with
19343 @code{--enable-libvidstab}.
19345 This filter accepts the following options:
19349 Set the path to the file used to write the transforms information.
19350 Default value is @file{transforms.trf}.
19353 Set how shaky the video is and how quick the camera is. It accepts an
19354 integer in the range 1-10, a value of 1 means little shakiness, a
19355 value of 10 means strong shakiness. Default value is 5.
19358 Set the accuracy of the detection process. It must be a value in the
19359 range 1-15. A value of 1 means low accuracy, a value of 15 means high
19360 accuracy. Default value is 15.
19363 Set stepsize of the search process. The region around minimum is
19364 scanned with 1 pixel resolution. Default value is 6.
19367 Set minimum contrast. Below this value a local measurement field is
19368 discarded. Must be a floating point value in the range 0-1. Default
19372 Set reference frame number for tripod mode.
19374 If enabled, the motion of the frames is compared to a reference frame
19375 in the filtered stream, identified by the specified number. The idea
19376 is to compensate all movements in a more-or-less static scene and keep
19377 the camera view absolutely still.
19379 If set to 0, it is disabled. The frames are counted starting from 1.
19382 Show fields and transforms in the resulting frames. It accepts an
19383 integer in the range 0-2. Default value is 0, which disables any
19387 @subsection Examples
19391 Use default values:
19397 Analyze strongly shaky movie and put the results in file
19398 @file{mytransforms.trf}:
19400 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
19404 Visualize the result of internal transformations in the resulting
19407 vidstabdetect=show=1
19411 Analyze a video with medium shakiness using @command{ffmpeg}:
19413 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
19417 @anchor{vidstabtransform}
19418 @section vidstabtransform
19420 Video stabilization/deshaking: pass 2 of 2,
19421 see @ref{vidstabdetect} for pass 1.
19423 Read a file with transform information for each frame and
19424 apply/compensate them. Together with the @ref{vidstabdetect}
19425 filter this can be used to deshake videos. See also
19426 @url{http://public.hronopik.de/vid.stab}. It is important to also use
19427 the @ref{unsharp} filter, see below.
19429 To enable compilation of this filter you need to configure FFmpeg with
19430 @code{--enable-libvidstab}.
19432 @subsection Options
19436 Set path to the file used to read the transforms. Default value is
19437 @file{transforms.trf}.
19440 Set the number of frames (value*2 + 1) used for lowpass filtering the
19441 camera movements. Default value is 10.
19443 For example a number of 10 means that 21 frames are used (10 in the
19444 past and 10 in the future) to smoothen the motion in the video. A
19445 larger value leads to a smoother video, but limits the acceleration of
19446 the camera (pan/tilt movements). 0 is a special case where a static
19447 camera is simulated.
19450 Set the camera path optimization algorithm.
19452 Accepted values are:
19455 gaussian kernel low-pass filter on camera motion (default)
19457 averaging on transformations
19461 Set maximal number of pixels to translate frames. Default value is -1,
19465 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
19466 value is -1, meaning no limit.
19469 Specify how to deal with borders that may be visible due to movement
19472 Available values are:
19475 keep image information from previous frame (default)
19477 fill the border black
19481 Invert transforms if set to 1. Default value is 0.
19484 Consider transforms as relative to previous frame if set to 1,
19485 absolute if set to 0. Default value is 0.
19488 Set percentage to zoom. A positive value will result in a zoom-in
19489 effect, a negative value in a zoom-out effect. Default value is 0 (no
19493 Set optimal zooming to avoid borders.
19495 Accepted values are:
19500 optimal static zoom value is determined (only very strong movements
19501 will lead to visible borders) (default)
19503 optimal adaptive zoom value is determined (no borders will be
19504 visible), see @option{zoomspeed}
19507 Note that the value given at zoom is added to the one calculated here.
19510 Set percent to zoom maximally each frame (enabled when
19511 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
19515 Specify type of interpolation.
19517 Available values are:
19522 linear only horizontal
19524 linear in both directions (default)
19526 cubic in both directions (slow)
19530 Enable virtual tripod mode if set to 1, which is equivalent to
19531 @code{relative=0:smoothing=0}. Default value is 0.
19533 Use also @code{tripod} option of @ref{vidstabdetect}.
19536 Increase log verbosity if set to 1. Also the detected global motions
19537 are written to the temporary file @file{global_motions.trf}. Default
19541 @subsection Examples
19545 Use @command{ffmpeg} for a typical stabilization with default values:
19547 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
19550 Note the use of the @ref{unsharp} filter which is always recommended.
19553 Zoom in a bit more and load transform data from a given file:
19555 vidstabtransform=zoom=5:input="mytransforms.trf"
19559 Smoothen the video even more:
19561 vidstabtransform=smoothing=30
19567 Flip the input video vertically.
19569 For example, to vertically flip a video with @command{ffmpeg}:
19571 ffmpeg -i in.avi -vf "vflip" out.avi
19576 Detect variable frame rate video.
19578 This filter tries to detect if the input is variable or constant frame rate.
19580 At end it will output number of frames detected as having variable delta pts,
19581 and ones with constant delta pts.
19582 If there was frames with variable delta, than it will also show min, max and
19583 average delta encountered.
19587 Boost or alter saturation.
19589 The filter accepts the following options:
19592 Set strength of boost if positive value or strength of alter if negative value.
19593 Default is 0. Allowed range is from -2 to 2.
19596 Set the red balance. Default is 1. Allowed range is from -10 to 10.
19599 Set the green balance. Default is 1. Allowed range is from -10 to 10.
19602 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
19605 Set the red luma coefficient.
19608 Set the green luma coefficient.
19611 Set the blue luma coefficient.
19614 If @code{intensity} is negative and this is set to 1, colors will change,
19615 otherwise colors will be less saturated, more towards gray.
19618 @subsection Commands
19620 This filter supports the all above options as @ref{commands}.
19625 Make or reverse a natural vignetting effect.
19627 The filter accepts the following options:
19631 Set lens angle expression as a number of radians.
19633 The value is clipped in the @code{[0,PI/2]} range.
19635 Default value: @code{"PI/5"}
19639 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
19643 Set forward/backward mode.
19645 Available modes are:
19648 The larger the distance from the central point, the darker the image becomes.
19651 The larger the distance from the central point, the brighter the image becomes.
19652 This can be used to reverse a vignette effect, though there is no automatic
19653 detection to extract the lens @option{angle} and other settings (yet). It can
19654 also be used to create a burning effect.
19657 Default value is @samp{forward}.
19660 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
19662 It accepts the following values:
19665 Evaluate expressions only once during the filter initialization.
19668 Evaluate expressions for each incoming frame. This is way slower than the
19669 @samp{init} mode since it requires all the scalers to be re-computed, but it
19670 allows advanced dynamic expressions.
19673 Default value is @samp{init}.
19676 Set dithering to reduce the circular banding effects. Default is @code{1}
19680 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
19681 Setting this value to the SAR of the input will make a rectangular vignetting
19682 following the dimensions of the video.
19684 Default is @code{1/1}.
19687 @subsection Expressions
19689 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
19690 following parameters.
19695 input width and height
19698 the number of input frame, starting from 0
19701 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
19702 @var{TB} units, NAN if undefined
19705 frame rate of the input video, NAN if the input frame rate is unknown
19708 the PTS (Presentation TimeStamp) of the filtered video frame,
19709 expressed in seconds, NAN if undefined
19712 time base of the input video
19716 @subsection Examples
19720 Apply simple strong vignetting effect:
19726 Make a flickering vignetting:
19728 vignette='PI/4+random(1)*PI/50':eval=frame
19733 @section vmafmotion
19735 Obtain the average VMAF motion score of a video.
19736 It is one of the component metrics of VMAF.
19738 The obtained average motion score is printed through the logging system.
19740 The filter accepts the following options:
19744 If specified, the filter will use the named file to save the motion score of
19745 each frame with respect to the previous frame.
19746 When filename equals "-" the data is sent to standard output.
19751 ffmpeg -i ref.mpg -vf vmafmotion -f null -
19755 Stack input videos vertically.
19757 All streams must be of same pixel format and of same width.
19759 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
19760 to create same output.
19762 The filter accepts the following options:
19766 Set number of input streams. Default is 2.
19769 If set to 1, force the output to terminate when the shortest input
19770 terminates. Default value is 0.
19775 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
19776 Deinterlacing Filter").
19778 Based on the process described by Martin Weston for BBC R&D, and
19779 implemented based on the de-interlace algorithm written by Jim
19780 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
19781 uses filter coefficients calculated by BBC R&D.
19783 This filter uses field-dominance information in frame to decide which
19784 of each pair of fields to place first in the output.
19785 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
19787 There are two sets of filter coefficients, so called "simple"
19788 and "complex". Which set of filter coefficients is used can
19789 be set by passing an optional parameter:
19793 Set the interlacing filter coefficients. Accepts one of the following values:
19797 Simple filter coefficient set.
19799 More-complex filter coefficient set.
19801 Default value is @samp{complex}.
19804 Specify which frames to deinterlace. Accepts one of the following values:
19808 Deinterlace all frames,
19810 Only deinterlace frames marked as interlaced.
19813 Default value is @samp{all}.
19817 Video waveform monitor.
19819 The waveform monitor plots color component intensity. By default luminance
19820 only. Each column of the waveform corresponds to a column of pixels in the
19823 It accepts the following options:
19827 Can be either @code{row}, or @code{column}. Default is @code{column}.
19828 In row mode, the graph on the left side represents color component value 0 and
19829 the right side represents value = 255. In column mode, the top side represents
19830 color component value = 0 and bottom side represents value = 255.
19833 Set intensity. Smaller values are useful to find out how many values of the same
19834 luminance are distributed across input rows/columns.
19835 Default value is @code{0.04}. Allowed range is [0, 1].
19838 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
19839 In mirrored mode, higher values will be represented on the left
19840 side for @code{row} mode and at the top for @code{column} mode. Default is
19841 @code{1} (mirrored).
19845 It accepts the following values:
19848 Presents information identical to that in the @code{parade}, except
19849 that the graphs representing color components are superimposed directly
19852 This display mode makes it easier to spot relative differences or similarities
19853 in overlapping areas of the color components that are supposed to be identical,
19854 such as neutral whites, grays, or blacks.
19857 Display separate graph for the color components side by side in
19858 @code{row} mode or one below the other in @code{column} mode.
19861 Display separate graph for the color components side by side in
19862 @code{column} mode or one below the other in @code{row} mode.
19864 Using this display mode makes it easy to spot color casts in the highlights
19865 and shadows of an image, by comparing the contours of the top and the bottom
19866 graphs of each waveform. Since whites, grays, and blacks are characterized
19867 by exactly equal amounts of red, green, and blue, neutral areas of the picture
19868 should display three waveforms of roughly equal width/height. If not, the
19869 correction is easy to perform by making level adjustments the three waveforms.
19871 Default is @code{stack}.
19873 @item components, c
19874 Set which color components to display. Default is 1, which means only luminance
19875 or red color component if input is in RGB colorspace. If is set for example to
19876 7 it will display all 3 (if) available color components.
19881 No envelope, this is default.
19884 Instant envelope, minimum and maximum values presented in graph will be easily
19885 visible even with small @code{step} value.
19888 Hold minimum and maximum values presented in graph across time. This way you
19889 can still spot out of range values without constantly looking at waveforms.
19892 Peak and instant envelope combined together.
19898 No filtering, this is default.
19901 Luma and chroma combined together.
19904 Similar as above, but shows difference between blue and red chroma.
19907 Similar as above, but use different colors.
19910 Similar as above, but again with different colors.
19913 Displays only chroma.
19916 Displays actual color value on waveform.
19919 Similar as above, but with luma showing frequency of chroma values.
19923 Set which graticule to display.
19927 Do not display graticule.
19930 Display green graticule showing legal broadcast ranges.
19933 Display orange graticule showing legal broadcast ranges.
19936 Display invert graticule showing legal broadcast ranges.
19940 Set graticule opacity.
19943 Set graticule flags.
19947 Draw numbers above lines. By default enabled.
19950 Draw dots instead of lines.
19954 Set scale used for displaying graticule.
19961 Default is digital.
19964 Set background opacity.
19968 Set tint for output.
19969 Only used with lowpass filter and when display is not overlay and input
19970 pixel formats are not RGB.
19973 @section weave, doubleweave
19975 The @code{weave} takes a field-based video input and join
19976 each two sequential fields into single frame, producing a new double
19977 height clip with half the frame rate and half the frame count.
