1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Commands
1091 This filter supports the all above options as @ref{commands}.
1093 @subsection Examples
1097 Fade in first 15 seconds of audio:
1099 afade=t=in:ss=0:d=15
1103 Fade out last 25 seconds of a 900 seconds audio:
1105 afade=t=out:st=875:d=25
1110 Denoise audio samples with FFT.
1112 A description of the accepted parameters follows.
1116 Set the noise reduction in dB, allowed range is 0.01 to 97.
1117 Default value is 12 dB.
1120 Set the noise floor in dB, allowed range is -80 to -20.
1121 Default value is -50 dB.
1126 It accepts the following values:
1135 Select shellac noise.
1138 Select custom noise, defined in @code{bn} option.
1140 Default value is white noise.
1144 Set custom band noise for every one of 15 bands.
1145 Bands are separated by ' ' or '|'.
1148 Set the residual floor in dB, allowed range is -80 to -20.
1149 Default value is -38 dB.
1152 Enable noise tracking. By default is disabled.
1153 With this enabled, noise floor is automatically adjusted.
1156 Enable residual tracking. By default is disabled.
1159 Set the output mode.
1161 It accepts the following values:
1164 Pass input unchanged.
1167 Pass noise filtered out.
1172 Default value is @var{o}.
1176 @subsection Commands
1178 This filter supports the following commands:
1180 @item sample_noise, sn
1181 Start or stop measuring noise profile.
1182 Syntax for the command is : "start" or "stop" string.
1183 After measuring noise profile is stopped it will be
1184 automatically applied in filtering.
1186 @item noise_reduction, nr
1187 Change noise reduction. Argument is single float number.
1188 Syntax for the command is : "@var{noise_reduction}"
1190 @item noise_floor, nf
1191 Change noise floor. Argument is single float number.
1192 Syntax for the command is : "@var{noise_floor}"
1194 @item output_mode, om
1195 Change output mode operation.
1196 Syntax for the command is : "i", "o" or "n" string.
1200 Apply arbitrary expressions to samples in frequency domain.
1204 Set frequency domain real expression for each separate channel separated
1205 by '|'. Default is "re".
1206 If the number of input channels is greater than the number of
1207 expressions, the last specified expression is used for the remaining
1211 Set frequency domain imaginary expression for each separate channel
1212 separated by '|'. Default is "im".
1214 Each expression in @var{real} and @var{imag} can contain the following
1215 constants and functions:
1222 current frequency bin number
1225 number of available bins
1228 channel number of the current expression
1237 current real part of frequency bin of current channel
1240 current imaginary part of frequency bin of current channel
1243 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1246 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1250 Set window size. Allowed range is from 16 to 131072.
1251 Default is @code{4096}
1254 Set window function. Default is @code{hann}.
1257 Set window overlap. If set to 1, the recommended overlap for selected
1258 window function will be picked. Default is @code{0.75}.
1261 @subsection Examples
1265 Leave almost only low frequencies in audio:
1267 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1271 Apply robotize effect:
1273 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1277 Apply whisper effect:
1279 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
1286 Apply an arbitrary Finite Impulse Response filter.
1288 This filter is designed for applying long FIR filters,
1289 up to 60 seconds long.
1291 It can be used as component for digital crossover filters,
1292 room equalization, cross talk cancellation, wavefield synthesis,
1293 auralization, ambiophonics, ambisonics and spatialization.
1295 This filter uses the streams higher than first one as FIR coefficients.
1296 If the non-first stream holds a single channel, it will be used
1297 for all input channels in the first stream, otherwise
1298 the number of channels in the non-first stream must be same as
1299 the number of channels in the first stream.
1301 It accepts the following parameters:
1305 Set dry gain. This sets input gain.
1308 Set wet gain. This sets final output gain.
1311 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1314 Enable applying gain measured from power of IR.
1316 Set which approach to use for auto gain measurement.
1320 Do not apply any gain.
1323 select peak gain, very conservative approach. This is default value.
1326 select DC gain, limited application.
1329 select gain to noise approach, this is most popular one.
1333 Set gain to be applied to IR coefficients before filtering.
1334 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1337 Set format of IR stream. Can be @code{mono} or @code{input}.
1338 Default is @code{input}.
1341 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1342 Allowed range is 0.1 to 60 seconds.
1345 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1346 By default it is disabled.
1349 Set for which IR channel to display frequency response. By default is first channel
1350 displayed. This option is used only when @var{response} is enabled.
1353 Set video stream size. This option is used only when @var{response} is enabled.
1356 Set video stream frame rate. This option is used only when @var{response} is enabled.
1359 Set minimal partition size used for convolution. Default is @var{8192}.
1360 Allowed range is from @var{1} to @var{32768}.
1361 Lower values decreases latency at cost of higher CPU usage.
1364 Set maximal partition size used for convolution. Default is @var{8192}.
1365 Allowed range is from @var{8} to @var{32768}.
1366 Lower values may increase CPU usage.
1369 Set number of input impulse responses streams which will be switchable at runtime.
1370 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1373 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1374 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1375 This option can be changed at runtime via @ref{commands}.
1378 @subsection Examples
1382 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1384 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1391 Set output format constraints for the input audio. The framework will
1392 negotiate the most appropriate format to minimize conversions.
1394 It accepts the following parameters:
1397 @item sample_fmts, f
1398 A '|'-separated list of requested sample formats.
1400 @item sample_rates, r
1401 A '|'-separated list of requested sample rates.
1403 @item channel_layouts, cl
1404 A '|'-separated list of requested channel layouts.
1406 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1407 for the required syntax.
1410 If a parameter is omitted, all values are allowed.
1412 Force the output to either unsigned 8-bit or signed 16-bit stereo
1414 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1418 Apply frequency shift to input audio samples.
1420 The filter accepts the following options:
1424 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1425 Default value is 0.0.
1428 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1429 Default value is 1.0.
1432 @subsection Commands
1434 This filter supports the all above options as @ref{commands}.
1438 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1439 processing reduces disturbing noise between useful signals.
1441 Gating is done by detecting the volume below a chosen level @var{threshold}
1442 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1443 floor is set via @var{range}. Because an exact manipulation of the signal
1444 would cause distortion of the waveform the reduction can be levelled over
1445 time. This is done by setting @var{attack} and @var{release}.
1447 @var{attack} determines how long the signal has to fall below the threshold
1448 before any reduction will occur and @var{release} sets the time the signal
1449 has to rise above the threshold to reduce the reduction again.
1450 Shorter signals than the chosen attack time will be left untouched.
1454 Set input level before filtering.
1455 Default is 1. Allowed range is from 0.015625 to 64.
1458 Set the mode of operation. Can be @code{upward} or @code{downward}.
1459 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1460 will be amplified, expanding dynamic range in upward direction.
1461 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1464 Set the level of gain reduction when the signal is below the threshold.
1465 Default is 0.06125. Allowed range is from 0 to 1.
1466 Setting this to 0 disables reduction and then filter behaves like expander.
1469 If a signal rises above this level the gain reduction is released.
1470 Default is 0.125. Allowed range is from 0 to 1.
1473 Set a ratio by which the signal is reduced.
1474 Default is 2. Allowed range is from 1 to 9000.
1477 Amount of milliseconds the signal has to rise above the threshold before gain
1479 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1482 Amount of milliseconds the signal has to fall below the threshold before the
1483 reduction is increased again. Default is 250 milliseconds.
1484 Allowed range is from 0.01 to 9000.
1487 Set amount of amplification of signal after processing.
1488 Default is 1. Allowed range is from 1 to 64.
1491 Curve the sharp knee around the threshold to enter gain reduction more softly.
1492 Default is 2.828427125. Allowed range is from 1 to 8.
1495 Choose if exact signal should be taken for detection or an RMS like one.
1496 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1499 Choose if the average level between all channels or the louder channel affects
1501 Default is @code{average}. Can be @code{average} or @code{maximum}.
1504 @subsection Commands
1506 This filter supports the all above options as @ref{commands}.
1510 Apply an arbitrary Infinite Impulse Response filter.
1512 It accepts the following parameters:
1516 Set B/numerator/zeros/reflection coefficients.
1519 Set A/denominator/poles/ladder coefficients.
1531 Set coefficients format.
1535 lattice-ladder function
1537 analog transfer function
1539 digital transfer function
1541 Z-plane zeros/poles, cartesian (default)
1543 Z-plane zeros/poles, polar radians
1545 Z-plane zeros/poles, polar degrees
1551 Set type of processing.
1563 Set filtering precision.
1567 double-precision floating-point (default)
1569 single-precision floating-point
1577 Normalize filter coefficients, by default is enabled.
1578 Enabling it will normalize magnitude response at DC to 0dB.
1581 How much to use filtered signal in output. Default is 1.
1582 Range is between 0 and 1.
1585 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1586 By default it is disabled.
1589 Set for which IR channel to display frequency response. By default is first channel
1590 displayed. This option is used only when @var{response} is enabled.
1593 Set video stream size. This option is used only when @var{response} is enabled.
1596 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1599 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1600 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1603 Different coefficients and gains can be provided for every channel, in such case
1604 use '|' to separate coefficients or gains. Last provided coefficients will be
1605 used for all remaining channels.
1607 @subsection Examples
1611 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1613 aiir=k=1:z=7.957584807809675810E-1 -2.575128568908332300 3.674839853930788710 -2.57512875289799137 7.957586296317130880E-1:p=1 -2.86950072432325953 3.63022088054647218 -2.28075678147272232 6.361362326477423500E-1:f=tf:r=d
1617 Same as above but in @code{zp} format:
1619 aiir=k=0.79575848078096756:z=0.80918701+0.58773007i 0.80918701-0.58773007i 0.80884700+0.58784055i 0.80884700-0.58784055i:p=0.63892345+0.59951235i 0.63892345-0.59951235i 0.79582691+0.44198673i 0.79582691-0.44198673i:f=zp:r=s
1623 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1625 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1631 The limiter prevents an input signal from rising over a desired threshold.
1632 This limiter uses lookahead technology to prevent your signal from distorting.
1633 It means that there is a small delay after the signal is processed. Keep in mind
1634 that the delay it produces is the attack time you set.
1636 The filter accepts the following options:
1640 Set input gain. Default is 1.
1643 Set output gain. Default is 1.
1646 Don't let signals above this level pass the limiter. Default is 1.
1649 The limiter will reach its attenuation level in this amount of time in
1650 milliseconds. Default is 5 milliseconds.
1653 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1654 Default is 50 milliseconds.
1657 When gain reduction is always needed ASC takes care of releasing to an
1658 average reduction level rather than reaching a reduction of 0 in the release
1662 Select how much the release time is affected by ASC, 0 means nearly no changes
1663 in release time while 1 produces higher release times.
1666 Auto level output signal. Default is enabled.
1667 This normalizes audio back to 0dB if enabled.
1670 Depending on picked setting it is recommended to upsample input 2x or 4x times
1671 with @ref{aresample} before applying this filter.
1675 Apply a two-pole all-pass filter with central frequency (in Hz)
1676 @var{frequency}, and filter-width @var{width}.
1677 An all-pass filter changes the audio's frequency to phase relationship
1678 without changing its frequency to amplitude relationship.
1680 The filter accepts the following options:
1684 Set frequency in Hz.
1687 Set method to specify band-width of filter.
1702 Specify the band-width of a filter in width_type units.
1705 How much to use filtered signal in output. Default is 1.
1706 Range is between 0 and 1.
1709 Specify which channels to filter, by default all available are filtered.
1712 Normalize biquad coefficients, by default is disabled.
1713 Enabling it will normalize magnitude response at DC to 0dB.
1716 Set the filter order, can be 1 or 2. Default is 2.
1719 Set transform type of IIR filter.
1728 Set precison of filtering.
1731 Pick automatic sample format depending on surround filters.
1733 Always use signed 16-bit.
1735 Always use signed 32-bit.
1737 Always use float 32-bit.
1739 Always use float 64-bit.
1743 @subsection Commands
1745 This filter supports the following commands:
1748 Change allpass frequency.
1749 Syntax for the command is : "@var{frequency}"
1752 Change allpass width_type.
1753 Syntax for the command is : "@var{width_type}"
1756 Change allpass width.
1757 Syntax for the command is : "@var{width}"
1761 Syntax for the command is : "@var{mix}"
1768 The filter accepts the following options:
1772 Set the number of loops. Setting this value to -1 will result in infinite loops.
1776 Set maximal number of samples. Default is 0.
1779 Set first sample of loop. Default is 0.
1785 Merge two or more audio streams into a single multi-channel stream.
1787 The filter accepts the following options:
1792 Set the number of inputs. Default is 2.
1796 If the channel layouts of the inputs are disjoint, and therefore compatible,
1797 the channel layout of the output will be set accordingly and the channels
1798 will be reordered as necessary. If the channel layouts of the inputs are not
1799 disjoint, the output will have all the channels of the first input then all
1800 the channels of the second input, in that order, and the channel layout of
1801 the output will be the default value corresponding to the total number of
1804 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1805 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1806 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1807 first input, b1 is the first channel of the second input).
1809 On the other hand, if both input are in stereo, the output channels will be
1810 in the default order: a1, a2, b1, b2, and the channel layout will be
1811 arbitrarily set to 4.0, which may or may not be the expected value.
1813 All inputs must have the same sample rate, and format.
1815 If inputs do not have the same duration, the output will stop with the
1818 @subsection Examples
1822 Merge two mono files into a stereo stream:
1824 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1828 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1830 ffmpeg -i input.mkv -filter_complex "[0:1][0:2][0:3][0:4][0:5][0:6] amerge=inputs=6" -c:a pcm_s16le output.mkv
1836 Mixes multiple audio inputs into a single output.
1838 Note that this filter only supports float samples (the @var{amerge}
1839 and @var{pan} audio filters support many formats). If the @var{amix}
1840 input has integer samples then @ref{aresample} will be automatically
1841 inserted to perform the conversion to float samples.
1845 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1847 will mix 3 input audio streams to a single output with the same duration as the
1848 first input and a dropout transition time of 3 seconds.
1850 It accepts the following parameters:
1854 The number of inputs. If unspecified, it defaults to 2.
1857 How to determine the end-of-stream.
1861 The duration of the longest input. (default)
1864 The duration of the shortest input.
1867 The duration of the first input.
1871 @item dropout_transition
1872 The transition time, in seconds, for volume renormalization when an input
1873 stream ends. The default value is 2 seconds.
1876 Specify weight of each input audio stream as sequence.
1877 Each weight is separated by space. By default all inputs have same weight.
1880 @subsection Commands
1882 This filter supports the following commands:
1885 Syntax is same as option with same name.
1890 Multiply first audio stream with second audio stream and store result
1891 in output audio stream. Multiplication is done by multiplying each
1892 sample from first stream with sample at same position from second stream.
1894 With this element-wise multiplication one can create amplitude fades and
1895 amplitude modulations.
1897 @section anequalizer
1899 High-order parametric multiband equalizer for each channel.
1901 It accepts the following parameters:
1905 This option string is in format:
1906 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1907 Each equalizer band is separated by '|'.
1911 Set channel number to which equalization will be applied.
1912 If input doesn't have that channel the entry is ignored.
1915 Set central frequency for band.
1916 If input doesn't have that frequency the entry is ignored.
1919 Set band width in Hertz.
1922 Set band gain in dB.
1925 Set filter type for band, optional, can be:
1929 Butterworth, this is default.
1940 With this option activated frequency response of anequalizer is displayed
1944 Set video stream size. Only useful if curves option is activated.
1947 Set max gain that will be displayed. Only useful if curves option is activated.
1948 Setting this to a reasonable value makes it possible to display gain which is derived from
1949 neighbour bands which are too close to each other and thus produce higher gain
1950 when both are activated.
1953 Set frequency scale used to draw frequency response in video output.
1954 Can be linear or logarithmic. Default is logarithmic.
1957 Set color for each channel curve which is going to be displayed in video stream.
1958 This is list of color names separated by space or by '|'.
1959 Unrecognised or missing colors will be replaced by white color.
1962 @subsection Examples
1966 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1967 for first 2 channels using Chebyshev type 1 filter:
1969 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1973 @subsection Commands
1975 This filter supports the following commands:
1978 Alter existing filter parameters.
1979 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1981 @var{fN} is existing filter number, starting from 0, if no such filter is available
1983 @var{freq} set new frequency parameter.
1984 @var{width} set new width parameter in Hertz.
1985 @var{gain} set new gain parameter in dB.
1987 Full filter invocation with asendcmd may look like this:
1988 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1993 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1995 Each sample is adjusted by looking for other samples with similar contexts. This
1996 context similarity is defined by comparing their surrounding patches of size
1997 @option{p}. Patches are searched in an area of @option{r} around the sample.
1999 The filter accepts the following options:
2003 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2006 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2007 Default value is 2 milliseconds.
2010 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2011 Default value is 6 milliseconds.
2014 Set the output mode.
2016 It accepts the following values:
2019 Pass input unchanged.
2022 Pass noise filtered out.
2027 Default value is @var{o}.
2031 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2034 @subsection Commands
2036 This filter supports the all above options as @ref{commands}.
2039 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2041 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2042 relate to producing the least mean square of the error signal (difference between the desired,
2043 2nd input audio stream and the actual signal, the 1st input audio stream).
2045 A description of the accepted options follows.
2058 Set the filter leakage.
2061 It accepts the following values:
2070 Pass filtered samples.
2073 Pass difference between desired and filtered samples.
2075 Default value is @var{o}.
2079 @subsection Examples
2083 One of many usages of this filter is noise reduction, input audio is filtered
2084 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2086 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2090 @subsection Commands
2092 This filter supports the same commands as options, excluding option @code{order}.
2096 Pass the audio source unchanged to the output.
2100 Pad the end of an audio stream with silence.
2102 This can be used together with @command{ffmpeg} @option{-shortest} to
2103 extend audio streams to the same length as the video stream.
2105 A description of the accepted options follows.
2109 Set silence packet size. Default value is 4096.
2112 Set the number of samples of silence to add to the end. After the
2113 value is reached, the stream is terminated. This option is mutually
2114 exclusive with @option{whole_len}.
2117 Set the minimum total number of samples in the output audio stream. If
2118 the value is longer than the input audio length, silence is added to
2119 the end, until the value is reached. This option is mutually exclusive
2120 with @option{pad_len}.
2123 Specify the duration of samples of silence to add. See
2124 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2125 for the accepted syntax. Used only if set to non-zero value.
2128 Specify the minimum total duration in the output audio stream. See
2129 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2130 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2131 the input audio length, silence is added to the end, until the value is reached.
2132 This option is mutually exclusive with @option{pad_dur}
2135 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2136 nor @option{whole_dur} option is set, the filter will add silence to the end of
2137 the input stream indefinitely.
2139 @subsection Examples
2143 Add 1024 samples of silence to the end of the input:
2149 Make sure the audio output will contain at least 10000 samples, pad
2150 the input with silence if required:
2152 apad=whole_len=10000
2156 Use @command{ffmpeg} to pad the audio input with silence, so that the
2157 video stream will always result the shortest and will be converted
2158 until the end in the output file when using the @option{shortest}
2161 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2166 Add a phasing effect to the input audio.
2168 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2169 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2171 A description of the accepted parameters follows.
2175 Set input gain. Default is 0.4.
2178 Set output gain. Default is 0.74
2181 Set delay in milliseconds. Default is 3.0.
2184 Set decay. Default is 0.4.
2187 Set modulation speed in Hz. Default is 0.5.
2190 Set modulation type. Default is triangular.
2192 It accepts the following values:
2199 @section aphaseshift
2200 Apply phase shift to input audio samples.
2202 The filter accepts the following options:
2206 Specify phase shift. Allowed range is from -1.0 to 1.0.
2207 Default value is 0.0.
2210 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2211 Default value is 1.0.
2214 @subsection Commands
2216 This filter supports the all above options as @ref{commands}.
2220 Audio pulsator is something between an autopanner and a tremolo.
2221 But it can produce funny stereo effects as well. Pulsator changes the volume
2222 of the left and right channel based on a LFO (low frequency oscillator) with
2223 different waveforms and shifted phases.
2224 This filter have the ability to define an offset between left and right
2225 channel. An offset of 0 means that both LFO shapes match each other.
2226 The left and right channel are altered equally - a conventional tremolo.
2227 An offset of 50% means that the shape of the right channel is exactly shifted
2228 in phase (or moved backwards about half of the frequency) - pulsator acts as
2229 an autopanner. At 1 both curves match again. Every setting in between moves the
2230 phase shift gapless between all stages and produces some "bypassing" sounds with
2231 sine and triangle waveforms. The more you set the offset near 1 (starting from
2232 the 0.5) the faster the signal passes from the left to the right speaker.
2234 The filter accepts the following options:
2238 Set input gain. By default it is 1. Range is [0.015625 - 64].
2241 Set output gain. By default it is 1. Range is [0.015625 - 64].
2244 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2245 sawup or sawdown. Default is sine.
2248 Set modulation. Define how much of original signal is affected by the LFO.
2251 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2254 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2257 Set pulse width. Default is 1. Allowed range is [0 - 2].
2260 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2263 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2267 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2271 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2272 if timing is set to hz.
2278 Resample the input audio to the specified parameters, using the
2279 libswresample library. If none are specified then the filter will
2280 automatically convert between its input and output.
2282 This filter is also able to stretch/squeeze the audio data to make it match
2283 the timestamps or to inject silence / cut out audio to make it match the
2284 timestamps, do a combination of both or do neither.
2286 The filter accepts the syntax
2287 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2288 expresses a sample rate and @var{resampler_options} is a list of
2289 @var{key}=@var{value} pairs, separated by ":". See the
2290 @ref{Resampler Options,,"Resampler Options" section in the
2291 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2292 for the complete list of supported options.
2294 @subsection Examples
2298 Resample the input audio to 44100Hz:
2304 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2305 samples per second compensation:
2307 aresample=async=1000
2313 Reverse an audio clip.
2315 Warning: This filter requires memory to buffer the entire clip, so trimming
2318 @subsection Examples
2322 Take the first 5 seconds of a clip, and reverse it.
2324 atrim=end=5,areverse
2330 Reduce noise from speech using Recurrent Neural Networks.
2332 This filter accepts the following options:
2336 Set train model file to load. This option is always required.
2339 Set how much to mix filtered samples into final output.
2340 Allowed range is from -1 to 1. Default value is 1.
2341 Negative values are special, they set how much to keep filtered noise
2342 in the final filter output. Set this option to -1 to hear actual
2343 noise removed from input signal.
2346 @section asetnsamples
2348 Set the number of samples per each output audio frame.
2350 The last output packet may contain a different number of samples, as
2351 the filter will flush all the remaining samples when the input audio
2354 The filter accepts the following options:
2358 @item nb_out_samples, n
2359 Set the number of frames per each output audio frame. The number is
2360 intended as the number of samples @emph{per each channel}.
2361 Default value is 1024.
2364 If set to 1, the filter will pad the last audio frame with zeroes, so
2365 that the last frame will contain the same number of samples as the
2366 previous ones. Default value is 1.
2369 For example, to set the number of per-frame samples to 1234 and
2370 disable padding for the last frame, use:
2372 asetnsamples=n=1234:p=0
2377 Set the sample rate without altering the PCM data.
2378 This will result in a change of speed and pitch.
2380 The filter accepts the following options:
2383 @item sample_rate, r
2384 Set the output sample rate. Default is 44100 Hz.
2389 Show a line containing various information for each input audio frame.
2390 The input audio is not modified.
2392 The shown line contains a sequence of key/value pairs of the form
2393 @var{key}:@var{value}.
2395 The following values are shown in the output:
2399 The (sequential) number of the input frame, starting from 0.
2402 The presentation timestamp of the input frame, in time base units; the time base
2403 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2406 The presentation timestamp of the input frame in seconds.
2409 position of the frame in the input stream, -1 if this information in
2410 unavailable and/or meaningless (for example in case of synthetic audio)
2419 The sample rate for the audio frame.
2422 The number of samples (per channel) in the frame.
2425 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2426 audio, the data is treated as if all the planes were concatenated.
2428 @item plane_checksums
2429 A list of Adler-32 checksums for each data plane.
2433 Apply audio soft clipping.
2435 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2436 along a smooth curve, rather than the abrupt shape of hard-clipping.
2438 This filter accepts the following options:
2442 Set type of soft-clipping.
2444 It accepts the following values:
2458 Set threshold from where to start clipping. Default value is 0dB or 1.
2461 Set gain applied to output. Default value is 0dB or 1.
2464 Set additional parameter which controls sigmoid function.
2467 Set oversampling factor.
2470 @subsection Commands
2472 This filter supports the all above options as @ref{commands}.
2475 Automatic Speech Recognition
2477 This filter uses PocketSphinx for speech recognition. To enable
2478 compilation of this filter, you need to configure FFmpeg with
2479 @code{--enable-pocketsphinx}.
2481 It accepts the following options:
2485 Set sampling rate of input audio. Defaults is @code{16000}.
2486 This need to match speech models, otherwise one will get poor results.
2489 Set dictionary containing acoustic model files.
2492 Set pronunciation dictionary.
2495 Set language model file.
2498 Set language model set.
2501 Set which language model to use.
2504 Set output for log messages.
2507 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2512 Display time domain statistical information about the audio channels.
2513 Statistics are calculated and displayed for each audio channel and,
2514 where applicable, an overall figure is also given.
2516 It accepts the following option:
2519 Short window length in seconds, used for peak and trough RMS measurement.
2520 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2524 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2525 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2528 Available keys for each channel are:
2574 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2575 this @code{lavfi.astats.Overall.Peak_count}.
2577 For description what each key means read below.
2580 Set number of frame after which stats are going to be recalculated.
2581 Default is disabled.
2583 @item measure_perchannel
2584 Select the entries which need to be measured per channel. The metadata keys can
2585 be used as flags, default is @option{all} which measures everything.
2586 @option{none} disables all per channel measurement.
2588 @item measure_overall
2589 Select the entries which need to be measured overall. The metadata keys can
2590 be used as flags, default is @option{all} which measures everything.
2591 @option{none} disables all overall measurement.
2595 A description of each shown parameter follows:
2599 Mean amplitude displacement from zero.
2602 Minimal sample level.
2605 Maximal sample level.
2607 @item Min difference
2608 Minimal difference between two consecutive samples.
2610 @item Max difference
2611 Maximal difference between two consecutive samples.
2613 @item Mean difference
2614 Mean difference between two consecutive samples.
2615 The average of each difference between two consecutive samples.
2617 @item RMS difference
2618 Root Mean Square difference between two consecutive samples.
2622 Standard peak and RMS level measured in dBFS.
2626 Peak and trough values for RMS level measured over a short window.
2629 Standard ratio of peak to RMS level (note: not in dB).
2632 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2633 (i.e. either @var{Min level} or @var{Max level}).
2636 Number of occasions (not the number of samples) that the signal attained either
2637 @var{Min level} or @var{Max level}.
2639 @item Noise floor dB
2640 Minimum local peak measured in dBFS over a short window.
2642 @item Noise floor count
2643 Number of occasions (not the number of samples) that the signal attained
2647 Overall bit depth of audio. Number of bits used for each sample.
2650 Measured dynamic range of audio in dB.
2652 @item Zero crossings
2653 Number of points where the waveform crosses the zero level axis.
2655 @item Zero crossings rate
2656 Rate of Zero crossings and number of audio samples.
2660 Boost subwoofer frequencies.
2662 The filter accepts the following options:
2666 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2667 Default value is 0.7.
2670 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2671 Default value is 0.7.
2674 Set delay line decay gain value. Allowed range is from 0 to 1.
2675 Default value is 0.7.
2678 Set delay line feedback gain value. Allowed range is from 0 to 1.
2679 Default value is 0.9.
2682 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2683 Default value is 100.
2686 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2687 Default value is 0.5.
2690 Set delay. Allowed range is from 1 to 100.
2691 Default value is 20.
2694 @subsection Commands
2696 This filter supports the all above options as @ref{commands}.
2699 Cut subwoofer frequencies.
2701 This filter allows to set custom, steeper
2702 roll off than highpass filter, and thus is able to more attenuate
2703 frequency content in stop-band.
2705 The filter accepts the following options:
2709 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2710 Default value is 20.
2713 Set filter order. Available values are from 3 to 20.
2714 Default value is 10.
2717 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2720 @subsection Commands
2722 This filter supports the all above options as @ref{commands}.
2725 Cut super frequencies.
2727 The filter accepts the following options:
2731 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2732 Default value is 20000.
2735 Set filter order. Available values are from 3 to 20.
2736 Default value is 10.
2739 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2742 @subsection Commands
2744 This filter supports the all above options as @ref{commands}.
2747 Apply high order Butterworth band-pass filter.
2749 The filter accepts the following options:
2753 Set center frequency in Hertz. Allowed range is 2 to 999999.
2754 Default value is 1000.
2757 Set filter order. Available values are from 4 to 20.
2761 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2764 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2767 @subsection Commands
2769 This filter supports the all above options as @ref{commands}.
2772 Apply high order Butterworth band-stop filter.
2774 The filter accepts the following options:
2778 Set center frequency in Hertz. Allowed range is 2 to 999999.
2779 Default value is 1000.
2782 Set filter order. Available values are from 4 to 20.
2786 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2789 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2792 @subsection Commands
2794 This filter supports the all above options as @ref{commands}.
2800 The filter accepts exactly one parameter, the audio tempo. If not
2801 specified then the filter will assume nominal 1.0 tempo. Tempo must
2802 be in the [0.5, 100.0] range.
2804 Note that tempo greater than 2 will skip some samples rather than
2805 blend them in. If for any reason this is a concern it is always
2806 possible to daisy-chain several instances of atempo to achieve the
2807 desired product tempo.
2809 @subsection Examples
2813 Slow down audio to 80% tempo:
2819 To speed up audio to 300% tempo:
2825 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2827 atempo=sqrt(3),atempo=sqrt(3)
2831 @subsection Commands
2833 This filter supports the following commands:
2836 Change filter tempo scale factor.
2837 Syntax for the command is : "@var{tempo}"
2842 Trim the input so that the output contains one continuous subpart of the input.
2844 It accepts the following parameters:
2847 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2848 sample with the timestamp @var{start} will be the first sample in the output.
2851 Specify time of the first audio sample that will be dropped, i.e. the
2852 audio sample immediately preceding the one with the timestamp @var{end} will be
2853 the last sample in the output.
2856 Same as @var{start}, except this option sets the start timestamp in samples
2860 Same as @var{end}, except this option sets the end timestamp in samples instead
2864 The maximum duration of the output in seconds.
2867 The number of the first sample that should be output.
2870 The number of the first sample that should be dropped.
2873 @option{start}, @option{end}, and @option{duration} are expressed as time
2874 duration specifications; see
2875 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2877 Note that the first two sets of the start/end options and the @option{duration}
2878 option look at the frame timestamp, while the _sample options simply count the
2879 samples that pass through the filter. So start/end_pts and start/end_sample will
2880 give different results when the timestamps are wrong, inexact or do not start at
2881 zero. Also note that this filter does not modify the timestamps. If you wish
2882 to have the output timestamps start at zero, insert the asetpts filter after the
2885 If multiple start or end options are set, this filter tries to be greedy and
2886 keep all samples that match at least one of the specified constraints. To keep
2887 only the part that matches all the constraints at once, chain multiple atrim
2890 The defaults are such that all the input is kept. So it is possible to set e.g.
2891 just the end values to keep everything before the specified time.
2896 Drop everything except the second minute of input:
2898 ffmpeg -i INPUT -af atrim=60:120
2902 Keep only the first 1000 samples:
2904 ffmpeg -i INPUT -af atrim=end_sample=1000
2909 @section axcorrelate
2910 Calculate normalized cross-correlation between two input audio streams.
2912 Resulted samples are always between -1 and 1 inclusive.
2913 If result is 1 it means two input samples are highly correlated in that selected segment.
2914 Result 0 means they are not correlated at all.
2915 If result is -1 it means two input samples are out of phase, which means they cancel each
2918 The filter accepts the following options:
2922 Set size of segment over which cross-correlation is calculated.
2923 Default is 256. Allowed range is from 2 to 131072.
2926 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2927 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2928 are always zero and thus need much less calculations to make.
2929 This is generally not true, but is valid for typical audio streams.
2932 @subsection Examples
2936 Calculate correlation between channels in stereo audio stream:
2938 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2944 Apply a two-pole Butterworth band-pass filter with central
2945 frequency @var{frequency}, and (3dB-point) band-width width.
2946 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2947 instead of the default: constant 0dB peak gain.
2948 The filter roll off at 6dB per octave (20dB per decade).
2950 The filter accepts the following options:
2954 Set the filter's central frequency. Default is @code{3000}.
2957 Constant skirt gain if set to 1. Defaults to 0.
2960 Set method to specify band-width of filter.
2975 Specify the band-width of a filter in width_type units.
2978 How much to use filtered signal in output. Default is 1.
2979 Range is between 0 and 1.
2982 Specify which channels to filter, by default all available are filtered.
2985 Normalize biquad coefficients, by default is disabled.
2986 Enabling it will normalize magnitude response at DC to 0dB.
2989 Set transform type of IIR filter.
2998 Set precison of filtering.
3001 Pick automatic sample format depending on surround filters.
3003 Always use signed 16-bit.
3005 Always use signed 32-bit.
3007 Always use float 32-bit.
3009 Always use float 64-bit.
3013 @subsection Commands
3015 This filter supports the following commands:
3018 Change bandpass frequency.
3019 Syntax for the command is : "@var{frequency}"
3022 Change bandpass width_type.
3023 Syntax for the command is : "@var{width_type}"
3026 Change bandpass width.
3027 Syntax for the command is : "@var{width}"
3030 Change bandpass mix.
3031 Syntax for the command is : "@var{mix}"
3036 Apply a two-pole Butterworth band-reject filter with central
3037 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
3038 The filter roll off at 6dB per octave (20dB per decade).
3040 The filter accepts the following options:
3044 Set the filter's central frequency. Default is @code{3000}.
3047 Set method to specify band-width of filter.
3062 Specify the band-width of a filter in width_type units.
3065 How much to use filtered signal in output. Default is 1.
3066 Range is between 0 and 1.
3069 Specify which channels to filter, by default all available are filtered.
3072 Normalize biquad coefficients, by default is disabled.
3073 Enabling it will normalize magnitude response at DC to 0dB.
3076 Set transform type of IIR filter.
3085 Set precison of filtering.
3088 Pick automatic sample format depending on surround filters.
3090 Always use signed 16-bit.
3092 Always use signed 32-bit.
3094 Always use float 32-bit.
3096 Always use float 64-bit.
3100 @subsection Commands
3102 This filter supports the following commands:
3105 Change bandreject frequency.
3106 Syntax for the command is : "@var{frequency}"
3109 Change bandreject width_type.
3110 Syntax for the command is : "@var{width_type}"
3113 Change bandreject width.
3114 Syntax for the command is : "@var{width}"
3117 Change bandreject mix.
3118 Syntax for the command is : "@var{mix}"
3121 @section bass, lowshelf
3123 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3124 shelving filter with a response similar to that of a standard
3125 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3127 The filter accepts the following options:
3131 Give the gain at 0 Hz. Its useful range is about -20
3132 (for a large cut) to +20 (for a large boost).
3133 Beware of clipping when using a positive gain.
3136 Set the filter's central frequency and so can be used
3137 to extend or reduce the frequency range to be boosted or cut.
3138 The default value is @code{100} Hz.
3141 Set method to specify band-width of filter.
3156 Determine how steep is the filter's shelf transition.
3159 Set number of poles. Default is 2.
3162 How much to use filtered signal in output. Default is 1.
3163 Range is between 0 and 1.
3166 Specify which channels to filter, by default all available are filtered.
3169 Normalize biquad coefficients, by default is disabled.
3170 Enabling it will normalize magnitude response at DC to 0dB.
3173 Set transform type of IIR filter.
3182 Set precison of filtering.
3185 Pick automatic sample format depending on surround filters.
3187 Always use signed 16-bit.
3189 Always use signed 32-bit.
3191 Always use float 32-bit.
3193 Always use float 64-bit.
3197 @subsection Commands
3199 This filter supports the following commands:
3202 Change bass frequency.
3203 Syntax for the command is : "@var{frequency}"
3206 Change bass width_type.
3207 Syntax for the command is : "@var{width_type}"
3211 Syntax for the command is : "@var{width}"
3215 Syntax for the command is : "@var{gain}"
3219 Syntax for the command is : "@var{mix}"
3224 Apply a biquad IIR filter with the given coefficients.
3225 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3226 are the numerator and denominator coefficients respectively.
3227 and @var{channels}, @var{c} specify which channels to filter, by default all
3228 available are filtered.
3230 @subsection Commands
3232 This filter supports the following commands:
3240 Change biquad parameter.
3241 Syntax for the command is : "@var{value}"
3244 How much to use filtered signal in output. Default is 1.
3245 Range is between 0 and 1.
3248 Specify which channels to filter, by default all available are filtered.
3251 Normalize biquad coefficients, by default is disabled.
3252 Enabling it will normalize magnitude response at DC to 0dB.
3255 Set transform type of IIR filter.
3264 Set precison of filtering.
3267 Pick automatic sample format depending on surround filters.
3269 Always use signed 16-bit.
3271 Always use signed 32-bit.
3273 Always use float 32-bit.
3275 Always use float 64-bit.
3280 Bauer stereo to binaural transformation, which improves headphone listening of
3281 stereo audio records.
3283 To enable compilation of this filter you need to configure FFmpeg with
3284 @code{--enable-libbs2b}.
3286 It accepts the following parameters:
3290 Pre-defined crossfeed level.
3294 Default level (fcut=700, feed=50).
3297 Chu Moy circuit (fcut=700, feed=60).
3300 Jan Meier circuit (fcut=650, feed=95).
3305 Cut frequency (in Hz).
3314 Remap input channels to new locations.
3316 It accepts the following parameters:
3319 Map channels from input to output. The argument is a '|'-separated list of
3320 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3321 @var{in_channel} form. @var{in_channel} can be either the name of the input
3322 channel (e.g. FL for front left) or its index in the input channel layout.
3323 @var{out_channel} is the name of the output channel or its index in the output
3324 channel layout. If @var{out_channel} is not given then it is implicitly an
3325 index, starting with zero and increasing by one for each mapping.
3327 @item channel_layout
3328 The channel layout of the output stream.
3331 If no mapping is present, the filter will implicitly map input channels to
3332 output channels, preserving indices.
3334 @subsection Examples
3338 For example, assuming a 5.1+downmix input MOV file,
3340 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3342 will create an output WAV file tagged as stereo from the downmix channels of
3346 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3348 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3352 @section channelsplit
3354 Split each channel from an input audio stream into a separate output stream.
3356 It accepts the following parameters:
3358 @item channel_layout
3359 The channel layout of the input stream. The default is "stereo".
3361 A channel layout describing the channels to be extracted as separate output streams
3362 or "all" to extract each input channel as a separate stream. The default is "all".
3364 Choosing channels not present in channel layout in the input will result in an error.
3367 @subsection Examples
3371 For example, assuming a stereo input MP3 file,
3373 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3375 will create an output Matroska file with two audio streams, one containing only
3376 the left channel and the other the right channel.
3379 Split a 5.1 WAV file into per-channel files:
3381 ffmpeg -i in.wav -filter_complex
3382 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3383 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3384 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3389 Extract only LFE from a 5.1 WAV file:
3391 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3392 -map '[LFE]' lfe.wav
3397 Add a chorus effect to the audio.
3399 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3401 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3402 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3403 The modulation depth defines the range the modulated delay is played before or after
3404 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3405 sound tuned around the original one, like in a chorus where some vocals are slightly
3408 It accepts the following parameters:
3411 Set input gain. Default is 0.4.
3414 Set output gain. Default is 0.4.
3417 Set delays. A typical delay is around 40ms to 60ms.
3429 @subsection Examples
3435 chorus=0.7:0.9:55:0.4:0.25:2
3441 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3445 Fuller sounding chorus with three delays:
3447 chorus=0.5:0.9:50|60|40:0.4|0.32|0.3:0.25|0.4|0.3:2|2.3|1.3
3452 Compress or expand the audio's dynamic range.
3454 It accepts the following parameters:
3460 A list of times in seconds for each channel over which the instantaneous level
3461 of the input signal is averaged to determine its volume. @var{attacks} refers to
3462 increase of volume and @var{decays} refers to decrease of volume. For most
3463 situations, the attack time (response to the audio getting louder) should be
3464 shorter than the decay time, because the human ear is more sensitive to sudden
3465 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3466 a typical value for decay is 0.8 seconds.
3467 If specified number of attacks & decays is lower than number of channels, the last
3468 set attack/decay will be used for all remaining channels.
3471 A list of points for the transfer function, specified in dB relative to the
3472 maximum possible signal amplitude. Each key points list must be defined using
3473 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3474 @code{x0/y0 x1/y1 x2/y2 ....}
3476 The input values must be in strictly increasing order but the transfer function
3477 does not have to be monotonically rising. The point @code{0/0} is assumed but
3478 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3479 function are @code{-70/-70|-60/-20|1/0}.
3482 Set the curve radius in dB for all joints. It defaults to 0.01.
3485 Set the additional gain in dB to be applied at all points on the transfer
3486 function. This allows for easy adjustment of the overall gain.
3490 Set an initial volume, in dB, to be assumed for each channel when filtering
3491 starts. This permits the user to supply a nominal level initially, so that, for
3492 example, a very large gain is not applied to initial signal levels before the
3493 companding has begun to operate. A typical value for audio which is initially
3494 quiet is -90 dB. It defaults to 0.
3497 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3498 delayed before being fed to the volume adjuster. Specifying a delay
3499 approximately equal to the attack/decay times allows the filter to effectively
3500 operate in predictive rather than reactive mode. It defaults to 0.
3504 @subsection Examples
3508 Make music with both quiet and loud passages suitable for listening to in a
3511 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3514 Another example for audio with whisper and explosion parts:
3516 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3520 A noise gate for when the noise is at a lower level than the signal:
3522 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3526 Here is another noise gate, this time for when the noise is at a higher level
3527 than the signal (making it, in some ways, similar to squelch):
3529 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3533 2:1 compression starting at -6dB:
3535 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3539 2:1 compression starting at -9dB:
3541 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3545 2:1 compression starting at -12dB:
3547 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3551 2:1 compression starting at -18dB:
3553 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3557 3:1 compression starting at -15dB:
3559 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3565 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3571 compand=attacks=0:points=-80/-169|-54/-80|-49.5/-64.6|-41.1/-41.1|-25.8/-15|-10.8/-4.5|0/0|20/8.3
3575 Hard limiter at -6dB:
3577 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3581 Hard limiter at -12dB:
3583 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3587 Hard noise gate at -35 dB:
3589 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3595 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3599 @section compensationdelay
3601 Compensation Delay Line is a metric based delay to compensate differing
3602 positions of microphones or speakers.
3604 For example, you have recorded guitar with two microphones placed in
3605 different locations. Because the front of sound wave has fixed speed in
3606 normal conditions, the phasing of microphones can vary and depends on
3607 their location and interposition. The best sound mix can be achieved when
3608 these microphones are in phase (synchronized). Note that a distance of
3609 ~30 cm between microphones makes one microphone capture the signal in
3610 antiphase to the other microphone. That makes the final mix sound moody.
3611 This filter helps to solve phasing problems by adding different delays
3612 to each microphone track and make them synchronized.
3614 The best result can be reached when you take one track as base and
3615 synchronize other tracks one by one with it.
3616 Remember that synchronization/delay tolerance depends on sample rate, too.
3617 Higher sample rates will give more tolerance.
3619 The filter accepts the following parameters:
3623 Set millimeters distance. This is compensation distance for fine tuning.
3627 Set cm distance. This is compensation distance for tightening distance setup.
3631 Set meters distance. This is compensation distance for hard distance setup.
3635 Set dry amount. Amount of unprocessed (dry) signal.
3639 Set wet amount. Amount of processed (wet) signal.
3643 Set temperature in degrees Celsius. This is the temperature of the environment.
3648 Apply headphone crossfeed filter.
3650 Crossfeed is the process of blending the left and right channels of stereo
3652 It is mainly used to reduce extreme stereo separation of low frequencies.
3654 The intent is to produce more speaker like sound to the listener.
3656 The filter accepts the following options:
3660 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3661 This sets gain of low shelf filter for side part of stereo image.
3662 Default is -6dB. Max allowed is -30db when strength is set to 1.
3665 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3666 This sets cut off frequency of low shelf filter. Default is cut off near
3667 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3670 Set curve slope of low shelf filter. Default is 0.5.
3671 Allowed range is from 0.01 to 1.
3674 Set input gain. Default is 0.9.
3677 Set output gain. Default is 1.
3680 @subsection Commands
3682 This filter supports the all above options as @ref{commands}.
3684 @section crystalizer
3685 Simple algorithm to expand audio dynamic range.
3687 The filter accepts the following options:
3691 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3692 (unchanged sound) to 10.0 (maximum effect).
3695 Enable clipping. By default is enabled.
3698 @subsection Commands
3700 This filter supports the all above options as @ref{commands}.
3703 Apply a DC shift to the audio.
3705 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3706 in the recording chain) from the audio. The effect of a DC offset is reduced
3707 headroom and hence volume. The @ref{astats} filter can be used to determine if
3708 a signal has a DC offset.
3712 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3716 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3717 used to prevent clipping.
3722 Apply de-essing to the audio samples.
3726 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3730 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3734 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3738 Set the output mode.
3740 It accepts the following values:
3743 Pass input unchanged.
3746 Pass ess filtered out.
3751 Default value is @var{o}.
3757 Measure audio dynamic range.
3759 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3760 is found in transition material. And anything less that 8 have very poor dynamics
3761 and is very compressed.
3763 The filter accepts the following options:
3767 Set window length in seconds used to split audio into segments of equal length.
3768 Default is 3 seconds.
3772 Dynamic Audio Normalizer.
3774 This filter applies a certain amount of gain to the input audio in order
3775 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3776 contrast to more "simple" normalization algorithms, the Dynamic Audio
3777 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3778 This allows for applying extra gain to the "quiet" sections of the audio
3779 while avoiding distortions or clipping the "loud" sections. In other words:
3780 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3781 sections, in the sense that the volume of each section is brought to the
3782 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3783 this goal *without* applying "dynamic range compressing". It will retain 100%
3784 of the dynamic range *within* each section of the audio file.
3788 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3789 Default is 500 milliseconds.
3790 The Dynamic Audio Normalizer processes the input audio in small chunks,
3791 referred to as frames. This is required, because a peak magnitude has no
3792 meaning for just a single sample value. Instead, we need to determine the
3793 peak magnitude for a contiguous sequence of sample values. While a "standard"
3794 normalizer would simply use the peak magnitude of the complete file, the
3795 Dynamic Audio Normalizer determines the peak magnitude individually for each
3796 frame. The length of a frame is specified in milliseconds. By default, the
3797 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3798 been found to give good results with most files.
3799 Note that the exact frame length, in number of samples, will be determined
3800 automatically, based on the sampling rate of the individual input audio file.
3803 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3804 number. Default is 31.
3805 Probably the most important parameter of the Dynamic Audio Normalizer is the
3806 @code{window size} of the Gaussian smoothing filter. The filter's window size
3807 is specified in frames, centered around the current frame. For the sake of
3808 simplicity, this must be an odd number. Consequently, the default value of 31
3809 takes into account the current frame, as well as the 15 preceding frames and
3810 the 15 subsequent frames. Using a larger window results in a stronger
3811 smoothing effect and thus in less gain variation, i.e. slower gain
3812 adaptation. Conversely, using a smaller window results in a weaker smoothing
3813 effect and thus in more gain variation, i.e. faster gain adaptation.
3814 In other words, the more you increase this value, the more the Dynamic Audio
3815 Normalizer will behave like a "traditional" normalization filter. On the
3816 contrary, the more you decrease this value, the more the Dynamic Audio
3817 Normalizer will behave like a dynamic range compressor.
3820 Set the target peak value. This specifies the highest permissible magnitude
3821 level for the normalized audio input. This filter will try to approach the
3822 target peak magnitude as closely as possible, but at the same time it also
3823 makes sure that the normalized signal will never exceed the peak magnitude.
3824 A frame's maximum local gain factor is imposed directly by the target peak
3825 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3826 It is not recommended to go above this value.
3829 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3830 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3831 factor for each input frame, i.e. the maximum gain factor that does not
3832 result in clipping or distortion. The maximum gain factor is determined by
3833 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3834 additionally bounds the frame's maximum gain factor by a predetermined
3835 (global) maximum gain factor. This is done in order to avoid excessive gain
3836 factors in "silent" or almost silent frames. By default, the maximum gain
3837 factor is 10.0, For most inputs the default value should be sufficient and
3838 it usually is not recommended to increase this value. Though, for input
3839 with an extremely low overall volume level, it may be necessary to allow even
3840 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3841 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3842 Instead, a "sigmoid" threshold function will be applied. This way, the
3843 gain factors will smoothly approach the threshold value, but never exceed that
3847 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3848 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3849 This means that the maximum local gain factor for each frame is defined
3850 (only) by the frame's highest magnitude sample. This way, the samples can
3851 be amplified as much as possible without exceeding the maximum signal
3852 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3853 Normalizer can also take into account the frame's root mean square,
3854 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3855 determine the power of a time-varying signal. It is therefore considered
3856 that the RMS is a better approximation of the "perceived loudness" than
3857 just looking at the signal's peak magnitude. Consequently, by adjusting all
3858 frames to a constant RMS value, a uniform "perceived loudness" can be
3859 established. If a target RMS value has been specified, a frame's local gain
3860 factor is defined as the factor that would result in exactly that RMS value.
3861 Note, however, that the maximum local gain factor is still restricted by the
3862 frame's highest magnitude sample, in order to prevent clipping.
3865 Enable channels coupling. By default is enabled.
3866 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3867 amount. This means the same gain factor will be applied to all channels, i.e.
3868 the maximum possible gain factor is determined by the "loudest" channel.
3869 However, in some recordings, it may happen that the volume of the different
3870 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3871 In this case, this option can be used to disable the channel coupling. This way,
3872 the gain factor will be determined independently for each channel, depending
3873 only on the individual channel's highest magnitude sample. This allows for
3874 harmonizing the volume of the different channels.
3877 Enable DC bias correction. By default is disabled.
3878 An audio signal (in the time domain) is a sequence of sample values.
3879 In the Dynamic Audio Normalizer these sample values are represented in the
3880 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3881 audio signal, or "waveform", should be centered around the zero point.
3882 That means if we calculate the mean value of all samples in a file, or in a
3883 single frame, then the result should be 0.0 or at least very close to that
3884 value. If, however, there is a significant deviation of the mean value from
3885 0.0, in either positive or negative direction, this is referred to as a
3886 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3887 Audio Normalizer provides optional DC bias correction.
3888 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3889 the mean value, or "DC correction" offset, of each input frame and subtract
3890 that value from all of the frame's sample values which ensures those samples
3891 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3892 boundaries, the DC correction offset values will be interpolated smoothly
3893 between neighbouring frames.
3895 @item altboundary, b
3896 Enable alternative boundary mode. By default is disabled.
3897 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3898 around each frame. This includes the preceding frames as well as the
3899 subsequent frames. However, for the "boundary" frames, located at the very
3900 beginning and at the very end of the audio file, not all neighbouring
3901 frames are available. In particular, for the first few frames in the audio
3902 file, the preceding frames are not known. And, similarly, for the last few
3903 frames in the audio file, the subsequent frames are not known. Thus, the
3904 question arises which gain factors should be assumed for the missing frames
3905 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3906 to deal with this situation. The default boundary mode assumes a gain factor
3907 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3908 "fade out" at the beginning and at the end of the input, respectively.
3911 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3912 By default, the Dynamic Audio Normalizer does not apply "traditional"
3913 compression. This means that signal peaks will not be pruned and thus the
3914 full dynamic range will be retained within each local neighbourhood. However,
3915 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3916 normalization algorithm with a more "traditional" compression.
3917 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3918 (thresholding) function. If (and only if) the compression feature is enabled,
3919 all input frames will be processed by a soft knee thresholding function prior
3920 to the actual normalization process. Put simply, the thresholding function is
3921 going to prune all samples whose magnitude exceeds a certain threshold value.
3922 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3923 value. Instead, the threshold value will be adjusted for each individual
3925 In general, smaller parameters result in stronger compression, and vice versa.
3926 Values below 3.0 are not recommended, because audible distortion may appear.
3929 Set the target threshold value. This specifies the lowest permissible
3930 magnitude level for the audio input which will be normalized.
3931 If input frame volume is above this value frame will be normalized.
3932 Otherwise frame may not be normalized at all. The default value is set
3933 to 0, which means all input frames will be normalized.
3934 This option is mostly useful if digital noise is not wanted to be amplified.
3937 @subsection Commands
3939 This filter supports the all above options as @ref{commands}.
3943 Make audio easier to listen to on headphones.
3945 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3946 so that when listened to on headphones the stereo image is moved from
3947 inside your head (standard for headphones) to outside and in front of
3948 the listener (standard for speakers).
3954 Apply a two-pole peaking equalisation (EQ) filter. With this
3955 filter, the signal-level at and around a selected frequency can
3956 be increased or decreased, whilst (unlike bandpass and bandreject
3957 filters) that at all other frequencies is unchanged.
3959 In order to produce complex equalisation curves, this filter can
3960 be given several times, each with a different central frequency.
3962 The filter accepts the following options:
3966 Set the filter's central frequency in Hz.
3969 Set method to specify band-width of filter.
3984 Specify the band-width of a filter in width_type units.
3987 Set the required gain or attenuation in dB.
3988 Beware of clipping when using a positive gain.
3991 How much to use filtered signal in output. Default is 1.
3992 Range is between 0 and 1.
3995 Specify which channels to filter, by default all available are filtered.
3998 Normalize biquad coefficients, by default is disabled.
3999 Enabling it will normalize magnitude response at DC to 0dB.
4002 Set transform type of IIR filter.
4011 Set precison of filtering.
4014 Pick automatic sample format depending on surround filters.
4016 Always use signed 16-bit.
4018 Always use signed 32-bit.
4020 Always use float 32-bit.
4022 Always use float 64-bit.
4026 @subsection Examples
4029 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4031 equalizer=f=1000:t=h:width=200:g=-10
4035 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4037 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4041 @subsection Commands
4043 This filter supports the following commands:
4046 Change equalizer frequency.
4047 Syntax for the command is : "@var{frequency}"
4050 Change equalizer width_type.
4051 Syntax for the command is : "@var{width_type}"
4054 Change equalizer width.
4055 Syntax for the command is : "@var{width}"
4058 Change equalizer gain.
4059 Syntax for the command is : "@var{gain}"
4062 Change equalizer mix.
4063 Syntax for the command is : "@var{mix}"
4066 @section extrastereo
4068 Linearly increases the difference between left and right channels which
4069 adds some sort of "live" effect to playback.
4071 The filter accepts the following options:
4075 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4076 (average of both channels), with 1.0 sound will be unchanged, with
4077 -1.0 left and right channels will be swapped.
4080 Enable clipping. By default is enabled.
4083 @subsection Commands
4085 This filter supports the all above options as @ref{commands}.
4087 @section firequalizer
4088 Apply FIR Equalization using arbitrary frequency response.
4090 The filter accepts the following option:
4094 Set gain curve equation (in dB). The expression can contain variables:
4097 the evaluated frequency
4101 channel number, set to 0 when multichannels evaluation is disabled
4103 channel id, see libavutil/channel_layout.h, set to the first channel id when
4104 multichannels evaluation is disabled
4108 channel_layout, see libavutil/channel_layout.h
4113 @item gain_interpolate(f)
4114 interpolate gain on frequency f based on gain_entry
4115 @item cubic_interpolate(f)
4116 same as gain_interpolate, but smoother
4118 This option is also available as command. Default is @code{gain_interpolate(f)}.
4121 Set gain entry for gain_interpolate function. The expression can
4125 store gain entry at frequency f with value g
4127 This option is also available as command.
4130 Set filter delay in seconds. Higher value means more accurate.
4131 Default is @code{0.01}.
4134 Set filter accuracy in Hz. Lower value means more accurate.
4135 Default is @code{5}.
4138 Set window function. Acceptable values are:
4141 rectangular window, useful when gain curve is already smooth
4143 hann window (default)
4149 3-terms continuous 1st derivative nuttall window
4151 minimum 3-terms discontinuous nuttall window
4153 4-terms continuous 1st derivative nuttall window
4155 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4157 blackman-harris window
4163 If enabled, use fixed number of audio samples. This improves speed when
4164 filtering with large delay. Default is disabled.
