1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Examples
1093 Fade in first 15 seconds of audio:
1095 afade=t=in:ss=0:d=15
1099 Fade out last 25 seconds of a 900 seconds audio:
1101 afade=t=out:st=875:d=25
1106 Denoise audio samples with FFT.
1108 A description of the accepted parameters follows.
1112 Set the noise reduction in dB, allowed range is 0.01 to 97.
1113 Default value is 12 dB.
1116 Set the noise floor in dB, allowed range is -80 to -20.
1117 Default value is -50 dB.
1122 It accepts the following values:
1131 Select shellac noise.
1134 Select custom noise, defined in @code{bn} option.
1136 Default value is white noise.
1140 Set custom band noise for every one of 15 bands.
1141 Bands are separated by ' ' or '|'.
1144 Set the residual floor in dB, allowed range is -80 to -20.
1145 Default value is -38 dB.
1148 Enable noise tracking. By default is disabled.
1149 With this enabled, noise floor is automatically adjusted.
1152 Enable residual tracking. By default is disabled.
1155 Set the output mode.
1157 It accepts the following values:
1160 Pass input unchanged.
1163 Pass noise filtered out.
1168 Default value is @var{o}.
1172 @subsection Commands
1174 This filter supports the following commands:
1176 @item sample_noise, sn
1177 Start or stop measuring noise profile.
1178 Syntax for the command is : "start" or "stop" string.
1179 After measuring noise profile is stopped it will be
1180 automatically applied in filtering.
1182 @item noise_reduction, nr
1183 Change noise reduction. Argument is single float number.
1184 Syntax for the command is : "@var{noise_reduction}"
1186 @item noise_floor, nf
1187 Change noise floor. Argument is single float number.
1188 Syntax for the command is : "@var{noise_floor}"
1190 @item output_mode, om
1191 Change output mode operation.
1192 Syntax for the command is : "i", "o" or "n" string.
1196 Apply arbitrary expressions to samples in frequency domain.
1200 Set frequency domain real expression for each separate channel separated
1201 by '|'. Default is "re".
1202 If the number of input channels is greater than the number of
1203 expressions, the last specified expression is used for the remaining
1207 Set frequency domain imaginary expression for each separate channel
1208 separated by '|'. Default is "im".
1210 Each expression in @var{real} and @var{imag} can contain the following
1211 constants and functions:
1218 current frequency bin number
1221 number of available bins
1224 channel number of the current expression
1233 current real part of frequency bin of current channel
1236 current imaginary part of frequency bin of current channel
1239 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1242 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1246 Set window size. Allowed range is from 16 to 131072.
1247 Default is @code{4096}
1250 Set window function. Default is @code{hann}.
1253 Set window overlap. If set to 1, the recommended overlap for selected
1254 window function will be picked. Default is @code{0.75}.
1257 @subsection Examples
1261 Leave almost only low frequencies in audio:
1263 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1267 Apply robotize effect:
1269 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1273 Apply whisper effect:
1275 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"
1282 Apply an arbitrary Finite Impulse Response filter.
1284 This filter is designed for applying long FIR filters,
1285 up to 60 seconds long.
1287 It can be used as component for digital crossover filters,
1288 room equalization, cross talk cancellation, wavefield synthesis,
1289 auralization, ambiophonics, ambisonics and spatialization.
1291 This filter uses the streams higher than first one as FIR coefficients.
1292 If the non-first stream holds a single channel, it will be used
1293 for all input channels in the first stream, otherwise
1294 the number of channels in the non-first stream must be same as
1295 the number of channels in the first stream.
1297 It accepts the following parameters:
1301 Set dry gain. This sets input gain.
1304 Set wet gain. This sets final output gain.
1307 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1310 Enable applying gain measured from power of IR.
1312 Set which approach to use for auto gain measurement.
1316 Do not apply any gain.
1319 select peak gain, very conservative approach. This is default value.
1322 select DC gain, limited application.
1325 select gain to noise approach, this is most popular one.
1329 Set gain to be applied to IR coefficients before filtering.
1330 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1333 Set format of IR stream. Can be @code{mono} or @code{input}.
1334 Default is @code{input}.
1337 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1338 Allowed range is 0.1 to 60 seconds.
1341 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1342 By default it is disabled.
1345 Set for which IR channel to display frequency response. By default is first channel
1346 displayed. This option is used only when @var{response} is enabled.
1349 Set video stream size. This option is used only when @var{response} is enabled.
1352 Set video stream frame rate. This option is used only when @var{response} is enabled.
1355 Set minimal partition size used for convolution. Default is @var{8192}.
1356 Allowed range is from @var{1} to @var{32768}.
1357 Lower values decreases latency at cost of higher CPU usage.
1360 Set maximal partition size used for convolution. Default is @var{8192}.
1361 Allowed range is from @var{8} to @var{32768}.
1362 Lower values may increase CPU usage.
1365 Set number of input impulse responses streams which will be switchable at runtime.
1366 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1369 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1370 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1371 This option can be changed at runtime via @ref{commands}.
1374 @subsection Examples
1378 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1380 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1387 Set output format constraints for the input audio. The framework will
1388 negotiate the most appropriate format to minimize conversions.
1390 It accepts the following parameters:
1393 @item sample_fmts, f
1394 A '|'-separated list of requested sample formats.
1396 @item sample_rates, r
1397 A '|'-separated list of requested sample rates.
1399 @item channel_layouts, cl
1400 A '|'-separated list of requested channel layouts.
1402 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1403 for the required syntax.
1406 If a parameter is omitted, all values are allowed.
1408 Force the output to either unsigned 8-bit or signed 16-bit stereo
1410 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1414 Apply frequency shift to input audio samples.
1416 The filter accepts the following options:
1420 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1421 Default value is 0.0.
1424 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1425 Default value is 1.0.
1428 @subsection Commands
1430 This filter supports the all above options as @ref{commands}.
1434 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1435 processing reduces disturbing noise between useful signals.
1437 Gating is done by detecting the volume below a chosen level @var{threshold}
1438 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1439 floor is set via @var{range}. Because an exact manipulation of the signal
1440 would cause distortion of the waveform the reduction can be levelled over
1441 time. This is done by setting @var{attack} and @var{release}.
1443 @var{attack} determines how long the signal has to fall below the threshold
1444 before any reduction will occur and @var{release} sets the time the signal
1445 has to rise above the threshold to reduce the reduction again.
1446 Shorter signals than the chosen attack time will be left untouched.
1450 Set input level before filtering.
1451 Default is 1. Allowed range is from 0.015625 to 64.
1454 Set the mode of operation. Can be @code{upward} or @code{downward}.
1455 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1456 will be amplified, expanding dynamic range in upward direction.
1457 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1460 Set the level of gain reduction when the signal is below the threshold.
1461 Default is 0.06125. Allowed range is from 0 to 1.
1462 Setting this to 0 disables reduction and then filter behaves like expander.
1465 If a signal rises above this level the gain reduction is released.
1466 Default is 0.125. Allowed range is from 0 to 1.
1469 Set a ratio by which the signal is reduced.
1470 Default is 2. Allowed range is from 1 to 9000.
1473 Amount of milliseconds the signal has to rise above the threshold before gain
1475 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1478 Amount of milliseconds the signal has to fall below the threshold before the
1479 reduction is increased again. Default is 250 milliseconds.
1480 Allowed range is from 0.01 to 9000.
1483 Set amount of amplification of signal after processing.
1484 Default is 1. Allowed range is from 1 to 64.
1487 Curve the sharp knee around the threshold to enter gain reduction more softly.
1488 Default is 2.828427125. Allowed range is from 1 to 8.
1491 Choose if exact signal should be taken for detection or an RMS like one.
1492 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1495 Choose if the average level between all channels or the louder channel affects
1497 Default is @code{average}. Can be @code{average} or @code{maximum}.
1500 @subsection Commands
1502 This filter supports the all above options as @ref{commands}.
1506 Apply an arbitrary Infinite Impulse Response filter.
1508 It accepts the following parameters:
1512 Set B/numerator/zeros/reflection coefficients.
1515 Set A/denominator/poles/ladder coefficients.
1527 Set coefficients format.
1531 lattice-ladder function
1533 analog transfer function
1535 digital transfer function
1537 Z-plane zeros/poles, cartesian (default)
1539 Z-plane zeros/poles, polar radians
1541 Z-plane zeros/poles, polar degrees
1547 Set type of processing.
1559 Set filtering precision.
1563 double-precision floating-point (default)
1565 single-precision floating-point
1573 Normalize filter coefficients, by default is enabled.
1574 Enabling it will normalize magnitude response at DC to 0dB.
1577 How much to use filtered signal in output. Default is 1.
1578 Range is between 0 and 1.
1581 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1582 By default it is disabled.
1585 Set for which IR channel to display frequency response. By default is first channel
1586 displayed. This option is used only when @var{response} is enabled.
1589 Set video stream size. This option is used only when @var{response} is enabled.
1592 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1595 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1596 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1599 Different coefficients and gains can be provided for every channel, in such case
1600 use '|' to separate coefficients or gains. Last provided coefficients will be
1601 used for all remaining channels.
1603 @subsection Examples
1607 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1609 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
1613 Same as above but in @code{zp} format:
1615 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
1619 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1621 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1627 The limiter prevents an input signal from rising over a desired threshold.
1628 This limiter uses lookahead technology to prevent your signal from distorting.
1629 It means that there is a small delay after the signal is processed. Keep in mind
1630 that the delay it produces is the attack time you set.
1632 The filter accepts the following options:
1636 Set input gain. Default is 1.
1639 Set output gain. Default is 1.
1642 Don't let signals above this level pass the limiter. Default is 1.
1645 The limiter will reach its attenuation level in this amount of time in
1646 milliseconds. Default is 5 milliseconds.
1649 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1650 Default is 50 milliseconds.
1653 When gain reduction is always needed ASC takes care of releasing to an
1654 average reduction level rather than reaching a reduction of 0 in the release
1658 Select how much the release time is affected by ASC, 0 means nearly no changes
1659 in release time while 1 produces higher release times.
1662 Auto level output signal. Default is enabled.
1663 This normalizes audio back to 0dB if enabled.
1666 Depending on picked setting it is recommended to upsample input 2x or 4x times
1667 with @ref{aresample} before applying this filter.
1671 Apply a two-pole all-pass filter with central frequency (in Hz)
1672 @var{frequency}, and filter-width @var{width}.
1673 An all-pass filter changes the audio's frequency to phase relationship
1674 without changing its frequency to amplitude relationship.
1676 The filter accepts the following options:
1680 Set frequency in Hz.
1683 Set method to specify band-width of filter.
1698 Specify the band-width of a filter in width_type units.
1701 How much to use filtered signal in output. Default is 1.
1702 Range is between 0 and 1.
1705 Specify which channels to filter, by default all available are filtered.
1708 Normalize biquad coefficients, by default is disabled.
1709 Enabling it will normalize magnitude response at DC to 0dB.
1712 Set the filter order, can be 1 or 2. Default is 2.
1715 Set transform type of IIR filter.
1724 Set precison of filtering.
1727 Pick automatic sample format depending on surround filters.
1729 Always use signed 16-bit.
1731 Always use signed 32-bit.
1733 Always use float 32-bit.
1735 Always use float 64-bit.
1739 @subsection Commands
1741 This filter supports the following commands:
1744 Change allpass frequency.
1745 Syntax for the command is : "@var{frequency}"
1748 Change allpass width_type.
1749 Syntax for the command is : "@var{width_type}"
1752 Change allpass width.
1753 Syntax for the command is : "@var{width}"
1757 Syntax for the command is : "@var{mix}"
1764 The filter accepts the following options:
1768 Set the number of loops. Setting this value to -1 will result in infinite loops.
1772 Set maximal number of samples. Default is 0.
1775 Set first sample of loop. Default is 0.
1781 Merge two or more audio streams into a single multi-channel stream.
1783 The filter accepts the following options:
1788 Set the number of inputs. Default is 2.
1792 If the channel layouts of the inputs are disjoint, and therefore compatible,
1793 the channel layout of the output will be set accordingly and the channels
1794 will be reordered as necessary. If the channel layouts of the inputs are not
1795 disjoint, the output will have all the channels of the first input then all
1796 the channels of the second input, in that order, and the channel layout of
1797 the output will be the default value corresponding to the total number of
1800 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1801 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1802 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1803 first input, b1 is the first channel of the second input).
1805 On the other hand, if both input are in stereo, the output channels will be
1806 in the default order: a1, a2, b1, b2, and the channel layout will be
1807 arbitrarily set to 4.0, which may or may not be the expected value.
1809 All inputs must have the same sample rate, and format.
1811 If inputs do not have the same duration, the output will stop with the
1814 @subsection Examples
1818 Merge two mono files into a stereo stream:
1820 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1824 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1826 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
1832 Mixes multiple audio inputs into a single output.
1834 Note that this filter only supports float samples (the @var{amerge}
1835 and @var{pan} audio filters support many formats). If the @var{amix}
1836 input has integer samples then @ref{aresample} will be automatically
1837 inserted to perform the conversion to float samples.
1841 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1843 will mix 3 input audio streams to a single output with the same duration as the
1844 first input and a dropout transition time of 3 seconds.
1846 It accepts the following parameters:
1850 The number of inputs. If unspecified, it defaults to 2.
1853 How to determine the end-of-stream.
1857 The duration of the longest input. (default)
1860 The duration of the shortest input.
1863 The duration of the first input.
1867 @item dropout_transition
1868 The transition time, in seconds, for volume renormalization when an input
1869 stream ends. The default value is 2 seconds.
1872 Specify weight of each input audio stream as sequence.
1873 Each weight is separated by space. By default all inputs have same weight.
1876 @subsection Commands
1878 This filter supports the following commands:
1881 Syntax is same as option with same name.
1886 Multiply first audio stream with second audio stream and store result
1887 in output audio stream. Multiplication is done by multiplying each
1888 sample from first stream with sample at same position from second stream.
1890 With this element-wise multiplication one can create amplitude fades and
1891 amplitude modulations.
1893 @section anequalizer
1895 High-order parametric multiband equalizer for each channel.
1897 It accepts the following parameters:
1901 This option string is in format:
1902 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1903 Each equalizer band is separated by '|'.
1907 Set channel number to which equalization will be applied.
1908 If input doesn't have that channel the entry is ignored.
1911 Set central frequency for band.
1912 If input doesn't have that frequency the entry is ignored.
1915 Set band width in Hertz.
1918 Set band gain in dB.
1921 Set filter type for band, optional, can be:
1925 Butterworth, this is default.
1936 With this option activated frequency response of anequalizer is displayed
1940 Set video stream size. Only useful if curves option is activated.
1943 Set max gain that will be displayed. Only useful if curves option is activated.
1944 Setting this to a reasonable value makes it possible to display gain which is derived from
1945 neighbour bands which are too close to each other and thus produce higher gain
1946 when both are activated.
1949 Set frequency scale used to draw frequency response in video output.
1950 Can be linear or logarithmic. Default is logarithmic.
1953 Set color for each channel curve which is going to be displayed in video stream.
1954 This is list of color names separated by space or by '|'.
1955 Unrecognised or missing colors will be replaced by white color.
1958 @subsection Examples
1962 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1963 for first 2 channels using Chebyshev type 1 filter:
1965 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1969 @subsection Commands
1971 This filter supports the following commands:
1974 Alter existing filter parameters.
1975 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1977 @var{fN} is existing filter number, starting from 0, if no such filter is available
1979 @var{freq} set new frequency parameter.
1980 @var{width} set new width parameter in Hertz.
1981 @var{gain} set new gain parameter in dB.
1983 Full filter invocation with asendcmd may look like this:
1984 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1989 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1991 Each sample is adjusted by looking for other samples with similar contexts. This
1992 context similarity is defined by comparing their surrounding patches of size
1993 @option{p}. Patches are searched in an area of @option{r} around the sample.
1995 The filter accepts the following options:
1999 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2002 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2003 Default value is 2 milliseconds.
2006 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2007 Default value is 6 milliseconds.
2010 Set the output mode.
2012 It accepts the following values:
2015 Pass input unchanged.
2018 Pass noise filtered out.
2023 Default value is @var{o}.
2027 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2030 @subsection Commands
2032 This filter supports the all above options as @ref{commands}.
2035 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2037 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2038 relate to producing the least mean square of the error signal (difference between the desired,
2039 2nd input audio stream and the actual signal, the 1st input audio stream).
2041 A description of the accepted options follows.
2054 Set the filter leakage.
2057 It accepts the following values:
2066 Pass filtered samples.
2069 Pass difference between desired and filtered samples.
2071 Default value is @var{o}.
2075 @subsection Examples
2079 One of many usages of this filter is noise reduction, input audio is filtered
2080 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2082 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2086 @subsection Commands
2088 This filter supports the same commands as options, excluding option @code{order}.
2092 Pass the audio source unchanged to the output.
2096 Pad the end of an audio stream with silence.
2098 This can be used together with @command{ffmpeg} @option{-shortest} to
2099 extend audio streams to the same length as the video stream.
2101 A description of the accepted options follows.
2105 Set silence packet size. Default value is 4096.
2108 Set the number of samples of silence to add to the end. After the
2109 value is reached, the stream is terminated. This option is mutually
2110 exclusive with @option{whole_len}.
2113 Set the minimum total number of samples in the output audio stream. If
2114 the value is longer than the input audio length, silence is added to
2115 the end, until the value is reached. This option is mutually exclusive
2116 with @option{pad_len}.
2119 Specify the duration of samples of silence to add. See
2120 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2121 for the accepted syntax. Used only if set to non-zero value.
2124 Specify the minimum total duration in the output audio stream. See
2125 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2126 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2127 the input audio length, silence is added to the end, until the value is reached.
2128 This option is mutually exclusive with @option{pad_dur}
2131 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2132 nor @option{whole_dur} option is set, the filter will add silence to the end of
2133 the input stream indefinitely.
2135 @subsection Examples
2139 Add 1024 samples of silence to the end of the input:
2145 Make sure the audio output will contain at least 10000 samples, pad
2146 the input with silence if required:
2148 apad=whole_len=10000
2152 Use @command{ffmpeg} to pad the audio input with silence, so that the
2153 video stream will always result the shortest and will be converted
2154 until the end in the output file when using the @option{shortest}
2157 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2162 Add a phasing effect to the input audio.
2164 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2165 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2167 A description of the accepted parameters follows.
2171 Set input gain. Default is 0.4.
2174 Set output gain. Default is 0.74
2177 Set delay in milliseconds. Default is 3.0.
2180 Set decay. Default is 0.4.
2183 Set modulation speed in Hz. Default is 0.5.
2186 Set modulation type. Default is triangular.
2188 It accepts the following values:
2195 @section aphaseshift
2196 Apply phase shift to input audio samples.
2198 The filter accepts the following options:
2202 Specify phase shift. Allowed range is from -1.0 to 1.0.
2203 Default value is 0.0.
2206 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2207 Default value is 1.0.
2210 @subsection Commands
2212 This filter supports the all above options as @ref{commands}.
2216 Audio pulsator is something between an autopanner and a tremolo.
2217 But it can produce funny stereo effects as well. Pulsator changes the volume
2218 of the left and right channel based on a LFO (low frequency oscillator) with
2219 different waveforms and shifted phases.
2220 This filter have the ability to define an offset between left and right
2221 channel. An offset of 0 means that both LFO shapes match each other.
2222 The left and right channel are altered equally - a conventional tremolo.
2223 An offset of 50% means that the shape of the right channel is exactly shifted
2224 in phase (or moved backwards about half of the frequency) - pulsator acts as
2225 an autopanner. At 1 both curves match again. Every setting in between moves the
2226 phase shift gapless between all stages and produces some "bypassing" sounds with
2227 sine and triangle waveforms. The more you set the offset near 1 (starting from
2228 the 0.5) the faster the signal passes from the left to the right speaker.
2230 The filter accepts the following options:
2234 Set input gain. By default it is 1. Range is [0.015625 - 64].
2237 Set output gain. By default it is 1. Range is [0.015625 - 64].
2240 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2241 sawup or sawdown. Default is sine.
2244 Set modulation. Define how much of original signal is affected by the LFO.
2247 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2250 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2253 Set pulse width. Default is 1. Allowed range is [0 - 2].
2256 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2259 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2263 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2267 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2268 if timing is set to hz.
2274 Resample the input audio to the specified parameters, using the
2275 libswresample library. If none are specified then the filter will
2276 automatically convert between its input and output.
2278 This filter is also able to stretch/squeeze the audio data to make it match
2279 the timestamps or to inject silence / cut out audio to make it match the
2280 timestamps, do a combination of both or do neither.
2282 The filter accepts the syntax
2283 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2284 expresses a sample rate and @var{resampler_options} is a list of
2285 @var{key}=@var{value} pairs, separated by ":". See the
2286 @ref{Resampler Options,,"Resampler Options" section in the
2287 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2288 for the complete list of supported options.
2290 @subsection Examples
2294 Resample the input audio to 44100Hz:
2300 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2301 samples per second compensation:
2303 aresample=async=1000
2309 Reverse an audio clip.
2311 Warning: This filter requires memory to buffer the entire clip, so trimming
2314 @subsection Examples
2318 Take the first 5 seconds of a clip, and reverse it.
2320 atrim=end=5,areverse
2326 Reduce noise from speech using Recurrent Neural Networks.
2328 This filter accepts the following options:
2332 Set train model file to load. This option is always required.
2335 Set how much to mix filtered samples into final output.
2336 Allowed range is from -1 to 1. Default value is 1.
2337 Negative values are special, they set how much to keep filtered noise
2338 in the final filter output. Set this option to -1 to hear actual
2339 noise removed from input signal.
2342 @section asetnsamples
2344 Set the number of samples per each output audio frame.
2346 The last output packet may contain a different number of samples, as
2347 the filter will flush all the remaining samples when the input audio
2350 The filter accepts the following options:
2354 @item nb_out_samples, n
2355 Set the number of frames per each output audio frame. The number is
2356 intended as the number of samples @emph{per each channel}.
2357 Default value is 1024.
2360 If set to 1, the filter will pad the last audio frame with zeroes, so
2361 that the last frame will contain the same number of samples as the
2362 previous ones. Default value is 1.
2365 For example, to set the number of per-frame samples to 1234 and
2366 disable padding for the last frame, use:
2368 asetnsamples=n=1234:p=0
2373 Set the sample rate without altering the PCM data.
2374 This will result in a change of speed and pitch.
2376 The filter accepts the following options:
2379 @item sample_rate, r
2380 Set the output sample rate. Default is 44100 Hz.
2385 Show a line containing various information for each input audio frame.
2386 The input audio is not modified.
2388 The shown line contains a sequence of key/value pairs of the form
2389 @var{key}:@var{value}.
2391 The following values are shown in the output:
2395 The (sequential) number of the input frame, starting from 0.
2398 The presentation timestamp of the input frame, in time base units; the time base
2399 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2402 The presentation timestamp of the input frame in seconds.
2405 position of the frame in the input stream, -1 if this information in
2406 unavailable and/or meaningless (for example in case of synthetic audio)
2415 The sample rate for the audio frame.
2418 The number of samples (per channel) in the frame.
2421 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2422 audio, the data is treated as if all the planes were concatenated.
2424 @item plane_checksums
2425 A list of Adler-32 checksums for each data plane.
2429 Apply audio soft clipping.
2431 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2432 along a smooth curve, rather than the abrupt shape of hard-clipping.
2434 This filter accepts the following options:
2438 Set type of soft-clipping.
2440 It accepts the following values:
2454 Set additional parameter which controls sigmoid function.
2457 Set oversampling factor.
2460 @subsection Commands
2462 This filter supports the all above options as @ref{commands}.
2465 Automatic Speech Recognition
2467 This filter uses PocketSphinx for speech recognition. To enable
2468 compilation of this filter, you need to configure FFmpeg with
2469 @code{--enable-pocketsphinx}.
2471 It accepts the following options:
2475 Set sampling rate of input audio. Defaults is @code{16000}.
2476 This need to match speech models, otherwise one will get poor results.
2479 Set dictionary containing acoustic model files.
2482 Set pronunciation dictionary.
2485 Set language model file.
2488 Set language model set.
2491 Set which language model to use.
2494 Set output for log messages.
2497 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2502 Display time domain statistical information about the audio channels.
2503 Statistics are calculated and displayed for each audio channel and,
2504 where applicable, an overall figure is also given.
2506 It accepts the following option:
2509 Short window length in seconds, used for peak and trough RMS measurement.
2510 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2514 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2515 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2518 Available keys for each channel are:
2564 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2565 this @code{lavfi.astats.Overall.Peak_count}.
2567 For description what each key means read below.
2570 Set number of frame after which stats are going to be recalculated.
2571 Default is disabled.
2573 @item measure_perchannel
2574 Select the entries which need to be measured per channel. The metadata keys can
2575 be used as flags, default is @option{all} which measures everything.
2576 @option{none} disables all per channel measurement.
2578 @item measure_overall
2579 Select the entries which need to be measured overall. The metadata keys can
2580 be used as flags, default is @option{all} which measures everything.
2581 @option{none} disables all overall measurement.
2585 A description of each shown parameter follows:
2589 Mean amplitude displacement from zero.
2592 Minimal sample level.
2595 Maximal sample level.
2597 @item Min difference
2598 Minimal difference between two consecutive samples.
2600 @item Max difference
2601 Maximal difference between two consecutive samples.
2603 @item Mean difference
2604 Mean difference between two consecutive samples.
2605 The average of each difference between two consecutive samples.
2607 @item RMS difference
2608 Root Mean Square difference between two consecutive samples.
2612 Standard peak and RMS level measured in dBFS.
2616 Peak and trough values for RMS level measured over a short window.
2619 Standard ratio of peak to RMS level (note: not in dB).
2622 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2623 (i.e. either @var{Min level} or @var{Max level}).
2626 Number of occasions (not the number of samples) that the signal attained either
2627 @var{Min level} or @var{Max level}.
2629 @item Noise floor dB
2630 Minimum local peak measured in dBFS over a short window.
2632 @item Noise floor count
2633 Number of occasions (not the number of samples) that the signal attained
2637 Overall bit depth of audio. Number of bits used for each sample.
2640 Measured dynamic range of audio in dB.
2642 @item Zero crossings
2643 Number of points where the waveform crosses the zero level axis.
2645 @item Zero crossings rate
2646 Rate of Zero crossings and number of audio samples.
2650 Boost subwoofer frequencies.
2652 The filter accepts the following options:
2656 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2657 Default value is 0.7.
2660 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2661 Default value is 0.7.
2664 Set delay line decay gain value. Allowed range is from 0 to 1.
2665 Default value is 0.7.
2668 Set delay line feedback gain value. Allowed range is from 0 to 1.
2669 Default value is 0.9.
2672 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2673 Default value is 100.
2676 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2677 Default value is 0.5.
2680 Set delay. Allowed range is from 1 to 100.
2681 Default value is 20.
2684 @subsection Commands
2686 This filter supports the all above options as @ref{commands}.
2689 Cut subwoofer frequencies.
2691 This filter allows to set custom, steeper
2692 roll off than highpass filter, and thus is able to more attenuate
2693 frequency content in stop-band.
2695 The filter accepts the following options:
2699 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2700 Default value is 20.
2703 Set filter order. Available values are from 3 to 20.
2704 Default value is 10.
2707 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2710 @subsection Commands
2712 This filter supports the all above options as @ref{commands}.
2715 Cut super frequencies.
2717 The filter accepts the following options:
2721 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2722 Default value is 20000.
2725 Set filter order. Available values are from 3 to 20.
2726 Default value is 10.
2729 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2732 @subsection Commands
2734 This filter supports the all above options as @ref{commands}.
2737 Apply high order Butterworth band-pass filter.
2739 The filter accepts the following options:
2743 Set center frequency in Hertz. Allowed range is 2 to 999999.
2744 Default value is 1000.
2747 Set filter order. Available values are from 4 to 20.
2751 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2754 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2757 @subsection Commands
2759 This filter supports the all above options as @ref{commands}.
2762 Apply high order Butterworth band-stop filter.
2764 The filter accepts the following options:
2768 Set center frequency in Hertz. Allowed range is 2 to 999999.
2769 Default value is 1000.
2772 Set filter order. Available values are from 4 to 20.
2776 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2779 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2782 @subsection Commands
2784 This filter supports the all above options as @ref{commands}.
2790 The filter accepts exactly one parameter, the audio tempo. If not
2791 specified then the filter will assume nominal 1.0 tempo. Tempo must
2792 be in the [0.5, 100.0] range.
2794 Note that tempo greater than 2 will skip some samples rather than
2795 blend them in. If for any reason this is a concern it is always
2796 possible to daisy-chain several instances of atempo to achieve the
2797 desired product tempo.
2799 @subsection Examples
2803 Slow down audio to 80% tempo:
2809 To speed up audio to 300% tempo:
2815 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2817 atempo=sqrt(3),atempo=sqrt(3)
2821 @subsection Commands
2823 This filter supports the following commands:
2826 Change filter tempo scale factor.
2827 Syntax for the command is : "@var{tempo}"
2832 Trim the input so that the output contains one continuous subpart of the input.
2834 It accepts the following parameters:
2837 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2838 sample with the timestamp @var{start} will be the first sample in the output.
2841 Specify time of the first audio sample that will be dropped, i.e. the
2842 audio sample immediately preceding the one with the timestamp @var{end} will be
2843 the last sample in the output.
2846 Same as @var{start}, except this option sets the start timestamp in samples
2850 Same as @var{end}, except this option sets the end timestamp in samples instead
2854 The maximum duration of the output in seconds.
2857 The number of the first sample that should be output.
2860 The number of the first sample that should be dropped.
2863 @option{start}, @option{end}, and @option{duration} are expressed as time
2864 duration specifications; see
2865 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2867 Note that the first two sets of the start/end options and the @option{duration}
2868 option look at the frame timestamp, while the _sample options simply count the
2869 samples that pass through the filter. So start/end_pts and start/end_sample will
2870 give different results when the timestamps are wrong, inexact or do not start at
2871 zero. Also note that this filter does not modify the timestamps. If you wish
2872 to have the output timestamps start at zero, insert the asetpts filter after the
2875 If multiple start or end options are set, this filter tries to be greedy and
2876 keep all samples that match at least one of the specified constraints. To keep
2877 only the part that matches all the constraints at once, chain multiple atrim
2880 The defaults are such that all the input is kept. So it is possible to set e.g.
2881 just the end values to keep everything before the specified time.
2886 Drop everything except the second minute of input:
2888 ffmpeg -i INPUT -af atrim=60:120
2892 Keep only the first 1000 samples:
2894 ffmpeg -i INPUT -af atrim=end_sample=1000
2899 @section axcorrelate
2900 Calculate normalized cross-correlation between two input audio streams.
2902 Resulted samples are always between -1 and 1 inclusive.
2903 If result is 1 it means two input samples are highly correlated in that selected segment.
2904 Result 0 means they are not correlated at all.
2905 If result is -1 it means two input samples are out of phase, which means they cancel each
2908 The filter accepts the following options:
2912 Set size of segment over which cross-correlation is calculated.
2913 Default is 256. Allowed range is from 2 to 131072.
2916 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2917 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2918 are always zero and thus need much less calculations to make.
2919 This is generally not true, but is valid for typical audio streams.
2922 @subsection Examples
2926 Calculate correlation between channels in stereo audio stream:
2928 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2934 Apply a two-pole Butterworth band-pass filter with central
2935 frequency @var{frequency}, and (3dB-point) band-width width.
2936 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2937 instead of the default: constant 0dB peak gain.
2938 The filter roll off at 6dB per octave (20dB per decade).
2940 The filter accepts the following options:
2944 Set the filter's central frequency. Default is @code{3000}.
2947 Constant skirt gain if set to 1. Defaults to 0.
2950 Set method to specify band-width of filter.
2965 Specify the band-width of a filter in width_type units.
2968 How much to use filtered signal in output. Default is 1.
2969 Range is between 0 and 1.
2972 Specify which channels to filter, by default all available are filtered.
2975 Normalize biquad coefficients, by default is disabled.
2976 Enabling it will normalize magnitude response at DC to 0dB.
2979 Set transform type of IIR filter.
2988 Set precison of filtering.
2991 Pick automatic sample format depending on surround filters.
2993 Always use signed 16-bit.
2995 Always use signed 32-bit.
2997 Always use float 32-bit.
2999 Always use float 64-bit.
3003 @subsection Commands
3005 This filter supports the following commands:
3008 Change bandpass frequency.
3009 Syntax for the command is : "@var{frequency}"
3012 Change bandpass width_type.
3013 Syntax for the command is : "@var{width_type}"
3016 Change bandpass width.
3017 Syntax for the command is : "@var{width}"
3020 Change bandpass mix.
3021 Syntax for the command is : "@var{mix}"
3026 Apply a two-pole Butterworth band-reject filter with central
3027 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
3028 The filter roll off at 6dB per octave (20dB per decade).
3030 The filter accepts the following options:
3034 Set the filter's central frequency. Default is @code{3000}.
3037 Set method to specify band-width of filter.
3052 Specify the band-width of a filter in width_type units.
3055 How much to use filtered signal in output. Default is 1.
3056 Range is between 0 and 1.
3059 Specify which channels to filter, by default all available are filtered.
3062 Normalize biquad coefficients, by default is disabled.
3063 Enabling it will normalize magnitude response at DC to 0dB.
3066 Set transform type of IIR filter.
3075 Set precison of filtering.
3078 Pick automatic sample format depending on surround filters.
3080 Always use signed 16-bit.
3082 Always use signed 32-bit.
3084 Always use float 32-bit.
3086 Always use float 64-bit.
3090 @subsection Commands
3092 This filter supports the following commands:
3095 Change bandreject frequency.
3096 Syntax for the command is : "@var{frequency}"
3099 Change bandreject width_type.
3100 Syntax for the command is : "@var{width_type}"
3103 Change bandreject width.
3104 Syntax for the command is : "@var{width}"
3107 Change bandreject mix.
3108 Syntax for the command is : "@var{mix}"
3111 @section bass, lowshelf
3113 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3114 shelving filter with a response similar to that of a standard
3115 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3117 The filter accepts the following options:
3121 Give the gain at 0 Hz. Its useful range is about -20
3122 (for a large cut) to +20 (for a large boost).
3123 Beware of clipping when using a positive gain.
3126 Set the filter's central frequency and so can be used
3127 to extend or reduce the frequency range to be boosted or cut.
3128 The default value is @code{100} Hz.
3131 Set method to specify band-width of filter.
3146 Determine how steep is the filter's shelf transition.
3149 How much to use filtered signal in output. Default is 1.
3150 Range is between 0 and 1.
3153 Specify which channels to filter, by default all available are filtered.
3156 Normalize biquad coefficients, by default is disabled.
3157 Enabling it will normalize magnitude response at DC to 0dB.
3160 Set transform type of IIR filter.
3169 Set precison of filtering.
3172 Pick automatic sample format depending on surround filters.
3174 Always use signed 16-bit.
3176 Always use signed 32-bit.
3178 Always use float 32-bit.
3180 Always use float 64-bit.
3184 @subsection Commands
3186 This filter supports the following commands:
3189 Change bass frequency.
3190 Syntax for the command is : "@var{frequency}"
3193 Change bass width_type.
3194 Syntax for the command is : "@var{width_type}"
3198 Syntax for the command is : "@var{width}"
3202 Syntax for the command is : "@var{gain}"
3206 Syntax for the command is : "@var{mix}"
3211 Apply a biquad IIR filter with the given coefficients.
3212 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3213 are the numerator and denominator coefficients respectively.
3214 and @var{channels}, @var{c} specify which channels to filter, by default all
3215 available are filtered.
3217 @subsection Commands
3219 This filter supports the following commands:
3227 Change biquad parameter.
3228 Syntax for the command is : "@var{value}"
3231 How much to use filtered signal in output. Default is 1.
3232 Range is between 0 and 1.
3235 Specify which channels to filter, by default all available are filtered.
3238 Normalize biquad coefficients, by default is disabled.
3239 Enabling it will normalize magnitude response at DC to 0dB.
3242 Set transform type of IIR filter.
3251 Set precison of filtering.
3254 Pick automatic sample format depending on surround filters.
3256 Always use signed 16-bit.
3258 Always use signed 32-bit.
3260 Always use float 32-bit.
3262 Always use float 64-bit.
3267 Bauer stereo to binaural transformation, which improves headphone listening of
3268 stereo audio records.
3270 To enable compilation of this filter you need to configure FFmpeg with
3271 @code{--enable-libbs2b}.
3273 It accepts the following parameters:
3277 Pre-defined crossfeed level.
3281 Default level (fcut=700, feed=50).
3284 Chu Moy circuit (fcut=700, feed=60).
3287 Jan Meier circuit (fcut=650, feed=95).
3292 Cut frequency (in Hz).
3301 Remap input channels to new locations.
3303 It accepts the following parameters:
3306 Map channels from input to output. The argument is a '|'-separated list of
3307 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3308 @var{in_channel} form. @var{in_channel} can be either the name of the input
3309 channel (e.g. FL for front left) or its index in the input channel layout.
3310 @var{out_channel} is the name of the output channel or its index in the output
3311 channel layout. If @var{out_channel} is not given then it is implicitly an
3312 index, starting with zero and increasing by one for each mapping.
3314 @item channel_layout
3315 The channel layout of the output stream.
3318 If no mapping is present, the filter will implicitly map input channels to
3319 output channels, preserving indices.
3321 @subsection Examples
3325 For example, assuming a 5.1+downmix input MOV file,
3327 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3329 will create an output WAV file tagged as stereo from the downmix channels of
3333 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3335 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3339 @section channelsplit
3341 Split each channel from an input audio stream into a separate output stream.
3343 It accepts the following parameters:
3345 @item channel_layout
3346 The channel layout of the input stream. The default is "stereo".
3348 A channel layout describing the channels to be extracted as separate output streams
3349 or "all" to extract each input channel as a separate stream. The default is "all".
3351 Choosing channels not present in channel layout in the input will result in an error.
3354 @subsection Examples
3358 For example, assuming a stereo input MP3 file,
3360 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3362 will create an output Matroska file with two audio streams, one containing only
3363 the left channel and the other the right channel.
3366 Split a 5.1 WAV file into per-channel files:
3368 ffmpeg -i in.wav -filter_complex
3369 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3370 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3371 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3376 Extract only LFE from a 5.1 WAV file:
3378 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3379 -map '[LFE]' lfe.wav
3384 Add a chorus effect to the audio.
3386 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3388 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3389 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3390 The modulation depth defines the range the modulated delay is played before or after
3391 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3392 sound tuned around the original one, like in a chorus where some vocals are slightly
3395 It accepts the following parameters:
3398 Set input gain. Default is 0.4.
3401 Set output gain. Default is 0.4.
3404 Set delays. A typical delay is around 40ms to 60ms.
3416 @subsection Examples
3422 chorus=0.7:0.9:55:0.4:0.25:2
3428 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3432 Fuller sounding chorus with three delays:
3434 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
3439 Compress or expand the audio's dynamic range.
3441 It accepts the following parameters:
3447 A list of times in seconds for each channel over which the instantaneous level
3448 of the input signal is averaged to determine its volume. @var{attacks} refers to
3449 increase of volume and @var{decays} refers to decrease of volume. For most
3450 situations, the attack time (response to the audio getting louder) should be
3451 shorter than the decay time, because the human ear is more sensitive to sudden
3452 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3453 a typical value for decay is 0.8 seconds.
3454 If specified number of attacks & decays is lower than number of channels, the last
3455 set attack/decay will be used for all remaining channels.
3458 A list of points for the transfer function, specified in dB relative to the
3459 maximum possible signal amplitude. Each key points list must be defined using
3460 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3461 @code{x0/y0 x1/y1 x2/y2 ....}
3463 The input values must be in strictly increasing order but the transfer function
3464 does not have to be monotonically rising. The point @code{0/0} is assumed but
3465 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3466 function are @code{-70/-70|-60/-20|1/0}.
3469 Set the curve radius in dB for all joints. It defaults to 0.01.
3472 Set the additional gain in dB to be applied at all points on the transfer
3473 function. This allows for easy adjustment of the overall gain.
3477 Set an initial volume, in dB, to be assumed for each channel when filtering
3478 starts. This permits the user to supply a nominal level initially, so that, for
3479 example, a very large gain is not applied to initial signal levels before the
3480 companding has begun to operate. A typical value for audio which is initially
3481 quiet is -90 dB. It defaults to 0.
3484 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3485 delayed before being fed to the volume adjuster. Specifying a delay
3486 approximately equal to the attack/decay times allows the filter to effectively
3487 operate in predictive rather than reactive mode. It defaults to 0.
3491 @subsection Examples
3495 Make music with both quiet and loud passages suitable for listening to in a
3498 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3501 Another example for audio with whisper and explosion parts:
3503 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3507 A noise gate for when the noise is at a lower level than the signal:
3509 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3513 Here is another noise gate, this time for when the noise is at a higher level
3514 than the signal (making it, in some ways, similar to squelch):
3516 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3520 2:1 compression starting at -6dB:
3522 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3526 2:1 compression starting at -9dB:
3528 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3532 2:1 compression starting at -12dB:
3534 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3538 2:1 compression starting at -18dB:
3540 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3544 3:1 compression starting at -15dB:
3546 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3552 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3558 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
3562 Hard limiter at -6dB:
3564 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3568 Hard limiter at -12dB:
3570 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3574 Hard noise gate at -35 dB:
3576 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3582 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3586 @section compensationdelay
3588 Compensation Delay Line is a metric based delay to compensate differing
3589 positions of microphones or speakers.
3591 For example, you have recorded guitar with two microphones placed in
3592 different locations. Because the front of sound wave has fixed speed in
3593 normal conditions, the phasing of microphones can vary and depends on
3594 their location and interposition. The best sound mix can be achieved when
3595 these microphones are in phase (synchronized). Note that a distance of
3596 ~30 cm between microphones makes one microphone capture the signal in
3597 antiphase to the other microphone. That makes the final mix sound moody.
3598 This filter helps to solve phasing problems by adding different delays
3599 to each microphone track and make them synchronized.
3601 The best result can be reached when you take one track as base and
3602 synchronize other tracks one by one with it.
3603 Remember that synchronization/delay tolerance depends on sample rate, too.
3604 Higher sample rates will give more tolerance.
3606 The filter accepts the following parameters:
3610 Set millimeters distance. This is compensation distance for fine tuning.
3614 Set cm distance. This is compensation distance for tightening distance setup.
3618 Set meters distance. This is compensation distance for hard distance setup.
3622 Set dry amount. Amount of unprocessed (dry) signal.
3626 Set wet amount. Amount of processed (wet) signal.
3630 Set temperature in degrees Celsius. This is the temperature of the environment.
3635 Apply headphone crossfeed filter.
3637 Crossfeed is the process of blending the left and right channels of stereo
3639 It is mainly used to reduce extreme stereo separation of low frequencies.
3641 The intent is to produce more speaker like sound to the listener.
3643 The filter accepts the following options:
3647 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3648 This sets gain of low shelf filter for side part of stereo image.
3649 Default is -6dB. Max allowed is -30db when strength is set to 1.
3652 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3653 This sets cut off frequency of low shelf filter. Default is cut off near
3654 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3657 Set curve slope of low shelf filter. Default is 0.5.
3658 Allowed range is from 0.01 to 1.
3661 Set input gain. Default is 0.9.
3664 Set output gain. Default is 1.
3667 @subsection Commands
3669 This filter supports the all above options as @ref{commands}.
3671 @section crystalizer
3672 Simple algorithm to expand audio dynamic range.
3674 The filter accepts the following options:
3678 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3679 (unchanged sound) to 10.0 (maximum effect).
3682 Enable clipping. By default is enabled.
3685 @subsection Commands
3687 This filter supports the all above options as @ref{commands}.
3690 Apply a DC shift to the audio.
3692 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3693 in the recording chain) from the audio. The effect of a DC offset is reduced
3694 headroom and hence volume. The @ref{astats} filter can be used to determine if
3695 a signal has a DC offset.
3699 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3703 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3704 used to prevent clipping.
3709 Apply de-essing to the audio samples.
3713 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3717 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3721 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3725 Set the output mode.
3727 It accepts the following values:
3730 Pass input unchanged.
3733 Pass ess filtered out.
3738 Default value is @var{o}.
3744 Measure audio dynamic range.
3746 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3747 is found in transition material. And anything less that 8 have very poor dynamics
3748 and is very compressed.
3750 The filter accepts the following options:
3754 Set window length in seconds used to split audio into segments of equal length.
3755 Default is 3 seconds.
3759 Dynamic Audio Normalizer.
3761 This filter applies a certain amount of gain to the input audio in order
3762 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3763 contrast to more "simple" normalization algorithms, the Dynamic Audio
3764 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3765 This allows for applying extra gain to the "quiet" sections of the audio
3766 while avoiding distortions or clipping the "loud" sections. In other words:
3767 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3768 sections, in the sense that the volume of each section is brought to the
3769 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3770 this goal *without* applying "dynamic range compressing". It will retain 100%
3771 of the dynamic range *within* each section of the audio file.
3775 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3776 Default is 500 milliseconds.
3777 The Dynamic Audio Normalizer processes the input audio in small chunks,
3778 referred to as frames. This is required, because a peak magnitude has no
3779 meaning for just a single sample value. Instead, we need to determine the
3780 peak magnitude for a contiguous sequence of sample values. While a "standard"
3781 normalizer would simply use the peak magnitude of the complete file, the
3782 Dynamic Audio Normalizer determines the peak magnitude individually for each
3783 frame. The length of a frame is specified in milliseconds. By default, the
3784 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3785 been found to give good results with most files.
3786 Note that the exact frame length, in number of samples, will be determined
3787 automatically, based on the sampling rate of the individual input audio file.
3790 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3791 number. Default is 31.
3792 Probably the most important parameter of the Dynamic Audio Normalizer is the
3793 @code{window size} of the Gaussian smoothing filter. The filter's window size
3794 is specified in frames, centered around the current frame. For the sake of
3795 simplicity, this must be an odd number. Consequently, the default value of 31
3796 takes into account the current frame, as well as the 15 preceding frames and
3797 the 15 subsequent frames. Using a larger window results in a stronger
3798 smoothing effect and thus in less gain variation, i.e. slower gain
3799 adaptation. Conversely, using a smaller window results in a weaker smoothing
3800 effect and thus in more gain variation, i.e. faster gain adaptation.
3801 In other words, the more you increase this value, the more the Dynamic Audio
3802 Normalizer will behave like a "traditional" normalization filter. On the
3803 contrary, the more you decrease this value, the more the Dynamic Audio
3804 Normalizer will behave like a dynamic range compressor.
3807 Set the target peak value. This specifies the highest permissible magnitude
3808 level for the normalized audio input. This filter will try to approach the
3809 target peak magnitude as closely as possible, but at the same time it also
3810 makes sure that the normalized signal will never exceed the peak magnitude.
3811 A frame's maximum local gain factor is imposed directly by the target peak
3812 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3813 It is not recommended to go above this value.
3816 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3817 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3818 factor for each input frame, i.e. the maximum gain factor that does not
3819 result in clipping or distortion. The maximum gain factor is determined by
3820 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3821 additionally bounds the frame's maximum gain factor by a predetermined
3822 (global) maximum gain factor. This is done in order to avoid excessive gain
3823 factors in "silent" or almost silent frames. By default, the maximum gain
3824 factor is 10.0, For most inputs the default value should be sufficient and
3825 it usually is not recommended to increase this value. Though, for input
3826 with an extremely low overall volume level, it may be necessary to allow even
3827 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3828 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3829 Instead, a "sigmoid" threshold function will be applied. This way, the
3830 gain factors will smoothly approach the threshold value, but never exceed that
3834 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3835 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3836 This means that the maximum local gain factor for each frame is defined
3837 (only) by the frame's highest magnitude sample. This way, the samples can
3838 be amplified as much as possible without exceeding the maximum signal
3839 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3840 Normalizer can also take into account the frame's root mean square,
3841 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3842 determine the power of a time-varying signal. It is therefore considered
3843 that the RMS is a better approximation of the "perceived loudness" than
3844 just looking at the signal's peak magnitude. Consequently, by adjusting all
3845 frames to a constant RMS value, a uniform "perceived loudness" can be
3846 established. If a target RMS value has been specified, a frame's local gain
3847 factor is defined as the factor that would result in exactly that RMS value.
3848 Note, however, that the maximum local gain factor is still restricted by the
3849 frame's highest magnitude sample, in order to prevent clipping.
3852 Enable channels coupling. By default is enabled.
3853 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3854 amount. This means the same gain factor will be applied to all channels, i.e.
3855 the maximum possible gain factor is determined by the "loudest" channel.
3856 However, in some recordings, it may happen that the volume of the different
3857 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3858 In this case, this option can be used to disable the channel coupling. This way,
3859 the gain factor will be determined independently for each channel, depending
3860 only on the individual channel's highest magnitude sample. This allows for
3861 harmonizing the volume of the different channels.
3864 Enable DC bias correction. By default is disabled.
3865 An audio signal (in the time domain) is a sequence of sample values.
3866 In the Dynamic Audio Normalizer these sample values are represented in the
3867 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3868 audio signal, or "waveform", should be centered around the zero point.
3869 That means if we calculate the mean value of all samples in a file, or in a
3870 single frame, then the result should be 0.0 or at least very close to that
3871 value. If, however, there is a significant deviation of the mean value from
3872 0.0, in either positive or negative direction, this is referred to as a
3873 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3874 Audio Normalizer provides optional DC bias correction.
3875 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3876 the mean value, or "DC correction" offset, of each input frame and subtract
3877 that value from all of the frame's sample values which ensures those samples
3878 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3879 boundaries, the DC correction offset values will be interpolated smoothly
3880 between neighbouring frames.
3882 @item altboundary, b
3883 Enable alternative boundary mode. By default is disabled.
3884 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3885 around each frame. This includes the preceding frames as well as the
3886 subsequent frames. However, for the "boundary" frames, located at the very
3887 beginning and at the very end of the audio file, not all neighbouring
3888 frames are available. In particular, for the first few frames in the audio
3889 file, the preceding frames are not known. And, similarly, for the last few
3890 frames in the audio file, the subsequent frames are not known. Thus, the
3891 question arises which gain factors should be assumed for the missing frames
3892 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3893 to deal with this situation. The default boundary mode assumes a gain factor
3894 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3895 "fade out" at the beginning and at the end of the input, respectively.
3898 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3899 By default, the Dynamic Audio Normalizer does not apply "traditional"
3900 compression. This means that signal peaks will not be pruned and thus the
3901 full dynamic range will be retained within each local neighbourhood. However,
3902 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3903 normalization algorithm with a more "traditional" compression.
3904 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3905 (thresholding) function. If (and only if) the compression feature is enabled,
3906 all input frames will be processed by a soft knee thresholding function prior
3907 to the actual normalization process. Put simply, the thresholding function is
3908 going to prune all samples whose magnitude exceeds a certain threshold value.
3909 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3910 value. Instead, the threshold value will be adjusted for each individual
3912 In general, smaller parameters result in stronger compression, and vice versa.
3913 Values below 3.0 are not recommended, because audible distortion may appear.
3916 Set the target threshold value. This specifies the lowest permissible
3917 magnitude level for the audio input which will be normalized.
3918 If input frame volume is above this value frame will be normalized.
3919 Otherwise frame may not be normalized at all. The default value is set
3920 to 0, which means all input frames will be normalized.
3921 This option is mostly useful if digital noise is not wanted to be amplified.
3924 @subsection Commands
3926 This filter supports the all above options as @ref{commands}.
3930 Make audio easier to listen to on headphones.
3932 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3933 so that when listened to on headphones the stereo image is moved from
3934 inside your head (standard for headphones) to outside and in front of
3935 the listener (standard for speakers).
3941 Apply a two-pole peaking equalisation (EQ) filter. With this
3942 filter, the signal-level at and around a selected frequency can
3943 be increased or decreased, whilst (unlike bandpass and bandreject
3944 filters) that at all other frequencies is unchanged.
3946 In order to produce complex equalisation curves, this filter can
3947 be given several times, each with a different central frequency.
3949 The filter accepts the following options:
3953 Set the filter's central frequency in Hz.
3956 Set method to specify band-width of filter.
3971 Specify the band-width of a filter in width_type units.
3974 Set the required gain or attenuation in dB.
3975 Beware of clipping when using a positive gain.
3978 How much to use filtered signal in output. Default is 1.
3979 Range is between 0 and 1.
3982 Specify which channels to filter, by default all available are filtered.
3985 Normalize biquad coefficients, by default is disabled.
3986 Enabling it will normalize magnitude response at DC to 0dB.
3989 Set transform type of IIR filter.
3998 Set precison of filtering.
4001 Pick automatic sample format depending on surround filters.
4003 Always use signed 16-bit.
4005 Always use signed 32-bit.
4007 Always use float 32-bit.
4009 Always use float 64-bit.
4013 @subsection Examples
4016 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4018 equalizer=f=1000:t=h:width=200:g=-10
4022 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4024 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4028 @subsection Commands
4030 This filter supports the following commands:
4033 Change equalizer frequency.
4034 Syntax for the command is : "@var{frequency}"
4037 Change equalizer width_type.
4038 Syntax for the command is : "@var{width_type}"
4041 Change equalizer width.
4042 Syntax for the command is : "@var{width}"
4045 Change equalizer gain.
4046 Syntax for the command is : "@var{gain}"
4049 Change equalizer mix.
4050 Syntax for the command is : "@var{mix}"
4053 @section extrastereo
4055 Linearly increases the difference between left and right channels which
4056 adds some sort of "live" effect to playback.
4058 The filter accepts the following options:
4062 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4063 (average of both channels), with 1.0 sound will be unchanged, with
4064 -1.0 left and right channels will be swapped.
4067 Enable clipping. By default is enabled.
4070 @subsection Commands
4072 This filter supports the all above options as @ref{commands}.
4074 @section firequalizer
4075 Apply FIR Equalization using arbitrary frequency response.
4077 The filter accepts the following option:
4081 Set gain curve equation (in dB). The expression can contain variables:
4084 the evaluated frequency
4088 channel number, set to 0 when multichannels evaluation is disabled
4090 channel id, see libavutil/channel_layout.h, set to the first channel id when
4091 multichannels evaluation is disabled
4095 channel_layout, see libavutil/channel_layout.h
4100 @item gain_interpolate(f)
4101 interpolate gain on frequency f based on gain_entry
4102 @item cubic_interpolate(f)
4103 same as gain_interpolate, but smoother
4105 This option is also available as command. Default is @code{gain_interpolate(f)}.
4108 Set gain entry for gain_interpolate function. The expression can
4112 store gain entry at frequency f with value g
4114 This option is also available as command.
4117 Set filter delay in seconds. Higher value means more accurate.
4118 Default is @code{0.01}.
4121 Set filter accuracy in Hz. Lower value means more accurate.
4122 Default is @code{5}.
4125 Set window function. Acceptable values are:
4128 rectangular window, useful when gain curve is already smooth
4130 hann window (default)
4136 3-terms continuous 1st derivative nuttall window
4138 minimum 3-terms discontinuous nuttall window
4140 4-terms continuous 1st derivative nuttall window
4142 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4144 blackman-harris window
4150 If enabled, use fixed number of audio samples. This improves speed when
4151 filtering with large delay. Default is disabled.
4154 Enable multichannels evaluation on gain. Default is disabled.
4157 Enable zero phase mode by subtracting timestamp to compensate delay.
4158 Default is disabled.
4161 Set scale used by gain. Acceptable values are:
4164 linear frequency, linear gain
4166 linear frequency, logarithmic (in dB) gain (default)
4168 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4170 logarithmic frequency, logarithmic gain
4174 Set file for dumping, suitable for gnuplot.
4177 Set scale for dumpfile. Acceptable values are same with scale option.
4181 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4182 Default is disabled.
4185 Enable minimum phase impulse response. Default is disabled.
4188 @subsection Examples
4193 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4196 lowpass at 1000 Hz with gain_entry:
4198 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4201 custom equalization:
4203 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4206 higher delay with zero phase to compensate delay:
4208 firequalizer=delay=0.1:fixed=on:zero_phase=on
4211 lowpass on left channel, highpass on right channel:
4213 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4214 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4219 Apply a flanging effect to the audio.
4221 The filter accepts the following options:
4225 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4228 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4231 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4235 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4236 Default value is 71.
4239 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4242 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4243 Default value is @var{sinusoidal}.
4246 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4247 Default value is 25.
4250 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4251 Default is @var{linear}.