19979 The @code{doubleweave} works same as @code{weave} but without
19980 halving frame rate and frame count.
19982 It accepts the following option:
19986 Set first field. Available values are:
19990 Set the frame as top-field-first.
19993 Set the frame as bottom-field-first.
19997 @subsection Examples
20001 Interlace video using @ref{select} and @ref{separatefields} filter:
20003 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
20008 Apply the xBR high-quality magnification filter which is designed for pixel
20009 art. It follows a set of edge-detection rules, see
20010 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
20012 It accepts the following option:
20016 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
20017 @code{3xBR} and @code{4} for @code{4xBR}.
20018 Default is @code{3}.
20022 Pick median pixels from several input videos.
20024 The filter accepts the following options:
20028 Set number of inputs.
20029 Default is 3. Allowed range is from 3 to 255.
20030 If number of inputs is even number, than result will be mean value between two median values.
20033 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
20037 Stack video inputs into custom layout.
20039 All streams must be of same pixel format.
20041 The filter accepts the following options:
20045 Set number of input streams. Default is 2.
20048 Specify layout of inputs.
20049 This option requires the desired layout configuration to be explicitly set by the user.
20050 This sets position of each video input in output. Each input
20051 is separated by '|'.
20052 The first number represents the column, and the second number represents the row.
20053 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
20054 where X is video input from which to take width or height.
20055 Multiple values can be used when separated by '+'. In such
20056 case values are summed together.
20058 Note that if inputs are of different sizes gaps may appear, as not all of
20059 the output video frame will be filled. Similarly, videos can overlap each
20060 other if their position doesn't leave enough space for the full frame of
20063 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
20064 a layout must be set by the user.
20067 If set to 1, force the output to terminate when the shortest input
20068 terminates. Default value is 0.
20071 If set to valid color, all unused pixels will be filled with that color.
20072 By default fill is set to none, so it is disabled.
20075 @subsection Examples
20079 Display 4 inputs into 2x2 grid.
20083 input1(0, 0) | input3(w0, 0)
20084 input2(0, h0) | input4(w0, h0)
20088 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
20091 Note that if inputs are of different sizes, gaps or overlaps may occur.
20094 Display 4 inputs into 1x4 grid.
20101 input4(0, h0+h1+h2)
20105 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
20108 Note that if inputs are of different widths, unused space will appear.
20111 Display 9 inputs into 3x3 grid.
20115 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
20116 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
20117 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
20121 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
20124 Note that if inputs are of different sizes, gaps or overlaps may occur.
20127 Display 16 inputs into 4x4 grid.
20131 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
20132 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
20133 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
20134 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
20138 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|
20139 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
20142 Note that if inputs are of different sizes, gaps or overlaps may occur.
20149 Deinterlace the input video ("yadif" means "yet another deinterlacing
20152 It accepts the following parameters:
20158 The interlacing mode to adopt. It accepts one of the following values:
20161 @item 0, send_frame
20162 Output one frame for each frame.
20163 @item 1, send_field
20164 Output one frame for each field.
20165 @item 2, send_frame_nospatial
20166 Like @code{send_frame}, but it skips the spatial interlacing check.
20167 @item 3, send_field_nospatial
20168 Like @code{send_field}, but it skips the spatial interlacing check.
20171 The default value is @code{send_frame}.
20174 The picture field parity assumed for the input interlaced video. It accepts one
20175 of the following values:
20179 Assume the top field is first.
20181 Assume the bottom field is first.
20183 Enable automatic detection of field parity.
20186 The default value is @code{auto}.
20187 If the interlacing is unknown or the decoder does not export this information,
20188 top field first will be assumed.
20191 Specify which frames to deinterlace. Accepts one of the following
20196 Deinterlace all frames.
20197 @item 1, interlaced
20198 Only deinterlace frames marked as interlaced.
20201 The default value is @code{all}.
20204 @section yadif_cuda
20206 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
20207 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
20210 It accepts the following parameters:
20216 The interlacing mode to adopt. It accepts one of the following values:
20219 @item 0, send_frame
20220 Output one frame for each frame.
20221 @item 1, send_field
20222 Output one frame for each field.
20223 @item 2, send_frame_nospatial
20224 Like @code{send_frame}, but it skips the spatial interlacing check.
20225 @item 3, send_field_nospatial
20226 Like @code{send_field}, but it skips the spatial interlacing check.
20229 The default value is @code{send_frame}.
20232 The picture field parity assumed for the input interlaced video. It accepts one
20233 of the following values:
20237 Assume the top field is first.
20239 Assume the bottom field is first.
20241 Enable automatic detection of field parity.
20244 The default value is @code{auto}.
20245 If the interlacing is unknown or the decoder does not export this information,
20246 top field first will be assumed.
20249 Specify which frames to deinterlace. Accepts one of the following
20254 Deinterlace all frames.
20255 @item 1, interlaced
20256 Only deinterlace frames marked as interlaced.
20259 The default value is @code{all}.
20264 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
20265 The algorithm is described in
20266 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
20268 It accepts the following parameters:
20272 Set the window radius. Default value is 3.
20275 Set which planes to filter. Default is only the first plane.
20278 Set blur strength. Default value is 128.
20281 @subsection Commands
20282 This filter supports same @ref{commands} as options.
20286 Apply Zoom & Pan effect.
20288 This filter accepts the following options:
20292 Set the zoom expression. Range is 1-10. Default is 1.
20296 Set the x and y expression. Default is 0.
20299 Set the duration expression in number of frames.
20300 This sets for how many number of frames effect will last for
20301 single input image.
20304 Set the output image size, default is 'hd720'.
20307 Set the output frame rate, default is '25'.
20310 Each expression can contain the following constants:
20329 Output frame count.
20333 Last calculated 'x' and 'y' position from 'x' and 'y' expression
20334 for current input frame.
20338 'x' and 'y' of last output frame of previous input frame or 0 when there was
20339 not yet such frame (first input frame).
20342 Last calculated zoom from 'z' expression for current input frame.
20345 Last calculated zoom of last output frame of previous input frame.
20348 Number of output frames for current input frame. Calculated from 'd' expression
20349 for each input frame.
20352 number of output frames created for previous input frame
20355 Rational number: input width / input height
20358 sample aspect ratio
20361 display aspect ratio
20365 @subsection Examples
20369 Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
20371 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
20375 Zoom-in up to 1.5 and pan always at center of picture:
20377 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
20381 Same as above but without pausing:
20383 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
20389 Scale (resize) the input video, using the z.lib library:
20390 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
20391 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
20393 The zscale filter forces the output display aspect ratio to be the same
20394 as the input, by changing the output sample aspect ratio.
20396 If the input image format is different from the format requested by
20397 the next filter, the zscale filter will convert the input to the
20400 @subsection Options
20401 The filter accepts the following options.
20406 Set the output video dimension expression. Default value is the input
20409 If the @var{width} or @var{w} value is 0, the input width is used for
20410 the output. If the @var{height} or @var{h} value is 0, the input height
20411 is used for the output.
20413 If one and only one of the values is -n with n >= 1, the zscale filter
20414 will use a value that maintains the aspect ratio of the input image,
20415 calculated from the other specified dimension. After that it will,
20416 however, make sure that the calculated dimension is divisible by n and
20417 adjust the value if necessary.
20419 If both values are -n with n >= 1, the behavior will be identical to
20420 both values being set to 0 as previously detailed.
20422 See below for the list of accepted constants for use in the dimension
20426 Set the video size. For the syntax of this option, check the
20427 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
20430 Set the dither type.
20432 Possible values are:
20437 @item error_diffusion
20443 Set the resize filter type.
20445 Possible values are:
20455 Default is bilinear.
20458 Set the color range.
20460 Possible values are:
20467 Default is same as input.
20470 Set the color primaries.
20472 Possible values are:
20482 Default is same as input.
20485 Set the transfer characteristics.
20487 Possible values are:
20501 Default is same as input.
20504 Set the colorspace matrix.
20506 Possible value are:
20517 Default is same as input.
20520 Set the input color range.
20522 Possible values are:
20529 Default is same as input.
20531 @item primariesin, pin
20532 Set the input color primaries.
20534 Possible values are:
20544 Default is same as input.
20546 @item transferin, tin
20547 Set the input transfer characteristics.
20549 Possible values are:
20560 Default is same as input.
20562 @item matrixin, min
20563 Set the input colorspace matrix.
20565 Possible value are:
20577 Set the output chroma location.
20579 Possible values are:
20590 @item chromalin, cin
20591 Set the input chroma location.
20593 Possible values are:
20605 Set the nominal peak luminance.
20608 The values of the @option{w} and @option{h} options are expressions
20609 containing the following constants:
20614 The input width and height
20618 These are the same as @var{in_w} and @var{in_h}.
20622 The output (scaled) width and height
20626 These are the same as @var{out_w} and @var{out_h}
20629 The same as @var{iw} / @var{ih}
20632 input sample aspect ratio
20635 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
20639 horizontal and vertical input chroma subsample values. For example for the
20640 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
20644 horizontal and vertical output chroma subsample values. For example for the
20645 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
20648 @subsection Commands
20650 This filter supports the following commands:
20654 Set the output video dimension expression.
20655 The command accepts the same syntax of the corresponding option.
20657 If the specified expression is not valid, it is kept at its current
20661 @c man end VIDEO FILTERS
20663 @chapter OpenCL Video Filters
20664 @c man begin OPENCL VIDEO FILTERS
20666 Below is a description of the currently available OpenCL video filters.
20668 To enable compilation of these filters you need to configure FFmpeg with
20669 @code{--enable-opencl}.
20671 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
20674 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
20675 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
20676 given device parameters.
20678 @item -filter_hw_device @var{name}
20679 Pass the hardware device called @var{name} to all filters in any filter graph.
20683 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
20687 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
20689 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
20693 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.
20695 @section avgblur_opencl
20697 Apply average blur filter.
20699 The filter accepts the following options:
20703 Set horizontal radius size.
20704 Range is @code{[1, 1024]} and default value is @code{1}.
20707 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
20710 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
20713 @subsection Example
20717 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.
20719 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
20723 @section boxblur_opencl
20725 Apply a boxblur algorithm to the input video.
20727 It accepts the following parameters:
20731 @item luma_radius, lr
20732 @item luma_power, lp
20733 @item chroma_radius, cr
20734 @item chroma_power, cp
20735 @item alpha_radius, ar
20736 @item alpha_power, ap
20740 A description of the accepted options follows.
20743 @item luma_radius, lr
20744 @item chroma_radius, cr
20745 @item alpha_radius, ar
20746 Set an expression for the box radius in pixels used for blurring the
20747 corresponding input plane.
20749 The radius value must be a non-negative number, and must not be
20750 greater than the value of the expression @code{min(w,h)/2} for the
20751 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
20754 Default value for @option{luma_radius} is "2". If not specified,
20755 @option{chroma_radius} and @option{alpha_radius} default to the
20756 corresponding value set for @option{luma_radius}.
20758 The expressions can contain the following constants:
20762 The input width and height in pixels.
20766 The input chroma image width and height in pixels.
20770 The horizontal and vertical chroma subsample values. For example, for the
20771 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
20774 @item luma_power, lp
20775 @item chroma_power, cp
20776 @item alpha_power, ap
20777 Specify how many times the boxblur filter is applied to the
20778 corresponding plane.
20780 Default value for @option{luma_power} is 2. If not specified,
20781 @option{chroma_power} and @option{alpha_power} default to the
20782 corresponding value set for @option{luma_power}.
20784 A value of 0 will disable the effect.
20787 @subsection Examples
20789 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.
20793 Apply a boxblur filter with the luma, chroma, and alpha radius
20794 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.
20796 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
20797 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
20801 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.
20803 For the luma plane, a 2x2 box radius will be run once.
20805 For the chroma plane, a 4x4 box radius will be run 5 times.
20807 For the alpha plane, a 3x3 box radius will be run 7 times.
20809 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
20813 @section convolution_opencl
20815 Apply convolution of 3x3, 5x5, 7x7 matrix.
20817 The filter accepts the following options:
20824 Set matrix for each plane.
20825 Matrix is sequence of 9, 25 or 49 signed numbers.
20826 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
20832 Set multiplier for calculated value for each plane.
20833 If unset or 0, it will be sum of all matrix elements.
20834 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
20840 Set bias for each plane. This value is added to the result of the multiplication.
20841 Useful for making the overall image brighter or darker.
20842 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
20846 @subsection Examples
20852 -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
20858 -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
20862 Apply edge enhance:
20864 -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
20870 -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
20874 Apply laplacian edge detector which includes diagonals:
20876 -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
20882 -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
20886 @section dilation_opencl
20888 Apply dilation effect to the video.
20890 This filter replaces the pixel by the local(3x3) maximum.
20892 It accepts the following options:
20899 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
20900 If @code{0}, plane will remain unchanged.
20903 Flag which specifies the pixel to refer to.
20904 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
20906 Flags to local 3x3 coordinates region centered on @code{x}:
20915 @subsection Example
20919 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.