4167 Enable multichannels evaluation on gain. Default is disabled.
4170 Enable zero phase mode by subtracting timestamp to compensate delay.
4171 Default is disabled.
4174 Set scale used by gain. Acceptable values are:
4177 linear frequency, linear gain
4179 linear frequency, logarithmic (in dB) gain (default)
4181 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4183 logarithmic frequency, logarithmic gain
4187 Set file for dumping, suitable for gnuplot.
4190 Set scale for dumpfile. Acceptable values are same with scale option.
4194 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4195 Default is disabled.
4198 Enable minimum phase impulse response. Default is disabled.
4201 @subsection Examples
4206 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4209 lowpass at 1000 Hz with gain_entry:
4211 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4214 custom equalization:
4216 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4219 higher delay with zero phase to compensate delay:
4221 firequalizer=delay=0.1:fixed=on:zero_phase=on
4224 lowpass on left channel, highpass on right channel:
4226 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4227 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4232 Apply a flanging effect to the audio.
4234 The filter accepts the following options:
4238 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4241 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4244 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4248 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4249 Default value is 71.
4252 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4255 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4256 Default value is @var{sinusoidal}.
4259 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4260 Default value is 25.
4263 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4264 Default is @var{linear}.
4268 Apply Haas effect to audio.
4270 Note that this makes most sense to apply on mono signals.
4271 With this filter applied to mono signals it give some directionality and
4272 stretches its stereo image.
4274 The filter accepts the following options:
4278 Set input level. By default is @var{1}, or 0dB
4281 Set output level. By default is @var{1}, or 0dB.
4284 Set gain applied to side part of signal. By default is @var{1}.
4287 Set kind of middle source. Can be one of the following:
4297 Pick middle part signal of stereo image.
4300 Pick side part signal of stereo image.
4304 Change middle phase. By default is disabled.
4307 Set left channel delay. By default is @var{2.05} milliseconds.
4310 Set left channel balance. By default is @var{-1}.
4313 Set left channel gain. By default is @var{1}.
4316 Change left phase. By default is disabled.
4319 Set right channel delay. By defaults is @var{2.12} milliseconds.
4322 Set right channel balance. By default is @var{1}.
4325 Set right channel gain. By default is @var{1}.
4328 Change right phase. By default is enabled.
4333 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4334 embedded HDCD codes is expanded into a 20-bit PCM stream.
4336 The filter supports the Peak Extend and Low-level Gain Adjustment features
4337 of HDCD, and detects the Transient Filter flag.
4340 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4343 When using the filter with wav, note the default encoding for wav is 16-bit,
4344 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4345 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4347 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4348 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4351 The filter accepts the following options:
4354 @item disable_autoconvert
4355 Disable any automatic format conversion or resampling in the filter graph.
4357 @item process_stereo
4358 Process the stereo channels together. If target_gain does not match between
4359 channels, consider it invalid and use the last valid target_gain.
4362 Set the code detect timer period in ms.
4365 Always extend peaks above -3dBFS even if PE isn't signaled.
4368 Replace audio with a solid tone and adjust the amplitude to signal some
4369 specific aspect of the decoding process. The output file can be loaded in
4370 an audio editor alongside the original to aid analysis.
4372 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4379 Gain adjustment level at each sample
4381 Samples where peak extend occurs
4383 Samples where the code detect timer is active
4385 Samples where the target gain does not match between channels
4391 Apply head-related transfer functions (HRTFs) to create virtual
4392 loudspeakers around the user for binaural listening via headphones.
4393 The HRIRs are provided via additional streams, for each channel
4394 one stereo input stream is needed.
4396 The filter accepts the following options:
4400 Set mapping of input streams for convolution.
4401 The argument is a '|'-separated list of channel names in order as they
4402 are given as additional stream inputs for filter.
4403 This also specify number of input streams. Number of input streams
4404 must be not less than number of channels in first stream plus one.
4407 Set gain applied to audio. Value is in dB. Default is 0.
4410 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4411 processing audio in time domain which is slow.
4412 @var{freq} is processing audio in frequency domain which is fast.
4413 Default is @var{freq}.
4416 Set custom gain for LFE channels. Value is in dB. Default is 0.
4419 Set size of frame in number of samples which will be processed at once.
4420 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4423 Set format of hrir stream.
4424 Default value is @var{stereo}. Alternative value is @var{multich}.
4425 If value is set to @var{stereo}, number of additional streams should
4426 be greater or equal to number of input channels in first input stream.
4427 Also each additional stream should have stereo number of channels.
4428 If value is set to @var{multich}, number of additional streams should
4429 be exactly one. Also number of input channels of additional stream
4430 should be equal or greater than twice number of channels of first input
4434 @subsection Examples
4438 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4439 each amovie filter use stereo file with IR coefficients as input.
4440 The files give coefficients for each position of virtual loudspeaker:
4443 -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"
4448 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4449 but now in @var{multich} @var{hrir} format.
4451 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"
4458 Apply a high-pass filter with 3dB point frequency.
4459 The filter can be either single-pole, or double-pole (the default).
4460 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4462 The filter accepts the following options:
4466 Set frequency in Hz. Default is 3000.
4469 Set number of poles. Default is 2.
4472 Set method to specify band-width of filter.
4487 Specify the band-width of a filter in width_type units.
4488 Applies only to double-pole filter.
4489 The default is 0.707q and gives a Butterworth response.
4492 How much to use filtered signal in output. Default is 1.
4493 Range is between 0 and 1.
4496 Specify which channels to filter, by default all available are filtered.
4499 Normalize biquad coefficients, by default is disabled.
4500 Enabling it will normalize magnitude response at DC to 0dB.
4503 Set transform type of IIR filter.
4512 Set precison of filtering.
4515 Pick automatic sample format depending on surround filters.
4517 Always use signed 16-bit.
4519 Always use signed 32-bit.
4521 Always use float 32-bit.
4523 Always use float 64-bit.
4527 @subsection Commands
4529 This filter supports the following commands:
4532 Change highpass frequency.
4533 Syntax for the command is : "@var{frequency}"
4536 Change highpass width_type.
4537 Syntax for the command is : "@var{width_type}"
4540 Change highpass width.
4541 Syntax for the command is : "@var{width}"
4544 Change highpass mix.
4545 Syntax for the command is : "@var{mix}"
4550 Join multiple input streams into one multi-channel stream.
4552 It accepts the following parameters:
4556 The number of input streams. It defaults to 2.
4558 @item channel_layout
4559 The desired output channel layout. It defaults to stereo.
4562 Map channels from inputs to output. The argument is a '|'-separated list of
4563 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4564 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4565 can be either the name of the input channel (e.g. FL for front left) or its
4566 index in the specified input stream. @var{out_channel} is the name of the output
4570 The filter will attempt to guess the mappings when they are not specified
4571 explicitly. It does so by first trying to find an unused matching input channel
4572 and if that fails it picks the first unused input channel.
4574 Join 3 inputs (with properly set channel layouts):
4576 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4579 Build a 5.1 output from 6 single-channel streams:
4581 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4582 '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'
4588 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4590 To enable compilation of this filter you need to configure FFmpeg with
4591 @code{--enable-ladspa}.
4595 Specifies the name of LADSPA plugin library to load. If the environment
4596 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4597 each one of the directories specified by the colon separated list in
4598 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4599 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4600 @file{/usr/lib/ladspa/}.
4603 Specifies the plugin within the library. Some libraries contain only
4604 one plugin, but others contain many of them. If this is not set filter
4605 will list all available plugins within the specified library.
4608 Set the '|' separated list of controls which are zero or more floating point
4609 values that determine the behavior of the loaded plugin (for example delay,
4611 Controls need to be defined using the following syntax:
4612 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4613 @var{valuei} is the value set on the @var{i}-th control.
4614 Alternatively they can be also defined using the following syntax:
4615 @var{value0}|@var{value1}|@var{value2}|..., where
4616 @var{valuei} is the value set on the @var{i}-th control.
4617 If @option{controls} is set to @code{help}, all available controls and
4618 their valid ranges are printed.
4620 @item sample_rate, s
4621 Specify the sample rate, default to 44100. Only used if plugin have
4625 Set the number of samples per channel per each output frame, default
4626 is 1024. Only used if plugin have zero inputs.
4629 Set the minimum duration of the sourced audio. See
4630 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4631 for the accepted syntax.
4632 Note that the resulting duration may be greater than the specified duration,
4633 as the generated audio is always cut at the end of a complete frame.
4634 If not specified, or the expressed duration is negative, the audio is
4635 supposed to be generated forever.
4636 Only used if plugin have zero inputs.
4639 Enable latency compensation, by default is disabled.
4640 Only used if plugin have inputs.
4643 @subsection Examples
4647 List all available plugins within amp (LADSPA example plugin) library:
4653 List all available controls and their valid ranges for @code{vcf_notch}
4654 plugin from @code{VCF} library:
4656 ladspa=f=vcf:p=vcf_notch:c=help
4660 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4663 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4667 Add reverberation to the audio using TAP-plugins
4668 (Tom's Audio Processing plugins):
4670 ladspa=file=tap_reverb:tap_reverb
4674 Generate white noise, with 0.2 amplitude:
4676 ladspa=file=cmt:noise_source_white:c=c0=.2
4680 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4681 @code{C* Audio Plugin Suite} (CAPS) library:
4683 ladspa=file=caps:Click:c=c1=20'
4687 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4689 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4693 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4694 @code{SWH Plugins} collection:
4696 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4700 Attenuate low frequencies using Multiband EQ from Steve Harris
4701 @code{SWH Plugins} collection:
4703 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4707 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4710 ladspa=caps:Narrower
4714 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4716 ladspa=caps:White:.2
4720 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4722 ladspa=caps:Fractal:c=c1=1
4726 Dynamic volume normalization using @code{VLevel} plugin:
4728 ladspa=vlevel-ladspa:vlevel_mono
4732 @subsection Commands
4734 This filter supports the following commands:
4737 Modify the @var{N}-th control value.
4739 If the specified value is not valid, it is ignored and prior one is kept.
4744 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4745 Support for both single pass (livestreams, files) and double pass (files) modes.
4746 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4747 detect true peaks, the audio stream will be upsampled to 192 kHz.
4748 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4750 The filter accepts the following options:
4754 Set integrated loudness target.
4755 Range is -70.0 - -5.0. Default value is -24.0.
4758 Set loudness range target.
4759 Range is 1.0 - 20.0. Default value is 7.0.
4762 Set maximum true peak.
4763 Range is -9.0 - +0.0. Default value is -2.0.
4765 @item measured_I, measured_i
4766 Measured IL of input file.
4767 Range is -99.0 - +0.0.
4769 @item measured_LRA, measured_lra
4770 Measured LRA of input file.
4771 Range is 0.0 - 99.0.
4773 @item measured_TP, measured_tp
4774 Measured true peak of input file.
4775 Range is -99.0 - +99.0.
4777 @item measured_thresh
4778 Measured threshold of input file.
4779 Range is -99.0 - +0.0.
4782 Set offset gain. Gain is applied before the true-peak limiter.
4783 Range is -99.0 - +99.0. Default is +0.0.
4786 Normalize by linearly scaling the source audio.
4787 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4788 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4789 be lower than source LRA and the change in integrated loudness shouldn't
4790 result in a true peak which exceeds the target TP. If any of these
4791 conditions aren't met, normalization mode will revert to @var{dynamic}.
4792 Options are @code{true} or @code{false}. Default is @code{true}.
4795 Treat mono input files as "dual-mono". If a mono file is intended for playback
4796 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4797 If set to @code{true}, this option will compensate for this effect.
4798 Multi-channel input files are not affected by this option.
4799 Options are true or false. Default is false.
4802 Set print format for stats. Options are summary, json, or none.
4803 Default value is none.
4808 Apply a low-pass filter with 3dB point frequency.
4809 The filter can be either single-pole or double-pole (the default).
4810 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4812 The filter accepts the following options:
4816 Set frequency in Hz. Default is 500.
4819 Set number of poles. Default is 2.
4822 Set method to specify band-width of filter.
4837 Specify the band-width of a filter in width_type units.
4838 Applies only to double-pole filter.
4839 The default is 0.707q and gives a Butterworth response.
4842 How much to use filtered signal in output. Default is 1.
4843 Range is between 0 and 1.
4846 Specify which channels to filter, by default all available are filtered.
4849 Normalize biquad coefficients, by default is disabled.
4850 Enabling it will normalize magnitude response at DC to 0dB.
4853 Set transform type of IIR filter.
4862 Set precison of filtering.
4865 Pick automatic sample format depending on surround filters.
4867 Always use signed 16-bit.
4869 Always use signed 32-bit.
4871 Always use float 32-bit.
4873 Always use float 64-bit.
4877 @subsection Examples
4880 Lowpass only LFE channel, it LFE is not present it does nothing:
4886 @subsection Commands
4888 This filter supports the following commands:
4891 Change lowpass frequency.
4892 Syntax for the command is : "@var{frequency}"
4895 Change lowpass width_type.
4896 Syntax for the command is : "@var{width_type}"
4899 Change lowpass width.
4900 Syntax for the command is : "@var{width}"
4904 Syntax for the command is : "@var{mix}"
4909 Load a LV2 (LADSPA Version 2) plugin.
4911 To enable compilation of this filter you need to configure FFmpeg with
4912 @code{--enable-lv2}.
4916 Specifies the plugin URI. You may need to escape ':'.
4919 Set the '|' separated list of controls which are zero or more floating point
4920 values that determine the behavior of the loaded plugin (for example delay,
4922 If @option{controls} is set to @code{help}, all available controls and
4923 their valid ranges are printed.
4925 @item sample_rate, s
4926 Specify the sample rate, default to 44100. Only used if plugin have
4930 Set the number of samples per channel per each output frame, default
4931 is 1024. Only used if plugin have zero inputs.
4934 Set the minimum duration of the sourced audio. See
4935 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4936 for the accepted syntax.
4937 Note that the resulting duration may be greater than the specified duration,
4938 as the generated audio is always cut at the end of a complete frame.
4939 If not specified, or the expressed duration is negative, the audio is
4940 supposed to be generated forever.
4941 Only used if plugin have zero inputs.
4944 @subsection Examples
4948 Apply bass enhancer plugin from Calf:
4950 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4954 Apply vinyl plugin from Calf:
4956 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4960 Apply bit crusher plugin from ArtyFX:
4962 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4967 Multiband Compress or expand the audio's dynamic range.
4969 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4970 This is akin to the crossover of a loudspeaker, and results in flat frequency
4971 response when absent compander action.
4973 It accepts the following parameters:
4977 This option syntax is:
4978 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4979 For explanation of each item refer to compand filter documentation.
4985 Mix channels with specific gain levels. The filter accepts the output
4986 channel layout followed by a set of channels definitions.
4988 This filter is also designed to efficiently remap the channels of an audio
4991 The filter accepts parameters of the form:
4992 "@var{l}|@var{outdef}|@var{outdef}|..."
4996 output channel layout or number of channels
4999 output channel specification, of the form:
5000 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
5003 output channel to define, either a channel name (FL, FR, etc.) or a channel
5004 number (c0, c1, etc.)
5007 multiplicative coefficient for the channel, 1 leaving the volume unchanged
5010 input channel to use, see out_name for details; it is not possible to mix
5011 named and numbered input channels
5014 If the `=' in a channel specification is replaced by `<', then the gains for
5015 that specification will be renormalized so that the total is 1, thus
5016 avoiding clipping noise.
5018 @subsection Mixing examples
5020 For example, if you want to down-mix from stereo to mono, but with a bigger
5021 factor for the left channel:
5023 pan=1c|c0=0.9*c0+0.1*c1
5026 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5027 7-channels surround:
5029 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5032 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5033 that should be preferred (see "-ac" option) unless you have very specific
5036 @subsection Remapping examples
5038 The channel remapping will be effective if, and only if:
5041 @item gain coefficients are zeroes or ones,
5042 @item only one input per channel output,
5045 If all these conditions are satisfied, the filter will notify the user ("Pure
5046 channel mapping detected"), and use an optimized and lossless method to do the
5049 For example, if you have a 5.1 source and want a stereo audio stream by
5050 dropping the extra channels:
5052 pan="stereo| c0=FL | c1=FR"
5055 Given the same source, you can also switch front left and front right channels
5056 and keep the input channel layout:
5058 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5061 If the input is a stereo audio stream, you can mute the front left channel (and
5062 still keep the stereo channel layout) with:
5067 Still with a stereo audio stream input, you can copy the right channel in both
5068 front left and right:
5070 pan="stereo| c0=FR | c1=FR"
5075 ReplayGain scanner filter. This filter takes an audio stream as an input and
5076 outputs it unchanged.
5077 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5081 Convert the audio sample format, sample rate and channel layout. It is
5082 not meant to be used directly.
5085 Apply time-stretching and pitch-shifting with librubberband.
5087 To enable compilation of this filter, you need to configure FFmpeg with
5088 @code{--enable-librubberband}.
5090 The filter accepts the following options:
5094 Set tempo scale factor.
5097 Set pitch scale factor.
5100 Set transients detector.
5101 Possible values are:
5110 Possible values are:
5119 Possible values are:
5126 Set processing window size.
5127 Possible values are:
5136 Possible values are:
5143 Enable formant preservation when shift pitching.
5144 Possible values are:
5152 Possible values are:
5161 Possible values are:
5168 @subsection Commands
5170 This filter supports the following commands:
5173 Change filter tempo scale factor.
5174 Syntax for the command is : "@var{tempo}"
5177 Change filter pitch scale factor.
5178 Syntax for the command is : "@var{pitch}"
5181 @section sidechaincompress
5183 This filter acts like normal compressor but has the ability to compress
5184 detected signal using second input signal.
5185 It needs two input streams and returns one output stream.
5186 First input stream will be processed depending on second stream signal.
5187 The filtered signal then can be filtered with other filters in later stages of
5188 processing. See @ref{pan} and @ref{amerge} filter.
5190 The filter accepts the following options:
5194 Set input gain. Default is 1. Range is between 0.015625 and 64.
5197 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5198 Default is @code{downward}.
5201 If a signal of second stream raises above this level it will affect the gain
5202 reduction of first stream.
5203 By default is 0.125. Range is between 0.00097563 and 1.
5206 Set a ratio about which the signal is reduced. 1:2 means that if the level
5207 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5208 Default is 2. Range is between 1 and 20.
5211 Amount of milliseconds the signal has to rise above the threshold before gain
5212 reduction starts. Default is 20. Range is between 0.01 and 2000.
5215 Amount of milliseconds the signal has to fall below the threshold before
5216 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5219 Set the amount by how much signal will be amplified after processing.
5220 Default is 1. Range is from 1 to 64.
5223 Curve the sharp knee around the threshold to enter gain reduction more softly.
5224 Default is 2.82843. Range is between 1 and 8.
5227 Choose if the @code{average} level between all channels of side-chain stream
5228 or the louder(@code{maximum}) channel of side-chain stream affects the
5229 reduction. Default is @code{average}.
5232 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5233 of @code{rms}. Default is @code{rms} which is mainly smoother.
5236 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5239 How much to use compressed signal in output. Default is 1.
5240 Range is between 0 and 1.
5243 @subsection Commands
5245 This filter supports the all above options as @ref{commands}.
5247 @subsection Examples
5251 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5252 depending on the signal of 2nd input and later compressed signal to be
5253 merged with 2nd input:
5255 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5259 @section sidechaingate
5261 A sidechain gate acts like a normal (wideband) gate but has the ability to
5262 filter the detected signal before sending it to the gain reduction stage.
5263 Normally a gate uses the full range signal to detect a level above the
5265 For example: If you cut all lower frequencies from your sidechain signal
5266 the gate will decrease the volume of your track only if not enough highs
5267 appear. With this technique you are able to reduce the resonation of a
5268 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5270 It needs two input streams and returns one output stream.
5271 First input stream will be processed depending on second stream signal.
5273 The filter accepts the following options:
5277 Set input level before filtering.
5278 Default is 1. Allowed range is from 0.015625 to 64.
5281 Set the mode of operation. Can be @code{upward} or @code{downward}.
5282 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5283 will be amplified, expanding dynamic range in upward direction.
5284 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5287 Set the level of gain reduction when the signal is below the threshold.
5288 Default is 0.06125. Allowed range is from 0 to 1.
5289 Setting this to 0 disables reduction and then filter behaves like expander.
5292 If a signal rises above this level the gain reduction is released.
5293 Default is 0.125. Allowed range is from 0 to 1.
5296 Set a ratio about which the signal is reduced.
5297 Default is 2. Allowed range is from 1 to 9000.
5300 Amount of milliseconds the signal has to rise above the threshold before gain
5302 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5305 Amount of milliseconds the signal has to fall below the threshold before the
5306 reduction is increased again. Default is 250 milliseconds.
5307 Allowed range is from 0.01 to 9000.
5310 Set amount of amplification of signal after processing.
5311 Default is 1. Allowed range is from 1 to 64.
5314 Curve the sharp knee around the threshold to enter gain reduction more softly.
5315 Default is 2.828427125. Allowed range is from 1 to 8.
5318 Choose if exact signal should be taken for detection or an RMS like one.
5319 Default is rms. Can be peak or rms.
5322 Choose if the average level between all channels or the louder channel affects
5324 Default is average. Can be average or maximum.
5327 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5330 @subsection Commands
5332 This filter supports the all above options as @ref{commands}.
5334 @section silencedetect
5336 Detect silence in an audio stream.
5338 This filter logs a message when it detects that the input audio volume is less
5339 or equal to a noise tolerance value for a duration greater or equal to the
5340 minimum detected noise duration.
5342 The printed times and duration are expressed in seconds. The
5343 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5344 is set on the first frame whose timestamp equals or exceeds the detection
5345 duration and it contains the timestamp of the first frame of the silence.
5347 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5348 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5349 keys are set on the first frame after the silence. If @option{mono} is
5350 enabled, and each channel is evaluated separately, the @code{.X}
5351 suffixed keys are used, and @code{X} corresponds to the channel number.
5353 The filter accepts the following options:
5357 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5358 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5361 Set silence duration until notification (default is 2 seconds). See
5362 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5363 for the accepted syntax.
5366 Process each channel separately, instead of combined. By default is disabled.
5369 @subsection Examples
5373 Detect 5 seconds of silence with -50dB noise tolerance:
5375 silencedetect=n=-50dB:d=5
5379 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5380 tolerance in @file{silence.mp3}:
5382 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5386 @section silenceremove
5388 Remove silence from the beginning, middle or end of the audio.
5390 The filter accepts the following options:
5394 This value is used to indicate if audio should be trimmed at beginning of
5395 the audio. A value of zero indicates no silence should be trimmed from the
5396 beginning. When specifying a non-zero value, it trims audio up until it
5397 finds non-silence. Normally, when trimming silence from beginning of audio
5398 the @var{start_periods} will be @code{1} but it can be increased to higher
5399 values to trim all audio up to specific count of non-silence periods.
5400 Default value is @code{0}.
5402 @item start_duration
5403 Specify the amount of time that non-silence must be detected before it stops
5404 trimming audio. By increasing the duration, bursts of noises can be treated
5405 as silence and trimmed off. Default value is @code{0}.
5407 @item start_threshold
5408 This indicates what sample value should be treated as silence. For digital
5409 audio, a value of @code{0} may be fine but for audio recorded from analog,
5410 you may wish to increase the value to account for background noise.
5411 Can be specified in dB (in case "dB" is appended to the specified value)
5412 or amplitude ratio. Default value is @code{0}.
5415 Specify max duration of silence at beginning that will be kept after
5416 trimming. Default is 0, which is equal to trimming all samples detected
5420 Specify mode of detection of silence end in start of multi-channel audio.
5421 Can be @var{any} or @var{all}. Default is @var{any}.
5422 With @var{any}, any sample that is detected as non-silence will cause
5423 stopped trimming of silence.
5424 With @var{all}, only if all channels are detected as non-silence will cause
5425 stopped trimming of silence.
5428 Set the count for trimming silence from the end of audio.
5429 To remove silence from the middle of a file, specify a @var{stop_periods}
5430 that is negative. This value is then treated as a positive value and is
5431 used to indicate the effect should restart processing as specified by
5432 @var{start_periods}, making it suitable for removing periods of silence
5433 in the middle of the audio.
5434 Default value is @code{0}.
5437 Specify a duration of silence that must exist before audio is not copied any
5438 more. By specifying a higher duration, silence that is wanted can be left in
5440 Default value is @code{0}.
5442 @item stop_threshold
5443 This is the same as @option{start_threshold} but for trimming silence from
5445 Can be specified in dB (in case "dB" is appended to the specified value)
5446 or amplitude ratio. Default value is @code{0}.
5449 Specify max duration of silence at end that will be kept after
5450 trimming. Default is 0, which is equal to trimming all samples detected
5454 Specify mode of detection of silence start in end of multi-channel audio.
5455 Can be @var{any} or @var{all}. Default is @var{any}.
5456 With @var{any}, any sample that is detected as non-silence will cause
5457 stopped trimming of silence.
5458 With @var{all}, only if all channels are detected as non-silence will cause
5459 stopped trimming of silence.
5462 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5463 and works better with digital silence which is exactly 0.
5464 Default value is @code{rms}.
5467 Set duration in number of seconds used to calculate size of window in number
5468 of samples for detecting silence.
5469 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5472 @subsection Examples
5476 The following example shows how this filter can be used to start a recording
5477 that does not contain the delay at the start which usually occurs between
5478 pressing the record button and the start of the performance:
5480 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5484 Trim all silence encountered from beginning to end where there is more than 1
5485 second of silence in audio:
5487 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5491 Trim all digital silence samples, using peak detection, from beginning to end
5492 where there is more than 0 samples of digital silence in audio and digital
5493 silence is detected in all channels at same positions in stream:
5495 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5501 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5502 loudspeakers around the user for binaural listening via headphones (audio
5503 formats up to 9 channels supported).
5504 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5505 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5506 Austrian Academy of Sciences.
5508 To enable compilation of this filter you need to configure FFmpeg with
5509 @code{--enable-libmysofa}.
5511 The filter accepts the following options:
5515 Set the SOFA file used for rendering.
5518 Set gain applied to audio. Value is in dB. Default is 0.
5521 Set rotation of virtual loudspeakers in deg. Default is 0.
5524 Set elevation of virtual speakers in deg. Default is 0.
5527 Set distance in meters between loudspeakers and the listener with near-field
5528 HRTFs. Default is 1.
5531 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5532 processing audio in time domain which is slow.
5533 @var{freq} is processing audio in frequency domain which is fast.
5534 Default is @var{freq}.
5537 Set custom positions of virtual loudspeakers. Syntax for this option is:
5538 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5539 Each virtual loudspeaker is described with short channel name following with
5540 azimuth and elevation in degrees.
5541 Each virtual loudspeaker description is separated by '|'.
5542 For example to override front left and front right channel positions use:
5543 'speakers=FL 45 15|FR 345 15'.
5544 Descriptions with unrecognised channel names are ignored.
5547 Set custom gain for LFE channels. Value is in dB. Default is 0.
5550 Set custom frame size in number of samples. Default is 1024.
5551 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5552 is set to @var{freq}.
5555 Should all IRs be normalized upon importing SOFA file.
5556 By default is enabled.
5559 Should nearest IRs be interpolated with neighbor IRs if exact position
5560 does not match. By default is disabled.
5563 Minphase all IRs upon loading of SOFA file. By default is disabled.
5566 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5569 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5572 @subsection Examples
5576 Using ClubFritz6 sofa file:
5578 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5582 Using ClubFritz12 sofa file and bigger radius with small rotation:
5584 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5588 Similar as above but with custom speaker positions for front left, front right, back left and back right
5589 and also with custom gain:
5591 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5598 This filter expands or compresses each half-cycle of audio samples
5599 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5600 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5602 The filter accepts the following options:
5606 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5607 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5610 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5611 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5612 would be such that local peak value reaches target peak value but never to surpass it and that
5613 ratio between new and previous peak value does not surpass this option value.
5615 @item compression, c
5616 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5617 This option controls maximum local half-cycle of samples compression. This option is used
5618 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5619 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5620 that peak's half-cycle will be compressed by current compression factor.
5623 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5624 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5625 Any half-cycle samples with their local peak value below or same as this option value will be
5626 compressed by current compression factor, otherwise, if greater than threshold value they will be
5627 expanded with expansion factor so that it could reach peak target value but never surpass it.
5630 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5631 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5632 each new half-cycle until it reaches @option{expansion} value.
5633 Setting this options too high may lead to distortions.
5636 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5637 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5638 each new half-cycle until it reaches @option{compression} value.
5641 Specify which channels to filter, by default all available channels are filtered.
5644 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5645 option. When enabled any half-cycle of samples with their local peak value below or same as
5646 @option{threshold} option will be expanded otherwise it will be compressed.
5649 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5650 When disabled each filtered channel gain calculation is independent, otherwise when this option
5651 is enabled the minimum of all possible gains for each filtered channel is used.
5654 @subsection Commands
5656 This filter supports the all above options as @ref{commands}.
5658 @section stereotools
5660 This filter has some handy utilities to manage stereo signals, for converting
5661 M/S stereo recordings to L/R signal while having control over the parameters
5662 or spreading the stereo image of master track.
5664 The filter accepts the following options:
5668 Set input level before filtering for both channels. Defaults is 1.
5669 Allowed range is from 0.015625 to 64.
5672 Set output level after filtering for both channels. Defaults is 1.
5673 Allowed range is from 0.015625 to 64.
5676 Set input balance between both channels. Default is 0.
5677 Allowed range is from -1 to 1.
5680 Set output balance between both channels. Default is 0.
5681 Allowed range is from -1 to 1.
5684 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5685 clipping. Disabled by default.
5688 Mute the left channel. Disabled by default.
5691 Mute the right channel. Disabled by default.
5694 Change the phase of the left channel. Disabled by default.
5697 Change the phase of the right channel. Disabled by default.
5700 Set stereo mode. Available values are:
5704 Left/Right to Left/Right, this is default.
5707 Left/Right to Mid/Side.
5710 Mid/Side to Left/Right.
5713 Left/Right to Left/Left.
5716 Left/Right to Right/Right.
5719 Left/Right to Left + Right.
5722 Left/Right to Right/Left.
5725 Mid/Side to Left/Left.
5728 Mid/Side to Right/Right.
5731 Mid/Side to Right/Left.
5734 Left/Right to Left - Right.
5738 Set level of side signal. Default is 1.
5739 Allowed range is from 0.015625 to 64.
5742 Set balance of side signal. Default is 0.
5743 Allowed range is from -1 to 1.
5746 Set level of the middle signal. Default is 1.
5747 Allowed range is from 0.015625 to 64.
5750 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5753 Set stereo base between mono and inversed channels. Default is 0.
5754 Allowed range is from -1 to 1.
5757 Set delay in milliseconds how much to delay left from right channel and
5758 vice versa. Default is 0. Allowed range is from -20 to 20.
5761 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5764 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5766 @item bmode_in, bmode_out
5767 Set balance mode for balance_in/balance_out option.
5769 Can be one of the following:
5773 Classic balance mode. Attenuate one channel at time.
5774 Gain is raised up to 1.
5777 Similar as classic mode above but gain is raised up to 2.
5780 Equal power distribution, from -6dB to +6dB range.
5784 @subsection Commands
5786 This filter supports the all above options as @ref{commands}.
5788 @subsection Examples
5792 Apply karaoke like effect:
5794 stereotools=mlev=0.015625
5798 Convert M/S signal to L/R:
5800 "stereotools=mode=ms>lr"
5804 @section stereowiden
5806 This filter enhance the stereo effect by suppressing signal common to both
5807 channels and by delaying the signal of left into right and vice versa,
5808 thereby widening the stereo effect.
5810 The filter accepts the following options:
5814 Time in milliseconds of the delay of left signal into right and vice versa.
5815 Default is 20 milliseconds.
5818 Amount of gain in delayed signal into right and vice versa. Gives a delay
5819 effect of left signal in right output and vice versa which gives widening
5820 effect. Default is 0.3.
5823 Cross feed of left into right with inverted phase. This helps in suppressing
5824 the mono. If the value is 1 it will cancel all the signal common to both
5825 channels. Default is 0.3.
5828 Set level of input signal of original channel. Default is 0.8.
5831 @subsection Commands
5833 This filter supports the all above options except @code{delay} as @ref{commands}.
5835 @section superequalizer
5836 Apply 18 band equalizer.
5838 The filter accepts the following options:
5845 Set 131Hz band gain.
5847 Set 185Hz band gain.
5849 Set 262Hz band gain.
5851 Set 370Hz band gain.
5853 Set 523Hz band gain.
5855 Set 740Hz band gain.
5857 Set 1047Hz band gain.
5859 Set 1480Hz band gain.
5861 Set 2093Hz band gain.
5863 Set 2960Hz band gain.
5865 Set 4186Hz band gain.
5867 Set 5920Hz band gain.
5869 Set 8372Hz band gain.
5871 Set 11840Hz band gain.
5873 Set 16744Hz band gain.
5875 Set 20000Hz band gain.
5879 Apply audio surround upmix filter.
5881 This filter allows to produce multichannel output from audio stream.
5883 The filter accepts the following options:
5887 Set output channel layout. By default, this is @var{5.1}.
5889 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5890 for the required syntax.
5893 Set input channel layout. By default, this is @var{stereo}.
5895 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5896 for the required syntax.
5899 Set input volume level. By default, this is @var{1}.
5902 Set output volume level. By default, this is @var{1}.
5905 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5908 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5911 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5914 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5915 In @var{add} mode, LFE channel is created from input audio and added to output.
5916 In @var{sub} mode, LFE channel is created from input audio and added to output but
5917 also all non-LFE output channels are subtracted with output LFE channel.
5920 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5921 Default is @var{90}.
5924 Set front center input volume. By default, this is @var{1}.
5927 Set front center output volume. By default, this is @var{1}.
5930 Set front left input volume. By default, this is @var{1}.
5933 Set front left output volume. By default, this is @var{1}.
5936 Set front right input volume. By default, this is @var{1}.
5939 Set front right output volume. By default, this is @var{1}.
5942 Set side left input volume. By default, this is @var{1}.
5945 Set side left output volume. By default, this is @var{1}.
5948 Set side right input volume. By default, this is @var{1}.
5951 Set side right output volume. By default, this is @var{1}.
5954 Set back left input volume. By default, this is @var{1}.
5957 Set back left output volume. By default, this is @var{1}.
5960 Set back right input volume. By default, this is @var{1}.
5963 Set back right output volume. By default, this is @var{1}.
5966 Set back center input volume. By default, this is @var{1}.
5969 Set back center output volume. By default, this is @var{1}.
5972 Set LFE input volume. By default, this is @var{1}.
5975 Set LFE output volume. By default, this is @var{1}.
5978 Set spread usage of stereo image across X axis for all channels.
5981 Set spread usage of stereo image across Y axis for all channels.
5983 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5984 Set spread usage of stereo image across X axis for each channel.
5986 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5987 Set spread usage of stereo image across Y axis for each channel.
5990 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5993 Set window function.
5995 It accepts the following values:
6018 Default is @code{hann}.
6021 Set window overlap. If set to 1, the recommended overlap for selected
6022 window function will be picked. Default is @code{0.5}.
6025 @section treble, highshelf
6027 Boost or cut treble (upper) frequencies of the audio using a two-pole
6028 shelving filter with a response similar to that of a standard
6029 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6031 The filter accepts the following options:
6035 Give the gain at whichever is the lower of ~22 kHz and the
6036 Nyquist frequency. Its useful range is about -20 (for a large cut)
6037 to +20 (for a large boost). Beware of clipping when using a positive gain.
6040 Set the filter's central frequency and so can be used
6041 to extend or reduce the frequency range to be boosted or cut.
6042 The default value is @code{3000} Hz.
6045 Set method to specify band-width of filter.
6060 Determine how steep is the filter's shelf transition.
6063 Set number of poles. Default is 2.
6066 How much to use filtered signal in output. Default is 1.
6067 Range is between 0 and 1.
6070 Specify which channels to filter, by default all available are filtered.
6073 Normalize biquad coefficients, by default is disabled.
6074 Enabling it will normalize magnitude response at DC to 0dB.
6077 Set transform type of IIR filter.
6086 Set precison of filtering.
6089 Pick automatic sample format depending on surround filters.
6091 Always use signed 16-bit.
6093 Always use signed 32-bit.
6095 Always use float 32-bit.
6097 Always use float 64-bit.
6101 @subsection Commands
6103 This filter supports the following commands:
6106 Change treble frequency.
6107 Syntax for the command is : "@var{frequency}"
6110 Change treble width_type.
6111 Syntax for the command is : "@var{width_type}"
6114 Change treble width.
6115 Syntax for the command is : "@var{width}"
6119 Syntax for the command is : "@var{gain}"
6123 Syntax for the command is : "@var{mix}"
6128 Sinusoidal amplitude modulation.
6130 The filter accepts the following options:
6134 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6135 (20 Hz or lower) will result in a tremolo effect.
6136 This filter may also be used as a ring modulator by specifying
6137 a modulation frequency higher than 20 Hz.
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 Sinusoidal phase modulation.
6149 The filter accepts the following options:
6153 Modulation frequency in Hertz.
6154 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6157 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6158 Default value is 0.5.
6163 Adjust the input audio volume.
6165 It accepts the following parameters:
6169 Set audio volume expression.
6171 Output values are clipped to the maximum value.
6173 The output audio volume is given by the relation:
6175 @var{output_volume} = @var{volume} * @var{input_volume}
6178 The default value for @var{volume} is "1.0".
6181 This parameter represents the mathematical precision.
6183 It determines which input sample formats will be allowed, which affects the
6184 precision of the volume scaling.
6188 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6190 32-bit floating-point; this limits input sample format to FLT. (default)
6192 64-bit floating-point; this limits input sample format to DBL.
6196 Choose the behaviour on encountering ReplayGain side data in input frames.
6200 Remove ReplayGain side data, ignoring its contents (the default).
6203 Ignore ReplayGain side data, but leave it in the frame.
6206 Prefer the track gain, if present.
6209 Prefer the album gain, if present.
6212 @item replaygain_preamp
6213 Pre-amplification gain in dB to apply to the selected replaygain gain.
6215 Default value for @var{replaygain_preamp} is 0.0.
6217 @item replaygain_noclip
6218 Prevent clipping by limiting the gain applied.
6220 Default value for @var{replaygain_noclip} is 1.
6223 Set when the volume expression is evaluated.
6225 It accepts the following values:
6228 only evaluate expression once during the filter initialization, or
6229 when the @samp{volume} command is sent
6232 evaluate expression for each incoming frame
6235 Default value is @samp{once}.
6238 The volume expression can contain the following parameters.
6242 frame number (starting at zero)
6245 @item nb_consumed_samples
6246 number of samples consumed by the filter
6248 number of samples in the current frame
6250 original frame position in the file
6256 PTS at start of stream
6258 time at start of stream
6264 last set volume value
6267 Note that when @option{eval} is set to @samp{once} only the
6268 @var{sample_rate} and @var{tb} variables are available, all other
6269 variables will evaluate to NAN.
6271 @subsection Commands
6273 This filter supports the following commands:
6276 Modify the volume expression.
6277 The command accepts the same syntax of the corresponding option.
6279 If the specified expression is not valid, it is kept at its current
6283 @subsection Examples
6287 Halve the input audio volume:
6291 volume=volume=-6.0206dB
6294 In all the above example the named key for @option{volume} can be
6295 omitted, for example like in:
6301 Increase input audio power by 6 decibels using fixed-point precision:
6303 volume=volume=6dB:precision=fixed
6307 Fade volume after time 10 with an annihilation period of 5 seconds:
6309 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6313 @section volumedetect
6315 Detect the volume of the input video.
6317 The filter has no parameters. The input is not modified. Statistics about
6318 the volume will be printed in the log when the input stream end is reached.
6320 In particular it will show the mean volume (root mean square), maximum
6321 volume (on a per-sample basis), and the beginning of a histogram of the
6322 registered volume values (from the maximum value to a cumulated 1/1000 of
6325 All volumes are in decibels relative to the maximum PCM value.
6327 @subsection Examples
6329 Here is an excerpt of the output:
6331 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6332 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6333 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6334 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6335 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6336 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6337 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6338 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6339 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6345 The mean square energy is approximately -27 dB, or 10^-2.7.
6347 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6349 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6352 In other words, raising the volume by +4 dB does not cause any clipping,
6353 raising it by +5 dB causes clipping for 6 samples, etc.
6355 @c man end AUDIO FILTERS
6357 @chapter Audio Sources
6358 @c man begin AUDIO SOURCES
6360 Below is a description of the currently available audio sources.
6364 Buffer audio frames, and make them available to the filter chain.
6366 This source is mainly intended for a programmatic use, in particular
6367 through the interface defined in @file{libavfilter/buffersrc.h}.
6369 It accepts the following parameters:
6373 The timebase which will be used for timestamps of submitted frames. It must be
6374 either a floating-point number or in @var{numerator}/@var{denominator} form.
6377 The sample rate of the incoming audio buffers.
6380 The sample format of the incoming audio buffers.
6381 Either a sample format name or its corresponding integer representation from
6382 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6384 @item channel_layout
6385 The channel layout of the incoming audio buffers.
6386 Either a channel layout name from channel_layout_map in
6387 @file{libavutil/channel_layout.c} or its corresponding integer representation
6388 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6391 The number of channels of the incoming audio buffers.
6392 If both @var{channels} and @var{channel_layout} are specified, then they
6397 @subsection Examples
6400 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6403 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6404 Since the sample format with name "s16p" corresponds to the number
6405 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6408 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6413 Generate an audio signal specified by an expression.
6415 This source accepts in input one or more expressions (one for each
6416 channel), which are evaluated and used to generate a corresponding
6419 This source accepts the following options:
6423 Set the '|'-separated expressions list for each separate channel. In case the
6424 @option{channel_layout} option is not specified, the selected channel layout
6425 depends on the number of provided expressions. Otherwise the last
6426 specified expression is applied to the remaining output channels.
6428 @item channel_layout, c
6429 Set the channel layout. The number of channels in the specified layout
6430 must be equal to the number of specified expressions.
6433 Set the minimum duration of the sourced audio. See
6434 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6435 for the accepted syntax.
6436 Note that the resulting duration may be greater than the specified
6437 duration, as the generated audio is always cut at the end of a
6440 If not specified, or the expressed duration is negative, the audio is
6441 supposed to be generated forever.
6444 Set the number of samples per channel per each output frame,
6447 @item sample_rate, s
6448 Specify the sample rate, default to 44100.
6451 Each expression in @var{exprs} can contain the following constants:
6455 number of the evaluated sample, starting from 0
6458 time of the evaluated sample expressed in seconds, starting from 0
6465 @subsection Examples
6475 Generate a sin signal with frequency of 440 Hz, set sample rate to
6478 aevalsrc="sin(440*2*PI*t):s=8000"
6482 Generate a two channels signal, specify the channel layout (Front
6483 Center + Back Center) explicitly:
6485 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6489 Generate white noise:
6491 aevalsrc="-2+random(0)"
6495 Generate an amplitude modulated signal:
6497 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6501 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6503 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6510 Generate a FIR coefficients using frequency sampling method.
6512 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6514 The filter accepts the following options:
6518 Set number of filter coefficents in output audio stream.
6519 Default value is 1025.
6522 Set frequency points from where magnitude and phase are set.
6523 This must be in non decreasing order, and first element must be 0, while last element
6524 must be 1. Elements are separated by white spaces.
6527 Set magnitude value for every frequency point set by @option{frequency}.
6528 Number of values must be same as number of frequency points.
6529 Values are separated by white spaces.
6532 Set phase value for every frequency point set by @option{frequency}.
6533 Number of values must be same as number of frequency points.
6534 Values are separated by white spaces.
6536 @item sample_rate, r
6537 Set sample rate, default is 44100.
6540 Set number of samples per each frame. Default is 1024.
6543 Set window function. Default is blackman.
6548 The null audio source, return unprocessed audio frames. It is mainly useful
6549 as a template and to be employed in analysis / debugging tools, or as
6550 the source for filters which ignore the input data (for example the sox
6553 This source accepts the following options:
6557 @item channel_layout, cl
6559 Specifies the channel layout, and can be either an integer or a string
6560 representing a channel layout. The default value of @var{channel_layout}
6563 Check the channel_layout_map definition in
6564 @file{libavutil/channel_layout.c} for the mapping between strings and
6565 channel layout values.
6567 @item sample_rate, r
6568 Specifies the sample rate, and defaults to 44100.
6571 Set the number of samples per requested frames.
6574 Set the duration of the sourced audio. See
6575 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6576 for the accepted syntax.
6578 If not specified, or the expressed duration is negative, the audio is
6579 supposed to be generated forever.
6582 @subsection Examples
6586 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6588 anullsrc=r=48000:cl=4
6592 Do the same operation with a more obvious syntax:
6594 anullsrc=r=48000:cl=mono
6598 All the parameters need to be explicitly defined.
6602 Synthesize a voice utterance using the libflite library.
6604 To enable compilation of this filter you need to configure FFmpeg with
6605 @code{--enable-libflite}.
6607 Note that versions of the flite library prior to 2.0 are not thread-safe.
6609 The filter accepts the following options:
6614 If set to 1, list the names of the available voices and exit
6615 immediately. Default value is 0.
6618 Set the maximum number of samples per frame. Default value is 512.
6621 Set the filename containing the text to speak.
6624 Set the text to speak.
6627 Set the voice to use for the speech synthesis. Default value is
6628 @code{kal}. See also the @var{list_voices} option.
6631 @subsection Examples
6635 Read from file @file{speech.txt}, and synthesize the text using the
6636 standard flite voice:
6638 flite=textfile=speech.txt
6642 Read the specified text selecting the @code{slt} voice:
6644 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6648 Input text to ffmpeg:
6650 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6654 Make @file{ffplay} speak the specified text, using @code{flite} and
6655 the @code{lavfi} device:
6657 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6661 For more information about libflite, check:
6662 @url{http://www.festvox.org/flite/}
6666 Generate a noise audio signal.
6668 The filter accepts the following options:
6671 @item sample_rate, r
6672 Specify the sample rate. Default value is 48000 Hz.
6675 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6679 Specify the duration of the generated audio stream. Not specifying this option
6680 results in noise with an infinite length.
6682 @item color, colour, c
6683 Specify the color of noise. Available noise colors are white, pink, brown,
6684 blue, violet and velvet. Default color is white.
6687 Specify a value used to seed the PRNG.
6690 Set the number of samples per each output frame, default is 1024.
6693 @subsection Examples
6698 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6700 anoisesrc=d=60:c=pink:r=44100:a=0.5
6706 Generate odd-tap Hilbert transform FIR coefficients.
6708 The resulting stream can be used with @ref{afir} filter for phase-shifting
6709 the signal by 90 degrees.
6711 This is used in many matrix coding schemes and for analytic signal generation.
6712 The process is often written as a multiplication by i (or j), the imaginary unit.
6714 The filter accepts the following options:
6718 @item sample_rate, s
6719 Set sample rate, default is 44100.
6722 Set length of FIR filter, default is 22051.
6725 Set number of samples per each frame.
6728 Set window function to be used when generating FIR coefficients.
6733 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6735 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6737 The filter accepts the following options:
6740 @item sample_rate, r
6741 Set sample rate, default is 44100.
6744 Set number of samples per each frame. Default is 1024.
6747 Set high-pass frequency. Default is 0.
6750 Set low-pass frequency. Default is 0.
6751 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6752 is higher than 0 then filter will create band-pass filter coefficients,
6753 otherwise band-reject filter coefficients.
6756 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6759 Set Kaiser window beta.
6762 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6765 Enable rounding, by default is disabled.
6768 Set number of taps for high-pass filter.
6771 Set number of taps for low-pass filter.
6776 Generate an audio signal made of a sine wave with amplitude 1/8.
6778 The audio signal is bit-exact.
6780 The filter accepts the following options:
6785 Set the carrier frequency. Default is 440 Hz.
6787 @item beep_factor, b
6788 Enable a periodic beep every second with frequency @var{beep_factor} times
6789 the carrier frequency. Default is 0, meaning the beep is disabled.
6791 @item sample_rate, r
6792 Specify the sample rate, default is 44100.
6795 Specify the duration of the generated audio stream.
6797 @item samples_per_frame
6798 Set the number of samples per output frame.
6800 The expression can contain the following constants:
6804 The (sequential) number of the output audio frame, starting from 0.
6807 The PTS (Presentation TimeStamp) of the output audio frame,
6808 expressed in @var{TB} units.
6811 The PTS of the output audio frame, expressed in seconds.
6814 The timebase of the output audio frames.
6817 Default is @code{1024}.
6820 @subsection Examples
6825 Generate a simple 440 Hz sine wave:
6831 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6835 sine=frequency=220:beep_factor=4:duration=5
6839 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6842 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6846 @c man end AUDIO SOURCES
6848 @chapter Audio Sinks
6849 @c man begin AUDIO SINKS
6851 Below is a description of the currently available audio sinks.
6853 @section abuffersink
6855 Buffer audio frames, and make them available to the end of filter chain.
6857 This sink is mainly intended for programmatic use, in particular
6858 through the interface defined in @file{libavfilter/buffersink.h}
6859 or the options system.
6861 It accepts a pointer to an AVABufferSinkContext structure, which
6862 defines the incoming buffers' formats, to be passed as the opaque
6863 parameter to @code{avfilter_init_filter} for initialization.
6866 Null audio sink; do absolutely nothing with the input audio. It is
6867 mainly useful as a template and for use in analysis / debugging
6870 @c man end AUDIO SINKS
6872 @chapter Video Filters
6873 @c man begin VIDEO FILTERS
6875 When you configure your FFmpeg build, you can disable any of the
6876 existing filters using @code{--disable-filters}.
6877 The configure output will show the video filters included in your
6880 Below is a description of the currently available video filters.
6884 Mark a region of interest in a video frame.
6886 The frame data is passed through unchanged, but metadata is attached
6887 to the frame indicating regions of interest which can affect the
6888 behaviour of later encoding. Multiple regions can be marked by
6889 applying the filter multiple times.
6893 Region distance in pixels from the left edge of the frame.
6895 Region distance in pixels from the top edge of the frame.
6897 Region width in pixels.
6899 Region height in pixels.
6901 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6902 and may contain the following variables:
6905 Width of the input frame.
6907 Height of the input frame.
6911 Quantisation offset to apply within the region.
6913 This must be a real value in the range -1 to +1. A value of zero
6914 indicates no quality change. A negative value asks for better quality
6915 (less quantisation), while a positive value asks for worse quality
6916 (greater quantisation).
6918 The range is calibrated so that the extreme values indicate the
6919 largest possible offset - if the rest of the frame is encoded with the
6920 worst possible quality, an offset of -1 indicates that this region
6921 should be encoded with the best possible quality anyway. Intermediate
6922 values are then interpolated in some codec-dependent way.
6924 For example, in 10-bit H.264 the quantisation parameter varies between
6925 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6926 this region should be encoded with a QP around one-tenth of the full
6927 range better than the rest of the frame. So, if most of the frame
6928 were to be encoded with a QP of around 30, this region would get a QP
6929 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6930 An extreme value of -1 would indicate that this region should be
6931 encoded with the best possible quality regardless of the treatment of
6932 the rest of the frame - that is, should be encoded at a QP of -12.
6934 If set to true, remove any existing regions of interest marked on the
6935 frame before adding the new one.
6938 @subsection Examples
6942 Mark the centre quarter of the frame as interesting.
6944 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6947 Mark the 100-pixel-wide region on the left edge of the frame as very
6948 uninteresting (to be encoded at much lower quality than the rest of
6951 addroi=0:0:100:ih:+1/5
6955 @section alphaextract
6957 Extract the alpha component from the input as a grayscale video. This
6958 is especially useful with the @var{alphamerge} filter.
6962 Add or replace the alpha component of the primary input with the
6963 grayscale value of a second input. This is intended for use with
6964 @var{alphaextract} to allow the transmission or storage of frame
6965 sequences that have alpha in a format that doesn't support an alpha
6968 For example, to reconstruct full frames from a normal YUV-encoded video
6969 and a separate video created with @var{alphaextract}, you might use:
6971 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6976 Amplify differences between current pixel and pixels of adjacent frames in
6977 same pixel location.
6979 This filter accepts the following options:
6983 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6984 For example radius of 3 will instruct filter to calculate average of 7 frames.
6987 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6990 Set threshold for difference amplification. Any difference greater or equal to
6991 this value will not alter source pixel. Default is 10.
6992 Allowed range is from 0 to 65535.
6995 Set tolerance for difference amplification. Any difference lower to
6996 this value will not alter source pixel. Default is 0.
6997 Allowed range is from 0 to 65535.
7000 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7001 This option controls maximum possible value that will decrease source pixel value.
7004 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7005 This option controls maximum possible value that will increase source pixel value.
7008 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
7011 @subsection Commands
7013 This filter supports the following @ref{commands} that corresponds to option of same name:
7025 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7026 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7027 Substation Alpha) subtitles files.
7029 This filter accepts the following option in addition to the common options from
7030 the @ref{subtitles} filter:
7034 Set the shaping engine
7036 Available values are:
7039 The default libass shaping engine, which is the best available.
7041 Fast, font-agnostic shaper that can do only substitutions
7043 Slower shaper using OpenType for substitutions and positioning
7046 The default is @code{auto}.
7050 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7052 The filter accepts the following options:
7056 Set threshold A for 1st plane. Default is 0.02.
7057 Valid range is 0 to 0.3.
7060 Set threshold B for 1st plane. Default is 0.04.
7061 Valid range is 0 to 5.
7064 Set threshold A for 2nd plane. Default is 0.02.
7065 Valid range is 0 to 0.3.
7068 Set threshold B for 2nd plane. Default is 0.04.
7069 Valid range is 0 to 5.
7072 Set threshold A for 3rd plane. Default is 0.02.
7073 Valid range is 0 to 0.3.
7076 Set threshold B for 3rd plane. Default is 0.04.
7077 Valid range is 0 to 5.
7079 Threshold A is designed to react on abrupt changes in the input signal and
7080 threshold B is designed to react on continuous changes in the input signal.
7083 Set number of frames filter will use for averaging. Default is 9. Must be odd
7084 number in range [5, 129].
7087 Set what planes of frame filter will use for averaging. Default is all.
7090 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7091 Alternatively can be set to @code{s} serial.
7093 Parallel can be faster then serial, while other way around is never true.
7094 Parallel will abort early on first change being greater then thresholds, while serial
7095 will continue processing other side of frames if they are equal or below thresholds.
7098 @subsection Commands
7099 This filter supports same @ref{commands} as options except option @code{s}.
7100 The command accepts the same syntax of the corresponding option.
7104 Apply average blur filter.
7106 The filter accepts the following options:
7110 Set horizontal radius size.
7113 Set which planes to filter. By default all planes are filtered.
7116 Set vertical radius size, if zero it will be same as @code{sizeX}.
7117 Default is @code{0}.
7120 @subsection Commands
7121 This filter supports same commands as options.
7122 The command accepts the same syntax of the corresponding option.
7124 If the specified expression is not valid, it is kept at its current
7129 Compute the bounding box for the non-black pixels in the input frame
7132 This filter computes the bounding box containing all the pixels with a
7133 luminance value greater than the minimum allowed value.
7134 The parameters describing the bounding box are printed on the filter
7137 The filter accepts the following option:
7141 Set the minimal luminance value. Default is @code{16}.
7145 Apply bilateral filter, spatial smoothing while preserving edges.
7147 The filter accepts the following options:
7150 Set sigma of gaussian function to calculate spatial weight.
7151 Allowed range is 0 to 512. Default is 0.1.
7154 Set sigma of gaussian function to calculate range weight.
7155 Allowed range is 0 to 1. Default is 0.1.
7158 Set planes to filter. Default is first only.
7161 @section bitplanenoise
7163 Show and measure bit plane noise.
7165 The filter accepts the following options:
7169 Set which plane to analyze. Default is @code{1}.
7172 Filter out noisy pixels from @code{bitplane} set above.
7173 Default is disabled.
7176 @section blackdetect
7178 Detect video intervals that are (almost) completely black. Can be
7179 useful to detect chapter transitions, commercials, or invalid
7182 The filter outputs its detection analysis to both the log as well as
7183 frame metadata. If a black segment of at least the specified minimum
7184 duration is found, a line with the start and end timestamps as well
7185 as duration is printed to the log with level @code{info}. In addition,
7186 a log line with level @code{debug} is printed per frame showing the
7187 black amount detected for that frame.