4255 Apply Haas effect to audio.
4257 Note that this makes most sense to apply on mono signals.
4258 With this filter applied to mono signals it give some directionality and
4259 stretches its stereo image.
4261 The filter accepts the following options:
4265 Set input level. By default is @var{1}, or 0dB
4268 Set output level. By default is @var{1}, or 0dB.
4271 Set gain applied to side part of signal. By default is @var{1}.
4274 Set kind of middle source. Can be one of the following:
4284 Pick middle part signal of stereo image.
4287 Pick side part signal of stereo image.
4291 Change middle phase. By default is disabled.
4294 Set left channel delay. By default is @var{2.05} milliseconds.
4297 Set left channel balance. By default is @var{-1}.
4300 Set left channel gain. By default is @var{1}.
4303 Change left phase. By default is disabled.
4306 Set right channel delay. By defaults is @var{2.12} milliseconds.
4309 Set right channel balance. By default is @var{1}.
4312 Set right channel gain. By default is @var{1}.
4315 Change right phase. By default is enabled.
4320 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4321 embedded HDCD codes is expanded into a 20-bit PCM stream.
4323 The filter supports the Peak Extend and Low-level Gain Adjustment features
4324 of HDCD, and detects the Transient Filter flag.
4327 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4330 When using the filter with wav, note the default encoding for wav is 16-bit,
4331 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4332 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4334 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4335 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4338 The filter accepts the following options:
4341 @item disable_autoconvert
4342 Disable any automatic format conversion or resampling in the filter graph.
4344 @item process_stereo
4345 Process the stereo channels together. If target_gain does not match between
4346 channels, consider it invalid and use the last valid target_gain.
4349 Set the code detect timer period in ms.
4352 Always extend peaks above -3dBFS even if PE isn't signaled.
4355 Replace audio with a solid tone and adjust the amplitude to signal some
4356 specific aspect of the decoding process. The output file can be loaded in
4357 an audio editor alongside the original to aid analysis.
4359 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4366 Gain adjustment level at each sample
4368 Samples where peak extend occurs
4370 Samples where the code detect timer is active
4372 Samples where the target gain does not match between channels
4378 Apply head-related transfer functions (HRTFs) to create virtual
4379 loudspeakers around the user for binaural listening via headphones.
4380 The HRIRs are provided via additional streams, for each channel
4381 one stereo input stream is needed.
4383 The filter accepts the following options:
4387 Set mapping of input streams for convolution.
4388 The argument is a '|'-separated list of channel names in order as they
4389 are given as additional stream inputs for filter.
4390 This also specify number of input streams. Number of input streams
4391 must be not less than number of channels in first stream plus one.
4394 Set gain applied to audio. Value is in dB. Default is 0.
4397 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4398 processing audio in time domain which is slow.
4399 @var{freq} is processing audio in frequency domain which is fast.
4400 Default is @var{freq}.
4403 Set custom gain for LFE channels. Value is in dB. Default is 0.
4406 Set size of frame in number of samples which will be processed at once.
4407 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4410 Set format of hrir stream.
4411 Default value is @var{stereo}. Alternative value is @var{multich}.
4412 If value is set to @var{stereo}, number of additional streams should
4413 be greater or equal to number of input channels in first input stream.
4414 Also each additional stream should have stereo number of channels.
4415 If value is set to @var{multich}, number of additional streams should
4416 be exactly one. Also number of input channels of additional stream
4417 should be equal or greater than twice number of channels of first input
4421 @subsection Examples
4425 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4426 each amovie filter use stereo file with IR coefficients as input.
4427 The files give coefficients for each position of virtual loudspeaker:
4430 -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"
4435 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4436 but now in @var{multich} @var{hrir} format.
4438 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"
4445 Apply a high-pass filter with 3dB point frequency.
4446 The filter can be either single-pole, or double-pole (the default).
4447 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4449 The filter accepts the following options:
4453 Set frequency in Hz. Default is 3000.
4456 Set number of poles. Default is 2.
4459 Set method to specify band-width of filter.
4474 Specify the band-width of a filter in width_type units.
4475 Applies only to double-pole filter.
4476 The default is 0.707q and gives a Butterworth response.
4479 How much to use filtered signal in output. Default is 1.
4480 Range is between 0 and 1.
4483 Specify which channels to filter, by default all available are filtered.
4486 Normalize biquad coefficients, by default is disabled.
4487 Enabling it will normalize magnitude response at DC to 0dB.
4490 Set transform type of IIR filter.
4499 Set precison of filtering.
4502 Pick automatic sample format depending on surround filters.
4504 Always use signed 16-bit.
4506 Always use signed 32-bit.
4508 Always use float 32-bit.
4510 Always use float 64-bit.
4514 @subsection Commands
4516 This filter supports the following commands:
4519 Change highpass frequency.
4520 Syntax for the command is : "@var{frequency}"
4523 Change highpass width_type.
4524 Syntax for the command is : "@var{width_type}"
4527 Change highpass width.
4528 Syntax for the command is : "@var{width}"
4531 Change highpass mix.
4532 Syntax for the command is : "@var{mix}"
4537 Join multiple input streams into one multi-channel stream.
4539 It accepts the following parameters:
4543 The number of input streams. It defaults to 2.
4545 @item channel_layout
4546 The desired output channel layout. It defaults to stereo.
4549 Map channels from inputs to output. The argument is a '|'-separated list of
4550 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4551 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4552 can be either the name of the input channel (e.g. FL for front left) or its
4553 index in the specified input stream. @var{out_channel} is the name of the output
4557 The filter will attempt to guess the mappings when they are not specified
4558 explicitly. It does so by first trying to find an unused matching input channel
4559 and if that fails it picks the first unused input channel.
4561 Join 3 inputs (with properly set channel layouts):
4563 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4566 Build a 5.1 output from 6 single-channel streams:
4568 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4569 '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'
4575 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4577 To enable compilation of this filter you need to configure FFmpeg with
4578 @code{--enable-ladspa}.
4582 Specifies the name of LADSPA plugin library to load. If the environment
4583 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4584 each one of the directories specified by the colon separated list in
4585 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4586 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4587 @file{/usr/lib/ladspa/}.
4590 Specifies the plugin within the library. Some libraries contain only
4591 one plugin, but others contain many of them. If this is not set filter
4592 will list all available plugins within the specified library.
4595 Set the '|' separated list of controls which are zero or more floating point
4596 values that determine the behavior of the loaded plugin (for example delay,
4598 Controls need to be defined using the following syntax:
4599 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4600 @var{valuei} is the value set on the @var{i}-th control.
4601 Alternatively they can be also defined using the following syntax:
4602 @var{value0}|@var{value1}|@var{value2}|..., where
4603 @var{valuei} is the value set on the @var{i}-th control.
4604 If @option{controls} is set to @code{help}, all available controls and
4605 their valid ranges are printed.
4607 @item sample_rate, s
4608 Specify the sample rate, default to 44100. Only used if plugin have
4612 Set the number of samples per channel per each output frame, default
4613 is 1024. Only used if plugin have zero inputs.
4616 Set the minimum duration of the sourced audio. See
4617 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4618 for the accepted syntax.
4619 Note that the resulting duration may be greater than the specified duration,
4620 as the generated audio is always cut at the end of a complete frame.
4621 If not specified, or the expressed duration is negative, the audio is
4622 supposed to be generated forever.
4623 Only used if plugin have zero inputs.
4626 Enable latency compensation, by default is disabled.
4627 Only used if plugin have inputs.
4630 @subsection Examples
4634 List all available plugins within amp (LADSPA example plugin) library:
4640 List all available controls and their valid ranges for @code{vcf_notch}
4641 plugin from @code{VCF} library:
4643 ladspa=f=vcf:p=vcf_notch:c=help
4647 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4650 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4654 Add reverberation to the audio using TAP-plugins
4655 (Tom's Audio Processing plugins):
4657 ladspa=file=tap_reverb:tap_reverb
4661 Generate white noise, with 0.2 amplitude:
4663 ladspa=file=cmt:noise_source_white:c=c0=.2
4667 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4668 @code{C* Audio Plugin Suite} (CAPS) library:
4670 ladspa=file=caps:Click:c=c1=20'
4674 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4676 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4680 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4681 @code{SWH Plugins} collection:
4683 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4687 Attenuate low frequencies using Multiband EQ from Steve Harris
4688 @code{SWH Plugins} collection:
4690 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4694 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4697 ladspa=caps:Narrower
4701 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4703 ladspa=caps:White:.2
4707 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4709 ladspa=caps:Fractal:c=c1=1
4713 Dynamic volume normalization using @code{VLevel} plugin:
4715 ladspa=vlevel-ladspa:vlevel_mono
4719 @subsection Commands
4721 This filter supports the following commands:
4724 Modify the @var{N}-th control value.
4726 If the specified value is not valid, it is ignored and prior one is kept.
4731 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4732 Support for both single pass (livestreams, files) and double pass (files) modes.
4733 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4734 detect true peaks, the audio stream will be upsampled to 192 kHz.
4735 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4737 The filter accepts the following options:
4741 Set integrated loudness target.
4742 Range is -70.0 - -5.0. Default value is -24.0.
4745 Set loudness range target.
4746 Range is 1.0 - 20.0. Default value is 7.0.
4749 Set maximum true peak.
4750 Range is -9.0 - +0.0. Default value is -2.0.
4752 @item measured_I, measured_i
4753 Measured IL of input file.
4754 Range is -99.0 - +0.0.
4756 @item measured_LRA, measured_lra
4757 Measured LRA of input file.
4758 Range is 0.0 - 99.0.
4760 @item measured_TP, measured_tp
4761 Measured true peak of input file.
4762 Range is -99.0 - +99.0.
4764 @item measured_thresh
4765 Measured threshold of input file.
4766 Range is -99.0 - +0.0.
4769 Set offset gain. Gain is applied before the true-peak limiter.
4770 Range is -99.0 - +99.0. Default is +0.0.
4773 Normalize by linearly scaling the source audio.
4774 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4775 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4776 be lower than source LRA and the change in integrated loudness shouldn't
4777 result in a true peak which exceeds the target TP. If any of these
4778 conditions aren't met, normalization mode will revert to @var{dynamic}.
4779 Options are @code{true} or @code{false}. Default is @code{true}.
4782 Treat mono input files as "dual-mono". If a mono file is intended for playback
4783 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4784 If set to @code{true}, this option will compensate for this effect.
4785 Multi-channel input files are not affected by this option.
4786 Options are true or false. Default is false.
4789 Set print format for stats. Options are summary, json, or none.
4790 Default value is none.
4795 Apply a low-pass filter with 3dB point frequency.
4796 The filter can be either single-pole or double-pole (the default).
4797 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4799 The filter accepts the following options:
4803 Set frequency in Hz. Default is 500.
4806 Set number of poles. Default is 2.
4809 Set method to specify band-width of filter.
4824 Specify the band-width of a filter in width_type units.
4825 Applies only to double-pole filter.
4826 The default is 0.707q and gives a Butterworth response.
4829 How much to use filtered signal in output. Default is 1.
4830 Range is between 0 and 1.
4833 Specify which channels to filter, by default all available are filtered.
4836 Normalize biquad coefficients, by default is disabled.
4837 Enabling it will normalize magnitude response at DC to 0dB.
4840 Set transform type of IIR filter.
4849 Set precison of filtering.
4852 Pick automatic sample format depending on surround filters.
4854 Always use signed 16-bit.
4856 Always use signed 32-bit.
4858 Always use float 32-bit.
4860 Always use float 64-bit.
4864 @subsection Examples
4867 Lowpass only LFE channel, it LFE is not present it does nothing:
4873 @subsection Commands
4875 This filter supports the following commands:
4878 Change lowpass frequency.
4879 Syntax for the command is : "@var{frequency}"
4882 Change lowpass width_type.
4883 Syntax for the command is : "@var{width_type}"
4886 Change lowpass width.
4887 Syntax for the command is : "@var{width}"
4891 Syntax for the command is : "@var{mix}"
4896 Load a LV2 (LADSPA Version 2) plugin.
4898 To enable compilation of this filter you need to configure FFmpeg with
4899 @code{--enable-lv2}.
4903 Specifies the plugin URI. You may need to escape ':'.
4906 Set the '|' separated list of controls which are zero or more floating point
4907 values that determine the behavior of the loaded plugin (for example delay,
4909 If @option{controls} is set to @code{help}, all available controls and
4910 their valid ranges are printed.
4912 @item sample_rate, s
4913 Specify the sample rate, default to 44100. Only used if plugin have
4917 Set the number of samples per channel per each output frame, default
4918 is 1024. Only used if plugin have zero inputs.
4921 Set the minimum duration of the sourced audio. See
4922 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4923 for the accepted syntax.
4924 Note that the resulting duration may be greater than the specified duration,
4925 as the generated audio is always cut at the end of a complete frame.
4926 If not specified, or the expressed duration is negative, the audio is
4927 supposed to be generated forever.
4928 Only used if plugin have zero inputs.
4931 @subsection Examples
4935 Apply bass enhancer plugin from Calf:
4937 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4941 Apply vinyl plugin from Calf:
4943 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4947 Apply bit crusher plugin from ArtyFX:
4949 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4954 Multiband Compress or expand the audio's dynamic range.
4956 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4957 This is akin to the crossover of a loudspeaker, and results in flat frequency
4958 response when absent compander action.
4960 It accepts the following parameters:
4964 This option syntax is:
4965 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4966 For explanation of each item refer to compand filter documentation.
4972 Mix channels with specific gain levels. The filter accepts the output
4973 channel layout followed by a set of channels definitions.
4975 This filter is also designed to efficiently remap the channels of an audio
4978 The filter accepts parameters of the form:
4979 "@var{l}|@var{outdef}|@var{outdef}|..."
4983 output channel layout or number of channels
4986 output channel specification, of the form:
4987 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4990 output channel to define, either a channel name (FL, FR, etc.) or a channel
4991 number (c0, c1, etc.)
4994 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4997 input channel to use, see out_name for details; it is not possible to mix
4998 named and numbered input channels
5001 If the `=' in a channel specification is replaced by `<', then the gains for
5002 that specification will be renormalized so that the total is 1, thus
5003 avoiding clipping noise.
5005 @subsection Mixing examples
5007 For example, if you want to down-mix from stereo to mono, but with a bigger
5008 factor for the left channel:
5010 pan=1c|c0=0.9*c0+0.1*c1
5013 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5014 7-channels surround:
5016 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5019 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5020 that should be preferred (see "-ac" option) unless you have very specific
5023 @subsection Remapping examples
5025 The channel remapping will be effective if, and only if:
5028 @item gain coefficients are zeroes or ones,
5029 @item only one input per channel output,
5032 If all these conditions are satisfied, the filter will notify the user ("Pure
5033 channel mapping detected"), and use an optimized and lossless method to do the
5036 For example, if you have a 5.1 source and want a stereo audio stream by
5037 dropping the extra channels:
5039 pan="stereo| c0=FL | c1=FR"
5042 Given the same source, you can also switch front left and front right channels
5043 and keep the input channel layout:
5045 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5048 If the input is a stereo audio stream, you can mute the front left channel (and
5049 still keep the stereo channel layout) with:
5054 Still with a stereo audio stream input, you can copy the right channel in both
5055 front left and right:
5057 pan="stereo| c0=FR | c1=FR"
5062 ReplayGain scanner filter. This filter takes an audio stream as an input and
5063 outputs it unchanged.
5064 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5068 Convert the audio sample format, sample rate and channel layout. It is
5069 not meant to be used directly.
5072 Apply time-stretching and pitch-shifting with librubberband.
5074 To enable compilation of this filter, you need to configure FFmpeg with
5075 @code{--enable-librubberband}.
5077 The filter accepts the following options:
5081 Set tempo scale factor.
5084 Set pitch scale factor.
5087 Set transients detector.
5088 Possible values are:
5097 Possible values are:
5106 Possible values are:
5113 Set processing window size.
5114 Possible values are:
5123 Possible values are:
5130 Enable formant preservation when shift pitching.
5131 Possible values are:
5139 Possible values are:
5148 Possible values are:
5155 @subsection Commands
5157 This filter supports the following commands:
5160 Change filter tempo scale factor.
5161 Syntax for the command is : "@var{tempo}"
5164 Change filter pitch scale factor.
5165 Syntax for the command is : "@var{pitch}"
5168 @section sidechaincompress
5170 This filter acts like normal compressor but has the ability to compress
5171 detected signal using second input signal.
5172 It needs two input streams and returns one output stream.
5173 First input stream will be processed depending on second stream signal.
5174 The filtered signal then can be filtered with other filters in later stages of
5175 processing. See @ref{pan} and @ref{amerge} filter.
5177 The filter accepts the following options:
5181 Set input gain. Default is 1. Range is between 0.015625 and 64.
5184 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5185 Default is @code{downward}.
5188 If a signal of second stream raises above this level it will affect the gain
5189 reduction of first stream.
5190 By default is 0.125. Range is between 0.00097563 and 1.
5193 Set a ratio about which the signal is reduced. 1:2 means that if the level
5194 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5195 Default is 2. Range is between 1 and 20.
5198 Amount of milliseconds the signal has to rise above the threshold before gain
5199 reduction starts. Default is 20. Range is between 0.01 and 2000.
5202 Amount of milliseconds the signal has to fall below the threshold before
5203 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5206 Set the amount by how much signal will be amplified after processing.
5207 Default is 1. Range is from 1 to 64.
5210 Curve the sharp knee around the threshold to enter gain reduction more softly.
5211 Default is 2.82843. Range is between 1 and 8.
5214 Choose if the @code{average} level between all channels of side-chain stream
5215 or the louder(@code{maximum}) channel of side-chain stream affects the
5216 reduction. Default is @code{average}.
5219 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5220 of @code{rms}. Default is @code{rms} which is mainly smoother.
5223 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5226 How much to use compressed signal in output. Default is 1.
5227 Range is between 0 and 1.
5230 @subsection Commands
5232 This filter supports the all above options as @ref{commands}.
5234 @subsection Examples
5238 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5239 depending on the signal of 2nd input and later compressed signal to be
5240 merged with 2nd input:
5242 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5246 @section sidechaingate
5248 A sidechain gate acts like a normal (wideband) gate but has the ability to
5249 filter the detected signal before sending it to the gain reduction stage.
5250 Normally a gate uses the full range signal to detect a level above the
5252 For example: If you cut all lower frequencies from your sidechain signal
5253 the gate will decrease the volume of your track only if not enough highs
5254 appear. With this technique you are able to reduce the resonation of a
5255 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5257 It needs two input streams and returns one output stream.
5258 First input stream will be processed depending on second stream signal.
5260 The filter accepts the following options:
5264 Set input level before filtering.
5265 Default is 1. Allowed range is from 0.015625 to 64.
5268 Set the mode of operation. Can be @code{upward} or @code{downward}.
5269 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5270 will be amplified, expanding dynamic range in upward direction.
5271 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5274 Set the level of gain reduction when the signal is below the threshold.
5275 Default is 0.06125. Allowed range is from 0 to 1.
5276 Setting this to 0 disables reduction and then filter behaves like expander.
5279 If a signal rises above this level the gain reduction is released.
5280 Default is 0.125. Allowed range is from 0 to 1.
5283 Set a ratio about which the signal is reduced.
5284 Default is 2. Allowed range is from 1 to 9000.
5287 Amount of milliseconds the signal has to rise above the threshold before gain
5289 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5292 Amount of milliseconds the signal has to fall below the threshold before the
5293 reduction is increased again. Default is 250 milliseconds.
5294 Allowed range is from 0.01 to 9000.
5297 Set amount of amplification of signal after processing.
5298 Default is 1. Allowed range is from 1 to 64.
5301 Curve the sharp knee around the threshold to enter gain reduction more softly.
5302 Default is 2.828427125. Allowed range is from 1 to 8.
5305 Choose if exact signal should be taken for detection or an RMS like one.
5306 Default is rms. Can be peak or rms.
5309 Choose if the average level between all channels or the louder channel affects
5311 Default is average. Can be average or maximum.
5314 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5317 @subsection Commands
5319 This filter supports the all above options as @ref{commands}.
5321 @section silencedetect
5323 Detect silence in an audio stream.
5325 This filter logs a message when it detects that the input audio volume is less
5326 or equal to a noise tolerance value for a duration greater or equal to the
5327 minimum detected noise duration.
5329 The printed times and duration are expressed in seconds. The
5330 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5331 is set on the first frame whose timestamp equals or exceeds the detection
5332 duration and it contains the timestamp of the first frame of the silence.
5334 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5335 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5336 keys are set on the first frame after the silence. If @option{mono} is
5337 enabled, and each channel is evaluated separately, the @code{.X}
5338 suffixed keys are used, and @code{X} corresponds to the channel number.
5340 The filter accepts the following options:
5344 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5345 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5348 Set silence duration until notification (default is 2 seconds). See
5349 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5350 for the accepted syntax.
5353 Process each channel separately, instead of combined. By default is disabled.
5356 @subsection Examples
5360 Detect 5 seconds of silence with -50dB noise tolerance:
5362 silencedetect=n=-50dB:d=5
5366 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5367 tolerance in @file{silence.mp3}:
5369 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5373 @section silenceremove
5375 Remove silence from the beginning, middle or end of the audio.
5377 The filter accepts the following options:
5381 This value is used to indicate if audio should be trimmed at beginning of
5382 the audio. A value of zero indicates no silence should be trimmed from the
5383 beginning. When specifying a non-zero value, it trims audio up until it
5384 finds non-silence. Normally, when trimming silence from beginning of audio
5385 the @var{start_periods} will be @code{1} but it can be increased to higher
5386 values to trim all audio up to specific count of non-silence periods.
5387 Default value is @code{0}.
5389 @item start_duration
5390 Specify the amount of time that non-silence must be detected before it stops
5391 trimming audio. By increasing the duration, bursts of noises can be treated
5392 as silence and trimmed off. Default value is @code{0}.
5394 @item start_threshold
5395 This indicates what sample value should be treated as silence. For digital
5396 audio, a value of @code{0} may be fine but for audio recorded from analog,
5397 you may wish to increase the value to account for background noise.
5398 Can be specified in dB (in case "dB" is appended to the specified value)
5399 or amplitude ratio. Default value is @code{0}.
5402 Specify max duration of silence at beginning that will be kept after
5403 trimming. Default is 0, which is equal to trimming all samples detected
5407 Specify mode of detection of silence end in start of multi-channel audio.
5408 Can be @var{any} or @var{all}. Default is @var{any}.
5409 With @var{any}, any sample that is detected as non-silence will cause
5410 stopped trimming of silence.
5411 With @var{all}, only if all channels are detected as non-silence will cause
5412 stopped trimming of silence.
5415 Set the count for trimming silence from the end of audio.
5416 To remove silence from the middle of a file, specify a @var{stop_periods}
5417 that is negative. This value is then treated as a positive value and is
5418 used to indicate the effect should restart processing as specified by
5419 @var{start_periods}, making it suitable for removing periods of silence
5420 in the middle of the audio.
5421 Default value is @code{0}.
5424 Specify a duration of silence that must exist before audio is not copied any
5425 more. By specifying a higher duration, silence that is wanted can be left in
5427 Default value is @code{0}.
5429 @item stop_threshold
5430 This is the same as @option{start_threshold} but for trimming silence from
5432 Can be specified in dB (in case "dB" is appended to the specified value)
5433 or amplitude ratio. Default value is @code{0}.
5436 Specify max duration of silence at end that will be kept after
5437 trimming. Default is 0, which is equal to trimming all samples detected
5441 Specify mode of detection of silence start in end of multi-channel audio.
5442 Can be @var{any} or @var{all}. Default is @var{any}.
5443 With @var{any}, any sample that is detected as non-silence will cause
5444 stopped trimming of silence.
5445 With @var{all}, only if all channels are detected as non-silence will cause
5446 stopped trimming of silence.
5449 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5450 and works better with digital silence which is exactly 0.
5451 Default value is @code{rms}.
5454 Set duration in number of seconds used to calculate size of window in number
5455 of samples for detecting silence.
5456 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5459 @subsection Examples
5463 The following example shows how this filter can be used to start a recording
5464 that does not contain the delay at the start which usually occurs between
5465 pressing the record button and the start of the performance:
5467 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5471 Trim all silence encountered from beginning to end where there is more than 1
5472 second of silence in audio:
5474 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5478 Trim all digital silence samples, using peak detection, from beginning to end
5479 where there is more than 0 samples of digital silence in audio and digital
5480 silence is detected in all channels at same positions in stream:
5482 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5488 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5489 loudspeakers around the user for binaural listening via headphones (audio
5490 formats up to 9 channels supported).
5491 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5492 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5493 Austrian Academy of Sciences.
5495 To enable compilation of this filter you need to configure FFmpeg with
5496 @code{--enable-libmysofa}.
5498 The filter accepts the following options:
5502 Set the SOFA file used for rendering.
5505 Set gain applied to audio. Value is in dB. Default is 0.
5508 Set rotation of virtual loudspeakers in deg. Default is 0.
5511 Set elevation of virtual speakers in deg. Default is 0.
5514 Set distance in meters between loudspeakers and the listener with near-field
5515 HRTFs. Default is 1.
5518 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5519 processing audio in time domain which is slow.
5520 @var{freq} is processing audio in frequency domain which is fast.
5521 Default is @var{freq}.
5524 Set custom positions of virtual loudspeakers. Syntax for this option is:
5525 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5526 Each virtual loudspeaker is described with short channel name following with
5527 azimuth and elevation in degrees.
5528 Each virtual loudspeaker description is separated by '|'.
5529 For example to override front left and front right channel positions use:
5530 'speakers=FL 45 15|FR 345 15'.
5531 Descriptions with unrecognised channel names are ignored.
5534 Set custom gain for LFE channels. Value is in dB. Default is 0.
5537 Set custom frame size in number of samples. Default is 1024.
5538 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5539 is set to @var{freq}.
5542 Should all IRs be normalized upon importing SOFA file.
5543 By default is enabled.
5546 Should nearest IRs be interpolated with neighbor IRs if exact position
5547 does not match. By default is disabled.
5550 Minphase all IRs upon loading of SOFA file. By default is disabled.
5553 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5556 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5559 @subsection Examples
5563 Using ClubFritz6 sofa file:
5565 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5569 Using ClubFritz12 sofa file and bigger radius with small rotation:
5571 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5575 Similar as above but with custom speaker positions for front left, front right, back left and back right
5576 and also with custom gain:
5578 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5585 This filter expands or compresses each half-cycle of audio samples
5586 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5587 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5589 The filter accepts the following options:
5593 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5594 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5597 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5598 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5599 would be such that local peak value reaches target peak value but never to surpass it and that
5600 ratio between new and previous peak value does not surpass this option value.
5602 @item compression, c
5603 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5604 This option controls maximum local half-cycle of samples compression. This option is used
5605 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5606 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5607 that peak's half-cycle will be compressed by current compression factor.
5610 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5611 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5612 Any half-cycle samples with their local peak value below or same as this option value will be
5613 compressed by current compression factor, otherwise, if greater than threshold value they will be
5614 expanded with expansion factor so that it could reach peak target value but never surpass it.
5617 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5618 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5619 each new half-cycle until it reaches @option{expansion} value.
5620 Setting this options too high may lead to distortions.
5623 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5624 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5625 each new half-cycle until it reaches @option{compression} value.
5628 Specify which channels to filter, by default all available channels are filtered.
5631 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5632 option. When enabled any half-cycle of samples with their local peak value below or same as
5633 @option{threshold} option will be expanded otherwise it will be compressed.
5636 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5637 When disabled each filtered channel gain calculation is independent, otherwise when this option
5638 is enabled the minimum of all possible gains for each filtered channel is used.
5641 @subsection Commands
5643 This filter supports the all above options as @ref{commands}.
5645 @section stereotools
5647 This filter has some handy utilities to manage stereo signals, for converting
5648 M/S stereo recordings to L/R signal while having control over the parameters
5649 or spreading the stereo image of master track.
5651 The filter accepts the following options:
5655 Set input level before filtering for both channels. Defaults is 1.
5656 Allowed range is from 0.015625 to 64.
5659 Set output level after filtering for both channels. Defaults is 1.
5660 Allowed range is from 0.015625 to 64.
5663 Set input balance between both channels. Default is 0.
5664 Allowed range is from -1 to 1.
5667 Set output balance between both channels. Default is 0.
5668 Allowed range is from -1 to 1.
5671 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5672 clipping. Disabled by default.
5675 Mute the left channel. Disabled by default.
5678 Mute the right channel. Disabled by default.
5681 Change the phase of the left channel. Disabled by default.
5684 Change the phase of the right channel. Disabled by default.
5687 Set stereo mode. Available values are:
5691 Left/Right to Left/Right, this is default.
5694 Left/Right to Mid/Side.
5697 Mid/Side to Left/Right.
5700 Left/Right to Left/Left.
5703 Left/Right to Right/Right.
5706 Left/Right to Left + Right.
5709 Left/Right to Right/Left.
5712 Mid/Side to Left/Left.
5715 Mid/Side to Right/Right.
5718 Mid/Side to Right/Left.
5721 Left/Right to Left - Right.
5725 Set level of side signal. Default is 1.
5726 Allowed range is from 0.015625 to 64.
5729 Set balance of side signal. Default is 0.
5730 Allowed range is from -1 to 1.
5733 Set level of the middle signal. Default is 1.
5734 Allowed range is from 0.015625 to 64.
5737 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5740 Set stereo base between mono and inversed channels. Default is 0.
5741 Allowed range is from -1 to 1.
5744 Set delay in milliseconds how much to delay left from right channel and
5745 vice versa. Default is 0. Allowed range is from -20 to 20.
5748 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5751 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5753 @item bmode_in, bmode_out
5754 Set balance mode for balance_in/balance_out option.
5756 Can be one of the following:
5760 Classic balance mode. Attenuate one channel at time.
5761 Gain is raised up to 1.
5764 Similar as classic mode above but gain is raised up to 2.
5767 Equal power distribution, from -6dB to +6dB range.
5771 @subsection Commands
5773 This filter supports the all above options as @ref{commands}.
5775 @subsection Examples
5779 Apply karaoke like effect:
5781 stereotools=mlev=0.015625
5785 Convert M/S signal to L/R:
5787 "stereotools=mode=ms>lr"
5791 @section stereowiden
5793 This filter enhance the stereo effect by suppressing signal common to both
5794 channels and by delaying the signal of left into right and vice versa,
5795 thereby widening the stereo effect.
5797 The filter accepts the following options:
5801 Time in milliseconds of the delay of left signal into right and vice versa.
5802 Default is 20 milliseconds.
5805 Amount of gain in delayed signal into right and vice versa. Gives a delay
5806 effect of left signal in right output and vice versa which gives widening
5807 effect. Default is 0.3.
5810 Cross feed of left into right with inverted phase. This helps in suppressing
5811 the mono. If the value is 1 it will cancel all the signal common to both
5812 channels. Default is 0.3.
5815 Set level of input signal of original channel. Default is 0.8.
5818 @subsection Commands
5820 This filter supports the all above options except @code{delay} as @ref{commands}.
5822 @section superequalizer
5823 Apply 18 band equalizer.
5825 The filter accepts the following options:
5832 Set 131Hz band gain.
5834 Set 185Hz band gain.
5836 Set 262Hz band gain.
5838 Set 370Hz band gain.
5840 Set 523Hz band gain.
5842 Set 740Hz band gain.
5844 Set 1047Hz band gain.
5846 Set 1480Hz band gain.
5848 Set 2093Hz band gain.
5850 Set 2960Hz band gain.
5852 Set 4186Hz band gain.
5854 Set 5920Hz band gain.
5856 Set 8372Hz band gain.
5858 Set 11840Hz band gain.
5860 Set 16744Hz band gain.
5862 Set 20000Hz band gain.
5866 Apply audio surround upmix filter.
5868 This filter allows to produce multichannel output from audio stream.
5870 The filter accepts the following options:
5874 Set output channel layout. By default, this is @var{5.1}.
5876 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5877 for the required syntax.
5880 Set input channel layout. By default, this is @var{stereo}.
5882 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5883 for the required syntax.
5886 Set input volume level. By default, this is @var{1}.
5889 Set output volume level. By default, this is @var{1}.
5892 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5895 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5898 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5901 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5902 In @var{add} mode, LFE channel is created from input audio and added to output.
5903 In @var{sub} mode, LFE channel is created from input audio and added to output but
5904 also all non-LFE output channels are subtracted with output LFE channel.
5907 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5908 Default is @var{90}.
5911 Set front center input volume. By default, this is @var{1}.
5914 Set front center output volume. By default, this is @var{1}.
5917 Set front left input volume. By default, this is @var{1}.
5920 Set front left output volume. By default, this is @var{1}.
5923 Set front right input volume. By default, this is @var{1}.
5926 Set front right output volume. By default, this is @var{1}.
5929 Set side left input volume. By default, this is @var{1}.
5932 Set side left output volume. By default, this is @var{1}.
5935 Set side right input volume. By default, this is @var{1}.
5938 Set side right output volume. By default, this is @var{1}.
5941 Set back left input volume. By default, this is @var{1}.
5944 Set back left output volume. By default, this is @var{1}.
5947 Set back right input volume. By default, this is @var{1}.
5950 Set back right output volume. By default, this is @var{1}.
5953 Set back center input volume. By default, this is @var{1}.
5956 Set back center output volume. By default, this is @var{1}.
5959 Set LFE input volume. By default, this is @var{1}.
5962 Set LFE output volume. By default, this is @var{1}.
5965 Set spread usage of stereo image across X axis for all channels.
5968 Set spread usage of stereo image across Y axis for all channels.
5970 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5971 Set spread usage of stereo image across X axis for each channel.
5973 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5974 Set spread usage of stereo image across Y axis for each channel.
5977 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5980 Set window function.
5982 It accepts the following values:
6005 Default is @code{hann}.
6008 Set window overlap. If set to 1, the recommended overlap for selected
6009 window function will be picked. Default is @code{0.5}.
6012 @section treble, highshelf
6014 Boost or cut treble (upper) frequencies of the audio using a two-pole
6015 shelving filter with a response similar to that of a standard
6016 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6018 The filter accepts the following options:
6022 Give the gain at whichever is the lower of ~22 kHz and the
6023 Nyquist frequency. Its useful range is about -20 (for a large cut)
6024 to +20 (for a large boost). Beware of clipping when using a positive gain.
6027 Set the filter's central frequency and so can be used
6028 to extend or reduce the frequency range to be boosted or cut.
6029 The default value is @code{3000} Hz.
6032 Set method to specify band-width of filter.
6047 Determine how steep is the filter's shelf transition.
6050 How much to use filtered signal in output. Default is 1.
6051 Range is between 0 and 1.
6054 Specify which channels to filter, by default all available are filtered.
6057 Normalize biquad coefficients, by default is disabled.
6058 Enabling it will normalize magnitude response at DC to 0dB.
6061 Set transform type of IIR filter.
6070 Set precison of filtering.
6073 Pick automatic sample format depending on surround filters.
6075 Always use signed 16-bit.
6077 Always use signed 32-bit.
6079 Always use float 32-bit.
6081 Always use float 64-bit.
6085 @subsection Commands
6087 This filter supports the following commands:
6090 Change treble frequency.
6091 Syntax for the command is : "@var{frequency}"
6094 Change treble width_type.
6095 Syntax for the command is : "@var{width_type}"
6098 Change treble width.
6099 Syntax for the command is : "@var{width}"
6103 Syntax for the command is : "@var{gain}"
6107 Syntax for the command is : "@var{mix}"
6112 Sinusoidal amplitude modulation.
6114 The filter accepts the following options:
6118 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6119 (20 Hz or lower) will result in a tremolo effect.
6120 This filter may also be used as a ring modulator by specifying
6121 a modulation frequency higher than 20 Hz.
6122 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6125 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6126 Default value is 0.5.
6131 Sinusoidal phase modulation.
6133 The filter accepts the following options:
6137 Modulation frequency in Hertz.
6138 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6141 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6142 Default value is 0.5.
6147 Adjust the input audio volume.
6149 It accepts the following parameters:
6153 Set audio volume expression.
6155 Output values are clipped to the maximum value.
6157 The output audio volume is given by the relation:
6159 @var{output_volume} = @var{volume} * @var{input_volume}
6162 The default value for @var{volume} is "1.0".
6165 This parameter represents the mathematical precision.
6167 It determines which input sample formats will be allowed, which affects the
6168 precision of the volume scaling.
6172 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6174 32-bit floating-point; this limits input sample format to FLT. (default)
6176 64-bit floating-point; this limits input sample format to DBL.
6180 Choose the behaviour on encountering ReplayGain side data in input frames.
6184 Remove ReplayGain side data, ignoring its contents (the default).
6187 Ignore ReplayGain side data, but leave it in the frame.
6190 Prefer the track gain, if present.
6193 Prefer the album gain, if present.
6196 @item replaygain_preamp
6197 Pre-amplification gain in dB to apply to the selected replaygain gain.
6199 Default value for @var{replaygain_preamp} is 0.0.
6201 @item replaygain_noclip
6202 Prevent clipping by limiting the gain applied.
6204 Default value for @var{replaygain_noclip} is 1.
6207 Set when the volume expression is evaluated.
6209 It accepts the following values:
6212 only evaluate expression once during the filter initialization, or
6213 when the @samp{volume} command is sent
6216 evaluate expression for each incoming frame
6219 Default value is @samp{once}.
6222 The volume expression can contain the following parameters.
6226 frame number (starting at zero)
6229 @item nb_consumed_samples
6230 number of samples consumed by the filter
6232 number of samples in the current frame
6234 original frame position in the file
6240 PTS at start of stream
6242 time at start of stream
6248 last set volume value
6251 Note that when @option{eval} is set to @samp{once} only the
6252 @var{sample_rate} and @var{tb} variables are available, all other
6253 variables will evaluate to NAN.
6255 @subsection Commands
6257 This filter supports the following commands:
6260 Modify the volume expression.
6261 The command accepts the same syntax of the corresponding option.
6263 If the specified expression is not valid, it is kept at its current
6267 @subsection Examples
6271 Halve the input audio volume:
6275 volume=volume=-6.0206dB
6278 In all the above example the named key for @option{volume} can be
6279 omitted, for example like in:
6285 Increase input audio power by 6 decibels using fixed-point precision:
6287 volume=volume=6dB:precision=fixed
6291 Fade volume after time 10 with an annihilation period of 5 seconds:
6293 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6297 @section volumedetect
6299 Detect the volume of the input video.
6301 The filter has no parameters. The input is not modified. Statistics about
6302 the volume will be printed in the log when the input stream end is reached.
6304 In particular it will show the mean volume (root mean square), maximum
6305 volume (on a per-sample basis), and the beginning of a histogram of the
6306 registered volume values (from the maximum value to a cumulated 1/1000 of
6309 All volumes are in decibels relative to the maximum PCM value.
6311 @subsection Examples
6313 Here is an excerpt of the output:
6315 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6316 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6317 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6318 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6319 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6320 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6321 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6322 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6323 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6329 The mean square energy is approximately -27 dB, or 10^-2.7.
6331 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6333 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6336 In other words, raising the volume by +4 dB does not cause any clipping,
6337 raising it by +5 dB causes clipping for 6 samples, etc.
6339 @c man end AUDIO FILTERS
6341 @chapter Audio Sources
6342 @c man begin AUDIO SOURCES
6344 Below is a description of the currently available audio sources.
6348 Buffer audio frames, and make them available to the filter chain.
6350 This source is mainly intended for a programmatic use, in particular
6351 through the interface defined in @file{libavfilter/buffersrc.h}.
6353 It accepts the following parameters:
6357 The timebase which will be used for timestamps of submitted frames. It must be
6358 either a floating-point number or in @var{numerator}/@var{denominator} form.
6361 The sample rate of the incoming audio buffers.
6364 The sample format of the incoming audio buffers.
6365 Either a sample format name or its corresponding integer representation from
6366 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6368 @item channel_layout
6369 The channel layout of the incoming audio buffers.
6370 Either a channel layout name from channel_layout_map in
6371 @file{libavutil/channel_layout.c} or its corresponding integer representation
6372 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6375 The number of channels of the incoming audio buffers.
6376 If both @var{channels} and @var{channel_layout} are specified, then they
6381 @subsection Examples
6384 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6387 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6388 Since the sample format with name "s16p" corresponds to the number
6389 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6392 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6397 Generate an audio signal specified by an expression.
6399 This source accepts in input one or more expressions (one for each
6400 channel), which are evaluated and used to generate a corresponding
6403 This source accepts the following options:
6407 Set the '|'-separated expressions list for each separate channel. In case the
6408 @option{channel_layout} option is not specified, the selected channel layout
6409 depends on the number of provided expressions. Otherwise the last
6410 specified expression is applied to the remaining output channels.
6412 @item channel_layout, c
6413 Set the channel layout. The number of channels in the specified layout
6414 must be equal to the number of specified expressions.
6417 Set the minimum duration of the sourced audio. See
6418 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6419 for the accepted syntax.
6420 Note that the resulting duration may be greater than the specified
6421 duration, as the generated audio is always cut at the end of a
6424 If not specified, or the expressed duration is negative, the audio is
6425 supposed to be generated forever.
6428 Set the number of samples per channel per each output frame,
6431 @item sample_rate, s
6432 Specify the sample rate, default to 44100.
6435 Each expression in @var{exprs} can contain the following constants:
6439 number of the evaluated sample, starting from 0
6442 time of the evaluated sample expressed in seconds, starting from 0
6449 @subsection Examples
6459 Generate a sin signal with frequency of 440 Hz, set sample rate to
6462 aevalsrc="sin(440*2*PI*t):s=8000"
6466 Generate a two channels signal, specify the channel layout (Front
6467 Center + Back Center) explicitly:
6469 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6473 Generate white noise:
6475 aevalsrc="-2+random(0)"
6479 Generate an amplitude modulated signal:
6481 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6485 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6487 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6494 Generate a FIR coefficients using frequency sampling method.
6496 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6498 The filter accepts the following options:
6502 Set number of filter coefficents in output audio stream.
6503 Default value is 1025.
6506 Set frequency points from where magnitude and phase are set.
6507 This must be in non decreasing order, and first element must be 0, while last element
6508 must be 1. Elements are separated by white spaces.
6511 Set magnitude value for every frequency point set by @option{frequency}.
6512 Number of values must be same as number of frequency points.
6513 Values are separated by white spaces.
6516 Set phase value for every frequency point set by @option{frequency}.
6517 Number of values must be same as number of frequency points.
6518 Values are separated by white spaces.
6520 @item sample_rate, r
6521 Set sample rate, default is 44100.
6524 Set number of samples per each frame. Default is 1024.
6527 Set window function. Default is blackman.
6532 The null audio source, return unprocessed audio frames. It is mainly useful
6533 as a template and to be employed in analysis / debugging tools, or as
6534 the source for filters which ignore the input data (for example the sox
6537 This source accepts the following options:
6541 @item channel_layout, cl
6543 Specifies the channel layout, and can be either an integer or a string
6544 representing a channel layout. The default value of @var{channel_layout}
6547 Check the channel_layout_map definition in
6548 @file{libavutil/channel_layout.c} for the mapping between strings and
6549 channel layout values.
6551 @item sample_rate, r
6552 Specifies the sample rate, and defaults to 44100.
6555 Set the number of samples per requested frames.
6558 Set the duration of the sourced audio. See
6559 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6560 for the accepted syntax.
6562 If not specified, or the expressed duration is negative, the audio is
6563 supposed to be generated forever.
6566 @subsection Examples
6570 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6572 anullsrc=r=48000:cl=4
6576 Do the same operation with a more obvious syntax:
6578 anullsrc=r=48000:cl=mono
6582 All the parameters need to be explicitly defined.
6586 Synthesize a voice utterance using the libflite library.
6588 To enable compilation of this filter you need to configure FFmpeg with
6589 @code{--enable-libflite}.
6591 Note that versions of the flite library prior to 2.0 are not thread-safe.
6593 The filter accepts the following options:
6598 If set to 1, list the names of the available voices and exit
6599 immediately. Default value is 0.
6602 Set the maximum number of samples per frame. Default value is 512.
6605 Set the filename containing the text to speak.
6608 Set the text to speak.
6611 Set the voice to use for the speech synthesis. Default value is
6612 @code{kal}. See also the @var{list_voices} option.
6615 @subsection Examples
6619 Read from file @file{speech.txt}, and synthesize the text using the
6620 standard flite voice:
6622 flite=textfile=speech.txt
6626 Read the specified text selecting the @code{slt} voice:
6628 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6632 Input text to ffmpeg:
6634 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6638 Make @file{ffplay} speak the specified text, using @code{flite} and
6639 the @code{lavfi} device:
6641 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6645 For more information about libflite, check:
6646 @url{http://www.festvox.org/flite/}
6650 Generate a noise audio signal.
6652 The filter accepts the following options:
6655 @item sample_rate, r
6656 Specify the sample rate. Default value is 48000 Hz.
6659 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6663 Specify the duration of the generated audio stream. Not specifying this option
6664 results in noise with an infinite length.
6666 @item color, colour, c
6667 Specify the color of noise. Available noise colors are white, pink, brown,
6668 blue, violet and velvet. Default color is white.
6671 Specify a value used to seed the PRNG.
6674 Set the number of samples per each output frame, default is 1024.
6677 @subsection Examples
6682 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6684 anoisesrc=d=60:c=pink:r=44100:a=0.5
6690 Generate odd-tap Hilbert transform FIR coefficients.
6692 The resulting stream can be used with @ref{afir} filter for phase-shifting
6693 the signal by 90 degrees.
6695 This is used in many matrix coding schemes and for analytic signal generation.
6696 The process is often written as a multiplication by i (or j), the imaginary unit.
6698 The filter accepts the following options:
6702 @item sample_rate, s
6703 Set sample rate, default is 44100.
6706 Set length of FIR filter, default is 22051.
6709 Set number of samples per each frame.
6712 Set window function to be used when generating FIR coefficients.
6717 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6719 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6721 The filter accepts the following options:
6724 @item sample_rate, r
6725 Set sample rate, default is 44100.
6728 Set number of samples per each frame. Default is 1024.
6731 Set high-pass frequency. Default is 0.
6734 Set low-pass frequency. Default is 0.
6735 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6736 is higher than 0 then filter will create band-pass filter coefficients,
6737 otherwise band-reject filter coefficients.
6740 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6743 Set Kaiser window beta.
6746 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6749 Enable rounding, by default is disabled.
6752 Set number of taps for high-pass filter.
6755 Set number of taps for low-pass filter.
6760 Generate an audio signal made of a sine wave with amplitude 1/8.
6762 The audio signal is bit-exact.
6764 The filter accepts the following options:
6769 Set the carrier frequency. Default is 440 Hz.
6771 @item beep_factor, b
6772 Enable a periodic beep every second with frequency @var{beep_factor} times
6773 the carrier frequency. Default is 0, meaning the beep is disabled.
6775 @item sample_rate, r
6776 Specify the sample rate, default is 44100.
6779 Specify the duration of the generated audio stream.
6781 @item samples_per_frame
6782 Set the number of samples per output frame.
6784 The expression can contain the following constants:
6788 The (sequential) number of the output audio frame, starting from 0.
6791 The PTS (Presentation TimeStamp) of the output audio frame,
6792 expressed in @var{TB} units.
6795 The PTS of the output audio frame, expressed in seconds.
6798 The timebase of the output audio frames.
6801 Default is @code{1024}.
6804 @subsection Examples
6809 Generate a simple 440 Hz sine wave:
6815 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6819 sine=frequency=220:beep_factor=4:duration=5
6823 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6826 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6830 @c man end AUDIO SOURCES
6832 @chapter Audio Sinks
6833 @c man begin AUDIO SINKS
6835 Below is a description of the currently available audio sinks.
6837 @section abuffersink
6839 Buffer audio frames, and make them available to the end of filter chain.
6841 This sink is mainly intended for programmatic use, in particular
6842 through the interface defined in @file{libavfilter/buffersink.h}
6843 or the options system.
6845 It accepts a pointer to an AVABufferSinkContext structure, which
6846 defines the incoming buffers' formats, to be passed as the opaque
6847 parameter to @code{avfilter_init_filter} for initialization.
6850 Null audio sink; do absolutely nothing with the input audio. It is
6851 mainly useful as a template and for use in analysis / debugging
6854 @c man end AUDIO SINKS
6856 @chapter Video Filters
6857 @c man begin VIDEO FILTERS
6859 When you configure your FFmpeg build, you can disable any of the
6860 existing filters using @code{--disable-filters}.
6861 The configure output will show the video filters included in your
6864 Below is a description of the currently available video filters.
6868 Mark a region of interest in a video frame.
6870 The frame data is passed through unchanged, but metadata is attached
6871 to the frame indicating regions of interest which can affect the
6872 behaviour of later encoding. Multiple regions can be marked by
6873 applying the filter multiple times.
6877 Region distance in pixels from the left edge of the frame.
6879 Region distance in pixels from the top edge of the frame.
6881 Region width in pixels.
6883 Region height in pixels.
6885 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6886 and may contain the following variables:
6889 Width of the input frame.
6891 Height of the input frame.
6895 Quantisation offset to apply within the region.
6897 This must be a real value in the range -1 to +1. A value of zero
6898 indicates no quality change. A negative value asks for better quality
6899 (less quantisation), while a positive value asks for worse quality
6900 (greater quantisation).
6902 The range is calibrated so that the extreme values indicate the
6903 largest possible offset - if the rest of the frame is encoded with the
6904 worst possible quality, an offset of -1 indicates that this region
6905 should be encoded with the best possible quality anyway. Intermediate
6906 values are then interpolated in some codec-dependent way.
6908 For example, in 10-bit H.264 the quantisation parameter varies between
6909 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6910 this region should be encoded with a QP around one-tenth of the full
6911 range better than the rest of the frame. So, if most of the frame
6912 were to be encoded with a QP of around 30, this region would get a QP
6913 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6914 An extreme value of -1 would indicate that this region should be
6915 encoded with the best possible quality regardless of the treatment of
6916 the rest of the frame - that is, should be encoded at a QP of -12.
6918 If set to true, remove any existing regions of interest marked on the
6919 frame before adding the new one.
6922 @subsection Examples
6926 Mark the centre quarter of the frame as interesting.
6928 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6931 Mark the 100-pixel-wide region on the left edge of the frame as very
6932 uninteresting (to be encoded at much lower quality than the rest of
6935 addroi=0:0:100:ih:+1/5
6939 @section alphaextract
6941 Extract the alpha component from the input as a grayscale video. This
6942 is especially useful with the @var{alphamerge} filter.
6946 Add or replace the alpha component of the primary input with the
6947 grayscale value of a second input. This is intended for use with
6948 @var{alphaextract} to allow the transmission or storage of frame
6949 sequences that have alpha in a format that doesn't support an alpha
6952 For example, to reconstruct full frames from a normal YUV-encoded video
6953 and a separate video created with @var{alphaextract}, you might use:
6955 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6960 Amplify differences between current pixel and pixels of adjacent frames in
6961 same pixel location.
6963 This filter accepts the following options:
6967 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6968 For example radius of 3 will instruct filter to calculate average of 7 frames.
6971 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6974 Set threshold for difference amplification. Any difference greater or equal to
6975 this value will not alter source pixel. Default is 10.
6976 Allowed range is from 0 to 65535.
6979 Set tolerance for difference amplification. Any difference lower to
6980 this value will not alter source pixel. Default is 0.
6981 Allowed range is from 0 to 65535.
6984 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6985 This option controls maximum possible value that will decrease source pixel value.
6988 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6989 This option controls maximum possible value that will increase source pixel value.
6992 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6995 @subsection Commands
6997 This filter supports the following @ref{commands} that corresponds to option of same name:
7009 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7010 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7011 Substation Alpha) subtitles files.
7013 This filter accepts the following option in addition to the common options from
7014 the @ref{subtitles} filter:
7018 Set the shaping engine
7020 Available values are:
7023 The default libass shaping engine, which is the best available.
7025 Fast, font-agnostic shaper that can do only substitutions
7027 Slower shaper using OpenType for substitutions and positioning
7030 The default is @code{auto}.
7034 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7036 The filter accepts the following options:
7040 Set threshold A for 1st plane. Default is 0.02.
7041 Valid range is 0 to 0.3.
7044 Set threshold B for 1st plane. Default is 0.04.
7045 Valid range is 0 to 5.
7048 Set threshold A for 2nd plane. Default is 0.02.
7049 Valid range is 0 to 0.3.
7052 Set threshold B for 2nd plane. Default is 0.04.
7053 Valid range is 0 to 5.
7056 Set threshold A for 3rd plane. Default is 0.02.
7057 Valid range is 0 to 0.3.
7060 Set threshold B for 3rd plane. Default is 0.04.
7061 Valid range is 0 to 5.
7063 Threshold A is designed to react on abrupt changes in the input signal and
7064 threshold B is designed to react on continuous changes in the input signal.
7067 Set number of frames filter will use for averaging. Default is 9. Must be odd
7068 number in range [5, 129].
7071 Set what planes of frame filter will use for averaging. Default is all.
7074 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7075 Alternatively can be set to @code{s} serial.
7077 Parallel can be faster then serial, while other way around is never true.
7078 Parallel will abort early on first change being greater then thresholds, while serial
7079 will continue processing other side of frames if they are equal or below thresholds.
7082 @subsection Commands
7083 This filter supports same @ref{commands} as options except option @code{s}.
7084 The command accepts the same syntax of the corresponding option.
7088 Apply average blur filter.
7090 The filter accepts the following options:
7094 Set horizontal radius size.
7097 Set which planes to filter. By default all planes are filtered.
7100 Set vertical radius size, if zero it will be same as @code{sizeX}.
7101 Default is @code{0}.
7104 @subsection Commands
7105 This filter supports same commands as options.
7106 The command accepts the same syntax of the corresponding option.
7108 If the specified expression is not valid, it is kept at its current
7113 Compute the bounding box for the non-black pixels in the input frame
7116 This filter computes the bounding box containing all the pixels with a
7117 luminance value greater than the minimum allowed value.
7118 The parameters describing the bounding box are printed on the filter
7121 The filter accepts the following option:
7125 Set the minimal luminance value. Default is @code{16}.
7129 Apply bilateral filter, spatial smoothing while preserving edges.
7131 The filter accepts the following options:
7134 Set sigma of gaussian function to calculate spatial weight.
7135 Allowed range is 0 to 512. Default is 0.1.
7138 Set sigma of gaussian function to calculate range weight.
7139 Allowed range is 0 to 1. Default is 0.1.
7142 Set planes to filter. Default is first only.
7145 @section bitplanenoise
7147 Show and measure bit plane noise.
7149 The filter accepts the following options:
7153 Set which plane to analyze. Default is @code{1}.
7156 Filter out noisy pixels from @code{bitplane} set above.
7157 Default is disabled.
7160 @section blackdetect
7162 Detect video intervals that are (almost) completely black. Can be
7163 useful to detect chapter transitions, commercials, or invalid
7166 The filter outputs its detection analysis to both the log as well as
7167 frame metadata. If a black segment of at least the specified minimum
7168 duration is found, a line with the start and end timestamps as well
7169 as duration is printed to the log with level @code{info}. In addition,
7170 a log line with level @code{debug} is printed per frame showing the
7171 black amount detected for that frame.
7173 The filter also attaches metadata to the first frame of a black
7174 segment with key @code{lavfi.black_start} and to the first frame
7175 after the black segment ends with key @code{lavfi.black_end}. The
7176 value is the frame's timestamp. This metadata is added regardless
7177 of the minimum duration specified.
7179 The filter accepts the following options:
7182 @item black_min_duration, d
7183 Set the minimum detected black duration expressed in seconds. It must
7184 be a non-negative floating point number.
7186 Default value is 2.0.
7188 @item picture_black_ratio_th, pic_th
7189 Set the threshold for considering a picture "black".
7190 Express the minimum value for the ratio:
7192 @var{nb_black_pixels} / @var{nb_pixels}
7195 for which a picture is considered black.
7196 Default value is 0.98.
7198 @item pixel_black_th, pix_th
7199 Set the threshold for considering a pixel "black".
7201 The threshold expresses the maximum pixel luminance value for which a
7202 pixel is considered "black". The provided value is scaled according to
7203 the following equation:
7205 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7208 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7209 the input video format, the range is [0-255] for YUV full-range
7210 formats and [16-235] for YUV non full-range formats.