20921 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
20925 @section erosion_opencl
20927 Apply erosion effect to the video.
20929 This filter replaces the pixel by the local(3x3) minimum.
20931 It accepts the following options:
20938 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
20939 If @code{0}, plane will remain unchanged.
20942 Flag which specifies the pixel to refer to.
20943 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
20945 Flags to local 3x3 coordinates region centered on @code{x}:
20954 @subsection Example
20958 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.
20960 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
20964 @section colorkey_opencl
20965 RGB colorspace color keying.
20967 The filter accepts the following options:
20971 The color which will be replaced with transparency.
20974 Similarity percentage with the key color.
20976 0.01 matches only the exact key color, while 1.0 matches everything.
20981 0.0 makes pixels either fully transparent, or not transparent at all.
20983 Higher values result in semi-transparent pixels, with a higher transparency
20984 the more similar the pixels color is to the key color.
20987 @subsection Examples
20991 Make every semi-green pixel in the input transparent with some slight blending:
20993 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
20997 @section deshake_opencl
20998 Feature-point based video stabilization filter.
21000 The filter accepts the following options:
21004 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
21007 Whether or not additional debug info should be displayed, both in the processed output and in the console.
21009 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
21011 Viewing point matches in the output video is only supported for RGB input.
21013 Defaults to @code{0}.
21015 @item adaptive_crop
21016 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
21018 Defaults to @code{1}.
21020 @item refine_features
21021 Whether or not feature points should be refined at a sub-pixel level.
21023 This can be turned off for a slight performance gain at the cost of precision.
21025 Defaults to @code{1}.
21027 @item smooth_strength
21028 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
21030 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
21032 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
21034 Defaults to @code{0.0}.
21036 @item smooth_window_multiplier
21037 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
21039 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
21041 Acceptable values range from @code{0.1} to @code{10.0}.
21043 Larger values increase the amount of motion data available for determining how to smooth the camera path,
21044 potentially improving smoothness, but also increase latency and memory usage.
21046 Defaults to @code{2.0}.
21050 @subsection Examples
21054 Stabilize a video with a fixed, medium smoothing strength:
21056 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
21060 Stabilize a video with debugging (both in console and in rendered video):
21062 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
21066 @section nlmeans_opencl
21068 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
21070 @section overlay_opencl
21072 Overlay one video on top of another.
21074 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
21075 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
21077 The filter accepts the following options:
21082 Set the x coordinate of the overlaid video on the main video.
21083 Default value is @code{0}.
21086 Set the y coordinate of the overlaid video on the main video.
21087 Default value is @code{0}.
21091 @subsection Examples
21095 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
21097 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21100 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
21102 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21107 @section prewitt_opencl
21109 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
21111 The filter accepts the following option:
21115 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21118 Set value which will be multiplied with filtered result.
21119 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21122 Set value which will be added to filtered result.
21123 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21126 @subsection Example
21130 Apply the Prewitt operator with scale set to 2 and delta set to 10.
21132 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
21136 @section roberts_opencl
21137 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
21139 The filter accepts the following option:
21143 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21146 Set value which will be multiplied with filtered result.
21147 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21150 Set value which will be added to filtered result.
21151 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21154 @subsection Example
21158 Apply the Roberts cross operator with scale set to 2 and delta set to 10
21160 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
21164 @section sobel_opencl
21166 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
21168 The filter accepts the following option:
21172 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21175 Set value which will be multiplied with filtered result.
21176 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
21179 Set value which will be added to filtered result.
21180 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
21183 @subsection Example
21187 Apply sobel operator with scale set to 2 and delta set to 10
21189 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
21193 @section tonemap_opencl
21195 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
21197 It accepts the following parameters:
21201 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
21204 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
21207 Apply desaturation for highlights that exceed this level of brightness. The
21208 higher the parameter, the more color information will be preserved. This
21209 setting helps prevent unnaturally blown-out colors for super-highlights, by
21210 (smoothly) turning into white instead. This makes images feel more natural,
21211 at the cost of reducing information about out-of-range colors.
21213 The default value is 0.5, and the algorithm here is a little different from
21214 the cpu version tonemap currently. A setting of 0.0 disables this option.
21217 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
21218 is used to detect whether the scene has changed or not. If the distance between
21219 the current frame average brightness and the current running average exceeds
21220 a threshold value, we would re-calculate scene average and peak brightness.
21221 The default value is 0.2.
21224 Specify the output pixel format.
21226 Currently supported formats are:
21233 Set the output color range.
21235 Possible values are:
21241 Default is same as input.
21244 Set the output color primaries.
21246 Possible values are:
21252 Default is same as input.
21255 Set the output transfer characteristics.
21257 Possible values are:
21266 Set the output colorspace matrix.
21268 Possible value are:
21274 Default is same as input.
21278 @subsection Example
21282 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
21284 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
21288 @section unsharp_opencl
21290 Sharpen or blur the input video.
21292 It accepts the following parameters:
21295 @item luma_msize_x, lx
21296 Set the luma matrix horizontal size.
21297 Range is @code{[1, 23]} and default value is @code{5}.
21299 @item luma_msize_y, ly
21300 Set the luma matrix vertical size.
21301 Range is @code{[1, 23]} and default value is @code{5}.
21303 @item luma_amount, la
21304 Set the luma effect strength.
21305 Range is @code{[-10, 10]} and default value is @code{1.0}.
21307 Negative values will blur the input video, while positive values will
21308 sharpen it, a value of zero will disable the effect.
21310 @item chroma_msize_x, cx
21311 Set the chroma matrix horizontal size.
21312 Range is @code{[1, 23]} and default value is @code{5}.
21314 @item chroma_msize_y, cy
21315 Set the chroma matrix vertical size.
21316 Range is @code{[1, 23]} and default value is @code{5}.
21318 @item chroma_amount, ca
21319 Set the chroma effect strength.
21320 Range is @code{[-10, 10]} and default value is @code{0.0}.
21322 Negative values will blur the input video, while positive values will
21323 sharpen it, a value of zero will disable the effect.
21327 All parameters are optional and default to the equivalent of the
21328 string '5:5:1.0:5:5:0.0'.
21330 @subsection Examples
21334 Apply strong luma sharpen effect:
21336 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
21340 Apply a strong blur of both luma and chroma parameters:
21342 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
21346 @c man end OPENCL VIDEO FILTERS
21348 @chapter VAAPI Video Filters
21349 @c man begin VAAPI VIDEO FILTERS
21351 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
21353 To enable compilation of these filters you need to configure FFmpeg with
21354 @code{--enable-vaapi}.
21356 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}
21358 @section tonemap_vappi
21360 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
21361 It maps the dynamic range of HDR10 content to the SDR content.
21362 It currently only accepts HDR10 as input.
21364 It accepts the following parameters:
21368 Specify the output pixel format.
21370 Currently supported formats are:
21379 Set the output color primaries.
21381 Default is same as input.
21384 Set the output transfer characteristics.
21389 Set the output colorspace matrix.
21391 Default is same as input.
21395 @subsection Example
21399 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
21401 tonemap_vaapi=format=p010:t=bt2020-10
21405 @c man end VAAPI VIDEO FILTERS
21407 @chapter Video Sources
21408 @c man begin VIDEO SOURCES
21410 Below is a description of the currently available video sources.
21414 Buffer video frames, and make them available to the filter chain.
21416 This source is mainly intended for a programmatic use, in particular
21417 through the interface defined in @file{libavfilter/vsrc_buffer.h}.
21419 It accepts the following parameters:
21424 Specify the size (width and height) of the buffered video frames. For the
21425 syntax of this option, check the
21426 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21429 The input video width.
21432 The input video height.
21435 A string representing the pixel format of the buffered video frames.
21436 It may be a number corresponding to a pixel format, or a pixel format
21440 Specify the timebase assumed by the timestamps of the buffered frames.
21443 Specify the frame rate expected for the video stream.
21445 @item pixel_aspect, sar
21446 The sample (pixel) aspect ratio of the input video.
21449 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
21450 to the filtergraph description to specify swscale flags for automatically
21451 inserted scalers. See @ref{Filtergraph syntax}.
21453 @item hw_frames_ctx
21454 When using a hardware pixel format, this should be a reference to an
21455 AVHWFramesContext describing input frames.
21460 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
21463 will instruct the source to accept video frames with size 320x240 and
21464 with format "yuv410p", assuming 1/24 as the timestamps timebase and
21465 square pixels (1:1 sample aspect ratio).
21466 Since the pixel format with name "yuv410p" corresponds to the number 6
21467 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
21468 this example corresponds to:
21470 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
21473 Alternatively, the options can be specified as a flat string, but this
21474 syntax is deprecated:
21476 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
21480 Create a pattern generated by an elementary cellular automaton.
21482 The initial state of the cellular automaton can be defined through the
21483 @option{filename} and @option{pattern} options. If such options are
21484 not specified an initial state is created randomly.
21486 At each new frame a new row in the video is filled with the result of
21487 the cellular automaton next generation. The behavior when the whole
21488 frame is filled is defined by the @option{scroll} option.
21490 This source accepts the following options:
21494 Read the initial cellular automaton state, i.e. the starting row, from
21495 the specified file.
21496 In the file, each non-whitespace character is considered an alive
21497 cell, a newline will terminate the row, and further characters in the
21498 file will be ignored.
21501 Read the initial cellular automaton state, i.e. the starting row, from
21502 the specified string.
21504 Each non-whitespace character in the string is considered an alive
21505 cell, a newline will terminate the row, and further characters in the
21506 string will be ignored.
21509 Set the video rate, that is the number of frames generated per second.
21512 @item random_fill_ratio, ratio
21513 Set the random fill ratio for the initial cellular automaton row. It
21514 is a floating point number value ranging from 0 to 1, defaults to
21517 This option is ignored when a file or a pattern is specified.
21519 @item random_seed, seed
21520 Set the seed for filling randomly the initial row, must be an integer
21521 included between 0 and UINT32_MAX. If not specified, or if explicitly
21522 set to -1, the filter will try to use a good random seed on a best
21526 Set the cellular automaton rule, it is a number ranging from 0 to 255.
21527 Default value is 110.
21530 Set the size of the output video. For the syntax of this option, check the
21531 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21533 If @option{filename} or @option{pattern} is specified, the size is set
21534 by default to the width of the specified initial state row, and the
21535 height is set to @var{width} * PHI.
21537 If @option{size} is set, it must contain the width of the specified
21538 pattern string, and the specified pattern will be centered in the
21541 If a filename or a pattern string is not specified, the size value
21542 defaults to "320x518" (used for a randomly generated initial state).
21545 If set to 1, scroll the output upward when all the rows in the output
21546 have been already filled. If set to 0, the new generated row will be
21547 written over the top row just after the bottom row is filled.
21550 @item start_full, full
21551 If set to 1, completely fill the output with generated rows before
21552 outputting the first frame.
21553 This is the default behavior, for disabling set the value to 0.
21556 If set to 1, stitch the left and right row edges together.
21557 This is the default behavior, for disabling set the value to 0.
21560 @subsection Examples
21564 Read the initial state from @file{pattern}, and specify an output of
21567 cellauto=f=pattern:s=200x400
21571 Generate a random initial row with a width of 200 cells, with a fill
21574 cellauto=ratio=2/3:s=200x200
21578 Create a pattern generated by rule 18 starting by a single alive cell
21579 centered on an initial row with width 100:
21581 cellauto=p=@@:s=100x400:full=0:rule=18
21585 Specify a more elaborated initial pattern:
21587 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
21592 @anchor{coreimagesrc}
21593 @section coreimagesrc
21594 Video source generated on GPU using Apple's CoreImage API on OSX.
21596 This video source is a specialized version of the @ref{coreimage} video filter.
21597 Use a core image generator at the beginning of the applied filterchain to
21598 generate the content.
21600 The coreimagesrc video source accepts the following options:
21602 @item list_generators
21603 List all available generators along with all their respective options as well as
21604 possible minimum and maximum values along with the default values.
21606 list_generators=true
21610 Specify the size of the sourced video. For the syntax of this option, check the
21611 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21612 The default value is @code{320x240}.