7189 The filter also attaches metadata to the first frame of a black
7190 segment with key @code{lavfi.black_start} and to the first frame
7191 after the black segment ends with key @code{lavfi.black_end}. The
7192 value is the frame's timestamp. This metadata is added regardless
7193 of the minimum duration specified.
7195 The filter accepts the following options:
7198 @item black_min_duration, d
7199 Set the minimum detected black duration expressed in seconds. It must
7200 be a non-negative floating point number.
7202 Default value is 2.0.
7204 @item picture_black_ratio_th, pic_th
7205 Set the threshold for considering a picture "black".
7206 Express the minimum value for the ratio:
7208 @var{nb_black_pixels} / @var{nb_pixels}
7211 for which a picture is considered black.
7212 Default value is 0.98.
7214 @item pixel_black_th, pix_th
7215 Set the threshold for considering a pixel "black".
7217 The threshold expresses the maximum pixel luminance value for which a
7218 pixel is considered "black". The provided value is scaled according to
7219 the following equation:
7221 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7224 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7225 the input video format, the range is [0-255] for YUV full-range
7226 formats and [16-235] for YUV non full-range formats.
7228 Default value is 0.10.
7231 The following example sets the maximum pixel threshold to the minimum
7232 value, and detects only black intervals of 2 or more seconds:
7234 blackdetect=d=2:pix_th=0.00
7239 Detect frames that are (almost) completely black. Can be useful to
7240 detect chapter transitions or commercials. Output lines consist of
7241 the frame number of the detected frame, the percentage of blackness,
7242 the position in the file if known or -1 and the timestamp in seconds.
7244 In order to display the output lines, you need to set the loglevel at
7245 least to the AV_LOG_INFO value.
7247 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7248 The value represents the percentage of pixels in the picture that
7249 are below the threshold value.
7251 It accepts the following parameters:
7256 The percentage of the pixels that have to be below the threshold; it defaults to
7259 @item threshold, thresh
7260 The threshold below which a pixel value is considered black; it defaults to
7268 Blend two video frames into each other.
7270 The @code{blend} filter takes two input streams and outputs one
7271 stream, the first input is the "top" layer and second input is
7272 "bottom" layer. By default, the output terminates when the longest input terminates.
7274 The @code{tblend} (time blend) filter takes two consecutive frames
7275 from one single stream, and outputs the result obtained by blending
7276 the new frame on top of the old frame.
7278 A description of the accepted options follows.
7286 Set blend mode for specific pixel component or all pixel components in case
7287 of @var{all_mode}. Default value is @code{normal}.
7289 Available values for component modes are:
7331 Set blend opacity for specific pixel component or all pixel components in case
7332 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7339 Set blend expression for specific pixel component or all pixel components in case
7340 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7342 The expressions can use the following variables:
7346 The sequential number of the filtered frame, starting from @code{0}.
7350 the coordinates of the current sample
7354 the width and height of currently filtered plane
7358 Width and height scale for the plane being filtered. It is the
7359 ratio between the dimensions of the current plane to the luma plane,
7360 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7361 the luma plane and @code{0.5,0.5} for the chroma planes.
7364 Time of the current frame, expressed in seconds.
7367 Value of pixel component at current location for first video frame (top layer).
7370 Value of pixel component at current location for second video frame (bottom layer).
7374 The @code{blend} filter also supports the @ref{framesync} options.
7376 @subsection Examples
7380 Apply transition from bottom layer to top layer in first 10 seconds:
7382 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7386 Apply linear horizontal transition from top layer to bottom layer:
7388 blend=all_expr='A*(X/W)+B*(1-X/W)'
7392 Apply 1x1 checkerboard effect:
7394 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7398 Apply uncover left effect:
7400 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7404 Apply uncover down effect:
7406 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7410 Apply uncover up-left effect:
7412 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7416 Split diagonally video and shows top and bottom layer on each side:
7418 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7422 Display differences between the current and the previous frame:
7424 tblend=all_mode=grainextract
7430 Denoise frames using Block-Matching 3D algorithm.
7432 The filter accepts the following options.
7436 Set denoising strength. Default value is 1.
7437 Allowed range is from 0 to 999.9.
7438 The denoising algorithm is very sensitive to sigma, so adjust it
7439 according to the source.
7442 Set local patch size. This sets dimensions in 2D.
7445 Set sliding step for processing blocks. Default value is 4.
7446 Allowed range is from 1 to 64.
7447 Smaller values allows processing more reference blocks and is slower.
7450 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7451 When set to 1, no block matching is done. Larger values allows more blocks
7453 Allowed range is from 1 to 256.
7456 Set radius for search block matching. Default is 9.
7457 Allowed range is from 1 to INT32_MAX.
7460 Set step between two search locations for block matching. Default is 1.
7461 Allowed range is from 1 to 64. Smaller is slower.
7464 Set threshold of mean square error for block matching. Valid range is 0 to
7468 Set thresholding parameter for hard thresholding in 3D transformed domain.
7469 Larger values results in stronger hard-thresholding filtering in frequency
7473 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7474 Default is @code{basic}.
7477 If enabled, filter will use 2nd stream for block matching.
7478 Default is disabled for @code{basic} value of @var{estim} option,
7479 and always enabled if value of @var{estim} is @code{final}.
7482 Set planes to filter. Default is all available except alpha.
7485 @subsection Examples
7489 Basic filtering with bm3d:
7491 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7495 Same as above, but filtering only luma:
7497 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7501 Same as above, but with both estimation modes:
7503 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
7507 Same as above, but prefilter with @ref{nlmeans} filter instead:
7509 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
7515 Apply a boxblur algorithm to the input video.
7517 It accepts the following parameters:
7521 @item luma_radius, lr
7522 @item luma_power, lp
7523 @item chroma_radius, cr
7524 @item chroma_power, cp
7525 @item alpha_radius, ar
7526 @item alpha_power, ap
7530 A description of the accepted options follows.
7533 @item luma_radius, lr
7534 @item chroma_radius, cr
7535 @item alpha_radius, ar
7536 Set an expression for the box radius in pixels used for blurring the
7537 corresponding input plane.
7539 The radius value must be a non-negative number, and must not be
7540 greater than the value of the expression @code{min(w,h)/2} for the
7541 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7544 Default value for @option{luma_radius} is "2". If not specified,
7545 @option{chroma_radius} and @option{alpha_radius} default to the
7546 corresponding value set for @option{luma_radius}.
7548 The expressions can contain the following constants:
7552 The input width and height in pixels.
7556 The input chroma image width and height in pixels.
7560 The horizontal and vertical chroma subsample values. For example, for the
7561 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7564 @item luma_power, lp
7565 @item chroma_power, cp
7566 @item alpha_power, ap
7567 Specify how many times the boxblur filter is applied to the
7568 corresponding plane.
7570 Default value for @option{luma_power} is 2. If not specified,
7571 @option{chroma_power} and @option{alpha_power} default to the
7572 corresponding value set for @option{luma_power}.
7574 A value of 0 will disable the effect.
7577 @subsection Examples
7581 Apply a boxblur filter with the luma, chroma, and alpha radii
7584 boxblur=luma_radius=2:luma_power=1
7589 Set the luma radius to 2, and alpha and chroma radius to 0:
7591 boxblur=2:1:cr=0:ar=0
7595 Set the luma and chroma radii to a fraction of the video dimension:
7597 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7603 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7604 Deinterlacing Filter").
7606 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7607 interpolation algorithms.
7608 It accepts the following parameters:
7612 The interlacing mode to adopt. It accepts one of the following values:
7616 Output one frame for each frame.
7618 Output one frame for each field.
7621 The default value is @code{send_field}.
7624 The picture field parity assumed for the input interlaced video. It accepts one
7625 of the following values:
7629 Assume the top field is first.
7631 Assume the bottom field is first.
7633 Enable automatic detection of field parity.
7636 The default value is @code{auto}.
7637 If the interlacing is unknown or the decoder does not export this information,
7638 top field first will be assumed.
7641 Specify which frames to deinterlace. Accepts one of the following
7646 Deinterlace all frames.
7648 Only deinterlace frames marked as interlaced.
7651 The default value is @code{all}.
7656 Apply Contrast Adaptive Sharpen filter to video stream.
7658 The filter accepts the following options:
7662 Set the sharpening strength. Default value is 0.
7665 Set planes to filter. Default value is to filter all
7666 planes except alpha plane.
7670 Remove all color information for all colors except for certain one.
7672 The filter accepts the following options:
7676 The color which will not be replaced with neutral chroma.
7679 Similarity percentage with the above color.
7680 0.01 matches only the exact key color, while 1.0 matches everything.
7684 0.0 makes pixels either fully gray, or not gray at all.
7685 Higher values result in more preserved color.
7688 Signals that the color passed is already in YUV instead of RGB.
7690 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7691 This can be used to pass exact YUV values as hexadecimal numbers.
7694 @subsection Commands
7695 This filter supports same @ref{commands} as options.
7696 The command accepts the same syntax of the corresponding option.
7698 If the specified expression is not valid, it is kept at its current
7702 YUV colorspace color/chroma keying.
7704 The filter accepts the following options:
7708 The color which will be replaced with transparency.
7711 Similarity percentage with the key color.
7713 0.01 matches only the exact key color, while 1.0 matches everything.
7718 0.0 makes pixels either fully transparent, or not transparent at all.
7720 Higher values result in semi-transparent pixels, with a higher transparency
7721 the more similar the pixels color is to the key color.
7724 Signals that the color passed is already in YUV instead of RGB.
7726 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7727 This can be used to pass exact YUV values as hexadecimal numbers.
7730 @subsection Commands
7731 This filter supports same @ref{commands} as options.
7732 The command accepts the same syntax of the corresponding option.
7734 If the specified expression is not valid, it is kept at its current
7737 @subsection Examples
7741 Make every green pixel in the input image transparent:
7743 ffmpeg -i input.png -vf chromakey=green out.png
7747 Overlay a greenscreen-video on top of a static black background.
7749 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
7754 Reduce chrominance noise.
7756 The filter accepts the following options:
7760 Set threshold for averaging chrominance values.
7761 Sum of absolute difference of Y, U and V pixel components of current
7762 pixel and neighbour pixels lower than this threshold will be used in
7763 averaging. Luma component is left unchanged and is copied to output.
7764 Default value is 30. Allowed range is from 1 to 200.
7767 Set horizontal radius of rectangle used for averaging.
7768 Allowed range is from 1 to 100. Default value is 5.
7771 Set vertical radius of rectangle used for averaging.
7772 Allowed range is from 1 to 100. Default value is 5.
7775 Set horizontal step when averaging. Default value is 1.
7776 Allowed range is from 1 to 50.
7777 Mostly useful to speed-up filtering.
7780 Set vertical step when averaging. Default value is 1.
7781 Allowed range is from 1 to 50.
7782 Mostly useful to speed-up filtering.
7785 Set Y threshold for averaging chrominance values.
7786 Set finer control for max allowed difference between Y components
7787 of current pixel and neigbour pixels.
7788 Default value is 200. Allowed range is from 1 to 200.
7791 Set U threshold for averaging chrominance values.
7792 Set finer control for max allowed difference between U components
7793 of current pixel and neigbour pixels.
7794 Default value is 200. Allowed range is from 1 to 200.
7797 Set V threshold for averaging chrominance values.
7798 Set finer control for max allowed difference between V components
7799 of current pixel and neigbour pixels.
7800 Default value is 200. Allowed range is from 1 to 200.
7803 @subsection Commands
7804 This filter supports same @ref{commands} as options.
7805 The command accepts the same syntax of the corresponding option.
7807 @section chromashift
7808 Shift chroma pixels horizontally and/or vertically.
7810 The filter accepts the following options:
7813 Set amount to shift chroma-blue horizontally.
7815 Set amount to shift chroma-blue vertically.
7817 Set amount to shift chroma-red horizontally.
7819 Set amount to shift chroma-red vertically.
7821 Set edge mode, can be @var{smear}, default, or @var{warp}.
7824 @subsection Commands
7826 This filter supports the all above options as @ref{commands}.
7830 Display CIE color diagram with pixels overlaid onto it.
7832 The filter accepts the following options:
7847 @item uhdtv, rec2020
7861 Set what gamuts to draw.
7863 See @code{system} option for available values.
7866 Set ciescope size, by default set to 512.
7869 Set intensity used to map input pixel values to CIE diagram.
7872 Set contrast used to draw tongue colors that are out of active color system gamut.
7875 Correct gamma displayed on scope, by default enabled.
7878 Show white point on CIE diagram, by default disabled.
7881 Set input gamma. Used only with XYZ input color space.
7886 Visualize information exported by some codecs.
7888 Some codecs can export information through frames using side-data or other
7889 means. For example, some MPEG based codecs export motion vectors through the
7890 @var{export_mvs} flag in the codec @option{flags2} option.
7892 The filter accepts the following option:
7896 Set motion vectors to visualize.
7898 Available flags for @var{mv} are:
7902 forward predicted MVs of P-frames
7904 forward predicted MVs of B-frames
7906 backward predicted MVs of B-frames
7910 Display quantization parameters using the chroma planes.
7913 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7915 Available flags for @var{mv_type} are:
7919 forward predicted MVs
7921 backward predicted MVs
7924 @item frame_type, ft
7925 Set frame type to visualize motion vectors of.
7927 Available flags for @var{frame_type} are:
7931 intra-coded frames (I-frames)
7933 predicted frames (P-frames)
7935 bi-directionally predicted frames (B-frames)
7939 @subsection Examples
7943 Visualize forward predicted MVs of all frames using @command{ffplay}:
7945 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7949 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7951 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7955 @section colorbalance
7956 Modify intensity of primary colors (red, green and blue) of input frames.
7958 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7959 regions for the red-cyan, green-magenta or blue-yellow balance.
7961 A positive adjustment value shifts the balance towards the primary color, a negative
7962 value towards the complementary color.
7964 The filter accepts the following options:
7970 Adjust red, green and blue shadows (darkest pixels).
7975 Adjust red, green and blue midtones (medium pixels).
7980 Adjust red, green and blue highlights (brightest pixels).
7982 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7985 Preserve lightness when changing color balance. Default is disabled.
7988 @subsection Examples
7992 Add red color cast to shadows:
7998 @subsection Commands
8000 This filter supports the all above options as @ref{commands}.
8002 @section colorchannelmixer
8004 Adjust video input frames by re-mixing color channels.
8006 This filter modifies a color channel by adding the values associated to
8007 the other channels of the same pixels. For example if the value to
8008 modify is red, the output value will be:
8010 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8013 The filter accepts the following options:
8020 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8021 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8027 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8028 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8034 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8035 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8041 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8042 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8044 Allowed ranges for options are @code{[-2.0, 2.0]}.
8047 @subsection Examples
8051 Convert source to grayscale:
8053 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8056 Simulate sepia tones:
8058 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8062 @subsection Commands
8064 This filter supports the all above options as @ref{commands}.
8067 RGB colorspace color keying.
8069 The filter accepts the following options:
8073 The color which will be replaced with transparency.
8076 Similarity percentage with the key color.
8078 0.01 matches only the exact key color, while 1.0 matches everything.
8083 0.0 makes pixels either fully transparent, or not transparent at all.
8085 Higher values result in semi-transparent pixels, with a higher transparency
8086 the more similar the pixels color is to the key color.
8089 @subsection Examples
8093 Make every green pixel in the input image transparent:
8095 ffmpeg -i input.png -vf colorkey=green out.png
8099 Overlay a greenscreen-video on top of a static background image.
8101 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
8105 @subsection Commands
8106 This filter supports same @ref{commands} as options.
8107 The command accepts the same syntax of the corresponding option.
8109 If the specified expression is not valid, it is kept at its current
8113 Remove all color information for all RGB colors except for certain one.
8115 The filter accepts the following options:
8119 The color which will not be replaced with neutral gray.
8122 Similarity percentage with the above color.
8123 0.01 matches only the exact key color, while 1.0 matches everything.
8126 Blend percentage. 0.0 makes pixels fully gray.
8127 Higher values result in more preserved color.
8130 @subsection Commands
8131 This filter supports same @ref{commands} as options.
8132 The command accepts the same syntax of the corresponding option.
8134 If the specified expression is not valid, it is kept at its current
8137 @section colorlevels
8139 Adjust video input frames using levels.
8141 The filter accepts the following options:
8148 Adjust red, green, blue and alpha input black point.
8149 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8155 Adjust red, green, blue and alpha input white point.
8156 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8158 Input levels are used to lighten highlights (bright tones), darken shadows
8159 (dark tones), change the balance of bright and dark tones.
8165 Adjust red, green, blue and alpha output black point.
8166 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8172 Adjust red, green, blue and alpha output white point.
8173 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8175 Output levels allows manual selection of a constrained output level range.
8178 @subsection Examples
8182 Make video output darker:
8184 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8190 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8194 Make video output lighter:
8196 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8200 Increase brightness:
8202 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8206 @subsection Commands
8208 This filter supports the all above options as @ref{commands}.
8210 @section colormatrix
8212 Convert color matrix.
8214 The filter accepts the following options:
8219 Specify the source and destination color matrix. Both values must be
8222 The accepted values are:
8250 For example to convert from BT.601 to SMPTE-240M, use the command:
8252 colormatrix=bt601:smpte240m
8257 Convert colorspace, transfer characteristics or color primaries.
8258 Input video needs to have an even size.
8260 The filter accepts the following options:
8265 Specify all color properties at once.
8267 The accepted values are:
8297 Specify output colorspace.
8299 The accepted values are:
8308 BT.470BG or BT.601-6 625
8311 SMPTE-170M or BT.601-6 525
8320 BT.2020 with non-constant luminance
8326 Specify output transfer characteristics.
8328 The accepted values are:
8340 Constant gamma of 2.2
8343 Constant gamma of 2.8
8346 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8364 BT.2020 for 10-bits content
8367 BT.2020 for 12-bits content
8373 Specify output color primaries.
8375 The accepted values are:
8384 BT.470BG or BT.601-6 625
8387 SMPTE-170M or BT.601-6 525
8411 Specify output color range.
8413 The accepted values are:
8416 TV (restricted) range
8419 MPEG (restricted) range
8430 Specify output color format.
8432 The accepted values are:
8435 YUV 4:2:0 planar 8-bits
8438 YUV 4:2:0 planar 10-bits
8441 YUV 4:2:0 planar 12-bits
8444 YUV 4:2:2 planar 8-bits
8447 YUV 4:2:2 planar 10-bits
8450 YUV 4:2:2 planar 12-bits
8453 YUV 4:4:4 planar 8-bits
8456 YUV 4:4:4 planar 10-bits
8459 YUV 4:4:4 planar 12-bits
8464 Do a fast conversion, which skips gamma/primary correction. This will take
8465 significantly less CPU, but will be mathematically incorrect. To get output
8466 compatible with that produced by the colormatrix filter, use fast=1.
8469 Specify dithering mode.
8471 The accepted values are:
8477 Floyd-Steinberg dithering
8481 Whitepoint adaptation mode.
8483 The accepted values are:
8486 Bradford whitepoint adaptation
8489 von Kries whitepoint adaptation
8492 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8496 Override all input properties at once. Same accepted values as @ref{all}.
8499 Override input colorspace. Same accepted values as @ref{space}.
8502 Override input color primaries. Same accepted values as @ref{primaries}.
8505 Override input transfer characteristics. Same accepted values as @ref{trc}.
8508 Override input color range. Same accepted values as @ref{range}.
8512 The filter converts the transfer characteristics, color space and color
8513 primaries to the specified user values. The output value, if not specified,
8514 is set to a default value based on the "all" property. If that property is
8515 also not specified, the filter will log an error. The output color range and
8516 format default to the same value as the input color range and format. The
8517 input transfer characteristics, color space, color primaries and color range
8518 should be set on the input data. If any of these are missing, the filter will
8519 log an error and no conversion will take place.
8521 For example to convert the input to SMPTE-240M, use the command:
8523 colorspace=smpte240m
8526 @section convolution
8528 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8530 The filter accepts the following options:
8537 Set matrix for each plane.
8538 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8539 and from 1 to 49 odd number of signed integers in @var{row} mode.
8545 Set multiplier for calculated value for each plane.
8546 If unset or 0, it will be sum of all matrix elements.
8552 Set bias for each plane. This value is added to the result of the multiplication.
8553 Useful for making the overall image brighter or darker. Default is 0.0.
8559 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8560 Default is @var{square}.
8563 @subsection Commands
8565 This filter supports the all above options as @ref{commands}.
8567 @subsection Examples
8573 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"
8579 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"
8585 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"
8591 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"
8595 Apply laplacian edge detector which includes diagonals:
8597 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"
8603 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"
8609 Apply 2D convolution of video stream in frequency domain using second stream
8612 The filter accepts the following options:
8616 Set which planes to process.
8619 Set which impulse video frames will be processed, can be @var{first}
8620 or @var{all}. Default is @var{all}.
8623 The @code{convolve} filter also supports the @ref{framesync} options.
8627 Copy the input video source unchanged to the output. This is mainly useful for
8632 Video filtering on GPU using Apple's CoreImage API on OSX.
8634 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8635 processed by video hardware. However, software-based OpenGL implementations
8636 exist which means there is no guarantee for hardware processing. It depends on
8639 There are many filters and image generators provided by Apple that come with a
8640 large variety of options. The filter has to be referenced by its name along
8643 The coreimage filter accepts the following options:
8646 List all available filters and generators along with all their respective
8647 options as well as possible minimum and maximum values along with the default
8654 Specify all filters by their respective name and options.
8655 Use @var{list_filters} to determine all valid filter names and options.
8656 Numerical options are specified by a float value and are automatically clamped
8657 to their respective value range. Vector and color options have to be specified
8658 by a list of space separated float values. Character escaping has to be done.
8659 A special option name @code{default} is available to use default options for a
8662 It is required to specify either @code{default} or at least one of the filter options.
8663 All omitted options are used with their default values.
8664 The syntax of the filter string is as follows:
8666 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8670 Specify a rectangle where the output of the filter chain is copied into the
8671 input image. It is given by a list of space separated float values:
8673 output_rect=x\ y\ width\ height
8675 If not given, the output rectangle equals the dimensions of the input image.
8676 The output rectangle is automatically cropped at the borders of the input
8677 image. Negative values are valid for each component.
8679 output_rect=25\ 25\ 100\ 100
8683 Several filters can be chained for successive processing without GPU-HOST
8684 transfers allowing for fast processing of complex filter chains.
8685 Currently, only filters with zero (generators) or exactly one (filters) input
8686 image and one output image are supported. Also, transition filters are not yet
8689 Some filters generate output images with additional padding depending on the
8690 respective filter kernel. The padding is automatically removed to ensure the
8691 filter output has the same size as the input image.
8693 For image generators, the size of the output image is determined by the
8694 previous output image of the filter chain or the input image of the whole
8695 filterchain, respectively. The generators do not use the pixel information of
8696 this image to generate their output. However, the generated output is
8697 blended onto this image, resulting in partial or complete coverage of the
8700 The @ref{coreimagesrc} video source can be used for generating input images
8701 which are directly fed into the filter chain. By using it, providing input
8702 images by another video source or an input video is not required.
8704 @subsection Examples
8709 List all filters available:
8711 coreimage=list_filters=true
8715 Use the CIBoxBlur filter with default options to blur an image:
8717 coreimage=filter=CIBoxBlur@@default
8721 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8722 its center at 100x100 and a radius of 50 pixels:
8724 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8728 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8729 given as complete and escaped command-line for Apple's standard bash shell:
8731 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8737 Cover a rectangular object
8739 It accepts the following options:
8743 Filepath of the optional cover image, needs to be in yuv420.
8748 It accepts the following values:
8751 cover it by the supplied image
8753 cover it by interpolating the surrounding pixels
8756 Default value is @var{blur}.
8759 @subsection Examples
8763 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8765 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8771 Crop the input video to given dimensions.
8773 It accepts the following parameters:
8777 The width of the output video. It defaults to @code{iw}.
8778 This expression is evaluated only once during the filter
8779 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8782 The height of the output video. It defaults to @code{ih}.
8783 This expression is evaluated only once during the filter
8784 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8787 The horizontal position, in the input video, of the left edge of the output
8788 video. It defaults to @code{(in_w-out_w)/2}.
8789 This expression is evaluated per-frame.
8792 The vertical position, in the input video, of the top edge of the output video.
8793 It defaults to @code{(in_h-out_h)/2}.
8794 This expression is evaluated per-frame.
8797 If set to 1 will force the output display aspect ratio
8798 to be the same of the input, by changing the output sample aspect
8799 ratio. It defaults to 0.
8802 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8803 width/height/x/y as specified and will not be rounded to nearest smaller value.
8807 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8808 expressions containing the following constants:
8813 The computed values for @var{x} and @var{y}. They are evaluated for
8818 The input width and height.
8822 These are the same as @var{in_w} and @var{in_h}.
8826 The output (cropped) width and height.
8830 These are the same as @var{out_w} and @var{out_h}.
8833 same as @var{iw} / @var{ih}
8836 input sample aspect ratio
8839 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8843 horizontal and vertical chroma subsample values. For example for the
8844 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8847 The number of the input frame, starting from 0.
8850 the position in the file of the input frame, NAN if unknown
8853 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8857 The expression for @var{out_w} may depend on the value of @var{out_h},
8858 and the expression for @var{out_h} may depend on @var{out_w}, but they
8859 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8860 evaluated after @var{out_w} and @var{out_h}.
8862 The @var{x} and @var{y} parameters specify the expressions for the
8863 position of the top-left corner of the output (non-cropped) area. They
8864 are evaluated for each frame. If the evaluated value is not valid, it
8865 is approximated to the nearest valid value.
8867 The expression for @var{x} may depend on @var{y}, and the expression
8868 for @var{y} may depend on @var{x}.
8870 @subsection Examples
8874 Crop area with size 100x100 at position (12,34).
8879 Using named options, the example above becomes:
8881 crop=w=100:h=100:x=12:y=34
8885 Crop the central input area with size 100x100:
8891 Crop the central input area with size 2/3 of the input video:
8893 crop=2/3*in_w:2/3*in_h
8897 Crop the input video central square:
8904 Delimit the rectangle with the top-left corner placed at position
8905 100:100 and the right-bottom corner corresponding to the right-bottom
8906 corner of the input image.
8908 crop=in_w-100:in_h-100:100:100
8912 Crop 10 pixels from the left and right borders, and 20 pixels from
8913 the top and bottom borders
8915 crop=in_w-2*10:in_h-2*20
8919 Keep only the bottom right quarter of the input image:
8921 crop=in_w/2:in_h/2:in_w/2:in_h/2
8925 Crop height for getting Greek harmony:
8927 crop=in_w:1/PHI*in_w
8931 Apply trembling effect:
8933 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)
8937 Apply erratic camera effect depending on timestamp:
8939 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)"
8943 Set x depending on the value of y:
8945 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8949 @subsection Commands
8951 This filter supports the following commands:
8957 Set width/height of the output video and the horizontal/vertical position
8959 The command accepts the same syntax of the corresponding option.
8961 If the specified expression is not valid, it is kept at its current
8967 Auto-detect the crop size.
8969 It calculates the necessary cropping parameters and prints the
8970 recommended parameters via the logging system. The detected dimensions
8971 correspond to the non-black area of the input video.
8973 It accepts the following parameters:
8978 Set higher black value threshold, which can be optionally specified
8979 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8980 value greater to the set value is considered non-black. It defaults to 24.
8981 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8982 on the bitdepth of the pixel format.
8985 The value which the width/height should be divisible by. It defaults to
8986 16. The offset is automatically adjusted to center the video. Use 2 to
8987 get only even dimensions (needed for 4:2:2 video). 16 is best when
8988 encoding to most video codecs.
8991 Set the number of initial frames for which evaluation is skipped.
8992 Default is 2. Range is 0 to INT_MAX.
8994 @item reset_count, reset
8995 Set the counter that determines after how many frames cropdetect will
8996 reset the previously detected largest video area and start over to
8997 detect the current optimal crop area. Default value is 0.
8999 This can be useful when channel logos distort the video area. 0
9000 indicates 'never reset', and returns the largest area encountered during
9007 Delay video filtering until a given wallclock timestamp. The filter first
9008 passes on @option{preroll} amount of frames, then it buffers at most
9009 @option{buffer} amount of frames and waits for the cue. After reaching the cue
9010 it forwards the buffered frames and also any subsequent frames coming in its
9013 The filter can be used synchronize the output of multiple ffmpeg processes for
9014 realtime output devices like decklink. By putting the delay in the filtering
9015 chain and pre-buffering frames the process can pass on data to output almost
9016 immediately after the target wallclock timestamp is reached.
9018 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9024 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9027 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9030 The maximum duration of content to buffer before waiting for the cue expressed
9031 in seconds. Default is 0.
9038 Apply color adjustments using curves.
9040 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9041 component (red, green and blue) has its values defined by @var{N} key points
9042 tied from each other using a smooth curve. The x-axis represents the pixel
9043 values from the input frame, and the y-axis the new pixel values to be set for
9046 By default, a component curve is defined by the two points @var{(0;0)} and
9047 @var{(1;1)}. This creates a straight line where each original pixel value is
9048 "adjusted" to its own value, which means no change to the image.
9050 The filter allows you to redefine these two points and add some more. A new
9051 curve (using a natural cubic spline interpolation) will be define to pass
9052 smoothly through all these new coordinates. The new defined points needs to be
9053 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9054 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9055 the vector spaces, the values will be clipped accordingly.
9057 The filter accepts the following options:
9061 Select one of the available color presets. This option can be used in addition
9062 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9063 options takes priority on the preset values.
9064 Available presets are:
9067 @item color_negative
9070 @item increase_contrast
9072 @item linear_contrast
9073 @item medium_contrast
9075 @item strong_contrast
9078 Default is @code{none}.
9080 Set the master key points. These points will define a second pass mapping. It
9081 is sometimes called a "luminance" or "value" mapping. It can be used with
9082 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9083 post-processing LUT.
9085 Set the key points for the red component.
9087 Set the key points for the green component.
9089 Set the key points for the blue component.
9091 Set the key points for all components (not including master).
9092 Can be used in addition to the other key points component
9093 options. In this case, the unset component(s) will fallback on this
9094 @option{all} setting.
9096 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9098 Save Gnuplot script of the curves in specified file.
9101 To avoid some filtergraph syntax conflicts, each key points list need to be
9102 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9104 @subsection Examples
9108 Increase slightly the middle level of blue:
9110 curves=blue='0/0 0.5/0.58 1/1'
9116 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'
9118 Here we obtain the following coordinates for each components:
9121 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9123 @code{(0;0) (0.50;0.48) (1;1)}
9125 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9129 The previous example can also be achieved with the associated built-in preset:
9131 curves=preset=vintage
9141 Use a Photoshop preset and redefine the points of the green component:
9143 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9147 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9148 and @command{gnuplot}:
9150 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9151 gnuplot -p /tmp/curves.plt
9157 Video data analysis filter.
9159 This filter shows hexadecimal pixel values of part of video.
9161 The filter accepts the following options:
9165 Set output video size.
9168 Set x offset from where to pick pixels.
9171 Set y offset from where to pick pixels.
9174 Set scope mode, can be one of the following:
9177 Draw hexadecimal pixel values with white color on black background.
9180 Draw hexadecimal pixel values with input video pixel color on black
9184 Draw hexadecimal pixel values on color background picked from input video,
9185 the text color is picked in such way so its always visible.
9189 Draw rows and columns numbers on left and top of video.
9192 Set background opacity.
9195 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9199 Apply Directional blur filter.
9201 The filter accepts the following options:
9205 Set angle of directional blur. Default is @code{45}.
9208 Set radius of directional blur. Default is @code{5}.
9211 Set which planes to filter. By default all planes are filtered.
9214 @subsection Commands
9215 This filter supports same @ref{commands} as options.
9216 The command accepts the same syntax of the corresponding option.
9218 If the specified expression is not valid, it is kept at its current
9223 Denoise frames using 2D DCT (frequency domain filtering).
9225 This filter is not designed for real time.
9227 The filter accepts the following options:
9231 Set the noise sigma constant.
9233 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9234 coefficient (absolute value) below this threshold with be dropped.
9236 If you need a more advanced filtering, see @option{expr}.
9238 Default is @code{0}.
9241 Set number overlapping pixels for each block. Since the filter can be slow, you
9242 may want to reduce this value, at the cost of a less effective filter and the
9243 risk of various artefacts.
9245 If the overlapping value doesn't permit processing the whole input width or
9246 height, a warning will be displayed and according borders won't be denoised.
9248 Default value is @var{blocksize}-1, which is the best possible setting.
9251 Set the coefficient factor expression.
9253 For each coefficient of a DCT block, this expression will be evaluated as a
9254 multiplier value for the coefficient.
9256 If this is option is set, the @option{sigma} option will be ignored.
9258 The absolute value of the coefficient can be accessed through the @var{c}
9262 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9263 @var{blocksize}, which is the width and height of the processed blocks.
9265 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9266 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9267 on the speed processing. Also, a larger block size does not necessarily means a
9271 @subsection Examples
9273 Apply a denoise with a @option{sigma} of @code{4.5}:
9278 The same operation can be achieved using the expression system:
9280 dctdnoiz=e='gte(c, 4.5*3)'
9283 Violent denoise using a block size of @code{16x16}:
9290 Remove banding artifacts from input video.
9291 It works by replacing banded pixels with average value of referenced pixels.
9293 The filter accepts the following options:
9300 Set banding detection threshold for each plane. Default is 0.02.
9301 Valid range is 0.00003 to 0.5.
9302 If difference between current pixel and reference pixel is less than threshold,
9303 it will be considered as banded.
9306 Banding detection range in pixels. Default is 16. If positive, random number
9307 in range 0 to set value will be used. If negative, exact absolute value
9309 The range defines square of four pixels around current pixel.
9312 Set direction in radians from which four pixel will be compared. If positive,
9313 random direction from 0 to set direction will be picked. If negative, exact of
9314 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9315 will pick only pixels on same row and -PI/2 will pick only pixels on same
9319 If enabled, current pixel is compared with average value of all four
9320 surrounding pixels. The default is enabled. If disabled current pixel is
9321 compared with all four surrounding pixels. The pixel is considered banded
9322 if only all four differences with surrounding pixels are less than threshold.
9325 If enabled, current pixel is changed if and only if all pixel components are banded,
9326 e.g. banding detection threshold is triggered for all color components.
9327 The default is disabled.
9332 Remove blocking artifacts from input video.
9334 The filter accepts the following options:
9338 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9339 This controls what kind of deblocking is applied.
9342 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9348 Set blocking detection thresholds. Allowed range is 0 to 1.
9349 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9350 Using higher threshold gives more deblocking strength.
9351 Setting @var{alpha} controls threshold detection at exact edge of block.
9352 Remaining options controls threshold detection near the edge. Each one for
9353 below/above or left/right. Setting any of those to @var{0} disables
9357 Set planes to filter. Default is to filter all available planes.
9360 @subsection Examples
9364 Deblock using weak filter and block size of 4 pixels.
9366 deblock=filter=weak:block=4
9370 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9371 deblocking more edges.
9373 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9377 Similar as above, but filter only first plane.
9379 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9383 Similar as above, but filter only second and third plane.
9385 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9392 Drop duplicated frames at regular intervals.
9394 The filter accepts the following options:
9398 Set the number of frames from which one will be dropped. Setting this to
9399 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9400 Default is @code{5}.
9403 Set the threshold for duplicate detection. If the difference metric for a frame
9404 is less than or equal to this value, then it is declared as duplicate. Default
9408 Set scene change threshold. Default is @code{15}.
9412 Set the size of the x and y-axis blocks used during metric calculations.
9413 Larger blocks give better noise suppression, but also give worse detection of
9414 small movements. Must be a power of two. Default is @code{32}.
9417 Mark main input as a pre-processed input and activate clean source input
9418 stream. This allows the input to be pre-processed with various filters to help
9419 the metrics calculation while keeping the frame selection lossless. When set to
9420 @code{1}, the first stream is for the pre-processed input, and the second
9421 stream is the clean source from where the kept frames are chosen. Default is
9425 Set whether or not chroma is considered in the metric calculations. Default is
9431 Apply 2D deconvolution of video stream in frequency domain using second stream
9434 The filter accepts the following options:
9438 Set which planes to process.
9441 Set which impulse video frames will be processed, can be @var{first}
9442 or @var{all}. Default is @var{all}.
9445 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9446 and height are not same and not power of 2 or if stream prior to convolving
9450 The @code{deconvolve} filter also supports the @ref{framesync} options.
9454 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9456 It accepts the following options:
9460 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9461 @var{rainbows} for cross-color reduction.
9464 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9467 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9470 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9473 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9478 Apply deflate effect to the video.
9480 This filter replaces the pixel by the local(3x3) average by taking into account
9481 only values lower than the pixel.
9483 It accepts the following options:
9490 Limit the maximum change for each plane, default is 65535.
9491 If 0, plane will remain unchanged.
9494 @subsection Commands
9496 This filter supports the all above options as @ref{commands}.
9500 Remove temporal frame luminance variations.
9502 It accepts the following options:
9506 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9509 Set averaging mode to smooth temporal luminance variations.
9511 Available values are:
9536 Do not actually modify frame. Useful when one only wants metadata.
9541 Remove judder produced by partially interlaced telecined content.
9543 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9544 source was partially telecined content then the output of @code{pullup,dejudder}
9545 will have a variable frame rate. May change the recorded frame rate of the
9546 container. Aside from that change, this filter will not affect constant frame
9549 The option available in this filter is:
9553 Specify the length of the window over which the judder repeats.
9555 Accepts any integer greater than 1. Useful values are:
9559 If the original was telecined from 24 to 30 fps (Film to NTSC).
9562 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9565 If a mixture of the two.
9568 The default is @samp{4}.
9573 Suppress a TV station logo by a simple interpolation of the surrounding
9574 pixels. Just set a rectangle covering the logo and watch it disappear
9575 (and sometimes something even uglier appear - your mileage may vary).
9577 It accepts the following parameters:
9582 Specify the top left corner coordinates of the logo. They must be
9587 Specify the width and height of the logo to clear. They must be
9591 Specify the thickness of the fuzzy edge of the rectangle (added to
9592 @var{w} and @var{h}). The default value is 1. This option is
9593 deprecated, setting higher values should no longer be necessary and
9597 When set to 1, a green rectangle is drawn on the screen to simplify
9598 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9599 The default value is 0.
9601 The rectangle is drawn on the outermost pixels which will be (partly)
9602 replaced with interpolated values. The values of the next pixels
9603 immediately outside this rectangle in each direction will be used to
9604 compute the interpolated pixel values inside the rectangle.
9608 @subsection Examples
9612 Set a rectangle covering the area with top left corner coordinates 0,0
9613 and size 100x77, and a band of size 10:
9615 delogo=x=0:y=0:w=100:h=77:band=10
9623 Remove the rain in the input image/video by applying the derain methods based on
9624 convolutional neural networks. Supported models:
9628 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9629 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9632 Training as well as model generation scripts are provided in
9633 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9635 Native model files (.model) can be generated from TensorFlow model
9636 files (.pb) by using tools/python/convert.py
9638 The filter accepts the following options:
9642 Specify which filter to use. This option accepts the following values:
9646 Derain filter. To conduct derain filter, you need to use a derain model.
9649 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9651 Default value is @samp{derain}.
9654 Specify which DNN backend to use for model loading and execution. This option accepts
9655 the following values:
9659 Native implementation of DNN loading and execution.
9662 TensorFlow backend. To enable this backend you
9663 need to install the TensorFlow for C library (see
9664 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9665 @code{--enable-libtensorflow}
9667 Default value is @samp{native}.
9670 Set path to model file specifying network architecture and its parameters.
9671 Note that different backends use different file formats. TensorFlow and native
9672 backend can load files for only its format.
9675 It can also be finished with @ref{dnn_processing} filter.
9679 Attempt to fix small changes in horizontal and/or vertical shift. This
9680 filter helps remove camera shake from hand-holding a camera, bumping a
9681 tripod, moving on a vehicle, etc.
9683 The filter accepts the following options:
9691 Specify a rectangular area where to limit the search for motion
9693 If desired the search for motion vectors can be limited to a
9694 rectangular area of the frame defined by its top left corner, width
9695 and height. These parameters have the same meaning as the drawbox
9696 filter which can be used to visualise the position of the bounding
9699 This is useful when simultaneous movement of subjects within the frame
9700 might be confused for camera motion by the motion vector search.
9702 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9703 then the full frame is used. This allows later options to be set
9704 without specifying the bounding box for the motion vector search.
9706 Default - search the whole frame.
9710 Specify the maximum extent of movement in x and y directions in the
9711 range 0-64 pixels. Default 16.
9714 Specify how to generate pixels to fill blanks at the edge of the
9715 frame. Available values are:
9718 Fill zeroes at blank locations
9720 Original image at blank locations
9722 Extruded edge value at blank locations
9724 Mirrored edge at blank locations
9726 Default value is @samp{mirror}.
9729 Specify the blocksize to use for motion search. Range 4-128 pixels,
9733 Specify the contrast threshold for blocks. Only blocks with more than
9734 the specified contrast (difference between darkest and lightest
9735 pixels) will be considered. Range 1-255, default 125.
9738 Specify the search strategy. Available values are:
9741 Set exhaustive search
9743 Set less exhaustive search.
9745 Default value is @samp{exhaustive}.
9748 If set then a detailed log of the motion search is written to the
9755 Remove unwanted contamination of foreground colors, caused by reflected color of
9756 greenscreen or bluescreen.
9758 This filter accepts the following options:
9762 Set what type of despill to use.
9765 Set how spillmap will be generated.
9768 Set how much to get rid of still remaining spill.
9771 Controls amount of red in spill area.
9774 Controls amount of green in spill area.
9775 Should be -1 for greenscreen.
9778 Controls amount of blue in spill area.
9779 Should be -1 for bluescreen.
9782 Controls brightness of spill area, preserving colors.
9785 Modify alpha from generated spillmap.
9788 @subsection Commands
9790 This filter supports the all above options as @ref{commands}.
9794 Apply an exact inverse of the telecine operation. It requires a predefined
9795 pattern specified using the pattern option which must be the same as that passed
9796 to the telecine filter.
9798 This filter accepts the following options:
9807 The default value is @code{top}.
9811 A string of numbers representing the pulldown pattern you wish to apply.
9812 The default value is @code{23}.
9815 A number representing position of the first frame with respect to the telecine
9816 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9821 Apply dilation effect to the video.
9823 This filter replaces the pixel by the local(3x3) maximum.
9825 It accepts the following options:
9832 Limit the maximum change for each plane, default is 65535.
9833 If 0, plane will remain unchanged.
9836 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9839 Flags to local 3x3 coordinates maps like this:
9846 @subsection Commands
9848 This filter supports the all above options as @ref{commands}.
9852 Displace pixels as indicated by second and third input stream.
9854 It takes three input streams and outputs one stream, the first input is the
9855 source, and second and third input are displacement maps.
9857 The second input specifies how much to displace pixels along the
9858 x-axis, while the third input specifies how much to displace pixels
9860 If one of displacement map streams terminates, last frame from that
9861 displacement map will be used.
9863 Note that once generated, displacements maps can be reused over and over again.
9865 A description of the accepted options follows.
9869 Set displace behavior for pixels that are out of range.
9871 Available values are:
9874 Missing pixels are replaced by black pixels.
9877 Adjacent pixels will spread out to replace missing pixels.
9880 Out of range pixels are wrapped so they point to pixels of other side.
9883 Out of range pixels will be replaced with mirrored pixels.
9885 Default is @samp{smear}.
9889 @subsection Examples
9893 Add ripple effect to rgb input of video size hd720:
9895 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
9899 Add wave effect to rgb input of video size hd720:
9901 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
9905 @anchor{dnn_processing}
9906 @section dnn_processing
9908 Do image processing with deep neural networks. It works together with another filter
9909 which converts the pixel format of the Frame to what the dnn network requires.
9911 The filter accepts the following options:
9915 Specify which DNN backend to use for model loading and execution. This option accepts
9916 the following values:
9920 Native implementation of DNN loading and execution.
9923 TensorFlow backend. To enable this backend you
9924 need to install the TensorFlow for C library (see
9925 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9926 @code{--enable-libtensorflow}
9929 OpenVINO backend. To enable this backend you
9930 need to build and install the OpenVINO for C library (see
9931 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9932 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9933 be needed if the header files and libraries are not installed into system path)
9937 Default value is @samp{native}.
9940 Set path to model file specifying network architecture and its parameters.
9941 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9942 backend can load files for only its format.
9944 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9947 Set the input name of the dnn network.
9950 Set the output name of the dnn network.
9954 @subsection Examples
9958 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9960 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9964 Halve the pixel value of the frame with format gray32f:
9966 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
9970 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9972 ./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
9976 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9978 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9985 Draw a colored box on the input image.
9987 It accepts the following parameters:
9992 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9996 The expressions which specify the width and height of the box; if 0 they are interpreted as
9997 the input width and height. It defaults to 0.
10000 Specify the color of the box to write. For the general syntax of this option,
10001 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10002 value @code{invert} is used, the box edge color is the same as the
10003 video with inverted luma.
10006 The expression which sets the thickness of the box edge.
10007 A value of @code{fill} will create a filled box. Default value is @code{3}.
10009 See below for the list of accepted constants.
10012 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10013 will overwrite the video's color and alpha pixels.
10014 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10017 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10018 following constants:
10022 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10026 horizontal and vertical chroma subsample values. For example for the
10027 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10031 The input width and height.
10034 The input sample aspect ratio.
10038 The x and y offset coordinates where the box is drawn.
10042 The width and height of the drawn box.
10045 The thickness of the drawn box.
10047 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10048 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10052 @subsection Examples
10056 Draw a black box around the edge of the input image:
10062 Draw a box with color red and an opacity of 50%:
10064 drawbox=10:20:200:60:red@@0.5
10067 The previous example can be specified as:
10069 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10073 Fill the box with pink color:
10075 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10079 Draw a 2-pixel red 2.40:1 mask:
10081 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
10085 @subsection Commands
10086 This filter supports same commands as options.
10087 The command accepts the same syntax of the corresponding option.
10089 If the specified expression is not valid, it is kept at its current
10094 Draw a graph using input video metadata.
10096 It accepts the following parameters:
10100 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10103 Set 1st foreground color expression.
10106 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10109 Set 2nd foreground color expression.
10112 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10115 Set 3rd foreground color expression.
10118 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10121 Set 4th foreground color expression.
10124 Set minimal value of metadata value.
10127 Set maximal value of metadata value.
10130 Set graph background color. Default is white.
10135 Available values for mode is:
10142 Default is @code{line}.
10147 Available values for slide is:
10150 Draw new frame when right border is reached.
10153 Replace old columns with new ones.
10156 Scroll from right to left.
10159 Scroll from left to right.
10162 Draw single picture.
10165 Default is @code{frame}.
10168 Set size of graph video. For the syntax of this option, check the
10169 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10170 The default value is @code{900x256}.
10173 Set the output frame rate. Default value is @code{25}.
10175 The foreground color expressions can use the following variables:
10178 Minimal value of metadata value.
10181 Maximal value of metadata value.
10184 Current metadata key value.
10187 The color is defined as 0xAABBGGRR.
10190 Example using metadata from @ref{signalstats} filter:
10192 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10195 Example using metadata from @ref{ebur128} filter:
10197 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10202 Draw a grid on the input image.
10204 It accepts the following parameters:
10209 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10213 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10214 input width and height, respectively, minus @code{thickness}, so image gets
10215 framed. Default to 0.
10218 Specify the color of the grid. For the general syntax of this option,
10219 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10220 value @code{invert} is used, the grid color is the same as the
10221 video with inverted luma.
10224 The expression which sets the thickness of the grid line. Default value is @code{1}.
10226 See below for the list of accepted constants.
10229 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10230 will overwrite the video's color and alpha pixels.
10231 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10234 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10235 following constants:
10239 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10243 horizontal and vertical chroma subsample values. For example for the
10244 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10248 The input grid cell width and height.
10251 The input sample aspect ratio.
10255 The x and y coordinates of some point of grid intersection (meant to configure offset).
10259 The width and height of the drawn cell.
10262 The thickness of the drawn cell.
10264 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10265 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10269 @subsection Examples
10273 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10275 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10279 Draw a white 3x3 grid with an opacity of 50%:
10281 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10285 @subsection Commands
10286 This filter supports same commands as options.
10287 The command accepts the same syntax of the corresponding option.
10289 If the specified expression is not valid, it is kept at its current
10295 Draw a text string or text from a specified file on top of a video, using the
10296 libfreetype library.
10298 To enable compilation of this filter, you need to configure FFmpeg with
10299 @code{--enable-libfreetype}.
10300 To enable default font fallback and the @var{font} option you need to
10301 configure FFmpeg with @code{--enable-libfontconfig}.
10302 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10303 @code{--enable-libfribidi}.
10307 It accepts the following parameters:
10312 Used to draw a box around text using the background color.
10313 The value must be either 1 (enable) or 0 (disable).
10314 The default value of @var{box} is 0.
10317 Set the width of the border to be drawn around the box using @var{boxcolor}.
10318 The default value of @var{boxborderw} is 0.
10321 The color to be used for drawing box around text. For the syntax of this
10322 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10324 The default value of @var{boxcolor} is "white".
10327 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10328 The default value of @var{line_spacing} is 0.
10331 Set the width of the border to be drawn around the text using @var{bordercolor}.
10332 The default value of @var{borderw} is 0.
10335 Set the color to be used for drawing border around text. For the syntax of this
10336 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10338 The default value of @var{bordercolor} is "black".
10341 Select how the @var{text} is expanded. Can be either @code{none},
10342 @code{strftime} (deprecated) or
10343 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10347 Set a start time for the count. Value is in microseconds. Only applied
10348 in the deprecated strftime expansion mode. To emulate in normal expansion
10349 mode use the @code{pts} function, supplying the start time (in seconds)
10350 as the second argument.
10353 If true, check and fix text coords to avoid clipping.
10356 The color to be used for drawing fonts. For the syntax of this option, check
10357 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10359 The default value of @var{fontcolor} is "black".
10361 @item fontcolor_expr
10362 String which is expanded the same way as @var{text} to obtain dynamic
10363 @var{fontcolor} value. By default this option has empty value and is not
10364 processed. When this option is set, it overrides @var{fontcolor} option.
10367 The font family to be used for drawing text. By default Sans.
10370 The font file to be used for drawing text. The path must be included.
10371 This parameter is mandatory if the fontconfig support is disabled.
10374 Draw the text applying alpha blending. The value can
10375 be a number between 0.0 and 1.0.
10376 The expression accepts the same variables @var{x, y} as well.
10377 The default value is 1.
10378 Please see @var{fontcolor_expr}.
10381 The font size to be used for drawing text.
10382 The default value of @var{fontsize} is 16.
10385 If set to 1, attempt to shape the text (for example, reverse the order of
10386 right-to-left text and join Arabic characters) before drawing it.
10387 Otherwise, just draw the text exactly as given.
10388 By default 1 (if supported).
10390 @item ft_load_flags
10391 The flags to be used for loading the fonts.
10393 The flags map the corresponding flags supported by libfreetype, and are
10394 a combination of the following values:
10401 @item vertical_layout
10402 @item force_autohint
10405 @item ignore_global_advance_width
10407 @item ignore_transform
10409 @item linear_design
10413 Default value is "default".
10415 For more information consult the documentation for the FT_LOAD_*
10419 The color to be used for drawing a shadow behind the drawn text. For the
10420 syntax of this option, check the @ref{color syntax,,"Color" section in the
10421 ffmpeg-utils manual,ffmpeg-utils}.
10423 The default value of @var{shadowcolor} is "black".
10427 The x and y offsets for the text shadow position with respect to the
10428 position of the text. They can be either positive or negative
10429 values. The default value for both is "0".
10432 The starting frame number for the n/frame_num variable. The default value
10436 The size in number of spaces to use for rendering the tab.
10437 Default value is 4.
10440 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10441 format. It can be used with or without text parameter. @var{timecode_rate}
10442 option must be specified.
10444 @item timecode_rate, rate, r
10445 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10446 integer. Minimum value is "1".
10447 Drop-frame timecode is supported for frame rates 30 & 60.
10450 If set to 1, the output of the timecode option will wrap around at 24 hours.
10451 Default is 0 (disabled).
10454 The text string to be drawn. The text must be a sequence of UTF-8
10455 encoded characters.
10456 This parameter is mandatory if no file is specified with the parameter
10460 A text file containing text to be drawn. The text must be a sequence
10461 of UTF-8 encoded characters.
10463 This parameter is mandatory if no text string is specified with the
10464 parameter @var{text}.
10466 If both @var{text} and @var{textfile} are specified, an error is thrown.
10469 If set to 1, the @var{textfile} will be reloaded before each frame.
10470 Be sure to update it atomically, or it may be read partially, or even fail.
10474 The expressions which specify the offsets where text will be drawn
10475 within the video frame. They are relative to the top/left border of the
10478 The default value of @var{x} and @var{y} is "0".
10480 See below for the list of accepted constants and functions.
10483 The parameters for @var{x} and @var{y} are expressions containing the
10484 following constants and functions:
10488 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10492 horizontal and vertical chroma subsample values. For example for the
10493 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10496 the height of each text line
10504 @item max_glyph_a, ascent
10505 the maximum distance from the baseline to the highest/upper grid
10506 coordinate used to place a glyph outline point, for all the rendered
10508 It is a positive value, due to the grid's orientation with the Y axis
10511 @item max_glyph_d, descent
10512 the maximum distance from the baseline to the lowest grid coordinate
10513 used to place a glyph outline point, for all the rendered glyphs.
10514 This is a negative value, due to the grid's orientation, with the Y axis
10518 maximum glyph height, that is the maximum height for all the glyphs
10519 contained in the rendered text, it is equivalent to @var{ascent} -
10523 maximum glyph width, that is the maximum width for all the glyphs
10524 contained in the rendered text
10527 the number of input frame, starting from 0
10529 @item rand(min, max)
10530 return a random number included between @var{min} and @var{max}
10533 The input sample aspect ratio.
10536 timestamp expressed in seconds, NAN if the input timestamp is unknown
10539 the height of the rendered text
10542 the width of the rendered text
10546 the x and y offset coordinates where the text is drawn.
10548 These parameters allow the @var{x} and @var{y} expressions to refer
10549 to each other, so you can for example specify @code{y=x/dar}.
10552 A one character description of the current frame's picture type.
10555 The current packet's position in the input file or stream
10556 (in bytes, from the start of the input). A value of -1 indicates
10557 this info is not available.
10560 The current packet's duration, in seconds.
10563 The current packet's size (in bytes).
10566 @anchor{drawtext_expansion}
10567 @subsection Text expansion
10569 If @option{expansion} is set to @code{strftime},
10570 the filter recognizes strftime() sequences in the provided text and
10571 expands them accordingly. Check the documentation of strftime(). This
10572 feature is deprecated.
10574 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10576 If @option{expansion} is set to @code{normal} (which is the default),
10577 the following expansion mechanism is used.
10579 The backslash character @samp{\}, followed by any character, always expands to
10580 the second character.
10582 Sequences of the form @code{%@{...@}} are expanded. The text between the
10583 braces is a function name, possibly followed by arguments separated by ':'.
10584 If the arguments contain special characters or delimiters (':' or '@}'),
10585 they should be escaped.
10587 Note that they probably must also be escaped as the value for the
10588 @option{text} option in the filter argument string and as the filter
10589 argument in the filtergraph description, and possibly also for the shell,
10590 that makes up to four levels of escaping; using a text file avoids these
10593 The following functions are available:
10598 The expression evaluation result.
10600 It must take one argument specifying the expression to be evaluated,
10601 which accepts the same constants and functions as the @var{x} and
10602 @var{y} values. Note that not all constants should be used, for
10603 example the text size is not known when evaluating the expression, so
10604 the constants @var{text_w} and @var{text_h} will have an undefined
10607 @item expr_int_format, eif
10608 Evaluate the expression's value and output as formatted integer.
10610 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10611 The second argument specifies the output format. Allowed values are @samp{x},
10612 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10613 @code{printf} function.
10614 The third parameter is optional and sets the number of positions taken by the output.
10615 It can be used to add padding with zeros from the left.
10618 The time at which the filter is running, expressed in UTC.
10619 It can accept an argument: a strftime() format string.
10622 The time at which the filter is running, expressed in the local time zone.
10623 It can accept an argument: a strftime() format string.
10626 Frame metadata. Takes one or two arguments.
10628 The first argument is mandatory and specifies the metadata key.
10630 The second argument is optional and specifies a default value, used when the
10631 metadata key is not found or empty.