7212 Default value is 0.10.
7215 The following example sets the maximum pixel threshold to the minimum
7216 value, and detects only black intervals of 2 or more seconds:
7218 blackdetect=d=2:pix_th=0.00
7223 Detect frames that are (almost) completely black. Can be useful to
7224 detect chapter transitions or commercials. Output lines consist of
7225 the frame number of the detected frame, the percentage of blackness,
7226 the position in the file if known or -1 and the timestamp in seconds.
7228 In order to display the output lines, you need to set the loglevel at
7229 least to the AV_LOG_INFO value.
7231 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7232 The value represents the percentage of pixels in the picture that
7233 are below the threshold value.
7235 It accepts the following parameters:
7240 The percentage of the pixels that have to be below the threshold; it defaults to
7243 @item threshold, thresh
7244 The threshold below which a pixel value is considered black; it defaults to
7252 Blend two video frames into each other.
7254 The @code{blend} filter takes two input streams and outputs one
7255 stream, the first input is the "top" layer and second input is
7256 "bottom" layer. By default, the output terminates when the longest input terminates.
7258 The @code{tblend} (time blend) filter takes two consecutive frames
7259 from one single stream, and outputs the result obtained by blending
7260 the new frame on top of the old frame.
7262 A description of the accepted options follows.
7270 Set blend mode for specific pixel component or all pixel components in case
7271 of @var{all_mode}. Default value is @code{normal}.
7273 Available values for component modes are:
7315 Set blend opacity for specific pixel component or all pixel components in case
7316 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7323 Set blend expression for specific pixel component or all pixel components in case
7324 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7326 The expressions can use the following variables:
7330 The sequential number of the filtered frame, starting from @code{0}.
7334 the coordinates of the current sample
7338 the width and height of currently filtered plane
7342 Width and height scale for the plane being filtered. It is the
7343 ratio between the dimensions of the current plane to the luma plane,
7344 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7345 the luma plane and @code{0.5,0.5} for the chroma planes.
7348 Time of the current frame, expressed in seconds.
7351 Value of pixel component at current location for first video frame (top layer).
7354 Value of pixel component at current location for second video frame (bottom layer).
7358 The @code{blend} filter also supports the @ref{framesync} options.
7360 @subsection Examples
7364 Apply transition from bottom layer to top layer in first 10 seconds:
7366 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7370 Apply linear horizontal transition from top layer to bottom layer:
7372 blend=all_expr='A*(X/W)+B*(1-X/W)'
7376 Apply 1x1 checkerboard effect:
7378 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7382 Apply uncover left effect:
7384 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7388 Apply uncover down effect:
7390 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7394 Apply uncover up-left effect:
7396 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7400 Split diagonally video and shows top and bottom layer on each side:
7402 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7406 Display differences between the current and the previous frame:
7408 tblend=all_mode=grainextract
7414 Denoise frames using Block-Matching 3D algorithm.
7416 The filter accepts the following options.
7420 Set denoising strength. Default value is 1.
7421 Allowed range is from 0 to 999.9.
7422 The denoising algorithm is very sensitive to sigma, so adjust it
7423 according to the source.
7426 Set local patch size. This sets dimensions in 2D.
7429 Set sliding step for processing blocks. Default value is 4.
7430 Allowed range is from 1 to 64.
7431 Smaller values allows processing more reference blocks and is slower.
7434 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7435 When set to 1, no block matching is done. Larger values allows more blocks
7437 Allowed range is from 1 to 256.
7440 Set radius for search block matching. Default is 9.
7441 Allowed range is from 1 to INT32_MAX.
7444 Set step between two search locations for block matching. Default is 1.
7445 Allowed range is from 1 to 64. Smaller is slower.
7448 Set threshold of mean square error for block matching. Valid range is 0 to
7452 Set thresholding parameter for hard thresholding in 3D transformed domain.
7453 Larger values results in stronger hard-thresholding filtering in frequency
7457 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7458 Default is @code{basic}.
7461 If enabled, filter will use 2nd stream for block matching.
7462 Default is disabled for @code{basic} value of @var{estim} option,
7463 and always enabled if value of @var{estim} is @code{final}.
7466 Set planes to filter. Default is all available except alpha.
7469 @subsection Examples
7473 Basic filtering with bm3d:
7475 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7479 Same as above, but filtering only luma:
7481 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7485 Same as above, but with both estimation modes:
7487 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
7491 Same as above, but prefilter with @ref{nlmeans} filter instead:
7493 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
7499 Apply a boxblur algorithm to the input video.
7501 It accepts the following parameters:
7505 @item luma_radius, lr
7506 @item luma_power, lp
7507 @item chroma_radius, cr
7508 @item chroma_power, cp
7509 @item alpha_radius, ar
7510 @item alpha_power, ap
7514 A description of the accepted options follows.
7517 @item luma_radius, lr
7518 @item chroma_radius, cr
7519 @item alpha_radius, ar
7520 Set an expression for the box radius in pixels used for blurring the
7521 corresponding input plane.
7523 The radius value must be a non-negative number, and must not be
7524 greater than the value of the expression @code{min(w,h)/2} for the
7525 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7528 Default value for @option{luma_radius} is "2". If not specified,
7529 @option{chroma_radius} and @option{alpha_radius} default to the
7530 corresponding value set for @option{luma_radius}.
7532 The expressions can contain the following constants:
7536 The input width and height in pixels.
7540 The input chroma image width and height in pixels.
7544 The horizontal and vertical chroma subsample values. For example, for the
7545 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7548 @item luma_power, lp
7549 @item chroma_power, cp
7550 @item alpha_power, ap
7551 Specify how many times the boxblur filter is applied to the
7552 corresponding plane.
7554 Default value for @option{luma_power} is 2. If not specified,
7555 @option{chroma_power} and @option{alpha_power} default to the
7556 corresponding value set for @option{luma_power}.
7558 A value of 0 will disable the effect.
7561 @subsection Examples
7565 Apply a boxblur filter with the luma, chroma, and alpha radii
7568 boxblur=luma_radius=2:luma_power=1
7573 Set the luma radius to 2, and alpha and chroma radius to 0:
7575 boxblur=2:1:cr=0:ar=0
7579 Set the luma and chroma radii to a fraction of the video dimension:
7581 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7587 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7588 Deinterlacing Filter").
7590 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7591 interpolation algorithms.
7592 It accepts the following parameters:
7596 The interlacing mode to adopt. It accepts one of the following values:
7600 Output one frame for each frame.
7602 Output one frame for each field.
7605 The default value is @code{send_field}.
7608 The picture field parity assumed for the input interlaced video. It accepts one
7609 of the following values:
7613 Assume the top field is first.
7615 Assume the bottom field is first.
7617 Enable automatic detection of field parity.
7620 The default value is @code{auto}.
7621 If the interlacing is unknown or the decoder does not export this information,
7622 top field first will be assumed.
7625 Specify which frames to deinterlace. Accepts one of the following
7630 Deinterlace all frames.
7632 Only deinterlace frames marked as interlaced.
7635 The default value is @code{all}.
7640 Apply Contrast Adaptive Sharpen filter to video stream.
7642 The filter accepts the following options:
7646 Set the sharpening strength. Default value is 0.
7649 Set planes to filter. Default value is to filter all
7650 planes except alpha plane.
7654 Remove all color information for all colors except for certain one.
7656 The filter accepts the following options:
7660 The color which will not be replaced with neutral chroma.
7663 Similarity percentage with the above color.
7664 0.01 matches only the exact key color, while 1.0 matches everything.
7668 0.0 makes pixels either fully gray, or not gray at all.
7669 Higher values result in more preserved color.
7672 Signals that the color passed is already in YUV instead of RGB.
7674 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7675 This can be used to pass exact YUV values as hexadecimal numbers.
7678 @subsection Commands
7679 This filter supports same @ref{commands} as options.
7680 The command accepts the same syntax of the corresponding option.
7682 If the specified expression is not valid, it is kept at its current
7686 YUV colorspace color/chroma keying.
7688 The filter accepts the following options:
7692 The color which will be replaced with transparency.
7695 Similarity percentage with the key color.
7697 0.01 matches only the exact key color, while 1.0 matches everything.
7702 0.0 makes pixels either fully transparent, or not transparent at all.
7704 Higher values result in semi-transparent pixels, with a higher transparency
7705 the more similar the pixels color is to the key color.
7708 Signals that the color passed is already in YUV instead of RGB.
7710 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7711 This can be used to pass exact YUV values as hexadecimal numbers.
7714 @subsection Commands
7715 This filter supports same @ref{commands} as options.
7716 The command accepts the same syntax of the corresponding option.
7718 If the specified expression is not valid, it is kept at its current
7721 @subsection Examples
7725 Make every green pixel in the input image transparent:
7727 ffmpeg -i input.png -vf chromakey=green out.png
7731 Overlay a greenscreen-video on top of a static black background.
7733 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
7738 Reduce chrominance noise.
7740 The filter accepts the following options:
7744 Set threshold for averaging chrominance values.
7745 Sum of absolute difference of Y, U and V pixel components of current
7746 pixel and neighbour pixels lower than this threshold will be used in
7747 averaging. Luma component is left unchanged and is copied to output.
7748 Default value is 30. Allowed range is from 1 to 200.
7751 Set horizontal radius of rectangle used for averaging.
7752 Allowed range is from 1 to 100. Default value is 5.
7755 Set vertical radius of rectangle used for averaging.
7756 Allowed range is from 1 to 100. Default value is 5.
7759 Set horizontal step when averaging. Default value is 1.
7760 Allowed range is from 1 to 50.
7761 Mostly useful to speed-up filtering.
7764 Set vertical step when averaging. Default value is 1.
7765 Allowed range is from 1 to 50.
7766 Mostly useful to speed-up filtering.
7769 Set Y threshold for averaging chrominance values.
7770 Set finer control for max allowed difference between Y components
7771 of current pixel and neigbour pixels.
7772 Default value is 200. Allowed range is from 1 to 200.
7775 Set U threshold for averaging chrominance values.
7776 Set finer control for max allowed difference between U components
7777 of current pixel and neigbour pixels.
7778 Default value is 200. Allowed range is from 1 to 200.
7781 Set V threshold for averaging chrominance values.
7782 Set finer control for max allowed difference between V components
7783 of current pixel and neigbour pixels.
7784 Default value is 200. Allowed range is from 1 to 200.
7787 @subsection Commands
7788 This filter supports same @ref{commands} as options.
7789 The command accepts the same syntax of the corresponding option.
7791 @section chromashift
7792 Shift chroma pixels horizontally and/or vertically.
7794 The filter accepts the following options:
7797 Set amount to shift chroma-blue horizontally.
7799 Set amount to shift chroma-blue vertically.
7801 Set amount to shift chroma-red horizontally.
7803 Set amount to shift chroma-red vertically.
7805 Set edge mode, can be @var{smear}, default, or @var{warp}.
7808 @subsection Commands
7810 This filter supports the all above options as @ref{commands}.
7814 Display CIE color diagram with pixels overlaid onto it.
7816 The filter accepts the following options:
7831 @item uhdtv, rec2020
7845 Set what gamuts to draw.
7847 See @code{system} option for available values.
7850 Set ciescope size, by default set to 512.
7853 Set intensity used to map input pixel values to CIE diagram.
7856 Set contrast used to draw tongue colors that are out of active color system gamut.
7859 Correct gamma displayed on scope, by default enabled.
7862 Show white point on CIE diagram, by default disabled.
7865 Set input gamma. Used only with XYZ input color space.
7870 Visualize information exported by some codecs.
7872 Some codecs can export information through frames using side-data or other
7873 means. For example, some MPEG based codecs export motion vectors through the
7874 @var{export_mvs} flag in the codec @option{flags2} option.
7876 The filter accepts the following option:
7880 Set motion vectors to visualize.
7882 Available flags for @var{mv} are:
7886 forward predicted MVs of P-frames
7888 forward predicted MVs of B-frames
7890 backward predicted MVs of B-frames
7894 Display quantization parameters using the chroma planes.
7897 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7899 Available flags for @var{mv_type} are:
7903 forward predicted MVs
7905 backward predicted MVs
7908 @item frame_type, ft
7909 Set frame type to visualize motion vectors of.
7911 Available flags for @var{frame_type} are:
7915 intra-coded frames (I-frames)
7917 predicted frames (P-frames)
7919 bi-directionally predicted frames (B-frames)
7923 @subsection Examples
7927 Visualize forward predicted MVs of all frames using @command{ffplay}:
7929 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7933 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7935 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7939 @section colorbalance
7940 Modify intensity of primary colors (red, green and blue) of input frames.
7942 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7943 regions for the red-cyan, green-magenta or blue-yellow balance.
7945 A positive adjustment value shifts the balance towards the primary color, a negative
7946 value towards the complementary color.
7948 The filter accepts the following options:
7954 Adjust red, green and blue shadows (darkest pixels).
7959 Adjust red, green and blue midtones (medium pixels).
7964 Adjust red, green and blue highlights (brightest pixels).
7966 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7969 Preserve lightness when changing color balance. Default is disabled.
7972 @subsection Examples
7976 Add red color cast to shadows:
7982 @subsection Commands
7984 This filter supports the all above options as @ref{commands}.
7986 @section colorchannelmixer
7988 Adjust video input frames by re-mixing color channels.
7990 This filter modifies a color channel by adding the values associated to
7991 the other channels of the same pixels. For example if the value to
7992 modify is red, the output value will be:
7994 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7997 The filter accepts the following options:
8004 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8005 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8011 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8012 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8018 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8019 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8025 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8026 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8028 Allowed ranges for options are @code{[-2.0, 2.0]}.
8031 @subsection Examples
8035 Convert source to grayscale:
8037 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8040 Simulate sepia tones:
8042 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8046 @subsection Commands
8048 This filter supports the all above options as @ref{commands}.
8051 RGB colorspace color keying.
8053 The filter accepts the following options:
8057 The color which will be replaced with transparency.
8060 Similarity percentage with the key color.
8062 0.01 matches only the exact key color, while 1.0 matches everything.
8067 0.0 makes pixels either fully transparent, or not transparent at all.
8069 Higher values result in semi-transparent pixels, with a higher transparency
8070 the more similar the pixels color is to the key color.
8073 @subsection Examples
8077 Make every green pixel in the input image transparent:
8079 ffmpeg -i input.png -vf colorkey=green out.png
8083 Overlay a greenscreen-video on top of a static background image.
8085 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
8089 @subsection Commands
8090 This filter supports same @ref{commands} as options.
8091 The command accepts the same syntax of the corresponding option.
8093 If the specified expression is not valid, it is kept at its current
8097 Remove all color information for all RGB colors except for certain one.
8099 The filter accepts the following options:
8103 The color which will not be replaced with neutral gray.
8106 Similarity percentage with the above color.
8107 0.01 matches only the exact key color, while 1.0 matches everything.
8110 Blend percentage. 0.0 makes pixels fully gray.
8111 Higher values result in more preserved color.
8114 @subsection Commands
8115 This filter supports same @ref{commands} as options.
8116 The command accepts the same syntax of the corresponding option.
8118 If the specified expression is not valid, it is kept at its current
8121 @section colorlevels
8123 Adjust video input frames using levels.
8125 The filter accepts the following options:
8132 Adjust red, green, blue and alpha input black point.
8133 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8139 Adjust red, green, blue and alpha input white point.
8140 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8142 Input levels are used to lighten highlights (bright tones), darken shadows
8143 (dark tones), change the balance of bright and dark tones.
8149 Adjust red, green, blue and alpha output black point.
8150 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8156 Adjust red, green, blue and alpha output white point.
8157 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8159 Output levels allows manual selection of a constrained output level range.
8162 @subsection Examples
8166 Make video output darker:
8168 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8174 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8178 Make video output lighter:
8180 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8184 Increase brightness:
8186 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8190 @subsection Commands
8192 This filter supports the all above options as @ref{commands}.
8194 @section colormatrix
8196 Convert color matrix.
8198 The filter accepts the following options:
8203 Specify the source and destination color matrix. Both values must be
8206 The accepted values are:
8234 For example to convert from BT.601 to SMPTE-240M, use the command:
8236 colormatrix=bt601:smpte240m
8241 Convert colorspace, transfer characteristics or color primaries.
8242 Input video needs to have an even size.
8244 The filter accepts the following options:
8249 Specify all color properties at once.
8251 The accepted values are:
8281 Specify output colorspace.
8283 The accepted values are:
8292 BT.470BG or BT.601-6 625
8295 SMPTE-170M or BT.601-6 525
8304 BT.2020 with non-constant luminance
8310 Specify output transfer characteristics.
8312 The accepted values are:
8324 Constant gamma of 2.2
8327 Constant gamma of 2.8
8330 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8348 BT.2020 for 10-bits content
8351 BT.2020 for 12-bits content
8357 Specify output color primaries.
8359 The accepted values are:
8368 BT.470BG or BT.601-6 625
8371 SMPTE-170M or BT.601-6 525
8395 Specify output color range.
8397 The accepted values are:
8400 TV (restricted) range
8403 MPEG (restricted) range
8414 Specify output color format.
8416 The accepted values are:
8419 YUV 4:2:0 planar 8-bits
8422 YUV 4:2:0 planar 10-bits
8425 YUV 4:2:0 planar 12-bits
8428 YUV 4:2:2 planar 8-bits
8431 YUV 4:2:2 planar 10-bits
8434 YUV 4:2:2 planar 12-bits
8437 YUV 4:4:4 planar 8-bits
8440 YUV 4:4:4 planar 10-bits
8443 YUV 4:4:4 planar 12-bits
8448 Do a fast conversion, which skips gamma/primary correction. This will take
8449 significantly less CPU, but will be mathematically incorrect. To get output
8450 compatible with that produced by the colormatrix filter, use fast=1.
8453 Specify dithering mode.
8455 The accepted values are:
8461 Floyd-Steinberg dithering
8465 Whitepoint adaptation mode.
8467 The accepted values are:
8470 Bradford whitepoint adaptation
8473 von Kries whitepoint adaptation
8476 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8480 Override all input properties at once. Same accepted values as @ref{all}.
8483 Override input colorspace. Same accepted values as @ref{space}.
8486 Override input color primaries. Same accepted values as @ref{primaries}.
8489 Override input transfer characteristics. Same accepted values as @ref{trc}.
8492 Override input color range. Same accepted values as @ref{range}.
8496 The filter converts the transfer characteristics, color space and color
8497 primaries to the specified user values. The output value, if not specified,
8498 is set to a default value based on the "all" property. If that property is
8499 also not specified, the filter will log an error. The output color range and
8500 format default to the same value as the input color range and format. The
8501 input transfer characteristics, color space, color primaries and color range
8502 should be set on the input data. If any of these are missing, the filter will
8503 log an error and no conversion will take place.
8505 For example to convert the input to SMPTE-240M, use the command:
8507 colorspace=smpte240m
8510 @section convolution
8512 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8514 The filter accepts the following options:
8521 Set matrix for each plane.
8522 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8523 and from 1 to 49 odd number of signed integers in @var{row} mode.
8529 Set multiplier for calculated value for each plane.
8530 If unset or 0, it will be sum of all matrix elements.
8536 Set bias for each plane. This value is added to the result of the multiplication.
8537 Useful for making the overall image brighter or darker. Default is 0.0.
8543 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8544 Default is @var{square}.
8547 @subsection Commands
8549 This filter supports the all above options as @ref{commands}.
8551 @subsection Examples
8557 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"
8563 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"
8569 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"
8575 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"
8579 Apply laplacian edge detector which includes diagonals:
8581 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"
8587 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"
8593 Apply 2D convolution of video stream in frequency domain using second stream
8596 The filter accepts the following options:
8600 Set which planes to process.
8603 Set which impulse video frames will be processed, can be @var{first}
8604 or @var{all}. Default is @var{all}.
8607 The @code{convolve} filter also supports the @ref{framesync} options.
8611 Copy the input video source unchanged to the output. This is mainly useful for
8616 Video filtering on GPU using Apple's CoreImage API on OSX.
8618 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8619 processed by video hardware. However, software-based OpenGL implementations
8620 exist which means there is no guarantee for hardware processing. It depends on
8623 There are many filters and image generators provided by Apple that come with a
8624 large variety of options. The filter has to be referenced by its name along
8627 The coreimage filter accepts the following options:
8630 List all available filters and generators along with all their respective
8631 options as well as possible minimum and maximum values along with the default
8638 Specify all filters by their respective name and options.
8639 Use @var{list_filters} to determine all valid filter names and options.
8640 Numerical options are specified by a float value and are automatically clamped
8641 to their respective value range. Vector and color options have to be specified
8642 by a list of space separated float values. Character escaping has to be done.
8643 A special option name @code{default} is available to use default options for a
8646 It is required to specify either @code{default} or at least one of the filter options.
8647 All omitted options are used with their default values.
8648 The syntax of the filter string is as follows:
8650 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8654 Specify a rectangle where the output of the filter chain is copied into the
8655 input image. It is given by a list of space separated float values:
8657 output_rect=x\ y\ width\ height
8659 If not given, the output rectangle equals the dimensions of the input image.
8660 The output rectangle is automatically cropped at the borders of the input
8661 image. Negative values are valid for each component.
8663 output_rect=25\ 25\ 100\ 100
8667 Several filters can be chained for successive processing without GPU-HOST
8668 transfers allowing for fast processing of complex filter chains.
8669 Currently, only filters with zero (generators) or exactly one (filters) input
8670 image and one output image are supported. Also, transition filters are not yet
8673 Some filters generate output images with additional padding depending on the
8674 respective filter kernel. The padding is automatically removed to ensure the
8675 filter output has the same size as the input image.
8677 For image generators, the size of the output image is determined by the
8678 previous output image of the filter chain or the input image of the whole
8679 filterchain, respectively. The generators do not use the pixel information of
8680 this image to generate their output. However, the generated output is
8681 blended onto this image, resulting in partial or complete coverage of the
8684 The @ref{coreimagesrc} video source can be used for generating input images
8685 which are directly fed into the filter chain. By using it, providing input
8686 images by another video source or an input video is not required.
8688 @subsection Examples
8693 List all filters available:
8695 coreimage=list_filters=true
8699 Use the CIBoxBlur filter with default options to blur an image:
8701 coreimage=filter=CIBoxBlur@@default
8705 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8706 its center at 100x100 and a radius of 50 pixels:
8708 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8712 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8713 given as complete and escaped command-line for Apple's standard bash shell:
8715 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8721 Cover a rectangular object
8723 It accepts the following options:
8727 Filepath of the optional cover image, needs to be in yuv420.
8732 It accepts the following values:
8735 cover it by the supplied image
8737 cover it by interpolating the surrounding pixels
8740 Default value is @var{blur}.
8743 @subsection Examples
8747 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8749 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8755 Crop the input video to given dimensions.
8757 It accepts the following parameters:
8761 The width of the output video. It defaults to @code{iw}.
8762 This expression is evaluated only once during the filter
8763 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8766 The height of the output video. It defaults to @code{ih}.
8767 This expression is evaluated only once during the filter
8768 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8771 The horizontal position, in the input video, of the left edge of the output
8772 video. It defaults to @code{(in_w-out_w)/2}.
8773 This expression is evaluated per-frame.
8776 The vertical position, in the input video, of the top edge of the output video.
8777 It defaults to @code{(in_h-out_h)/2}.
8778 This expression is evaluated per-frame.
8781 If set to 1 will force the output display aspect ratio
8782 to be the same of the input, by changing the output sample aspect
8783 ratio. It defaults to 0.
8786 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8787 width/height/x/y as specified and will not be rounded to nearest smaller value.
8791 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8792 expressions containing the following constants:
8797 The computed values for @var{x} and @var{y}. They are evaluated for
8802 The input width and height.
8806 These are the same as @var{in_w} and @var{in_h}.
8810 The output (cropped) width and height.
8814 These are the same as @var{out_w} and @var{out_h}.
8817 same as @var{iw} / @var{ih}
8820 input sample aspect ratio
8823 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8827 horizontal and vertical chroma subsample values. For example for the
8828 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8831 The number of the input frame, starting from 0.
8834 the position in the file of the input frame, NAN if unknown
8837 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8841 The expression for @var{out_w} may depend on the value of @var{out_h},
8842 and the expression for @var{out_h} may depend on @var{out_w}, but they
8843 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8844 evaluated after @var{out_w} and @var{out_h}.
8846 The @var{x} and @var{y} parameters specify the expressions for the
8847 position of the top-left corner of the output (non-cropped) area. They
8848 are evaluated for each frame. If the evaluated value is not valid, it
8849 is approximated to the nearest valid value.
8851 The expression for @var{x} may depend on @var{y}, and the expression
8852 for @var{y} may depend on @var{x}.
8854 @subsection Examples
8858 Crop area with size 100x100 at position (12,34).
8863 Using named options, the example above becomes:
8865 crop=w=100:h=100:x=12:y=34
8869 Crop the central input area with size 100x100:
8875 Crop the central input area with size 2/3 of the input video:
8877 crop=2/3*in_w:2/3*in_h
8881 Crop the input video central square:
8888 Delimit the rectangle with the top-left corner placed at position
8889 100:100 and the right-bottom corner corresponding to the right-bottom
8890 corner of the input image.
8892 crop=in_w-100:in_h-100:100:100
8896 Crop 10 pixels from the left and right borders, and 20 pixels from
8897 the top and bottom borders
8899 crop=in_w-2*10:in_h-2*20
8903 Keep only the bottom right quarter of the input image:
8905 crop=in_w/2:in_h/2:in_w/2:in_h/2
8909 Crop height for getting Greek harmony:
8911 crop=in_w:1/PHI*in_w
8915 Apply trembling effect:
8917 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)
8921 Apply erratic camera effect depending on timestamp:
8923 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)"
8927 Set x depending on the value of y:
8929 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8933 @subsection Commands
8935 This filter supports the following commands:
8941 Set width/height of the output video and the horizontal/vertical position
8943 The command accepts the same syntax of the corresponding option.
8945 If the specified expression is not valid, it is kept at its current
8951 Auto-detect the crop size.
8953 It calculates the necessary cropping parameters and prints the
8954 recommended parameters via the logging system. The detected dimensions
8955 correspond to the non-black area of the input video.
8957 It accepts the following parameters:
8962 Set higher black value threshold, which can be optionally specified
8963 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8964 value greater to the set value is considered non-black. It defaults to 24.
8965 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8966 on the bitdepth of the pixel format.
8969 The value which the width/height should be divisible by. It defaults to
8970 16. The offset is automatically adjusted to center the video. Use 2 to
8971 get only even dimensions (needed for 4:2:2 video). 16 is best when
8972 encoding to most video codecs.
8975 Set the number of initial frames for which evaluation is skipped.
8976 Default is 2. Range is 0 to INT_MAX.
8978 @item reset_count, reset
8979 Set the counter that determines after how many frames cropdetect will
8980 reset the previously detected largest video area and start over to
8981 detect the current optimal crop area. Default value is 0.
8983 This can be useful when channel logos distort the video area. 0
8984 indicates 'never reset', and returns the largest area encountered during
8991 Delay video filtering until a given wallclock timestamp. The filter first
8992 passes on @option{preroll} amount of frames, then it buffers at most
8993 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8994 it forwards the buffered frames and also any subsequent frames coming in its
8997 The filter can be used synchronize the output of multiple ffmpeg processes for
8998 realtime output devices like decklink. By putting the delay in the filtering
8999 chain and pre-buffering frames the process can pass on data to output almost
9000 immediately after the target wallclock timestamp is reached.
9002 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9008 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9011 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9014 The maximum duration of content to buffer before waiting for the cue expressed
9015 in seconds. Default is 0.
9022 Apply color adjustments using curves.
9024 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9025 component (red, green and blue) has its values defined by @var{N} key points
9026 tied from each other using a smooth curve. The x-axis represents the pixel
9027 values from the input frame, and the y-axis the new pixel values to be set for
9030 By default, a component curve is defined by the two points @var{(0;0)} and
9031 @var{(1;1)}. This creates a straight line where each original pixel value is
9032 "adjusted" to its own value, which means no change to the image.
9034 The filter allows you to redefine these two points and add some more. A new
9035 curve (using a natural cubic spline interpolation) will be define to pass
9036 smoothly through all these new coordinates. The new defined points needs to be
9037 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9038 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9039 the vector spaces, the values will be clipped accordingly.
9041 The filter accepts the following options:
9045 Select one of the available color presets. This option can be used in addition
9046 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9047 options takes priority on the preset values.
9048 Available presets are:
9051 @item color_negative
9054 @item increase_contrast
9056 @item linear_contrast
9057 @item medium_contrast
9059 @item strong_contrast
9062 Default is @code{none}.
9064 Set the master key points. These points will define a second pass mapping. It
9065 is sometimes called a "luminance" or "value" mapping. It can be used with
9066 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9067 post-processing LUT.
9069 Set the key points for the red component.
9071 Set the key points for the green component.
9073 Set the key points for the blue component.
9075 Set the key points for all components (not including master).
9076 Can be used in addition to the other key points component
9077 options. In this case, the unset component(s) will fallback on this
9078 @option{all} setting.
9080 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9082 Save Gnuplot script of the curves in specified file.
9085 To avoid some filtergraph syntax conflicts, each key points list need to be
9086 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9088 @subsection Examples
9092 Increase slightly the middle level of blue:
9094 curves=blue='0/0 0.5/0.58 1/1'
9100 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'
9102 Here we obtain the following coordinates for each components:
9105 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9107 @code{(0;0) (0.50;0.48) (1;1)}
9109 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9113 The previous example can also be achieved with the associated built-in preset:
9115 curves=preset=vintage
9125 Use a Photoshop preset and redefine the points of the green component:
9127 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9131 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9132 and @command{gnuplot}:
9134 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9135 gnuplot -p /tmp/curves.plt
9141 Video data analysis filter.
9143 This filter shows hexadecimal pixel values of part of video.
9145 The filter accepts the following options:
9149 Set output video size.
9152 Set x offset from where to pick pixels.
9155 Set y offset from where to pick pixels.
9158 Set scope mode, can be one of the following:
9161 Draw hexadecimal pixel values with white color on black background.
9164 Draw hexadecimal pixel values with input video pixel color on black
9168 Draw hexadecimal pixel values on color background picked from input video,
9169 the text color is picked in such way so its always visible.
9173 Draw rows and columns numbers on left and top of video.
9176 Set background opacity.
9179 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9183 Apply Directional blur filter.
9185 The filter accepts the following options:
9189 Set angle of directional blur. Default is @code{45}.
9192 Set radius of directional blur. Default is @code{5}.
9195 Set which planes to filter. By default all planes are filtered.
9198 @subsection Commands
9199 This filter supports same @ref{commands} as options.
9200 The command accepts the same syntax of the corresponding option.
9202 If the specified expression is not valid, it is kept at its current
9207 Denoise frames using 2D DCT (frequency domain filtering).
9209 This filter is not designed for real time.
9211 The filter accepts the following options:
9215 Set the noise sigma constant.
9217 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9218 coefficient (absolute value) below this threshold with be dropped.
9220 If you need a more advanced filtering, see @option{expr}.
9222 Default is @code{0}.
9225 Set number overlapping pixels for each block. Since the filter can be slow, you
9226 may want to reduce this value, at the cost of a less effective filter and the
9227 risk of various artefacts.
9229 If the overlapping value doesn't permit processing the whole input width or
9230 height, a warning will be displayed and according borders won't be denoised.
9232 Default value is @var{blocksize}-1, which is the best possible setting.
9235 Set the coefficient factor expression.
9237 For each coefficient of a DCT block, this expression will be evaluated as a
9238 multiplier value for the coefficient.
9240 If this is option is set, the @option{sigma} option will be ignored.
9242 The absolute value of the coefficient can be accessed through the @var{c}
9246 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9247 @var{blocksize}, which is the width and height of the processed blocks.
9249 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9250 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9251 on the speed processing. Also, a larger block size does not necessarily means a
9255 @subsection Examples
9257 Apply a denoise with a @option{sigma} of @code{4.5}:
9262 The same operation can be achieved using the expression system:
9264 dctdnoiz=e='gte(c, 4.5*3)'
9267 Violent denoise using a block size of @code{16x16}:
9274 Remove banding artifacts from input video.
9275 It works by replacing banded pixels with average value of referenced pixels.
9277 The filter accepts the following options:
9284 Set banding detection threshold for each plane. Default is 0.02.
9285 Valid range is 0.00003 to 0.5.
9286 If difference between current pixel and reference pixel is less than threshold,
9287 it will be considered as banded.
9290 Banding detection range in pixels. Default is 16. If positive, random number
9291 in range 0 to set value will be used. If negative, exact absolute value
9293 The range defines square of four pixels around current pixel.
9296 Set direction in radians from which four pixel will be compared. If positive,
9297 random direction from 0 to set direction will be picked. If negative, exact of
9298 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9299 will pick only pixels on same row and -PI/2 will pick only pixels on same
9303 If enabled, current pixel is compared with average value of all four
9304 surrounding pixels. The default is enabled. If disabled current pixel is
9305 compared with all four surrounding pixels. The pixel is considered banded
9306 if only all four differences with surrounding pixels are less than threshold.
9309 If enabled, current pixel is changed if and only if all pixel components are banded,
9310 e.g. banding detection threshold is triggered for all color components.
9311 The default is disabled.
9316 Remove blocking artifacts from input video.
9318 The filter accepts the following options:
9322 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9323 This controls what kind of deblocking is applied.
9326 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9332 Set blocking detection thresholds. Allowed range is 0 to 1.
9333 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9334 Using higher threshold gives more deblocking strength.
9335 Setting @var{alpha} controls threshold detection at exact edge of block.
9336 Remaining options controls threshold detection near the edge. Each one for
9337 below/above or left/right. Setting any of those to @var{0} disables
9341 Set planes to filter. Default is to filter all available planes.
9344 @subsection Examples
9348 Deblock using weak filter and block size of 4 pixels.
9350 deblock=filter=weak:block=4
9354 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9355 deblocking more edges.
9357 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9361 Similar as above, but filter only first plane.
9363 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9367 Similar as above, but filter only second and third plane.
9369 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9376 Drop duplicated frames at regular intervals.
9378 The filter accepts the following options:
9382 Set the number of frames from which one will be dropped. Setting this to
9383 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9384 Default is @code{5}.
9387 Set the threshold for duplicate detection. If the difference metric for a frame
9388 is less than or equal to this value, then it is declared as duplicate. Default
9392 Set scene change threshold. Default is @code{15}.
9396 Set the size of the x and y-axis blocks used during metric calculations.
9397 Larger blocks give better noise suppression, but also give worse detection of
9398 small movements. Must be a power of two. Default is @code{32}.
9401 Mark main input as a pre-processed input and activate clean source input
9402 stream. This allows the input to be pre-processed with various filters to help
9403 the metrics calculation while keeping the frame selection lossless. When set to
9404 @code{1}, the first stream is for the pre-processed input, and the second
9405 stream is the clean source from where the kept frames are chosen. Default is
9409 Set whether or not chroma is considered in the metric calculations. Default is
9415 Apply 2D deconvolution of video stream in frequency domain using second stream
9418 The filter accepts the following options:
9422 Set which planes to process.
9425 Set which impulse video frames will be processed, can be @var{first}
9426 or @var{all}. Default is @var{all}.
9429 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9430 and height are not same and not power of 2 or if stream prior to convolving
9434 The @code{deconvolve} filter also supports the @ref{framesync} options.
9438 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9440 It accepts the following options:
9444 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9445 @var{rainbows} for cross-color reduction.
9448 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9451 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9454 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9457 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9462 Apply deflate effect to the video.
9464 This filter replaces the pixel by the local(3x3) average by taking into account
9465 only values lower than the pixel.
9467 It accepts the following options:
9474 Limit the maximum change for each plane, default is 65535.
9475 If 0, plane will remain unchanged.
9478 @subsection Commands
9480 This filter supports the all above options as @ref{commands}.
9484 Remove temporal frame luminance variations.
9486 It accepts the following options:
9490 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9493 Set averaging mode to smooth temporal luminance variations.
9495 Available values are:
9520 Do not actually modify frame. Useful when one only wants metadata.
9525 Remove judder produced by partially interlaced telecined content.
9527 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9528 source was partially telecined content then the output of @code{pullup,dejudder}
9529 will have a variable frame rate. May change the recorded frame rate of the
9530 container. Aside from that change, this filter will not affect constant frame
9533 The option available in this filter is:
9537 Specify the length of the window over which the judder repeats.
9539 Accepts any integer greater than 1. Useful values are:
9543 If the original was telecined from 24 to 30 fps (Film to NTSC).
9546 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9549 If a mixture of the two.
9552 The default is @samp{4}.
9557 Suppress a TV station logo by a simple interpolation of the surrounding
9558 pixels. Just set a rectangle covering the logo and watch it disappear
9559 (and sometimes something even uglier appear - your mileage may vary).
9561 It accepts the following parameters:
9566 Specify the top left corner coordinates of the logo. They must be
9571 Specify the width and height of the logo to clear. They must be
9575 Specify the thickness of the fuzzy edge of the rectangle (added to
9576 @var{w} and @var{h}). The default value is 1. This option is
9577 deprecated, setting higher values should no longer be necessary and
9581 When set to 1, a green rectangle is drawn on the screen to simplify
9582 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9583 The default value is 0.
9585 The rectangle is drawn on the outermost pixels which will be (partly)
9586 replaced with interpolated values. The values of the next pixels
9587 immediately outside this rectangle in each direction will be used to
9588 compute the interpolated pixel values inside the rectangle.
9592 @subsection Examples
9596 Set a rectangle covering the area with top left corner coordinates 0,0
9597 and size 100x77, and a band of size 10:
9599 delogo=x=0:y=0:w=100:h=77:band=10
9607 Remove the rain in the input image/video by applying the derain methods based on
9608 convolutional neural networks. Supported models:
9612 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9613 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9616 Training as well as model generation scripts are provided in
9617 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9619 Native model files (.model) can be generated from TensorFlow model
9620 files (.pb) by using tools/python/convert.py
9622 The filter accepts the following options:
9626 Specify which filter to use. This option accepts the following values:
9630 Derain filter. To conduct derain filter, you need to use a derain model.
9633 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9635 Default value is @samp{derain}.
9638 Specify which DNN backend to use for model loading and execution. This option accepts
9639 the following values:
9643 Native implementation of DNN loading and execution.
9646 TensorFlow backend. To enable this backend you
9647 need to install the TensorFlow for C library (see
9648 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9649 @code{--enable-libtensorflow}
9651 Default value is @samp{native}.
9654 Set path to model file specifying network architecture and its parameters.
9655 Note that different backends use different file formats. TensorFlow and native
9656 backend can load files for only its format.
9659 It can also be finished with @ref{dnn_processing} filter.
9663 Attempt to fix small changes in horizontal and/or vertical shift. This
9664 filter helps remove camera shake from hand-holding a camera, bumping a
9665 tripod, moving on a vehicle, etc.
9667 The filter accepts the following options:
9675 Specify a rectangular area where to limit the search for motion
9677 If desired the search for motion vectors can be limited to a
9678 rectangular area of the frame defined by its top left corner, width
9679 and height. These parameters have the same meaning as the drawbox
9680 filter which can be used to visualise the position of the bounding
9683 This is useful when simultaneous movement of subjects within the frame
9684 might be confused for camera motion by the motion vector search.
9686 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9687 then the full frame is used. This allows later options to be set
9688 without specifying the bounding box for the motion vector search.
9690 Default - search the whole frame.
9694 Specify the maximum extent of movement in x and y directions in the
9695 range 0-64 pixels. Default 16.
9698 Specify how to generate pixels to fill blanks at the edge of the
9699 frame. Available values are:
9702 Fill zeroes at blank locations
9704 Original image at blank locations
9706 Extruded edge value at blank locations
9708 Mirrored edge at blank locations
9710 Default value is @samp{mirror}.
9713 Specify the blocksize to use for motion search. Range 4-128 pixels,
9717 Specify the contrast threshold for blocks. Only blocks with more than
9718 the specified contrast (difference between darkest and lightest
9719 pixels) will be considered. Range 1-255, default 125.
9722 Specify the search strategy. Available values are:
9725 Set exhaustive search
9727 Set less exhaustive search.
9729 Default value is @samp{exhaustive}.
9732 If set then a detailed log of the motion search is written to the
9739 Remove unwanted contamination of foreground colors, caused by reflected color of
9740 greenscreen or bluescreen.
9742 This filter accepts the following options:
9746 Set what type of despill to use.
9749 Set how spillmap will be generated.
9752 Set how much to get rid of still remaining spill.
9755 Controls amount of red in spill area.
9758 Controls amount of green in spill area.
9759 Should be -1 for greenscreen.
9762 Controls amount of blue in spill area.
9763 Should be -1 for bluescreen.
9766 Controls brightness of spill area, preserving colors.
9769 Modify alpha from generated spillmap.
9772 @subsection Commands
9774 This filter supports the all above options as @ref{commands}.
9778 Apply an exact inverse of the telecine operation. It requires a predefined
9779 pattern specified using the pattern option which must be the same as that passed
9780 to the telecine filter.
9782 This filter accepts the following options:
9791 The default value is @code{top}.
9795 A string of numbers representing the pulldown pattern you wish to apply.
9796 The default value is @code{23}.
9799 A number representing position of the first frame with respect to the telecine
9800 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9805 Apply dilation effect to the video.
9807 This filter replaces the pixel by the local(3x3) maximum.
9809 It accepts the following options:
9816 Limit the maximum change for each plane, default is 65535.
9817 If 0, plane will remain unchanged.
9820 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9823 Flags to local 3x3 coordinates maps like this:
9830 @subsection Commands
9832 This filter supports the all above options as @ref{commands}.
9836 Displace pixels as indicated by second and third input stream.
9838 It takes three input streams and outputs one stream, the first input is the
9839 source, and second and third input are displacement maps.
9841 The second input specifies how much to displace pixels along the
9842 x-axis, while the third input specifies how much to displace pixels
9844 If one of displacement map streams terminates, last frame from that
9845 displacement map will be used.
9847 Note that once generated, displacements maps can be reused over and over again.
9849 A description of the accepted options follows.
9853 Set displace behavior for pixels that are out of range.
9855 Available values are:
9858 Missing pixels are replaced by black pixels.
9861 Adjacent pixels will spread out to replace missing pixels.
9864 Out of range pixels are wrapped so they point to pixels of other side.
9867 Out of range pixels will be replaced with mirrored pixels.
9869 Default is @samp{smear}.
9873 @subsection Examples
9877 Add ripple effect to rgb input of video size hd720:
9879 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
9883 Add wave effect to rgb input of video size hd720:
9885 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
9889 @anchor{dnn_processing}
9890 @section dnn_processing
9892 Do image processing with deep neural networks. It works together with another filter
9893 which converts the pixel format of the Frame to what the dnn network requires.
9895 The filter accepts the following options:
9899 Specify which DNN backend to use for model loading and execution. This option accepts
9900 the following values:
9904 Native implementation of DNN loading and execution.
9907 TensorFlow backend. To enable this backend you
9908 need to install the TensorFlow for C library (see
9909 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9910 @code{--enable-libtensorflow}
9913 OpenVINO backend. To enable this backend you
9914 need to build and install the OpenVINO for C library (see
9915 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9916 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9917 be needed if the header files and libraries are not installed into system path)
9921 Default value is @samp{native}.
9924 Set path to model file specifying network architecture and its parameters.
9925 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9926 backend can load files for only its format.
9928 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9931 Set the input name of the dnn network.
9934 Set the output name of the dnn network.
9938 @subsection Examples
9942 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9944 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9948 Halve the pixel value of the frame with format gray32f:
9950 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
9954 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9956 ./ffmpeg -i 480p.jpg -vf format=yuv420p,scale=w=iw*2:h=ih*2,dnn_processing=dnn_backend=tensorflow:model=srcnn.pb:input=x:output=y -y srcnn.jpg
9960 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9962 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9969 Draw a colored box on the input image.
9971 It accepts the following parameters:
9976 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9980 The expressions which specify the width and height of the box; if 0 they are interpreted as
9981 the input width and height. It defaults to 0.
9984 Specify the color of the box to write. For the general syntax of this option,
9985 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9986 value @code{invert} is used, the box edge color is the same as the
9987 video with inverted luma.
9990 The expression which sets the thickness of the box edge.
9991 A value of @code{fill} will create a filled box. Default value is @code{3}.
9993 See below for the list of accepted constants.
9996 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
9997 will overwrite the video's color and alpha pixels.
9998 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10001 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10002 following constants:
10006 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10010 horizontal and vertical chroma subsample values. For example for the
10011 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10015 The input width and height.
10018 The input sample aspect ratio.
10022 The x and y offset coordinates where the box is drawn.
10026 The width and height of the drawn box.
10029 The thickness of the drawn box.
10031 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10032 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10036 @subsection Examples
10040 Draw a black box around the edge of the input image:
10046 Draw a box with color red and an opacity of 50%:
10048 drawbox=10:20:200:60:red@@0.5
10051 The previous example can be specified as:
10053 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10057 Fill the box with pink color:
10059 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10063 Draw a 2-pixel red 2.40:1 mask:
10065 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
10069 @subsection Commands
10070 This filter supports same commands as options.
10071 The command accepts the same syntax of the corresponding option.
10073 If the specified expression is not valid, it is kept at its current
10078 Draw a graph using input video metadata.
10080 It accepts the following parameters:
10084 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10087 Set 1st foreground color expression.
10090 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10093 Set 2nd foreground color expression.
10096 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10099 Set 3rd foreground color expression.
10102 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10105 Set 4th foreground color expression.
10108 Set minimal value of metadata value.
10111 Set maximal value of metadata value.
10114 Set graph background color. Default is white.
10119 Available values for mode is:
10126 Default is @code{line}.
10131 Available values for slide is:
10134 Draw new frame when right border is reached.
10137 Replace old columns with new ones.
10140 Scroll from right to left.
10143 Scroll from left to right.
10146 Draw single picture.
10149 Default is @code{frame}.
10152 Set size of graph video. For the syntax of this option, check the
10153 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10154 The default value is @code{900x256}.
10157 Set the output frame rate. Default value is @code{25}.
10159 The foreground color expressions can use the following variables:
10162 Minimal value of metadata value.
10165 Maximal value of metadata value.
10168 Current metadata key value.
10171 The color is defined as 0xAABBGGRR.
10174 Example using metadata from @ref{signalstats} filter:
10176 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10179 Example using metadata from @ref{ebur128} filter:
10181 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10186 Draw a grid on the input image.
10188 It accepts the following parameters:
10193 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10197 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10198 input width and height, respectively, minus @code{thickness}, so image gets
10199 framed. Default to 0.
10202 Specify the color of the grid. For the general syntax of this option,
10203 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10204 value @code{invert} is used, the grid color is the same as the
10205 video with inverted luma.
10208 The expression which sets the thickness of the grid line. Default value is @code{1}.
10210 See below for the list of accepted constants.
10213 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10214 will overwrite the video's color and alpha pixels.
10215 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10218 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10219 following constants:
10223 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10227 horizontal and vertical chroma subsample values. For example for the
10228 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10232 The input grid cell width and height.
10235 The input sample aspect ratio.
10239 The x and y coordinates of some point of grid intersection (meant to configure offset).
10243 The width and height of the drawn cell.
10246 The thickness of the drawn cell.
10248 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10249 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10253 @subsection Examples
10257 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10259 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10263 Draw a white 3x3 grid with an opacity of 50%:
10265 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10269 @subsection Commands
10270 This filter supports same commands as options.
10271 The command accepts the same syntax of the corresponding option.
10273 If the specified expression is not valid, it is kept at its current
10279 Draw a text string or text from a specified file on top of a video, using the
10280 libfreetype library.
10282 To enable compilation of this filter, you need to configure FFmpeg with
10283 @code{--enable-libfreetype}.
10284 To enable default font fallback and the @var{font} option you need to
10285 configure FFmpeg with @code{--enable-libfontconfig}.
10286 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10287 @code{--enable-libfribidi}.
10291 It accepts the following parameters:
10296 Used to draw a box around text using the background color.
10297 The value must be either 1 (enable) or 0 (disable).
10298 The default value of @var{box} is 0.
10301 Set the width of the border to be drawn around the box using @var{boxcolor}.
10302 The default value of @var{boxborderw} is 0.
10305 The color to be used for drawing box around text. For the syntax of this
10306 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10308 The default value of @var{boxcolor} is "white".
10311 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10312 The default value of @var{line_spacing} is 0.
10315 Set the width of the border to be drawn around the text using @var{bordercolor}.
10316 The default value of @var{borderw} is 0.
10319 Set the color to be used for drawing border around text. For the syntax of this
10320 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10322 The default value of @var{bordercolor} is "black".
10325 Select how the @var{text} is expanded. Can be either @code{none},
10326 @code{strftime} (deprecated) or
10327 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10331 Set a start time for the count. Value is in microseconds. Only applied
10332 in the deprecated strftime expansion mode. To emulate in normal expansion
10333 mode use the @code{pts} function, supplying the start time (in seconds)
10334 as the second argument.
10337 If true, check and fix text coords to avoid clipping.
10340 The color to be used for drawing fonts. For the syntax of this option, check
10341 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10343 The default value of @var{fontcolor} is "black".
10345 @item fontcolor_expr
10346 String which is expanded the same way as @var{text} to obtain dynamic
10347 @var{fontcolor} value. By default this option has empty value and is not
10348 processed. When this option is set, it overrides @var{fontcolor} option.
10351 The font family to be used for drawing text. By default Sans.
10354 The font file to be used for drawing text. The path must be included.
10355 This parameter is mandatory if the fontconfig support is disabled.
10358 Draw the text applying alpha blending. The value can
10359 be a number between 0.0 and 1.0.
10360 The expression accepts the same variables @var{x, y} as well.
10361 The default value is 1.
10362 Please see @var{fontcolor_expr}.
10365 The font size to be used for drawing text.
10366 The default value of @var{fontsize} is 16.
10369 If set to 1, attempt to shape the text (for example, reverse the order of
10370 right-to-left text and join Arabic characters) before drawing it.
10371 Otherwise, just draw the text exactly as given.
10372 By default 1 (if supported).
10374 @item ft_load_flags
10375 The flags to be used for loading the fonts.
10377 The flags map the corresponding flags supported by libfreetype, and are
10378 a combination of the following values:
10385 @item vertical_layout
10386 @item force_autohint
10389 @item ignore_global_advance_width
10391 @item ignore_transform
10393 @item linear_design
10397 Default value is "default".
10399 For more information consult the documentation for the FT_LOAD_*
10403 The color to be used for drawing a shadow behind the drawn text. For the
10404 syntax of this option, check the @ref{color syntax,,"Color" section in the
10405 ffmpeg-utils manual,ffmpeg-utils}.
10407 The default value of @var{shadowcolor} is "black".
10411 The x and y offsets for the text shadow position with respect to the
10412 position of the text. They can be either positive or negative
10413 values. The default value for both is "0".
10416 The starting frame number for the n/frame_num variable. The default value
10420 The size in number of spaces to use for rendering the tab.
10421 Default value is 4.
10424 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10425 format. It can be used with or without text parameter. @var{timecode_rate}
10426 option must be specified.
10428 @item timecode_rate, rate, r
10429 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10430 integer. Minimum value is "1".
10431 Drop-frame timecode is supported for frame rates 30 & 60.
10434 If set to 1, the output of the timecode option will wrap around at 24 hours.
10435 Default is 0 (disabled).
10438 The text string to be drawn. The text must be a sequence of UTF-8
10439 encoded characters.
10440 This parameter is mandatory if no file is specified with the parameter
10444 A text file containing text to be drawn. The text must be a sequence
10445 of UTF-8 encoded characters.
10447 This parameter is mandatory if no text string is specified with the
10448 parameter @var{text}.
10450 If both @var{text} and @var{textfile} are specified, an error is thrown.
10453 If set to 1, the @var{textfile} will be reloaded before each frame.
10454 Be sure to update it atomically, or it may be read partially, or even fail.
10458 The expressions which specify the offsets where text will be drawn
10459 within the video frame. They are relative to the top/left border of the
10462 The default value of @var{x} and @var{y} is "0".
10464 See below for the list of accepted constants and functions.
10467 The parameters for @var{x} and @var{y} are expressions containing the
10468 following constants and functions:
10472 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10476 horizontal and vertical chroma subsample values. For example for the
10477 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10480 the height of each text line
10488 @item max_glyph_a, ascent
10489 the maximum distance from the baseline to the highest/upper grid
10490 coordinate used to place a glyph outline point, for all the rendered
10492 It is a positive value, due to the grid's orientation with the Y axis
10495 @item max_glyph_d, descent
10496 the maximum distance from the baseline to the lowest grid coordinate
10497 used to place a glyph outline point, for all the rendered glyphs.
10498 This is a negative value, due to the grid's orientation, with the Y axis
10502 maximum glyph height, that is the maximum height for all the glyphs
10503 contained in the rendered text, it is equivalent to @var{ascent} -
10507 maximum glyph width, that is the maximum width for all the glyphs
10508 contained in the rendered text
10511 the number of input frame, starting from 0
10513 @item rand(min, max)
10514 return a random number included between @var{min} and @var{max}
10517 The input sample aspect ratio.
10520 timestamp expressed in seconds, NAN if the input timestamp is unknown
10523 the height of the rendered text
10526 the width of the rendered text
10530 the x and y offset coordinates where the text is drawn.
10532 These parameters allow the @var{x} and @var{y} expressions to refer
10533 to each other, so you can for example specify @code{y=x/dar}.
10536 A one character description of the current frame's picture type.
10539 The current packet's position in the input file or stream
10540 (in bytes, from the start of the input). A value of -1 indicates
10541 this info is not available.
10544 The current packet's duration, in seconds.
10547 The current packet's size (in bytes).
10550 @anchor{drawtext_expansion}
10551 @subsection Text expansion
10553 If @option{expansion} is set to @code{strftime},
10554 the filter recognizes strftime() sequences in the provided text and
10555 expands them accordingly. Check the documentation of strftime(). This
10556 feature is deprecated.
10558 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10560 If @option{expansion} is set to @code{normal} (which is the default),
10561 the following expansion mechanism is used.
10563 The backslash character @samp{\}, followed by any character, always expands to
10564 the second character.
10566 Sequences of the form @code{%@{...@}} are expanded. The text between the
10567 braces is a function name, possibly followed by arguments separated by ':'.
10568 If the arguments contain special characters or delimiters (':' or '@}'),
10569 they should be escaped.
10571 Note that they probably must also be escaped as the value for the
10572 @option{text} option in the filter argument string and as the filter
10573 argument in the filtergraph description, and possibly also for the shell,
10574 that makes up to four levels of escaping; using a text file avoids these
10577 The following functions are available:
10582 The expression evaluation result.
10584 It must take one argument specifying the expression to be evaluated,
10585 which accepts the same constants and functions as the @var{x} and
10586 @var{y} values. Note that not all constants should be used, for
10587 example the text size is not known when evaluating the expression, so
10588 the constants @var{text_w} and @var{text_h} will have an undefined
10591 @item expr_int_format, eif
10592 Evaluate the expression's value and output as formatted integer.
10594 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10595 The second argument specifies the output format. Allowed values are @samp{x},
10596 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10597 @code{printf} function.
10598 The third parameter is optional and sets the number of positions taken by the output.
10599 It can be used to add padding with zeros from the left.
10602 The time at which the filter is running, expressed in UTC.
10603 It can accept an argument: a strftime() format string.
10606 The time at which the filter is running, expressed in the local time zone.
10607 It can accept an argument: a strftime() format string.
10610 Frame metadata. Takes one or two arguments.
10612 The first argument is mandatory and specifies the metadata key.
10614 The second argument is optional and specifies a default value, used when the
10615 metadata key is not found or empty.
10617 Available metadata can be identified by inspecting entries
10618 starting with TAG included within each frame section
10619 printed by running @code{ffprobe -show_frames}.
10621 String metadata generated in filters leading to
10622 the drawtext filter are also available.
10625 The frame number, starting from 0.
10628 A one character description of the current picture type.
10631 The timestamp of the current frame.
10632 It can take up to three arguments.
10634 The first argument is the format of the timestamp; it defaults to @code{flt}
10635 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10636 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10637 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10638 @code{localtime} stands for the timestamp of the frame formatted as
10639 local time zone time.
10641 The second argument is an offset added to the timestamp.
10643 If the format is set to @code{hms}, a third argument @code{24HH} may be
10644 supplied to present the hour part of the formatted timestamp in 24h format
10647 If the format is set to @code{localtime} or @code{gmtime},
10648 a third argument may be supplied: a strftime() format string.