21615 Specify the frame rate of the sourced video, as the number of frames
21616 generated per second. It has to be a string in the format
21617 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
21618 number or a valid video frame rate abbreviation. The default value is
21622 Set the sample aspect ratio of the sourced video.
21625 Set the duration of the sourced video. See
21626 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
21627 for the accepted syntax.
21629 If not specified, or the expressed duration is negative, the video is
21630 supposed to be generated forever.
21633 Additionally, all options of the @ref{coreimage} video filter are accepted.
21634 A complete filterchain can be used for further processing of the
21635 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
21636 and examples for details.
21638 @subsection Examples
21643 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
21644 given as complete and escaped command-line for Apple's standard bash shell:
21646 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
21648 This example is equivalent to the QRCode example of @ref{coreimage} without the
21649 need for a nullsrc video source.
21653 @section mandelbrot
21655 Generate a Mandelbrot set fractal, and progressively zoom towards the
21656 point specified with @var{start_x} and @var{start_y}.
21658 This source accepts the following options:
21663 Set the terminal pts value. Default value is 400.
21666 Set the terminal scale value.
21667 Must be a floating point value. Default value is 0.3.
21670 Set the inner coloring mode, that is the algorithm used to draw the
21671 Mandelbrot fractal internal region.
21673 It shall assume one of the following values:
21678 Show time until convergence.
21680 Set color based on point closest to the origin of the iterations.
21685 Default value is @var{mincol}.
21688 Set the bailout value. Default value is 10.0.
21691 Set the maximum of iterations performed by the rendering
21692 algorithm. Default value is 7189.
21695 Set outer coloring mode.
21696 It shall assume one of following values:
21698 @item iteration_count
21699 Set iteration count mode.
21700 @item normalized_iteration_count
21701 set normalized iteration count mode.
21703 Default value is @var{normalized_iteration_count}.
21706 Set frame rate, expressed as number of frames per second. Default
21710 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
21711 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
21714 Set the initial scale value. Default value is 3.0.
21717 Set the initial x position. Must be a floating point value between
21718 -100 and 100. Default value is -0.743643887037158704752191506114774.
21721 Set the initial y position. Must be a floating point value between
21722 -100 and 100. Default value is -0.131825904205311970493132056385139.
21727 Generate various test patterns, as generated by the MPlayer test filter.
21729 The size of the generated video is fixed, and is 256x256.
21730 This source is useful in particular for testing encoding features.
21732 This source accepts the following options:
21737 Specify the frame rate of the sourced video, as the number of frames
21738 generated per second. It has to be a string in the format
21739 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
21740 number or a valid video frame rate abbreviation. The default value is
21744 Set the duration of the sourced video. See
21745 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
21746 for the accepted syntax.
21748 If not specified, or the expressed duration is negative, the video is
21749 supposed to be generated forever.
21753 Set the number or the name of the test to perform. Supported tests are:
21767 @item max_frames, m
21768 Set the maximum number of frames generated for each test, default value is 30.
21772 Default value is "all", which will cycle through the list of all tests.
21777 mptestsrc=t=dc_luma
21780 will generate a "dc_luma" test pattern.
21782 @section frei0r_src
21784 Provide a frei0r source.
21786 To enable compilation of this filter you need to install the frei0r
21787 header and configure FFmpeg with @code{--enable-frei0r}.
21789 This source accepts the following parameters:
21794 The size of the video to generate. For the syntax of this option, check the
21795 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21798 The framerate of the generated video. It may be a string of the form
21799 @var{num}/@var{den} or a frame rate abbreviation.
21802 The name to the frei0r source to load. For more information regarding frei0r and
21803 how to set the parameters, read the @ref{frei0r} section in the video filters
21806 @item filter_params
21807 A '|'-separated list of parameters to pass to the frei0r source.
21811 For example, to generate a frei0r partik0l source with size 200x200
21812 and frame rate 10 which is overlaid on the overlay filter main input:
21814 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
21819 Generate a life pattern.
21821 This source is based on a generalization of John Conway's life game.
21823 The sourced input represents a life grid, each pixel represents a cell
21824 which can be in one of two possible states, alive or dead. Every cell
21825 interacts with its eight neighbours, which are the cells that are
21826 horizontally, vertically, or diagonally adjacent.
21828 At each interaction the grid evolves according to the adopted rule,
21829 which specifies the number of neighbor alive cells which will make a
21830 cell stay alive or born. The @option{rule} option allows one to specify
21833 This source accepts the following options:
21837 Set the file from which to read the initial grid state. In the file,
21838 each non-whitespace character is considered an alive cell, and newline
21839 is used to delimit the end of each row.
21841 If this option is not specified, the initial grid is generated
21845 Set the video rate, that is the number of frames generated per second.
21848 @item random_fill_ratio, ratio
21849 Set the random fill ratio for the initial random grid. It is a
21850 floating point number value ranging from 0 to 1, defaults to 1/PHI.
21851 It is ignored when a file is specified.
21853 @item random_seed, seed
21854 Set the seed for filling the initial random grid, must be an integer
21855 included between 0 and UINT32_MAX. If not specified, or if explicitly
21856 set to -1, the filter will try to use a good random seed on a best
21862 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
21863 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
21864 @var{NS} specifies the number of alive neighbor cells which make a
21865 live cell stay alive, and @var{NB} the number of alive neighbor cells
21866 which make a dead cell to become alive (i.e. to "born").
21867 "s" and "b" can be used in place of "S" and "B", respectively.
21869 Alternatively a rule can be specified by an 18-bits integer. The 9
21870 high order bits are used to encode the next cell state if it is alive
21871 for each number of neighbor alive cells, the low order bits specify
21872 the rule for "borning" new cells. Higher order bits encode for an
21873 higher number of neighbor cells.
21874 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
21875 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
21877 Default value is "S23/B3", which is the original Conway's game of life
21878 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
21879 cells, and will born a new cell if there are three alive cells around
21883 Set the size of the output video. For the syntax of this option, check the
21884 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21886 If @option{filename} is specified, the size is set by default to the
21887 same size of the input file. If @option{size} is set, it must contain
21888 the size specified in the input file, and the initial grid defined in
21889 that file is centered in the larger resulting area.
21891 If a filename is not specified, the size value defaults to "320x240"
21892 (used for a randomly generated initial grid).
21895 If set to 1, stitch the left and right grid edges together, and the
21896 top and bottom edges also. Defaults to 1.
21899 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
21900 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
21901 value from 0 to 255.
21904 Set the color of living (or new born) cells.
21907 Set the color of dead cells. If @option{mold} is set, this is the first color
21908 used to represent a dead cell.
21911 Set mold color, for definitely dead and moldy cells.
21913 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
21914 ffmpeg-utils manual,ffmpeg-utils}.
21917 @subsection Examples
21921 Read a grid from @file{pattern}, and center it on a grid of size
21924 life=f=pattern:s=300x300
21928 Generate a random grid of size 200x200, with a fill ratio of 2/3:
21930 life=ratio=2/3:s=200x200
21934 Specify a custom rule for evolving a randomly generated grid:
21940 Full example with slow death effect (mold) using @command{ffplay}:
21942 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
21949 @anchor{haldclutsrc}
21952 @anchor{pal100bars}
21953 @anchor{rgbtestsrc}
21955 @anchor{smptehdbars}
21958 @anchor{yuvtestsrc}
21959 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
21961 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
21963 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
21965 The @code{color} source provides an uniformly colored input.
21967 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
21968 @ref{haldclut} filter.
21970 The @code{nullsrc} source returns unprocessed video frames. It is
21971 mainly useful to be employed in analysis / debugging tools, or as the
21972 source for filters which ignore the input data.
21974 The @code{pal75bars} source generates a color bars pattern, based on
21975 EBU PAL recommendations with 75% color levels.
21977 The @code{pal100bars} source generates a color bars pattern, based on
21978 EBU PAL recommendations with 100% color levels.
21980 The @code{rgbtestsrc} source generates an RGB test pattern useful for
21981 detecting RGB vs BGR issues. You should see a red, green and blue
21982 stripe from top to bottom.
21984 The @code{smptebars} source generates a color bars pattern, based on
21985 the SMPTE Engineering Guideline EG 1-1990.
21987 The @code{smptehdbars} source generates a color bars pattern, based on
21988 the SMPTE RP 219-2002.
21990 The @code{testsrc} source generates a test video pattern, showing a
21991 color pattern, a scrolling gradient and a timestamp. This is mainly
21992 intended for testing purposes.
21994 The @code{testsrc2} source is similar to testsrc, but supports more
21995 pixel formats instead of just @code{rgb24}. This allows using it as an
21996 input for other tests without requiring a format conversion.
21998 The @code{yuvtestsrc} source generates an YUV test pattern. You should
21999 see a y, cb and cr stripe from top to bottom.
22001 The sources accept the following parameters:
22006 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
22007 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
22008 pixels to be used as identity matrix for 3D lookup tables. Each component is
22009 coded on a @code{1/(N*N)} scale.
22012 Specify the color of the source, only available in the @code{color}
22013 source. For the syntax of this option, check the
22014 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
22017 Specify the size of the sourced video. For the syntax of this option, check the
22018 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22019 The default value is @code{320x240}.
22021 This option is not available with the @code{allrgb}, @code{allyuv}, and
22022 @code{haldclutsrc} filters.
22025 Specify the frame rate of the sourced video, as the number of frames
22026 generated per second. It has to be a string in the format
22027 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22028 number or a valid video frame rate abbreviation. The default value is
22032 Set the duration of the sourced video. See
22033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22034 for the accepted syntax.
22036 If not specified, or the expressed duration is negative, the video is
22037 supposed to be generated forever.
22040 Set the sample aspect ratio of the sourced video.
22043 Specify the alpha (opacity) of the background, only available in the
22044 @code{testsrc2} source. The value must be between 0 (fully transparent) and
22045 255 (fully opaque, the default).
22048 Set the number of decimals to show in the timestamp, only available in the
22049 @code{testsrc} source.
22051 The displayed timestamp value will correspond to the original
22052 timestamp value multiplied by the power of 10 of the specified
22053 value. Default value is 0.
22056 @subsection Examples
22060 Generate a video with a duration of 5.3 seconds, with size
22061 176x144 and a frame rate of 10 frames per second:
22063 testsrc=duration=5.3:size=qcif:rate=10
22067 The following graph description will generate a red source
22068 with an opacity of 0.2, with size "qcif" and a frame rate of 10
22071 color=c=red@@0.2:s=qcif:r=10
22075 If the input content is to be ignored, @code{nullsrc} can be used. The
22076 following command generates noise in the luminance plane by employing
22077 the @code{geq} filter:
22079 nullsrc=s=256x256, geq=random(1)*255:128:128
22083 @subsection Commands
22085 The @code{color} source supports the following commands:
22089 Set the color of the created image. Accepts the same syntax of the
22090 corresponding @option{color} option.
22095 Generate video using an OpenCL program.
22100 OpenCL program source file.
22103 Kernel name in program.
22106 Size of frames to generate. This must be set.
22109 Pixel format to use for the generated frames. This must be set.
22112 Number of frames generated every second. Default value is '25'.
22116 For details of how the program loading works, see the @ref{program_opencl}
22123 Generate a colour ramp by setting pixel values from the position of the pixel
22124 in the output image. (Note that this will work with all pixel formats, but
22125 the generated output will not be the same.)
22127 __kernel void ramp(__write_only image2d_t dst,
22128 unsigned int index)
22130 int2 loc = (int2)(get_global_id(0), get_global_id(1));
22133 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
22135 write_imagef(dst, loc, val);
22140 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
22142 __kernel void sierpinski_carpet(__write_only image2d_t dst,
22143 unsigned int index)
22145 int2 loc = (int2)(get_global_id(0), get_global_id(1));
22147 float4 value = 0.0f;
22148 int x = loc.x + index;
22149 int y = loc.y + index;
22150 while (x > 0 || y > 0) {
22151 if (x % 3 == 1 && y % 3 == 1) {
22159 write_imagef(dst, loc, value);
22165 @section sierpinski
22167 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
22169 This source accepts the following options:
22173 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22174 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22177 Set frame rate, expressed as number of frames per second. Default
22181 Set seed which is used for random panning.
22184 Set max jump for single pan destination. Allowed range is from 1 to 10000.
22187 Set fractal type, can be default @code{carpet} or @code{triangle}.
22190 @c man end VIDEO SOURCES
22192 @chapter Video Sinks
22193 @c man begin VIDEO SINKS
22195 Below is a description of the currently available video sinks.
22197 @section buffersink
22199 Buffer video frames, and make them available to the end of the filter
22202 This sink is mainly intended for programmatic use, in particular
22203 through the interface defined in @file{libavfilter/buffersink.h}
22204 or the options system.