10633 Available metadata can be identified by inspecting entries
10634 starting with TAG included within each frame section
10635 printed by running @code{ffprobe -show_frames}.
10637 String metadata generated in filters leading to
10638 the drawtext filter are also available.
10641 The frame number, starting from 0.
10644 A one character description of the current picture type.
10647 The timestamp of the current frame.
10648 It can take up to three arguments.
10650 The first argument is the format of the timestamp; it defaults to @code{flt}
10651 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10652 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10653 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10654 @code{localtime} stands for the timestamp of the frame formatted as
10655 local time zone time.
10657 The second argument is an offset added to the timestamp.
10659 If the format is set to @code{hms}, a third argument @code{24HH} may be
10660 supplied to present the hour part of the formatted timestamp in 24h format
10663 If the format is set to @code{localtime} or @code{gmtime},
10664 a third argument may be supplied: a strftime() format string.
10665 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10668 @subsection Commands
10670 This filter supports altering parameters via commands:
10673 Alter existing filter parameters.
10675 Syntax for the argument is the same as for filter invocation, e.g.
10678 fontsize=56:fontcolor=green:text='Hello World'
10681 Full filter invocation with sendcmd would look like this:
10684 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10688 If the entire argument can't be parsed or applied as valid values then the filter will
10689 continue with its existing parameters.
10691 @subsection Examples
10695 Draw "Test Text" with font FreeSerif, using the default values for the
10696 optional parameters.
10699 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10703 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10704 and y=50 (counting from the top-left corner of the screen), text is
10705 yellow with a red box around it. Both the text and the box have an
10709 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10710 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10713 Note that the double quotes are not necessary if spaces are not used
10714 within the parameter list.
10717 Show the text at the center of the video frame:
10719 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10723 Show the text at a random position, switching to a new position every 30 seconds:
10725 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)"
10729 Show a text line sliding from right to left in the last row of the video
10730 frame. The file @file{LONG_LINE} is assumed to contain a single line
10733 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10737 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10739 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10743 Draw a single green letter "g", at the center of the input video.
10744 The glyph baseline is placed at half screen height.
10746 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10750 Show text for 1 second every 3 seconds:
10752 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10756 Use fontconfig to set the font. Note that the colons need to be escaped.
10758 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10762 Draw "Test Text" with font size dependent on height of the video.
10764 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10768 Print the date of a real-time encoding (see strftime(3)):
10770 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10774 Show text fading in and out (appearing/disappearing):
10777 DS=1.0 # display start
10778 DE=10.0 # display end
10779 FID=1.5 # fade in duration
10780 FOD=5 # fade out duration
10781 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 @}"
10785 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10786 and the @option{fontsize} value are included in the @option{y} offset.
10788 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10789 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10793 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10794 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10795 must have option @option{-export_path_metadata 1} for the special metadata fields
10796 to be available for filters.
10798 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10803 For more information about libfreetype, check:
10804 @url{http://www.freetype.org/}.
10806 For more information about fontconfig, check:
10807 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10809 For more information about libfribidi, check:
10810 @url{http://fribidi.org/}.
10812 @section edgedetect
10814 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10816 The filter accepts the following options:
10821 Set low and high threshold values used by the Canny thresholding
10824 The high threshold selects the "strong" edge pixels, which are then
10825 connected through 8-connectivity with the "weak" edge pixels selected
10826 by the low threshold.
10828 @var{low} and @var{high} threshold values must be chosen in the range
10829 [0,1], and @var{low} should be lesser or equal to @var{high}.
10831 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10835 Define the drawing mode.
10839 Draw white/gray wires on black background.
10842 Mix the colors to create a paint/cartoon effect.
10845 Apply Canny edge detector on all selected planes.
10847 Default value is @var{wires}.
10850 Select planes for filtering. By default all available planes are filtered.
10853 @subsection Examples
10857 Standard edge detection with custom values for the hysteresis thresholding:
10859 edgedetect=low=0.1:high=0.4
10863 Painting effect without thresholding:
10865 edgedetect=mode=colormix:high=0
10871 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10873 For each input image, the filter will compute the optimal mapping from
10874 the input to the output given the codebook length, that is the number
10875 of distinct output colors.
10877 This filter accepts the following options.
10880 @item codebook_length, l
10881 Set codebook length. The value must be a positive integer, and
10882 represents the number of distinct output colors. Default value is 256.
10885 Set the maximum number of iterations to apply for computing the optimal
10886 mapping. The higher the value the better the result and the higher the
10887 computation time. Default value is 1.
10890 Set a random seed, must be an integer included between 0 and
10891 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10892 will try to use a good random seed on a best effort basis.
10895 Set pal8 output pixel format. This option does not work with codebook
10896 length greater than 256.
10901 Measure graylevel entropy in histogram of color channels of video frames.
10903 It accepts the following parameters:
10907 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10909 @var{diff} mode measures entropy of histogram delta values, absolute differences
10910 between neighbour histogram values.
10914 Set brightness, contrast, saturation and approximate gamma adjustment.
10916 The filter accepts the following options:
10920 Set the contrast expression. The value must be a float value in range
10921 @code{-1000.0} to @code{1000.0}. The default value is "1".
10924 Set the brightness expression. The value must be a float value in
10925 range @code{-1.0} to @code{1.0}. The default value is "0".
10928 Set the saturation expression. The value must be a float in
10929 range @code{0.0} to @code{3.0}. The default value is "1".
10932 Set the gamma expression. The value must be a float in range
10933 @code{0.1} to @code{10.0}. The default value is "1".
10936 Set the gamma expression for red. The value must be a float in
10937 range @code{0.1} to @code{10.0}. The default value is "1".
10940 Set the gamma expression for green. The value must be a float in range
10941 @code{0.1} to @code{10.0}. The default value is "1".
10944 Set the gamma expression for blue. The value must be a float in range
10945 @code{0.1} to @code{10.0}. The default value is "1".
10948 Set the gamma weight expression. It can be used to reduce the effect
10949 of a high gamma value on bright image areas, e.g. keep them from
10950 getting overamplified and just plain white. The value must be a float
10951 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10952 gamma correction all the way down while @code{1.0} leaves it at its
10953 full strength. Default is "1".
10956 Set when the expressions for brightness, contrast, saturation and
10957 gamma expressions are evaluated.
10959 It accepts the following values:
10962 only evaluate expressions once during the filter initialization or
10963 when a command is processed
10966 evaluate expressions for each incoming frame
10969 Default value is @samp{init}.
10972 The expressions accept the following parameters:
10975 frame count of the input frame starting from 0
10978 byte position of the corresponding packet in the input file, NAN if
10982 frame rate of the input video, NAN if the input frame rate is unknown
10985 timestamp expressed in seconds, NAN if the input timestamp is unknown
10988 @subsection Commands
10989 The filter supports the following commands:
10993 Set the contrast expression.
10996 Set the brightness expression.
10999 Set the saturation expression.
11002 Set the gamma expression.
11005 Set the gamma_r expression.
11008 Set gamma_g expression.
11011 Set gamma_b expression.
11014 Set gamma_weight expression.
11016 The command accepts the same syntax of the corresponding option.
11018 If the specified expression is not valid, it is kept at its current
11025 Apply erosion effect to the video.
11027 This filter replaces the pixel by the local(3x3) minimum.
11029 It accepts the following options:
11036 Limit the maximum change for each plane, default is 65535.
11037 If 0, plane will remain unchanged.
11040 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11043 Flags to local 3x3 coordinates maps like this:
11050 @subsection Commands
11052 This filter supports the all above options as @ref{commands}.
11054 @section extractplanes
11056 Extract color channel components from input video stream into
11057 separate grayscale video streams.
11059 The filter accepts the following option:
11063 Set plane(s) to extract.
11065 Available values for planes are:
11076 Choosing planes not available in the input will result in an error.
11077 That means you cannot select @code{r}, @code{g}, @code{b} planes
11078 with @code{y}, @code{u}, @code{v} planes at same time.
11081 @subsection Examples
11085 Extract luma, u and v color channel component from input video frame
11086 into 3 grayscale outputs:
11088 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
11094 Apply a fade-in/out effect to the input video.
11096 It accepts the following parameters:
11100 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11102 Default is @code{in}.
11104 @item start_frame, s
11105 Specify the number of the frame to start applying the fade
11106 effect at. Default is 0.
11109 The number of frames that the fade effect lasts. 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}.
11116 If set to 1, fade only alpha channel, if one exists on the input.
11117 Default value is 0.
11119 @item start_time, st
11120 Specify the timestamp (in seconds) of the frame to start to apply the fade
11121 effect. If both start_frame and start_time are specified, the fade will start at
11122 whichever comes last. Default is 0.
11125 The number of seconds for which the fade effect has to last. At the end of the
11126 fade-in effect the output video will have the same intensity as the input video,
11127 at the end of the fade-out transition the output video will be filled with the
11128 selected @option{color}.
11129 If both duration and nb_frames are specified, duration is used. Default is 0
11130 (nb_frames is used by default).
11133 Specify the color of the fade. Default is "black".
11136 @subsection Examples
11140 Fade in the first 30 frames of video:
11145 The command above is equivalent to:
11151 Fade out the last 45 frames of a 200-frame video:
11154 fade=type=out:start_frame=155:nb_frames=45
11158 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11160 fade=in:0:25, fade=out:975:25
11164 Make the first 5 frames yellow, then fade in from frame 5-24:
11166 fade=in:5:20:color=yellow
11170 Fade in alpha over first 25 frames of video:
11172 fade=in:0:25:alpha=1
11176 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11178 fade=t=in:st=5.5:d=0.5
11184 Denoise frames using 3D FFT (frequency domain filtering).
11186 The filter accepts the following options:
11190 Set the noise sigma constant. This sets denoising strength.
11191 Default value is 1. Allowed range is from 0 to 30.
11192 Using very high sigma with low overlap may give blocking artifacts.
11195 Set amount of denoising. By default all detected noise is reduced.
11196 Default value is 1. Allowed range is from 0 to 1.
11199 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11200 Actual size of block in pixels is 2 to power of @var{block}, so by default
11201 block size in pixels is 2^4 which is 16.
11204 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11207 Set number of previous frames to use for denoising. By default is set to 0.
11210 Set number of next frames to to use for denoising. By default is set to 0.
11213 Set planes which will be filtered, by default are all available filtered
11218 Apply arbitrary expressions to samples in frequency domain
11222 Adjust the dc value (gain) of the luma plane of the image. The filter
11223 accepts an integer value in range @code{0} to @code{1000}. The default
11224 value is set to @code{0}.
11227 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11228 filter accepts an integer value in range @code{0} to @code{1000}. The
11229 default value is set to @code{0}.
11232 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11233 filter accepts an integer value in range @code{0} to @code{1000}. The
11234 default value is set to @code{0}.
11237 Set the frequency domain weight expression for the luma plane.
11240 Set the frequency domain weight expression for the 1st chroma plane.
11243 Set the frequency domain weight expression for the 2nd chroma plane.
11246 Set when the expressions are evaluated.
11248 It accepts the following values:
11251 Only evaluate expressions once during the filter initialization.
11254 Evaluate expressions for each incoming frame.
11257 Default value is @samp{init}.
11259 The filter accepts the following variables:
11262 The coordinates of the current sample.
11266 The width and height of the image.
11269 The number of input frame, starting from 0.
11272 @subsection Examples
11278 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11284 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11290 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11296 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11303 Extract a single field from an interlaced image using stride
11304 arithmetic to avoid wasting CPU time. The output frames are marked as
11307 The filter accepts the following options:
11311 Specify whether to extract the top (if the value is @code{0} or
11312 @code{top}) or the bottom field (if the value is @code{1} or
11318 Create new frames by copying the top and bottom fields from surrounding frames
11319 supplied as numbers by the hint file.
11323 Set file containing hints: absolute/relative frame numbers.
11325 There must be one line for each frame in a clip. Each line must contain two
11326 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11327 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11328 is current frame number for @code{absolute} mode or out of [-1, 1] range
11329 for @code{relative} mode. First number tells from which frame to pick up top
11330 field and second number tells from which frame to pick up bottom field.
11332 If optionally followed by @code{+} output frame will be marked as interlaced,
11333 else if followed by @code{-} output frame will be marked as progressive, else
11334 it will be marked same as input frame.
11335 If optionally followed by @code{t} output frame will use only top field, or in
11336 case of @code{b} it will use only bottom field.
11337 If line starts with @code{#} or @code{;} that line is skipped.
11340 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11343 Example of first several lines of @code{hint} file for @code{relative} mode:
11345 0,0 - # first frame
11346 1,0 - # second frame, use third's frame top field and second's frame bottom field
11347 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11362 @section fieldmatch
11364 Field matching filter for inverse telecine. It is meant to reconstruct the
11365 progressive frames from a telecined stream. The filter does not drop duplicated
11366 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11367 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11369 The separation of the field matching and the decimation is notably motivated by
11370 the possibility of inserting a de-interlacing filter fallback between the two.
11371 If the source has mixed telecined and real interlaced content,
11372 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11373 But these remaining combed frames will be marked as interlaced, and thus can be
11374 de-interlaced by a later filter such as @ref{yadif} before decimation.
11376 In addition to the various configuration options, @code{fieldmatch} can take an
11377 optional second stream, activated through the @option{ppsrc} option. If
11378 enabled, the frames reconstruction will be based on the fields and frames from
11379 this second stream. This allows the first input to be pre-processed in order to
11380 help the various algorithms of the filter, while keeping the output lossless
11381 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11382 or brightness/contrast adjustments can help.
11384 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11385 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11386 which @code{fieldmatch} is based on. While the semantic and usage are very
11387 close, some behaviour and options names can differ.
11389 The @ref{decimate} filter currently only works for constant frame rate input.
11390 If your input has mixed telecined (30fps) and progressive content with a lower
11391 framerate like 24fps use the following filterchain to produce the necessary cfr
11392 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11394 The filter accepts the following options:
11398 Specify the assumed field order of the input stream. Available values are:
11402 Auto detect parity (use FFmpeg's internal parity value).
11404 Assume bottom field first.
11406 Assume top field first.
11409 Note that it is sometimes recommended not to trust the parity announced by the
11412 Default value is @var{auto}.
11415 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11416 sense that it won't risk creating jerkiness due to duplicate frames when
11417 possible, but if there are bad edits or blended fields it will end up
11418 outputting combed frames when a good match might actually exist. On the other
11419 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11420 but will almost always find a good frame if there is one. The other values are
11421 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11422 jerkiness and creating duplicate frames versus finding good matches in sections
11423 with bad edits, orphaned fields, blended fields, etc.
11425 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11427 Available values are:
11431 2-way matching (p/c)
11433 2-way matching, and trying 3rd match if still combed (p/c + n)
11435 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11437 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11438 still combed (p/c + n + u/b)
11440 3-way matching (p/c/n)
11442 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11443 detected as combed (p/c/n + u/b)
11446 The parenthesis at the end indicate the matches that would be used for that
11447 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11450 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11453 Default value is @var{pc_n}.
11456 Mark the main input stream as a pre-processed input, and enable the secondary
11457 input stream as the clean source to pick the fields from. See the filter
11458 introduction for more details. It is similar to the @option{clip2} feature from
11461 Default value is @code{0} (disabled).
11464 Set the field to match from. It is recommended to set this to the same value as
11465 @option{order} unless you experience matching failures with that setting. In
11466 certain circumstances changing the field that is used to match from can have a
11467 large impact on matching performance. Available values are:
11471 Automatic (same value as @option{order}).
11473 Match from the bottom field.
11475 Match from the top field.
11478 Default value is @var{auto}.
11481 Set whether or not chroma is included during the match comparisons. In most
11482 cases it is recommended to leave this enabled. You should set this to @code{0}
11483 only if your clip has bad chroma problems such as heavy rainbowing or other
11484 artifacts. Setting this to @code{0} could also be used to speed things up at
11485 the cost of some accuracy.
11487 Default value is @code{1}.
11491 These define an exclusion band which excludes the lines between @option{y0} and
11492 @option{y1} from being included in the field matching decision. An exclusion
11493 band can be used to ignore subtitles, a logo, or other things that may
11494 interfere with the matching. @option{y0} sets the starting scan line and
11495 @option{y1} sets the ending line; all lines in between @option{y0} and
11496 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11497 @option{y0} and @option{y1} to the same value will disable the feature.
11498 @option{y0} and @option{y1} defaults to @code{0}.
11501 Set the scene change detection threshold as a percentage of maximum change on
11502 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11503 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11504 @option{scthresh} is @code{[0.0, 100.0]}.
11506 Default value is @code{12.0}.
11509 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11510 account the combed scores of matches when deciding what match to use as the
11511 final match. Available values are:
11515 No final matching based on combed scores.
11517 Combed scores are only used when a scene change is detected.
11519 Use combed scores all the time.
11522 Default is @var{sc}.
11525 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11526 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11527 Available values are:
11531 No forced calculation.
11533 Force p/c/n calculations.
11535 Force p/c/n/u/b calculations.
11538 Default value is @var{none}.
11541 This is the area combing threshold used for combed frame detection. This
11542 essentially controls how "strong" or "visible" combing must be to be detected.
11543 Larger values mean combing must be more visible and smaller values mean combing
11544 can be less visible or strong and still be detected. Valid settings are from
11545 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11546 be detected as combed). This is basically a pixel difference value. A good
11547 range is @code{[8, 12]}.
11549 Default value is @code{9}.
11552 Sets whether or not chroma is considered in the combed frame decision. Only
11553 disable this if your source has chroma problems (rainbowing, etc.) that are
11554 causing problems for the combed frame detection with chroma enabled. Actually,
11555 using @option{chroma}=@var{0} is usually more reliable, except for the case
11556 where there is chroma only combing in the source.
11558 Default value is @code{0}.
11562 Respectively set the x-axis and y-axis size of the window used during combed
11563 frame detection. This has to do with the size of the area in which
11564 @option{combpel} pixels are required to be detected as combed for a frame to be
11565 declared combed. See the @option{combpel} parameter description for more info.
11566 Possible values are any number that is a power of 2 starting at 4 and going up
11569 Default value is @code{16}.
11572 The number of combed pixels inside any of the @option{blocky} by
11573 @option{blockx} size blocks on the frame for the frame to be detected as
11574 combed. While @option{cthresh} controls how "visible" the combing must be, this
11575 setting controls "how much" combing there must be in any localized area (a
11576 window defined by the @option{blockx} and @option{blocky} settings) on the
11577 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11578 which point no frames will ever be detected as combed). This setting is known
11579 as @option{MI} in TFM/VFM vocabulary.
11581 Default value is @code{80}.
11584 @anchor{p/c/n/u/b meaning}
11585 @subsection p/c/n/u/b meaning
11587 @subsubsection p/c/n
11589 We assume the following telecined stream:
11592 Top fields: 1 2 2 3 4
11593 Bottom fields: 1 2 3 4 4
11596 The numbers correspond to the progressive frame the fields relate to. Here, the
11597 first two frames are progressive, the 3rd and 4th are combed, and so on.
11599 When @code{fieldmatch} is configured to run a matching from bottom
11600 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11605 B 1 2 3 4 4 <-- matching reference
11614 As a result of the field matching, we can see that some frames get duplicated.
11615 To perform a complete inverse telecine, you need to rely on a decimation filter
11616 after this operation. See for instance the @ref{decimate} filter.
11618 The same operation now matching from top fields (@option{field}=@var{top})
11623 T 1 2 2 3 4 <-- matching reference
11633 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11634 basically, they refer to the frame and field of the opposite parity:
11637 @item @var{p} matches the field of the opposite parity in the previous frame
11638 @item @var{c} matches the field of the opposite parity in the current frame
11639 @item @var{n} matches the field of the opposite parity in the next frame
11644 The @var{u} and @var{b} matching are a bit special in the sense that they match
11645 from the opposite parity flag. In the following examples, we assume that we are
11646 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11647 'x' is placed above and below each matched fields.
11649 With bottom matching (@option{field}=@var{bottom}):
11654 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11655 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11663 With top matching (@option{field}=@var{top}):
11668 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11669 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11677 @subsection Examples
11679 Simple IVTC of a top field first telecined stream:
11681 fieldmatch=order=tff:combmatch=none, decimate
11684 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11686 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11689 @section fieldorder
11691 Transform the field order of the input video.
11693 It accepts the following parameters:
11698 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11699 for bottom field first.
11702 The default value is @samp{tff}.
11704 The transformation is done by shifting the picture content up or down
11705 by one line, and filling the remaining line with appropriate picture content.
11706 This method is consistent with most broadcast field order converters.
11708 If the input video is not flagged as being interlaced, or it is already
11709 flagged as being of the required output field order, then this filter does
11710 not alter the incoming video.
11712 It is very useful when converting to or from PAL DV material,
11713 which is bottom field first.
11717 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11720 @section fifo, afifo
11722 Buffer input images and send them when they are requested.
11724 It is mainly useful when auto-inserted by the libavfilter
11727 It does not take parameters.
11729 @section fillborders
11731 Fill borders of the input video, without changing video stream dimensions.
11732 Sometimes video can have garbage at the four edges and you may not want to
11733 crop video input to keep size multiple of some number.
11735 This filter accepts the following options:
11739 Number of pixels to fill from left border.
11742 Number of pixels to fill from right border.
11745 Number of pixels to fill from top border.
11748 Number of pixels to fill from bottom border.
11753 It accepts the following values:
11756 fill pixels using outermost pixels
11759 fill pixels using mirroring
11762 fill pixels with constant value
11765 Default is @var{smear}.
11768 Set color for pixels in fixed mode. Default is @var{black}.
11771 @subsection Commands
11772 This filter supports same @ref{commands} as options.
11773 The command accepts the same syntax of the corresponding option.
11775 If the specified expression is not valid, it is kept at its current
11780 Find a rectangular object
11782 It accepts the following options:
11786 Filepath of the object image, needs to be in gray8.
11789 Detection threshold, default is 0.5.
11792 Number of mipmaps, default is 3.
11794 @item xmin, ymin, xmax, ymax
11795 Specifies the rectangle in which to search.
11798 @subsection Examples
11802 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11804 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11810 Flood area with values of same pixel components with another values.
11812 It accepts the following options:
11815 Set pixel x coordinate.
11818 Set pixel y coordinate.
11821 Set source #0 component value.
11824 Set source #1 component value.
11827 Set source #2 component value.
11830 Set source #3 component value.
11833 Set destination #0 component value.
11836 Set destination #1 component value.
11839 Set destination #2 component value.
11842 Set destination #3 component value.
11848 Convert the input video to one of the specified pixel formats.
11849 Libavfilter will try to pick one that is suitable as input to
11852 It accepts the following parameters:
11856 A '|'-separated list of pixel format names, such as
11857 "pix_fmts=yuv420p|monow|rgb24".
11861 @subsection Examples
11865 Convert the input video to the @var{yuv420p} format
11867 format=pix_fmts=yuv420p
11870 Convert the input video to any of the formats in the list
11872 format=pix_fmts=yuv420p|yuv444p|yuv410p
11879 Convert the video to specified constant frame rate by duplicating or dropping
11880 frames as necessary.
11882 It accepts the following parameters:
11886 The desired output frame rate. The default is @code{25}.
11889 Assume the first PTS should be the given value, in seconds. This allows for
11890 padding/trimming at the start of stream. By default, no assumption is made
11891 about the first frame's expected PTS, so no padding or trimming is done.
11892 For example, this could be set to 0 to pad the beginning with duplicates of
11893 the first frame if a video stream starts after the audio stream or to trim any
11894 frames with a negative PTS.
11897 Timestamp (PTS) rounding method.
11899 Possible values are:
11906 round towards -infinity
11908 round towards +infinity
11912 The default is @code{near}.
11915 Action performed when reading the last frame.
11917 Possible values are:
11920 Use same timestamp rounding method as used for other frames.
11922 Pass through last frame if input duration has not been reached yet.
11924 The default is @code{round}.
11928 Alternatively, the options can be specified as a flat string:
11929 @var{fps}[:@var{start_time}[:@var{round}]].
11931 See also the @ref{setpts} filter.
11933 @subsection Examples
11937 A typical usage in order to set the fps to 25:
11943 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11945 fps=fps=film:round=near
11951 Pack two different video streams into a stereoscopic video, setting proper
11952 metadata on supported codecs. The two views should have the same size and
11953 framerate and processing will stop when the shorter video ends. Please note
11954 that you may conveniently adjust view properties with the @ref{scale} and
11957 It accepts the following parameters:
11961 The desired packing format. Supported values are:
11966 The views are next to each other (default).
11969 The views are on top of each other.
11972 The views are packed by line.
11975 The views are packed by column.
11978 The views are temporally interleaved.
11987 # Convert left and right views into a frame-sequential video
11988 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11990 # Convert views into a side-by-side video with the same output resolution as the input
11991 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
11996 Change the frame rate by interpolating new video output frames from the source
11999 This filter is not designed to function correctly with interlaced media. If
12000 you wish to change the frame rate of interlaced media then you are required
12001 to deinterlace before this filter and re-interlace after this filter.
12003 A description of the accepted options follows.
12007 Specify the output frames per second. This option can also be specified
12008 as a value alone. The default is @code{50}.
12011 Specify the start of a range where the output frame will be created as a
12012 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12013 the default is @code{15}.
12016 Specify the end of a range where the output frame will be created as a
12017 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12018 the default is @code{240}.
12021 Specify the level at which a scene change is detected as a value between
12022 0 and 100 to indicate a new scene; a low value reflects a low
12023 probability for the current frame to introduce a new scene, while a higher
12024 value means the current frame is more likely to be one.
12025 The default is @code{8.2}.
12028 Specify flags influencing the filter process.
12030 Available value for @var{flags} is:
12033 @item scene_change_detect, scd
12034 Enable scene change detection using the value of the option @var{scene}.
12035 This flag is enabled by default.
12041 Select one frame every N-th frame.
12043 This filter accepts the following option:
12046 Select frame after every @code{step} frames.
12047 Allowed values are positive integers higher than 0. Default value is @code{1}.
12050 @section freezedetect
12052 Detect frozen video.
12054 This filter logs a message and sets frame metadata when it detects that the
12055 input video has no significant change in content during a specified duration.
12056 Video freeze detection calculates the mean average absolute difference of all
12057 the components of video frames and compares it to a noise floor.
12059 The printed times and duration are expressed in seconds. The
12060 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12061 whose timestamp equals or exceeds the detection duration and it contains the
12062 timestamp of the first frame of the freeze. The
12063 @code{lavfi.freezedetect.freeze_duration} and
12064 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12067 The filter accepts the following options:
12071 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12072 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12076 Set freeze duration until notification (default is 2 seconds).
12079 @section freezeframes
12081 Freeze video frames.
12083 This filter freezes video frames using frame from 2nd input.
12085 The filter accepts the following options:
12089 Set number of first frame from which to start freeze.
12092 Set number of last frame from which to end freeze.
12095 Set number of frame from 2nd input which will be used instead of replaced frames.
12101 Apply a frei0r effect to the input video.
12103 To enable the compilation of this filter, you need to install the frei0r
12104 header and configure FFmpeg with @code{--enable-frei0r}.
12106 It accepts the following parameters:
12111 The name of the frei0r effect to load. If the environment variable
12112 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12113 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12114 Otherwise, the standard frei0r paths are searched, in this order:
12115 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12116 @file{/usr/lib/frei0r-1/}.
12118 @item filter_params
12119 A '|'-separated list of parameters to pass to the frei0r effect.
12123 A frei0r effect parameter can be a boolean (its value is either
12124 "y" or "n"), a double, a color (specified as
12125 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12126 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12127 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12128 a position (specified as @var{X}/@var{Y}, where
12129 @var{X} and @var{Y} are floating point numbers) and/or a string.
12131 The number and types of parameters depend on the loaded effect. If an
12132 effect parameter is not specified, the default value is set.
12134 @subsection Examples
12138 Apply the distort0r effect, setting the first two double parameters:
12140 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12144 Apply the colordistance effect, taking a color as the first parameter:
12146 frei0r=colordistance:0.2/0.3/0.4
12147 frei0r=colordistance:violet
12148 frei0r=colordistance:0x112233
12152 Apply the perspective effect, specifying the top left and top right image
12155 frei0r=perspective:0.2/0.2|0.8/0.2
12159 For more information, see
12160 @url{http://frei0r.dyne.org}
12162 @subsection Commands
12164 This filter supports the @option{filter_params} option as @ref{commands}.
12168 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12170 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12171 processing filter, one of them is performed once per block, not per pixel.
12172 This allows for much higher speed.
12174 The filter accepts the following options:
12178 Set quality. This option defines the number of levels for averaging. It accepts
12179 an integer in the range 4-5. Default value is @code{4}.
12182 Force a constant quantization parameter. It accepts an integer in range 0-63.
12183 If not set, the filter will use the QP from the video stream (if available).
12186 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12187 more details but also more artifacts, while higher values make the image smoother
12188 but also blurrier. Default value is @code{0} − PSNR optimal.
12190 @item use_bframe_qp
12191 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12192 option may cause flicker since the B-Frames have often larger QP. Default is
12193 @code{0} (not enabled).
12199 Apply Gaussian blur filter.
12201 The filter accepts the following options:
12205 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12208 Set number of steps for Gaussian approximation. Default is @code{1}.
12211 Set which planes to filter. By default all planes are filtered.
12214 Set vertical sigma, if negative it will be same as @code{sigma}.
12215 Default is @code{-1}.
12218 @subsection Commands
12219 This filter supports same commands as options.
12220 The command accepts the same syntax of the corresponding option.
12222 If the specified expression is not valid, it is kept at its current
12227 Apply generic equation to each pixel.
12229 The filter accepts the following options:
12232 @item lum_expr, lum
12233 Set the luminance expression.
12235 Set the chrominance blue expression.
12237 Set the chrominance red expression.
12238 @item alpha_expr, a
12239 Set the alpha expression.
12241 Set the red expression.
12242 @item green_expr, g
12243 Set the green expression.
12245 Set the blue expression.
12248 The colorspace is selected according to the specified options. If one
12249 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12250 options is specified, the filter will automatically select a YCbCr
12251 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12252 @option{blue_expr} options is specified, it will select an RGB
12255 If one of the chrominance expression is not defined, it falls back on the other
12256 one. If no alpha expression is specified it will evaluate to opaque value.
12257 If none of chrominance expressions are specified, they will evaluate
12258 to the luminance expression.
12260 The expressions can use the following variables and functions:
12264 The sequential number of the filtered frame, starting from @code{0}.
12268 The coordinates of the current sample.
12272 The width and height of the image.
12276 Width and height scale depending on the currently filtered plane. It is the
12277 ratio between the corresponding luma plane number of pixels and the current
12278 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12279 @code{0.5,0.5} for chroma planes.
12282 Time of the current frame, expressed in seconds.
12285 Return the value of the pixel at location (@var{x},@var{y}) of the current
12289 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12293 Return the value of the pixel at location (@var{x},@var{y}) of the
12294 blue-difference chroma plane. Return 0 if there is no such plane.
12297 Return the value of the pixel at location (@var{x},@var{y}) of the
12298 red-difference chroma plane. Return 0 if there is no such plane.
12303 Return the value of the pixel at location (@var{x},@var{y}) of the
12304 red/green/blue component. Return 0 if there is no such component.
12307 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12308 plane. Return 0 if there is no such plane.
12310 @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)
12311 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12312 sums of samples within a rectangle. See the functions without the sum postfix.
12314 @item interpolation
12315 Set one of interpolation methods:
12320 Default is bilinear.
12323 For functions, if @var{x} and @var{y} are outside the area, the value will be
12324 automatically clipped to the closer edge.
12326 Please note that this filter can use multiple threads in which case each slice
12327 will have its own expression state. If you want to use only a single expression
12328 state because your expressions depend on previous state then you should limit
12329 the number of filter threads to 1.
12331 @subsection Examples
12335 Flip the image horizontally:
12341 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12342 wavelength of 100 pixels:
12344 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12348 Generate a fancy enigmatic moving light:
12350 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
12354 Generate a quick emboss effect:
12356 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12360 Modify RGB components depending on pixel position:
12362 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12366 Create a radial gradient that is the same size as the input (also see
12367 the @ref{vignette} filter):
12369 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12375 Fix the banding artifacts that are sometimes introduced into nearly flat
12376 regions by truncation to 8-bit color depth.
12377 Interpolate the gradients that should go where the bands are, and
12380 It is designed for playback only. Do not use it prior to
12381 lossy compression, because compression tends to lose the dither and
12382 bring back the bands.
12384 It accepts the following parameters:
12389 The maximum amount by which the filter will change any one pixel. This is also
12390 the threshold for detecting nearly flat regions. Acceptable values range from
12391 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12395 The neighborhood to fit the gradient to. A larger radius makes for smoother
12396 gradients, but also prevents the filter from modifying the pixels near detailed
12397 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12398 values will be clipped to the valid range.
12402 Alternatively, the options can be specified as a flat string:
12403 @var{strength}[:@var{radius}]
12405 @subsection Examples
12409 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12415 Specify radius, omitting the strength (which will fall-back to the default
12423 @anchor{graphmonitor}
12424 @section graphmonitor
12425 Show various filtergraph stats.
12427 With this filter one can debug complete filtergraph.
12428 Especially issues with links filling with queued frames.
12430 The filter accepts the following options:
12434 Set video output size. Default is @var{hd720}.
12437 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12440 Set output mode, can be @var{fulll} or @var{compact}.
12441 In @var{compact} mode only filters with some queued frames have displayed stats.
12444 Set flags which enable which stats are shown in video.
12446 Available values for flags are:
12449 Display number of queued frames in each link.
12451 @item frame_count_in
12452 Display number of frames taken from filter.
12454 @item frame_count_out
12455 Display number of frames given out from filter.
12458 Display current filtered frame pts.
12461 Display current filtered frame time.
12464 Display time base for filter link.
12467 Display used format for filter link.
12470 Display video size or number of audio channels in case of audio used by filter link.
12473 Display video frame rate or sample rate in case of audio used by filter link.
12476 Display link output status.
12480 Set upper limit for video rate of output stream, Default value is @var{25}.
12481 This guarantee that output video frame rate will not be higher than this value.
12485 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12486 and corrects the scene colors accordingly.
12488 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12490 The filter accepts the following options:
12494 The order of differentiation to be applied on the scene. Must be chosen in the range
12495 [0,2] and default value is 1.
12498 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12499 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12500 max value instead of calculating Minkowski distance.
12503 The standard deviation of Gaussian blur to be applied on the scene. Must be
12504 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12505 can't be equal to 0 if @var{difford} is greater than 0.
12508 @subsection Examples
12514 greyedge=difford=1:minknorm=5:sigma=2
12520 greyedge=difford=1:minknorm=0:sigma=2
12528 Apply a Hald CLUT to a video stream.
12530 First input is the video stream to process, and second one is the Hald CLUT.
12531 The Hald CLUT input can be a simple picture or a complete video stream.
12533 The filter accepts the following options:
12537 Force termination when the shortest input terminates. Default is @code{0}.
12539 Continue applying the last CLUT after the end of the stream. A value of
12540 @code{0} disable the filter after the last frame of the CLUT is reached.
12541 Default is @code{1}.
12544 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12545 filters share the same internals).
12547 This filter also supports the @ref{framesync} options.
12549 More information about the Hald CLUT can be found on Eskil Steenberg's website
12550 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12552 @subsection Workflow examples
12554 @subsubsection Hald CLUT video stream
12556 Generate an identity Hald CLUT stream altered with various effects:
12558 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
12561 Note: make sure you use a lossless codec.
12563 Then use it with @code{haldclut} to apply it on some random stream:
12565 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12568 The Hald CLUT will be applied to the 10 first seconds (duration of
12569 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12570 to the remaining frames of the @code{mandelbrot} stream.
12572 @subsubsection Hald CLUT with preview
12574 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12575 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12576 biggest possible square starting at the top left of the picture. The remaining
12577 padding pixels (bottom or right) will be ignored. This area can be used to add
12578 a preview of the Hald CLUT.
12580 Typically, the following generated Hald CLUT will be supported by the
12581 @code{haldclut} filter:
12584 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12585 pad=iw+320 [padded_clut];
12586 smptebars=s=320x256, split [a][b];
12587 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12588 [main][b] overlay=W-320" -frames:v 1 clut.png
12591 It contains the original and a preview of the effect of the CLUT: SMPTE color
12592 bars are displayed on the right-top, and below the same color bars processed by
12595 Then, the effect of this Hald CLUT can be visualized with:
12597 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12602 Flip the input video horizontally.
12604 For example, to horizontally flip the input video with @command{ffmpeg}:
12606 ffmpeg -i in.avi -vf "hflip" out.avi
12610 This filter applies a global color histogram equalization on a
12613 It can be used to correct video that has a compressed range of pixel
12614 intensities. The filter redistributes the pixel intensities to
12615 equalize their distribution across the intensity range. It may be
12616 viewed as an "automatically adjusting contrast filter". This filter is
12617 useful only for correcting degraded or poorly captured source
12620 The filter accepts the following options:
12624 Determine the amount of equalization to be applied. As the strength
12625 is reduced, the distribution of pixel intensities more-and-more
12626 approaches that of the input frame. The value must be a float number
12627 in the range [0,1] and defaults to 0.200.
12630 Set the maximum intensity that can generated and scale the output
12631 values appropriately. The strength should be set as desired and then
12632 the intensity can be limited if needed to avoid washing-out. The value
12633 must be a float number in the range [0,1] and defaults to 0.210.
12636 Set the antibanding level. If enabled the filter will randomly vary
12637 the luminance of output pixels by a small amount to avoid banding of
12638 the histogram. Possible values are @code{none}, @code{weak} or
12639 @code{strong}. It defaults to @code{none}.
12645 Compute and draw a color distribution histogram for the input video.
12647 The computed histogram is a representation of the color component
12648 distribution in an image.
12650 Standard histogram displays the color components distribution in an image.
12651 Displays color graph for each color component. Shows distribution of
12652 the Y, U, V, A or R, G, B components, depending on input format, in the
12653 current frame. Below each graph a color component scale meter is shown.
12655 The filter accepts the following options:
12659 Set height of level. Default value is @code{200}.
12660 Allowed range is [50, 2048].
12663 Set height of color scale. Default value is @code{12}.
12664 Allowed range is [0, 40].
12668 It accepts the following values:
12671 Per color component graphs are placed below each other.
12674 Per color component graphs are placed side by side.
12677 Presents information identical to that in the @code{parade}, except
12678 that the graphs representing color components are superimposed directly
12681 Default is @code{stack}.
12684 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12685 Default is @code{linear}.
12688 Set what color components to display.
12689 Default is @code{7}.
12692 Set foreground opacity. Default is @code{0.7}.
12695 Set background opacity. Default is @code{0.5}.
12698 @subsection Examples
12703 Calculate and draw histogram:
12705 ffplay -i input -vf histogram
12713 This is a high precision/quality 3d denoise filter. It aims to reduce
12714 image noise, producing smooth images and making still images really
12715 still. It should enhance compressibility.
12717 It accepts the following optional parameters:
12721 A non-negative floating point number which specifies spatial luma strength.
12722 It defaults to 4.0.
12724 @item chroma_spatial
12725 A non-negative floating point number which specifies spatial chroma strength.
12726 It defaults to 3.0*@var{luma_spatial}/4.0.
12729 A floating point number which specifies luma temporal strength. It defaults to
12730 6.0*@var{luma_spatial}/4.0.
12733 A floating point number which specifies chroma temporal strength. It defaults to
12734 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12737 @subsection Commands
12738 This filter supports same @ref{commands} as options.
12739 The command accepts the same syntax of the corresponding option.
12741 If the specified expression is not valid, it is kept at its current
12744 @anchor{hwdownload}
12745 @section hwdownload
12747 Download hardware frames to system memory.
12749 The input must be in hardware frames, and the output a non-hardware format.
12750 Not all formats will be supported on the output - it may be necessary to insert
12751 an additional @option{format} filter immediately following in the graph to get
12752 the output in a supported format.
12756 Map hardware frames to system memory or to another device.
12758 This filter has several different modes of operation; which one is used depends
12759 on the input and output formats:
12762 Hardware frame input, normal frame output
12764 Map the input frames to system memory and pass them to the output. If the
12765 original hardware frame is later required (for example, after overlaying
12766 something else on part of it), the @option{hwmap} filter can be used again
12767 in the next mode to retrieve it.
12769 Normal frame input, hardware frame output
12771 If the input is actually a software-mapped hardware frame, then unmap it -
12772 that is, return the original hardware frame.
12774 Otherwise, a device must be provided. Create new hardware surfaces on that
12775 device for the output, then map them back to the software format at the input
12776 and give those frames to the preceding filter. This will then act like the
12777 @option{hwupload} filter, but may be able to avoid an additional copy when
12778 the input is already in a compatible format.
12780 Hardware frame input and output
12782 A device must be supplied for the output, either directly or with the
12783 @option{derive_device} option. The input and output devices must be of
12784 different types and compatible - the exact meaning of this is
12785 system-dependent, but typically it means that they must refer to the same
12786 underlying hardware context (for example, refer to the same graphics card).
12788 If the input frames were originally created on the output device, then unmap
12789 to retrieve the original frames.
12791 Otherwise, map the frames to the output device - create new hardware frames
12792 on the output corresponding to the frames on the input.
12795 The following additional parameters are accepted:
12799 Set the frame mapping mode. Some combination of:
12802 The mapped frame should be readable.
12804 The mapped frame should be writeable.
12806 The mapping will always overwrite the entire frame.
12808 This may improve performance in some cases, as the original contents of the
12809 frame need not be loaded.
12811 The mapping must not involve any copying.
12813 Indirect mappings to copies of frames are created in some cases where either
12814 direct mapping is not possible or it would have unexpected properties.
12815 Setting this flag ensures that the mapping is direct and will fail if that is
12818 Defaults to @var{read+write} if not specified.
12820 @item derive_device @var{type}
12821 Rather than using the device supplied at initialisation, instead derive a new
12822 device of type @var{type} from the device the input frames exist on.
12825 In a hardware to hardware mapping, map in reverse - create frames in the sink
12826 and map them back to the source. This may be necessary in some cases where
12827 a mapping in one direction is required but only the opposite direction is
12828 supported by the devices being used.
12830 This option is dangerous - it may break the preceding filter in undefined
12831 ways if there are any additional constraints on that filter's output.
12832 Do not use it without fully understanding the implications of its use.
12838 Upload system memory frames to hardware surfaces.
12840 The device to upload to must be supplied when the filter is initialised. If
12841 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12842 option or with the @option{derive_device} option. The input and output devices
12843 must be of different types and compatible - the exact meaning of this is
12844 system-dependent, but typically it means that they must refer to the same
12845 underlying hardware context (for example, refer to the same graphics card).
12847 The following additional parameters are accepted:
12850 @item derive_device @var{type}
12851 Rather than using the device supplied at initialisation, instead derive a new
12852 device of type @var{type} from the device the input frames exist on.
12855 @anchor{hwupload_cuda}
12856 @section hwupload_cuda
12858 Upload system memory frames to a CUDA device.
12860 It accepts the following optional parameters:
12864 The number of the CUDA device to use
12869 Apply a high-quality magnification filter designed for pixel art. This filter
12870 was originally created by Maxim Stepin.
12872 It accepts the following option:
12876 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12877 @code{hq3x} and @code{4} for @code{hq4x}.
12878 Default is @code{3}.
12882 Stack input videos horizontally.
12884 All streams must be of same pixel format and of same height.
12886 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12887 to create same output.
12889 The filter accepts the following option:
12893 Set number of input streams. Default is 2.
12896 If set to 1, force the output to terminate when the shortest input
12897 terminates. Default value is 0.
12902 Modify the hue and/or the saturation of the input.
12904 It accepts the following parameters:
12908 Specify the hue angle as a number of degrees. It accepts an expression,
12909 and defaults to "0".
12912 Specify the saturation in the [-10,10] range. It accepts an expression and
12916 Specify the hue angle as a number of radians. It accepts an
12917 expression, and defaults to "0".
12920 Specify the brightness in the [-10,10] range. It accepts an expression and
12924 @option{h} and @option{H} are mutually exclusive, and can't be
12925 specified at the same time.
12927 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12928 expressions containing the following constants:
12932 frame count of the input frame starting from 0
12935 presentation timestamp of the input frame expressed in time base units
12938 frame rate of the input video, NAN if the input frame rate is unknown
12941 timestamp expressed in seconds, NAN if the input timestamp is unknown
12944 time base of the input video
12947 @subsection Examples
12951 Set the hue to 90 degrees and the saturation to 1.0:
12957 Same command but expressing the hue in radians:
12963 Rotate hue and make the saturation swing between 0
12964 and 2 over a period of 1 second:
12966 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12970 Apply a 3 seconds saturation fade-in effect starting at 0:
12972 hue="s=min(t/3\,1)"
12975 The general fade-in expression can be written as:
12977 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12981 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12983 hue="s=max(0\, min(1\, (8-t)/3))"
12986 The general fade-out expression can be written as:
12988 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12993 @subsection Commands
12995 This filter supports the following commands:
13001 Modify the hue and/or the saturation and/or brightness of the input video.
13002 The command accepts the same syntax of the corresponding option.
13004 If the specified expression is not valid, it is kept at its current
13008 @section hysteresis
13010 Grow first stream into second stream by connecting components.
13011 This makes it possible to build more robust edge masks.
13013 This filter accepts the following options:
13017 Set which planes will be processed as bitmap, unprocessed planes will be
13018 copied from first stream.
13019 By default value 0xf, all planes will be processed.
13022 Set threshold which is used in filtering. If pixel component value is higher than
13023 this value filter algorithm for connecting components is activated.
13024 By default value is 0.
13027 The @code{hysteresis} filter also supports the @ref{framesync} options.
13031 Detect video interlacing type.
13033 This filter tries to detect if the input frames are interlaced, progressive,
13034 top or bottom field first. It will also try to detect fields that are
13035 repeated between adjacent frames (a sign of telecine).
13037 Single frame detection considers only immediately adjacent frames when classifying each frame.
13038 Multiple frame detection incorporates the classification history of previous frames.
13040 The filter will log these metadata values:
13043 @item single.current_frame
13044 Detected type of current frame using single-frame detection. One of:
13045 ``tff'' (top field first), ``bff'' (bottom field first),
13046 ``progressive'', or ``undetermined''
13049 Cumulative number of frames detected as top field first using single-frame detection.
13052 Cumulative number of frames detected as top field first using multiple-frame detection.
13055 Cumulative number of frames detected as bottom field first using single-frame detection.
13057 @item multiple.current_frame
13058 Detected type of current frame using multiple-frame detection. One of:
13059 ``tff'' (top field first), ``bff'' (bottom field first),
13060 ``progressive'', or ``undetermined''
13063 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13065 @item single.progressive
13066 Cumulative number of frames detected as progressive using single-frame detection.
13068 @item multiple.progressive
13069 Cumulative number of frames detected as progressive using multiple-frame detection.
13071 @item single.undetermined
13072 Cumulative number of frames that could not be classified using single-frame detection.
13074 @item multiple.undetermined
13075 Cumulative number of frames that could not be classified using multiple-frame detection.
13077 @item repeated.current_frame
13078 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13080 @item repeated.neither
13081 Cumulative number of frames with no repeated field.
13084 Cumulative number of frames with the top field repeated from the previous frame's top field.
13086 @item repeated.bottom
13087 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13090 The filter accepts the following options:
13094 Set interlacing threshold.
13096 Set progressive threshold.
13098 Threshold for repeated field detection.
13100 Number of frames after which a given frame's contribution to the
13101 statistics is halved (i.e., it contributes only 0.5 to its
13102 classification). The default of 0 means that all frames seen are given
13103 full weight of 1.0 forever.
13104 @item analyze_interlaced_flag
13105 When this is not 0 then idet will use the specified number of frames to determine
13106 if the interlaced flag is accurate, it will not count undetermined frames.
13107 If the flag is found to be accurate it will be used without any further
13108 computations, if it is found to be inaccurate it will be cleared without any
13109 further computations. This allows inserting the idet filter as a low computational
13110 method to clean up the interlaced flag
13115 Deinterleave or interleave fields.
13117 This filter allows one to process interlaced images fields without
13118 deinterlacing them. Deinterleaving splits the input frame into 2
13119 fields (so called half pictures). Odd lines are moved to the top
13120 half of the output image, even lines to the bottom half.
13121 You can process (filter) them independently and then re-interleave them.
13123 The filter accepts the following options:
13127 @item chroma_mode, c
13128 @item alpha_mode, a
13129 Available values for @var{luma_mode}, @var{chroma_mode} and
13130 @var{alpha_mode} are:
13136 @item deinterleave, d
13137 Deinterleave fields, placing one above the other.
13139 @item interleave, i
13140 Interleave fields. Reverse the effect of deinterleaving.
13142 Default value is @code{none}.
13144 @item luma_swap, ls
13145 @item chroma_swap, cs
13146 @item alpha_swap, as
13147 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13150 @subsection Commands
13152 This filter supports the all above options as @ref{commands}.
13156 Apply inflate effect to the video.
13158 This filter replaces the pixel by the local(3x3) average by taking into account
13159 only values higher than the pixel.
13161 It accepts the following options:
13168 Limit the maximum change for each plane, default is 65535.
13169 If 0, plane will remain unchanged.
13172 @subsection Commands
13174 This filter supports the all above options as @ref{commands}.
13178 Simple interlacing filter from progressive contents. This interleaves upper (or
13179 lower) lines from odd frames with lower (or upper) lines from even frames,
13180 halving the frame rate and preserving image height.
13183 Original Original New Frame
13184 Frame 'j' Frame 'j+1' (tff)
13185 ========== =========== ==================
13186 Line 0 --------------------> Frame 'j' Line 0
13187 Line 1 Line 1 ----> Frame 'j+1' Line 1
13188 Line 2 ---------------------> Frame 'j' Line 2
13189 Line 3 Line 3 ----> Frame 'j+1' Line 3
13191 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13194 It accepts the following optional parameters:
13198 This determines whether the interlaced frame is taken from the even
13199 (tff - default) or odd (bff) lines of the progressive frame.
13202 Vertical lowpass filter to avoid twitter interlacing and
13203 reduce moire patterns.
13207 Disable vertical lowpass filter
13210 Enable linear filter (default)
13213 Enable complex filter. This will slightly less reduce twitter and moire
13214 but better retain detail and subjective sharpness impression.
13221 Deinterlace input video by applying Donald Graft's adaptive kernel
13222 deinterling. Work on interlaced parts of a video to produce
13223 progressive frames.
13225 The description of the accepted parameters follows.
13229 Set the threshold which affects the filter's tolerance when
13230 determining if a pixel line must be processed. It must be an integer
13231 in the range [0,255] and defaults to 10. A value of 0 will result in
13232 applying the process on every pixels.
13235 Paint pixels exceeding the threshold value to white if set to 1.
13239 Set the fields order. Swap fields if set to 1, leave fields alone if
13243 Enable additional sharpening if set to 1. Default is 0.
13246 Enable twoway sharpening if set to 1. Default is 0.
13249 @subsection Examples
13253 Apply default values:
13255 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13259 Enable additional sharpening:
13265 Paint processed pixels in white:
13273 Slowly update darker pixels.
13275 This filter makes short flashes of light appear longer.
13276 This filter accepts the following options:
13280 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13283 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13286 @section lenscorrection
13288 Correct radial lens distortion
13290 This filter can be used to correct for radial distortion as can result from the use
13291 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13292 one can use tools available for example as part of opencv or simply trial-and-error.
13293 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13294 and extract the k1 and k2 coefficients from the resulting matrix.
13296 Note that effectively the same filter is available in the open-source tools Krita and
13297 Digikam from the KDE project.
13299 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13300 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13301 brightness distribution, so you may want to use both filters together in certain
13302 cases, though you will have to take care of ordering, i.e. whether vignetting should
13303 be applied before or after lens correction.
13305 @subsection Options
13307 The filter accepts the following options:
13311 Relative x-coordinate of the focal point of the image, and thereby the center of the
13312 distortion. This value has a range [0,1] and is expressed as fractions of the image
13313 width. Default is 0.5.
13315 Relative y-coordinate of the focal point of the image, and thereby the center of the
13316 distortion. This value has a range [0,1] and is expressed as fractions of the image
13317 height. Default is 0.5.
13319 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13320 no correction. Default is 0.
13322 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13323 0 means no correction. Default is 0.
13326 The formula that generates the correction is:
13328 @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)
13330 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13331 distances from the focal point in the source and target images, respectively.
13335 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13337 The @code{lensfun} filter requires the camera make, camera model, and lens model
13338 to apply the lens correction. The filter will load the lensfun database and
13339 query it to find the corresponding camera and lens entries in the database. As
13340 long as these entries can be found with the given options, the filter can
13341 perform corrections on frames. Note that incomplete strings will result in the
13342 filter choosing the best match with the given options, and the filter will
13343 output the chosen camera and lens models (logged with level "info"). You must
13344 provide the make, camera model, and lens model as they are required.
13346 The filter accepts the following options:
13350 The make of the camera (for example, "Canon"). This option is required.
13353 The model of the camera (for example, "Canon EOS 100D"). This option is
13357 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13358 option is required.
13361 The type of correction to apply. The following values are valid options:
13365 Enables fixing lens vignetting.
13368 Enables fixing lens geometry. This is the default.
13371 Enables fixing chromatic aberrations.
13374 Enables fixing lens vignetting and lens geometry.
13377 Enables fixing lens vignetting and chromatic aberrations.
13380 Enables fixing both lens geometry and chromatic aberrations.
13383 Enables all possible corrections.
13387 The focal length of the image/video (zoom; expected constant for video). For
13388 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13389 range should be chosen when using that lens. Default 18.
13392 The aperture of the image/video (expected constant for video). Note that
13393 aperture is only used for vignetting correction. Default 3.5.
13395 @item focus_distance
13396 The focus distance of the image/video (expected constant for video). Note that
13397 focus distance is only used for vignetting and only slightly affects the
13398 vignetting correction process. If unknown, leave it at the default value (which
13402 The scale factor which is applied after transformation. After correction the
13403 video is no longer necessarily rectangular. This parameter controls how much of
13404 the resulting image is visible. The value 0 means that a value will be chosen
13405 automatically such that there is little or no unmapped area in the output
13406 image. 1.0 means that no additional scaling is done. Lower values may result
13407 in more of the corrected image being visible, while higher values may avoid
13408 unmapped areas in the output.
13410 @item target_geometry
13411 The target geometry of the output image/video. The following values are valid
13415 @item rectilinear (default)
13418 @item equirectangular
13419 @item fisheye_orthographic
13420 @item fisheye_stereographic
13421 @item fisheye_equisolid
13422 @item fisheye_thoby
13425 Apply the reverse of image correction (instead of correcting distortion, apply
13428 @item interpolation
13429 The type of interpolation used when correcting distortion. The following values
13434 @item linear (default)
13439 @subsection Examples
13443 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13444 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13448 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
13452 Apply the same as before, but only for the first 5 seconds of video.
13455 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
13462 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13463 score between two input videos.
13465 The obtained VMAF score is printed through the logging system.
13467 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13468 After installing the library it can be enabled using:
13469 @code{./configure --enable-libvmaf}.
13470 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13472 The filter has following options:
13476 Set the model path which is to be used for SVM.
13477 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13480 Set the file path to be used to store logs.
13483 Set the format of the log file (csv, json or xml).
13485 @item enable_transform
13486 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13487 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13488 Default value: @code{false}
13491 Invokes the phone model which will generate VMAF scores higher than in the
13492 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13493 Default value: @code{false}
13496 Enables computing psnr along with vmaf.
13497 Default value: @code{false}
13500 Enables computing ssim along with vmaf.
13501 Default value: @code{false}
13504 Enables computing ms_ssim along with vmaf.
13505 Default value: @code{false}
13508 Set the pool method to be used for computing vmaf.
13509 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13512 Set number of threads to be used when computing vmaf.
13513 Default value: @code{0}, which makes use of all available logical processors.
13516 Set interval for frame subsampling used when computing vmaf.
13517 Default value: @code{1}
13519 @item enable_conf_interval
13520 Enables confidence interval.
13521 Default value: @code{false}
13524 This filter also supports the @ref{framesync} options.
13526 @subsection Examples
13529 On the below examples the input file @file{main.mpg} being processed is
13530 compared with the reference file @file{ref.mpg}.