10649 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10652 @subsection Commands
10654 This filter supports altering parameters via commands:
10657 Alter existing filter parameters.
10659 Syntax for the argument is the same as for filter invocation, e.g.
10662 fontsize=56:fontcolor=green:text='Hello World'
10665 Full filter invocation with sendcmd would look like this:
10668 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10672 If the entire argument can't be parsed or applied as valid values then the filter will
10673 continue with its existing parameters.
10675 @subsection Examples
10679 Draw "Test Text" with font FreeSerif, using the default values for the
10680 optional parameters.
10683 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10687 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10688 and y=50 (counting from the top-left corner of the screen), text is
10689 yellow with a red box around it. Both the text and the box have an
10693 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10694 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10697 Note that the double quotes are not necessary if spaces are not used
10698 within the parameter list.
10701 Show the text at the center of the video frame:
10703 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10707 Show the text at a random position, switching to a new position every 30 seconds:
10709 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)"
10713 Show a text line sliding from right to left in the last row of the video
10714 frame. The file @file{LONG_LINE} is assumed to contain a single line
10717 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10721 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10723 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10727 Draw a single green letter "g", at the center of the input video.
10728 The glyph baseline is placed at half screen height.
10730 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10734 Show text for 1 second every 3 seconds:
10736 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10740 Use fontconfig to set the font. Note that the colons need to be escaped.
10742 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10746 Draw "Test Text" with font size dependent on height of the video.
10748 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10752 Print the date of a real-time encoding (see strftime(3)):
10754 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10758 Show text fading in and out (appearing/disappearing):
10761 DS=1.0 # display start
10762 DE=10.0 # display end
10763 FID=1.5 # fade in duration
10764 FOD=5 # fade out duration
10765 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 @}"
10769 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10770 and the @option{fontsize} value are included in the @option{y} offset.
10772 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10773 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10777 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10778 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10779 must have option @option{-export_path_metadata 1} for the special metadata fields
10780 to be available for filters.
10782 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10787 For more information about libfreetype, check:
10788 @url{http://www.freetype.org/}.
10790 For more information about fontconfig, check:
10791 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10793 For more information about libfribidi, check:
10794 @url{http://fribidi.org/}.
10796 @section edgedetect
10798 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10800 The filter accepts the following options:
10805 Set low and high threshold values used by the Canny thresholding
10808 The high threshold selects the "strong" edge pixels, which are then
10809 connected through 8-connectivity with the "weak" edge pixels selected
10810 by the low threshold.
10812 @var{low} and @var{high} threshold values must be chosen in the range
10813 [0,1], and @var{low} should be lesser or equal to @var{high}.
10815 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10819 Define the drawing mode.
10823 Draw white/gray wires on black background.
10826 Mix the colors to create a paint/cartoon effect.
10829 Apply Canny edge detector on all selected planes.
10831 Default value is @var{wires}.
10834 Select planes for filtering. By default all available planes are filtered.
10837 @subsection Examples
10841 Standard edge detection with custom values for the hysteresis thresholding:
10843 edgedetect=low=0.1:high=0.4
10847 Painting effect without thresholding:
10849 edgedetect=mode=colormix:high=0
10855 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10857 For each input image, the filter will compute the optimal mapping from
10858 the input to the output given the codebook length, that is the number
10859 of distinct output colors.
10861 This filter accepts the following options.
10864 @item codebook_length, l
10865 Set codebook length. The value must be a positive integer, and
10866 represents the number of distinct output colors. Default value is 256.
10869 Set the maximum number of iterations to apply for computing the optimal
10870 mapping. The higher the value the better the result and the higher the
10871 computation time. Default value is 1.
10874 Set a random seed, must be an integer included between 0 and
10875 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10876 will try to use a good random seed on a best effort basis.
10879 Set pal8 output pixel format. This option does not work with codebook
10880 length greater than 256.
10885 Measure graylevel entropy in histogram of color channels of video frames.
10887 It accepts the following parameters:
10891 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10893 @var{diff} mode measures entropy of histogram delta values, absolute differences
10894 between neighbour histogram values.
10898 Set brightness, contrast, saturation and approximate gamma adjustment.
10900 The filter accepts the following options:
10904 Set the contrast expression. The value must be a float value in range
10905 @code{-1000.0} to @code{1000.0}. The default value is "1".
10908 Set the brightness expression. The value must be a float value in
10909 range @code{-1.0} to @code{1.0}. The default value is "0".
10912 Set the saturation expression. The value must be a float in
10913 range @code{0.0} to @code{3.0}. The default value is "1".
10916 Set the gamma expression. The value must be a float in range
10917 @code{0.1} to @code{10.0}. The default value is "1".
10920 Set the gamma expression for red. The value must be a float in
10921 range @code{0.1} to @code{10.0}. The default value is "1".
10924 Set the gamma expression for green. The value must be a float in range
10925 @code{0.1} to @code{10.0}. The default value is "1".
10928 Set the gamma expression for blue. The value must be a float in range
10929 @code{0.1} to @code{10.0}. The default value is "1".
10932 Set the gamma weight expression. It can be used to reduce the effect
10933 of a high gamma value on bright image areas, e.g. keep them from
10934 getting overamplified and just plain white. The value must be a float
10935 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10936 gamma correction all the way down while @code{1.0} leaves it at its
10937 full strength. Default is "1".
10940 Set when the expressions for brightness, contrast, saturation and
10941 gamma expressions are evaluated.
10943 It accepts the following values:
10946 only evaluate expressions once during the filter initialization or
10947 when a command is processed
10950 evaluate expressions for each incoming frame
10953 Default value is @samp{init}.
10956 The expressions accept the following parameters:
10959 frame count of the input frame starting from 0
10962 byte position of the corresponding packet in the input file, NAN if
10966 frame rate of the input video, NAN if the input frame rate is unknown
10969 timestamp expressed in seconds, NAN if the input timestamp is unknown
10972 @subsection Commands
10973 The filter supports the following commands:
10977 Set the contrast expression.
10980 Set the brightness expression.
10983 Set the saturation expression.
10986 Set the gamma expression.
10989 Set the gamma_r expression.
10992 Set gamma_g expression.
10995 Set gamma_b expression.
10998 Set gamma_weight expression.
11000 The command accepts the same syntax of the corresponding option.
11002 If the specified expression is not valid, it is kept at its current
11009 Apply erosion effect to the video.
11011 This filter replaces the pixel by the local(3x3) minimum.
11013 It accepts the following options:
11020 Limit the maximum change for each plane, default is 65535.
11021 If 0, plane will remain unchanged.
11024 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11027 Flags to local 3x3 coordinates maps like this:
11034 @subsection Commands
11036 This filter supports the all above options as @ref{commands}.
11038 @section extractplanes
11040 Extract color channel components from input video stream into
11041 separate grayscale video streams.
11043 The filter accepts the following option:
11047 Set plane(s) to extract.
11049 Available values for planes are:
11060 Choosing planes not available in the input will result in an error.
11061 That means you cannot select @code{r}, @code{g}, @code{b} planes
11062 with @code{y}, @code{u}, @code{v} planes at same time.
11065 @subsection Examples
11069 Extract luma, u and v color channel component from input video frame
11070 into 3 grayscale outputs:
11072 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
11078 Apply a fade-in/out effect to the input video.
11080 It accepts the following parameters:
11084 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11086 Default is @code{in}.
11088 @item start_frame, s
11089 Specify the number of the frame to start applying the fade
11090 effect at. Default is 0.
11093 The number of frames that the fade effect lasts. At the end of the
11094 fade-in effect, the output video will have the same intensity as the input video.
11095 At the end of the fade-out transition, the output video will be filled with the
11096 selected @option{color}.
11100 If set to 1, fade only alpha channel, if one exists on the input.
11101 Default value is 0.
11103 @item start_time, st
11104 Specify the timestamp (in seconds) of the frame to start to apply the fade
11105 effect. If both start_frame and start_time are specified, the fade will start at
11106 whichever comes last. Default is 0.
11109 The number of seconds for which the fade effect has to last. At the end of the
11110 fade-in effect the output video will have the same intensity as the input video,
11111 at the end of the fade-out transition the output video will be filled with the
11112 selected @option{color}.
11113 If both duration and nb_frames are specified, duration is used. Default is 0
11114 (nb_frames is used by default).
11117 Specify the color of the fade. Default is "black".
11120 @subsection Examples
11124 Fade in the first 30 frames of video:
11129 The command above is equivalent to:
11135 Fade out the last 45 frames of a 200-frame video:
11138 fade=type=out:start_frame=155:nb_frames=45
11142 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11144 fade=in:0:25, fade=out:975:25
11148 Make the first 5 frames yellow, then fade in from frame 5-24:
11150 fade=in:5:20:color=yellow
11154 Fade in alpha over first 25 frames of video:
11156 fade=in:0:25:alpha=1
11160 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11162 fade=t=in:st=5.5:d=0.5
11168 Denoise frames using 3D FFT (frequency domain filtering).
11170 The filter accepts the following options:
11174 Set the noise sigma constant. This sets denoising strength.
11175 Default value is 1. Allowed range is from 0 to 30.
11176 Using very high sigma with low overlap may give blocking artifacts.
11179 Set amount of denoising. By default all detected noise is reduced.
11180 Default value is 1. Allowed range is from 0 to 1.
11183 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11184 Actual size of block in pixels is 2 to power of @var{block}, so by default
11185 block size in pixels is 2^4 which is 16.
11188 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11191 Set number of previous frames to use for denoising. By default is set to 0.
11194 Set number of next frames to to use for denoising. By default is set to 0.
11197 Set planes which will be filtered, by default are all available filtered
11202 Apply arbitrary expressions to samples in frequency domain
11206 Adjust the dc value (gain) of the luma plane of the image. The filter
11207 accepts an integer value in range @code{0} to @code{1000}. The default
11208 value is set to @code{0}.
11211 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11212 filter accepts an integer value in range @code{0} to @code{1000}. The
11213 default value is set to @code{0}.
11216 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11217 filter accepts an integer value in range @code{0} to @code{1000}. The
11218 default value is set to @code{0}.
11221 Set the frequency domain weight expression for the luma plane.
11224 Set the frequency domain weight expression for the 1st chroma plane.
11227 Set the frequency domain weight expression for the 2nd chroma plane.
11230 Set when the expressions are evaluated.
11232 It accepts the following values:
11235 Only evaluate expressions once during the filter initialization.
11238 Evaluate expressions for each incoming frame.
11241 Default value is @samp{init}.
11243 The filter accepts the following variables:
11246 The coordinates of the current sample.
11250 The width and height of the image.
11253 The number of input frame, starting from 0.
11256 @subsection Examples
11262 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11268 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11274 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11280 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11287 Extract a single field from an interlaced image using stride
11288 arithmetic to avoid wasting CPU time. The output frames are marked as
11291 The filter accepts the following options:
11295 Specify whether to extract the top (if the value is @code{0} or
11296 @code{top}) or the bottom field (if the value is @code{1} or
11302 Create new frames by copying the top and bottom fields from surrounding frames
11303 supplied as numbers by the hint file.
11307 Set file containing hints: absolute/relative frame numbers.
11309 There must be one line for each frame in a clip. Each line must contain two
11310 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11311 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11312 is current frame number for @code{absolute} mode or out of [-1, 1] range
11313 for @code{relative} mode. First number tells from which frame to pick up top
11314 field and second number tells from which frame to pick up bottom field.
11316 If optionally followed by @code{+} output frame will be marked as interlaced,
11317 else if followed by @code{-} output frame will be marked as progressive, else
11318 it will be marked same as input frame.
11319 If optionally followed by @code{t} output frame will use only top field, or in
11320 case of @code{b} it will use only bottom field.
11321 If line starts with @code{#} or @code{;} that line is skipped.
11324 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11327 Example of first several lines of @code{hint} file for @code{relative} mode:
11329 0,0 - # first frame
11330 1,0 - # second frame, use third's frame top field and second's frame bottom field
11331 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11346 @section fieldmatch
11348 Field matching filter for inverse telecine. It is meant to reconstruct the
11349 progressive frames from a telecined stream. The filter does not drop duplicated
11350 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11351 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11353 The separation of the field matching and the decimation is notably motivated by
11354 the possibility of inserting a de-interlacing filter fallback between the two.
11355 If the source has mixed telecined and real interlaced content,
11356 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11357 But these remaining combed frames will be marked as interlaced, and thus can be
11358 de-interlaced by a later filter such as @ref{yadif} before decimation.
11360 In addition to the various configuration options, @code{fieldmatch} can take an
11361 optional second stream, activated through the @option{ppsrc} option. If
11362 enabled, the frames reconstruction will be based on the fields and frames from
11363 this second stream. This allows the first input to be pre-processed in order to
11364 help the various algorithms of the filter, while keeping the output lossless
11365 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11366 or brightness/contrast adjustments can help.
11368 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11369 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11370 which @code{fieldmatch} is based on. While the semantic and usage are very
11371 close, some behaviour and options names can differ.
11373 The @ref{decimate} filter currently only works for constant frame rate input.
11374 If your input has mixed telecined (30fps) and progressive content with a lower
11375 framerate like 24fps use the following filterchain to produce the necessary cfr
11376 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11378 The filter accepts the following options:
11382 Specify the assumed field order of the input stream. Available values are:
11386 Auto detect parity (use FFmpeg's internal parity value).
11388 Assume bottom field first.
11390 Assume top field first.
11393 Note that it is sometimes recommended not to trust the parity announced by the
11396 Default value is @var{auto}.
11399 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11400 sense that it won't risk creating jerkiness due to duplicate frames when
11401 possible, but if there are bad edits or blended fields it will end up
11402 outputting combed frames when a good match might actually exist. On the other
11403 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11404 but will almost always find a good frame if there is one. The other values are
11405 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11406 jerkiness and creating duplicate frames versus finding good matches in sections
11407 with bad edits, orphaned fields, blended fields, etc.
11409 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11411 Available values are:
11415 2-way matching (p/c)
11417 2-way matching, and trying 3rd match if still combed (p/c + n)
11419 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11421 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11422 still combed (p/c + n + u/b)
11424 3-way matching (p/c/n)
11426 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11427 detected as combed (p/c/n + u/b)
11430 The parenthesis at the end indicate the matches that would be used for that
11431 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11434 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11437 Default value is @var{pc_n}.
11440 Mark the main input stream as a pre-processed input, and enable the secondary
11441 input stream as the clean source to pick the fields from. See the filter
11442 introduction for more details. It is similar to the @option{clip2} feature from
11445 Default value is @code{0} (disabled).
11448 Set the field to match from. It is recommended to set this to the same value as
11449 @option{order} unless you experience matching failures with that setting. In
11450 certain circumstances changing the field that is used to match from can have a
11451 large impact on matching performance. Available values are:
11455 Automatic (same value as @option{order}).
11457 Match from the bottom field.
11459 Match from the top field.
11462 Default value is @var{auto}.
11465 Set whether or not chroma is included during the match comparisons. In most
11466 cases it is recommended to leave this enabled. You should set this to @code{0}
11467 only if your clip has bad chroma problems such as heavy rainbowing or other
11468 artifacts. Setting this to @code{0} could also be used to speed things up at
11469 the cost of some accuracy.
11471 Default value is @code{1}.
11475 These define an exclusion band which excludes the lines between @option{y0} and
11476 @option{y1} from being included in the field matching decision. An exclusion
11477 band can be used to ignore subtitles, a logo, or other things that may
11478 interfere with the matching. @option{y0} sets the starting scan line and
11479 @option{y1} sets the ending line; all lines in between @option{y0} and
11480 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11481 @option{y0} and @option{y1} to the same value will disable the feature.
11482 @option{y0} and @option{y1} defaults to @code{0}.
11485 Set the scene change detection threshold as a percentage of maximum change on
11486 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11487 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11488 @option{scthresh} is @code{[0.0, 100.0]}.
11490 Default value is @code{12.0}.
11493 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11494 account the combed scores of matches when deciding what match to use as the
11495 final match. Available values are:
11499 No final matching based on combed scores.
11501 Combed scores are only used when a scene change is detected.
11503 Use combed scores all the time.
11506 Default is @var{sc}.
11509 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11510 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11511 Available values are:
11515 No forced calculation.
11517 Force p/c/n calculations.
11519 Force p/c/n/u/b calculations.
11522 Default value is @var{none}.
11525 This is the area combing threshold used for combed frame detection. This
11526 essentially controls how "strong" or "visible" combing must be to be detected.
11527 Larger values mean combing must be more visible and smaller values mean combing
11528 can be less visible or strong and still be detected. Valid settings are from
11529 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11530 be detected as combed). This is basically a pixel difference value. A good
11531 range is @code{[8, 12]}.
11533 Default value is @code{9}.
11536 Sets whether or not chroma is considered in the combed frame decision. Only
11537 disable this if your source has chroma problems (rainbowing, etc.) that are
11538 causing problems for the combed frame detection with chroma enabled. Actually,
11539 using @option{chroma}=@var{0} is usually more reliable, except for the case
11540 where there is chroma only combing in the source.
11542 Default value is @code{0}.
11546 Respectively set the x-axis and y-axis size of the window used during combed
11547 frame detection. This has to do with the size of the area in which
11548 @option{combpel} pixels are required to be detected as combed for a frame to be
11549 declared combed. See the @option{combpel} parameter description for more info.
11550 Possible values are any number that is a power of 2 starting at 4 and going up
11553 Default value is @code{16}.
11556 The number of combed pixels inside any of the @option{blocky} by
11557 @option{blockx} size blocks on the frame for the frame to be detected as
11558 combed. While @option{cthresh} controls how "visible" the combing must be, this
11559 setting controls "how much" combing there must be in any localized area (a
11560 window defined by the @option{blockx} and @option{blocky} settings) on the
11561 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11562 which point no frames will ever be detected as combed). This setting is known
11563 as @option{MI} in TFM/VFM vocabulary.
11565 Default value is @code{80}.
11568 @anchor{p/c/n/u/b meaning}
11569 @subsection p/c/n/u/b meaning
11571 @subsubsection p/c/n
11573 We assume the following telecined stream:
11576 Top fields: 1 2 2 3 4
11577 Bottom fields: 1 2 3 4 4
11580 The numbers correspond to the progressive frame the fields relate to. Here, the
11581 first two frames are progressive, the 3rd and 4th are combed, and so on.
11583 When @code{fieldmatch} is configured to run a matching from bottom
11584 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11589 B 1 2 3 4 4 <-- matching reference
11598 As a result of the field matching, we can see that some frames get duplicated.
11599 To perform a complete inverse telecine, you need to rely on a decimation filter
11600 after this operation. See for instance the @ref{decimate} filter.
11602 The same operation now matching from top fields (@option{field}=@var{top})
11607 T 1 2 2 3 4 <-- matching reference
11617 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11618 basically, they refer to the frame and field of the opposite parity:
11621 @item @var{p} matches the field of the opposite parity in the previous frame
11622 @item @var{c} matches the field of the opposite parity in the current frame
11623 @item @var{n} matches the field of the opposite parity in the next frame
11628 The @var{u} and @var{b} matching are a bit special in the sense that they match
11629 from the opposite parity flag. In the following examples, we assume that we are
11630 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11631 'x' is placed above and below each matched fields.
11633 With bottom matching (@option{field}=@var{bottom}):
11638 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11639 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11647 With top matching (@option{field}=@var{top}):
11652 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11653 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11661 @subsection Examples
11663 Simple IVTC of a top field first telecined stream:
11665 fieldmatch=order=tff:combmatch=none, decimate
11668 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11670 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11673 @section fieldorder
11675 Transform the field order of the input video.
11677 It accepts the following parameters:
11682 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11683 for bottom field first.
11686 The default value is @samp{tff}.
11688 The transformation is done by shifting the picture content up or down
11689 by one line, and filling the remaining line with appropriate picture content.
11690 This method is consistent with most broadcast field order converters.
11692 If the input video is not flagged as being interlaced, or it is already
11693 flagged as being of the required output field order, then this filter does
11694 not alter the incoming video.
11696 It is very useful when converting to or from PAL DV material,
11697 which is bottom field first.
11701 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11704 @section fifo, afifo
11706 Buffer input images and send them when they are requested.
11708 It is mainly useful when auto-inserted by the libavfilter
11711 It does not take parameters.
11713 @section fillborders
11715 Fill borders of the input video, without changing video stream dimensions.
11716 Sometimes video can have garbage at the four edges and you may not want to
11717 crop video input to keep size multiple of some number.
11719 This filter accepts the following options:
11723 Number of pixels to fill from left border.
11726 Number of pixels to fill from right border.
11729 Number of pixels to fill from top border.
11732 Number of pixels to fill from bottom border.
11737 It accepts the following values:
11740 fill pixels using outermost pixels
11743 fill pixels using mirroring
11746 fill pixels with constant value
11749 Default is @var{smear}.
11752 Set color for pixels in fixed mode. Default is @var{black}.
11755 @subsection Commands
11756 This filter supports same @ref{commands} as options.
11757 The command accepts the same syntax of the corresponding option.
11759 If the specified expression is not valid, it is kept at its current
11764 Find a rectangular object
11766 It accepts the following options:
11770 Filepath of the object image, needs to be in gray8.
11773 Detection threshold, default is 0.5.
11776 Number of mipmaps, default is 3.
11778 @item xmin, ymin, xmax, ymax
11779 Specifies the rectangle in which to search.
11782 @subsection Examples
11786 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11788 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11794 Flood area with values of same pixel components with another values.
11796 It accepts the following options:
11799 Set pixel x coordinate.
11802 Set pixel y coordinate.
11805 Set source #0 component value.
11808 Set source #1 component value.
11811 Set source #2 component value.
11814 Set source #3 component value.
11817 Set destination #0 component value.
11820 Set destination #1 component value.
11823 Set destination #2 component value.
11826 Set destination #3 component value.
11832 Convert the input video to one of the specified pixel formats.
11833 Libavfilter will try to pick one that is suitable as input to
11836 It accepts the following parameters:
11840 A '|'-separated list of pixel format names, such as
11841 "pix_fmts=yuv420p|monow|rgb24".
11845 @subsection Examples
11849 Convert the input video to the @var{yuv420p} format
11851 format=pix_fmts=yuv420p
11854 Convert the input video to any of the formats in the list
11856 format=pix_fmts=yuv420p|yuv444p|yuv410p
11863 Convert the video to specified constant frame rate by duplicating or dropping
11864 frames as necessary.
11866 It accepts the following parameters:
11870 The desired output frame rate. The default is @code{25}.
11873 Assume the first PTS should be the given value, in seconds. This allows for
11874 padding/trimming at the start of stream. By default, no assumption is made
11875 about the first frame's expected PTS, so no padding or trimming is done.
11876 For example, this could be set to 0 to pad the beginning with duplicates of
11877 the first frame if a video stream starts after the audio stream or to trim any
11878 frames with a negative PTS.
11881 Timestamp (PTS) rounding method.
11883 Possible values are:
11890 round towards -infinity
11892 round towards +infinity
11896 The default is @code{near}.
11899 Action performed when reading the last frame.
11901 Possible values are:
11904 Use same timestamp rounding method as used for other frames.
11906 Pass through last frame if input duration has not been reached yet.
11908 The default is @code{round}.
11912 Alternatively, the options can be specified as a flat string:
11913 @var{fps}[:@var{start_time}[:@var{round}]].
11915 See also the @ref{setpts} filter.
11917 @subsection Examples
11921 A typical usage in order to set the fps to 25:
11927 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11929 fps=fps=film:round=near
11935 Pack two different video streams into a stereoscopic video, setting proper
11936 metadata on supported codecs. The two views should have the same size and
11937 framerate and processing will stop when the shorter video ends. Please note
11938 that you may conveniently adjust view properties with the @ref{scale} and
11941 It accepts the following parameters:
11945 The desired packing format. Supported values are:
11950 The views are next to each other (default).
11953 The views are on top of each other.
11956 The views are packed by line.
11959 The views are packed by column.
11962 The views are temporally interleaved.
11971 # Convert left and right views into a frame-sequential video
11972 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11974 # Convert views into a side-by-side video with the same output resolution as the input
11975 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
11980 Change the frame rate by interpolating new video output frames from the source
11983 This filter is not designed to function correctly with interlaced media. If
11984 you wish to change the frame rate of interlaced media then you are required
11985 to deinterlace before this filter and re-interlace after this filter.
11987 A description of the accepted options follows.
11991 Specify the output frames per second. This option can also be specified
11992 as a value alone. The default is @code{50}.
11995 Specify the start of a range where the output frame will be created as a
11996 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11997 the default is @code{15}.
12000 Specify the end of a range where the output frame will be created as a
12001 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12002 the default is @code{240}.
12005 Specify the level at which a scene change is detected as a value between
12006 0 and 100 to indicate a new scene; a low value reflects a low
12007 probability for the current frame to introduce a new scene, while a higher
12008 value means the current frame is more likely to be one.
12009 The default is @code{8.2}.
12012 Specify flags influencing the filter process.
12014 Available value for @var{flags} is:
12017 @item scene_change_detect, scd
12018 Enable scene change detection using the value of the option @var{scene}.
12019 This flag is enabled by default.
12025 Select one frame every N-th frame.
12027 This filter accepts the following option:
12030 Select frame after every @code{step} frames.
12031 Allowed values are positive integers higher than 0. Default value is @code{1}.
12034 @section freezedetect
12036 Detect frozen video.
12038 This filter logs a message and sets frame metadata when it detects that the
12039 input video has no significant change in content during a specified duration.
12040 Video freeze detection calculates the mean average absolute difference of all
12041 the components of video frames and compares it to a noise floor.
12043 The printed times and duration are expressed in seconds. The
12044 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12045 whose timestamp equals or exceeds the detection duration and it contains the
12046 timestamp of the first frame of the freeze. The
12047 @code{lavfi.freezedetect.freeze_duration} and
12048 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12051 The filter accepts the following options:
12055 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12056 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12060 Set freeze duration until notification (default is 2 seconds).
12063 @section freezeframes
12065 Freeze video frames.
12067 This filter freezes video frames using frame from 2nd input.
12069 The filter accepts the following options:
12073 Set number of first frame from which to start freeze.
12076 Set number of last frame from which to end freeze.
12079 Set number of frame from 2nd input which will be used instead of replaced frames.
12085 Apply a frei0r effect to the input video.
12087 To enable the compilation of this filter, you need to install the frei0r
12088 header and configure FFmpeg with @code{--enable-frei0r}.
12090 It accepts the following parameters:
12095 The name of the frei0r effect to load. If the environment variable
12096 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12097 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12098 Otherwise, the standard frei0r paths are searched, in this order:
12099 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12100 @file{/usr/lib/frei0r-1/}.
12102 @item filter_params
12103 A '|'-separated list of parameters to pass to the frei0r effect.
12107 A frei0r effect parameter can be a boolean (its value is either
12108 "y" or "n"), a double, a color (specified as
12109 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12110 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12111 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12112 a position (specified as @var{X}/@var{Y}, where
12113 @var{X} and @var{Y} are floating point numbers) and/or a string.
12115 The number and types of parameters depend on the loaded effect. If an
12116 effect parameter is not specified, the default value is set.
12118 @subsection Examples
12122 Apply the distort0r effect, setting the first two double parameters:
12124 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12128 Apply the colordistance effect, taking a color as the first parameter:
12130 frei0r=colordistance:0.2/0.3/0.4
12131 frei0r=colordistance:violet
12132 frei0r=colordistance:0x112233
12136 Apply the perspective effect, specifying the top left and top right image
12139 frei0r=perspective:0.2/0.2|0.8/0.2
12143 For more information, see
12144 @url{http://frei0r.dyne.org}
12146 @subsection Commands
12148 This filter supports the @option{filter_params} option as @ref{commands}.
12152 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12154 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12155 processing filter, one of them is performed once per block, not per pixel.
12156 This allows for much higher speed.
12158 The filter accepts the following options:
12162 Set quality. This option defines the number of levels for averaging. It accepts
12163 an integer in the range 4-5. Default value is @code{4}.
12166 Force a constant quantization parameter. It accepts an integer in range 0-63.
12167 If not set, the filter will use the QP from the video stream (if available).
12170 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12171 more details but also more artifacts, while higher values make the image smoother
12172 but also blurrier. Default value is @code{0} − PSNR optimal.
12174 @item use_bframe_qp
12175 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12176 option may cause flicker since the B-Frames have often larger QP. Default is
12177 @code{0} (not enabled).
12183 Apply Gaussian blur filter.
12185 The filter accepts the following options:
12189 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12192 Set number of steps for Gaussian approximation. Default is @code{1}.
12195 Set which planes to filter. By default all planes are filtered.
12198 Set vertical sigma, if negative it will be same as @code{sigma}.
12199 Default is @code{-1}.
12202 @subsection Commands
12203 This filter supports same commands as options.
12204 The command accepts the same syntax of the corresponding option.
12206 If the specified expression is not valid, it is kept at its current
12211 Apply generic equation to each pixel.
12213 The filter accepts the following options:
12216 @item lum_expr, lum
12217 Set the luminance expression.
12219 Set the chrominance blue expression.
12221 Set the chrominance red expression.
12222 @item alpha_expr, a
12223 Set the alpha expression.
12225 Set the red expression.
12226 @item green_expr, g
12227 Set the green expression.
12229 Set the blue expression.
12232 The colorspace is selected according to the specified options. If one
12233 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12234 options is specified, the filter will automatically select a YCbCr
12235 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12236 @option{blue_expr} options is specified, it will select an RGB
12239 If one of the chrominance expression is not defined, it falls back on the other
12240 one. If no alpha expression is specified it will evaluate to opaque value.
12241 If none of chrominance expressions are specified, they will evaluate
12242 to the luminance expression.
12244 The expressions can use the following variables and functions:
12248 The sequential number of the filtered frame, starting from @code{0}.
12252 The coordinates of the current sample.
12256 The width and height of the image.
12260 Width and height scale depending on the currently filtered plane. It is the
12261 ratio between the corresponding luma plane number of pixels and the current
12262 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12263 @code{0.5,0.5} for chroma planes.
12266 Time of the current frame, expressed in seconds.
12269 Return the value of the pixel at location (@var{x},@var{y}) of the current
12273 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12277 Return the value of the pixel at location (@var{x},@var{y}) of the
12278 blue-difference chroma plane. Return 0 if there is no such plane.
12281 Return the value of the pixel at location (@var{x},@var{y}) of the
12282 red-difference chroma plane. Return 0 if there is no such plane.
12287 Return the value of the pixel at location (@var{x},@var{y}) of the
12288 red/green/blue component. Return 0 if there is no such component.
12291 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12292 plane. Return 0 if there is no such plane.
12294 @item psum(x,y), lumsum(x, y), cbsum(x,y), crsum(x,y), rsum(x,y), gsum(x,y), bsum(x,y), alphasum(x,y)
12295 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12296 sums of samples within a rectangle. See the functions without the sum postfix.
12298 @item interpolation
12299 Set one of interpolation methods:
12304 Default is bilinear.
12307 For functions, if @var{x} and @var{y} are outside the area, the value will be
12308 automatically clipped to the closer edge.
12310 Please note that this filter can use multiple threads in which case each slice
12311 will have its own expression state. If you want to use only a single expression
12312 state because your expressions depend on previous state then you should limit
12313 the number of filter threads to 1.
12315 @subsection Examples
12319 Flip the image horizontally:
12325 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12326 wavelength of 100 pixels:
12328 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12332 Generate a fancy enigmatic moving light:
12334 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
12338 Generate a quick emboss effect:
12340 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12344 Modify RGB components depending on pixel position:
12346 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12350 Create a radial gradient that is the same size as the input (also see
12351 the @ref{vignette} filter):
12353 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12359 Fix the banding artifacts that are sometimes introduced into nearly flat
12360 regions by truncation to 8-bit color depth.
12361 Interpolate the gradients that should go where the bands are, and
12364 It is designed for playback only. Do not use it prior to
12365 lossy compression, because compression tends to lose the dither and
12366 bring back the bands.
12368 It accepts the following parameters:
12373 The maximum amount by which the filter will change any one pixel. This is also
12374 the threshold for detecting nearly flat regions. Acceptable values range from
12375 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12379 The neighborhood to fit the gradient to. A larger radius makes for smoother
12380 gradients, but also prevents the filter from modifying the pixels near detailed
12381 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12382 values will be clipped to the valid range.
12386 Alternatively, the options can be specified as a flat string:
12387 @var{strength}[:@var{radius}]
12389 @subsection Examples
12393 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12399 Specify radius, omitting the strength (which will fall-back to the default
12407 @anchor{graphmonitor}
12408 @section graphmonitor
12409 Show various filtergraph stats.
12411 With this filter one can debug complete filtergraph.
12412 Especially issues with links filling with queued frames.
12414 The filter accepts the following options:
12418 Set video output size. Default is @var{hd720}.
12421 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12424 Set output mode, can be @var{fulll} or @var{compact}.
12425 In @var{compact} mode only filters with some queued frames have displayed stats.
12428 Set flags which enable which stats are shown in video.
12430 Available values for flags are:
12433 Display number of queued frames in each link.
12435 @item frame_count_in
12436 Display number of frames taken from filter.
12438 @item frame_count_out
12439 Display number of frames given out from filter.
12442 Display current filtered frame pts.
12445 Display current filtered frame time.
12448 Display time base for filter link.
12451 Display used format for filter link.
12454 Display video size or number of audio channels in case of audio used by filter link.
12457 Display video frame rate or sample rate in case of audio used by filter link.
12460 Display link output status.
12464 Set upper limit for video rate of output stream, Default value is @var{25}.
12465 This guarantee that output video frame rate will not be higher than this value.
12469 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12470 and corrects the scene colors accordingly.
12472 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12474 The filter accepts the following options:
12478 The order of differentiation to be applied on the scene. Must be chosen in the range
12479 [0,2] and default value is 1.
12482 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12483 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12484 max value instead of calculating Minkowski distance.
12487 The standard deviation of Gaussian blur to be applied on the scene. Must be
12488 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12489 can't be equal to 0 if @var{difford} is greater than 0.
12492 @subsection Examples
12498 greyedge=difford=1:minknorm=5:sigma=2
12504 greyedge=difford=1:minknorm=0:sigma=2
12512 Apply a Hald CLUT to a video stream.
12514 First input is the video stream to process, and second one is the Hald CLUT.
12515 The Hald CLUT input can be a simple picture or a complete video stream.
12517 The filter accepts the following options:
12521 Force termination when the shortest input terminates. Default is @code{0}.
12523 Continue applying the last CLUT after the end of the stream. A value of
12524 @code{0} disable the filter after the last frame of the CLUT is reached.
12525 Default is @code{1}.
12528 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12529 filters share the same internals).
12531 This filter also supports the @ref{framesync} options.
12533 More information about the Hald CLUT can be found on Eskil Steenberg's website
12534 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12536 @subsection Workflow examples
12538 @subsubsection Hald CLUT video stream
12540 Generate an identity Hald CLUT stream altered with various effects:
12542 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
12545 Note: make sure you use a lossless codec.
12547 Then use it with @code{haldclut} to apply it on some random stream:
12549 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12552 The Hald CLUT will be applied to the 10 first seconds (duration of
12553 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12554 to the remaining frames of the @code{mandelbrot} stream.
12556 @subsubsection Hald CLUT with preview
12558 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12559 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12560 biggest possible square starting at the top left of the picture. The remaining
12561 padding pixels (bottom or right) will be ignored. This area can be used to add
12562 a preview of the Hald CLUT.
12564 Typically, the following generated Hald CLUT will be supported by the
12565 @code{haldclut} filter:
12568 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12569 pad=iw+320 [padded_clut];
12570 smptebars=s=320x256, split [a][b];
12571 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12572 [main][b] overlay=W-320" -frames:v 1 clut.png
12575 It contains the original and a preview of the effect of the CLUT: SMPTE color
12576 bars are displayed on the right-top, and below the same color bars processed by
12579 Then, the effect of this Hald CLUT can be visualized with:
12581 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12586 Flip the input video horizontally.
12588 For example, to horizontally flip the input video with @command{ffmpeg}:
12590 ffmpeg -i in.avi -vf "hflip" out.avi
12594 This filter applies a global color histogram equalization on a
12597 It can be used to correct video that has a compressed range of pixel
12598 intensities. The filter redistributes the pixel intensities to
12599 equalize their distribution across the intensity range. It may be
12600 viewed as an "automatically adjusting contrast filter". This filter is
12601 useful only for correcting degraded or poorly captured source
12604 The filter accepts the following options:
12608 Determine the amount of equalization to be applied. As the strength
12609 is reduced, the distribution of pixel intensities more-and-more
12610 approaches that of the input frame. The value must be a float number
12611 in the range [0,1] and defaults to 0.200.
12614 Set the maximum intensity that can generated and scale the output
12615 values appropriately. The strength should be set as desired and then
12616 the intensity can be limited if needed to avoid washing-out. The value
12617 must be a float number in the range [0,1] and defaults to 0.210.
12620 Set the antibanding level. If enabled the filter will randomly vary
12621 the luminance of output pixels by a small amount to avoid banding of
12622 the histogram. Possible values are @code{none}, @code{weak} or
12623 @code{strong}. It defaults to @code{none}.
12629 Compute and draw a color distribution histogram for the input video.
12631 The computed histogram is a representation of the color component
12632 distribution in an image.
12634 Standard histogram displays the color components distribution in an image.
12635 Displays color graph for each color component. Shows distribution of
12636 the Y, U, V, A or R, G, B components, depending on input format, in the
12637 current frame. Below each graph a color component scale meter is shown.
12639 The filter accepts the following options:
12643 Set height of level. Default value is @code{200}.
12644 Allowed range is [50, 2048].
12647 Set height of color scale. Default value is @code{12}.
12648 Allowed range is [0, 40].
12652 It accepts the following values:
12655 Per color component graphs are placed below each other.
12658 Per color component graphs are placed side by side.
12661 Presents information identical to that in the @code{parade}, except
12662 that the graphs representing color components are superimposed directly
12665 Default is @code{stack}.
12668 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12669 Default is @code{linear}.
12672 Set what color components to display.
12673 Default is @code{7}.
12676 Set foreground opacity. Default is @code{0.7}.
12679 Set background opacity. Default is @code{0.5}.
12682 @subsection Examples
12687 Calculate and draw histogram:
12689 ffplay -i input -vf histogram
12697 This is a high precision/quality 3d denoise filter. It aims to reduce
12698 image noise, producing smooth images and making still images really
12699 still. It should enhance compressibility.
12701 It accepts the following optional parameters:
12705 A non-negative floating point number which specifies spatial luma strength.
12706 It defaults to 4.0.
12708 @item chroma_spatial
12709 A non-negative floating point number which specifies spatial chroma strength.
12710 It defaults to 3.0*@var{luma_spatial}/4.0.
12713 A floating point number which specifies luma temporal strength. It defaults to
12714 6.0*@var{luma_spatial}/4.0.
12717 A floating point number which specifies chroma temporal strength. It defaults to
12718 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12721 @subsection Commands
12722 This filter supports same @ref{commands} as options.
12723 The command accepts the same syntax of the corresponding option.
12725 If the specified expression is not valid, it is kept at its current
12728 @anchor{hwdownload}
12729 @section hwdownload
12731 Download hardware frames to system memory.
12733 The input must be in hardware frames, and the output a non-hardware format.
12734 Not all formats will be supported on the output - it may be necessary to insert
12735 an additional @option{format} filter immediately following in the graph to get
12736 the output in a supported format.
12740 Map hardware frames to system memory or to another device.
12742 This filter has several different modes of operation; which one is used depends
12743 on the input and output formats:
12746 Hardware frame input, normal frame output
12748 Map the input frames to system memory and pass them to the output. If the
12749 original hardware frame is later required (for example, after overlaying
12750 something else on part of it), the @option{hwmap} filter can be used again
12751 in the next mode to retrieve it.
12753 Normal frame input, hardware frame output
12755 If the input is actually a software-mapped hardware frame, then unmap it -
12756 that is, return the original hardware frame.
12758 Otherwise, a device must be provided. Create new hardware surfaces on that
12759 device for the output, then map them back to the software format at the input
12760 and give those frames to the preceding filter. This will then act like the
12761 @option{hwupload} filter, but may be able to avoid an additional copy when
12762 the input is already in a compatible format.
12764 Hardware frame input and output
12766 A device must be supplied for the output, either directly or with the
12767 @option{derive_device} option. The input and output devices must be of
12768 different types and compatible - the exact meaning of this is
12769 system-dependent, but typically it means that they must refer to the same
12770 underlying hardware context (for example, refer to the same graphics card).
12772 If the input frames were originally created on the output device, then unmap
12773 to retrieve the original frames.
12775 Otherwise, map the frames to the output device - create new hardware frames
12776 on the output corresponding to the frames on the input.
12779 The following additional parameters are accepted:
12783 Set the frame mapping mode. Some combination of:
12786 The mapped frame should be readable.
12788 The mapped frame should be writeable.
12790 The mapping will always overwrite the entire frame.
12792 This may improve performance in some cases, as the original contents of the
12793 frame need not be loaded.
12795 The mapping must not involve any copying.
12797 Indirect mappings to copies of frames are created in some cases where either
12798 direct mapping is not possible or it would have unexpected properties.
12799 Setting this flag ensures that the mapping is direct and will fail if that is
12802 Defaults to @var{read+write} if not specified.
12804 @item derive_device @var{type}
12805 Rather than using the device supplied at initialisation, instead derive a new
12806 device of type @var{type} from the device the input frames exist on.
12809 In a hardware to hardware mapping, map in reverse - create frames in the sink
12810 and map them back to the source. This may be necessary in some cases where
12811 a mapping in one direction is required but only the opposite direction is
12812 supported by the devices being used.
12814 This option is dangerous - it may break the preceding filter in undefined
12815 ways if there are any additional constraints on that filter's output.
12816 Do not use it without fully understanding the implications of its use.
12822 Upload system memory frames to hardware surfaces.
12824 The device to upload to must be supplied when the filter is initialised. If
12825 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12826 option or with the @option{derive_device} option. The input and output devices
12827 must be of different types and compatible - the exact meaning of this is
12828 system-dependent, but typically it means that they must refer to the same
12829 underlying hardware context (for example, refer to the same graphics card).
12831 The following additional parameters are accepted:
12834 @item derive_device @var{type}
12835 Rather than using the device supplied at initialisation, instead derive a new
12836 device of type @var{type} from the device the input frames exist on.
12839 @anchor{hwupload_cuda}
12840 @section hwupload_cuda
12842 Upload system memory frames to a CUDA device.
12844 It accepts the following optional parameters:
12848 The number of the CUDA device to use
12853 Apply a high-quality magnification filter designed for pixel art. This filter
12854 was originally created by Maxim Stepin.
12856 It accepts the following option:
12860 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12861 @code{hq3x} and @code{4} for @code{hq4x}.
12862 Default is @code{3}.
12866 Stack input videos horizontally.
12868 All streams must be of same pixel format and of same height.
12870 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12871 to create same output.
12873 The filter accepts the following option:
12877 Set number of input streams. Default is 2.
12880 If set to 1, force the output to terminate when the shortest input
12881 terminates. Default value is 0.
12886 Modify the hue and/or the saturation of the input.
12888 It accepts the following parameters:
12892 Specify the hue angle as a number of degrees. It accepts an expression,
12893 and defaults to "0".
12896 Specify the saturation in the [-10,10] range. It accepts an expression and
12900 Specify the hue angle as a number of radians. It accepts an
12901 expression, and defaults to "0".
12904 Specify the brightness in the [-10,10] range. It accepts an expression and
12908 @option{h} and @option{H} are mutually exclusive, and can't be
12909 specified at the same time.
12911 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12912 expressions containing the following constants:
12916 frame count of the input frame starting from 0
12919 presentation timestamp of the input frame expressed in time base units
12922 frame rate of the input video, NAN if the input frame rate is unknown
12925 timestamp expressed in seconds, NAN if the input timestamp is unknown
12928 time base of the input video
12931 @subsection Examples
12935 Set the hue to 90 degrees and the saturation to 1.0:
12941 Same command but expressing the hue in radians:
12947 Rotate hue and make the saturation swing between 0
12948 and 2 over a period of 1 second:
12950 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12954 Apply a 3 seconds saturation fade-in effect starting at 0:
12956 hue="s=min(t/3\,1)"
12959 The general fade-in expression can be written as:
12961 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12965 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12967 hue="s=max(0\, min(1\, (8-t)/3))"
12970 The general fade-out expression can be written as:
12972 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12977 @subsection Commands
12979 This filter supports the following commands:
12985 Modify the hue and/or the saturation and/or brightness of the input video.
12986 The command accepts the same syntax of the corresponding option.
12988 If the specified expression is not valid, it is kept at its current
12992 @section hysteresis
12994 Grow first stream into second stream by connecting components.
12995 This makes it possible to build more robust edge masks.
12997 This filter accepts the following options:
13001 Set which planes will be processed as bitmap, unprocessed planes will be
13002 copied from first stream.
13003 By default value 0xf, all planes will be processed.
13006 Set threshold which is used in filtering. If pixel component value is higher than
13007 this value filter algorithm for connecting components is activated.
13008 By default value is 0.
13011 The @code{hysteresis} filter also supports the @ref{framesync} options.
13015 Detect video interlacing type.
13017 This filter tries to detect if the input frames are interlaced, progressive,
13018 top or bottom field first. It will also try to detect fields that are
13019 repeated between adjacent frames (a sign of telecine).
13021 Single frame detection considers only immediately adjacent frames when classifying each frame.
13022 Multiple frame detection incorporates the classification history of previous frames.
13024 The filter will log these metadata values:
13027 @item single.current_frame
13028 Detected type of current frame using single-frame detection. One of:
13029 ``tff'' (top field first), ``bff'' (bottom field first),
13030 ``progressive'', or ``undetermined''
13033 Cumulative number of frames detected as top field first using single-frame detection.
13036 Cumulative number of frames detected as top field first using multiple-frame detection.
13039 Cumulative number of frames detected as bottom field first using single-frame detection.
13041 @item multiple.current_frame
13042 Detected type of current frame using multiple-frame detection. One of:
13043 ``tff'' (top field first), ``bff'' (bottom field first),
13044 ``progressive'', or ``undetermined''
13047 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13049 @item single.progressive
13050 Cumulative number of frames detected as progressive using single-frame detection.
13052 @item multiple.progressive
13053 Cumulative number of frames detected as progressive using multiple-frame detection.
13055 @item single.undetermined
13056 Cumulative number of frames that could not be classified using single-frame detection.
13058 @item multiple.undetermined
13059 Cumulative number of frames that could not be classified using multiple-frame detection.
13061 @item repeated.current_frame
13062 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13064 @item repeated.neither
13065 Cumulative number of frames with no repeated field.
13068 Cumulative number of frames with the top field repeated from the previous frame's top field.
13070 @item repeated.bottom
13071 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13074 The filter accepts the following options:
13078 Set interlacing threshold.
13080 Set progressive threshold.
13082 Threshold for repeated field detection.
13084 Number of frames after which a given frame's contribution to the
13085 statistics is halved (i.e., it contributes only 0.5 to its
13086 classification). The default of 0 means that all frames seen are given
13087 full weight of 1.0 forever.
13088 @item analyze_interlaced_flag
13089 When this is not 0 then idet will use the specified number of frames to determine
13090 if the interlaced flag is accurate, it will not count undetermined frames.
13091 If the flag is found to be accurate it will be used without any further
13092 computations, if it is found to be inaccurate it will be cleared without any
13093 further computations. This allows inserting the idet filter as a low computational
13094 method to clean up the interlaced flag
13099 Deinterleave or interleave fields.
13101 This filter allows one to process interlaced images fields without
13102 deinterlacing them. Deinterleaving splits the input frame into 2
13103 fields (so called half pictures). Odd lines are moved to the top
13104 half of the output image, even lines to the bottom half.
13105 You can process (filter) them independently and then re-interleave them.
13107 The filter accepts the following options:
13111 @item chroma_mode, c
13112 @item alpha_mode, a
13113 Available values for @var{luma_mode}, @var{chroma_mode} and
13114 @var{alpha_mode} are:
13120 @item deinterleave, d
13121 Deinterleave fields, placing one above the other.
13123 @item interleave, i
13124 Interleave fields. Reverse the effect of deinterleaving.
13126 Default value is @code{none}.
13128 @item luma_swap, ls
13129 @item chroma_swap, cs
13130 @item alpha_swap, as
13131 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13134 @subsection Commands
13136 This filter supports the all above options as @ref{commands}.
13140 Apply inflate effect to the video.
13142 This filter replaces the pixel by the local(3x3) average by taking into account
13143 only values higher than the pixel.
13145 It accepts the following options:
13152 Limit the maximum change for each plane, default is 65535.
13153 If 0, plane will remain unchanged.
13156 @subsection Commands
13158 This filter supports the all above options as @ref{commands}.
13162 Simple interlacing filter from progressive contents. This interleaves upper (or
13163 lower) lines from odd frames with lower (or upper) lines from even frames,
13164 halving the frame rate and preserving image height.
13167 Original Original New Frame
13168 Frame 'j' Frame 'j+1' (tff)
13169 ========== =========== ==================
13170 Line 0 --------------------> Frame 'j' Line 0
13171 Line 1 Line 1 ----> Frame 'j+1' Line 1
13172 Line 2 ---------------------> Frame 'j' Line 2
13173 Line 3 Line 3 ----> Frame 'j+1' Line 3
13175 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13178 It accepts the following optional parameters:
13182 This determines whether the interlaced frame is taken from the even
13183 (tff - default) or odd (bff) lines of the progressive frame.
13186 Vertical lowpass filter to avoid twitter interlacing and
13187 reduce moire patterns.
13191 Disable vertical lowpass filter
13194 Enable linear filter (default)
13197 Enable complex filter. This will slightly less reduce twitter and moire
13198 but better retain detail and subjective sharpness impression.
13205 Deinterlace input video by applying Donald Graft's adaptive kernel
13206 deinterling. Work on interlaced parts of a video to produce
13207 progressive frames.
13209 The description of the accepted parameters follows.
13213 Set the threshold which affects the filter's tolerance when
13214 determining if a pixel line must be processed. It must be an integer
13215 in the range [0,255] and defaults to 10. A value of 0 will result in
13216 applying the process on every pixels.
13219 Paint pixels exceeding the threshold value to white if set to 1.
13223 Set the fields order. Swap fields if set to 1, leave fields alone if
13227 Enable additional sharpening if set to 1. Default is 0.
13230 Enable twoway sharpening if set to 1. Default is 0.
13233 @subsection Examples
13237 Apply default values:
13239 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13243 Enable additional sharpening:
13249 Paint processed pixels in white:
13257 Slowly update darker pixels.
13259 This filter makes short flashes of light appear longer.
13260 This filter accepts the following options:
13264 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13267 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13270 @section lenscorrection
13272 Correct radial lens distortion
13274 This filter can be used to correct for radial distortion as can result from the use
13275 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13276 one can use tools available for example as part of opencv or simply trial-and-error.
13277 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13278 and extract the k1 and k2 coefficients from the resulting matrix.
13280 Note that effectively the same filter is available in the open-source tools Krita and
13281 Digikam from the KDE project.
13283 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13284 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13285 brightness distribution, so you may want to use both filters together in certain
13286 cases, though you will have to take care of ordering, i.e. whether vignetting should
13287 be applied before or after lens correction.
13289 @subsection Options
13291 The filter accepts the following options:
13295 Relative x-coordinate of the focal point of the image, and thereby the center of the
13296 distortion. This value has a range [0,1] and is expressed as fractions of the image
13297 width. Default is 0.5.
13299 Relative y-coordinate of the focal point of the image, and thereby the center of the
13300 distortion. This value has a range [0,1] and is expressed as fractions of the image
13301 height. Default is 0.5.
13303 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13304 no correction. Default is 0.
13306 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13307 0 means no correction. Default is 0.
13310 The formula that generates the correction is:
13312 @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)
13314 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13315 distances from the focal point in the source and target images, respectively.
13319 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13321 The @code{lensfun} filter requires the camera make, camera model, and lens model
13322 to apply the lens correction. The filter will load the lensfun database and
13323 query it to find the corresponding camera and lens entries in the database. As
13324 long as these entries can be found with the given options, the filter can
13325 perform corrections on frames. Note that incomplete strings will result in the
13326 filter choosing the best match with the given options, and the filter will
13327 output the chosen camera and lens models (logged with level "info"). You must
13328 provide the make, camera model, and lens model as they are required.
13330 The filter accepts the following options:
13334 The make of the camera (for example, "Canon"). This option is required.
13337 The model of the camera (for example, "Canon EOS 100D"). This option is
13341 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13342 option is required.
13345 The type of correction to apply. The following values are valid options:
13349 Enables fixing lens vignetting.
13352 Enables fixing lens geometry. This is the default.
13355 Enables fixing chromatic aberrations.
13358 Enables fixing lens vignetting and lens geometry.
13361 Enables fixing lens vignetting and chromatic aberrations.
13364 Enables fixing both lens geometry and chromatic aberrations.
13367 Enables all possible corrections.
13371 The focal length of the image/video (zoom; expected constant for video). For
13372 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13373 range should be chosen when using that lens. Default 18.
13376 The aperture of the image/video (expected constant for video). Note that
13377 aperture is only used for vignetting correction. Default 3.5.
13379 @item focus_distance
13380 The focus distance of the image/video (expected constant for video). Note that
13381 focus distance is only used for vignetting and only slightly affects the
13382 vignetting correction process. If unknown, leave it at the default value (which
13386 The scale factor which is applied after transformation. After correction the
13387 video is no longer necessarily rectangular. This parameter controls how much of
13388 the resulting image is visible. The value 0 means that a value will be chosen
13389 automatically such that there is little or no unmapped area in the output
13390 image. 1.0 means that no additional scaling is done. Lower values may result
13391 in more of the corrected image being visible, while higher values may avoid
13392 unmapped areas in the output.
13394 @item target_geometry
13395 The target geometry of the output image/video. The following values are valid
13399 @item rectilinear (default)
13402 @item equirectangular
13403 @item fisheye_orthographic
13404 @item fisheye_stereographic
13405 @item fisheye_equisolid
13406 @item fisheye_thoby
13409 Apply the reverse of image correction (instead of correcting distortion, apply
13412 @item interpolation
13413 The type of interpolation used when correcting distortion. The following values
13418 @item linear (default)
13423 @subsection Examples
13427 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13428 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13432 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
13436 Apply the same as before, but only for the first 5 seconds of video.
13439 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
13446 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13447 score between two input videos.
13449 The obtained VMAF score is printed through the logging system.
13451 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13452 After installing the library it can be enabled using:
13453 @code{./configure --enable-libvmaf}.
13454 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13456 The filter has following options:
13460 Set the model path which is to be used for SVM.
13461 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13464 Set the file path to be used to store logs.
13467 Set the format of the log file (csv, json or xml).
13469 @item enable_transform
13470 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13471 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13472 Default value: @code{false}
13475 Invokes the phone model which will generate VMAF scores higher than in the
13476 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13477 Default value: @code{false}
13480 Enables computing psnr along with vmaf.
13481 Default value: @code{false}
13484 Enables computing ssim along with vmaf.
13485 Default value: @code{false}
13488 Enables computing ms_ssim along with vmaf.
13489 Default value: @code{false}
13492 Set the pool method to be used for computing vmaf.
13493 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13496 Set number of threads to be used when computing vmaf.
13497 Default value: @code{0}, which makes use of all available logical processors.
13500 Set interval for frame subsampling used when computing vmaf.
13501 Default value: @code{1}
13503 @item enable_conf_interval
13504 Enables confidence interval.
13505 Default value: @code{false}
13508 This filter also supports the @ref{framesync} options.
13510 @subsection Examples
13513 On the below examples the input file @file{main.mpg} being processed is
13514 compared with the reference file @file{ref.mpg}.