22206 It accepts a pointer to an AVBufferSinkContext structure, which
22207 defines the incoming buffers' formats, to be passed as the opaque
22208 parameter to @code{avfilter_init_filter} for initialization.
22212 Null video sink: do absolutely nothing with the input video. It is
22213 mainly useful as a template and for use in analysis / debugging
22216 @c man end VIDEO SINKS
22218 @chapter Multimedia Filters
22219 @c man begin MULTIMEDIA FILTERS
22221 Below is a description of the currently available multimedia filters.
22225 Convert input audio to a video output, displaying the audio bit scope.
22227 The filter accepts the following options:
22231 Set frame rate, expressed as number of frames per second. Default
22235 Specify the video size for the output. For the syntax of this option, check the
22236 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22237 Default value is @code{1024x256}.
22240 Specify list of colors separated by space or by '|' which will be used to
22241 draw channels. Unrecognized or missing colors will be replaced
22245 @section adrawgraph
22246 Draw a graph using input audio metadata.
22248 See @ref{drawgraph}
22250 @section agraphmonitor
22252 See @ref{graphmonitor}.
22254 @section ahistogram
22256 Convert input audio to a video output, displaying the volume histogram.
22258 The filter accepts the following options:
22262 Specify how histogram is calculated.
22264 It accepts the following values:
22267 Use single histogram for all channels.
22269 Use separate histogram for each channel.
22271 Default is @code{single}.
22274 Set frame rate, expressed as number of frames per second. Default
22278 Specify the video size for the output. For the syntax of this option, check the
22279 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22280 Default value is @code{hd720}.
22285 It accepts the following values:
22296 reverse logarithmic
22298 Default is @code{log}.
22301 Set amplitude scale.
22303 It accepts the following values:
22310 Default is @code{log}.
22313 Set how much frames to accumulate in histogram.
22314 Default is 1. Setting this to -1 accumulates all frames.
22317 Set histogram ratio of window height.
22320 Set sonogram sliding.
22322 It accepts the following values:
22325 replace old rows with new ones.
22327 scroll from top to bottom.
22329 Default is @code{replace}.
22332 @section aphasemeter
22334 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
22335 representing mean phase of current audio frame. A video output can also be produced and is
22336 enabled by default. The audio is passed through as first output.
22338 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
22339 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
22340 and @code{1} means channels are in phase.
22342 The filter accepts the following options, all related to its video output:
22346 Set the output frame rate. Default value is @code{25}.
22349 Set the video size for the output. For the syntax of this option, check the
22350 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22351 Default value is @code{800x400}.
22356 Specify the red, green, blue contrast. Default values are @code{2},
22357 @code{7} and @code{1}.
22358 Allowed range is @code{[0, 255]}.
22361 Set color which will be used for drawing median phase. If color is
22362 @code{none} which is default, no median phase value will be drawn.
22365 Enable video output. Default is enabled.
22368 @section avectorscope
22370 Convert input audio to a video output, representing the audio vector
22373 The filter is used to measure the difference between channels of stereo
22374 audio stream. A monaural signal, consisting of identical left and right
22375 signal, results in straight vertical line. Any stereo separation is visible
22376 as a deviation from this line, creating a Lissajous figure.
22377 If the straight (or deviation from it) but horizontal line appears this
22378 indicates that the left and right channels are out of phase.
22380 The filter accepts the following options:
22384 Set the vectorscope mode.
22386 Available values are:
22389 Lissajous rotated by 45 degrees.
22392 Same as above but not rotated.
22395 Shape resembling half of circle.
22398 Default value is @samp{lissajous}.
22401 Set the video size for the output. For the syntax of this option, check the
22402 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22403 Default value is @code{400x400}.
22406 Set the output frame rate. Default value is @code{25}.
22412 Specify the red, green, blue and alpha contrast. Default values are @code{40},
22413 @code{160}, @code{80} and @code{255}.
22414 Allowed range is @code{[0, 255]}.
22420 Specify the red, green, blue and alpha fade. Default values are @code{15},
22421 @code{10}, @code{5} and @code{5}.
22422 Allowed range is @code{[0, 255]}.
22425 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
22426 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
22429 Set the vectorscope drawing mode.
22431 Available values are:
22434 Draw dot for each sample.
22437 Draw line between previous and current sample.
22440 Default value is @samp{dot}.
22443 Specify amplitude scale of audio samples.
22445 Available values are:
22461 Swap left channel axis with right channel axis.
22471 Mirror only x axis.
22474 Mirror only y axis.
22482 @subsection Examples
22486 Complete example using @command{ffplay}:
22488 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
22489 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
22493 @section bench, abench
22495 Benchmark part of a filtergraph.
22497 The filter accepts the following options:
22501 Start or stop a timer.
22503 Available values are:
22506 Get the current time, set it as frame metadata (using the key
22507 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
22510 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
22511 the input frame metadata to get the time difference. Time difference, average,
22512 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
22513 @code{min}) are then printed. The timestamps are expressed in seconds.
22517 @subsection Examples
22521 Benchmark @ref{selectivecolor} filter:
22523 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
22529 Concatenate audio and video streams, joining them together one after the
22532 The filter works on segments of synchronized video and audio streams. All
22533 segments must have the same number of streams of each type, and that will
22534 also be the number of streams at output.
22536 The filter accepts the following options:
22541 Set the number of segments. Default is 2.
22544 Set the number of output video streams, that is also the number of video
22545 streams in each segment. Default is 1.
22548 Set the number of output audio streams, that is also the number of audio
22549 streams in each segment. Default is 0.
22552 Activate unsafe mode: do not fail if segments have a different format.
22556 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
22557 @var{a} audio outputs.
22559 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
22560 segment, in the same order as the outputs, then the inputs for the second
22563 Related streams do not always have exactly the same duration, for various
22564 reasons including codec frame size or sloppy authoring. For that reason,
22565 related synchronized streams (e.g. a video and its audio track) should be
22566 concatenated at once. The concat filter will use the duration of the longest
22567 stream in each segment (except the last one), and if necessary pad shorter
22568 audio streams with silence.
22570 For this filter to work correctly, all segments must start at timestamp 0.
22572 All corresponding streams must have the same parameters in all segments; the
22573 filtering system will automatically select a common pixel format for video
22574 streams, and a common sample format, sample rate and channel layout for
22575 audio streams, but other settings, such as resolution, must be converted
22576 explicitly by the user.
22578 Different frame rates are acceptable but will result in variable frame rate
22579 at output; be sure to configure the output file to handle it.
22581 @subsection Examples
22585 Concatenate an opening, an episode and an ending, all in bilingual version
22586 (video in stream 0, audio in streams 1 and 2):
22588 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
22589 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
22590 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
22591 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
22595 Concatenate two parts, handling audio and video separately, using the
22596 (a)movie sources, and adjusting the resolution:
22598 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
22599 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
22600 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
22602 Note that a desync will happen at the stitch if the audio and video streams
22603 do not have exactly the same duration in the first file.
22607 @subsection Commands
22609 This filter supports the following commands:
22612 Close the current segment and step to the next one
22618 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
22619 level. By default, it logs a message at a frequency of 10Hz with the
22620 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
22621 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
22623 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
22624 sample format is double-precision floating point. The input stream will be converted to
22625 this specification, if needed. Users may need to insert aformat and/or aresample filters
22626 after this filter to obtain the original parameters.
22628 The filter also has a video output (see the @var{video} option) with a real
22629 time graph to observe the loudness evolution. The graphic contains the logged
22630 message mentioned above, so it is not printed anymore when this option is set,
22631 unless the verbose logging is set. The main graphing area contains the
22632 short-term loudness (3 seconds of analysis), and the gauge on the right is for
22633 the momentary loudness (400 milliseconds), but can optionally be configured
22634 to instead display short-term loudness (see @var{gauge}).
22636 The green area marks a +/- 1LU target range around the target loudness
22637 (-23LUFS by default, unless modified through @var{target}).
22639 More information about the Loudness Recommendation EBU R128 on
22640 @url{http://tech.ebu.ch/loudness}.
22642 The filter accepts the following options:
22647 Activate the video output. The audio stream is passed unchanged whether this
22648 option is set or no. The video stream will be the first output stream if
22649 activated. Default is @code{0}.
22652 Set the video size. This option is for video only. For the syntax of this
22654 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22655 Default and minimum resolution is @code{640x480}.
22658 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
22659 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
22660 other integer value between this range is allowed.
22663 Set metadata injection. If set to @code{1}, the audio input will be segmented
22664 into 100ms output frames, each of them containing various loudness information
22665 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
22667 Default is @code{0}.
22670 Force the frame logging level.
22672 Available values are:
22675 information logging level
22677 verbose logging level
22680 By default, the logging level is set to @var{info}. If the @option{video} or
22681 the @option{metadata} options are set, it switches to @var{verbose}.
22686 Available modes can be cumulated (the option is a @code{flag} type). Possible
22690 Disable any peak mode (default).
22692 Enable sample-peak mode.
22694 Simple peak mode looking for the higher sample value. It logs a message
22695 for sample-peak (identified by @code{SPK}).
22697 Enable true-peak mode.
22699 If enabled, the peak lookup is done on an over-sampled version of the input
22700 stream for better peak accuracy. It logs a message for true-peak.
22701 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
22702 This mode requires a build with @code{libswresample}.
22706 Treat mono input files as "dual mono". If a mono file is intended for playback
22707 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
22708 If set to @code{true}, this option will compensate for this effect.
22709 Multi-channel input files are not affected by this option.
22712 Set a specific pan law to be used for the measurement of dual mono files.
22713 This parameter is optional, and has a default value of -3.01dB.
22716 Set a specific target level (in LUFS) used as relative zero in the visualization.
22717 This parameter is optional and has a default value of -23LUFS as specified
22718 by EBU R128. However, material published online may prefer a level of -16LUFS
22719 (e.g. for use with podcasts or video platforms).
22722 Set the value displayed by the gauge. Valid values are @code{momentary} and s
22723 @code{shortterm}. By default the momentary value will be used, but in certain
22724 scenarios it may be more useful to observe the short term value instead (e.g.
22728 Sets the display scale for the loudness. Valid parameters are @code{absolute}
22729 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
22730 video output, not the summary or continuous log output.
22733 @subsection Examples
22737 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
22739 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
22743 Run an analysis with @command{ffmpeg}:
22745 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
22749 @section interleave, ainterleave
22751 Temporally interleave frames from several inputs.
22753 @code{interleave} works with video inputs, @code{ainterleave} with audio.
22755 These filters read frames from several inputs and send the oldest
22756 queued frame to the output.
22758 Input streams must have well defined, monotonically increasing frame
22761 In order to submit one frame to output, these filters need to enqueue
22762 at least one frame for each input, so they cannot work in case one
22763 input is not yet terminated and will not receive incoming frames.
22765 For example consider the case when one input is a @code{select} filter
22766 which always drops input frames. The @code{interleave} filter will keep
22767 reading from that input, but it will never be able to send new frames
22768 to output until the input sends an end-of-stream signal.
22770 Also, depending on inputs synchronization, the filters will drop
22771 frames in case one input receives more frames than the other ones, and
22772 the queue is already filled.
22774 These filters accept the following options:
22778 Set the number of different inputs, it is 2 by default.
22781 @subsection Examples
22785 Interleave frames belonging to different streams using @command{ffmpeg}:
22787 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
22791 Add flickering blur effect:
22793 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
22797 @section metadata, ametadata
22799 Manipulate frame metadata.
22801 This filter accepts the following options:
22805 Set mode of operation of the filter.
22807 Can be one of the following:
22811 If both @code{value} and @code{key} is set, select frames
22812 which have such metadata. If only @code{key} is set, select
22813 every frame that has such key in metadata.
22816 Add new metadata @code{key} and @code{value}. If key is already available
22820 Modify value of already present key.
22823 If @code{value} is set, delete only keys that have such value.
22824 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
22828 Print key and its value if metadata was found. If @code{key} is not set print all
22829 metadata values available in frame.
22833 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
22836 Set metadata value which will be used. This option is mandatory for
22837 @code{modify} and @code{add} mode.
22840 Which function to use when comparing metadata value and @code{value}.
22842 Can be one of following:
22846 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
22849 Values are interpreted as strings, returns true if metadata value starts with
22850 the @code{value} option string.