13533 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13537 Example with options:
13539 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13543 Example with options and different containers:
13545 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 -
13551 Limits the pixel components values to the specified range [min, max].
13553 The filter accepts the following options:
13557 Lower bound. Defaults to the lowest allowed value for the input.
13560 Upper bound. Defaults to the highest allowed value for the input.
13563 Specify which planes will be processed. Defaults to all available.
13566 @subsection Commands
13568 This filter supports the all above options as @ref{commands}.
13574 The filter accepts the following options:
13578 Set the number of loops. Setting this value to -1 will result in infinite loops.
13582 Set maximal size in number of frames. Default is 0.
13585 Set first frame of loop. Default is 0.
13588 @subsection Examples
13592 Loop single first frame infinitely:
13594 loop=loop=-1:size=1:start=0
13598 Loop single first frame 10 times:
13600 loop=loop=10:size=1:start=0
13604 Loop 10 first frames 5 times:
13606 loop=loop=5:size=10:start=0
13612 Apply a 1D LUT to an input video.
13614 The filter accepts the following options:
13618 Set the 1D LUT file name.
13620 Currently supported formats:
13629 Select interpolation mode.
13631 Available values are:
13635 Use values from the nearest defined point.
13637 Interpolate values using the linear interpolation.
13639 Interpolate values using the cosine interpolation.
13641 Interpolate values using the cubic interpolation.
13643 Interpolate values using the spline interpolation.
13650 Apply a 3D LUT to an input video.
13652 The filter accepts the following options:
13656 Set the 3D LUT file name.
13658 Currently supported formats:
13672 Select interpolation mode.
13674 Available values are:
13678 Use values from the nearest defined point.
13680 Interpolate values using the 8 points defining a cube.
13682 Interpolate values using a tetrahedron.
13688 Turn certain luma values into transparency.
13690 The filter accepts the following options:
13694 Set the luma which will be used as base for transparency.
13695 Default value is @code{0}.
13698 Set the range of luma values to be keyed out.
13699 Default value is @code{0.01}.
13702 Set the range of softness. Default value is @code{0}.
13703 Use this to control gradual transition from zero to full transparency.
13706 @subsection Commands
13707 This filter supports same @ref{commands} as options.
13708 The command accepts the same syntax of the corresponding option.
13710 If the specified expression is not valid, it is kept at its current
13713 @section lut, lutrgb, lutyuv
13715 Compute a look-up table for binding each pixel component input value
13716 to an output value, and apply it to the input video.
13718 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13719 to an RGB input video.
13721 These filters accept the following parameters:
13724 set first pixel component expression
13726 set second pixel component expression
13728 set third pixel component expression
13730 set fourth pixel component expression, corresponds to the alpha component
13733 set red component expression
13735 set green component expression
13737 set blue component expression
13739 alpha component expression
13742 set Y/luminance component expression
13744 set U/Cb component expression
13746 set V/Cr component expression
13749 Each of them specifies the expression to use for computing the lookup table for
13750 the corresponding pixel component values.
13752 The exact component associated to each of the @var{c*} options depends on the
13755 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13756 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13758 The expressions can contain the following constants and functions:
13763 The input width and height.
13766 The input value for the pixel component.
13769 The input value, clipped to the @var{minval}-@var{maxval} range.
13772 The maximum value for the pixel component.
13775 The minimum value for the pixel component.
13778 The negated value for the pixel component value, clipped to the
13779 @var{minval}-@var{maxval} range; it corresponds to the expression
13780 "maxval-clipval+minval".
13783 The computed value in @var{val}, clipped to the
13784 @var{minval}-@var{maxval} range.
13786 @item gammaval(gamma)
13787 The computed gamma correction value of the pixel component value,
13788 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13790 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13794 All expressions default to "val".
13796 @subsection Examples
13800 Negate input video:
13802 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13803 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13806 The above is the same as:
13808 lutrgb="r=negval:g=negval:b=negval"
13809 lutyuv="y=negval:u=negval:v=negval"
13819 Remove chroma components, turning the video into a graytone image:
13821 lutyuv="u=128:v=128"
13825 Apply a luma burning effect:
13831 Remove green and blue components:
13837 Set a constant alpha channel value on input:
13839 format=rgba,lutrgb=a="maxval-minval/2"
13843 Correct luminance gamma by a factor of 0.5:
13845 lutyuv=y=gammaval(0.5)
13849 Discard least significant bits of luma:
13851 lutyuv=y='bitand(val, 128+64+32)'
13855 Technicolor like effect:
13857 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13861 @section lut2, tlut2
13863 The @code{lut2} filter takes two input streams and outputs one
13866 The @code{tlut2} (time lut2) filter takes two consecutive frames
13867 from one single stream.
13869 This filter accepts the following parameters:
13872 set first pixel component expression
13874 set second pixel component expression
13876 set third pixel component expression
13878 set fourth pixel component expression, corresponds to the alpha component
13881 set output bit depth, only available for @code{lut2} filter. By default is 0,
13882 which means bit depth is automatically picked from first input format.
13885 The @code{lut2} filter also supports the @ref{framesync} options.
13887 Each of them specifies the expression to use for computing the lookup table for
13888 the corresponding pixel component values.
13890 The exact component associated to each of the @var{c*} options depends on the
13893 The expressions can contain the following constants:
13898 The input width and height.
13901 The first input value for the pixel component.
13904 The second input value for the pixel component.
13907 The first input video bit depth.
13910 The second input video bit depth.
13913 All expressions default to "x".
13915 @subsection Examples
13919 Highlight differences between two RGB video streams:
13921 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)'
13925 Highlight differences between two YUV video streams:
13927 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)'
13931 Show max difference between two video streams:
13933 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)))'
13937 @section maskedclamp
13939 Clamp the first input stream with the second input and third input stream.
13941 Returns the value of first stream to be between second input
13942 stream - @code{undershoot} and third input stream + @code{overshoot}.
13944 This filter accepts the following options:
13947 Default value is @code{0}.
13950 Default value is @code{0}.
13953 Set which planes will be processed as bitmap, unprocessed planes will be
13954 copied from first stream.
13955 By default value 0xf, all planes will be processed.
13960 Merge the second and third input stream into output stream using absolute differences
13961 between second input stream and first input stream and absolute difference between
13962 third input stream and first input stream. The picked value will be from second input
13963 stream if second absolute difference is greater than first one or from third input stream
13966 This filter accepts the following options:
13969 Set which planes will be processed as bitmap, unprocessed planes will be
13970 copied from first stream.
13971 By default value 0xf, all planes will be processed.
13974 @section maskedmerge
13976 Merge the first input stream with the second input stream using per pixel
13977 weights in the third input stream.
13979 A value of 0 in the third stream pixel component means that pixel component
13980 from first stream is returned unchanged, while maximum value (eg. 255 for
13981 8-bit videos) means that pixel component from second stream is returned
13982 unchanged. Intermediate values define the amount of merging between both
13983 input stream's pixel components.
13985 This filter accepts the following options:
13988 Set which planes will be processed as bitmap, unprocessed planes will be
13989 copied from first stream.
13990 By default value 0xf, all planes will be processed.
13995 Merge the second and third input stream into output stream using absolute differences
13996 between second input stream and first input stream and absolute difference between
13997 third input stream and first input stream. The picked value will be from second input
13998 stream if second absolute difference is less than first one or from third input stream
14001 This filter accepts the following options:
14004 Set which planes will be processed as bitmap, unprocessed planes will be
14005 copied from first stream.
14006 By default value 0xf, all planes will be processed.
14009 @section maskedthreshold
14010 Pick pixels comparing absolute difference of two video streams with fixed
14013 If absolute difference between pixel component of first and second video
14014 stream is equal or lower than user supplied threshold than pixel component
14015 from first video stream is picked, otherwise pixel component from second
14016 video stream is picked.
14018 This filter accepts the following options:
14021 Set threshold used when picking pixels from absolute difference from two input
14025 Set which planes will be processed as bitmap, unprocessed planes will be
14026 copied from second stream.
14027 By default value 0xf, all planes will be processed.
14031 Create mask from input video.
14033 For example it is useful to create motion masks after @code{tblend} filter.
14035 This filter accepts the following options:
14039 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14042 Set high threshold. Any pixel component higher than this value will be set to max value
14043 allowed for current pixel format.
14046 Set planes to filter, by default all available planes are filtered.
14049 Fill all frame pixels with this value.
14052 Set max average pixel value for frame. If sum of all pixel components is higher that this
14053 average, output frame will be completely filled with value set by @var{fill} option.
14054 Typically useful for scene changes when used in combination with @code{tblend} filter.
14059 Apply motion-compensation deinterlacing.
14061 It needs one field per frame as input and must thus be used together
14062 with yadif=1/3 or equivalent.
14064 This filter accepts the following options:
14067 Set the deinterlacing mode.
14069 It accepts one of the following values:
14074 use iterative motion estimation
14076 like @samp{slow}, but use multiple reference frames.
14078 Default value is @samp{fast}.
14081 Set the picture field parity assumed for the input video. It must be
14082 one of the following values:
14086 assume top field first
14088 assume bottom field first
14091 Default value is @samp{bff}.
14094 Set per-block quantization parameter (QP) used by the internal
14097 Higher values should result in a smoother motion vector field but less
14098 optimal individual vectors. Default value is 1.
14103 Pick median pixel from certain rectangle defined by radius.
14105 This filter accepts the following options:
14109 Set horizontal radius size. Default value is @code{1}.
14110 Allowed range is integer from 1 to 127.
14113 Set which planes to process. Default is @code{15}, which is all available planes.
14116 Set vertical radius size. Default value is @code{0}.
14117 Allowed range is integer from 0 to 127.
14118 If it is 0, value will be picked from horizontal @code{radius} option.
14121 Set median percentile. Default value is @code{0.5}.
14122 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14123 minimum values, and @code{1} maximum values.
14126 @subsection Commands
14127 This filter supports same @ref{commands} as options.
14128 The command accepts the same syntax of the corresponding option.
14130 If the specified expression is not valid, it is kept at its current
14133 @section mergeplanes
14135 Merge color channel components from several video streams.
14137 The filter accepts up to 4 input streams, and merge selected input
14138 planes to the output video.
14140 This filter accepts the following options:
14143 Set input to output plane mapping. Default is @code{0}.
14145 The mappings is specified as a bitmap. It should be specified as a
14146 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14147 mapping for the first plane of the output stream. 'A' sets the number of
14148 the input stream to use (from 0 to 3), and 'a' the plane number of the
14149 corresponding input to use (from 0 to 3). The rest of the mappings is
14150 similar, 'Bb' describes the mapping for the output stream second
14151 plane, 'Cc' describes the mapping for the output stream third plane and
14152 'Dd' describes the mapping for the output stream fourth plane.
14155 Set output pixel format. Default is @code{yuva444p}.
14158 @subsection Examples
14162 Merge three gray video streams of same width and height into single video stream:
14164 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14168 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14170 [a0][a1]mergeplanes=0x00010210:yuva444p
14174 Swap Y and A plane in yuva444p stream:
14176 format=yuva444p,mergeplanes=0x03010200:yuva444p
14180 Swap U and V plane in yuv420p stream:
14182 format=yuv420p,mergeplanes=0x000201:yuv420p
14186 Cast a rgb24 clip to yuv444p:
14188 format=rgb24,mergeplanes=0x000102:yuv444p
14194 Estimate and export motion vectors using block matching algorithms.
14195 Motion vectors are stored in frame side data to be used by other filters.
14197 This filter accepts the following options:
14200 Specify the motion estimation method. Accepts one of the following values:
14204 Exhaustive search algorithm.
14206 Three step search algorithm.
14208 Two dimensional logarithmic search algorithm.
14210 New three step search algorithm.
14212 Four step search algorithm.
14214 Diamond search algorithm.
14216 Hexagon-based search algorithm.
14218 Enhanced predictive zonal search algorithm.
14220 Uneven multi-hexagon search algorithm.
14222 Default value is @samp{esa}.
14225 Macroblock size. Default @code{16}.
14228 Search parameter. Default @code{7}.
14231 @section midequalizer
14233 Apply Midway Image Equalization effect using two video streams.
14235 Midway Image Equalization adjusts a pair of images to have the same
14236 histogram, while maintaining their dynamics as much as possible. It's
14237 useful for e.g. matching exposures from a pair of stereo cameras.
14239 This filter has two inputs and one output, which must be of same pixel format, but
14240 may be of different sizes. The output of filter is first input adjusted with
14241 midway histogram of both inputs.
14243 This filter accepts the following option:
14247 Set which planes to process. Default is @code{15}, which is all available planes.
14250 @section minterpolate
14252 Convert the video to specified frame rate using motion interpolation.
14254 This filter accepts the following options:
14257 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}.
14260 Motion interpolation mode. Following values are accepted:
14263 Duplicate previous or next frame for interpolating new ones.
14265 Blend source frames. Interpolated frame is mean of previous and next frames.
14267 Motion compensated interpolation. Following options are effective when this mode is selected:
14271 Motion compensation mode. Following values are accepted:
14274 Overlapped block motion compensation.
14276 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14278 Default mode is @samp{obmc}.
14281 Motion estimation mode. Following values are accepted:
14284 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14286 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14288 Default mode is @samp{bilat}.
14291 The algorithm to be used for motion estimation. Following values are accepted:
14294 Exhaustive search algorithm.
14296 Three step search algorithm.
14298 Two dimensional logarithmic search algorithm.
14300 New three step search algorithm.
14302 Four step search algorithm.
14304 Diamond search algorithm.
14306 Hexagon-based search algorithm.
14308 Enhanced predictive zonal search algorithm.
14310 Uneven multi-hexagon search algorithm.
14312 Default algorithm is @samp{epzs}.
14315 Macroblock size. Default @code{16}.
14318 Motion estimation search parameter. Default @code{32}.
14321 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).
14326 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:
14329 Disable scene change detection.
14331 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14333 Default method is @samp{fdiff}.
14335 @item scd_threshold
14336 Scene change detection threshold. Default is @code{10.}.
14341 Mix several video input streams into one video stream.
14343 A description of the accepted options follows.
14347 The number of inputs. If unspecified, it defaults to 2.
14350 Specify weight of each input video stream as sequence.
14351 Each weight is separated by space. If number of weights
14352 is smaller than number of @var{frames} last specified
14353 weight will be used for all remaining unset weights.
14356 Specify scale, if it is set it will be multiplied with sum
14357 of each weight multiplied with pixel values to give final destination
14358 pixel value. By default @var{scale} is auto scaled to sum of weights.
14361 Specify how end of stream is determined.
14364 The duration of the longest input. (default)
14367 The duration of the shortest input.
14370 The duration of the first input.
14374 @section mpdecimate
14376 Drop frames that do not differ greatly from the previous frame in
14377 order to reduce frame rate.
14379 The main use of this filter is for very-low-bitrate encoding
14380 (e.g. streaming over dialup modem), but it could in theory be used for
14381 fixing movies that were inverse-telecined incorrectly.
14383 A description of the accepted options follows.
14387 Set the maximum number of consecutive frames which can be dropped (if
14388 positive), or the minimum interval between dropped frames (if
14389 negative). If the value is 0, the frame is dropped disregarding the
14390 number of previous sequentially dropped frames.
14392 Default value is 0.
14397 Set the dropping threshold values.
14399 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14400 represent actual pixel value differences, so a threshold of 64
14401 corresponds to 1 unit of difference for each pixel, or the same spread
14402 out differently over the block.
14404 A frame is a candidate for dropping if no 8x8 blocks differ by more
14405 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14406 meaning the whole image) differ by more than a threshold of @option{lo}.
14408 Default value for @option{hi} is 64*12, default value for @option{lo} is
14409 64*5, and default value for @option{frac} is 0.33.
14415 Negate (invert) the input video.
14417 It accepts the following option:
14422 With value 1, it negates the alpha component, if present. Default value is 0.
14428 Denoise frames using Non-Local Means algorithm.
14430 Each pixel is adjusted by looking for other pixels with similar contexts. This
14431 context similarity is defined by comparing their surrounding patches of size
14432 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14435 Note that the research area defines centers for patches, which means some
14436 patches will be made of pixels outside that research area.
14438 The filter accepts the following options.
14442 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14445 Set patch size. Default is 7. Must be odd number in range [0, 99].
14448 Same as @option{p} but for chroma planes.
14450 The default value is @var{0} and means automatic.
14453 Set research size. Default is 15. Must be odd number in range [0, 99].
14456 Same as @option{r} but for chroma planes.
14458 The default value is @var{0} and means automatic.
14463 Deinterlace video using neural network edge directed interpolation.
14465 This filter accepts the following options:
14469 Mandatory option, without binary file filter can not work.
14470 Currently file can be found here:
14471 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14474 Set which frames to deinterlace, by default it is @code{all}.
14475 Can be @code{all} or @code{interlaced}.
14478 Set mode of operation.
14480 Can be one of the following:
14484 Use frame flags, both fields.
14486 Use frame flags, single field.
14488 Use top field only.
14490 Use bottom field only.
14492 Use both fields, top first.
14494 Use both fields, bottom first.
14498 Set which planes to process, by default filter process all frames.
14501 Set size of local neighborhood around each pixel, used by the predictor neural
14504 Can be one of the following:
14517 Set the number of neurons in predictor neural network.
14518 Can be one of the following:
14529 Controls the number of different neural network predictions that are blended
14530 together to compute the final output value. Can be @code{fast}, default or
14534 Set which set of weights to use in the predictor.
14535 Can be one of the following:
14539 weights trained to minimize absolute error
14541 weights trained to minimize squared error
14545 Controls whether or not the prescreener neural network is used to decide
14546 which pixels should be processed by the predictor neural network and which
14547 can be handled by simple cubic interpolation.
14548 The prescreener is trained to know whether cubic interpolation will be
14549 sufficient for a pixel or whether it should be predicted by the predictor nn.
14550 The computational complexity of the prescreener nn is much less than that of
14551 the predictor nn. Since most pixels can be handled by cubic interpolation,
14552 using the prescreener generally results in much faster processing.
14553 The prescreener is pretty accurate, so the difference between using it and not
14554 using it is almost always unnoticeable.
14556 Can be one of the following:
14564 Default is @code{new}.
14567 Set various debugging flags.
14572 Force libavfilter not to use any of the specified pixel formats for the
14573 input to the next filter.
14575 It accepts the following parameters:
14579 A '|'-separated list of pixel format names, such as
14580 pix_fmts=yuv420p|monow|rgb24".
14584 @subsection Examples
14588 Force libavfilter to use a format different from @var{yuv420p} for the
14589 input to the vflip filter:
14591 noformat=pix_fmts=yuv420p,vflip
14595 Convert the input video to any of the formats not contained in the list:
14597 noformat=yuv420p|yuv444p|yuv410p
14603 Add noise on video input frame.
14605 The filter accepts the following options:
14613 Set noise seed for specific pixel component or all pixel components in case
14614 of @var{all_seed}. Default value is @code{123457}.
14616 @item all_strength, alls
14617 @item c0_strength, c0s
14618 @item c1_strength, c1s
14619 @item c2_strength, c2s
14620 @item c3_strength, c3s
14621 Set noise strength for specific pixel component or all pixel components in case
14622 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14624 @item all_flags, allf
14625 @item c0_flags, c0f
14626 @item c1_flags, c1f
14627 @item c2_flags, c2f
14628 @item c3_flags, c3f
14629 Set pixel component flags or set flags for all components if @var{all_flags}.
14630 Available values for component flags are:
14633 averaged temporal noise (smoother)
14635 mix random noise with a (semi)regular pattern
14637 temporal noise (noise pattern changes between frames)
14639 uniform noise (gaussian otherwise)
14643 @subsection Examples
14645 Add temporal and uniform noise to input video:
14647 noise=alls=20:allf=t+u
14652 Normalize RGB video (aka histogram stretching, contrast stretching).
14653 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14655 For each channel of each frame, the filter computes the input range and maps
14656 it linearly to the user-specified output range. The output range defaults
14657 to the full dynamic range from pure black to pure white.
14659 Temporal smoothing can be used on the input range to reduce flickering (rapid
14660 changes in brightness) caused when small dark or bright objects enter or leave
14661 the scene. This is similar to the auto-exposure (automatic gain control) on a
14662 video camera, and, like a video camera, it may cause a period of over- or
14663 under-exposure of the video.
14665 The R,G,B channels can be normalized independently, which may cause some
14666 color shifting, or linked together as a single channel, which prevents
14667 color shifting. Linked normalization preserves hue. Independent normalization
14668 does not, so it can be used to remove some color casts. Independent and linked
14669 normalization can be combined in any ratio.
14671 The normalize filter accepts the following options:
14676 Colors which define the output range. The minimum input value is mapped to
14677 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14678 The defaults are black and white respectively. Specifying white for
14679 @var{blackpt} and black for @var{whitept} will give color-inverted,
14680 normalized video. Shades of grey can be used to reduce the dynamic range
14681 (contrast). Specifying saturated colors here can create some interesting
14685 The number of previous frames to use for temporal smoothing. The input range
14686 of each channel is smoothed using a rolling average over the current frame
14687 and the @var{smoothing} previous frames. The default is 0 (no temporal
14691 Controls the ratio of independent (color shifting) channel normalization to
14692 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14693 independent. Defaults to 1.0 (fully independent).
14696 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14697 expensive no-op. Defaults to 1.0 (full strength).
14701 @subsection Commands
14702 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14703 The command accepts the same syntax of the corresponding option.
14705 If the specified expression is not valid, it is kept at its current
14708 @subsection Examples
14710 Stretch video contrast to use the full dynamic range, with no temporal
14711 smoothing; may flicker depending on the source content:
14713 normalize=blackpt=black:whitept=white:smoothing=0
14716 As above, but with 50 frames of temporal smoothing; flicker should be
14717 reduced, depending on the source content:
14719 normalize=blackpt=black:whitept=white:smoothing=50
14722 As above, but with hue-preserving linked channel normalization:
14724 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14727 As above, but with half strength:
14729 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14732 Map the darkest input color to red, the brightest input color to cyan:
14734 normalize=blackpt=red:whitept=cyan
14739 Pass the video source unchanged to the output.
14742 Optical Character Recognition
14744 This filter uses Tesseract for optical character recognition. To enable
14745 compilation of this filter, you need to configure FFmpeg with
14746 @code{--enable-libtesseract}.
14748 It accepts the following options:
14752 Set datapath to tesseract data. Default is to use whatever was
14753 set at installation.
14756 Set language, default is "eng".
14759 Set character whitelist.
14762 Set character blacklist.
14765 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14766 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14770 Apply a video transform using libopencv.
14772 To enable this filter, install the libopencv library and headers and
14773 configure FFmpeg with @code{--enable-libopencv}.
14775 It accepts the following parameters:
14780 The name of the libopencv filter to apply.
14782 @item filter_params
14783 The parameters to pass to the libopencv filter. If not specified, the default
14784 values are assumed.
14788 Refer to the official libopencv documentation for more precise
14790 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14792 Several libopencv filters are supported; see the following subsections.
14797 Dilate an image by using a specific structuring element.
14798 It corresponds to the libopencv function @code{cvDilate}.
14800 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14802 @var{struct_el} represents a structuring element, and has the syntax:
14803 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14805 @var{cols} and @var{rows} represent the number of columns and rows of
14806 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14807 point, and @var{shape} the shape for the structuring element. @var{shape}
14808 must be "rect", "cross", "ellipse", or "custom".
14810 If the value for @var{shape} is "custom", it must be followed by a
14811 string of the form "=@var{filename}". The file with name
14812 @var{filename} is assumed to represent a binary image, with each
14813 printable character corresponding to a bright pixel. When a custom
14814 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14815 or columns and rows of the read file are assumed instead.
14817 The default value for @var{struct_el} is "3x3+0x0/rect".
14819 @var{nb_iterations} specifies the number of times the transform is
14820 applied to the image, and defaults to 1.
14824 # Use the default values
14827 # Dilate using a structuring element with a 5x5 cross, iterating two times
14828 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14830 # Read the shape from the file diamond.shape, iterating two times.
14831 # The file diamond.shape may contain a pattern of characters like this
14837 # The specified columns and rows are ignored
14838 # but the anchor point coordinates are not
14839 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14844 Erode an image by using a specific structuring element.
14845 It corresponds to the libopencv function @code{cvErode}.
14847 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14848 with the same syntax and semantics as the @ref{dilate} filter.
14852 Smooth the input video.
14854 The filter takes the following parameters:
14855 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14857 @var{type} is the type of smooth filter to apply, and must be one of
14858 the following values: "blur", "blur_no_scale", "median", "gaussian",
14859 or "bilateral". The default value is "gaussian".
14861 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14862 depends on the smooth type. @var{param1} and
14863 @var{param2} accept integer positive values or 0. @var{param3} and
14864 @var{param4} accept floating point values.
14866 The default value for @var{param1} is 3. The default value for the
14867 other parameters is 0.
14869 These parameters correspond to the parameters assigned to the
14870 libopencv function @code{cvSmooth}.
14872 @section oscilloscope
14874 2D Video Oscilloscope.
14876 Useful to measure spatial impulse, step responses, chroma delays, etc.
14878 It accepts the following parameters:
14882 Set scope center x position.
14885 Set scope center y position.
14888 Set scope size, relative to frame diagonal.
14891 Set scope tilt/rotation.
14897 Set trace center x position.
14900 Set trace center y position.
14903 Set trace width, relative to width of frame.
14906 Set trace height, relative to height of frame.
14909 Set which components to trace. By default it traces first three components.
14912 Draw trace grid. By default is enabled.
14915 Draw some statistics. By default is enabled.
14918 Draw scope. By default is enabled.
14921 @subsection Commands
14922 This filter supports same @ref{commands} as options.
14923 The command accepts the same syntax of the corresponding option.
14925 If the specified expression is not valid, it is kept at its current
14928 @subsection Examples
14932 Inspect full first row of video frame.
14934 oscilloscope=x=0.5:y=0:s=1
14938 Inspect full last row of video frame.
14940 oscilloscope=x=0.5:y=1:s=1
14944 Inspect full 5th line of video frame of height 1080.
14946 oscilloscope=x=0.5:y=5/1080:s=1
14950 Inspect full last column of video frame.
14952 oscilloscope=x=1:y=0.5:s=1:t=1
14960 Overlay one video on top of another.
14962 It takes two inputs and has one output. The first input is the "main"
14963 video on which the second input is overlaid.
14965 It accepts the following parameters:
14967 A description of the accepted options follows.
14972 Set the expression for the x and y coordinates of the overlaid video
14973 on the main video. Default value is "0" for both expressions. In case
14974 the expression is invalid, it is set to a huge value (meaning that the
14975 overlay will not be displayed within the output visible area).
14978 See @ref{framesync}.
14981 Set when the expressions for @option{x}, and @option{y} are evaluated.
14983 It accepts the following values:
14986 only evaluate expressions once during the filter initialization or
14987 when a command is processed
14990 evaluate expressions for each incoming frame
14993 Default value is @samp{frame}.
14996 See @ref{framesync}.
14999 Set the format for the output video.
15001 It accepts the following values:
15004 force YUV420 output
15007 force YUV420p10 output
15010 force YUV422 output
15013 force YUV422p10 output
15016 force YUV444 output
15019 force packed RGB output
15022 force planar RGB output
15025 automatically pick format
15028 Default value is @samp{yuv420}.
15031 See @ref{framesync}.
15034 Set format of alpha of the overlaid video, it can be @var{straight} or
15035 @var{premultiplied}. Default is @var{straight}.
15038 The @option{x}, and @option{y} expressions can contain the following
15044 The main input width and height.
15048 The overlay input width and height.
15052 The computed values for @var{x} and @var{y}. They are evaluated for
15057 horizontal and vertical chroma subsample values of the output
15058 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15062 the number of input frame, starting from 0
15065 the position in the file of the input frame, NAN if unknown
15068 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15072 This filter also supports the @ref{framesync} options.
15074 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15075 when evaluation is done @emph{per frame}, and will evaluate to NAN
15076 when @option{eval} is set to @samp{init}.
15078 Be aware that frames are taken from each input video in timestamp
15079 order, hence, if their initial timestamps differ, it is a good idea
15080 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15081 have them begin in the same zero timestamp, as the example for
15082 the @var{movie} filter does.
15084 You can chain together more overlays but you should test the
15085 efficiency of such approach.
15087 @subsection Commands
15089 This filter supports the following commands:
15093 Modify the x and y of the overlay input.
15094 The command accepts the same syntax of the corresponding option.
15096 If the specified expression is not valid, it is kept at its current
15100 @subsection Examples
15104 Draw the overlay at 10 pixels from the bottom right corner of the main
15107 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15110 Using named options the example above becomes:
15112 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15116 Insert a transparent PNG logo in the bottom left corner of the input,
15117 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15119 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15123 Insert 2 different transparent PNG logos (second logo on bottom
15124 right corner) using the @command{ffmpeg} tool:
15126 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
15130 Add a transparent color layer on top of the main video; @code{WxH}
15131 must specify the size of the main input to the overlay filter:
15133 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15137 Play an original video and a filtered version (here with the deshake
15138 filter) side by side using the @command{ffplay} tool:
15140 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15143 The above command is the same as:
15145 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15149 Make a sliding overlay appearing from the left to the right top part of the
15150 screen starting since time 2:
15152 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15156 Compose output by putting two input videos side to side:
15158 ffmpeg -i left.avi -i right.avi -filter_complex "
15159 nullsrc=size=200x100 [background];
15160 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15161 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15162 [background][left] overlay=shortest=1 [background+left];
15163 [background+left][right] overlay=shortest=1:x=100 [left+right]
15168 Mask 10-20 seconds of a video by applying the delogo filter to a section
15170 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15171 -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]'
15176 Chain several overlays in cascade:
15178 nullsrc=s=200x200 [bg];
15179 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15180 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15181 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15182 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15183 [in3] null, [mid2] overlay=100:100 [out0]
15188 @anchor{overlay_cuda}
15189 @section overlay_cuda
15191 Overlay one video on top of another.
15193 This is the CUDA variant of the @ref{overlay} filter.
15194 It only accepts CUDA frames. The underlying input pixel formats have to match.
15196 It takes two inputs and has one output. The first input is the "main"
15197 video on which the second input is overlaid.
15199 It accepts the following parameters:
15204 Set the x and y coordinates of the overlaid video on the main video.
15205 Default value is "0" for both expressions.
15208 See @ref{framesync}.
15211 See @ref{framesync}.
15214 See @ref{framesync}.
15218 This filter also supports the @ref{framesync} options.
15222 Apply Overcomplete Wavelet denoiser.
15224 The filter accepts the following options:
15230 Larger depth values will denoise lower frequency components more, but
15231 slow down filtering.
15233 Must be an int in the range 8-16, default is @code{8}.
15235 @item luma_strength, ls
15238 Must be a double value in the range 0-1000, default is @code{1.0}.
15240 @item chroma_strength, cs
15241 Set chroma strength.
15243 Must be a double value in the range 0-1000, default is @code{1.0}.
15249 Add paddings to the input image, and place the original input at the
15250 provided @var{x}, @var{y} coordinates.
15252 It accepts the following parameters:
15257 Specify an expression for the size of the output image with the
15258 paddings added. If the value for @var{width} or @var{height} is 0, the
15259 corresponding input size is used for the output.
15261 The @var{width} expression can reference the value set by the
15262 @var{height} expression, and vice versa.
15264 The default value of @var{width} and @var{height} is 0.
15268 Specify the offsets to place the input image at within the padded area,
15269 with respect to the top/left border of the output image.
15271 The @var{x} expression can reference the value set by the @var{y}
15272 expression, and vice versa.
15274 The default value of @var{x} and @var{y} is 0.
15276 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15277 so the input image is centered on the padded area.
15280 Specify the color of the padded area. For the syntax of this option,
15281 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15282 manual,ffmpeg-utils}.
15284 The default value of @var{color} is "black".
15287 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15289 It accepts the following values:
15293 Only evaluate expressions once during the filter initialization or when
15294 a command is processed.
15297 Evaluate expressions for each incoming frame.
15301 Default value is @samp{init}.
15304 Pad to aspect instead to a resolution.
15308 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15309 options are expressions containing the following constants:
15314 The input video width and height.
15318 These are the same as @var{in_w} and @var{in_h}.
15322 The output width and height (the size of the padded area), as
15323 specified by the @var{width} and @var{height} expressions.
15327 These are the same as @var{out_w} and @var{out_h}.
15331 The x and y offsets as specified by the @var{x} and @var{y}
15332 expressions, or NAN if not yet specified.
15335 same as @var{iw} / @var{ih}
15338 input sample aspect ratio
15341 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15345 The horizontal and vertical chroma subsample values. For example for the
15346 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15349 @subsection Examples
15353 Add paddings with the color "violet" to the input video. The output video
15354 size is 640x480, and the top-left corner of the input video is placed at
15357 pad=640:480:0:40:violet
15360 The example above is equivalent to the following command:
15362 pad=width=640:height=480:x=0:y=40:color=violet
15366 Pad the input to get an output with dimensions increased by 3/2,
15367 and put the input video at the center of the padded area:
15369 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15373 Pad the input to get a squared output with size equal to the maximum
15374 value between the input width and height, and put the input video at
15375 the center of the padded area:
15377 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15381 Pad the input to get a final w/h ratio of 16:9:
15383 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15387 In case of anamorphic video, in order to set the output display aspect
15388 correctly, it is necessary to use @var{sar} in the expression,
15389 according to the relation:
15391 (ih * X / ih) * sar = output_dar
15392 X = output_dar / sar
15395 Thus the previous example needs to be modified to:
15397 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15401 Double the output size and put the input video in the bottom-right
15402 corner of the output padded area:
15404 pad="2*iw:2*ih:ow-iw:oh-ih"
15408 @anchor{palettegen}
15409 @section palettegen
15411 Generate one palette for a whole video stream.
15413 It accepts the following options:
15417 Set the maximum number of colors to quantize in the palette.
15418 Note: the palette will still contain 256 colors; the unused palette entries
15421 @item reserve_transparent
15422 Create a palette of 255 colors maximum and reserve the last one for
15423 transparency. Reserving the transparency color is useful for GIF optimization.
15424 If not set, the maximum of colors in the palette will be 256. You probably want
15425 to disable this option for a standalone image.
15428 @item transparency_color
15429 Set the color that will be used as background for transparency.
15432 Set statistics mode.
15434 It accepts the following values:
15437 Compute full frame histograms.
15439 Compute histograms only for the part that differs from previous frame. This
15440 might be relevant to give more importance to the moving part of your input if
15441 the background is static.
15443 Compute new histogram for each frame.
15446 Default value is @var{full}.
15449 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15450 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15451 color quantization of the palette. This information is also visible at
15452 @var{info} logging level.
15454 @subsection Examples
15458 Generate a representative palette of a given video using @command{ffmpeg}:
15460 ffmpeg -i input.mkv -vf palettegen palette.png
15464 @section paletteuse
15466 Use a palette to downsample an input video stream.
15468 The filter takes two inputs: one video stream and a palette. The palette must
15469 be a 256 pixels image.
15471 It accepts the following options:
15475 Select dithering mode. Available algorithms are:
15478 Ordered 8x8 bayer dithering (deterministic)
15480 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15481 Note: this dithering is sometimes considered "wrong" and is included as a
15483 @item floyd_steinberg
15484 Floyd and Steingberg dithering (error diffusion)
15486 Frankie Sierra dithering v2 (error diffusion)
15488 Frankie Sierra dithering v2 "Lite" (error diffusion)
15491 Default is @var{sierra2_4a}.
15494 When @var{bayer} dithering is selected, this option defines the scale of the
15495 pattern (how much the crosshatch pattern is visible). A low value means more
15496 visible pattern for less banding, and higher value means less visible pattern
15497 at the cost of more banding.
15499 The option must be an integer value in the range [0,5]. Default is @var{2}.
15502 If set, define the zone to process
15506 Only the changing rectangle will be reprocessed. This is similar to GIF
15507 cropping/offsetting compression mechanism. This option can be useful for speed
15508 if only a part of the image is changing, and has use cases such as limiting the
15509 scope of the error diffusal @option{dither} to the rectangle that bounds the
15510 moving scene (it leads to more deterministic output if the scene doesn't change
15511 much, and as a result less moving noise and better GIF compression).
15514 Default is @var{none}.
15517 Take new palette for each output frame.
15519 @item alpha_threshold
15520 Sets the alpha threshold for transparency. Alpha values above this threshold
15521 will be treated as completely opaque, and values below this threshold will be
15522 treated as completely transparent.
15524 The option must be an integer value in the range [0,255]. Default is @var{128}.
15527 @subsection Examples
15531 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15532 using @command{ffmpeg}:
15534 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15538 @section perspective
15540 Correct perspective of video not recorded perpendicular to the screen.
15542 A description of the accepted parameters follows.
15553 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15554 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15555 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15556 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15557 then the corners of the source will be sent to the specified coordinates.
15559 The expressions can use the following variables:
15564 the width and height of video frame.
15568 Output frame count.
15571 @item interpolation
15572 Set interpolation for perspective correction.
15574 It accepts the following values:
15580 Default value is @samp{linear}.
15583 Set interpretation of coordinate options.
15585 It accepts the following values:
15589 Send point in the source specified by the given coordinates to
15590 the corners of the destination.
15592 @item 1, destination
15594 Send the corners of the source to the point in the destination specified
15595 by the given coordinates.
15597 Default value is @samp{source}.
15601 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15603 It accepts the following values:
15606 only evaluate expressions once during the filter initialization or
15607 when a command is processed
15610 evaluate expressions for each incoming frame
15613 Default value is @samp{init}.
15618 Delay interlaced video by one field time so that the field order changes.
15620 The intended use is to fix PAL movies that have been captured with the
15621 opposite field order to the film-to-video transfer.
15623 A description of the accepted parameters follows.
15629 It accepts the following values:
15632 Capture field order top-first, transfer bottom-first.
15633 Filter will delay the bottom field.
15636 Capture field order bottom-first, transfer top-first.
15637 Filter will delay the top field.
15640 Capture and transfer with the same field order. This mode only exists
15641 for the documentation of the other options to refer to, but if you
15642 actually select it, the filter will faithfully do nothing.
15645 Capture field order determined automatically by field flags, transfer
15647 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15648 basis using field flags. If no field information is available,
15649 then this works just like @samp{u}.
15652 Capture unknown or varying, transfer opposite.
15653 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15654 analyzing the images and selecting the alternative that produces best
15655 match between the fields.
15658 Capture top-first, transfer unknown or varying.
15659 Filter selects among @samp{t} and @samp{p} using image analysis.
15662 Capture bottom-first, transfer unknown or varying.
15663 Filter selects among @samp{b} and @samp{p} using image analysis.
15666 Capture determined by field flags, transfer unknown or varying.
15667 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15668 image analysis. If no field information is available, then this works just
15669 like @samp{U}. This is the default mode.
15672 Both capture and transfer unknown or varying.
15673 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15677 @section photosensitivity
15678 Reduce various flashes in video, so to help users with epilepsy.
15680 It accepts the following options:
15683 Set how many frames to use when filtering. Default is 30.
15686 Set detection threshold factor. Default is 1.
15690 Set how many pixels to skip when sampling frames. Default is 1.
15691 Allowed range is from 1 to 1024.
15694 Leave frames unchanged. Default is disabled.
15697 @section pixdesctest
15699 Pixel format descriptor test filter, mainly useful for internal
15700 testing. The output video should be equal to the input video.
15704 format=monow, pixdesctest
15707 can be used to test the monowhite pixel format descriptor definition.
15711 Display sample values of color channels. Mainly useful for checking color
15712 and levels. Minimum supported resolution is 640x480.
15714 The filters accept the following options:
15718 Set scope X position, relative offset on X axis.
15721 Set scope Y position, relative offset on Y axis.
15730 Set window opacity. This window also holds statistics about pixel area.
15733 Set window X position, relative offset on X axis.
15736 Set window Y position, relative offset on Y axis.
15741 Enable the specified chain of postprocessing subfilters using libpostproc. This
15742 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15743 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15744 Each subfilter and some options have a short and a long name that can be used
15745 interchangeably, i.e. dr/dering are the same.
15747 The filters accept the following options:
15751 Set postprocessing subfilters string.
15754 All subfilters share common options to determine their scope:
15758 Honor the quality commands for this subfilter.
15761 Do chrominance filtering, too (default).
15764 Do luminance filtering only (no chrominance).
15767 Do chrominance filtering only (no luminance).
15770 These options can be appended after the subfilter name, separated by a '|'.
15772 Available subfilters are:
15775 @item hb/hdeblock[|difference[|flatness]]
15776 Horizontal deblocking filter
15779 Difference factor where higher values mean more deblocking (default: @code{32}).
15781 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15784 @item vb/vdeblock[|difference[|flatness]]
15785 Vertical deblocking filter
15788 Difference factor where higher values mean more deblocking (default: @code{32}).
15790 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15793 @item ha/hadeblock[|difference[|flatness]]
15794 Accurate horizontal deblocking filter
15797 Difference factor where higher values mean more deblocking (default: @code{32}).
15799 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15802 @item va/vadeblock[|difference[|flatness]]
15803 Accurate vertical deblocking filter
15806 Difference factor where higher values mean more deblocking (default: @code{32}).
15808 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15812 The horizontal and vertical deblocking filters share the difference and
15813 flatness values so you cannot set different horizontal and vertical
15817 @item h1/x1hdeblock
15818 Experimental horizontal deblocking filter
15820 @item v1/x1vdeblock
15821 Experimental vertical deblocking filter
15826 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15829 larger -> stronger filtering
15831 larger -> stronger filtering
15833 larger -> stronger filtering
15836 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15839 Stretch luminance to @code{0-255}.
15842 @item lb/linblenddeint
15843 Linear blend deinterlacing filter that deinterlaces the given block by
15844 filtering all lines with a @code{(1 2 1)} filter.
15846 @item li/linipoldeint
15847 Linear interpolating deinterlacing filter that deinterlaces the given block by
15848 linearly interpolating every second line.
15850 @item ci/cubicipoldeint
15851 Cubic interpolating deinterlacing filter deinterlaces the given block by
15852 cubically interpolating every second line.
15854 @item md/mediandeint
15855 Median deinterlacing filter that deinterlaces the given block by applying a
15856 median filter to every second line.
15858 @item fd/ffmpegdeint
15859 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15860 second line with a @code{(-1 4 2 4 -1)} filter.
15863 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15864 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15866 @item fq/forceQuant[|quantizer]
15867 Overrides the quantizer table from the input with the constant quantizer you
15875 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15878 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15881 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15884 @subsection Examples
15888 Apply horizontal and vertical deblocking, deringing and automatic
15889 brightness/contrast:
15895 Apply default filters without brightness/contrast correction:
15901 Apply default filters and temporal denoiser:
15903 pp=default/tmpnoise|1|2|3
15907 Apply deblocking on luminance only, and switch vertical deblocking on or off
15908 automatically depending on available CPU time:
15915 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15916 similar to spp = 6 with 7 point DCT, where only the center sample is
15919 The filter accepts the following options:
15923 Force a constant quantization parameter. It accepts an integer in range
15924 0 to 63. If not set, the filter will use the QP from the video stream
15928 Set thresholding mode. Available modes are:
15932 Set hard thresholding.
15934 Set soft thresholding (better de-ringing effect, but likely blurrier).
15936 Set medium thresholding (good results, default).
15940 @section premultiply
15941 Apply alpha premultiply effect to input video stream using first plane
15942 of second stream as alpha.
15944 Both streams must have same dimensions and same pixel format.
15946 The filter accepts the following option:
15950 Set which planes will be processed, unprocessed planes will be copied.
15951 By default value 0xf, all planes will be processed.
15954 Do not require 2nd input for processing, instead use alpha plane from input stream.
15958 Apply prewitt operator to input video stream.
15960 The filter accepts the following option:
15964 Set which planes will be processed, unprocessed planes will be copied.
15965 By default value 0xf, all planes will be processed.
15968 Set value which will be multiplied with filtered result.
15971 Set value which will be added to filtered result.
15974 @subsection Commands
15976 This filter supports the all above options as @ref{commands}.
15978 @section pseudocolor
15980 Alter frame colors in video with pseudocolors.
15982 This filter accepts the following options:
15986 set pixel first component expression
15989 set pixel second component expression
15992 set pixel third component expression
15995 set pixel fourth component expression, corresponds to the alpha component
15998 set component to use as base for altering colors
16001 Each of them specifies the expression to use for computing the lookup table for
16002 the corresponding pixel component values.
16004 The expressions can contain the following constants and functions:
16009 The input width and height.
16012 The input value for the pixel component.
16014 @item ymin, umin, vmin, amin
16015 The minimum allowed component value.
16017 @item ymax, umax, vmax, amax
16018 The maximum allowed component value.
16021 All expressions default to "val".
16023 @subsection Examples
16027 Change too high luma values to gradient:
16029 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'"
16035 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16036 Ratio) between two input videos.
16038 This filter takes in input two input videos, the first input is
16039 considered the "main" source and is passed unchanged to the
16040 output. The second input is used as a "reference" video for computing
16043 Both video inputs must have the same resolution and pixel format for
16044 this filter to work correctly. Also it assumes that both inputs
16045 have the same number of frames, which are compared one by one.
16047 The obtained average PSNR is printed through the logging system.
16049 The filter stores the accumulated MSE (mean squared error) of each
16050 frame, and at the end of the processing it is averaged across all frames
16051 equally, and the following formula is applied to obtain the PSNR:
16054 PSNR = 10*log10(MAX^2/MSE)
16057 Where MAX is the average of the maximum values of each component of the
16060 The description of the accepted parameters follows.
16063 @item stats_file, f
16064 If specified the filter will use the named file to save the PSNR of
16065 each individual frame. When filename equals "-" the data is sent to
16068 @item stats_version
16069 Specifies which version of the stats file format to use. Details of
16070 each format are written below.
16071 Default value is 1.
16073 @item stats_add_max
16074 Determines whether the max value is output to the stats log.
16075 Default value is 0.
16076 Requires stats_version >= 2. If this is set and stats_version < 2,
16077 the filter will return an error.
16080 This filter also supports the @ref{framesync} options.
16082 The file printed if @var{stats_file} is selected, contains a sequence of
16083 key/value pairs of the form @var{key}:@var{value} for each compared
16086 If a @var{stats_version} greater than 1 is specified, a header line precedes
16087 the list of per-frame-pair stats, with key value pairs following the frame
16088 format with the following parameters:
16091 @item psnr_log_version
16092 The version of the log file format. Will match @var{stats_version}.
16095 A comma separated list of the per-frame-pair parameters included in
16099 A description of each shown per-frame-pair parameter follows:
16103 sequential number of the input frame, starting from 1
16106 Mean Square Error pixel-by-pixel average difference of the compared
16107 frames, averaged over all the image components.
16109 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16110 Mean Square Error pixel-by-pixel average difference of the compared
16111 frames for the component specified by the suffix.
16113 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16114 Peak Signal to Noise ratio of the compared frames for the component
16115 specified by the suffix.
16117 @item max_avg, max_y, max_u, max_v
16118 Maximum allowed value for each channel, and average over all
16122 @subsection Examples
16127 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16128 [main][ref] psnr="stats_file=stats.log" [out]
16131 On this example the input file being processed is compared with the
16132 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16133 is stored in @file{stats.log}.
16136 Another example with different containers:
16138 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 -
16145 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16146 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16149 The pullup filter is designed to take advantage of future context in making
16150 its decisions. This filter is stateless in the sense that it does not lock
16151 onto a pattern to follow, but it instead looks forward to the following
16152 fields in order to identify matches and rebuild progressive frames.
16154 To produce content with an even framerate, insert the fps filter after
16155 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16156 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16158 The filter accepts the following options:
16165 These options set the amount of "junk" to ignore at the left, right, top, and
16166 bottom of the image, respectively. Left and right are in units of 8 pixels,
16167 while top and bottom are in units of 2 lines.
16168 The default is 8 pixels on each side.
16171 Set the strict breaks. Setting this option to 1 will reduce the chances of
16172 filter generating an occasional mismatched frame, but it may also cause an
16173 excessive number of frames to be dropped during high motion sequences.
16174 Conversely, setting it to -1 will make filter match fields more easily.
16175 This may help processing of video where there is slight blurring between
16176 the fields, but may also cause there to be interlaced frames in the output.
16177 Default value is @code{0}.
16180 Set the metric plane to use. It accepts the following values:
16186 Use chroma blue plane.
16189 Use chroma red plane.
16192 This option may be set to use chroma plane instead of the default luma plane
16193 for doing filter's computations. This may improve accuracy on very clean
16194 source material, but more likely will decrease accuracy, especially if there
16195 is chroma noise (rainbow effect) or any grayscale video.
16196 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16197 load and make pullup usable in realtime on slow machines.
16200 For best results (without duplicated frames in the output file) it is
16201 necessary to change the output frame rate. For example, to inverse
16202 telecine NTSC input:
16204 ffmpeg -i input -vf pullup -r 24000/1001 ...
16209 Change video quantization parameters (QP).
16211 The filter accepts the following option:
16215 Set expression for quantization parameter.
16218 The expression is evaluated through the eval API and can contain, among others,
16219 the following constants:
16223 1 if index is not 129, 0 otherwise.
16226 Sequential index starting from -129 to 128.
16229 @subsection Examples
16233 Some equation like:
16241 Flush video frames from internal cache of frames into a random order.
16242 No frame is discarded.
16243 Inspired by @ref{frei0r} nervous filter.
16247 Set size in number of frames of internal cache, in range from @code{2} to
16248 @code{512}. Default is @code{30}.
16251 Set seed for random number generator, must be an integer included between
16252 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16253 less than @code{0}, the filter will try to use a good random seed on a
16257 @section readeia608
16259 Read closed captioning (EIA-608) information from the top lines of a video frame.
16261 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16262 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16263 with EIA-608 data (starting from 0). A description of each metadata value follows:
16266 @item lavfi.readeia608.X.cc
16267 The two bytes stored as EIA-608 data (printed in hexadecimal).
16269 @item lavfi.readeia608.X.line
16270 The number of the line on which the EIA-608 data was identified and read.
16273 This filter accepts the following options:
16277 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16280 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16283 Set the ratio of width reserved for sync code detection.
16284 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16287 Enable checking the parity bit. In the event of a parity error, the filter will output
16288 @code{0x00} for that character. Default is false.
16291 Lowpass lines prior to further processing. Default is enabled.
16294 @subsection Commands
16296 This filter supports the all above options as @ref{commands}.
16298 @subsection Examples
16302 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16304 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
16310 Read vertical interval timecode (VITC) information from the top lines of a
16313 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16314 timecode value, if a valid timecode has been detected. Further metadata key
16315 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16316 timecode data has been found or not.
16318 This filter accepts the following options:
16322 Set the maximum number of lines to scan for VITC data. If the value is set to
16323 @code{-1} the full video frame is scanned. Default is @code{45}.
16326 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16327 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16330 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16331 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16334 @subsection Examples
16338 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16339 draw @code{--:--:--:--} as a placeholder:
16341 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16347 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16349 Destination pixel at position (X, Y) will be picked from source (x, y) position
16350 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16351 value for pixel will be used for destination pixel.
16353 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16354 will have Xmap/Ymap video stream dimensions.
16355 Xmap and Ymap input video streams are 16bit depth, single channel.
16359 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16360 Default is @code{color}.
16363 Specify the color of the unmapped pixels. For the syntax of this option,
16364 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16365 manual,ffmpeg-utils}. Default color is @code{black}.
16368 @section removegrain
16370 The removegrain filter is a spatial denoiser for progressive video.
16374 Set mode for the first plane.
16377 Set mode for the second plane.
16380 Set mode for the third plane.
16383 Set mode for the fourth plane.
16386 Range of mode is from 0 to 24. Description of each mode follows:
16390 Leave input plane unchanged. Default.
16393 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16396 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16399 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16402 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16403 This is equivalent to a median filter.
16406 Line-sensitive clipping giving the minimal change.
16409 Line-sensitive clipping, intermediate.
16412 Line-sensitive clipping, intermediate.
16415 Line-sensitive clipping, intermediate.
16418 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16421 Replaces the target pixel with the closest neighbour.
16424 [1 2 1] horizontal and vertical kernel blur.
16430 Bob mode, interpolates top field from the line where the neighbours
16431 pixels are the closest.
16434 Bob mode, interpolates bottom field from the line where the neighbours
16435 pixels are the closest.
16438 Bob mode, interpolates top field. Same as 13 but with a more complicated
16439 interpolation formula.
16442 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16443 interpolation formula.
16446 Clips the pixel with the minimum and maximum of respectively the maximum and
16447 minimum of each pair of opposite neighbour pixels.
16450 Line-sensitive clipping using opposite neighbours whose greatest distance from
16451 the current pixel is minimal.
16454 Replaces the pixel with the average of its 8 neighbours.
16457 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16460 Clips pixels using the averages of opposite neighbour.
16463 Same as mode 21 but simpler and faster.
16466 Small edge and halo removal, but reputed useless.
16472 @section removelogo
16474 Suppress a TV station logo, using an image file to determine which
16475 pixels comprise the logo. It works by filling in the pixels that
16476 comprise the logo with neighboring pixels.
16478 The filter accepts the following options:
16482 Set the filter bitmap file, which can be any image format supported by
16483 libavformat. The width and height of the image file must match those of the
16484 video stream being processed.
16487 Pixels in the provided bitmap image with a value of zero are not
16488 considered part of the logo, non-zero pixels are considered part of
16489 the logo. If you use white (255) for the logo and black (0) for the
16490 rest, you will be safe. For making the filter bitmap, it is
16491 recommended to take a screen capture of a black frame with the logo
16492 visible, and then using a threshold filter followed by the erode
16493 filter once or twice.
16495 If needed, little splotches can be fixed manually. Remember that if
16496 logo pixels are not covered, the filter quality will be much
16497 reduced. Marking too many pixels as part of the logo does not hurt as
16498 much, but it will increase the amount of blurring needed to cover over
16499 the image and will destroy more information than necessary, and extra
16500 pixels will slow things down on a large logo.
16502 @section repeatfields
16504 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16505 fields based on its value.
16509 Reverse a video clip.
16511 Warning: This filter requires memory to buffer the entire clip, so trimming
16514 @subsection Examples
16518 Take the first 5 seconds of a clip, and reverse it.
16525 Shift R/G/B/A pixels horizontally and/or vertically.
16527 The filter accepts the following options:
16530 Set amount to shift red horizontally.
16532 Set amount to shift red vertically.
16534 Set amount to shift green horizontally.
16536 Set amount to shift green vertically.
16538 Set amount to shift blue horizontally.
16540 Set amount to shift blue vertically.
16542 Set amount to shift alpha horizontally.
16544 Set amount to shift alpha vertically.
16546 Set edge mode, can be @var{smear}, default, or @var{warp}.
16549 @subsection Commands
16551 This filter supports the all above options as @ref{commands}.
16554 Apply roberts cross operator to input video stream.
16556 The filter accepts the following option:
16560 Set which planes will be processed, unprocessed planes will be copied.
16561 By default value 0xf, all planes will be processed.
16564 Set value which will be multiplied with filtered result.
16567 Set value which will be added to filtered result.
16570 @subsection Commands
16572 This filter supports the all above options as @ref{commands}.
16576 Rotate video by an arbitrary angle expressed in radians.
16578 The filter accepts the following options:
16580 A description of the optional parameters follows.
16583 Set an expression for the angle by which to rotate the input video
16584 clockwise, expressed as a number of radians. A negative value will
16585 result in a counter-clockwise rotation. By default it is set to "0".
16587 This expression is evaluated for each frame.
16590 Set the output width expression, default value is "iw".
16591 This expression is evaluated just once during configuration.
16594 Set the output height expression, default value is "ih".
16595 This expression is evaluated just once during configuration.
16598 Enable bilinear interpolation if set to 1, a value of 0 disables
16599 it. Default value is 1.
16602 Set the color used to fill the output area not covered by the rotated
16603 image. For the general syntax of this option, check the
16604 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16605 If the special value "none" is selected then no
16606 background is printed (useful for example if the background is never shown).
16608 Default value is "black".
16611 The expressions for the angle and the output size can contain the
16612 following constants and functions:
16616 sequential number of the input frame, starting from 0. It is always NAN
16617 before the first frame is filtered.
16620 time in seconds of the input frame, it is set to 0 when the filter is
16621 configured. It is always NAN before the first frame is filtered.
16625 horizontal and vertical chroma subsample values. For example for the
16626 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16630 the input video width and height
16634 the output width and height, that is the size of the padded area as
16635 specified by the @var{width} and @var{height} expressions
16639 the minimal width/height required for completely containing the input
16640 video rotated by @var{a} radians.
16642 These are only available when computing the @option{out_w} and
16643 @option{out_h} expressions.