13517 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13521 Example with options:
13523 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13527 Example with options and different containers:
13529 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 -
13535 Limits the pixel components values to the specified range [min, max].
13537 The filter accepts the following options:
13541 Lower bound. Defaults to the lowest allowed value for the input.
13544 Upper bound. Defaults to the highest allowed value for the input.
13547 Specify which planes will be processed. Defaults to all available.
13554 The filter accepts the following options:
13558 Set the number of loops. Setting this value to -1 will result in infinite loops.
13562 Set maximal size in number of frames. Default is 0.
13565 Set first frame of loop. Default is 0.
13568 @subsection Examples
13572 Loop single first frame infinitely:
13574 loop=loop=-1:size=1:start=0
13578 Loop single first frame 10 times:
13580 loop=loop=10:size=1:start=0
13584 Loop 10 first frames 5 times:
13586 loop=loop=5:size=10:start=0
13592 Apply a 1D LUT to an input video.
13594 The filter accepts the following options:
13598 Set the 1D LUT file name.
13600 Currently supported formats:
13609 Select interpolation mode.
13611 Available values are:
13615 Use values from the nearest defined point.
13617 Interpolate values using the linear interpolation.
13619 Interpolate values using the cosine interpolation.
13621 Interpolate values using the cubic interpolation.
13623 Interpolate values using the spline interpolation.
13630 Apply a 3D LUT to an input video.
13632 The filter accepts the following options:
13636 Set the 3D LUT file name.
13638 Currently supported formats:
13652 Select interpolation mode.
13654 Available values are:
13658 Use values from the nearest defined point.
13660 Interpolate values using the 8 points defining a cube.
13662 Interpolate values using a tetrahedron.
13668 Turn certain luma values into transparency.
13670 The filter accepts the following options:
13674 Set the luma which will be used as base for transparency.
13675 Default value is @code{0}.
13678 Set the range of luma values to be keyed out.
13679 Default value is @code{0.01}.
13682 Set the range of softness. Default value is @code{0}.
13683 Use this to control gradual transition from zero to full transparency.
13686 @subsection Commands
13687 This filter supports same @ref{commands} as options.
13688 The command accepts the same syntax of the corresponding option.
13690 If the specified expression is not valid, it is kept at its current
13693 @section lut, lutrgb, lutyuv
13695 Compute a look-up table for binding each pixel component input value
13696 to an output value, and apply it to the input video.
13698 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13699 to an RGB input video.
13701 These filters accept the following parameters:
13704 set first pixel component expression
13706 set second pixel component expression
13708 set third pixel component expression
13710 set fourth pixel component expression, corresponds to the alpha component
13713 set red component expression
13715 set green component expression
13717 set blue component expression
13719 alpha component expression
13722 set Y/luminance component expression
13724 set U/Cb component expression
13726 set V/Cr component expression
13729 Each of them specifies the expression to use for computing the lookup table for
13730 the corresponding pixel component values.
13732 The exact component associated to each of the @var{c*} options depends on the
13735 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13736 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13738 The expressions can contain the following constants and functions:
13743 The input width and height.
13746 The input value for the pixel component.
13749 The input value, clipped to the @var{minval}-@var{maxval} range.
13752 The maximum value for the pixel component.
13755 The minimum value for the pixel component.
13758 The negated value for the pixel component value, clipped to the
13759 @var{minval}-@var{maxval} range; it corresponds to the expression
13760 "maxval-clipval+minval".
13763 The computed value in @var{val}, clipped to the
13764 @var{minval}-@var{maxval} range.
13766 @item gammaval(gamma)
13767 The computed gamma correction value of the pixel component value,
13768 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13770 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13774 All expressions default to "val".
13776 @subsection Examples
13780 Negate input video:
13782 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13783 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13786 The above is the same as:
13788 lutrgb="r=negval:g=negval:b=negval"
13789 lutyuv="y=negval:u=negval:v=negval"
13799 Remove chroma components, turning the video into a graytone image:
13801 lutyuv="u=128:v=128"
13805 Apply a luma burning effect:
13811 Remove green and blue components:
13817 Set a constant alpha channel value on input:
13819 format=rgba,lutrgb=a="maxval-minval/2"
13823 Correct luminance gamma by a factor of 0.5:
13825 lutyuv=y=gammaval(0.5)
13829 Discard least significant bits of luma:
13831 lutyuv=y='bitand(val, 128+64+32)'
13835 Technicolor like effect:
13837 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13841 @section lut2, tlut2
13843 The @code{lut2} filter takes two input streams and outputs one
13846 The @code{tlut2} (time lut2) filter takes two consecutive frames
13847 from one single stream.
13849 This filter accepts the following parameters:
13852 set first pixel component expression
13854 set second pixel component expression
13856 set third pixel component expression
13858 set fourth pixel component expression, corresponds to the alpha component
13861 set output bit depth, only available for @code{lut2} filter. By default is 0,
13862 which means bit depth is automatically picked from first input format.
13865 The @code{lut2} filter also supports the @ref{framesync} options.
13867 Each of them specifies the expression to use for computing the lookup table for
13868 the corresponding pixel component values.
13870 The exact component associated to each of the @var{c*} options depends on the
13873 The expressions can contain the following constants:
13878 The input width and height.
13881 The first input value for the pixel component.
13884 The second input value for the pixel component.
13887 The first input video bit depth.
13890 The second input video bit depth.
13893 All expressions default to "x".
13895 @subsection Examples
13899 Highlight differences between two RGB video streams:
13901 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)'
13905 Highlight differences between two YUV video streams:
13907 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)'
13911 Show max difference between two video streams:
13913 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)))'
13917 @section maskedclamp
13919 Clamp the first input stream with the second input and third input stream.
13921 Returns the value of first stream to be between second input
13922 stream - @code{undershoot} and third input stream + @code{overshoot}.
13924 This filter accepts the following options:
13927 Default value is @code{0}.
13930 Default value is @code{0}.
13933 Set which planes will be processed as bitmap, unprocessed planes will be
13934 copied from first stream.
13935 By default value 0xf, all planes will be processed.
13940 Merge the second and third input stream into output stream using absolute differences
13941 between second input stream and first input stream and absolute difference between
13942 third input stream and first input stream. The picked value will be from second input
13943 stream if second absolute difference is greater than first one or from third input stream
13946 This filter accepts the following options:
13949 Set which planes will be processed as bitmap, unprocessed planes will be
13950 copied from first stream.
13951 By default value 0xf, all planes will be processed.
13954 @section maskedmerge
13956 Merge the first input stream with the second input stream using per pixel
13957 weights in the third input stream.
13959 A value of 0 in the third stream pixel component means that pixel component
13960 from first stream is returned unchanged, while maximum value (eg. 255 for
13961 8-bit videos) means that pixel component from second stream is returned
13962 unchanged. Intermediate values define the amount of merging between both
13963 input stream's pixel components.
13965 This filter accepts the following options:
13968 Set which planes will be processed as bitmap, unprocessed planes will be
13969 copied from first stream.
13970 By default value 0xf, all planes will be processed.
13975 Merge the second and third input stream into output stream using absolute differences
13976 between second input stream and first input stream and absolute difference between
13977 third input stream and first input stream. The picked value will be from second input
13978 stream if second absolute difference is less than first one or from third input stream
13981 This filter accepts the following options:
13984 Set which planes will be processed as bitmap, unprocessed planes will be
13985 copied from first stream.
13986 By default value 0xf, all planes will be processed.
13989 @section maskedthreshold
13990 Pick pixels comparing absolute difference of two video streams with fixed
13993 If absolute difference between pixel component of first and second video
13994 stream is equal or lower than user supplied threshold than pixel component
13995 from first video stream is picked, otherwise pixel component from second
13996 video stream is picked.
13998 This filter accepts the following options:
14001 Set threshold used when picking pixels from absolute difference from two input
14005 Set which planes will be processed as bitmap, unprocessed planes will be
14006 copied from second stream.
14007 By default value 0xf, all planes will be processed.
14011 Create mask from input video.
14013 For example it is useful to create motion masks after @code{tblend} filter.
14015 This filter accepts the following options:
14019 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14022 Set high threshold. Any pixel component higher than this value will be set to max value
14023 allowed for current pixel format.
14026 Set planes to filter, by default all available planes are filtered.
14029 Fill all frame pixels with this value.
14032 Set max average pixel value for frame. If sum of all pixel components is higher that this
14033 average, output frame will be completely filled with value set by @var{fill} option.
14034 Typically useful for scene changes when used in combination with @code{tblend} filter.
14039 Apply motion-compensation deinterlacing.
14041 It needs one field per frame as input and must thus be used together
14042 with yadif=1/3 or equivalent.
14044 This filter accepts the following options:
14047 Set the deinterlacing mode.
14049 It accepts one of the following values:
14054 use iterative motion estimation
14056 like @samp{slow}, but use multiple reference frames.
14058 Default value is @samp{fast}.
14061 Set the picture field parity assumed for the input video. It must be
14062 one of the following values:
14066 assume top field first
14068 assume bottom field first
14071 Default value is @samp{bff}.
14074 Set per-block quantization parameter (QP) used by the internal
14077 Higher values should result in a smoother motion vector field but less
14078 optimal individual vectors. Default value is 1.
14083 Pick median pixel from certain rectangle defined by radius.
14085 This filter accepts the following options:
14089 Set horizontal radius size. Default value is @code{1}.
14090 Allowed range is integer from 1 to 127.
14093 Set which planes to process. Default is @code{15}, which is all available planes.
14096 Set vertical radius size. Default value is @code{0}.
14097 Allowed range is integer from 0 to 127.
14098 If it is 0, value will be picked from horizontal @code{radius} option.
14101 Set median percentile. Default value is @code{0.5}.
14102 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14103 minimum values, and @code{1} maximum values.
14106 @subsection Commands
14107 This filter supports same @ref{commands} as options.
14108 The command accepts the same syntax of the corresponding option.
14110 If the specified expression is not valid, it is kept at its current
14113 @section mergeplanes
14115 Merge color channel components from several video streams.
14117 The filter accepts up to 4 input streams, and merge selected input
14118 planes to the output video.
14120 This filter accepts the following options:
14123 Set input to output plane mapping. Default is @code{0}.
14125 The mappings is specified as a bitmap. It should be specified as a
14126 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14127 mapping for the first plane of the output stream. 'A' sets the number of
14128 the input stream to use (from 0 to 3), and 'a' the plane number of the
14129 corresponding input to use (from 0 to 3). The rest of the mappings is
14130 similar, 'Bb' describes the mapping for the output stream second
14131 plane, 'Cc' describes the mapping for the output stream third plane and
14132 'Dd' describes the mapping for the output stream fourth plane.
14135 Set output pixel format. Default is @code{yuva444p}.
14138 @subsection Examples
14142 Merge three gray video streams of same width and height into single video stream:
14144 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14148 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14150 [a0][a1]mergeplanes=0x00010210:yuva444p
14154 Swap Y and A plane in yuva444p stream:
14156 format=yuva444p,mergeplanes=0x03010200:yuva444p
14160 Swap U and V plane in yuv420p stream:
14162 format=yuv420p,mergeplanes=0x000201:yuv420p
14166 Cast a rgb24 clip to yuv444p:
14168 format=rgb24,mergeplanes=0x000102:yuv444p
14174 Estimate and export motion vectors using block matching algorithms.
14175 Motion vectors are stored in frame side data to be used by other filters.
14177 This filter accepts the following options:
14180 Specify the motion estimation method. Accepts one of the following values:
14184 Exhaustive search algorithm.
14186 Three step search algorithm.
14188 Two dimensional logarithmic search algorithm.
14190 New three step search algorithm.
14192 Four step search algorithm.
14194 Diamond search algorithm.
14196 Hexagon-based search algorithm.
14198 Enhanced predictive zonal search algorithm.
14200 Uneven multi-hexagon search algorithm.
14202 Default value is @samp{esa}.
14205 Macroblock size. Default @code{16}.
14208 Search parameter. Default @code{7}.
14211 @section midequalizer
14213 Apply Midway Image Equalization effect using two video streams.
14215 Midway Image Equalization adjusts a pair of images to have the same
14216 histogram, while maintaining their dynamics as much as possible. It's
14217 useful for e.g. matching exposures from a pair of stereo cameras.
14219 This filter has two inputs and one output, which must be of same pixel format, but
14220 may be of different sizes. The output of filter is first input adjusted with
14221 midway histogram of both inputs.
14223 This filter accepts the following option:
14227 Set which planes to process. Default is @code{15}, which is all available planes.
14230 @section minterpolate
14232 Convert the video to specified frame rate using motion interpolation.
14234 This filter accepts the following options:
14237 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}.
14240 Motion interpolation mode. Following values are accepted:
14243 Duplicate previous or next frame for interpolating new ones.
14245 Blend source frames. Interpolated frame is mean of previous and next frames.
14247 Motion compensated interpolation. Following options are effective when this mode is selected:
14251 Motion compensation mode. Following values are accepted:
14254 Overlapped block motion compensation.
14256 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14258 Default mode is @samp{obmc}.
14261 Motion estimation mode. Following values are accepted:
14264 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14266 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14268 Default mode is @samp{bilat}.
14271 The algorithm to be used for motion estimation. Following values are accepted:
14274 Exhaustive search algorithm.
14276 Three step search algorithm.
14278 Two dimensional logarithmic search algorithm.
14280 New three step search algorithm.
14282 Four step search algorithm.
14284 Diamond search algorithm.
14286 Hexagon-based search algorithm.
14288 Enhanced predictive zonal search algorithm.
14290 Uneven multi-hexagon search algorithm.
14292 Default algorithm is @samp{epzs}.
14295 Macroblock size. Default @code{16}.
14298 Motion estimation search parameter. Default @code{32}.
14301 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).
14306 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:
14309 Disable scene change detection.
14311 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14313 Default method is @samp{fdiff}.
14315 @item scd_threshold
14316 Scene change detection threshold. Default is @code{10.}.
14321 Mix several video input streams into one video stream.
14323 A description of the accepted options follows.
14327 The number of inputs. If unspecified, it defaults to 2.
14330 Specify weight of each input video stream as sequence.
14331 Each weight is separated by space. If number of weights
14332 is smaller than number of @var{frames} last specified
14333 weight will be used for all remaining unset weights.
14336 Specify scale, if it is set it will be multiplied with sum
14337 of each weight multiplied with pixel values to give final destination
14338 pixel value. By default @var{scale} is auto scaled to sum of weights.
14341 Specify how end of stream is determined.
14344 The duration of the longest input. (default)
14347 The duration of the shortest input.
14350 The duration of the first input.
14354 @section mpdecimate
14356 Drop frames that do not differ greatly from the previous frame in
14357 order to reduce frame rate.
14359 The main use of this filter is for very-low-bitrate encoding
14360 (e.g. streaming over dialup modem), but it could in theory be used for
14361 fixing movies that were inverse-telecined incorrectly.
14363 A description of the accepted options follows.
14367 Set the maximum number of consecutive frames which can be dropped (if
14368 positive), or the minimum interval between dropped frames (if
14369 negative). If the value is 0, the frame is dropped disregarding the
14370 number of previous sequentially dropped frames.
14372 Default value is 0.
14377 Set the dropping threshold values.
14379 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14380 represent actual pixel value differences, so a threshold of 64
14381 corresponds to 1 unit of difference for each pixel, or the same spread
14382 out differently over the block.
14384 A frame is a candidate for dropping if no 8x8 blocks differ by more
14385 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14386 meaning the whole image) differ by more than a threshold of @option{lo}.
14388 Default value for @option{hi} is 64*12, default value for @option{lo} is
14389 64*5, and default value for @option{frac} is 0.33.
14395 Negate (invert) the input video.
14397 It accepts the following option:
14402 With value 1, it negates the alpha component, if present. Default value is 0.
14408 Denoise frames using Non-Local Means algorithm.
14410 Each pixel is adjusted by looking for other pixels with similar contexts. This
14411 context similarity is defined by comparing their surrounding patches of size
14412 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14415 Note that the research area defines centers for patches, which means some
14416 patches will be made of pixels outside that research area.
14418 The filter accepts the following options.
14422 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14425 Set patch size. Default is 7. Must be odd number in range [0, 99].
14428 Same as @option{p} but for chroma planes.
14430 The default value is @var{0} and means automatic.
14433 Set research size. Default is 15. Must be odd number in range [0, 99].
14436 Same as @option{r} but for chroma planes.
14438 The default value is @var{0} and means automatic.
14443 Deinterlace video using neural network edge directed interpolation.
14445 This filter accepts the following options:
14449 Mandatory option, without binary file filter can not work.
14450 Currently file can be found here:
14451 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14454 Set which frames to deinterlace, by default it is @code{all}.
14455 Can be @code{all} or @code{interlaced}.
14458 Set mode of operation.
14460 Can be one of the following:
14464 Use frame flags, both fields.
14466 Use frame flags, single field.
14468 Use top field only.
14470 Use bottom field only.
14472 Use both fields, top first.
14474 Use both fields, bottom first.
14478 Set which planes to process, by default filter process all frames.
14481 Set size of local neighborhood around each pixel, used by the predictor neural
14484 Can be one of the following:
14497 Set the number of neurons in predictor neural network.
14498 Can be one of the following:
14509 Controls the number of different neural network predictions that are blended
14510 together to compute the final output value. Can be @code{fast}, default or
14514 Set which set of weights to use in the predictor.
14515 Can be one of the following:
14519 weights trained to minimize absolute error
14521 weights trained to minimize squared error
14525 Controls whether or not the prescreener neural network is used to decide
14526 which pixels should be processed by the predictor neural network and which
14527 can be handled by simple cubic interpolation.
14528 The prescreener is trained to know whether cubic interpolation will be
14529 sufficient for a pixel or whether it should be predicted by the predictor nn.
14530 The computational complexity of the prescreener nn is much less than that of
14531 the predictor nn. Since most pixels can be handled by cubic interpolation,
14532 using the prescreener generally results in much faster processing.
14533 The prescreener is pretty accurate, so the difference between using it and not
14534 using it is almost always unnoticeable.
14536 Can be one of the following:
14544 Default is @code{new}.
14547 Set various debugging flags.
14552 Force libavfilter not to use any of the specified pixel formats for the
14553 input to the next filter.
14555 It accepts the following parameters:
14559 A '|'-separated list of pixel format names, such as
14560 pix_fmts=yuv420p|monow|rgb24".
14564 @subsection Examples
14568 Force libavfilter to use a format different from @var{yuv420p} for the
14569 input to the vflip filter:
14571 noformat=pix_fmts=yuv420p,vflip
14575 Convert the input video to any of the formats not contained in the list:
14577 noformat=yuv420p|yuv444p|yuv410p
14583 Add noise on video input frame.
14585 The filter accepts the following options:
14593 Set noise seed for specific pixel component or all pixel components in case
14594 of @var{all_seed}. Default value is @code{123457}.
14596 @item all_strength, alls
14597 @item c0_strength, c0s
14598 @item c1_strength, c1s
14599 @item c2_strength, c2s
14600 @item c3_strength, c3s
14601 Set noise strength for specific pixel component or all pixel components in case
14602 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14604 @item all_flags, allf
14605 @item c0_flags, c0f
14606 @item c1_flags, c1f
14607 @item c2_flags, c2f
14608 @item c3_flags, c3f
14609 Set pixel component flags or set flags for all components if @var{all_flags}.
14610 Available values for component flags are:
14613 averaged temporal noise (smoother)
14615 mix random noise with a (semi)regular pattern
14617 temporal noise (noise pattern changes between frames)
14619 uniform noise (gaussian otherwise)
14623 @subsection Examples
14625 Add temporal and uniform noise to input video:
14627 noise=alls=20:allf=t+u
14632 Normalize RGB video (aka histogram stretching, contrast stretching).
14633 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14635 For each channel of each frame, the filter computes the input range and maps
14636 it linearly to the user-specified output range. The output range defaults
14637 to the full dynamic range from pure black to pure white.
14639 Temporal smoothing can be used on the input range to reduce flickering (rapid
14640 changes in brightness) caused when small dark or bright objects enter or leave
14641 the scene. This is similar to the auto-exposure (automatic gain control) on a
14642 video camera, and, like a video camera, it may cause a period of over- or
14643 under-exposure of the video.
14645 The R,G,B channels can be normalized independently, which may cause some
14646 color shifting, or linked together as a single channel, which prevents
14647 color shifting. Linked normalization preserves hue. Independent normalization
14648 does not, so it can be used to remove some color casts. Independent and linked
14649 normalization can be combined in any ratio.
14651 The normalize filter accepts the following options:
14656 Colors which define the output range. The minimum input value is mapped to
14657 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14658 The defaults are black and white respectively. Specifying white for
14659 @var{blackpt} and black for @var{whitept} will give color-inverted,
14660 normalized video. Shades of grey can be used to reduce the dynamic range
14661 (contrast). Specifying saturated colors here can create some interesting
14665 The number of previous frames to use for temporal smoothing. The input range
14666 of each channel is smoothed using a rolling average over the current frame
14667 and the @var{smoothing} previous frames. The default is 0 (no temporal
14671 Controls the ratio of independent (color shifting) channel normalization to
14672 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14673 independent. Defaults to 1.0 (fully independent).
14676 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14677 expensive no-op. Defaults to 1.0 (full strength).
14681 @subsection Commands
14682 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14683 The command accepts the same syntax of the corresponding option.
14685 If the specified expression is not valid, it is kept at its current
14688 @subsection Examples
14690 Stretch video contrast to use the full dynamic range, with no temporal
14691 smoothing; may flicker depending on the source content:
14693 normalize=blackpt=black:whitept=white:smoothing=0
14696 As above, but with 50 frames of temporal smoothing; flicker should be
14697 reduced, depending on the source content:
14699 normalize=blackpt=black:whitept=white:smoothing=50
14702 As above, but with hue-preserving linked channel normalization:
14704 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14707 As above, but with half strength:
14709 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14712 Map the darkest input color to red, the brightest input color to cyan:
14714 normalize=blackpt=red:whitept=cyan
14719 Pass the video source unchanged to the output.
14722 Optical Character Recognition
14724 This filter uses Tesseract for optical character recognition. To enable
14725 compilation of this filter, you need to configure FFmpeg with
14726 @code{--enable-libtesseract}.
14728 It accepts the following options:
14732 Set datapath to tesseract data. Default is to use whatever was
14733 set at installation.
14736 Set language, default is "eng".
14739 Set character whitelist.
14742 Set character blacklist.
14745 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14746 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14750 Apply a video transform using libopencv.
14752 To enable this filter, install the libopencv library and headers and
14753 configure FFmpeg with @code{--enable-libopencv}.
14755 It accepts the following parameters:
14760 The name of the libopencv filter to apply.
14762 @item filter_params
14763 The parameters to pass to the libopencv filter. If not specified, the default
14764 values are assumed.
14768 Refer to the official libopencv documentation for more precise
14770 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14772 Several libopencv filters are supported; see the following subsections.
14777 Dilate an image by using a specific structuring element.
14778 It corresponds to the libopencv function @code{cvDilate}.
14780 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14782 @var{struct_el} represents a structuring element, and has the syntax:
14783 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14785 @var{cols} and @var{rows} represent the number of columns and rows of
14786 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14787 point, and @var{shape} the shape for the structuring element. @var{shape}
14788 must be "rect", "cross", "ellipse", or "custom".
14790 If the value for @var{shape} is "custom", it must be followed by a
14791 string of the form "=@var{filename}". The file with name
14792 @var{filename} is assumed to represent a binary image, with each
14793 printable character corresponding to a bright pixel. When a custom
14794 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14795 or columns and rows of the read file are assumed instead.
14797 The default value for @var{struct_el} is "3x3+0x0/rect".
14799 @var{nb_iterations} specifies the number of times the transform is
14800 applied to the image, and defaults to 1.
14804 # Use the default values
14807 # Dilate using a structuring element with a 5x5 cross, iterating two times
14808 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14810 # Read the shape from the file diamond.shape, iterating two times.
14811 # The file diamond.shape may contain a pattern of characters like this
14817 # The specified columns and rows are ignored
14818 # but the anchor point coordinates are not
14819 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14824 Erode an image by using a specific structuring element.
14825 It corresponds to the libopencv function @code{cvErode}.
14827 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14828 with the same syntax and semantics as the @ref{dilate} filter.
14832 Smooth the input video.
14834 The filter takes the following parameters:
14835 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14837 @var{type} is the type of smooth filter to apply, and must be one of
14838 the following values: "blur", "blur_no_scale", "median", "gaussian",
14839 or "bilateral". The default value is "gaussian".
14841 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14842 depends on the smooth type. @var{param1} and
14843 @var{param2} accept integer positive values or 0. @var{param3} and
14844 @var{param4} accept floating point values.
14846 The default value for @var{param1} is 3. The default value for the
14847 other parameters is 0.
14849 These parameters correspond to the parameters assigned to the
14850 libopencv function @code{cvSmooth}.
14852 @section oscilloscope
14854 2D Video Oscilloscope.
14856 Useful to measure spatial impulse, step responses, chroma delays, etc.
14858 It accepts the following parameters:
14862 Set scope center x position.
14865 Set scope center y position.
14868 Set scope size, relative to frame diagonal.
14871 Set scope tilt/rotation.
14877 Set trace center x position.
14880 Set trace center y position.
14883 Set trace width, relative to width of frame.
14886 Set trace height, relative to height of frame.
14889 Set which components to trace. By default it traces first three components.
14892 Draw trace grid. By default is enabled.
14895 Draw some statistics. By default is enabled.
14898 Draw scope. By default is enabled.
14901 @subsection Commands
14902 This filter supports same @ref{commands} as options.
14903 The command accepts the same syntax of the corresponding option.
14905 If the specified expression is not valid, it is kept at its current
14908 @subsection Examples
14912 Inspect full first row of video frame.
14914 oscilloscope=x=0.5:y=0:s=1
14918 Inspect full last row of video frame.
14920 oscilloscope=x=0.5:y=1:s=1
14924 Inspect full 5th line of video frame of height 1080.
14926 oscilloscope=x=0.5:y=5/1080:s=1
14930 Inspect full last column of video frame.
14932 oscilloscope=x=1:y=0.5:s=1:t=1
14940 Overlay one video on top of another.
14942 It takes two inputs and has one output. The first input is the "main"
14943 video on which the second input is overlaid.
14945 It accepts the following parameters:
14947 A description of the accepted options follows.
14952 Set the expression for the x and y coordinates of the overlaid video
14953 on the main video. Default value is "0" for both expressions. In case
14954 the expression is invalid, it is set to a huge value (meaning that the
14955 overlay will not be displayed within the output visible area).
14958 See @ref{framesync}.
14961 Set when the expressions for @option{x}, and @option{y} are evaluated.
14963 It accepts the following values:
14966 only evaluate expressions once during the filter initialization or
14967 when a command is processed
14970 evaluate expressions for each incoming frame
14973 Default value is @samp{frame}.
14976 See @ref{framesync}.
14979 Set the format for the output video.
14981 It accepts the following values:
14984 force YUV420 output
14987 force YUV420p10 output
14990 force YUV422 output
14993 force YUV422p10 output
14996 force YUV444 output
14999 force packed RGB output
15002 force planar RGB output
15005 automatically pick format
15008 Default value is @samp{yuv420}.
15011 See @ref{framesync}.
15014 Set format of alpha of the overlaid video, it can be @var{straight} or
15015 @var{premultiplied}. Default is @var{straight}.
15018 The @option{x}, and @option{y} expressions can contain the following
15024 The main input width and height.
15028 The overlay input width and height.
15032 The computed values for @var{x} and @var{y}. They are evaluated for
15037 horizontal and vertical chroma subsample values of the output
15038 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15042 the number of input frame, starting from 0
15045 the position in the file of the input frame, NAN if unknown
15048 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15052 This filter also supports the @ref{framesync} options.
15054 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15055 when evaluation is done @emph{per frame}, and will evaluate to NAN
15056 when @option{eval} is set to @samp{init}.
15058 Be aware that frames are taken from each input video in timestamp
15059 order, hence, if their initial timestamps differ, it is a good idea
15060 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15061 have them begin in the same zero timestamp, as the example for
15062 the @var{movie} filter does.
15064 You can chain together more overlays but you should test the
15065 efficiency of such approach.
15067 @subsection Commands
15069 This filter supports the following commands:
15073 Modify the x and y of the overlay input.
15074 The command accepts the same syntax of the corresponding option.
15076 If the specified expression is not valid, it is kept at its current
15080 @subsection Examples
15084 Draw the overlay at 10 pixels from the bottom right corner of the main
15087 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15090 Using named options the example above becomes:
15092 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15096 Insert a transparent PNG logo in the bottom left corner of the input,
15097 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15099 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15103 Insert 2 different transparent PNG logos (second logo on bottom
15104 right corner) using the @command{ffmpeg} tool:
15106 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
15110 Add a transparent color layer on top of the main video; @code{WxH}
15111 must specify the size of the main input to the overlay filter:
15113 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15117 Play an original video and a filtered version (here with the deshake
15118 filter) side by side using the @command{ffplay} tool:
15120 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15123 The above command is the same as:
15125 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15129 Make a sliding overlay appearing from the left to the right top part of the
15130 screen starting since time 2:
15132 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15136 Compose output by putting two input videos side to side:
15138 ffmpeg -i left.avi -i right.avi -filter_complex "
15139 nullsrc=size=200x100 [background];
15140 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15141 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15142 [background][left] overlay=shortest=1 [background+left];
15143 [background+left][right] overlay=shortest=1:x=100 [left+right]
15148 Mask 10-20 seconds of a video by applying the delogo filter to a section
15150 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15151 -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]'
15156 Chain several overlays in cascade:
15158 nullsrc=s=200x200 [bg];
15159 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15160 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15161 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15162 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15163 [in3] null, [mid2] overlay=100:100 [out0]
15168 @anchor{overlay_cuda}
15169 @section overlay_cuda
15171 Overlay one video on top of another.
15173 This is the CUDA variant of the @ref{overlay} filter.
15174 It only accepts CUDA frames. The underlying input pixel formats have to match.
15176 It takes two inputs and has one output. The first input is the "main"
15177 video on which the second input is overlaid.
15179 It accepts the following parameters:
15184 Set the x and y coordinates of the overlaid video on the main video.
15185 Default value is "0" for both expressions.
15188 See @ref{framesync}.
15191 See @ref{framesync}.
15194 See @ref{framesync}.
15198 This filter also supports the @ref{framesync} options.
15202 Apply Overcomplete Wavelet denoiser.
15204 The filter accepts the following options:
15210 Larger depth values will denoise lower frequency components more, but
15211 slow down filtering.
15213 Must be an int in the range 8-16, default is @code{8}.
15215 @item luma_strength, ls
15218 Must be a double value in the range 0-1000, default is @code{1.0}.
15220 @item chroma_strength, cs
15221 Set chroma strength.
15223 Must be a double value in the range 0-1000, default is @code{1.0}.
15229 Add paddings to the input image, and place the original input at the
15230 provided @var{x}, @var{y} coordinates.
15232 It accepts the following parameters:
15237 Specify an expression for the size of the output image with the
15238 paddings added. If the value for @var{width} or @var{height} is 0, the
15239 corresponding input size is used for the output.
15241 The @var{width} expression can reference the value set by the
15242 @var{height} expression, and vice versa.
15244 The default value of @var{width} and @var{height} is 0.
15248 Specify the offsets to place the input image at within the padded area,
15249 with respect to the top/left border of the output image.
15251 The @var{x} expression can reference the value set by the @var{y}
15252 expression, and vice versa.
15254 The default value of @var{x} and @var{y} is 0.
15256 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15257 so the input image is centered on the padded area.
15260 Specify the color of the padded area. For the syntax of this option,
15261 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15262 manual,ffmpeg-utils}.
15264 The default value of @var{color} is "black".
15267 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15269 It accepts the following values:
15273 Only evaluate expressions once during the filter initialization or when
15274 a command is processed.
15277 Evaluate expressions for each incoming frame.
15281 Default value is @samp{init}.
15284 Pad to aspect instead to a resolution.
15288 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15289 options are expressions containing the following constants:
15294 The input video width and height.
15298 These are the same as @var{in_w} and @var{in_h}.
15302 The output width and height (the size of the padded area), as
15303 specified by the @var{width} and @var{height} expressions.
15307 These are the same as @var{out_w} and @var{out_h}.
15311 The x and y offsets as specified by the @var{x} and @var{y}
15312 expressions, or NAN if not yet specified.
15315 same as @var{iw} / @var{ih}
15318 input sample aspect ratio
15321 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15325 The horizontal and vertical chroma subsample values. For example for the
15326 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15329 @subsection Examples
15333 Add paddings with the color "violet" to the input video. The output video
15334 size is 640x480, and the top-left corner of the input video is placed at
15337 pad=640:480:0:40:violet
15340 The example above is equivalent to the following command:
15342 pad=width=640:height=480:x=0:y=40:color=violet
15346 Pad the input to get an output with dimensions increased by 3/2,
15347 and put the input video at the center of the padded area:
15349 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15353 Pad the input to get a squared output with size equal to the maximum
15354 value between the input width and height, and put the input video at
15355 the center of the padded area:
15357 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15361 Pad the input to get a final w/h ratio of 16:9:
15363 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15367 In case of anamorphic video, in order to set the output display aspect
15368 correctly, it is necessary to use @var{sar} in the expression,
15369 according to the relation:
15371 (ih * X / ih) * sar = output_dar
15372 X = output_dar / sar
15375 Thus the previous example needs to be modified to:
15377 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15381 Double the output size and put the input video in the bottom-right
15382 corner of the output padded area:
15384 pad="2*iw:2*ih:ow-iw:oh-ih"
15388 @anchor{palettegen}
15389 @section palettegen
15391 Generate one palette for a whole video stream.
15393 It accepts the following options:
15397 Set the maximum number of colors to quantize in the palette.
15398 Note: the palette will still contain 256 colors; the unused palette entries
15401 @item reserve_transparent
15402 Create a palette of 255 colors maximum and reserve the last one for
15403 transparency. Reserving the transparency color is useful for GIF optimization.
15404 If not set, the maximum of colors in the palette will be 256. You probably want
15405 to disable this option for a standalone image.
15408 @item transparency_color
15409 Set the color that will be used as background for transparency.
15412 Set statistics mode.
15414 It accepts the following values:
15417 Compute full frame histograms.
15419 Compute histograms only for the part that differs from previous frame. This
15420 might be relevant to give more importance to the moving part of your input if
15421 the background is static.
15423 Compute new histogram for each frame.
15426 Default value is @var{full}.
15429 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15430 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15431 color quantization of the palette. This information is also visible at
15432 @var{info} logging level.
15434 @subsection Examples
15438 Generate a representative palette of a given video using @command{ffmpeg}:
15440 ffmpeg -i input.mkv -vf palettegen palette.png
15444 @section paletteuse
15446 Use a palette to downsample an input video stream.
15448 The filter takes two inputs: one video stream and a palette. The palette must
15449 be a 256 pixels image.
15451 It accepts the following options:
15455 Select dithering mode. Available algorithms are:
15458 Ordered 8x8 bayer dithering (deterministic)
15460 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15461 Note: this dithering is sometimes considered "wrong" and is included as a
15463 @item floyd_steinberg
15464 Floyd and Steingberg dithering (error diffusion)
15466 Frankie Sierra dithering v2 (error diffusion)
15468 Frankie Sierra dithering v2 "Lite" (error diffusion)
15471 Default is @var{sierra2_4a}.
15474 When @var{bayer} dithering is selected, this option defines the scale of the
15475 pattern (how much the crosshatch pattern is visible). A low value means more
15476 visible pattern for less banding, and higher value means less visible pattern
15477 at the cost of more banding.
15479 The option must be an integer value in the range [0,5]. Default is @var{2}.
15482 If set, define the zone to process
15486 Only the changing rectangle will be reprocessed. This is similar to GIF
15487 cropping/offsetting compression mechanism. This option can be useful for speed
15488 if only a part of the image is changing, and has use cases such as limiting the
15489 scope of the error diffusal @option{dither} to the rectangle that bounds the
15490 moving scene (it leads to more deterministic output if the scene doesn't change
15491 much, and as a result less moving noise and better GIF compression).
15494 Default is @var{none}.
15497 Take new palette for each output frame.
15499 @item alpha_threshold
15500 Sets the alpha threshold for transparency. Alpha values above this threshold
15501 will be treated as completely opaque, and values below this threshold will be
15502 treated as completely transparent.
15504 The option must be an integer value in the range [0,255]. Default is @var{128}.
15507 @subsection Examples
15511 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15512 using @command{ffmpeg}:
15514 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15518 @section perspective
15520 Correct perspective of video not recorded perpendicular to the screen.
15522 A description of the accepted parameters follows.
15533 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15534 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15535 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15536 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15537 then the corners of the source will be sent to the specified coordinates.
15539 The expressions can use the following variables:
15544 the width and height of video frame.
15548 Output frame count.
15551 @item interpolation
15552 Set interpolation for perspective correction.
15554 It accepts the following values:
15560 Default value is @samp{linear}.
15563 Set interpretation of coordinate options.
15565 It accepts the following values:
15569 Send point in the source specified by the given coordinates to
15570 the corners of the destination.
15572 @item 1, destination
15574 Send the corners of the source to the point in the destination specified
15575 by the given coordinates.
15577 Default value is @samp{source}.
15581 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15583 It accepts the following values:
15586 only evaluate expressions once during the filter initialization or
15587 when a command is processed
15590 evaluate expressions for each incoming frame
15593 Default value is @samp{init}.
15598 Delay interlaced video by one field time so that the field order changes.
15600 The intended use is to fix PAL movies that have been captured with the
15601 opposite field order to the film-to-video transfer.
15603 A description of the accepted parameters follows.
15609 It accepts the following values:
15612 Capture field order top-first, transfer bottom-first.
15613 Filter will delay the bottom field.
15616 Capture field order bottom-first, transfer top-first.
15617 Filter will delay the top field.
15620 Capture and transfer with the same field order. This mode only exists
15621 for the documentation of the other options to refer to, but if you
15622 actually select it, the filter will faithfully do nothing.
15625 Capture field order determined automatically by field flags, transfer
15627 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15628 basis using field flags. If no field information is available,
15629 then this works just like @samp{u}.
15632 Capture unknown or varying, transfer opposite.
15633 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15634 analyzing the images and selecting the alternative that produces best
15635 match between the fields.
15638 Capture top-first, transfer unknown or varying.
15639 Filter selects among @samp{t} and @samp{p} using image analysis.
15642 Capture bottom-first, transfer unknown or varying.
15643 Filter selects among @samp{b} and @samp{p} using image analysis.
15646 Capture determined by field flags, transfer unknown or varying.
15647 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15648 image analysis. If no field information is available, then this works just
15649 like @samp{U}. This is the default mode.
15652 Both capture and transfer unknown or varying.
15653 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15657 @section photosensitivity
15658 Reduce various flashes in video, so to help users with epilepsy.
15660 It accepts the following options:
15663 Set how many frames to use when filtering. Default is 30.
15666 Set detection threshold factor. Default is 1.
15670 Set how many pixels to skip when sampling frames. Default is 1.
15671 Allowed range is from 1 to 1024.
15674 Leave frames unchanged. Default is disabled.
15677 @section pixdesctest
15679 Pixel format descriptor test filter, mainly useful for internal
15680 testing. The output video should be equal to the input video.
15684 format=monow, pixdesctest
15687 can be used to test the monowhite pixel format descriptor definition.
15691 Display sample values of color channels. Mainly useful for checking color
15692 and levels. Minimum supported resolution is 640x480.
15694 The filters accept the following options:
15698 Set scope X position, relative offset on X axis.
15701 Set scope Y position, relative offset on Y axis.
15710 Set window opacity. This window also holds statistics about pixel area.
15713 Set window X position, relative offset on X axis.
15716 Set window Y position, relative offset on Y axis.
15721 Enable the specified chain of postprocessing subfilters using libpostproc. This
15722 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15723 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15724 Each subfilter and some options have a short and a long name that can be used
15725 interchangeably, i.e. dr/dering are the same.
15727 The filters accept the following options:
15731 Set postprocessing subfilters string.
15734 All subfilters share common options to determine their scope:
15738 Honor the quality commands for this subfilter.
15741 Do chrominance filtering, too (default).
15744 Do luminance filtering only (no chrominance).
15747 Do chrominance filtering only (no luminance).
15750 These options can be appended after the subfilter name, separated by a '|'.
15752 Available subfilters are:
15755 @item hb/hdeblock[|difference[|flatness]]
15756 Horizontal deblocking filter
15759 Difference factor where higher values mean more deblocking (default: @code{32}).
15761 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15764 @item vb/vdeblock[|difference[|flatness]]
15765 Vertical deblocking filter
15768 Difference factor where higher values mean more deblocking (default: @code{32}).
15770 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15773 @item ha/hadeblock[|difference[|flatness]]
15774 Accurate horizontal deblocking filter
15777 Difference factor where higher values mean more deblocking (default: @code{32}).
15779 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15782 @item va/vadeblock[|difference[|flatness]]
15783 Accurate vertical deblocking filter
15786 Difference factor where higher values mean more deblocking (default: @code{32}).
15788 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15792 The horizontal and vertical deblocking filters share the difference and
15793 flatness values so you cannot set different horizontal and vertical
15797 @item h1/x1hdeblock
15798 Experimental horizontal deblocking filter
15800 @item v1/x1vdeblock
15801 Experimental vertical deblocking filter
15806 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15809 larger -> stronger filtering
15811 larger -> stronger filtering
15813 larger -> stronger filtering
15816 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15819 Stretch luminance to @code{0-255}.
15822 @item lb/linblenddeint
15823 Linear blend deinterlacing filter that deinterlaces the given block by
15824 filtering all lines with a @code{(1 2 1)} filter.
15826 @item li/linipoldeint
15827 Linear interpolating deinterlacing filter that deinterlaces the given block by
15828 linearly interpolating every second line.
15830 @item ci/cubicipoldeint
15831 Cubic interpolating deinterlacing filter deinterlaces the given block by
15832 cubically interpolating every second line.
15834 @item md/mediandeint
15835 Median deinterlacing filter that deinterlaces the given block by applying a
15836 median filter to every second line.
15838 @item fd/ffmpegdeint
15839 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15840 second line with a @code{(-1 4 2 4 -1)} filter.
15843 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15844 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15846 @item fq/forceQuant[|quantizer]
15847 Overrides the quantizer table from the input with the constant quantizer you
15855 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15858 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15861 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15864 @subsection Examples
15868 Apply horizontal and vertical deblocking, deringing and automatic
15869 brightness/contrast:
15875 Apply default filters without brightness/contrast correction:
15881 Apply default filters and temporal denoiser:
15883 pp=default/tmpnoise|1|2|3
15887 Apply deblocking on luminance only, and switch vertical deblocking on or off
15888 automatically depending on available CPU time:
15895 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15896 similar to spp = 6 with 7 point DCT, where only the center sample is
15899 The filter accepts the following options:
15903 Force a constant quantization parameter. It accepts an integer in range
15904 0 to 63. If not set, the filter will use the QP from the video stream
15908 Set thresholding mode. Available modes are:
15912 Set hard thresholding.
15914 Set soft thresholding (better de-ringing effect, but likely blurrier).
15916 Set medium thresholding (good results, default).
15920 @section premultiply
15921 Apply alpha premultiply effect to input video stream using first plane
15922 of second stream as alpha.
15924 Both streams must have same dimensions and same pixel format.
15926 The filter accepts the following option:
15930 Set which planes will be processed, unprocessed planes will be copied.
15931 By default value 0xf, all planes will be processed.
15934 Do not require 2nd input for processing, instead use alpha plane from input stream.
15938 Apply prewitt operator to input video stream.
15940 The filter accepts the following option:
15944 Set which planes will be processed, unprocessed planes will be copied.
15945 By default value 0xf, all planes will be processed.
15948 Set value which will be multiplied with filtered result.
15951 Set value which will be added to filtered result.
15954 @subsection Commands
15956 This filter supports the all above options as @ref{commands}.
15958 @section pseudocolor
15960 Alter frame colors in video with pseudocolors.
15962 This filter accepts the following options:
15966 set pixel first component expression
15969 set pixel second component expression
15972 set pixel third component expression
15975 set pixel fourth component expression, corresponds to the alpha component
15978 set component to use as base for altering colors
15981 Each of them specifies the expression to use for computing the lookup table for
15982 the corresponding pixel component values.
15984 The expressions can contain the following constants and functions:
15989 The input width and height.
15992 The input value for the pixel component.
15994 @item ymin, umin, vmin, amin
15995 The minimum allowed component value.
15997 @item ymax, umax, vmax, amax
15998 The maximum allowed component value.
16001 All expressions default to "val".
16003 @subsection Examples
16007 Change too high luma values to gradient:
16009 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'"
16015 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16016 Ratio) between two input videos.
16018 This filter takes in input two input videos, the first input is
16019 considered the "main" source and is passed unchanged to the
16020 output. The second input is used as a "reference" video for computing
16023 Both video inputs must have the same resolution and pixel format for
16024 this filter to work correctly. Also it assumes that both inputs
16025 have the same number of frames, which are compared one by one.
16027 The obtained average PSNR is printed through the logging system.
16029 The filter stores the accumulated MSE (mean squared error) of each
16030 frame, and at the end of the processing it is averaged across all frames
16031 equally, and the following formula is applied to obtain the PSNR:
16034 PSNR = 10*log10(MAX^2/MSE)
16037 Where MAX is the average of the maximum values of each component of the
16040 The description of the accepted parameters follows.
16043 @item stats_file, f
16044 If specified the filter will use the named file to save the PSNR of
16045 each individual frame. When filename equals "-" the data is sent to
16048 @item stats_version
16049 Specifies which version of the stats file format to use. Details of
16050 each format are written below.
16051 Default value is 1.
16053 @item stats_add_max
16054 Determines whether the max value is output to the stats log.
16055 Default value is 0.
16056 Requires stats_version >= 2. If this is set and stats_version < 2,
16057 the filter will return an error.
16060 This filter also supports the @ref{framesync} options.
16062 The file printed if @var{stats_file} is selected, contains a sequence of
16063 key/value pairs of the form @var{key}:@var{value} for each compared
16066 If a @var{stats_version} greater than 1 is specified, a header line precedes
16067 the list of per-frame-pair stats, with key value pairs following the frame
16068 format with the following parameters:
16071 @item psnr_log_version
16072 The version of the log file format. Will match @var{stats_version}.
16075 A comma separated list of the per-frame-pair parameters included in
16079 A description of each shown per-frame-pair parameter follows:
16083 sequential number of the input frame, starting from 1
16086 Mean Square Error pixel-by-pixel average difference of the compared
16087 frames, averaged over all the image components.
16089 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16090 Mean Square Error pixel-by-pixel average difference of the compared
16091 frames for the component specified by the suffix.
16093 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16094 Peak Signal to Noise ratio of the compared frames for the component
16095 specified by the suffix.
16097 @item max_avg, max_y, max_u, max_v
16098 Maximum allowed value for each channel, and average over all
16102 @subsection Examples
16107 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16108 [main][ref] psnr="stats_file=stats.log" [out]
16111 On this example the input file being processed is compared with the
16112 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16113 is stored in @file{stats.log}.
16116 Another example with different containers:
16118 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 -
16125 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16126 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16129 The pullup filter is designed to take advantage of future context in making
16130 its decisions. This filter is stateless in the sense that it does not lock
16131 onto a pattern to follow, but it instead looks forward to the following
16132 fields in order to identify matches and rebuild progressive frames.
16134 To produce content with an even framerate, insert the fps filter after
16135 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16136 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16138 The filter accepts the following options:
16145 These options set the amount of "junk" to ignore at the left, right, top, and
16146 bottom of the image, respectively. Left and right are in units of 8 pixels,
16147 while top and bottom are in units of 2 lines.
16148 The default is 8 pixels on each side.
16151 Set the strict breaks. Setting this option to 1 will reduce the chances of
16152 filter generating an occasional mismatched frame, but it may also cause an
16153 excessive number of frames to be dropped during high motion sequences.
16154 Conversely, setting it to -1 will make filter match fields more easily.
16155 This may help processing of video where there is slight blurring between
16156 the fields, but may also cause there to be interlaced frames in the output.
16157 Default value is @code{0}.
16160 Set the metric plane to use. It accepts the following values:
16166 Use chroma blue plane.
16169 Use chroma red plane.
16172 This option may be set to use chroma plane instead of the default luma plane
16173 for doing filter's computations. This may improve accuracy on very clean
16174 source material, but more likely will decrease accuracy, especially if there
16175 is chroma noise (rainbow effect) or any grayscale video.
16176 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16177 load and make pullup usable in realtime on slow machines.
16180 For best results (without duplicated frames in the output file) it is
16181 necessary to change the output frame rate. For example, to inverse
16182 telecine NTSC input:
16184 ffmpeg -i input -vf pullup -r 24000/1001 ...
16189 Change video quantization parameters (QP).
16191 The filter accepts the following option:
16195 Set expression for quantization parameter.
16198 The expression is evaluated through the eval API and can contain, among others,
16199 the following constants:
16203 1 if index is not 129, 0 otherwise.
16206 Sequential index starting from -129 to 128.
16209 @subsection Examples
16213 Some equation like:
16221 Flush video frames from internal cache of frames into a random order.
16222 No frame is discarded.
16223 Inspired by @ref{frei0r} nervous filter.
16227 Set size in number of frames of internal cache, in range from @code{2} to
16228 @code{512}. Default is @code{30}.
16231 Set seed for random number generator, must be an integer included between
16232 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16233 less than @code{0}, the filter will try to use a good random seed on a
16237 @section readeia608
16239 Read closed captioning (EIA-608) information from the top lines of a video frame.
16241 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16242 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16243 with EIA-608 data (starting from 0). A description of each metadata value follows:
16246 @item lavfi.readeia608.X.cc
16247 The two bytes stored as EIA-608 data (printed in hexadecimal).
16249 @item lavfi.readeia608.X.line
16250 The number of the line on which the EIA-608 data was identified and read.
16253 This filter accepts the following options:
16257 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16260 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16263 Set the ratio of width reserved for sync code detection.
16264 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16267 Enable checking the parity bit. In the event of a parity error, the filter will output
16268 @code{0x00} for that character. Default is false.
16271 Lowpass lines prior to further processing. Default is enabled.
16274 @subsection Commands
16276 This filter supports the all above options as @ref{commands}.
16278 @subsection Examples
16282 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16284 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
16290 Read vertical interval timecode (VITC) information from the top lines of a
16293 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16294 timecode value, if a valid timecode has been detected. Further metadata key
16295 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16296 timecode data has been found or not.
16298 This filter accepts the following options:
16302 Set the maximum number of lines to scan for VITC data. If the value is set to
16303 @code{-1} the full video frame is scanned. Default is @code{45}.
16306 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16307 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16310 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16311 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16314 @subsection Examples
16318 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16319 draw @code{--:--:--:--} as a placeholder:
16321 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16327 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16329 Destination pixel at position (X, Y) will be picked from source (x, y) position
16330 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16331 value for pixel will be used for destination pixel.
16333 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16334 will have Xmap/Ymap video stream dimensions.
16335 Xmap and Ymap input video streams are 16bit depth, single channel.
16339 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16340 Default is @code{color}.
16343 Specify the color of the unmapped pixels. For the syntax of this option,
16344 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16345 manual,ffmpeg-utils}. Default color is @code{black}.
16348 @section removegrain
16350 The removegrain filter is a spatial denoiser for progressive video.
16354 Set mode for the first plane.
16357 Set mode for the second plane.
16360 Set mode for the third plane.
16363 Set mode for the fourth plane.
16366 Range of mode is from 0 to 24. Description of each mode follows:
16370 Leave input plane unchanged. Default.
16373 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16376 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16379 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16382 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16383 This is equivalent to a median filter.
16386 Line-sensitive clipping giving the minimal change.
16389 Line-sensitive clipping, intermediate.
16392 Line-sensitive clipping, intermediate.
16395 Line-sensitive clipping, intermediate.
16398 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16401 Replaces the target pixel with the closest neighbour.
16404 [1 2 1] horizontal and vertical kernel blur.
16410 Bob mode, interpolates top field from the line where the neighbours
16411 pixels are the closest.
16414 Bob mode, interpolates bottom field from the line where the neighbours
16415 pixels are the closest.
16418 Bob mode, interpolates top field. Same as 13 but with a more complicated
16419 interpolation formula.
16422 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16423 interpolation formula.
16426 Clips the pixel with the minimum and maximum of respectively the maximum and
16427 minimum of each pair of opposite neighbour pixels.
16430 Line-sensitive clipping using opposite neighbours whose greatest distance from
16431 the current pixel is minimal.
16434 Replaces the pixel with the average of its 8 neighbours.
16437 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16440 Clips pixels using the averages of opposite neighbour.
16443 Same as mode 21 but simpler and faster.
16446 Small edge and halo removal, but reputed useless.
16452 @section removelogo
16454 Suppress a TV station logo, using an image file to determine which
16455 pixels comprise the logo. It works by filling in the pixels that
16456 comprise the logo with neighboring pixels.
16458 The filter accepts the following options:
16462 Set the filter bitmap file, which can be any image format supported by
16463 libavformat. The width and height of the image file must match those of the
16464 video stream being processed.
16467 Pixels in the provided bitmap image with a value of zero are not
16468 considered part of the logo, non-zero pixels are considered part of
16469 the logo. If you use white (255) for the logo and black (0) for the
16470 rest, you will be safe. For making the filter bitmap, it is
16471 recommended to take a screen capture of a black frame with the logo
16472 visible, and then using a threshold filter followed by the erode
16473 filter once or twice.
16475 If needed, little splotches can be fixed manually. Remember that if
16476 logo pixels are not covered, the filter quality will be much
16477 reduced. Marking too many pixels as part of the logo does not hurt as
16478 much, but it will increase the amount of blurring needed to cover over
16479 the image and will destroy more information than necessary, and extra
16480 pixels will slow things down on a large logo.
16482 @section repeatfields
16484 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16485 fields based on its value.
16489 Reverse a video clip.
16491 Warning: This filter requires memory to buffer the entire clip, so trimming
16494 @subsection Examples
16498 Take the first 5 seconds of a clip, and reverse it.
16505 Shift R/G/B/A pixels horizontally and/or vertically.
16507 The filter accepts the following options:
16510 Set amount to shift red horizontally.
16512 Set amount to shift red vertically.
16514 Set amount to shift green horizontally.
16516 Set amount to shift green vertically.
16518 Set amount to shift blue horizontally.
16520 Set amount to shift blue vertically.
16522 Set amount to shift alpha horizontally.
16524 Set amount to shift alpha vertically.
16526 Set edge mode, can be @var{smear}, default, or @var{warp}.
16529 @subsection Commands
16531 This filter supports the all above options as @ref{commands}.
16534 Apply roberts cross operator to input video stream.
16536 The filter accepts the following option:
16540 Set which planes will be processed, unprocessed planes will be copied.
16541 By default value 0xf, all planes will be processed.
16544 Set value which will be multiplied with filtered result.
16547 Set value which will be added to filtered result.
16550 @subsection Commands
16552 This filter supports the all above options as @ref{commands}.
16556 Rotate video by an arbitrary angle expressed in radians.
16558 The filter accepts the following options:
16560 A description of the optional parameters follows.
16563 Set an expression for the angle by which to rotate the input video
16564 clockwise, expressed as a number of radians. A negative value will
16565 result in a counter-clockwise rotation. By default it is set to "0".
16567 This expression is evaluated for each frame.
16570 Set the output width expression, default value is "iw".
16571 This expression is evaluated just once during configuration.
16574 Set the output height expression, default value is "ih".
16575 This expression is evaluated just once during configuration.
16578 Enable bilinear interpolation if set to 1, a value of 0 disables
16579 it. Default value is 1.
16582 Set the color used to fill the output area not covered by the rotated
16583 image. For the general syntax of this option, check the
16584 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16585 If the special value "none" is selected then no
16586 background is printed (useful for example if the background is never shown).
16588 Default value is "black".
16591 The expressions for the angle and the output size can contain the
16592 following constants and functions:
16596 sequential number of the input frame, starting from 0. It is always NAN
16597 before the first frame is filtered.
16600 time in seconds of the input frame, it is set to 0 when the filter is
16601 configured. It is always NAN before the first frame is filtered.
16605 horizontal and vertical chroma subsample values. For example for the
16606 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16610 the input video width and height
16614 the output width and height, that is the size of the padded area as
16615 specified by the @var{width} and @var{height} expressions
16619 the minimal width/height required for completely containing the input
16620 video rotated by @var{a} radians.
16622 These are only available when computing the @option{out_w} and
16623 @option{out_h} expressions.
16626 @subsection Examples
16630 Rotate the input by PI/6 radians clockwise:
16636 Rotate the input by PI/6 radians counter-clockwise:
16642 Rotate the input by 45 degrees clockwise:
16648 Apply a constant rotation with period T, starting from an angle of PI/3:
16650 rotate=PI/3+2*PI*t/T
16654 Make the input video rotation oscillating with a period of T
16655 seconds and an amplitude of A radians:
16657 rotate=A*sin(2*PI/T*t)
16661 Rotate the video, output size is chosen so that the whole rotating
16662 input video is always completely contained in the output:
16664 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16668 Rotate the video, reduce the output size so that no background is ever
16671 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16675 @subsection Commands
16677 The filter supports the following commands:
16681 Set the angle expression.
16682 The command accepts the same syntax of the corresponding option.
16684 If the specified expression is not valid, it is kept at its current
16690 Apply Shape Adaptive Blur.
16692 The filter accepts the following options:
16695 @item luma_radius, lr
16696 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16697 value is 1.0. A greater value will result in a more blurred image, and
16698 in slower processing.