22853 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
22856 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
22859 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
22862 Values are interpreted as floats, returns true if expression from option @code{expr}
22866 Values are interpreted as strings, returns true if metadata value ends with
22867 the @code{value} option string.
22871 Set expression which is used when @code{function} is set to @code{expr}.
22872 The expression is evaluated through the eval API and can contain the following
22877 Float representation of @code{value} from metadata key.
22880 Float representation of @code{value} as supplied by user in @code{value} option.
22884 If specified in @code{print} mode, output is written to the named file. Instead of
22885 plain filename any writable url can be specified. Filename ``-'' is a shorthand
22886 for standard output. If @code{file} option is not set, output is written to the log
22887 with AV_LOG_INFO loglevel.
22890 Reduces buffering in print mode when output is written to a URL set using @var{file}.
22894 @subsection Examples
22898 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
22901 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
22904 Print silencedetect output to file @file{metadata.txt}.
22906 silencedetect,ametadata=mode=print:file=metadata.txt
22909 Direct all metadata to a pipe with file descriptor 4.
22911 metadata=mode=print:file='pipe\:4'
22915 @section perms, aperms
22917 Set read/write permissions for the output frames.
22919 These filters are mainly aimed at developers to test direct path in the
22920 following filter in the filtergraph.
22922 The filters accept the following options:
22926 Select the permissions mode.
22928 It accepts the following values:
22931 Do nothing. This is the default.
22933 Set all the output frames read-only.
22935 Set all the output frames directly writable.
22937 Make the frame read-only if writable, and writable if read-only.
22939 Set each output frame read-only or writable randomly.
22943 Set the seed for the @var{random} mode, must be an integer included between
22944 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
22945 @code{-1}, the filter will try to use a good random seed on a best effort
22949 Note: in case of auto-inserted filter between the permission filter and the
22950 following one, the permission might not be received as expected in that
22951 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
22952 perms/aperms filter can avoid this problem.
22954 @section realtime, arealtime
22956 Slow down filtering to match real time approximately.
22958 These filters will pause the filtering for a variable amount of time to
22959 match the output rate with the input timestamps.
22960 They are similar to the @option{re} option to @code{ffmpeg}.
22962 They accept the following options:
22966 Time limit for the pauses. Any pause longer than that will be considered
22967 a timestamp discontinuity and reset the timer. Default is 2 seconds.
22969 Speed factor for processing. The value must be a float larger than zero.
22970 Values larger than 1.0 will result in faster than realtime processing,
22971 smaller will slow processing down. The @var{limit} is automatically adapted
22972 accordingly. Default is 1.0.
22974 A processing speed faster than what is possible without these filters cannot
22979 @section select, aselect
22981 Select frames to pass in output.
22983 This filter accepts the following options:
22988 Set expression, which is evaluated for each input frame.
22990 If the expression is evaluated to zero, the frame is discarded.
22992 If the evaluation result is negative or NaN, the frame is sent to the
22993 first output; otherwise it is sent to the output with index
22994 @code{ceil(val)-1}, assuming that the input index starts from 0.
22996 For example a value of @code{1.2} corresponds to the output with index
22997 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
23000 Set the number of outputs. The output to which to send the selected
23001 frame is based on the result of the evaluation. Default value is 1.
23004 The expression can contain the following constants:
23008 The (sequential) number of the filtered frame, starting from 0.
23011 The (sequential) number of the selected frame, starting from 0.
23013 @item prev_selected_n
23014 The sequential number of the last selected frame. It's NAN if undefined.
23017 The timebase of the input timestamps.
23020 The PTS (Presentation TimeStamp) of the filtered video frame,
23021 expressed in @var{TB} units. It's NAN if undefined.
23024 The PTS of the filtered video frame,
23025 expressed in seconds. It's NAN if undefined.
23028 The PTS of the previously filtered video frame. It's NAN if undefined.
23030 @item prev_selected_pts
23031 The PTS of the last previously filtered video frame. It's NAN if undefined.
23033 @item prev_selected_t
23034 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
23037 The PTS of the first video frame in the video. It's NAN if undefined.
23040 The time of the first video frame in the video. It's NAN if undefined.
23042 @item pict_type @emph{(video only)}
23043 The type of the filtered frame. It can assume one of the following
23055 @item interlace_type @emph{(video only)}
23056 The frame interlace type. It can assume one of the following values:
23059 The frame is progressive (not interlaced).
23061 The frame is top-field-first.
23063 The frame is bottom-field-first.
23066 @item consumed_sample_n @emph{(audio only)}
23067 the number of selected samples before the current frame
23069 @item samples_n @emph{(audio only)}
23070 the number of samples in the current frame
23072 @item sample_rate @emph{(audio only)}
23073 the input sample rate
23076 This is 1 if the filtered frame is a key-frame, 0 otherwise.
23079 the position in the file of the filtered frame, -1 if the information
23080 is not available (e.g. for synthetic video)
23082 @item scene @emph{(video only)}
23083 value between 0 and 1 to indicate a new scene; a low value reflects a low
23084 probability for the current frame to introduce a new scene, while a higher
23085 value means the current frame is more likely to be one (see the example below)
23087 @item concatdec_select
23088 The concat demuxer can select only part of a concat input file by setting an
23089 inpoint and an outpoint, but the output packets may not be entirely contained
23090 in the selected interval. By using this variable, it is possible to skip frames
23091 generated by the concat demuxer which are not exactly contained in the selected
23094 This works by comparing the frame pts against the @var{lavf.concat.start_time}
23095 and the @var{lavf.concat.duration} packet metadata values which are also
23096 present in the decoded frames.
23098 The @var{concatdec_select} variable is -1 if the frame pts is at least
23099 start_time and either the duration metadata is missing or the frame pts is less
23100 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
23103 That basically means that an input frame is selected if its pts is within the
23104 interval set by the concat demuxer.
23108 The default value of the select expression is "1".
23110 @subsection Examples
23114 Select all frames in input:
23119 The example above is the same as:
23131 Select only I-frames:
23133 select='eq(pict_type\,I)'
23137 Select one frame every 100:
23139 select='not(mod(n\,100))'
23143 Select only frames contained in the 10-20 time interval:
23145 select=between(t\,10\,20)
23149 Select only I-frames contained in the 10-20 time interval:
23151 select=between(t\,10\,20)*eq(pict_type\,I)
23155 Select frames with a minimum distance of 10 seconds:
23157 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
23161 Use aselect to select only audio frames with samples number > 100:
23163 aselect='gt(samples_n\,100)'
23167 Create a mosaic of the first scenes:
23169 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
23172 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
23176 Send even and odd frames to separate outputs, and compose them:
23178 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
23182 Select useful frames from an ffconcat file which is using inpoints and
23183 outpoints but where the source files are not intra frame only.
23185 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
23189 @section sendcmd, asendcmd
23191 Send commands to filters in the filtergraph.
23193 These filters read commands to be sent to other filters in the
23196 @code{sendcmd} must be inserted between two video filters,
23197 @code{asendcmd} must be inserted between two audio filters, but apart
23198 from that they act the same way.
23200 The specification of commands can be provided in the filter arguments
23201 with the @var{commands} option, or in a file specified by the
23202 @var{filename} option.
23204 These filters accept the following options:
23207 Set the commands to be read and sent to the other filters.
23209 Set the filename of the commands to be read and sent to the other
23213 @subsection Commands syntax
23215 A commands description consists of a sequence of interval
23216 specifications, comprising a list of commands to be executed when a
23217 particular event related to that interval occurs. The occurring event
23218 is typically the current frame time entering or leaving a given time
23221 An interval is specified by the following syntax:
23223 @var{START}[-@var{END}] @var{COMMANDS};
23226 The time interval is specified by the @var{START} and @var{END} times.
23227 @var{END} is optional and defaults to the maximum time.
23229 The current frame time is considered within the specified interval if
23230 it is included in the interval [@var{START}, @var{END}), that is when
23231 the time is greater or equal to @var{START} and is lesser than
23234 @var{COMMANDS} consists of a sequence of one or more command
23235 specifications, separated by ",", relating to that interval. The
23236 syntax of a command specification is given by:
23238 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
23241 @var{FLAGS} is optional and specifies the type of events relating to
23242 the time interval which enable sending the specified command, and must
23243 be a non-null sequence of identifier flags separated by "+" or "|" and
23244 enclosed between "[" and "]".
23246 The following flags are recognized:
23249 The command is sent when the current frame timestamp enters the
23250 specified interval. In other words, the command is sent when the
23251 previous frame timestamp was not in the given interval, and the
23255 The command is sent when the current frame timestamp leaves the
23256 specified interval. In other words, the command is sent when the
23257 previous frame timestamp was in the given interval, and the
23261 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
23264 @var{TARGET} specifies the target of the command, usually the name of
23265 the filter class or a specific filter instance name.
23267 @var{COMMAND} specifies the name of the command for the target filter.
23269 @var{ARG} is optional and specifies the optional list of argument for
23270 the given @var{COMMAND}.
23272 Between one interval specification and another, whitespaces, or
23273 sequences of characters starting with @code{#} until the end of line,
23274 are ignored and can be used to annotate comments.
23276 A simplified BNF description of the commands specification syntax
23279 @var{COMMAND_FLAG} ::= "enter" | "leave"
23280 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
23281 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
23282 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
23283 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
23284 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
23287 @subsection Examples
23291 Specify audio tempo change at second 4:
23293 asendcmd=c='4.0 atempo tempo 1.5',atempo
23297 Target a specific filter instance:
23299 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
23303 Specify a list of drawtext and hue commands in a file.
23305 # show text in the interval 5-10
23306 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
23307 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
23309 # desaturate the image in the interval 15-20
23310 15.0-20.0 [enter] hue s 0,
23311 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
23313 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
23315 # apply an exponential saturation fade-out effect, starting from time 25
23316 25 [enter] hue s exp(25-t)
23319 A filtergraph allowing to read and process the above command list
23320 stored in a file @file{test.cmd}, can be specified with:
23322 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
23327 @section setpts, asetpts
23329 Change the PTS (presentation timestamp) of the input frames.
23331 @code{setpts} works on video frames, @code{asetpts} on audio frames.
23333 This filter accepts the following options:
23338 The expression which is evaluated for each frame to construct its timestamp.
23342 The expression is evaluated through the eval API and can contain the following
23346 @item FRAME_RATE, FR
23347 frame rate, only defined for constant frame-rate video
23350 The presentation timestamp in input
23353 The count of the input frame for video or the number of consumed samples,
23354 not including the current frame for audio, starting from 0.
23356 @item NB_CONSUMED_SAMPLES
23357 The number of consumed samples, not including the current frame (only
23360 @item NB_SAMPLES, S
23361 The number of samples in the current frame (only audio)
23363 @item SAMPLE_RATE, SR
23364 The audio sample rate.
23367 The PTS of the first frame.
23370 the time in seconds of the first frame
23373 State whether the current frame is interlaced.
23376 the time in seconds of the current frame
23379 original position in the file of the frame, or undefined if undefined
23380 for the current frame
23383 The previous input PTS.
23386 previous input time in seconds
23389 The previous output PTS.
23392 previous output time in seconds
23395 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
23399 The wallclock (RTC) time at the start of the movie in microseconds.
23402 The timebase of the input timestamps.
23406 @subsection Examples
23410 Start counting PTS from zero
23412 setpts=PTS-STARTPTS
23416 Apply fast motion effect:
23422 Apply slow motion effect:
23428 Set fixed rate of 25 frames per second:
23434 Set fixed rate 25 fps with some jitter:
23436 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
23440 Apply an offset of 10 seconds to the input PTS:
23446 Generate timestamps from a "live source" and rebase onto the current timebase:
23448 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
23452 Generate timestamps by counting samples:
23461 Force color range for the output video frame.
23463 The @code{setrange} filter marks the color range property for the
23464 output frames. It does not change the input frame, but only sets the
23465 corresponding property, which affects how the frame is treated by
23468 The filter accepts the following options:
23473 Available values are:
23477 Keep the same color range property.
23479 @item unspecified, unknown
23480 Set the color range as unspecified.
23482 @item limited, tv, mpeg
23483 Set the color range as limited.
23485 @item full, pc, jpeg
23486 Set the color range as full.
23490 @section settb, asettb
23492 Set the timebase to use for the output frames timestamps.
23493 It is mainly useful for testing timebase configuration.
23495 It accepts the following parameters:
23500 The expression which is evaluated into the output timebase.
23504 The value for @option{tb} is an arithmetic expression representing a
23505 rational. The expression can contain the constants "AVTB" (the default
23506 timebase), "intb" (the input timebase) and "sr" (the sample rate,
23507 audio only). Default value is "intb".