16646 @subsection Examples
16650 Rotate the input by PI/6 radians clockwise:
16656 Rotate the input by PI/6 radians counter-clockwise:
16662 Rotate the input by 45 degrees clockwise:
16668 Apply a constant rotation with period T, starting from an angle of PI/3:
16670 rotate=PI/3+2*PI*t/T
16674 Make the input video rotation oscillating with a period of T
16675 seconds and an amplitude of A radians:
16677 rotate=A*sin(2*PI/T*t)
16681 Rotate the video, output size is chosen so that the whole rotating
16682 input video is always completely contained in the output:
16684 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16688 Rotate the video, reduce the output size so that no background is ever
16691 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16695 @subsection Commands
16697 The filter supports the following commands:
16701 Set the angle expression.
16702 The command accepts the same syntax of the corresponding option.
16704 If the specified expression is not valid, it is kept at its current
16710 Apply Shape Adaptive Blur.
16712 The filter accepts the following options:
16715 @item luma_radius, lr
16716 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16717 value is 1.0. A greater value will result in a more blurred image, and
16718 in slower processing.
16720 @item luma_pre_filter_radius, lpfr
16721 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16724 @item luma_strength, ls
16725 Set luma maximum difference between pixels to still be considered, must
16726 be a value in the 0.1-100.0 range, default value is 1.0.
16728 @item chroma_radius, cr
16729 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16730 greater value will result in a more blurred image, and in slower
16733 @item chroma_pre_filter_radius, cpfr
16734 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16736 @item chroma_strength, cs
16737 Set chroma maximum difference between pixels to still be considered,
16738 must be a value in the -0.9-100.0 range.
16741 Each chroma option value, if not explicitly specified, is set to the
16742 corresponding luma option value.
16747 Scale (resize) the input video, using the libswscale library.
16749 The scale filter forces the output display aspect ratio to be the same
16750 of the input, by changing the output sample aspect ratio.
16752 If the input image format is different from the format requested by
16753 the next filter, the scale filter will convert the input to the
16756 @subsection Options
16757 The filter accepts the following options, or any of the options
16758 supported by the libswscale scaler.
16760 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16761 the complete list of scaler options.
16766 Set the output video dimension expression. Default value is the input
16769 If the @var{width} or @var{w} value is 0, the input width is used for
16770 the output. If the @var{height} or @var{h} value is 0, the input height
16771 is used for the output.
16773 If one and only one of the values is -n with n >= 1, the scale filter
16774 will use a value that maintains the aspect ratio of the input image,
16775 calculated from the other specified dimension. After that it will,
16776 however, make sure that the calculated dimension is divisible by n and
16777 adjust the value if necessary.
16779 If both values are -n with n >= 1, the behavior will be identical to
16780 both values being set to 0 as previously detailed.
16782 See below for the list of accepted constants for use in the dimension
16786 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16790 Only evaluate expressions once during the filter initialization or when a command is processed.
16793 Evaluate expressions for each incoming frame.
16797 Default value is @samp{init}.
16801 Set the interlacing mode. It accepts the following values:
16805 Force interlaced aware scaling.
16808 Do not apply interlaced scaling.
16811 Select interlaced aware scaling depending on whether the source frames
16812 are flagged as interlaced or not.
16815 Default value is @samp{0}.
16818 Set libswscale scaling flags. See
16819 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16820 complete list of values. If not explicitly specified the filter applies
16824 @item param0, param1
16825 Set libswscale input parameters for scaling algorithms that need them. See
16826 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16827 complete documentation. If not explicitly specified the filter applies
16833 Set the video size. For the syntax of this option, check the
16834 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16836 @item in_color_matrix
16837 @item out_color_matrix
16838 Set in/output YCbCr color space type.
16840 This allows the autodetected value to be overridden as well as allows forcing
16841 a specific value used for the output and encoder.
16843 If not specified, the color space type depends on the pixel format.
16849 Choose automatically.
16852 Format conforming to International Telecommunication Union (ITU)
16853 Recommendation BT.709.
16856 Set color space conforming to the United States Federal Communications
16857 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16862 Set color space conforming to:
16866 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16869 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16872 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16877 Set color space conforming to SMPTE ST 240:1999.
16880 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16885 Set in/output YCbCr sample range.
16887 This allows the autodetected value to be overridden as well as allows forcing
16888 a specific value used for the output and encoder. If not specified, the
16889 range depends on the pixel format. Possible values:
16893 Choose automatically.
16896 Set full range (0-255 in case of 8-bit luma).
16898 @item mpeg/limited/tv
16899 Set "MPEG" range (16-235 in case of 8-bit luma).
16902 @item force_original_aspect_ratio
16903 Enable decreasing or increasing output video width or height if necessary to
16904 keep the original aspect ratio. Possible values:
16908 Scale the video as specified and disable this feature.
16911 The output video dimensions will automatically be decreased if needed.
16914 The output video dimensions will automatically be increased if needed.
16918 One useful instance of this option is that when you know a specific device's
16919 maximum allowed resolution, you can use this to limit the output video to
16920 that, while retaining the aspect ratio. For example, device A allows
16921 1280x720 playback, and your video is 1920x800. Using this option (set it to
16922 decrease) and specifying 1280x720 to the command line makes the output
16925 Please note that this is a different thing than specifying -1 for @option{w}
16926 or @option{h}, you still need to specify the output resolution for this option
16929 @item force_divisible_by
16930 Ensures that both the output dimensions, width and height, are divisible by the
16931 given integer when used together with @option{force_original_aspect_ratio}. This
16932 works similar to using @code{-n} in the @option{w} and @option{h} options.
16934 This option respects the value set for @option{force_original_aspect_ratio},
16935 increasing or decreasing the resolution accordingly. The video's aspect ratio
16936 may be slightly modified.
16938 This option can be handy if you need to have a video fit within or exceed
16939 a defined resolution using @option{force_original_aspect_ratio} but also have
16940 encoder restrictions on width or height divisibility.
16944 The values of the @option{w} and @option{h} options are expressions
16945 containing the following constants:
16950 The input width and height
16954 These are the same as @var{in_w} and @var{in_h}.
16958 The output (scaled) width and height
16962 These are the same as @var{out_w} and @var{out_h}
16965 The same as @var{iw} / @var{ih}
16968 input sample aspect ratio
16971 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16975 horizontal and vertical input chroma subsample values. For example for the
16976 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16980 horizontal and vertical output chroma subsample values. For example for the
16981 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16984 The (sequential) number of the input frame, starting from 0.
16985 Only available with @code{eval=frame}.
16988 The presentation timestamp of the input frame, expressed as a number of
16989 seconds. Only available with @code{eval=frame}.
16992 The position (byte offset) of the frame in the input stream, or NaN if
16993 this information is unavailable and/or meaningless (for example in case of synthetic video).
16994 Only available with @code{eval=frame}.
16997 @subsection Examples
17001 Scale the input video to a size of 200x100
17006 This is equivalent to:
17017 Specify a size abbreviation for the output size:
17022 which can also be written as:
17028 Scale the input to 2x:
17030 scale=w=2*iw:h=2*ih
17034 The above is the same as:
17036 scale=2*in_w:2*in_h
17040 Scale the input to 2x with forced interlaced scaling:
17042 scale=2*iw:2*ih:interl=1
17046 Scale the input to half size:
17048 scale=w=iw/2:h=ih/2
17052 Increase the width, and set the height to the same size:
17058 Seek Greek harmony:
17065 Increase the height, and set the width to 3/2 of the height:
17067 scale=w=3/2*oh:h=3/5*ih
17071 Increase the size, making the size a multiple of the chroma
17074 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17078 Increase the width to a maximum of 500 pixels,
17079 keeping the same aspect ratio as the input:
17081 scale=w='min(500\, iw*3/2):h=-1'
17085 Make pixels square by combining scale and setsar:
17087 scale='trunc(ih*dar):ih',setsar=1/1
17091 Make pixels square by combining scale and setsar,
17092 making sure the resulting resolution is even (required by some codecs):
17094 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17098 @subsection Commands
17100 This filter supports the following commands:
17104 Set the output video dimension expression.
17105 The command accepts the same syntax of the corresponding option.
17107 If the specified expression is not valid, it is kept at its current
17113 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17114 format conversion on CUDA video frames. Setting the output width and height
17115 works in the same way as for the @var{scale} filter.
17117 The following additional options are accepted:
17120 The pixel format of the output CUDA frames. If set to the string "same" (the
17121 default), the input format will be kept. Note that automatic format negotiation
17122 and conversion is not yet supported for hardware frames
17125 The interpolation algorithm used for resizing. One of the following:
17132 @item cubic2p_bspline
17133 2-parameter cubic (B=1, C=0)
17135 @item cubic2p_catmullrom
17136 2-parameter cubic (B=0, C=1/2)
17138 @item cubic2p_b05c03
17139 2-parameter cubic (B=1/2, C=3/10)
17147 @item force_original_aspect_ratio
17148 Enable decreasing or increasing output video width or height if necessary to
17149 keep the original aspect ratio. Possible values:
17153 Scale the video as specified and disable this feature.
17156 The output video dimensions will automatically be decreased if needed.
17159 The output video dimensions will automatically be increased if needed.
17163 One useful instance of this option is that when you know a specific device's
17164 maximum allowed resolution, you can use this to limit the output video to
17165 that, while retaining the aspect ratio. For example, device A allows
17166 1280x720 playback, and your video is 1920x800. Using this option (set it to
17167 decrease) and specifying 1280x720 to the command line makes the output
17170 Please note that this is a different thing than specifying -1 for @option{w}
17171 or @option{h}, you still need to specify the output resolution for this option
17174 @item force_divisible_by
17175 Ensures that both the output dimensions, width and height, are divisible by the
17176 given integer when used together with @option{force_original_aspect_ratio}. This
17177 works similar to using @code{-n} in the @option{w} and @option{h} options.
17179 This option respects the value set for @option{force_original_aspect_ratio},
17180 increasing or decreasing the resolution accordingly. The video's aspect ratio
17181 may be slightly modified.
17183 This option can be handy if you need to have a video fit within or exceed
17184 a defined resolution using @option{force_original_aspect_ratio} but also have
17185 encoder restrictions on width or height divisibility.
17191 Scale (resize) the input video, based on a reference video.
17193 See the scale filter for available options, scale2ref supports the same but
17194 uses the reference video instead of the main input as basis. scale2ref also
17195 supports the following additional constants for the @option{w} and
17196 @option{h} options:
17201 The main input video's width and height
17204 The same as @var{main_w} / @var{main_h}
17207 The main input video's sample aspect ratio
17209 @item main_dar, mdar
17210 The main input video's display aspect ratio. Calculated from
17211 @code{(main_w / main_h) * main_sar}.
17215 The main input video's horizontal and vertical chroma subsample values.
17216 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17220 The (sequential) number of the main input frame, starting from 0.
17221 Only available with @code{eval=frame}.
17224 The presentation timestamp of the main input frame, expressed as a number of
17225 seconds. Only available with @code{eval=frame}.
17228 The position (byte offset) of the frame in the main input stream, or NaN if
17229 this information is unavailable and/or meaningless (for example in case of synthetic video).
17230 Only available with @code{eval=frame}.
17233 @subsection Examples
17237 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17239 'scale2ref[b][a];[a][b]overlay'
17243 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17245 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17249 @subsection Commands
17251 This filter supports the following commands:
17255 Set the output video dimension expression.
17256 The command accepts the same syntax of the corresponding option.
17258 If the specified expression is not valid, it is kept at its current
17263 Scroll input video horizontally and/or vertically by constant speed.
17265 The filter accepts the following options:
17267 @item horizontal, h
17268 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17269 Negative values changes scrolling direction.
17272 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17273 Negative values changes scrolling direction.
17276 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17279 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17282 @subsection Commands
17284 This filter supports the following @ref{commands}:
17286 @item horizontal, h
17287 Set the horizontal scrolling speed.
17289 Set the vertical scrolling speed.
17295 Detect video scene change.
17297 This filter sets frame metadata with mafd between frame, the scene score, and
17298 forward the frame to the next filter, so they can use these metadata to detect
17299 scene change or others.
17301 In addition, this filter logs a message and sets frame metadata when it detects
17302 a scene change by @option{threshold}.
17304 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17306 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17307 to detect scene change.
17309 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17310 detect scene change with @option{threshold}.
17312 The filter accepts the following options:
17316 Set the scene change detection threshold as a percentage of maximum change. Good
17317 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17320 Default value is @code{10.}.
17323 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17324 You can enable it if you want to get snapshot of scene change frames only.
17327 @anchor{selectivecolor}
17328 @section selectivecolor
17330 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17331 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17332 by the "purity" of the color (that is, how saturated it already is).
17334 This filter is similar to the Adobe Photoshop Selective Color tool.
17336 The filter accepts the following options:
17339 @item correction_method
17340 Select color correction method.
17342 Available values are:
17345 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17348 Specified adjustments are relative to the original component value.
17350 Default is @code{absolute}.
17352 Adjustments for red pixels (pixels where the red component is the maximum)
17354 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17356 Adjustments for green pixels (pixels where the green component is the maximum)
17358 Adjustments for cyan pixels (pixels where the red component is the minimum)
17360 Adjustments for blue pixels (pixels where the blue component is the maximum)
17362 Adjustments for magenta pixels (pixels where the green component is the minimum)
17364 Adjustments for white pixels (pixels where all components are greater than 128)
17366 Adjustments for all pixels except pure black and pure white
17368 Adjustments for black pixels (pixels where all components are lesser than 128)
17370 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17373 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17374 4 space separated floating point adjustment values in the [-1,1] range,
17375 respectively to adjust the amount of cyan, magenta, yellow and black for the
17376 pixels of its range.
17378 @subsection Examples
17382 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17383 increase magenta by 27% in blue areas:
17385 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17389 Use a Photoshop selective color preset:
17391 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17395 @anchor{separatefields}
17396 @section separatefields
17398 The @code{separatefields} takes a frame-based video input and splits
17399 each frame into its components fields, producing a new half height clip
17400 with twice the frame rate and twice the frame count.
17402 This filter use field-dominance information in frame to decide which
17403 of each pair of fields to place first in the output.
17404 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17406 @section setdar, setsar
17408 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17411 This is done by changing the specified Sample (aka Pixel) Aspect
17412 Ratio, according to the following equation:
17414 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17417 Keep in mind that the @code{setdar} filter does not modify the pixel
17418 dimensions of the video frame. Also, the display aspect ratio set by
17419 this filter may be changed by later filters in the filterchain,
17420 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17423 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17424 the filter output video.
17426 Note that as a consequence of the application of this filter, the
17427 output display aspect ratio will change according to the equation
17430 Keep in mind that the sample aspect ratio set by the @code{setsar}
17431 filter may be changed by later filters in the filterchain, e.g. if
17432 another "setsar" or a "setdar" filter is applied.
17434 It accepts the following parameters:
17437 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17438 Set the aspect ratio used by the filter.
17440 The parameter can be a floating point number string, an expression, or
17441 a string of the form @var{num}:@var{den}, where @var{num} and
17442 @var{den} are the numerator and denominator of the aspect ratio. If
17443 the parameter is not specified, it is assumed the value "0".
17444 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17448 Set the maximum integer value to use for expressing numerator and
17449 denominator when reducing the expressed aspect ratio to a rational.
17450 Default value is @code{100}.
17454 The parameter @var{sar} is an expression containing
17455 the following constants:
17459 These are approximated values for the mathematical constants e
17460 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17463 The input width and height.
17466 These are the same as @var{w} / @var{h}.
17469 The input sample aspect ratio.
17472 The input display aspect ratio. It is the same as
17473 (@var{w} / @var{h}) * @var{sar}.
17476 Horizontal and vertical chroma subsample values. For example, for the
17477 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17480 @subsection Examples
17485 To change the display aspect ratio to 16:9, specify one of the following:
17492 To change the sample aspect ratio to 10:11, specify:
17498 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17499 1000 in the aspect ratio reduction, use the command:
17501 setdar=ratio=16/9:max=1000
17509 Force field for the output video frame.
17511 The @code{setfield} filter marks the interlace type field for the
17512 output frames. It does not change the input frame, but only sets the
17513 corresponding property, which affects how the frame is treated by
17514 following filters (e.g. @code{fieldorder} or @code{yadif}).
17516 The filter accepts the following options:
17521 Available values are:
17525 Keep the same field property.
17528 Mark the frame as bottom-field-first.
17531 Mark the frame as top-field-first.
17534 Mark the frame as progressive.
17541 Force frame parameter for the output video frame.
17543 The @code{setparams} filter marks interlace and color range for the
17544 output frames. It does not change the input frame, but only sets the
17545 corresponding property, which affects how the frame is treated by
17550 Available values are:
17554 Keep the same field property (default).
17557 Mark the frame as bottom-field-first.
17560 Mark the frame as top-field-first.
17563 Mark the frame as progressive.
17567 Available values are:
17571 Keep the same color range property (default).
17573 @item unspecified, unknown
17574 Mark the frame as unspecified color range.
17576 @item limited, tv, mpeg
17577 Mark the frame as limited range.
17579 @item full, pc, jpeg
17580 Mark the frame as full range.
17583 @item color_primaries
17584 Set the color primaries.
17585 Available values are:
17589 Keep the same color primaries property (default).
17606 Set the color transfer.
17607 Available values are:
17611 Keep the same color trc property (default).
17633 Set the colorspace.
17634 Available values are:
17638 Keep the same colorspace property (default).
17651 @item chroma-derived-nc
17652 @item chroma-derived-c
17659 Show a line containing various information for each input video frame.
17660 The input video is not modified.
17662 This filter supports the following options:
17666 Calculate checksums of each plane. By default enabled.
17669 The shown line contains a sequence of key/value pairs of the form
17670 @var{key}:@var{value}.
17672 The following values are shown in the output:
17676 The (sequential) number of the input frame, starting from 0.
17679 The Presentation TimeStamp of the input frame, expressed as a number of
17680 time base units. The time base unit depends on the filter input pad.
17683 The Presentation TimeStamp of the input frame, expressed as a number of
17687 The position of the frame in the input stream, or -1 if this information is
17688 unavailable and/or meaningless (for example in case of synthetic video).
17691 The pixel format name.
17694 The sample aspect ratio of the input frame, expressed in the form
17695 @var{num}/@var{den}.
17698 The size of the input frame. For the syntax of this option, check the
17699 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17702 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17703 for bottom field first).
17706 This is 1 if the frame is a key frame, 0 otherwise.
17709 The picture type of the input frame ("I" for an I-frame, "P" for a
17710 P-frame, "B" for a B-frame, or "?" for an unknown type).
17711 Also refer to the documentation of the @code{AVPictureType} enum and of
17712 the @code{av_get_picture_type_char} function defined in
17713 @file{libavutil/avutil.h}.
17716 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17718 @item plane_checksum
17719 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17720 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17723 The mean value of pixels in each plane of the input frame, expressed in the form
17724 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17727 The standard deviation of pixel values in each plane of the input frame, expressed
17728 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17732 @section showpalette
17734 Displays the 256 colors palette of each frame. This filter is only relevant for
17735 @var{pal8} pixel format frames.
17737 It accepts the following option:
17741 Set the size of the box used to represent one palette color entry. Default is
17742 @code{30} (for a @code{30x30} pixel box).
17745 @section shuffleframes
17747 Reorder and/or duplicate and/or drop video frames.
17749 It accepts the following parameters:
17753 Set the destination indexes of input frames.
17754 This is space or '|' separated list of indexes that maps input frames to output
17755 frames. Number of indexes also sets maximal value that each index may have.
17756 '-1' index have special meaning and that is to drop frame.
17759 The first frame has the index 0. The default is to keep the input unchanged.
17761 @subsection Examples
17765 Swap second and third frame of every three frames of the input:
17767 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17771 Swap 10th and 1st frame of every ten frames of the input:
17773 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17777 @section shuffleplanes
17779 Reorder and/or duplicate video planes.
17781 It accepts the following parameters:
17786 The index of the input plane to be used as the first output plane.
17789 The index of the input plane to be used as the second output plane.
17792 The index of the input plane to be used as the third output plane.
17795 The index of the input plane to be used as the fourth output plane.
17799 The first plane has the index 0. The default is to keep the input unchanged.
17801 @subsection Examples
17805 Swap the second and third planes of the input:
17807 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17811 @anchor{signalstats}
17812 @section signalstats
17813 Evaluate various visual metrics that assist in determining issues associated
17814 with the digitization of analog video media.
17816 By default the filter will log these metadata values:
17820 Display the minimal Y value contained within the input frame. Expressed in
17824 Display the Y value at the 10% percentile within the input frame. Expressed in
17828 Display the average Y value within the input frame. Expressed in range of
17832 Display the Y value at the 90% percentile within the input frame. Expressed in
17836 Display the maximum Y value contained within the input frame. Expressed in
17840 Display the minimal U value contained within the input frame. Expressed in
17844 Display the U value at the 10% percentile within the input frame. Expressed in
17848 Display the average U value within the input frame. Expressed in range of
17852 Display the U value at the 90% percentile within the input frame. Expressed in
17856 Display the maximum U value contained within the input frame. Expressed in
17860 Display the minimal V value contained within the input frame. Expressed in
17864 Display the V value at the 10% percentile within the input frame. Expressed in
17868 Display the average V value within the input frame. Expressed in range of
17872 Display the V value at the 90% percentile within the input frame. Expressed in
17876 Display the maximum V value contained within the input frame. Expressed in
17880 Display the minimal saturation value contained within the input frame.
17881 Expressed in range of [0-~181.02].
17884 Display the saturation value at the 10% percentile within the input frame.
17885 Expressed in range of [0-~181.02].
17888 Display the average saturation value within the input frame. Expressed in range
17892 Display the saturation value at the 90% percentile within the input frame.
17893 Expressed in range of [0-~181.02].
17896 Display the maximum saturation value contained within the input frame.
17897 Expressed in range of [0-~181.02].
17900 Display the median value for hue within the input frame. Expressed in range of
17904 Display the average value for hue within the input frame. Expressed in range of
17908 Display the average of sample value difference between all values of the Y
17909 plane in the current frame and corresponding values of the previous input frame.
17910 Expressed in range of [0-255].
17913 Display the average of sample value difference between all values of the U
17914 plane in the current frame and corresponding values of the previous input frame.
17915 Expressed in range of [0-255].
17918 Display the average of sample value difference between all values of the V
17919 plane in the current frame and corresponding values of the previous input frame.
17920 Expressed in range of [0-255].
17923 Display bit depth of Y plane in current frame.
17924 Expressed in range of [0-16].
17927 Display bit depth of U plane in current frame.
17928 Expressed in range of [0-16].
17931 Display bit depth of V plane in current frame.
17932 Expressed in range of [0-16].
17935 The filter accepts the following options:
17941 @option{stat} specify an additional form of image analysis.
17942 @option{out} output video with the specified type of pixel highlighted.
17944 Both options accept the following values:
17948 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17949 unlike the neighboring pixels of the same field. Examples of temporal outliers
17950 include the results of video dropouts, head clogs, or tape tracking issues.
17953 Identify @var{vertical line repetition}. Vertical line repetition includes
17954 similar rows of pixels within a frame. In born-digital video vertical line
17955 repetition is common, but this pattern is uncommon in video digitized from an
17956 analog source. When it occurs in video that results from the digitization of an
17957 analog source it can indicate concealment from a dropout compensator.
17960 Identify pixels that fall outside of legal broadcast range.
17964 Set the highlight color for the @option{out} option. The default color is
17968 @subsection Examples
17972 Output data of various video metrics:
17974 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17978 Output specific data about the minimum and maximum values of the Y plane per frame:
17980 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17984 Playback video while highlighting pixels that are outside of broadcast range in red.
17986 ffplay example.mov -vf signalstats="out=brng:color=red"
17990 Playback video with signalstats metadata drawn over the frame.
17992 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17995 The contents of signalstat_drawtext.txt used in the command are:
17998 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17999 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18000 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18001 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18009 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18010 input. In this case the matching between the inputs can be calculated additionally.
18011 The filter always passes through the first input. The signature of each stream can
18012 be written into a file.
18014 It accepts the following options:
18018 Enable or disable the matching process.
18020 Available values are:
18024 Disable the calculation of a matching (default).
18026 Calculate the matching for the whole video and output whether the whole video
18027 matches or only parts.
18029 Calculate only until a matching is found or the video ends. Should be faster in
18034 Set the number of inputs. The option value must be a non negative integer.
18035 Default value is 1.
18038 Set the path to which the output is written. If there is more than one input,
18039 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18040 integer), that will be replaced with the input number. If no filename is
18041 specified, no output will be written. This is the default.
18044 Choose the output format.
18046 Available values are:
18050 Use the specified binary representation (default).
18052 Use the specified xml representation.
18056 Set threshold to detect one word as similar. The option value must be an integer
18057 greater than zero. The default value is 9000.
18060 Set threshold to detect all words as similar. The option value must be an integer
18061 greater than zero. The default value is 60000.
18064 Set threshold to detect frames as similar. The option value must be an integer
18065 greater than zero. The default value is 116.
18068 Set the minimum length of a sequence in frames to recognize it as matching
18069 sequence. The option value must be a non negative integer value.
18070 The default value is 0.
18073 Set the minimum relation, that matching frames to all frames must have.
18074 The option value must be a double value between 0 and 1. The default value is 0.5.
18077 @subsection Examples
18081 To calculate the signature of an input video and store it in signature.bin:
18083 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18087 To detect whether two videos match and store the signatures in XML format in
18088 signature0.xml and signature1.xml:
18090 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 -
18098 Blur the input video without impacting the outlines.
18100 It accepts the following options:
18103 @item luma_radius, lr
18104 Set the luma radius. The option value must be a float number in
18105 the range [0.1,5.0] that specifies the variance of the gaussian filter
18106 used to blur the image (slower if larger). Default value is 1.0.
18108 @item luma_strength, ls
18109 Set the luma strength. The option value must be a float number
18110 in the range [-1.0,1.0] that configures the blurring. A value included
18111 in [0.0,1.0] will blur the image whereas a value included in
18112 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18114 @item luma_threshold, lt
18115 Set the luma threshold used as a coefficient to determine
18116 whether a pixel should be blurred or not. The option value must be an
18117 integer in the range [-30,30]. A value of 0 will filter all the image,
18118 a value included in [0,30] will filter flat areas and a value included
18119 in [-30,0] will filter edges. Default value is 0.
18121 @item chroma_radius, cr
18122 Set the chroma radius. The option value must be a float number in
18123 the range [0.1,5.0] that specifies the variance of the gaussian filter
18124 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18126 @item chroma_strength, cs
18127 Set the chroma strength. The option value must be a float number
18128 in the range [-1.0,1.0] that configures the blurring. A value included
18129 in [0.0,1.0] will blur the image whereas a value included in
18130 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18132 @item chroma_threshold, ct
18133 Set the chroma threshold used as a coefficient to determine
18134 whether a pixel should be blurred or not. The option value must be an
18135 integer in the range [-30,30]. A value of 0 will filter all the image,
18136 a value included in [0,30] will filter flat areas and a value included
18137 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18140 If a chroma option is not explicitly set, the corresponding luma value
18144 Apply sobel operator to input video stream.
18146 The filter accepts the following option:
18150 Set which planes will be processed, unprocessed planes will be copied.
18151 By default value 0xf, all planes will be processed.
18154 Set value which will be multiplied with filtered result.
18157 Set value which will be added to filtered result.
18160 @subsection Commands
18162 This filter supports the all above options as @ref{commands}.
18167 Apply a simple postprocessing filter that compresses and decompresses the image
18168 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18169 and average the results.
18171 The filter accepts the following options:
18175 Set quality. This option defines the number of levels for averaging. It accepts
18176 an integer in the range 0-6. If set to @code{0}, the filter will have no
18177 effect. A value of @code{6} means the higher quality. For each increment of
18178 that value the speed drops by a factor of approximately 2. Default value is
18182 Force a constant quantization parameter. If not set, the filter will use the QP
18183 from the video stream (if available).
18186 Set thresholding mode. Available modes are:
18190 Set hard thresholding (default).
18192 Set soft thresholding (better de-ringing effect, but likely blurrier).
18195 @item use_bframe_qp
18196 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18197 option may cause flicker since the B-Frames have often larger QP. Default is
18198 @code{0} (not enabled).
18201 @subsection Commands
18203 This filter supports the following commands:
18205 @item quality, level
18206 Set quality level. The value @code{max} can be used to set the maximum level,
18207 currently @code{6}.
18213 Scale the input by applying one of the super-resolution methods based on
18214 convolutional neural networks. Supported models:
18218 Super-Resolution Convolutional Neural Network model (SRCNN).
18219 See @url{https://arxiv.org/abs/1501.00092}.
18222 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18223 See @url{https://arxiv.org/abs/1609.05158}.
18226 Training scripts as well as scripts for model file (.pb) saving can be found at
18227 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18228 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18230 Native model files (.model) can be generated from TensorFlow model
18231 files (.pb) by using tools/python/convert.py
18233 The filter accepts the following options:
18237 Specify which DNN backend to use for model loading and execution. This option accepts
18238 the following values:
18242 Native implementation of DNN loading and execution.
18245 TensorFlow backend. To enable this backend you
18246 need to install the TensorFlow for C library (see
18247 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18248 @code{--enable-libtensorflow}
18251 Default value is @samp{native}.
18254 Set path to model file specifying network architecture and its parameters.
18255 Note that different backends use different file formats. TensorFlow backend
18256 can load files for both formats, while native backend can load files for only
18260 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18261 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18262 input upscaled using bicubic upscaling with proper scale factor.
18265 This feature can also be finished with @ref{dnn_processing} filter.
18269 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18271 This filter takes in input two input videos, the first input is
18272 considered the "main" source and is passed unchanged to the
18273 output. The second input is used as a "reference" video for computing
18276 Both video inputs must have the same resolution and pixel format for
18277 this filter to work correctly. Also it assumes that both inputs
18278 have the same number of frames, which are compared one by one.
18280 The filter stores the calculated SSIM of each frame.
18282 The description of the accepted parameters follows.
18285 @item stats_file, f
18286 If specified the filter will use the named file to save the SSIM of
18287 each individual frame. When filename equals "-" the data is sent to
18291 The file printed if @var{stats_file} is selected, contains a sequence of
18292 key/value pairs of the form @var{key}:@var{value} for each compared
18295 A description of each shown parameter follows:
18299 sequential number of the input frame, starting from 1
18301 @item Y, U, V, R, G, B
18302 SSIM of the compared frames for the component specified by the suffix.
18305 SSIM of the compared frames for the whole frame.
18308 Same as above but in dB representation.
18311 This filter also supports the @ref{framesync} options.
18313 @subsection Examples
18318 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18319 [main][ref] ssim="stats_file=stats.log" [out]
18322 On this example the input file being processed is compared with the
18323 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18324 is stored in @file{stats.log}.
18327 Another example with both psnr and ssim at same time:
18329 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18333 Another example with different containers:
18335 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 -
18341 Convert between different stereoscopic image formats.
18343 The filters accept the following options:
18347 Set stereoscopic image format of input.
18349 Available values for input image formats are:
18352 side by side parallel (left eye left, right eye right)
18355 side by side crosseye (right eye left, left eye right)
18358 side by side parallel with half width resolution
18359 (left eye left, right eye right)
18362 side by side crosseye with half width resolution
18363 (right eye left, left eye right)
18367 above-below (left eye above, right eye below)
18371 above-below (right eye above, left eye below)
18375 above-below with half height resolution
18376 (left eye above, right eye below)
18380 above-below with half height resolution
18381 (right eye above, left eye below)
18384 alternating frames (left eye first, right eye second)
18387 alternating frames (right eye first, left eye second)
18390 interleaved rows (left eye has top row, right eye starts on next row)
18393 interleaved rows (right eye has top row, left eye starts on next row)
18396 interleaved columns, left eye first
18399 interleaved columns, right eye first
18401 Default value is @samp{sbsl}.
18405 Set stereoscopic image format of output.
18409 side by side parallel (left eye left, right eye right)
18412 side by side crosseye (right eye left, left eye right)
18415 side by side parallel with half width resolution
18416 (left eye left, right eye right)
18419 side by side crosseye with half width resolution
18420 (right eye left, left eye right)
18424 above-below (left eye above, right eye below)
18428 above-below (right eye above, left eye below)
18432 above-below with half height resolution
18433 (left eye above, right eye below)
18437 above-below with half height resolution
18438 (right eye above, left eye below)
18441 alternating frames (left eye first, right eye second)
18444 alternating frames (right eye first, left eye second)
18447 interleaved rows (left eye has top row, right eye starts on next row)
18450 interleaved rows (right eye has top row, left eye starts on next row)
18453 anaglyph red/blue gray
18454 (red filter on left eye, blue filter on right eye)
18457 anaglyph red/green gray
18458 (red filter on left eye, green filter on right eye)
18461 anaglyph red/cyan gray
18462 (red filter on left eye, cyan filter on right eye)
18465 anaglyph red/cyan half colored
18466 (red filter on left eye, cyan filter on right eye)
18469 anaglyph red/cyan color
18470 (red filter on left eye, cyan filter on right eye)
18473 anaglyph red/cyan color optimized with the least squares projection of dubois
18474 (red filter on left eye, cyan filter on right eye)
18477 anaglyph green/magenta gray
18478 (green filter on left eye, magenta filter on right eye)
18481 anaglyph green/magenta half colored
18482 (green filter on left eye, magenta filter on right eye)
18485 anaglyph green/magenta colored
18486 (green filter on left eye, magenta filter on right eye)
18489 anaglyph green/magenta color optimized with the least squares projection of dubois
18490 (green filter on left eye, magenta filter on right eye)
18493 anaglyph yellow/blue gray
18494 (yellow filter on left eye, blue filter on right eye)
18497 anaglyph yellow/blue half colored
18498 (yellow filter on left eye, blue filter on right eye)
18501 anaglyph yellow/blue colored
18502 (yellow filter on left eye, blue filter on right eye)
18505 anaglyph yellow/blue color optimized with the least squares projection of dubois
18506 (yellow filter on left eye, blue filter on right eye)
18509 mono output (left eye only)
18512 mono output (right eye only)
18515 checkerboard, left eye first
18518 checkerboard, right eye first
18521 interleaved columns, left eye first
18524 interleaved columns, right eye first
18530 Default value is @samp{arcd}.
18533 @subsection Examples
18537 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18543 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18549 @section streamselect, astreamselect
18550 Select video or audio streams.
18552 The filter accepts the following options:
18556 Set number of inputs. Default is 2.
18559 Set input indexes to remap to outputs.
18562 @subsection Commands
18564 The @code{streamselect} and @code{astreamselect} filter supports the following
18569 Set input indexes to remap to outputs.
18572 @subsection Examples
18576 Select first 5 seconds 1st stream and rest of time 2nd stream:
18578 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18582 Same as above, but for audio:
18584 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18591 Draw subtitles on top of input video using the libass library.
18593 To enable compilation of this filter you need to configure FFmpeg with
18594 @code{--enable-libass}. This filter also requires a build with libavcodec and
18595 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18596 Alpha) subtitles format.
18598 The filter accepts the following options:
18602 Set the filename of the subtitle file to read. It must be specified.
18604 @item original_size
18605 Specify the size of the original video, the video for which the ASS file
18606 was composed. For the syntax of this option, check the
18607 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18608 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18609 correctly scale the fonts if the aspect ratio has been changed.
18612 Set a directory path containing fonts that can be used by the filter.
18613 These fonts will be used in addition to whatever the font provider uses.
18616 Process alpha channel, by default alpha channel is untouched.
18619 Set subtitles input character encoding. @code{subtitles} filter only. Only
18620 useful if not UTF-8.
18622 @item stream_index, si
18623 Set subtitles stream index. @code{subtitles} filter only.
18626 Override default style or script info parameters of the subtitles. It accepts a
18627 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18630 If the first key is not specified, it is assumed that the first value
18631 specifies the @option{filename}.
18633 For example, to render the file @file{sub.srt} on top of the input
18634 video, use the command:
18639 which is equivalent to:
18641 subtitles=filename=sub.srt
18644 To render the default subtitles stream from file @file{video.mkv}, use:
18646 subtitles=video.mkv
18649 To render the second subtitles stream from that file, use:
18651 subtitles=video.mkv:si=1
18654 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18655 @code{DejaVu Serif}, use:
18657 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18660 @section super2xsai
18662 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18663 Interpolate) pixel art scaling algorithm.
18665 Useful for enlarging pixel art images without reducing sharpness.
18669 Swap two rectangular objects in video.
18671 This filter accepts the following options:
18681 Set 1st rect x coordinate.
18684 Set 1st rect y coordinate.
18687 Set 2nd rect x coordinate.
18690 Set 2nd rect y coordinate.
18692 All expressions are evaluated once for each frame.
18695 The all options are expressions containing the following constants:
18700 The input width and height.
18703 same as @var{w} / @var{h}
18706 input sample aspect ratio
18709 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18712 The number of the input frame, starting from 0.
18715 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18718 the position in the file of the input frame, NAN if unknown
18725 Blend successive video frames.
18731 Apply telecine process to the video.
18733 This filter accepts the following options:
18742 The default value is @code{top}.
18746 A string of numbers representing the pulldown pattern you wish to apply.
18747 The default value is @code{23}.
18751 Some typical patterns:
18756 24p: 2332 (preferred)
18763 24p: 222222222223 ("Euro pulldown")
18768 @section thistogram
18770 Compute and draw a color distribution histogram for the input video across time.
18772 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18773 at certain time, this filter shows also past histograms of number of frames defined
18774 by @code{width} option.
18776 The computed histogram is a representation of the color component
18777 distribution in an image.
18779 The filter accepts the following options:
18783 Set width of single color component output. Default value is @code{0}.
18784 Value of @code{0} means width will be picked from input video.
18785 This also set number of passed histograms to keep.
18786 Allowed range is [0, 8192].
18788 @item display_mode, d
18790 It accepts the following values:
18793 Per color component graphs are placed below each other.
18796 Per color component graphs are placed side by side.
18799 Presents information identical to that in the @code{parade}, except
18800 that the graphs representing color components are superimposed directly
18803 Default is @code{stack}.
18805 @item levels_mode, m
18806 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18807 Default is @code{linear}.
18809 @item components, c
18810 Set what color components to display.
18811 Default is @code{7}.
18814 Set background opacity. Default is @code{0.9}.
18817 Show envelope. Default is disabled.
18820 Set envelope color. Default is @code{gold}.
18825 Available values for slide is:
18828 Draw new frame when right border is reached.
18831 Replace old columns with new ones.
18834 Scroll from right to left.
18837 Scroll from left to right.
18840 Draw single picture.
18843 Default is @code{replace}.
18848 Apply threshold effect to video stream.
18850 This filter needs four video streams to perform thresholding.
18851 First stream is stream we are filtering.
18852 Second stream is holding threshold values, third stream is holding min values,
18853 and last, fourth stream is holding max values.
18855 The filter accepts the following option:
18859 Set which planes will be processed, unprocessed planes will be copied.
18860 By default value 0xf, all planes will be processed.
18863 For example if first stream pixel's component value is less then threshold value
18864 of pixel component from 2nd threshold stream, third stream value will picked,
18865 otherwise fourth stream pixel component value will be picked.
18867 Using color source filter one can perform various types of thresholding:
18869 @subsection Examples
18873 Binary threshold, using gray color as threshold:
18875 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18879 Inverted binary threshold, using gray color as threshold:
18881 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18885 Truncate binary threshold, using gray color as threshold:
18887 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18891 Threshold to zero, using gray color as threshold:
18893 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18897 Inverted threshold to zero, using gray color as threshold:
18899 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18904 Select the most representative frame in a given sequence of consecutive frames.
18906 The filter accepts the following options:
18910 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18911 will pick one of them, and then handle the next batch of @var{n} frames until
18912 the end. Default is @code{100}.
18915 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18916 value will result in a higher memory usage, so a high value is not recommended.
18918 @subsection Examples
18922 Extract one picture each 50 frames:
18928 Complete example of a thumbnail creation with @command{ffmpeg}:
18930 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18937 Tile several successive frames together.
18939 The @ref{untile} filter can do the reverse.
18941 The filter accepts the following options:
18946 Set the grid size (i.e. the number of lines and columns). For the syntax of
18947 this option, check the
18948 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18951 Set the maximum number of frames to render in the given area. It must be less
18952 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18953 the area will be used.
18956 Set the outer border margin in pixels.
18959 Set the inner border thickness (i.e. the number of pixels between frames). For
18960 more advanced padding options (such as having different values for the edges),
18961 refer to the pad video filter.
18964 Specify the color of the unused area. For the syntax of this option, check the
18965 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18966 The default value of @var{color} is "black".
18969 Set the number of frames to overlap when tiling several successive frames together.
18970 The value must be between @code{0} and @var{nb_frames - 1}.
18973 Set the number of frames to initially be empty before displaying first output frame.
18974 This controls how soon will one get first output frame.
18975 The value must be between @code{0} and @var{nb_frames - 1}.
18978 @subsection Examples
18982 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18984 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18986 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18987 duplicating each output frame to accommodate the originally detected frame
18991 Display @code{5} pictures in an area of @code{3x2} frames,
18992 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18993 mixed flat and named options:
18995 tile=3x2:nb_frames=5:padding=7:margin=2
18999 @section tinterlace
19001 Perform various types of temporal field interlacing.
19003 Frames are counted starting from 1, so the first input frame is
19006 The filter accepts the following options:
19011 Specify the mode of the interlacing. This option can also be specified
19012 as a value alone. See below for a list of values for this option.
19014 Available values are:
19018 Move odd frames into the upper field, even into the lower field,
19019 generating a double height frame at half frame rate.
19023 Frame 1 Frame 2 Frame 3 Frame 4
19025 11111 22222 33333 44444
19026 11111 22222 33333 44444
19027 11111 22222 33333 44444
19028 11111 22222 33333 44444
19042 Only output odd frames, even frames are dropped, generating a frame with
19043 unchanged height at half frame rate.
19048 Frame 1 Frame 2 Frame 3 Frame 4
19050 11111 22222 33333 44444
19051 11111 22222 33333 44444
19052 11111 22222 33333 44444
19053 11111 22222 33333 44444
19063 Only output even frames, odd frames are dropped, generating a frame with
19064 unchanged height at half frame rate.
19069 Frame 1 Frame 2 Frame 3 Frame 4
19071 11111 22222 33333 44444
19072 11111 22222 33333 44444
19073 11111 22222 33333 44444
19074 11111 22222 33333 44444
19084 Expand each frame to full height, but pad alternate lines with black,
19085 generating a frame with double height at the same input frame rate.
19090 Frame 1 Frame 2 Frame 3 Frame 4
19092 11111 22222 33333 44444
19093 11111 22222 33333 44444
19094 11111 22222 33333 44444
19095 11111 22222 33333 44444
19098 11111 ..... 33333 .....
19099 ..... 22222 ..... 44444
19100 11111 ..... 33333 .....
19101 ..... 22222 ..... 44444
19102 11111 ..... 33333 .....
19103 ..... 22222 ..... 44444
19104 11111 ..... 33333 .....
19105 ..... 22222 ..... 44444
19109 @item interleave_top, 4
19110 Interleave the upper field from odd frames with the lower field from
19111 even frames, generating a frame with unchanged height at half frame rate.
19116 Frame 1 Frame 2 Frame 3 Frame 4
19118 11111<- 22222 33333<- 44444
19119 11111 22222<- 33333 44444<-
19120 11111<- 22222 33333<- 44444
19121 11111 22222<- 33333 44444<-
19131 @item interleave_bottom, 5
19132 Interleave the lower field from odd frames with the upper field from
19133 even frames, generating a frame with unchanged height at half frame rate.
19138 Frame 1 Frame 2 Frame 3 Frame 4
19140 11111 22222<- 33333 44444<-
19141 11111<- 22222 33333<- 44444
19142 11111 22222<- 33333 44444<-
19143 11111<- 22222 33333<- 44444
19153 @item interlacex2, 6
19154 Double frame rate with unchanged height. Frames are inserted each
19155 containing the second temporal field from the previous input frame and
19156 the first temporal field from the next input frame. This mode relies on
19157 the top_field_first flag. Useful for interlaced video displays with no
19158 field synchronisation.
19163 Frame 1 Frame 2 Frame 3 Frame 4
19165 11111 22222 33333 44444
19166 11111 22222 33333 44444
19167 11111 22222 33333 44444
19168 11111 22222 33333 44444
19171 11111 22222 22222 33333 33333 44444 44444
19172 11111 11111 22222 22222 33333 33333 44444
19173 11111 22222 22222 33333 33333 44444 44444
19174 11111 11111 22222 22222 33333 33333 44444
19179 Move odd frames into the upper field, even into the lower field,
19180 generating a double height frame at same frame rate.
19185 Frame 1 Frame 2 Frame 3 Frame 4
19187 11111 22222 33333 44444
19188 11111 22222 33333 44444
19189 11111 22222 33333 44444
19190 11111 22222 33333 44444
19193 11111 33333 33333 55555
19194 22222 22222 44444 44444
19195 11111 33333 33333 55555
19196 22222 22222 44444 44444
19197 11111 33333 33333 55555
19198 22222 22222 44444 44444
19199 11111 33333 33333 55555
19200 22222 22222 44444 44444
19205 Numeric values are deprecated but are accepted for backward
19206 compatibility reasons.
19208 Default mode is @code{merge}.
19211 Specify flags influencing the filter process.
19213 Available value for @var{flags} is:
19216 @item low_pass_filter, vlpf
19217 Enable linear vertical low-pass filtering in the filter.
19218 Vertical low-pass filtering is required when creating an interlaced
19219 destination from a progressive source which contains high-frequency
19220 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19223 @item complex_filter, cvlpf
19224 Enable complex vertical low-pass filtering.
19225 This will slightly less reduce interlace 'twitter' and Moire
19226 patterning but better retain detail and subjective sharpness impression.
19229 Bypass already interlaced frames, only adjust the frame rate.
19232 Vertical low-pass filtering and bypassing already interlaced frames can only be
19233 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19238 Pick median pixels from several successive input video frames.
19240 The filter accepts the following options:
19244 Set radius of median filter.
19245 Default is 1. Allowed range is from 1 to 127.
19248 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19251 Set median percentile. Default value is @code{0.5}.
19252 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19253 minimum values, and @code{1} maximum values.
19258 Mix successive video frames.
19260 A description of the accepted options follows.
19264 The number of successive frames to mix. If unspecified, it defaults to 3.
19267 Specify weight of each input video frame.
19268 Each weight is separated by space. If number of weights is smaller than
19269 number of @var{frames} last specified weight will be used for all remaining
19273 Specify scale, if it is set it will be multiplied with sum
19274 of each weight multiplied with pixel values to give final destination
19275 pixel value. By default @var{scale} is auto scaled to sum of weights.
19278 @subsection Examples
19282 Average 7 successive frames:
19284 tmix=frames=7:weights="1 1 1 1 1 1 1"
19288 Apply simple temporal convolution:
19290 tmix=frames=3:weights="-1 3 -1"
19294 Similar as above but only showing temporal differences:
19296 tmix=frames=3:weights="-1 2 -1":scale=1
19302 Tone map colors from different dynamic ranges.
19304 This filter expects data in single precision floating point, as it needs to
19305 operate on (and can output) out-of-range values. Another filter, such as
19306 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19308 The tonemapping algorithms implemented only work on linear light, so input
19309 data should be linearized beforehand (and possibly correctly tagged).
19312 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19315 @subsection Options
19316 The filter accepts the following options.
19320 Set the tone map algorithm to use.
19322 Possible values are:
19325 Do not apply any tone map, only desaturate overbright pixels.
19328 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19329 in-range values, while distorting out-of-range values.
19332 Stretch the entire reference gamut to a linear multiple of the display.
19335 Fit a logarithmic transfer between the tone curves.
19338 Preserve overall image brightness with a simple curve, using nonlinear
19339 contrast, which results in flattening details and degrading color accuracy.
19342 Preserve both dark and bright details better than @var{reinhard}, at the cost
19343 of slightly darkening everything. Use it when detail preservation is more
19344 important than color and brightness accuracy.
19347 Smoothly map out-of-range values, while retaining contrast and colors for
19348 in-range material as much as possible. Use it when color accuracy is more
19349 important than detail preservation.
19355 Tune the tone mapping algorithm.
19357 This affects the following algorithms:
19363 Specifies the scale factor to use while stretching.
19367 Specifies the exponent of the function.
19371 Specify an extra linear coefficient to multiply into the signal before clipping.
19375 Specify the local contrast coefficient at the display peak.
19376 Default to 0.5, which means that in-gamut values will be about half as bright
19383 Specify the transition point from linear to mobius transform. Every value
19384 below this point is guaranteed to be mapped 1:1. The higher the value, the
19385 more accurate the result will be, at the cost of losing bright details.
19386 Default to 0.3, which due to the steep initial slope still preserves in-range
19387 colors fairly accurately.
19391 Apply desaturation for highlights that exceed this level of brightness. The
19392 higher the parameter, the more color information will be preserved. This
19393 setting helps prevent unnaturally blown-out colors for super-highlights, by
19394 (smoothly) turning into white instead. This makes images feel more natural,
19395 at the cost of reducing information about out-of-range colors.
19397 The default of 2.0 is somewhat conservative and will mostly just apply to
19398 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19400 This option works only if the input frame has a supported color tag.
19403 Override signal/nominal/reference peak with this value. Useful when the
19404 embedded peak information in display metadata is not reliable or when tone
19405 mapping from a lower range to a higher range.
19410 Temporarily pad video frames.
19412 The filter accepts the following options:
19416 Specify number of delay frames before input video stream. Default is 0.
19419 Specify number of padding frames after input video stream.
19420 Set to -1 to pad indefinitely. Default is 0.
19423 Set kind of frames added to beginning of stream.
19424 Can be either @var{add} or @var{clone}.
19425 With @var{add} frames of solid-color are added.
19426 With @var{clone} frames are clones of first frame.
19427 Default is @var{add}.
19430 Set kind of frames added to end of stream.
19431 Can be either @var{add} or @var{clone}.
19432 With @var{add} frames of solid-color are added.
19433 With @var{clone} frames are clones of last frame.
19434 Default is @var{add}.
19436 @item start_duration, stop_duration
19437 Specify the duration of the start/stop delay. See
19438 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19439 for the accepted syntax.
19440 These options override @var{start} and @var{stop}. Default is 0.
19443 Specify the color of the padded area. For the syntax of this option,
19444 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19445 manual,ffmpeg-utils}.
19447 The default value of @var{color} is "black".
19453 Transpose rows with columns in the input video and optionally flip it.
19455 It accepts the following parameters:
19460 Specify the transposition direction.
19462 Can assume the following values:
19464 @item 0, 4, cclock_flip
19465 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19473 Rotate by 90 degrees clockwise, that is:
19481 Rotate by 90 degrees counterclockwise, that is:
19488 @item 3, 7, clock_flip
19489 Rotate by 90 degrees clockwise and vertically flip, that is:
19497 For values between 4-7, the transposition is only done if the input
19498 video geometry is portrait and not landscape. These values are
19499 deprecated, the @code{passthrough} option should be used instead.
19501 Numerical values are deprecated, and should be dropped in favor of
19502 symbolic constants.
19505 Do not apply the transposition if the input geometry matches the one
19506 specified by the specified value. It accepts the following values:
19509 Always apply transposition.
19511 Preserve portrait geometry (when @var{height} >= @var{width}).
19513 Preserve landscape geometry (when @var{width} >= @var{height}).
19516 Default value is @code{none}.
19519 For example to rotate by 90 degrees clockwise and preserve portrait
19522 transpose=dir=1:passthrough=portrait
19525 The command above can also be specified as:
19527 transpose=1:portrait
19530 @section transpose_npp
19532 Transpose rows with columns in the input video and optionally flip it.
19533 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19535 It accepts the following parameters:
19540 Specify the transposition direction.
19542 Can assume the following values:
19545 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19548 Rotate by 90 degrees clockwise.
19551 Rotate by 90 degrees counterclockwise.
19554 Rotate by 90 degrees clockwise and vertically flip.
19558 Do not apply the transposition if the input geometry matches the one
19559 specified by the specified value. It accepts the following values:
19562 Always apply transposition. (default)
19564 Preserve portrait geometry (when @var{height} >= @var{width}).
19566 Preserve landscape geometry (when @var{width} >= @var{height}).
19572 Trim the input so that the output contains one continuous subpart of the input.
19574 It accepts the following parameters:
19577 Specify the time of the start of the kept section, i.e. the frame with the
19578 timestamp @var{start} will be the first frame in the output.
19581 Specify the time of the first frame that will be dropped, i.e. the frame
19582 immediately preceding the one with the timestamp @var{end} will be the last
19583 frame in the output.
19586 This is the same as @var{start}, except this option sets the start timestamp
19587 in timebase units instead of seconds.
19590 This is the same as @var{end}, except this option sets the end timestamp
19591 in timebase units instead of seconds.
19594 The maximum duration of the output in seconds.
19597 The number of the first frame that should be passed to the output.
19600 The number of the first frame that should be dropped.
19603 @option{start}, @option{end}, and @option{duration} are expressed as time
19604 duration specifications; see
19605 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19606 for the accepted syntax.
19608 Note that the first two sets of the start/end options and the @option{duration}
19609 option look at the frame timestamp, while the _frame variants simply count the
19610 frames that pass through the filter. Also note that this filter does not modify
19611 the timestamps. If you wish for the output timestamps to start at zero, insert a
19612 setpts filter after the trim filter.
19614 If multiple start or end options are set, this filter tries to be greedy and
19615 keep all the frames that match at least one of the specified constraints. To keep
19616 only the part that matches all the constraints at once, chain multiple trim
19619 The defaults are such that all the input is kept. So it is possible to set e.g.
19620 just the end values to keep everything before the specified time.
19625 Drop everything except the second minute of input:
19627 ffmpeg -i INPUT -vf trim=60:120
19631 Keep only the first second:
19633 ffmpeg -i INPUT -vf trim=duration=1
19638 @section unpremultiply
19639 Apply alpha unpremultiply effect to input video stream using first plane
19640 of second stream as alpha.
19642 Both streams must have same dimensions and same pixel format.
19644 The filter accepts the following option:
19648 Set which planes will be processed, unprocessed planes will be copied.
19649 By default value 0xf, all planes will be processed.
19651 If the format has 1 or 2 components, then luma is bit 0.
19652 If the format has 3 or 4 components:
19653 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19654 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19655 If present, the alpha channel is always the last bit.
19658 Do not require 2nd input for processing, instead use alpha plane from input stream.
19664 Sharpen or blur the input video.
19666 It accepts the following parameters:
19669 @item luma_msize_x, lx
19670 Set the luma matrix horizontal size. It must be an odd integer between
19671 3 and 23. The default value is 5.
19673 @item luma_msize_y, ly
19674 Set the luma matrix vertical size. It must be an odd integer between 3
19675 and 23. The default value is 5.
19677 @item luma_amount, la
19678 Set the luma effect strength. It must be a floating point number, reasonable
19679 values lay between -1.5 and 1.5.
19681 Negative values will blur the input video, while positive values will
19682 sharpen it, a value of zero will disable the effect.
19684 Default value is 1.0.
19686 @item chroma_msize_x, cx
19687 Set the chroma matrix horizontal size. It must be an odd integer
19688 between 3 and 23. The default value is 5.
19690 @item chroma_msize_y, cy
19691 Set the chroma matrix vertical size. It must be an odd integer
19692 between 3 and 23. The default value is 5.
19694 @item chroma_amount, ca
19695 Set the chroma effect strength. It must be a floating point number, reasonable
19696 values lay between -1.5 and 1.5.
19698 Negative values will blur the input video, while positive values will
19699 sharpen it, a value of zero will disable the effect.
19701 Default value is 0.0.
19705 All parameters are optional and default to the equivalent of the
19706 string '5:5:1.0:5:5:0.0'.
19708 @subsection Examples
19712 Apply strong luma sharpen effect:
19714 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19718 Apply a strong blur of both luma and chroma parameters:
19720 unsharp=7:7:-2:7:7:-2
19727 Decompose a video made of tiled images into the individual images.
19729 The frame rate of the output video is the frame rate of the input video
19730 multiplied by the number of tiles.
19732 This filter does the reverse of @ref{tile}.
19734 The filter accepts the following options:
19739 Set the grid size (i.e. the number of lines and columns). For the syntax of
19740 this option, check the
19741 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19744 @subsection Examples
19748 Produce a 1-second video from a still image file made of 25 frames stacked
19749 vertically, like an analogic film reel:
19751 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19757 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19758 the image at several (or - in the case of @option{quality} level @code{8} - all)
19759 shifts and average the results.