16700 @item luma_pre_filter_radius, lpfr
16701 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16704 @item luma_strength, ls
16705 Set luma maximum difference between pixels to still be considered, must
16706 be a value in the 0.1-100.0 range, default value is 1.0.
16708 @item chroma_radius, cr
16709 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16710 greater value will result in a more blurred image, and in slower
16713 @item chroma_pre_filter_radius, cpfr
16714 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16716 @item chroma_strength, cs
16717 Set chroma maximum difference between pixels to still be considered,
16718 must be a value in the -0.9-100.0 range.
16721 Each chroma option value, if not explicitly specified, is set to the
16722 corresponding luma option value.
16727 Scale (resize) the input video, using the libswscale library.
16729 The scale filter forces the output display aspect ratio to be the same
16730 of the input, by changing the output sample aspect ratio.
16732 If the input image format is different from the format requested by
16733 the next filter, the scale filter will convert the input to the
16736 @subsection Options
16737 The filter accepts the following options, or any of the options
16738 supported by the libswscale scaler.
16740 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16741 the complete list of scaler options.
16746 Set the output video dimension expression. Default value is the input
16749 If the @var{width} or @var{w} value is 0, the input width is used for
16750 the output. If the @var{height} or @var{h} value is 0, the input height
16751 is used for the output.
16753 If one and only one of the values is -n with n >= 1, the scale filter
16754 will use a value that maintains the aspect ratio of the input image,
16755 calculated from the other specified dimension. After that it will,
16756 however, make sure that the calculated dimension is divisible by n and
16757 adjust the value if necessary.
16759 If both values are -n with n >= 1, the behavior will be identical to
16760 both values being set to 0 as previously detailed.
16762 See below for the list of accepted constants for use in the dimension
16766 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16770 Only evaluate expressions once during the filter initialization or when a command is processed.
16773 Evaluate expressions for each incoming frame.
16777 Default value is @samp{init}.
16781 Set the interlacing mode. It accepts the following values:
16785 Force interlaced aware scaling.
16788 Do not apply interlaced scaling.
16791 Select interlaced aware scaling depending on whether the source frames
16792 are flagged as interlaced or not.
16795 Default value is @samp{0}.
16798 Set libswscale scaling flags. See
16799 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16800 complete list of values. If not explicitly specified the filter applies
16804 @item param0, param1
16805 Set libswscale input parameters for scaling algorithms that need them. See
16806 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16807 complete documentation. If not explicitly specified the filter applies
16813 Set the video size. For the syntax of this option, check the
16814 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16816 @item in_color_matrix
16817 @item out_color_matrix
16818 Set in/output YCbCr color space type.
16820 This allows the autodetected value to be overridden as well as allows forcing
16821 a specific value used for the output and encoder.
16823 If not specified, the color space type depends on the pixel format.
16829 Choose automatically.
16832 Format conforming to International Telecommunication Union (ITU)
16833 Recommendation BT.709.
16836 Set color space conforming to the United States Federal Communications
16837 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16842 Set color space conforming to:
16846 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16849 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16852 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16857 Set color space conforming to SMPTE ST 240:1999.
16860 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16865 Set in/output YCbCr sample range.
16867 This allows the autodetected value to be overridden as well as allows forcing
16868 a specific value used for the output and encoder. If not specified, the
16869 range depends on the pixel format. Possible values:
16873 Choose automatically.
16876 Set full range (0-255 in case of 8-bit luma).
16878 @item mpeg/limited/tv
16879 Set "MPEG" range (16-235 in case of 8-bit luma).
16882 @item force_original_aspect_ratio
16883 Enable decreasing or increasing output video width or height if necessary to
16884 keep the original aspect ratio. Possible values:
16888 Scale the video as specified and disable this feature.
16891 The output video dimensions will automatically be decreased if needed.
16894 The output video dimensions will automatically be increased if needed.
16898 One useful instance of this option is that when you know a specific device's
16899 maximum allowed resolution, you can use this to limit the output video to
16900 that, while retaining the aspect ratio. For example, device A allows
16901 1280x720 playback, and your video is 1920x800. Using this option (set it to
16902 decrease) and specifying 1280x720 to the command line makes the output
16905 Please note that this is a different thing than specifying -1 for @option{w}
16906 or @option{h}, you still need to specify the output resolution for this option
16909 @item force_divisible_by
16910 Ensures that both the output dimensions, width and height, are divisible by the
16911 given integer when used together with @option{force_original_aspect_ratio}. This
16912 works similar to using @code{-n} in the @option{w} and @option{h} options.
16914 This option respects the value set for @option{force_original_aspect_ratio},
16915 increasing or decreasing the resolution accordingly. The video's aspect ratio
16916 may be slightly modified.
16918 This option can be handy if you need to have a video fit within or exceed
16919 a defined resolution using @option{force_original_aspect_ratio} but also have
16920 encoder restrictions on width or height divisibility.
16924 The values of the @option{w} and @option{h} options are expressions
16925 containing the following constants:
16930 The input width and height
16934 These are the same as @var{in_w} and @var{in_h}.
16938 The output (scaled) width and height
16942 These are the same as @var{out_w} and @var{out_h}
16945 The same as @var{iw} / @var{ih}
16948 input sample aspect ratio
16951 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16955 horizontal and vertical input chroma subsample values. For example for the
16956 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16960 horizontal and vertical output chroma subsample values. For example for the
16961 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16964 The (sequential) number of the input frame, starting from 0.
16965 Only available with @code{eval=frame}.
16968 The presentation timestamp of the input frame, expressed as a number of
16969 seconds. Only available with @code{eval=frame}.
16972 The position (byte offset) of the frame in the input stream, or NaN if
16973 this information is unavailable and/or meaningless (for example in case of synthetic video).
16974 Only available with @code{eval=frame}.
16977 @subsection Examples
16981 Scale the input video to a size of 200x100
16986 This is equivalent to:
16997 Specify a size abbreviation for the output size:
17002 which can also be written as:
17008 Scale the input to 2x:
17010 scale=w=2*iw:h=2*ih
17014 The above is the same as:
17016 scale=2*in_w:2*in_h
17020 Scale the input to 2x with forced interlaced scaling:
17022 scale=2*iw:2*ih:interl=1
17026 Scale the input to half size:
17028 scale=w=iw/2:h=ih/2
17032 Increase the width, and set the height to the same size:
17038 Seek Greek harmony:
17045 Increase the height, and set the width to 3/2 of the height:
17047 scale=w=3/2*oh:h=3/5*ih
17051 Increase the size, making the size a multiple of the chroma
17054 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17058 Increase the width to a maximum of 500 pixels,
17059 keeping the same aspect ratio as the input:
17061 scale=w='min(500\, iw*3/2):h=-1'
17065 Make pixels square by combining scale and setsar:
17067 scale='trunc(ih*dar):ih',setsar=1/1
17071 Make pixels square by combining scale and setsar,
17072 making sure the resulting resolution is even (required by some codecs):
17074 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17078 @subsection Commands
17080 This filter supports the following commands:
17084 Set the output video dimension expression.
17085 The command accepts the same syntax of the corresponding option.
17087 If the specified expression is not valid, it is kept at its current
17093 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17094 format conversion on CUDA video frames. Setting the output width and height
17095 works in the same way as for the @var{scale} filter.
17097 The following additional options are accepted:
17100 The pixel format of the output CUDA frames. If set to the string "same" (the
17101 default), the input format will be kept. Note that automatic format negotiation
17102 and conversion is not yet supported for hardware frames
17105 The interpolation algorithm used for resizing. One of the following:
17112 @item cubic2p_bspline
17113 2-parameter cubic (B=1, C=0)
17115 @item cubic2p_catmullrom
17116 2-parameter cubic (B=0, C=1/2)
17118 @item cubic2p_b05c03
17119 2-parameter cubic (B=1/2, C=3/10)
17127 @item force_original_aspect_ratio
17128 Enable decreasing or increasing output video width or height if necessary to
17129 keep the original aspect ratio. Possible values:
17133 Scale the video as specified and disable this feature.
17136 The output video dimensions will automatically be decreased if needed.
17139 The output video dimensions will automatically be increased if needed.
17143 One useful instance of this option is that when you know a specific device's
17144 maximum allowed resolution, you can use this to limit the output video to
17145 that, while retaining the aspect ratio. For example, device A allows
17146 1280x720 playback, and your video is 1920x800. Using this option (set it to
17147 decrease) and specifying 1280x720 to the command line makes the output
17150 Please note that this is a different thing than specifying -1 for @option{w}
17151 or @option{h}, you still need to specify the output resolution for this option
17154 @item force_divisible_by
17155 Ensures that both the output dimensions, width and height, are divisible by the
17156 given integer when used together with @option{force_original_aspect_ratio}. This
17157 works similar to using @code{-n} in the @option{w} and @option{h} options.
17159 This option respects the value set for @option{force_original_aspect_ratio},
17160 increasing or decreasing the resolution accordingly. The video's aspect ratio
17161 may be slightly modified.
17163 This option can be handy if you need to have a video fit within or exceed
17164 a defined resolution using @option{force_original_aspect_ratio} but also have
17165 encoder restrictions on width or height divisibility.
17171 Scale (resize) the input video, based on a reference video.
17173 See the scale filter for available options, scale2ref supports the same but
17174 uses the reference video instead of the main input as basis. scale2ref also
17175 supports the following additional constants for the @option{w} and
17176 @option{h} options:
17181 The main input video's width and height
17184 The same as @var{main_w} / @var{main_h}
17187 The main input video's sample aspect ratio
17189 @item main_dar, mdar
17190 The main input video's display aspect ratio. Calculated from
17191 @code{(main_w / main_h) * main_sar}.
17195 The main input video's horizontal and vertical chroma subsample values.
17196 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17200 The (sequential) number of the main input frame, starting from 0.
17201 Only available with @code{eval=frame}.
17204 The presentation timestamp of the main input frame, expressed as a number of
17205 seconds. Only available with @code{eval=frame}.
17208 The position (byte offset) of the frame in the main input stream, or NaN if
17209 this information is unavailable and/or meaningless (for example in case of synthetic video).
17210 Only available with @code{eval=frame}.
17213 @subsection Examples
17217 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17219 'scale2ref[b][a];[a][b]overlay'
17223 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17225 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17229 @subsection Commands
17231 This filter supports the following commands:
17235 Set the output video dimension expression.
17236 The command accepts the same syntax of the corresponding option.
17238 If the specified expression is not valid, it is kept at its current
17243 Scroll input video horizontally and/or vertically by constant speed.
17245 The filter accepts the following options:
17247 @item horizontal, h
17248 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17249 Negative values changes scrolling direction.
17252 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17253 Negative values changes scrolling direction.
17256 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17259 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17262 @subsection Commands
17264 This filter supports the following @ref{commands}:
17266 @item horizontal, h
17267 Set the horizontal scrolling speed.
17269 Set the vertical scrolling speed.
17275 Detect video scene change.
17277 This filter sets frame metadata with mafd between frame, the scene score, and
17278 forward the frame to the next filter, so they can use these metadata to detect
17279 scene change or others.
17281 In addition, this filter logs a message and sets frame metadata when it detects
17282 a scene change by @option{threshold}.
17284 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17286 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17287 to detect scene change.
17289 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17290 detect scene change with @option{threshold}.
17292 The filter accepts the following options:
17296 Set the scene change detection threshold as a percentage of maximum change. Good
17297 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17300 Default value is @code{10.}.
17303 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17304 You can enable it if you want to get snapshot of scene change frames only.
17307 @anchor{selectivecolor}
17308 @section selectivecolor
17310 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17311 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17312 by the "purity" of the color (that is, how saturated it already is).
17314 This filter is similar to the Adobe Photoshop Selective Color tool.
17316 The filter accepts the following options:
17319 @item correction_method
17320 Select color correction method.
17322 Available values are:
17325 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17328 Specified adjustments are relative to the original component value.
17330 Default is @code{absolute}.
17332 Adjustments for red pixels (pixels where the red component is the maximum)
17334 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17336 Adjustments for green pixels (pixels where the green component is the maximum)
17338 Adjustments for cyan pixels (pixels where the red component is the minimum)
17340 Adjustments for blue pixels (pixels where the blue component is the maximum)
17342 Adjustments for magenta pixels (pixels where the green component is the minimum)
17344 Adjustments for white pixels (pixels where all components are greater than 128)
17346 Adjustments for all pixels except pure black and pure white
17348 Adjustments for black pixels (pixels where all components are lesser than 128)
17350 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17353 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17354 4 space separated floating point adjustment values in the [-1,1] range,
17355 respectively to adjust the amount of cyan, magenta, yellow and black for the
17356 pixels of its range.
17358 @subsection Examples
17362 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17363 increase magenta by 27% in blue areas:
17365 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17369 Use a Photoshop selective color preset:
17371 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17375 @anchor{separatefields}
17376 @section separatefields
17378 The @code{separatefields} takes a frame-based video input and splits
17379 each frame into its components fields, producing a new half height clip
17380 with twice the frame rate and twice the frame count.
17382 This filter use field-dominance information in frame to decide which
17383 of each pair of fields to place first in the output.
17384 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17386 @section setdar, setsar
17388 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17391 This is done by changing the specified Sample (aka Pixel) Aspect
17392 Ratio, according to the following equation:
17394 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17397 Keep in mind that the @code{setdar} filter does not modify the pixel
17398 dimensions of the video frame. Also, the display aspect ratio set by
17399 this filter may be changed by later filters in the filterchain,
17400 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17403 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17404 the filter output video.
17406 Note that as a consequence of the application of this filter, the
17407 output display aspect ratio will change according to the equation
17410 Keep in mind that the sample aspect ratio set by the @code{setsar}
17411 filter may be changed by later filters in the filterchain, e.g. if
17412 another "setsar" or a "setdar" filter is applied.
17414 It accepts the following parameters:
17417 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17418 Set the aspect ratio used by the filter.
17420 The parameter can be a floating point number string, an expression, or
17421 a string of the form @var{num}:@var{den}, where @var{num} and
17422 @var{den} are the numerator and denominator of the aspect ratio. If
17423 the parameter is not specified, it is assumed the value "0".
17424 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17428 Set the maximum integer value to use for expressing numerator and
17429 denominator when reducing the expressed aspect ratio to a rational.
17430 Default value is @code{100}.
17434 The parameter @var{sar} is an expression containing
17435 the following constants:
17439 These are approximated values for the mathematical constants e
17440 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17443 The input width and height.
17446 These are the same as @var{w} / @var{h}.
17449 The input sample aspect ratio.
17452 The input display aspect ratio. It is the same as
17453 (@var{w} / @var{h}) * @var{sar}.
17456 Horizontal and vertical chroma subsample values. For example, for the
17457 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17460 @subsection Examples
17465 To change the display aspect ratio to 16:9, specify one of the following:
17472 To change the sample aspect ratio to 10:11, specify:
17478 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17479 1000 in the aspect ratio reduction, use the command:
17481 setdar=ratio=16/9:max=1000
17489 Force field for the output video frame.
17491 The @code{setfield} filter marks the interlace type field for the
17492 output frames. It does not change the input frame, but only sets the
17493 corresponding property, which affects how the frame is treated by
17494 following filters (e.g. @code{fieldorder} or @code{yadif}).
17496 The filter accepts the following options:
17501 Available values are:
17505 Keep the same field property.
17508 Mark the frame as bottom-field-first.
17511 Mark the frame as top-field-first.
17514 Mark the frame as progressive.
17521 Force frame parameter for the output video frame.
17523 The @code{setparams} filter marks interlace and color range for the
17524 output frames. It does not change the input frame, but only sets the
17525 corresponding property, which affects how the frame is treated by
17530 Available values are:
17534 Keep the same field property (default).
17537 Mark the frame as bottom-field-first.
17540 Mark the frame as top-field-first.
17543 Mark the frame as progressive.
17547 Available values are:
17551 Keep the same color range property (default).
17553 @item unspecified, unknown
17554 Mark the frame as unspecified color range.
17556 @item limited, tv, mpeg
17557 Mark the frame as limited range.
17559 @item full, pc, jpeg
17560 Mark the frame as full range.
17563 @item color_primaries
17564 Set the color primaries.
17565 Available values are:
17569 Keep the same color primaries property (default).
17586 Set the color transfer.
17587 Available values are:
17591 Keep the same color trc property (default).
17613 Set the colorspace.
17614 Available values are:
17618 Keep the same colorspace property (default).
17631 @item chroma-derived-nc
17632 @item chroma-derived-c
17639 Show a line containing various information for each input video frame.
17640 The input video is not modified.
17642 This filter supports the following options:
17646 Calculate checksums of each plane. By default enabled.
17649 The shown line contains a sequence of key/value pairs of the form
17650 @var{key}:@var{value}.
17652 The following values are shown in the output:
17656 The (sequential) number of the input frame, starting from 0.
17659 The Presentation TimeStamp of the input frame, expressed as a number of
17660 time base units. The time base unit depends on the filter input pad.
17663 The Presentation TimeStamp of the input frame, expressed as a number of
17667 The position of the frame in the input stream, or -1 if this information is
17668 unavailable and/or meaningless (for example in case of synthetic video).
17671 The pixel format name.
17674 The sample aspect ratio of the input frame, expressed in the form
17675 @var{num}/@var{den}.
17678 The size of the input frame. For the syntax of this option, check the
17679 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17682 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17683 for bottom field first).
17686 This is 1 if the frame is a key frame, 0 otherwise.
17689 The picture type of the input frame ("I" for an I-frame, "P" for a
17690 P-frame, "B" for a B-frame, or "?" for an unknown type).
17691 Also refer to the documentation of the @code{AVPictureType} enum and of
17692 the @code{av_get_picture_type_char} function defined in
17693 @file{libavutil/avutil.h}.
17696 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17698 @item plane_checksum
17699 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17700 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17703 The mean value of pixels in each plane of the input frame, expressed in the form
17704 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17707 The standard deviation of pixel values in each plane of the input frame, expressed
17708 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17712 @section showpalette
17714 Displays the 256 colors palette of each frame. This filter is only relevant for
17715 @var{pal8} pixel format frames.
17717 It accepts the following option:
17721 Set the size of the box used to represent one palette color entry. Default is
17722 @code{30} (for a @code{30x30} pixel box).
17725 @section shuffleframes
17727 Reorder and/or duplicate and/or drop video frames.
17729 It accepts the following parameters:
17733 Set the destination indexes of input frames.
17734 This is space or '|' separated list of indexes that maps input frames to output
17735 frames. Number of indexes also sets maximal value that each index may have.
17736 '-1' index have special meaning and that is to drop frame.
17739 The first frame has the index 0. The default is to keep the input unchanged.
17741 @subsection Examples
17745 Swap second and third frame of every three frames of the input:
17747 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17751 Swap 10th and 1st frame of every ten frames of the input:
17753 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17757 @section shuffleplanes
17759 Reorder and/or duplicate video planes.
17761 It accepts the following parameters:
17766 The index of the input plane to be used as the first output plane.
17769 The index of the input plane to be used as the second output plane.
17772 The index of the input plane to be used as the third output plane.
17775 The index of the input plane to be used as the fourth output plane.
17779 The first plane has the index 0. The default is to keep the input unchanged.
17781 @subsection Examples
17785 Swap the second and third planes of the input:
17787 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17791 @anchor{signalstats}
17792 @section signalstats
17793 Evaluate various visual metrics that assist in determining issues associated
17794 with the digitization of analog video media.
17796 By default the filter will log these metadata values:
17800 Display the minimal Y value contained within the input frame. Expressed in
17804 Display the Y value at the 10% percentile within the input frame. Expressed in
17808 Display the average Y value within the input frame. Expressed in range of
17812 Display the Y value at the 90% percentile within the input frame. Expressed in
17816 Display the maximum Y value contained within the input frame. Expressed in
17820 Display the minimal U value contained within the input frame. Expressed in
17824 Display the U value at the 10% percentile within the input frame. Expressed in
17828 Display the average U value within the input frame. Expressed in range of
17832 Display the U value at the 90% percentile within the input frame. Expressed in
17836 Display the maximum U value contained within the input frame. Expressed in
17840 Display the minimal V value contained within the input frame. Expressed in
17844 Display the V value at the 10% percentile within the input frame. Expressed in
17848 Display the average V value within the input frame. Expressed in range of
17852 Display the V value at the 90% percentile within the input frame. Expressed in
17856 Display the maximum V value contained within the input frame. Expressed in
17860 Display the minimal saturation value contained within the input frame.
17861 Expressed in range of [0-~181.02].
17864 Display the saturation value at the 10% percentile within the input frame.
17865 Expressed in range of [0-~181.02].
17868 Display the average saturation value within the input frame. Expressed in range
17872 Display the saturation value at the 90% percentile within the input frame.
17873 Expressed in range of [0-~181.02].
17876 Display the maximum saturation value contained within the input frame.
17877 Expressed in range of [0-~181.02].
17880 Display the median value for hue within the input frame. Expressed in range of
17884 Display the average value for hue within the input frame. Expressed in range of
17888 Display the average of sample value difference between all values of the Y
17889 plane in the current frame and corresponding values of the previous input frame.
17890 Expressed in range of [0-255].
17893 Display the average of sample value difference between all values of the U
17894 plane in the current frame and corresponding values of the previous input frame.
17895 Expressed in range of [0-255].
17898 Display the average of sample value difference between all values of the V
17899 plane in the current frame and corresponding values of the previous input frame.
17900 Expressed in range of [0-255].
17903 Display bit depth of Y plane in current frame.
17904 Expressed in range of [0-16].
17907 Display bit depth of U plane in current frame.
17908 Expressed in range of [0-16].
17911 Display bit depth of V plane in current frame.
17912 Expressed in range of [0-16].
17915 The filter accepts the following options:
17921 @option{stat} specify an additional form of image analysis.
17922 @option{out} output video with the specified type of pixel highlighted.
17924 Both options accept the following values:
17928 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17929 unlike the neighboring pixels of the same field. Examples of temporal outliers
17930 include the results of video dropouts, head clogs, or tape tracking issues.
17933 Identify @var{vertical line repetition}. Vertical line repetition includes
17934 similar rows of pixels within a frame. In born-digital video vertical line
17935 repetition is common, but this pattern is uncommon in video digitized from an
17936 analog source. When it occurs in video that results from the digitization of an
17937 analog source it can indicate concealment from a dropout compensator.
17940 Identify pixels that fall outside of legal broadcast range.
17944 Set the highlight color for the @option{out} option. The default color is
17948 @subsection Examples
17952 Output data of various video metrics:
17954 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17958 Output specific data about the minimum and maximum values of the Y plane per frame:
17960 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17964 Playback video while highlighting pixels that are outside of broadcast range in red.
17966 ffplay example.mov -vf signalstats="out=brng:color=red"
17970 Playback video with signalstats metadata drawn over the frame.
17972 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17975 The contents of signalstat_drawtext.txt used in the command are:
17978 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17979 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
17980 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
17981 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
17989 Calculates the MPEG-7 Video Signature. The filter can handle more than one
17990 input. In this case the matching between the inputs can be calculated additionally.
17991 The filter always passes through the first input. The signature of each stream can
17992 be written into a file.
17994 It accepts the following options:
17998 Enable or disable the matching process.
18000 Available values are:
18004 Disable the calculation of a matching (default).
18006 Calculate the matching for the whole video and output whether the whole video
18007 matches or only parts.
18009 Calculate only until a matching is found or the video ends. Should be faster in
18014 Set the number of inputs. The option value must be a non negative integer.
18015 Default value is 1.
18018 Set the path to which the output is written. If there is more than one input,
18019 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18020 integer), that will be replaced with the input number. If no filename is
18021 specified, no output will be written. This is the default.
18024 Choose the output format.
18026 Available values are:
18030 Use the specified binary representation (default).
18032 Use the specified xml representation.
18036 Set threshold to detect one word as similar. The option value must be an integer
18037 greater than zero. The default value is 9000.
18040 Set threshold to detect all words as similar. The option value must be an integer
18041 greater than zero. The default value is 60000.
18044 Set threshold to detect frames as similar. The option value must be an integer
18045 greater than zero. The default value is 116.
18048 Set the minimum length of a sequence in frames to recognize it as matching
18049 sequence. The option value must be a non negative integer value.
18050 The default value is 0.
18053 Set the minimum relation, that matching frames to all frames must have.
18054 The option value must be a double value between 0 and 1. The default value is 0.5.
18057 @subsection Examples
18061 To calculate the signature of an input video and store it in signature.bin:
18063 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18067 To detect whether two videos match and store the signatures in XML format in
18068 signature0.xml and signature1.xml:
18070 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 -
18078 Blur the input video without impacting the outlines.
18080 It accepts the following options:
18083 @item luma_radius, lr
18084 Set the luma radius. The option value must be a float number in
18085 the range [0.1,5.0] that specifies the variance of the gaussian filter
18086 used to blur the image (slower if larger). Default value is 1.0.
18088 @item luma_strength, ls
18089 Set the luma strength. The option value must be a float number
18090 in the range [-1.0,1.0] that configures the blurring. A value included
18091 in [0.0,1.0] will blur the image whereas a value included in
18092 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18094 @item luma_threshold, lt
18095 Set the luma threshold used as a coefficient to determine
18096 whether a pixel should be blurred or not. The option value must be an
18097 integer in the range [-30,30]. A value of 0 will filter all the image,
18098 a value included in [0,30] will filter flat areas and a value included
18099 in [-30,0] will filter edges. Default value is 0.
18101 @item chroma_radius, cr
18102 Set the chroma radius. The option value must be a float number in
18103 the range [0.1,5.0] that specifies the variance of the gaussian filter
18104 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18106 @item chroma_strength, cs
18107 Set the chroma strength. The option value must be a float number
18108 in the range [-1.0,1.0] that configures the blurring. A value included
18109 in [0.0,1.0] will blur the image whereas a value included in
18110 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18112 @item chroma_threshold, ct
18113 Set the chroma threshold used as a coefficient to determine
18114 whether a pixel should be blurred or not. The option value must be an
18115 integer in the range [-30,30]. A value of 0 will filter all the image,
18116 a value included in [0,30] will filter flat areas and a value included
18117 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18120 If a chroma option is not explicitly set, the corresponding luma value
18124 Apply sobel operator to input video stream.
18126 The filter accepts the following option:
18130 Set which planes will be processed, unprocessed planes will be copied.
18131 By default value 0xf, all planes will be processed.
18134 Set value which will be multiplied with filtered result.
18137 Set value which will be added to filtered result.
18140 @subsection Commands
18142 This filter supports the all above options as @ref{commands}.
18147 Apply a simple postprocessing filter that compresses and decompresses the image
18148 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18149 and average the results.
18151 The filter accepts the following options:
18155 Set quality. This option defines the number of levels for averaging. It accepts
18156 an integer in the range 0-6. If set to @code{0}, the filter will have no
18157 effect. A value of @code{6} means the higher quality. For each increment of
18158 that value the speed drops by a factor of approximately 2. Default value is
18162 Force a constant quantization parameter. If not set, the filter will use the QP
18163 from the video stream (if available).
18166 Set thresholding mode. Available modes are:
18170 Set hard thresholding (default).
18172 Set soft thresholding (better de-ringing effect, but likely blurrier).
18175 @item use_bframe_qp
18176 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18177 option may cause flicker since the B-Frames have often larger QP. Default is
18178 @code{0} (not enabled).
18181 @subsection Commands
18183 This filter supports the following commands:
18185 @item quality, level
18186 Set quality level. The value @code{max} can be used to set the maximum level,
18187 currently @code{6}.
18193 Scale the input by applying one of the super-resolution methods based on
18194 convolutional neural networks. Supported models:
18198 Super-Resolution Convolutional Neural Network model (SRCNN).
18199 See @url{https://arxiv.org/abs/1501.00092}.
18202 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18203 See @url{https://arxiv.org/abs/1609.05158}.
18206 Training scripts as well as scripts for model file (.pb) saving can be found at
18207 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18208 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18210 Native model files (.model) can be generated from TensorFlow model
18211 files (.pb) by using tools/python/convert.py
18213 The filter accepts the following options:
18217 Specify which DNN backend to use for model loading and execution. This option accepts
18218 the following values:
18222 Native implementation of DNN loading and execution.
18225 TensorFlow backend. To enable this backend you
18226 need to install the TensorFlow for C library (see
18227 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18228 @code{--enable-libtensorflow}
18231 Default value is @samp{native}.
18234 Set path to model file specifying network architecture and its parameters.
18235 Note that different backends use different file formats. TensorFlow backend
18236 can load files for both formats, while native backend can load files for only
18240 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18241 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18242 input upscaled using bicubic upscaling with proper scale factor.
18245 This feature can also be finished with @ref{dnn_processing} filter.
18249 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18251 This filter takes in input two input videos, the first input is
18252 considered the "main" source and is passed unchanged to the
18253 output. The second input is used as a "reference" video for computing
18256 Both video inputs must have the same resolution and pixel format for
18257 this filter to work correctly. Also it assumes that both inputs
18258 have the same number of frames, which are compared one by one.
18260 The filter stores the calculated SSIM of each frame.
18262 The description of the accepted parameters follows.
18265 @item stats_file, f
18266 If specified the filter will use the named file to save the SSIM of
18267 each individual frame. When filename equals "-" the data is sent to
18271 The file printed if @var{stats_file} is selected, contains a sequence of
18272 key/value pairs of the form @var{key}:@var{value} for each compared
18275 A description of each shown parameter follows:
18279 sequential number of the input frame, starting from 1
18281 @item Y, U, V, R, G, B
18282 SSIM of the compared frames for the component specified by the suffix.
18285 SSIM of the compared frames for the whole frame.
18288 Same as above but in dB representation.
18291 This filter also supports the @ref{framesync} options.
18293 @subsection Examples
18298 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18299 [main][ref] ssim="stats_file=stats.log" [out]
18302 On this example the input file being processed is compared with the
18303 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18304 is stored in @file{stats.log}.
18307 Another example with both psnr and ssim at same time:
18309 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18313 Another example with different containers:
18315 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 -
18321 Convert between different stereoscopic image formats.
18323 The filters accept the following options:
18327 Set stereoscopic image format of input.
18329 Available values for input image formats are:
18332 side by side parallel (left eye left, right eye right)
18335 side by side crosseye (right eye left, left eye right)
18338 side by side parallel with half width resolution
18339 (left eye left, right eye right)
18342 side by side crosseye with half width resolution
18343 (right eye left, left eye right)
18347 above-below (left eye above, right eye below)
18351 above-below (right eye above, left eye below)
18355 above-below with half height resolution
18356 (left eye above, right eye below)
18360 above-below with half height resolution
18361 (right eye above, left eye below)
18364 alternating frames (left eye first, right eye second)
18367 alternating frames (right eye first, left eye second)
18370 interleaved rows (left eye has top row, right eye starts on next row)
18373 interleaved rows (right eye has top row, left eye starts on next row)
18376 interleaved columns, left eye first
18379 interleaved columns, right eye first
18381 Default value is @samp{sbsl}.
18385 Set stereoscopic image format of output.
18389 side by side parallel (left eye left, right eye right)
18392 side by side crosseye (right eye left, left eye right)
18395 side by side parallel with half width resolution
18396 (left eye left, right eye right)
18399 side by side crosseye with half width resolution
18400 (right eye left, left eye right)
18404 above-below (left eye above, right eye below)
18408 above-below (right eye above, left eye below)
18412 above-below with half height resolution
18413 (left eye above, right eye below)
18417 above-below with half height resolution
18418 (right eye above, left eye below)
18421 alternating frames (left eye first, right eye second)
18424 alternating frames (right eye first, left eye second)
18427 interleaved rows (left eye has top row, right eye starts on next row)
18430 interleaved rows (right eye has top row, left eye starts on next row)
18433 anaglyph red/blue gray
18434 (red filter on left eye, blue filter on right eye)
18437 anaglyph red/green gray
18438 (red filter on left eye, green filter on right eye)
18441 anaglyph red/cyan gray
18442 (red filter on left eye, cyan filter on right eye)
18445 anaglyph red/cyan half colored
18446 (red filter on left eye, cyan filter on right eye)
18449 anaglyph red/cyan color
18450 (red filter on left eye, cyan filter on right eye)
18453 anaglyph red/cyan color optimized with the least squares projection of dubois
18454 (red filter on left eye, cyan filter on right eye)
18457 anaglyph green/magenta gray
18458 (green filter on left eye, magenta filter on right eye)
18461 anaglyph green/magenta half colored
18462 (green filter on left eye, magenta filter on right eye)
18465 anaglyph green/magenta colored
18466 (green filter on left eye, magenta filter on right eye)
18469 anaglyph green/magenta color optimized with the least squares projection of dubois
18470 (green filter on left eye, magenta filter on right eye)
18473 anaglyph yellow/blue gray
18474 (yellow filter on left eye, blue filter on right eye)
18477 anaglyph yellow/blue half colored
18478 (yellow filter on left eye, blue filter on right eye)
18481 anaglyph yellow/blue colored
18482 (yellow filter on left eye, blue filter on right eye)
18485 anaglyph yellow/blue color optimized with the least squares projection of dubois
18486 (yellow filter on left eye, blue filter on right eye)
18489 mono output (left eye only)
18492 mono output (right eye only)
18495 checkerboard, left eye first
18498 checkerboard, right eye first
18501 interleaved columns, left eye first
18504 interleaved columns, right eye first
18510 Default value is @samp{arcd}.
18513 @subsection Examples
18517 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18523 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18529 @section streamselect, astreamselect
18530 Select video or audio streams.
18532 The filter accepts the following options:
18536 Set number of inputs. Default is 2.
18539 Set input indexes to remap to outputs.
18542 @subsection Commands
18544 The @code{streamselect} and @code{astreamselect} filter supports the following
18549 Set input indexes to remap to outputs.
18552 @subsection Examples
18556 Select first 5 seconds 1st stream and rest of time 2nd stream:
18558 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18562 Same as above, but for audio:
18564 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18571 Draw subtitles on top of input video using the libass library.
18573 To enable compilation of this filter you need to configure FFmpeg with
18574 @code{--enable-libass}. This filter also requires a build with libavcodec and
18575 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18576 Alpha) subtitles format.
18578 The filter accepts the following options:
18582 Set the filename of the subtitle file to read. It must be specified.
18584 @item original_size
18585 Specify the size of the original video, the video for which the ASS file
18586 was composed. For the syntax of this option, check the
18587 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18588 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18589 correctly scale the fonts if the aspect ratio has been changed.
18592 Set a directory path containing fonts that can be used by the filter.
18593 These fonts will be used in addition to whatever the font provider uses.
18596 Process alpha channel, by default alpha channel is untouched.
18599 Set subtitles input character encoding. @code{subtitles} filter only. Only
18600 useful if not UTF-8.
18602 @item stream_index, si
18603 Set subtitles stream index. @code{subtitles} filter only.
18606 Override default style or script info parameters of the subtitles. It accepts a
18607 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18610 If the first key is not specified, it is assumed that the first value
18611 specifies the @option{filename}.
18613 For example, to render the file @file{sub.srt} on top of the input
18614 video, use the command:
18619 which is equivalent to:
18621 subtitles=filename=sub.srt
18624 To render the default subtitles stream from file @file{video.mkv}, use:
18626 subtitles=video.mkv
18629 To render the second subtitles stream from that file, use:
18631 subtitles=video.mkv:si=1
18634 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18635 @code{DejaVu Serif}, use:
18637 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18640 @section super2xsai
18642 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18643 Interpolate) pixel art scaling algorithm.
18645 Useful for enlarging pixel art images without reducing sharpness.
18649 Swap two rectangular objects in video.
18651 This filter accepts the following options:
18661 Set 1st rect x coordinate.
18664 Set 1st rect y coordinate.
18667 Set 2nd rect x coordinate.
18670 Set 2nd rect y coordinate.
18672 All expressions are evaluated once for each frame.
18675 The all options are expressions containing the following constants:
18680 The input width and height.
18683 same as @var{w} / @var{h}
18686 input sample aspect ratio
18689 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18692 The number of the input frame, starting from 0.
18695 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18698 the position in the file of the input frame, NAN if unknown
18705 Blend successive video frames.
18711 Apply telecine process to the video.
18713 This filter accepts the following options:
18722 The default value is @code{top}.
18726 A string of numbers representing the pulldown pattern you wish to apply.
18727 The default value is @code{23}.
18731 Some typical patterns:
18736 24p: 2332 (preferred)
18743 24p: 222222222223 ("Euro pulldown")
18748 @section thistogram
18750 Compute and draw a color distribution histogram for the input video across time.
18752 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18753 at certain time, this filter shows also past histograms of number of frames defined
18754 by @code{width} option.
18756 The computed histogram is a representation of the color component
18757 distribution in an image.
18759 The filter accepts the following options:
18763 Set width of single color component output. Default value is @code{0}.
18764 Value of @code{0} means width will be picked from input video.
18765 This also set number of passed histograms to keep.
18766 Allowed range is [0, 8192].
18768 @item display_mode, d
18770 It accepts the following values:
18773 Per color component graphs are placed below each other.
18776 Per color component graphs are placed side by side.
18779 Presents information identical to that in the @code{parade}, except
18780 that the graphs representing color components are superimposed directly
18783 Default is @code{stack}.
18785 @item levels_mode, m
18786 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18787 Default is @code{linear}.
18789 @item components, c
18790 Set what color components to display.
18791 Default is @code{7}.
18794 Set background opacity. Default is @code{0.9}.
18797 Show envelope. Default is disabled.
18800 Set envelope color. Default is @code{gold}.
18805 Available values for slide is:
18808 Draw new frame when right border is reached.
18811 Replace old columns with new ones.
18814 Scroll from right to left.
18817 Scroll from left to right.
18820 Draw single picture.
18823 Default is @code{replace}.
18828 Apply threshold effect to video stream.
18830 This filter needs four video streams to perform thresholding.
18831 First stream is stream we are filtering.
18832 Second stream is holding threshold values, third stream is holding min values,
18833 and last, fourth stream is holding max values.
18835 The filter accepts the following option:
18839 Set which planes will be processed, unprocessed planes will be copied.
18840 By default value 0xf, all planes will be processed.
18843 For example if first stream pixel's component value is less then threshold value
18844 of pixel component from 2nd threshold stream, third stream value will picked,
18845 otherwise fourth stream pixel component value will be picked.
18847 Using color source filter one can perform various types of thresholding:
18849 @subsection Examples
18853 Binary threshold, using gray color as threshold:
18855 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18859 Inverted binary threshold, using gray color as threshold:
18861 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18865 Truncate binary threshold, using gray color as threshold:
18867 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18871 Threshold to zero, using gray color as threshold:
18873 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18877 Inverted threshold to zero, using gray color as threshold:
18879 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18884 Select the most representative frame in a given sequence of consecutive frames.
18886 The filter accepts the following options:
18890 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18891 will pick one of them, and then handle the next batch of @var{n} frames until
18892 the end. Default is @code{100}.
18895 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18896 value will result in a higher memory usage, so a high value is not recommended.
18898 @subsection Examples
18902 Extract one picture each 50 frames:
18908 Complete example of a thumbnail creation with @command{ffmpeg}:
18910 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18917 Tile several successive frames together.
18919 The @ref{untile} filter can do the reverse.
18921 The filter accepts the following options:
18926 Set the grid size (i.e. the number of lines and columns). For the syntax of
18927 this option, check the
18928 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18931 Set the maximum number of frames to render in the given area. It must be less
18932 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18933 the area will be used.
18936 Set the outer border margin in pixels.
18939 Set the inner border thickness (i.e. the number of pixels between frames). For
18940 more advanced padding options (such as having different values for the edges),
18941 refer to the pad video filter.
18944 Specify the color of the unused area. For the syntax of this option, check the
18945 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18946 The default value of @var{color} is "black".
18949 Set the number of frames to overlap when tiling several successive frames together.
18950 The value must be between @code{0} and @var{nb_frames - 1}.
18953 Set the number of frames to initially be empty before displaying first output frame.
18954 This controls how soon will one get first output frame.
18955 The value must be between @code{0} and @var{nb_frames - 1}.
18958 @subsection Examples
18962 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18964 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18966 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18967 duplicating each output frame to accommodate the originally detected frame
18971 Display @code{5} pictures in an area of @code{3x2} frames,
18972 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18973 mixed flat and named options:
18975 tile=3x2:nb_frames=5:padding=7:margin=2
18979 @section tinterlace
18981 Perform various types of temporal field interlacing.
18983 Frames are counted starting from 1, so the first input frame is
18986 The filter accepts the following options:
18991 Specify the mode of the interlacing. This option can also be specified
18992 as a value alone. See below for a list of values for this option.
18994 Available values are:
18998 Move odd frames into the upper field, even into the lower field,
18999 generating a double height frame at half frame rate.
19003 Frame 1 Frame 2 Frame 3 Frame 4
19005 11111 22222 33333 44444
19006 11111 22222 33333 44444
19007 11111 22222 33333 44444
19008 11111 22222 33333 44444
19022 Only output odd frames, even frames are dropped, generating a frame with
19023 unchanged height at half frame rate.
19028 Frame 1 Frame 2 Frame 3 Frame 4
19030 11111 22222 33333 44444
19031 11111 22222 33333 44444
19032 11111 22222 33333 44444
19033 11111 22222 33333 44444
19043 Only output even frames, odd frames are dropped, generating a frame with
19044 unchanged height at half frame rate.
19049 Frame 1 Frame 2 Frame 3 Frame 4
19051 11111 22222 33333 44444
19052 11111 22222 33333 44444
19053 11111 22222 33333 44444
19054 11111 22222 33333 44444
19064 Expand each frame to full height, but pad alternate lines with black,
19065 generating a frame with double height at the same input frame rate.
19070 Frame 1 Frame 2 Frame 3 Frame 4
19072 11111 22222 33333 44444
19073 11111 22222 33333 44444
19074 11111 22222 33333 44444
19075 11111 22222 33333 44444
19078 11111 ..... 33333 .....
19079 ..... 22222 ..... 44444
19080 11111 ..... 33333 .....
19081 ..... 22222 ..... 44444
19082 11111 ..... 33333 .....
19083 ..... 22222 ..... 44444
19084 11111 ..... 33333 .....
19085 ..... 22222 ..... 44444
19089 @item interleave_top, 4
19090 Interleave the upper field from odd frames with the lower field from
19091 even frames, generating a frame with unchanged height at half frame rate.
19096 Frame 1 Frame 2 Frame 3 Frame 4
19098 11111<- 22222 33333<- 44444
19099 11111 22222<- 33333 44444<-
19100 11111<- 22222 33333<- 44444
19101 11111 22222<- 33333 44444<-
19111 @item interleave_bottom, 5
19112 Interleave the lower field from odd frames with the upper field from
19113 even frames, generating a frame with unchanged height at half frame rate.
19118 Frame 1 Frame 2 Frame 3 Frame 4
19120 11111 22222<- 33333 44444<-
19121 11111<- 22222 33333<- 44444
19122 11111 22222<- 33333 44444<-
19123 11111<- 22222 33333<- 44444
19133 @item interlacex2, 6
19134 Double frame rate with unchanged height. Frames are inserted each
19135 containing the second temporal field from the previous input frame and
19136 the first temporal field from the next input frame. This mode relies on
19137 the top_field_first flag. Useful for interlaced video displays with no
19138 field synchronisation.
19143 Frame 1 Frame 2 Frame 3 Frame 4
19145 11111 22222 33333 44444
19146 11111 22222 33333 44444
19147 11111 22222 33333 44444
19148 11111 22222 33333 44444
19151 11111 22222 22222 33333 33333 44444 44444
19152 11111 11111 22222 22222 33333 33333 44444
19153 11111 22222 22222 33333 33333 44444 44444
19154 11111 11111 22222 22222 33333 33333 44444
19159 Move odd frames into the upper field, even into the lower field,
19160 generating a double height frame at same frame rate.
19165 Frame 1 Frame 2 Frame 3 Frame 4
19167 11111 22222 33333 44444
19168 11111 22222 33333 44444
19169 11111 22222 33333 44444
19170 11111 22222 33333 44444
19173 11111 33333 33333 55555
19174 22222 22222 44444 44444
19175 11111 33333 33333 55555
19176 22222 22222 44444 44444
19177 11111 33333 33333 55555
19178 22222 22222 44444 44444
19179 11111 33333 33333 55555
19180 22222 22222 44444 44444
19185 Numeric values are deprecated but are accepted for backward
19186 compatibility reasons.
19188 Default mode is @code{merge}.
19191 Specify flags influencing the filter process.
19193 Available value for @var{flags} is:
19196 @item low_pass_filter, vlpf
19197 Enable linear vertical low-pass filtering in the filter.
19198 Vertical low-pass filtering is required when creating an interlaced
19199 destination from a progressive source which contains high-frequency
19200 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19203 @item complex_filter, cvlpf
19204 Enable complex vertical low-pass filtering.
19205 This will slightly less reduce interlace 'twitter' and Moire
19206 patterning but better retain detail and subjective sharpness impression.
19209 Bypass already interlaced frames, only adjust the frame rate.
19212 Vertical low-pass filtering and bypassing already interlaced frames can only be
19213 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19218 Pick median pixels from several successive input video frames.
19220 The filter accepts the following options:
19224 Set radius of median filter.
19225 Default is 1. Allowed range is from 1 to 127.
19228 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19231 Set median percentile. Default value is @code{0.5}.
19232 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19233 minimum values, and @code{1} maximum values.
19238 Mix successive video frames.
19240 A description of the accepted options follows.
19244 The number of successive frames to mix. If unspecified, it defaults to 3.
19247 Specify weight of each input video frame.
19248 Each weight is separated by space. If number of weights is smaller than
19249 number of @var{frames} last specified weight will be used for all remaining
19253 Specify scale, if it is set it will be multiplied with sum
19254 of each weight multiplied with pixel values to give final destination
19255 pixel value. By default @var{scale} is auto scaled to sum of weights.
19258 @subsection Examples
19262 Average 7 successive frames:
19264 tmix=frames=7:weights="1 1 1 1 1 1 1"
19268 Apply simple temporal convolution:
19270 tmix=frames=3:weights="-1 3 -1"
19274 Similar as above but only showing temporal differences:
19276 tmix=frames=3:weights="-1 2 -1":scale=1
19282 Tone map colors from different dynamic ranges.
19284 This filter expects data in single precision floating point, as it needs to
19285 operate on (and can output) out-of-range values. Another filter, such as
19286 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19288 The tonemapping algorithms implemented only work on linear light, so input
19289 data should be linearized beforehand (and possibly correctly tagged).
19292 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19295 @subsection Options
19296 The filter accepts the following options.
19300 Set the tone map algorithm to use.
19302 Possible values are:
19305 Do not apply any tone map, only desaturate overbright pixels.
19308 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19309 in-range values, while distorting out-of-range values.
19312 Stretch the entire reference gamut to a linear multiple of the display.
19315 Fit a logarithmic transfer between the tone curves.
19318 Preserve overall image brightness with a simple curve, using nonlinear
19319 contrast, which results in flattening details and degrading color accuracy.
19322 Preserve both dark and bright details better than @var{reinhard}, at the cost
19323 of slightly darkening everything. Use it when detail preservation is more
19324 important than color and brightness accuracy.
19327 Smoothly map out-of-range values, while retaining contrast and colors for
19328 in-range material as much as possible. Use it when color accuracy is more
19329 important than detail preservation.
19335 Tune the tone mapping algorithm.
19337 This affects the following algorithms:
19343 Specifies the scale factor to use while stretching.
19347 Specifies the exponent of the function.
19351 Specify an extra linear coefficient to multiply into the signal before clipping.
19355 Specify the local contrast coefficient at the display peak.
19356 Default to 0.5, which means that in-gamut values will be about half as bright
19363 Specify the transition point from linear to mobius transform. Every value
19364 below this point is guaranteed to be mapped 1:1. The higher the value, the
19365 more accurate the result will be, at the cost of losing bright details.
19366 Default to 0.3, which due to the steep initial slope still preserves in-range
19367 colors fairly accurately.
19371 Apply desaturation for highlights that exceed this level of brightness. The
19372 higher the parameter, the more color information will be preserved. This
19373 setting helps prevent unnaturally blown-out colors for super-highlights, by
19374 (smoothly) turning into white instead. This makes images feel more natural,
19375 at the cost of reducing information about out-of-range colors.
19377 The default of 2.0 is somewhat conservative and will mostly just apply to
19378 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19380 This option works only if the input frame has a supported color tag.
19383 Override signal/nominal/reference peak with this value. Useful when the
19384 embedded peak information in display metadata is not reliable or when tone
19385 mapping from a lower range to a higher range.
19390 Temporarily pad video frames.
19392 The filter accepts the following options:
19396 Specify number of delay frames before input video stream. Default is 0.
19399 Specify number of padding frames after input video stream.
19400 Set to -1 to pad indefinitely. Default is 0.
19403 Set kind of frames added to beginning of stream.
19404 Can be either @var{add} or @var{clone}.
19405 With @var{add} frames of solid-color are added.
19406 With @var{clone} frames are clones of first frame.
19407 Default is @var{add}.
19410 Set kind of frames added to end of stream.
19411 Can be either @var{add} or @var{clone}.
19412 With @var{add} frames of solid-color are added.
19413 With @var{clone} frames are clones of last frame.
19414 Default is @var{add}.
19416 @item start_duration, stop_duration
19417 Specify the duration of the start/stop delay. See
19418 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19419 for the accepted syntax.
19420 These options override @var{start} and @var{stop}. Default is 0.
19423 Specify the color of the padded area. For the syntax of this option,
19424 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19425 manual,ffmpeg-utils}.
19427 The default value of @var{color} is "black".
19433 Transpose rows with columns in the input video and optionally flip it.
19435 It accepts the following parameters:
19440 Specify the transposition direction.
19442 Can assume the following values:
19444 @item 0, 4, cclock_flip
19445 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19453 Rotate by 90 degrees clockwise, that is:
19461 Rotate by 90 degrees counterclockwise, that is:
19468 @item 3, 7, clock_flip
19469 Rotate by 90 degrees clockwise and vertically flip, that is:
19477 For values between 4-7, the transposition is only done if the input
19478 video geometry is portrait and not landscape. These values are
19479 deprecated, the @code{passthrough} option should be used instead.
19481 Numerical values are deprecated, and should be dropped in favor of
19482 symbolic constants.
19485 Do not apply the transposition if the input geometry matches the one
19486 specified by the specified value. It accepts the following values:
19489 Always apply transposition.
19491 Preserve portrait geometry (when @var{height} >= @var{width}).
19493 Preserve landscape geometry (when @var{width} >= @var{height}).
19496 Default value is @code{none}.
19499 For example to rotate by 90 degrees clockwise and preserve portrait
19502 transpose=dir=1:passthrough=portrait
19505 The command above can also be specified as:
19507 transpose=1:portrait
19510 @section transpose_npp
19512 Transpose rows with columns in the input video and optionally flip it.
19513 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19515 It accepts the following parameters:
19520 Specify the transposition direction.
19522 Can assume the following values:
19525 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19528 Rotate by 90 degrees clockwise.
19531 Rotate by 90 degrees counterclockwise.
19534 Rotate by 90 degrees clockwise and vertically flip.
19538 Do not apply the transposition if the input geometry matches the one
19539 specified by the specified value. It accepts the following values:
19542 Always apply transposition. (default)
19544 Preserve portrait geometry (when @var{height} >= @var{width}).
19546 Preserve landscape geometry (when @var{width} >= @var{height}).
19552 Trim the input so that the output contains one continuous subpart of the input.
19554 It accepts the following parameters:
19557 Specify the time of the start of the kept section, i.e. the frame with the
19558 timestamp @var{start} will be the first frame in the output.
19561 Specify the time of the first frame that will be dropped, i.e. the frame
19562 immediately preceding the one with the timestamp @var{end} will be the last
19563 frame in the output.
19566 This is the same as @var{start}, except this option sets the start timestamp
19567 in timebase units instead of seconds.
19570 This is the same as @var{end}, except this option sets the end timestamp
19571 in timebase units instead of seconds.
19574 The maximum duration of the output in seconds.
19577 The number of the first frame that should be passed to the output.
19580 The number of the first frame that should be dropped.
19583 @option{start}, @option{end}, and @option{duration} are expressed as time
19584 duration specifications; see
19585 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19586 for the accepted syntax.
19588 Note that the first two sets of the start/end options and the @option{duration}
19589 option look at the frame timestamp, while the _frame variants simply count the
19590 frames that pass through the filter. Also note that this filter does not modify
19591 the timestamps. If you wish for the output timestamps to start at zero, insert a
19592 setpts filter after the trim filter.
19594 If multiple start or end options are set, this filter tries to be greedy and
19595 keep all the frames that match at least one of the specified constraints. To keep
19596 only the part that matches all the constraints at once, chain multiple trim
19599 The defaults are such that all the input is kept. So it is possible to set e.g.
19600 just the end values to keep everything before the specified time.
19605 Drop everything except the second minute of input:
19607 ffmpeg -i INPUT -vf trim=60:120
19611 Keep only the first second:
19613 ffmpeg -i INPUT -vf trim=duration=1
19618 @section unpremultiply
19619 Apply alpha unpremultiply effect to input video stream using first plane
19620 of second stream as alpha.
19622 Both streams must have same dimensions and same pixel format.
19624 The filter accepts the following option:
19628 Set which planes will be processed, unprocessed planes will be copied.
19629 By default value 0xf, all planes will be processed.
19631 If the format has 1 or 2 components, then luma is bit 0.
19632 If the format has 3 or 4 components:
19633 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19634 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19635 If present, the alpha channel is always the last bit.
19638 Do not require 2nd input for processing, instead use alpha plane from input stream.
19644 Sharpen or blur the input video.
19646 It accepts the following parameters:
19649 @item luma_msize_x, lx
19650 Set the luma matrix horizontal size. It must be an odd integer between
19651 3 and 23. The default value is 5.
19653 @item luma_msize_y, ly
19654 Set the luma matrix vertical size. It must be an odd integer between 3
19655 and 23. The default value is 5.
19657 @item luma_amount, la
19658 Set the luma effect strength. It must be a floating point number, reasonable
19659 values lay between -1.5 and 1.5.
19661 Negative values will blur the input video, while positive values will
19662 sharpen it, a value of zero will disable the effect.
19664 Default value is 1.0.
19666 @item chroma_msize_x, cx
19667 Set the chroma matrix horizontal size. It must be an odd integer
19668 between 3 and 23. The default value is 5.
19670 @item chroma_msize_y, cy
19671 Set the chroma matrix vertical size. It must be an odd integer
19672 between 3 and 23. The default value is 5.
19674 @item chroma_amount, ca
19675 Set the chroma effect strength. It must be a floating point number, reasonable
19676 values lay between -1.5 and 1.5.
19678 Negative values will blur the input video, while positive values will
19679 sharpen it, a value of zero will disable the effect.
19681 Default value is 0.0.
19685 All parameters are optional and default to the equivalent of the
19686 string '5:5:1.0:5:5:0.0'.
19688 @subsection Examples
19692 Apply strong luma sharpen effect:
19694 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19698 Apply a strong blur of both luma and chroma parameters:
19700 unsharp=7:7:-2:7:7:-2
19707 Decompose a video made of tiled images into the individual images.
19709 The frame rate of the output video is the frame rate of the input video
19710 multiplied by the number of tiles.
19712 This filter does the reverse of @ref{tile}.
19714 The filter accepts the following options:
19719 Set the grid size (i.e. the number of lines and columns). For the syntax of
19720 this option, check the
19721 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19724 @subsection Examples
19728 Produce a 1-second video from a still image file made of 25 frames stacked
19729 vertically, like an analogic film reel:
19731 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19737 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19738 the image at several (or - in the case of @option{quality} level @code{8} - all)
19739 shifts and average the results.
19741 The way this differs from the behavior of spp is that uspp actually encodes &
19742 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19743 DCT similar to MJPEG.
19745 The filter accepts the following options:
19749 Set quality. This option defines the number of levels for averaging. It accepts
19750 an integer in the range 0-8. If set to @code{0}, the filter will have no
19751 effect. A value of @code{8} means the higher quality. For each increment of
19752 that value the speed drops by a factor of approximately 2. Default value is
19756 Force a constant quantization parameter. If not set, the filter will use the QP
19757 from the video stream (if available).