23509 @subsection Examples
23513 Set the timebase to 1/25:
23519 Set the timebase to 1/10:
23525 Set the timebase to 1001/1000:
23531 Set the timebase to 2*intb:
23537 Set the default timebase value:
23544 Convert input audio to a video output representing frequency spectrum
23545 logarithmically using Brown-Puckette constant Q transform algorithm with
23546 direct frequency domain coefficient calculation (but the transform itself
23547 is not really constant Q, instead the Q factor is actually variable/clamped),
23548 with musical tone scale, from E0 to D#10.
23550 The filter accepts the following options:
23554 Specify the video size for the output. It must be even. For the syntax of this option,
23555 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23556 Default value is @code{1920x1080}.
23559 Set the output frame rate. Default value is @code{25}.
23562 Set the bargraph height. It must be even. Default value is @code{-1} which
23563 computes the bargraph height automatically.
23566 Set the axis height. It must be even. Default value is @code{-1} which computes
23567 the axis height automatically.
23570 Set the sonogram height. It must be even. Default value is @code{-1} which
23571 computes the sonogram height automatically.
23574 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
23575 instead. Default value is @code{1}.
23577 @item sono_v, volume
23578 Specify the sonogram volume expression. It can contain variables:
23581 the @var{bar_v} evaluated expression
23582 @item frequency, freq, f
23583 the frequency where it is evaluated
23584 @item timeclamp, tc
23585 the value of @var{timeclamp} option
23589 @item a_weighting(f)
23590 A-weighting of equal loudness
23591 @item b_weighting(f)
23592 B-weighting of equal loudness
23593 @item c_weighting(f)
23594 C-weighting of equal loudness.
23596 Default value is @code{16}.
23598 @item bar_v, volume2
23599 Specify the bargraph volume expression. It can contain variables:
23602 the @var{sono_v} evaluated expression
23603 @item frequency, freq, f
23604 the frequency where it is evaluated
23605 @item timeclamp, tc
23606 the value of @var{timeclamp} option
23610 @item a_weighting(f)
23611 A-weighting of equal loudness
23612 @item b_weighting(f)
23613 B-weighting of equal loudness
23614 @item c_weighting(f)
23615 C-weighting of equal loudness.
23617 Default value is @code{sono_v}.
23619 @item sono_g, gamma
23620 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
23621 higher gamma makes the spectrum having more range. Default value is @code{3}.
23622 Acceptable range is @code{[1, 7]}.
23624 @item bar_g, gamma2
23625 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
23629 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
23630 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
23632 @item timeclamp, tc
23633 Specify the transform timeclamp. At low frequency, there is trade-off between
23634 accuracy in time domain and frequency domain. If timeclamp is lower,
23635 event in time domain is represented more accurately (such as fast bass drum),
23636 otherwise event in frequency domain is represented more accurately
23637 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
23640 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
23641 limits future samples by applying asymmetric windowing in time domain, useful
23642 when low latency is required. Accepted range is @code{[0, 1]}.
23645 Specify the transform base frequency. Default value is @code{20.01523126408007475},
23646 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
23649 Specify the transform end frequency. Default value is @code{20495.59681441799654},
23650 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
23653 This option is deprecated and ignored.
23656 Specify the transform length in time domain. Use this option to control accuracy
23657 trade-off between time domain and frequency domain at every frequency sample.
23658 It can contain variables:
23660 @item frequency, freq, f
23661 the frequency where it is evaluated
23662 @item timeclamp, tc
23663 the value of @var{timeclamp} option.
23665 Default value is @code{384*tc/(384+tc*f)}.
23668 Specify the transform count for every video frame. Default value is @code{6}.
23669 Acceptable range is @code{[1, 30]}.
23672 Specify the transform count for every single pixel. Default value is @code{0},
23673 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
23676 Specify font file for use with freetype to draw the axis. If not specified,
23677 use embedded font. Note that drawing with font file or embedded font is not
23678 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
23682 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
23683 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
23687 Specify font color expression. This is arithmetic expression that should return
23688 integer value 0xRRGGBB. It can contain variables:
23690 @item frequency, freq, f
23691 the frequency where it is evaluated
23692 @item timeclamp, tc
23693 the value of @var{timeclamp} option
23698 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
23699 @item r(x), g(x), b(x)
23700 red, green, and blue value of intensity x.
23702 Default value is @code{st(0, (midi(f)-59.5)/12);
23703 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
23704 r(1-ld(1)) + b(ld(1))}.
23707 Specify image file to draw the axis. This option override @var{fontfile} and
23708 @var{fontcolor} option.
23711 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
23712 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
23713 Default value is @code{1}.
23716 Set colorspace. The accepted values are:
23719 Unspecified (default)
23728 BT.470BG or BT.601-6 625
23731 SMPTE-170M or BT.601-6 525
23737 BT.2020 with non-constant luminance
23742 Set spectrogram color scheme. This is list of floating point values with format
23743 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
23744 The default is @code{1|0.5|0|0|0.5|1}.
23748 @subsection Examples
23752 Playing audio while showing the spectrum:
23754 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
23758 Same as above, but with frame rate 30 fps:
23760 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
23764 Playing at 1280x720:
23766 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
23770 Disable sonogram display:
23776 A1 and its harmonics: A1, A2, (near)E3, A3:
23778 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),
23779 asplit[a][out1]; [a] showcqt [out0]'
23783 Same as above, but with more accuracy in frequency domain:
23785 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),
23786 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
23792 bar_v=10:sono_v=bar_v*a_weighting(f)
23796 Custom gamma, now spectrum is linear to the amplitude.
23802 Custom tlength equation:
23804 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)))'
23808 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
23810 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
23814 Custom font using fontconfig:
23816 font='Courier New,Monospace,mono|bold'
23820 Custom frequency range with custom axis using image file:
23822 axisfile=myaxis.png:basefreq=40:endfreq=10000
23828 Convert input audio to video output representing the audio power spectrum.
23829 Audio amplitude is on Y-axis while frequency is on X-axis.
23831 The filter accepts the following options:
23835 Specify size of video. For the syntax of this option, check the
23836 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23837 Default is @code{1024x512}.
23841 This set how each frequency bin will be represented.
23843 It accepts the following values:
23849 Default is @code{bar}.
23852 Set amplitude scale.
23854 It accepts the following values:
23868 Default is @code{log}.
23871 Set frequency scale.
23873 It accepts the following values:
23882 Reverse logarithmic scale.
23884 Default is @code{lin}.
23887 Set window size. Allowed range is from 16 to 65536.
23889 Default is @code{2048}
23892 Set windowing function.
23894 It accepts the following values:
23917 Default is @code{hanning}.
23920 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
23921 which means optimal overlap for selected window function will be picked.
23924 Set time averaging. Setting this to 0 will display current maximal peaks.
23925 Default is @code{1}, which means time averaging is disabled.
23928 Specify list of colors separated by space or by '|' which will be used to
23929 draw channel frequencies. Unrecognized or missing colors will be replaced
23933 Set channel display mode.
23935 It accepts the following values:
23940 Default is @code{combined}.
23943 Set minimum amplitude used in @code{log} amplitude scaler.
23947 @section showspatial
23949 Convert stereo input audio to a video output, representing the spatial relationship
23950 between two channels.
23952 The filter accepts the following options:
23956 Specify the video size for the output. For the syntax of this option, check the
23957 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23958 Default value is @code{512x512}.
23961 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
23964 Set window function.
23966 It accepts the following values:
23991 Default value is @code{hann}.
23994 Set ratio of overlap window. Default value is @code{0.5}.
23995 When value is @code{1} overlap is set to recommended size for specific
23996 window function currently used.
23999 @anchor{showspectrum}
24000 @section showspectrum
24002 Convert input audio to a video output, representing the audio frequency
24005 The filter accepts the following options:
24009 Specify the video size for the output. For the syntax of this option, check the
24010 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24011 Default value is @code{640x512}.
24014 Specify how the spectrum should slide along the window.
24016 It accepts the following values:
24019 the samples start again on the left when they reach the right
24021 the samples scroll from right to left
24023 frames are only produced when the samples reach the right
24025 the samples scroll from left to right
24028 Default value is @code{replace}.
24031 Specify display mode.
24033 It accepts the following values:
24036 all channels are displayed in the same row
24038 all channels are displayed in separate rows
24041 Default value is @samp{combined}.
24044 Specify display color mode.
24046 It accepts the following values:
24049 each channel is displayed in a separate color
24051 each channel is displayed using the same color scheme
24053 each channel is displayed using the rainbow color scheme
24055 each channel is displayed using the moreland color scheme
24057 each channel is displayed using the nebulae color scheme
24059 each channel is displayed using the fire color scheme
24061 each channel is displayed using the fiery color scheme
24063 each channel is displayed using the fruit color scheme
24065 each channel is displayed using the cool color scheme
24067 each channel is displayed using the magma color scheme
24069 each channel is displayed using the green color scheme
24071 each channel is displayed using the viridis color scheme
24073 each channel is displayed using the plasma color scheme
24075 each channel is displayed using the cividis color scheme
24077 each channel is displayed using the terrain color scheme
24080 Default value is @samp{channel}.
24083 Specify scale used for calculating intensity color values.
24085 It accepts the following values:
24090 square root, default
24101 Default value is @samp{sqrt}.
24104 Specify frequency scale.
24106 It accepts the following values:
24114 Default value is @samp{lin}.
24117 Set saturation modifier for displayed colors. Negative values provide
24118 alternative color scheme. @code{0} is no saturation at all.
24119 Saturation must be in [-10.0, 10.0] range.
24120 Default value is @code{1}.
24123 Set window function.
24125 It accepts the following values:
24150 Default value is @code{hann}.
24153 Set orientation of time vs frequency axis. Can be @code{vertical} or
24154 @code{horizontal}. Default is @code{vertical}.
24157 Set ratio of overlap window. Default value is @code{0}.
24158 When value is @code{1} overlap is set to recommended size for specific
24159 window function currently used.
24162 Set scale gain for calculating intensity color values.
24163 Default value is @code{1}.
24166 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
24169 Set color rotation, must be in [-1.0, 1.0] range.
24170 Default value is @code{0}.
24173 Set start frequency from which to display spectrogram. Default is @code{0}.
24176 Set stop frequency to which to display spectrogram. Default is @code{0}.
24179 Set upper frame rate limit. Default is @code{auto}, unlimited.
24182 Draw time and frequency axes and legends. Default is disabled.
24185 The usage is very similar to the showwaves filter; see the examples in that
24188 @subsection Examples
24192 Large window with logarithmic color scaling:
24194 showspectrum=s=1280x480:scale=log
24198 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
24200 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24201 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
24205 @section showspectrumpic
24207 Convert input audio to a single video frame, representing the audio frequency
24210 The filter accepts the following options:
24214 Specify the video size for the output. For the syntax of this option, check the
24215 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24216 Default value is @code{4096x2048}.
24219 Specify display mode.
24221 It accepts the following values:
24224 all channels are displayed in the same row
24226 all channels are displayed in separate rows
24228 Default value is @samp{combined}.
24231 Specify display color mode.
24233 It accepts the following values:
24236 each channel is displayed in a separate color
24238 each channel is displayed using the same color scheme
24240 each channel is displayed using the rainbow color scheme
24242 each channel is displayed using the moreland color scheme
24244 each channel is displayed using the nebulae color scheme
24246 each channel is displayed using the fire color scheme
24248 each channel is displayed using the fiery color scheme
24250 each channel is displayed using the fruit color scheme
24252 each channel is displayed using the cool color scheme
24254 each channel is displayed using the magma color scheme
24256 each channel is displayed using the green color scheme
24258 each channel is displayed using the viridis color scheme
24260 each channel is displayed using the plasma color scheme
24262 each channel is displayed using the cividis color scheme
24264 each channel is displayed using the terrain color scheme
24266 Default value is @samp{intensity}.
24269 Specify scale used for calculating intensity color values.
24271 It accepts the following values:
24276 square root, default
24286 Default value is @samp{log}.
24289 Specify frequency scale.
24291 It accepts the following values:
24299 Default value is @samp{lin}.
24302 Set saturation modifier for displayed colors. Negative values provide
24303 alternative color scheme. @code{0} is no saturation at all.