19761 The way this differs from the behavior of spp is that uspp actually encodes &
19762 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19763 DCT similar to MJPEG.
19765 The filter accepts the following options:
19769 Set quality. This option defines the number of levels for averaging. It accepts
19770 an integer in the range 0-8. If set to @code{0}, the filter will have no
19771 effect. A value of @code{8} means the higher quality. For each increment of
19772 that value the speed drops by a factor of approximately 2. Default value is
19776 Force a constant quantization parameter. If not set, the filter will use the QP
19777 from the video stream (if available).
19782 Convert 360 videos between various formats.
19784 The filter accepts the following options:
19790 Set format of the input/output video.
19798 Equirectangular projection.
19803 Cubemap with 3x2/6x1/1x6 layout.
19805 Format specific options:
19810 Set padding proportion for the input/output cubemap. Values in decimals.
19817 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)
19820 Default value is @b{@samp{0}}.
19821 Maximum value is @b{@samp{0.1}}.
19825 Set fixed padding for the input/output cubemap. Values in pixels.
19827 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19831 Set order of faces for the input/output cubemap. Choose one direction for each position.
19833 Designation of directions:
19849 Default value is @b{@samp{rludfb}}.
19853 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
19855 Designation of angles:
19858 0 degrees clockwise
19860 90 degrees clockwise
19862 180 degrees clockwise
19864 270 degrees clockwise
19867 Default value is @b{@samp{000000}}.
19871 Equi-Angular Cubemap.
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 Format specific options:
19903 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19905 If diagonal field of view is set it overrides horizontal and vertical field of view.
19910 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19912 If diagonal field of view is set it overrides horizontal and vertical field of view.
19918 Facebook's 360 formats.
19921 Stereographic format.
19923 Format specific options:
19928 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19930 If diagonal field of view is set it overrides horizontal and vertical field of view.
19935 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19937 If diagonal field of view is set it overrides horizontal and vertical field of view.
19944 Ball format, gives significant distortion toward the back.
19947 Hammer-Aitoff map projection format.
19950 Sinusoidal map projection format.
19953 Fisheye projection.
19955 Format specific options:
19960 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19962 If diagonal field of view is set it overrides horizontal and vertical field of view.
19967 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19969 If diagonal field of view is set it overrides horizontal and vertical field of view.
19973 Pannini projection.
19975 Format specific options:
19978 Set output pannini parameter.
19981 Set input pannini parameter.
19985 Cylindrical projection.
19987 Format specific options:
19992 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19994 If diagonal field of view is set it overrides horizontal and vertical field of view.
19999 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20001 If diagonal field of view is set it overrides horizontal and vertical field of view.
20005 Perspective projection. @i{(output only)}
20007 Format specific options:
20010 Set perspective parameter.
20014 Tetrahedron projection.
20017 Truncated square pyramid projection.
20021 Half equirectangular projection.
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 Orthographic format.
20046 Format specific options:
20051 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20053 If diagonal field of view is set it overrides horizontal and vertical field of view.
20058 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20060 If diagonal field of view is set it overrides horizontal and vertical field of view.
20064 Octahedron projection.
20068 Set interpolation method.@*
20069 @i{Note: more complex interpolation methods require much more memory to run.}
20079 Bilinear interpolation.
20081 Lagrange9 interpolation.
20084 Bicubic interpolation.
20087 Lanczos interpolation.
20090 Spline16 interpolation.
20093 Gaussian interpolation.
20095 Mitchell interpolation.
20098 Default value is @b{@samp{line}}.
20102 Set the output video resolution.
20104 Default resolution depends on formats.
20108 Set the input/output stereo format.
20119 Default value is @b{@samp{2d}} for input and output format.
20124 Set rotation for the output video. Values in degrees.
20127 Set rotation order for the output video. Choose one item for each position.
20138 Default value is @b{@samp{ypr}}.
20143 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20147 Set if input video is flipped horizontally/vertically. Boolean values.
20150 Set if input video is transposed. Boolean value, by default disabled.
20153 Set if output video needs to be transposed. Boolean value, by default disabled.
20156 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20159 @subsection Examples
20163 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20165 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20168 Extract back view of Equi-Angular Cubemap:
20170 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20173 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20175 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20179 @subsection Commands
20181 This filter supports subset of above options as @ref{commands}.
20183 @section vaguedenoiser
20185 Apply a wavelet based denoiser.
20187 It transforms each frame from the video input into the wavelet domain,
20188 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20189 the obtained coefficients. It does an inverse wavelet transform after.
20190 Due to wavelet properties, it should give a nice smoothed result, and
20191 reduced noise, without blurring picture features.
20193 This filter accepts the following options:
20197 The filtering strength. The higher, the more filtered the video will be.
20198 Hard thresholding can use a higher threshold than soft thresholding
20199 before the video looks overfiltered. Default value is 2.
20202 The filtering method the filter will use.
20204 It accepts the following values:
20207 All values under the threshold will be zeroed.
20210 All values under the threshold will be zeroed. All values above will be
20211 reduced by the threshold.
20214 Scales or nullifies coefficients - intermediary between (more) soft and
20215 (less) hard thresholding.
20218 Default is garrote.
20221 Number of times, the wavelet will decompose the picture. Picture can't
20222 be decomposed beyond a particular point (typically, 8 for a 640x480
20223 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20226 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20229 A list of the planes to process. By default all planes are processed.
20232 The threshold type the filter will use.
20234 It accepts the following values:
20237 Threshold used is same for all decompositions.
20240 Threshold used depends also on each decomposition coefficients.
20243 Default is universal.
20246 @section vectorscope
20248 Display 2 color component values in the two dimensional graph (which is called
20251 This filter accepts the following options:
20255 Set vectorscope mode.
20257 It accepts the following values:
20261 Gray values are displayed on graph, higher brightness means more pixels have
20262 same component color value on location in graph. This is the default mode.
20265 Gray values are displayed on graph. Surrounding pixels values which are not
20266 present in video frame are drawn in gradient of 2 color components which are
20267 set by option @code{x} and @code{y}. The 3rd color component is static.
20270 Actual color components values present in video frame are displayed on graph.
20273 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20274 on graph increases value of another color component, which is luminance by
20275 default values of @code{x} and @code{y}.
20278 Actual colors present in video frame are displayed on graph. If two different
20279 colors map to same position on graph then color with higher value of component
20280 not present in graph is picked.
20283 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20284 component picked from radial gradient.
20288 Set which color component will be represented on X-axis. Default is @code{1}.
20291 Set which color component will be represented on Y-axis. Default is @code{2}.
20294 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20295 of color component which represents frequency of (X, Y) location in graph.
20300 No envelope, this is default.
20303 Instant envelope, even darkest single pixel will be clearly highlighted.
20306 Hold maximum and minimum values presented in graph over time. This way you
20307 can still spot out of range values without constantly looking at vectorscope.
20310 Peak and instant envelope combined together.
20314 Set what kind of graticule to draw.
20323 Set graticule opacity.
20326 Set graticule flags.
20330 Draw graticule for white point.
20333 Draw graticule for black point.
20336 Draw color points short names.
20340 Set background opacity.
20342 @item lthreshold, l
20343 Set low threshold for color component not represented on X or Y axis.
20344 Values lower than this value will be ignored. Default is 0.
20345 Note this value is multiplied with actual max possible value one pixel component
20346 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20349 @item hthreshold, h
20350 Set high threshold for color component not represented on X or Y axis.
20351 Values higher than this value will be ignored. Default is 1.
20352 Note this value is multiplied with actual max possible value one pixel component
20353 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20354 is 0.9 * 255 = 230.
20356 @item colorspace, c
20357 Set what kind of colorspace to use when drawing graticule.
20367 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20368 This means no tint, and output will remain gray.
20371 @anchor{vidstabdetect}
20372 @section vidstabdetect
20374 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20375 @ref{vidstabtransform} for pass 2.
20377 This filter generates a file with relative translation and rotation
20378 transform information about subsequent frames, which is then used by
20379 the @ref{vidstabtransform} filter.
20381 To enable compilation of this filter you need to configure FFmpeg with
20382 @code{--enable-libvidstab}.
20384 This filter accepts the following options:
20388 Set the path to the file used to write the transforms information.
20389 Default value is @file{transforms.trf}.
20392 Set how shaky the video is and how quick the camera is. It accepts an
20393 integer in the range 1-10, a value of 1 means little shakiness, a
20394 value of 10 means strong shakiness. Default value is 5.
20397 Set the accuracy of the detection process. It must be a value in the
20398 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20399 accuracy. Default value is 15.
20402 Set stepsize of the search process. The region around minimum is
20403 scanned with 1 pixel resolution. Default value is 6.
20406 Set minimum contrast. Below this value a local measurement field is
20407 discarded. Must be a floating point value in the range 0-1. Default
20411 Set reference frame number for tripod mode.
20413 If enabled, the motion of the frames is compared to a reference frame
20414 in the filtered stream, identified by the specified number. The idea
20415 is to compensate all movements in a more-or-less static scene and keep
20416 the camera view absolutely still.
20418 If set to 0, it is disabled. The frames are counted starting from 1.
20421 Show fields and transforms in the resulting frames. It accepts an
20422 integer in the range 0-2. Default value is 0, which disables any
20426 @subsection Examples
20430 Use default values:
20436 Analyze strongly shaky movie and put the results in file
20437 @file{mytransforms.trf}:
20439 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20443 Visualize the result of internal transformations in the resulting
20446 vidstabdetect=show=1
20450 Analyze a video with medium shakiness using @command{ffmpeg}:
20452 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20456 @anchor{vidstabtransform}
20457 @section vidstabtransform
20459 Video stabilization/deshaking: pass 2 of 2,
20460 see @ref{vidstabdetect} for pass 1.
20462 Read a file with transform information for each frame and
20463 apply/compensate them. Together with the @ref{vidstabdetect}
20464 filter this can be used to deshake videos. See also
20465 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20466 the @ref{unsharp} filter, see below.
20468 To enable compilation of this filter you need to configure FFmpeg with
20469 @code{--enable-libvidstab}.
20471 @subsection Options
20475 Set path to the file used to read the transforms. Default value is
20476 @file{transforms.trf}.
20479 Set the number of frames (value*2 + 1) used for lowpass filtering the
20480 camera movements. Default value is 10.
20482 For example a number of 10 means that 21 frames are used (10 in the
20483 past and 10 in the future) to smoothen the motion in the video. A
20484 larger value leads to a smoother video, but limits the acceleration of
20485 the camera (pan/tilt movements). 0 is a special case where a static
20486 camera is simulated.
20489 Set the camera path optimization algorithm.
20491 Accepted values are:
20494 gaussian kernel low-pass filter on camera motion (default)
20496 averaging on transformations
20500 Set maximal number of pixels to translate frames. Default value is -1,
20504 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20505 value is -1, meaning no limit.
20508 Specify how to deal with borders that may be visible due to movement
20511 Available values are:
20514 keep image information from previous frame (default)
20516 fill the border black
20520 Invert transforms if set to 1. Default value is 0.
20523 Consider transforms as relative to previous frame if set to 1,
20524 absolute if set to 0. Default value is 0.
20527 Set percentage to zoom. A positive value will result in a zoom-in
20528 effect, a negative value in a zoom-out effect. Default value is 0 (no
20532 Set optimal zooming to avoid borders.
20534 Accepted values are:
20539 optimal static zoom value is determined (only very strong movements
20540 will lead to visible borders) (default)
20542 optimal adaptive zoom value is determined (no borders will be
20543 visible), see @option{zoomspeed}
20546 Note that the value given at zoom is added to the one calculated here.
20549 Set percent to zoom maximally each frame (enabled when
20550 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20554 Specify type of interpolation.
20556 Available values are:
20561 linear only horizontal
20563 linear in both directions (default)
20565 cubic in both directions (slow)
20569 Enable virtual tripod mode if set to 1, which is equivalent to
20570 @code{relative=0:smoothing=0}. Default value is 0.
20572 Use also @code{tripod} option of @ref{vidstabdetect}.
20575 Increase log verbosity if set to 1. Also the detected global motions
20576 are written to the temporary file @file{global_motions.trf}. Default
20580 @subsection Examples
20584 Use @command{ffmpeg} for a typical stabilization with default values:
20586 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20589 Note the use of the @ref{unsharp} filter which is always recommended.
20592 Zoom in a bit more and load transform data from a given file:
20594 vidstabtransform=zoom=5:input="mytransforms.trf"
20598 Smoothen the video even more:
20600 vidstabtransform=smoothing=30
20606 Flip the input video vertically.
20608 For example, to vertically flip a video with @command{ffmpeg}:
20610 ffmpeg -i in.avi -vf "vflip" out.avi
20615 Detect variable frame rate video.
20617 This filter tries to detect if the input is variable or constant frame rate.
20619 At end it will output number of frames detected as having variable delta pts,
20620 and ones with constant delta pts.
20621 If there was frames with variable delta, than it will also show min, max and
20622 average delta encountered.
20626 Boost or alter saturation.
20628 The filter accepts the following options:
20631 Set strength of boost if positive value or strength of alter if negative value.
20632 Default is 0. Allowed range is from -2 to 2.
20635 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20638 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20641 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20644 Set the red luma coefficient.
20647 Set the green luma coefficient.
20650 Set the blue luma coefficient.
20653 If @code{intensity} is negative and this is set to 1, colors will change,
20654 otherwise colors will be less saturated, more towards gray.
20657 @subsection Commands
20659 This filter supports the all above options as @ref{commands}.
20664 Make or reverse a natural vignetting effect.
20666 The filter accepts the following options:
20670 Set lens angle expression as a number of radians.
20672 The value is clipped in the @code{[0,PI/2]} range.
20674 Default value: @code{"PI/5"}
20678 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20682 Set forward/backward mode.
20684 Available modes are:
20687 The larger the distance from the central point, the darker the image becomes.
20690 The larger the distance from the central point, the brighter the image becomes.
20691 This can be used to reverse a vignette effect, though there is no automatic
20692 detection to extract the lens @option{angle} and other settings (yet). It can
20693 also be used to create a burning effect.
20696 Default value is @samp{forward}.
20699 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20701 It accepts the following values:
20704 Evaluate expressions only once during the filter initialization.
20707 Evaluate expressions for each incoming frame. This is way slower than the
20708 @samp{init} mode since it requires all the scalers to be re-computed, but it
20709 allows advanced dynamic expressions.
20712 Default value is @samp{init}.
20715 Set dithering to reduce the circular banding effects. Default is @code{1}
20719 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20720 Setting this value to the SAR of the input will make a rectangular vignetting
20721 following the dimensions of the video.
20723 Default is @code{1/1}.
20726 @subsection Expressions
20728 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20729 following parameters.
20734 input width and height
20737 the number of input frame, starting from 0
20740 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20741 @var{TB} units, NAN if undefined
20744 frame rate of the input video, NAN if the input frame rate is unknown
20747 the PTS (Presentation TimeStamp) of the filtered video frame,
20748 expressed in seconds, NAN if undefined
20751 time base of the input video
20755 @subsection Examples
20759 Apply simple strong vignetting effect:
20765 Make a flickering vignetting:
20767 vignette='PI/4+random(1)*PI/50':eval=frame
20772 @section vmafmotion
20774 Obtain the average VMAF motion score of a video.
20775 It is one of the component metrics of VMAF.
20777 The obtained average motion score is printed through the logging system.
20779 The filter accepts the following options:
20783 If specified, the filter will use the named file to save the motion score of
20784 each frame with respect to the previous frame.
20785 When filename equals "-" the data is sent to standard output.
20790 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20794 Stack input videos vertically.
20796 All streams must be of same pixel format and of same width.
20798 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20799 to create same output.
20801 The filter accepts the following options:
20805 Set number of input streams. Default is 2.
20808 If set to 1, force the output to terminate when the shortest input
20809 terminates. Default value is 0.
20814 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20815 Deinterlacing Filter").
20817 Based on the process described by Martin Weston for BBC R&D, and
20818 implemented based on the de-interlace algorithm written by Jim
20819 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20820 uses filter coefficients calculated by BBC R&D.
20822 This filter uses field-dominance information in frame to decide which
20823 of each pair of fields to place first in the output.
20824 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20826 There are two sets of filter coefficients, so called "simple"
20827 and "complex". Which set of filter coefficients is used can
20828 be set by passing an optional parameter:
20832 Set the interlacing filter coefficients. Accepts one of the following values:
20836 Simple filter coefficient set.
20838 More-complex filter coefficient set.
20840 Default value is @samp{complex}.
20843 Specify which frames to deinterlace. Accepts one of the following values:
20847 Deinterlace all frames,
20849 Only deinterlace frames marked as interlaced.
20852 Default value is @samp{all}.
20856 Video waveform monitor.
20858 The waveform monitor plots color component intensity. By default luminance
20859 only. Each column of the waveform corresponds to a column of pixels in the
20862 It accepts the following options:
20866 Can be either @code{row}, or @code{column}. Default is @code{column}.
20867 In row mode, the graph on the left side represents color component value 0 and
20868 the right side represents value = 255. In column mode, the top side represents
20869 color component value = 0 and bottom side represents value = 255.
20872 Set intensity. Smaller values are useful to find out how many values of the same
20873 luminance are distributed across input rows/columns.
20874 Default value is @code{0.04}. Allowed range is [0, 1].
20877 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
20878 In mirrored mode, higher values will be represented on the left
20879 side for @code{row} mode and at the top for @code{column} mode. Default is
20880 @code{1} (mirrored).
20884 It accepts the following values:
20887 Presents information identical to that in the @code{parade}, except
20888 that the graphs representing color components are superimposed directly
20891 This display mode makes it easier to spot relative differences or similarities
20892 in overlapping areas of the color components that are supposed to be identical,
20893 such as neutral whites, grays, or blacks.
20896 Display separate graph for the color components side by side in
20897 @code{row} mode or one below the other in @code{column} mode.
20900 Display separate graph for the color components side by side in
20901 @code{column} mode or one below the other in @code{row} mode.
20903 Using this display mode makes it easy to spot color casts in the highlights
20904 and shadows of an image, by comparing the contours of the top and the bottom
20905 graphs of each waveform. Since whites, grays, and blacks are characterized
20906 by exactly equal amounts of red, green, and blue, neutral areas of the picture
20907 should display three waveforms of roughly equal width/height. If not, the
20908 correction is easy to perform by making level adjustments the three waveforms.
20910 Default is @code{stack}.
20912 @item components, c
20913 Set which color components to display. Default is 1, which means only luminance
20914 or red color component if input is in RGB colorspace. If is set for example to
20915 7 it will display all 3 (if) available color components.
20920 No envelope, this is default.
20923 Instant envelope, minimum and maximum values presented in graph will be easily
20924 visible even with small @code{step} value.
20927 Hold minimum and maximum values presented in graph across time. This way you
20928 can still spot out of range values without constantly looking at waveforms.
20931 Peak and instant envelope combined together.
20937 No filtering, this is default.
20940 Luma and chroma combined together.
20943 Similar as above, but shows difference between blue and red chroma.
20946 Similar as above, but use different colors.
20949 Similar as above, but again with different colors.
20952 Displays only chroma.
20955 Displays actual color value on waveform.
20958 Similar as above, but with luma showing frequency of chroma values.
20962 Set which graticule to display.
20966 Do not display graticule.
20969 Display green graticule showing legal broadcast ranges.
20972 Display orange graticule showing legal broadcast ranges.
20975 Display invert graticule showing legal broadcast ranges.
20979 Set graticule opacity.
20982 Set graticule flags.
20986 Draw numbers above lines. By default enabled.
20989 Draw dots instead of lines.
20993 Set scale used for displaying graticule.
21000 Default is digital.
21003 Set background opacity.
21007 Set tint for output.
21008 Only used with lowpass filter and when display is not overlay and input
21009 pixel formats are not RGB.
21012 @section weave, doubleweave
21014 The @code{weave} takes a field-based video input and join
21015 each two sequential fields into single frame, producing a new double
21016 height clip with half the frame rate and half the frame count.
21018 The @code{doubleweave} works same as @code{weave} but without
21019 halving frame rate and frame count.
21021 It accepts the following option:
21025 Set first field. Available values are:
21029 Set the frame as top-field-first.
21032 Set the frame as bottom-field-first.
21036 @subsection Examples
21040 Interlace video using @ref{select} and @ref{separatefields} filter:
21042 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21047 Apply the xBR high-quality magnification filter which is designed for pixel
21048 art. It follows a set of edge-detection rules, see
21049 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21051 It accepts the following option:
21055 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21056 @code{3xBR} and @code{4} for @code{4xBR}.
21057 Default is @code{3}.
21062 Apply cross fade from one input video stream to another input video stream.
21063 The cross fade is applied for specified duration.
21065 The filter accepts the following options:
21069 Set one of available transition effects:
21117 Default transition effect is fade.
21120 Set cross fade duration in seconds.
21121 Default duration is 1 second.
21124 Set cross fade start relative to first input stream in seconds.
21125 Default offset is 0.
21128 Set expression for custom transition effect.
21130 The expressions can use the following variables and functions:
21135 The coordinates of the current sample.
21139 The width and height of the image.
21142 Progress of transition effect.
21145 Currently processed plane.
21148 Return value of first input at current location and plane.
21151 Return value of second input at current location and plane.
21157 Return the value of the pixel at location (@var{x},@var{y}) of the
21158 first/second/third/fourth component of first input.
21164 Return the value of the pixel at location (@var{x},@var{y}) of the
21165 first/second/third/fourth component of second input.
21169 @subsection Examples
21173 Cross fade from one input video to another input video, with fade transition and duration of transition
21174 of 2 seconds starting at offset of 5 seconds:
21176 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21181 Pick median pixels from several input videos.
21183 The filter accepts the following options:
21187 Set number of inputs.
21188 Default is 3. Allowed range is from 3 to 255.
21189 If number of inputs is even number, than result will be mean value between two median values.
21192 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21195 Set median percentile. Default value is @code{0.5}.
21196 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21197 minimum values, and @code{1} maximum values.
21201 Stack video inputs into custom layout.
21203 All streams must be of same pixel format.
21205 The filter accepts the following options:
21209 Set number of input streams. Default is 2.
21212 Specify layout of inputs.
21213 This option requires the desired layout configuration to be explicitly set by the user.
21214 This sets position of each video input in output. Each input
21215 is separated by '|'.
21216 The first number represents the column, and the second number represents the row.
21217 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21218 where X is video input from which to take width or height.
21219 Multiple values can be used when separated by '+'. In such
21220 case values are summed together.
21222 Note that if inputs are of different sizes gaps may appear, as not all of
21223 the output video frame will be filled. Similarly, videos can overlap each
21224 other if their position doesn't leave enough space for the full frame of
21227 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21228 a layout must be set by the user.
21231 If set to 1, force the output to terminate when the shortest input
21232 terminates. Default value is 0.
21235 If set to valid color, all unused pixels will be filled with that color.
21236 By default fill is set to none, so it is disabled.
21239 @subsection Examples
21243 Display 4 inputs into 2x2 grid.
21247 input1(0, 0) | input3(w0, 0)
21248 input2(0, h0) | input4(w0, h0)
21252 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21255 Note that if inputs are of different sizes, gaps or overlaps may occur.
21258 Display 4 inputs into 1x4 grid.
21265 input4(0, h0+h1+h2)
21269 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21272 Note that if inputs are of different widths, unused space will appear.
21275 Display 9 inputs into 3x3 grid.
21279 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21280 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21281 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21285 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
21288 Note that if inputs are of different sizes, gaps or overlaps may occur.
21291 Display 16 inputs into 4x4 grid.
21295 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21296 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21297 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21298 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21302 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|
21303 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
21306 Note that if inputs are of different sizes, gaps or overlaps may occur.
21313 Deinterlace the input video ("yadif" means "yet another deinterlacing
21316 It accepts the following parameters:
21322 The interlacing mode to adopt. It accepts one of the following values:
21325 @item 0, send_frame
21326 Output one frame for each frame.
21327 @item 1, send_field
21328 Output one frame for each field.
21329 @item 2, send_frame_nospatial
21330 Like @code{send_frame}, but it skips the spatial interlacing check.
21331 @item 3, send_field_nospatial
21332 Like @code{send_field}, but it skips the spatial interlacing check.
21335 The default value is @code{send_frame}.
21338 The picture field parity assumed for the input interlaced video. It accepts one
21339 of the following values:
21343 Assume the top field is first.
21345 Assume the bottom field is first.
21347 Enable automatic detection of field parity.
21350 The default value is @code{auto}.
21351 If the interlacing is unknown or the decoder does not export this information,
21352 top field first will be assumed.
21355 Specify which frames to deinterlace. Accepts one of the following
21360 Deinterlace all frames.
21361 @item 1, interlaced
21362 Only deinterlace frames marked as interlaced.
21365 The default value is @code{all}.
21368 @section yadif_cuda
21370 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21371 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21374 It accepts the following parameters:
21380 The interlacing mode to adopt. It accepts one of the following values:
21383 @item 0, send_frame
21384 Output one frame for each frame.
21385 @item 1, send_field
21386 Output one frame for each field.
21387 @item 2, send_frame_nospatial
21388 Like @code{send_frame}, but it skips the spatial interlacing check.
21389 @item 3, send_field_nospatial
21390 Like @code{send_field}, but it skips the spatial interlacing check.
21393 The default value is @code{send_frame}.
21396 The picture field parity assumed for the input interlaced video. It accepts one
21397 of the following values:
21401 Assume the top field is first.
21403 Assume the bottom field is first.
21405 Enable automatic detection of field parity.
21408 The default value is @code{auto}.
21409 If the interlacing is unknown or the decoder does not export this information,
21410 top field first will be assumed.
21413 Specify which frames to deinterlace. Accepts one of the following
21418 Deinterlace all frames.
21419 @item 1, interlaced
21420 Only deinterlace frames marked as interlaced.
21423 The default value is @code{all}.
21428 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21429 The algorithm is described in
21430 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21432 It accepts the following parameters:
21436 Set the window radius. Default value is 3.
21439 Set which planes to filter. Default is only the first plane.
21442 Set blur strength. Default value is 128.
21445 @subsection Commands
21446 This filter supports same @ref{commands} as options.
21450 Apply Zoom & Pan effect.
21452 This filter accepts the following options:
21456 Set the zoom expression. Range is 1-10. Default is 1.
21460 Set the x and y expression. Default is 0.
21463 Set the duration expression in number of frames.
21464 This sets for how many number of frames effect will last for
21465 single input image.
21468 Set the output image size, default is 'hd720'.
21471 Set the output frame rate, default is '25'.
21474 Each expression can contain the following constants:
21493 Output frame count.
21496 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21498 @item out_time, time, ot
21499 The output timestamp expressed in seconds.
21503 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21504 for current input frame.
21508 'x' and 'y' of last output frame of previous input frame or 0 when there was
21509 not yet such frame (first input frame).
21512 Last calculated zoom from 'z' expression for current input frame.
21515 Last calculated zoom of last output frame of previous input frame.
21518 Number of output frames for current input frame. Calculated from 'd' expression
21519 for each input frame.
21522 number of output frames created for previous input frame
21525 Rational number: input width / input height
21528 sample aspect ratio
21531 display aspect ratio
21535 @subsection Examples
21539 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21541 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
21545 Zoom in up to 1.5x and pan always at center of picture:
21547 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21551 Same as above but without pausing:
21553 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21557 Zoom in 2x into center of picture only for the first second of the input video:
21559 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21566 Scale (resize) the input video, using the z.lib library:
21567 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21568 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21570 The zscale filter forces the output display aspect ratio to be the same
21571 as the input, by changing the output sample aspect ratio.
21573 If the input image format is different from the format requested by
21574 the next filter, the zscale filter will convert the input to the
21577 @subsection Options
21578 The filter accepts the following options.
21583 Set the output video dimension expression. Default value is the input
21586 If the @var{width} or @var{w} value is 0, the input width is used for
21587 the output. If the @var{height} or @var{h} value is 0, the input height
21588 is used for the output.
21590 If one and only one of the values is -n with n >= 1, the zscale filter
21591 will use a value that maintains the aspect ratio of the input image,
21592 calculated from the other specified dimension. After that it will,
21593 however, make sure that the calculated dimension is divisible by n and
21594 adjust the value if necessary.
21596 If both values are -n with n >= 1, the behavior will be identical to
21597 both values being set to 0 as previously detailed.
21599 See below for the list of accepted constants for use in the dimension
21603 Set the video size. For the syntax of this option, check the
21604 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21607 Set the dither type.
21609 Possible values are:
21614 @item error_diffusion
21620 Set the resize filter type.
21622 Possible values are:
21632 Default is bilinear.
21635 Set the color range.
21637 Possible values are:
21644 Default is same as input.
21647 Set the color primaries.
21649 Possible values are:
21659 Default is same as input.
21662 Set the transfer characteristics.
21664 Possible values are:
21678 Default is same as input.
21681 Set the colorspace matrix.
21683 Possible value are:
21694 Default is same as input.
21697 Set the input color range.
21699 Possible values are:
21706 Default is same as input.
21708 @item primariesin, pin
21709 Set the input color primaries.
21711 Possible values are:
21721 Default is same as input.
21723 @item transferin, tin
21724 Set the input transfer characteristics.
21726 Possible values are:
21737 Default is same as input.
21739 @item matrixin, min
21740 Set the input colorspace matrix.
21742 Possible value are:
21754 Set the output chroma location.
21756 Possible values are:
21767 @item chromalin, cin
21768 Set the input chroma location.
21770 Possible values are:
21782 Set the nominal peak luminance.
21785 The values of the @option{w} and @option{h} options are expressions
21786 containing the following constants:
21791 The input width and height
21795 These are the same as @var{in_w} and @var{in_h}.
21799 The output (scaled) width and height
21803 These are the same as @var{out_w} and @var{out_h}
21806 The same as @var{iw} / @var{ih}
21809 input sample aspect ratio
21812 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
21816 horizontal and vertical input chroma subsample values. For example for the
21817 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21821 horizontal and vertical output chroma subsample values. For example for the
21822 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21825 @subsection Commands
21827 This filter supports the following commands:
21831 Set the output video dimension expression.
21832 The command accepts the same syntax of the corresponding option.
21834 If the specified expression is not valid, it is kept at its current
21838 @c man end VIDEO FILTERS
21840 @chapter OpenCL Video Filters
21841 @c man begin OPENCL VIDEO FILTERS
21843 Below is a description of the currently available OpenCL video filters.
21845 To enable compilation of these filters you need to configure FFmpeg with
21846 @code{--enable-opencl}.
21848 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
21851 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
21852 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
21853 given device parameters.
21855 @item -filter_hw_device @var{name}
21856 Pass the hardware device called @var{name} to all filters in any filter graph.
21860 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
21864 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
21866 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
21870 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.
21872 @section avgblur_opencl
21874 Apply average blur filter.
21876 The filter accepts the following options:
21880 Set horizontal radius size.
21881 Range is @code{[1, 1024]} and default value is @code{1}.
21884 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21887 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
21890 @subsection Example
21894 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.
21896 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
21900 @section boxblur_opencl
21902 Apply a boxblur algorithm to the input video.
21904 It accepts the following parameters:
21908 @item luma_radius, lr
21909 @item luma_power, lp
21910 @item chroma_radius, cr
21911 @item chroma_power, cp
21912 @item alpha_radius, ar
21913 @item alpha_power, ap
21917 A description of the accepted options follows.
21920 @item luma_radius, lr
21921 @item chroma_radius, cr
21922 @item alpha_radius, ar
21923 Set an expression for the box radius in pixels used for blurring the
21924 corresponding input plane.
21926 The radius value must be a non-negative number, and must not be
21927 greater than the value of the expression @code{min(w,h)/2} for the
21928 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
21931 Default value for @option{luma_radius} is "2". If not specified,
21932 @option{chroma_radius} and @option{alpha_radius} default to the
21933 corresponding value set for @option{luma_radius}.
21935 The expressions can contain the following constants:
21939 The input width and height in pixels.
21943 The input chroma image width and height in pixels.
21947 The horizontal and vertical chroma subsample values. For example, for the
21948 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
21951 @item luma_power, lp
21952 @item chroma_power, cp
21953 @item alpha_power, ap
21954 Specify how many times the boxblur filter is applied to the
21955 corresponding plane.
21957 Default value for @option{luma_power} is 2. If not specified,
21958 @option{chroma_power} and @option{alpha_power} default to the
21959 corresponding value set for @option{luma_power}.
21961 A value of 0 will disable the effect.
21964 @subsection Examples
21966 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.
21970 Apply a boxblur filter with the luma, chroma, and alpha radius
21971 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.
21973 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
21974 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
21978 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.
21980 For the luma plane, a 2x2 box radius will be run once.
21982 For the chroma plane, a 4x4 box radius will be run 5 times.
21984 For the alpha plane, a 3x3 box radius will be run 7 times.
21986 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
21990 @section colorkey_opencl
21991 RGB colorspace color keying.
21993 The filter accepts the following options:
21997 The color which will be replaced with transparency.
22000 Similarity percentage with the key color.
22002 0.01 matches only the exact key color, while 1.0 matches everything.
22007 0.0 makes pixels either fully transparent, or not transparent at all.
22009 Higher values result in semi-transparent pixels, with a higher transparency
22010 the more similar the pixels color is to the key color.
22013 @subsection Examples
22017 Make every semi-green pixel in the input transparent with some slight blending:
22019 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22023 @section convolution_opencl
22025 Apply convolution of 3x3, 5x5, 7x7 matrix.
22027 The filter accepts the following options:
22034 Set matrix for each plane.
22035 Matrix is sequence of 9, 25 or 49 signed numbers.
22036 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22042 Set multiplier for calculated value for each plane.
22043 If unset or 0, it will be sum of all matrix elements.
22044 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22050 Set bias for each plane. This value is added to the result of the multiplication.
22051 Useful for making the overall image brighter or darker.
22052 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22056 @subsection Examples
22062 -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
22068 -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
22072 Apply edge enhance:
22074 -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
22080 -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
22084 Apply laplacian edge detector which includes diagonals:
22086 -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
22092 -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
22096 @section erosion_opencl
22098 Apply erosion effect to the video.
22100 This filter replaces the pixel by the local(3x3) minimum.
22102 It accepts the following options:
22109 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22110 If @code{0}, plane will remain unchanged.
22113 Flag which specifies the pixel to refer to.
22114 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22116 Flags to local 3x3 coordinates region centered on @code{x}:
22125 @subsection Example
22129 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.
22131 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22135 @section deshake_opencl
22136 Feature-point based video stabilization filter.
22138 The filter accepts the following options:
22142 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22145 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22147 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22149 Viewing point matches in the output video is only supported for RGB input.
22151 Defaults to @code{0}.
22153 @item adaptive_crop
22154 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22156 Defaults to @code{1}.
22158 @item refine_features
22159 Whether or not feature points should be refined at a sub-pixel level.
22161 This can be turned off for a slight performance gain at the cost of precision.
22163 Defaults to @code{1}.
22165 @item smooth_strength
22166 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22168 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22170 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22172 Defaults to @code{0.0}.
22174 @item smooth_window_multiplier
22175 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22177 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22179 Acceptable values range from @code{0.1} to @code{10.0}.
22181 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22182 potentially improving smoothness, but also increase latency and memory usage.
22184 Defaults to @code{2.0}.
22188 @subsection Examples
22192 Stabilize a video with a fixed, medium smoothing strength:
22194 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22198 Stabilize a video with debugging (both in console and in rendered video):
22200 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22204 @section dilation_opencl
22206 Apply dilation effect to the video.
22208 This filter replaces the pixel by the local(3x3) maximum.
22210 It accepts the following options:
22217 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22218 If @code{0}, plane will remain unchanged.
22221 Flag which specifies the pixel to refer to.
22222 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22224 Flags to local 3x3 coordinates region centered on @code{x}:
22233 @subsection Example
22237 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.
22239 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22243 @section nlmeans_opencl
22245 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22247 @section overlay_opencl
22249 Overlay one video on top of another.
22251 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22252 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22254 The filter accepts the following options:
22259 Set the x coordinate of the overlaid video on the main video.
22260 Default value is @code{0}.
22263 Set the y coordinate of the overlaid video on the main video.
22264 Default value is @code{0}.
22268 @subsection Examples
22272 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22274 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22277 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22279 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22284 @section pad_opencl
22286 Add paddings to the input image, and place the original input at the
22287 provided @var{x}, @var{y} coordinates.
22289 It accepts the following options:
22294 Specify an expression for the size of the output image with the
22295 paddings added. If the value for @var{width} or @var{height} is 0, the
22296 corresponding input size is used for the output.
22298 The @var{width} expression can reference the value set by the
22299 @var{height} expression, and vice versa.
22301 The default value of @var{width} and @var{height} is 0.
22305 Specify the offsets to place the input image at within the padded area,
22306 with respect to the top/left border of the output image.
22308 The @var{x} expression can reference the value set by the @var{y}
22309 expression, and vice versa.
22311 The default value of @var{x} and @var{y} is 0.
22313 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22314 so the input image is centered on the padded area.
22317 Specify the color of the padded area. For the syntax of this option,
22318 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22319 manual,ffmpeg-utils}.
22322 Pad to an aspect instead to a resolution.
22325 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22326 options are expressions containing the following constants:
22331 The input video width and height.
22335 These are the same as @var{in_w} and @var{in_h}.
22339 The output width and height (the size of the padded area), as
22340 specified by the @var{width} and @var{height} expressions.
22344 These are the same as @var{out_w} and @var{out_h}.
22348 The x and y offsets as specified by the @var{x} and @var{y}
22349 expressions, or NAN if not yet specified.
22352 same as @var{iw} / @var{ih}
22355 input sample aspect ratio
22358 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22361 @section prewitt_opencl
22363 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22365 The filter accepts the following option:
22369 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22372 Set value which will be multiplied with filtered result.
22373 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22376 Set value which will be added to filtered result.
22377 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22380 @subsection Example
22384 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22386 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22390 @anchor{program_opencl}
22391 @section program_opencl
22393 Filter video using an OpenCL program.
22398 OpenCL program source file.
22401 Kernel name in program.
22404 Number of inputs to the filter. Defaults to 1.
22407 Size of output frames. Defaults to the same as the first input.
22411 The @code{program_opencl} filter also supports the @ref{framesync} options.
22413 The program source file must contain a kernel function with the given name,
22414 which will be run once for each plane of the output. Each run on a plane
22415 gets enqueued as a separate 2D global NDRange with one work-item for each
22416 pixel to be generated. The global ID offset for each work-item is therefore
22417 the coordinates of a pixel in the destination image.
22419 The kernel function needs to take the following arguments:
22422 Destination image, @var{__write_only image2d_t}.
22424 This image will become the output; the kernel should write all of it.
22426 Frame index, @var{unsigned int}.
22428 This is a counter starting from zero and increasing by one for each frame.
22430 Source images, @var{__read_only image2d_t}.
22432 These are the most recent images on each input. The kernel may read from
22433 them to generate the output, but they can't be written to.
22440 Copy the input to the output (output must be the same size as the input).
22442 __kernel void copy(__write_only image2d_t destination,
22443 unsigned int index,
22444 __read_only image2d_t source)
22446 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22448 int2 location = (int2)(get_global_id(0), get_global_id(1));
22450 float4 value = read_imagef(source, sampler, location);
22452 write_imagef(destination, location, value);
22457 Apply a simple transformation, rotating the input by an amount increasing
22458 with the index counter. Pixel values are linearly interpolated by the
22459 sampler, and the output need not have the same dimensions as the input.
22461 __kernel void rotate_image(__write_only image2d_t dst,
22462 unsigned int index,
22463 __read_only image2d_t src)
22465 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22466 CLK_FILTER_LINEAR);
22468 float angle = (float)index / 100.0f;
22470 float2 dst_dim = convert_float2(get_image_dim(dst));
22471 float2 src_dim = convert_float2(get_image_dim(src));
22473 float2 dst_cen = dst_dim / 2.0f;
22474 float2 src_cen = src_dim / 2.0f;
22476 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22478 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22480 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22481 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22483 src_pos = src_pos * src_dim / dst_dim;
22485 float2 src_loc = src_pos + src_cen;
22487 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22488 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22489 write_imagef(dst, dst_loc, 0.5f);
22491 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22496 Blend two inputs together, with the amount of each input used varying
22497 with the index counter.
22499 __kernel void blend_images(__write_only image2d_t dst,
22500 unsigned int index,
22501 __read_only image2d_t src1,
22502 __read_only image2d_t src2)
22504 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22505 CLK_FILTER_LINEAR);
22507 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22509 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22510 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22511 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22513 float4 val1 = read_imagef(src1, sampler, src1_loc);
22514 float4 val2 = read_imagef(src2, sampler, src2_loc);
22516 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22522 @section roberts_opencl
22523 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22525 The filter accepts the following option:
22529 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22532 Set value which will be multiplied with filtered result.
22533 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22536 Set value which will be added to filtered result.
22537 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22540 @subsection Example
22544 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22546 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22550 @section sobel_opencl
22552 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22554 The filter accepts the following option:
22558 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22561 Set value which will be multiplied with filtered result.
22562 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22565 Set value which will be added to filtered result.
22566 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22569 @subsection Example
22573 Apply sobel operator with scale set to 2 and delta set to 10
22575 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22579 @section tonemap_opencl
22581 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22583 It accepts the following parameters:
22587 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22590 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22593 Apply desaturation for highlights that exceed this level of brightness. The
22594 higher the parameter, the more color information will be preserved. This
22595 setting helps prevent unnaturally blown-out colors for super-highlights, by
22596 (smoothly) turning into white instead. This makes images feel more natural,
22597 at the cost of reducing information about out-of-range colors.
22599 The default value is 0.5, and the algorithm here is a little different from
22600 the cpu version tonemap currently. A setting of 0.0 disables this option.
22603 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22604 is used to detect whether the scene has changed or not. If the distance between
22605 the current frame average brightness and the current running average exceeds
22606 a threshold value, we would re-calculate scene average and peak brightness.
22607 The default value is 0.2.
22610 Specify the output pixel format.
22612 Currently supported formats are:
22619 Set the output color range.
22621 Possible values are:
22627 Default is same as input.
22630 Set the output color primaries.
22632 Possible values are:
22638 Default is same as input.
22641 Set the output transfer characteristics.
22643 Possible values are:
22652 Set the output colorspace matrix.
22654 Possible value are:
22660 Default is same as input.
22664 @subsection Example
22668 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22670 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22674 @section unsharp_opencl
22676 Sharpen or blur the input video.
22678 It accepts the following parameters:
22681 @item luma_msize_x, lx
22682 Set the luma matrix horizontal size.
22683 Range is @code{[1, 23]} and default value is @code{5}.
22685 @item luma_msize_y, ly
22686 Set the luma matrix vertical size.
22687 Range is @code{[1, 23]} and default value is @code{5}.
22689 @item luma_amount, la
22690 Set the luma effect strength.
22691 Range is @code{[-10, 10]} and default value is @code{1.0}.
22693 Negative values will blur the input video, while positive values will
22694 sharpen it, a value of zero will disable the effect.
22696 @item chroma_msize_x, cx
22697 Set the chroma matrix horizontal size.
22698 Range is @code{[1, 23]} and default value is @code{5}.
22700 @item chroma_msize_y, cy
22701 Set the chroma matrix vertical size.
22702 Range is @code{[1, 23]} and default value is @code{5}.
22704 @item chroma_amount, ca
22705 Set the chroma effect strength.
22706 Range is @code{[-10, 10]} and default value is @code{0.0}.
22708 Negative values will blur the input video, while positive values will
22709 sharpen it, a value of zero will disable the effect.
22713 All parameters are optional and default to the equivalent of the
22714 string '5:5:1.0:5:5:0.0'.
22716 @subsection Examples
22720 Apply strong luma sharpen effect:
22722 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22726 Apply a strong blur of both luma and chroma parameters:
22728 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22732 @section xfade_opencl
22734 Cross fade two videos with custom transition effect by using OpenCL.
22736 It accepts the following options:
22740 Set one of possible transition effects.
22744 Select custom transition effect, the actual transition description
22745 will be picked from source and kernel options.
22757 Default transition is fade.
22761 OpenCL program source file for custom transition.
22764 Set name of kernel to use for custom transition from program source file.
22767 Set duration of video transition.
22770 Set time of start of transition relative to first video.
22773 The program source file must contain a kernel function with the given name,
22774 which will be run once for each plane of the output. Each run on a plane
22775 gets enqueued as a separate 2D global NDRange with one work-item for each
22776 pixel to be generated. The global ID offset for each work-item is therefore
22777 the coordinates of a pixel in the destination image.
22779 The kernel function needs to take the following arguments:
22782 Destination image, @var{__write_only image2d_t}.
22784 This image will become the output; the kernel should write all of it.
22787 First Source image, @var{__read_only image2d_t}.
22788 Second Source image, @var{__read_only image2d_t}.
22790 These are the most recent images on each input. The kernel may read from
22791 them to generate the output, but they can't be written to.
22794 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22801 Apply dots curtain transition effect:
22803 __kernel void blend_images(__write_only image2d_t dst,
22804 __read_only image2d_t src1,
22805 __read_only image2d_t src2,
22808 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22809 CLK_FILTER_LINEAR);
22810 int2 p = (int2)(get_global_id(0), get_global_id(1));
22811 float2 rp = (float2)(get_global_id(0), get_global_id(1));
22812 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
22815 float2 dots = (float2)(20.0, 20.0);
22816 float2 center = (float2)(0,0);
22819 float4 val1 = read_imagef(src1, sampler, p);
22820 float4 val2 = read_imagef(src2, sampler, p);
22821 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
22823 write_imagef(dst, p, next ? val1 : val2);
22829 @c man end OPENCL VIDEO FILTERS
22831 @chapter VAAPI Video Filters
22832 @c man begin VAAPI VIDEO FILTERS
22834 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
22836 To enable compilation of these filters you need to configure FFmpeg with
22837 @code{--enable-vaapi}.
22839 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}
22841 @section tonemap_vaapi
22843 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
22844 It maps the dynamic range of HDR10 content to the SDR content.
22845 It currently only accepts HDR10 as input.
22847 It accepts the following parameters:
22851 Specify the output pixel format.
22853 Currently supported formats are:
22862 Set the output color primaries.
22864 Default is same as input.
22867 Set the output transfer characteristics.
22872 Set the output colorspace matrix.
22874 Default is same as input.
22878 @subsection Example
22882 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
22884 tonemap_vaapi=format=p010:t=bt2020-10
22888 @c man end VAAPI VIDEO FILTERS
22890 @chapter Video Sources
22891 @c man begin VIDEO SOURCES
22893 Below is a description of the currently available video sources.
22897 Buffer video frames, and make them available to the filter chain.
22899 This source is mainly intended for a programmatic use, in particular
22900 through the interface defined in @file{libavfilter/buffersrc.h}.
22902 It accepts the following parameters:
22907 Specify the size (width and height) of the buffered video frames. For the
22908 syntax of this option, check the
22909 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22912 The input video width.
22915 The input video height.
22918 A string representing the pixel format of the buffered video frames.
22919 It may be a number corresponding to a pixel format, or a pixel format
22923 Specify the timebase assumed by the timestamps of the buffered frames.
22926 Specify the frame rate expected for the video stream.
22928 @item pixel_aspect, sar
22929 The sample (pixel) aspect ratio of the input video.
22932 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
22933 to the filtergraph description to specify swscale flags for automatically
22934 inserted scalers. See @ref{Filtergraph syntax}.
22936 @item hw_frames_ctx
22937 When using a hardware pixel format, this should be a reference to an
22938 AVHWFramesContext describing input frames.
22943 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
22946 will instruct the source to accept video frames with size 320x240 and
22947 with format "yuv410p", assuming 1/24 as the timestamps timebase and
22948 square pixels (1:1 sample aspect ratio).
22949 Since the pixel format with name "yuv410p" corresponds to the number 6
22950 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
22951 this example corresponds to:
22953 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
22956 Alternatively, the options can be specified as a flat string, but this
22957 syntax is deprecated:
22959 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
22963 Create a pattern generated by an elementary cellular automaton.
22965 The initial state of the cellular automaton can be defined through the
22966 @option{filename} and @option{pattern} options. If such options are
22967 not specified an initial state is created randomly.
22969 At each new frame a new row in the video is filled with the result of
22970 the cellular automaton next generation. The behavior when the whole
22971 frame is filled is defined by the @option{scroll} option.
22973 This source accepts the following options:
22977 Read the initial cellular automaton state, i.e. the starting row, from
22978 the specified file.
22979 In the file, each non-whitespace character is considered an alive
22980 cell, a newline will terminate the row, and further characters in the
22981 file will be ignored.
22984 Read the initial cellular automaton state, i.e. the starting row, from
22985 the specified string.
22987 Each non-whitespace character in the string is considered an alive
22988 cell, a newline will terminate the row, and further characters in the
22989 string will be ignored.
22992 Set the video rate, that is the number of frames generated per second.
22995 @item random_fill_ratio, ratio
22996 Set the random fill ratio for the initial cellular automaton row. It
22997 is a floating point number value ranging from 0 to 1, defaults to
23000 This option is ignored when a file or a pattern is specified.
23002 @item random_seed, seed
23003 Set the seed for filling randomly the initial row, must be an integer
23004 included between 0 and UINT32_MAX. If not specified, or if explicitly
23005 set to -1, the filter will try to use a good random seed on a best
23009 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23010 Default value is 110.
23013 Set the size of the output video. For the syntax of this option, check the
23014 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23016 If @option{filename} or @option{pattern} is specified, the size is set
23017 by default to the width of the specified initial state row, and the
23018 height is set to @var{width} * PHI.
23020 If @option{size} is set, it must contain the width of the specified
23021 pattern string, and the specified pattern will be centered in the
23024 If a filename or a pattern string is not specified, the size value
23025 defaults to "320x518" (used for a randomly generated initial state).
23028 If set to 1, scroll the output upward when all the rows in the output
23029 have been already filled. If set to 0, the new generated row will be
23030 written over the top row just after the bottom row is filled.
23033 @item start_full, full
23034 If set to 1, completely fill the output with generated rows before
23035 outputting the first frame.
23036 This is the default behavior, for disabling set the value to 0.
23039 If set to 1, stitch the left and right row edges together.
23040 This is the default behavior, for disabling set the value to 0.
23043 @subsection Examples
23047 Read the initial state from @file{pattern}, and specify an output of
23050 cellauto=f=pattern:s=200x400
23054 Generate a random initial row with a width of 200 cells, with a fill
23057 cellauto=ratio=2/3:s=200x200
23061 Create a pattern generated by rule 18 starting by a single alive cell
23062 centered on an initial row with width 100:
23064 cellauto=p=@@:s=100x400:full=0:rule=18
23068 Specify a more elaborated initial pattern:
23070 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23075 @anchor{coreimagesrc}
23076 @section coreimagesrc
23077 Video source generated on GPU using Apple's CoreImage API on OSX.
23079 This video source is a specialized version of the @ref{coreimage} video filter.
23080 Use a core image generator at the beginning of the applied filterchain to
23081 generate the content.
23083 The coreimagesrc video source accepts the following options:
23085 @item list_generators
23086 List all available generators along with all their respective options as well as
23087 possible minimum and maximum values along with the default values.
23089 list_generators=true
23093 Specify the size of the sourced video. For the syntax of this option, check the
23094 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23095 The default value is @code{320x240}.
23098 Specify the frame rate of the sourced video, as the number of frames
23099 generated per second. It has to be a string in the format
23100 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23101 number or a valid video frame rate abbreviation. The default value is
23105 Set the sample aspect ratio of the sourced video.
23108 Set the duration of the sourced video. See
23109 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23110 for the accepted syntax.
23112 If not specified, or the expressed duration is negative, the video is
23113 supposed to be generated forever.
23116 Additionally, all options of the @ref{coreimage} video filter are accepted.
23117 A complete filterchain can be used for further processing of the
23118 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23119 and examples for details.
23121 @subsection Examples
23126 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23127 given as complete and escaped command-line for Apple's standard bash shell:
23129 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23131 This example is equivalent to the QRCode example of @ref{coreimage} without the
23132 need for a nullsrc video source.
23137 Generate several gradients.
23141 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23142 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23145 Set frame rate, expressed as number of frames per second. Default
23148 @item c0, c1, c2, c3, c4, c5, c6, c7
23149 Set 8 colors. Default values for colors is to pick random one.
23151 @item x0, y0, y0, y1
23152 Set gradient line source and destination points. If negative or out of range, random ones
23156 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23159 Set seed for picking gradient line points.
23162 Set the duration of the sourced video. See
23163 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23164 for the accepted syntax.
23166 If not specified, or the expressed duration is negative, the video is
23167 supposed to be generated forever.
23170 Set speed of gradients rotation.
23174 @section mandelbrot
23176 Generate a Mandelbrot set fractal, and progressively zoom towards the
23177 point specified with @var{start_x} and @var{start_y}.
23179 This source accepts the following options:
23184 Set the terminal pts value. Default value is 400.
23187 Set the terminal scale value.
23188 Must be a floating point value. Default value is 0.3.
23191 Set the inner coloring mode, that is the algorithm used to draw the
23192 Mandelbrot fractal internal region.
23194 It shall assume one of the following values:
23199 Show time until convergence.
23201 Set color based on point closest to the origin of the iterations.
23206 Default value is @var{mincol}.
23209 Set the bailout value. Default value is 10.0.
23212 Set the maximum of iterations performed by the rendering
23213 algorithm. Default value is 7189.
23216 Set outer coloring mode.
23217 It shall assume one of following values:
23219 @item iteration_count
23220 Set iteration count mode.
23221 @item normalized_iteration_count
23222 set normalized iteration count mode.
23224 Default value is @var{normalized_iteration_count}.
23227 Set frame rate, expressed as number of frames per second. Default
23231 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23232 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23235 Set the initial scale value. Default value is 3.0.
23238 Set the initial x position. Must be a floating point value between
23239 -100 and 100. Default value is -0.743643887037158704752191506114774.
23242 Set the initial y position. Must be a floating point value between
23243 -100 and 100. Default value is -0.131825904205311970493132056385139.
23248 Generate various test patterns, as generated by the MPlayer test filter.
23250 The size of the generated video is fixed, and is 256x256.
23251 This source is useful in particular for testing encoding features.
23253 This source accepts the following options:
23258 Specify the frame rate of the sourced video, as the number of frames
23259 generated per second. It has to be a string in the format
23260 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23261 number or a valid video frame rate abbreviation. The default value is
23265 Set the duration of the sourced video. See
23266 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23267 for the accepted syntax.
23269 If not specified, or the expressed duration is negative, the video is
23270 supposed to be generated forever.
23274 Set the number or the name of the test to perform. Supported tests are:
23288 @item max_frames, m
23289 Set the maximum number of frames generated for each test, default value is 30.
23293 Default value is "all", which will cycle through the list of all tests.
23298 mptestsrc=t=dc_luma
23301 will generate a "dc_luma" test pattern.
23303 @section frei0r_src
23305 Provide a frei0r source.
23307 To enable compilation of this filter you need to install the frei0r
23308 header and configure FFmpeg with @code{--enable-frei0r}.
23310 This source accepts the following parameters:
23315 The size of the video to generate. For the syntax of this option, check the
23316 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23319 The framerate of the generated video. It may be a string of the form
23320 @var{num}/@var{den} or a frame rate abbreviation.
23323 The name to the frei0r source to load. For more information regarding frei0r and
23324 how to set the parameters, read the @ref{frei0r} section in the video filters
23327 @item filter_params
23328 A '|'-separated list of parameters to pass to the frei0r source.
23332 For example, to generate a frei0r partik0l source with size 200x200
23333 and frame rate 10 which is overlaid on the overlay filter main input:
23335 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23340 Generate a life pattern.
23342 This source is based on a generalization of John Conway's life game.
23344 The sourced input represents a life grid, each pixel represents a cell
23345 which can be in one of two possible states, alive or dead. Every cell
23346 interacts with its eight neighbours, which are the cells that are
23347 horizontally, vertically, or diagonally adjacent.
23349 At each interaction the grid evolves according to the adopted rule,
23350 which specifies the number of neighbor alive cells which will make a
23351 cell stay alive or born. The @option{rule} option allows one to specify
23354 This source accepts the following options:
23358 Set the file from which to read the initial grid state. In the file,
23359 each non-whitespace character is considered an alive cell, and newline
23360 is used to delimit the end of each row.
23362 If this option is not specified, the initial grid is generated
23366 Set the video rate, that is the number of frames generated per second.
23369 @item random_fill_ratio, ratio
23370 Set the random fill ratio for the initial random grid. It is a
23371 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23372 It is ignored when a file is specified.