19762 Convert 360 videos between various formats.
19764 The filter accepts the following options:
19770 Set format of the input/output video.
19778 Equirectangular projection.
19783 Cubemap with 3x2/6x1/1x6 layout.
19785 Format specific options:
19790 Set padding proportion for the input/output cubemap. Values in decimals.
19797 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)
19800 Default value is @b{@samp{0}}.
19801 Maximum value is @b{@samp{0.1}}.
19805 Set fixed padding for the input/output cubemap. Values in pixels.
19807 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19811 Set order of faces for the input/output cubemap. Choose one direction for each position.
19813 Designation of directions:
19829 Default value is @b{@samp{rludfb}}.
19833 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
19835 Designation of angles:
19838 0 degrees clockwise
19840 90 degrees clockwise
19842 180 degrees clockwise
19844 270 degrees clockwise
19847 Default value is @b{@samp{000000}}.
19851 Equi-Angular Cubemap.
19858 Format specific options:
19863 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19865 If diagonal field of view is set it overrides horizontal and vertical field of view.
19870 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19872 If diagonal field of view is set it overrides horizontal and vertical field of view.
19878 Format specific options:
19883 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19885 If diagonal field of view is set it overrides horizontal and vertical field of view.
19890 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19892 If diagonal field of view is set it overrides horizontal and vertical field of view.
19898 Facebook's 360 formats.
19901 Stereographic format.
19903 Format specific options:
19908 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19910 If diagonal field of view is set it overrides horizontal and vertical field of view.
19915 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19917 If diagonal field of view is set it overrides horizontal and vertical field of view.
19924 Ball format, gives significant distortion toward the back.
19927 Hammer-Aitoff map projection format.
19930 Sinusoidal map projection format.
19933 Fisheye projection.
19935 Format specific options:
19940 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19942 If diagonal field of view is set it overrides horizontal and vertical field of view.
19947 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19949 If diagonal field of view is set it overrides horizontal and vertical field of view.
19953 Pannini projection.
19955 Format specific options:
19958 Set output pannini parameter.
19961 Set input pannini parameter.
19965 Cylindrical projection.
19967 Format specific options:
19972 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19974 If diagonal field of view is set it overrides horizontal and vertical field of view.
19979 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19981 If diagonal field of view is set it overrides horizontal and vertical field of view.
19985 Perspective projection. @i{(output only)}
19987 Format specific options:
19990 Set perspective parameter.
19994 Tetrahedron projection.
19997 Truncated square pyramid projection.
20001 Half equirectangular projection.
20006 Format specific options:
20011 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20013 If diagonal field of view is set it overrides horizontal and vertical field of view.
20018 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20020 If diagonal field of view is set it overrides horizontal and vertical field of view.
20024 Orthographic format.
20026 Format specific options:
20031 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20033 If diagonal field of view is set it overrides horizontal and vertical field of view.
20038 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20040 If diagonal field of view is set it overrides horizontal and vertical field of view.
20044 Octahedron projection.
20048 Set interpolation method.@*
20049 @i{Note: more complex interpolation methods require much more memory to run.}
20059 Bilinear interpolation.
20061 Lagrange9 interpolation.
20064 Bicubic interpolation.
20067 Lanczos interpolation.
20070 Spline16 interpolation.
20073 Gaussian interpolation.
20075 Mitchell interpolation.
20078 Default value is @b{@samp{line}}.
20082 Set the output video resolution.
20084 Default resolution depends on formats.
20088 Set the input/output stereo format.
20099 Default value is @b{@samp{2d}} for input and output format.
20104 Set rotation for the output video. Values in degrees.
20107 Set rotation order for the output video. Choose one item for each position.
20118 Default value is @b{@samp{ypr}}.
20123 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20127 Set if input video is flipped horizontally/vertically. Boolean values.
20130 Set if input video is transposed. Boolean value, by default disabled.
20133 Set if output video needs to be transposed. Boolean value, by default disabled.
20136 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20139 @subsection Examples
20143 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20145 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20148 Extract back view of Equi-Angular Cubemap:
20150 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20153 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20155 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20159 @subsection Commands
20161 This filter supports subset of above options as @ref{commands}.
20163 @section vaguedenoiser
20165 Apply a wavelet based denoiser.
20167 It transforms each frame from the video input into the wavelet domain,
20168 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20169 the obtained coefficients. It does an inverse wavelet transform after.
20170 Due to wavelet properties, it should give a nice smoothed result, and
20171 reduced noise, without blurring picture features.
20173 This filter accepts the following options:
20177 The filtering strength. The higher, the more filtered the video will be.
20178 Hard thresholding can use a higher threshold than soft thresholding
20179 before the video looks overfiltered. Default value is 2.
20182 The filtering method the filter will use.
20184 It accepts the following values:
20187 All values under the threshold will be zeroed.
20190 All values under the threshold will be zeroed. All values above will be
20191 reduced by the threshold.
20194 Scales or nullifies coefficients - intermediary between (more) soft and
20195 (less) hard thresholding.
20198 Default is garrote.
20201 Number of times, the wavelet will decompose the picture. Picture can't
20202 be decomposed beyond a particular point (typically, 8 for a 640x480
20203 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20206 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20209 A list of the planes to process. By default all planes are processed.
20212 The threshold type the filter will use.
20214 It accepts the following values:
20217 Threshold used is same for all decompositions.
20220 Threshold used depends also on each decomposition coefficients.
20223 Default is universal.
20226 @section vectorscope
20228 Display 2 color component values in the two dimensional graph (which is called
20231 This filter accepts the following options:
20235 Set vectorscope mode.
20237 It accepts the following values:
20241 Gray values are displayed on graph, higher brightness means more pixels have
20242 same component color value on location in graph. This is the default mode.
20245 Gray values are displayed on graph. Surrounding pixels values which are not
20246 present in video frame are drawn in gradient of 2 color components which are
20247 set by option @code{x} and @code{y}. The 3rd color component is static.
20250 Actual color components values present in video frame are displayed on graph.
20253 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20254 on graph increases value of another color component, which is luminance by
20255 default values of @code{x} and @code{y}.
20258 Actual colors present in video frame are displayed on graph. If two different
20259 colors map to same position on graph then color with higher value of component
20260 not present in graph is picked.
20263 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20264 component picked from radial gradient.
20268 Set which color component will be represented on X-axis. Default is @code{1}.
20271 Set which color component will be represented on Y-axis. Default is @code{2}.
20274 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20275 of color component which represents frequency of (X, Y) location in graph.
20280 No envelope, this is default.
20283 Instant envelope, even darkest single pixel will be clearly highlighted.
20286 Hold maximum and minimum values presented in graph over time. This way you
20287 can still spot out of range values without constantly looking at vectorscope.
20290 Peak and instant envelope combined together.
20294 Set what kind of graticule to draw.
20303 Set graticule opacity.
20306 Set graticule flags.
20310 Draw graticule for white point.
20313 Draw graticule for black point.
20316 Draw color points short names.
20320 Set background opacity.
20322 @item lthreshold, l
20323 Set low threshold for color component not represented on X or Y axis.
20324 Values lower than this value will be ignored. Default is 0.
20325 Note this value is multiplied with actual max possible value one pixel component
20326 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20329 @item hthreshold, h
20330 Set high threshold for color component not represented on X or Y axis.
20331 Values higher than this value will be ignored. Default is 1.
20332 Note this value is multiplied with actual max possible value one pixel component
20333 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20334 is 0.9 * 255 = 230.
20336 @item colorspace, c
20337 Set what kind of colorspace to use when drawing graticule.
20347 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20348 This means no tint, and output will remain gray.
20351 @anchor{vidstabdetect}
20352 @section vidstabdetect
20354 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20355 @ref{vidstabtransform} for pass 2.
20357 This filter generates a file with relative translation and rotation
20358 transform information about subsequent frames, which is then used by
20359 the @ref{vidstabtransform} filter.
20361 To enable compilation of this filter you need to configure FFmpeg with
20362 @code{--enable-libvidstab}.
20364 This filter accepts the following options:
20368 Set the path to the file used to write the transforms information.
20369 Default value is @file{transforms.trf}.
20372 Set how shaky the video is and how quick the camera is. It accepts an
20373 integer in the range 1-10, a value of 1 means little shakiness, a
20374 value of 10 means strong shakiness. Default value is 5.
20377 Set the accuracy of the detection process. It must be a value in the
20378 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20379 accuracy. Default value is 15.
20382 Set stepsize of the search process. The region around minimum is
20383 scanned with 1 pixel resolution. Default value is 6.
20386 Set minimum contrast. Below this value a local measurement field is
20387 discarded. Must be a floating point value in the range 0-1. Default
20391 Set reference frame number for tripod mode.
20393 If enabled, the motion of the frames is compared to a reference frame
20394 in the filtered stream, identified by the specified number. The idea
20395 is to compensate all movements in a more-or-less static scene and keep
20396 the camera view absolutely still.
20398 If set to 0, it is disabled. The frames are counted starting from 1.
20401 Show fields and transforms in the resulting frames. It accepts an
20402 integer in the range 0-2. Default value is 0, which disables any
20406 @subsection Examples
20410 Use default values:
20416 Analyze strongly shaky movie and put the results in file
20417 @file{mytransforms.trf}:
20419 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20423 Visualize the result of internal transformations in the resulting
20426 vidstabdetect=show=1
20430 Analyze a video with medium shakiness using @command{ffmpeg}:
20432 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20436 @anchor{vidstabtransform}
20437 @section vidstabtransform
20439 Video stabilization/deshaking: pass 2 of 2,
20440 see @ref{vidstabdetect} for pass 1.
20442 Read a file with transform information for each frame and
20443 apply/compensate them. Together with the @ref{vidstabdetect}
20444 filter this can be used to deshake videos. See also
20445 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20446 the @ref{unsharp} filter, see below.
20448 To enable compilation of this filter you need to configure FFmpeg with
20449 @code{--enable-libvidstab}.
20451 @subsection Options
20455 Set path to the file used to read the transforms. Default value is
20456 @file{transforms.trf}.
20459 Set the number of frames (value*2 + 1) used for lowpass filtering the
20460 camera movements. Default value is 10.
20462 For example a number of 10 means that 21 frames are used (10 in the
20463 past and 10 in the future) to smoothen the motion in the video. A
20464 larger value leads to a smoother video, but limits the acceleration of
20465 the camera (pan/tilt movements). 0 is a special case where a static
20466 camera is simulated.
20469 Set the camera path optimization algorithm.
20471 Accepted values are:
20474 gaussian kernel low-pass filter on camera motion (default)
20476 averaging on transformations
20480 Set maximal number of pixels to translate frames. Default value is -1,
20484 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20485 value is -1, meaning no limit.
20488 Specify how to deal with borders that may be visible due to movement
20491 Available values are:
20494 keep image information from previous frame (default)
20496 fill the border black
20500 Invert transforms if set to 1. Default value is 0.
20503 Consider transforms as relative to previous frame if set to 1,
20504 absolute if set to 0. Default value is 0.
20507 Set percentage to zoom. A positive value will result in a zoom-in
20508 effect, a negative value in a zoom-out effect. Default value is 0 (no
20512 Set optimal zooming to avoid borders.
20514 Accepted values are:
20519 optimal static zoom value is determined (only very strong movements
20520 will lead to visible borders) (default)
20522 optimal adaptive zoom value is determined (no borders will be
20523 visible), see @option{zoomspeed}
20526 Note that the value given at zoom is added to the one calculated here.
20529 Set percent to zoom maximally each frame (enabled when
20530 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20534 Specify type of interpolation.
20536 Available values are:
20541 linear only horizontal
20543 linear in both directions (default)
20545 cubic in both directions (slow)
20549 Enable virtual tripod mode if set to 1, which is equivalent to
20550 @code{relative=0:smoothing=0}. Default value is 0.
20552 Use also @code{tripod} option of @ref{vidstabdetect}.
20555 Increase log verbosity if set to 1. Also the detected global motions
20556 are written to the temporary file @file{global_motions.trf}. Default
20560 @subsection Examples
20564 Use @command{ffmpeg} for a typical stabilization with default values:
20566 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20569 Note the use of the @ref{unsharp} filter which is always recommended.
20572 Zoom in a bit more and load transform data from a given file:
20574 vidstabtransform=zoom=5:input="mytransforms.trf"
20578 Smoothen the video even more:
20580 vidstabtransform=smoothing=30
20586 Flip the input video vertically.
20588 For example, to vertically flip a video with @command{ffmpeg}:
20590 ffmpeg -i in.avi -vf "vflip" out.avi
20595 Detect variable frame rate video.
20597 This filter tries to detect if the input is variable or constant frame rate.
20599 At end it will output number of frames detected as having variable delta pts,
20600 and ones with constant delta pts.
20601 If there was frames with variable delta, than it will also show min, max and
20602 average delta encountered.
20606 Boost or alter saturation.
20608 The filter accepts the following options:
20611 Set strength of boost if positive value or strength of alter if negative value.
20612 Default is 0. Allowed range is from -2 to 2.
20615 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20618 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20621 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20624 Set the red luma coefficient.
20627 Set the green luma coefficient.
20630 Set the blue luma coefficient.
20633 If @code{intensity} is negative and this is set to 1, colors will change,
20634 otherwise colors will be less saturated, more towards gray.
20637 @subsection Commands
20639 This filter supports the all above options as @ref{commands}.
20644 Make or reverse a natural vignetting effect.
20646 The filter accepts the following options:
20650 Set lens angle expression as a number of radians.
20652 The value is clipped in the @code{[0,PI/2]} range.
20654 Default value: @code{"PI/5"}
20658 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20662 Set forward/backward mode.
20664 Available modes are:
20667 The larger the distance from the central point, the darker the image becomes.
20670 The larger the distance from the central point, the brighter the image becomes.
20671 This can be used to reverse a vignette effect, though there is no automatic
20672 detection to extract the lens @option{angle} and other settings (yet). It can
20673 also be used to create a burning effect.
20676 Default value is @samp{forward}.
20679 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20681 It accepts the following values:
20684 Evaluate expressions only once during the filter initialization.
20687 Evaluate expressions for each incoming frame. This is way slower than the
20688 @samp{init} mode since it requires all the scalers to be re-computed, but it
20689 allows advanced dynamic expressions.
20692 Default value is @samp{init}.
20695 Set dithering to reduce the circular banding effects. Default is @code{1}
20699 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20700 Setting this value to the SAR of the input will make a rectangular vignetting
20701 following the dimensions of the video.
20703 Default is @code{1/1}.
20706 @subsection Expressions
20708 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20709 following parameters.
20714 input width and height
20717 the number of input frame, starting from 0
20720 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20721 @var{TB} units, NAN if undefined
20724 frame rate of the input video, NAN if the input frame rate is unknown
20727 the PTS (Presentation TimeStamp) of the filtered video frame,
20728 expressed in seconds, NAN if undefined
20731 time base of the input video
20735 @subsection Examples
20739 Apply simple strong vignetting effect:
20745 Make a flickering vignetting:
20747 vignette='PI/4+random(1)*PI/50':eval=frame
20752 @section vmafmotion
20754 Obtain the average VMAF motion score of a video.
20755 It is one of the component metrics of VMAF.
20757 The obtained average motion score is printed through the logging system.
20759 The filter accepts the following options:
20763 If specified, the filter will use the named file to save the motion score of
20764 each frame with respect to the previous frame.
20765 When filename equals "-" the data is sent to standard output.
20770 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20774 Stack input videos vertically.
20776 All streams must be of same pixel format and of same width.
20778 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20779 to create same output.
20781 The filter accepts the following options:
20785 Set number of input streams. Default is 2.
20788 If set to 1, force the output to terminate when the shortest input
20789 terminates. Default value is 0.
20794 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20795 Deinterlacing Filter").
20797 Based on the process described by Martin Weston for BBC R&D, and
20798 implemented based on the de-interlace algorithm written by Jim
20799 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20800 uses filter coefficients calculated by BBC R&D.
20802 This filter uses field-dominance information in frame to decide which
20803 of each pair of fields to place first in the output.
20804 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20806 There are two sets of filter coefficients, so called "simple"
20807 and "complex". Which set of filter coefficients is used can
20808 be set by passing an optional parameter:
20812 Set the interlacing filter coefficients. Accepts one of the following values:
20816 Simple filter coefficient set.
20818 More-complex filter coefficient set.
20820 Default value is @samp{complex}.
20823 Specify which frames to deinterlace. Accepts one of the following values:
20827 Deinterlace all frames,
20829 Only deinterlace frames marked as interlaced.
20832 Default value is @samp{all}.
20836 Video waveform monitor.
20838 The waveform monitor plots color component intensity. By default luminance
20839 only. Each column of the waveform corresponds to a column of pixels in the
20842 It accepts the following options:
20846 Can be either @code{row}, or @code{column}. Default is @code{column}.
20847 In row mode, the graph on the left side represents color component value 0 and
20848 the right side represents value = 255. In column mode, the top side represents
20849 color component value = 0 and bottom side represents value = 255.
20852 Set intensity. Smaller values are useful to find out how many values of the same
20853 luminance are distributed across input rows/columns.
20854 Default value is @code{0.04}. Allowed range is [0, 1].
20857 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
20858 In mirrored mode, higher values will be represented on the left
20859 side for @code{row} mode and at the top for @code{column} mode. Default is
20860 @code{1} (mirrored).
20864 It accepts the following values:
20867 Presents information identical to that in the @code{parade}, except
20868 that the graphs representing color components are superimposed directly
20871 This display mode makes it easier to spot relative differences or similarities
20872 in overlapping areas of the color components that are supposed to be identical,
20873 such as neutral whites, grays, or blacks.
20876 Display separate graph for the color components side by side in
20877 @code{row} mode or one below the other in @code{column} mode.
20880 Display separate graph for the color components side by side in
20881 @code{column} mode or one below the other in @code{row} mode.
20883 Using this display mode makes it easy to spot color casts in the highlights
20884 and shadows of an image, by comparing the contours of the top and the bottom
20885 graphs of each waveform. Since whites, grays, and blacks are characterized
20886 by exactly equal amounts of red, green, and blue, neutral areas of the picture
20887 should display three waveforms of roughly equal width/height. If not, the
20888 correction is easy to perform by making level adjustments the three waveforms.
20890 Default is @code{stack}.
20892 @item components, c
20893 Set which color components to display. Default is 1, which means only luminance
20894 or red color component if input is in RGB colorspace. If is set for example to
20895 7 it will display all 3 (if) available color components.
20900 No envelope, this is default.
20903 Instant envelope, minimum and maximum values presented in graph will be easily
20904 visible even with small @code{step} value.
20907 Hold minimum and maximum values presented in graph across time. This way you
20908 can still spot out of range values without constantly looking at waveforms.
20911 Peak and instant envelope combined together.
20917 No filtering, this is default.
20920 Luma and chroma combined together.
20923 Similar as above, but shows difference between blue and red chroma.
20926 Similar as above, but use different colors.
20929 Similar as above, but again with different colors.
20932 Displays only chroma.
20935 Displays actual color value on waveform.
20938 Similar as above, but with luma showing frequency of chroma values.
20942 Set which graticule to display.
20946 Do not display graticule.
20949 Display green graticule showing legal broadcast ranges.
20952 Display orange graticule showing legal broadcast ranges.
20955 Display invert graticule showing legal broadcast ranges.
20959 Set graticule opacity.
20962 Set graticule flags.
20966 Draw numbers above lines. By default enabled.
20969 Draw dots instead of lines.
20973 Set scale used for displaying graticule.
20980 Default is digital.
20983 Set background opacity.
20987 Set tint for output.
20988 Only used with lowpass filter and when display is not overlay and input
20989 pixel formats are not RGB.
20992 @section weave, doubleweave
20994 The @code{weave} takes a field-based video input and join
20995 each two sequential fields into single frame, producing a new double
20996 height clip with half the frame rate and half the frame count.
20998 The @code{doubleweave} works same as @code{weave} but without
20999 halving frame rate and frame count.
21001 It accepts the following option:
21005 Set first field. Available values are:
21009 Set the frame as top-field-first.
21012 Set the frame as bottom-field-first.
21016 @subsection Examples
21020 Interlace video using @ref{select} and @ref{separatefields} filter:
21022 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21027 Apply the xBR high-quality magnification filter which is designed for pixel
21028 art. It follows a set of edge-detection rules, see
21029 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21031 It accepts the following option:
21035 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21036 @code{3xBR} and @code{4} for @code{4xBR}.
21037 Default is @code{3}.
21042 Apply cross fade from one input video stream to another input video stream.
21043 The cross fade is applied for specified duration.
21045 The filter accepts the following options:
21049 Set one of available transition effects:
21097 Default transition effect is fade.
21100 Set cross fade duration in seconds.
21101 Default duration is 1 second.
21104 Set cross fade start relative to first input stream in seconds.
21105 Default offset is 0.
21108 Set expression for custom transition effect.
21110 The expressions can use the following variables and functions:
21115 The coordinates of the current sample.
21119 The width and height of the image.
21122 Progress of transition effect.
21125 Currently processed plane.
21128 Return value of first input at current location and plane.
21131 Return value of second input at current location and plane.
21137 Return the value of the pixel at location (@var{x},@var{y}) of the
21138 first/second/third/fourth component of first input.
21144 Return the value of the pixel at location (@var{x},@var{y}) of the
21145 first/second/third/fourth component of second input.
21149 @subsection Examples
21153 Cross fade from one input video to another input video, with fade transition and duration of transition
21154 of 2 seconds starting at offset of 5 seconds:
21156 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21161 Pick median pixels from several input videos.
21163 The filter accepts the following options:
21167 Set number of inputs.
21168 Default is 3. Allowed range is from 3 to 255.
21169 If number of inputs is even number, than result will be mean value between two median values.
21172 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21175 Set median percentile. Default value is @code{0.5}.
21176 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21177 minimum values, and @code{1} maximum values.
21181 Stack video inputs into custom layout.
21183 All streams must be of same pixel format.
21185 The filter accepts the following options:
21189 Set number of input streams. Default is 2.
21192 Specify layout of inputs.
21193 This option requires the desired layout configuration to be explicitly set by the user.
21194 This sets position of each video input in output. Each input
21195 is separated by '|'.
21196 The first number represents the column, and the second number represents the row.
21197 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21198 where X is video input from which to take width or height.
21199 Multiple values can be used when separated by '+'. In such
21200 case values are summed together.
21202 Note that if inputs are of different sizes gaps may appear, as not all of
21203 the output video frame will be filled. Similarly, videos can overlap each
21204 other if their position doesn't leave enough space for the full frame of
21207 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21208 a layout must be set by the user.
21211 If set to 1, force the output to terminate when the shortest input
21212 terminates. Default value is 0.
21215 If set to valid color, all unused pixels will be filled with that color.
21216 By default fill is set to none, so it is disabled.
21219 @subsection Examples
21223 Display 4 inputs into 2x2 grid.
21227 input1(0, 0) | input3(w0, 0)
21228 input2(0, h0) | input4(w0, h0)
21232 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21235 Note that if inputs are of different sizes, gaps or overlaps may occur.
21238 Display 4 inputs into 1x4 grid.
21245 input4(0, h0+h1+h2)
21249 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21252 Note that if inputs are of different widths, unused space will appear.
21255 Display 9 inputs into 3x3 grid.
21259 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21260 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21261 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21265 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
21268 Note that if inputs are of different sizes, gaps or overlaps may occur.
21271 Display 16 inputs into 4x4 grid.
21275 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21276 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21277 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21278 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21282 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|
21283 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
21286 Note that if inputs are of different sizes, gaps or overlaps may occur.
21293 Deinterlace the input video ("yadif" means "yet another deinterlacing
21296 It accepts the following parameters:
21302 The interlacing mode to adopt. It accepts one of the following values:
21305 @item 0, send_frame
21306 Output one frame for each frame.
21307 @item 1, send_field
21308 Output one frame for each field.
21309 @item 2, send_frame_nospatial
21310 Like @code{send_frame}, but it skips the spatial interlacing check.
21311 @item 3, send_field_nospatial
21312 Like @code{send_field}, but it skips the spatial interlacing check.
21315 The default value is @code{send_frame}.
21318 The picture field parity assumed for the input interlaced video. It accepts one
21319 of the following values:
21323 Assume the top field is first.
21325 Assume the bottom field is first.
21327 Enable automatic detection of field parity.
21330 The default value is @code{auto}.
21331 If the interlacing is unknown or the decoder does not export this information,
21332 top field first will be assumed.
21335 Specify which frames to deinterlace. Accepts one of the following
21340 Deinterlace all frames.
21341 @item 1, interlaced
21342 Only deinterlace frames marked as interlaced.
21345 The default value is @code{all}.
21348 @section yadif_cuda
21350 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21351 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21354 It accepts the following parameters:
21360 The interlacing mode to adopt. It accepts one of the following values:
21363 @item 0, send_frame
21364 Output one frame for each frame.
21365 @item 1, send_field
21366 Output one frame for each field.
21367 @item 2, send_frame_nospatial
21368 Like @code{send_frame}, but it skips the spatial interlacing check.
21369 @item 3, send_field_nospatial
21370 Like @code{send_field}, but it skips the spatial interlacing check.
21373 The default value is @code{send_frame}.
21376 The picture field parity assumed for the input interlaced video. It accepts one
21377 of the following values:
21381 Assume the top field is first.
21383 Assume the bottom field is first.
21385 Enable automatic detection of field parity.
21388 The default value is @code{auto}.
21389 If the interlacing is unknown or the decoder does not export this information,
21390 top field first will be assumed.
21393 Specify which frames to deinterlace. Accepts one of the following
21398 Deinterlace all frames.
21399 @item 1, interlaced
21400 Only deinterlace frames marked as interlaced.
21403 The default value is @code{all}.
21408 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21409 The algorithm is described in
21410 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21412 It accepts the following parameters:
21416 Set the window radius. Default value is 3.
21419 Set which planes to filter. Default is only the first plane.
21422 Set blur strength. Default value is 128.
21425 @subsection Commands
21426 This filter supports same @ref{commands} as options.
21430 Apply Zoom & Pan effect.
21432 This filter accepts the following options:
21436 Set the zoom expression. Range is 1-10. Default is 1.
21440 Set the x and y expression. Default is 0.
21443 Set the duration expression in number of frames.
21444 This sets for how many number of frames effect will last for
21445 single input image.
21448 Set the output image size, default is 'hd720'.
21451 Set the output frame rate, default is '25'.
21454 Each expression can contain the following constants:
21473 Output frame count.
21476 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21478 @item out_time, time, ot
21479 The output timestamp expressed in seconds.
21483 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21484 for current input frame.
21488 'x' and 'y' of last output frame of previous input frame or 0 when there was
21489 not yet such frame (first input frame).
21492 Last calculated zoom from 'z' expression for current input frame.
21495 Last calculated zoom of last output frame of previous input frame.
21498 Number of output frames for current input frame. Calculated from 'd' expression
21499 for each input frame.
21502 number of output frames created for previous input frame
21505 Rational number: input width / input height
21508 sample aspect ratio
21511 display aspect ratio
21515 @subsection Examples
21519 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21521 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
21525 Zoom in up to 1.5x and pan always at center of picture:
21527 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21531 Same as above but without pausing:
21533 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21537 Zoom in 2x into center of picture only for the first second of the input video:
21539 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21546 Scale (resize) the input video, using the z.lib library:
21547 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21548 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21550 The zscale filter forces the output display aspect ratio to be the same
21551 as the input, by changing the output sample aspect ratio.
21553 If the input image format is different from the format requested by
21554 the next filter, the zscale filter will convert the input to the
21557 @subsection Options
21558 The filter accepts the following options.
21563 Set the output video dimension expression. Default value is the input
21566 If the @var{width} or @var{w} value is 0, the input width is used for
21567 the output. If the @var{height} or @var{h} value is 0, the input height
21568 is used for the output.
21570 If one and only one of the values is -n with n >= 1, the zscale filter
21571 will use a value that maintains the aspect ratio of the input image,
21572 calculated from the other specified dimension. After that it will,
21573 however, make sure that the calculated dimension is divisible by n and
21574 adjust the value if necessary.
21576 If both values are -n with n >= 1, the behavior will be identical to
21577 both values being set to 0 as previously detailed.
21579 See below for the list of accepted constants for use in the dimension
21583 Set the video size. For the syntax of this option, check the
21584 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21587 Set the dither type.
21589 Possible values are:
21594 @item error_diffusion
21600 Set the resize filter type.
21602 Possible values are:
21612 Default is bilinear.
21615 Set the color range.
21617 Possible values are:
21624 Default is same as input.
21627 Set the color primaries.
21629 Possible values are:
21639 Default is same as input.
21642 Set the transfer characteristics.
21644 Possible values are:
21658 Default is same as input.
21661 Set the colorspace matrix.
21663 Possible value are:
21674 Default is same as input.
21677 Set the input color range.
21679 Possible values are:
21686 Default is same as input.
21688 @item primariesin, pin
21689 Set the input color primaries.
21691 Possible values are:
21701 Default is same as input.
21703 @item transferin, tin
21704 Set the input transfer characteristics.
21706 Possible values are:
21717 Default is same as input.
21719 @item matrixin, min
21720 Set the input colorspace matrix.
21722 Possible value are:
21734 Set the output chroma location.
21736 Possible values are:
21747 @item chromalin, cin
21748 Set the input chroma location.
21750 Possible values are:
21762 Set the nominal peak luminance.
21765 The values of the @option{w} and @option{h} options are expressions
21766 containing the following constants:
21771 The input width and height
21775 These are the same as @var{in_w} and @var{in_h}.
21779 The output (scaled) width and height
21783 These are the same as @var{out_w} and @var{out_h}
21786 The same as @var{iw} / @var{ih}
21789 input sample aspect ratio
21792 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
21796 horizontal and vertical input chroma subsample values. For example for the
21797 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21801 horizontal and vertical output chroma subsample values. For example for the
21802 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21805 @subsection Commands
21807 This filter supports the following commands:
21811 Set the output video dimension expression.
21812 The command accepts the same syntax of the corresponding option.
21814 If the specified expression is not valid, it is kept at its current
21818 @c man end VIDEO FILTERS
21820 @chapter OpenCL Video Filters
21821 @c man begin OPENCL VIDEO FILTERS
21823 Below is a description of the currently available OpenCL video filters.
21825 To enable compilation of these filters you need to configure FFmpeg with
21826 @code{--enable-opencl}.
21828 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
21831 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
21832 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
21833 given device parameters.
21835 @item -filter_hw_device @var{name}
21836 Pass the hardware device called @var{name} to all filters in any filter graph.
21840 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
21844 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
21846 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
21850 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.
21852 @section avgblur_opencl
21854 Apply average blur filter.
21856 The filter accepts the following options:
21860 Set horizontal radius size.
21861 Range is @code{[1, 1024]} and default value is @code{1}.
21864 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21867 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
21870 @subsection Example
21874 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.
21876 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
21880 @section boxblur_opencl
21882 Apply a boxblur algorithm to the input video.
21884 It accepts the following parameters:
21888 @item luma_radius, lr
21889 @item luma_power, lp
21890 @item chroma_radius, cr
21891 @item chroma_power, cp
21892 @item alpha_radius, ar
21893 @item alpha_power, ap
21897 A description of the accepted options follows.
21900 @item luma_radius, lr
21901 @item chroma_radius, cr
21902 @item alpha_radius, ar
21903 Set an expression for the box radius in pixels used for blurring the
21904 corresponding input plane.
21906 The radius value must be a non-negative number, and must not be
21907 greater than the value of the expression @code{min(w,h)/2} for the
21908 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
21911 Default value for @option{luma_radius} is "2". If not specified,
21912 @option{chroma_radius} and @option{alpha_radius} default to the
21913 corresponding value set for @option{luma_radius}.
21915 The expressions can contain the following constants:
21919 The input width and height in pixels.
21923 The input chroma image width and height in pixels.
21927 The horizontal and vertical chroma subsample values. For example, for the
21928 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
21931 @item luma_power, lp
21932 @item chroma_power, cp
21933 @item alpha_power, ap
21934 Specify how many times the boxblur filter is applied to the
21935 corresponding plane.
21937 Default value for @option{luma_power} is 2. If not specified,
21938 @option{chroma_power} and @option{alpha_power} default to the
21939 corresponding value set for @option{luma_power}.
21941 A value of 0 will disable the effect.
21944 @subsection Examples
21946 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.
21950 Apply a boxblur filter with the luma, chroma, and alpha radius
21951 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.
21953 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
21954 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
21958 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.
21960 For the luma plane, a 2x2 box radius will be run once.
21962 For the chroma plane, a 4x4 box radius will be run 5 times.
21964 For the alpha plane, a 3x3 box radius will be run 7 times.
21966 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
21970 @section colorkey_opencl
21971 RGB colorspace color keying.
21973 The filter accepts the following options:
21977 The color which will be replaced with transparency.
21980 Similarity percentage with the key color.
21982 0.01 matches only the exact key color, while 1.0 matches everything.
21987 0.0 makes pixels either fully transparent, or not transparent at all.
21989 Higher values result in semi-transparent pixels, with a higher transparency
21990 the more similar the pixels color is to the key color.
21993 @subsection Examples
21997 Make every semi-green pixel in the input transparent with some slight blending:
21999 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22003 @section convolution_opencl
22005 Apply convolution of 3x3, 5x5, 7x7 matrix.
22007 The filter accepts the following options:
22014 Set matrix for each plane.
22015 Matrix is sequence of 9, 25 or 49 signed numbers.
22016 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22022 Set multiplier for calculated value for each plane.
22023 If unset or 0, it will be sum of all matrix elements.
22024 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22030 Set bias for each plane. This value is added to the result of the multiplication.
22031 Useful for making the overall image brighter or darker.
22032 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22036 @subsection Examples
22042 -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
22048 -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
22052 Apply edge enhance:
22054 -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
22060 -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
22064 Apply laplacian edge detector which includes diagonals:
22066 -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
22072 -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
22076 @section erosion_opencl
22078 Apply erosion effect to the video.
22080 This filter replaces the pixel by the local(3x3) minimum.
22082 It accepts the following options:
22089 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22090 If @code{0}, plane will remain unchanged.
22093 Flag which specifies the pixel to refer to.
22094 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22096 Flags to local 3x3 coordinates region centered on @code{x}:
22105 @subsection Example
22109 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.
22111 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22115 @section deshake_opencl
22116 Feature-point based video stabilization filter.
22118 The filter accepts the following options:
22122 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22125 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22127 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22129 Viewing point matches in the output video is only supported for RGB input.
22131 Defaults to @code{0}.
22133 @item adaptive_crop
22134 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22136 Defaults to @code{1}.
22138 @item refine_features
22139 Whether or not feature points should be refined at a sub-pixel level.
22141 This can be turned off for a slight performance gain at the cost of precision.
22143 Defaults to @code{1}.
22145 @item smooth_strength
22146 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22148 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22150 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22152 Defaults to @code{0.0}.
22154 @item smooth_window_multiplier
22155 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22157 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22159 Acceptable values range from @code{0.1} to @code{10.0}.
22161 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22162 potentially improving smoothness, but also increase latency and memory usage.
22164 Defaults to @code{2.0}.
22168 @subsection Examples
22172 Stabilize a video with a fixed, medium smoothing strength:
22174 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22178 Stabilize a video with debugging (both in console and in rendered video):
22180 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22184 @section dilation_opencl
22186 Apply dilation effect to the video.
22188 This filter replaces the pixel by the local(3x3) maximum.
22190 It accepts the following options:
22197 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22198 If @code{0}, plane will remain unchanged.
22201 Flag which specifies the pixel to refer to.
22202 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22204 Flags to local 3x3 coordinates region centered on @code{x}:
22213 @subsection Example
22217 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.
22219 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22223 @section nlmeans_opencl
22225 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22227 @section overlay_opencl
22229 Overlay one video on top of another.
22231 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22232 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22234 The filter accepts the following options:
22239 Set the x coordinate of the overlaid video on the main video.
22240 Default value is @code{0}.
22243 Set the y coordinate of the overlaid video on the main video.
22244 Default value is @code{0}.
22248 @subsection Examples
22252 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22254 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22257 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22259 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22264 @section pad_opencl
22266 Add paddings to the input image, and place the original input at the
22267 provided @var{x}, @var{y} coordinates.
22269 It accepts the following options:
22274 Specify an expression for the size of the output image with the
22275 paddings added. If the value for @var{width} or @var{height} is 0, the
22276 corresponding input size is used for the output.
22278 The @var{width} expression can reference the value set by the
22279 @var{height} expression, and vice versa.
22281 The default value of @var{width} and @var{height} is 0.
22285 Specify the offsets to place the input image at within the padded area,
22286 with respect to the top/left border of the output image.
22288 The @var{x} expression can reference the value set by the @var{y}
22289 expression, and vice versa.
22291 The default value of @var{x} and @var{y} is 0.
22293 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22294 so the input image is centered on the padded area.
22297 Specify the color of the padded area. For the syntax of this option,
22298 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22299 manual,ffmpeg-utils}.
22302 Pad to an aspect instead to a resolution.
22305 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22306 options are expressions containing the following constants:
22311 The input video width and height.
22315 These are the same as @var{in_w} and @var{in_h}.
22319 The output width and height (the size of the padded area), as
22320 specified by the @var{width} and @var{height} expressions.
22324 These are the same as @var{out_w} and @var{out_h}.
22328 The x and y offsets as specified by the @var{x} and @var{y}
22329 expressions, or NAN if not yet specified.
22332 same as @var{iw} / @var{ih}
22335 input sample aspect ratio
22338 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22341 @section prewitt_opencl
22343 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22345 The filter accepts the following option:
22349 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22352 Set value which will be multiplied with filtered result.
22353 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22356 Set value which will be added to filtered result.
22357 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22360 @subsection Example
22364 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22366 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22370 @anchor{program_opencl}
22371 @section program_opencl
22373 Filter video using an OpenCL program.
22378 OpenCL program source file.
22381 Kernel name in program.
22384 Number of inputs to the filter. Defaults to 1.
22387 Size of output frames. Defaults to the same as the first input.
22391 The @code{program_opencl} filter also supports the @ref{framesync} options.
22393 The program source file must contain a kernel function with the given name,
22394 which will be run once for each plane of the output. Each run on a plane
22395 gets enqueued as a separate 2D global NDRange with one work-item for each
22396 pixel to be generated. The global ID offset for each work-item is therefore
22397 the coordinates of a pixel in the destination image.
22399 The kernel function needs to take the following arguments:
22402 Destination image, @var{__write_only image2d_t}.
22404 This image will become the output; the kernel should write all of it.
22406 Frame index, @var{unsigned int}.
22408 This is a counter starting from zero and increasing by one for each frame.
22410 Source images, @var{__read_only image2d_t}.
22412 These are the most recent images on each input. The kernel may read from
22413 them to generate the output, but they can't be written to.
22420 Copy the input to the output (output must be the same size as the input).
22422 __kernel void copy(__write_only image2d_t destination,
22423 unsigned int index,
22424 __read_only image2d_t source)
22426 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22428 int2 location = (int2)(get_global_id(0), get_global_id(1));
22430 float4 value = read_imagef(source, sampler, location);
22432 write_imagef(destination, location, value);
22437 Apply a simple transformation, rotating the input by an amount increasing
22438 with the index counter. Pixel values are linearly interpolated by the
22439 sampler, and the output need not have the same dimensions as the input.
22441 __kernel void rotate_image(__write_only image2d_t dst,
22442 unsigned int index,
22443 __read_only image2d_t src)
22445 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22446 CLK_FILTER_LINEAR);
22448 float angle = (float)index / 100.0f;
22450 float2 dst_dim = convert_float2(get_image_dim(dst));
22451 float2 src_dim = convert_float2(get_image_dim(src));
22453 float2 dst_cen = dst_dim / 2.0f;
22454 float2 src_cen = src_dim / 2.0f;
22456 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22458 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22460 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22461 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22463 src_pos = src_pos * src_dim / dst_dim;
22465 float2 src_loc = src_pos + src_cen;
22467 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22468 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22469 write_imagef(dst, dst_loc, 0.5f);
22471 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22476 Blend two inputs together, with the amount of each input used varying
22477 with the index counter.
22479 __kernel void blend_images(__write_only image2d_t dst,
22480 unsigned int index,
22481 __read_only image2d_t src1,
22482 __read_only image2d_t src2)
22484 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22485 CLK_FILTER_LINEAR);
22487 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22489 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22490 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22491 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22493 float4 val1 = read_imagef(src1, sampler, src1_loc);
22494 float4 val2 = read_imagef(src2, sampler, src2_loc);
22496 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22502 @section roberts_opencl
22503 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22505 The filter accepts the following option:
22509 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22512 Set value which will be multiplied with filtered result.
22513 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22516 Set value which will be added to filtered result.
22517 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22520 @subsection Example
22524 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22526 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22530 @section sobel_opencl
22532 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22534 The filter accepts the following option:
22538 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22541 Set value which will be multiplied with filtered result.
22542 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22545 Set value which will be added to filtered result.
22546 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22549 @subsection Example
22553 Apply sobel operator with scale set to 2 and delta set to 10
22555 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22559 @section tonemap_opencl
22561 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22563 It accepts the following parameters:
22567 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22570 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22573 Apply desaturation for highlights that exceed this level of brightness. The
22574 higher the parameter, the more color information will be preserved. This
22575 setting helps prevent unnaturally blown-out colors for super-highlights, by
22576 (smoothly) turning into white instead. This makes images feel more natural,
22577 at the cost of reducing information about out-of-range colors.
22579 The default value is 0.5, and the algorithm here is a little different from
22580 the cpu version tonemap currently. A setting of 0.0 disables this option.
22583 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22584 is used to detect whether the scene has changed or not. If the distance between
22585 the current frame average brightness and the current running average exceeds
22586 a threshold value, we would re-calculate scene average and peak brightness.
22587 The default value is 0.2.
22590 Specify the output pixel format.
22592 Currently supported formats are:
22599 Set the output color range.
22601 Possible values are:
22607 Default is same as input.
22610 Set the output color primaries.
22612 Possible values are:
22618 Default is same as input.
22621 Set the output transfer characteristics.
22623 Possible values are:
22632 Set the output colorspace matrix.
22634 Possible value are:
22640 Default is same as input.
22644 @subsection Example
22648 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22650 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22654 @section unsharp_opencl
22656 Sharpen or blur the input video.
22658 It accepts the following parameters:
22661 @item luma_msize_x, lx
22662 Set the luma matrix horizontal size.
22663 Range is @code{[1, 23]} and default value is @code{5}.
22665 @item luma_msize_y, ly
22666 Set the luma matrix vertical size.
22667 Range is @code{[1, 23]} and default value is @code{5}.
22669 @item luma_amount, la
22670 Set the luma effect strength.
22671 Range is @code{[-10, 10]} and default value is @code{1.0}.
22673 Negative values will blur the input video, while positive values will
22674 sharpen it, a value of zero will disable the effect.
22676 @item chroma_msize_x, cx
22677 Set the chroma matrix horizontal size.
22678 Range is @code{[1, 23]} and default value is @code{5}.
22680 @item chroma_msize_y, cy
22681 Set the chroma matrix vertical size.
22682 Range is @code{[1, 23]} and default value is @code{5}.
22684 @item chroma_amount, ca
22685 Set the chroma effect strength.
22686 Range is @code{[-10, 10]} and default value is @code{0.0}.
22688 Negative values will blur the input video, while positive values will
22689 sharpen it, a value of zero will disable the effect.
22693 All parameters are optional and default to the equivalent of the
22694 string '5:5:1.0:5:5:0.0'.
22696 @subsection Examples
22700 Apply strong luma sharpen effect:
22702 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22706 Apply a strong blur of both luma and chroma parameters:
22708 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22712 @section xfade_opencl
22714 Cross fade two videos with custom transition effect by using OpenCL.
22716 It accepts the following options:
22720 Set one of possible transition effects.
22724 Select custom transition effect, the actual transition description
22725 will be picked from source and kernel options.
22737 Default transition is fade.
22741 OpenCL program source file for custom transition.
22744 Set name of kernel to use for custom transition from program source file.
22747 Set duration of video transition.
22750 Set time of start of transition relative to first video.
22753 The program source file must contain a kernel function with the given name,
22754 which will be run once for each plane of the output. Each run on a plane
22755 gets enqueued as a separate 2D global NDRange with one work-item for each
22756 pixel to be generated. The global ID offset for each work-item is therefore
22757 the coordinates of a pixel in the destination image.
22759 The kernel function needs to take the following arguments:
22762 Destination image, @var{__write_only image2d_t}.
22764 This image will become the output; the kernel should write all of it.
22767 First Source image, @var{__read_only image2d_t}.
22768 Second Source image, @var{__read_only image2d_t}.
22770 These are the most recent images on each input. The kernel may read from
22771 them to generate the output, but they can't be written to.
22774 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22781 Apply dots curtain transition effect:
22783 __kernel void blend_images(__write_only image2d_t dst,
22784 __read_only image2d_t src1,
22785 __read_only image2d_t src2,
22788 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22789 CLK_FILTER_LINEAR);
22790 int2 p = (int2)(get_global_id(0), get_global_id(1));
22791 float2 rp = (float2)(get_global_id(0), get_global_id(1));
22792 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
22795 float2 dots = (float2)(20.0, 20.0);
22796 float2 center = (float2)(0,0);
22799 float4 val1 = read_imagef(src1, sampler, p);
22800 float4 val2 = read_imagef(src2, sampler, p);
22801 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
22803 write_imagef(dst, p, next ? val1 : val2);
22809 @c man end OPENCL VIDEO FILTERS
22811 @chapter VAAPI Video Filters
22812 @c man begin VAAPI VIDEO FILTERS
22814 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
22816 To enable compilation of these filters you need to configure FFmpeg with
22817 @code{--enable-vaapi}.
22819 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}
22821 @section tonemap_vaapi
22823 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
22824 It maps the dynamic range of HDR10 content to the SDR content.
22825 It currently only accepts HDR10 as input.
22827 It accepts the following parameters:
22831 Specify the output pixel format.
22833 Currently supported formats are:
22842 Set the output color primaries.
22844 Default is same as input.
22847 Set the output transfer characteristics.
22852 Set the output colorspace matrix.
22854 Default is same as input.
22858 @subsection Example
22862 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
22864 tonemap_vaapi=format=p010:t=bt2020-10
22868 @c man end VAAPI VIDEO FILTERS
22870 @chapter Video Sources
22871 @c man begin VIDEO SOURCES
22873 Below is a description of the currently available video sources.
22877 Buffer video frames, and make them available to the filter chain.
22879 This source is mainly intended for a programmatic use, in particular
22880 through the interface defined in @file{libavfilter/buffersrc.h}.
22882 It accepts the following parameters:
22887 Specify the size (width and height) of the buffered video frames. For the
22888 syntax of this option, check the
22889 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22892 The input video width.
22895 The input video height.
22898 A string representing the pixel format of the buffered video frames.
22899 It may be a number corresponding to a pixel format, or a pixel format
22903 Specify the timebase assumed by the timestamps of the buffered frames.
22906 Specify the frame rate expected for the video stream.
22908 @item pixel_aspect, sar
22909 The sample (pixel) aspect ratio of the input video.
22912 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
22913 to the filtergraph description to specify swscale flags for automatically
22914 inserted scalers. See @ref{Filtergraph syntax}.
22916 @item hw_frames_ctx
22917 When using a hardware pixel format, this should be a reference to an
22918 AVHWFramesContext describing input frames.
22923 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
22926 will instruct the source to accept video frames with size 320x240 and
22927 with format "yuv410p", assuming 1/24 as the timestamps timebase and
22928 square pixels (1:1 sample aspect ratio).
22929 Since the pixel format with name "yuv410p" corresponds to the number 6
22930 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
22931 this example corresponds to:
22933 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
22936 Alternatively, the options can be specified as a flat string, but this
22937 syntax is deprecated:
22939 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
22943 Create a pattern generated by an elementary cellular automaton.
22945 The initial state of the cellular automaton can be defined through the
22946 @option{filename} and @option{pattern} options. If such options are
22947 not specified an initial state is created randomly.
22949 At each new frame a new row in the video is filled with the result of
22950 the cellular automaton next generation. The behavior when the whole
22951 frame is filled is defined by the @option{scroll} option.
22953 This source accepts the following options:
22957 Read the initial cellular automaton state, i.e. the starting row, from
22958 the specified file.
22959 In the file, each non-whitespace character is considered an alive
22960 cell, a newline will terminate the row, and further characters in the
22961 file will be ignored.
22964 Read the initial cellular automaton state, i.e. the starting row, from
22965 the specified string.
22967 Each non-whitespace character in the string is considered an alive
22968 cell, a newline will terminate the row, and further characters in the
22969 string will be ignored.
22972 Set the video rate, that is the number of frames generated per second.
22975 @item random_fill_ratio, ratio
22976 Set the random fill ratio for the initial cellular automaton row. It
22977 is a floating point number value ranging from 0 to 1, defaults to
22980 This option is ignored when a file or a pattern is specified.
22982 @item random_seed, seed
22983 Set the seed for filling randomly the initial row, must be an integer
22984 included between 0 and UINT32_MAX. If not specified, or if explicitly
22985 set to -1, the filter will try to use a good random seed on a best
22989 Set the cellular automaton rule, it is a number ranging from 0 to 255.
22990 Default value is 110.
22993 Set the size of the output video. For the syntax of this option, check the
22994 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22996 If @option{filename} or @option{pattern} is specified, the size is set
22997 by default to the width of the specified initial state row, and the
22998 height is set to @var{width} * PHI.
23000 If @option{size} is set, it must contain the width of the specified
23001 pattern string, and the specified pattern will be centered in the
23004 If a filename or a pattern string is not specified, the size value
23005 defaults to "320x518" (used for a randomly generated initial state).
23008 If set to 1, scroll the output upward when all the rows in the output
23009 have been already filled. If set to 0, the new generated row will be
23010 written over the top row just after the bottom row is filled.
23013 @item start_full, full
23014 If set to 1, completely fill the output with generated rows before
23015 outputting the first frame.
23016 This is the default behavior, for disabling set the value to 0.
23019 If set to 1, stitch the left and right row edges together.
23020 This is the default behavior, for disabling set the value to 0.
23023 @subsection Examples
23027 Read the initial state from @file{pattern}, and specify an output of
23030 cellauto=f=pattern:s=200x400
23034 Generate a random initial row with a width of 200 cells, with a fill
23037 cellauto=ratio=2/3:s=200x200
23041 Create a pattern generated by rule 18 starting by a single alive cell
23042 centered on an initial row with width 100:
23044 cellauto=p=@@:s=100x400:full=0:rule=18
23048 Specify a more elaborated initial pattern:
23050 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23055 @anchor{coreimagesrc}
23056 @section coreimagesrc
23057 Video source generated on GPU using Apple's CoreImage API on OSX.
23059 This video source is a specialized version of the @ref{coreimage} video filter.
23060 Use a core image generator at the beginning of the applied filterchain to
23061 generate the content.
23063 The coreimagesrc video source accepts the following options:
23065 @item list_generators
23066 List all available generators along with all their respective options as well as
23067 possible minimum and maximum values along with the default values.
23069 list_generators=true
23073 Specify the size of the sourced video. For the syntax of this option, check the
23074 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23075 The default value is @code{320x240}.
23078 Specify the frame rate of the sourced video, as the number of frames
23079 generated per second. It has to be a string in the format
23080 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23081 number or a valid video frame rate abbreviation. The default value is
23085 Set the sample aspect ratio of the sourced video.
23088 Set the duration of the sourced video. See
23089 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23090 for the accepted syntax.
23092 If not specified, or the expressed duration is negative, the video is
23093 supposed to be generated forever.
23096 Additionally, all options of the @ref{coreimage} video filter are accepted.
23097 A complete filterchain can be used for further processing of the
23098 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23099 and examples for details.
23101 @subsection Examples
23106 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23107 given as complete and escaped command-line for Apple's standard bash shell:
23109 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23111 This example is equivalent to the QRCode example of @ref{coreimage} without the
23112 need for a nullsrc video source.
23117 Generate several gradients.
23121 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23122 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23125 Set frame rate, expressed as number of frames per second. Default
23128 @item c0, c1, c2, c3, c4, c5, c6, c7
23129 Set 8 colors. Default values for colors is to pick random one.
23131 @item x0, y0, y0, y1
23132 Set gradient line source and destination points. If negative or out of range, random ones
23136 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23139 Set seed for picking gradient line points.
23142 Set the duration of the sourced video. See
23143 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23144 for the accepted syntax.
23146 If not specified, or the expressed duration is negative, the video is
23147 supposed to be generated forever.
23150 Set speed of gradients rotation.
23154 @section mandelbrot
23156 Generate a Mandelbrot set fractal, and progressively zoom towards the
23157 point specified with @var{start_x} and @var{start_y}.
23159 This source accepts the following options:
23164 Set the terminal pts value. Default value is 400.
23167 Set the terminal scale value.
23168 Must be a floating point value. Default value is 0.3.
23171 Set the inner coloring mode, that is the algorithm used to draw the
23172 Mandelbrot fractal internal region.
23174 It shall assume one of the following values:
23179 Show time until convergence.
23181 Set color based on point closest to the origin of the iterations.
23186 Default value is @var{mincol}.
23189 Set the bailout value. Default value is 10.0.
23192 Set the maximum of iterations performed by the rendering
23193 algorithm. Default value is 7189.
23196 Set outer coloring mode.
23197 It shall assume one of following values:
23199 @item iteration_count
23200 Set iteration count mode.
23201 @item normalized_iteration_count
23202 set normalized iteration count mode.
23204 Default value is @var{normalized_iteration_count}.
23207 Set frame rate, expressed as number of frames per second. Default
23211 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23212 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23215 Set the initial scale value. Default value is 3.0.
23218 Set the initial x position. Must be a floating point value between
23219 -100 and 100. Default value is -0.743643887037158704752191506114774.
23222 Set the initial y position. Must be a floating point value between
23223 -100 and 100. Default value is -0.131825904205311970493132056385139.
23228 Generate various test patterns, as generated by the MPlayer test filter.
23230 The size of the generated video is fixed, and is 256x256.
23231 This source is useful in particular for testing encoding features.
23233 This source accepts the following options:
23238 Specify the frame rate of the sourced video, as the number of frames
23239 generated per second. It has to be a string in the format
23240 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23241 number or a valid video frame rate abbreviation. The default value is
23245 Set the duration of the sourced video. See
23246 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23247 for the accepted syntax.
23249 If not specified, or the expressed duration is negative, the video is
23250 supposed to be generated forever.
23254 Set the number or the name of the test to perform. Supported tests are:
23268 @item max_frames, m
23269 Set the maximum number of frames generated for each test, default value is 30.
23273 Default value is "all", which will cycle through the list of all tests.
23278 mptestsrc=t=dc_luma
23281 will generate a "dc_luma" test pattern.
23283 @section frei0r_src
23285 Provide a frei0r source.
23287 To enable compilation of this filter you need to install the frei0r
23288 header and configure FFmpeg with @code{--enable-frei0r}.
23290 This source accepts the following parameters:
23295 The size of the video to generate. For the syntax of this option, check the
23296 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23299 The framerate of the generated video. It may be a string of the form
23300 @var{num}/@var{den} or a frame rate abbreviation.
23303 The name to the frei0r source to load. For more information regarding frei0r and
23304 how to set the parameters, read the @ref{frei0r} section in the video filters
23307 @item filter_params
23308 A '|'-separated list of parameters to pass to the frei0r source.
23312 For example, to generate a frei0r partik0l source with size 200x200
23313 and frame rate 10 which is overlaid on the overlay filter main input:
23315 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23320 Generate a life pattern.
23322 This source is based on a generalization of John Conway's life game.
23324 The sourced input represents a life grid, each pixel represents a cell
23325 which can be in one of two possible states, alive or dead. Every cell
23326 interacts with its eight neighbours, which are the cells that are
23327 horizontally, vertically, or diagonally adjacent.
23329 At each interaction the grid evolves according to the adopted rule,
23330 which specifies the number of neighbor alive cells which will make a
23331 cell stay alive or born. The @option{rule} option allows one to specify
23334 This source accepts the following options:
23338 Set the file from which to read the initial grid state. In the file,
23339 each non-whitespace character is considered an alive cell, and newline
23340 is used to delimit the end of each row.
23342 If this option is not specified, the initial grid is generated
23346 Set the video rate, that is the number of frames generated per second.
23349 @item random_fill_ratio, ratio
23350 Set the random fill ratio for the initial random grid. It is a
23351 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23352 It is ignored when a file is specified.