24304 Saturation must be in [-10.0, 10.0] range.
24305 Default value is @code{1}.
24308 Set window function.
24310 It accepts the following values:
24334 Default value is @code{hann}.
24337 Set orientation of time vs frequency axis. Can be @code{vertical} or
24338 @code{horizontal}. Default is @code{vertical}.
24341 Set scale gain for calculating intensity color values.
24342 Default value is @code{1}.
24345 Draw time and frequency axes and legends. Default is enabled.
24348 Set color rotation, must be in [-1.0, 1.0] range.
24349 Default value is @code{0}.
24352 Set start frequency from which to display spectrogram. Default is @code{0}.
24355 Set stop frequency to which to display spectrogram. Default is @code{0}.
24358 @subsection Examples
24362 Extract an audio spectrogram of a whole audio track
24363 in a 1024x1024 picture using @command{ffmpeg}:
24365 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
24369 @section showvolume
24371 Convert input audio volume to a video output.
24373 The filter accepts the following options:
24380 Set border width, allowed range is [0, 5]. Default is 1.
24383 Set channel width, allowed range is [80, 8192]. Default is 400.
24386 Set channel height, allowed range is [1, 900]. Default is 20.
24389 Set fade, allowed range is [0, 1]. Default is 0.95.
24392 Set volume color expression.
24394 The expression can use the following variables:
24398 Current max volume of channel in dB.
24404 Current channel number, starting from 0.
24408 If set, displays channel names. Default is enabled.
24411 If set, displays volume values. Default is enabled.
24414 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
24415 default is @code{h}.
24418 Set step size, allowed range is [0, 5]. Default is 0, which means
24422 Set background opacity, allowed range is [0, 1]. Default is 0.
24425 Set metering mode, can be peak: @code{p} or rms: @code{r},
24426 default is @code{p}.
24429 Set display scale, can be linear: @code{lin} or log: @code{log},
24430 default is @code{lin}.
24434 If set to > 0., display a line for the max level
24435 in the previous seconds.
24436 default is disabled: @code{0.}
24439 The color of the max line. Use when @code{dm} option is set to > 0.
24440 default is: @code{orange}
24445 Convert input audio to a video output, representing the samples waves.
24447 The filter accepts the following options:
24451 Specify the video size for the output. For the syntax of this option, check the
24452 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24453 Default value is @code{600x240}.
24458 Available values are:
24461 Draw a point for each sample.
24464 Draw a vertical line for each sample.
24467 Draw a point for each sample and a line between them.
24470 Draw a centered vertical line for each sample.
24473 Default value is @code{point}.
24476 Set the number of samples which are printed on the same column. A
24477 larger value will decrease the frame rate. Must be a positive
24478 integer. This option can be set only if the value for @var{rate}
24479 is not explicitly specified.
24482 Set the (approximate) output frame rate. This is done by setting the
24483 option @var{n}. Default value is "25".
24485 @item split_channels
24486 Set if channels should be drawn separately or overlap. Default value is 0.
24489 Set colors separated by '|' which are going to be used for drawing of each channel.
24492 Set amplitude scale.
24494 Available values are:
24512 Set the draw mode. This is mostly useful to set for high @var{n}.
24514 Available values are:
24517 Scale pixel values for each drawn sample.
24520 Draw every sample directly.
24523 Default value is @code{scale}.
24526 @subsection Examples
24530 Output the input file audio and the corresponding video representation
24533 amovie=a.mp3,asplit[out0],showwaves[out1]
24537 Create a synthetic signal and show it with showwaves, forcing a
24538 frame rate of 30 frames per second:
24540 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
24544 @section showwavespic
24546 Convert input audio to a single video frame, representing the samples waves.
24548 The filter accepts the following options:
24552 Specify the video size for the output. For the syntax of this option, check the
24553 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24554 Default value is @code{600x240}.
24556 @item split_channels
24557 Set if channels should be drawn separately or overlap. Default value is 0.
24560 Set colors separated by '|' which are going to be used for drawing of each channel.
24563 Set amplitude scale.
24565 Available values are:
24585 Available values are:
24588 Scale pixel values for each drawn sample.
24591 Draw every sample directly.
24594 Default value is @code{scale}.
24597 @subsection Examples
24601 Extract a channel split representation of the wave form of a whole audio track
24602 in a 1024x800 picture using @command{ffmpeg}:
24604 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
24608 @section sidedata, asidedata
24610 Delete frame side data, or select frames based on it.
24612 This filter accepts the following options:
24616 Set mode of operation of the filter.
24618 Can be one of the following:
24622 Select every frame with side data of @code{type}.
24625 Delete side data of @code{type}. If @code{type} is not set, delete all side
24631 Set side data type used with all modes. Must be set for @code{select} mode. For
24632 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
24633 in @file{libavutil/frame.h}. For example, to choose
24634 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
24638 @section spectrumsynth
24640 Synthesize audio from 2 input video spectrums, first input stream represents
24641 magnitude across time and second represents phase across time.
24642 The filter will transform from frequency domain as displayed in videos back
24643 to time domain as presented in audio output.
24645 This filter is primarily created for reversing processed @ref{showspectrum}
24646 filter outputs, but can synthesize sound from other spectrograms too.
24647 But in such case results are going to be poor if the phase data is not
24648 available, because in such cases phase data need to be recreated, usually
24649 it's just recreated from random noise.
24650 For best results use gray only output (@code{channel} color mode in
24651 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
24652 @code{lin} scale for phase video. To produce phase, for 2nd video, use
24653 @code{data} option. Inputs videos should generally use @code{fullframe}
24654 slide mode as that saves resources needed for decoding video.
24656 The filter accepts the following options:
24660 Specify sample rate of output audio, the sample rate of audio from which
24661 spectrum was generated may differ.
24664 Set number of channels represented in input video spectrums.
24667 Set scale which was used when generating magnitude input spectrum.
24668 Can be @code{lin} or @code{log}. Default is @code{log}.
24671 Set slide which was used when generating inputs spectrums.
24672 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
24673 Default is @code{fullframe}.
24676 Set window function used for resynthesis.
24679 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
24680 which means optimal overlap for selected window function will be picked.
24683 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
24684 Default is @code{vertical}.
24687 @subsection Examples
24691 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
24692 then resynthesize videos back to audio with spectrumsynth:
24694 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
24695 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
24696 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
24700 @section split, asplit
24702 Split input into several identical outputs.
24704 @code{asplit} works with audio input, @code{split} with video.
24706 The filter accepts a single parameter which specifies the number of outputs. If
24707 unspecified, it defaults to 2.
24709 @subsection Examples
24713 Create two separate outputs from the same input:
24715 [in] split [out0][out1]
24719 To create 3 or more outputs, you need to specify the number of
24722 [in] asplit=3 [out0][out1][out2]
24726 Create two separate outputs from the same input, one cropped and
24729 [in] split [splitout1][splitout2];
24730 [splitout1] crop=100:100:0:0 [cropout];
24731 [splitout2] pad=200:200:100:100 [padout];
24735 Create 5 copies of the input audio with @command{ffmpeg}:
24737 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
24743 Receive commands sent through a libzmq client, and forward them to
24744 filters in the filtergraph.
24746 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
24747 must be inserted between two video filters, @code{azmq} between two
24748 audio filters. Both are capable to send messages to any filter type.
24750 To enable these filters you need to install the libzmq library and
24751 headers and configure FFmpeg with @code{--enable-libzmq}.
24753 For more information about libzmq see:
24754 @url{http://www.zeromq.org/}
24756 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
24757 receives messages sent through a network interface defined by the
24758 @option{bind_address} (or the abbreviation "@option{b}") option.
24759 Default value of this option is @file{tcp://localhost:5555}. You may
24760 want to alter this value to your needs, but do not forget to escape any
24761 ':' signs (see @ref{filtergraph escaping}).
24763 The received message must be in the form:
24765 @var{TARGET} @var{COMMAND} [@var{ARG}]
24768 @var{TARGET} specifies the target of the command, usually the name of
24769 the filter class or a specific filter instance name. The default
24770 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
24771 but you can override this by using the @samp{filter_name@@id} syntax
24772 (see @ref{Filtergraph syntax}).
24774 @var{COMMAND} specifies the name of the command for the target filter.
24776 @var{ARG} is optional and specifies the optional argument list for the
24777 given @var{COMMAND}.
24779 Upon reception, the message is processed and the corresponding command
24780 is injected into the filtergraph. Depending on the result, the filter
24781 will send a reply to the client, adopting the format:
24783 @var{ERROR_CODE} @var{ERROR_REASON}
24787 @var{MESSAGE} is optional.
24789 @subsection Examples
24791 Look at @file{tools/zmqsend} for an example of a zmq client which can
24792 be used to send commands processed by these filters.
24794 Consider the following filtergraph generated by @command{ffplay}.
24795 In this example the last overlay filter has an instance name. All other
24796 filters will have default instance names.
24799 ffplay -dumpgraph 1 -f lavfi "
24800 color=s=100x100:c=red [l];
24801 color=s=100x100:c=blue [r];
24802 nullsrc=s=200x100, zmq [bg];
24803 [bg][l] overlay [bg+l];
24804 [bg+l][r] overlay@@my=x=100 "
24807 To change the color of the left side of the video, the following
24808 command can be used:
24810 echo Parsed_color_0 c yellow | tools/zmqsend
24813 To change the right side:
24815 echo Parsed_color_1 c pink | tools/zmqsend
24818 To change the position of the right side:
24820 echo overlay@@my x 150 | tools/zmqsend
24824 @c man end MULTIMEDIA FILTERS
24826 @chapter Multimedia Sources
24827 @c man begin MULTIMEDIA SOURCES
24829 Below is a description of the currently available multimedia sources.
24833 This is the same as @ref{movie} source, except it selects an audio
24839 Read audio and/or video stream(s) from a movie container.
24841 It accepts the following parameters:
24845 The name of the resource to read (not necessarily a file; it can also be a
24846 device or a stream accessed through some protocol).
24848 @item format_name, f
24849 Specifies the format assumed for the movie to read, and can be either
24850 the name of a container or an input device. If not specified, the
24851 format is guessed from @var{movie_name} or by probing.
24853 @item seek_point, sp
24854 Specifies the seek point in seconds. The frames will be output
24855 starting from this seek point. The parameter is evaluated with
24856 @code{av_strtod}, so the numerical value may be suffixed by an IS
24857 postfix. The default value is "0".
24860 Specifies the streams to read. Several streams can be specified,
24861 separated by "+". The source will then have as many outputs, in the
24862 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
24863 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
24864 respectively the default (best suited) video and audio stream. Default
24865 is "dv", or "da" if the filter is called as "amovie".
24867 @item stream_index, si
24868 Specifies the index of the video stream to read. If the value is -1,
24869 the most suitable video stream will be automatically selected. The default
24870 value is "-1". Deprecated. If the filter is called "amovie", it will select
24871 audio instead of video.
24874 Specifies how many times to read the stream in sequence.
24875 If the value is 0, the stream will be looped infinitely.
24876 Default value is "1".
24878 Note that when the movie is looped the source timestamps are not
24879 changed, so it will generate non monotonically increasing timestamps.
24881 @item discontinuity
24882 Specifies the time difference between frames above which the point is
24883 considered a timestamp discontinuity which is removed by adjusting the later
24887 It allows overlaying a second video on top of the main input of
24888 a filtergraph, as shown in this graph:
24890 input -----------> deltapts0 --> overlay --> output
24893 movie --> scale--> deltapts1 -------+
24895 @subsection Examples
24899 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
24900 on top of the input labelled "in":
24902 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
24903 [in] setpts=PTS-STARTPTS [main];
24904 [main][over] overlay=16:16 [out]
24908 Read from a video4linux2 device, and overlay it on top of the input
24911 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
24912 [in] setpts=PTS-STARTPTS [main];
24913 [main][over] overlay=16:16 [out]
24917 Read the first video stream and the audio stream with id 0x81 from
24918 dvd.vob; the video is connected to the pad named "video" and the audio is
24919 connected to the pad named "audio":
24921 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
24925 @subsection Commands
24927 Both movie and amovie support the following commands:
24930 Perform seek using "av_seek_frame".
24931 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
24934 @var{stream_index}: If stream_index is -1, a default
24935 stream is selected, and @var{timestamp} is automatically converted
24936 from AV_TIME_BASE units to the stream specific time_base.
24938 @var{timestamp}: Timestamp in AVStream.time_base units
24939 or, if no stream is specified, in AV_TIME_BASE units.
24941 @var{flags}: Flags which select direction and seeking mode.
24945 Get movie duration in AV_TIME_BASE units.
24949 @c man end MULTIMEDIA SOURCES