23374 @item random_seed, seed
23375 Set the seed for filling the initial random grid, must be an integer
23376 included between 0 and UINT32_MAX. If not specified, or if explicitly
23377 set to -1, the filter will try to use a good random seed on a best
23383 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23384 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23385 @var{NS} specifies the number of alive neighbor cells which make a
23386 live cell stay alive, and @var{NB} the number of alive neighbor cells
23387 which make a dead cell to become alive (i.e. to "born").
23388 "s" and "b" can be used in place of "S" and "B", respectively.
23390 Alternatively a rule can be specified by an 18-bits integer. The 9
23391 high order bits are used to encode the next cell state if it is alive
23392 for each number of neighbor alive cells, the low order bits specify
23393 the rule for "borning" new cells. Higher order bits encode for an
23394 higher number of neighbor cells.
23395 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23396 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23398 Default value is "S23/B3", which is the original Conway's game of life
23399 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23400 cells, and will born a new cell if there are three alive cells around
23404 Set the size of the output video. For the syntax of this option, check the
23405 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23407 If @option{filename} is specified, the size is set by default to the
23408 same size of the input file. If @option{size} is set, it must contain
23409 the size specified in the input file, and the initial grid defined in
23410 that file is centered in the larger resulting area.
23412 If a filename is not specified, the size value defaults to "320x240"
23413 (used for a randomly generated initial grid).
23416 If set to 1, stitch the left and right grid edges together, and the
23417 top and bottom edges also. Defaults to 1.
23420 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23421 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23422 value from 0 to 255.
23425 Set the color of living (or new born) cells.
23428 Set the color of dead cells. If @option{mold} is set, this is the first color
23429 used to represent a dead cell.
23432 Set mold color, for definitely dead and moldy cells.
23434 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23435 ffmpeg-utils manual,ffmpeg-utils}.
23438 @subsection Examples
23442 Read a grid from @file{pattern}, and center it on a grid of size
23445 life=f=pattern:s=300x300
23449 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23451 life=ratio=2/3:s=200x200
23455 Specify a custom rule for evolving a randomly generated grid:
23461 Full example with slow death effect (mold) using @command{ffplay}:
23463 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23470 @anchor{haldclutsrc}
23473 @anchor{pal100bars}
23474 @anchor{rgbtestsrc}
23476 @anchor{smptehdbars}
23479 @anchor{yuvtestsrc}
23480 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23482 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23484 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23486 The @code{color} source provides an uniformly colored input.
23488 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23489 @ref{haldclut} filter.
23491 The @code{nullsrc} source returns unprocessed video frames. It is
23492 mainly useful to be employed in analysis / debugging tools, or as the
23493 source for filters which ignore the input data.
23495 The @code{pal75bars} source generates a color bars pattern, based on
23496 EBU PAL recommendations with 75% color levels.
23498 The @code{pal100bars} source generates a color bars pattern, based on
23499 EBU PAL recommendations with 100% color levels.
23501 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23502 detecting RGB vs BGR issues. You should see a red, green and blue
23503 stripe from top to bottom.
23505 The @code{smptebars} source generates a color bars pattern, based on
23506 the SMPTE Engineering Guideline EG 1-1990.
23508 The @code{smptehdbars} source generates a color bars pattern, based on
23509 the SMPTE RP 219-2002.
23511 The @code{testsrc} source generates a test video pattern, showing a
23512 color pattern, a scrolling gradient and a timestamp. This is mainly
23513 intended for testing purposes.
23515 The @code{testsrc2} source is similar to testsrc, but supports more
23516 pixel formats instead of just @code{rgb24}. This allows using it as an
23517 input for other tests without requiring a format conversion.
23519 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23520 see a y, cb and cr stripe from top to bottom.
23522 The sources accept the following parameters:
23527 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23528 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23529 pixels to be used as identity matrix for 3D lookup tables. Each component is
23530 coded on a @code{1/(N*N)} scale.
23533 Specify the color of the source, only available in the @code{color}
23534 source. For the syntax of this option, check the
23535 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23538 Specify the size of the sourced video. For the syntax of this option, check the
23539 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23540 The default value is @code{320x240}.
23542 This option is not available with the @code{allrgb}, @code{allyuv}, and
23543 @code{haldclutsrc} filters.
23546 Specify the frame rate of the sourced video, as the number of frames
23547 generated per second. It has to be a string in the format
23548 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23549 number or a valid video frame rate abbreviation. The default value is
23553 Set the duration of the sourced video. See
23554 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23555 for the accepted syntax.
23557 If not specified, or the expressed duration is negative, the video is
23558 supposed to be generated forever.
23560 Since the frame rate is used as time base, all frames including the last one
23561 will have their full duration. If the specified duration is not a multiple
23562 of the frame duration, it will be rounded up.
23565 Set the sample aspect ratio of the sourced video.
23568 Specify the alpha (opacity) of the background, only available in the
23569 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23570 255 (fully opaque, the default).
23573 Set the number of decimals to show in the timestamp, only available in the
23574 @code{testsrc} source.
23576 The displayed timestamp value will correspond to the original
23577 timestamp value multiplied by the power of 10 of the specified
23578 value. Default value is 0.
23581 @subsection Examples
23585 Generate a video with a duration of 5.3 seconds, with size
23586 176x144 and a frame rate of 10 frames per second:
23588 testsrc=duration=5.3:size=qcif:rate=10
23592 The following graph description will generate a red source
23593 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23596 color=c=red@@0.2:s=qcif:r=10
23600 If the input content is to be ignored, @code{nullsrc} can be used. The
23601 following command generates noise in the luminance plane by employing
23602 the @code{geq} filter:
23604 nullsrc=s=256x256, geq=random(1)*255:128:128
23608 @subsection Commands
23610 The @code{color} source supports the following commands:
23614 Set the color of the created image. Accepts the same syntax of the
23615 corresponding @option{color} option.
23620 Generate video using an OpenCL program.
23625 OpenCL program source file.
23628 Kernel name in program.
23631 Size of frames to generate. This must be set.
23634 Pixel format to use for the generated frames. This must be set.
23637 Number of frames generated every second. Default value is '25'.
23641 For details of how the program loading works, see the @ref{program_opencl}
23648 Generate a colour ramp by setting pixel values from the position of the pixel
23649 in the output image. (Note that this will work with all pixel formats, but
23650 the generated output will not be the same.)
23652 __kernel void ramp(__write_only image2d_t dst,
23653 unsigned int index)
23655 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23658 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23660 write_imagef(dst, loc, val);
23665 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23667 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23668 unsigned int index)
23670 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23672 float4 value = 0.0f;
23673 int x = loc.x + index;
23674 int y = loc.y + index;
23675 while (x > 0 || y > 0) {
23676 if (x % 3 == 1 && y % 3 == 1) {
23684 write_imagef(dst, loc, value);
23690 @section sierpinski
23692 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23694 This source accepts the following options:
23698 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23699 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23702 Set frame rate, expressed as number of frames per second. Default
23706 Set seed which is used for random panning.
23709 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23712 Set fractal type, can be default @code{carpet} or @code{triangle}.
23715 @c man end VIDEO SOURCES
23717 @chapter Video Sinks
23718 @c man begin VIDEO SINKS
23720 Below is a description of the currently available video sinks.
23722 @section buffersink
23724 Buffer video frames, and make them available to the end of the filter
23727 This sink is mainly intended for programmatic use, in particular
23728 through the interface defined in @file{libavfilter/buffersink.h}
23729 or the options system.
23731 It accepts a pointer to an AVBufferSinkContext structure, which
23732 defines the incoming buffers' formats, to be passed as the opaque
23733 parameter to @code{avfilter_init_filter} for initialization.
23737 Null video sink: do absolutely nothing with the input video. It is
23738 mainly useful as a template and for use in analysis / debugging
23741 @c man end VIDEO SINKS
23743 @chapter Multimedia Filters
23744 @c man begin MULTIMEDIA FILTERS
23746 Below is a description of the currently available multimedia filters.
23750 Convert input audio to a video output, displaying the audio bit scope.
23752 The filter accepts the following options:
23756 Set frame rate, expressed as number of frames per second. Default
23760 Specify the video size for the output. For the syntax of this option, check the
23761 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23762 Default value is @code{1024x256}.
23765 Specify list of colors separated by space or by '|' which will be used to
23766 draw channels. Unrecognized or missing colors will be replaced
23770 @section adrawgraph
23771 Draw a graph using input audio metadata.
23773 See @ref{drawgraph}
23775 @section agraphmonitor
23777 See @ref{graphmonitor}.
23779 @section ahistogram
23781 Convert input audio to a video output, displaying the volume histogram.
23783 The filter accepts the following options:
23787 Specify how histogram is calculated.
23789 It accepts the following values:
23792 Use single histogram for all channels.
23794 Use separate histogram for each channel.
23796 Default is @code{single}.
23799 Set frame rate, expressed as number of frames per second. Default
23803 Specify the video size for the output. For the syntax of this option, check the
23804 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23805 Default value is @code{hd720}.
23810 It accepts the following values:
23821 reverse logarithmic
23823 Default is @code{log}.
23826 Set amplitude scale.
23828 It accepts the following values:
23835 Default is @code{log}.
23838 Set how much frames to accumulate in histogram.
23839 Default is 1. Setting this to -1 accumulates all frames.
23842 Set histogram ratio of window height.
23845 Set sonogram sliding.
23847 It accepts the following values:
23850 replace old rows with new ones.
23852 scroll from top to bottom.
23854 Default is @code{replace}.
23857 @section aphasemeter
23859 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
23860 representing mean phase of current audio frame. A video output can also be produced and is
23861 enabled by default. The audio is passed through as first output.
23863 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
23864 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
23865 and @code{1} means channels are in phase.
23867 The filter accepts the following options, all related to its video output:
23871 Set the output frame rate. Default value is @code{25}.
23874 Set the video size for the output. For the syntax of this option, check the
23875 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23876 Default value is @code{800x400}.
23881 Specify the red, green, blue contrast. Default values are @code{2},
23882 @code{7} and @code{1}.
23883 Allowed range is @code{[0, 255]}.
23886 Set color which will be used for drawing median phase. If color is
23887 @code{none} which is default, no median phase value will be drawn.
23890 Enable video output. Default is enabled.
23893 @subsection phasing detection
23895 The filter also detects out of phase and mono sequences in stereo streams.
23896 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
23898 The filter accepts the following options for this detection:
23902 Enable mono and out of phase detection. Default is disabled.
23905 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
23906 Allowed range is @code{[0, 1]}.
23909 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
23910 Allowed range is @code{[90, 180]}.
23913 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
23916 @subsection Examples
23920 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
23922 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
23926 @section avectorscope
23928 Convert input audio to a video output, representing the audio vector
23931 The filter is used to measure the difference between channels of stereo
23932 audio stream. A monaural signal, consisting of identical left and right
23933 signal, results in straight vertical line. Any stereo separation is visible
23934 as a deviation from this line, creating a Lissajous figure.
23935 If the straight (or deviation from it) but horizontal line appears this
23936 indicates that the left and right channels are out of phase.
23938 The filter accepts the following options:
23942 Set the vectorscope mode.
23944 Available values are:
23947 Lissajous rotated by 45 degrees.
23950 Same as above but not rotated.
23953 Shape resembling half of circle.
23956 Default value is @samp{lissajous}.
23959 Set the video size for the output. For the syntax of this option, check the
23960 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23961 Default value is @code{400x400}.
23964 Set the output frame rate. Default value is @code{25}.
23970 Specify the red, green, blue and alpha contrast. Default values are @code{40},
23971 @code{160}, @code{80} and @code{255}.
23972 Allowed range is @code{[0, 255]}.
23978 Specify the red, green, blue and alpha fade. Default values are @code{15},
23979 @code{10}, @code{5} and @code{5}.
23980 Allowed range is @code{[0, 255]}.
23983 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
23984 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
23987 Set the vectorscope drawing mode.
23989 Available values are:
23992 Draw dot for each sample.
23995 Draw line between previous and current sample.
23998 Default value is @samp{dot}.
24001 Specify amplitude scale of audio samples.
24003 Available values are:
24019 Swap left channel axis with right channel axis.
24029 Mirror only x axis.
24032 Mirror only y axis.
24040 @subsection Examples
24044 Complete example using @command{ffplay}:
24046 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24047 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24051 @section bench, abench
24053 Benchmark part of a filtergraph.
24055 The filter accepts the following options:
24059 Start or stop a timer.
24061 Available values are:
24064 Get the current time, set it as frame metadata (using the key
24065 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24068 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24069 the input frame metadata to get the time difference. Time difference, average,
24070 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24071 @code{min}) are then printed. The timestamps are expressed in seconds.
24075 @subsection Examples
24079 Benchmark @ref{selectivecolor} filter:
24081 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24087 Concatenate audio and video streams, joining them together one after the
24090 The filter works on segments of synchronized video and audio streams. All
24091 segments must have the same number of streams of each type, and that will
24092 also be the number of streams at output.
24094 The filter accepts the following options:
24099 Set the number of segments. Default is 2.
24102 Set the number of output video streams, that is also the number of video
24103 streams in each segment. Default is 1.
24106 Set the number of output audio streams, that is also the number of audio
24107 streams in each segment. Default is 0.
24110 Activate unsafe mode: do not fail if segments have a different format.
24114 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24115 @var{a} audio outputs.
24117 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24118 segment, in the same order as the outputs, then the inputs for the second
24121 Related streams do not always have exactly the same duration, for various
24122 reasons including codec frame size or sloppy authoring. For that reason,
24123 related synchronized streams (e.g. a video and its audio track) should be
24124 concatenated at once. The concat filter will use the duration of the longest
24125 stream in each segment (except the last one), and if necessary pad shorter
24126 audio streams with silence.
24128 For this filter to work correctly, all segments must start at timestamp 0.
24130 All corresponding streams must have the same parameters in all segments; the
24131 filtering system will automatically select a common pixel format for video
24132 streams, and a common sample format, sample rate and channel layout for
24133 audio streams, but other settings, such as resolution, must be converted
24134 explicitly by the user.
24136 Different frame rates are acceptable but will result in variable frame rate
24137 at output; be sure to configure the output file to handle it.
24139 @subsection Examples
24143 Concatenate an opening, an episode and an ending, all in bilingual version
24144 (video in stream 0, audio in streams 1 and 2):
24146 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24147 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24148 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24149 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24153 Concatenate two parts, handling audio and video separately, using the
24154 (a)movie sources, and adjusting the resolution:
24156 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24157 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24158 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24160 Note that a desync will happen at the stitch if the audio and video streams
24161 do not have exactly the same duration in the first file.
24165 @subsection Commands
24167 This filter supports the following commands:
24170 Close the current segment and step to the next one
24176 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24177 level. By default, it logs a message at a frequency of 10Hz with the
24178 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24179 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24181 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24182 sample format is double-precision floating point. The input stream will be converted to
24183 this specification, if needed. Users may need to insert aformat and/or aresample filters
24184 after this filter to obtain the original parameters.
24186 The filter also has a video output (see the @var{video} option) with a real
24187 time graph to observe the loudness evolution. The graphic contains the logged
24188 message mentioned above, so it is not printed anymore when this option is set,
24189 unless the verbose logging is set. The main graphing area contains the
24190 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24191 the momentary loudness (400 milliseconds), but can optionally be configured
24192 to instead display short-term loudness (see @var{gauge}).
24194 The green area marks a +/- 1LU target range around the target loudness
24195 (-23LUFS by default, unless modified through @var{target}).
24197 More information about the Loudness Recommendation EBU R128 on
24198 @url{http://tech.ebu.ch/loudness}.
24200 The filter accepts the following options:
24205 Activate the video output. The audio stream is passed unchanged whether this
24206 option is set or no. The video stream will be the first output stream if
24207 activated. Default is @code{0}.
24210 Set the video size. This option is for video only. For the syntax of this
24212 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24213 Default and minimum resolution is @code{640x480}.
24216 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24217 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24218 other integer value between this range is allowed.
24221 Set metadata injection. If set to @code{1}, the audio input will be segmented
24222 into 100ms output frames, each of them containing various loudness information
24223 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24225 Default is @code{0}.
24228 Force the frame logging level.
24230 Available values are:
24233 information logging level
24235 verbose logging level
24238 By default, the logging level is set to @var{info}. If the @option{video} or
24239 the @option{metadata} options are set, it switches to @var{verbose}.
24244 Available modes can be cumulated (the option is a @code{flag} type). Possible
24248 Disable any peak mode (default).
24250 Enable sample-peak mode.
24252 Simple peak mode looking for the higher sample value. It logs a message
24253 for sample-peak (identified by @code{SPK}).
24255 Enable true-peak mode.
24257 If enabled, the peak lookup is done on an over-sampled version of the input
24258 stream for better peak accuracy. It logs a message for true-peak.
24259 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24260 This mode requires a build with @code{libswresample}.
24264 Treat mono input files as "dual mono". If a mono file is intended for playback
24265 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24266 If set to @code{true}, this option will compensate for this effect.
24267 Multi-channel input files are not affected by this option.
24270 Set a specific pan law to be used for the measurement of dual mono files.
24271 This parameter is optional, and has a default value of -3.01dB.
24274 Set a specific target level (in LUFS) used as relative zero in the visualization.
24275 This parameter is optional and has a default value of -23LUFS as specified
24276 by EBU R128. However, material published online may prefer a level of -16LUFS
24277 (e.g. for use with podcasts or video platforms).
24280 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24281 @code{shortterm}. By default the momentary value will be used, but in certain
24282 scenarios it may be more useful to observe the short term value instead (e.g.
24286 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24287 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24288 video output, not the summary or continuous log output.
24291 @subsection Examples
24295 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24297 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24301 Run an analysis with @command{ffmpeg}:
24303 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24307 @section interleave, ainterleave
24309 Temporally interleave frames from several inputs.
24311 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24313 These filters read frames from several inputs and send the oldest
24314 queued frame to the output.
24316 Input streams must have well defined, monotonically increasing frame
24319 In order to submit one frame to output, these filters need to enqueue
24320 at least one frame for each input, so they cannot work in case one
24321 input is not yet terminated and will not receive incoming frames.
24323 For example consider the case when one input is a @code{select} filter
24324 which always drops input frames. The @code{interleave} filter will keep
24325 reading from that input, but it will never be able to send new frames
24326 to output until the input sends an end-of-stream signal.
24328 Also, depending on inputs synchronization, the filters will drop
24329 frames in case one input receives more frames than the other ones, and
24330 the queue is already filled.
24332 These filters accept the following options:
24336 Set the number of different inputs, it is 2 by default.
24339 How to determine the end-of-stream.
24343 The duration of the longest input. (default)
24346 The duration of the shortest input.
24349 The duration of the first input.
24354 @subsection Examples
24358 Interleave frames belonging to different streams using @command{ffmpeg}:
24360 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24364 Add flickering blur effect:
24366 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24370 @section metadata, ametadata
24372 Manipulate frame metadata.
24374 This filter accepts the following options:
24378 Set mode of operation of the filter.
24380 Can be one of the following:
24384 If both @code{value} and @code{key} is set, select frames
24385 which have such metadata. If only @code{key} is set, select
24386 every frame that has such key in metadata.
24389 Add new metadata @code{key} and @code{value}. If key is already available
24393 Modify value of already present key.
24396 If @code{value} is set, delete only keys that have such value.
24397 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24401 Print key and its value if metadata was found. If @code{key} is not set print all
24402 metadata values available in frame.
24406 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24409 Set metadata value which will be used. This option is mandatory for
24410 @code{modify} and @code{add} mode.
24413 Which function to use when comparing metadata value and @code{value}.
24415 Can be one of following:
24419 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24422 Values are interpreted as strings, returns true if metadata value starts with
24423 the @code{value} option string.
24426 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24429 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24432 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24435 Values are interpreted as floats, returns true if expression from option @code{expr}
24439 Values are interpreted as strings, returns true if metadata value ends with
24440 the @code{value} option string.
24444 Set expression which is used when @code{function} is set to @code{expr}.
24445 The expression is evaluated through the eval API and can contain the following
24450 Float representation of @code{value} from metadata key.
24453 Float representation of @code{value} as supplied by user in @code{value} option.
24457 If specified in @code{print} mode, output is written to the named file. Instead of
24458 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24459 for standard output. If @code{file} option is not set, output is written to the log
24460 with AV_LOG_INFO loglevel.
24463 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24467 @subsection Examples
24471 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24474 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24477 Print silencedetect output to file @file{metadata.txt}.
24479 silencedetect,ametadata=mode=print:file=metadata.txt
24482 Direct all metadata to a pipe with file descriptor 4.
24484 metadata=mode=print:file='pipe\:4'
24488 @section perms, aperms
24490 Set read/write permissions for the output frames.
24492 These filters are mainly aimed at developers to test direct path in the
24493 following filter in the filtergraph.
24495 The filters accept the following options:
24499 Select the permissions mode.
24501 It accepts the following values:
24504 Do nothing. This is the default.
24506 Set all the output frames read-only.
24508 Set all the output frames directly writable.
24510 Make the frame read-only if writable, and writable if read-only.
24512 Set each output frame read-only or writable randomly.
24516 Set the seed for the @var{random} mode, must be an integer included between
24517 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24518 @code{-1}, the filter will try to use a good random seed on a best effort
24522 Note: in case of auto-inserted filter between the permission filter and the
24523 following one, the permission might not be received as expected in that
24524 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24525 perms/aperms filter can avoid this problem.
24527 @section realtime, arealtime
24529 Slow down filtering to match real time approximately.
24531 These filters will pause the filtering for a variable amount of time to
24532 match the output rate with the input timestamps.
24533 They are similar to the @option{re} option to @code{ffmpeg}.
24535 They accept the following options:
24539 Time limit for the pauses. Any pause longer than that will be considered
24540 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24542 Speed factor for processing. The value must be a float larger than zero.
24543 Values larger than 1.0 will result in faster than realtime processing,
24544 smaller will slow processing down. The @var{limit} is automatically adapted
24545 accordingly. Default is 1.0.
24547 A processing speed faster than what is possible without these filters cannot
24552 @section select, aselect
24554 Select frames to pass in output.
24556 This filter accepts the following options:
24561 Set expression, which is evaluated for each input frame.
24563 If the expression is evaluated to zero, the frame is discarded.
24565 If the evaluation result is negative or NaN, the frame is sent to the
24566 first output; otherwise it is sent to the output with index
24567 @code{ceil(val)-1}, assuming that the input index starts from 0.
24569 For example a value of @code{1.2} corresponds to the output with index
24570 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24573 Set the number of outputs. The output to which to send the selected
24574 frame is based on the result of the evaluation. Default value is 1.
24577 The expression can contain the following constants:
24581 The (sequential) number of the filtered frame, starting from 0.
24584 The (sequential) number of the selected frame, starting from 0.
24586 @item prev_selected_n
24587 The sequential number of the last selected frame. It's NAN if undefined.
24590 The timebase of the input timestamps.
24593 The PTS (Presentation TimeStamp) of the filtered video frame,
24594 expressed in @var{TB} units. It's NAN if undefined.
24597 The PTS of the filtered video frame,
24598 expressed in seconds. It's NAN if undefined.
24601 The PTS of the previously filtered video frame. It's NAN if undefined.
24603 @item prev_selected_pts
24604 The PTS of the last previously filtered video frame. It's NAN if undefined.
24606 @item prev_selected_t
24607 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24610 The PTS of the first video frame in the video. It's NAN if undefined.
24613 The time of the first video frame in the video. It's NAN if undefined.
24615 @item pict_type @emph{(video only)}
24616 The type of the filtered frame. It can assume one of the following
24628 @item interlace_type @emph{(video only)}
24629 The frame interlace type. It can assume one of the following values:
24632 The frame is progressive (not interlaced).
24634 The frame is top-field-first.
24636 The frame is bottom-field-first.
24639 @item consumed_sample_n @emph{(audio only)}
24640 the number of selected samples before the current frame
24642 @item samples_n @emph{(audio only)}
24643 the number of samples in the current frame
24645 @item sample_rate @emph{(audio only)}
24646 the input sample rate
24649 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24652 the position in the file of the filtered frame, -1 if the information
24653 is not available (e.g. for synthetic video)
24655 @item scene @emph{(video only)}
24656 value between 0 and 1 to indicate a new scene; a low value reflects a low
24657 probability for the current frame to introduce a new scene, while a higher
24658 value means the current frame is more likely to be one (see the example below)
24660 @item concatdec_select
24661 The concat demuxer can select only part of a concat input file by setting an
24662 inpoint and an outpoint, but the output packets may not be entirely contained
24663 in the selected interval. By using this variable, it is possible to skip frames
24664 generated by the concat demuxer which are not exactly contained in the selected
24667 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24668 and the @var{lavf.concat.duration} packet metadata values which are also
24669 present in the decoded frames.
24671 The @var{concatdec_select} variable is -1 if the frame pts is at least
24672 start_time and either the duration metadata is missing or the frame pts is less
24673 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24676 That basically means that an input frame is selected if its pts is within the
24677 interval set by the concat demuxer.
24681 The default value of the select expression is "1".
24683 @subsection Examples
24687 Select all frames in input:
24692 The example above is the same as:
24704 Select only I-frames:
24706 select='eq(pict_type\,I)'
24710 Select one frame every 100:
24712 select='not(mod(n\,100))'
24716 Select only frames contained in the 10-20 time interval:
24718 select=between(t\,10\,20)
24722 Select only I-frames contained in the 10-20 time interval:
24724 select=between(t\,10\,20)*eq(pict_type\,I)
24728 Select frames with a minimum distance of 10 seconds:
24730 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24734 Use aselect to select only audio frames with samples number > 100:
24736 aselect='gt(samples_n\,100)'
24740 Create a mosaic of the first scenes:
24742 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24745 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24749 Send even and odd frames to separate outputs, and compose them:
24751 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24755 Select useful frames from an ffconcat file which is using inpoints and
24756 outpoints but where the source files are not intra frame only.
24758 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24762 @section sendcmd, asendcmd
24764 Send commands to filters in the filtergraph.
24766 These filters read commands to be sent to other filters in the
24769 @code{sendcmd} must be inserted between two video filters,
24770 @code{asendcmd} must be inserted between two audio filters, but apart
24771 from that they act the same way.
24773 The specification of commands can be provided in the filter arguments
24774 with the @var{commands} option, or in a file specified by the
24775 @var{filename} option.
24777 These filters accept the following options:
24780 Set the commands to be read and sent to the other filters.
24782 Set the filename of the commands to be read and sent to the other
24786 @subsection Commands syntax
24788 A commands description consists of a sequence of interval
24789 specifications, comprising a list of commands to be executed when a
24790 particular event related to that interval occurs. The occurring event
24791 is typically the current frame time entering or leaving a given time
24794 An interval is specified by the following syntax:
24796 @var{START}[-@var{END}] @var{COMMANDS};
24799 The time interval is specified by the @var{START} and @var{END} times.
24800 @var{END} is optional and defaults to the maximum time.
24802 The current frame time is considered within the specified interval if
24803 it is included in the interval [@var{START}, @var{END}), that is when
24804 the time is greater or equal to @var{START} and is lesser than
24807 @var{COMMANDS} consists of a sequence of one or more command
24808 specifications, separated by ",", relating to that interval. The
24809 syntax of a command specification is given by:
24811 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
24814 @var{FLAGS} is optional and specifies the type of events relating to
24815 the time interval which enable sending the specified command, and must
24816 be a non-null sequence of identifier flags separated by "+" or "|" and
24817 enclosed between "[" and "]".
24819 The following flags are recognized:
24822 The command is sent when the current frame timestamp enters the
24823 specified interval. In other words, the command is sent when the
24824 previous frame timestamp was not in the given interval, and the
24828 The command is sent when the current frame timestamp leaves the
24829 specified interval. In other words, the command is sent when the
24830 previous frame timestamp was in the given interval, and the
24834 The command @var{ARG} is interpreted as expression and result of
24835 expression is passed as @var{ARG}.
24837 The expression is evaluated through the eval API and can contain the following
24842 Original position in the file of the frame, or undefined if undefined
24843 for the current frame.
24846 The presentation timestamp in input.
24849 The count of the input frame for video or audio, starting from 0.
24852 The time in seconds of the current frame.
24855 The start time in seconds of the current command interval.
24858 The end time in seconds of the current command interval.
24861 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
24866 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
24869 @var{TARGET} specifies the target of the command, usually the name of
24870 the filter class or a specific filter instance name.
24872 @var{COMMAND} specifies the name of the command for the target filter.
24874 @var{ARG} is optional and specifies the optional list of argument for
24875 the given @var{COMMAND}.
24877 Between one interval specification and another, whitespaces, or
24878 sequences of characters starting with @code{#} until the end of line,
24879 are ignored and can be used to annotate comments.
24881 A simplified BNF description of the commands specification syntax
24884 @var{COMMAND_FLAG} ::= "enter" | "leave"
24885 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
24886 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
24887 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
24888 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
24889 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
24892 @subsection Examples
24896 Specify audio tempo change at second 4:
24898 asendcmd=c='4.0 atempo tempo 1.5',atempo
24902 Target a specific filter instance:
24904 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
24908 Specify a list of drawtext and hue commands in a file.
24910 # show text in the interval 5-10
24911 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
24912 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
24914 # desaturate the image in the interval 15-20
24915 15.0-20.0 [enter] hue s 0,
24916 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
24918 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
24920 # apply an exponential saturation fade-out effect, starting from time 25
24921 25 [enter] hue s exp(25-t)
24924 A filtergraph allowing to read and process the above command list
24925 stored in a file @file{test.cmd}, can be specified with:
24927 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
24932 @section setpts, asetpts
24934 Change the PTS (presentation timestamp) of the input frames.
24936 @code{setpts} works on video frames, @code{asetpts} on audio frames.
24938 This filter accepts the following options:
24943 The expression which is evaluated for each frame to construct its timestamp.
24947 The expression is evaluated through the eval API and can contain the following
24951 @item FRAME_RATE, FR
24952 frame rate, only defined for constant frame-rate video
24955 The presentation timestamp in input
24958 The count of the input frame for video or the number of consumed samples,
24959 not including the current frame for audio, starting from 0.
24961 @item NB_CONSUMED_SAMPLES
24962 The number of consumed samples, not including the current frame (only
24965 @item NB_SAMPLES, S
24966 The number of samples in the current frame (only audio)
24968 @item SAMPLE_RATE, SR
24969 The audio sample rate.
24972 The PTS of the first frame.
24975 the time in seconds of the first frame
24978 State whether the current frame is interlaced.
24981 the time in seconds of the current frame
24984 original position in the file of the frame, or undefined if undefined
24985 for the current frame
24988 The previous input PTS.
24991 previous input time in seconds
24994 The previous output PTS.
24997 previous output time in seconds
25000 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25004 The wallclock (RTC) time at the start of the movie in microseconds.
25007 The timebase of the input timestamps.
25011 @subsection Examples
25015 Start counting PTS from zero
25017 setpts=PTS-STARTPTS
25021 Apply fast motion effect:
25027 Apply slow motion effect:
25033 Set fixed rate of 25 frames per second:
25039 Set fixed rate 25 fps with some jitter:
25041 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25045 Apply an offset of 10 seconds to the input PTS:
25051 Generate timestamps from a "live source" and rebase onto the current timebase:
25053 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25057 Generate timestamps by counting samples:
25066 Force color range for the output video frame.
25068 The @code{setrange} filter marks the color range property for the
25069 output frames. It does not change the input frame, but only sets the
25070 corresponding property, which affects how the frame is treated by
25073 The filter accepts the following options:
25078 Available values are:
25082 Keep the same color range property.
25084 @item unspecified, unknown
25085 Set the color range as unspecified.
25087 @item limited, tv, mpeg
25088 Set the color range as limited.
25090 @item full, pc, jpeg
25091 Set the color range as full.
25095 @section settb, asettb
25097 Set the timebase to use for the output frames timestamps.
25098 It is mainly useful for testing timebase configuration.
25100 It accepts the following parameters:
25105 The expression which is evaluated into the output timebase.
25109 The value for @option{tb} is an arithmetic expression representing a
25110 rational. The expression can contain the constants "AVTB" (the default
25111 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25112 audio only). Default value is "intb".
25114 @subsection Examples
25118 Set the timebase to 1/25:
25124 Set the timebase to 1/10:
25130 Set the timebase to 1001/1000:
25136 Set the timebase to 2*intb:
25142 Set the default timebase value:
25149 Convert input audio to a video output representing frequency spectrum
25150 logarithmically using Brown-Puckette constant Q transform algorithm with
25151 direct frequency domain coefficient calculation (but the transform itself
25152 is not really constant Q, instead the Q factor is actually variable/clamped),
25153 with musical tone scale, from E0 to D#10.
25155 The filter accepts the following options:
25159 Specify the video size for the output. It must be even. For the syntax of this option,
25160 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25161 Default value is @code{1920x1080}.
25164 Set the output frame rate. Default value is @code{25}.
25167 Set the bargraph height. It must be even. Default value is @code{-1} which
25168 computes the bargraph height automatically.
25171 Set the axis height. It must be even. Default value is @code{-1} which computes
25172 the axis height automatically.
25175 Set the sonogram height. It must be even. Default value is @code{-1} which
25176 computes the sonogram height automatically.
25179 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25180 instead. Default value is @code{1}.
25182 @item sono_v, volume
25183 Specify the sonogram volume expression. It can contain variables:
25186 the @var{bar_v} evaluated expression
25187 @item frequency, freq, f
25188 the frequency where it is evaluated
25189 @item timeclamp, tc
25190 the value of @var{timeclamp} option
25194 @item a_weighting(f)
25195 A-weighting of equal loudness
25196 @item b_weighting(f)
25197 B-weighting of equal loudness
25198 @item c_weighting(f)
25199 C-weighting of equal loudness.
25201 Default value is @code{16}.
25203 @item bar_v, volume2
25204 Specify the bargraph volume expression. It can contain variables:
25207 the @var{sono_v} evaluated expression
25208 @item frequency, freq, f
25209 the frequency where it is evaluated
25210 @item timeclamp, tc
25211 the value of @var{timeclamp} option
25215 @item a_weighting(f)
25216 A-weighting of equal loudness
25217 @item b_weighting(f)
25218 B-weighting of equal loudness
25219 @item c_weighting(f)
25220 C-weighting of equal loudness.
25222 Default value is @code{sono_v}.
25224 @item sono_g, gamma
25225 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25226 higher gamma makes the spectrum having more range. Default value is @code{3}.
25227 Acceptable range is @code{[1, 7]}.
25229 @item bar_g, gamma2
25230 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25234 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25235 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25237 @item timeclamp, tc
25238 Specify the transform timeclamp. At low frequency, there is trade-off between
25239 accuracy in time domain and frequency domain. If timeclamp is lower,
25240 event in time domain is represented more accurately (such as fast bass drum),
25241 otherwise event in frequency domain is represented more accurately
25242 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25245 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25246 limits future samples by applying asymmetric windowing in time domain, useful
25247 when low latency is required. Accepted range is @code{[0, 1]}.
25250 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25251 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25254 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25255 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25258 This option is deprecated and ignored.
25261 Specify the transform length in time domain. Use this option to control accuracy
25262 trade-off between time domain and frequency domain at every frequency sample.
25263 It can contain variables:
25265 @item frequency, freq, f
25266 the frequency where it is evaluated
25267 @item timeclamp, tc
25268 the value of @var{timeclamp} option.
25270 Default value is @code{384*tc/(384+tc*f)}.
25273 Specify the transform count for every video frame. Default value is @code{6}.
25274 Acceptable range is @code{[1, 30]}.
25277 Specify the transform count for every single pixel. Default value is @code{0},
25278 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25281 Specify font file for use with freetype to draw the axis. If not specified,
25282 use embedded font. Note that drawing with font file or embedded font is not
25283 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25287 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25288 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25292 Specify font color expression. This is arithmetic expression that should return
25293 integer value 0xRRGGBB. It can contain variables:
25295 @item frequency, freq, f
25296 the frequency where it is evaluated
25297 @item timeclamp, tc
25298 the value of @var{timeclamp} option
25303 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25304 @item r(x), g(x), b(x)
25305 red, green, and blue value of intensity x.
25307 Default value is @code{st(0, (midi(f)-59.5)/12);
25308 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25309 r(1-ld(1)) + b(ld(1))}.
25312 Specify image file to draw the axis. This option override @var{fontfile} and
25313 @var{fontcolor} option.
25316 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25317 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25318 Default value is @code{1}.
25321 Set colorspace. The accepted values are:
25324 Unspecified (default)
25333 BT.470BG or BT.601-6 625
25336 SMPTE-170M or BT.601-6 525
25342 BT.2020 with non-constant luminance
25347 Set spectrogram color scheme. This is list of floating point values with format
25348 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25349 The default is @code{1|0.5|0|0|0.5|1}.
25353 @subsection Examples
25357 Playing audio while showing the spectrum:
25359 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25363 Same as above, but with frame rate 30 fps:
25365 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25369 Playing at 1280x720:
25371 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25375 Disable sonogram display:
25381 A1 and its harmonics: A1, A2, (near)E3, A3:
25383 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),
25384 asplit[a][out1]; [a] showcqt [out0]'
25388 Same as above, but with more accuracy in frequency domain:
25390 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),
25391 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25397 bar_v=10:sono_v=bar_v*a_weighting(f)
25401 Custom gamma, now spectrum is linear to the amplitude.
25407 Custom tlength equation:
25409 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)))'
25413 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25415 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25419 Custom font using fontconfig:
25421 font='Courier New,Monospace,mono|bold'
25425 Custom frequency range with custom axis using image file:
25427 axisfile=myaxis.png:basefreq=40:endfreq=10000
25433 Convert input audio to video output representing the audio power spectrum.
25434 Audio amplitude is on Y-axis while frequency is on X-axis.
25436 The filter accepts the following options:
25440 Specify size of video. For the syntax of this option, check the
25441 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25442 Default is @code{1024x512}.
25446 This set how each frequency bin will be represented.
25448 It accepts the following values:
25454 Default is @code{bar}.
25457 Set amplitude scale.
25459 It accepts the following values:
25473 Default is @code{log}.
25476 Set frequency scale.
25478 It accepts the following values:
25487 Reverse logarithmic scale.
25489 Default is @code{lin}.
25492 Set window size. Allowed range is from 16 to 65536.
25494 Default is @code{2048}
25497 Set windowing function.
25499 It accepts the following values:
25522 Default is @code{hanning}.
25525 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25526 which means optimal overlap for selected window function will be picked.
25529 Set time averaging. Setting this to 0 will display current maximal peaks.
25530 Default is @code{1}, which means time averaging is disabled.
25533 Specify list of colors separated by space or by '|' which will be used to
25534 draw channel frequencies. Unrecognized or missing colors will be replaced
25538 Set channel display mode.
25540 It accepts the following values:
25545 Default is @code{combined}.
25548 Set minimum amplitude used in @code{log} amplitude scaler.
25551 Set data display mode.
25553 It accepts the following values:
25559 Default is @code{magnitude}.
25562 @section showspatial
25564 Convert stereo input audio to a video output, representing the spatial relationship
25565 between two channels.
25567 The filter accepts the following options:
25571 Specify the video size for the output. For the syntax of this option, check the
25572 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25573 Default value is @code{512x512}.
25576 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25579 Set window function.
25581 It accepts the following values:
25606 Default value is @code{hann}.
25609 Set ratio of overlap window. Default value is @code{0.5}.
25610 When value is @code{1} overlap is set to recommended size for specific
25611 window function currently used.
25614 @anchor{showspectrum}
25615 @section showspectrum
25617 Convert input audio to a video output, representing the audio frequency
25620 The filter accepts the following options:
25624 Specify the video size for the output. For the syntax of this option, check the
25625 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25626 Default value is @code{640x512}.
25629 Specify how the spectrum should slide along the window.
25631 It accepts the following values:
25634 the samples start again on the left when they reach the right
25636 the samples scroll from right to left
25638 frames are only produced when the samples reach the right
25640 the samples scroll from left to right
25643 Default value is @code{replace}.
25646 Specify display mode.
25648 It accepts the following values:
25651 all channels are displayed in the same row
25653 all channels are displayed in separate rows
25656 Default value is @samp{combined}.
25659 Specify display color mode.
25661 It accepts the following values:
25664 each channel is displayed in a separate color
25666 each channel is displayed using the same color scheme
25668 each channel is displayed using the rainbow color scheme
25670 each channel is displayed using the moreland color scheme
25672 each channel is displayed using the nebulae color scheme
25674 each channel is displayed using the fire color scheme
25676 each channel is displayed using the fiery color scheme
25678 each channel is displayed using the fruit color scheme
25680 each channel is displayed using the cool color scheme
25682 each channel is displayed using the magma color scheme
25684 each channel is displayed using the green color scheme
25686 each channel is displayed using the viridis color scheme
25688 each channel is displayed using the plasma color scheme
25690 each channel is displayed using the cividis color scheme
25692 each channel is displayed using the terrain color scheme
25695 Default value is @samp{channel}.
25698 Specify scale used for calculating intensity color values.
25700 It accepts the following values:
25705 square root, default
25716 Default value is @samp{sqrt}.
25719 Specify frequency scale.
25721 It accepts the following values:
25729 Default value is @samp{lin}.
25732 Set saturation modifier for displayed colors. Negative values provide
25733 alternative color scheme. @code{0} is no saturation at all.
25734 Saturation must be in [-10.0, 10.0] range.
25735 Default value is @code{1}.
25738 Set window function.
25740 It accepts the following values:
25765 Default value is @code{hann}.
25768 Set orientation of time vs frequency axis. Can be @code{vertical} or
25769 @code{horizontal}. Default is @code{vertical}.
25772 Set ratio of overlap window. Default value is @code{0}.
25773 When value is @code{1} overlap is set to recommended size for specific
25774 window function currently used.
25777 Set scale gain for calculating intensity color values.
25778 Default value is @code{1}.
25781 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25784 Set color rotation, must be in [-1.0, 1.0] range.
25785 Default value is @code{0}.
25788 Set start frequency from which to display spectrogram. Default is @code{0}.
25791 Set stop frequency to which to display spectrogram. Default is @code{0}.
25794 Set upper frame rate limit. Default is @code{auto}, unlimited.
25797 Draw time and frequency axes and legends. Default is disabled.
25800 The usage is very similar to the showwaves filter; see the examples in that
25803 @subsection Examples
25807 Large window with logarithmic color scaling:
25809 showspectrum=s=1280x480:scale=log
25813 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
25815 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
25816 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
25820 @section showspectrumpic
25822 Convert input audio to a single video frame, representing the audio frequency
25825 The filter accepts the following options:
25829 Specify the video size for the output. For the syntax of this option, check the
25830 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25831 Default value is @code{4096x2048}.
25834 Specify display mode.
25836 It accepts the following values:
25839 all channels are displayed in the same row
25841 all channels are displayed in separate rows
25843 Default value is @samp{combined}.
25846 Specify display color mode.
25848 It accepts the following values:
25851 each channel is displayed in a separate color
25853 each channel is displayed using the same color scheme
25855 each channel is displayed using the rainbow color scheme
25857 each channel is displayed using the moreland color scheme
25859 each channel is displayed using the nebulae color scheme
25861 each channel is displayed using the fire color scheme
25863 each channel is displayed using the fiery color scheme
25865 each channel is displayed using the fruit color scheme
25867 each channel is displayed using the cool color scheme
25869 each channel is displayed using the magma color scheme
25871 each channel is displayed using the green color scheme
25873 each channel is displayed using the viridis color scheme
25875 each channel is displayed using the plasma color scheme
25877 each channel is displayed using the cividis color scheme
25879 each channel is displayed using the terrain color scheme
25881 Default value is @samp{intensity}.
25884 Specify scale used for calculating intensity color values.
25886 It accepts the following values:
25891 square root, default
25901 Default value is @samp{log}.
25904 Specify frequency scale.
25906 It accepts the following values:
25914 Default value is @samp{lin}.
25917 Set saturation modifier for displayed colors. Negative values provide
25918 alternative color scheme. @code{0} is no saturation at all.
25919 Saturation must be in [-10.0, 10.0] range.
25920 Default value is @code{1}.
25923 Set window function.
25925 It accepts the following values:
25949 Default value is @code{hann}.
25952 Set orientation of time vs frequency axis. Can be @code{vertical} or
25953 @code{horizontal}. Default is @code{vertical}.
25956 Set scale gain for calculating intensity color values.
25957 Default value is @code{1}.
25960 Draw time and frequency axes and legends. Default is enabled.
25963 Set color rotation, must be in [-1.0, 1.0] range.
25964 Default value is @code{0}.
25967 Set start frequency from which to display spectrogram. Default is @code{0}.
25970 Set stop frequency to which to display spectrogram. Default is @code{0}.
25973 @subsection Examples
25977 Extract an audio spectrogram of a whole audio track
25978 in a 1024x1024 picture using @command{ffmpeg}:
25980 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
25984 @section showvolume
25986 Convert input audio volume to a video output.
25988 The filter accepts the following options:
25995 Set border width, allowed range is [0, 5]. Default is 1.
25998 Set channel width, allowed range is [80, 8192]. Default is 400.
26001 Set channel height, allowed range is [1, 900]. Default is 20.
26004 Set fade, allowed range is [0, 1]. Default is 0.95.
26007 Set volume color expression.
26009 The expression can use the following variables:
26013 Current max volume of channel in dB.
26019 Current channel number, starting from 0.
26023 If set, displays channel names. Default is enabled.
26026 If set, displays volume values. Default is enabled.
26029 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26030 default is @code{h}.
26033 Set step size, allowed range is [0, 5]. Default is 0, which means
26037 Set background opacity, allowed range is [0, 1]. Default is 0.
26040 Set metering mode, can be peak: @code{p} or rms: @code{r},
26041 default is @code{p}.
26044 Set display scale, can be linear: @code{lin} or log: @code{log},
26045 default is @code{lin}.
26049 If set to > 0., display a line for the max level
26050 in the previous seconds.
26051 default is disabled: @code{0.}
26054 The color of the max line. Use when @code{dm} option is set to > 0.
26055 default is: @code{orange}
26060 Convert input audio to a video output, representing the samples waves.
26062 The filter accepts the following options:
26066 Specify the video size for the output. For the syntax of this option, check the
26067 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26068 Default value is @code{600x240}.
26073 Available values are:
26076 Draw a point for each sample.
26079 Draw a vertical line for each sample.
26082 Draw a point for each sample and a line between them.
26085 Draw a centered vertical line for each sample.
26088 Default value is @code{point}.
26091 Set the number of samples which are printed on the same column. A
26092 larger value will decrease the frame rate. Must be a positive
26093 integer. This option can be set only if the value for @var{rate}
26094 is not explicitly specified.
26097 Set the (approximate) output frame rate. This is done by setting the
26098 option @var{n}. Default value is "25".
26100 @item split_channels
26101 Set if channels should be drawn separately or overlap. Default value is 0.
26104 Set colors separated by '|' which are going to be used for drawing of each channel.
26107 Set amplitude scale.
26109 Available values are:
26127 Set the draw mode. This is mostly useful to set for high @var{n}.
26129 Available values are:
26132 Scale pixel values for each drawn sample.
26135 Draw every sample directly.
26138 Default value is @code{scale}.
26141 @subsection Examples
26145 Output the input file audio and the corresponding video representation
26148 amovie=a.mp3,asplit[out0],showwaves[out1]
26152 Create a synthetic signal and show it with showwaves, forcing a
26153 frame rate of 30 frames per second:
26155 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26159 @section showwavespic
26161 Convert input audio to a single video frame, representing the samples waves.
26163 The filter accepts the following options:
26167 Specify the video size for the output. For the syntax of this option, check the
26168 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26169 Default value is @code{600x240}.
26171 @item split_channels
26172 Set if channels should be drawn separately or overlap. Default value is 0.
26175 Set colors separated by '|' which are going to be used for drawing of each channel.
26178 Set amplitude scale.
26180 Available values are:
26200 Available values are:
26203 Scale pixel values for each drawn sample.
26206 Draw every sample directly.
26209 Default value is @code{scale}.
26212 Set the filter mode.
26214 Available values are:
26217 Use average samples values for each drawn sample.
26220 Use peak samples values for each drawn sample.
26223 Default value is @code{average}.
26226 @subsection Examples
26230 Extract a channel split representation of the wave form of a whole audio track
26231 in a 1024x800 picture using @command{ffmpeg}:
26233 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26237 @section sidedata, asidedata
26239 Delete frame side data, or select frames based on it.
26241 This filter accepts the following options:
26245 Set mode of operation of the filter.
26247 Can be one of the following:
26251 Select every frame with side data of @code{type}.
26254 Delete side data of @code{type}. If @code{type} is not set, delete all side
26260 Set side data type used with all modes. Must be set for @code{select} mode. For
26261 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26262 in @file{libavutil/frame.h}. For example, to choose
26263 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26267 @section spectrumsynth
26269 Synthesize audio from 2 input video spectrums, first input stream represents
26270 magnitude across time and second represents phase across time.
26271 The filter will transform from frequency domain as displayed in videos back
26272 to time domain as presented in audio output.
26274 This filter is primarily created for reversing processed @ref{showspectrum}
26275 filter outputs, but can synthesize sound from other spectrograms too.
26276 But in such case results are going to be poor if the phase data is not
26277 available, because in such cases phase data need to be recreated, usually
26278 it's just recreated from random noise.
26279 For best results use gray only output (@code{channel} color mode in
26280 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26281 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26282 @code{data} option. Inputs videos should generally use @code{fullframe}
26283 slide mode as that saves resources needed for decoding video.
26285 The filter accepts the following options:
26289 Specify sample rate of output audio, the sample rate of audio from which
26290 spectrum was generated may differ.
26293 Set number of channels represented in input video spectrums.
26296 Set scale which was used when generating magnitude input spectrum.
26297 Can be @code{lin} or @code{log}. Default is @code{log}.
26300 Set slide which was used when generating inputs spectrums.
26301 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26302 Default is @code{fullframe}.
26305 Set window function used for resynthesis.
26308 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26309 which means optimal overlap for selected window function will be picked.
26312 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26313 Default is @code{vertical}.
26316 @subsection Examples
26320 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26321 then resynthesize videos back to audio with spectrumsynth:
26323 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
26324 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
26325 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26329 @section split, asplit
26331 Split input into several identical outputs.
26333 @code{asplit} works with audio input, @code{split} with video.
26335 The filter accepts a single parameter which specifies the number of outputs. If
26336 unspecified, it defaults to 2.
26338 @subsection Examples
26342 Create two separate outputs from the same input:
26344 [in] split [out0][out1]
26348 To create 3 or more outputs, you need to specify the number of
26351 [in] asplit=3 [out0][out1][out2]
26355 Create two separate outputs from the same input, one cropped and
26358 [in] split [splitout1][splitout2];
26359 [splitout1] crop=100:100:0:0 [cropout];
26360 [splitout2] pad=200:200:100:100 [padout];
26364 Create 5 copies of the input audio with @command{ffmpeg}:
26366 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26372 Receive commands sent through a libzmq client, and forward them to
26373 filters in the filtergraph.
26375 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26376 must be inserted between two video filters, @code{azmq} between two
26377 audio filters. Both are capable to send messages to any filter type.
26379 To enable these filters you need to install the libzmq library and
26380 headers and configure FFmpeg with @code{--enable-libzmq}.
26382 For more information about libzmq see:
26383 @url{http://www.zeromq.org/}
26385 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26386 receives messages sent through a network interface defined by the
26387 @option{bind_address} (or the abbreviation "@option{b}") option.
26388 Default value of this option is @file{tcp://localhost:5555}. You may
26389 want to alter this value to your needs, but do not forget to escape any
26390 ':' signs (see @ref{filtergraph escaping}).
26392 The received message must be in the form:
26394 @var{TARGET} @var{COMMAND} [@var{ARG}]
26397 @var{TARGET} specifies the target of the command, usually the name of
26398 the filter class or a specific filter instance name. The default
26399 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26400 but you can override this by using the @samp{filter_name@@id} syntax
26401 (see @ref{Filtergraph syntax}).
26403 @var{COMMAND} specifies the name of the command for the target filter.
26405 @var{ARG} is optional and specifies the optional argument list for the
26406 given @var{COMMAND}.
26408 Upon reception, the message is processed and the corresponding command
26409 is injected into the filtergraph. Depending on the result, the filter
26410 will send a reply to the client, adopting the format:
26412 @var{ERROR_CODE} @var{ERROR_REASON}
26416 @var{MESSAGE} is optional.
26418 @subsection Examples
26420 Look at @file{tools/zmqsend} for an example of a zmq client which can
26421 be used to send commands processed by these filters.
26423 Consider the following filtergraph generated by @command{ffplay}.
26424 In this example the last overlay filter has an instance name. All other
26425 filters will have default instance names.
26428 ffplay -dumpgraph 1 -f lavfi "
26429 color=s=100x100:c=red [l];
26430 color=s=100x100:c=blue [r];
26431 nullsrc=s=200x100, zmq [bg];
26432 [bg][l] overlay [bg+l];
26433 [bg+l][r] overlay@@my=x=100 "
26436 To change the color of the left side of the video, the following
26437 command can be used:
26439 echo Parsed_color_0 c yellow | tools/zmqsend
26442 To change the right side:
26444 echo Parsed_color_1 c pink | tools/zmqsend
26447 To change the position of the right side:
26449 echo overlay@@my x 150 | tools/zmqsend
26453 @c man end MULTIMEDIA FILTERS
26455 @chapter Multimedia Sources
26456 @c man begin MULTIMEDIA SOURCES
26458 Below is a description of the currently available multimedia sources.
26462 This is the same as @ref{movie} source, except it selects an audio
26468 Read audio and/or video stream(s) from a movie container.
26470 It accepts the following parameters:
26474 The name of the resource to read (not necessarily a file; it can also be a
26475 device or a stream accessed through some protocol).
26477 @item format_name, f
26478 Specifies the format assumed for the movie to read, and can be either
26479 the name of a container or an input device. If not specified, the
26480 format is guessed from @var{movie_name} or by probing.
26482 @item seek_point, sp
26483 Specifies the seek point in seconds. The frames will be output
26484 starting from this seek point. The parameter is evaluated with
26485 @code{av_strtod}, so the numerical value may be suffixed by an IS
26486 postfix. The default value is "0".
26489 Specifies the streams to read. Several streams can be specified,
26490 separated by "+". The source will then have as many outputs, in the
26491 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26492 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26493 respectively the default (best suited) video and audio stream. Default
26494 is "dv", or "da" if the filter is called as "amovie".
26496 @item stream_index, si
26497 Specifies the index of the video stream to read. If the value is -1,
26498 the most suitable video stream will be automatically selected. The default
26499 value is "-1". Deprecated. If the filter is called "amovie", it will select
26500 audio instead of video.
26503 Specifies how many times to read the stream in sequence.
26504 If the value is 0, the stream will be looped infinitely.
26505 Default value is "1".
26507 Note that when the movie is looped the source timestamps are not
26508 changed, so it will generate non monotonically increasing timestamps.
26510 @item discontinuity
26511 Specifies the time difference between frames above which the point is
26512 considered a timestamp discontinuity which is removed by adjusting the later
26516 It allows overlaying a second video on top of the main input of
26517 a filtergraph, as shown in this graph:
26519 input -----------> deltapts0 --> overlay --> output
26522 movie --> scale--> deltapts1 -------+
26524 @subsection Examples
26528 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26529 on top of the input labelled "in":
26531 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26532 [in] setpts=PTS-STARTPTS [main];
26533 [main][over] overlay=16:16 [out]
26537 Read from a video4linux2 device, and overlay it on top of the input
26540 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26541 [in] setpts=PTS-STARTPTS [main];
26542 [main][over] overlay=16:16 [out]
26546 Read the first video stream and the audio stream with id 0x81 from
26547 dvd.vob; the video is connected to the pad named "video" and the audio is
26548 connected to the pad named "audio":
26550 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26554 @subsection Commands
26556 Both movie and amovie support the following commands:
26559 Perform seek using "av_seek_frame".
26560 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26563 @var{stream_index}: If stream_index is -1, a default
26564 stream is selected, and @var{timestamp} is automatically converted
26565 from AV_TIME_BASE units to the stream specific time_base.
26567 @var{timestamp}: Timestamp in AVStream.time_base units
26568 or, if no stream is specified, in AV_TIME_BASE units.
26570 @var{flags}: Flags which select direction and seeking mode.
26574 Get movie duration in AV_TIME_BASE units.
26578 @c man end MULTIMEDIA SOURCES