23354 @item random_seed, seed
23355 Set the seed for filling the initial random grid, must be an integer
23356 included between 0 and UINT32_MAX. If not specified, or if explicitly
23357 set to -1, the filter will try to use a good random seed on a best
23363 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23364 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23365 @var{NS} specifies the number of alive neighbor cells which make a
23366 live cell stay alive, and @var{NB} the number of alive neighbor cells
23367 which make a dead cell to become alive (i.e. to "born").
23368 "s" and "b" can be used in place of "S" and "B", respectively.
23370 Alternatively a rule can be specified by an 18-bits integer. The 9
23371 high order bits are used to encode the next cell state if it is alive
23372 for each number of neighbor alive cells, the low order bits specify
23373 the rule for "borning" new cells. Higher order bits encode for an
23374 higher number of neighbor cells.
23375 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23376 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23378 Default value is "S23/B3", which is the original Conway's game of life
23379 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23380 cells, and will born a new cell if there are three alive cells around
23384 Set the size of the output video. For the syntax of this option, check the
23385 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23387 If @option{filename} is specified, the size is set by default to the
23388 same size of the input file. If @option{size} is set, it must contain
23389 the size specified in the input file, and the initial grid defined in
23390 that file is centered in the larger resulting area.
23392 If a filename is not specified, the size value defaults to "320x240"
23393 (used for a randomly generated initial grid).
23396 If set to 1, stitch the left and right grid edges together, and the
23397 top and bottom edges also. Defaults to 1.
23400 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23401 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23402 value from 0 to 255.
23405 Set the color of living (or new born) cells.
23408 Set the color of dead cells. If @option{mold} is set, this is the first color
23409 used to represent a dead cell.
23412 Set mold color, for definitely dead and moldy cells.
23414 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23415 ffmpeg-utils manual,ffmpeg-utils}.
23418 @subsection Examples
23422 Read a grid from @file{pattern}, and center it on a grid of size
23425 life=f=pattern:s=300x300
23429 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23431 life=ratio=2/3:s=200x200
23435 Specify a custom rule for evolving a randomly generated grid:
23441 Full example with slow death effect (mold) using @command{ffplay}:
23443 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23450 @anchor{haldclutsrc}
23453 @anchor{pal100bars}
23454 @anchor{rgbtestsrc}
23456 @anchor{smptehdbars}
23459 @anchor{yuvtestsrc}
23460 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23462 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23464 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23466 The @code{color} source provides an uniformly colored input.
23468 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23469 @ref{haldclut} filter.
23471 The @code{nullsrc} source returns unprocessed video frames. It is
23472 mainly useful to be employed in analysis / debugging tools, or as the
23473 source for filters which ignore the input data.
23475 The @code{pal75bars} source generates a color bars pattern, based on
23476 EBU PAL recommendations with 75% color levels.
23478 The @code{pal100bars} source generates a color bars pattern, based on
23479 EBU PAL recommendations with 100% color levels.
23481 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23482 detecting RGB vs BGR issues. You should see a red, green and blue
23483 stripe from top to bottom.
23485 The @code{smptebars} source generates a color bars pattern, based on
23486 the SMPTE Engineering Guideline EG 1-1990.
23488 The @code{smptehdbars} source generates a color bars pattern, based on
23489 the SMPTE RP 219-2002.
23491 The @code{testsrc} source generates a test video pattern, showing a
23492 color pattern, a scrolling gradient and a timestamp. This is mainly
23493 intended for testing purposes.
23495 The @code{testsrc2} source is similar to testsrc, but supports more
23496 pixel formats instead of just @code{rgb24}. This allows using it as an
23497 input for other tests without requiring a format conversion.
23499 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23500 see a y, cb and cr stripe from top to bottom.
23502 The sources accept the following parameters:
23507 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23508 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23509 pixels to be used as identity matrix for 3D lookup tables. Each component is
23510 coded on a @code{1/(N*N)} scale.
23513 Specify the color of the source, only available in the @code{color}
23514 source. For the syntax of this option, check the
23515 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23518 Specify the size of the sourced video. For the syntax of this option, check the
23519 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23520 The default value is @code{320x240}.
23522 This option is not available with the @code{allrgb}, @code{allyuv}, and
23523 @code{haldclutsrc} filters.
23526 Specify the frame rate of the sourced video, as the number of frames
23527 generated per second. It has to be a string in the format
23528 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23529 number or a valid video frame rate abbreviation. The default value is
23533 Set the duration of the sourced video. See
23534 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23535 for the accepted syntax.
23537 If not specified, or the expressed duration is negative, the video is
23538 supposed to be generated forever.
23540 Since the frame rate is used as time base, all frames including the last one
23541 will have their full duration. If the specified duration is not a multiple
23542 of the frame duration, it will be rounded up.
23545 Set the sample aspect ratio of the sourced video.
23548 Specify the alpha (opacity) of the background, only available in the
23549 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23550 255 (fully opaque, the default).
23553 Set the number of decimals to show in the timestamp, only available in the
23554 @code{testsrc} source.
23556 The displayed timestamp value will correspond to the original
23557 timestamp value multiplied by the power of 10 of the specified
23558 value. Default value is 0.
23561 @subsection Examples
23565 Generate a video with a duration of 5.3 seconds, with size
23566 176x144 and a frame rate of 10 frames per second:
23568 testsrc=duration=5.3:size=qcif:rate=10
23572 The following graph description will generate a red source
23573 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23576 color=c=red@@0.2:s=qcif:r=10
23580 If the input content is to be ignored, @code{nullsrc} can be used. The
23581 following command generates noise in the luminance plane by employing
23582 the @code{geq} filter:
23584 nullsrc=s=256x256, geq=random(1)*255:128:128
23588 @subsection Commands
23590 The @code{color} source supports the following commands:
23594 Set the color of the created image. Accepts the same syntax of the
23595 corresponding @option{color} option.
23600 Generate video using an OpenCL program.
23605 OpenCL program source file.
23608 Kernel name in program.
23611 Size of frames to generate. This must be set.
23614 Pixel format to use for the generated frames. This must be set.
23617 Number of frames generated every second. Default value is '25'.
23621 For details of how the program loading works, see the @ref{program_opencl}
23628 Generate a colour ramp by setting pixel values from the position of the pixel
23629 in the output image. (Note that this will work with all pixel formats, but
23630 the generated output will not be the same.)
23632 __kernel void ramp(__write_only image2d_t dst,
23633 unsigned int index)
23635 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23638 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23640 write_imagef(dst, loc, val);
23645 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23647 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23648 unsigned int index)
23650 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23652 float4 value = 0.0f;
23653 int x = loc.x + index;
23654 int y = loc.y + index;
23655 while (x > 0 || y > 0) {
23656 if (x % 3 == 1 && y % 3 == 1) {
23664 write_imagef(dst, loc, value);
23670 @section sierpinski
23672 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23674 This source accepts the following options:
23678 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23679 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23682 Set frame rate, expressed as number of frames per second. Default
23686 Set seed which is used for random panning.
23689 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23692 Set fractal type, can be default @code{carpet} or @code{triangle}.
23695 @c man end VIDEO SOURCES
23697 @chapter Video Sinks
23698 @c man begin VIDEO SINKS
23700 Below is a description of the currently available video sinks.
23702 @section buffersink
23704 Buffer video frames, and make them available to the end of the filter
23707 This sink is mainly intended for programmatic use, in particular
23708 through the interface defined in @file{libavfilter/buffersink.h}
23709 or the options system.
23711 It accepts a pointer to an AVBufferSinkContext structure, which
23712 defines the incoming buffers' formats, to be passed as the opaque
23713 parameter to @code{avfilter_init_filter} for initialization.
23717 Null video sink: do absolutely nothing with the input video. It is
23718 mainly useful as a template and for use in analysis / debugging
23721 @c man end VIDEO SINKS
23723 @chapter Multimedia Filters
23724 @c man begin MULTIMEDIA FILTERS
23726 Below is a description of the currently available multimedia filters.
23730 Convert input audio to a video output, displaying the audio bit scope.
23732 The filter accepts the following options:
23736 Set frame rate, expressed as number of frames per second. Default
23740 Specify the video size for the output. For the syntax of this option, check the
23741 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23742 Default value is @code{1024x256}.
23745 Specify list of colors separated by space or by '|' which will be used to
23746 draw channels. Unrecognized or missing colors will be replaced
23750 @section adrawgraph
23751 Draw a graph using input audio metadata.
23753 See @ref{drawgraph}
23755 @section agraphmonitor
23757 See @ref{graphmonitor}.
23759 @section ahistogram
23761 Convert input audio to a video output, displaying the volume histogram.
23763 The filter accepts the following options:
23767 Specify how histogram is calculated.
23769 It accepts the following values:
23772 Use single histogram for all channels.
23774 Use separate histogram for each channel.
23776 Default is @code{single}.
23779 Set frame rate, expressed as number of frames per second. Default
23783 Specify the video size for the output. For the syntax of this option, check the
23784 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23785 Default value is @code{hd720}.
23790 It accepts the following values:
23801 reverse logarithmic
23803 Default is @code{log}.
23806 Set amplitude scale.
23808 It accepts the following values:
23815 Default is @code{log}.
23818 Set how much frames to accumulate in histogram.
23819 Default is 1. Setting this to -1 accumulates all frames.
23822 Set histogram ratio of window height.
23825 Set sonogram sliding.
23827 It accepts the following values:
23830 replace old rows with new ones.
23832 scroll from top to bottom.
23834 Default is @code{replace}.
23837 @section aphasemeter
23839 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
23840 representing mean phase of current audio frame. A video output can also be produced and is
23841 enabled by default. The audio is passed through as first output.
23843 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
23844 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
23845 and @code{1} means channels are in phase.
23847 The filter accepts the following options, all related to its video output:
23851 Set the output frame rate. Default value is @code{25}.
23854 Set the video size for the output. For the syntax of this option, check the
23855 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23856 Default value is @code{800x400}.
23861 Specify the red, green, blue contrast. Default values are @code{2},
23862 @code{7} and @code{1}.
23863 Allowed range is @code{[0, 255]}.
23866 Set color which will be used for drawing median phase. If color is
23867 @code{none} which is default, no median phase value will be drawn.
23870 Enable video output. Default is enabled.
23873 @subsection phasing detection
23875 The filter also detects out of phase and mono sequences in stereo streams.
23876 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
23878 The filter accepts the following options for this detection:
23882 Enable mono and out of phase detection. Default is disabled.
23885 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
23886 Allowed range is @code{[0, 1]}.
23889 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
23890 Allowed range is @code{[90, 180]}.
23893 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
23896 @subsection Examples
23900 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
23902 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
23906 @section avectorscope
23908 Convert input audio to a video output, representing the audio vector
23911 The filter is used to measure the difference between channels of stereo
23912 audio stream. A monaural signal, consisting of identical left and right
23913 signal, results in straight vertical line. Any stereo separation is visible
23914 as a deviation from this line, creating a Lissajous figure.
23915 If the straight (or deviation from it) but horizontal line appears this
23916 indicates that the left and right channels are out of phase.
23918 The filter accepts the following options:
23922 Set the vectorscope mode.
23924 Available values are:
23927 Lissajous rotated by 45 degrees.
23930 Same as above but not rotated.
23933 Shape resembling half of circle.
23936 Default value is @samp{lissajous}.
23939 Set the video size for the output. For the syntax of this option, check the
23940 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23941 Default value is @code{400x400}.
23944 Set the output frame rate. Default value is @code{25}.
23950 Specify the red, green, blue and alpha contrast. Default values are @code{40},
23951 @code{160}, @code{80} and @code{255}.
23952 Allowed range is @code{[0, 255]}.
23958 Specify the red, green, blue and alpha fade. Default values are @code{15},
23959 @code{10}, @code{5} and @code{5}.
23960 Allowed range is @code{[0, 255]}.
23963 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
23964 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
23967 Set the vectorscope drawing mode.
23969 Available values are:
23972 Draw dot for each sample.
23975 Draw line between previous and current sample.
23978 Default value is @samp{dot}.
23981 Specify amplitude scale of audio samples.
23983 Available values are:
23999 Swap left channel axis with right channel axis.
24009 Mirror only x axis.
24012 Mirror only y axis.
24020 @subsection Examples
24024 Complete example using @command{ffplay}:
24026 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24027 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24031 @section bench, abench
24033 Benchmark part of a filtergraph.
24035 The filter accepts the following options:
24039 Start or stop a timer.
24041 Available values are:
24044 Get the current time, set it as frame metadata (using the key
24045 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24048 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24049 the input frame metadata to get the time difference. Time difference, average,
24050 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24051 @code{min}) are then printed. The timestamps are expressed in seconds.
24055 @subsection Examples
24059 Benchmark @ref{selectivecolor} filter:
24061 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24067 Concatenate audio and video streams, joining them together one after the
24070 The filter works on segments of synchronized video and audio streams. All
24071 segments must have the same number of streams of each type, and that will
24072 also be the number of streams at output.
24074 The filter accepts the following options:
24079 Set the number of segments. Default is 2.
24082 Set the number of output video streams, that is also the number of video
24083 streams in each segment. Default is 1.
24086 Set the number of output audio streams, that is also the number of audio
24087 streams in each segment. Default is 0.
24090 Activate unsafe mode: do not fail if segments have a different format.
24094 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24095 @var{a} audio outputs.
24097 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24098 segment, in the same order as the outputs, then the inputs for the second
24101 Related streams do not always have exactly the same duration, for various
24102 reasons including codec frame size or sloppy authoring. For that reason,
24103 related synchronized streams (e.g. a video and its audio track) should be
24104 concatenated at once. The concat filter will use the duration of the longest
24105 stream in each segment (except the last one), and if necessary pad shorter
24106 audio streams with silence.
24108 For this filter to work correctly, all segments must start at timestamp 0.
24110 All corresponding streams must have the same parameters in all segments; the
24111 filtering system will automatically select a common pixel format for video
24112 streams, and a common sample format, sample rate and channel layout for
24113 audio streams, but other settings, such as resolution, must be converted
24114 explicitly by the user.
24116 Different frame rates are acceptable but will result in variable frame rate
24117 at output; be sure to configure the output file to handle it.
24119 @subsection Examples
24123 Concatenate an opening, an episode and an ending, all in bilingual version
24124 (video in stream 0, audio in streams 1 and 2):
24126 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24127 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24128 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24129 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24133 Concatenate two parts, handling audio and video separately, using the
24134 (a)movie sources, and adjusting the resolution:
24136 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24137 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24138 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24140 Note that a desync will happen at the stitch if the audio and video streams
24141 do not have exactly the same duration in the first file.
24145 @subsection Commands
24147 This filter supports the following commands:
24150 Close the current segment and step to the next one
24156 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24157 level. By default, it logs a message at a frequency of 10Hz with the
24158 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24159 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24161 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24162 sample format is double-precision floating point. The input stream will be converted to
24163 this specification, if needed. Users may need to insert aformat and/or aresample filters
24164 after this filter to obtain the original parameters.
24166 The filter also has a video output (see the @var{video} option) with a real
24167 time graph to observe the loudness evolution. The graphic contains the logged
24168 message mentioned above, so it is not printed anymore when this option is set,
24169 unless the verbose logging is set. The main graphing area contains the
24170 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24171 the momentary loudness (400 milliseconds), but can optionally be configured
24172 to instead display short-term loudness (see @var{gauge}).
24174 The green area marks a +/- 1LU target range around the target loudness
24175 (-23LUFS by default, unless modified through @var{target}).
24177 More information about the Loudness Recommendation EBU R128 on
24178 @url{http://tech.ebu.ch/loudness}.
24180 The filter accepts the following options:
24185 Activate the video output. The audio stream is passed unchanged whether this
24186 option is set or no. The video stream will be the first output stream if
24187 activated. Default is @code{0}.
24190 Set the video size. This option is for video only. For the syntax of this
24192 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24193 Default and minimum resolution is @code{640x480}.
24196 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24197 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24198 other integer value between this range is allowed.
24201 Set metadata injection. If set to @code{1}, the audio input will be segmented
24202 into 100ms output frames, each of them containing various loudness information
24203 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24205 Default is @code{0}.
24208 Force the frame logging level.
24210 Available values are:
24213 information logging level
24215 verbose logging level
24218 By default, the logging level is set to @var{info}. If the @option{video} or
24219 the @option{metadata} options are set, it switches to @var{verbose}.
24224 Available modes can be cumulated (the option is a @code{flag} type). Possible
24228 Disable any peak mode (default).
24230 Enable sample-peak mode.
24232 Simple peak mode looking for the higher sample value. It logs a message
24233 for sample-peak (identified by @code{SPK}).
24235 Enable true-peak mode.
24237 If enabled, the peak lookup is done on an over-sampled version of the input
24238 stream for better peak accuracy. It logs a message for true-peak.
24239 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24240 This mode requires a build with @code{libswresample}.
24244 Treat mono input files as "dual mono". If a mono file is intended for playback
24245 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24246 If set to @code{true}, this option will compensate for this effect.
24247 Multi-channel input files are not affected by this option.
24250 Set a specific pan law to be used for the measurement of dual mono files.
24251 This parameter is optional, and has a default value of -3.01dB.
24254 Set a specific target level (in LUFS) used as relative zero in the visualization.
24255 This parameter is optional and has a default value of -23LUFS as specified
24256 by EBU R128. However, material published online may prefer a level of -16LUFS
24257 (e.g. for use with podcasts or video platforms).
24260 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24261 @code{shortterm}. By default the momentary value will be used, but in certain
24262 scenarios it may be more useful to observe the short term value instead (e.g.
24266 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24267 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24268 video output, not the summary or continuous log output.
24271 @subsection Examples
24275 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24277 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24281 Run an analysis with @command{ffmpeg}:
24283 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24287 @section interleave, ainterleave
24289 Temporally interleave frames from several inputs.
24291 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24293 These filters read frames from several inputs and send the oldest
24294 queued frame to the output.
24296 Input streams must have well defined, monotonically increasing frame
24299 In order to submit one frame to output, these filters need to enqueue
24300 at least one frame for each input, so they cannot work in case one
24301 input is not yet terminated and will not receive incoming frames.
24303 For example consider the case when one input is a @code{select} filter
24304 which always drops input frames. The @code{interleave} filter will keep
24305 reading from that input, but it will never be able to send new frames
24306 to output until the input sends an end-of-stream signal.
24308 Also, depending on inputs synchronization, the filters will drop
24309 frames in case one input receives more frames than the other ones, and
24310 the queue is already filled.
24312 These filters accept the following options:
24316 Set the number of different inputs, it is 2 by default.
24319 How to determine the end-of-stream.
24323 The duration of the longest input. (default)
24326 The duration of the shortest input.
24329 The duration of the first input.
24334 @subsection Examples
24338 Interleave frames belonging to different streams using @command{ffmpeg}:
24340 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24344 Add flickering blur effect:
24346 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24350 @section metadata, ametadata
24352 Manipulate frame metadata.
24354 This filter accepts the following options:
24358 Set mode of operation of the filter.
24360 Can be one of the following:
24364 If both @code{value} and @code{key} is set, select frames
24365 which have such metadata. If only @code{key} is set, select
24366 every frame that has such key in metadata.
24369 Add new metadata @code{key} and @code{value}. If key is already available
24373 Modify value of already present key.
24376 If @code{value} is set, delete only keys that have such value.
24377 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24381 Print key and its value if metadata was found. If @code{key} is not set print all
24382 metadata values available in frame.
24386 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24389 Set metadata value which will be used. This option is mandatory for
24390 @code{modify} and @code{add} mode.
24393 Which function to use when comparing metadata value and @code{value}.
24395 Can be one of following:
24399 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24402 Values are interpreted as strings, returns true if metadata value starts with
24403 the @code{value} option string.
24406 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24409 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24412 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24415 Values are interpreted as floats, returns true if expression from option @code{expr}
24419 Values are interpreted as strings, returns true if metadata value ends with
24420 the @code{value} option string.
24424 Set expression which is used when @code{function} is set to @code{expr}.
24425 The expression is evaluated through the eval API and can contain the following
24430 Float representation of @code{value} from metadata key.
24433 Float representation of @code{value} as supplied by user in @code{value} option.
24437 If specified in @code{print} mode, output is written to the named file. Instead of
24438 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24439 for standard output. If @code{file} option is not set, output is written to the log
24440 with AV_LOG_INFO loglevel.
24443 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24447 @subsection Examples
24451 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24454 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24457 Print silencedetect output to file @file{metadata.txt}.
24459 silencedetect,ametadata=mode=print:file=metadata.txt
24462 Direct all metadata to a pipe with file descriptor 4.
24464 metadata=mode=print:file='pipe\:4'
24468 @section perms, aperms
24470 Set read/write permissions for the output frames.
24472 These filters are mainly aimed at developers to test direct path in the
24473 following filter in the filtergraph.
24475 The filters accept the following options:
24479 Select the permissions mode.
24481 It accepts the following values:
24484 Do nothing. This is the default.
24486 Set all the output frames read-only.
24488 Set all the output frames directly writable.
24490 Make the frame read-only if writable, and writable if read-only.
24492 Set each output frame read-only or writable randomly.
24496 Set the seed for the @var{random} mode, must be an integer included between
24497 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24498 @code{-1}, the filter will try to use a good random seed on a best effort
24502 Note: in case of auto-inserted filter between the permission filter and the
24503 following one, the permission might not be received as expected in that
24504 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24505 perms/aperms filter can avoid this problem.
24507 @section realtime, arealtime
24509 Slow down filtering to match real time approximately.
24511 These filters will pause the filtering for a variable amount of time to
24512 match the output rate with the input timestamps.
24513 They are similar to the @option{re} option to @code{ffmpeg}.
24515 They accept the following options:
24519 Time limit for the pauses. Any pause longer than that will be considered
24520 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24522 Speed factor for processing. The value must be a float larger than zero.
24523 Values larger than 1.0 will result in faster than realtime processing,
24524 smaller will slow processing down. The @var{limit} is automatically adapted
24525 accordingly. Default is 1.0.
24527 A processing speed faster than what is possible without these filters cannot
24532 @section select, aselect
24534 Select frames to pass in output.
24536 This filter accepts the following options:
24541 Set expression, which is evaluated for each input frame.
24543 If the expression is evaluated to zero, the frame is discarded.
24545 If the evaluation result is negative or NaN, the frame is sent to the
24546 first output; otherwise it is sent to the output with index
24547 @code{ceil(val)-1}, assuming that the input index starts from 0.
24549 For example a value of @code{1.2} corresponds to the output with index
24550 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24553 Set the number of outputs. The output to which to send the selected
24554 frame is based on the result of the evaluation. Default value is 1.
24557 The expression can contain the following constants:
24561 The (sequential) number of the filtered frame, starting from 0.
24564 The (sequential) number of the selected frame, starting from 0.
24566 @item prev_selected_n
24567 The sequential number of the last selected frame. It's NAN if undefined.
24570 The timebase of the input timestamps.
24573 The PTS (Presentation TimeStamp) of the filtered video frame,
24574 expressed in @var{TB} units. It's NAN if undefined.
24577 The PTS of the filtered video frame,
24578 expressed in seconds. It's NAN if undefined.
24581 The PTS of the previously filtered video frame. It's NAN if undefined.
24583 @item prev_selected_pts
24584 The PTS of the last previously filtered video frame. It's NAN if undefined.
24586 @item prev_selected_t
24587 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24590 The PTS of the first video frame in the video. It's NAN if undefined.
24593 The time of the first video frame in the video. It's NAN if undefined.
24595 @item pict_type @emph{(video only)}
24596 The type of the filtered frame. It can assume one of the following
24608 @item interlace_type @emph{(video only)}
24609 The frame interlace type. It can assume one of the following values:
24612 The frame is progressive (not interlaced).
24614 The frame is top-field-first.
24616 The frame is bottom-field-first.
24619 @item consumed_sample_n @emph{(audio only)}
24620 the number of selected samples before the current frame
24622 @item samples_n @emph{(audio only)}
24623 the number of samples in the current frame
24625 @item sample_rate @emph{(audio only)}
24626 the input sample rate
24629 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24632 the position in the file of the filtered frame, -1 if the information
24633 is not available (e.g. for synthetic video)
24635 @item scene @emph{(video only)}
24636 value between 0 and 1 to indicate a new scene; a low value reflects a low
24637 probability for the current frame to introduce a new scene, while a higher
24638 value means the current frame is more likely to be one (see the example below)
24640 @item concatdec_select
24641 The concat demuxer can select only part of a concat input file by setting an
24642 inpoint and an outpoint, but the output packets may not be entirely contained
24643 in the selected interval. By using this variable, it is possible to skip frames
24644 generated by the concat demuxer which are not exactly contained in the selected
24647 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24648 and the @var{lavf.concat.duration} packet metadata values which are also
24649 present in the decoded frames.
24651 The @var{concatdec_select} variable is -1 if the frame pts is at least
24652 start_time and either the duration metadata is missing or the frame pts is less
24653 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24656 That basically means that an input frame is selected if its pts is within the
24657 interval set by the concat demuxer.
24661 The default value of the select expression is "1".
24663 @subsection Examples
24667 Select all frames in input:
24672 The example above is the same as:
24684 Select only I-frames:
24686 select='eq(pict_type\,I)'
24690 Select one frame every 100:
24692 select='not(mod(n\,100))'
24696 Select only frames contained in the 10-20 time interval:
24698 select=between(t\,10\,20)
24702 Select only I-frames contained in the 10-20 time interval:
24704 select=between(t\,10\,20)*eq(pict_type\,I)
24708 Select frames with a minimum distance of 10 seconds:
24710 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24714 Use aselect to select only audio frames with samples number > 100:
24716 aselect='gt(samples_n\,100)'
24720 Create a mosaic of the first scenes:
24722 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24725 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24729 Send even and odd frames to separate outputs, and compose them:
24731 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24735 Select useful frames from an ffconcat file which is using inpoints and
24736 outpoints but where the source files are not intra frame only.
24738 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24742 @section sendcmd, asendcmd
24744 Send commands to filters in the filtergraph.
24746 These filters read commands to be sent to other filters in the
24749 @code{sendcmd} must be inserted between two video filters,
24750 @code{asendcmd} must be inserted between two audio filters, but apart
24751 from that they act the same way.
24753 The specification of commands can be provided in the filter arguments
24754 with the @var{commands} option, or in a file specified by the
24755 @var{filename} option.
24757 These filters accept the following options:
24760 Set the commands to be read and sent to the other filters.
24762 Set the filename of the commands to be read and sent to the other
24766 @subsection Commands syntax
24768 A commands description consists of a sequence of interval
24769 specifications, comprising a list of commands to be executed when a
24770 particular event related to that interval occurs. The occurring event
24771 is typically the current frame time entering or leaving a given time
24774 An interval is specified by the following syntax:
24776 @var{START}[-@var{END}] @var{COMMANDS};
24779 The time interval is specified by the @var{START} and @var{END} times.
24780 @var{END} is optional and defaults to the maximum time.
24782 The current frame time is considered within the specified interval if
24783 it is included in the interval [@var{START}, @var{END}), that is when
24784 the time is greater or equal to @var{START} and is lesser than
24787 @var{COMMANDS} consists of a sequence of one or more command
24788 specifications, separated by ",", relating to that interval. The
24789 syntax of a command specification is given by:
24791 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
24794 @var{FLAGS} is optional and specifies the type of events relating to
24795 the time interval which enable sending the specified command, and must
24796 be a non-null sequence of identifier flags separated by "+" or "|" and
24797 enclosed between "[" and "]".
24799 The following flags are recognized:
24802 The command is sent when the current frame timestamp enters the
24803 specified interval. In other words, the command is sent when the
24804 previous frame timestamp was not in the given interval, and the
24808 The command is sent when the current frame timestamp leaves the
24809 specified interval. In other words, the command is sent when the
24810 previous frame timestamp was in the given interval, and the
24814 The command @var{ARG} is interpreted as expression and result of
24815 expression is passed as @var{ARG}.
24817 The expression is evaluated through the eval API and can contain the following
24822 Original position in the file of the frame, or undefined if undefined
24823 for the current frame.
24826 The presentation timestamp in input.
24829 The count of the input frame for video or audio, starting from 0.
24832 The time in seconds of the current frame.
24835 The start time in seconds of the current command interval.
24838 The end time in seconds of the current command interval.
24841 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
24846 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
24849 @var{TARGET} specifies the target of the command, usually the name of
24850 the filter class or a specific filter instance name.
24852 @var{COMMAND} specifies the name of the command for the target filter.
24854 @var{ARG} is optional and specifies the optional list of argument for
24855 the given @var{COMMAND}.
24857 Between one interval specification and another, whitespaces, or
24858 sequences of characters starting with @code{#} until the end of line,
24859 are ignored and can be used to annotate comments.
24861 A simplified BNF description of the commands specification syntax
24864 @var{COMMAND_FLAG} ::= "enter" | "leave"
24865 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
24866 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
24867 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
24868 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
24869 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
24872 @subsection Examples
24876 Specify audio tempo change at second 4:
24878 asendcmd=c='4.0 atempo tempo 1.5',atempo
24882 Target a specific filter instance:
24884 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
24888 Specify a list of drawtext and hue commands in a file.
24890 # show text in the interval 5-10
24891 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
24892 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
24894 # desaturate the image in the interval 15-20
24895 15.0-20.0 [enter] hue s 0,
24896 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
24898 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
24900 # apply an exponential saturation fade-out effect, starting from time 25
24901 25 [enter] hue s exp(25-t)
24904 A filtergraph allowing to read and process the above command list
24905 stored in a file @file{test.cmd}, can be specified with:
24907 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
24912 @section setpts, asetpts
24914 Change the PTS (presentation timestamp) of the input frames.
24916 @code{setpts} works on video frames, @code{asetpts} on audio frames.
24918 This filter accepts the following options:
24923 The expression which is evaluated for each frame to construct its timestamp.
24927 The expression is evaluated through the eval API and can contain the following
24931 @item FRAME_RATE, FR
24932 frame rate, only defined for constant frame-rate video
24935 The presentation timestamp in input
24938 The count of the input frame for video or the number of consumed samples,
24939 not including the current frame for audio, starting from 0.
24941 @item NB_CONSUMED_SAMPLES
24942 The number of consumed samples, not including the current frame (only
24945 @item NB_SAMPLES, S
24946 The number of samples in the current frame (only audio)
24948 @item SAMPLE_RATE, SR
24949 The audio sample rate.
24952 The PTS of the first frame.
24955 the time in seconds of the first frame
24958 State whether the current frame is interlaced.
24961 the time in seconds of the current frame
24964 original position in the file of the frame, or undefined if undefined
24965 for the current frame
24968 The previous input PTS.
24971 previous input time in seconds
24974 The previous output PTS.
24977 previous output time in seconds
24980 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
24984 The wallclock (RTC) time at the start of the movie in microseconds.
24987 The timebase of the input timestamps.
24991 @subsection Examples
24995 Start counting PTS from zero
24997 setpts=PTS-STARTPTS
25001 Apply fast motion effect:
25007 Apply slow motion effect:
25013 Set fixed rate of 25 frames per second:
25019 Set fixed rate 25 fps with some jitter:
25021 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25025 Apply an offset of 10 seconds to the input PTS:
25031 Generate timestamps from a "live source" and rebase onto the current timebase:
25033 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25037 Generate timestamps by counting samples:
25046 Force color range for the output video frame.
25048 The @code{setrange} filter marks the color range property for the
25049 output frames. It does not change the input frame, but only sets the
25050 corresponding property, which affects how the frame is treated by
25053 The filter accepts the following options:
25058 Available values are:
25062 Keep the same color range property.
25064 @item unspecified, unknown
25065 Set the color range as unspecified.
25067 @item limited, tv, mpeg
25068 Set the color range as limited.
25070 @item full, pc, jpeg
25071 Set the color range as full.
25075 @section settb, asettb
25077 Set the timebase to use for the output frames timestamps.
25078 It is mainly useful for testing timebase configuration.
25080 It accepts the following parameters:
25085 The expression which is evaluated into the output timebase.
25089 The value for @option{tb} is an arithmetic expression representing a
25090 rational. The expression can contain the constants "AVTB" (the default
25091 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25092 audio only). Default value is "intb".
25094 @subsection Examples
25098 Set the timebase to 1/25:
25104 Set the timebase to 1/10:
25110 Set the timebase to 1001/1000:
25116 Set the timebase to 2*intb:
25122 Set the default timebase value:
25129 Convert input audio to a video output representing frequency spectrum
25130 logarithmically using Brown-Puckette constant Q transform algorithm with
25131 direct frequency domain coefficient calculation (but the transform itself
25132 is not really constant Q, instead the Q factor is actually variable/clamped),
25133 with musical tone scale, from E0 to D#10.
25135 The filter accepts the following options:
25139 Specify the video size for the output. It must be even. For the syntax of this option,
25140 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25141 Default value is @code{1920x1080}.
25144 Set the output frame rate. Default value is @code{25}.
25147 Set the bargraph height. It must be even. Default value is @code{-1} which
25148 computes the bargraph height automatically.
25151 Set the axis height. It must be even. Default value is @code{-1} which computes
25152 the axis height automatically.
25155 Set the sonogram height. It must be even. Default value is @code{-1} which
25156 computes the sonogram height automatically.
25159 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25160 instead. Default value is @code{1}.
25162 @item sono_v, volume
25163 Specify the sonogram volume expression. It can contain variables:
25166 the @var{bar_v} evaluated expression
25167 @item frequency, freq, f
25168 the frequency where it is evaluated
25169 @item timeclamp, tc
25170 the value of @var{timeclamp} option
25174 @item a_weighting(f)
25175 A-weighting of equal loudness
25176 @item b_weighting(f)
25177 B-weighting of equal loudness
25178 @item c_weighting(f)
25179 C-weighting of equal loudness.
25181 Default value is @code{16}.
25183 @item bar_v, volume2
25184 Specify the bargraph volume expression. It can contain variables:
25187 the @var{sono_v} evaluated expression
25188 @item frequency, freq, f
25189 the frequency where it is evaluated
25190 @item timeclamp, tc
25191 the value of @var{timeclamp} option
25195 @item a_weighting(f)
25196 A-weighting of equal loudness
25197 @item b_weighting(f)
25198 B-weighting of equal loudness
25199 @item c_weighting(f)
25200 C-weighting of equal loudness.
25202 Default value is @code{sono_v}.
25204 @item sono_g, gamma
25205 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25206 higher gamma makes the spectrum having more range. Default value is @code{3}.
25207 Acceptable range is @code{[1, 7]}.
25209 @item bar_g, gamma2
25210 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25214 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25215 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25217 @item timeclamp, tc
25218 Specify the transform timeclamp. At low frequency, there is trade-off between
25219 accuracy in time domain and frequency domain. If timeclamp is lower,
25220 event in time domain is represented more accurately (such as fast bass drum),
25221 otherwise event in frequency domain is represented more accurately
25222 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25225 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25226 limits future samples by applying asymmetric windowing in time domain, useful
25227 when low latency is required. Accepted range is @code{[0, 1]}.
25230 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25231 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25234 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25235 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25238 This option is deprecated and ignored.
25241 Specify the transform length in time domain. Use this option to control accuracy
25242 trade-off between time domain and frequency domain at every frequency sample.
25243 It can contain variables:
25245 @item frequency, freq, f
25246 the frequency where it is evaluated
25247 @item timeclamp, tc
25248 the value of @var{timeclamp} option.
25250 Default value is @code{384*tc/(384+tc*f)}.
25253 Specify the transform count for every video frame. Default value is @code{6}.
25254 Acceptable range is @code{[1, 30]}.
25257 Specify the transform count for every single pixel. Default value is @code{0},
25258 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25261 Specify font file for use with freetype to draw the axis. If not specified,
25262 use embedded font. Note that drawing with font file or embedded font is not
25263 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25267 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25268 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25272 Specify font color expression. This is arithmetic expression that should return
25273 integer value 0xRRGGBB. It can contain variables:
25275 @item frequency, freq, f
25276 the frequency where it is evaluated
25277 @item timeclamp, tc
25278 the value of @var{timeclamp} option
25283 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25284 @item r(x), g(x), b(x)
25285 red, green, and blue value of intensity x.
25287 Default value is @code{st(0, (midi(f)-59.5)/12);
25288 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25289 r(1-ld(1)) + b(ld(1))}.
25292 Specify image file to draw the axis. This option override @var{fontfile} and
25293 @var{fontcolor} option.
25296 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25297 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25298 Default value is @code{1}.
25301 Set colorspace. The accepted values are:
25304 Unspecified (default)
25313 BT.470BG or BT.601-6 625
25316 SMPTE-170M or BT.601-6 525
25322 BT.2020 with non-constant luminance
25327 Set spectrogram color scheme. This is list of floating point values with format
25328 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25329 The default is @code{1|0.5|0|0|0.5|1}.
25333 @subsection Examples
25337 Playing audio while showing the spectrum:
25339 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25343 Same as above, but with frame rate 30 fps:
25345 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25349 Playing at 1280x720:
25351 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25355 Disable sonogram display:
25361 A1 and its harmonics: A1, A2, (near)E3, A3:
25363 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),
25364 asplit[a][out1]; [a] showcqt [out0]'
25368 Same as above, but with more accuracy in frequency domain:
25370 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),
25371 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25377 bar_v=10:sono_v=bar_v*a_weighting(f)
25381 Custom gamma, now spectrum is linear to the amplitude.
25387 Custom tlength equation:
25389 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)))'
25393 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25395 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25399 Custom font using fontconfig:
25401 font='Courier New,Monospace,mono|bold'
25405 Custom frequency range with custom axis using image file:
25407 axisfile=myaxis.png:basefreq=40:endfreq=10000
25413 Convert input audio to video output representing the audio power spectrum.
25414 Audio amplitude is on Y-axis while frequency is on X-axis.
25416 The filter accepts the following options:
25420 Specify size of video. For the syntax of this option, check the
25421 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25422 Default is @code{1024x512}.
25426 This set how each frequency bin will be represented.
25428 It accepts the following values:
25434 Default is @code{bar}.
25437 Set amplitude scale.
25439 It accepts the following values:
25453 Default is @code{log}.
25456 Set frequency scale.
25458 It accepts the following values:
25467 Reverse logarithmic scale.
25469 Default is @code{lin}.
25472 Set window size. Allowed range is from 16 to 65536.
25474 Default is @code{2048}
25477 Set windowing function.
25479 It accepts the following values:
25502 Default is @code{hanning}.
25505 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25506 which means optimal overlap for selected window function will be picked.
25509 Set time averaging. Setting this to 0 will display current maximal peaks.
25510 Default is @code{1}, which means time averaging is disabled.
25513 Specify list of colors separated by space or by '|' which will be used to
25514 draw channel frequencies. Unrecognized or missing colors will be replaced
25518 Set channel display mode.
25520 It accepts the following values:
25525 Default is @code{combined}.
25528 Set minimum amplitude used in @code{log} amplitude scaler.
25531 Set data display mode.
25533 It accepts the following values:
25539 Default is @code{magnitude}.
25542 @section showspatial
25544 Convert stereo input audio to a video output, representing the spatial relationship
25545 between two channels.
25547 The filter accepts the following options:
25551 Specify the video size for the output. For the syntax of this option, check the
25552 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25553 Default value is @code{512x512}.
25556 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25559 Set window function.
25561 It accepts the following values:
25586 Default value is @code{hann}.
25589 Set ratio of overlap window. Default value is @code{0.5}.
25590 When value is @code{1} overlap is set to recommended size for specific
25591 window function currently used.
25594 @anchor{showspectrum}
25595 @section showspectrum
25597 Convert input audio to a video output, representing the audio frequency
25600 The filter accepts the following options:
25604 Specify the video size for the output. For the syntax of this option, check the
25605 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25606 Default value is @code{640x512}.
25609 Specify how the spectrum should slide along the window.
25611 It accepts the following values:
25614 the samples start again on the left when they reach the right
25616 the samples scroll from right to left
25618 frames are only produced when the samples reach the right
25620 the samples scroll from left to right
25623 Default value is @code{replace}.
25626 Specify display mode.
25628 It accepts the following values:
25631 all channels are displayed in the same row
25633 all channels are displayed in separate rows
25636 Default value is @samp{combined}.
25639 Specify display color mode.
25641 It accepts the following values:
25644 each channel is displayed in a separate color
25646 each channel is displayed using the same color scheme
25648 each channel is displayed using the rainbow color scheme
25650 each channel is displayed using the moreland color scheme
25652 each channel is displayed using the nebulae color scheme
25654 each channel is displayed using the fire color scheme
25656 each channel is displayed using the fiery color scheme
25658 each channel is displayed using the fruit color scheme
25660 each channel is displayed using the cool color scheme
25662 each channel is displayed using the magma color scheme
25664 each channel is displayed using the green color scheme
25666 each channel is displayed using the viridis color scheme
25668 each channel is displayed using the plasma color scheme
25670 each channel is displayed using the cividis color scheme
25672 each channel is displayed using the terrain color scheme
25675 Default value is @samp{channel}.
25678 Specify scale used for calculating intensity color values.
25680 It accepts the following values:
25685 square root, default
25696 Default value is @samp{sqrt}.
25699 Specify frequency scale.
25701 It accepts the following values:
25709 Default value is @samp{lin}.
25712 Set saturation modifier for displayed colors. Negative values provide
25713 alternative color scheme. @code{0} is no saturation at all.
25714 Saturation must be in [-10.0, 10.0] range.
25715 Default value is @code{1}.
25718 Set window function.
25720 It accepts the following values:
25745 Default value is @code{hann}.
25748 Set orientation of time vs frequency axis. Can be @code{vertical} or
25749 @code{horizontal}. Default is @code{vertical}.
25752 Set ratio of overlap window. Default value is @code{0}.
25753 When value is @code{1} overlap is set to recommended size for specific
25754 window function currently used.
25757 Set scale gain for calculating intensity color values.
25758 Default value is @code{1}.
25761 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25764 Set color rotation, must be in [-1.0, 1.0] range.
25765 Default value is @code{0}.
25768 Set start frequency from which to display spectrogram. Default is @code{0}.
25771 Set stop frequency to which to display spectrogram. Default is @code{0}.
25774 Set upper frame rate limit. Default is @code{auto}, unlimited.
25777 Draw time and frequency axes and legends. Default is disabled.
25780 The usage is very similar to the showwaves filter; see the examples in that
25783 @subsection Examples
25787 Large window with logarithmic color scaling:
25789 showspectrum=s=1280x480:scale=log
25793 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
25795 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
25796 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
25800 @section showspectrumpic
25802 Convert input audio to a single video frame, representing the audio frequency
25805 The filter accepts the following options:
25809 Specify the video size for the output. For the syntax of this option, check the
25810 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25811 Default value is @code{4096x2048}.
25814 Specify display mode.
25816 It accepts the following values:
25819 all channels are displayed in the same row
25821 all channels are displayed in separate rows
25823 Default value is @samp{combined}.
25826 Specify display color mode.
25828 It accepts the following values:
25831 each channel is displayed in a separate color
25833 each channel is displayed using the same color scheme
25835 each channel is displayed using the rainbow color scheme
25837 each channel is displayed using the moreland color scheme
25839 each channel is displayed using the nebulae color scheme
25841 each channel is displayed using the fire color scheme
25843 each channel is displayed using the fiery color scheme
25845 each channel is displayed using the fruit color scheme
25847 each channel is displayed using the cool color scheme
25849 each channel is displayed using the magma color scheme
25851 each channel is displayed using the green color scheme
25853 each channel is displayed using the viridis color scheme
25855 each channel is displayed using the plasma color scheme
25857 each channel is displayed using the cividis color scheme
25859 each channel is displayed using the terrain color scheme
25861 Default value is @samp{intensity}.
25864 Specify scale used for calculating intensity color values.
25866 It accepts the following values:
25871 square root, default
25881 Default value is @samp{log}.
25884 Specify frequency scale.
25886 It accepts the following values:
25894 Default value is @samp{lin}.
25897 Set saturation modifier for displayed colors. Negative values provide
25898 alternative color scheme. @code{0} is no saturation at all.
25899 Saturation must be in [-10.0, 10.0] range.
25900 Default value is @code{1}.
25903 Set window function.
25905 It accepts the following values:
25929 Default value is @code{hann}.
25932 Set orientation of time vs frequency axis. Can be @code{vertical} or
25933 @code{horizontal}. Default is @code{vertical}.
25936 Set scale gain for calculating intensity color values.
25937 Default value is @code{1}.
25940 Draw time and frequency axes and legends. Default is enabled.
25943 Set color rotation, must be in [-1.0, 1.0] range.
25944 Default value is @code{0}.
25947 Set start frequency from which to display spectrogram. Default is @code{0}.
25950 Set stop frequency to which to display spectrogram. Default is @code{0}.
25953 @subsection Examples
25957 Extract an audio spectrogram of a whole audio track
25958 in a 1024x1024 picture using @command{ffmpeg}:
25960 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
25964 @section showvolume
25966 Convert input audio volume to a video output.
25968 The filter accepts the following options:
25975 Set border width, allowed range is [0, 5]. Default is 1.
25978 Set channel width, allowed range is [80, 8192]. Default is 400.
25981 Set channel height, allowed range is [1, 900]. Default is 20.
25984 Set fade, allowed range is [0, 1]. Default is 0.95.
25987 Set volume color expression.
25989 The expression can use the following variables:
25993 Current max volume of channel in dB.
25999 Current channel number, starting from 0.
26003 If set, displays channel names. Default is enabled.
26006 If set, displays volume values. Default is enabled.
26009 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26010 default is @code{h}.
26013 Set step size, allowed range is [0, 5]. Default is 0, which means
26017 Set background opacity, allowed range is [0, 1]. Default is 0.
26020 Set metering mode, can be peak: @code{p} or rms: @code{r},
26021 default is @code{p}.
26024 Set display scale, can be linear: @code{lin} or log: @code{log},
26025 default is @code{lin}.
26029 If set to > 0., display a line for the max level
26030 in the previous seconds.
26031 default is disabled: @code{0.}
26034 The color of the max line. Use when @code{dm} option is set to > 0.
26035 default is: @code{orange}
26040 Convert input audio to a video output, representing the samples waves.
26042 The filter accepts the following options:
26046 Specify the video size for the output. For the syntax of this option, check the
26047 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26048 Default value is @code{600x240}.
26053 Available values are:
26056 Draw a point for each sample.
26059 Draw a vertical line for each sample.
26062 Draw a point for each sample and a line between them.
26065 Draw a centered vertical line for each sample.
26068 Default value is @code{point}.
26071 Set the number of samples which are printed on the same column. A
26072 larger value will decrease the frame rate. Must be a positive
26073 integer. This option can be set only if the value for @var{rate}
26074 is not explicitly specified.
26077 Set the (approximate) output frame rate. This is done by setting the
26078 option @var{n}. Default value is "25".
26080 @item split_channels
26081 Set if channels should be drawn separately or overlap. Default value is 0.
26084 Set colors separated by '|' which are going to be used for drawing of each channel.
26087 Set amplitude scale.
26089 Available values are:
26107 Set the draw mode. This is mostly useful to set for high @var{n}.
26109 Available values are:
26112 Scale pixel values for each drawn sample.
26115 Draw every sample directly.
26118 Default value is @code{scale}.
26121 @subsection Examples
26125 Output the input file audio and the corresponding video representation
26128 amovie=a.mp3,asplit[out0],showwaves[out1]
26132 Create a synthetic signal and show it with showwaves, forcing a
26133 frame rate of 30 frames per second:
26135 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26139 @section showwavespic
26141 Convert input audio to a single video frame, representing the samples waves.
26143 The filter accepts the following options:
26147 Specify the video size for the output. For the syntax of this option, check the
26148 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26149 Default value is @code{600x240}.
26151 @item split_channels
26152 Set if channels should be drawn separately or overlap. Default value is 0.
26155 Set colors separated by '|' which are going to be used for drawing of each channel.
26158 Set amplitude scale.
26160 Available values are:
26180 Available values are:
26183 Scale pixel values for each drawn sample.
26186 Draw every sample directly.
26189 Default value is @code{scale}.
26192 Set the filter mode.
26194 Available values are:
26197 Use average samples values for each drawn sample.
26200 Use peak samples values for each drawn sample.
26203 Default value is @code{average}.
26206 @subsection Examples
26210 Extract a channel split representation of the wave form of a whole audio track
26211 in a 1024x800 picture using @command{ffmpeg}:
26213 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26217 @section sidedata, asidedata
26219 Delete frame side data, or select frames based on it.
26221 This filter accepts the following options:
26225 Set mode of operation of the filter.
26227 Can be one of the following:
26231 Select every frame with side data of @code{type}.
26234 Delete side data of @code{type}. If @code{type} is not set, delete all side
26240 Set side data type used with all modes. Must be set for @code{select} mode. For
26241 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26242 in @file{libavutil/frame.h}. For example, to choose
26243 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26247 @section spectrumsynth
26249 Synthesize audio from 2 input video spectrums, first input stream represents
26250 magnitude across time and second represents phase across time.
26251 The filter will transform from frequency domain as displayed in videos back
26252 to time domain as presented in audio output.
26254 This filter is primarily created for reversing processed @ref{showspectrum}
26255 filter outputs, but can synthesize sound from other spectrograms too.
26256 But in such case results are going to be poor if the phase data is not
26257 available, because in such cases phase data need to be recreated, usually
26258 it's just recreated from random noise.
26259 For best results use gray only output (@code{channel} color mode in
26260 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26261 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26262 @code{data} option. Inputs videos should generally use @code{fullframe}
26263 slide mode as that saves resources needed for decoding video.
26265 The filter accepts the following options:
26269 Specify sample rate of output audio, the sample rate of audio from which
26270 spectrum was generated may differ.
26273 Set number of channels represented in input video spectrums.
26276 Set scale which was used when generating magnitude input spectrum.
26277 Can be @code{lin} or @code{log}. Default is @code{log}.
26280 Set slide which was used when generating inputs spectrums.
26281 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26282 Default is @code{fullframe}.
26285 Set window function used for resynthesis.
26288 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26289 which means optimal overlap for selected window function will be picked.
26292 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26293 Default is @code{vertical}.
26296 @subsection Examples
26300 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26301 then resynthesize videos back to audio with spectrumsynth:
26303 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
26304 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
26305 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26309 @section split, asplit
26311 Split input into several identical outputs.
26313 @code{asplit} works with audio input, @code{split} with video.
26315 The filter accepts a single parameter which specifies the number of outputs. If
26316 unspecified, it defaults to 2.
26318 @subsection Examples
26322 Create two separate outputs from the same input:
26324 [in] split [out0][out1]
26328 To create 3 or more outputs, you need to specify the number of
26331 [in] asplit=3 [out0][out1][out2]
26335 Create two separate outputs from the same input, one cropped and
26338 [in] split [splitout1][splitout2];
26339 [splitout1] crop=100:100:0:0 [cropout];
26340 [splitout2] pad=200:200:100:100 [padout];
26344 Create 5 copies of the input audio with @command{ffmpeg}:
26346 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26352 Receive commands sent through a libzmq client, and forward them to
26353 filters in the filtergraph.
26355 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26356 must be inserted between two video filters, @code{azmq} between two
26357 audio filters. Both are capable to send messages to any filter type.
26359 To enable these filters you need to install the libzmq library and
26360 headers and configure FFmpeg with @code{--enable-libzmq}.
26362 For more information about libzmq see:
26363 @url{http://www.zeromq.org/}
26365 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26366 receives messages sent through a network interface defined by the
26367 @option{bind_address} (or the abbreviation "@option{b}") option.
26368 Default value of this option is @file{tcp://localhost:5555}. You may
26369 want to alter this value to your needs, but do not forget to escape any
26370 ':' signs (see @ref{filtergraph escaping}).
26372 The received message must be in the form:
26374 @var{TARGET} @var{COMMAND} [@var{ARG}]
26377 @var{TARGET} specifies the target of the command, usually the name of
26378 the filter class or a specific filter instance name. The default
26379 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26380 but you can override this by using the @samp{filter_name@@id} syntax
26381 (see @ref{Filtergraph syntax}).
26383 @var{COMMAND} specifies the name of the command for the target filter.
26385 @var{ARG} is optional and specifies the optional argument list for the
26386 given @var{COMMAND}.
26388 Upon reception, the message is processed and the corresponding command
26389 is injected into the filtergraph. Depending on the result, the filter
26390 will send a reply to the client, adopting the format:
26392 @var{ERROR_CODE} @var{ERROR_REASON}
26396 @var{MESSAGE} is optional.
26398 @subsection Examples
26400 Look at @file{tools/zmqsend} for an example of a zmq client which can
26401 be used to send commands processed by these filters.
26403 Consider the following filtergraph generated by @command{ffplay}.
26404 In this example the last overlay filter has an instance name. All other
26405 filters will have default instance names.
26408 ffplay -dumpgraph 1 -f lavfi "
26409 color=s=100x100:c=red [l];
26410 color=s=100x100:c=blue [r];
26411 nullsrc=s=200x100, zmq [bg];
26412 [bg][l] overlay [bg+l];
26413 [bg+l][r] overlay@@my=x=100 "
26416 To change the color of the left side of the video, the following
26417 command can be used:
26419 echo Parsed_color_0 c yellow | tools/zmqsend
26422 To change the right side:
26424 echo Parsed_color_1 c pink | tools/zmqsend
26427 To change the position of the right side:
26429 echo overlay@@my x 150 | tools/zmqsend
26433 @c man end MULTIMEDIA FILTERS
26435 @chapter Multimedia Sources
26436 @c man begin MULTIMEDIA SOURCES
26438 Below is a description of the currently available multimedia sources.
26442 This is the same as @ref{movie} source, except it selects an audio
26448 Read audio and/or video stream(s) from a movie container.
26450 It accepts the following parameters:
26454 The name of the resource to read (not necessarily a file; it can also be a
26455 device or a stream accessed through some protocol).
26457 @item format_name, f
26458 Specifies the format assumed for the movie to read, and can be either
26459 the name of a container or an input device. If not specified, the
26460 format is guessed from @var{movie_name} or by probing.
26462 @item seek_point, sp
26463 Specifies the seek point in seconds. The frames will be output
26464 starting from this seek point. The parameter is evaluated with
26465 @code{av_strtod}, so the numerical value may be suffixed by an IS
26466 postfix. The default value is "0".
26469 Specifies the streams to read. Several streams can be specified,
26470 separated by "+". The source will then have as many outputs, in the
26471 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26472 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26473 respectively the default (best suited) video and audio stream. Default
26474 is "dv", or "da" if the filter is called as "amovie".
26476 @item stream_index, si
26477 Specifies the index of the video stream to read. If the value is -1,
26478 the most suitable video stream will be automatically selected. The default
26479 value is "-1". Deprecated. If the filter is called "amovie", it will select
26480 audio instead of video.
26483 Specifies how many times to read the stream in sequence.
26484 If the value is 0, the stream will be looped infinitely.
26485 Default value is "1".
26487 Note that when the movie is looped the source timestamps are not
26488 changed, so it will generate non monotonically increasing timestamps.
26490 @item discontinuity
26491 Specifies the time difference between frames above which the point is
26492 considered a timestamp discontinuity which is removed by adjusting the later
26496 It allows overlaying a second video on top of the main input of
26497 a filtergraph, as shown in this graph:
26499 input -----------> deltapts0 --> overlay --> output
26502 movie --> scale--> deltapts1 -------+
26504 @subsection Examples
26508 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26509 on top of the input labelled "in":
26511 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26512 [in] setpts=PTS-STARTPTS [main];
26513 [main][over] overlay=16:16 [out]
26517 Read from a video4linux2 device, and overlay it on top of the input
26520 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26521 [in] setpts=PTS-STARTPTS [main];
26522 [main][over] overlay=16:16 [out]
26526 Read the first video stream and the audio stream with id 0x81 from
26527 dvd.vob; the video is connected to the pad named "video" and the audio is
26528 connected to the pad named "audio":
26530 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26534 @subsection Commands
26536 Both movie and amovie support the following commands:
26539 Perform seek using "av_seek_frame".
26540 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26543 @var{stream_index}: If stream_index is -1, a default
26544 stream is selected, and @var{timestamp} is automatically converted
26545 from AV_TIME_BASE units to the stream specific time_base.
26547 @var{timestamp}: Timestamp in AVStream.time_base units
26548 or, if no stream is specified, in AV_TIME_BASE units.
26550 @var{flags}: Flags which select direction and seeking mode.
26554 Get movie duration in AV_TIME_BASE units.
26558 @c man end MULTIMEDIA SOURCES